Previous Page

448 RISK ANALYSIS OF BURIED WASTES FROM ELECTRICITY GENERATION

Case Study 2 this procedure, a key element is to estimate the probability

per year, P, for an atom of this buried waste to be dissolved
In this case, the resistivity method has been used to map
by groundwater and eventually enter a human stomach.
the extent of contamination induced by an open waste
Obviously, this probability is site dependent and depends
disposal site and to determine where to locate future
on present underground characteristics and future events
monitoring wells. The site is located SE of the city Isparta,
in the region, many of them unpredictable, including
Turkey, on a nearly 100 m thick Quaternary alluvial
climate changes, geological land uplifting causing rivers to
deposit of gravel, sand, and clay. The Schlumberger
change their courses, volcanoes, earthquakes, and human
sounding resistivity method was used to map resistivities
intrusion.
from 0.5 down to 25 meters in depth. Figure 11 shows the
apparent resistivity maps for different electrode spacings.
As expected, leachate from the waste disposal site has INTRODUCTION
decreased groundwater resistivity. The deeper layers are
characterized by low resistivities, and the shallower ones A first estimate of P can be obtained by considering that
by high resistivities. all properties of the region are the U.S. average (lower
A resistivity pseudosection along measurement line 9 48 states); the purpose of this article is to obtain such an
is performed, and the result is shown in Fig. 12. As can estimate. Most relevant characteristics and future changes
be seen, the low resistivity front from the waste disposal in them are occurring somewhere in the United States
site base is characterized by its low resistivity and is and are hence taken into account with or including some
developing northward. So, the extent of groundwater estimate of their probability of occurrence. The value of
pollution can easily be manifested using geoelectrical P we obtain can be interpreted as the average value of
mapping (8). P for a large number of randomly chosen sites. It seems
reasonable to assume that the very elaborate site selection
BIBLIOGRAPHY efforts by geological and hydrologic experts should choose
a site at least as secure as a randomly chosen site.
1. Rao, T.G. (Ed.). (1992). Groundwater Exploration Techniques. We assume that the waste converted to a rock-like
National Geophysical Research Institute, Hyderabad, India. material behaves like average rock—differences between
2. Jahani, H.R. (2005). Applied Geophysical Methods for Ground- them will be considered later—and calculate P for average
water Investigations. Iranian Hydraulic Association, Tehran, rock. P can be calculated as the dissolution rate, R, the
Iran. probability per year for an atom of the rock to be dissolved
3. Griffiths, D.H. and King, R.F. (1988). Applied Geophysics for in groundwater, times the probability, p, for an atom once
Geologists and Engineers. Pergamon Press, Oxford. dissolved into groundwater to enter a human stomach:
4. U.S. Soil Conservation Service. (1971). Seismic and Resistivity
Methods of Geophysical Exploration. Water Resources Publi- P=R×p (1)
cations, United States Soil Conservation Service, Washington,
DC.
5. Zohdy, A.A.R. et al. (1990). Application of Surface Geo- Dissolution Rate for Rock, R
physics to Groundwater Investigations. U.S. Geological Survey,
Denver, CO. We begin by calculating R. From the rate at which rivers
6. Iranian National Karst Research Center. (1997). Resistivity carry dissolved and suspended material into the oceans, it
Methods of Geophysical Explorations in Karstic Areas in West is estimated (1) that the surface of the continent is eroding
Azerbaijan Province. Shiraz. at an average rate of 5 × 10−5 (hereafter, we use the
7. Hotzl, H. and Werner, A. (Eds.). (1992). Tracer Hydrology. A.A. notation 5 E−5) meters per year (hereafter, m/y). About
Balkema, Rotterdam, The Netherlands. 28% of this material is in solution, corresponding to 1.4
8. Karlýk, G. and Ali Kaya, M. (2001). Investigation of groundwa- E−5 m/y removed by chemical dissolution. It is estimated
ter contamination using electric and electromagnetic methods (2,3) that 15% of the water flow in rivers is derived
at an open waste-disposal site: A case study from Isparta, from groundwater (aquifers), and the rest comes from
Turkey. Environ. Geol. 40(6): 725–731. surface runoff. If the concentrations of dissolved materials
were the same in rivers as in aquifers, this would mean
that the latter dissolve (0.15 × 1.4 E−5) = 2.1 E−6 m/y
RISK ANALYSIS OF BURIED WASTES FROM of rock thickness. The average concentrations of a few
ELECTRICITY GENERATION dissolved materials in aquifers (4) and in river water (5)
are listed in Table 1. Silica and calcium are such important
BERNARD L. COHEN
University of Pittsburgh
Pittsburgh, Pennsylvania Table 1. Average Concentrations in mg/kg of Dissolved
Materials in Aquifersa and in River Waterb
Material Silica Calcium Magnesium Potassium Iron Uranium
An important aspect of generating electricity from nuclear
power is the disposal of the high level radioactive waste. Aquifers 20 30 3 2 0.3 3 E−4
The most widely anticipated plan is to convert it into a Rivers 7 15 4 2.2 0.5 4 E−4
rock-like material and bury it deep underground, typically a
Ref. 4.
600 meters below the surface. In evaluating the safety of b
Ref. 5.
 RISK ANALYSIS OF BURIED WASTES FROM ELECTRICITY GENERATION 449

components of rock that we interpret these data to indicate Table 2. Calculation of Fractional Removal per year of
that aquifers carry about twice the concentration of Rock Materials by a Typical Aquifer. Columns (2) and (3)
dissolved material in rivers, which leads to the conclusion refer to a 1 Square Meter Cross Section of the Aquifer.
that [30/(30 + 85)] = 26% of the dissolved material in See Discussion in Text
rivers, (0.26 × 1.4 E−5) = 3.6 E−6 m/y of rock depth, is (3) (4)
contributed by aquifer dissolution of rock. (1) (2) Kg in Rock Fraction Removed
The next problem is to estimate what fraction of this Material Kg/y into River (× 1 E6) (× 1 E−8)
material, f , is removed per meter of depth at 600 m; let us Ca 0.3 15 2
say from between 599 m and 600 m. As a crude gross over- Mg 0.03 3 1
estimate, we might assume that rock erosion is constant K 0.02 3 0.7
with depth down to 600 m and zero below 600 m, in which Fe 0.003 9 0.03
case f = 1/600 = 16 E−4. A more reasonable approach is U 3 E−6 8 E−4 0.3
as follows: The annual circulation of groundwater, defined Silica 0.2 150 0.13
as the quantity per year entering or leaving aquifers at Carbonate 1.3 18 8
the stated depth, has been given (2,3) as

shallow (<800 m) : 310 E9 cubic meters per year data for other materials, obtained analogously to that
for Ca.
deep (>800 m) : 6.2 E9 cubic meters per year
Of the materials listed in Table 2, the materials in the
radioactive waste are most similar to Fe and U and not
If we assume that the flow decreases exponentially with at all similar to carbonate ions. Thus, the numbers in
depth and that the rate of rock dissolution is proportional Column (4) of Table 2 might give an estimate of about
to this groundwater flow, determining the constants in the R = 7 E−9, seven times higher than Eq. 2. But the aquifer
relationship from these data, we find (6) that f = 2.6 E−4, we have assumed in this calculation is substantially more
one-sixth of our crude gross overestimate. robust than the groundwater encountered by average rock
Using this gives the quantity of rock eroded per at 600 m depth, so it seems reasonable to conclude that
year per meter of depth at 600 m below the surface Eq. 2 is roughly verified.
as (3.6 E−6 × 2.6 E−4) = 1.0 E−9 m/y. If 1.0 E−9 m/y of
depth is dissolved from one meter of rock depth, the Probability of Transport from Groundwater to Human
probability per year for an atom of that rock to be dissolved Stomachs, p
must be 1.0 E−9. Thus,
Next we turn to the problem of estimating the probability
R = 1.0 E−9 (2) of transferring an atom of material dissolved in groundwa-
ter into human stomachs, p in Eq. 1. We assume that once
the material is dissolved in groundwater, it moves with
This result plays a key role in our risk analysis,
groundwater, eventually reaching shallow aquifers that
so it is interesting to seek alternative approaches to
feed into rivers. Contributions to p derive from our use
calculating it. One such approach follows: A typical aquifer
of rivers and wells drilled into aquifers for potable water,
reaching to the waste burial depth of 600 m may be
from fish removed from rivers and used for food, and from
about 100 km long and have a flow velocity of 100 m/y
use of well and river water for irrigating food crops. The
through rock of 10% porosity, discharging into a river
potable water path is the simplest to calculate and, it
The water discharged from it annually per square meter
turns out, the most important. The average person ingests
of cross-sectional area is then 10% of the volume of
2 L/day of potable water which corresponds to ingestion
a column of water 100 m long and one square meter
by the U.S. population of 2 × 365 × 2.8 E8 = 2.0 E11 L/y
in cross section, which is 10 cubic meters, or 10,000
(liters per year). Of this, 45% is derived from wells, and
liters (L).
55% comes from rivers (7), corresponding to 9 E10 and
Chemical analyses of groundwater (4) indicate that it
1.1 E11 L/y respectively entering human stomachs. The
typically contains 30 mg/L or 30 E−6 kg/L (milligrams
estimated water flow in U.S. rivers is 1.7 E15 L/y (1) and
per liter, or kilograms per liter) of Calcium (Ca), so it
1.9 E15 L/y (2,3); we use 1.8 E15 L/y. The estimated water
discharges (30 E−6 kg/L × 10,000L =) 0.3 kg of Ca into
flow in aquifers is 16% of the flow in rivers (3), or 2.9 E14
the river each year. This is the first entry in Column (2)
L/y The contributions to p from the use of rivers and wells
of Table 2. This calcium was derived by dissolution from
for potable water are then
the rock through which the aquifer had passed, 100 km,
or 1E 5 m, long, one square meter in cross section, and p(rivers) = (1.1 E11/1.8 E15) = 6.1 E−5
a specific gravity of about 3.0, which gives it a mass of
(3000 × 1 E5) = 3 E8 Kg. Typical rock contains 5% Ca (1), p(well water) = (9 E10/2.9 E14) = 3.1 E−4
so the Ca contained in this source rock is (0.05 × 3 E8) =
15 E6 kg; this is the first entry in Column (3) of Table 2. Note that we have ignored removal of material by filtration
If 0.3 kg of Ca per year is dissolved from 15 E6 kg of processes, making these conservative estimates, that is,
Ca in the rock, the probability for a Ca atom to be more likely to be high than low.
dissolved must be (0.3/15 E6) = 2 E−8 per year; this is Analyses of the fish and irrigation contributions
the first entry in Column (4) of Table 2. Table 2 includes to p will be sketched below. They give results that
450 RISK ANALYSIS OF BURIED WASTES FROM ELECTRICITY GENERATION

vary considerably from element to element because of H

variations in bioaccumulative factors (8). For the fish Human oral
pathway, p(fish) is rarely larger than 1 E−5, and for most intake
important elements in the waste, it is much smaller. For
0.9 0.1
the irrigation pathway, p(irrigation) is rarely more than 0.02 F W 0.45
1 E−4 and somewhat less for the most important elements Food Drinking
in the waste. We therefore take as conservative estimates water
0.88 0.1 0.55
p(fish) = 1 E−5
G E
D
p(irrigation) = 1 E−4 Sea-
Soil Irrigation
food water 0.37

Summing these four contributions gives p = 5 E−4. 1.0 0.63

Using this and Eq. 2 in Eq. 1 then gives the result we 0.6
are seeking for P: B C
Rivers Wells
P = (1 E−9 × 5 E−4) = 5 E−13 (3)
0.4 1.0
Alternative Approach for Transfer from Rock to Human
A
Stomachs Rock
We now present an essentially independent method (9)
Figure 1. Pathway diagram for transmission of materials from
for calculating P, based on our knowledge of quantities of
rock and soil into food and water ingested by people. Numbers
elements entering human stomachs each year (10,11) and attached to each line show the fraction of the input to the target
the quantities of these materials in rock (5), supplemented box that is derived from that line.
by an estimation of the fraction of these that derive from
soil rather than from rock.
We begin with the very crude assumption that all their fractional contributions to human intake are then
material entering human stomachs derives from the top calculated as
600 m of rock. We know how much of each chemical
element is contained in this rock (5), and we know how ABWH 0.4 × 0.55 × 0.1 = 0.022;
much of each chemical element enters human stomachs ACWH 1.0 × 0.45 × 0.1 = 0.045;
each year (10,11), so the ratio of these gives a crude ABEFH 0.4 × 0.63 × 0.1 × 0.9 = 0.023;
estimate of P for each element. An average of these, ACEFH 1.0 × 0.37 × 0.1 × 0.9 = 0.033;
weighting for the importance of each element in the ABGFH 0.4 × 1.0 × 0.02 × 0.9 = 0.007;
waste, gives TOTAL = 0.13
P(very crude) = 1 E−11
Our final value for P is then
Some of the crudeness of this estimate may be reduced
by dropping the implication that all depths of rock down to P(final) = (0.13 × 1.6 E−12) = 2 E−13 (4)
600 m are equally likely to contribute and assuming that
the contribution from various depths is proportional to the This result, Eq. 4, is in substantial agreement with
groundwater flow at that depth. Using the exponential the estimate from Equation 3 derived from essentially
decrease in this flow alluded to before, it turns out that independent approaches.
the flow at 600 m depth is 16% of the average for the top
600 m, giving an improved, but still rough estimate:
HIGH LEVEL WASTE VERSUS AVERAGE ROCK
P(rough) = (0.16 × 1 E−11) = 1.6 E−12
This result applies to average rock, but our interest
This result is still an overestimate because most of is in applying it to the rock-like material, probably
the material entering human stomachs derives from a borosilicate glass, into which the waste has been
the top layers of soil rather than from the underlying converted. The principal difference is in leachability
rock. To analyze this problem, a pathway analysis was by groundwater. Leach rates by distilled water for
developed (9) as shown in Fig. 1. The numbers attached various rocks and glasses have been measured, and
to the arrows are the fraction of the input into the the results were correlated with the Gibbs free energy
compartment derived from that arrow. For example, 90% of hydration (12). It was concluded that the leach rate
of human oral intake is from food, and 10% is from for borosilicate glass is similar to that for basalt and
drinking water; 2% of the minerals in food comes from about three times more rapid than the leach rate for
seafood (derived from rivers), 88% comes directly from average rock. Applying this to the results in Eqs. 3 and 4
soil, and 10% comes from soil via irrigation water (which is gives the rate of transfer for an atom of buried high
derived 63% from rivers and 37% from wells). Using Fig. 1, level waste into a human stomach as about 1.0 E−12
the various pathways from rock to human stomachs and per year.
 GROUNDWATER CONTAMINATION FROM RUNOFF 451

Concerns have been raised that the high level waste GROUNDWATER CONTAMINATION FROM
glass may be much less secure than average rock because it RUNOFF
is not in chemical equilibrium with the local rock–ground
water regime. However, chemical equilibrium is a
NITISH PRIYADARSHI
surface phenomenon (13,14). Groundwater is saturated
Ranchi University
with dissolved silica. If a foreign silica-based material
Ranchi, Jharkhand, India
such as glass is emplaced, there is initially rapid
dissolution, removing a very thin surface layer. But this
brings more silica into the water which then becomes
INTRODUCTION
supersaturated with silica, resulting in precipitation of
silica from the water onto the surface—this would be
Groundwater is water that lies below the soil surface
the most insoluble component of the silica originally in
and fills the pore spaces in and around rock, sand,
the water. The result is a buildup of highly insoluble
gravel, and other materials. Groundwater, under most
silica on the surface. Waste can then be dissolved only
conditions, is safer and more reliable for use than surface
by diffusion through this surface layer, a very slow
water because surface water is more readily exposed to
process. Moreover, this leads to further thickening of
pollutants from factories, for example, than groundwater.
the surface layer, which further retards the diffusion
This by no means says that groundwater is invulnerable
process. Consequently, the rate of dissolution decreases
to contamination. Once groundwater is contaminated,
exponentially, only a tiny fraction of the waste becomes
it is extremely costly to remove the contaminant. Any
involved in the process, and dissolution proceeds only as
chemicals that are easily soluble and penetrate the soil
for the surrounding rock.
are prime candidates as groundwater pollutants.

BIBLIOGRAPHY
HOW MUCH DO WE DEPEND ON GROUNDWATER?
1. Garrels, R.M. and MacKenzie, F.T. (1971). Evolution of
Sedimentary Rocks. W.W. Norton Co., New York. According to 1985 U.S. figures, groundwater provides
2. Todd, D.K. (1980). Groundwater Hydrology. John Wiley & an estimated
Sons, New York.
3. Ad Hoc Panel on Hydrology. (1962). Scientific Hydrology, • 22% of all freshwater withdrawals
Federal Council for Science and Technology. Washington,
• 53% of drinking water for the total population and
DC.
97% of drinking water for the rural population
4. White, D.E., Hein, J.D., and Waring, G.A. (1963). Chemical
Composition of Sub-Surface Waters, Chap. F in Data of • 40% of public water supply withdrawals
Geochemistry, 6th Edn. M. Fleischer (Ed.). U.S. Government • 46% of domestic and commercial use
Printing Office, Washington, DC. • 24% of industrial and mining use
5. Bowen, H.J.M. (1979). Environmental Chemistry of the
• 34% of agricultural use (mostly for irrigation)
Elements. Academic Press, New York.
6. Cohen, B.L. (1985). A simple probabilistic risk analysis
for high level waste repositories. Nuclear Technology 68: CAUSES OF URBAN RUNOFF
73–76.
7. Byrne, F. (1974). Earth and Man. Wm. C. Brown Co., Urban runoff can be attributed to many things, including
Dubuque, IA.
the amount of rainfall, the soil conditions, and the degree
8. Cohen, B.L. (1984). Probability for human intake of an atom of urbanization. Rainwater can go in many directions
randomly released into ground, rivers, oceans, and air. Health
once it has reached the earth’s surface. Rainwater can
Physics 47: 281–292.
be absorbed by the soil on the land surface; absorbed
9. Cohen, B.L. (1986). A generic probabilistic risk analysis for a
by surrounding vegetation; directly deposited into oceans,
high level waste repository. Health Physics 51: 519–528.
streams, and rivers; and infiltrate through the surface and
10. International Commission on Radiological Protection. (1975).
subsurface soils into the groundwater.
Report of the Task Group on Reference Man. ICRP Publication
23, Pergamon Press, New York.
Urban runoff can happen anytime of the year when
excessive water use from irrigation, car washing, and other
11. Pennington, J.A.T. et al. (1984). Selected minerals in food
survey, 1974–1982. Jour. Am. Dietetic Assn. 84: 771ff.
sources carries litter, lawn clippings, and other urban
pollutants into storm drains.
12. Plodinek, M.J., Jantzen, C.M., and Wicks, G.C.. (1983). Sta-
bility of Radioactive Waste Glasses Assessed from Hydration
The majority of people live in cities that are becoming
Thermodynamics. Savannah River Laboratory Reports DP- larger and larger and are sprawling into vast suburban
MS-83-61, DP-MS-83-21, and DP-MS-83-66. neighborhoods. Roofs, parking lots, streets, and other
13. Cohen, B.L. (1980). Analysis, critique, and evaluation of impervious surfaces of an urban environment cause
high level waste water intrusion scenario studies. Nuclear rainwater to collect and be forced out through a storm
Technology 48: 63ff. drain system. If the drainage system does not connect to
14. Cohen, B.L. (1985). Critique of the National Academy of a wastewater treatment facility, then the rainwater and
Sciences study of the isolation system for geologic disposal of everything that is carried with it travels into groundwater,
radioactive waste. Nuclear Technology 20: 433ff. local streams, and rivers.
452 GROUNDWATER CONTAMINATION FROM RUNOFF

GROUNDWATER CONTAMINATION FROM URBAN Landfills are another major source of contamination.
RUNOFF Landfills are the places where our garbage is taken to be
buried. Landfills are supposed to have a protective bottom
Not too long ago, urban runoff was considered an insignif- layer to prevent contaminants from getting into the water.
icant contributor to groundwater contamination. Urban After rainfall, contaminants from the landfill (car battery
runoff is now recognized as a significant source of contam- acid, paint, household cleaners, etc.) are washed away
ination in water. Effluents from urban areas contain large with surface runoff and make their way down into the
concentration of oils, greases, nutrients, heavy metals, and groundwater. The problem of pollution from landfills is
detergents. Detergents that are soluble can pass through greatest where high rainfall and shallow water tables
the soil and pollute groundwater. Raw sewage dumped in occur. Important pollutants frequently found in leachate
shallow soakpits and seepage from polluted lakes, ponds, include BOD, COD, iron, manganese, chloride, nitrate,
and streams also pollute groundwater. Rainfall could pick hardness, and trace elements. Hardness, alkalinity, and
up substantial contaminants from dust and air and join total dissolved solids are often increased.
the aquifer below. The infiltration of liquids containing Chemicals include products used on lawns and farm
toxic pollutants may cause pollution in sandy soils and fields to fertilize plants. The primary fertilizers are
well waters. compounds of nitrogen, phosphorus, and potassium.
Some of the contaminants expected in urban runoff are Phosphate and potassium fertilizers are readily adsorbed
shown in Table 1. Groundwater moves through rocks and on soil particles and seldom constitute a pollution problem.
subsurface soil, so it has lots of opportunity to dissolve But nitrogen in solution is only partially used by plants or
substances as it moves. For that reason, groundwater adsorbed by the soils, and it is the primary fertilizer
often has more dissolved substances than surface water. pollutant (1). When the rain comes, this chemical is
Even though the ground is an excellent mechanism washed away with the surface runoff and eventually goes
for filtering out particulate matter, such as leaves, into the groundwater.
soil, and bugs, dissolved chemicals and gases can still Animal wastes are confined within a limited area, in
occur in large enough concentrations in groundwater to beef or milk production, so large amounts of waste are
cause problems. Underground water can be contaminated deposited on the ground. Thus, for the 120 to 150 days that
through urban runoff containing industrial, domestic, and a beef animal remains in a feedlot, it produces more than a
agricultural chemical wastes, which includes chemicals half ton of manure on a dry weight basis. When thousands
such as pesticides and herbicides that many homeowners of animals are in a single feedlot, the natural assimilative
apply to their lawns. Contamination of groundwater by capacity of the soil becomes overtaxed. Surface runoff
road salt is of major concern in northern areas of the in contact with the manure carries highly concentrated
United States. Salt is spread on the roads to melt ice, and pollutants to subsurface waters. Animal wastes may
salt is so soluble in water that excess sodium and chloride transport salts, organic loads, and bacteria into the water
is transported through runoff after ice melts and finally and soil. Nitrate-nitrogen is the most important persistent
contaminates subsurface groundwater. pollutant that may reach the groundwater (2).
In the United States today, there may be more than Sanitary systems have a significant impact on shallow
20,000 known abandoned and uncontrolled hazardous aquifers in urban areas. Field observations revealed that
waste sites, and numbers grow every year. If there is contamination of groundwater is caused by infiltration of
a leak, these contaminants may be carried away with surface runoff polluted by municipal waste and sewage
runoff and finally make their way down through the soil and/or leakage of sanitary effluents. It has increased
and into the groundwater. the salt content of groundwater, particularly nitrates

Table 1. The Contaminants that may be Expected in Runoffa

1. Asbestos from brake and clutch linings
2. Bacteria and viruses from animals and birds, soils, litter, livestock hauling, livestock waste hauling and on-site sewage tanks and
fields
3. Bromide from auto exhaust
4. Cadmium from tire fillers and insecticides
5. Chromium from moving engine parts and brake linings
6. Copper from bearing and bushing wear, moving engine parts, brake linings, and radiator repair
7. Cyanide and chloride from deicing road salts
8. Pesticides (fungicides, herbicides, and insecticides) from roadside maintenance
9. Iron from auto bodies, moving engine parts, bridges, guardrails, overpasses, lamp standards, and other structures
10. Manganese from moving engine parts and gasoline additives
11. Nickel from diesel fuel, lubricating oil, bushing wear, brake linings, and asphalt paving
12. Nitrogen from animal wastes, on-site sewage systems, vegetative matter, and fertilizers
13. Petroleum compounds from paving, fuels spills, engine blow-by, lubricant leaks, antifreeze, and hydraulic fluids
14. Phosphorus from animal wastes, on-site sewage systems, vegetative matter, and fertilizers
15. Potassium from fertilizers
16. Sulfate from roadbeds, road salts, and fuels
17. Zinc and lead from tire fillers, motor oil additives, automotive and radiator repairs, grease, and paint manufacturing
a
Source: Ministry of Water, Land and Air Protection, Government of British Columbia.
 SALINE SEEP 453

and chlorides, and has reduced the oxygen content. eliminates crop growth. Human induced salinization of
In addition, introduction of pathogenic bacteria and land and water resources due to water table rise is as old as
viruses into groundwater has caused many outbreaks of the history of human settlement and irrigation (2). Early
waterborne disease. settlements in the valleys of the Tigris and Euphrates
Solid materials are frequently stockpiled near indus- flourished about 4000 B.C.: however, the irrigated farming
trial plants and construction sites. These may be the raw of wheat resulted in salt accumulation, and in time
materials awaiting use, or they may be solid wastes placed the area was abandoned. Historic evidence from Russia,
for temporary or permanent storage. Precipitation falling China, India, Pakistan, South American, United States,
on unsheltered stockpiles may transport heavy metals, Canada, and Australia show that saline seeps are a
salts, and other inorganic and organic constituents as worldwide problem. The number of saline seeps will
pollutants to the groundwater. probably increase in the future because an increase
Urban runoff pollutants are many and varied, depend- in world population will be accompanied by increases
ing on land uses and pollutant sources in an urban area. in irrigation.
Typically, loadings of urban pollutants are greatest from Saline seeps result from a combination of geologic,
industrial and commercial areas. Although sources of spe-
climatic, and management factors. Development of a seep
cific pollutants may vary widely in urban areas, motor
starts at the recharge area. As the water moves through
vehicles are recognized as a major source of pollutants;
the aquifer, the water dissolves and accumulates salts.
they contribute oils, greases, hydrocarbons, and toxic met-
Saline seeps are formed by mobilization of salt stored in
als such as lead, zinc, copper, chromium, and arsenic.
the soil profile. The very act of irrigation is the cause
for salinity because every irrigation event adds some
BIBLIOGRAPHY salt to the soil. Land clearing or irrigation can cause
the groundwater to rise. The additional water moves the
1. Todd, D.K. (1995). Groundwater Hydrology. Wiley, Toronto, water table closer to the land surface. When the water
Canada, p. 332. table is within 2–3 feet of the surface, capillary action can
2. Gillham, R.W. and Webber, L.R. (1969). Nitrogen contamina- lift the water to the surface where it evaporates and leaves
tion of groundwater by barnyard leachates. J. Water Pollut. salts behind (Figs. 1 and 2).
Control Fed. 41(10): 1752–1762.
The salt can also move laterally to affect streams.
The common soil forming minerals calcium, sodium,
magnesium, and potassium form water-soluble salts in
SALINE SEEP
the soil.
JOHN E. MOORE Techniques for controlling seeps originated in Montana
USGS (Refired) and Alberta (3). The two major methods to control seeps
Denver, Colorado are (1) cut off the recharge water and (2) reclaim the
seep area. Installing subsurface tiles to intercept the
groundwater flow might cut off recharge water. The second
Jackson (1) defines saline seep as an intermittent or method is to increase water use in the recharge area by
continuous saline water discharge at or near the changing the type of crop. Alfalfa is an excellent crop
soil surface under dryland conditions that reduces or substitute because it is a deep-rooted perennial capable of

Discharge area
Excess water

Evapotranspiration
Infiltration

Deep percolation

Water
table XXX
XXXXXXXXXXXXXXXXXX

XXX Saline seep

Salt particles Low permeability

Figure 1. Water movement to seep.

454 GROUNDWATER SAMPLING TECHNIQUES FOR ENVIRONMENTAL PROJECTS

Water evaporates 3. Brown, P.L. et al. (1983). Saline seep diagnosis control and
reclamation. USDA-ARS conservation Research Report No. 30.
4. Lamon, R.E. and Whitney, D.A. (1992). Management of Saline
and Sodic Soils:MF-1022 Cooperative Extension Service.
Salts remain behind xxxxxxxxxxxxxxxxxx
Soil surface Kansas State University, Manhattan, KS.
Salty water moves 5. Wentz, D. (1997). Dryland Saline Seeps-Types and Causes.
upward from the Alberta Agriculture.
water table by
capillary action

Water table GROUNDWATER SAMPLING TECHNIQUES

Groundwater FOR ENVIRONMENTAL PROJECTS
JAMES A. JACOBS
Figure 2. Evaporation and salt concentration.
Environmental Bio-Systems, Inc.
Mill Valley, California

using large amounts of water. Seeps have been reclaimed Groundwater sampling techniques are important for
within 5 years by planting alfalfa in the recharge area (4). accurate project results and understanding aquifers
In southeast Australia, the replacement of deep-rooted whether the project is sampling a groundwater monitoring
native vegetation by dryland and irrigated agriculture well as part of an environmental assessment or regular
has resulted in a major change in surface water and sampling of a potable water supply well. Detailed
groundwater systems in the Murray–Darling basin. The groundwater sampling information can be obtained
Murray River has a salinity of less than 25 mg/L in the from Testa (1), Jacobs (2), the American Society for
headwaters and 480 mg/L downstream. Water that was Testing Materials (3), the California Regional Water
previously used by vegetation for transpiration is now Quality Control Board (4), and Driscoll (5). Although
leaking past the vadose zone and entering the water table. there are numerous differences in groundwater sampling
The salinization process is closely linked to groundwater and record keeping techniques, regional variations, and
changes. In both irrigation and dryland farming, salts specific regulatory requirements that vary from state to
are remobilized into the production zone. The salts were state, a generalized summary of selected environmental
originally stored in the aquifer or the vadose zone. groundwater sampling techniques is given below.
Evaporation concentrates the salts in the near surface
zone to levels where the plants are. For salinization to PURPOSE OF GROUNDWATER SAMPLES
occur, it is necessary to have both a hydrologic change and
a source of salt. Groundwater samples are collected for a variety of
Saline soils and sodic soils are quite common in parts reasons. Groundwater can be collected as a ‘‘grab’’
of western North Dakota and Montana. Most saline seeps sample or from a monitoring well. Groundwater samples
have developed recently (postsettlement). The formation are used to define the vertical and aerial extent of
of saline seeps is closely related to the practice of summer groundwater contamination, to monitor target chemical
fallow for moisture conservation (5). concentrations over time, to provide a measurement
Colorado River has 25 mg/L salt in the headwaters for detecting a contaminant plume front, to evalvate
and 825 mg/L downstream at Imperial Dam. In the unexpected changes in plume size or direction of flow,
San Joaquin River in California, salinity increased from to determine the extent of interaquifer movement of
330 mg/L in the 1930s to 600 mg/L in the 1970s. In the contaminants, to determine aquifer characteristics such
former Soviet Union, the salinity of the Syr Darya River as permeability, transmissivity, etc.), to estimate the rate
which discharges to the Aral Sea increases from 300 mg/L of contaminant plume movement, to develop a data base
to 2000 mg/L downstream. for designing remedial measures, to determine the effects
The water table rise in agricultural areas of the world of the remedial measures, to assist in performing the
has contributed to the salinization of large areas. The remedial work (provide hydraulic control and contaminant
development of land and water resources of the earth will removal), to provide data base for groundwater modeling,
be more difficult and more expensive than in the past. The and to evaluate the aquifer during regular sampling over
control of population and the efficient use of developed the yearly hydrologic cycle (6).
resources will be essential in the future.
GRAB GROUNDWATER SAMPLING

BIBLIOGRAPHY A ‘‘grab’’ groundwater sample is collected without the use

of a fully installed groundwater monitoring well. Methods
1. Jackson, J. (1997). Glossary of Geology. American Geological include using a bailer or jar to collect groundwater from an
Institute, p. 769. excavation or trench made with a backhoe or excavator.
2. Ghasemi, F. et al. (1994). Water table rise and the expansion Other ‘‘grab’’ samples can be collected using a drilling
of land and water resources salinsation: a global view: rig wherein a sampler is drilled or pushed to the target
IAH Congress Proceedings. Water Down Under, Adelaide, depth and the sampler is pushed a bit further into the
Australia, pp. 607–610. aquifer. Water is allowed to collect in the sampler. The
 GROUNDWATER SAMPLING TECHNIQUES FOR ENVIRONMENTAL PROJECTS 455

water is removed from the sampler using a bailer or The low-flow sampling method, typically used in 2-
by using a peristaltic pump and tubing, check-ball valve inch diameter groundwater monitoring wells, relates to
with tubing or other similar methods. These methods, the amount of drawdown in a well during purging. The
frequently called punch methods, are commonly used with indicator parameters are collected before and during
hollow stem auger rigs or direct push technology rigs. Cone purging and allowed to stabilize prior to groundwater
penetrometer (CPT) rigs can use specialized groundwater sample collection. Purge rates may be higher than
containers that are inserted into a cavity of the water sample rates to maximize purge efficiency. Prior to the
sampler tip. The groundwater containers are under a collection of the groundwater sample, the pump rate
vacuum. At the proper depth, the sampler is opened, may be lowered to lower sample turbidity and entrained
the groundwater container is pierced, and water enters air in the sample, and to mimic conditions that may
the container. exist in the natural aquifer. High turbidity in water
Temporary wells are used occasionally in environmen- samples being analyzed for metals requiores filtering
tal investigations. A temporary well consists of a well to remove colloidaland suspended sediments that may
screen below a casing riser placed in an open borehole. contain metals.
Several well volumes can be purged prior to sampling. Discrete sampling systems are designed to collect
A bailer or other method is used to bring the water to groundwater samples form the formation at a specific
the surface. After groundwater sampling is completed, the depth or multiple target depths without significant
borehole is usually tremie grouted to seal the annulus. purging prior to sample collection.

WATER WELLS GROUNDWATER SAMPLING PROTOCOL

Water wells are installed using a variety of rigs: hollow Depth to the top of the water and free-product thickness
stem auger, mud or air-rotary, air percussion, diamond within wells are monitored by using a water level meter or
coring, cable tool, and others. After well installation, an interface probe. Water level data collected from areas
wells are developed within a day to several days of that have multiple wells are used to develop groundwater
installation to remove residual drilling materials from contour maps for the project site. Groundwater flow is
the well bore and to improve well performance by perpendicular to equipotential lines drawn on the map. If
removing any fine material in the filter pack that can free-product is detected, a product sample is sometimes
pass from the native soil into the well. Well development collected for source identification. In environmental
techniques include pumping, bailing, surging, jetting, projects, where several chemicals are to be analyzed for
and airlifting. In most cases, surging and pumping are a given well, individual samples are collected in order
satisfactory. Development water is inspected for turbidity, of decreasing volatility. When the results from previous
product sheen, odors, and sediment. A minimum of sampling are known, it is recommended to start sampling
several casing volumes is removed during development. the cleanest wells first, moving to the most contaminated
The well is considered fully developed when consistent wells later in the sampling event.
pH, temperature, and conductivity readings indicate Each chemical analysis requires specific sample preser-
characteristic groundwater for the aquifer. If the aquifer vation techniques or additives. Sealed chemical ice is
is slow to recharge, development will continue until the placed in the coolers to maintain samples at a temper-
well is pumped dry. Cross contamination of wells from ature of 4 ◦ C. Where several types of chemicals are to be
pumps is avoided by using dedicated equipment and proper analyzed for a given well, individual samples are collected
decontamination procedures. in the following order: volatile organic compounds, purge-
Due to the large volumes of purge water generated able organic compounds, purgeable organic halogens, total
during groundwater sampling and the costs for stor- organic compounds, total organic halogens, extractable
age and disposal of the water, some regulatory agencies organic compounds, total metals, dissolved metals, phe-
are now allowing various low-flow groundwater sampling nols, and cyanide.
techniques for compliance groundwater monitoring. Propo-
nents of micropurging suggest that evacuation of a specific QUALITY CONTROL OF GROUNDWATER SAMPLES
volume of water may increase turbidity from agitation or
mixing of the well water column. Larger purge volume may A QC program, which is independent of the laboratory’s
increase the risk of colloid mobilization. Proper well devel- program, verifies sample integrity and includes the
opment immediately after well construction is essential submittal of duplicates, field blanks, and travel blanks to
when using low-flow sampling (7). the laboratory. The QC samples are packaged and sealed
Micropurging is the evacuation of tubing, sampling in the same manner as the other samples and are assigned
devices, and other items from the well prior to sampling independent sample numbers.
groundwater. There is no purging prior to sample A duplicate sample is collected for about 5% of the
collection with micropurging, so there is no method for samples or one per sampling round, whichever is greater.
determining if the well contains stagnant well water or The duplicate sample is acquired by filling separate
water representative of the aquifer. For this reason, many containers from the same well bailer as the actual sample.
regulatory agencies do not approve micropurging. Instead, The contents of the bailer are evenly divided between
when minimal purging is requested, low-flow sampling is the actual and duplicate samples to ensure duplication.
preferred. The field blank is submitted to the laboratory for the
456 GROUNDWATER SAMPLING WITH PASSIVE DIFFUSION SAMPLERS

same analyses as the rest of the sampling set. The semipermeable membrane (such as dialysis-type cellulose
field blank is acquired by dispensing deionized water or a synthetic polymer). On deployment, contaminants dif-
from a clean sampling bailer into the containers in the fuse through the membrane and are subsequently retained
same manner as groundwater samples. When sampling in the trapping phase until analysis. After the SPMD is
groundwater for analyzing volatile compounds, travel removed, the contaminants must typically be extracted
blanks are used to detect the introduction of contaminants from the trapping phase before analysis. A significant dis-
during transportation from the field to the laboratory. The advantage with SPMDs and dosimeter-type devices for
travel blank generally contains deionized water, which groundwater sampling is that back-calculation to actual
has been poured into the container prior to being taken in situ aqueous concentrations may not be straightforward
into the field. due to nonlinear and/or nonequilibrium sorption behavior
of the trapping phase (5).
BIBLIOGRAPHY Passive diffusion (PD) samplers are similar to
dosimeter-type and SPMD sampling methods in that they
1. Testa, S.M. (1994). Geologic Aspects of Hazardous Waste generally consist of a sealed container with a semiper-
Management. CRC Press, Boca Raton, FL, pp. 145–187.
meable membrane. The PD sampler is suspended in
2. Jacobs, J. (2000). Monitoring well construction and sampling a monitoring well, and the solutes of interest diffuse
techniques. In: Standard Handbook of Environmental Science,
across the membrane into the sampler. However, unlike
Health, and Technology. J. Lehr (Ed.). McGraw-Hill, New
dosimeter-type devices and SPMDs, PD samplers do not
York, pp. 11.46–11.68.
contain a trapping phase but are instead filled with water
3. American Society for Testing Materials. (1984). Standard
or air. Therefore, once the PD sampler is retrieved from
Practice for Description and Identification of Soils (Visual-
Manual Procedure), Method D 2488-84. Philadelphia, PA. the well (after equilibration), the concentration measured
in the sampler is assumed to relate directly to in situ
4. California Regional Water Quality Control Board. (1989).
Leaking Underground Fuel Tank (LUFT) Field Manual: analyte concentrations at the deployment location. Pas-
Guidelines for Site Assessment, Cleanup, and Underground sive diffusion samplers are less expensive to construct
Storage Tank Closure, Sacramento, CA. than SPMDs and easier to analyze because no extraction
5. Driscoll, F.G. (1986). Groundwater and Wells, 2nd Edn. step is required. Furthermore, the correlation between
Johnson Filtration Systems, St. Paul, MN, pp. 268–339. the contaminant concentration measured in the PD sam-
6. Sisk, S.J. (1981). NEIC Manual for Groundwater/Subsurface pler and the surrounding environment is straightforward:
Investigations at Hazardous Waste Sites. EPA-330/9-81-002. for water-filled PD samplers deployed in groundwater the
U.S. Environmental Protection Agency, Denver, CO. concentrations are equivalent.
7. Puls, R.W. and Barcelona, M.J. (1996). Low-flow (minimal Passive diffusion-type sampling methods have been
drawdown) ground water sampling procedures. EPA Ground- used in a variety of applications for many years,
water Issue, April. including dissolved gas measurement in seawater and
groundwater (3,6,7). Generally, PD samplers offer several
potential advantages for monitoring many common
GROUNDWATER SAMPLING WITH PASSIVE groundwater contaminants, compared to conventional
DIFFUSION SAMPLERS sampling techniques. For sites with ongoing monitoring
programs, significant cost savings may be realized due to
CRAIG E. DIVINE reduced sampling-related labor and minimal sampling-
Colorado School of Mines related waste generation. For dissolved constituents
Golden, Colorado with large Henry’s Law constants, such as volatile
organic compounds (VOCs) and dissolved gases, analyte
loss due to volatilization can be greatly minimized.
INTRODUCTION: DIFFUSION-BASED GROUNDWATER
In some cases, analytical matrix interferences may be
SAMPLING METHODS
reduced if the membrane is selective against interfering
Passive diffusion-based sampling methods have been used constituents (i.e., the matrix in the PD sampler may be
to monitor environmental contaminants, particularly trace comparatively less complex). Additionally, multiple PD
metals and certain lipophilic compounds including poly- samplers can be deployed to provide solute concentration
chlorinated biphenyls (PCBs), polynuclear aromatic hydro- measurements across very discrete intervals. For example,
carbons (PAHs), and pesticides, for many years. In particu- PD samplers with dialysis-type membranes have been
lar, dosimeter-type and semipermeable membrane devices used to characterize ion and trace metal concentration
(SPMDs) are commonly used in limnology, oceanography, profiles on the scale of a few centimeters in sediment (3).
and ecotoxicology for evaluating ultralow concentrations
and potential bioavailability of contaminants (1–4). These MEMBRANE MATERIALS AND SPECIFIC CONSTRUCTION
types of samplers are most frequently used for sam- EXAMPLES
pling sediment porewater and surface water (i.e., rivers,
estuaries, harbors) and have occasionally been used for Because of the potentially significant costs savings and
groundwater characterization in contaminant toxicology technical advantages noted above, the use of PD samplers
studies (5). These types of sampling devices are composed for groundwater monitoring at contaminated sites is
of a ‘‘trapping’’ phase, such as hexane, activated carbon, greatly increasing [particularly water-filled PD samplers
or triolein (a prominent fish lipid), contained within a using low-density polyethylene (LDPE) membranes].
 GROUNDWATER SAMPLING WITH PASSIVE DIFFUSION SAMPLERS 457

(a) Grommet

LDPE membrane
(0.010 cm thick)

~0.6 m
Diameter = ~3.2 cm

Lay-flat LDPE tubing

filled with ~220 mL
deionized H2O

Grommet

(b)
LDPE membrane

Open mouth

Tig
hte
n to
Rubber band sea
Threads (seal gasket) l

Nylon cable tie

LDPE membrane
Glass vial
(gas-or water-filled)

Figure 1. Schematics of passive diffusion

samplers.

Laboratory and field compatibility tests indicate that made of lay-flat LDPE tubing filled with deionized water
LDPE membranes are suitable for most fuel and solvent are commercially available from EON Products, Inc.
VOCs; however, it has been demonstrated that the (Snellville, Georgia) and Columbia Analytical, Inc. (Kelso,
LDPE membrane is generally not suitable for measuring Washington). A schematic of this type of sampler, which
inorganic ions, metals, pesticides, explosives, semivolatile is also known as a PD ‘‘bag’’ sampler is shown in Fig. 1b.
organics, alcohols, organic acids, and ethers [including For a description of peepers and other PD samplers with
methyl-t-butyl ether (MtBE)] with PD samplers (8,9). dialysis-type membranes, see Davison et al. (3).
Dissolved gases, such as He, Ne, H2 , O2 , and N2 ,
can be measured by PD samplers with LDPE (10) or EQUILIBRATION TIME FOR PASSIVE DIFFUSION
silicone (7) membranes. Passive diffusion-type samplers SAMPLERS
(also known as dialysis samplers or ‘‘peepers’’) that use
other membrane materials, such as cellulose acetate, can To prevent biased-low measurements, PD samplers must
be used to monitor major ions, metals, nutrients, dissolved be deployed long enough in a well for equilibration to
gases, and pH (3,11). occur. Based on empirical laboratory and field studies,
Passive diffusion samplers can be constructed easily a minimum of 2–3 days is required for equilibration
from inexpensive materials. As shown in Fig. 1a, the open for most VOCs using commercially available LDPE
mouth of a glass laboratory vial can be covered with a PD bag samplers (8). Carnigan (12) found that some
membrane and sealed with a rubber band gasket and dialysis-type samplers require equilibration times of 3–20
nylon zip tie. Drop sheeting, which is widely available in days for measuring inorganics in porewater. Based on
the painting supply sections of hardware stores, can be these general results, a minimum equilibration time of
used for LDPE membranes and is offered in a range approximately 2 weeks is commonly recommended as a
of thicknesses. PD samplers filled with water can be conservative approach in field studies (8,9). However, note
constructed while submerged in deionized water to prevent that this is a generic rule of thumb. As equilibration is a
entrapment of air bubbles. Passive diffusion samplers diffusion process, the specific equilibration time depends
458 GROUNDWATER SAMPLING WITH PASSIVE DIFFUSION SAMPLERS

on sampler volume, membrane surface area, membrane PD samplers, including both constructed (Fig. 1a) and
thickness, and the diffusion coefficient of the solute in the commercially available (Fig. 1b) samplers. The specific
membrane material. parameter values and resulting equilibration times are
Sanford et al. (2) present the following solution that summarized in Table 1, and the predicted PD sampler
describes the concentration within the PD sampler as it concentration histories are shown in Fig. 2.
equilibrates with the surrounding groundwater: These results show the relative importance of the
   various parameters for equilibration times. For example,
−Dm At
increasing the A/∀ ratio reduces the equilibration time,
Cs (t) = Cr H 1 − e ∀Lm  (1) and this can be accomplished for constructed PD samplers
(Fig. 1a) by filling the sampler with inert sand or glass
where beads (i.e., compare results of Scenario 3 with Scenario
4). It is also important to note that Eq. 1 is valid only for
Cs (t) concentration within the sampler at time t after advection-dominated systems. If the groundwater velocity
PD sampler deployment [M L−3 ], is very low (diffusion-dominated systems), a concentration
Cr resident concentration at the measurement location gradient will develop outside of the sampler, and
[M L−3 ], equilibration times can be much longer (13). Equation 1 is
appropriate, however, for most shallow aquifer systems.
Dm effective constituent diffusion coefficient for the
membrane material [L2 T−1 ],
CASE STUDY: APPLICATION OF PASSIVE DIFFUSIONS
H solute dimensionless Henry’s law constant [ ], SAMPLERS AT A HETEROGENEOUS SITE
A surface area of the PD sampler membrane [L2 ],
t deployment time [T], Passive diffusion samplers measure water quality directly
∀ internal volume of the PD sample [L3 ], and in contact with the membrane surface; therefore, they
generally measure concentrations at relatively discrete
Lm membrane thickness [L].
intervals within a well. Conversely, samples collected by
This solution assumes Fickian solute diffusion across conventional purge-and-bail methods tend to integrate
a thin membrane, a constant Cr concentration at the water quality across the entire well screen region and
measurement location, and a well-mixed reservoir in therefore are most significantly influenced by water
the PD sampler. As presented, Equation 1 incorporates quality conditions in the highest permeable zones across
air–water partitioning and is applicable to gas-filled PD the well screen interval. In other words, PD samplers
samplers deployed in groundwater, such as those described tend to measure porewater concentrations over relatively
by Sanford et al. (7) and Divine and McCray (10). For small intervals, and purge-and-bail type methods tend
conditions where no phase partitioning occurs (such as to measure flux-weighted concentrations. For vertically
a water-filled PD sampler deployed in groundwater), uniform plume distributions and homogeneous aquifer
Henry’s law constant (i.e., the gas-water partition
coefficient) is treated as unity. 1.2
Normalized PD sampler

As clearly shown in Eq. 1, the rate at which a PD 1.0

concentration

sampler equilibrates relates directly to the Dm value

0.8
of the solute, the sampler dimensions (A, ∀), and the Scenario 1
thickness of the membrane (Lm ). Therefore, PD samplers 0.6
Scenario 2
can be constructed to minimize equilibration times by 0.4
using a thin membrane and increasing the A/∀ ratio. Scenario 3
0.2
The effective Dm value is temperature-dependent and Scenario 4
specific to both the membrane and solute. Divine and 0.0
0 1 2 3 4 5 6 7
McCray (10) estimated that a practical range is 1–4 × Deployment time, day
10−7 cm2 s−1 for Dm values of several dissolved gases and Figure 2. Passive diffusion sampler concentrations after deploy-
VOCs for LDPE membranes. Sanford et al. (2) measured ment, predicted by Equation 1. The normalized concentration is
Dm values for helium and neon of 4 × 10−7 cm2 s−1 defined as Cs /Cr (at equilibrium Cs = Cr ; therefore, Cs /Cr = 1.0).
for silicone membranes. Based on these Dm ranges, The parameter values used in the models are summarized in
equilibration times are predicted using Eq. 1 for several Table 1.

Table 1. Passive Diffusion Sampler Dimensions, Membrane Thickness, Membrane Diffusion Coefficients, and Predicted
Equilibration Times for Various Constructed (Fig. 1a) and Commercially Available (Fig. 1b) PD Samplers. The
Equilibration Time is Defined by Equation 1 when Cs ≥ 0.95Cr )
A, ∀, Lm , Dm , Equilibration
cm2 mL cm (cm2 s−1 ) Time, Day

Scenario 1 Commercial ‘‘bag’’ sampler 350 220 0.010 4.0 × 10−7 0.7
Scenario 2 Commercial ‘‘bag’’ sampler 350 220 0.010 2.5 × 10−7 2.7
Scenario 3 Constructed sampler: sand-filled 20-mL glass vial 3.2 12.9 0.005 2.5 × 10−7 2.8
Scenario 4 Constructed sampler: 20-mL glass vial 3.2 23 0.005 2.5 × 10−7 5.1
 GROUNDWATER SAMPLING WITH PASSIVE DIFFUSION SAMPLERS 459

1,1-DCE concentration, µg/L

Lithology
0 200 400 600 800 1000 1200 1400 1600 1800 2000
16 Weathered
110 sandstone
17 110 1300
Sandstone/
18 claystone
19 140 380
Sandstone
Depth, ft bls

20
21 910 1500
Claystone,
22
1100 unfractured
23 1600
Purge-and-bail
Sandstone
24
Low-flow sampling
25
Figure 3. Comparison of sampling results for
Passive diffusion sampler Claystone, few
fractures
purge-and-bail, low-flow, and passive diffusion sam-
26 pling methods for a single monitoring well.

conditions, results from these two sampling methods may 10000

be equivalent. However, these two methods may yield
significantly different concentration values for a specific
well at a heterogeneous site with plume stratification.
To illustrate this potential behavior, the concentrations Purge-and-bail method, µg/L
1000
of 1,1-dichloroethene (1,1-DCE) (a common VOC) mea-
sured at a monitoring well by purge-and-bail, low-flow 1:1 line
[another discrete-type sampling technique; see Puls and
Barcelona (14)]. PD sampling methods are shown in Fig. 3.
100
The aquifer at this study site is highly heterogeneous, and
this is clearly expressed in the results of the various sam-
pling methods. Only one value (1300 µg/L) is obtained by
the purge-and-bail method; however, the low-flow and PD
sampling methods characterize concentrations at various 10
discrete depth intervals and show significant contaminant Not detected
stratification (approximately an order of magnitude over by purge-and-bail
several vertical feet).
Figure 4 compares the results of purge-and-bail with 1
1 10 100 1000 10000
PD sampling methods at several wells at the same
Passive diffusion sampler method, µg/L
heterogeneous site. The purge-and-bail measurements are
generally higher than the average discrete PD sampler Figure 4. Comparison of sampling results (various VOCs) for
purge-and-bail and passive diffusion sampling methods for a
measurements, possibly indicating that the majority of the
series of wells at a heterogeneous site.
contaminant mass exists in highly permeable zones. These
data provide additional characterization information and
have potentially important implications for the site. At measurement within site objectives or regulatory require-
some sites, a similar targeted dual-method sampling ments may be challenging. For example, an analytical
strategy (i.e., use of both purge-and-bail and PD sampling water quality goal or regulatory limit may be met by one
methods) may be useful for evaluating contaminant method and not by another at the same well simply due
distribution, remedial system design and performance, to the inherent averaging characteristics of the sampling
and long-term plume behavior. methods. For additional discussion of the results of this
Clearly, using PD samplers under heterogeneous study, see Divine et al. (15).
aquifer conditions, particularly in an ongoing compli-
ance monitoring program, requires careful consideration BIBLIOGRAPHY
of ways to interpret measurement data and satisfy mon-
1. Sordergren, A. (1987). Solvent-filled dialysis membranes
itoring program objectives. For example, determining the
simulate uptake of pollutants by aquatic organisms. Environ.
appropriate intervals for PD sampler placement may be
Sci. Technol. 21(9): 855–859.
challenging (i.e., Fig. 3). Additionally, as indicated by the
2. DiGiano, F.A., Elliot, D., and Leith, D. (1988). Application of
circled values in Fig. 4, the PD sampling method mea-
passive dosimetery to the detection of trace organic contami-
sured low concentrations of some nontarget VOCs at a few nants in water. Environ. Sci. Technol. 22(11): 1365–1367.
discrete intervals in certain wells, but these compounds 3. Davidson, W., Fones, G., Harper, M., Teasdale, P., and
were not measured above the analytical quantitation limit Zhang, H. (2000). Dialysis, DET, and DGT: In situ diffusional
by the purge-and-bail method. Assuming that these dif- techniques for studying water, sediments, and soils. In: In
ferences are related to the differing averaging natures of Situ Monitoring of Aquatic Systems. J. Buffle and G. Horvai
the sampling methods, interpreting the discrete interval (Eds.). John Wiley & Sons, New York, Chapter 11.
460 SPECIFIC CAPACITY

4. Dunn, R.K., Teasdale, P.R., Warnken, J., and Schleich, R.R. capacity to estimate the optimum depth for installing a
(2003). Evaluation of diffusive gradient in a thin film pump to obtain the desired yield. Units typically associated
technique for monitoring trace metal concentration in with specific capacity are liters per second per meter
estuarine waters. Environ. Sci. Technol. 37(12): 2794–2800.
of drawdown (l/s/m), cubic meters per day per meter
5. Gustavson, K.E. and Harkine, J.M. (2000). Comparison of of drawdown (m3 /day/m), cubic feet per day per foot of
sampling techniques and evaluation of semipermeable
drawdown (ft3 /day/ft), or gallons per minute per foot of
membrane devices (SPMDs) for monitoring polynuclear
aromatic hydrocarbons (PAHs) in groundwater. Environ. Sci.
drawdown (gpm/ft).
Technol. 34(20): 4445–4451. When taking measurements to compute specific capac-
6. Dyck, W. and Da Silva, F.G. (1981). The use of ping-pong
ity, the pumping rate should be constant, and the period
balls and latex tubing for sampling the helium content of lake of pumping should be sufficiently long that the rate of
sediments. J. Geochem. Exploration 14: 41–48. change in drawdown is small. The specific capacity of
7. Sanford, W.E., Shropshire, R.G., and Solomon, D.K. (1996). a well tends to become smaller as the pumping rate and
Dissolved gas tracers in groundwater: Simplified injection, duration of pumping increase. At high pumping rates, non-
sampling, and analysis. Water Resour. Res. 32(6): 1635–1642. laminar flow losses are larger and result in an increased
8. Vroblesky, D.A. and Campbell, T.R. (2001). Equilibration rate of drawdown. Water level drawdown in unconfined
times, compound selectivity, and stability of diffusion aquifers results in reduced aquifer transmissivity near
samplers for collection of ground-water VOC concentrations. pumping wells; larger drawdown at high pumping rates
Adv. Environ. Res. 5: 1–12. then results in progressively smaller transmissivity and
9. Vroblesky, D.A. (2001). User’s Guide for Polyethylene-Based smaller specific capacity. A specific capacity value remark-
Passive Diffusion Bag Samplers to Obtain Volatile Organic ably smaller than expected may result when the well is
Compound Concentrations in Wells, Part 1: Deployment, not screened in the correct water-bearing zone or when
Recovery, Data Interpretation, and Quality Control and
well screen openings are clogged due to incomplete well
Assurance. United States Geological Survey, Columbia, SC.
development, precipitation of minerals in the screen open-
10. Divine, C.E. and McCray, J.E. (2004). Estimation of diffusion
ings, or biofouling. After many years of pumping, if the
coefficients and equilibration times for LDPE passive
diffusion samplers. Environ. Sci. Technol. 38(6): 1849–1857.
specific capacity declines to about half the original value,
well rehabilitation should be considered.
11. Webster, I.T., Teasdale, P.R., and Grigg, N.J. (1998). The-
oretical and experimental analysis of peeper equilibration Specific capacity may be used to estimate the maximum
dynamics. Environ. Sci. Technol. 32(11): 1727–1733. potential pumping rate of a well by multiplying the
12. Carnigan, R. (1984). Interstitial water sampling by dialysis, available drawdown by specific capacity. The available
methodological notes. Limnol. Oceanogr. 29: 667–670. drawdown is the difference between the depth at which
13. Harrington, G.A., Cook, P.G., and Robinson, N.I. (2000). the pump is set and the depth to the nonpumping water
Equilibration times of gas-field diffusion samplers in slow- level in the well.
moving ground water systems. Ground Water Monitoring Specific capacity may be used to estimate the
Remediation 20(2): 60. transmissivity of aquifer media in the vicinity of the
14. Puls, R.W. and Barcelona, M.J. (1996). Low-Flow (minimal pumping well. Theis (1) suggested that specific capacity
Drawdown) Ground-Water Sampling Procedures. EPA/540/S- could be used to estimate the transmissivity of an aquifer
95/504, Ada, OK. using the equation,
15. Divine, C.E., Madsen, L., Andrews, S.D., and Santangelo-
Drieling, T. (in press). Passive diffusion ground water Q 2.3 2.25T ∗ t
samplers at a heterogeneous site: Pilot study results. Ground T∗ = log
h 4π r2 S
Water Monitoring and Remediation.
Q
where =specific capacity
h
SPECIFIC CAPACITY t=time
r=radius of pumping well
ERIC S. WILSON S=aquifer storativity
E. L. Montgomery & Associates, T ∗ =transmissivity; an initial guess of the value
Inc. of T is made and substituted in the
Tucson, Arizona equation. The equation is then solved for
h. After the initial solution, the value of T
must then be adjusted until the calculated
The specific capacity of a well is the pumping rate of the value of h is reasonably close to the
well divided by the drawdown at that rate: measured value.
Q
Sc = Driscoll (2) indicated that transmissivity could be
h
estimated as
where Sc =specific capacity
Q=constant pumping rate, and
T = Sc × 2, 000 for confined aquifers
h=drawdown

Water well contractors often report specific capacity to or

characterize the performance of a well and use specific T = Sc × 1, 500 for unconfined aquifers
 SOIL WATER 461

where T is in gpd/ft and where the range in water content is from a minimum,
Sc is specific capacity. or residual, saturation, θr , to a maximum equal to the
saturated porosity, θs , so that θr < θ < θs .
Razack and Huntley (3) gave an empirical relationship The water content can also be defined with a
between transmissivity and specific capacity as follows: gravimetric, or mass, measurement, θg , such that

T = 15.3(Sc )0.67 Mw
θg = (2)
2
Ms
where T is in m /day
Sc in m3 /day/m, where Mw is the water mass and Ms is the soil mass. We
can readily convert between these two measures by using:
or
T = 33.6(Sc )0.67
ρw θg
θ = θg = (3)
where T is in ft2 /day ρs ρs
Sc in ft3 /day/ft.
where ρw = Mw /Vw = 1 kg/L is the water density and
Using specific capacity data to estimate transmissivity ρs = Ms /Vs (kg/L) is the bulk density of the soil.
can be useful when trying to estimate potential pumping The soil water content can also be reported as a relative
rates or estimating how much available head is needed to saturation, , which is defined using
produce a specified pumping rate. But, as in any estimation
technique, it is not always applicable or advisable. θ − θr
= (4)
These instances may include situations where substantial θs − θr
drawdown occurs because the pumping duration is too
short for the specific capacity test, too small a pumping where θr and θs are the residual and saturated water
rate is used for an adequate specific capacity test, or well contents, respectively. The relative saturation varies from
construction is inefficient. zero for a dry soil to a maximum of one for a fully saturated
soil, 0 < < 1.

BIBLIOGRAPHY
MEASURING WATER CONTENTS
1. Theis, C.V. (1963). Estimating the Transmissibility of Aquifers
from the Specific Capacity of Wells. In: Methods of Determining Water contents are commonly measured with time-domain
Permeability, Transmissibility, and Drawdown. R. Bentall, reflectometry (TDR) devices, as shown in (Fig. 1). TDRs
compiler, U.S. Geological Survey Water-Supply Paper 1536-I,
determine the water content with the electromagnetic
pp. 331–340.
properties of a wave pulse passing through a conducting
2. Driscoll, F.G. (1986). Groundwater and Wells, 2nd Edn.
set of rods (such as 3-mm stainless-steel welding rods)
Johnson Filtration Systems, Inc., St. Paul, MN.
placed in the soil.
3. Razack, M. and Huntley, D. (1991). Assessing transmissivity
The principle of TDRs is that the velocity of electro-
from specific capacity in a large and heterogeneous alluvial
aquifer. J. Ground Water 29(6): 856–861.
magnetic waves along a conductor is a function of the
dielectric coefficient of the media around the conductor.
Liquid water has a dielectric constant of 80.2 at 20 ◦ C,
SOIL WATER whereas ice is 3.2, petroleum is 1.8 to 2.2, quartz is 4.3,
and air is only 1.00.
TODD RASMUSSEN
The University of Georgia
Athens, Georgia Exterior
shield
Coaxial cable

Soil water is a measure of the quantity of water present

within a unit volume of soil. Soil-water content can vary Block
Conductor
from fully saturated, in which case the water content
equals the soil porosity, to oven-dry, in which case all free
water has been evaporated. Soil
TDR rods
Wet soils are characterized as having high water
contents, which means that more soil pores are filled
with water, whereas dry soils have more pores that are
filled with air.
The volumetric water content, θ , is defined as the Figure 1. Time-domain reflectometer (TDR) probe that mea-
volume of water, Vw , per unit volume of soil, Vs : sures soil-water content. The three-rod assembly is placed within
the soil. Coaxial cable connects probe to the water content tester.
Vw Center wire of coaxial cable connects to the central rod, whereas
θ= (1) the exterior shield is connected to the two outer rods.
Vs
462 SOIL WATER

As the water dielectric constant is so much greater for equilibrium with the soil fluids using a semiporous ceramic
water than for soil or air, the wave velocity is substantially or metal filter that prevents air from entering the probe.
retarded as the water content of a soil increases, because Psychrometry is a useful technique in drier soils,
larger dielectric constants cause slower wave velocities ψ > 5 bar (2). The dewpoint temperature of the pore atmo-
and, hence, longer travel times. sphere is measured with a thermocouple psychrometer.
Before the use of TDRs, neutron probes used the The dewpoint temperature can then, along with the pore
thermalization of neutrons to determine the water temperature, determine the relative humidity of the pore
content (1). This older method has fallen out of favor, atmosphere. The matric tension can be readily determined
however, because of the rigorous licensing and inspection from the relative humidity, as shown in Table 1.
requirements associated with radioactive materials.
MOISTURE CHARACTERISTIC CURVES
MATRIC TENSION
The amount of water contained within soil can be related
Water molecules in the unsaturated zone are bound by to the fluid tension, or negative pressure, present within
adhesive forces to soil surfaces and by cohesive forces to the water. Dry soils contain water that is held by strong
other water molecules, which are surface tension forces, adhesive forces to soil surfaces along with strong cohesive
and they result in fluid pressures that are less than forces to other water molecules. These forces are reflected
atmospheric. in the tension that holds the water against the force of
Rather than use negative pressure heads, p < 0, to gravity.
describe the soil potential, we employ soil, or matric, Each soil has a unique relationship between soil tension
tensions, ψ > 0. The matric tension is the negative fluid and water content. This relationship, termed the soil
pressure head, ψ = −p, which increases as the pressure moisture characteristic curve, is a function that describes
head becomes more negative (more tightly bound) and how water content and matric tension are related. The
decreases as the pressure head becomes less negative (less characteristic curve for each material varies with the pore
tightly bound). size distribution—a soil with larger pores releases water
The total head within the soil pore fluids, h, is the at a lower tension than does a soil with smaller pores.
difference between the elevation, z, and the matric tension: A number of formulations have been proposed that
relate the relative saturation, , to matric tension, ψ. The
H =z+p=z−ψ (5) Brooks and Corey equation has the form:

−λ
ψ
Capillary theory can relate the matric tension, ψ, to the = (7)
fluid potential associated with the radius of curvature, r, ψb
of fluid films within pores:
where ψb is the bubbling pressure and λ is a pore-
2σ cos α size distribution index. This equation only holds for
ψ= (6) ψ > ψo − = 1 otherwise.
γr
Another common formulation is the van Genuchten
equation:
where σ (Pa·m) is the surface tension and α is the contact  (1−n)/n
angle at the air–water–solid interface. = 1 + (αψ)n (8)
The matric tension of a wet soil, 0 < ψ < 0.7 bar, can be
measured with tensiometers, as shown in (Fig. 2), which where α and n are empirical coefficients.
determine the negative pressure head of fluids inside a Moisture characteristic curves are commonly estimated
probe. The fluid within the probe is allowed to reach with pressure plates (shown in (Fig. 3)) or Tempe pressure
cells. In both methods, one or more soil samples are placed
on top of a porous plate within a sealed chamber. Nitrogen
gas under high pressure is forced into the chamber, and
Vacuum gage
water is forced out of the soil samples through the porous

Table 1. Measures of Soil Tension

Water-filled tube Fluid Pressure Soil Tension Relative
(bars) (m) Pore Radius Humidity (%)
Soil
0 0 ∞ 100
0.1 −1 15 µm 99.993
Porous cup 1 −10 1.5 µm 99.93
15 −150 100 nm 98.9
100 −1000 15 nm 93.
Figure 2. Soil-water tensiometer probe that monitors soil- 500 −5,000 3 nm 70.
tension. A porous cup allows water to exchange between the 1,000 −10,000 1.5 nm 48.
soil and a reservoir of water within the tube assembly. A vacuum 5,000 −50,000 3 Å 2.6
pressure gauge monitors the water pressure inside the water 10,000 −100,000 1.5 Å 0.068
reservoir.
 SOIL AND GROUNDWATER GEOCHEMISTRY AND MICROBIOLOGY 463

in, a geological matrix, including soil, that complex

Pressure chamber
mixture of minerals, life, and organic material on the
surface of the earth. Groundwater, of course, is water
located in the saturated zone of the earth. Microbiology
Gas inflow
is the study of small forms of life, usually including the
bacteria and archaea domains and protists and fungi of the
eukaryaceae domain, and naturally also covers numerous
subcategories.
The interdisciplinary studies of the interactions among
microflora, earth materials, and chemical reactions are
Soil samples Porous plate often linked under geomicrobiology or biogeochemistry.

ORIGINS
Water outflow
Figure 3. Pressure plate procedure for determining the soil The relationship between life in soil and earth materials
moisture characteristic curve. A gas pressure is applied to soil
appears to go back to the very origins of life on
samples within the chamber, and water is released through the
earth. Bacterial and archaeans (superficially bacteria-like
porous plate. The water content, θ, is found for each pressure, ψ.
microflora that have entirely different genetic operations,
biochemistry, and cell membranes) were interacting with
plate, which is permeable to water but not to air. The iron and other minerals as part of energy transformations
resulting water content of the samples can be determined since within a few 100 million years of the coalescence of
once the outflow has stopped. the planet. Many archaeans specialize in methanogenesis
and adaptation to extreme environments at the present
SAMPLING SOIL WATER time, such as geothermal waters, undersea hydrothermal
vents, and hypersaline lakes (e.g., Dead Sea, Great
Soil water samples are more difficult to collect in the Salt Lake, and solar salt lagoons). Biochemical evidence
unsaturated zone than in the saturated zone because suggests that they may have developed in association
water does not freely flow into a well or piezometer. with such conditions, which were common early in
Commonly, a soil lysimeter collects water samples by earth history.
placing a suction within a hollow tube that has a porous Bacteria and cyanobacteria (colloquially known as blue-
cup placed within the soil. Water flows through the porous green algae) strongly affected the earth environment
cup into the hollow tube where it can then be collected. from early in the Archean era (±4 billion years before
Soil lysimeters are limited to conditions where the soils present, b.y. BP). Cyanobacteria, with a photosynthetic
are wet, ψ < 0.7 bar. For drier conditions, soil samples mechanism similar to algae and plants, used carbon
can be collected and returned to the laboratory, where dioxide and expended oxygen as a waste product in
mechanical or gas pressure can be applied to the samples shallow anoxic seas. Ferrous iron (Fe), common in water
to force pore fluids out of the sample. and the earth’s crust, scavenged the oxygen, forming
various FeIII minerals, including magnetite, hematite
BIBLIOGRAPHY (recently confirmed on Mars), goethite, and less-organized
mineral forms. Such iron-containing Archean sediments
1. Elder, A.N. and Rasmussen, T.C. (1994). Neutron probe were deposited as banded iron formations, with iron
calibration in unsaturated tuff. Soil Sci. Soc. Am. J. 58(5): bands in a cherty matrix, which are important iron ores
1301–1307.
worldwide where they occur (such as the Mesabi Range in
2. Rasmussen, T.C. and Rhodes, S.C. (1994). Energy-related Minnesota). When much of the available free oceanic FeII
methods: Psychrometers. In: Vadose Zone Characterization
was depleted, oxygen began to build up in the atmosphere,
and Monitoring; Principles, Methods and Case Studies.
L.G. Wilson, L. Everett, and S. Cullen (Eds.). Lewis Publish-
with a recognizably modern atmosphere dating from 1.7
ers, Boca Raton, FL. b.y. BP.
Bacteria have the most diverse biochemical and
respiratory capabilities among forms of life. Although
SOIL AND GROUNDWATER GEOCHEMISTRY others require oxygen for respiration, bacteria can respire
using sulfur, forms of nitrogen, carbon dioxide, and
AND MICROBIOLOGY
metals (primarily Fe and Mn) as electron acceptors. Such
STUART SMITH respiration is used to process a wide variety of organic
Smith-Comeskey GroundWater compounds, including recalcitrant ring structures. Such
Science LLC versatility appears to be a very ancient capability.
Upper Sandusky, Ohio In ancient shallow marine settings, cyanobacteria
formed stromatolites, laminations of organic and sedi-
DEFINITIONS mentary structures formed as sand and debris, and copre-
cipitated minerals were trapped in cyanobacterial masses.
Geochemistry can be defined as the study of the chemical Although many stromatolites are found as fossil struc-
composition of, and of actual or possible chemical changes tures, a number of locations exist on the modern-day Earth
464 SOIL AND GROUNDWATER GEOCHEMISTRY AND MICROBIOLOGY

where stromatolites are still forming, which are one form OVERCOMING CHALLENGES IN MICROBIOLOGICAL
of the common microbial phenomenon, biofilm formation. METHODS IN GROUNDWATER STUDIES
From there on, microbial processes and mineral-
biological interactions, and the environment-manipulating Background
‘‘behavior’’ known as biofilm formation (all developed in the
Environmental microbiology and geomicrobiology are the
Archean era), have continued in action until the present.
oldest disciplines within the broader field of microbiology.
However, industrial, public health, and clinical interests
• Bog ores (important to the development of an iron quickly overshadowed the study of the microbiology
industry in the United States during the nation- of natural environments. Until the 1970s, scientific
forming period) and gold and copper placer ores preconceptions in many circles, based on observations
are younger manifestations of the ancient microbial of engineered filters, limited methods (and limited
practice of depositing metals. reading) and placed limits on groundwater microbiology.
• Bacteria reduce oxides of Fe, Mn, N, and S to accu- Considerations of the chemistry of groundwater assumed
mulate ‘‘reducing power’’ to oxidize (‘‘digest’’) complex that chemical transformations occurred in an environment
organic compounds. These capacities are exploited in devoid of life. A notable exception was that of the
groundwater contamination remediation. The pres- microbiologists in the Soviet petroleum industry of the
ence of high levels of Fe and Mn in well water samples 1950s and 1960s, who described the microbiology of
is also an indicator that organic carbon compounds petroleum reservoirs and systems.
are present in groundwater accessed by a well.
• Microbe-mediated biogeochemical transformations Analysis
alter groundwater quality by oxidizing or reducing Fe
Methods for the analysis of groundwater environments
and Mn, and mobilizing them into solution. Recently,
for microbes were very limited prior to the 1980s. The
the reduction of insoluble arsenic minerals to soluble
principal analytical method was culturing, which recovers
forms and As mobilization by microbes has attracted
only a small fraction of bacteria. It was generally assumed
interest as a factor in As presence in groundwater.
that water passing through the soil was purified by active
• Biofilm-forming microflora coat, clog, and corrode microbial processes and by filtration.
engineered structures such as wells and drains in Interest in the occurrence of biofouling microflora in
the saturated zone, resulting in biofouling (further groundwater and concerns about groundwater contam-
discussed in the article BIOFOULING IN WATER WELLS). ination in the 1970s and 1980s motivated innovation
in methods suitable for groundwater microbiology. As
SOIL MICROBIOLOGY groundwater contamination became more and more evi-
dent in the United States and other western nations during
the 1980s, the motivation for understanding groundwa-
Soil microbiology is a special category of geomicrobiology
ter environments increased. In addition, new methods
focused on the earth’s regolith, and specifically soil,
in microbiology, based on advances in molecular biol-
which is a complex of mineral, organic, and biological
ogy, provided microbiologists with new tools to explore
components. Although beyond the scope of a short article,
this difficult-to-sample microbial habitat. Among these
soil microbiology has traditionally been distinguished
are phospholipid fatty acid methyl ester (PL-FAME) and
from groundwater or geomicrobiology. This distinction
nucleic acid testing, described further in EVALUATION OF
has origins in academic discipline; however, studies
MICROBIAL COMPONENTS OF BIOFOULING. Such methods provide
indicate that surficial soils have species richness and total
the means to describe microbial associations in a mean-
microbial biomass values that are much higher than in
ingful way, as cultivation and phenotypic classification
earth materials below the soil zone. Protists and fungi are
typically fail to do so.
common in soil but limited in type and numbers (but not
Classification, in particular, is a problem in description.
entirely absent) in aquifers.
With micro-organisms, it often depends on making
Soil microbiology is much more thoroughly known com-
comparisons with libraries of known genotypes or
pared with that of less-accessible earth materials. It is
phenotypes. Some deep subsurface communities are so
also much more thoroughly influenced by human activi-
isolated and specifically adapted that comparisons with
ties, which include cultivation, fertilization, construction,
described forms are weak. A general recognition exists in
and application of contaminants (both deliberately and
the geobiological community that the task of describing
accidentally). Soil microflora and soil structure experience
aquifer and other deep subsurface communities compares
close interactions with animals and plants. Changes in
with the task of describing the recently discovered deep-
such macroflora and fauna communities can trigger sweep-
sea hydrothermal vent ecosystems.
ing changes in soil microbiology. Clearing and cultivation
of forest and prairie soils result in simplification in micro-
Sampling
bial community structure. The introduction of Eurasian
earthworms into North American boreal soils (where they Acquiring samples is perhaps the most difficult aspect of
have been absent since the Pleistocene glaciations) has the microbiology of the deep subsurface. Such sampling
resulted in measurable shifts in the availability of organic is relatively expensive, at times technically challenging,
matter and nutrients to native plants. In turn, changes in and prone to sample compromise. Typically, methods for
vegetation affect soil and watershed hydrology. both solid and liquid sample acquisitions are adapted
 CHARACTERIZING SOIL SPATIAL VARIABILITY 465

from those developed for other purposes (borehole and READING LIST
core drilling, pumped or zonal sampling). Only coring can
deliver solid samples isolated from drilling contamination. Ehrlich, H.L. (2002). Geomicrobiology, 4th Edn. Marcel Dekker,
A variety of water sampling techniques exist to obtain New York.
discrete samples. Krumbein, W.E. (Ed.). (1983). Microbial Geochemistry. Blackwell
However, if properly interpreted, available sampling Scientific Publications, Oxford, UK.
and analytical techniques appear to be sufficient to gain Bitton, G. and Gerba, C.P. (Eds.). (1984). Groundwater Pollution
Microbiology. Wiley-Interscience, New York.
an understanding of deep subsurface microbial ecology.
Sinclair, J.L. and Ghiorse, W.C. (1989). Distribution of aerobic
bacteria, protozoa, algae and fungi in deep subsurface
NOTABLE DISCOVERIES sediments. Geomicrobiology. 7(1–2): 15–32.
Chapelle, F.H. (1993). Ground-Water Microbiology and Geochem-
istry. John Wiley & Sons, New York.
1. Microbial community diversity: PL-FAME and rRNA
Chapelle, F.H. (2000). Ground-Water Microbiology and Geochem-
analysis reveal subsurface communities that are
istry, 2nd Edn. John Wiley & Sons, New York.
diverse and typically adapted to coping with
Madsen, E.L. and Ghiorse, W.C. (1993). Groundwater microbiol-
starvation and exploiting resource ‘‘windfalls’’ when
ogy: subsurface ecosystem processes. In: Aquatic Microbiology.
they occur. Bacteria can persist in viable but T.E. Ford (Ed.). Blackwell Scientific Publications, Boston, MA.
nonculturable and inactive states when conditions
LeClerc, H. (2004). Microbiology of Natural Mineral Waters.
dictate. Microbial communities can have biomass Centre Evian. Available: http://www.centre-evian.com.
that rivals surface soils and may include fungi and
Madsen, E. (2004). Bibliography of important works in envi-
protozoa where pore sizes permit. ronmental microbiology. Available: http://www.micro.cornell.
2. Microbial communities and numbers vary signifi- edu/faculty/Madsen/Bibliography of important works
cantly between hydraulically transmissive and con- in Environmental Microbiology.htm.
fining formations in the same wells and boreholes. Sylvia, D.M., Fuhrmann, J.J., Hartel, P.G., and Zuberer, D.A.
3. Community biochemical adaptability: Bacteria that (2005). Principles and Applications of Soil Microbiology, 2nd
Edn. Prentice-Hall, New York.
form biofilms and readily exchange genetic poten-
tial can be quite plastic in their ability to exploit
resources. Early in the effort to employ bioreme-
diation in treating contaminated groundwater and CHARACTERIZING SOIL SPATIAL VARIABILITY
soil, it was recognized that natural communities
DENNIS L. CORWIN
were capable of degrading a variety of anthropogenic
compounds, especially hydrocarbons. SCOTT M. LESCH
USDA-ARS George E. Brown,
4. Iron and sulfur biogeochemical cycles are particu- Jr., Salinity Laboratory
larly important in deeper systems, whereas nitrogen Riverside, California
is strongly cycled in the soil and shallow saturated
flow zones. In deeper systems, carbon and nitrogen
are cycled locally within biofilms. SPATIAL VARIABILITY: DEFINITION, ORIGIN,
5. Groundwater contamination locally alters oxidation- SIGNIFICANCE, AND MEASUREMENT METHODS
reduction systems, nutrient cycling, and community
Spatial variability refers to the variation in soil physical,
structure. Microbial communities react to influxes
chemical, mineralogical, and biological properties both
of organic carbon by increasing biomass and tipping
laterally across the landscape and vertically through
toward anaerobic respiration as oxygen is depleted
the soil profile. The spatial heterogeneity of soils is a
by hydrocarbon-using or other aerobic heterotrophs,
consequence of interacting meteorological, pedological,
which is closely analogous to eutrophication in
and biological processes and anthropogenic factors over
surface water systems. For energetic reasons, the
time. Within-field spatial variability of soil properties
following are favored sequentially, switching to the
has been a primary concern within soil science ever
next when the electron acceptor is depleted: nitrate,
since the classic paper by Nielson et al. (1) concerning
iron (and other metal) oxides, sulfate, and carbon
the variability of field-measured soil water properties.
dioxide (reduced to methane).
The significance of soil spatial variability stems from its
6. Microbial communities may include very long-lived complex influence on any landscape-scale soil-related issue
and ancient lineages. Studies of communities in including solute transport in the vadose zone, precision
deep, isolated systems reveal communities that agriculture, and soil quality assessment, to mention a few.
may be very ancient and existing in isolation from A variety of measurement methods are available for
the surface. potentially characterizing soil spatial variability includ-
ing ground penetrating radar (GPR), aerial photography,
An article of this length is insufficient to consider multispectral and hyperspectral imagery, time-domain
the subject comprehensively (considering that entire reflectometry (TDR), and apparent soil electrical conduc-
textbooks, conferences, and hundreds of webpages are tivity (ECa ). However, none of these approaches has been
devoted to the topic). Related subjects are considered as extensively investigated with respect to characteriza-
elsewhere in the Encyclopedia of Water. tion of within-field soil variability as ECa (2).
466 CHARACTERIZING SOIL SPATIAL VARIABILITY

Since its early agricultural use for measuring soil to representatively characterize the spatial distribution
salinity, the application of ECa has evolved into a widely of the existing soil conditions that influence the soil’s
accepted means of establishing the spatial variability intended use. Therefore, assessing soil quality requires
of several soil physicochemical properties that influence quantitative knowledge of each indicator property asso-
the ECa measurement. Geospatial measurements of ECa ciated with a soil’s quality and the spatial variability of
are well suited for characterizing spatial variability for those indicator properties.
several reasons. Geospatial measurements of ECa are Furthermore, spatial variability has a profound influ-
reliable, quick, and easy to take. The mobilization of ECa ence on environmental impacts, particularly of non-point
measurement equipment is easy and can be accomplished source pollutants. This result is because of its influence
at a reasonable cost. Finally, and most importantly, on the fate and transport of agrochemical inputs (e.g.,
ECa is influenced by a variety of soil properties for pesticides, fertilizers, etc.) in an agroecosystem. In fact,
which the spatial variability of each could be potentially it has become clear that the real constraint on modeling
established. solute transport is not the detail of the model structures,
but defining the characteristics of individual places (7).
Jury (3) provides an excellent fundamental discussion of
NEED FOR CHARACTERIZING SOIL SPATIAL VARIABILITY
the spatial variability of soil properties and its impact
on solute transport in the vadose zone. As Jury points
The prospect of feeding a projected additional 3 billion
out, ‘‘any hope of estimating a continuous spatial pattern
people over the next 30 years poses formidable, but not
of chemical emissions at each point in space within a
insurmountable, challenges. Feeding the ever-increasing
field must be abandoned due to field-scale variability of
world population will require a sustainable agricultural
soils.’’
system that can keep pace with population growth.
The characterization of spatial variability is without
Sustainable agriculture hinges on a delicate balance
a doubt one of the most significant areas of concern
of economic sustainability, crop productivity, resource
in soil science because of its broad reaching influence
utilization, and minimized detrimental environmental
on all field- and landscape-scale issues. Research in
impacts. These factors are influenced either directly or
the characterization of spatial variability has advanced
indirectly by soil spatial variability.
tremendously over the past decade in large part because
Spatial variability within a field amounts to an order
of georeferenced ECa measurements.
of magnitude or more for soils (3,4) and to a factor of
2–4 or more for crops (5). Spatial variation in crops is the
result of a complex interaction of biological (e.g., pests, EDAPHIC FACTORS INFLUENCING ECa MEASUREMENTS
earthworms, microbes), edaphic (e.g., salinity, organic
matter, nutrients, texture), anthropogenic (e.g., irrigation The earliest field applications of geophysical measure-
efficiency, soil compaction due to farm equipment), ments of ECa in soil science, which occurred in the early
topographic (e.g., slope, elevation), and climatic (e.g., 1980s, involved the determination of salinity within the
relative humidity, temperature, rainfall) factors. To a soil profile of arid zone soils. However, it soon became
varying extent from one field to the next, crop patterns apparent that the measurement of ECa in the field as a
are influenced by soil-related properties. Conventional quantification of soil salinity was more complicated than
farming is not as efficient as it could be because it initially anticipated because of the complexity of current
currently treats a field uniformly. To maintain uniform flow pathways originating from the spatial heterogeneity
crop production, conventional farming overcomes the of properties influencing current flow in soil.
inherent variability of soil and crop conditions between Three pathways of current flow contribute to the ECa
and within fields with an over application of inputs (i.e., of soil: (1) a liquid phase pathway via dissolved solids
irrigation water, pesticides, fertilizers, etc.) intended to contained in the soil water occupying the large pores, (2) a
meet the needs of areas of lowest productivity. Bullock solid–liquid phase pathway primarily via exchangeable
and Bullock (6) pointed out the imminent need for efficient cations associated with clay minerals, and (3) a solid
methods to accurately measure within-field variation in pathway via soil particles that are in direct and continuous
soil physical and chemical properties as a key component contact with one another (8). Of these three pathways,
for precision agriculture. the solid pathway is usually negligible; hence, most
The effective utilization of soil as a resource depends soil can be approximated with a dual-pathway parallel
on assessing and maintaining its quality. A fundamental conductance equation.
component of assessing field-scale soil quality is establish- An analysis of this dual-pathway system indicates that
ing the spatial distribution of the soil properties affecting ECa will be influenced by a complex interaction of soil
the soil’s intended management goal (e.g., maximize agri- properties including salinity, soil texture, water content,
cultural productivity, minimize environmental impact, ρb , and temperature. The ρb is directly influenced by clay
and/or maximize waste recycling) and its intended func- content (soil texture) and organic matter (OM). Further-
tion (e.g., biodiversity, filtering and buffering, nutrient more, the exchange surfaces on clays and OM provide a
cycling, physical stability and structural support, resis- solid–liquid phase pathway primarily via exchangeable
tance and resilience, and water and solute flow). It is not cations; consequently, the cation exchange capacity (CEC)
sufficient to take a single measurement within a field to and OM are recognized as additional factors that can influ-
characterize its soil quality. Rather, a sufficient number ence ECa measurements. Measurements of ECa must be
of measurements must be taken and at specific locations interpreted with these influencing factors in mind.
 CHARACTERIZING SOIL SPATIAL VARIABILITY 467

MOBILE ECa MEASUREMENT EQUIPMENT spatial variability with ECa -directed soil sampling is
based on the hypothesis that when ECa correlates with
An ECa survey to characterize spatial variability uses a soil property or properties, then spatial ECa information
mobile geophysical equipment coupled with a global can identify sites that reflect the range and variability
positioning system (GPS) to georeference each ECa of the property or properties. In instances where ECa
measurement. Mobile ECa equipment has been developed correlates well with a particular soil property, an ECa -
by a variety of researchers (9–11). The development of directed soil sampling approach can optimally establish
mobile ECa measurement equipment has made it possible the spatial distribution of that property while minimizing
to produce ECa maps with measurements taken every the corresponding sampling and labor costs (2,15,16). Also,
few meters. if ECa is correlated with crop yield, then an ECa -directed
Mobile ECa measurement equipment has been devel- soil sampling approach can identify those soil properties
oped for both electrical resistivity (ER) and electromag- that are causing the variability in crop yield (13). Details
netic induction (EMI) geophysical approaches. In the case for conducting a field-scale ECa survey for the purpose
of ER, four stainless-steel electrodes are inserted into the of characterizing the spatial variability of soil properties
soil generally at equal distances and connected to a resis- influencing soil quality or crop yield variation can be found
tivity meter. Current is applied to the two outer electrodes in Corwin and Lesch (16).
with the two inner electrodes serving as the potential Figure 1 shows EMh and EMv maps of an ECa survey
electrodes. Fixing the electrode array and mounting it performed on a 6.5-ha field in Maricopa, AZ, along with the
on or pulling it from a vehicle saves considerable time associated soil sample sites selected with a specific type
for measurements. By coupling the fixed electrode array of ECa -directed sampling methodology known as spatial
to a datalogger and GPS georeferences the ECa mea- response surface sampling design (15,16). Once soil cores
surement, geospatial ECa measurements can be taken. are taken at the selected sites and the corresponding soil
Veris Technologies1 (Salinas, KS; www.veristech.com) has physicochemical properties are analyzed, the properties
developed a commercial mobile system for measuring ECa that are well correlated with ECa can be identified.
using the principles of ER. In a similar fashion, an EMI Statistical prediction models can then be derived that
unit can be mobilized by pulling a sled-mounted EM-38 spatially map (predict) these correlated soil properties
unit (Geonics Ltd., Mississaugua, Ontario, Canada) behind with the calibrated ECa survey data. Figure 2 shows
a vehicle and logging the EMI data simultaneously along the corresponding ECa -predicted 0–0.9-m percent water
with real-time GPS position information. Measurements content patterns for the field shown in Fig. 1.
of ECa with EMI generally involve taking readings in
the horizontal coil configuration (EMh ), which is a shallow GUIDELINES FOR CONDUCTING AN ECA -DIRECTED SOIL
reading penetrating to a depth of about 0.75–1.0 m, and/or SAMPLING SURVEY TO CHARACTERIZE SPATIAL
in the vertical coil configuration (EMv ), which is a deep VARIABILITY
reading penetrating to a depth of about 1.5 m.
The basic elements of a field-scale ECa survey for
characterizing spatial variability include (1) ECa survey
ECa -DIRECTED SAMPLING DESIGNS design, (2) georeferenced ECa data collection, (3) soil sam-
ple design based on georeferenced ECa data, (4) soil
Despite the complexities of the ECa measurement, the sample collection, (5) physicochemical analysis of per-
fact that ECa is influenced by several soil properties tinent soil properties, (6) spatial statistical analysis,
has broadened its utility and enabled it to play a (7) determination of the dominant soil properties influenc-
major role in addressing the issue of characterizing ing the ECa measurements at the study site, and (8) GIS
spatial variability. Geospatial measurements of ECa development. The outlined steps of an ECa -directed soil
have been successfully used for (1) identifying the soil sampling survey are provided in Table 1.
physicochemical properties influencing crop yield patterns
and soil condition, (2) establishing the spatial variation LIMITATIONS AND PRECAUTIONS
of these soil properties, and (3) characterizing the spatial
distribution of soil properties influencing solute transport At present, no other soil measurement provides a greater
through the vadose zone (4,12–14). level of spatial information than that of geospatial ECa
In a typical field survey, the spatial distribution of measurements. Even so, are limitations exist to the use
various soil properties can often be adequately determined of ECa to characterize soil spatial variability. Awareness
with conductivity survey information in conjunction with of these weaknesses is crucial for the proper use and
some type of soil sampling plan. Either probability interpretation of geospatial ECa measurements. The
(design) or prediction (model) based sampling strategies complex spatial heterogeneity of the soil system has
can identify the soil properties that correlate best with ECa subtle influences on geospatial ECa measurements that
in a particular field. Prediction-based sampling strategies can have significant interpretive impacts. The ability
objectively select sampling locations based on the observed to recognize and interpret these influences can be the
spatial ECa pattern and thus are often referred to difference between the successful or failed application of
as ‘‘sensor-directed’’ sampling plans (15). Characterizing ECa measurements for characterizing spatial variability.
First and foremost, in the absence of soil sample
1
Product identification is provided solely for the benefit of the information, geospatial measurements of ECa by them-
reader and does not imply the endorsement of the USDA. selves do not directly quantify soil spatial variability.
 (a) 3661004

EMh
mS/m
< 35 3660904
35 - 40
40 - 45
> 45
3660805

Coord system
UTM (m) 3660705
X: Easting
Y: Northing

3660606
408638 408681 408725 408768 408812

(b) 3661004

EMv
mS/m
< 80 3660904
80 - 85
85 - 90
> 90 3660805

Coord system
UTM (m) 3660705
X: Easting
Figure 1. Maps of ECa survey data and ECa -directed soil
Y: Northing
sampling sites for a field in Maricopa, AZ: (a) EM38 hor-
izontal and (b) EM38 vertical. White square symbol rep-
resents the soil sample site locations. UTM = universal 3660606
transverse mercator grid system in meters. 408638 408681 408725 408768 408812

GWC (0-90 cm) 3661004

%

3660904
< 12
12 - 14
14 - 16
> 16 3660805

Coord system
UTM (m)
Figure 2. Map of spatial variability of percent gravi- 3660705
metric water content (GWC) for the 0–0.9-m sam- X: Easting
pling depth, as determined from the survey data Y: Northing
and ECa -directed soil sampling sites shown in
Fig. 1 (R2 = 0.85 for predicted and measured GWC).
UTM = universal transverse mercator grid system 3660606
in meters. 408638 408681 408725 408768 408812

468
 CHARACTERIZING SOIL SPATIAL VARIABILITY 469

Table 1. Outline of Steps to Conduct an ECa Field Survey

1. Site description and ECa survey design

a. record site metadata
b. define the project’s/survey’s objective
c. establish site boundaries
d. select GPS coordinate system
e. establish ECa measurement intensity
2. ECa data collection with mobile GPS-based equipment
a. georeference site boundaries and significant physical geographic features with GPS
b. measure georeferenced ECa data at the predetermined spatial intensity and record associated metadata
3. Soil sample design based on georeferenced ECa data
a. statistically analyze ECa data using an appropriate statistical sampling design to establish the soil sample site
locations
b. establish site locations, depth of sampling, sample depth increments, and number of cores per site
4. Soil core sampling at specified sites designated by the sample design
a. obtain measurements of soil temperature through the profile at selected sites
b. at randomly selected locations, obtain duplicate soil cores within a 1-m distance of one another to establish
local-scale variation of soil properties
c. record soil core observations (e.g., mottling, horizonation, textural discontinuities, etc.)
5. Laboratory analysis of appropriate soil physicochemical properties defined by project objectives
6. If needed, stochastic and/or deterministic calibration of ECa to ECe or to other soil properties (e.g., water content
and texture)
7. Spatial statistical analysis to determine the soil properties influencing ECa and/or crop yield
a. soil quality assessment:
1. perform a basic statistical analysis of physicochemical data by depth increment and by composite depth over
the depth of measurement of ECa
2. determine the correlation between ECa and physicochemical soil properties by composite depth over the depth
of measurement of ECa
b. precision agriculture applications (if ECa correlates with crop yield, then):
1. perform a basic statistical analysis of physicochemical data by depth increment and by composite depths
2. determine the correlation between ECa and physicochemical soil properties by depth increment and by
composite depths
3. determine the correlation between crop yield and physicochemical soil properties by depth and by composite
depths to determine depth of concern (i.e., depth with consistently highest correlation, whether positive or
negative, of soil properties to yield) and the significant soil properties influencing crop yield (or crop quality)
4. conduct an exploratory graphical analysis to determine the relationship between the significant
physicochemical properties and crop yield (or crop quality)
5. formulate a spatial linear regression (SLR) model that relates soil properties (independent variables) to crop
yield or crop quality (dependent variable)
6. adjust this model for spatial autocorrelation, if necessary, using restricted maximum likelihood or some other
technique
7. Conduct a sensitivity analysis to establish dominant soil property influencing yield or quality
8. GIS database development and graphic display of spatial distribution of soil properties

Source: Ref. 16. With permission.

Rather, ECa -survey measurements provide the spatial of those soil properties influencing ECa in the landscape
information necessary to direct soil sampling. It is the of interest.
combined use of both the ECa -survey and the corre- Second, ECa -directed soil sampling can only spatially
sponding sample data that makes this sort of surveying characterize soil properties that correlate with ECa . This
technology invaluable for characterizing spatial variabil- correlation may be caused by a direct or indirect influence
ity. In other words, the primary strength of geospatial on the ECa measurement, or the correlation may be
ECa measurements lies in their effectiveness as a means a complete artifact. For example, salinity and water
to direct soil sampling (using a minimum number of sam- content will directly influence ECa and CEC will indirectly
ple sites) that best characterizes the spatial variability influence ECa through its influence on current flow at the
470 CHARACTERIZING SOIL SPATIAL VARIABILITY

surface of soil minerals. Likewise, in many instances, B grid sampling, which makes grid sampling impractical
and salinity distributions are similar. Consequently, a because of labor and cost intensiveness. Soil sampling
strong correlation between B and salinity (and thus ECa ) directed by geospatial ECa measurements provides a
can result, even though no cause-and-effect relationship viable alternative for characterizing spatial variability
exists between B and ECa . Thus, a solid understanding of a variety of soil-related physicochemical properties.
of the soil properties that directly influence and/or are Geospatial ECa measurements provide a means of
indirectly measured by ECa at a specific site is particularly significantly reducing the number of soil samples needed to
essential in temporal surveying applications, because over characterize spatial variability, provided that the target
time these correlations may or may not persist. soil properties are well correlated with the conductivity
Third, as previously mentioned, often a dynamic survey data.
and complex relationship exists between ECa and those Directed soil sampling with geospatial ECa measure-
properties that influence ECa . Ground-truth soil samples ments has its limitations in characterizing spatial vari-
are obligatory if spatial measurements of ECa are ability, particularly when the soil properties of interest
to be used for quantifying soil variability. Geospatial are not measured either directly or indirectly by the ECa
measurements of ECa do not supplant the need for soil survey data. In these instances, other types of sensors
sampling, but they do minimize the number of samples should be used (if available) or a more intensive soil sam-
necessary to characterize spatial variability. Additionally, pling strategy must be adopted. A robust methodology for
users of ECa survey data must exercise caution and the integrated use of multiple remote and ground-based
be aware of what ECa is actually measuring at the sensors has yet to be developed. The successful integra-
site of interest. Normally, the only way to establish tion of multispectral and hyperspectral imagery, TDR,
those soil properties that influence ECa at a specific GPR, aerial photography, and ECa sensors is needed to
location is by collecting representative soil samples and provide the redundant and supplemental data necessary
then determining the relationship between ECa and the to unravel the spatial complexity of soil. This type of inte-
particular property or properties of interest. This process grated methodology represents the future direction of soil
requires that every ECa survey has an associated soil spatial variability research.
sampling design. The location and number of sites can be
established (after acquiring the ECa survey data) using BIBLIOGRAPHY
model-based sample design software such as ESAP (17).
Finally, the temporal stability of ECa measurements 1. Nielsen, D.R., Biggar, J.W., and Erh, K.T. (1973). Spatial
at a site may be of potential concern because ECa is a variability of field-measured soil-water properties. Hilgardia
product of both static and dynamic factors, which adds 42(7): 215–259.
another dimension to the complexity of understanding 2. Corwin, D.L. and Lesch, S.M. (In press). Apparent soil
and interpreting geospatial ECa measurements. For this electrical conductivity measurements in agriculture. Comput.
reason, caution must be taken to characterize spatial Electron. Agric.
variability with ECa when dynamic conditions influencing 3. Jury, W.A. (1986). Spatial variability of soil properties. In:
ECa are significant. Apparent soil electrical conductivity Vadose Zone Modeling of Organic Pollutants. S.C. Hern
and S.M. Melancon (Eds.). Lewis Publishers, Chelsea, MI,
surveys are generally conducted (1) within a set time
pp. 245–269.
frame to minimize the effects of dynamic properties (such
4. Corwin, D.L. et al. (2003). Assessment and field-scale map-
as temperature and water content), (2) when the soil is
ping of soil quality properties of a saline-sodic soil. Geoderma
at or near field capacity (to avoid current-flow problems), 114(3–4): 231–259.
and (3) with regard for subtle topographic effects (e.g.,
5. Verhagen, A., Booltink, H.W.G., and Bouma, J. (1995). Site-
bed-furrow). Protocols for conducting an ECa survey that specific management: Balancing production and environmen-
consider all previously discussed limitations are presented tal requirements at farm level. Agric. Syst. 49: 369–384.
by Corwin and Lesch (16). 6. Bullock, D.S. and Bullock, D.G. (2000). Economic optimality
of input application rates in precision farming. Prec. Agric. 2:
FUTURE DIRECTION 71–101.
7. Beven, K.J. (2002). Towards an alternative blueprint for
Because of the heterogeneous nature of soils, characteriza- a physically-based digitally simulated hydrologic response
tion of soil spatial variability is a fundamental component modeling system. Hydrol. Process. 16(2): 189–206.
of any landscape-scale process that cannot be overlooked or 8. Rhoades, J.D., Chanduvi, F., and Lesch, S. (1999). Soil
superficially addressed. Significant technological advances Salinity Assessment: Methods and Interpretation of Electrical
Conductivity Measurements, FAO Irrigation and Drainage
have occurred over the past two decades, particularly in
Paper #57, Food and Agriculture Organization of the United
the area of sensor technology and in precisely locating Nations, Rome, Italy, pp. 1–150.
a geographic position with GPS. These advances have
9. McNeill, J.D. (1992). Rapid, accurate mapping of soil salinity
helped the progress of research in characterizing spatial by electromagnetic ground conductivity meters. In: Advances
variability at field and landscape scales. in Measurements of Soil Physical Properties: Bringing Theory
Geospatial measurement of ECa stands out among into Practice. G.C. Topp, W.D. Reynolds, and R.E. Green
the ground-based sensor technologies, contributing to (Eds.). SSSA Special Publication No. 30, ASA-CSSA-SSSA,
an improved ability to characterize spatial variability. Madison, WI, pp. 201–229.
Numerous soil samples are required for representative 10. Rhoades, J.D. (1993). Electrical conductivity methods for
estimates of field-scale spatial variability using traditional measuring and mapping soil salinity. In: Advances in
 DEEP SOIL-WATER MOVEMENT 471

Agronomy. D.L. Sparks (Ed.). Vol. 49, Academic Press, San of soil-water pressures are usually negative—water does
Diego, CA, pp. 201–251. not freely drain into a well. Some soil water may exist
11. Jaynes, D.B., Colvin, T.S., and Ambuel, J. (1993). Soil type with positive pressures, such as water that is perched in a
and crop yield determinations from ground conductivity saturated zone above a layer of lower permeability. Water
surveys. ASAE Paper No. 933552, 1993 ASAE Winter under positive pressure flows freely into a well.
Meetings, 14–17 Dec. 1993, Chicago, IL, ASAE, St. Joseph,
MI.
12. Corwin, D.L. et al. (1999). Evaluation of GIS-linked model of FORCES THAT CONTROL WATER MOVEMENT
salt loading to groundwater. J. Environ. Qual. 28: 471–480.
13. Corwin, D.L. et al. (2003). Identifying soil properties that Gravity is the dominant force that controls soil-water
influence cotton yield using soil sampling directed by movement. The downward force of gravity acts in
apparent soil electrical conductivity. Agron. J. 95(2): conjunction with geologic media to control the magnitude
352–364. and direction of fluid flow.
14. Corwin, D.L., Lesch, S.M., Oster, J.D., and Kaffka, S.R. (In A second important force is the matric tension (or soil-
press). Characterizing spatio-temporal variability with soil water tension), which causes water to move from zones of
sampling directed by apparent soil electrical conductivity. lower tension to zones of higher tension. The soil tension,
Geoderma. ψ, is the negative of the water pressure head, ψ = −P/γ ,
15. Lesch, S.M. (In press). Sensor-directed response surface where P is the fluid pressure relative to atmospheric
sampling designs for characterizing spatial variation in soil pressure and γ is the freshwater specific weight.
properties. Comput. Electron. Agric. The total head, H, combines the gravitational and
16. Corwin, D.L. and Lesch, S.M. (In press). Characterizing soil tension forces using H = z − ψ, where z is the elevation of
spatial variability with apparent soil electrical conductivity: the point of measurement.
I. Survey protocols. Comput. Electron. Agric.
In some circumstances, osmotic forces can also affect
17. Lesch, S.M., Rhoades, J.D., and Corwin, D.L. (2000). ESAP- water movement. Osmotic forces develop when solute
95 Version 2.10R: User Manual and Tutorial Guide,
concentration gradients are present, causing water to
Research Rpt. 146, USDA-ARS George E. Brown, Jr. Salinity
Laboratory, Riverside, CA.
move from zones of low salt concentration to zones of
higher concentration. When osmotic forces are present,
the total head must include the osmotic potential, φ, as
noted in Section GW456 (head).
DEEP SOIL-WATER MOVEMENT
TODD RASMUSSEN MATERIAL PROPERTIES
The University of Georgia
Athens, Georgia In addition to the total head, the characteristics of the
geologic medium are also important for controlling the
magnitude and direction of deep soil-water movement (1).
Soil water refers to subsurface water between the The important material properties that affect soil-water
Earth’s surface and the regional water table, called the movement are the unsaturated hydraulic conductivity
unsaturated, or vadose zone. Figure 1 illustrates the functions and the soil-water characteristic curves.
components of flow within this zone. Deep soil-water
movement usually implies the migration of water below Unsaturated Hydraulic Conductivity
the root zone, and is also called deep percolation.
Soil water is held by adhesive forces to soil surfaces and The unsaturated hydraulic conductivity, K(ψ), is com-
by cohesive forces to other water molecules. Measurements monly written as:

K(ψ) = Kr (ψ) Ks (1)

Precipitation
Evapotranspiration where Kr is the relative unsaturated hydraulic conductiv-
ity function that varies with matric tension, ψ, and Ks is
the saturated hydraulic conductivity.
As soil dries, the tension increases and the relative
Percolation hydraulic conductivity decreases, whereas wetting soils
Overland flow show a decrease in tension and an increase in relative
Infiltration conductivity. The relative conductivity ranges from a
Clay theoretical minimum of zero in dry soil, Kr → 0 as ψ → 0,
to a maximum of one in saturated soil, Kr → 1 as ψ = 0,
so that 0 < Kr ≤ 1.
Recharge Perched water
Both the relative and saturated hydraulic conductiv-
ities vary with the pore-size distribution of the geologic
Watertable media. Sandy materials may have a large saturated
Saturated zone
hydraulic conductivity, but a rapidly declining relative
hydraulic conductivity, whereas finer materials may have
Figure 1. Flow components in the unsaturated zone. a smaller saturated hydraulic conductivity but a slowly
472 DEEP SOIL-WATER MOVEMENT

Coarse-grained Coarse
Hydraulic conductivity, K

Water content, q
Fine-grained Fine

Wet Dry Wet Dry

Tension, y Tension, y
Figure 2. Unsaturated hydraulic conductivity, K, as a function Figure 3. Soil water content, θ, as a function of tension, ψ, for
of tension, ψ, for two materials. The coarser material is more two materials. The coarser material holds more water under wet
conductive under wet conditions (smaller values of tension), conditions (smaller values of tension), whereas the finer material
whereas the finer material is more conductive under drier holds more water under drier conditions.
conditions.

soils mediate the pulse of water. This water is retained

declining relative conductivity. This effect is shown if by the soil and then slowly released over time. The initial
Fig. 2. pulse of water at the surface is transformed into a slow
The crossover of the two curves shown in Fig. 2 results drainage at depth.
in the interesting phenomenon that coarser materials have Only in materials with little water storage capacity
a larger unsaturated hydraulic conductivity than finer (i.e., in wet soils or in materials with limited water holding
material in the wet range, but a smaller conductivity in capacity) will the downward movement of water be rapid.
the drier range. In these cases, the water that reaches the water table
Thus, the flow behavior for wet conditions may be may or may not be the same water that was added at
very different from the behavior under dry conditions the surface. In some situations, the added water merely
in heterogeneous materials. Coarser materials (such as displaces the water below it, causing an equivalent volume
sands, fractures, and other large openings) generally of water to move at depth. In other situations, the added
dominate flow under wet conditions, whereas finer water moves quickly through the unsaturated zone.
materials (such as silts, clays, and consolidated materials)
can dominate flow under drier conditions. Spatial Variability
Moisture Characteristic Curves Local subsurface conditions can also affect deep soil-water
movement. Hydrogeologic units that dip in one direction,
The ability of the unsaturated zone to hold water as
such as bedded sedimentary materials, can cause water
a function of the soil tension is called the moisture
to flow in the down-dip direction. Microscopic layering in
characteristic curve. The characteristic curve for each
features such as sand dunes can also affect the direction
material varies with the pore size distribution—a soil
of movement.
with larger pores releases water at a lower tension than a
Vertical features, such as fractures and joints, may
soil with smaller pores.
focus water movement during wet conditions. Localized
Figure 3 shows two soil characteristic curves, one for a
zones of preferential flow can also be found in soils with a
coarser material and a second for a finer material. Like
defined structure, such as in soil with aggregates. These
the unsaturated hydraulic conductivity function, the finer
soils form a dual-porosity network consisting of fine pores
soil can hold more water under dry conditions, whereas
within the aggregates and larger macropores between the
the coarser soil holds more water under wetter conditions.
aggregates.
The water retention properties of soils also affect water
Many explanations exist for preferential flow within
movement. Changes in water flux with depth can be
soils, including bypass flow, macropore flow, fracture flow,
attributed to the uptake and release of water as the soil
boundary layer flow, mobile zone flow, finger flow, funnel
tension changes. For example, a wetting front may move
flow, media heterogeneities, ion exclusion, and colloid
downward with time from the surface after a large rain
transport (2).
event, only to be absorbed by the dry soil underneath.
The ability to absorb the moisture is a function of the
ambient moisture tension—an already wet soil will have WATER VAPOR MOVEMENT
little ability to absorb additional water.
In some cases, water percolating down from above may Water vapor movement in deep systems can also be
not immediately reach the water table because the dry important, especially in very dry media where liquid flow
 SPECIFIC GRAVITY 473

is small. Water vapor is generally in equilibrium with the density, i.e., 1 kg/L. Thus, variations in the specific gravity
liquid phase and can be related to the matric tension of are a direct measure of the variation in density. The
the pore fluids (3): specific weight of a material equals the product of its

 density and the gravitational constant. The specific gravity
e gψ is important because it is needed to correct for the
RH = = exp (2)
es RT water density in wells to determine the freshwater head.
Well construction also relies on the specific gravity of
where RH is the relative humidity, e is the actual drilling muds to prevent the collapse of borehole walls in
vapor pressure, es is the saturated vapor pressure, R unconsolidated media.
is the water vapor gas constant, and T is the absolute The specific gravity is a measure of the density, or
temperature. Dissolved solids in soil water can lower the weight, of a fluid or solid. Denser materials have a larger
relative humidity—and increase the tension—because of specific gravity than do less dense materials. The specific
osmotic forces. gravity, , is defined using
The diffusion of water vapor through the subsurface
can be predicted using a variety of approaches that ρ
= (1)
depend on the specific conditions within the subsurface, ρo
which is because additional forces become important
where ρ is the density (mass per unit volume) of the
when considering subsurface vapor movement, including
material and ρo is the density of a reference material.
thermal effects and barometric pressure changes.
The density of water at 4 ◦ C, ρo = 1 kg/L, is generally
Temperature variation affects water movement because
the reference density for liquids and solids. Table 1
the saturated vapor pressure increases with temperature.
provides the specific gravity for a range of elements,
The natural temperature increase with depth is called the
including very light metals, such as aluminum and
geothermal gradient. Deeper water is warmer and has
titanium, to very dense metals, such as gold and platinum.
a higher vapor pressure than water at a shallow depth.
Note that for liquids and solids, the specific gravity equals
This vapor pressure gradient causes water vapor to move
the density because the standard density of water is 1 kg/L.
upward toward the cooler zones, where it may condense.
The density of hydrogen is common for gasses.
Also, barometric fluctuations cause oscillations in the
The specific gravity of a pure substance can be affected
vertical movement of soil gasses, leading to the ver-
by temperature. Table 2 provides the specific gravity and
tical advective transport of water vapor. The vertical
specific weight for water from 0 ◦ C to 100 ◦ C. Note that the
movement may be amplified when extreme barometric
specific gravity decreases more rapidly as the temperature
pressure conditions are present, such as during a hur-
increases.
ricane when extremely low surface barometric pressures
Changes in physical properties of solids, such as their
cause an upward air pressure gradient within the unsatu-
crystalline structure or porosity, can also result in a change
rated zone.
in their density. The specific gravity can also be affected
by any components present within the material, including
BIBLIOGRAPHY dissolved gasses, liquids, and solids, as well as suspended
gasses, liquids, and solids.
1. Evans, D., Rasmussen, T.C., and Nicholson, T.J. (2001). Flow For example, water with high concentrations of
and transport through unsaturated fractured rock: An suspended solids (such as drilling fluids) or dissolved solids
overview. In: Flow and Transport Through Unsaturated (such as brines) has higher specific gravities than does
Fractured Rock, 2nd Edn. D.D. Evans, T.J. Nicholson, and clear water. Also, a column of air bubbles rising within a
T.C. Rasmussen (Eds.). Geophysical Monograph 42. American
water column can reduce the fluid density.
Geophysical Union, Washington, DC.
The specific gravity is also related to the specific weight,
2. Rasmussen, T.C., Baldwin, R.H., Dowd, J.F., and Williams,
γ = ρg, where g = 9.807 m/s2 is the gravitational constant.
A.G. (2000). Tracer vs. pressure wave velocities through
unsaturated saprolite. Soil Sci. Soc. Amer. J. 64(1): 75–85.
The specific weight commonly converts measured pres-
sures to an equivalent hydraulic head because an increase
3. Hillel, D. (1971). Soil and Water: Physical Principles and
Processes. Academic Press, New York.
Table 1. Specific Gravity
of Selected Elements
SPECIFIC GRAVITY Material Specific Gravity
TODD RASMUSSEN Aluminum 2.70
The University of Georgia Titanium 4.51
Athens, Georgia Iron 7.87
Copper 8.96
Silver 10.50
The density of water and other materials varies with Lead 11.30
temperature as well as with the amount of dissolved and Mercury 13.50
Uranium 19.10
suspended matter within them. The specific gravity is
Gold 19.30
the observed density of a material divided by a reference Platinum 21.50
density, which is usually taken to be water at its maximum
474 SPECIFIC GRAVITY

Table 2. Variation of Pure Water with found using

Temperature f − 1
C= (2)
Temperature Density Specific Weight s − 1
( ◦ C) (kg/L) (Pa/m)
where f is the measured specific gravity of the fluid and
0 0.99984 9,805.4 s is the specific gravity of the suspended sediments.
10 0.99970 9,804.1
20 0.99821 9,789.4
IMPORTANCE OF SPECIFIC GRAVITY
30 0.99565 9,764.3
40 0.99222 9,730.7
Head Measurement
50 0.98803 9,689.6
60 0.98320 9,642.2 The use of water levels in wells to determine the hydraulic
70 0.97778 9,589.1 head within an aquifer is complicated when the density
80 0.97182 9,530.6 of the fluid within the borehole is not a constant (2). As
90 0.96535 9,467.2 noted in HEAD, changes in density affect the head according
100 0.95840 9,399.0 to the relationship:

H = zo + z (3)
in depth, z, is accompanied by a concomitant increase in where H is the freshwater head, zo is the midpoint
pressure, P, which yields γ = P/z. elevation of the screened zone within the aquifer,  = γ /γo
is the specific gravity of the fluid within the well,
MEASURING THE SPECIFIC GRAVITY z = z − zo is the height of the column of water within
the monitoring well above the screened zone, and z is the
water surface elevation in the well.
A traditional technique for measuring the fluid density
uses a hydrometer, which compares the weight of the Drilling Fluids
fluid to a standard fluid. The hydrometer is allowed to
float within the fluid, and the fluid density is determined The specific gravity of the drilling fluid is important during
from how much higher or lower the hydrometer level is well construction. In loose materials, the process of drilling
compared with pure water (Fig. 1). Heavier fluids cause a well is complicated by the tendency of the unsupported
the hydrometer to rise relative to clear water because of walls to collapse.
buoyancy. Borehole walls can be temporarily supported until
An alternative method for determining the fluid density a casing is emplaced by using drilling fluids that are
is to employ a differential pressure transducer to measure denser than water. The weight of the drilling fluid places
the change in fluid pressure over a known depth. The an outward force on the borehole walls that helps to
resulting ratio between the measured pressure and the support them.
depth yields an estimate of the specific weight, γ = P/z. Heavier drilling fluids place a correspondingly heavier
Pressure transducers have recently measured the force on the borehole wall. One important drawback is the
concentration and particle size distribution of suspended tendency of the drilling fluid to accumulate on the walls
sediments (1). The particle size distribution is determined and within the formation, which causes a reduction in well
using Stokes’s law, whereas the sediment concentration is yield if it is not removed during well development.
Lighter drilling fluids, such as foams and air, can also be
used. In this case, the inflow of water from the formation
into the borehole prevents the buildup of materials on
Hydrometer
the borehole wall, which eases well development. This
technique is suitable in consolidated media where borehole
collapse is not a problem.

Table 3. Specific Gravities of

Selected Drilling Fluids
Drilling Fluid Specific Gravity

Air 0.001
Mist 0.020
Stiff foam 0.100
Wet foam 0.200
Polymers 1.020
Clear water Water with mud Polymers w/ NaCl 1.200
Bentonite clay 1.300
Figure 1. Hydrometer technique for measuring fluid density. Polymers w/ CaCl 1.400
Clear water causes the hydrometer to float at a constant level. Bentonite w/ barite 2.200
Denser fluids cause the hydrometer to rise, whereas lighter fluids Barite (BaSO4 ) 4.200
cause the hydrometer to sink lower in the water column.
 HOT SPRINGS 475

Table 3 presents approximate specific gravities for a they had to rely on springs. Many early towns in Greece,
range of drilling fluids. Notice the wide range in specific Egypt, Mesopotamia, India, and China developed around
gravities, from a low of 0.01 for air to a high of 4.2 for springs. Many towns in the United States have taken their
barite. names from springs: Steamboat Springs, Colorado; Glen-
The values in Table 3 are approximate because the final wood Springs, Colorado; Springfield, Missouri; and Rock
density is a function of the relative concentrations of the Springs, Wyoming, are a few examples.
liquids, solids, and gasses within the drilling fluid. Also For centuries, hot springs and mineral springs were
note that the specific gravity of the air and mist are related considered of medicinal or therapeutic value. Bottled
to the density of water. ‘‘springwater’’ is still considered by many people of
higher quality than ordinary tap water. This belief is
BIBLIOGRAPHY at least partially responsible for the current high-volume
sale of bottled water. Unfortunately, this association of
1. Lewis, A.J. and Rasmussen, T.C. (1999). Determination of springwater with purity is not based on fact. Springwater
suspended sediment concentration and particle size distri- can contain higher concentrations of dissolved solids than
butions using pressure measurements. J. Environ. Qual. 28: local public water supplies, and springs become easily
1490–1496. contaminated. Springs become contaminated because they
2. Driscoll, F.G. (1986). Groundwater and Wells, 2nd Edn. are open, unprotected, and accessible to human beings and
Johnson Filtration Systems, Inc., St. Paul, MN. animals. Contamination can also occur from surface water
flow or flooding into spring collection systems.
More than 3400 public water supply systems (1992)
HOT SPRINGS in the United States obtain part or all of their drinking
JOHN E. MOORE water from springs. These systems provide drinking water
USGS (Retired) for more than 7 million people. However, the average
Denver, Colorado number of people served by a single public supply system
using springs as a water source is small. Springs are
an important component of the drinking water supply in
Springs have been important in the social, cultural, and many states. Public water supplies that use springs are
economic history of humankind. They have been an impor- more numerous in the western United States. Wyoming
tant source of water for humans. Our ancestors did not has 80 springs that are used for public water supply. A box
have drilling equipment, so in the absence of surface water, used to collect water from springs is shown in Fig. 1 (1).

Figure 1. Diagram of a spring collection box (1).

476 HOT SPRINGS

A spring is a source of water that flows naturally For example, springs are common in mountains and in
from an aquifer or soil onto the land or into a body watersheds that are underlain by fractured rock aquifers.
of surface water. Its occurrence depends on the nature Springs can be regional (long flow paths) or local (short
and relationship of rocks (especially permeable and flow paths). Local springs are comparatively small, can
impermeable strata), on the position of the water table, have low flow, and are typically from shallow aquifers.
and on topography. The discharge from these springs often fluctuates either
Springs are generally classified as gravity springs and seasonally or in greater cycles, sometimes in response to
artesian springs. Thermal springs are typically considered local precipitation. Local aquifers are quickly recharged,
a type of artesian spring. The vast majority of springs are and water movement through them is comparatively rapid,
gravity springs. Gravity springs are created by water resulting in low water mineralization. Springs supported
that moves downgradient (downhill) and emerges at the by local aquifers are more likely than regional flow
surface. Two gravity springs, contact and fault springs springs to stop flowing periodically. Regional springs more
are shown in Fig. 2. Artesian (nongravitational) springs typically have large discharges. Regional springs typically
occur when the potentiometric level of the groundwater have nearly constant discharge and are more mineralized
system is above the land surface and water flows at the than local springs. Regional springs rarely stop flowing
land surface under pressure either at the aquifer outcrop even during long droughts.
or from fractures or faults. The discharge rate of a spring is a function of three
Springs are replenished by precipitation that recharges main variables:
aquifers. The precipitation seeps into the soil and enters
fractures, joints, bedding planes, or pore spaces in sand 1. hydraulic conductivity (permeability) of the aquifer
aquifers and sedimentary rocks. Springs occur when water
flows through aquifers and discharges at the land surface 2. area contributing recharge to the aquifer
through faults, fractures, or by flow along an impermeable 3. quantity of recharge.
layer. They can also occur where water flows from large
orifices that result when the water dissolves carbonate Springs are also classified based on hydrogeologic
rock (karst) and enlarges fractures or joints to create characteristics. The Water Resources Division of the
a passage (Fig. 2). U.S. Geological Survey recognizes eight principal types
Springwater is sometimes forced to the surface along of springs based on hydrogeologic characteristics:
a fault from deep sources by thermal pressure gradients.
Springs are associated with volcanism and fractures that
1. An artesian spring is a release of pressurized water
extend to great depths in the earth’s crust. Springs
from a confined aquifer at the aquifer outcrop or
are a common feature in some hydrogeologic settings.
through an opening in the confining unit.
2. A depression spring is formed when the water
Water table
table intersects a steeply sloping land surface. This
type of spring is sensitive to seasonal fluctuations
Sandstone Spring in groundwater storage and frequently disappears
during dry periods.
Seep
3. A fracture spring (fault spring) is formed in fractured
Stream or jointed rocks. Water movement is through
fractures, and springs form where the fractures
Shale intersect the land surface. This type of spring is
particularly sensitive to seasonal fluctuations in
groundwater storage and frequently disappears due
to reduced flow during dry periods. Daily fluctuations
Contact spring of discharge of small springs are commonly the result
of use of water by vegetation.
4. A contact spring occurs where a permeable water-
Sandstone bearing unit overlies a less permeable unit that
intersects the ground surface.
Water table
5. A geyser spring is a periodic thermal spring resulting
Spring
from the expansive force of superheated steam
within constricted subsurface channels.
Fault zone Swamp
Stream 6. A perched spring occurs where infiltrating water
discharges above the regional water table.
Slitstone 7. A seep spring discharges from numerous small
openings in permeable material. These springs
typically have very low discharge rates.
Fault spring
8. A tubular spring discharges from rounded channels
Figure 2. Contact and fault springs (2). (karst solution openings, lava tubes).
 SQUEEZING WATER FROM ROCK 477

Springs occur in many sizes, types of discharge points, Moore, J.E. et al. (in press). A Field Guide to Evaluate Springs.
and locations with respect to topography. They occur in USEPA, Washington, DC.
the highest elevations of mountainous areas and on valley Sanders, L.L. (1998). A Manual of Field Hydrogeology. Prentice-
floors. Many of the public water supply springs in the Hall, Upper Saddle River, NJ.
western United States are small; discharges are less than
5 gallons per minute. Springs vary in their physical and
chemical compositions. They can be cold (50 ◦ F) or hot SQUEEZING WATER FROM ROCK
(more than 10 ◦ C above the air temperature). Shallow
groundwater has a temperature within a few degrees of the
ambient air temperature. Higher temperatures usually NASA Earth Science Enterprise
indicate deeper circulation. Thermal springs gain their Data and Services
temperature increases when the water comes in contact
with recently emplaced igneous masses. Thermal and hot On December 16, 1811, an earthquake estimated at
springs are due to deep-seated thermal sources and are magnitude 7.0 to 8.0 on the Richter scale rocked
classed as volcanic springs or fissure springs. Springs the frontier town of New Madrid, Missouri, in the
may occur singly or in groups that can include dozens of central United States. It was the first of three powerful
habitats in various sizes. earthquakes that would shake the central Mississippi
Springs may be highly mineralized, especially thermal River Valley that winter. Damages were reported as far
springs and sometimes region springs that have a very long away as Charleston, South Carolina, and Washington, DC.
flow path. Thermal springs have pH values ranging from According to the U.S. Geological Survey, the New Madrid
7.2–7.6. Electrical conductance of spring water ranges events were the most powerful earthquakes in recorded
from 2–10,000 milligrams per million. U.S. history.
A list of question that could be asked to evaluate spring Survivors of the New Madrid earthquakes reported
conditions is shown below. not only intense ground shaking and land movement,
as would be expected during an earthquake, but also
• Are there standing water and surface drainage an unfamiliar phenomenon: water and sand spouting up
around the spring? through fissures, or cracks, in the Earth’s surface. In a
letter published in Lorenzo Dow’s Journal, New Madrid
• Is there deeply rooted vegetation in the spring
resident Eliza Bryan (1) wrote in 1816:
collection area?
• Are there roots in the collection pits?
. . . the surface of hundreds of acres was, from time to time,
• Is there a lock-tight fence around the collection area? covered over in various depths by the sand which issued from
• Is there a diversion channel capable of diverting the fissures, which were made in great numbers all over this
surface water away from the collection area? country, some of which closed up immediately after they had
vomited forth their sand and water. . . 25 June 2003
• Are there privies, septic systems, underground
storage tanks, or barnyards near the spring?
• Is the collection chamber properly constructed
(lockable)?

BIBLIOGRAPHY

1. U.S. Environmental Protection Agency. (1999). Guidance

Manual for Conducting Sanitary Surveys of Public Water
Systems.
2. Moore, J.E. et al. (1995). Groundwater-a primer. American
Geological Institute, Arlington, VA.

READING LIST

Davis, S.N. and De Wiest, R.J.M. (1966). Hydrogeology. John

Wiley & Sons, New York.
U.S. Environmental Protection Agency. (1997). Investigation of
Hydrogeologic Mapping to Delineate Protection Zones Around
Springs-Report of Two Case Studies. Structures in the village of Bhuj, India, which was 80
kilometers (50 miles) from the earthquake epicenter, sustained
Fetter, C.W. (2000). Applied Hydrogeology. Prentice-Hall, New
considerable damage. (Image Copyright  James Mori, Research
York.
Center for Earthquake Prediction, Disaster Prevention Institute,
LaMoreaux, P.E. and Tanner, J.T. (2001). Springs and Bottled Kyoto University)
Waters of the World. Springer, New York.
Meinzer, O.E. (1923). Outline of Ground Water Hydrology. U.S.
Geology Survey Water Supply Paper 494, p. 71. This article is a US Government work and, as such, is in the
Moore, J.E. (2002). Field Hydrogeology. Lewis, Boca Raton, FL. public domain in the United States of America.
478 SQUEEZING WATER FROM ROCK

Modern earthquake researchers refer to this expulsion

of water and sand from the ground as earthquake
dewatering, which results from liquefaction. Liquefaction
occurs when an earthquake vigorously shakes and
compacts water-saturated sediments. As the compaction The Bhuj earthquake struck the Kachchh region of northwestern
process displaces the water between the sediment pores, India on January 26, 2001. The bottom image is a detail of the
water and sand shoot upward and out of the ground. area indicated in red in the top map. (Images adapted from maps
‘‘Imagine a cube full of sand and water. If you press from the University of Texas at Austin General Libraries)
it in from both sides (compressing and releasing it,
then compressing it again), you build up what’s called
pore-water pressure,’’ said Martitia Tuttle, geologist and Earthquake Severity Table Adapted from What is Richter
consultant for the U.S. Geological Survey. ‘‘It’s like Magnitude? (A New Browser Window Will Open.)
shaking a coke can—when the pressure builds up and Richter Magnitude Earthquake Effects
you release it, the fluid comes shooting to the surface.’’
Less than 3.5 Recorded but generally not felt
These fountains of water can sometimes shoot as high as
3.5 to 5.4 Often felt but rarely causes damage
30 feet into the air, according to Tuttle (2). Less than 6.0 Slight damage at most to
Tuttle is part of a team that has been studying well-constructed buildings; over
liquefaction in the New Madrid region in an effort to small regions, can cause major
identify large prehistoric earthquakes in the geologic damage to poorly
record. ‘‘Interpreting historic and prehistoric earthquake constructed buildings
events involves a lot of uncertainty, due to the lack of 6.1 to 6.9 Can be destructive in populated
data,’’ Tuttle said. ‘‘We often don’t know precisely where regions up to 100 kilometers across
the earthquakes occurred or how big they were.’’ What 7.0 to 7.9 Major earthquake that can cause
serious damage over areas larger
researchers needed was a modern-day event in a similar
than 100 kilometers across
environment with which to compare the New Madrid
8 or more Great earthquake that can cause
earthquakes. serious damage over areas several
On January 26, 2001, a magnitude 7.7 earthquake hundred kilometers across
near the village of Bhuj, located in the Kachchh
region of northwestern India, provided scientists with
just that analogue. Considered one of the two most
Bhuj earthquake killed about 20,000 people and caused
damaging earthquakes in India’s recorded history, the
an estimated $3.3 billion in damages.
According to Tuttle and co-authors (3) of a paper
recently published in Eos, local residents and survey
teams working in the area after the earthquake reported
fountains of water and sediments surging from the ground
during and immediately following the Bhuj earthquake.
The water flow was significant enough to reactivate
streams in previously dry river channels.
‘‘The Bhuj earthquake provided an opportunity to study
an earthquake that was very similar to the New Madrid
event—it was a very large earthquake, and it occurred in
an intraplate-like setting,’’ said Tuttle. The New Madrid
and Kachchh regions are both located more than 185 miles
(300 kilometers) from active plate boundaries.
‘‘Before the Bhuj earthquake, there was a lot of debate
in the seismological community about the magnitude of the
New Madrid earthquakes, but based on the similarity in
intensity between the New Madrid and Bhuj events, there
is a growing consensus among scientists that the New
 SQUEEZING WATER FROM ROCK 479

Madrid earthquakes were probably larger than magnitude exposing the presence of free water close to the Pakistani
7.5,’’ said Tuttle. ‘‘The Bhuj earthquake is a modern border, a region where it is difficult, if not impossible, to
event that helps us better understand the historic and take ground measurements,’’ he said.
prehistoric earthquake record in the New Madrid region.’’ Most other satellite instruments only look straight
But despite the unique study opportunity presented by down or towards the edge of the Earth. ‘‘It would take
the Bhuj earthquake, the area presented the researchers several more passes for another sensor to capture the
with some grave challenges. First, the Kachchh region of entire region that the MISR imagery covers,’’ said Tuttle.
India is remote and difficult to reach. Second, it shares ‘‘The MISR images enabled us to see a large area both
a politically sensitive border with Pakistan. These factors during and immediately after the earthquake, and we
made it difficult for research teams to conduct field studies could see that dewatering was occurring over a huge
to verify liquefaction effects or the presence of water in region.’’
some areas. Scientists now know that the New Madrid earthquakes
Bernard Pinty (4), research scientist at the Institute were not geologic flukes; similar events have occurred
for Environment and Sustainability in the Joint Research repeatedly in the geologic past. In fact, the central
Centre of the European Commission in Ispra, Italy, and Mississippi Valley experiences more earthquakes than
colleagues found that water expelled during the earth- any other region in the United States east of the
quake could be detected in imagery from NASA’s Multi- Rocky Mountains.
angle Imaging SpectroRadiometer (MISR) satellite sensor. The New Madrid Seismic Zone, which includes parts
The MISR instrument, which flies aboard NASA’s Terra of Tennessee, Kentucky, Missouri, Arkansas, Illinois,
satellite, views the sunlit face of the Earth at nine Indiana, and Mississippi, harbors a large liquefaction field.
widely spaced angles simultaneously. As it passes over Scientists estimate the probability of a magnitude 6.0 or
the Earth, its nine cameras successively view each piece greater earthquake occurring in the New Madrid Seismic
of the planet’s surface at a spatial resolution of 300 yards Zone within the next 50 years at 25–40 percent, according
(275 meters). to the U.S. Geological Survey. But a similar event would
The change in reflection at different angles provides a have much more serious consequences today.
means for distinguishing different types of land surface In 1811–1812, the central Mississippi Valley was a
covers. Since bright soils in India’s Rann of Kachchh (also remote frontier. Today, the region is home to millions
known as the Rann of Kutch) reflect most of the sun’s near- of people and includes the metropolitan cities of St.
infrared radiation, and water bodies absorb near-infrared Louis, Missouri, and Memphis, Tennessee. In addition,
radiation, MISR is able to detect the contrast and indicate most buildings in the region were not built to withstand
where dewatering occurred. earthquake shaking, as they often are in California
According to Pinty, these features made MISR an ideal and Japan.
tool for studying liquefaction in the Bhuj earthquake. Buildings and engineered structures are at much higher
‘‘MISR’s multi-angle capability was instrumental in risk in areas where liquefaction occurs, due to increased
ground motion. ‘‘The liquefaction process is extremely
damaging to all sorts of engineered structures; it has

This false-color MISR image was acquired after the Bhuj

earthquake, on January 31, 2001. The image shows numerous
These cracks in the ground surface north of Chobari, India areas where groundwater flowed up to the surface, including
were caused by liquefaction induced during the Bhuj earthquake. within the Rann of Kachchh, as well as near the Indo-Pakistani
(Image Copyright  James Mori, Research Center for Earthquake border. (Image courtesy of NASA/GSFC/LaRC/JPL, MISR Team)
Prediction, Disaster Prevention Institute, Kyoto University) Click here for more information.
480 STORAGE COEFFICIENT

caused major destruction in both Anchorage and San ‘‘The interesting thing is that this project started as an
Francisco,’’ said Pinty. ‘after hours’ sort of activity,’’ said Pinty. ‘‘But it soon
‘‘The ground failure that resulted from liquefaction acquired its own momentum, giving us results that are
during the New Madrid earthquakes was severe. We’re of significant relevance to earthquake study applications.
talking about vertical displacement of 3 to 6 feet (1 Our findings revealed that MISR can detect water and
to 2 meters), and lateral displacement up to 33 feet moist conditions on terrestrial surfaces, even under
(10 meters),’’ said Tuttle. ‘‘A recurrence of that type of unfavorable spectral conditions.’’
event would have severe consequences for engineered ‘‘If we study modern earthquakes that produce
structures.’’ liquefaction, we can better interpret the geologic record of
According to Tuttle, in 1988 a magnitude 5.9 earth- liquefaction during past events. This helps us anticipate
quake in Quebec, Canada, produced liquefaction. ‘‘Base- what is likely to happen in the future so that we can
ments cracked, septic fields were disrupted, and people make informed decisions about reducing and mitigating
described water and sand shooting into their basements, hazards,’’ said Tuttle.
out of their toilets, and into their bathtubs,’’ she said. ‘‘And For more information, visit the Langley Atmospheric
that was just a moderate-sized earthquake, not a big one.’’ Sciences Data Center DAAC. (A new browser window
MISR imagery, which is archived at the Langley will open.)
Research Center DAAC, promises to be a boon to ‘‘It’s one of those things where people tend to think—if
earthquake researchers seeking to minimize hazard risks. it hasn’t happened during my lifetime, then it can’t
happen here,’’ said Tuttle. ‘‘But the liquefaction field in
the New Madrid region is very large. We’re talking about
a huge earthquake that could have a significant impact on
society.’’

BIBLIOGRAPHY

1. Bryan, E. (1849). New Madrid, Territory of Missouri, March 22,

1816. In Lorenzo Dow’s Journal. Published by Joshua Martin:
344–346. Quoted by The Virtual Times, The New Madrid
Earthquake. Accessed June 5, 2003.
2. Tuttle, M.P. (2001). The use of liquefaction features in
paleoseismology: Lessons learned in the New Madrid seismic
zone, central United States. Journal of Seismology. 5: 361–380.
3. Tuttle, M.P. et al. (2002). Observations and comparisons of
liquefaction features and related effects induced by the Bhuj
The MISR instrument views the sunlit Earth simultaneously at earthquake. Earthquake Spectra. 18(Supp. A): 79–100.
nine widely spaced angles, providing ongoing global coverage at 4. Pinty, B. et al. (2003). Observing earthquake-related dewa-
high spatial detail. (Image courtesy of NASA) tering using MISR/Terra satellite data. EOS, Transactions,
American Geophysical Union. 84(5): 37–48.

This article contributed from Distributed Active Archive

Center (DAAC) Alliance: Supporting Earth Observing
Science 2003
For a copy, e-mail mailto:nasadaacs@nsidc.org?subject
=DA 2003 annual request via EO.

STORAGE COEFFICIENT
MATTHEW M. ULIANA
Texas State University—San
Marcos
San Marcos, Texas

Although earthquakes in the central and eastern United States DEFINITIONS

are less frequent than in the western United States, they
affect much larger areas. This is shown by two areas affected
A coefficient is a mathematical quantity that relates two
by earthquakes of similar magnitude—the 1895 Charleston,
Missouri, earthquake in the New Madrid Seismic Zone and the or more variables in an equation describing some physical
1994 Northridge, California, earthquake. Red indicates regions of process. In equations describing the flow of fluids in a
minor to major structural damage; yellow indicates areas where porous medium (e.g., groundwater in an aquifer), the
shaking was felt, but little or no damage to objects occurred. storage coefficient is the quantity that relates the
(Image courtesy of the U.S. Geological Survey) change in fluid potential (i.e., hydraulic head) in the
 STORAGE COEFFICIENT 481

aquifer to the change in the amount of water stored in the can be substituted for q to get the expanded version of the
aquifer at a given point. In practical terms, the storage flow equation:
coefficient tells us the volume of water that can be removed



from an aquifer for a given drop in hydraulic head. ∂h ∂h ∂h
∂ Kx ∂ Ky ∂ Kz
There are three storage coefficients used in groundwa- ∂x ∂y ∂z ∂h
+ + = Ss (3)
ter flow equations (the units of each are given in [brackets] ∂x ∂y ∂z ∂t
following each definition):
Equations 1 and 3 are specific versions of the continu-
Specific Storage (Ss ) is the volume of water removed ity (or conservation of mass) equation,
from a unit volume of a confined aquifer for a unit
drop in hydraulic head [L−1 ]. I − O = S (4)
Storativity (S) is the volume of water removed from
a unit area of an aquifer for a unit drop in where I = flux into a volume of a system [L3 · t−1 ]
hydraulic head; in confined aquifers, it is equal to the O = flux out of a volume of a system [L3 · t−1 ]
specific storage times the thickness of the aquifer, in S = change in the amount stored in that same
unconfined aquifers, it is equal to the specific storage volume of the system [L3 · t−1 ]
times the thickness of the aquifer plus the specific
The left side of Eq. 3 corresponds to the left side of
yield [−].
the continuity equation and describes the divergence of
Specific Yield (Sy ) is the volume of water removed flux at a point in the aquifer. In other words, it describes
from a unit volume of an unconfined aquifer for a the groundwater flow vector resolved into x, y, and z
unit drop in water level; this quantity is also defined components (see Fig. 1).
as the volume of water that will drain by gravity The right side of Eq. 3 corresponds to the right side
from a saturated unit volume of an aquifer and is of the continuity equation and describes the change in
equivalent to the effective porosity [−]. storage at the given point in the aquifer. The term ∂h/∂t
Specific Storage (Ss ) refers to the change in hydraulic head, with respect to
time, at the same point in the aquifer. The term Ss is the
The main equation that governs the flow of a fluid in a specific storage, and it relates the changes in head in the
porous medium is given by aquifer to the change in the amount of water stored in the
∂qx ∂qy ∂qz ∂h aquifer at a certain point. In other words, it is a quantity
+ + = Ss (1) that tells us the volume of water that can be removed from
∂x ∂y ∂z ∂t
a unit volume of a confined aquifer for a unit drop in head.
where q = specific discharge (i.e., groundwater flow) in the
x, y, or z direction [L · t−1 ]
CONFINED AQUIFERS—A CONCEPTUAL
h = hydraulic head [L]
UNDERSTANDING
t = time [t]
x, y, z = lengths in the principal directions in our Consider a confined aquifer that consists of a layer of
coordinate systems [L] sand overlain by an impermeable layer of clay that acts
Ss = specific storage [L−1 ] as a confining layer. If a well is drilled into the aquifer
Darcy’s law and water is pumped from the well, the heads around
dh the well will lower and create a cone of depression in the
q=K (2)
dl potentiometric surface (Fig. 2). The more water that is
pumped from the well, the larger the cone of depression.
where K = hydraulic conductivity [L · t−1 ] In the example shown in Fig. 2, the pumping well has
dh/dl = hydraulic gradient, or change in head with removed a volume of water (represented by the volume
length [−] in the barrel), and in response, the water levels in

Flow vector in 3-D Flow vector in 3-D

z z z

Resolved
vector in
z-direction
y y y
Resolved
vector in y-direction
x x x
Resolved
vector in
x-direction

Figure 1. Flow vector resolved into component vectors in three dimensions.

482 STORAGE COEFFICIENT

the well and in the surrounding aquifer have dropped.

However, the saturated thickness of the aquifer itself has Unit volume of a
not changed (i.e., the aquifer is still full to the base of confined aquifer
the confining layer). What is the source of this water?
Additional
The conceptual answer deals with elastic deformation and
component
compressibility of the aquifer materials and the water. of pressure
This aquifer is made up of sand grains and pore spaces
that are full of water. If there is no confining layer and
the aquifer is simply recharged by water infiltrating from
above, the pore spaces are full of water, and the pore
fluid pressure is equal to atmospheric pressure plus the
overlying column of water. In this case, the volume of
water in the pore space is equal to the volume of the
pore space.
However, if the aquifer is confined and an additional
component of pressure is added to the water, that pressure
will force a greater volume of water into the pores of the
sand. It does this by compressing the sand grains slightly,
pushing the sand grains around so that pore volume
is increased, and compressing the water itself slightly.
Increasing the pore fluid pressure causes the aquifer to
undergo elastic (i.e., recoverable) deformation, so that 1-meter
1 meter
a subsequent decrease in fluid pressure will cause an
Figure 4. Unit volume of a confined aquifer.
associated decrease in volume (Fig. 3).
This is the source of the additional volume of water.
Pumping water from the confined aquifer lowers the excess
fluid pressure (i.e., the pressure head above the top of the aquifer), causing the aquifer to ‘relax’ and yield water that
is stored by the excess pressure.

Q CONFINED AQUIFERS—A MATHEMATICAL

UNDERSTANDING
Potentiometric
surface
Consider a unit volume of a confined aquifer. This unit
volume is saturated with water, and the water has an
additional component of pressure that would cause the
Confining layer water level in a manometer to rise above the top of the
cube (Fig. 4). If a volume of water is removed from the
cube, it will reduce the additional component of pressure
and lower the head in the cube. Removing more water
Confined aquifer
will cause more head loss; less water will cause less head
loss. Therefore, there is a specific volume of water that,
if removed, will cause a unit drop in head. This volume
of water that results in a unit drop in head in the unit
Figure 2. Pumping well in a confined aquifer. volume of aquifer is the specific storage.
The specific storage is also called the elastic storage
coefficient; it deals only with the amount of water
removed from elastic storage. The specific storage is
a function of the compressibility of the aquifer matrix
∆h (i.e., the material that makes up the aquifer) and the
compressibility of the water and is strictly defined as

Ss = ρw g(α + nβ) (5)

where ρw = density of water [M/L3 ]

g = gravitational constant [L/t2 ]
α = compressibility of the aquifer matrix [1/M/Lt2 ]
Sand grains in an aquifer
n = porosity [−]
Figure 3. Elastic deformation in a sand aquifer. β = compressibility of water [1/M/Lt2 ]
 QANATS: AN INGENIOUS SUSTAINABLE GROUNDWATER RESOURCE SYSTEM 483

STORATIVITY IN CONFINED AQUIFERS aquifer, the storativity of the aquifer is the sum of the
specific yield and the thickness times the specific storage.
The specific storage actually refers to an infinitesimal
point in the aquifer. However, in real aquifers, the S = Sy + bSs (7)
thickness of the aquifer influences how much water the
aquifer will yield; a thicker aquifer will transmit and yield However, the magnitudes of these parameters are
more water than a thinner aquifer. Multiplying the specific drastically different: specific yield values in an aquifer
storage by the saturated thickness of the aquifer results generally range from 1% to 50%, or 0.01 to 0.5; specific
in storativity (S): storage values generally range from 1 × 10−3 to 1 × 10−6 ,
S = Ss b (6) or 0.001 to 0.000001, which means that, in an unconfined
aquifer, the volume of water provided by release from
elastic storage is orders of magnitude less than that
where b = saturated thickness of the aquifer [L]. provided by dewatering the pores; so the specific storage is
In other words, the storativity describes the volume of generally considered negligible, and unconfined storativity
water removed from a unit area of aquifer for a unit drop is a function only of the specific yield.
in head.

QANATS: AN INGENIOUS SUSTAINABLE

UNCONFINED AQUIFERS AND SPECIFIC YIELD
GROUNDWATER RESOURCE SYSTEM
In a confined aquifer, all the water pumped out through BAHRAM SAGHAFIAN
a well is provided by the release of water from elastic Soil Conservation and
storage. In an unconfined aquifer, the same thing Watershed Management
occurs—the changes in head result in a release of water Research Institute
from elastic storage in the aquifer. However, in an Tehran, Iran
unconfined aquifer, the pumping well is dewatering a
section of the aquifer (Fig. 5), and this part of the aquifer’s
response adds an additional volume of water to the volume A qanat system consists of an underground tunnel that
withdrawn from elastic storage. conveys water by gravity from highland aquifers to
This additional volume of water is released from the the ground surface at lower elevations. Blended with
pore spaces and is due to the specific yield of the aquifer. ingenuity, its purpose is simply to collect and bring
Specific yield is the volume of water that drains from a underground water to the surface for use. Qanat, invented
rock under the force of gravity and is directly related to and originated in Iran more than 2700 years ago, is
the effective porosity of the aquifer matrix. The volume considered a hydraulic engineering system of an Old World
of water released from a unit volume of aquifer due to era providing water to settlements since its invention. The
dewatering of the aquifer is therefore the unit volume overwhelming climate of Iran is arid and semiarid; two
times the specific yield. thirds of the country area in central, southern, and eastern
So, in unconfined aquifers, the volume of water released parts receive an average of about 150 mm of rainfall
from a unit volume of aquifer is a function of both annually. Yet, Iran is considered a farming country where
the specific yield (i.e., the water removed from the pore diverse crops are grown. This remarkable achievement in
space) and the specific storage (the water released from agricultural development of arid lands has materialized
elastic storage in the aquifer). Therefore, in an unconfined partly by building qanats to tap underground water in
regions where surface water is lacking or unreliable during
seasons of high water demand. Numerous urban and rural
qanat-watered settlements are vivid examples that not all
Water is removed civilizations were rooted by the sides of rivers. Qanats
from the pore spaces Q
have helped to transform barren lands into productive
assets. Thanks to its simple and efficient technique, the
Potentiometric idea behind this sustainable water resource system was
surface
passed from one generation to the next and traversed
or water table
geographic boundaries. Qanat’s use is not confined to Iran
and has been adopted in many regions across the world.

QANAT DESCRIPTION
Unconfined aquifer
Britannica characterizes a qanat as, ‘‘A qanat taps
underground mountain water sources trapped in and
beneath the upper reaches of alluvial fans and channels
the water downhill through a series of tunnels, often
several kilometres long, to the places where it is needed
Figure 5. Pumping well in an unconfined aquifer. for irrigation and domestic use . . ..’’ The dictionary of
484 QANATS: AN INGENIOUS SUSTAINABLE GROUNDWATER RESOURCE SYSTEM

(a)

Shafts
Mother well

Shafts

Qanat
outlet
Qanat gallery

Water table

(b)
Impermeable layer

Figure 1. Schematic plan (a) and cross section (b) of a qanat system.

Islamic Architecture defines qanat as a ‘‘subterranean normally recharged from a seasonal river, stretching over
canal system usually used to bring water some distance a long distance where the tunnel meets the ground surface.
from a river or mountains.’’ More technically, Lightfoot (1) Other terms for qanat, pronounced kanat, as it is
describes a qanat as ‘‘a form of subterranean aqueduct, or called in different parts of the world, are karez (e.g., in
subsurface canal, engineered to collect groundwater and Afghanistan and eastern Iran), foggora (e.g., in Algeria),
direct it through a gently sloping underground conduit to hattara (e.g., in Morrocco), and falaj (e.g., in Oman).
surface canals which provide water to agricultural fields.’’
As shown in Fig. 1, a qanat consists of a gently sloped HISTORY AND GEOGRAPHIC EXTENT OF QANATS
underground tunnel, or gallery, penetrating the upland
aquifer at one end and appearing in the ground surface Written records and recent excavations prove that qanats
at the other end. The gallery, is also vertically connected first appeared in ancient Iran (then Persia). Wulff (2)
to the ground by several wells, or shafts, dug along its describes the origin of qanat: ‘‘As early as the seventh
length, often at 20–30 m to a maximum of 100 m intervals, century B.C., the Assyrian king Sargon II reported that
providing ventilation and access during construction and during a campaign in Persia he had found an underground
maintenance. Put differently, a qanat consists of a series system for tapping water in operation near Lake Urmia.
of vertical shafts in sloping ground, inter-connected at the His son, King Sennacherib, applied the secret of using
underground conduits in building an irrigation system
bottom by a tunnel with a gradient flatter than that of the
around Nineveh, and he constructed a qanat on the
ground. The excavated soil is piled up around the entrance
Persian model to supply water for the city of Arbela.’’
of the wells to give some protection in times of storm
The earliest reports of qanats indicated that the water
surface runoff. The first shaft, called the mother well or
for the city of Ula northwest of Lake Urmia, located
main shaft, is usually sunk into the alluvial fan to a level
in northwest of Iran, was provided by qanats some 700
below the groundwater table. The line of qanat wells looks B.C. (3). English (4) reports that, based on early mining, the
like anthills in a distinctive plan view. The typical cross mountains of Kurdistan in western Iran, eastern Turkey,
section of a qanat gallery is 1.5 m high and about 1 m and northern Iraq were the birthplace of the qanat more
wide, which is just sufficient to allow men and boys to do than 2500 years ago.
the digging and cleaning. The technique was then widely diffused during the
Qanats can be classified into hill and river-based. Hill expansion of Achaemenid dynasty, ruling the Persian
qanats have their main shafts based in the aquifers located Empire, beginning sometime in the fifth century B.C.
in hill slopes. This type of qanat is not very long. If Qanats were constructed in Mesopotamia to the shores
this type of qanat taps a deep aquifer, it will enjoy of the Mediterranean to the west, in Egypt and the
a relatively constant discharge. A river-based qanat is Arabian Peninsula to the south, and in Afghanistan,
 QANATS: AN INGENIOUS SUSTAINABLE GROUNDWATER RESOURCE SYSTEM 485

Central Asia, and western China to the east. The Arabs

introduced the idea further to the west during the second
diffusion phase in the eighth century A.D. The greatest
developments of qanat technology outside the Persian
core area are in southern Morocco (Marrakesh and the
Tafilalet) and southern Algeria (5). Wulff (2) reported
the excavation of several particularly interesting qanats
constructed by Arab rulers of early medieval times in the
Middle East. These include a small qanat built by the
caliph of Damascus to supply water for his palace in 728
A.D. A century later, presumably with the aid of Persian
engineers, the caliph Mutawakkil in Iraq constructed a
qanat that brought water to his residence at Samarra
from the upper Tigris River 300 miles away.
Qanats are reported to exist in Spain, Sicily, Cyprus,
the Canary Islands, and in South American countries such
as Mexico, Peru, and Chile. The Madrid qanat system in Figure 2. A top view of a qanat during construction.
Spain, bringing water for urban use, still functions. The
qanat systems of Tehuacan and Parras in Mexico are
believed of Spanish origin, but those of Parras (Coahuila) requires great skill and endurance. Figure 2 shows a top
and Pica (Chile) may be pre-Columbian (6). English (5) view of qanat during construction.
and Beekman et al. (7) discuss scenarios of the diffusion Muqannis may encounter a number of dangers in their
of qanat technology throughout the world. work. Soft soil in the tunnel can cause collapse of the roof.
In such cases, the tunnel cross section is braced with baked
CONSTRUCTING A QANAT clay collars to control cave-ins (Fig. 3). Another danger is
air low in oxygen. Any change in the burning of the
The laborious construction of a new qanat is a task to illuminating oil lamps is an indication of such danger. And
be engineered by an expert, who usually heads a team of last, but not least, is the hazard during the breakthrough
diggers. The diggers, called muqannis, are responsible for into the mother well, which must be emptied.
building and repairing qanats. The expert must decide on
a number of key issues prior to qanat construction: the
site of the mother well and the alignment and the slope
of the underground gallery. In his work, the expert uses
simple instruments such as a long rope and a level. As the
mother well of a qanat is usually dug in an alluvial fan
at the foot of a mountain or hillside, the expert watches
for signs of seepage on the ground, slight variations in
the vegetation, and available groundwater. Once the site
is chosen, one or sometimes more trial wells are sunk.
The muqannis then start digging the trial well, normally
1 m in diameter. They use a windlass at the top and haul
up the soil excavated by their teammates. They continue
the digging and spoil the material around the rim of the
well until the water table is struck. The expert then should
determine whether the aquifer water has acceptable water
yield. If so, he measures the distance from the mother well
to the qanat outlet, where water is planned for use. Once
the general alignment of the qanat is determined, the
expert calculates the slope of the qanat gallery. This slope
must allow the water to flow, yet prevent erosion of the
bottom and walls of the gallery. The slope for short qanats
may run from 0.5 to 2 in 1000, whereas longer qanats
usually have gentler slopes. In steep land, the slope of
the gallery is maintained in a looping course. The work
then proceeds from the outlet of qanat by excavating the
tunnel back toward the mother well. Vertical shafts at a
distance of 30 m, or so, are dug to provide ventilation and
access. For shafts deeper than 100 m, the muqannis set up
another windlass halfway within the shaft for a two-stage
haul up of spoil. The muqannis maintain a straight course
by sighting along a pair of burning oil lamps. The task Figure 3. Cross section of a qanat gallery protected by a lining.
486 QANATS: AN INGENIOUS SUSTAINABLE GROUNDWATER RESOURCE SYSTEM

Muqannis are highly respected for their courage, hard additional aquifers or excavate the floor of the existing
work, and know-how. They are paid well for their work, tunnel to tap water at a deeper level (2).
and older muqannis are considered blessed or lucky. Note that, although agricultural production cannot
repay the investment and maintenance costs of new qanats
in some regions, existing qanats are still the second
QANATS AND WATER REGULATIONS
water supply system in volume providing for drinking
and domestic use in Iran. Keeping the qanat cost and
In Iran, the allotment of water delivered by qanats is
the enormous building effort in mind, individual farmers
regulated by certain laws, which are rooted in a traditional
are unwilling and unable to fund the construction of new
system of water rights. Normally a water master or bailiff,
qanats or even maintain old ones. Thus, farmers have
appointed by a qanat owner(s), supervises the water
turned to drilled wells as the more affordable alternative.
distribution according to each farmer’s time share. As
Wulff (2) has noted, ‘‘the bailiff is guided by an allocation
system that has been fixed for hundreds of years. For QANAT VERSUS DRILLED WELL
instance, three hamlets in a region west of Iran, still
receive the shares that were allotted to them in the 17th People in arid regions rely on underground aquifers for
century by the civil engineer in the reign of Shah Abbass. their water supply. The aquifers can be tapped by a
The hamlets of Dastgerd and Parvar in central Iran are qanat, or alternatively by a drilled well, and the water
entitled to eight shares apiece. These allocations are built is directed to an area where it is consumed. The volume of
into outlets from the qanat distribution basin.’’ English (4) water produced by a qanat depends upon several factors:
provides a detailed overview of qanat ownership, the extent, type, yield, and recharge rate and area of the
complex water distribution system, and, through some aquifer. Comparing a drilled well with a qanat system,
examples, how fragmentation of qanat water rights has there are advantages for each. The following compare the
taken place. He makes a note of strict and unforgiving two systems:
adherence to communal methods of water rotation and
upkeep of the water supply system serving to maintain • As history proves it, qanats are sustainable renew-
the social cohesion of qanat-watered villages. able water resource systems. However, they cannot
meet today’s increasing demand. Wells are the alter-
native at the expense of inducing negative water
QANATS: COSTLY AND TIME-CONSUMING
balance in the aquifers they exploit.
Qanats are labor-intensive and time-consuming to build. • Qanats need no extra energy to harvest water,
Both the cost and time to build a new qanat varies with whereas wells need diesel or electric power to bring
topography, depth of the water table, underground soil water to the surface.
type, length of the qanat, type of ownership, availability • Qanats can be built and maintained using low-
and experience of muqannis, and local conditions. Thus, technology tools and material locally available.
reports differ in figures provided for the cost and duration • Qanats can be used as a passive cooling system in
of construction of qanats. conjunction with a wind tower (9).
Noting that qanats are long-lasting water resources • Qanats reinforce social cooperation, an important
systems if properly maintained, Wulff (2) believes that the issue in traditional societies.
agricultural production sustained by qanat water repays
• Wells are dug in a short time, whereas qanat
the investment in construction and maintenance by 10%
construction is slow, investment-intensive, and quite
to 25% depending on the size of the qanat, the water
time-consuming. However, technology can be put to
yield, and the crop type. Wulff’s assessment of the cost
work to accelerate qanat construction by cutting the
in the 1960s showed that a 6-mile long qanat required
time and the cost.
US$13,500 to US$34,000 to build. The cost runs to about
$90,000 for a qanat 10 to 15 miles long. • Only a portion of a qanat’s water is used during
English (8), on the other hand, reported that it took nonirrigation season, unless conserved by other
27 years to construct a 1-km qanat in Kerman, in central means. During periods of low water use in fall
Iran, with a mother well 45 meter deep, due to three and winter, watertight gates can seal off the qanat
changes in ownership. Another 3-km qanat with two outlet damming up the water. In spring and summer,
mother wells, 50 and 55 m deep, and a bifurcated tunnel night flow may be stored in small ponds for daytime
took 17 years to build. The costs of these two qanats, use (10). A well pump can be switched off when water
respectively, were approximately $10,000 to $11,000 per is not needed.
km in the 1960s. As another example, the cost of • A typical qanat’s discharge is less than that of
constructing a 40-km long qanat near the city of Kerman pumped wells. The latter are favored to meet growing
with a mother well 90 m deep was roughly $213,000 when demands, though, in a nonsustainable manner.
completed in 1950 (4). Such expenses adjusted for inflation
can be raised only through government subsidies or large Experts in Iran now suggest an integrated approach of
corporations in modern Iran. Due to high construction conjunctive use of qanats and wells. The location of a
investment, qanat owners often try to increase the yield. well must be optimized with respect to an existing qanat
They may extend branches from the main tunnel to reach recharge area.
 LYSIMETERS 487

STATISTICS OF QANATS 2. Wulff, H.E. (1968). The qanats of Iran. Sci. Am. April:
94–105.
No one knows how many qanats were ever built, how many 3. Laessøe, J. (1951). The irrigation system at Ulhu, 8th century
are still in operation, and how much water they deliver. BC. J. Cuneiform Stud. 5: 21–32.
Available reports vary in numbers, albeit they have 4. English, P.W. (1998). Qanats and lifeworlds in Iranian
collected qanat statistics in different times. It is estimated plateau villages. Yale F&ES Bull. 103: 187–205.
that, as of the 1960s, between 40,000 to 50,000 qanats had 5. English, P.W. (1968). The origin and spread of qanats in the
been built in Iran during its history, of which some 27,000 Old World. Proc. Am. Philos. Soc. 112: 170–181.
running qanats were still used (11). Wulff (2) reported 6. Humlum, J. (1965). Underjordiske vandingskanaler: kareza,
that some 22000 qanats in Iran, totaling 170,000 miles qanat, foggara. Kultergeografi 16: 81–132.
of underground conduits built by manual labor, had a 7. Beekman, C.S., Weigand, P.S., and Pint, J.J. (1999). Old
total yield of 19,500 cubic feet of water per second, which World irrigation technology in a New World context: qanats
amounts to 17 billion cubic meters of water annually. Two in Spanish Colonial Western Mexico. Antiquity 73(279):
recent reports prepared in Iran put the total number of 440–446.
qanats, as of 1995, between 25,000 and 33,000, but it 8. English, P.W. (1966). City and Village in Iran. University of
is unclear whether all are currently used. The reports, Wisconsin Press, Madison, WI.
however, agree on the total volume of water provided by 9. Bahadori, M.N. (1978). Passive cooling systems in Iranian
qanats at around 9 billion cubic meters. Having accepted architecture. Sci. Am. February: 144–154.
this figure and the figure given by Wulff in 1968, an 10. Beaumont, P. (1989). The qanat: a means of water provision
alarming decline of about 50% in water yield of qanats is from groundwater sources. In: Qanat, Kariz, and Khettara,
inferred, which must be partly due to an excessive number Wishbech. P. Beaumont, M. Bonine, and K. McLachlan (Eds.).
Menas Press, pp. 13–31.
of wells causing underground water tables to drop and
affect qanat yield. The declining groundwater tables and 11. Goblot, H. (1979). Les Qanats: Une Technique d ’Acquisition
de l ’Eau. Mouton, Paris.
increasing number of abandoned qanats have shrunk the
share of water supplied by qanats to less than 10% of total
use.
The discharge of individual qanats varies widely with LYSIMETERS
the location and, in some cases, time of the year. Qanat
KRISTOFOR R. BRYE
discharges of a few liters to as much as 1 cubic meter per
University of Arkansas
second have been reported. Fayetteville, Arkansas
There are some unique qanats in Iran described
as follows:
Lysimetry is a measurement technique that employs
1. Vazvan Qanat in Meimeh, Isfahan province, is sampling devices called lysimeters to measure the
1800 m long. At a distance of 600 meter from its movement of water and solutes through a porous
outlet, a stone dam has been built to store the water.
medium. Lysimeters have most commonly been used to
A total of six gates have been installed at different
quantify water movement (i.e., drainage) through soil and
elevations to control the flow of water, depending on
typically consist of a porous plate surface of varying pore
the time and amount needed. This qanat system is
sizes through which water moves from the soil into a
essentially an underground dam.
collection reservoir.
2. The well-known and unique Moun Qanat in
Ardestan, Isfahan province, is a twin-gallery qanat
system, the only one of its kind reported in the world. BRIEF HISTORY OF LYSIMETERS
The two galleries are independent conduits, about
3 m apart in elevation. A thin impermeable layer According to Joffe (1), lysimeter studies were initiated in
is believed to separate the two qanat conduits. The the late eighteenth century to address the fate of pre-
water in the layers is not mixed, as the upper conduit cipitation in soil. Many lysimeter studies then shifted to
is bent around the shafts. The qanat is about 2 km consider leaching of soil and precipitation-deposited con-
long, and its total discharge waters 25 ha. stituents, such as ammonium and sulfur (1). Gradually,
3. Gonabad region, in Khorasan province northeast of lysimeters were adopted to investigate fertilizer losses
Iran, is known for some of the oldest qanats in from agricultural and nonagricultural soil (1).
Iran. Reports indicate that one qanat has a mother Early lysimeter research involved greatly disturbed
shaft with a depth of 140 meters and a length of soil profiles, unrepresentative of natural soil properties.
about 70 kilometers. Another qanat is reported with Lysimeter tanks were dug and refilled with soil after
a 270-m mother well. construction was complete (1). For this reason, the
validity of drainage and leaching measurements were
BIBLIOGRAPHY severely questioned, as they are today. Nonetheless, many
technological and methodological advancements have been
1. Lightfoot, D.R. (1997). QANATS in the Levant: Hydraulic made since the late 1700s, but the same fundamental
technology at the periphery of early empires. Technology and question remains today as it did then: how to measure
Culture 38(2): 451. natural water movement through undisturbed soil?
488 LYSIMETERS

DRAINAGE LYSIMETERS for general lysimeter installation may generate biased and
uninterpretable results (4).
Two general types of lysimeters exist for monitoring water Kung (6) described another disadvantage of nonweigh-
and solute fluxes through soil, weighing and nonweighing ing lysimeters, in which funnel flow may divert percolating
lysimeters. Weighing lysimeters are large-scale sampling water based on the angle of inclination of coarse textured
devices typically used for measuring evapotranspiration, soil lenses. Lenses, or soil textural discontinuities, may
drainage, and solute leaching. Inputs and losses of cause water to bypass the lysimeter or channel water
moisture are evaluated on the basis of changes in mass in directly to it. In either case, results may be erroneous.
a containerized column of soil within the lysimeter (2). Tension or tensionless column lysimeters that have
Weighing lysimeters are expensive, labor-intensive to sidewalls extending above the soil surface pose additional
construct, and usually involve filling the lysimeter by challenges (7–9). Impeding runoff, as the design of
reconstructing a disturbed soil profile (3,4), although large extended-walled column lysimeters do, unnaturally alters
weighing lysimeters that have undisturbed soil columns the water status of the contained soil column by permitting
have been constructed, such as the monolith lysimeters ponding during winter months and following intense
near Coshocton, Ohio. However, nonweighing lysimeters rainfall. Furthermore, crop management becomes more
are better suited for less burdensome, replicable field difficult as the extended sidewalls act as obstacles to
measurements of in situ water and solute fluxes. plowing, seeding, and cultivating (9).
Nonweighing lysimeters are well adapted to monitoring Zero-tension lysimeters rely on the formation of a
drainage and leaching from soils. Nonweighing lysimeters, saturated soil zone above the porous plate before drainage
also known as drainage or pan lysimeters, are typically occurs by gravitational flow, which is an unlikely natural
much smaller than weighing lysimeters. An advantage condition in most deep, well-drained and moderately
of a nonweighing pan lysimeter is that replicable field well-drained soils and alters the water content and soil
measurements can be obtained below undisturbed soil water potential of the rooting zone (7). The prerequisite
profiles. As a result, site-specific information can be of saturation above these sampling devices may cause
gathered relative to the mass flux of water and leachate drainage water to bypass the lysimeter (3) and create
concentrations of soluble organic and inorganic chemical unnatural matric potential gradients that cause water
species (3). Coupling drainage measurements with solute movement from the saturated zone to the surrounding
concentrations provides a means of estimating solute mass unsaturated soil matrix (10). Zero-tension lysimeters that
fluxes through an undisturbed soil profile. have upwardly extended sidewalls (7,8,11) may promote
preferential flow down the side of the soil column upon
DRAINAGE MEASUREMENTS USING NONWEIGHING drying (11). In addition, lateral flow that normally occurs
LYSIMETERS under field conditions would be restricted in containerized,
column lysimeters with fine textured, clayey horizons (12).
The collection of leachate through the porous surface Fixed-tension lysimeters improve the flow patterns
of a lysimeter depends on two conditions: (a) minimal around and through the lysimeter’s porous plate surface.
discontinuities in the flow path of moving water and (b) a Stored water often moves through soil under some tension.
driving force for water movement. Maintaining a capillary Applying tension (i.e., suction or vacuum) to a porous
connection between the soil and the porous surface of the surface produces more natural flow patterns through a
lysimeter is imperative for high-quality results. Neglecting soil profile than relying on gravity to generate drainage.
to minimize potential sources of error may cause ponding, However, fixed-tension lysimeters typically experience
preferential, and bypass flow around the lysimeter (3). vacuum depletion shortly after vacuum application unless
Drainage lysimeters have been designed to operate constant vacuum is maintained. Fixed-tension lysimeters
where a variety of forces act on soil water. Drainage and do not reproduce natural soil water potential variations
solute fluxes are measured and replicated in the field by that can occur during a relatively short period of
some variation of a nonweighing pan lysimeter operating time. The transient nature of soil water potentials
under tension (i.e., tension lysimeters) or tensionless influences unsaturated hydraulic conductivity and thus
(i.e., zero-tension lysimeters). In soil or any other porous alters drainage fluxes.
material, the bulk of the water moves in response to
gradients in gravitational and/or matric potentials. Zero- RECENT USAGE OF NONWEIGHING DRAINAGE
tension lysimeters rely on the force of gravity for water LYSIMETERS
movement through the lysimeter’s porous plate surface.
The effect of a matric potential gradient is created Drainage lysimeters, both tension and tensionless, have
by applying suction to a tension lysimeter to quantify been frequently employed in numerous studies to accom-
drainage and leaching. plish a variety of research goals. Zero-tension lysime-
Both types of nonweighing lysimeters have several ters have been used to study salt and water movement
significant disadvantages. Nonweighing lysimeters are and leaching losses under conventional and no-tillage
commonly operated after the lysimeter has been filled maize [Zea mays; (13)]; collect mobile colloids transport-
with soil, similar to weighing lysimeters. As a result, soil ing organic and heavy metal compounds (11); investigate
horizonation, structure, pore size distributions, and many nitrate (NO− 3 ) and pesticide leaching from removed, but
other physical properties no longer reflect natural soil intact, soil columns (12,14); monitor the fate and cycling
conditions. Therefore, using a disturbed profile technique of 15 N on revegetated mine spoils in containerized, soil
 LYSIMETERS 489

mini-columns (7); determine the effect of irrigation on the

Equilibrium - tension lysimeter
nitrate leaching potential of agroecosystems (4); measure
nitrate leaching from maize agroecosystems under zero,
intermediate, and high fertilizer N rates (3); and elucidate
the effect of various cropping systems on solute leaching 25 cm
and 15 N migration (8,9).
Jemison and Fox (10) evaluated the collection efficiency
of zero-tension lysimeters. A bromide (Br− ) tracer method
was used to determine an overall mean collection effi- 75 cm
ciency, based on lysimeter recovery of applied Br− , of 52%,
ranging from 13–92%, and high plot-to-plot variability of
Vacuum
volumes were collected. Cumulative observed drainage, tube
uncorrected for Br− collection efficiency, measured from
zero-tension lysimeters for 3 years, was consistently lower Spring plate
than LEACHM-predicted drainage (10). LEACHM (Leach- Drain tube
ing Estimation And Chemistry Model) is a complex com-
puter model capable of simulating water infiltration and
redistribution, solute transport and distribution in the Figure 1. Stainless steel equilibrium-tension lysimeter (ETL)
for measuring drainage and solute leaching below an undisturbed
soil, and tracer ion uptake by plants (15).
soil column.
The use of fixed-tension drainage lysimeters has
been somewhat more limited. Karnok and Kucharski (16)
described the use of an underground rhizotron-lysimeter
facility constructed to monitor root growth, evapotran-
spiration, and leaching under turf grass cover. Leaching
losses from observation cells were collected under a con-
tinuous, fixed tension of −33.3 kpa. Dolan et al. (17) used
wedge-shaped, fixed-tension microlysimeters to ascertain
the effect of nitrogen placement on nitrate leaching below
maize monoculture grown in sandy soil. Adams et al. (18)
evaluated nitrate leaching from poultry litter and manure-
amended pasture soil using pan lysimeters at a fixed
tension of ∼ −10 kpa. Despite several disadvantages,
fixed-tension and tensionless nonweighing lysimeters have
many potential uses and applications.

EQUILIBRIUM-TENSION LYSIMETERS

Equilibrium-tension lysimetry is an innovative and im-

proved nonweighing drainage lysimeter technique that
allows applied-lysimeter suction to mimic the surrounding
bulk soil matric potential fluctuations; thus it minimizes
the potential for bypass flow and preserves natural
drainage patterns (19). A heat dissipation sensor (19,20)
is used to monitor the bulk-soil matric potential, and
suction applied to the lysimeters is varied as the
bulk-soil matric potential varies. Lysimeter suction was
originally set manually every several days (19), but a
recently developed automated vacuum system maintains
equilibrium between applied-lysimeter suction and bulk-
soil matric potential fluctuations within seconds (21).
The use of equilibrium-tension lysimeters is relatively
Figure 2. Backhoe digging trench for lysimeter installation.
new and even more limited compared to the other types
of nonweighing drainage lysimeters (Fig. 1). Equilibrium-
tension lysimeters have been used to quantify drainage
losses for a multiyear period in a water balance study of a
restored prairie and maize agroecosystems in southern
Wisconsin (22). At the same study site in southern agroecosystems. Equilibrium-tension lysimeters have also
Wisconsin, equilibrium-tension lysimeters have been used been used to quantify heavy metal and antibiotic leaching
to quantify multiple years of nitrogen and carbon (23,24) from the root zone of tall fescue (Festuca arundinacea)
and soluble phosphorus fluxes (25) in drainage water from amended with poultry litter in northwest Arkansas
below the root zones of a restored prairie and maize (unpublished data).
490 LYSIMETERS

GENERAL TENSION-LYSIMETER INSTALLATION

AND SAMPLING PROCEDURES

For the most representative results in a field setting,

nonweighing drainage lysimeters are typically installed
below an undisturbed soil column. A trench is dug by
hand or with a backhoe, then a small cavity that extends
into the plot or measurement area is excavated in the
long sidewall of the trench (Fig. 2). The lysimeter is forced
upward against the base of the undisturbed soil column.
Some apparatus (e.g., a spring plate, shim, or screw jack)
is installed below the lysimeter to provide the necessary
upward force to maintain the capillary connection between
the lysimeter’s porous plate surface and the soil above
(Fig. 3). Once the cavity is outfitted with the lysimeter, the
trench can either be left open and accessible or completely
backfilled. If the trench will be backfilled, tube extensions
Figure 3. Rectangular cavity excavated in sidewall of the trench
that extend upward out of the soil can be connected to the
with nearly completed lysimeter installation.
drain and vacuum tubes attached to the lysimeter. The
lysimeter cavity itself is typically left open but protected
with plywood from damage during backfilling (Fig. 4). The
trench is backfilled by adding soil in the opposite order in
which it was removed and is revegetated, if necessary, so
that only the tube extensions and heat dissipation sensor
wires if used, are visible above the soil surface (Fig. 5).
Water collected in the lysimeter’s collection reservoir
is typically removed under vacuum. The leachate volume
is recorded, and, if necessary, a portion is saved and
refrigerated for subsequent chemical analysis.

BIBLIOGRAPHY

1. Joffe, J.S. (1932). Lysimeter studies. I. Moisture percolation

through soil profiles. Soil Sci. 34: 123–143.
2. Soil Science Society of America (SSSA). (1996). Glossary of
Soil Science Terms. Soil Sci. Soc. Am., Madison, WI.
3. Jemison, J.M. and Fox, R.H. (1994). Nitrate leaching from
Figure 4. Complete lysimeter installation with cavity protected nitrogen-fertilized and manured corn measured with zero-
from damage upon backfilling with treat plywood. tension pan lysimeters. J. Environ. Qual. 23: 337–343.
4. Martin, E.C., Loudon, T.L., Ritchie, J.T., and Werner, A.
(1994). Use of drainage lysimeters to evaluate nitrogen
and irrigation management strategies to minimize nitrate
leaching in maize production. Trans. ASAE. 37: 79–83.
5. Owens, L.B. (1987). Nitrate leaching losses from monolith
lysimeters as influenced by nitrapyrin. J. Environ. Qual. 16:
34–38.
6. Kung, K.J.S. (1993). Laboratory observation of funnel flow
mechanism and its influence on solute transport. J. Environ.
Qual. 22: 91–102.
7. Reeder, J.D. (1986). A non-weighing lysimeter design for
field studies using nitrogen-15. Soil Sci. Soc. Am. J. 50:
1224–1227.
8. Shipitalo, M.J. and Edwards, W.M. (1993). Seasonal patterns
of water and chemical movement in tilled and no-till column
lysimeters. Soil Sci. Soc. Am. J. 57: 218–223.
9. Moyer, J.W., Saporito, L.S., and Janke, R.R. (1996). Design,
construction, and installation of an intact soil core lysimeter.
Agron. J. 88: 253–256.
10. Jemison, J.M. and Fox, R.H. (1992). Estimation of zero-
Figure 5. Soil surface following lysimeter installation and trench tension pan lysimeter collection efficiency. Soil Sci. 154:
backfilling. 85–94.
 STEADY-STATE FLOW AQUIFER TESTS 491

11. Thompson, M.L. and Scharf, R.L. (1994). An improved zero- groundwater recharge, seepage from canals and reser-
tension lysimeter to monitor colloid transport in soils. J. voirs, and drainage that require knowledge of the
Environ. Qual. 23: 378–383. hydraulic characteristics of aquifers. Two basic aquifer
12. Bergstrom, L. and Johansson, R. (1991). Leaching of nitrate properties needed to make a quantitative assessment of
from monolith lysimeters of different types of agricultural such applications are transmissivity and storage coeffi-
soils. J. Environ. Qual. 20: 801–807.
cient which represent, in a general sense, its abilities
13. Tyler, D.D. and Thomas, G.W. (1977). Lysimeter measure- to transmit and store water, respectively. The other
ments of nitrate and chloride losses from soil under conven-
hydraulic properties of an aquifer are hydraulic resistance,
tional and no-tillage corn. J. Environ. Qual. 6: 63–66.
leakage factor, and drainage factor. Performing an aquifer
14. Bergstrom, L. (1990). Leaching of chlorsulfuron and met-
test is an effective way of determining these hydraulic
sulfuron methyl in three Swedish soils measured in field
lysimeters. J. Environ. Qual. 19: 701–706.
properties of aquifers. Steady-state flow aquifer tests are
conducted to determine transmissivity, hydraulic conduc-
15. Jemison, J.M., Jabro, J.D., and Fox, R.H. (1994). Evaluation
of LEACHM: I. Simulation of drainage, bromide leaching, and
tivity, and hydraulic resistance; more detailed unsteady-
corn bromide uptake. Agron. J. 86: 843–851. state tests are conducted if other hydraulic properties
16. Karnok, K.J. and Kucharski, R.T. (1982). Design and con- are also to be evaluated. The objective of this article is
struction of a rhizotron-lysimeter facility at the Ohio State to discuss steady-state methods of determining aquifer
University. Agron. J. 74: 152–156. characteristics.
17. Dolan, P.W., Lowery, B., Fermanich, K.J., Wollenhaupt, N.C.,
and McSweeney, K. (1993). Nitrogen placement and leaching
in a ridge-tillage system. pp. 176–183. In: Conference Pro- AQUIFER TESTS
ceedings on Agricultural Research to Protect Water Quality.
Feb. 21–24, 1993. Minneapolis, MN. Soil Water Conserv. Soc., The procedure involves pumping water from a well for
Ankeny, IA.
a certain time and at a certain rate and measuring its
18. Adams, P.L. et al. (1994). Poultry litter and manure con- effect on the water table in the well and at a number
tributions to nitrate leaching through the vadose zone. Soil
of piezometers installed in the vicinity. When pumping
Sci. Soc. Am. J. 58: 1206–1211.
is carried out in a well, the water level in its vicinity is
19. Brye, K.R., Norman, J.M., Bundy, L.G., and Gower, S.T.
lowered; the extent of lowering at any point is termed
(1999). An equilibrium tension lysimeter for measuring
drainage through soil. Soil Sci. Soc. Am. J. 63: 536–542. the, drawdown. The drawdown is maximum at the well
point and reduces gradually with distance from the well,
20. Reece, C.F. (1996). Evaluation of a line heat dissipation
sensor for measuring soil matric potential. Soil Sci. Soc. becoming zero at a certain distance termed the radius
Am. J. 60: 1022–1028. of influence. This distance varies with the well and
21. Masarik, K.C., Norman, J.M., Brye, K.R., and Baker, J.M. depends on the well discharge, formation constants, and
(2004). An improved automated approach to equilibrium the duration of pumping. The lowering of the water level
tension lysimeters. J. Environ. Qual. In Press. in the well during pumping creates a head difference or
22. Brye, K.R., Norman, J.M., Bundy, L.G., and Gower, S.T. hydraulic gradient from the surroundings which causes
(2000). Water budget evaluation of prairie and maize the flow of water in the well. The hydraulic gradient
ecosystems. Soil Sci. Soc. Am. J. 64: 715–724. increases as flow converges toward the well. The rate of
23. Brye, K.R., Norman, J.M., Bundy, L.G., and Gower, S.T. flow is same through all the concentric cylindrical surfaces
(2001). Nitrogen and carbon leaching in agroecosystems and around the well, and the surface area decreases as one
their role in denitrification potential. J. Environ. Qual. 30: approaches the well. The form of the surface resembles a
58–70. cone, and therefore it is known as the cone of depression,
24. Brye, K.R., Gower, S.T., Norman, J.M., and Bundy, L.G. which is the characteristic feature of every pumped well.
(2002). Carbon budgets for a prairie and agroecosystems: Its size and shape depend on the discharge, length of
Effects of land use and inter-annual variability. Ecol. Appl. pumping, aquifer characteristic, and other factors.
12: 962–979.
25. Brye, K.R. et al. (2002). Phosphorus leaching under a restored
tallgrass prairie and corn agroecosystems. J. Environ. Qual. Steady-State and Unsteady-State Flows
31: 769–781.
Both steady-state and unsteady-state groundwater flow
conditions occur in nature and cannot be very distinctly
STEADY-STATE FLOW AQUIFER TESTS separated. It is a matter of interpretation and choice as to
which concept is used for practical utility. Although every
HARENDRA S. CHAUHAN phenomenon in nature is transient, it could be considered
G.B. Pant University of a series of successive steady-state events.
Agriculture and Technology Steady-state groundwater movement requires that the
Uttar Pradesh, India boundary conditions of flow and aquifer characteristics
remain unchanging with time. Even though the aquifer
characteristics do not remain constant, they are assumed
Groundwater studies are conducted to evaluate the so for practical purposes. Thus, if a hydraulic disturbance
occurrence, availability, and quality of groundwater. like well pumping is created, even though the process
There are numerous practical applications like optimum is initially unsteady, eventually it becomes steady
development of groundwater basins, safe withdrawal, with time.
492 STEADY-STATE FLOW AQUIFER TESTS

DEFINITION OF AQUIFER PROPERTIES sinks or source. The methods of determining aquifer

constants depend on the solution of the selected continuity
The transmissivity (T) of an aquifer is the rate of flow equation and the specific boundary conditions. Steady-
under a unit hydraulic gradient through a vertical strip state solutions can be used to estimate transmissivity;
of the aquifer of unit width and extending through the the storage coefficient (S) cannot be determined because
whole saturated thickness of the aquifer. It has the it does not appear in the solution. The following methods
dimensions of length2 /time and, for example, may be based on steady-state approaches can be used to determine
expressed in m2 /day. T is the product of the average the transmissivity of an aquifer.
hydraulic conductivity (k) and the saturated thickness (b)
of the aquifer. Unconfined Aquifer Flows
Hydraulic conductivity (k) is the proportionality
constant k in Darcy’s law v = ki, where v is the specific Radial Flow to a Fully Penetrating Constant Discharge
discharge or Darcy’s velocity (L T−1 ) and i is the hydraulic Well (Fig. 1). The flow is considered assuming that the
gradient (dimensionless), which is the volume of water aquifer is homogeneous, isotropic, of uniform thickness,
that will flow through a porous medium in unit time and has an infinite aerial extent. The differential equation
under a unit hydraulic gradient through a unit area for groundwater flow in an unconfined aquifer may be
measured at right angles to the direction of flow. k written in cylindrical coordinates as
is a parameter that depends on both the properties of

the porous medium and of the fluid and has units of d r dh
=0 (1)
length/time, for example, m/day. dr dr
Saturated thickness (b) is equal to the physical
thickness of the aquifer between two confining layers in a In a well that fully penetrates the gravity aquifer,
confined aquifer. The saturated thickness for unconfined the flow will be axisymmetric. Neglecting the vertical
aquifers, equal to the difference between the free water component of the velocity as small, the equipotential
table and the impervious layer, is not constant but varies surfaces may be considered coaxial cylinders. The flow
with pumping. across
these cylinders
 at a distance r is given by Q = av =
Coefficient of storage (S) of an aquifer is the volume of dh
2π rh −k . Equation (1) may be integrated twice. The
water released from or taken into storage per unit surface dr
area of the aquifer per unit change in head. In unconfined two constants may be obtained from an assumed boundary
aquifers, S is known as the specific yield which represents condition of a piezometric head at the well as h(r0 ) = h0
the volume of water released from storage per unit surface and the relationship of the well flow at a given distance r.
area of aquifer per unit decline of the water table. The solution of Eq. 1 under the boundary conditions gives
Specific yield is also sometimes called effective porosity the phreatic surface of the flow as
or drainable pore space. S is a dimensionless quantity.
Hydraulic resistance (c) is a property of semiconfined Q r
h2 = ln + h2 (2)
aquifers that characterizes the resistance of a water- πk r0
bearing layer to vertical flow. It is also known as the
reciprocal leakage coefficient and is given by b /k where Three methods may be developed from Eq. 2 to
b and k are the thickness and the hydraulic conductivity, determine transmissivity.
respectively, of a semipervious layer. The dimension of the
hydraulic resistance is time and is, for example, expressed Determination of T Using Two Observation Wells:
in days. Dupuit–Thiem Method. It can be seen from Eq. 2 that there
Leakage factor (L) describes the spatial distribution of is no fixed saturated thickness in gravity aquifers and only
leakage through one or more semipervious layers into √ a k appears in the solution describing phreatic surfaces.
leaky aquifer or vice versa. It is defined as L = kbc, Thus, only k can be obtained from the solution. To obtain
has the dimension of length, and it may be expressed the transmissivity, k has to be multiplied by some average
in meters. depth. Measuring heads at two distances r1 , r2 (r2 ≥
Of these aquifer parameters, transmissivity (T),
hydraulic conductivity (k), and hydraulic resistance (c) are
the properties that can be determined from steady-state
aquifer tests. Ground surfaceQ

Water table
STEADY-STATE METHODS OF DETERMINING s
AQUIFER PARAMETERS Drawdown
curve Unconfined aquifer
Different well flow theories have been derived based on r H
h
the consideration of a continuity equation in cylindrical or h R
spherical coordinates, well depth, the fully or partially 2r

penetrating nature of wells; steady-state or unsteady-

state flow conditions; gravity, artesian, leaky or nonleaky
aquifer; constant or variable discharge conditions; and Figure 1. Flow to an unconfined aquifer.
 STEADY-STATE FLOW AQUIFER TESTS 493

r1 ), h(r1 ) = h1 , h(r2 ) = h2 in Eq. 2, an expression for k may

Ground surface Q
be written as
Q r2 Piezometric
k= ln (3) surface
π h(h22 − h21 ) r1
s
Drawdown
Contrary to Dupuit’s assumption, the vertical compo- curve
nents of velocity are large in gravity flow and cannot be r
ignored. Equation 3, consequently, does not give correct H
R
results. Setting h = H − s, Eq. 3 can be rewritten as h 2r h

Q r2
kH = 

 ln (4) b Confined aquifer Hydraulic
s2 s2 r1 conductivity k
2π s1 − 1 − s2 − 2
2H 2H
Figure 2. Flow to a confined aquifer.
where H is the saturated thickness of the aquifer, s1 and
s2 s2
s2 are drawdowns and s1 − 1 and s2 − 2 are corrected
2H 2H Writing h = H − s, transmissibility may be written as
drawdowns s1 and s2 ,
Transmissibility may be written in terms of corrected Q R Q R
drawdowns or, neglecting correction terms, it may also be kH =
 ln ≈ ln (9)
s20 r0 2π s0 r0
expressed as 2π s0 −
2H
Q r2 Q r2
T = kH = ln ≈ ln Confined Aquifer Flows (Fig. 2)
2π(s1 − s2 ) r1 2π(s1 − s2 ) r1
Similar to the unconfined aquifer, flow is considered to take
Q r2
≈ ln (5) place assuming that the aquifer is homogeneous, isotropic,
2π(h2 − h1 ) r1 of uniform thickness, and of infinite aerial extent.

Determination of T Using a Single Observation Well. Due

Radial Flow to a Fully Penetrating, Constant Discharge
to constraints on funds for pumping tests, it may be
Well. Differential equations of the groundwater flow to a
possible to construct only one observation well. For
fully penetrating well in a confined aquifer may be written
such a situation, one may have observations only in
in cylindrical coordinates as
a single well. For such conditions, setting h(r0 ) = h0
and h(r1 ) = h1 , (r1 ≥ r0 ), an expression for k may be

d dh
written as r =0 (10)
Q r1 dr dr
k= ln (6)
π(h21 − h20 ) r0
In a fully penetrating aquifer of thickness b, the
flow is axisymmetric, and the equipotential surfaces are
In the same way for such boundary conditions, T
coaxial circular cylinders. Flow across these cylinders
may be written in terms of corrected drawdowns. Con-
in a fully penetrating artesian aquifer of thickness b
sidering that corrections have small values, they may at a given distance r from the well center is given as
be neglected to give approximate values of transmissiv-

dh
ity as Q = av = 2π rb −k . Equation 10 may be integrated
dr
twice under the assumed condition of the piezometric head
Q r1 Q at the well, h(r0 ) = h0 , and the relationship of the well at
kH = 

 ln ≈
s20 s21 r0 2π(s0 − s1 ) a given distance r to give
2π s0 − − s1 −
2H 2H
Q r
r1 Q r1 h= ln + h0 (11)
ln ≈ ln (7) 2π kb r0
r0 2π(h1 − h0 ) r0
Thiem’s (1906) Equilibrium Equation Using Two Observa-
Determination of T Using the Head in the Pumped tion Wells. To reduce errors because of losses in the well
Well. Due to further constraints on funds, one may not and through the well screen, Thiem (1) proposed mea-
be able to make even a single observation well. For such a suring piezometric heads h1 , h2 in two observation wells
condition, measuring the head in the pumped well itself, located at distances r1 , r2 from a constantly discharging
h(r0 ) = h0 , and considering an arbitrary radius of influence well Q. Thus, setting h(r1 ) = h1 , h(r2 ) = h2 in Eq. 11 of an
h(R) = H, the hydraulic conductivity may be obtained as expression of transmissivity may be derived as

Q R Q r2
k= ln (8) T = kb = ln (12)
π(H2 − h20 ) r0 2π(h2 − h1 ) r1
494 STEADY-STATE FLOW AQUIFER TESTS

Determination of T Using a Single Observation Well. Con- Considering heads at two observation wells located
sidering the head as h(r1 ) = h1 at a single piezometer at distances r1 , r2 as h1 , h2 , an expression for k may be
installed at distance r1 from the well, T may be obtained obtained as

from Eq. 11 as 1 1
Q −
r1 r2
k= (17)
Q r1 2π(h2 − h1 )
T = kb = ln (13)
2π(h1 − h0 ) r0
Determination of k Considering the Head in the Well.
Using the measurement of the head in the pumped well
Determination of T Using the Piezometric Head in the itself, h(r0 ) = h0 , and considering an arbitrary radius of
Pumped Well. By measuring the piezometric head in the influence R, an expression for k in terms of the head and
pumped well itself and at an arbitrary radius of influence drawdown s = H − h may be written as
R, h(R) = H, T can be obtained from Eq. 11 as



1 1 1 1
Q R Q R Q − Q −
T = kb = ln = ln r0 R r0 R
2π(H − h0 ) r0 2π s0 r0
(14) k= = (18)
2π(H − h0 ) 2π s0

Spherical Flow to a Nonpenetrating, Constant Discharging Considering that R > r0 , 1/R is negligible and can be
Well in An Extensively Thick Artesian Aquifer (Fig. 3). The neglected; this gives
groundwater flow equation for such a flow system is
described by a continuity equation in spherical coordi- Q Q
k≈ ≈ (19)
nates as 2π r0 (H − h0 ) 2π r0 s0


d 2 dh
r =0 (15) Semiconfined Aquifer Flows
dr dr
Aquifer Constants from the Steady-State Flow of a Well
The isopiestic surfaces to a nonpenetrating well consist in a Semiconfined Aquifer: Hantush and Jacob method (2)
of concentric hemispheres. The flow across a hemispherical (Fig. 4). The solution is based on the assumption that the
surface at distance r is given by phreatic surface remains constant, so that the leakage
through the semiconfining layer takes place in proportion
Q = av = 2π r2 (−kdh/dr) to the drawdown of the piezometric level. Hantush and
Jacob (2) obtained a solution for L > 3D, as
Considering the piezometric head at the well, h(r0 ) = Q r
h0 , and the flow equation above, the piezometric surface sm = K0 (20)
2π kb L
can be obtained as

 where sm = steady-state drawdown in meters at
1 1
Q − distance r;
r0 r Q = discharge of the pumped well in cubic
h= + h0 (16)
4π k meters per day;
√
L = kbc, leakage factor in meters;
Determination of k Using Two Observation Wells. It may b
be observed that from Eq. 16, only hydraulic conductivity c =  , hydraulic resistance of a semipervious
k
can be determined as the thickness of the aquifer is layer in meters; and
not finite. K0 (x) = modified Bessel’s function of the first kind
and zero order.
Hantush observed that if r/L is <0.05, Eq. 20 can be
Ground surfaceQ approximated as
Piezometric

2.3Q L
surface sm = log 1.12 (21)
2π kb r
s
Drawdown
curve A plot of sm against r on semilog paper, with r on the
r H logarithmic scale, will show a straight line relationship in
h the range where r/L is small. The slope of the straight line,
h R
2r the drawdown difference per log cycle of r, is expressed as

r 2.3Q
r Thick confined aquifer
sm =
2π kb
2.3Q
Figure 3. Spherical flow to a thick confined aquifer. kb = (22)
2π sm
 STEADY-STATE FLOW AQUIFER TESTS 495

Approximate Methods
Ground surface Q
Initial water Logan’s (3) Method for confined aquifers. Equation 14
table using the observation in the pumped well itself is also
s known as Logan’s method. According to Logan (3), the
Drawdown accuracy of the calculation depends on the accuracy in
curve Ponded water measuring the maximum drawdown in the well s0 and on
r the accuracy of the ratio R/r0 . The ratio R/r0 may have a
H
R large value, but the ratio of the logarithm may be small. He
h h Piezometric head suggested an approximate value of 3.33 for the log ratio.
2r in main aquifer
Putting this value in Eq. 14 gives the transmissivity as
Semi confining Leak- Hydraulic
b′ 2.3Q R 1.22Q
stratum age conductivity k ′
kb = log ≈ (24)
2π s0 r0 s0

b Main aquifer Hydraulic

conductivity k Logan’s (3) Method for Unconfined Aquifers. Using a
similar approach and considering observation in the
pumped well itself, the transmissivity in an unconfined
Figure 4. Flow to a semiconfined aquifer. aquifer may be obtained from Eq. 9 as

2, 3Q R 1.22Q
For any point sm , r is selected on the straight line. kH = log ≈ (25)
2π s0 r0 s0
Putting r in Eq. 21 gives L.c can be obtained as

b L2 s20
c= = where s0 ≡ s0 − .
k kb 2H

Alternatively the extended straight line portion of the Transmissivity from Other Analytical Solutions
curve intercepts the r axis where the drawdown is zero. At
the interception point, sm = 0, r = ro , and Eq. 21 reduces to In principle, any analytical solution of a boundary value

 problem of groundwater flow can be used to estimate
L aquifer parameters, if other variables can be measured
log 1.12 =0
r or estimated. A few steady-state solutions for estimating
transmissivity are presented.
or √
L kbc
1.12 = 1.12 =1
r0 r0 Two Equal, Constant Discharge, Mutually Interfering
Wells. The relationships between two equal and constant
or discharge interfering wells distance B apart may be
(r0 /1.12)2
c= written as
kb 2π kb(ha − h0 )
Q1 = Q2 = (26)
Transmissivity from Steady Flow in a Semiconfined Aquifer: R2
ln
Ernst’s Modification of the Thiem Method. The relationship r0 B
is based on assuming that the well discharge is corrected
by subtracting, from the total discharge, the amount of where r0 = well radius, ha = average piezometric head at
phreatic water obtained from the semipervious layer. the external boundary, h0 = piezometric head at the well,
The determination requires two piezometers installed and R = radius of influence (R > B).
in semipervious layers at distances r1 , r2 . The coaxial Thus if the values of the constant equal discharges.
phreatic water from a circular area may be written Q, h0 , ha , R, and B are known, the transmissivity may be
approximately as r0 = r1 + 1/3(r2 − r1 ). determined from the relationship
The phreatic water delivered from a circular area is
R2
Q = π r20 s S × 24 Q ln
r0 B
kb = (27)

2π(ha − h0 )
where s = average drop of phreatic level in the
semipervious layer during the last hour (m) and S =
Steady-State Flow to a Partially Penetrating Well. A
specific yield of the semiconfined aquifer (to be estimated
relationship was obtained for such a flow system by De
from well logs). The transmissivity may be found from
Glee (4) as
r2
(Q − Q ) ln

r1 Qp 2 π hs 0.20
kb = (23) h2b − h0 = ln +
2π(s1 − s2 ) 4π k hs 2r0 b
496 STEADY-STATE FLOW AQUIFER TESTS

where hs is the depth of the partially penetrating well, b the average hydraulic conductivity, transmissibility, and
is the aquifer thickness, h2b is the initial piezometric head coefficient of storage of the aquifer. If two or more
at distance 2b, and h0 is the piezometric head in the well. observation wells are placed at different distances, the test
data can be analyzed both in terms of the time–drawdown
hs and the distance–drawdown relationships. It is always
Also 1.3 hs ≤ b; ≥5
2rw better to have as many observation wells as conditions may
allow. A minimum of three placed at varying distances and
From this relationship and measuring other parame- radially at equal angles around the well is desirable.
ters, the transmissivity may be found as
  Measurement of Water Levels. The depth to water may
Qp b 2 π hs 0.20
kb = ln + (28) be measured many times during the course of a pumping
4π(h2b − h0 ) hs 2r0 b test. Observations should be taken at close intervals
during the first 2 hours, and the time between readings
For a partially penetrating well in an unconfined should gradually increase as the test continues. The water
aquifer where the drawdown is also small in relation to level measurement may be recorded to the nearest 0.5 cm
the saturated thickness, Eq. 28 can be used by replacing b in the observation wells. Measurements may be made
by H. every half-minute during the first 5 minutes after starting
Kozeny (5) proposed a relationship, by which transmis- the pump, then every 5 minutes for an hour, then every 20
sivity may be determined from minutes for about two hours. From this point on, readings
r at hourly intervals should be sufficient.
2.3Qp log Water-level measurements can be taken in various
r0
T= (29) ways, the wetted-tape method, mechanical sounder, elec-
2π(h − h0 ) G
tric water-level indicator, pressure gauge, or pressure
logger. For detailed information on these devices, see
where
 
r0 πα Kruisman and De Ridder (7). Fairly accurate measure-
G=α 1+7 cos ments of water levels can be made manually.
2hs 2

Qp is the discharge of a partially penetrating well, Qf Measurements of Discharges. The other parameter that
is the discharge of a fully penetrating well, G = Qp /Qf is needs to be measured is the discharge from wells during
a correction factor giving the ratio of the two discharges, pumping tests. To avoid complicated calculations later,
α is the fractional penetration of the well screen hs /b, and the discharge rate should preferably be kept constant
hs is the depth of penetration of the aquifer. throughout the test. The discharge rate can be measured
using various devices like commercial water meters,
flumes, orifice weirs, containers or the jet-stream method.
PRACTICAL ASPECTS OF PUMPING TESTS (after Ref. 6)
The procedures for installation and measurements using
these devices are available in standard texts.
Observation Well Size, Location, and Numbers
An observation well should be just large enough to allow
SUMMARY
accurate and rapid measurement of water levels. Small
wells provide a fast response. Five-cm observation wells
An overview of steady-state approaches for determining
are usually effective when hand methods of measuring
the transmissivity of confined, unconfined, and semicon-
water levels are used. Larger pipes may be used with
fined aquifers has been presented. Detailed procedures
automatic recorders. Observation wells are commonly
for determining transmissivity using drawdown data in
installed with 1–2 m long screens at the bottom embedded
a pumped well and one or two observation wells are
in the aquifer to measure the pressure head. Observation
discussed in the context of two most common methods of
wells should be installed at about the same depth as the
Thiem and Dupuit–Thiem. A few approximate methods
middle of the well screen in the pumped well.
and analytical solutions for determining aquifer con-
Observation wells should be located at a reasonably
stants have been briefly introduced. General guidelines for
large distance and not too close to the pumped well.
conducting pumping tests including measurement of dis-
Setting observation wells at distances of 30–100 m from
charge and groundwater levels and numbers and locations
the pumped well may work out best in most cases.
of observation wells have been furnished.
When observation wells are too close to the pumped well,
correct drawdown readings may be difficult. The cone of
depression in a water table spreads at a slower rate than BIBLIOGRAPHY
in an artesian formation, so the observation wells for
a water table aquifer should be relatively closer to the 1. Thiem, G. (1906). Hydrologische Methoden. Leipzig, p. 56.
pumped well. 2. Hantush, M.S. and Jacob, C.E. (1955). Non steady radial flow
The number of observation wells depend upon the in an infinite leaky aquifer. Trans. Amer. Geophysical Union.
amount of information required and the funds available 36: 95–100.
for the test program. Drawdown data from the well 3. Logan, J. (1064). Estimating Transmissibility from Routine
itself or from a single observation well permit estimating Production of Tests of Water Wells Ground Water 2(1): 35–37.
 TIDAL EFFICIENCY 497

4. De Glee, G.J. (1930). Over Grundwaterstroomingen bij aquifer, which is similar to the effect that trains have as
Wateronttrekking Door Middle van Patten. J. Waltman Jr., they pass a well in some confined aquifers—water levels
Delft, the Netherlands, p. 175. rise and then fall as a train passes by (3).
5. Kozeny, J. (1933). Theorie und Berech nug der Brunneu Wasser The loading effect is greater in poorly consolidated
Kraft and Wasser Wartschift. 28: 88–92,113–116. aquifers where a greater proportion of the overlying load
6. Johnson Div. UDP, Inc. (1972). Ground Water and Wells. 2nd is borne by the fluid. Pore fluids in consolidated aquifers
Edn. Edward E. Johnson, St. Paul, MN, p. 440. are less likely to carry the load, and so these aquifers have
7. Kruisman, G.P. and De Ridder, N.A. (1970). Analysis and lower tidal efficiencies.
Evaluation of Pumping Test Data, International Institute Coastal aquifers are not the only type of aquifer that
for Land Reclamation and Improvement. Wageningen, the
shows a response to surface loads. The loading effect
Netherlands.
on confined aquifers can be observed in response to
the placement or removal of any type of mass on the
TIDAL EFFICIENCY ground surface. Components of the water budget (such as
precipitation, evaporatranspiration, and soil moisture) can
TODD RASMUSSEN immediately affect water levels in confined aquifers—even
The University of Georgia when no direct hydraulic communication is present (4).
Athens, Georgia An increase in water levels in a confined aquifer
under a waste lagoon may be mistakenly attributed to
contamination from the lagoon. The cause could be from
Ocean tides often cause water levels in coastal aquifers downward leakage of water from above, or it could also be
to fluctuate. These fluctuations result from two different from the increased surface load causing a rapid increase
causes. In confined aquifers, the additional weight of the in fluid pressure.
water on the surface increases the pressure on the water Another example of surface loading develops when
at depth, which causes an increase in water levels in barometric pressure changes cause water levels in wells
observation wells located near the coast. In this case, no to fluctuate. In these cases, the response is somewhat
movement of water between the ocean and the aquifer complicated because the barometric pressure also exerts a
may have occurred. In unconfined aquifers, however, the force on water levels in open wells. The tidal efficiency in
loading effect does not occur. Instead, water that has confined aquifers is related to the barometric efficiency,
moved directly through the pores within the aquifer BE, in that the two efficiencies must sum to unity,
has caused the water level response. Regardless of the TE + BE = 1. Thus, aquifers with low tidal efficiencies
mechanism, both processes can be described with the have large barometric efficiencies, and vice versa. The
phase lag, or delay, between the tidal peak and the aquifer reader is referred to BAROMETRIC EFFICIENCY for additional
response peak, as well as the amplitude ratio between the discussion of this topic.
magnitudes of the response and the tides
Water levels in wells located near the coast often
fluctuate with a rhythm that mimics the rise and fall UNCONFINED AQUIFERS
of the ocean. The tidal efficiency, TE, is the ratio of the
change in water levels in wells, W, to the change in tidal A second tidal influence occurs in unconfined, surficial
levels, T: coastal aquifers. In these aquifers, water levels change
W as the hydraulic perturbation created by changing ocean
TE =
T levels moves horizontally through the saturated zone
toward the well. Unlike the confined case in which the
where both water levels are measured with the same type
load is transmitted without concomitant fluid movement,
of units (e.g., feet, meters, psi). Tidal efficiencies can range
water is free to move through the unconfined aquifer.
from a maximum of one in poorly consolidated geologic
For a simple, one-dimensional flow geometry (i.e., a
media to a minimum of zero in rigid aquifers (1,2).
well placed at some distance from a linear shoreline), the
The tidal response in an aquifer can result from two
response is a function of the distance of the well from
types of phenomena, one in confined and a second in
the shoreline, L, and the aquifer hydraulic diffusivity,
unconfined aquifers. As described in greater detail below,
D = T/S, where T and S are the aquifer transmissivity and
the confined aquifer response is caused by the increased
storativity, repsectively. Nearby wells in more diffusive
load placed on the surface, whereas the unconfined aquifer
aquifers respond more rapidly than do more distant wells
response is caused by the movement of water in response
in aquifers with lower diffusivity. The response is a
to changing tidal water levels.
function of a dimensionless variable, κ = Dτ/L2 , where
τ is the response time.
CONFINED AQUIFERS This type of response is not limited to coastal areas,
however. A similar type of response is found in surficial
In confined aquifers, coastal tides affect water levels aquifers next to streams. In this case, changing water
because the surface load changes as ocean levels fluctuate. levels during floods cause a loss of water from the stream
The weight of the ocean compresses the underlying aquifer, into the adjacent aquifer, with a reversal of flow later in
which causes a change in the fluid pressure. In this the flood hydrograph. This type of response is also referred
situation, no water moves between the ocean and the to as bank storage (5).
498 COMBINED FREE AND POROUS FLOW IN THE SUBSURFACE

SINUSOIDAL OSCILLATIONS and petroleum engineering problems. The flow behavior in

the fluid layer is determined by the properties of the fluid
Many types of loading phenomena display periodic (e.g., viscosity and density), and surface (e.g., pressure
variation, tides being a common example. In addition to the and shear) and body (e.g., gravity and electric field) forces.
gravitational affects of the sun and moon on the oceans, However, flow in the porous medium depends additionally
gravitational forces also cause changes in atmospheric on the properties of the medium which include, among
pressure, called atmospheric tides, and even in the solid others, porosity and permeability.
earth, called earth tides. In fluid dynamic analysis of groundwater systems, it
These periodic variations can be characterized with the is commonly found that water flows through combined
period of the oscillation (e.g., 24 hours for solar tides) free (nonporous) and porous pathways between surface
and the amplitude, or magnitude, of the oscillation. When and subsurface systems. Such coupled regimes may also
observing the tidal response in confined aquifers, the tidal be present due to construction of industrial utilities and
efficiency is equal to the ratio of the magnitude of the structures in the ground. Important engineering processes
water level fluctuation in the well to the magnitude of the that may involve combined free and porous flow domains
ocean tides. For example, if peak daily ocean levels vary are, for example, fluid losses/leaks from underground pipes
by 1 m and the peak daily groundwater levels vary by 10 and storage tanks in old gas works sites; dig-and-treat,
cm, then the tidal efficiency is 0.10, or 10%. pump-and-treat, and permeable reactive barrier technolo-
Another influence is the phase lag, which is the delay gies for groundwater treatment; and drilling/extraction
between when the ocean tide and the aquifer water levels of oil from underground reservoirs. However, in most
peak. Normally, an insignificant delay occurs between the cases, the associated transport phenomena are deter-
ocean tide and a confined aquifer response; yet this delay mined by natural hydroenvironmental conditions, which
can be large in cases in which significant borehole storage include, for example, combined surface and subsurface
exists. flow, lake–groundwater interactions, seepage through
The tidal efficiency and phase lag in unconfined aquifers preferential flow channels and macropores, circulatory
can determine the aquifer diffusivity. The transmissivity flows and rise and fall of groundwater, glaciology, and
and storativity can be only determined, however, if the flow multiple karstic regions interfaced with granular porous
of water can be determined, which requires measurements sections. The subdomains in coupled flow systems are
of the flow rate between the ocean (or stream) and the
generally distinguished by an interfacial surface, which
aquifer.
represents the transitional zone for contaminants/fluid
mobility from free to porous sections or vice versa. To
BIBLIOGRAPHY gain on understanding of the combined flow behavior, it is
necessary to develop realistic methodologies for evaluat-
1. Bear, J. (1972). Dynamics of Fluids in Porous Media. Dover ing mass and momentum transfer across the free/porous
Publications, New York.
flow interfaces.
2. Walton, W.C. (1970). Groundwater Resource Evaluation.
The construction of well-posed mathematical formu-
McGraw-Hill, New York.
lations for combining different flow regimes can present
3. Jacob, C.E. (1940). On the flow of water in an elastic artesian
difficulties. Two general approaches have evolved for this
aquifer. Am. Geophys. Union Trans. 21: 574–586.
purpose: (1) methodologies based on treating the combined
4. vanderKamp, G. and Schmidt, R. (1997). Monitoring of total
flow domain as a single continuum and (2) formulations
soil moisture on a scale of hectares using groundwater
piezometers. Geophys. Res. Lett. 24(6): 719–722.
in which the flow domain is considered as a combi-
nation of discrete sections corresponding to different
5. Todd, D.K. (1959). Ground Water Hydrology. John Wiley &
Sons, New York.
flow regimes. In the former, the porous section is effec-
tively regarded as a pseudofluid layer, and the whole
domain is analyzed as a whole. As such, a suitably for-
COMBINED FREE AND POROUS FLOW IN THE mulated momentum equation for the entire domain in
conjunction with mass conservation equations is solved.
SUBSURFACE The mathematical model identifies fluid and porous
DIGANTA BHUSAN DAS domains based on spatially varying permeability. Such
Oxford University
a single-domain approach is usually applied where the
Oxford, United Kingdom flow transition from free to porous flow zones is not
distinct and properties of the porous domain change
VAHID NASSEHI progressively (e.g., metal solidification problems involv-
Loughborough University ing mushy zones). However, if the permeability of the
Loughborough, United porous medium is constant and remains above a rel-
Kingdom atively high level, then at defined interfaces between
adjacent zones, transitions between free/porous regimes
occur. Under these conditions, the domain should be
INTRODUCTION viewed as a combination of adjacent flow fields rather
than a geometric continuum. In these situations, the sec-
The presence of a fluid layer adjacent to a porous medium ond approach should be used where appropriate equations
is common in many environmental, chemical, mechanical of motion describing the flow in different subdomains are
 COMBINED FREE AND POROUS FLOW IN THE SUBSURFACE 499

solved. In general, a groundwater system corresponds to In these equations, the subscript p refers to the porous
this case. section; P is the interstitial averaged pressure, v is the
volume averaged seepage velocity; and ρ and µ are the
constant density and viscosity of the fluid, respectively.
GOVERNING MODEL EQUATION
µ/ is the effective viscosity which depends on the fluid
viscosity, µ, and the porosity. For flow of an incompressible
The main aim of this article is to present a general
fluid in a continuum porous medium, the equation of
methodology that can be used to simulate hydrodynamic
continuity is
behavior in combined free (nonporous sections) and porous
∇ · vp = 0 (5)
flow zones in the subsurface based on the assumption of
discrete flow domains, which involves selecting governing
model equations for each subdomain and the matching where v is the volume averaged seepage velocity of fluid.
conditions at the interface for linking them. Although
fluid mobility in combined free and porous flow on the Matching Boundary Condition for the Darcy Equation
surface has been the focus of intense research for more as a Porous Flow Equation. To couple the equations
than two decades, analyses of subsurface flow have representing the two domains, it is necessary to specify
not treated the possible interactions between free and the matching hydrodynamic conditions at the interfacial
porous flow regimes efficiently. Addressing these issues, region separating the two subdomains. Keeping the
the hydrodynamic conditions in underground zones of three-dimensional (3-D) subsurface problem in mind,
coupled fluid and porous domains have been analyzed on the matching conditions for pressure (continuous or
a 3-D spatial scale by Das et al. (1). This methodology discontinuous) and velocity components (normal and
is based on the principles of mass and momentum tangential) must be specified at the interface. The
conservation and primitive variable formulations (i.e., the Darcy equation is a first-order differential equation with
equations are expressed in terms of the pressure and respect to velocity and pressure, so it is not possible to
velocity). Conservative forms of the equations have also describe the velocity components fully at the free/porous
been adopted to enhance the accuracy of the numerical domain interface, which makes it impossible to link
calculations. The model equations are presented below. the Darcy and the N-S equations directly because the
latter requires specifying all velocity components to solve
Free Flow Region them. Special techniques have been invoked, henceforth,
The momentum transfer for constant density and viscosity for this purpose. Addressing this issue, Beavers and
of a fluid at a reference temperature in a free flow region Joseph (4) performed their well-known experiment on the
can be represented by the well known Navier–Stokes measurement of mass efflux of a Poiseulle fluid flowing
(N-S) equation, over a permeable block. They hypothesized the existence
of a slip velocity, commonly known as the Darcy slip
dvf phenomenon, at the interface and a shear velocity in the
∇Pf = −ρ + µ∇ 2 vf (1) porous boundary layer. Consequently, an ad hoc boundary
dt
condition was proposed based on the assumption that the
where the subscript f refers to the free flowing fluid. P and slip velocity is proportional to the shear effects. It allowed
v are the pressure and velocity terms, and ρ and µ are the jumps in the tangential velocities and gave a much-
constant density and viscosity of the fluid, respectively. In needed relationship between the flow within and outside
Eq. 1, the body forces are automatically incorporated in porous media. The Beavers and Joseph (4) formulation
the pressure term as the resultant flow potential. The was represented as
following continuity equation that describes the mass

conservation for an incompressible fluid can be used for a ∂u γ
= √ (uf − up ) (6)
free flow regime: ∂y f K
∇ · vf = 0 (2)
where the x direction was taken as parallel and the
Porous Flow Region y direction as perpendicular to the interface. K is
The corresponding macroscopic equations representing the direction-independent geometric permeability of the
the flow field in the porous section are more uncertain. porous medium (isotropic), and γ is a slip coefficient that
The momentum equations have received considerable depends only on the properties of the porous medium and is
attention in the past. The two most suitable are the independent of the fluid. The Beavers and Joseph condition
Darcy (2) and Brinkman (3) equations shown below in has now been modified for cases where the flow and the
the transient form for anisotropic porous media where the interface are perpendicular rather than parallel, as in
body forces have been incorporated in the pressure terms the original formulation. Jones (5) derived the following
to represent the driving flow potential: expression for the interfacial condition, assuming that
the Darcy slip phenomenon is due to shear stress at the
∂vp interface rather than a simple shear velocity:
Darcy equation: ∇Pp = −ρ − µK −1 žvp (3)
∂t

∂vp ∂v ∂u γ
Brinkman equation: ∇Pp = −ρ − µK −1 žvp µ ∇ 2 vp (4) + = √ (uf − up ) (7)
∂t ∂x ∂y f K
500 COMBINED FREE AND POROUS FLOW IN THE SUBSURFACE

z This interfacial condition for pressure jump arises from

Interface
the fact that stress in the porous side is carried by the
porous materials, whereas in the free flow domain, the
stress is borne by the fluid medium, which causes a
discontinuity in stress, and hence fluid pressure, across
y the fluid/porous interface.
w
Free Porous
flow flow v Matching Boundary Condition for the Brinkman Equation
as a Porous Flow Equation. The use of the Brinkman
u
equation for mathematically describing the flow in
x porous flow zones apparently does not cause any
Figure 1. Model of combined free and porous flow domain for mathematical difficulty. The second-order derivatives of
extending the Beavers and Joseph (1967) formulation to the 3-D the N-S equations and the Brinkman equation with
subsurface region. respect to velocity enable specifying all flow properties
as continuous across the interface, which facilitates
the task of coupling the equations of motion for the
Direct extension of the above formulations to 3-D free and permeable domains. The continuity of the
cases poses further challenges. However, by resolving flow variables across the interface yields the following
the 3-D framework into two planes (x–y and x–z), it matching boundary conditions:
is possible to use the Beavers and Joseph formulation
for three-dimensional modeling of combined flow in the uf = up (13a)
subsurface (1,2). The extension can be done based on
Fig. 1 where the x component of the velocity is normal vf = vp (13b)
to the interfacial plane and the y and z components wf = wp (13c)
are tangential (parallel). For the 3-D situation in Fig. 1,
postulating that the Darcy slip phenomenon takes place at Pf = Pp (13d)
both x–y and x–z planes, the following relationships can be
obtained heuristically based on the Jones (5) modification
of Beavers and Joseph’s original formulation (1,2). The DISCUSSION AND CONCLUSIONS
velocity components that are parallel to the interface can
be calculated by using these expressions: Although there are superficial similarities in mathemat-

 ical formulations of different coupled free and porous
∂v ∂u γ flow problems, Das et al. (1) clearly demonstrated that
Velocity slip at x–y plane: + =  (vf − vp )
∂x ∂y f Kyy combined flow behavior in the subsurface may differ sig-
(8) nificantly from other combined flow processes encountered


∂w ∂u γ in chemical, petroleum and mechanical engineering appli-
Velocity slip at x–z plane: + = √ (wf − wp ) cations. Therefore, models to represent specific scenarios
∂x ∂z f Kzz
(9) need to be developed. In the above studies by Das et al.,
natural flow behavior has been defined at the exit of the
The slip coefficient is represented in Eqs. 8 and 9 by γyz domain by imposing stress-free boundary conditions. Das
to indicate the plane on which it is active. Equation 8 et al. demonstrated that at the open end (exit) of the
is used to simulate the jump in the transverse (y) domain, coexisting inflow and outflow might take place. In
velocity component across the interface, and Eq. 9 is other words, fluid from the surrounding media might come
used for the vertical (z) velocity component. Kyy and inside the domain through the open end in addition to the
Kzz are the permeability in the y and z directions, defined inlet. For the particular domain investigated, Das
respectively. The above equations, therefore, represent et al. observed that although flow in the fluid domain is
the slip-flow phenomena at the permeable interface mostly unidirectional toward the free/porous interface and
with respect to directional geometric permeability. Also, appears similar to pipe flow, it might become circulatory
the velocity component normal to the interface, the on the porous side and reverse its direction. Evidence of
longitudinal component, can be taken as continuous across flow circulation, consistent with the simulated results of
the interface: Das et al., can also be found in the literature dealing with
uf = up (10) various scenarios of subsurface flow (6). However, fluid
circulation is a complex transient process often involving
However, the pressure does not necessarily remain a moving center of circulation and shifts of the flow rever-
continuous across the interface and only under static sal front to or from the interface. After a certain time
conditions, can it be said to be strictly continuous. If interval, the circulation may disappear, and the porous
the pressure is to be modeled as discontinuous at the y–z flow may become unidirectional toward the exit of the
interfacial plane during flow, the following relationship porous section. The main cause of such flow behavior has
is used:
 been attributed to complex/variable pressure distributions
∂u within the porous domain. The effects of a partially imper-
−P + 2µ = −Pp (12)
∂x f meable interface between the free and porous domains has
 GROUNDWATER TRACING 501

been explored by Das and Nassehi (7). It has been demon- READING LIST
strated that due to the impermeable interface, there is a
possibility of buildup of fluid in the free flow region. The Das, D.B. (2001). Computer Modelling of Contaminant Mobility
influence of having a the multiple number of fluid/porous in Subsurface Regions. Ph.D. Thesis, Chemical Engineering
interfaces in a flow domain has been demonstrated by Department, Loughborough University, Loughborough, UK.
Das and Nassehi (8). This study showed that, in gen-
eral, the presence of a multiple number of interfaces
in the free flow domain has a minimal effect on flow GROUNDWATER TRACING
behavior. However, the flow patterns on the porous side
may vary depending on the direction of the free/porous HAMID JAHANI R.
interfaces. Das and Nassehi (8) also examined concentra- Water Research Institute
tion profiles of a conservative pollutant in associated free Hakimieh, Tehran, Iran
flow and porous domains. The effects of different aspect
ratios of the subdomains on the hydrodynamic behav-
ior of free/porous regimes have been investigated by Das INTRODUCTION
and Nassehi (9). This study shows the influence of the
extent of free and porous domains, which govern the entire Groundwater tracing is now a major field in groundwater
flow behavior. hydrology aiming mainly at demonstrating the direction
The study of combined free and porous flow in of groundwater flow and secondly, at giving as much
the subsurface using different methodologies is still information on the hydrologic setting as possible. It has
relatively new in scientific literature. Although a number been used for several decades to study flow conditions,
of issues have been studied, future research aimed at for groundwater reconnaissance, and for managing water
better understanding of combined flow behavior must be resources. During the last decades, water protection has
continued. For example, investigations into the effects of been one of the main issues due to strong pollution of the
heterogeneity in porous media and multiphase flow (e.g., environment. A groundwater tracer is a substance or a sort
oil and water) in combined domains should be carried out. of energy conveyed by water to be detected at a sampling
A general methodology, which is not restricted to systems point and to be used in analyzing the temporal and
where free/porous interfaces are aligned to coordinate spatial distribution of water and the pollutant. Artificial
axes, also needs to be developed. and natural tracers can help in various ways to assess
and to predict contaminant transport in surface and
groundwater.
BIBLIOGRAPHY Groundwater tracing can be a component of a large
hydrogeological program, or it can be a hydrogeological
1. Das, D.B., Nassehi, V., and Wakeman, R.J. (2002). A finite program by itself. Nowadays, new field methods, new trac-
volume model for the hydrodynamics of combined free and ers, and new measurement techniques are being developed
porous flow in sub-surface regions. Adv. Environ. Res. 7(1): to make hydrogeologists more satisfied with results.
35–58. The advantage of groundwater tracing is that it
2. Darcy, H.P.G. (1856). Les Fontaines Publiques de la Ville de provides the possibility of a direct view of a groundwater
Dijon, Exposition et Application des Principes à Suivre et des system, so the results can be relied on with assurance. In a
Furmules à Employer dans les Questions de Distribution d’Eau. groundwater tracing program, depending on the question
Victor Dalmont, Paris.
and technical problem, both ‘‘natural’’ (those existing in
3. Brinkman, H.C. (1947). A calculation of the viscous force the water cycle) and ‘‘artificial’’ (those to be introduced
exerted by a flowing fluid on a dense swarm of particles.
into groundwater system) tracers can be selected. The
Appl. Sci. Res. A1: 27–34.
main group of artificial tracers, is soluble tracers, such
4. Beavers, G.S. and Joseph, D.D. (1967). Boundary conditions at
as fluorescent dyes that are highly detectable, as well as
naturally permeable wall. J. Fluid Mech. 30(1): 197–207.
different salts. Another group of artificial tracers includes
5. Jones, I.P. (1973). Low Reynolds number flow past a porous
drift bodies, such as spores, bacteria, or phages. The
spherical shell. Proc. Cambridge Philos. Soc. 73: 231–
applicability of each tracer depends on its properties
238.
and on the condition of the flow environment. The use
6. Devito, K.J., Waddington, J.M., and Branfireun, B.A. (1997).
of certain tracers, especially for groundwater flow, is
Flow reversal in peatlands influenced by local groundwater
systems. Hydrological Processes 11: 103–110. strongly restricted due to hydrogeologic conditions and
aquifer properties.
7. Das, D.B. and Nassehi, V. (2001). LANDFLOW: A finite
volume model of combined free and porous flow of water in For artificial tracers, the methodology comprises
contaminated land sites. Water Sci. Technol. 43(7): 55–64. introducing the tracer into the groundwater system,
8. Das, D.B. and Nassehi, V. (2003). Modelling of contaminant letting it travel together with the water, sampling, and
mobility in underground domains with multiple free/porous finally analyzing the water samples. For natural tracers,
interfaces. Water Resour. Res. 39(3), doi: 10. 1029/2002 the natural physical, chemical, or biological composition
WR001506. of water is traced. For example, the natural temperature
9. Das, D.B. and Nassehi, V. (2002). Development of a new math- of cold water from a spring could be a guide to identifying
ematical model for subsurface water quality management. snowmelt as the main water source and differentiating it
Water Sci. Technol. 45(9): 301–307. from local groundwater.
502 GROUNDWATER TRACING

THE HISTORICAL BACKGROUND The tracer

The groundwater tracing method is as old as the interest of 1. should be detectable in low concentrations. In
humans in nature. It was used while people were curious most cases, the tracer material is to be diluted
about finding the source of undiscovered or unknown caves in a big volume of groundwater storage, so it
or springs and used chaff, wood bits, or other drifts to must be easily detectable at low concentrations
reveal the route of the water. The way to perform the test in water samples. The tracer concentration should
is now almost as simple as it was in the past, at least be easily differentiated from background values in
in most cases. But the way to analyze samples and the sampled water.
results evolved into complicated analytical methods. The
2. should be applicable as a nontoxic additive to water.
development of new tracers and new analytical methods
It is compulsory for the tracer material not to harm
during the last 10 years has led to more accuracy in the
flora and fauna.
results and also more ability to cover longer distances
in tracing tests. While using the method during the last 3. should not be an absorptive material. The tracer
decades, it was aimed to find the direction of groundwater substance should not be easily adsorbed to aquifer
flow and its velocity; now, its major application is in material; otherwise, it is possible that a considerable
studying groundwater pollution problems and protecting portion of the injected tracer is lost during the test,
the environment. and it may be a major source of error in interpreting
the results. The tracer should move in the same
manner as water, and should not interact with
GOALS OF GROUNDWATER TRACING aquifer material.
4. should be economically available. In some tracing
The most important goals to be achieved in a tracing
experiments, the cost of tracer is a decisive factor.
program can be categorized into two groups:
Quantitative goals: To determine 5. should not change the hydraulic conductivity of the
medium under investigation.
• groundwater recharge amount 6. should be stable enough during the test period from
• groundwater discharge amount injection until detection.
• components of groundwater recharge
• age of groundwater It is obvious that one could not find an ideal tracer that
meets all of these specifications. Although there are some
• elevation of recharge
tracers developed specifically for groundwater tracing,
• groundwater velocity selecting a suitable tracer, as well as applying a suitable
• volume of aquifer dynamic storage amount of the tracer, is a state-of-the-art task and cannot
• dispersion coefficient be done without sufficient experience, favorable expertise,
• travel time and knowledge about the medium under study.

Quantitative goals: To determine

GROUNDWATER TRACER TYPES
• origin of water
• hydraulic connection between different surface water Groundwater tracers can be categorized into two main
and groundwater bodies groups of artificial (to be introduced into the system)
and natural (environmental or preexisting in the system)
• hydraulic connection between different basins
tracers (Fig. 1). Natural tracers are those present in water
• spring catchment area molecules or accompany water molecules naturally along
• aquifer recharge areas their path, such as deuterium (2 H), carbon-13, or water
• aquifer discharge areas temperature. Discharge pulses are often good tracers
• aquifer boundaries for water especially in the response of springs to flash
• groundwater flow direction and path rainfalls. Artificial tracers are much more diverse in type,
and there are many more tracers not mentioned.
• flow regime
• role of structural features in groundwater flow
Natural (Environmental) Tracers
• saline water–freshwater interface
• source of pollution Natural or environmental tracers are those that already
exist in the natural environment, the hydrologic cycle.
• pollution distribution pattern
They could be substances or some kinds of energy.
• degree of karstification
• seepage/leakage zones in dams and reservoirs Environmental Isotopes. Environmental isotopes used as
groundwater tracers are isotopes of hydrogen, oxygen, and
CHARACTERISTICS OF IDEAL TRACERS carbon. They could be part of a water molecule (2 H or
deuterium, 18 O, and 3 H or tritium), or they might enter the
Among the characteristics for selecting a tracer as an ideal hydrologic cycle and move together with water molecules
applicable tracer are the following: (13 C, 14 C, 14 N, 15 N, 32 S, 34 S, 36 S . . .).
 GROUNDWATER TRACING 503

sulfur isotopes (32 S, 34 S and 36 S) found in the sulfate ion

• Environmental isotopes
• Major ions vary quite widely and, under certain circumstances, can
Natural be useful indicators of the origin of the sulfate, which is
• Temperature
(environmental) particularly true if, for example, one wishes to distinguish
• Pulses
• ... sulfate originating from natural dissolution of gypsum (Ca
SO4 : 2H2 O) from sulfate originating from an industrial
spill of sulfuric acid.
Groundwater tracers
Active Isotopes
Tritium. Tritium (3 H) is an unstable isotope of hydrogen
whose half-life is 12.32 years. It enters the hydrologic
• Dyes cycle in the same manner as stable hydrogen in the H2 O
• Fluorescent molecule. This isotope is naturally produced at low levels
• Nonfluorescent
Artificial
• Drifts
in the atmosphere. During and after the nuclear tests
(introduced) in the Northern Hemisphere in the 1960s, the tritium
• Salts
• Detergents concentration (referred to as TU: Tritium Unit) in the
• ... atmosphere increased from a level of 5–15 TU to up to
4000 TU. 3 H can be used qualitatively to date groundwater
Figure 1. Groundwater tracer types. groundwater that has less than 2 to 4 TU is dated prior
to 1953; if the amount is significantly greater than 10 to
20 TU, it has been in contact with the atmosphere since
Stable Isotopes 1953. Tritium analysis in groundwater studies is mainly
Hydrogen and Oxygen Isotopes. Analysis of the natural for absolute groundwater dating for ages up to 40 years.
concentration of these isotopes in water has helped
in understanding groundwater flow. Because of the Carbon-14. 14 C is the radioactive isotope of carbon.
large differences in mass between the two hydrogen It has a half-life of 5730 years and is mainly used to
isotopes, they tend to fractionate whenever evaporation date groundwaters up to 30,000 years of age. Most 14 C
or condensation takes place. Other factors being equal, in potable groundwater is contained in the HCO3 − ion
waters of higher deuterium content are found near in water. As a first approximation, the initial number of
14
coastlines, at low elevations, in warm rains, and where C nuclei per total carbon nuclei, or X0 in Eq. 1, in a
partial evaporation occurs, as in saline lakes. The water sample is considered to have been constant prior to
variations in oxygen isotope almost follow those of 1950 due to the almost constant natural production of 14 C
deuterium. The most common use of studies of 2 H and 18 O in the atmosphere by cosmic radiation interacting with
has been to trace the regional movement of groundwater the atmosphere.
and to locate recharge areas. Xt = X0 e−λt (1)

Nitrogen. The two abundant isotopes of nitrogen (14 N where λ = decay constant
and 15 N) can vary significantly in nature. Ammonia t = time
escaping as vapor from decomposing animal wastes, for X0 = number of nuclei at zero time
example, tends to remove the lighter nitrogen (14 N) and Xt = number of nuclei at time t
leaves behind a residue rich in nitrogen. In contrast, many
fertilizers that have an ammonia base are isotopically If the only source of 14 C in water is originally from
light. Natural soil nitrate is somewhat between these two the active biosphere, then the 14 C that is measured in
extremes. As a consequence, nitrogen isotopes have been carbon from a water sample can be considered Xt in the
useful in helping to determine the origin of unusually high equation because λ is known from experimental work, the
amounts of nitrate in groundwater. age of the sample (t) can be determined directly. Due to
Most nitrogen in groundwater is in the form of nitrate some complexities, the use of 14 C in age determination
anion (NO3 − ) or dissolved nitrogen gas (N2 ) from the is not so simple. The existence of carbon from carbonate
atmosphere. Locally, in zones devoid of dissolved oxygen, formations, or formation of CH4 gas, are factors that affect
the chemically reduced form (NH4 + ) may predominate. In the results of age determination. But despite that, 14 C
general, nitrate moves as a conservative tracer and is an dating is highly useful in determining the approximate
important indicator of pollution. If nitrate concentrations residence time of old groundwater.
exceed about 10 mg/L, the health of infant mammals
including humans may be adversely affected. Also the Major ions. The chemical composition of groundwater
presence of more than about 5 mg/L of nitrate commonly is in most cases a good indicator of the environment
is an indirect indication of other forms of contamination, through which the water has passed. In almost all
including those from fertilizers and sewage. groundwater tracing and isotopic hydrologic programs,
water is analyzed for calcium, sodium, magnesium,
Sulfur. Most dissolved sulfur within shallow groundwa- potassium (major cations), chloride, sulfur, carbonate,
ter is bound within the sulfate ion (SO4 2− ). The stable bicarbonate, and nitrate (major anions).
504 GROUNDWATER TRACING

Figure 2. Isochlor map for Tehran

plain, 1977, Tehran, Iran. The general
groundwater flow direction is southward.
As can be seen, groundwater gets richer
in ions (specifically in chlor) as it flows
downstream.

Typically, if we follow a recharge event and track the etc. A pulse is normally an indicator of a hydrologic event
water down-gradient in the flow path, it can be seen upstream. For example, a sudden rise in the amount of
that the ionic concentration in the water increases as it turbidity in a spring could be an indication of a storm in a
migrates down-gradient (Fig. 2). Generally, groundwater spring recharge area. It could also be a sign of a hydraulic
is the bicarbonate type in recharge areas the sulfate type connection between the storm area and the spring.
midway, and the chloride type in terminal zones (e.g.,
basin outlet). The chemical composition of groundwater Artificial Tracers
would finally be demonstrated on specific presentation
charts such as Piper, Durov, Schuler, . . . and different Artificial tracers are those that are introduced intention-
groundwater types could be easily recognized. ally into groundwater. So an ‘‘injection point’’ is necessary
in artificial tracing in addition to a sampling network.1
Temperature. Water temperature is almost a good While in environmental tracing, the study is performed
indicator or tracer where there is an abrupt difference only through a sampling network. Normally, artificial
in the physical properties of the aquifer (e.g., a change tracing and natural tracing are complementary, and some-
in hydraulic conductivity, fracture properties . . .) or times both methods are applied to the same area to achieve
where there is a significant temperature gradient (Fig. 3) more reliable results. Among all artificial tracers, the most
Thermometry is a study tool for investigating changes in applicable are fluorescent dye tracers which are discussed
rock properties in a borehole column. It comprises a heat at the end for more emphasis.
sensor that goes down through the borehole and records
the water temperature along the borehole column. The Drifts. In areas such as karstified limestone terrains
record is called a temperature log. or highly faulted volcanic rocks where groundwater flow
follows a turbulent pattern, drifts usually are good choices
Pulses. Pulses are sudden changes in the amount of a to be selected as tracers. Signal-emitting floats, yeasts,
specific parameter that can be easily seen as a peak in a
graph of the parameter versus time. It could be the graph 1
Sometimes the injection point and sampling point are the same.
of a spring discharge, electrical conductivity, turbidity, This is the case in the so-called single-well tracing method.
 GROUNDWATER TRACING 505

bacteria, viruses, and spores have reportedly been used as 1000 to 10000) than salt tracers. So a small amount
tracers in many cases. of tracer is enough.
• Instrument detection is quick and easy without time-
Salts. Most tracers that have already been used consuming sample preparation.
extensively are common salts. This category of tracers
• The chemical properties of the dyes are in some cases
includes those compounds that ionize in water, resulting
better because salts (particularly the cationic part)
in separation into charged elements possessing a positive
are influenced by ion exchange in clayey sediments.
charge (cations) or a negative charge (anions). The most
common ionic tracers are chloride (Cl− ), bromide (Br− ), • Fluorescent dyes do not influence the density of water
lithium (Li+ ), and potassium (K+ ). These tracers have near the injection point.
been successfully used to determine flow paths, residence • Depending on the type of analytical instrumentation,
time, and aquifer properties. The most important factors the costs to determine fluorescent dyes are cheaper
that should be considered before applying such tracers (using an ordinary photometer) or comparable to
are aquifer material texture and natural background a flame photometer which can be used to analyze
concentrations in the aquifer. salt tracers. The detection of salts by electrical
The advantages of salt tracers are their availability, conductivity measurement is obviously cheaper
low cost, and stability, but they must be injected in large but cannot distinguish between different types of
amounts, hundreds of kilograms or tons, and this causes salt tracers.
problems in handling and transportation. As they are • A disadvantage is the relatively high cost of
injected in huge amounts, they make water denser around fluorescent dyes, but, on the other hand, up to
the injection point, and it may affect the hydrodynamic 1000 times less dye than salt is necessary for a
behavior of the water. tracing experiment.

Fluorescent Dyes. For many reasons, fluorescent dyes These are the most commonly used fluorescent dyes:
have become very important as tracers in hydrology and
hydrogeology. They can be used in hydrology to measure Uranine. For hydrologic or hydrogeologic purposes,
transit times in streams and rivers, to measure discharge uranine (C20 H10 O5 Na2 , Color index name: Acid Yellow
in open channels by the tracer dilution method, and 73) is the most important tracer (Fig. 4). It is also the
to investigate dispersion processes in rivers, lakes, and most sensitive of all. The dye is supplied as a hygroscopic
seas. Of course their main application is in hydrogeologic orange-red powder: it is very soluble in water and has
studies for estimating hydrodynamic parameters as well a yellowish-red color and an intense yellowish green
as studying sources of pollution. fluorescence. In solutions at concentrations higher than
Application of fluorescent tracers has some great 1 g/m3 (equal to 1 ppm), the color is orange-red. The
advantages and a few disadvantages: fluorescence is visible without any instrument above
concentrations of 50 mg/m3 (50 ppb) and approximately
• They are visible without the use of any instrument at above 10 mg/m3 (10 ppb) using a UV lamp, depending
higher concentrations (say more than 10 ppb). This on the quality of the dye. It has no specific odor. At
property can be useful for qualitative experiments. high concentrations, the fluorescent light is absorbed by
• The limit of detection for the dyes is in the sub-ppb the sample itself. The intensity of the fluorescence can
range and therefore much better (up to a factor of be measured exactly only if the concentration is much
lower (less than 100 ppb); otherwise, the solution has to
be diluted.
The solubility of Uranine is 500 mg/L, and its detection
limit is 0.001 mg/m3 . Its fluorescence intensity in solution
is very much dependent on the pH. At pH lower than 6,
the intensity tends to zero. Uranine is an anionic tracer,
Flow lines and it is adsorbed by clay minerals only to a minor extent.
Isotherms
It has also good temperature stability which is important
for analytics (a temperature increase in the instrument)
and for injection, for example, into thermal systems. From
26.1 25 24
26.8 24.8
25.5
26 NaO O O

COONa

Figure 3. Sketch showing a section of a temperature gradient

due to freshwater recharged from the river (up) and the plan of
isotherms at both sides of the river (down). Figure 4. Molecular structure of uranine.
506 GROUNDWATER TRACING

the view point of toxicity, it is not dangerous to humans or Tracer Injection. By injecting the tracer into the aquifer,
animals, in particular, at concentrations normally used in it will enter the hydrologic cycle and (it is assumed that)
tracing experiments. will travel in the same manner as water. The injection
method varies with the type of tracer, study purpose, site
Eosin. Eosin (C20 H6 Br4 Na2 O5 , Acid Red 87) (Fig. 5) conditions, etc.
is a powder made of red crystals of a bluish tinge. A
concentrated aqueous solution is deep brownish-red, dilute Sampling. Sampling is done for the purpose of analyzing
solutions are yellowish-red with a greenish fluorescence. water to find if there is tracer in it. Actually the sampling
The fluorescence is visible in more concentrated solutions work starts before injecting the tracer because there might
(above 10 ppb). be some background concentration of the tracer in the
It is very soluble in water, less in alcohol, and water; this is called background sampling.
insoluble in ether. At high concentrations, fluorescent Because of the very high dissemination potential
light is absorbed by the sample itself. For an exact of fluorescent tracers, it is very important to avoid
analysis, the concentration should not be higher than any contact between injecting personnel/equipment and
100 ppb, otherwise, dilution is recommended. Its solubility sampling. The sampling time interval for each sampling
is 300 g/L and the detection limit is 0.01 mg/m3 . The pH point should be scheduled according to the hydrogeologic
effect for eosin is not as significant as that for uranine. factors and distance to the injection point.
Eosin is an anionic tracer and is generally little adsorbed
by clay minerals. It can be used in karstic systems as well Analysis. All samples should be carefully packed and
as in porous groundwater aquifers. It is not dangerous to sent to the lab to be analyzed as soon as possible. The
humans or animals. analytical method and the instrumentation for measuring
Among other fluorescent dyes, sulforhodamine G, the fluorescence of the samples are discussed following. As
Sulforhodamine B, rhodamine WT, rhodamine B, and mentioned before, concentrations higher than 10 ppb of
pyranine are used most groundwater tracing. most fluorescent dyes may be visible and can be detected
without any instrument.
Tracing Experiment
Detection with a Two Monochromator Spectrofluorom-
While planning a tracing test, there are some factors eter. To analyze for tracers, the best method is to use
that must be considered. First, the objectives of the a spectrofluorometer with monochromators for excitation
study should be defined well because they affect other and emission (Fig 6). The light of the excitation source is
study components such as selecting the tracer. Then transmitted through a filter of maximum transmittance
all available resources (labor, money, time,) should be near or at the absorption maximum of the dye. The dye
considered to facilitate formatting the program. Third, all molecules in the solution will be excited, and the fluo-
relevant information and data (maps, reports, well logs) rescence can then be dispersed by a secondary filter or a
should be available so that a good idea of the geology and monochromator (prism or grating). Table 1 is a summary
hydrogeology of the site is gained. of spectral data of some fluorescent dyes, and Table 2 is a
The general direction of groundwater flow and also its summary of their approximate detection limits in water.
velocity should be estimated before starting the test, which The final result of an analysis is a time–concentration
will allow configuring the injection and sampling network (or breakthrough) curve, which presents the temporal
effectively. evolution of the sampling point tracer concentration
(Fig. 7).
Different Phases of the Test
Planning and General Assessment. Before starting a Interpretation of Results. With regard to the break-
tracing test, the test site should be visited and examined through curve plotted for each sampling point, one should
carefully, which is to schedule the program activities
and other components of the work such as selecting the
injection and sampling points, selecting the tracer (which
is among the most important and critical components of the Excitation Sample
Source
monochromator cell
work), designing the sampling program, transportation
(equipment, tracers, personnel, etc.

Slits
Br Br
NaO O O
Emission
monochromator
Br Br

COONa

Recorder Amplifier Detector

Figure 5. Molecular structure of eosin. Figure 6. Scheme of a spectrofluorometer.

 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY 507

Table 1. Summary of Spectral Data of Some Fluorescent Jahani, H.R. (1998). Groundwater Tracing and Isotope Hydrolog-
Dyes ical Studies in Rudbar-e Lorestan Area, Iran. Water Research
Center, Tehran, Iran, p. 83.
Wavelength
Fluorescent Fluorescent Difference Standards Office. (2003). Standard Guidelines on Groundwater
Dye Excitation Emission (nm) Tracing in Karst and Hardrock Formations (Draft Edition-
Farsi). Iranian Ministry of Energy, Tehran, Iran, p. 60.
Uranine 491.0 512.5 21.5
Eosin 515.5 536.0 20.5
RhodamineB 553.5 576.0 22.5 HYDRAULIC CONDUCTIVITY/
Pyranine 463.5 512.5 49.0
TRANSMISSIBILITY

STEVEN A. DIELMAN
Table 2. Summary of Approximate Detection Limits for ENVIRON International
some Fluorescent Dyes Corporation
Dye Limit of Detection (mg/m3 ) Arlington, Virginia

Uranine 0.001–0.025
Eosin 0.050–0.5 Hydraulic conductivity is a fundamental parameter that
Rhodamine B 0.010–0.5
governs the flow of liquids such as groundwater through
Pyranine 0.008–0.5
aquifers and other porous media. Specifically, hydraulic
conductivity is a quantitative measure of the capacity of
a geologic formation or other porous media to transmit
14 a specific fluid. It is determined by the characteristics of
Concentration (ppb)

12 both the porous medium and the fluid of interest.

10 Transmissibility (or transmissivity) is a property
8 closely related to hydraulic conductivity that describes
6 the capacity of a specific water-bearing unit of a
4 given thickness, such as an aquifer, to transmit water.
2 Transmissibility is most simply defined as the effective
0 hydraulic conductivity of an aquifer or other water-bearing
0 10 20 30 40 50 60 70 80 90 unit multiplied by the thickness of that unit.
Days after injection
Figure 7. The breakthrough curve for a spring in the Tang-e ORIGINS IN DARCY’S LAW
Shemiran area, western Iran.
The concept of hydraulic conductivity dates back to Henry
Darcy’s famous 1856 paper regarding the water supply for
first of all focus on the first arrival time of the tracer. the city of Dijon, France. In that paper, Darcy described the
It reveals first, that there is an underground hydraulic results of laboratory experiments that he had conducted
connection between the injection point and the sampling to investigate the flow of water through sand. From the
point and, second, the maximum groundwater flow veloc- results of those experiments came the relationship now
ity in that zone. The peak concentration also leads to the referred to as Darcy’s law, it is valid for laminar flow
average transit time. The shape of the curve is a guide conditions that occur in most groundwater systems and
to the approximate volume of water in which the tracer can be expressed as
has been diluted. The area under the curve is equivalent
to the total mass of tracer recovered from the sampling ∂h
point (e.g., a spring). The summation of this figure for all Q = −K A (1)
∂l
sampling points should be close to the injected amount.
Otherwise, there is another outlet for the aquifer which where Q = volumetric flow rate (L3 /t)
was unknown during planning. K = hydraulic conductivity (L/t)
∂h
= hydraulic gradient (L/L)
∂l
READING LIST A = Cross-sectional area perpendicular to the flow
direction (L2 )
Benischke, R. (2002). Fluorescent Tracers in Hydrology. Institute
of Hydrogeology and Geothermics, Graz, Austria, p. 60. As evident from Eq. 1, the volumetric flow of a fluid
Davis, S.N. (1985). Groundwater Tracers. National Groundwater through a porous medium is directly proportional to the
Association, Westerville, OH, p. 200. coefficient K; Darcy recognized that it is dependent on the
Ighanian, R. (1977). Hydrochemical Studies of Tehran Plain. type of soil through which water flows. For this reason, the
Ministry of Energy, Tehran, Iran, p. 32. coefficient K has historically been referred to as Darcy’s
Jahani, H.R. (2000). Groundwater Tracing in Tang-e Shemiran, coefficient of proportionality. Today, the coefficient K is
Western Iran. Water Research Center, Tehran, Iran, p. 85. more commonly called hydraulic conductivity.
508 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY

RELATIONSHIP TO PERMEABILITY Because transmissibility is defined as hydraulic conduc-

tivity multiplied by aquifer thickness, its units are (L2 /t),
It is intuitive that the flow of a liquid through soil also often expressed as volume per unit width of the aquifer
depends on the characteristics of the liquid. For example, per unit time (e.g., gallons per day per foot or gpd/ft).
if all other factors are equal, the volumetric flow rate
(Q in Darcy’s equation above), and hence the hydraulic
REPRESENTATIVE HYDRAULIC CONDUCTIVITY VALUES
conductivity, will be greater for water than for a more
viscous liquid such as a heavy oil.
Local hydraulic conductivity values can vary over an
The relationship of hydraulic conductivity (K) to
extremely wide range. Ranges in representative values
permeability (k)1 is useful in distinguishing the effects
of hydraulic conductivity for various geologic media are
of the porous medium from the effects of the fluid:
summarized below.
Figure 1 illustrates that hydraulic conductivity values
kρg
K= (2) may range over approximately 12 orders of magnitude,
µ
depending on the type of soil or rock. Even for specific
types of geologic deposits, variations over several orders of
where ρ = density of fluid;
magnitude are not uncommon.
µ = dynamic viscosity of fluid; and In general, geologic deposits that exhibit the lowest
g = gravitational acceleration constant. hydraulic conductivity values are fine-grained unconsoli-
Permeability k is a function only of the medium and is dated materials (e.g., clays) and unfractured sedimentary
commonly used in petroleum studies where various fluids and crystalline rocks. The presence of fractures in either
of differing characteristics such as oil, water, and natural consolidated or unconsolidated materials of otherwise low
gas may be present together in multiphase flow systems. permeability can significantly increase the hydraulic con-
From Eq. 2, it is clear that hydraulic conductivity is ductivity of these materials.
a function of the characteristics of the fluid as well as Geologic deposits that exhibit the highest hydraulic
the permeability of the medium (k).2 Specifically, a soil conductivity values are typically coarse-grained unconsol-
with a given permeability will exhibit relatively greater idated materials (e.g., coarse sands and gravels) and rocks
hydraulic conductivity to fluids of lower viscosities and/or that are highly fractured or otherwise exhibit significant
greater densities. secondary permeability (e.g., solution-enhanced void space
Rearranging Eq. 2 also allows determining the hydrau- in carbonate formations or basalt shrinkage cracks).
lic conductivity to one fluid if the hydraulic conductivity to
another is known: GEOLOGIC FACTORS INFLUENCING HYDRAULIC
µ2 ρ1 CONDUCTIVITY
K1 = K2 (3)
µ1 ρ2
For unconsolidated materials, the factors that have the
greatest influence on hydraulic conductivity are generally
where K1 , µ1 , and ρ1 are properties of fluid 1 and K2 , µ2 ,
particle size and particle sorting. It can be theoretically
and ρ2 are properties of fluid 2.
shown that hydraulic conductivity is proportional to the
square of the particle diameter for a uniform, porous
UNITS OF MEASUREMENT medium, that is, doubling the particle diameter results in
quadrupling hydraulic conductivity. In practice, however,
The units of hydraulic conductivity are (L/t), and the units geologic formations are not uniform but contain particles
of permeability are (L2 ). In the context of groundwater of varying sizes, and the definition of a representative
flow, the units of permeability are sometimes expressed particle diameter is not clear-cut. In general, poorly sorted
as ‘‘darcys’’; one darcy approximately equals 1 × 108 cm2 . deposits (i.e., those with greater variation in particle
Conversion factors for commonly used units of hydraulic sizes) will exhibit lesser relative hydraulic conductivity
conductivity and permeability are provided in Table 1. compared to that of well-sorted deposits, due to infilling of
For most applications, the viscosity and density of water pore space by finer grained particles.
vary over a narrow range, and permeability values can be In consolidated crystalline and sedimentary rocks, the
converted to hydraulic conductivity (of water), assuming hydraulic conductivity of the rock matrix tends to be low to
standard temperature and pressure, without introducing extremely low. Transmission of water through these rocks
significant error. is generally via interconnected voids such as fractures,
solution channels, shrinkage cracks or metamorphic (e.g.,
1
In older publications, K was sometimes referred to as the schist) foliations. Transmission of water through such
coefficient of permeability and, therefore, k become known as voids is said to occur via ‘‘secondary permeability.’’
‘‘specific’’ permeability or ‘‘intrinsic’’ permeability to differentiate
it from K. More recently, it has become conventional to refer to K
as hydraulic conductivity, allowing k to be referred to simply as HETEROGENEITY AND ANISOTROPY
permeability.
2
It is common practice in groundwater hydrology publications to Geologic depositional environments typically exhibit
use the term hydraulic conductivity as it relates to water, unless significant variability across very short distances. As a
a different fluid is specified. result, hydraulic conductivity values generally exhibit
 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY 509

Table 1. Hydraulic Conductivity and Permeability Conversion Factorsa

Hydraulic Conductivity Permeability

Multiply
cm/s m/s ft/s ft/day US gal/day/ft2 cm2 Darcy
To obtain by by by by by by by

cm/s 1 100 30.48 3.53E−04 4.72E−05 9.71E+04 9.66E−04

Hydraulic Conductivity

m/s 0.01 1 0.305 3.52E−06 4.72E−07 9.71E−02 9.66E−06

ft/s 3.28E−02 3.28 1 1.16E−05 1.55E−06 3.19E+03 3.17E−05

ft/day 2.84E+03 2.84E+05 8.64E+04 1 0.134 2.75E+08 2.74

US gal/day/ft2 2.12E+04 2.12E+06 6.45E+05 7.48 1 2.06E+09 1.48E+01

Permeability

cm2 1.03E−05 1.03E−03 3.14E−04 3.63E−09 2.71E−08 1 9.87E−09

Darcy 1.04E+03 1.04E+05 3.15E+04 3.65E−01 2.72 1.01E+08 1

a
Conversions between units of hydraulic conductivity (K) and permeability (k) assume the properties of water at approximately 15 ◦ C.

spatial variation across several orders of magnitude within to geologic conditions. Freeze and Cherry (2) identified
an area of hydrogeologic interest. three types of common heterogeneity patterns: trending,
The terms heterogeneity and anisotropy are often discontinuous, and layered.
employed to describe variability in hydraulic conduc- Trending heterogeneity refers to spatial trends result-
tivity. Heterogeneity refers to variability in space, and ing from, for example, gradational coarsening or fining
anisotropy refers to variability in direction from a given depositional sequences such as those that typically result
point. Figure 2 illustrates the relationship between het- from encroaching or receding seas. Trending heterogene-
erogeneity and anisotropy. ity may also result from postdepositional processes such
The distinction between anisotropy and heterogeneity as vertical trends in rock fracture frequency.
often depends on scale and context. For example, Discontinuous heterogeneity refers to an abrupt change
on one scale, layered heterogeneity (see Fig. 4) may in local hydraulic conductivity values, such as those
require consideration as a series of distinct geologic across a geologic discontinuity, fault, or other sharp
formations that have different representative hydraulic stratigraphic contact.
conductivity values. On another scale, it may be more Layered heterogeneity is a commonly encountered
convenient and appropriate to describe this system as one type of heterogeneity, particularly for unconsolidated
anisotropic layer. or consolidated sedimentary sequences formed in time-
varying depositional environments. An idealized example
Heterogeneity
of layered heterogeneity is illustrated in Fig. 4. A more
A porous medium of uniform characteristics is said realistic depiction of layered heterogeneity is illustrated in
to be homogenous. In reality, no geologic system is the Borden aquifer hydraulic conductivity plot depicted in
truly homogenous with respect to hydraulic conductivity. Fig. 3.
Each system exhibits varying types and degrees of In reality, distinguishing between types of hetero-
heterogeneity. An illustration of actual heterogeneity is geneities is often difficult. For example, layered het-
shown in Fig. 3. This two-dimensional cross-section was erogeneity may be considered a series of discontinuous
created from more than 700 measurements within an heterogeneities. In addition, multiple patterns in hetero-
area approximately 2 meters deep by 20 meters long geneity may be present at any given location.
within the intensely studied Borden aquifer, which is
generally recognized as ‘‘mildly’’ heterogenous compared Anisotropy
to most groundwater systems. Even in this small area
of a relatively uniform aquifer, measured hydraulic If hydraulic conductivity is independent of the direction
conductivity values range across approximately three of measurement, the medium is described as isotropic;
orders of magnitude. if, however, the hydraulic conductivity value depends on
In some cases, heterogeneity in hydraulic conductivity the direction of measurement, it is considered anisotropic.
values may exhibit one or more measurable trends due Mathematically, hydraulic conductivity is represented as
510 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY

Unconsolidated deposits Sedimentary rocks Crystalline rocks

101

100
Gravel

Karst and reef limestone

10−1

Permeable basalt
Coarse sand

Fractured igneous & metamorphic rock

10−2
Medium sand

Weathered
granite
Fine sand

10−3

Weathered
gabbro
10−4
Hydraulic conductivity (cm/s)

Limestone, Dolomite
Silt, loess

Sandstone
10−5

10−6

Basalt
Till

10−7
Siltstone
Unweathered marine clay

Unfractured igneous & metamorphic rocks

Clay

10−8
Anhydrite

Shale

10−9
Salt

10−10

10−11

10−12

Figure 1. Representative values of hydraulic conductivity for various rock types (adapted from: Ref. 1).

a scalar (direction-independent) variable in the form of hydraulic conductivity (Kv ) between 3:1 and 10:1 for
Darcy’s law presented in Eq. 1 In anisotropic settings, clays and shales due to mineral grain orientation. On
hydraulic conductivity must be considered a tensor a larger scale, however, Kh :Kv anisotropic ratios for
variable, leading to more complex forms of Darcy’s law sedimentary deposits may be extremely high due to
and other equations that describe the flow of groundwater stratified deposits of greatly different characteristics.
and other fluids through porous media. High Kh :Kv anisotropic ratios for stratified deposits
On a small scale, anisotropy can result from mineral are not surprising, considering the large differences in
grain orientation. On a larger scale, anisotropy can result typical hydraulic conductivities of coarse-grained and fine-
from sedimentary bedding planes, metamorphic structure grained deposits, which may differ by several orders of
(e.g., shistocity) or directional fracturing of crystalline magnitude.
rocks resulting from postdepositional stresses.
In sedimentary deposits of horizontal or near-horizontal EFFECTIVE HYDRAULIC CONDUCTIVITY: STATISTICAL
orientation, hydraulic conductivity is frequently greater TREATMENT OF HYDRAULIC CONDUCTIVITY VALUES
in the lateral directions compared to the vertical
direction. On a small scale, such anisotropy may be Groundwater hydrologists have used a variety of tech-
relatively modest. Freeze and Cherry (2) report ratios niques to characterize the spatial variability of hydraulic
of horizontal hydraulic conductivity (Kh ) to vertical conductivity values and to assess groundwater flow in
 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY 511

Homogeneous, isotropic Homogeneous, anisotropic

KZ

KZ
KZ
KX KZ KX

Z KX KX

Heterogeneous, isotropic Heterogeneous, anisotropic

KZ
KZ

KZ

KZ
KX KX

KX KX Figure 2. Four possible combina-

tions of heterogeneity and anisotropy
(adapted from Ref. 2).

219.14
4.5 5.5
4.0 4.0 5.0
4.0
Elevation, masl

218.64 4.0
5.0
5.0 5.0
5.0 5.0
4.5
218.14 5.0 4.5
4.0
4.5 4.5
5.0 5.0 5.0
217.64 5.0 5.0 4.5
4.0 5.0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Distance along section
5.0 In(K ) (cm/s)
Location of core
K < 10−3 cm/s

Figure 3. Hydraulic conductivity distribution at Canadian Forces Base Borden (from Ref. 3, v. 22, p. 2069–2082).

heterogenous geologic settings. One such approach is to expressed as

identify an ‘‘effective’’ hydraulic conductivity that, applied   dh
to a heterogenous water-bearing geologic unit, will allow Q= Qi = Ki Di W (4)
dx
an accurate representation of its ability to transmit water
as a whole. where
Consider horizontal groundwater flow through an Ki = hydraulic conductivity of layer i,
idealized layered geologic system such as that shown in dh/dx = horizontal hydraulic gradient (assumed
Fig. 4 where the a number of layers equals n. The total uniform in all layers),
flow through the system (Q) is equal to the sum of the Di = thickness of layer i, and
flow in all layers (Qi ). From Darcy’s law, this can be W = unit width.
512 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY

For combined flow through the layered system, an effective system with respect to groundwater flow through that
horizontal hydraulic conductivity (KH-EFF ) can be defined system.
for groundwater flow such that Frequently, hydraulic conductivity values and well
yields within a specific geologic setting are best character-
dh ized by a log-normal distribution. As such, the geometric
Q = KH-EFF DW (5)
dx mean is often preferred as a descriptive statistic for the
central tendency (e.g., median and mode) of a population
From Eqs. 4 and 5, the following relationship for KH-EFF
of log-normally distributed hydraulic conductivity values.
is obtained:  Further, several studies have concluded that the geometric
Ki Di
KH-EFF = (6) mean provides a better estimate of the effective hydraulic
D conductivity of a heterogenous groundwater system than
the harmonic mean or arithmetic means, which are less
This shows that, for flow parallel to the idealized layered
than and greater than the geometric mean, respectively.
system shown in Fig. 4, KH-EFF is equivalent to the
The preceding provides a simplified discussion of
arithmetic mean of Ki weighted by the thickness of each
common statistical treatments for hydraulic conductivity
layer Di . If all layers are of the same thickness (or all K
values. In recent decades, much work has been conducted
values are uniformly distributed), KH-EFF is equal to the
to apply more complex statistical models and stochastic
arithmetic mean of the individual Ki values.
techniques to characterize groundwater flow through
Consider, however, vertical groundwater flow through
heterogenous systems.
the same system. Using a similar analysis, it can be shown
that the effective horizontal conductivity in the vertical
direction (KV-EFF) is equal to METHODS OF ESTIMATION AND MEASUREMENT

D Empirical Relationships
KV-EFF =  (7)
Di /Ki Several efforts have been made over the years to correlate
hydraulic conductivity and/or permeability empirically
If all layers are of equal thickness (or are uniformly with characteristics of porous media through which water
distributed), KV-EFF for the idealized layered system is flows. Some of the more commonly used relationships are
equivalent to the harmonic mean of Ki values, which is provided in Table 2.
always lower than or equal to the arithmetic mean. Although such empirical relationships correlate
These two simplified conceptual models represent two hydraulic conductivity with various factors, nearly all can
extreme possibilities of flow uniformly parallel to or be simplified to the form
perpendicular to geologic layers, respectively. Therefore,
the arithmetic and harmonic means of a population of K = Cdx (8)
hydraulic conductivity values are generally considered to
be upper and lower bounding values, respectively, for the where C is a constant or shape factor that is a function
effective hydraulic conductivity of a heterogenous geologic of the fluid density and viscosity and may be correlated
with properties of the porous medium such as porosity
and d is some representative grain size (L). It has been
Silt mathematically shown (e.g., by Ref. 5) that for idealized
systems of uniform grain size, x = 2. For all the empirical
relationships shown in Table 2, x ranges between 1.5 and
Clay 2, which indicates a disproportionate increase in hydraulic
conductivity as grain size increases, but suggests an
exponent value of something less than 2 for many real-
world situations.

Sand and gravel Laboratory Testing

The hydraulic conductivity of geologic deposits can be
measured in the laboratory by using a permeameter, a
Weathered bedrock device similar to that originally devised by Darcy. Various
configurations of permeameters have been used; the most
common are the constant-head permeameter and the
falling-head permeameter shown in Fig. 5.
Permeameters consist of a chamber to contain a sample
Crystalline bedrock of soil or rock and allow the application of a hydraulic
gradient across the sample. Flow through the sample is
measured, and hydraulic conductivity is calculated based
on principles founded by Darcy. Careful consideration
must be used in evaluating permeameter test data. For
Figure 4. Example of layered heterogeneity. example, when testing unconsolidated material, was the
 HYDRAULIC CONDUCTIVITY/TRANSMISSIBILITY 513

test conducted on an intact core or on a repacked sample? If There are many different types of aquifer tests and
the former, the test likely represents the vertical hydraulic methods for evaluating the test results, but the most
conductivity, which may be much lower than the horizontal common types of tests employed in groundwater studies
hydraulic conductivity. If the latter, the disturbance of the are pumping tests and slug tests.
sample during repacking may result in a value that is Pumping tests involve pumping water from a well
neither representative of the horizontal nor the vertical at a known rate and measuring the changes in water
hydraulic conductivity. level through time at the pumping well and/or nearby
observation points. Slug tests involve rapid introduction
Field Measurements—Aquifer Tests or removal of a volume of water or other material into or
Hydraulic conductivity and transmissibility on a larger out of a well and measuring the changes in water level
scale are estimated by evaluating data from aquifer tests. through time at the well being tested.

Table 2. Examples of Empirical and Semiempirical Relationships for Estimating Hydraulic Conductivity and
Permeabilitya
Source Equation Parameters

Hazen (1911) K = Cd10 2

K = hydraulic conductivity (cm/s)
C = constant; 100 to 150 (cm/s)−1 for loose sand
d10 = effective grain size (cm)
(10% particles are finer, 90% coarser)

Harleman et al. (1963) k = (6.54 × 10−4 ) d10 2 k = permeability, cm2

d10 = effective grain size, cm
(10% particles are finer, 90% coarser)
Krumbein and Monk (1943) k = 760 d2 e−1.31σ k = permeability (Darcys)
d = geometric mean grain diameter, mm
σ = log standard deviation of the size distribution
Kozeny-Carmen Bear (1972) K = (ρw g/µ)n3 /(1 − n)2 (dm 2 /180) K = hydraulic conductivity
ρw = fluid density
µ = fluid viscosity
g = gravitational constant
dm = representative grain size
n = porosity

Shepard (1989) K = 40,000 d2 glass spheres K = hydraulic conductivity, ft/day

K = 1,600 d1.75 beach deposits d = mean grain size, mm
K = 450 d1.65 channel deposits
texturally immature
K = 100 d1.5 consolidated sediments
a
Adapted from Refs. 1 and 4.

dt
Constant water level

h0
h
h
Overflow
Falling-head tube

Q Q
dc

Porous plate Sample L Porous plate Sample L

Constant-head permeameter apparatus Falling-head permeameter apparatus

Figure 5. Permeameters (adapted from Ref. 4).

514 GROUNDWATER FLOW AND TRANSPORT PROCESS

In general, pumping tests affect a significantly larger the best management of groundwater resources, appraisal
volume of the aquifer than slug tests, and the resulting of the sustainable yield of aquifers, reliable evaluation
hydraulic conductivity and transmissibility estimates are of the extent of groundwater contamination, assessment
considered more representative of conditions across a of the corresponding hazards and risks, and design of
broad area. Slug tests can be used to target specific effective remediation systems and protective alternatives.
zones and points within the aquifer but are generally This chapter provides an overview of the main princi-
considered ‘‘point estimates’’ of hydraulic conductivity and ples of groundwater flow and transport processes, includ-
transmissibility. ing the basic concepts, definitions and terms, physical
governing equations, and mathematical modeling.
BIBLIOGRAPHY
VERTICAL DISTRIBUTION OF SUBSURFACE WATER
1. Domenico, P.A. and Schwartz, F.W. (1997). Physical and
Chemical Hydrogeology. John Wiley & Sons, New York. The subsurface occurrence of water may be divided into
2. Freeze, R.A. and Cherry, J.A. (1979). Groundwater. Prentice- zones of aeration and saturation (see Fig. 1). The zone of
Hall, Englewood Cliffs, NJ. aeration is partially occupied by water and partially by air.
3. Sudicky, E.A. (1986). A natural gradient experiment on solute In the zone of saturation, all pores are filled with water.
transport in a sand aquifer: spatial variability of hydraulic The zone of aeration may be further subdivided into the
conductivity and its role in the dispersion process. Water soil–water zone, the intermediate vadose zone, and the
Resour. Res. 22: 2069–2082. capillary zone (see Fig. 1).
4. Fetter, C.W. (1994). Applied Hydrogeology. Macmillan College The saturated zone extends from the upper surface of
Publishing Company, Inc., New York. saturation down to underlying impermeable rock. In the
5. Hubbert, M.K. (1940). The theory of groundwater motion. The absence of an overlying impermeable formation, the water
Journal of Geology 48: 785–944. table forms the upper surface of the zone of saturation,
which is defined as the surface of atmospheric pressure.
Water in the soil–water zone exists at less than saturation
GROUNDWATER FLOW AND TRANSPORT except temporarily when excessive water reaches the
PROCESS ground surface from rainfall or irrigation. The zone
extends from the ground surface down through the major
MOHAMMAD N. ALMASRI root zone. The water content and distribution within this
An-Najah National University zone depends on soil texture, vegetation, and atmospheric
Nablus, Palestine conditions. The intermediate vadose zone extends from
the lower edge of the soil–water zone to the upper limit
JAGATH J. KALUARACHCHI of the capillary zone. The capillary zone extends from the
Utah State University
Logan, Utah
Ground surface

Soil-water
INTRODUCTION zone

Groundwater is an important source of water for

agriculture and domestic drinking water. Two-thirds of Vadose water
aeration
Zone of

the world’s freshwater is found in underground aquifers. Intermediate

vadose
Groundwater occurs mainly in the tiny spaces between zone
soil particles that are called voids or pores. In its natural
state, groundwater is perpetually moving. This movement
is governed by conventional hydraulic principles such as
Darcy’s law, the concept of hydraulic gradient, and the
permeability of the medium (1). With the increasing water Capillary
zone Water table
demand due to population growth, management of water
resources is becoming a challenging task. Management
of groundwater resources must consider both sustainable
or phreatic water

development and prevention and control of groundwater

Groundwater
saturation
Zone of

contamination. Groundwater contamination is a critical

issue because of the importance of groundwater supply
and quality in sustainable development. In general,
contaminants found in groundwater are associated with
adverse health, social, environmental, and economic
impacts (2). Exposure to contaminated groundwater may
occur accidentally because contaminated water might Impermeable rock
be colorless, odorless, and tasteless. It is essential to
understand groundwater flow and transport processes for Figure 1. Illustration of saturated and unsaturated zones (1).
 GROUNDWATER FLOW AND TRANSPORT PROCESS 515

Recharge
area
Water
Water Piezometric surface table Artesian
table Ground Flowing well well
surface well

Water table

Unconfined
aquifer

Confining stratum
Impermeable
strata
Confined aquifer
Figure 2. Schematic cross section of unconfined
and confined aquifers (1).

water table up to the limit of capillary rise of water. The the fluid passing through it. In addition to hydraulic
thickness of the capillary zone varies inversely with the conductivity, the term transmissivity T (L2 T−1 ) is widely
pore size of a soil. employed in groundwater hydraulics. Transmissivity
equals the hydraulic conductivity multiplied by the
saturated thickness of the aquifer.
AQUIFER TYPES
Darcy flux or specific discharge is given by the following
An aquifer is a formation that contains sufficient saturated relationship:
Q dh
permeable material to yield significant quantities of water q= = −K (2)
A dl
to wells and springs (3). Therefore, aquifers have the
ability to store and transmit water. Aquifers might be The specific discharge q has the dimensions of velocity
bounded by confining beds that are relatively impermeable (L/T). Finally, the cross-sectional area of flow for a porous
materials located adjacent to the aquifers. Confined medium is actually much smaller than the dimensions of
aquifers occur where groundwater is confined under the aquifer. To find the velocity at which water is actually
pressure greater than atmospheric by overlying relatively moving, the specific discharge is divided by the effective
impermeable formations. In a well penetrating such an porosity to account for the actual open space available for
aquifer, the water level will rise above the bottom of the flow. The result is the average velocity at which water
confining bed, as shown by the artesian and flowing wells moves between two points as in the following equation:
of Fig. 2. Unconfined aquifers do not have confining beds
above and the aquifer can be directly recharged by rainfall q Q Kdh
v= = =− (3)
(see Fig. 2). An unconfined aquifer is one in which the ne ne A ne dl
water table varies in a rising and falling form.
STORAGE IN CONFINED AND UNCONFINED AQUIFERS
DARCY’S LAW
Storage in confined aquifers is associated with the
In 1856, the French engineer, Henry Darcy, developed compressibility of water and the porous matrix. The
the principal equation in groundwater hydrology based on change in the potentiometric head affects the arrangement
laboratory experiments he had carried out to analyze the of mineral grains as well as the density of water in the
flow of water through sand. Darcy’s law can be written as voids. If the water pressure increases, the mineral skeleton
will expand; if it drops, the mineral skeleton will contract.
dh Likewise, water will contract with an increase in pressure
Q = −KA (1) and expand if the pressure drops. The specific storage (Ss )
dl
is the amount of water per unit volume of a saturated
Q is the quantity of water per unit of time (L3 /T), A formation that is stored or expelled from storage owing to
is the area perpendicular to flow (L2 ), K is a constant compressibility of the mineral skeleton and the pore water
of proportionality called the hydraulic conductivity (L/T), per unit change in head. The value of specific storage is
h is the hydraulic head (L), and dh/dl is the hydraulic very small, generally 0.0001 ft−1 or less. In unconfined
gradient. The hydraulic conductivity expresses the ease aquifers, pumpage causes drawdown in the water table
by which water can pass through soil pores and openings itself, and the induced hydraulic gradient causes both a
and serves as a measure of the permeability of the horizontal and vertical flow component. Thus, the drop in
porous medium. The hydraulic conductivity of a soil is the water table creates a vertical hydraulic gradient which
a function of the properties of the porous medium and causes drainage of the porous medium (3–7).
516 GROUNDWATER FLOW AND TRANSPORT PROCESS

The storativity (S) of a confined aquifer is the product of sources are identifiable, located at single locations, and
the specific storage Ss and the aquifer thickness. The value have local impacts on groundwater quality. Conversely,
of the storativity of confined aquifers is of the order 0.005 nonpoint sources are diffuse, originating from many
or less. Storativity under unconfined conditions is referred smaller sources, largely distributed, lack a well-defined
to as the specific yield, defined as the ratio of the volume single point of origin, and have a regional influence on
of water an unconfined aquifer drains by gravity due to groundwater quality. Examples of nonpoint sources are
the unit decline in the water table to the total volume agricultural activities such as fertilizer and manure appli-
of the aquifer. The coarser the material, the more closely cations; and polluted atmospheric deposition, including
the specific yield approaches the porosity. The volume of dry fallout, precipitation, and runoff from roads and park-
water drained from an aquifer as the head is lowered may ing lots. The loading history defines how the concentration
be found from the formula, of a contaminant or its rate of application changed with
time at the source location.
Vw = S A h (4)
TRANSPORT PROCESSES
where Vw is the volume of water drained or released (L3 ),
S is the storativity (L0 ), A is the surface area overlying
The physical processes that control contaminant flux
the drained aquifer (L2 ), and h is the average decline in
into and out of a groundwater system are advection,
head (L). Note that h for a confined aquifer is the decline
mechanical dispersion, and molecular diffusion. In the
in the potentiometric head and for unconfined aquifers is
following subsections, the basic concepts related to these
the decline in the water table.
physical processes are illustrated.

MATHEMATICAL MODEL OF GROUNDWATER FLOW Advection

Transport by advection or convection entails movement
Groundwater flow is described using Darcy’s law and and carrying contaminants with flowing groundwater (6).
the mass conservation equation. The governing equation, In advective transport, contaminants travel at a rate equal
describing transient groundwater flow in a three- to the average linear velocity of the groundwater (5).
dimensional flow domain, is given as Mathematically, advective flux is expressed by the



 following equation (6):
∂ ∂h ∂ ∂h ∂ ∂h ∂h
Kx + Ky + Kz + W  = Ss
∂x ∂x ∂y ∂y ∂z ∂z ∂t J = vx ne C (6)
(5)
where Kx , Ky , Kz are the hydraulic conductivity values
where Jx is the advective mass flux (ML2 T−1 ) per unit
in the x, y, and z coordinates, respectively; Ss is the spe-
area per unit time, vx is the average linear velocity of
cific storage; t is time; h is the hydraulic head; and W  is
groundwater(LT−1 ), ne is the effective porosity through
the source or sink strength in units of volume/time/area.
which flow occurs, and C is the contaminant concentration
To solve Eq. 5, initial and boundary conditions should be
(ML−3 ). The average linear velocity can be determined
specified for the study area. Solving this equation yields
from Eq. 3. The product of the average linear velocity
the head distribution in time and space. Computer codes
and time since contaminants were introduced into the
can be used to solve Eq. 5 numerically. MODFLOW devel-
groundwater system defines the position of the advective
oped by the U.S. Geological Survey (8), is one of the most
front, as depicted in Fig. 3, where the vertical line
commonly used groundwater softwares for solving one-
represents the advective front without considering the
, two-, or three-dimensional groundwater flow problems.
effects of dispersion and diffusion (6).
MODFLOW is divided into packages that deal with single
simulation aspects. For example, the Well Package sim-
Diffusion
ulates the effects of injection and abstraction wells and
the River Package simulates the effect of rivers, etc. The The movement of contaminants from an area of high
WWW address for MODFLOW model and documenta- concentration to an area of lower concentration is called
tion is: http://water.usgs.gov/sofware/ground water.htm. molecular diffusion or diffusion. This diffusive flux is
A brief description of MODFLOW can also be found in
Charbeneau (7).
Advection only

SOURCES OF GROUNDWATER CONTAMINATION 1

concentration
Relative

(C /C0)

There are three key features that differentiate ground- 0.5 Advection + diffusion Advection +
water contamination sources: (1) degree of localization, dispersion
(2) loading history, and (3) types of contaminants. Sources 0
of groundwater contamination can be classified into point Distance x
and nonpoint sources. Point sources include leaking under- Figure 3. Advective transport and the influence of longitudinal
ground pipelines or tanks, wastewater lagoons, septic dispersion and diffusion on the transport of a solute in
systems, and leaking barrels of chemical wastes. These one-dimensional flow.
 GROUNDWATER FLOW AND TRANSPORT PROCESS 517

directly proportional to the diffusion coefficient and the C3 < C2

L2 = Vt2
concentration gradient (9). This is known as Fick’s first C3 < C 2 C2 < C1
L1 = Vt1
law and can be expressed in a one-dimensional system as C2 < C 1 C1
in the following: C1
dC V
Fx = −Dd (7) Tracer
dx injected Contours of C = constant
at t = 0 at t = t1 Contours of C = constant
where Fx is the diffusive mass flux (ML2 T−1 ), Dd is at t = t2
the diffusion coefficient for open water (L2 T−1 ), C is C max C max
the contaminant concentration (ML−3 ); and dC/dx is the
Figure 4. Longitudinal and transversal spreading of a tracer
concentration gradient (ML−4 ). The negative sign signifies from an instantaneous point injection at t = 0.
that the diffusive flux is in the direction of decreasing
contaminant concentration.
that increases with the characteristic transport length
Mechanical Dispersion
distance of measurement. This phenomenon is attributed
Dispersion describes the gradual spreading of contam- to the macroscale effects resulting from the heterogeneity
inants beyond the region they are expected to occupy of the subsurface on the field scale. Many researchers have
when considering advection alone. The phenomenon of dis- developed relationships between longitudinal dispersivity
persion is attributed mainly to the discrepancy between and flow length (6,12–14). The ratio of longitudinal to
contaminant velocities and the average linear velocity transverse dispersivity (αL /αT ) controls the shape of the
caused by subsurface heterogeneity (5,10,11). According contaminant plume. As the ratio decreases, the wider the
to Fetter (3), this discrepancy is attributed to three key plume will be (6). Bear (10) provided a range from 5 to 24
factors: (1) the velocity of groundwater flow is higher in for αL /αT .
the center of pores than at the edges, (2) the contami-
nants can pass through different flow paths to reach the
same location, and (3) groundwater moves faster in large MATHEMATICAL MODEL OF CONTAMINANT TRANSPORT
pores. Therefore, contaminants will dilute due to mixing
with noncontaminated water. Mixing, called mechanical The advection–dispersion–reaction partial differential
dispersion, may occur along the flow path, is called longitu- equation that governs three-dimensional transport of a
dinal dispersion, and is attributed to the aforementioned single chemical constituent in groundwater, considering
key factors. The groundwater flow path may diverge at advection, dispersion, fluid sinks/sources, equilibrium-
the pore scale, so the contaminant will spread in direc- controlled sorption, and first-order irreversible rate
tions normal to the flow path. This mixing is called lateral reactions, is given in the following (11):
or transverse dispersion (3). The difference between the
 
∂C ∂ ∂ ∂C qs ρb 
straight dashed line and the solid line in Fig. 3 depicts R = Dij (vi C) + Cs − λ C + S
−
the influence of longitudinal dispersion on the transport of ∂t ∂xi ∂xi ∂xj n n
a contaminant. Molecular diffusion and mechanical dis- (10)
persion contribute jointly to the dispersive process of where the retardation factor is defined as
contaminant transport. They are combined to define a ρb ∂S
coefficient called hydrodynamic dispersion as given in the R=1+ (11)
n ∂C
following equations (6):
and C is the dissolved concentration (ML−3 ); S is the sorbed
DL = αL vi + D∗ (8)
concentration (MM−1 ) which is a function of the dissolved
DT = αT vi + D∗ (9) concentration, C, as defined by a sorption isotherm; t is
time (T); Dij is the dispersion coefficient tensor (L2 T−1 );
where DL and DT are the longitudinal and transverse vi is the average linear groundwater velocity (LT−1 ); qs
hydrodynamic dispersion coefficients, respectively; and αL is the volumetric flow rate per unit volume of aquifer
and αT are the longitudinal and transverse dispersivities, and represents fluid sources and sinks (T−1 ); Cs is the
respectively. concentration of the fluid source or sink flux (ML−3 ); λ is
Figure 4 illustrates the process of hydrodynamic the reaction rate constant (T−1 ); R is the retardation factor
dispersion in two-dimensional groundwater flow at a (L0 ); ρb is the bulk density of the porous medium (ML−3 );
constant uniform velocity, V. A tracer was instantaneously and n is the porosity (L0 ). A complete derivation of this
injected into the aquifer at t = 0. The groundwater flow equation is provided in Freeze and Cherry (4) and Zheng
carries the tracer with it. Due to the effect of dispersion, the and Bennett (11).
solute spreads out such that the maximum concentration Solving time-dependent partial differential equations
decreases with time, as shown for times t1 and t2 when t2 > governing contaminant transport requires identification of
t1 . The spread of the tracer results in an elliptically shaped the initial and boundary conditions. The initial condition
concentration distribution that is normally distributed in describes the distribution of the contaminant in the
both the longitudinal and transverse directions. groundwater within the domain at the beginning of
As mentioned earlier, longitudinal dispersivity or more simulation. The boundary conditions can be classified
specifically, dispersivity, is a scale-dependent property into three categories: (1) specified concentration boundary,
518 REACTIVE TRANSPORT IN THE SATURATED ZONE: CASE HISTORIES FOR PERMEABLE REACTIVE BARRIERS

called the Dirichlet condition; (2) a specified concentration 8. Harbaugh, A.W. and McDonald, M.G. (1996), User’s doc-
gradient boundary, called the Neuman condition; and (3) a umentation for MODFLOW-96, an update to the U.S.
variable flux boundary, called the Cauchy condition (5). Geological Survey modular finite-difference ground-water
flow model: U.S. Geological Survey Open-File Report
Solving the mathematical model yields the contami-
96–485.
nant concentration as a function of space and time. The
9. Ingebritsen, S.E. and Sanford, W.E. (1998). Groundwater in
approach used in formulating and solving the mathe-
Geologic Processes. Cambridge University Press, Cambridge,
matical model can be classified generally into numerical UK.
and analytical approaches. Analytical solutions of the
10. Bear, J. (1979). Hydraulics of Groundwater. McGraw-Hill,
advection–dispersion equations in one-, two-, and three- New York.
dimensional space can be found in Baetslé (14), Bear (10),
11. Zheng, C. and Bennett, G.D. (1995). Applied Contaminant
Domenico (15), Domenico and Schwartz (5), and Fetter (6) Transport Modeling: Theory and Practice. International
among others. Thomson Publishing, Stamford, CT.
Numerical methods provide an approximate solution 12. Neuman, S.P. (1990). Universal scaling of hydraulic conduc-
of the mathematical model by solving a set of alge- tivities and dispersivities in geologic media. Water Resour.
braic equations that replace the original partial differ- Res. 26(8): 1749–1758.
ential equations. Numerical solutions are widely used 13. Xu, M. and Eckstein, Y. (1995). Use of weighted least-squares
and preferred because they have the capability to handle method in evaluation of the relationship between dispersivity
more general conditions. Numerical methods for solving and field scale. Ground Water 33(6): 905–908.
the advection–dispersion–reaction equation divide the 14. Baestlé, L.H. (1999). Migration of radionuclides in porous
model domain into small cells, approximate the governing media. Progress in Nuclear Energy Series XII, Health
equation over each cell at time t, and then predict a new Physics. A.M.F. Duhamel (Ed.). Pergamon Press, Elmsford,
value at time t + t. This process continues forward in NY, pp. 707–760.
time for small t values. The most common numerical 15. Domenico, P.A. (1987). An analytical model for multidi-
methods for solving the advection–dispersion–reaction mensional transport of a decaying contaminant species. J.
equation include the finite-difference method, the finite- Hydrolo. 91: 49–58.
element method, and the random walk method. A detailed 16. Sun, N. (1996). Mathematical Modeling of Ground Water
illustration of these methods can be found in Zheng Pollution. Springer, New York.
and Bennett (11) and Sun (16). Bedient et al. (17) cited 17. Bedient, P.B., Rifai, H.S., and Newell, C.J. (1999). Ground-
many references that provide detailed descriptions of water Contamination: Transport and Remediation. Prentice-
these numerical methods. One of the common com- Hall, New York.
puter codes used in simulating the fate and transport 18. Zheng, C. (1990). MT3D, a modular three-dimensional
of contaminants in groundwater is MT3D (18). MT3D transport model for simulation of advection, dispersion and
chemical reactions of contaminants in ground water systems.
is a three-dimensional contaminant transport model for
Report to the US Environmental Protection Agency.
the simulation of advection, dispersion, and chemical
19. Zheng, C. and Wang, P.P. (1999). MT3DMS, a modular three-
reactions of dissolved constituents in groundwater sys-
dimensional multi-species transport model for simulation of
tems. MT3D is based on a modular structure to per- advection, dispersion and chemical reactions of contaminants
mit simulation of transport components independently in groundwater systems; documentation and user’s guide.
or jointly with finite-difference groundwater flow mod- U.S. Army Engineer Research and Development Center,
els, such as MODFLOW. MT3D can be downloaded from Contract Report SERDP-99-1, Vicksburg, MS.
http://hydro.geo.ua.edu. Further information on MT3D can
be found in Zheng (18), Zheng and Bennett (11), and Zheng
and Wang (19). REACTIVE TRANSPORT IN THE SATURATED
ZONE: CASE HISTORIES FOR PERMEABLE
BIBLIOGRAPHY REACTIVE BARRIERS

1. Todd, D.K. (1980). Groundwater Hydrology. John Wiley & LEN LI

Sons, New York. CRAIG H. BENSON
2. Office of Technology Assessment. (1984). Protecting the University of
nation’s groundwater from contamination, Office of Tech- Wisconsin-Madison
nology Assessment, OTA-0-233, Washington, DC. Madison, Wisconsin
3. Fetter, C.W. (1994). Applied Hydrogeology. Prentice-Hall,
New York.
4. Freeze, R.A. and Cherry, J.A. (1979). Groundwater. Prentice- INTRODUCTION
Hall, New York.
5. Domenico, P.A. and Schwartz, F.W. (1990). Physical and The permeable reactive barrier (PRB) is an in situ
Chemical Hydrology. John Wiley & Sons, New York. remediation technology where contaminated groundwa-
6. Fetter, C.W. (1999). Contaminant Hydrogeology. Prentice- ter is treated passively as it flows through a reactive
Hall, New York. medium. Most PRBs use granular zero valent iron (ZVI)
7. Charbeneau, R.J. (2000). Groundwater Hydraulics and to create redox conditions that result in degradation
Pollutant Transport. Prentice-Hall, Upper Saddle River, or immobilization of groundwater contaminants, such
NJ. as chlorinated solvents and herbicides, heavy metals,
 REACTIVE TRANSPORT IN THE SATURATED ZONE: CASE HISTORIES FOR PERMEABLE REACTIVE BARRIERS 519

and radionuclides (1–5). These redox conditions promote Conceptual Model

precipitation of secondary minerals from ions typically
The conceptual model used by Li et al. (8) consists of
in groundwater as well as some groundwater contam-
a continuous-wall PRB containing ZVI that is placed
inants (2,6,7). The most common minerals formed in
orthogonal to the groundwater flow. Groundwater flowing
PRBs are magnetite, hematite, goethite, lepidocrocite,
into the PRB is assumed to contain dissolved oxygen,
calcite, aragonite, siderite, green rust, ferrous hydrox-
Fe2+ , Ca2+ , Mg2+ , SO4 2− , CO3 2− , OH− , HCO3 − , and
ide, ferrous sulfide, and marcasite (8). Calcium carbon-
NO3 − . These dissolved species are assumed to form
ates and siderite typically are found near the entrance
CaCO3 , CaMg(CO3 )2 , FeCO3 , Fe(OH)2 (am), FeS (am),
face of a PRB, whereas magnetite, ferrous hydrox-
and Mg(OH)2 in response to the redox conditions
ide, green rust, and iron oxyhydroxides form through-
induced by corrosion of the ZVI. In an actual PRB,
out a PRB (6,7). Accumulation of secondary minerals
additional ions may be present and minerals other than
causes loss of pore space and reactive surface area
those included in the model may also form. However,
of the medium in PRBs, which can alter flow paths,
the reactions in the model include the most common
residence times, and the treatment effectiveness of a
minerals observed in laboratory and field studies of
PRB (2,7,9).
PRBs (6,7,10,11). Contaminant transformation reactions
Developing a general assessment of mineral precip-
were not included in the model because the volume
itation in PRBs with field and/or laboratory data is
of mineral precipitates formed as a consequence of
complicated by site-specific geochemical and hydrogeologi-
contaminant treatment generally is negligible, even
cal conditions (7), and the relatively long period over which
though the contaminant transformations are intrinsic to
mineral deposition occurs (9). An alternative approach is
the remediation process (2,11).
to conduct analyses with a reactive transport simula-
tor that predicts mineral deposition and its effects on
Codes
hydraulics (10–12). Li (12) developed a reactive trans-
port simulator specifically for this purpose. This chapter The reactive transport simulator is created with the
describes a comparison between ion concentrations pre- public domain codes MODFLOW and RT3D (15,16). MOD-
dicted with this model and concentrations measured in FLOW simulated steady-state flow in the aquifer and
the field at three PRBs. PRB. Geochemical processes within the PRB were sim-
ulated with RT3D, a three-dimensional multicomponent
model that describes advective–dispersive-reactive trans-
REACTIVE TRANSPORT SIMULATOR port (16). Heads from MODLOW are input to RT3D. A
detailed description of the codes that comprise the simu-
Equilibrium and Kinetic Reactive Transport Models lator can be found in Li (12) and Li et al. (8).
Equilibrium and kinetic models have simulated geochemi- Li et al. (8) developed a geochemical module for
cal processes and mineral formation in PRBs (10,11,13,14). RT3D that simulated the mineral-forming reactions.
Equilibrium models can estimate the aqueous species and Geochemical reactions included in the module are
solid phases that exist in a state of thermodynamic equi- summarized in Table 1. Three redox reactions are related
librium, but they do not account for the rate at which to iron corrosion, along with six mineral precipitation
reactions occur and generally do not involve the move- reactions and one reaction for microbial sulfate reduction.
ment of groundwater. In contrast, kinetic models describe A pseudo-first-order reaction rate was assumed for iron
the rate at which geochemical reactions occur and predict corrosion by oxygen or nitrate with the reaction rate
proportional to the iron surface area and the concentration
aqueous and solid phase concentrations as a function of
of oxygen or nitrate (11). For iron corrosion by water,
time. In kinetic models, the geochemical reactions usually
the reaction rate was assumed to be proportional
are specified by the user along with reaction rate expres-
only to the iron surface area (10,11,17). The rate of
sions and the rate coefficients. Thus, when a kinetic model
degradation of sulfate to HS− by sulfate-reducing bacteria
is developed, an equilibrium model often defines the most
was assumed to follow the Monod equation (11), and
important reactions to be included in the kinetic model
the kinetics of mineral precipitation were assumed
(e.g., see Yabusaki et al. (10)).
to follow transition state theory (18). The kinetic rate
Equilibrium models are appropriate when the reaction
expressions and solubility constants for the mineral
rate is fast compared with the flow rate, whereas
reactions are summarized in Table 1. All geochemical
kinetic models are appropriate when flow rates are
reactions are assumed to occur in parallel and are solved
not negligible relative to reaction rates. For most
simultaneously.
PRBs, flow rates are fast enough such that reaction
kinetics must be considered when evaluating mineral
precipitation (10). Reactive transport codes incorporating FIELD DATA SOURCES AND MODEL SETUP
kinetic geochemical algorithms have been developed by
Mayer et al. (11) and Yabusaki et al. (10) for simulating Field data from three PRBs verified that the MODFLOW-
contaminant degradation and mineral precipitation in RT3D simulator produces reasonable predictions of major
PRBs. The code developed by Li et al. (8) is similar to those ion concentrations. Comparisons were not made between
developed by Mayer et al. (11) and Yabusaki et al. (10), measured and predicted solid phase concentrations at
but it focuses solely on geochemical reactions that result any of the sites because of a lack of field data. The
in mineral precipitation. selected PRBs are located at Moffett Federal Airfield in
520 REACTIVE TRANSPORT IN THE SATURATED ZONE: CASE HISTORIES FOR PERMEABLE REACTIVE BARRIERS

Table 1. Geochemical Reactions and Solubility Constants

Solubility
Mineral Kinetic Rate Constant
Reaction Formed Expressiona Log(Keq )b

Iron Corrosion:
Fe0 + H2 O + 0.5O2 (aq) → Fe2+ + 2OH− — kan,Fe SaDO —
Fe0 + 2H2 O → Fe2+ + H2 (aq) + 2OH− — kan,Fe S —
4Fe0 + 7H2 O + NO− 3 → 4Fe
2+
+ 10OH− + NH+
4 — kNO3 SaNO−3 —
Microbial Sulfate-Reduction:  

− − [SO2−
4 ] [H2 ]
SO2−
4 + 4H2 (aq) → HS + OH + 3H2 O — k —
KSO2− + [SO2−
4 ]
KH2 + [H2 ]
4
Secondary Mineral Precipitation-Dissolution:
CaCO3 (s) ↔ Ca2+ + CO2−
3 Aragonite/Calcite −kCa (1 − aCa2+ aCO2− /Keq ) −8.1
3
CaMg(CO3 )2 (s) ↔ Ca2+ + Mg2+ + 2CO2−
3 Dolomite −kCaMg (1 − aCa2+ aMg2+ a2CO2− /Keq ) −17.7
3
Mg(OH)2 (s) ↔ Mg2+ + 2OH− Brucite −kMg2 (1 − aMg2+ a2OH− /Keq ) −11.2
FeCO3 (s) ↔ Fe2+ + CO2−
3 Siderite −kFe1 (1 − aFe2+ aCO2− /Keq ) −10.5
3
Fe(OH)2 (am) ↔ Fe2+ + 2OH− Ferrous hydroxide −kFe (1 − aFe2+ a2OH− /Keq ) −15.2
FeS(am) + H2 O ↔ Fe2+ + HS− + OH− Ferrous sulfide −kFeS (1 − aFe2+ aHS− aOH− /Keq ) −18.4
a
k is rate coefficient, S is reactive surface area of ZVI, a is activity of aqueous species, and Keq is solubility constant. Activity coefficients were calculated with
the extended Debye—Huckel equation.
b
At 15 ◦ C and 101 kPa.

Mountain View, CA; Monticello, UT; and Copenhagen, in Yabusaki et al. (10). MODFLOW simulated steady-state
Denmark. These PRBs were selected because hydrological flow in the pea gravel zones and iron media zone. The
and geochemical data were sufficiently well defined for initial hydraulic conductivity in the PRB was 186 m/d,
each PRB to permit an assessment of the simulator. and the porosity was 0.66. The hydraulic conductivity in
For all sites, the simulations were set up assuming the pea gravel zones was 86.4 m/d. A hydraulic gradient
homogeneous hydraulic conductivity in each medium (i.e., of 0.0003 was applied across the PRB to yield the Darcy
aquifer, reactive zone, etc.). Constant-head boundaries velocity (0.04 m/d) reported in Yabusaki et al. (10). Five
were applied along the upstream and downstream faces of minerals reported by Yabusaki et al. (10) were included
the domain to establish a hydraulic gradient orthogonal in the model: CaCO3 , FeCO3 , Fe(OH)2 (am), Mg(OH)2 ,
to the PRB. No flow conditions were imposed at the other and FeS (am). Composition of groundwater entering the
boundaries. Each species in the PRBs was assumed to have PRB was based on information provided in Yabusaki
an initial concentration of zero. A constant concentration et al. (10) (Table 2).
boundary was applied at the upstream face of the PRBs,
and other boundaries were assigned the Cauchy boundary Monticello
condition with no dispersive flux. Influent groundwater A full-scale funnel and gate PRB containing granular iron
concentrations entering the PRB were set at values was installed in 1999 to treat uranium in groundwater at
measured in the field.
a former uranium mill site near Monticello, UT (14). The
gate is 30.0 m long, 2.4 m wide (along flow direction), and
Moffett Federal Airfield
4.5∼6.0 m deep. A 1.2-m-thick reactive zone consisting of
A pilot-scale funnel-and-gate PRB consisting of ZVI was 100% granular ZVI was sandwiched between two 0.6-m-
installed in 1996 at Moffett Federal Airfield to remove thick layers of pea gravel (up gradient and down gradient)
TCE and DCE from groundwater (10). The gate is 3.0 m in the gate. Samples of groundwater were collected from
long (orthogonal to flow direction), 3.0 m wide (along the June 1999 (installation) to September 1999 and analyzed
flow direction), and 5.5 m deep. The gate consists of a for target contaminants, pH, major cations and anions,
1.8-m-thick barrier of granular ZVI sandwiched between and indicator parameters. Core samples showed that
two 0.6-m-thick layers of pea gravel (up gradient and magnetite and CaCO3 formed within the PRB.
down gradient). Groundwater samples collected within the Reactive transport in the gate at the Monticello PRB
PRB between April 1996 (installation) and October 1997 was simulated with the MODFLOW-RT3D simulator
were analyzed for target contaminants, pH, major cations using a one-dimensional model similar to that employed
and anions, and indicator parameters (10). Analysis of for the PRB at Moffett Federal Airfield (see previous
core samples showed that magnetite, hematite, aragonite, section). A one-dimensional model was selected because
and marcasite formed in the iron media. No direct flow through the pea gravel zones and ZVI was nearly one-
measurements were made of siderite, FeS, Fe(OH)2 (am), dimensional and steady state. Hydraulic conductivities of
Fe(OH)3 , brucite, or green rust. the ZVI and pea gravel zones were set at 30.9 m/d and 8.64
The MODFLOW-RT3D simulator simulated the nearly m/d. Based on data in Morrison et al. (14), the porosity
one-dimensional conditions in the gate at the Moffett was set at 0.50 in the ZVI and 0.30 in the pea gravel
Federal Airfield site. The input was based on data provided and the hydraulic gradient was set at 0.142 to yield a
 REACTIVE TRANSPORT IN THE SATURATED ZONE: CASE HISTORIES FOR PERMEABLE REACTIVE BARRIERS 521

Table 2. Concentrations of Aqueous Species in Influent Ground Water

Aqueous Species Concentrations (M)

Aqueous Species in Moffett Federal Airfield, Former Mill Site, Copenhagen Freight
Influent Ground Mountain View, Monticello, Yard, Copenhagen,
Water Californiaa Utahb Denmarkc

Fe2+ 1.0 × 10−10 1.0 × 10−10 1.0 × 10−10

Ca2+ 3.9 × 10−3 8.5 × 10−3 3.3 × 10−3
Mg2+ 2.7 × 10−3 3.6 × 10−3 1.2 × 10−3
OH− 1.3 × 10−7 5.2 × 10−8 5.0 × 10−7
HCO− 3 4.7 × 10−3 2.5 × 10−3 1.0 × 10−2
CO3 2− 1.4 × 10−6 1.0 × 10−6 1.0 × 10−6
O2 (aq) 1.0 × 10−10 1.0 × 10−10 1.0 × 10−10
NO3 − 3.9 × 10−5 1.0 × 10−10 1.0 × 10−10
SO4 2− 3.6 × 10−3 1.2 × 10−2 1.2 × 10−3
a
From Yabusaki et al. (10).
b
From Morrison et al. (14).
c
From Killerich et al. (19).

groundwater velocity of 3 m/d. Five minerals were included at Moffett Field after 12 months of operation. The reaction
in the geochemical model: CaCO3 , FeCO3 , Fe(OH)2 (am), rate coefficients initially were estimated from data in
CaMg(CO3 )2 , and FeS. Composition of the groundwater the literature, but ultimately they were adjusted by
entering the PRB was defined based on data reported in trial-and-error to reproduce the measured pH, alkalinity,
Morrison et al. (14), as summarized in Table 2. and concentrations of SO4 2− , Ca2+ , and Mg2+ . A similar
calibration procedure was used at the other two sites. The
Copenhagen calibrated reaction rate coefficients are summarized in
A full-scale PRB containing granular ZVI was installed Table 3 along with typical ranges based on data reported
in 1998 at the Copenhagen Freight Yard in Copen- in the literature.
hagen, Denmark to treat chlorinated solvents in ground The predicted pH is in general agreement with the
water (19). The PRB is a continuous trench 15.0 m long, observed pH [Fig. 1(a)], including the rapid jump near
1.3 m wide (along flow direction), and 6.0 m deep and the entrance face and the leveling out in the interior.
consists of a 0.9-m-thick zone of 100% granular ZVI sand- The drop in alkalinity just up gradient of the PRB is
wiched between 0.2-m-thick layers of pea gravel. Samples not captured [Fig. 1(b)], but the alkalinity predicted by
of groundwater were collected from July 1998 (installation) RT3D inside the PRB is similar to that measured in the
to September 1999 and were analyzed for target contam- field. The sulfate concentration predicted by RT3D also
inants, pH, and major cations and anions (19). No direct is in general agreement with the trends observed in the
measurements were made of minerals forming within the field [Fig. 1(b)], as are the Ca2+ and Mg2+ concentrations
PRB (19). [Fig. 1(c)]. However, the predicted Ca2+ concentrations
The MODFLOW-RT3D simulator was set up to in the PRB are near the upper bound of the measured
simulate one-dimensional flow and reactive transport concentrations [Fig. 1(c)]. The field data are also too sparse
along the centerline of the Copenhagen PRB. The model to evaluate whether the simulator accurately predicts the
predicted steady-state flow in the surrounding aquifer, pea rapid changes in geochemical conditions occurring at the
gravel zones, and ZVI. Hydraulic conductivity of the ZVI, entrance face.
pea gravel, and aquifer were set at 6.1, 2.1, and 0.6 m/d,
respectively, based on site conditions (19). The porosity Monticello Mill Site
was set at 0.60 for the ZVI, 0.40 for the pea gravel zones,
Comparisons of measured and predicted pH, alkalinity,
and 0.3 in the aquifer (19). A hydraulic gradient of 0.011
and concentrations of Ca2+ , Mg2+ , and SO4 2− for conditions
was applied across the aquifer based on data reported in
at 3 months of operation are shown in Fig. 2. Calibrated
Killerich et al. (19). Five minerals were included in the
rate coefficients for the PRB at Monticello are summarized
model: CaCO3 , FeCO3 , Fe(OH)2 (am), CaMg(CO3 )2 , and
FeS. Composition of the groundwater entering the PRB in Table 3. As with the PRB at Moffett Field, insufficient
was defined based on data reported in Killerich et al. (19), data exist to determine whether the rapid changes
as summarized in Table 2. predicted at the entrance face accurately replicate the
conditions occurring in the PRB at Monticello.
The measured and predicted pH, HCO3 − , Mg2+ , and
COMPARISON OF PREDICTED AND MEASURED SO4 2− concentrations in the Monticello PRB are in close
CONCENTRATIONS agreement (Fig. 2). However, the predicted Ca2+ concen-
trations in the PRB are higher than are the measured
Moffett Federal Airfield
concentrations [Fig. 2(b)]. One reason for this discrepancy
Comparisons of the predicted and measured pH and may be that minerals other than CaCO3 and CaMg(CO3 )2
concentration profiles are shown in Fig. 1 for the PRB are precipitating (these minerals only account for 15%
522 REACTIVE TRANSPORT IN THE SATURATED ZONE: CASE HISTORIES FOR PERMEABLE REACTIVE BARRIERS

Table 3. Calibrated Rate Coefficients in PRB Case Studies

Rate Coefficients
Geochemical Moffett Federal Former Mill Copenhagen Freight Reported in
Reactions Terms Units Airfieldc Sitec Yardc Literatured

Oxygen Corrosiona m3 /m2 -d 2.8 × 10−2 2.8 × 10−2 2.8 × 10−2 < 2.8 × 104
Water Corrosiona mole/m2 -d 7.0 × 10−7 1.5 × 10−5 1.9 × 10−7 3.0 × 10−8 − 5.4 × 10−3
Nitrate Corrosiona m3 /m2 -d 2.6 × 10−5 — 2.8 × 10−5 1.8 × 10−8 − 2.8 × 10−5
Microbial sulfate reductionb M/d 1.0 × 10−3 1.0 × 10−3 4.0 × 10−3 5.0 × 10−6 − 5.0 × 10−3
CaCO3 M/d 1.0 × 10−4 4.0 × 10−4 1.0 × 10−5 2.7 × 10−9 − 1.4 × 10−4
FeCO3 M/d 1.1 × 10−5 2.1 × 10−5 1.0 × 10−5 1.1 × 10−5 − 2.7 × 10−4
Fe(OH)2 (am) M/d 5.1 × 10−5 5.1 × 10−5 5.0 × 10−5 <2.2 × 10−4
FeS (am) M/d 1.0 × 10−6 1.0 × 10−6 1.0 × 10−6 1.1 × 10−8 − 2.2 × 10−5
CaMg(CO3 )2 M/d — 1.0 × 10−6 1.0 × 10−9 6.9 × 10−6
Mg(OH)2 M/d 2.0 × 10−3 NA NA 3.4 × 10−6
a
Reactive surface area of ZVI set at 3.9 × 106 m2 /m3 .
b
Half-saturation constants for SO4 2− and H2 (aq) were KSO4 = 1.0 × 10−4 M, and KH2 = 1.0 × 10−7 M, respectively.
c
After calibration with field measurements.
d
Li et al. (8).

(a) 12 (a) Pea gravel ZVI

ZVI
10
11 0.015
Concentration (M)

pH
SO42−
10 9
0.010
pH

pH

pH
9 PH - Field data
Pea gravel HCO3− - Field data 8
0.005
SO42− - Field data
8 Field pH data
HCO3−

7 0.000 7

(b) ZVI
0.005
Field alkalinity data
(b) 0.010
Field SO42− data
Concentration (M)

0.004 Alkalinity Ca2+ - Field data

Ca2+ Mg2+ - Field data
0.003 0.008
Concentration (M)

0.002 SO42−
0.006
0.001
Pea gravel
0.004 Mg2+
0.000

(c) 0.005
ZVI 0.002
Field Ca2+ data Pea gravel
0.004 Ca 2+ ZVI
Field Mg2+ data
Concentration (M)

0.000
0.003 0.0 0.6 1.2 1.8
Mg2+ Distance (m)
0.002 Figure 2. Measured and predicted concentrations for the PRB at
the former Uranium Mill Site in Monticello, UT, after 3 months
0.001 of operation: (a) pH, HCO3 − , and SO4 2− and (b) Ca2+ and Mg2+ .
Pea gravel

0.000
0.0 0.6 1.2 1.8 2.4
removal in the ZVI. However, neither of these minerals
Distance into PRB (m)
has been observed in the core samples from the PRB at
Figure 1. Measured and predicted concentrations for the PRB Monticello. The surface complexation of calcium on the
at Moffett Federal Airfield, Mountain View, CA, after 12 months ZVI and secondary minerals surfaces, neither of which are
of operation: (a) pH, (b) alkalinity and SO4 2− , and (c) Ca2+ and
considered in the simulator, may also be removing calcium
Mg2+ .
from the solution (10).

Copenhagen Site
of aqueous calcium). For example, Morrison et al. (14)
included iron-rich calcite (Fe0.4 Ca0.6 CO3 ) and gypsum Comparisons of measured and predicted pH and Ca2+ ,
(CaSO4 ) in their equilibrium model to account for calcium Mg2+ , SO4 2− , and HCO3 − concentrations after 15 month
 REACTIVE TRANSPORT IN THE SATURATED ZONE: CASE HISTORIES FOR PERMEABLE REACTIVE BARRIERS 523

of operation of the Copenhagen PRB are shown in Fig. 3. Although good agreement between measured and
The calibrated rate coefficients corresponding to these predicted concentrations has been shown in Figs. 1–3, the
predictions are summarized in Table 3. accuracy of the simulator when applied to other problems
Close agreement exists between the measured and will depend greatly on the rate coefficients used as input.
predicted pH and the HCO3 − , SO4 2− , Ca2+ , and Mg2+ For example, the effect of varying the rate coefficient for
concentrations in the ZVI of the Copenhagen PRB iron corrosion by water (Table 1) within a reasonable range
[Figs. 3(a) and (b)], but the field data are too sparse to (1.0 × 10−7 to 1.0 × 10−5 mole/m2 -d) on pH is illustrated
evaluate whether the trends in geochemistry predicted by in Fig. 4 with the model for the PRB at Moffett Field. The
the model are accurate. In contrast to the PRBs at Moffett influent concentrations used for the simulation are shown
Field and Monticello, the predicted pH climbs gradually in Table 2, and the rate coefficients for the reactions other
at the Copenhagen site. This more graduate change in pH than iron corrosion by water are shown in Table 3. When
may be caused by a lower rate of iron corrosion by water the rate coefficient for iron corrosion by water is 10−7
(1.9 × 10−7 mole/m2 -d, Table 3), which results in less OH− mole/m2 -d, the predicted pH is much lower than the pH
released into solution. observed in the field, especially near the entrance. In
contrast, when the rate coefficient for iron corrosion by
Comparison of Calibrated Rate Coefficients and Literature water is 10−5 mole/m2 -d, the predicted pH increases too
Values rapidly near the entrance and levels out approximately
one pH unit higher than observed in the field.
The calibrated rates coefficients for the three PRB case In situ rate coefficients can vary considerably and
studies are summarized in Table 3. Most of the calibrated depend on a variety of factors, and no method currently
rate coefficients fall within the range of rate coefficients is available to predict rate coefficients accurately a
from the literature (Table 3). The exceptions are the priori. Thus, at the current time, simulators must be
rate coefficient for Mg(OH)2 at the Moffett PRB and the calibrated with field data if reliable predictions of mineral
rate coefficient for CaCO3 at the Monticello PRB. Both precipitation are to be obtained. Alternatively, parametric
rate coefficients are higher than are the maximum values analyses can be conducted to evaluate the potential range
reported in the literature. However, the rate coefficient for of conditions that may be realized in the field.
CaCO3 is only slightly above the upper bound from the
literature, and only one rate coefficient was available in
SUMMARY
the literature for Mg(OH)2 . Thus, these exceptions are not
considered unusual. Comparisons between predicted and measured concentra-
tions of common groundwater ions in PRBs have been
described in this article. The simulator used for the pre-
(a) 0.012 Pea gravel 11 dictions is based on the public domain codes MODFLOW
ZVI pH - Field data
HCO3− - Field data and RT3D, the latter being augmented with a geochemical
0.010
SO42− - Field data algorithm specifically for simulating geochemical reac-
HCO3− 10
Concentration (M)

0.008 tions in PRBs. Field data from three PRBs were used for
the comparison.
0.006 The comparisons showed that pH profiles and anion and
pH

9
pH cation concentrations in PRBs can be predicted accurately.
0.004 The ability to predict solid phase concentrations has not
8 been evaluated because of lack of field data. However,
0.002
SO42−
0.000
7 12
Pea gravel ZVI
(b)
Ca2+ - Field data
Ca2+ Mg2+ - Field data 11
0.003
Concentration (M)

10
pH

0.002

9
Mg2+
0.001 Field pH data
Pea gravel 8 k = 10−7 mole/m2-d
ZVI k = 7 × 10−7 mole/m2-d
0.000 k = 10−5 mole/m2-d
1.0 1.2 1.4 1.6 1.8 2.0 7
Distance (m) 0.0 0.6 1.2 1.8 2.4
Figure 3. Measured and predicted concentrations at the center- Distance into PRB (m)
line of the PRB at the Copenhagen Freight Yard, Copenhagen, Figure 4. Measured and predicted pH for the PRB at Moffett
Denmark, after 15 months of operation: (a) pH, HCO3 − , and Federal Airfield, Mountain View, CA, after 12 months of operation
SO4 2− and (b) Ca2+ and Mg2+ . using various rate coefficients (k) for iron corrosion by water.
524 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER

the accuracy of the predictions depends to a great extent by progressive irreversible dissolution of zerovalent iron.
on the reaction rate coefficients used as input. In this Environ. Sci. Technol. 35: 385–390.
study, the rate coefficients were obtained by calibration 15. McDonald, M.G. and Harbaugh, A.W. (1988). MODFLOW: A
because the actual in situ rate coefficients were unknown. Modular Three-Dimensional Finite-Difference Ground-Water
Thus, predictions made a priori are prone to considerable Flow Model. U.S. Geological Survey Techniques of Water-
uncertainty until methods are developed to accurately Resources Investigations, Washington, DC, book 6, chap. A1.
define reaction rate coefficients. 16. Clement, T.P. (1997). RT3D: A Modular Computer Code
In many cases, the simulator predicted that the greatest for Simulating Reactive Multi-Species Transport in 3-
Dimensional Groundwater Aquifers, PNNL-SA-11720. Pacific
changes in geochemistry generally occur at the entrance
Northwest National Laboratory, Richland, WA.
face of the PRB. However, the current field data sets
17. Reardon, E.J. (1995). Anaerobic corrosion of granular iron:
contain insufficient detail to confirm that these changes
Measurement and interpretation of hydrogen evolution rates.
in geochemistry are being predicted accurately. More
Environ. Sci. Technol. 29: 2936–2945.
detailed field observations are needed at the entrance
18. Lasaga, A.C. (1998). Kinetic Theory in the Earth Sciences.
face of the PRBs to more fully evaluate whether the
Princeton series in geochemistry. Princeton University Press,
simulator can accurately capture these rapid changes in Princeton, NJ.
geochemistry.
19. Killerich, O., Larsen, J.W., and Nielsen, C. (2000). Field
results from the use of a permeable reactive wall. Proceedings
BIBLIOGRAPHY of the Second International Conference on Remediation of
Chlorinated and Recalcitrant Compounds, Monterey, CA.
1. Blowes, D.W. et al. (2000). Treatment of inorganic contam-
inants using permeable reactive barriers. J. Contaminant
Hydrol. 45: 123–137. TRANSPORT OF REACTIVE SOLUTE IN SOIL
2. Wilkin, R.T., Puls, R., and Sewell, G.W. (2003). Long-term
AND GROUNDWATER
performance of permeable reactive barriers using zero-valent
iron: geochemical and microbiological effects. Ground Water
NATALIE L. CAPIRO
41: 493–503.
PHILIP B. BEDIENT
3. EPA. (1998). Permeable Reactive Barrier Technologies for
Contaminant Remediation. United States Environmental Rice University
Protection Agency, Washington, DC. Houston, Texas
4. Lee, T. and Benson, C.H. (2004). Sorption and degradation of
alachlor and metolachlor in ground water using green sands.
J. Environ. Quality 33: 1682–1693. INTRODUCTION
5. Lee, T., Benson, C.H., and Eykholt, G.R. (2004). Waste green
sands as reactive media for groundwater contaminated Remediation of contaminated aquifers, both vadose zone
with trichloroethylene (TCE). J. Hazardous Materials B109: (unsaturated portion of an aquifer) soils and groundwater,
25–36. has proven to be a challenging environmental endeavor.
6. Mackenzie, P.D., Horney, D.P., and Sivavec, T.M. (1999). Groundwater is the source of about 60% of the United
Mineral precipitation and porosity losses in granular iron States’ water usage and, with contaminant plumes
columns. J. Hazardous Materials 68: 1–17. impacting many states, associated human health concerns
7. Phillips, D.H. et al. (2003). Mineralogical characteristics and of great importance exist. Groundwater can become
transformations during long-term operation of a zerovalent polluted both directly by leaking underground storage
iron reactive barrier. J. Environ. Quality 32: 2033–2045. tanks and indirectly by leakage of contaminants on the
8. Li, L., Benson, C.H., and Mergener, E.A. (2004). Impact of soil surface that leach to the groundwater. Common
mineral fouling on hydraulic behavior of permeable reactive pollutants include agricultural chemicals applied to crops,
barriers. Ground Water. In press.
industrial spills, landfill leachate, disposal sites, and waste
9. Vikesland, P.J. et al. (2003). Longevity of granular iron in ponds. Subsurface remediation is complicated by not only
groundwater treatment processes: changes in solute transport
the behavior of the contaminants themselves but by the
properties over time. J. Contaminant Hydrol. 64: 3–33.
heterogeneities found in typical aquifer formations. The
10. Yabusaki, S. et al. (2001). Multicomponent reactive transport
processes that govern the fate and transport of solute
in an in situ zero-valent iron cell. Environ. Sci. Technol. 35:
in aquifer systems generally fall into the categories of
1493–1503.
hydrodynamic, abiotic, and biotic.
11. Mayer, K.U., Blowes, D.W., and Frind, E.O. (2001). Reactive
transport modeling of an in situ reactive barrier for the
The universal process that impacts all forms of
treatment of hexavalent chromium and trichloroethylene in solute include the hydrodynamic processes, which affect
groundwater. Water Resour. Res. 37: 3091–3103. contaminant transport by impacting the flow paths of
12. Li, L. (2004). Fouling and the Long-Term Performance of groundwater in the subsurface, including both the vadose
Permeable Reactive Barriers. Ph.D. Dissertation, University zone and the saturated aquifer. Abiotic processes affect
of Wisconsin-Madison, Madison, WI. contaminant transport physically by causing interactions
13. Liang, L. et al. (2002). Integration of Geochemical Data and between the contaminant and the stationary subsurface
Hydraulic Changes at Permeable Reactive Barriers Sites. Oak materials (i.e., sorption) or by affecting the form of the
Ridge National Laboratory, Oak Ridge, TN. contaminant (i.e., through hydrolysis and redox reactions).
14. Morrison, S.J., Metzler, D.R., and Carpenter, C.E. (2001). Biotic processes are driven by microbial activities and
Uranium precipitation in a permeable reactive barrier can affect contaminant transport by metabolizing or
 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER 525

mineralizing the contaminant (i.e., organic contaminants), concentration to an area of lower concentration. Diffusive
or possibly by using the contaminant in the metabolic transport can occur in the absence of velocity. Mass
process (i.e., nitrate under denitrifying conditions). transport in the subsurface because of diffusion in one
dimension can be described by Fick’s first law of diffusion,
PHYSICAL MECHANISMS OF SOLUTE TRANSPORT
dC
fx = −Dd (3)
The three basic hydrodynamic processes affecting con- dx
taminant transport in saturated and unsaturated porous
media are advection, diffusion, and hydrodynamic disper- where fx is the mass flux [M/L2 /T], Dd is the diffusion
sion. Advection represents the movement of a contaminant coefficient [L2 /T], and dC/dx is the concentration gradi-
with the flowing groundwater according to the seepage ent [M/L3 /L].
velocity in the pore space. Advection is the most important Typical values of Dd are relatively constant and range
process driving the down gradient migration of dissolved from 10−4 to 10−5 cm2 /sec (3). In cases of very low
contaminants in the saturated subsurface (1). If advection velocities, such as in a tight soil or clay liner, or in
was the only process acting on a tracer, a sharp inter- the case of mass transport involving very long time
face would exist continuously as the tracer flushes out periods, diffusion is the dominant hydrodynamic process.
the solute-free water. In essence, the tracer would behave On the other hand, in areas of high groundwater velocity,
as a plug. A breakthrough curve, illustrating the rela- molecular diffusion can actually be ignored compared with
tive effluent concentration (C/Co) vs. time in a continuous advection and other dispersion mechanisms.
tracer input, would appear as a step function (Fig. 1a). Freeze and Cherry (4) defined hydrodynamic dispersion
The average velocity of advective transport is described by as the process in which solutes spread out and are diluted
the amended Darcy equation: compared with simple advection alone. It is defined as
the sum of molecular diffusion and mechanical dispersion,
K dh where diffusion is caused by concentration gradients alone
vx = − (1) and dispersion is generally caused by local variations in
n dL
velocity around some mean velocity of flow. The net effect of
where vx is the average linear groundwater velocity the dispersive phenomena is to decrease the concentration
[L/t], K is the hydraulic conductivity (a measure of a gradient between two different solutions of the same
porous media’s ability to transmit water) [L/t], n is the solute. In essence, dispersion reduces the sharp interface
effective porosity [unit less], and dh/dL is the hydraulic between the two solutions. Mass transport because of
gradient [L/L]. dispersion can also occur normal to the direction of flow.
The average linear velocity or seepage velocity, vx , This transverse dispersion, Dy , is caused by diverging
equals the Darcy velocity divided by effective porosity, flowpaths in the porous media and concentration gradients
n, associated with the pore space through which water that cause mass to spread laterally from the main direction
can actually flow. The average linear velocity is less of flow. In most cases involving a two-dimensional plume
than the microscopic velocities of water molecules moving of contamination, Dy is much less than Dx , and the shape of
along individual flowpaths, because of tortuosity. The one- the plume tends to be elongated in the direction of flow (5).
dimensional (1-D) mass flux (Fx ) because of advection Many typical contaminant plumes in groundwater are
equals the product of water flow and concentration (C) of represented by 2-D advective-dispersive mechanisms.
solute, or Fx = vx nC.
Darcy’s law is also still valid for the macroscopic view One-Dimensional Transport Equations
(through the entire pore body) of unsaturated flow, except
that hydraulic conductivity is now a function of moisture The governing mass transport equation is difficult to solve
content. Darcy’s law is then used with the unsaturated in field cases of practical interest because of boundary
value for K and can be written as irregularities and variations in aquifer characteristics,
so numerical methods must generally be employed (6).
v = −K(θ )dh/dz (2) However, a limited number of relatively simple, 1-D
analytical solutions exist. The simplifying assumptions
where v is Darcy velocity, z is depth below surface [L], K(θ ) include the following: (a) the tracer is ideal, with constant
is unsaturated hydraulic conductivity [L/T], θ is volumetric density and viscosity; (b) the fluid is incompressible;
moisture content, h is potential or head (h = z + ψ) [L/T], (c) the medium is homogeneous and isotropic; and (d) only
and ψ is tension or suction [L]. For saturated soils, saturated flow is considered. Initial conditions (t = 0) in a
the hydraulic conductivity was assumed to be constant; soil column are usually set to zero (C(x, 0) = 0) or to some
however, in unsaturated soil, it decreases sharply with constant background concentration. Boundary conditions
decreasing water content, because as water drains from must be specified at the two ends of the 1-D column. For
pores, the flow decreases (2). a continuous source load at x = 0, the concentration is set
Although advection is associated with bulk macroscopic to C(0, t) = C0 for t > 0. The concentration at the other
groundwater movement, diffusion is a molecular-based boundary, x = ∞ is set to zero C(x, ∞) = 0 for t > 0.
phenomenon. The process involves spreading because of For an infinite 1-D column with background concentra-
concentration gradients and random motion. Diffusion tion of zero and a continuous input tracer concentration C0
causes a solute in water to move from an area of higher at −∞ ≤ x ≤ 0 for t ≥ 0, Bear (7) solves the problem using
526 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER

the Laplace transform at x = L, length of the column which continues to decrease in maximum concentration

 
   because of spreading out with time in the direction of flow.
C(x, t) 1 L − vx t vx L L + vx t For the spike source, the amount of mass under each curve
= erfc √ + exp erfc √
C0 2 2 Dx t Dx 2 Dx t is identical if the tracer is conservative.
(4) Additional information can be found on transport
where erfc is the complementary error function, erfc(x) = equations in several sources (3–6,8), including three-
√ x dimensional derivations and those incorporating degra-
1 − erf(x) = 1 − (2/ π ) e−u du, Dx is the coefficient
2

0 dation.
of hydrodynamic dispersion, and vx is the average
seepage velocity. The center of mass (C/C0 = 0.5) of Techniques for Determining Hydrodynamic Parameters
the breakthrough curve travels with the average linear Tracers and tracer tests are valuable methods for
velocity, vx , and corresponds to the point where x = vx t. estimating dispersion and advection below the water
Note that the second term on the right-hand side of Eq. (4) table. In a tracer test, a chemical species is introduced
can generally be neglected for most practical problems. to the subsurface and monitored as it is carried by
The error functions erf (β) and erfc(β) are tabulated groundwater through the subsurface in the direction of
in Domenico and Schawartz (3). flow. Average seepage velocity can be determined by
The corresponding solution can be derived for the analyzing the resulting curves in time and space, and
injection of a tracer pulse (instantaneous input) at x = 0 hydraulic conductivity (K) values can be computed from
with background concentration equal to zero in the 1-D Darcy’s Law (Eq. 1). One of the most successful field sites
column. As the slug moves downstream with vx in the +x for tracer studies has been the Borden landfill site in
direction, it spreads out according to Canada, which resulted in a series of classic papers on
  dispersion and adsorption processes measured during a
M (x − vx t)2 multiyear field experiment (9,10).
C(x, t) = exp − (5)
(4π Dx t)1/2 4Dx t Other methods of determining hydraulic conductivity
exist. Permeameters may be used in a laboratory setting
where M is the injected mass per unit cross-sectional area. to determine K, whereas slug tests and pump tests are
Figure 1b shows the resulting Gaussian distribution of recommended for field sites. A slug test involves suddenly
concentration for the instantaneous pulse source in one adding or removing a volume of water (called a slug) from
dimension. a well and monitoring the time it takes for the water
Plots comparing the shapes of Eqs. (4) and (5) are level to return to normal. This test is convenient, but
shown in Figs. 1a and 1b, respectively, at x = L at the end not as accurate as a tracer test to determine the speed
of a soil column. The differences between instantaneous with which a contaminant moves through an aquifer. The
(spike) source and continuous source transport problems pump test involves the constant removal of water from a
are obvious in one dimension. The continuous source single well and the subsequent observation of drawdown
produces a response or breakthrough curve, which starts in adjacent wells. This test gives the K value for a large
at a low concentration and eventually levels off to the part of an aquifer, but may produce a large volume of water
initial input concentration C0 as a function of time. The that requires proper handling and treatment (if performed
spike source produces a normal or Gaussian distribution, in a contaminated area) (3).

(a)
Step function, no dispersion
Co
1.0
Effects of dispersion
C/C0

vx 0.5
Center of mass
L 0.0
t
t = L/vx

(b)
No dispersion, only advection

With dispersion
C/C0

L
t
Figure 1. Breakthrough curves in one dimen-
Center of mass
sion for: (a) continuous and (b) instantaneous
t = L/vx
tracer inputs.
 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER 527

ABIOTIC SOLUTE TRANSPORT Cosolvation and Ionization

Cosolvation and ionization are two examples of processes

Abiotic processes affect contaminant transport by causing that act to increase the chemical solubility, by altering
interactions between the contaminant and the aquifer the solvent phase or by altering the chemical form,
media or by changing the form of the contaminant respectively. Cosolvation is the process by which the
that may subsequently interact with the subsurface mobility of one contaminant is enhanced by the presence
material. The interfaces of most interest in subsurface of a solvent (19), whereas ionization refers to the gaining
fate and transport are liquid/solid interfaces and gas/solid or losing of protons as a function of pH and thus going
interfaces. The processes that detail these interactions are from a neutral form to an ionic form. These processes
discussed in this section. are solvent motivated; when a contaminant is in the
presence of a solvent, it will have decreased sorption and
Sorption an increased solubility leading to elevated concentrations
in the surrounding solute. Based on these assumptions,
Sorption includes the phenomena of adsorption, absorp- it would be expected that as the fraction of cosolvent
tion, and ion exchange. Weber (11) defines adsorption as increases in solution, the level of sorption will decrease.
the accumulation occurring at the interface, absorption These attributes of cosolvents lead to increased mobility
as the partitioning between two phases (accumulation of originally slow dissolving hydrophobic (lacking affinity
from groundwater into organic carbon), and sorption as for water) contaminants. An example of this is the affect of
including both adsorption and absorption. Often, the terms ethanol on petroleum hydrocarbons. The issue concerning
adsorption and sorption are interchangeable. an ethanol releases’ potential to cosolubilize benzene and
Many organic contaminants are removed from solution other aromatic compounds in groundwater (20) has been
by sorption onto the aquifer matrix. Sorption of dissolved addressed in recent groundwater modeling studies (21,22)
contaminants onto the aquifer matrix results in slowing and in laboratory studies (23,24). The reason that
(retardation) of the contaminant relative to the average tremendous efforts are being employed to understand this
groundwater seepage velocity and a reduction of dissolved phenomenon is that benzene, which is known to increase
organic concentrations in groundwater. In some cases, cancer risks, has a maximum contaminant level of only
sorption is a reversible process; at a given solute concen- 0.005 mg/L in drinking water as set by the U.S. EPA and
tration (steady state), some portion of the contaminant is ethanol is now being used as a fuel additive (an oxygenate).
partitioning out of solution onto the aquifer matrix and
some portion is desorbing from the matrix back to solution; Oxidation and Reduction
this process is called desorption.
The ability of chemicals to completely desorb from a Oxidation and reduction (or redox) refer to the transfer
solid is an area of current study (12–15). At high con- of electrons and contaminant species change of ions or
taminant concentration, desorption patterns are generally compounds. Oxidation refers to the loss of electrons and
predictable through equilibrium analysis. However, the reduction refers to the gain of electrons in a reaction.
phenomena of solid phase contaminant concentrations The two processes often occur in tandem; electrons
exceeding the predicted concentration, based on the sur- given by one compound must be gained by another
rounding aqueous phase concentrations, is observed in compound. Inorganic compounds, such as reduced metals
materials that have been contaminated for long time and reduced sulfur, are possible mediators of oxidation-
periods and contain low solid-phase contaminant concen- reduction reactions (25–27). The presence or absence of
trations (nominally less than 20 mg/kg). The importance inorganic species is often dependent on aquifer conditions.
of this observation is potentially enormous, because site A common condition is that many anaerobic environments
remediation requirements include the removal of sorbed will contain elevated levels of Fe(II) or Fe2+ , H2 S/HS− ,
contamination. If a contaminant does not desorb as read- and other reduced species that are capable of donating
ily as it sorbed, the ability to meet cleanup levels may be electrons and reducing certain contaminants. For example,
seriously threatened (16). the two half reaction equations for the oxidation-reduction
Ion exchange is a specific category of sorption that of ferrous iron (Fe2+ ) to ferric iron (Fe3+ ) are:
occurs when the absorbent charge deficiency of a surface
can be neutralized more efficiently by ions in solution 4 Fe2+ ⇔ 4 Fe3+ + 4e− [oxidation]
than those ions currently absorbed (17). An example of ion
exchange that is familiar is a home water softener. Home O2 + 4H+ + 4e− ⇔ 2H2 O[reduction]
water softeners operate by accumulating the divalent
cations from the groundwater onto the absorbent (ion Biologically mediated oxidation-reduction reactions are
exchange resin) while releasing monovalent ions, which often more rapid than abiotic reactions for many con-
reduces water hardness. In the subsurface, the resin would taminants of interest in groundwater systems, and thus
be natural minerals such as those found in clays. The most abiotic oxidation-reduction are often neglected in trans-
common ions occurring naturally are as follows (listed port calculations. In cases where they become important,
in order of occurrence from high to low): cations, Ca2+ , oxidation-reduction processes are generally introduced as
Mg2+ , Na+ , and K+ ; anions, SO4 2− , Cl− , PO4 3− , and NO3 − a first-order reaction (rate = k1 X) with respect to contam-
(18). inant concentration, analogous to hydrolysis.
528 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER

Hydrolysis Additionally, volatilization from NAPLs in the vadose zone

or floating on the water table can yield high concentrations
Hydrolysis refers to a chemical reaction between a
of contaminants in the gas phase (16).
contaminant molecule and water. Hydrolysis may function
when direct biodegradation (microbially driven) of the
contaminant cannot, and hydrolysis may produce a BIOTIC PROCESSES AFFECTING SOLUTE TRANSPORT
product that is susceptible to biodegradation, which can
be significant in affecting fate and transport (17). With Biotic processes include those biological phenomena that
respect to organics, this process frequently results in the affect the transport and/or fate of compounds (organics,
formation of alcohols and alkenes (28). inorganics, or the microbes themselves) in the subsurface.
The rate that a chemical undergoes hydrolysis is Biodegradation of organics is one process that is of partic-
strongly influenced by the temperature and pH of the ular interest to minimize human health risks associated
system (27). As temperature increases, hydrolysis rates with the widespread contamination of chlorinated ethenes
increase. As pH becomes either acidic or alkaline (e.g., and hydrocarbons in the environment, because this process
greater than or less than 7.0), rates of acid-catalyzed is able to achieve mineralization (endproducts CO2 ,H2 O,
or base-catalyzed hydrolysis can increase. Typically, etc.) and effectively render the contaminant innocuous
hydrolysis is described with a first-order rate expression (harmless). It has only been in the last two decades that
as shown in Eq. 6 the variety and magnitude of micro-organisms present
C in aquifers, and thus the ability of aquifers to support
= e−khyd t (6) biodegradation, has been realized (33). Laboratory col-
C0
umn studies (34) and improved field sampling methods
where C is the concentration of the chemical species and are chiefly responsible for these advances.
khyd is a first-order hydrolysis rate coefficient. Knowing
factors, such as pH and temperature, is important Energy Flow and Metabolism
in determining the appropriate rate coefficient for a All living organisms require energy to survive, and
particular contaminant. the capture of usable energy through the process of
catabolism is a significant part of an organism’s overall
Volatilization metabolism. Bacteria that obtain energy from chemical
The transfer of a contaminant from the aqueous phase, forms are classified as chemotrophs. Chemotrophic
nonaqueous phase liquid (NAPL), or sorbed phase organisms are further differentiated as either lithotrophs,
directly to the gas phase is a process referred to as which use inorganic sources of energy, or organotrophs,
volatilization. The rate and extent to which volatilization which use organic compounds. Examples of lithotrophic
occurs is strongly influenced by a number of parameters substrates include ammonia (NH3 ), hydrogen (H2 ), ferrous
including the contaminant phase, the contaminant’s vapor iron (Fe2+ ), and sulfide (HS− ). Examples of common
pressure, environmental factors (e.g., temperature and groundwater contaminants that serve as organotrophic
others), proximity in respect to the vadose zone, and substrates include fuel components and chlorinated
other site-specific parameters (29,30). As a result of the solvents (35).
range of factors, that play an important role in vapor Biodegradation involves the process of extracting
transport, which include the physical transport processes energy from organic chemicals via oxidation of the organic
of advection and dispersion, it is difficult to calculate the chemicals during catabolism. As a chemical is oxidized, it
contribution of volatilization to the fate of chemicals in loses electrons. For this reason, chemotrophic substrates
groundwater. are often referred to as primary electron donors. Electrons
For the evaluation of volatilization, however, consider a lost during oxidation are coupled with the reduction of
closed bottle that contains water and air. If a contaminant electron acceptors. Eventually, the complex sequence of
is introduced to the bottle and allowed to equilibrate, some oxidation-reduction reactions that occur during catabolism
fraction of the contaminant added will reside in the gas results in the reduction of a terminal electron acceptor.
phase. The distribution of a chemical between the water Common terminal electron acceptors in groundwater
and the gas at equilibrium is described by the Henry’s systems include oxygen (O2 ), nitrate (NO3 − ), ferric iron
law coefficient (31,32). Mathematically, Henry’s law is (Fe3+ ), sulfate (SO4 2− ), and carbon dioxide (CO2 ). Aerobic
presented in Eq. (7), where Pc is the partial pressure of catabolism occurs when oxygen is the terminal electron
the contaminant, [C]aq is the aqueous phase concentration acceptor, and anaerobic conditions are in place when
of the contaminant, and Hc is the Henry’s constant. all other terminal electron acceptors are used. Typically,
an individual bacterial strain is capable of using only
Pc one terminal electron acceptor. Facultative aerobes are
Hc = (7)
[C]aq the exception to this rule; these organisms can use
nitrate as a terminal electron acceptor in the absence
In groundwater systems that are isolated from the of the preferred oxygen. An important evaluation of
vadose zone, Henry’s law can be applied to estimate the contaminant biodegradation potential involves an analysis
distribution of contaminants between the aqueous phase of the availability of primary electron donors and terminal
and the gas phase. If direct exchange is possible with the electron acceptors, as their presence is required for
vadose zone, equilibrium may not be achieved and the organisms to obtain energy for survival and selects the
calculation of gas phase concentrations is more difficult. types of organisms that will be present (16).
 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER 529

The metabolism of organic contaminants can be broadly accidental spills, improper storage, and disposal tech-
differentiated by the ability of organisms to use the niques, these chemicals have become major subsurface
contaminant for catabolic processes. If the compound is contaminants (41). The extensive impacts observed are
a primary electron donor and provides the bacterium with primarily because of their widespread usage by industry,
energy for cell growth, the contaminant is referred to their general mobility and persistence in the environment,
as a primary substrate, whereas secondary substrates and the health risks associated with compound exposure.
are contaminants that provide the cell with energy, but The metabolism of chlorinated solvents is perhaps
are present at concentrations that are not sufficient to more diverse than any other group of environmental
support the energy requirements of the organism. If contaminants. Depending on the compound of inter-
a compound is metabolized fortuitously as the cell is est, the electron acceptor condition, and the presence
obtaining energy from another primary electron donor, of inducing substrates, the metabolism of chlorinated
the transformation is referred to as cometabolic (36,37). solvents may occur through primary metabolism, sec-
However, cometabolism is generally a slow process and ondary metabolism, cometabolism, or through termi-
not all chlorinated solvents are susceptible (e.g., PCE and nal electron acceptor processes (16). Laboratory stud-
carbon tetrachloride). In the case of chlorinated ethenes, ies (34,36,37,42–44) have shown that, in the presence
the contaminant serves as an anaerobic terminal electron of a suitable substrate, certain microbes are capable of
acceptor in a process referred to as dehalorespiration, anaerobically metabolizing PCE by using it as an electron
which is faster than cometabolism. acceptor. In this process, known as reductive dechlorina-
tion, PCE may be dechlorinated to 1,1,2-trichloroethene
(TCE), to typically cis-1,2 dichloroethene (DCE), to vinyl
Metabolic Pathways of Common Contaminants
chloride, and eventually to ethene and ethane, an innocu-
An overview of how bacteria are capable of metabolizing ous compound (45).
two of the primary groundwater contaminants (hydrocar-
bons and chlorinated solvents) is presented in this section. SUMMARY OF SOLUTE TRANSPORT PHENOMENA
It is important to note that this continues to be an area
of extensive research. Theoretical frameworks continue Fate and transport of solutes in subsurface soils and
to change and it is difficult to assert definitive analysis groundwater is impacted by a combination of many
regarding the totality of metabolic processes that may processes that are difficult to quantify. No two field sites or
impact an individual contaminant’s fate. aquifers have the exact same characteristics. Soil type, pH,
As a result of accidental spills and leaking under- temperature, groundwater flow velocity, the availability of
ground storage tanks, hydrocarbon compounds consisting electron donors for microbes, and the contaminant itself
of gasoline, diesel, and jet fuels are common groundwater (among other parameters) all influence the path and
contaminants. Of particular concern are benzene, toluene, lifetime of the pollution source. However, measurement
ethylbenzene, and xylene isomers (BTEX), which, under techniques are improving, and understanding these
aerobic conditions, are rapidly biodegraded as primary processes is an area of on going research.
substrates (38). Unfortunately, the oxygen demand result- In summary, nonreactive (conservative) chemicals tend
ing from the aerobic oxidation of these compounds can to move unimpeded through the subsurface and are sub-
exceed the solubility of oxygen in water. Dissolved oxy- ject to only hydrodynamic processes such as advection,
gen concentrations are rapidly depleted because of the dispersion, and diffusion. On the other hand, many actual
biodegradation of these and other fuel constituents, yield- contaminants (organic solvent and fuels, and inorganic
ing anaerobic conditions in contaminated aquifers (39). metals) can be affected during groundwater transport by a
Only recently have laboratory studies confirmed that number of subsurface mechanisms and chemical reactions.
the biodegradation of all BTEX compounds occurs under The processes that have the largest impact on contaminant
all anaerobic electron acceptor conditions (listed above). fate and transport include advection, dispersion, sorption,
From these studies, it appears that toluene is the most cosolvation, volatilization, biological oxidation-reduction
degradable under all electron acceptor conditions and reactions, and cometabolism/mineralization. These pro-
benzene is the least degradable. Certain modeling studies cesses are reviewed in Table 1, and Fig. 2 provides an
of certain well-characterized contaminated field sites infer illustration of these processes acting on a typical contami-
that anaerobic benzene degradation may occur in situ. nant spill.
Only several recent laboratory studies conducted under
anaerobic conditions have demonstrated that benzene
BIBLIOGRAPHY
may degrade in the absence of oxygen (40). Despite the
possibility of anaerobic degradation of BTEX compounds,
1. Wiedemeier, T.H., Rifai, H.S., et al. (1999). Natural Attenu-
it is still generally slower than rates that are observed ation of Fuels and Chlorinated Solvents in the Subsurface.
through aerobic processes. John Wiley & Sons, New York.
Chlorinated methanes, ethanes, and ethenes comprise 2. Wierenga, P.J. (1995). Water and solute transport and storage,
a group of compounds commonly referred to as chlori- Handbook of Vadose Zone Characterization and Monitoring.
nated solvents. Chlorinated ethenes, particularly tetra- CRC Press, Boca Raton, FL, pp. 41–60.
chloroethene (PCE) and 1,1,2-trichloroethene (TCE), are 3. Domenico, P.A. and Schwartz, F.W. (1998). Physical and
commonly used as degreasers and cleaning solvents in the Chemical Hydrogeology. 2nd Edn. John Wiley & Sons, New
dry cleaning, aviation, and computer industries. Through York.
530 TRANSPORT OF REACTIVE SOLUTE IN SOIL AND GROUNDWATER

Table 1. Subsurface Processes Affecting Solute Transport

Category Process Contaminant (Solute) Behavior

Hydrodynamic Solute Transport Advection Plug flow; movement with bulk fluid
Diffusion Movement from high to low concentration
Dispersion Spreads out and dilutes
Abiotic Solute Transport Sorption Decreased solubility
Ion exchange Neutralizes the charge and sorbs
Cosolvation Increased solubility by solvent
Oxidation-Reduction Transfer of electrons
Hydrolysis More susceptible to biodegradation
Volatilization Phase transfer from aqueous or sorbed to gas
Biotic Transport Influences Cometabolism Metabolized as cell obtains energy
Mineralization Broken down to benign products
Oxidation-Reduction Microbial driven transfer of electrons

Surface spill of gasoline 11. Weber, W.J., Jr. (1972). Physiochemical Process for Water
Volatilization Quality Control. John Wiley & Sons, New York.
12. Fu, G., Kan, A.T., and Tomson, M.B. (1994). Environ. Toxicol.
Chem. 13(10): 1559–1567.
Diffusion
13. Kan, A.T., Fu, G., and Tomson, M.B. (1994). Environ. Sci.
Unsaturated Technol. 28: 859–867.
(Vadose) zone Dispersion
14. Kan, A.T., Fu, G., Hunter, M.A., and Tomson, M.B. (1997).
Environ. Sci. Technol. 31: 2176–2185.
15. Pignatello, J.P. and Xing, B. (1995). Environ. Sci. Technol.
Sorbed contaminant on soil 10: 1–11.
Volatilization 16. Hughes, J.B. (1999). Ground Water Contamination, Trans-
port and Remediation. Prentice-Hall, Upper Saddle River,
Capillary Fiinge NJ, pp. 203–240.
Concentrated LNAPL 17. Knox, R.C., Sabatini, D.A., and Canter, L.W. (1993). Subsur-
Dissolved plume face Transport and Fate Processes. Lewis Publishers, Boca
Water table
Raton, FL.
Diffusion Oxidation-reduction reactions
18. Mitchell, J.K. (1976). Fundamentals of Soil Behavior. John
Ground water flow Advection Wiley & Sons, New York.
Dispersion
Desorption 19. Gorelick, S.M., Freeze, R.A., Donohue, D., and Keely, J.F.
Hydrolysis +
biode gradation (1993). Groundwater Contamination: Optimal Capture and
Saturated zone Sorption Containment. Lewis Publishers, Boca Raton, FL.
20. Rice, D., Powers, S.E., and Alvarez, P.J.J. (1999). Potential
scenarios for ethanol-containing gasoline released into
Clay layer Iron exchange surface and subsurface waters. In: Health and Environmental
Assessment of the Use of Ethanol as a Fuel Oxygenate. Report
Figure 2. A typical scenario showing the processes impacting a to the California Environmental Policy Council in Response
surface gasoline spill, and its transport through the vadose zone to Executive Order D-5-99. Vol. 4.
to the groundwater. 21. Molsen, J.W., Barker, J.F., Frind, E.O., and Schirmer, M.
(2002). Water. Res. Research 38(1).
22. Powers, S.E. et al. (2001). Crit. Rev. Environ. Sci. Technol.
31: 79–123.
4. Freeze, R.A. and Cherry, J.A. (1979). Groundwater. Prentice- 23. Heermann, S.E. and Powers, S.E. (1998). J. Contam. Hydrol.
Hall, Englewood Cliffs, NJ. 34: 315–341.
5. Charbeneau, R.J. (2000). Groundwater Hydraulic and Pollu- 24. Rixey, W.G. and He, X. (2001). Dissolution characteristics
tant Transport. Prentice-Hall, Upper Saddle River, NJ. of ethanol from NAPL sources and the impact on BTEX
6. Bedient, P.B., Rifai, H.S., and Newell, C.J. (1999). Ground groundwater concentrations. In: Proceedings of the Conference
Water Contamination, Transport and Remediation. Prentice- on Petroleum Hydrocarbons and Organic Chemicals in
Hall, Upper Saddle River, NJ. Groundwater. November 14–16, Houston, TX, Ground Water
7. Bear, J. (1979). Hydraulics of Ground Water. McGraw-Hill, Publishing Company, Dublin, OH, pp. 41–52.
New York. 25. Kleeka and Gonsior. (1984). S.J. Chemosphere 13: 391–402.
8. Fetter, C.W. (1993). Contaminant Hydrogeology. Macmillian 26. Macalady, D.L. and Wolfe, N.L. (1984). Abiotic hydrolysis
Publishing Company, New York. of Sorbed Pesticides. ASC Symposium Series, No. 259, pp.
9. Mackay, D.M., Freyberg, D.L., Roberts, P.V., and Cherry, J.A. 221–244.
(1986). Water Resources Res. 22(13): 2017–2029. 27. Schwartzenbach, R.P., Gschwend, P.M., and Imboden, D.M.
10. Roberts, P.V., Goltz, M.N., and Mackay, D.M. (1986). Water (1993). Environmental Organic Chemistry. John Wiley &
Resources Res. 22(13): 2047–2058. Sons, New York.
 WATER IN THE UNSATURATED ZONE 531

28. Johnson, R.L., Palmer, C.D., and Fish, W. (1989). Transport Land surface Well
and Fate of Contaminants in the Subsurface. U.S. Environ-
mental Protection Agency, Ada, OK, EPA/625/4-89-019.
29. Mackay, D. and Leinonen, P.J. (1975). Environ. Sci. Technol.
9: 1178–1180.
30. Mackay, D. and Wolkoff, A.W. (1973). Environ. Sci. Technol.
7: 611–614.
31. Mackay, D., Shiu, W.Y., and Sutherland, R.P. (1979). Envi- Unsaturated
ron. Sci. Technol. 13: 333–337. zone
32. Thibodeaux, L.T. (1996). Environmental Chemodynamics:
Movement of Chemicals in Air, Water, and Soil. John Wiley
& Sons, New York.
33. Young, L.Y. and Cerniglia, C.E. (1995). Microbial Transfor-
mation and Degradation of Toxic Organic Chemicals. John Capilary fringe
Wiley & Sons, New York.
34. Maymo-Gatell, X., Tandoi, V., Gossett, J.M., and Zinder, S.H.
(1995). Appl. Environ. Microbiol. 61: 3928–3933. Saturated Water
35. Rittman, B.E. and McCarty, P.L. (2001). Environmental zone table
Biotechnology: Principles and Applications. McGraw-Hill,
New York.
36. Hollinger, C., Schraa, G., Stams, A.J.M., and Zehnder, A.B. Figure 1. A schematic diagram of the unsaturated zone.
(1993). Appl. Environ. Microbiol. 59(9): 2991–2997.
37. Carr, C.S. and Hughes, J.B. (1998). Environ. Sci. Technol. 32:
1817–1824. interactions, transport processes of various kinds, and
38. Alvarez, P.J.J. and Vogel, T. (1991). Appl. Environ. Microbiol. chemical reactions involving both natural and artificial
57: 2981–2985. substances. Several biological activities involving plant
39. Lee, M.D. et al. (1988). CRC Cri. Rev. Environ. Control. 18: roots, rodents, worms, microbiota, and other organisms
29–89. also take place in this zone. The zone is used for cultivating
40. Lovely, D.R., Coates, J.D., Woodward, J.C., and Phillips, plants, constructing structures, and disposal of waste.
E.J.P. (1995). Appl. Environ. Microbiol. 61: 953–958. To a large degree, the unsaturated zone controls the
41. Pankow, J.F. and Cherry, J.A. (1996). Dense Chlorinated transmission of water to aquifers (water bearing strata),
Solvents and Other DNAPLs in Groundwater. Waterloo Press, as well as to the land surface, to water on the surface,
Portland, OR. and to the atmosphere. It also acts as a filter as it
42. DiStefano, T.D., Gossett, J.M., and Zinder, S.H. (1992). Appl. removes undesirable substances before they reach and
Environ. Microbiol. 58(11): 3622–3629. affect aquifers. Figure 2 gives a pictorial representation of
43. Smatlak, C.R., Gossett, J.M., and Zinder, S.H. (1996). Envi- the unsaturated zone.
ron. Sci. Technol. 30(9): 2850–2858. Though the unsaturated zone stores water, plant
44. Yang, Y. and McCarty, P.L. (2000). Environ. Sci. Technol. 34: nutrients, and other substances, it is not always considered
2979–2984. a major storage component of the hydrologic cycle because
45. Freedman, D.L. and Gossett, J.M. (1989). Appl. Environ. it holds only a tiny fraction of the earth’s fresh water and
Microbiol. 55: 2144–2151. this water is usually difficult to extract. But it is of great
importance for storing water and nutrients in ways that
are vital to the biosphere.
WATER IN THE UNSATURATED ZONE
MUKAND SINGH BABEL SOIL WATER
Asian Institute of Technology
Parthumthani, Thailand Soil water is normally considered to include both the water
contained in the soil profile and the subsurface water in
the unsaturated subsoil layers. Soil water may be present
The soil profile immediately below the land surface and as gravity water in transit in larger pores, as capillary
above the water table is called the unsaturated zone, water in smaller pores, as hygroscopic water adhering
where the pores contain both water and air. Figure 1
shows a schematic diagram of the unsaturated zone. This
zone differs from the saturated zone, where the pores are
saturated with water. The unsaturated zone is sometimes
called the vadose zone. Land
The thickness of an unsaturated zone can range from Unsaturated zone surface
zero (e.g., when a lake or marsh is at the surface) Water table Surface water
Saturated zone
to hundreds of meters, as is common in arid regions. Groundwater
It is a zone of natural and human-induced activity. It
experiences physical phenomena such as thermodynamic Figure 2. Unsaturated zone.
532 WATER IN THE UNSATURATED ZONE

in a thin film to soil grains, and as water vapor. After components consists of gravitational potential, osmotic
a rain, gravitational water moves downward in larger potential (also sometimes called solute potential), matric
pores, but it is dispersed into capillary pores or passes potential, vapor potential, hydrostatic pressure potential,
through the unsaturated zone to the groundwater or to and overburden pressure potential. The major potential
a stream channel. Hygroscopic water is held very tightly components of interest in the unsaturated zone are
to the soil surface by molecular attraction and is not gravitational potential, osmotic potential, and matric
normally removed from soil. Capillary water, therefore, is potential that give rise to the total potential as follows:
the important variable element of soil water.
t =  g +  o +  p + . . . (1)
SOIL WATER POTENTIAL where t is the total potential, g is the gravitational
potential, o is the osmotic potential, and p is the
The most basic measure of the water in an unsaturated matric potential.
soil is water content or wetness, defined as the volume Gravitational potential is defined as the energy of water
of water per bulk volume of the soil. Water is held in per unit volume that is required to move a specific amount
unsaturated soil by forces whose effect is expressed in of pure, free water from an arbitrary reference point to the
terms of the energy state or pressure of the water. Like soil-water elevation. It is independent of soil properties
all matter, water can possess both kinetic and potential and solely dependent on the vertical distance between the
energy. However, in general, the flow of water in soils arbitrary reference point and the interest point. Often, the
is too small to consider kinetic energy. Therefore, the soil surface is chosen as the point of reference, whereby
energy state of soil water is represented by its equivalent the distance z (m) to each point of interest is negative by
potential energy. Potential energy is defined as the ability convention. Gravitational potential is given as
to do work. Work is done whenever an object is accelerated
through a distance or heat is produced or transferred. Soil g = ρl gz (2)
water potential thus measures the ability of soil water to
move and/or produce heat. Water potential is important to where ρl is the density of the soil water, g is the
any process in which soil water moves, such as infiltration acceleration due to gravity (m/s2 ) and g is the
and redistribution within the soil or removal of water from gravitational potential (J/m3 ).
the soil by evaporation or plant uptake. The osmotic potential component arises due to the salts
To define the soil water potential parameter, a reference in the soil water. It may be thought of more a suction than
body of water is defined, and the soil water potential of a pressure. In a dynamic soil–water–plant system, in the
any other body of water is defined as the difference in presence of a selective permeable membrane as found in
potential energy concentration (potential energy per unit the roots of plants, the pressure or energy potential (a
quantity of water) between the soil water and the reference suction) on the solution side is less than that of pure,
body of water. The reference body of water is usually free water. Water will pass through the membrane to the
called the reference state, which is defined as ‘‘pure (no solution side until equilibrium is reached. The osmotic
dissolved solutes), free (free of adsorptive forces) water at potential, o , is given by
a specified temperature and specified elevation exposed to
atmospheric pressure.’’ The water potential of water in the 
o = − (3)
reference state is zero. ρw
Total soil water potential may formally be defined
as the amount of work per unit quantity of pure water where  is the osmotic pressure and ρw is the density
that must be done by external forces to transfer an of water.
infinitesimal amount of water reversibly and isothermally Soil is porous, so it acts as a semipermeable barrier
from the reference state to the soil at the point under to salt solutions. An individual particle has an electric
consideration (1). field that causes a decrease in the concentration of the
soil solution near the particle surface, which results
in an increased cross-sectional area for water transport
FACTORS AFFECTING WATER POTENTIAL
compared to the area available for transport of dissolved
salts, causing water to move due to the osmotic difference.
A change in the potential energy of the soil water brings
The amount of flow depends on the particle size of the
about a change in the soil water potential. There are
soil and the gradient of the osmotic pressure. In coarse,
several factors that affect the potential energy of soil
saturated sand, soil pores are relatively larger than those
water, including adsorption of water onto soil particles,
in clays, and consequently, the thickness of fluid layer with
solutes dissolved in the soil water, the elevation of the
decreased salt concentration is small compared to the total
soil water in the earth’s gravitational field, and applied
area available for transport. Thus, the semipermeability
pressure (both positive and negative).
is small, and the transport of soil solution due to osmotic
potential is negligible. In clays, with many more pores,
WATER POTENTIAL COMPONENTS the solutes in the soil solution can be completely repelled
from the soil and restricted from flow. In such soils, water
The total soil water potential can be broken into several movement in response to an osmotic pressure gradient is
components. A complete list of soil water potential significant.
 GROUNDWATER AND VADOSE ZONE HYDROLOGY 533

The matric potential component results from the soil widely used indirect way of measuring soil water. In this
matrix. Water molecules can form hydrogen bonds with method, a source of high-energy neutrons is used that
the surface of soil minerals that give rise to adsorption is lowered into an aluminum access tube in the soil.
and with other water molecules that result in cohesion. Fast neutrons lose energy through collision with atoms
In unsaturated soil, these forces of attraction between the of low atomic weight, such as the hydrogen in water,
soil mineral surfaces and the soil water exert a ‘‘pull’’ on and are converted to slow neutrons. These slow neutrons
the soil water reducing the free energy of the water. It is are counted with a high count indicating high moisture
usual to deal with the combined effect of both capillary content. Time-domain reflectometry (TDR) determines soil
and adsorptive forces in the way in which water is held water content by measuring the dielectric constant of the
in the soil matrix. The matric potential is the potential of soil. TDR has two significant advantages over the neutron
the water in a pore of the medium relative to the potential probe method; it is nonradioactive and can be set up for an
of the air that results primarily from the adsorptive and automatic, nonmanual operation. The satellite-mounted
capillary forces of the soil matrix. When a medium is synthetic aperture radar (SAR) method has the potential
unsaturated, the water generally is at lower pressure for measuring soil moisture content on a drainage basin
than the air, so the matric potential is negative. scale. It determines soil water content by monitoring
Moisture content increases as the matric potential changes in soil dielectric properties.
increases. Zero matric potential is associated with high Tensiometers are the oldest and most widely used
(saturated or nearly saturated) water content. As the technique for measuring matric potential. They comprise a
matric potential decreases, the water content decreases. liquid-filled porous cup connected to a pressure-measuring
Figure 3 shows the relation between matric potential device such as a mercury manometer or a pressure
and water content, called a retention curve. It is transducer. The cup is embedded in the soil, and water can
a characteristic of a porous medium that depends flow between the soil and cup until the pressure potential
on the nature of its pores. This relation strongly inside the cup equalizes with that of the soil water. In fine-
influences the movement of water and other substances in textured soils, the resistance block method may be useful.
unsaturated media. In this method, two electrodes are embedded in a porous
block, which is buried in the soil. The matric potential
in the block comes into equilibrium with that of the soil
MEASUREMENT OF SOIL WATER
water, and the resistance across the electrodes varies
with the resulting water content of the block. Gypsum
Soil water can be measured as either soil water content resistance blocks can be used for measuring potentials as
or soil water potential; conversion from one to the other low as −15,000 cm (1500 kPa).
uses the moisture retention curve. Their combined use
provides a powerful approach to the study of soil fluxes in
unsaturated soil. BIBLIOGRAPHY
The most commonly used method for measuring the
water content of soil directly is the gravimetric method, 1. Tindall, J.A. and Kunkel, J.R. (1999). Unsaturated Zone
which involves taking a number of soil samples of known Hydrology for Scientists and Engineers. Prentice-Hall, Engle-
volume and determining their weight loss when dried wood, Cliffs, NJ.
in an oven at 105 ◦ C. The neutron probe is the most

GROUNDWATER AND VADOSE ZONE

Moisture retention curves HYDROLOGY
UWE DANNWOLF
URS Australia Pty Ltd.
Turner, Australia
Clayey
soil
Moisture content

The vadose zone is the geologic media between the land

surface and the regional water table.
Sandy
soil The vadose zone

• is the foundation for most man-made structures;

• provides nutrients and water to all plant and
animal life;
• protects drinking water reservoirs (aquifers);
• governs how much groundwater is replenished;
• provides the pathway for contaminants to penetrate
Matric potential (−cm water)
into groundwater, and
Figure 3. Moisture retention curves. • is one of the least understood media.
534 GROUNDWATER AND VADOSE ZONE HYDROLOGY

The latter point becomes evident when a definition for catchment, vegetative cover, relief [(slope or gradient) and
the vadose zone is needed. All publications include the depressions (pond formation)], soil cover and soil type,
unsaturated soil zone, but some publications include the initial soil moisture content, and rainfall intensity. For
capillary fringe; others do not. The term ‘‘vadose’’ is instance, if the supply rate of water to the soil surface is
derived from the Latin noun vadosus, which translates greater than the soil’s ability to allow the water to enter,
to ‘‘shallow’’. excess water either accumulates on the soil surface or
The vadose zone includes the weathered soil horizon becomes runoff. Steep slopes cause greater runoff rates,
and the unconsolidated sediments and rocks below. The allowing the water a shorter time to infiltrate. In contrast,
capillary fringe separates the water saturated zone from gentle slopes have less impact on infiltration due to
the unsaturated zone. The water saturated zone is often decreased runoff.
referred to as an aquifer if it allows significant amounts Surface water runoff is even more difficult to measure
of water to be transmitted. Several processes take place in than any of the above parameters. Hydrological models
the vadose zone. aid in estimating it.

Effective Recharge or Infiltration

PROCESSES IN THE VADOSE ZONE
Both evapotranspiration and surface water runoff limit the
The processes, which influence the vadose zone and the amount of total precipitation that enters the subsurface,
recharge to groundwater include the effective recharge or infiltration into the vadose zone.
The term infiltrability is generally used in soil physics and
• precipitation, hydrology to define the maximum rate at which rain or
• evapotranspiration, irrigation water can be transmitted through a soil under
• surface water runoff, and a given condition.
• infiltration and groundwater recharge. As a general rule of thumb, a total of four water relevant
zones exist in the vadose zone (after Reference 6):
Precipitation
1. a near-surface zone to shallow (<1 m) depths
Precipitation is the water that falls on the land surface affected by episodic rain;
as rain, snow, and other forms such as dew, frost, and
2. a zone below which shows signs of seasonal water
irrigation water. It is measured as volume per time
infiltration;
period (e.g., mm/a) or as volume per area and time
(e.g., L/km2 s), which is a relatively simple parameter 3. a zone of virtually constant moisture content, in
to measure and is often available from meteorological which the hydraulic gradient is unity (i.e., 1), and
stations. However, care should be exercised when using there is a predominantly downward flow toward the
this parameter for climatic water balancing because there water table; and
is great variability even between neighboring stations, due 4. a capillary zone above the water table, in which the
to the precipitation and wind characteristics as well as the water moves in the groundwater flow direction.
various measuring devices.
This simplified zoning is complicated by the fact
Evapotranspiration that irregularities (heterogeneities) in the soil profile
such as root channels, earthworm burrows, soil fissures
Evapotranspiration, the reverse of precipitation, repre-
or cracks, spatially varying sediment size therefore
sents the transport of water from the earth back to the
exhibiting variable hydraulic and physical characteristics
atmosphere through evaporation and plant transpiration.
This process depends on the microclimatic conditions at
a given site. This parameter depends on many factors
Coarse sand Silty sand Clay
(air temperature and humidity, solar irradiation, vegeta-
tion, wind speed, etc.). It is more difficult to measure and
therefore not available from all meteorological stations. Silty
Simple (1) and complex calculations (2–5) can estimate its sand
range. Its units are generally reported as millimeter per Coarse
time period. The vegetation type and cover greatly influ- sand
ence this parameter and are often neglected in standard Clay h
formulas, which can lead to significant variations.
Coarse
sand
Surface Water Runoff
Silty Capillary
Surface water runoff is the amount of precipitation that sand fringe
does not immediately enter the subsurface and contributes Water table 0 1.0
to surface water discharge (e.g., into local creeks). A Sw
portion of this surface water runoff appears on the surface Figure 1. Saturated water content (Sw ) above the water table
again after a short underground passage (interflow). The for a layered sequence. Dashed lines represent water content for
amount of surface water runoff depends on the size of the homogeneous sediments (from Reference 7, altered).
 GROUNDWATER AND VADOSE ZONE HYDROLOGY 535

(see Fig. 1), and distinct layering allow preferential (i.e., silt, clay, colloid, and ion. Clays are smaller than 2 µm,
faster) flow of water in selected areas. For instance, clay whereas colloids consist of particles less than 0.1 µm
layers generally impede flow due to their lower saturated in size. On the other hand, gravels start at 2 mm. The
hydraulic conductivity, whereas sand layers have a lower distribution of soil grain size and bulk densities (often
tendency to retard the movement of infiltrating water. referred to as ρ) of soil and rock are commonly measured
This process, which is also referred to as ‘‘fingering,’’ can during geotechnical investigations.
occur in less than a few percent and to up to 70% of the With regard to water and contaminant transport, a
soil profile. Such layers generally retard water movement critical parameter of vadose zone materials is the type
during infiltration. For the remainder of the soil profile, and distribution of void spaces (pores). Loose materials
the entering moisture results in a semicontinuous wetting contain widely distributed void spaces that typically are
front, a zone of increased moisture content and a zone interconnected. These pores are usually individually very
below of significantly lower moisture content. Beyond the small, but numerous, and the total volume of pores usually
wetting front, there is no visible penetration of water. varies from 30–50% of the total volume. However, not all
Infiltrating precipitation can appear as of these pores are available for water circulation. A clay
may exhibit a total porosity of >50%, however, only less
1. retarded interflow discharging as surface runoff than a few percent are available for water circulation. In
after a brief underground passage, contrast, in gravel sediments, almost all of the pore space
2. increased soil moisture, and (roughly 20% of the total volume) allows water transport.
3. effective recharge. As can be seen from Fig. 2, classifications exist for pore
spaces. The largest pores are obviously present in the
coarsest materials. Pores of the order of 60–1000 µm are
The effective recharge can be measured by using
present in sandy sediments. Small and smallest pores are
lysimeters, but this requires a substantial effort. It
present in clay grain size and below.
is furthermore dynamic, changing constantly over time
Also depicted in Fig. 2b is the approximate rate of water
and space.
movement in meters per year. The coarser the sediments,
The measuring units are volume or m per time period.
the higher the rate of vertical water movement. It is,
Infiltrating water also changes the moisture content of
however, important to note that even in silts, the vertical
the soil. If rain infiltrates a previously dry soil (typical
water movement under unit gradient can be as high as
moisture content of less than 10% saturation), then the
2 m per year.
rainwater is first absorbed into the soil pores, increasing
the moisture content. The soil gas in the soil profile
needs to be replaced, and as a result the infiltration Capillary Rise and Capillary Pressure
velocity is reduced initially. Later in the process when When water evaporates near the surface, it generates
the soil gas dissolves gradually in the infiltrating water, a low hydraulic potential, which is equilibrated by the
the infiltration rate increases again before a renewed capillary rise at the groundwater table. The rate at
reduction occurs. This reduction is from increased solid which this supply is determined is governed by the
particles washed into the soil profile from the rain and the permeability (and size, respectively) of the sediments.
growth of organisms near the soil surface. Additionally, Consequently, the height of the capillary fringe depends
the increase in soil moisture leads to a reduction in on the sediment’s permeability, as Fig. 2a outlines. The
capillary pressure. water in the capillary fringe is held by capillary forces
Snow and frost also impact the infiltration rate. In against gravity induced flow.
frost periods, soils of low initial water content may become The capillary pressure (Cp , also named matrix or soil
porous, and the infiltration rate is increased. Soils of suction) for sands and silts can be estimated roughly from
high initial water content may freeze and become nearly
impermeable. Cp (kPa) = 0.3/grain size (mm) (1)
In detail, infiltration depends on material properties,
such as The higher the soil suction, the higher the residual
saturation of a sediment.
• grain size,
• porosity, Soil–Water Characteristic Curve
• capillary forces and soil suction,
The soil–water characteristic curve (also referred to as
• permeability, and the water retention curve) for a soil is defined as the
• fluid densities and viscosities. relationship between the water content and the capillary
pressure. This curve is characteristic for each soil, it
governs the rate at which water is taken up or released by
MATERIAL DEPENDENCIES
the soil.
As can be seen from Fig. 3, when the water content
Grain Size and Porosity
decreases (i.e., the soil suction increases), the permeability
Unconsolidated materials are distinguished by their of a sediment (here, a silty soil) decreases. The order
grain sizes. Various classifications exist, the grain sizes of decrease depends on pore sizes. Sands have a much
commonly decrease in the order block, cobble, gravel, sand, slighter decrease in permeability. Tight soils stay wet
536 GROUNDWATER AND VADOSE ZONE HYDROLOGY

(a) 100000 1000

Thickness of cap. fringe, m

10000
100
Saturated horizontal

1000
permeability, m/d

100
10
10
Thickness of capillary fringe, m
1 1
0.1 Saturated horizontal permeability,
0.1 m/d
0.01
0.001
0.01
0.0001
0.00001 0.001
10000 1000 100 10 1 0.1 0.01 0.001 0.0001 0.00001 0.000001 1E-07
Vertical water movement, m/year

12 10000
(b)
10 1000

Capillary pressure, kPa

100
8
10 Capillary pressure, kPa
6
1 Water movement, m/year
4
0.1

2 0.01

0 0.001
10000 1000 100 10 1 0.1 0.01 0.001 0.0001 0.00001 0.000001 1E-07 Grain size, mm

Block Cobble Gravel Sand Silt Clay Colloid Ion

Largest pores Large pores Medium Small pores Smallest pores

60–1000 µm <60 µm <10 µm

Figure 2. (a) Generalized diagram depicting the thickness of the capillary fringe and saturated
horizontal permeability. (b) Generalized diagram depicting pore sizes, capillary pressure, and
the vertical annual water movement rate of sediments.

for much longer, however, suffer much more from 1. the drainage of an individual pore, which is
evaporative losses. controlled by the smallest radius. When the
Figure 3 also introduces terms often used by soil meniscus retreats under increasing tension, the
scientists to describe the wetting and drying behavior pore’s radius is decreasing. Rewetting (imbibition),
of a soil. The term field capacity is used when the water on the other hand, is controlled by the pore’s
content of a soil has reached a maximum (often referred to largest radius.
as sat). Any further increase in soil moisture results in a 2. contact angle hysteresis. This is a well-recognized
water release (i.e., gravity induced flow). When the water phenomenon, but it is only significant at relatively
content nearly equals its minimum, then the wilting point high flow rates. The slower the process, the less
is reached. By definition, the wilting point equals a soil contact angle hysteresis. Because drainage and
suction between 500 and 6000 kPa. A soil is referred to as rewetting curves are generally measured at slow
‘‘air dry’’ at or above a capillary pressure of 70,000 kPa. (‘‘quasi-equilibrium’’) rates, this is not likely to be an
Furthermore, two terms are defined through this water important contributor.
retention curve: the air entry value and the residual water 3. Air entrapment effects. During imbibition, air can
content. The air entry value is the meeting point of the become trapped inside the larger soil pores. Given
two tangents shown in Fig. 3. The residual water content sufficient time, the entrapped air bubbles slowly
(referred to as res) is defined similarly by using the dissolve in the water and cause hysteresis
lower tangent. Figure 3 further introduces a fundamental
feature soils display when drying out or becoming wet (the Determining the Soil–Water Characteristic Curve. The
hysteresis effect). most commonly used direct way to measure the soil–water
Hysteresis describes the effect of retardation of forces potential is with a tensiometer. A body of pure water is
acting on a body to change (as if from viscosity or hydraulically connected to the soil water via a permeable
internal friction in magnetic induction and other physical ceramic cup. If the soil water is under tension, then a
phenomena). In soil science, when a soil is dried and then small fraction of water in the tensiometer is pulled out,
rewetted, the drying and wetting soil–water characteristic lowering the pressure inside. This negative pressure can
curves are not identical. The wetting curve is drier than the then be measured by a gauge or pressure transducer.
drying curve over a wide range of soil–water potentials. Alternatively, mathematical models can help in deter-
This hysteresis is caused by: mining the soil–water characteristics of a soil.
 GROUNDWATER AND VADOSE ZONE HYDROLOGY 537
Coefficient of permeability,

1.E-04
1.E-05
1.E-06
k, m/s

1.E-07
1.E-08
1.E-09
1.E-10
1 10 100 1000 10000 100000 1000000 10000000
Soil suction, kPa

Air-Entry Value

50
Drying curve
40 Wetting curve
Water content, %

Field capacity
Wa

Wilting point
te

Air dry
30
r re
ten

20
tio
n
cu

Residual water
rve

10
content
0
1 10 100 1000 10000 100000 1000000 10000000
Soil suction, kPa
Figure 3. Visualization of the relationship between the permeability function and the soil water
characteristic curve (from Reference 8, altered).

Hydraulic Conductivity pumping tests from groundwater monitoring wells, slug

tests, and other indirect measurements like grain-size
The basic formula to determine hydraulic conductivity
distribution. In vadose zone studies, tensiometers are used
was first developed by Henry Darcy in 1856 when he
for determining the field capacity of a soil.
determined that the flux of water in sand columns depends
on hydraulic conductivity and the hydraulic gradient. It
can further be shown that hydraulic conductivity depends FLOW PROCESSES IN THE VADOSE ZONE
on both the water and the porous media properties
according to Two processes control water in the vadose zone. The first
kϕg is gravity, which moves water downward. The second is
K= (2)
µ a capillary process that, similar to water dripped slowly
on a sponge, moves water in all directions and stores
where µ = viscosity, and releases it. The capillary process is controlled by
φ = density the nature of the sediment and rock and the pores and
g = acceleration of gravity fractures within it. In most cases, capillary processes
k = intrinsic permeability of the soil (independent are dominant in fine-grained sediments (clay, silt, and
of fluid). the like), whereas gravity is dominant in coarse-grained
sediment and large fractures. Pores and fractures can be
The intrinsic permeability is generally a function of the completely filled with water in places within the vadose
porous media, and it is proportional to the square of the zone. These saturated portions of the vadose zone are
grain diameter, k ∼ d2 . Consequently, capillary pressure typically referred to as ‘‘perched water’’ or ‘‘capillary
depends on the soil’s conductivity: fringe,’’ depending on where they occur. At most sites,
√ the water content is changing over time throughout the
Cp ∼ 1/ K (3) system representing a dynamic system.
Once saturation is continuous, this groundwater moves
This relationship is commonly used in numerical modeling through the system from the highest potential to the
studies to scale the magnitude of the capillary pressure lowest. Numerical modeling studies show that a range of
curves in the absence of direct measurements of infiltration patterns occur, which can be related to the ratio
capillary pressure. of infiltration rate to unsaturated hydraulic conductivity.
A high value of this ratio reflects a prevailing hydraulic
Measurements of Hydraulic Conductivity. The hydraulic conductivity, which cannot readily redistribute the newly
conductivity of an aquifer with respect to water can infiltrated moisture. Moisture accumulates in the near-
be determined by a variety of field tests, including surface region before advancing down through the soil as
538 VADOSE ZONE MONITORING TECHNIQUES

a distinct wetting front. In contrast, low values of the ratio • UnSat: Waterloo Hydrogeologic Inc;. http://www.
of rainfall to unsaturated hydraulic conductivity show flowpath.com.
minimal moisture accumulation, as the relatively small
volumes of infiltrating moisture are readily redistributed BIBLIOGRAPHY
through the soil profile.
1. Linacre, E. (1993). A three-resistance model of crop and forest
evaporation. J. Theor. Appl. Climatology 48: 41–48.
MODELS
2. Penman, H.L. (1948). Natural evaporation from open water,
bare soil and grass. Proc. R. Soc.London A193: 120–146.
Accurate in situ measurement of unsaturated hydraulic
3. Thorntwaite, C.W. and Mather, J.R. (1955). The water
conductivity is cumbersome and time-consuming. One
balance, Publication in Climat. VIII, Nr.1. Drexel Inst.
alternative to direct measurement of unsaturated Laboratory of Climatology, Centerton, NJ.
hydraulic conductivity is to use theoretical methods, which
4. Turc, L. (1954). Le bilan d’eau des sols, relations entre les
predict the conductivity from more easily measured soil precipitations, l’evaporation et l’ecoulement. Trois Journée
water retention data. Methods of this type are generally d’Hydraulique. Alger.
based on statistical pore-size distribution models, a large 5. Haude, W. (1955). Zur Bestimmung der Verdunstung auf
number of which have appeared in the soil science and möglichst einfache Weise. Mitt. dt. Wetterd. 2: 11, Bad
petroleum engineering literature during the past several Kissingen.
decades (see Reference 9 for a review). Implementation 6. Faybishenko, B. (2000). Vadose Zone Characterization and
of these predictive conductivity models still requires inde- Monitoring—Current Technologies, Applications, and Future
pendently measured soil–water retention data. Measured Developments. In: Vadosc Zone Science and Technology
input retention data may be given either in tabular form Solutions. Battelle Press, Columbus, OH.
or by closed-form analytical expressions, which contain 7. Looney, B.B. and Falta, R.W. (Eds.). (2000). Vadose Zone
parameters that are fitted to the observed data. Science and Technology Solutions. Battelle Press, Columbus,
Water flow in variably saturated soils is traditionally OH, pp. 1–1465.
described by the Richards equation: 8. Fredlund, M.D. (1988). Unsaturated seepage modelling made
  easy. Geotechnical News June.
∂θ ∂ ∂h ∂K(θ ) 9. Mualem, Y. (1986). Hydraulic conductivity of unsaturated
= K(θ ) + (4)
∂t ∂z ∂z ∂z soils: Prediction and formulas. In: Methods of Soil Analysis.
A. Klute (Ed.), Part 1. Physical and Mineralogical Methods.
where h is the soil water pressure head (or fluid pressure Agron. Monogr. 9, 2nd Edn. American Society of Agronomy,
with dimension L), θ is the moisture content [−], t is time Madison, WI, pp. 799–823.
(T), z is soil depth (L), and K is the hydraulic conductivity 10. van Genuchten, M.Th. (1980). A closed-form equation for
[LT−1]. The solution of the Richards equation requires predicting the hydraulic conductivity of unsaturated soils.
knowledge of the unsaturated soil hydraulic functions θ (h) Soil Sci. Soc. Am. J. 44.
and K(h) or K(θ ). 11. Brooks, R.H. and Corey, A.T. (1964). Hydraulic Properties
Several equations have been proposed to describe the of Porous Media. Hydrology Papers, No. 3, Colorado State
soil–water retention curve near saturation. Most of these University, Fort Collins, CO.
functions are mathematically too complicated to be easily
incorporated into predictive pore-size distribution models
for hydraulic conductivity. A smooth function with attrac- VADOSE ZONE MONITORING TECHNIQUES
tive properties is the equation of van Genuchten (10):
PHILIP R. RYKWALDER
Seff (h) = [1 + (ah)n ]−m (5) University of Texas at San
Antonio
San Antonio, Texas
where a, n, and m are empirical constants affecting the
shape of the retention curve (i.e., curve fitting parameters).
Another mathematical solution is provided by Brooks and WILLIAM J. BLANFORD
Corey (11). Louisiana State University
Baton Rouge, Louisiana
Various computer programs exist to model unsaturated
soil properties, including (but not limited to)

• Soil Vision: www.soilvision.com. INTRODUCTION

• Rosetta: U.S. Salinity Laboratory USDA; http://www.
Monitoring fluids of the vadose zone has multiple
ussl.ars.usda.gov/models/rosetta/rosetta.
motives, including detecting contaminants, measuring
• RETC: Code for Quantifying the Hydraulic Functions soil properties and water conditions for construction or
of Unsaturated Soils, U.S. Salinity Laboratory monitoring geologic hazards, and determinations of water
USDA; http://www.ussl.ars.usda.gov/models/retc. saturation that may be used for optimizing irrigation.
• SEEP/W: http://www.geo-slope.com/products/seepw. Because fluids must move through the vadose zone to
asp. reach the water table, monitoring of the vadose zone allows
• Hydrus: http://www.ussl.ars.usda.gov/models/tu predicting possible sources and processes of groundwater
torials/hydrus2d/tutorial2d.htm. contamination. Geologic hazards such as landslides,
 VADOSE ZONE MONITORING TECHNIQUES 539

rising water tables, or land subsidence can be partially

predetermined by monitoring the vadose zone. And in
agriculture, monitoring soil water content to irrigate most Air inlet
efficiently has a long history. Water reservoir
Monitoring the temporal and spatial variance in the Bubble tower
vadose zone is vitally important to engineers, geologists,
meteorologists and other environmental scientists inter-
ested in studying water and contaminants. There are many
techniques that directly or indirectly assess water con-
tent, sorbed mass, gaseous and liquid contaminants, flow
direction, and other parameters that may be monitored.
The methods discussed here are divided into two cate-
gories, direct and indirect monitoring techniques. Direct
monitoring techniques require a degree of sampling or
large-scale soil disturbances. Conversely, indirect tech-
niques are analogous to remote sensing in which no
disturbance to the soil or water is required, although a
well may be required in some techniques. This section dis-
cusses in short detail some of the attributes and shortfalls
of most of the currently used techniques for monitoring
the three phases of the vadose zone: liquid, soil, and gas. Porous baseplate
For a more in-depth understanding, readers are directed Figure 1. Tension infiltrometer. Tension measurements may be
to the references listed after each entry. made using a handheld tensiometer on nongraduated models or
on the side of graduated models (pictured).
INDIRECT MONITORING METHODS
reservoir. A thin layer of moist sand may be placed under
Indirect methods for monitoring the vadose zone commonly the baseplate to ensure better hydraulic conductivity
have links to geophysical tools, principles, and techniques. between it and the soil. The reservoir supplies water to
Many vadose zone characteristics such as water content, the baseplate and the bubble tower; the latter consists
the presence of contamination, and flow direction, among of a series of tubes that relieve the vacuum in the
others may be determined without soil sampling. These reservoir when water infiltrates into the soil. Tension
techniques may also be used to determine the nature and measurements correspond to the level of the water in the
extent of soil stratification for site delineation. reservoir. An important point to note is that hydraulic
In practice, these methods are generally nonintrusive conductivity must exist between the baseplate and the
and do not require disturbing the soil, but some wet porous media for the device to operate.
require well installation. When drilling or augering is The main advantage of disk infiltrometers is that
necessitated, proper techniques must be ensured so that in situ measurements of vadose zone hydraulic properties
mud linings, drilling fluids, or other drilling by-products are rapid, reliable, and replicable and the devices are
do not affect hydraulic conductivity, which may then affect inexpensive and easy to operate. There exists, however,
measurements. It is also important to clean any grease or a limited range of tensions (generally 0 to 25 cm below
oil from drilling tools to avoid contamination and to clean the ground surface (BGS) in which the infiltrometers
tools between use. can operate, and measurements can be made only at
the soil surface unless trenches are dug to the desired
Tension Infiltrometers measurement level and the device used (1–3). Ideal soil
properties are also assumed (soil is uniform, homogenous,
Tension infiltrometers are used to characterize hydraulic
and nonswelling), and pore closure may result from the
properties of saturated and partially saturated porous
weight of the device if used on freshly cultivated soils.
media, including saturated and unsaturated hydraulic
There are other limitations that may be observed in
conductivities. Infiltrometers are commonly used to
Hussen and Warrick (4).
measure hydraulic conductivity as a function of matrix
potential in partially saturated porous media, sorptivity,
Neutron Moisture Probe
steady-state volumetric flow rate (Q), and water content
(Fig. 1). Initially designed for the agricultural industry, the neu-
There are variations of the disk infiltrometer and tron moisture probe has been adapted to environmental
multiple methods of measuring hydraulic properties using monitoring to obtain soil moisture levels. It is a nonde-
infiltrometers, but all consist of three basic parts: the structive method of obtaining soil moisture levels when
bubble tower, the reservoir, and a porous baseplate. The the probe is inserted in a well. The only restriction to
porous base plate lies flat on a wetted soil surface, depth of use is the length of the well used and the probe
and both the bubble tower and reservoir are connected cord length.
perpendicularly to it. The purpose of the baseplate is Once lowered down a well, the probe begins to emit
to establish hydraulic contact between the soil and the neutrons. The velocity at which neutrons return to the
540 VADOSE ZONE MONITORING TECHNIQUES

probe is correlative to the substance by which they calibration is necessary due to corrosion of metals and
were reflected. Neutrons reflected by heavy atoms of the accumulation of salts on the thermocouple junction.
the soil matrix return with higher velocities compared Also, psychrometer equations assume that water is the
to those of neutrons reflected by lighter hydrogen atoms. first liquid to condense, and, at a site containing volatile
Hydrogen may be found in organic matter or in mineral liquids, this may not always be true (9,10).
lattices, but water is the greatest source of hydrogen in
the subsurface. Due to this, the counts of reflected low- Tensiometry
velocity neutrons are correlative to soil moisture content.
A complete moisture stratification along the length of a Tensiometers measure the force which holds water to the
well may be obtained, given a well-calibrated instrument. soil, or the soil–water potential, at shallow depths (0 to 8 m
Attributes of the neutron probe include ease of use BGS). Because water moves from areas of high potential
and high accuracy, given a well-calibrated probe. It may energy to low, evaluation of water’s energy may be used
be operated by one person, and soil moisture changes to estimate the direction of water flow and the direction
of the order of 1–5 volume % may be measured. Water of contaminant plume migration. Rapid measurements
flow pathways may also be located with the instrument. at different locations may be obtained with tensiometers,
The largest drawback of the neutron moisture probes which may then be plotted to obtain a spatial distribution
lies with calibration difficulties that may be extreme. of water and an estimation of the water flow direction.
Improper neutron probe calibration nearly renders the A tensiometer measures matric potential using a porous
probe useless. Ideally, the probe is calibrated using soils ceramic cup placed in the soil, a connecting tube, and a
collected in a drum with a known quantity of water vacuum gauge. The cup and tube are filled with water
from the field site of interest. Another drawback is that to provide hydraulic contact from the tensiometer to the
neutron probes cannot distinguish between water and water in the soil. Once placed in the ground, water flows
other substances with similar hydrogen densities in the out of the cup and into the soil until equilibrium is reached.
subsurface, such as gasoline. Problems of applying single Because the device is sealed from the atmosphere, the drop
calibrations to multiple media types also arise, which is of the water level in the tube creates a vacuum which is
the case in heterogeneous soils. Inaccuracies may also measured by a vacuum gauge or pressure transducer from
be result from improper well packing and the presence the surface.
of certain elements, such as migrating boron or chlorine. Tensiomenters represent a low-cost, simple, and
Finally, care should be taken to mitigate exposure to the accurate measurement of matric potential (11–13). It is
neutron source (5–8). one of the most widely used devices for monitoring soil
moisture levels in the vadose zone, though tensiometry
Psychrometers values are limited to tension values below 0.85 bar at
atmospheric pressures near 1 bar. For field applications,
Psychrometers represent another tool adapted from air coming out of solution from water or air trapped in the
the agricultural industry for environmental monitoring. tube are the major shortcomings. During installation, care
Estimation of the hydraulic gradient using the matric must be taken to ensure good hydraulic contact between
suction, osmotic potential, and water vapor potential the cup and the soil.
are possible.
The generic class of instruments used to measure Time-Domain Reflectometry
relative humidity using both wet and dry bulb ther-
mometers are called psychrometers; they may be used for Time-domain reflectometry (TDR) allows shallow, rapid,
both surface (agriculture) and subsurface (environmental) in situ measurements of water content and electrical
applications. For the latter, they are placed in a small conductivity in porous media (14–17). It is commonly
diameter well and surrounded by silica flour or air-filled used to locate underground cables. It is a relatively new
packers. A dry bulb thermometer is employed to measure development in vadose zone monitoring that was first
the ambient temperature, and a wet bulb thermometer is described by Topp et al. (18). Accuracy within 1–2% of
then used to measure the temperature decrease resulting volumetric water content is feasible with TDR.
from fluid evaporation in the subsurface (evaporative cool- TDR uses a probe that is inserted into the soil equipped
ing). Liquid-filled thermometers are not accurate enough with 2 or 3 metallic rods that transmit electromagnetic
for the precision required, however, and thus more precise pulses. The travel time of these pulses across the rods
thermocouple thermometers are employed. The junction through the soil is used to compute the soil’s bulk dielectric
of two wires, a thermocouple, generates an electrical cur- constant from which the water content is inferred, which
rent that is a function of the junction temperature. The provides an average soil water content over the depth of
temperature of the thermocouple determines the activity probe insertion.
of the water surrounding it, which is then correlated with The main advantage of TDR is that simple and rapid
the evaporation rate. accumulation of shallow soil moisture content with high
The direction and magnitude of water and contaminant volumetric accuracy is possible. Calibration requirements
migration may be observed easily and inexpensively with are minimal compared to other indirect monitoring
psychrometers, but regular calibration and monitoring methods, and measurements for spatial and temporal
of the thermocouple are necessary as temperature analyses are easily obtained. The drawbacks associated
deviations of ±0.001 ◦ C between the thermocouple and with TDR include temperature-dependent readings and
the liquid phase can introduce significant errors. Regular probe insertion in rocky or hard substrates because
 VADOSE ZONE MONITORING TECHNIQUES 541

readings are limited to depth of insertion. Additionally, moisture potentials. Standard calibration curves exist, but
water molecules bound by interfacial forces in high surface as in many indirect techniques, soil-specific calibration
area substances, for example, clays, result in lowered of the instrument is recommended. Other limitations
bulk dielectric constant compared to other soils at similar are associated with dissolution of the gypsum block and
water content. salinity changes in vadose zone water.

Ground-Penetrating Radar
DIRECT MONITORING METHODS
Ground-penetrating radar (GPR) is an adaptation of
seismic reflection for shallow use that images subsurface Direct monitoring methods, in contrast to indirect
features by deploying radar waves. It may be used to monitoring methods, employ sampling of the vadose zone
image any homogeneity in the vadose zone, including for liquid, soil, or gas. Obtaining liquid samples is fairly
contaminant plumes and soil stratification. straightforward, using the most widely practiced method,
GPR is used to observe the electromagnetic impedance suction lysimeters, but techniques for soil and gas sample
of soils using high-frequency radar waves to measure acquisition are generally more complicated. Soil sampling
subsurface properties. It consists of an enclosed wave commonly involves trenching or digging of some sort,
emitter and antenna unit that is dragged across the which may be subject to Occupational Safety and Health
ground, generally in a grid pattern. Wave pulses are Administration (OSHA) guidelines (24). Trenches deeper
emitted and when an inhomogeneity is encountered, part than 5 feet are subject to OSHA regulations and
of the incident energy is reflected back to the radar diggings deeper than 20 feet require the approval of a
antenna within the unit. The reflected signal is amplified professional engineer. Soils are also subject to maximum
and transformed into a viewable image. The data are in allowable slope protocol, depending on their nature. In
visual form, not a quantified form, so results must be addition, certain cases may require confined space, heavy
interpreted. machinery, or decontamination permits. Readers are
Any inhomogeneity in the vadose zone can be imaged referred to OSHA literature for proper digging protocol. In
with GPR. For vadose zone monitoring, contaminant gas sampling, problems stem from the nature of the phase
plumes of both light nonaqueous phase liquids (LNAPLs) being sampled, and, to a lesser extent, safety concerns.
and dense nonaqueous phase liquids (DNAPLs), as well Vadose zone gas readily exchanges with any other gas
as buried utility lines and pipes may be imaged with with which it comes in contact, and thus samplers must
GPR. Additionally, both underground storage tanks (UST) be fully enclosed. Special safety procedures are required
and leaking underground storage tanks (LUST) may be when considering volatile substances, such as methane.
imaged and discerned. The most obvious drawback to GPR, Additionally, direct monitoring methods have a much
however, is the fact that all data must be interpreted from higher risk of introducing contamination into the aquifer
visual images. Heterogeneities are viewed as anomalies or exposing workers to contamination, and thus proper
in imaged data, which must be deciphered according field procedures are absolutely necessary. In contaminated
to available information on subsurface conditions. Due areas, all equipment must be constructed of nonreactive
to this fact, GPR benefits from use in combination and low sorbing materials and be decontaminated between
with other vadose zone monitoring techniques to avoid each use (25–27).
misinterpretations (19,20).
Direct Liquid Monitoring Techniques
Electrical Resistance Blocks
Directly monitoring the liquid phase of the vadose zone
Electrical resistance blocks have been used in agriculture involves obtaining a representative sample, whether in
for more than 50 years to measure soil moisture in dry the form of water or NAPL. These may be important
areas. In the environmental field, they are employed for soil–water chemistry analyses or contaminant plume
where other methods, including tensiometers, lysimeters, delineation, among others. The primary method of
and manometers, are not operable due to low soil–water collection is by using suction lysimeters placed into
tension, generally between 0.5 and 15 bars. unconsolidated soil. As a suction device, lysimeters rely
The method consists of two or more electrodes that on applying a vacuum greater than the soil–water tension
are placed in a porous block, commonly made of gypsum. to draw a water sample up to the surface. A porous cup is
The block is buried in the soil and allowed to remain inserted into the soil or in a shallow well, and a vacuum
until moisture equilibrium with the native soil is tube connects it to the surface. Once the vacuum is applied,
achieved. Changes in the electrical properties of the block, water is drawn to the surface and a sample is obtained.
measured with the electrodes, reflect changes in the water Outside of lysimeters, other methods involve driving
content, which are measured with a wheatstone bridge devices into unconsolidated sediment or lowering devices
resistance meter. down wells above the water table to acquire liquid samples.
The benefits of this technique include low-cost, Lysimeters have the advantage of being relatively cheap
extensive historical use in agriculture research, and and simple to install. Virtually no maintenance is required,
simplicity of use. It is generally useful for approximate although sample volumes may be small.
soil moisture level changes. The disadvantages include Described by Gardner (28), wavelengths of light from
limited historical use in the environmental industry gamma to X ray may be used to monitor the water content
and measurement and calibration difficulties at wet soil of the vadose zone by direct soil sampling. In practice,
542 VADOSE ZONE MONITORING TECHNIQUES

however, only gamma rays are employed. In this method, 3. Reynolds, W.D. and Elrick, D.E. (1991). Determination of
a sample is collected and a narrow beam of gamma hydraulic conductivity using a tension infiltrometer. Soil Sci.
radiation is sent through it. A recorder on the opposite Soc. Am. J. 55: 633–639.
side of the sample records only those rays that passed 4. Hussen, A.A. and Warrick, A.W. (1993). Algebraic models
freely through the sample with no reflection along the for disc tension permeameters. Water Resour. Res. 29:
way. By this method, the volumetric water content can be 2779–2786.
measured (29). 5. Amoozegar, A., Martin, K.C., and Hoover, M.T. (1989). Effect
of access hole properties on soil water content determination
Direct Soil Monitoring Techniques by neutron thermalization. Soil Sci. Soc. Am. J. 53: 330–335.
Due to the sorbing nature of many contaminants, soil 6. Arslan, A., Razzouk, A.K., and Alain, F. (1997). The perfor-
mance and radiation exposure of some neutron probes in
samples are used to monitor the spatial and temporal
measuring the water content of the topsoil layer. Aust. J. Soil
distributions of contaminants in the soil. For certain appli-
Res. 35(6): 1397–1407.
cations, undisturbed samples are desired, in which case
7. Dillel, D. (1982). Introduction to Soil Physics. Academic Press,
special precautions must be taken, and certain sampling
London.
tools and techniques must be employed. Generally, soil
8. Grismer, M.E., Bali, K.M., and Robinson, F.E. (1995). Field
sampling involves inserting a device into the ground to
scale neutron probe calibration and variable analysis for clay
recover a soil sample, although more intensive methods
soil. J. Irrig. Drainage Eng.—ASCE 121(5): 354–362.
may be necessary in certain situations. Hand-operated
9. Comstock, J.P. (2000). Correction of thermocouple psychrom-
samplers are typically used for shallow sampling and
eter readings for the interaction of temperature and actual
mechanically powered devices for deeper sampling. There water potential. Crop Sci. 40(3): 709–712.
are five basic hand-operated samplers, as described by Dor-
10. Rawlins, S.L. and Campbell, G.S. (1986). Water potential:
rance et al. (30), that include screw augers, barrel augers,
Thermocouple psychrometry. In: Methods of Soil Analysis.
tube samplers, hand-powered augers, and bulk samplers A. Klute (Ed.). Part 1, 2nd Edn. pp. 597–618.
(shovels, etc). Each type has its own pros and cons depend-
11. Nyhan, J.W. and Drennon, B.J. (1990). Tensiometer data
ing on cost, ease of use, availability, size of matrix grains, acquisition system for hydrologic studies requiring high
degree of sediment cohesiveness, and other factors. Due temporal resolution. Soil Sci. Soc. Am. J. 54(1): 293–296.
to these, sampler choice is often site specific. For deeper 12. Williams, T.H.L. (1978). Automatic scanning and recording
monitoring applications, mechanically powered tools such tensiometer system. J. Hydrol. 39(1–2): 175–183.
as a drill rig must be employed. 13. Whalley, W.R., Leedsharrison, P.B., Joy, P., and Hoefsloot, P.
(1994). Time-domain reflectometry and tensiometry combined
Direct Vadose Zone Gas Monitoring
in an integrated soil–water monitoring system. J. Agric. Eng.
Monitoring the gas phase of the vadose zone is impor- Res. 59(2): 141–144.
tant because many contaminants volatilize to a significant 14. Jones, S.B., Wraith, J.M., and Or., D. (2002). Time domain
degree. Unlike liquid and soil phases, the gaseous compo- reflectometry measurement principles and applications.
nent of the vadose zone is problematic because atmospheric Hydrologic Processes 16(1): 141–153.
air and foreign vadose zone gases readily exchange and 15. Long, D.S., Wraith, J.M., and Kegel, G. (2002). A heavy-duty
react upon exposure. When sampling, all gaseous vadose time domain reflectometry soil moisture probe for use in
zone samples must avoid exposure to foreign air, and all intensive field sampling. Soil Sci. Soc. Am. J. 66(2): 396–401.
sampling methods must be designed to avoid it. In situ 16. Lee, J., Horton, R., and Jaynes, D.B. (2002). The feasibility of
sampling and devices with internal liners designed to seal shallow time domain reflectometry probes to describe solute
samples from the atmosphere are preferred (31). transport through undisturbed soil cores. Soil Sci. Soc. Am.
Monitoring through sampling the gas phase typically J. 66(1): 53–57.
involves hydraulically ramming a specially designed 17. Noborio, K. (2001). Measurement of soil water content and
device into the soil where it collects samples via a vacuum electrical conductivity by time domain reflectometry: A
and pumps them to the surface for collection. Passive gas review. Comput. Electron. Agric. 31: 213–237.
sampling by an in situ absorbent, such as an activated 18. Topp, G.C., Davis, J.L., and Annan, A.P. (1980). Electromag-
carbon trap placed in the vadose zone, may be used, netic determination of soil–water content: measurement in
but less information is provided by this method. Direct coaxial transmission lines. Water Resour. Res. 16: 574.
sampling of gases is more desirable, but it is more difficult 19. Daniels, J.J., Roberts, R., and Vendl, M. (1995). Ground
and costly. An alternative is monitoring gas advection penetrating radar for detection of liquid contaminants. Appl.
by using groundwater tracers. A tracer that partitions to Geophys. 33: 195–207.
the gas phase is injected, samples are taken at other 20. Weil, G.J., Graf, R.J., and Leann, M.F. (1994). Investigations
sites, typically at wells, and analyzed for the tracer of hazardous waste sites using thermal IR and ground
concentration (32–35). penetrating radar. Photogrammetric Eng. Remote Sensing
60(8): 999–1005.
BIBLIOGRAPHY 21. Everett, L.G., Wilson, G.G., and Hoylman, E.W. (1984).
Vadose Zone Monitoring for Hazardous Waste Sites. Noyes,
1. Ankeny, M.M., Ahed, M., Kaspar, T.C., and Horton, R. (1991). Park Ridge, NJ.
Simple method for determining unsaturated hydraulic 22. Fishback, P.E. (1977). Scheduling irrigations by electri-
conductivity. Soil Sci. Soc. Am. J. 55: 467–470. cal resistance blocks. NebGuide G 77–340, Institute of
2. Lin, H.S. and McInnes, K.J. (1995). Water-flow in clay soil Agriculture and Natural Resources, University of Nebraska,
beneath a tension infiltrometer. Soil Sci. 159(6): 375–382. Lincoln, NE.
 VAPOR TRANSPORT IN THE UNSATURATED ZONE 543

23. Strangeways, I.C. (1983). Interfacing soil moisture gypsum Soil and groundwater that are contaminated by organic
blocks with a modern data-logging system using a simple, low solvents and fuel hydrocarbons may emit vapors that can
cost, dc method. Soil Sci. 136: 322. move through the subsurface and reach the land surface.
24. http://www.osha.gov/ (Occupational Safety and Health Human exposure to these harmful vapors has become an
Administration). increasingly important topic in the evaluation and remedi-
25. Fetter, C.W. (1999). Contaminant Hydrogeology, 2nd Edn. ation of contaminated soil and groundwater. The potential
Prentice-Hall, Upper Saddle River, NJ. for inhalation of vapors transported from a groundwa-
26. Hudak, P.F. and Loaiciga, H.A. (1999). Conjunctive vadose ter plume to the interior of buildings often controls the
and saturated zone monitoring for subsurface contamination. outcome of a risk based corrective action (RBCA) anal-
Environ. Monitoring Assessment 59(1): 15–29. ysis (1). Additionally, vapors emanating from subsurface
27. Mickam, J.E., Bellandi, R., and Tifft, E.C., Jr. (1989). Equip- contamination contribute to the reduction of contaminant
ment decontamination procedures for ground water and mass and may provide a beneficial mechanism of natural
vadose zone monitoring programs: Status and prospects.
attenuation at sites where there are no human receptors.
Ground Water Monitoring Rev. 9(2): 100–121.
Subsurface processes affecting the transport of organic
28. Gardner, W.H. (1986). Water content. In: Methods of Soil
vapors through the soil above the water table (here defined
Analysis. Part 1. A. Klute (Ed.). Monograph 9, American
Society of Agronomy, Madison, WI.
as the unsaturated zone) may slow, reduce, or enhance the
release of these vapors to potential receptors at the land
29. Wilson, L.G. (1990). Methods for sampling fluids in the vadose
surface. A careful investigation of organic vapor trans-
zone. In: Ground Water and Vadose Zone Monitoring. ASTM
STP 1053. D.M. Nielsen and Z.I. Johnson (Eds.) American port in the unsaturated zone is therefore important in
Society for Testing and Materials, Philadelphia, PA, pp. 7–24. assessing the potential for natural attenuation and the
30. Dorrance, D.W., Wilson, L.G., Everett, L.G., and Cullen, S.J. impact of even deep contamination on receptors at the
(1994). A compendium of soil samplers for the vadose zone. In: land surface. This article describes potential sources, fate
Handbook of Vadose Zone Characterization and Monitoring. and transport, and measurement of organic vapors in
L.G. Wilson, G.E. Lorne, and S.J. Cullen (Eds.). CRC Press, the unsaturated zone. Although this article deals exclu-
Boca Raton, FL, pp. 401–428. sively with organic vapors, many of the fate and transport
31. Lewis, T.E., Crockett, A.B., Siegrist, R.L., and Zarrabi, K. properties discussed are also applicable to other vapors,
(1991). Soil sampling and analysis for volatile organic including metals and moisture.
compounds. E.P.A. Ground Water Forum Issue, EPA/540/4-
91-001.
SOURCES OF ORGANIC VAPORS
32. Fetter, C.W. (1999). Contaminant Hydrogeology, 2nd Edn.
Prentice-Hall, Upper Saddle River, NJ.
Contaminants of concern in vapor transport in the
33. Hubbell, J.M., Wood, T.R., Higgs, B., Wylie, A.H., and McEl-
roy, D.L. (1998). Design, installation, and uses of combination unsaturated zone are typically volatile organic compounds
groundwater and gas sampling wells. Ground Water Monitor- (VOCs), although vapors emanating from inorganic
ing and Remediation 18(3): 151–157. sources such as mercury vapor may be of concern as
34. Marrin, D.L. (1988). Soil-gas sampling and misinterpretation. well. A chemical is considered volatile if its Henry’s law
Ground Water Monitoring Rev. 8(2): 51–54. constant is 1 × 10−5 atm-m3 /mol or greater (2). Examples
35. Thompson, G. and Marrin, D. (1987). Soil gas contaminant of VOCs that are important in impacted environmental
investigations: A dynamic approach. Ground Water Monitor- systems include chlorinated solvents such as carbon
ing Rev. 8(4): 88–93. tetrachloride, tetrachloroethylene, and trichloroethylene
(and their degradation compounds), fuel hydrocarbons
such as benzene, toluene, ethylbenzene, and o,m,p-
VAPOR TRANSPORT IN THE UNSATURATED xylenes as well as volatile pesticides such as chlordane,
ZONE aldrin, and lindane. The U.S. Environmental Protection
Agency lists 107 compounds whose toxicity and volatility
FRED D. TILLMAN produce a potentially unacceptable inhalation risk to
U.S. Environmental Protection receptors (2). These VOCs can be released into the
Agency subsurface environment from leaking landfill liners,
Athens, Georgia improper disposal, accidental spillage, or leaking storage
tanks (LSTs). Once in the subsurface, these compounds can
JAMES A. SMITH become bound to the soil matrix, dissolved in groundwater
University of Virginia (or soil water), and/or exist as a separate, residual
Charlottesville, Virginia phase known as a nonaqueous phase liquid (NAPL).
Soil, aqueous, and NAPL-phase organics may all be
sources of organic vapors in the subsurface. Therefore,
INTRODUCTION organic vapor transport in the unsaturated zone requires
understanding of interphase mass-transfer processes
The term ‘‘vapor’’ describes the gaseous phase of a sub- because the contaminant can be distributed among soil
stance that is normally a liquid or solid at standard atmo- gas, water, soil, and NAPL phases. In modeling organic
spheric pressure and temperature. Examples of vapors vapor transport, the equilibrium vapor concentration at
that are important in environmental systems include the source of contamination is often estimated from the
humidity (water vapor), gasoline vapor, and solvent vapor. following relationships (3):
544 VAPOR TRANSPORT IN THE UNSATURATED ZONE

Soil contamination: [L2 T−1 ], and dC/dx is the concentration gradient of

the compound in one dimension [ML−3 L−1 ]. From this
HTS CR ρb equation, it is apparent that the rate of molecular
Csource = (4) (1)
θw + Kd ρb + HTS θa diffusion in the gas phase depends on the concentration
gradient and the effective diffusion coefficient of the
Groundwater contamination: compound of interest. Several relations relate the effective
diffusion coefficient to the free-air diffusion coefficient
Csource = HTS Cw Henry’s Law (2) of the compound and the total and air-filled porosities
of the diffusing media (5–14). An increase in diffusive
NAPL contamination: flux is seen in soils with greater air passageways (i.e.,
greater porosities and air-filled porosities). Therefore, in
xi pi MWi a layered unsaturated zone, vapor diffusion from depth
Csource = Raoult’s Law (3)
RT to the land surface is limited by the wettest, least
porous soil layer. Because free-air diffusion coefficients are
where Csource is the vapor concentration at the source of compound dependent, ranging from 2.5 × 10−3 cm2 s−1 for
contamination (ML−3 ), HTS is the dimensionless Henry’s hexachloroethane to 2.71 × 10−1 cm2 s−1 for chloroethane
law constant at the system temperature (unitless), CR among the 107 volatile compounds of concern to the
is the soil concentration (MM−1 ), ρb is the soil dry U.S.EPA (2), different chemicals diffuse at different
bulk density (ML−3 ), θw is the soil water-filled porosity rates under the same concentration gradients. Increased
(L3 L−3 ), θa is the soil air-filled porosity (L3 L−3 ), Kd temperature also produces an increased free-air diffusion
is the soil–water partition coefficient (L3 M−1 ), Cw is coefficient, leading to a greater rate of diffusion relative to
concentration of the contaminant in the groundwater the same system at lower temperatures. Fick’s First Law
(ML−3 ), xi is the mole fraction of component i in the NAPL may not sufficiently describe vapor diffusion in systems
mixture (unitless), pi is the vapor pressure of component i where pore sizes are very small (Knudsen diffusion)
at the soil temperature (ML−1 T−2 ), MWi is the molecular and when volatile species constitute a substantial
weight of component i (M), R is the ideal gas constant fraction of the total soil-gas concentration (nonequimolar
(ML2 T−2 T−1 M−1 , e.g., 8.314570 J◦ K−1 mol−1 ), and T is the diffusion) (15–17). In these situations, it may be necessary
soil temperature (T). Both Henry’s law constant and vapor to employ the Stefan–Maxwell equation.
pressure rise with temperature for organic compounds of
interest, so a higher source vapor concentration is expected Advection
from higher soil and groundwater temperature.
The flow of soil gas (advection) in the subsurface may
be caused by gas-pressure gradients or, in certain cases,
TRANSPORT AND FATE OF ORGANIC VAPORS IN THE gas density gradients. Pressure-driven advection is pro-
UNSATURATED ZONE duced when differences in soil-gas pressure form, causing
soil gas to flow and carrying any vapors present with
Organic vapors emanating from contaminated soil or it. Air pressure gradients in the subsurface of natural
groundwater or from a residual phase such as gasoline systems may result from several phenomena. As diurnal
floating on the water table may move through unsaturated or weather related atmospheric pressure cycles occur at
zone soil gas by diffusion or soil-gas advection due the land surface, pressure waves are transmitted into the
to pressure or density gradients or a combination of unsaturated zone and air may flow in response—a pro-
these processes. cess known as ‘‘barometric pumping.’’ Barometric pump-
ing may cause soil gas to flow either toward the land
Diffusion surface carrying soil vapor or away from the land sur-
Molecular diffusion is the spreading out of compounds from face bringing in fresh atmospheric air (13,15,18–25). The
random collisions resulting from thermal motion of atoms. underpressurization of an overlying building will pro-
These collisions may be between molecules themselves or duce gas pressure differences in subsurface soils. This
between molecules and their surroundings. Under most underpressurization may be caused by thermal differ-
environmental conditions, molecular diffusion in natural ences between indoor and outdoor air (stack effects), wind
systems proceeds from locations of higher concentration loading on the building superstructure, and unbalanced
toward locations of lower concentrations. In a typical building ventilation (26,27). Soil-gas pressure gradients
scenario, organic vapors above a contaminated water table may also be produced by a rapidly rising or falling water
(high concentration) diffuse toward the land surface (lower table, as in coastal zones (28), or through the buildup
concentration). The well-known relation describing the of gas pressure from decomposing organic matter inside
diffusion of a compound across a unit of cross-sectional a landfill (29). Finally, natural temperature differences
area is Fick’s first law: between warmer deep and cooler shallow soil gas will
cause soil gas to rise (30,31). Density-driven flow of organic
dC vapors may occur in the vicinity of residual-phase organic
Fx = −Deff (4) compounds whose saturation gas densities are greater
dx
than that of air. As organic liquids with high vapor pres-
where Fx is the mass flux [ML−2 T−1 ], Deff is the effective sures and molecular weights volatilize, the density of the
diffusion coefficient of the compound in the gas phase soil gas surrounding the liquid changes. In almost all
 VAPOR TRANSPORT IN THE UNSATURATED ZONE 545

cases, organic liquids have molecular weights that are VAPOR SAMPLING AND FLUX MEASUREMENT
greater than air, so the resulting density-driven flow is
in a downward direction and is proportional to soil per- Organic vapors in the unsaturated zone above contami-
meability and density differences between the vapor and nated soil or groundwater are typically sampled to pro-
air (32–34). Organic compounds for which density-driven vide evidence of biodegradation (49,55), calibrate vapor
advection may be significant include methylene chloride, flow and transport models for estimating natural atten-
1,2-dichloroethylene, 1,1,1-trichloroethane, carbon tetra- uation (23), or to estimate the region of contaminated
chloride, and 1,1-dichloroethane, among others (32–34). groundwater (42,56,57). Soil gas is often sampled through
low volume, inert pipe or tubing penetrating the sub-
surface to the desired sampling depth (42,44,49,56,58,59).
Sorption This tubing can be installed with the help of a Geoprobe-
type device or inserted in a previously drilled borehole.
As organic vapors move through the unsaturated zone by
The sampling portion of the tubing must be sealed from
diffusion and advection, they come in contact with soil
the atmosphere and other sampling depths to prevent
moisture, infiltrating rainwater, and the soil matrix itself.
preferential flow down the borehole during sampling. Sam-
Each of these interactions may affect the concentration
ples may be collected from these vapor probes in various
of the contaminant in the soil gas. Depending on the ways, including direct syringe withdrawal (49), pumping
compound, organic vapors may adsorb to soil grain of soil gas into Tedlar bags (57,60,61), or pumping of
surfaces or partition into soil organic matter (35–40). soil gas through adsorbent traps (58,62,63). Analyses of
Adsorption of relatively nonpolar organic vapors is samples collected by syringe withdrawal require on-site
suppressed by high humidity in the subsurface because as analysis. Samples collected in Tedlar bags or onto adsor-
polar water molecules can effectively out-compete organic bent traps may be transported to an off-site laboratory for
vapors for mineral-surface adsorption sites (41,42). For analysis. Alternatively, there are field-portable gas detec-
these high-humidity conditions, sorption may be limited tors and gas chromatographs with flame-ionization and
to organic vapor partition into soil organic matter (41,42). photoionization detectors capable of measuring real time
Soil moisture trapped in unsaturated-zone pore space hydrocarbon concentrations (47,57,64).
or infiltrating rain water may also sorb organic vapors Efforts to measure the flux of VOCs from the unsat-
to varying degrees (43,44). Finally, gas-phase organic urated zone to the atmosphere at the land surface
compounds may adsorb to the air–water interface have included using the aerodynamic method, which
in unsaturated zones (45,46). Each of these sorption necessitates the measurement of trace concentrations
processes may act as both a source and a sink of organic and atmospheric conditions such as wind speed, tem-
vapors in the unsaturated zone. perature and pressure (65); using groundsheets sampled
through a nylon sampling tube (66); and several cham-
Biodegradation ber designs (58,59,67–71). Surface measurements of vapor
flux are important because they provide a direct measure-
Under favorable conditions, organic vapors in the unsatu- ment of total natural flux that can be used to estimate
rated zone that partition into soil moisture may be biode- volatilization rates for a field site. This natural attenua-
graded by indigenous bacteria in oxidation/reduction reac- tion process for contaminants in soil and groundwater may
tions. The aerobic biodegradation of petroleum hydrocar- be significant, and, at times, approach the removal rate of
bons requires an abundant oxygen supply as well as neces- active remediation systems (58). Complexities involved in
sary nutrients of nitrogen and phosphorus (47–51). When the measurement of vapor transport from the subsurface
oxygen is depleted, other possible electron acceptors for to the atmosphere include possible preferential flow paths
biodegradation of petroleum hydrocarbons include nitrate (e.g., underground utilities, macropores, etc.) and spatial
(NO3 − ), iron oxides (e.g., Fe(OH)3 ), sulfate (SO4 2− ), and and temporal variability in vapor concentrations. Spatial
carbon dioxide (CO2 ) (48). Lightly chlorinated compounds variability may be produced by subsurface heterogeneities
(e.g., chlorobenzene, dichlorobenzene) may be biodegraded that are geologic as well as constructed barriers (e.g.,
under aerobic conditions. The more highly chlorinated storm drains, basement foundations). Temporal variability
hydrocarbons are recalcitrant to aerobic biodegradation in vapor concentrations is produced by changing soil mois-
but may undergo direct or cometabolic anaerobic reduc- ture content, temperature, and gas-pressure gradients.
tive dechlorination. Reductive dechlorination has been
observed to be most effective under sulfate-reducing and SUMMARY
methanogenic conditions (52). In direct reductive dechlori-
nation, the chlorinated hydrocarbon is used as an electron Analysis of the transport of organic vapors in the unsatu-
acceptor and the bacteria gain energy and grow as a result rated zone is fundamental in evaluating potential human
of the reaction (53). In cometabolic reductive dechlori- exposure to harmful compounds from a contaminated sub-
nation, enzymes produced during microbial metabolism surface as well as determining a site’s potential for natural
of another hydrocarbon fortuitously reduce and dehalo- attenuation. Analysis of soil gas vapors may also aid in
genate the chlorinated contaminant. For either reductive estimating the areal extent of a contaminated plume. Soil,
dechlorination process to be successful, a primary sub- aqueous, and NAPL-phase organics may all be sources
strate (electron donor) such as soil organic matter, lactate, of organic vapors in the subsurface, and interphase mass
acetate, methanol, or glucose is necessary (54). transfer to and from soil, soil moisture, groundwater, and
546 VAPOR TRANSPORT IN THE UNSATURATED ZONE

soil gas all play a role in vapor transport in the unsaturated 14. Rolston, D.E. and Moldrup, P. (2002). Gas diffusivity. In:
zone. Diffusion of the vapor phase contaminant proceeds Methods of Soil Analysis: Part 4-Physical Methods. J.H. Dane
from higher concentrations near the source toward lower and G.C. Topp (Eds.). Soil Science Society of America,
concentrations at the land surface. Advection of soil gas Madison, WI, pp. 1113–1158.
results from pressure gradients in the subsurface and 15. Thorstenson, D.C. and Pollock, D.W. (1989). Gas transport
from density-driven flow for certain high molecular weight in unsaturated zones: Multicomponent systems and the
adequacy of Fick’s laws. Water Resour. Res. 25: 477–507.
compounds. Under favorable conditions, organic vapors in
16. Baehr, A.L. and Bruell, C.J. (1990). Application of the Stefan-
the unsaturated zone may be biodegraded by indigenous
Maxwell equations to determine limitations of Fick’s law
bacteria in oxidation/reduction reactions. All biodegrada-
when modeling organic vapor transport in sand columns.
tion reactions of organic compounds require nutrients and Water Resour. Res. 26: 1155–1163.
electron acceptors in sufficient, usable quantities. Soil-gas
17. Partridge, G.P., McLeary, K.S., Showers, T.R., Huebner, R.S.,
concentrations may be quantified through the use of vapor and Elliott, H.A. (2002). Single and multicomponent gas
probes, adsorbent traps, Tedlar bags, direct syringe with- phase diffusion in porous media: Modeling and laboratory
drawal with field-portable analytical equipment, and the measurements. Soil Sediment Contam. 11: 555–581.
use of gas chromatography. The total flux of contaminants 18. Sleep, B.E. and Sykes, J.F. (1989). Modeling the transport of
from the subsurface to the land surface may be estimated volatile organics in variably saturated media. Water Resour.
by the aerodynamic method, through the use of ground Res. 25: 81–92.
sheets, or with flux chambers. Surface flux measurement 19. Massmann, J. and Farrier, D.F. (1992). Effects of atmospheric
is useful for estimating the total volatilization of contam- pressures on gas transport in the vadose zone. Water Resour.
inants from a site, but all vapor-phase measurements are Res. 28: 777–791.
complicated by spatial and temporal variability. 20. Auer, L.H., Rosenberg, N.D., Birdsell, K.H., and Whiteny,
E.M. (1996). The effects of barometric pumping on contami-
nant transport. J. Contam. Hydrol. 24: 145–166.
BIBLIOGRAPHY
21. Elberling, B., Larsen, F., Christensen, S., and Postma, D.
(1998). Gas transport in a confined unsaturated zone
1. Laubacher, R.C., Bartholomae, P., Velasco, P., and Reisinger,
during atmospheric pressure cycles. Water Resour. Res. 34:
H.J. (1997). An evaluation of the vapor profile in the vadose
2855–2862.
zone above a gasoline plume. In: Pet. Hydrocarbons Org.
Chem. Ground Water: Prev. Detection Remediation Conf., 22. Tillman, F., Choi, J.C., Katchmark, W., Smith, J.A., and
Houston, TX, pp. 396–409. Wood, H.G.I. (2001). Unsaturated-zone airflow: Implications
for natural remediation of ground water by contaminant
2. Environmental Quality Management. (2003). User’s Guide for
transport through the subsurface. In: Physicochemical
Evaluating Subsurface Vapor Intrusion into Buildings. U.S.
Groundwater Remediation. J.A. Smith and S.E. Burns (Eds.).
Environmental Protection Agency, Office of Emergency and
Kluwer Academic/Plenum, New York, pp. 307–339.
Remedial Response, Durham, NC.
23. Choi, J.W., Tillman, F.D., and Smith, J.A. (2002). Relative
3. U.S. Environmental Protection Agency. (2002). Draft Guid-
importance of gas-phase diffusive and advective trichloroethy-
ance for Evaluating the Vapor Intrusion to Indoor Air Pathway
lene (TCE) fluxes in the unsaturated zone under natural
from Groundwater And Soils. USEPA Office of Solid Waste
conditions. Environ. Sci. Technol. 36: 3157–3164.
and Emergency Response.
24. Neeper, D.A. (2002). Investigation of the vadose zone using
4. Johnson et al. (1990).
barometric pressure cycles. J. Contam. Hydrol. 54: 59–80.
5. Buckingham, E. (1904). Contribution to Our Knowledge of the
25. Rossabi, J. and Falta, R.W. (2002). Analytical solution for
Aeration of Soils. USDA Bureau of Soils Bulletin No. 25.
subsurface gas flow to a well induced by surface pressure
6. Penman, H.L. (1940). Gas and vapor movements in soil I. The fluctuations. Ground Water 40: 67–75.
diffusion of vapours through porous solids. J. Agric. Sci. 30:
26. Nazaroff, W.W., Lewis, S.R., Doyle, S.M., Moed, B.A., and
463–462.
Nero, A.V. (1987). Experiments on pollutant transport from
7. Van Bavel, C.H.M. (1952). Gaseous diffusion and porosity in soil into residential basements by pressure-driven airflow.
porous media. Soil Sci. 73: 91–105. Environ. Sci. Technol. 21: 459–466.
8. Rust, R.H., Klute, A., and Geiseking, G.E. (1957). Diffusion- 27. Garbesi, K. and Sextro, R.G. (1989). Modeling and field
porosity measurements using a non-steady state system. Soil evidence of pressure-driven entry of soil gas into a house
Sci. 84: 453–465. through permeable below-grade walls. Environ. Sci. Technol.
9. Dye, R.F. and Dallavale, J.M. (1958). Diffusion of gases in 23: 1481–1487.
porous media. Ind. Eng. Chem. 50: 1195–1200. 28. Li, L., Dong, P., and Barry, D.A. (2002). Tide-induced water
10. Millington, R.J. (1959). Gas diffusion in porous media. Science table fluctuations in coastal aquifers bounded by rhythmic
130: 100–102. shorelines. J. Hydraul. Eng. 128: 925–933.
11. Currie, J.A. (1970). Movement of gases in soil respiration. 29. Little, J.C., Daisey, J.M., and Nazaroff, W.W. (1992). Trans-
Sorption and Transport Processes in Soils. SCI Monogr. Ser. port of subsurface contaminants into buildings: An exposure
37: 152–171. pathway for volatile organics. Environ. Sci. Technol. 26:
12. Nilson, R.H., Peterson, E.W., and Lie, K.H. (1991). Atmo- 2058–2066.
spheric pumping: A mechanism causing vertical transport 30. Gustin, M.S., Taylor, G.E., and Maxey, R.A. (1997). Effect
of contaminated gases through fractured permeable media. of temperature and air movement on the flux of elemental
J. Geophys. Res. 96: 21933–21948. mercury from substrate to the atmosphere. J. Geophys. Res.
13. Bartelt-Hunt, S.L. and Smith, J.A. (2002). Measurement of 102: 3891–3898.
effective air diffusion coefficients for trichloroethylene in 31. Krylov, V.V. and Ferguson, C.C. (1998). Contamination of
heterogeneous soil. J. Contam. Hydrol. 56: 193–208. indoor air by toxic soil vapours: The effects of subfloor
 VAPOR TRANSPORT IN THE UNSATURATED ZONE 547

ventilation and other protective measures. Build. Environ. transport in the unsaturated zone. Water Resour. Res. 32:
33: 331–347. 2231–2249.
32. Falta, R.W., Javandel, I., Pruess, K., and Witherspoon, P.A. 50. Lahvis, M.A., Baehr, A.L., and Baker, R.J. (1999). Quantifi-
(1989). Density-driven flow of gas in the unsaturated zone cation of aerobic biodegradation and volatilization rates of
due to the evaporation of volatile organic compounds. Water gasoline hydrocarbons near the water table under natural
Resour. Res. 25: 2159–2169. attenuation conditions. Water Resour. Res. 35: 753–765.
33. Mendoza, C.A. and Frind, E.O. (1990). Advective-dispersive 51. Hers, I., Atwater, J., Li, L., and Zapf-Gilje, R. (2000). Eval-
transport of dense organic vapors in the unsaturated zone 1. uation of vadose zone biodegradation of BTX vapours. J.
Model development. Water Resour. Res. 26: 379–387. Contam. Hydrol. 46: 233–264.
34. Mendoza, C.A. and Frind, E.O. (1990). Advective-dispersive 52. U.S. Environmental Protection Agency. (2000). Engineered
transport of dense organic vapors in the unsaturated zone: 2. approaches to in situ bioremediation of chlorinated solvents:
Sensitivity analysis. Water Resour. Res. 26: 387–398. Fundamentals and field applications. EPA 542-R-00-008,
35. Goss, K.U. (1994). Adsorption of organic vapors on polar Washington, DC.
mineral surfaces and on a bulk water surface: Development 53. McCarty, P.L. (1997). Breathing with chlorinated solvents.
of an empirical predictive model. Environ. Sci. Technol. 28: Science 276: 1521–1522.
640–645. 54. Vogel, T.M., Criddle, C.S., and McCarty, P.L. (1987). Trans-
36. Goss, K.U. (1994). Effects of temperature and relative formations of halogenated aliphatic compounds. Environ. Sci.
humidity on the sorption of vapors on quartz sand. Environ. Technol. 21: 722–736.
Sci. Technol. 26: 2287–2294. 55. Roggemans, S., Bruce, C.L., Johnson, P.C., and Johnson, R.L.
37. Goss, K.-U. and Eisenreich, S.J. (1996). Adsorption of VOCs (2001). Vadose zone natural attenuation of hydrocarbon
from the gas phase to different minerals and a mineral vapors: An empirical assessment of soil gas vertical profile
mixture. Environ. Sci. Technol. 30: 2135–2142. data. Am. Pet. Inst. Tech. Bull. 15: 12.
38. Popovicova, J. and Brusseau, M.L. (1998). Contaminant mass 56. Kerfoot, H.B. (1987). Soil-gas measurement for detection of
transfer during gas-phase transport in unsaturated porous groundwater contamination by volatile organic compounds.
media. Water Resour. Res. 34: 83–92. Environ. Sci. Technol. 21: 1022–1024.
39. Ruiz, J., Bilbao, R., and Murillo, M. (1998). Adsorption of 57. Moseley, C.L. and Meyer, M.R. (1992). Petroleum Contami-
different VOC onto soil minerals from gas phase: Influence nation of an elementary school: A case history involving air,
of mineral, type of VOC, and air humidity. Environ. Sci. soil-gas and groundwater monitoring. Environ. Sci. Technol.
Technol. 32: 1079–1084. 26: 185–192.
40. Goss, K.-U., Buschmann, J., and Schwarzenbach, R.P. (2004). 58. Smith, J.A., Tisdale, A.K., and Cho, H.J. (1996). Quantifi-
Adsorption of organic vapors to air-dry soils: Model cation of natural vapor fluxes of trichloroethene in the
predictions and experimental validation. Environ. Sci. unsaturated zone at Picatinny Arsenal, New Jersey. Environ.
Technol. 38: 3667–3673. Sci. Technol. 30: 2243–2250.
41. Chiou, C.T. and Shoup, T.D. (1985). Soil sorption of organic 59. Tillman, F., Choi, J.-W., and Smith, J.A. (2003). A comparison
vapors and effects of humidity on sorptive mechanism and of direct measurement and model simulation of total flux
capacity. Environ. Sci. Technol. 19: 1196–1200. of volatile organic compounds from the subsurface to the
42. Smith, J.A., Chiou, C.T., Kammer, J.A., and Kile, D.E. (1990). atmosphere under natural field conditions. Water Resour.
Effect of soil moisture on the sorption of trichloroethene Res. 39: 1284–1294.
vapor to vadose-zone soil at Picatinny Arsenal, New Jersey. 60. Fischer, M.L. et al. (1996). Factors affecting indoor air
Environ. Sci. Technol. 24: 676–683. concentrations of volatile organic compounds at a site of
43. Cho, H.J. and Jaffé, P.R. (1990). The volatilization of organic subsurface gasoline contamination. Environ. Sci. Technol.
compounds in unsaturated porous media during infiltration. 30: 2948–2957.
J. Contam. Hydrol. 6: 387–410. 61. Jeng, C.Y., Kremesec, V.J., and Primack, H.S. (1996). Models
44. Cho, H.J., Jaffé, P.R., and Smith, J.A. (1993). Simulating of hydrocarbon vapor diffusion through soil and transport into
the volatilization of solvents in unsaturated soils during buildings. In: Pet. Hydrocarbons Org. Chem. Ground Water:
laboratory and field infiltration experiments. Water Resour. Prev. Detection Remediation Conf. National Ground Water
Res. 29: 3329–3342. Association, Houston, TX, pp. 319–338.
45. Kim, H., Rao, P.S., and Annable, M.D. (1997). Determination 62. Tillman, F. (2003). Design and testing of a chamber device
of effective air-water interfacial area in partially saturated to measure total flux of volatile organic compounds from
porous media using surfactant adsorption. Water Resour. Res. the unsaturated zone under natural conditions. Ph.D.
33: 2705–2711. Dissertation. University of Virginia, Charlottesville, VA.
46. Kim, H., Annable, M.D., and Rao, P.S.C. (1998). Influence of 63. Hodgson, A.T., Garbesi, K., Sextro, R.G., and Daisey, J.M.
air-water interfacial adsorption and gas-phase partitioning (1992). Soil-gas contamination and entry of volatile organic
on the transport of organic chemicals in unsaturated porous compounds into a house near a landfill. J. Air Waste Manage.
media. Environ. Sci. Technol. 32: 1253–1259. Assoc. 42: 277–283.
47. Ostendorf, D.W. and Kampbell, D.H. (1991). Biodegradation 64. Ririe, G.T., Sweeney, R.E., Daugherty, S.J., and Peuron, P.M.
of hydrocarbon vapors in the unsaturated zone. Water Resour. (1998). A vapor transport model that is consistent with field
Res. 27: 453–462. and laboratory data. Pet. Hydrocarbons Org. Chem. Ground
48. Norris, R.D., Hinchee, R.E., Brown, R., McCarty, P.L., Wil- Water: Prev. Detection Remediation Conf., Ground Water
son, J.T., Reinhard, M., Bouwer, E.J., Borden, R.C., Vogel, Association, Houston, TX, pp. 299–309.
T.M., Thomas, J.M., and Ward, C.H. (1994). Handbook of 65. Chen, C., Green, R.E., Thomas, D.M., and Knuteson, J.A.
Bioremediation. Lewis, Boca Raton, FL. (1995). Modeling 1,3-dichloropropene fumigant volatilization
49. Lahvis, M.A. and Baehr, A.L. (1996). Estimation of rates of with vapor-phase advection in the soil profile. Environ. Sci.
aerobic hydrocarbon biodegradation by simulation of gas Technol. 29: 1816–1821.
548 APPLICATIONS OF SOIL VAPOR DATA TO GROUNDWATER INVESTIGATIONS

66. Wyatt, D.E., Richers, D.M., and Pirkle, R.J. (1995). Baromet- Soil vapor samples are most commonly analyzed for
ric pumping effects on soil gas studies for geological and volatile organic compounds (VOCs). VOCs include simple
environmental characterization. Environ. Geol. 25: 243–250. hydrocarbons such as methane; lighter hydrocarbons such
67. Eklund, B., Balfour, W.D., and Schmidt, C.E. (1985). Mea- as gasoline; gasoline constituents such as benzene and
surement of fugitive volatile organic emission rates. Environ. methyl tert-butyl ether (MtBE); and chlorinated solvents
Prog. 4: 199–202. such as trichloroethylene (TCE), trichloroethane (TCA),
68. Batterman, S.A., McQuown, B.C., Murthy, P.N., and McFar- and tetrachlorethylene (PCE).
land, A.R. (1992). Design and evaluation of a long-term soil Soil vapor is surveyed for a variety of reasons.
gas flux sampler. Environ. Sci. Technol. 26: 709–714.
Historically, soil vapor surveys have been used primarily
69. Reinhart, D.R., Cooper, D.C., and Walker, B.L. (1992). Flux for site assessment to identify soil and groundwater
chamber design and operation for the measurement of
contamination (1–4). More recent applications include
municipal solid waste landfill gas emission rates. J. Air Waste
Manage. Assoc. 42: 1067–1070.
evaluation of upward and downward vapor transport to
monitor site remediation (natural and engineered) and to
70. Eklund, B., Anderson, E.P., Walker, B.L., and Burrows, D.B.
(1998). Characterization of landfill gas composition at the
monitor landfills and migrating VOCs (5–7).
Fresh Kills municipal solid-waste landfill. Environ. Sci. Soil vapor surveys are quick, relatively nondisruptive,
Technol. 32: 2233–2237. and generally inexpensive. The surveys are frequently
71. Tillman, F. and Smith, J.A. (2004). Design and laboratory performed early in the site investigation, when investiga-
testing of a chamber device to measure total flux of volatile tion for numerous VOCs is performed. The data obtained
organic compounds from the unsaturated zone under natural are often used to focus subsequent soil and groundwater
conditions. J. Contam. Hydrol. 75: 71–90. investigatory sampling programs.
Regulatory agencies across the country are increasingly
concerned with human health risk due to the upward
APPLICATIONS OF SOIL VAPOR DATA TO migration of contaminant vapors through the vadose
GROUNDWATER INVESTIGATIONS zone into buildings. Vertical soil vapor profiles enable
measuring actual soil vapor concentrations that give a
BLAYNE HARTMAN more accurate calculation of the upward vapor migration
than values calculated from soil or groundwater data.
H&P Mobile Geochemistry
Solana Beach, California These data are used in risk based corrective action (RBCA)
computer models (8).
JAMES A. JACOBS Use of soil vapor data for monitoring site remediation
Environmental Bio-Systems, Inc.
takes many forms. Most commonly, concentrations of
Mill Valley, California vapor contaminants are monitored to measure the
progress of a vapor extraction system. Remediation of
groundwater or soil by natural attenuation requires
This article describes the uses and interpretation of soil monitoring the concentrations of contaminants, by-
vapor data for a variety of common applications related products, and oxygen in the soil vapor to measure the
primarily to groundwater. It begins with an overview of progress of the degradation of the contaminant plume.
the applications of soil vapor surveys, a brief discussion Soil vapor is frequently measured at the fringes
of the various soil vapor methods, and then focuses on a of landfills to ensure that harmful compounds such
number of specific issues relating to soil vapor surveys for as methane, chlorinated compounds (primarily vinyl
groundwater investigations. The intent of this article is to chloride), and petroleum related carcinogens such as
guide professionals on when, why, and how soil vapor data benzene are not migrating from landfills to neighboring
can be useful for groundwater-related investigations and developments.
how to interpret the results.
SOIL VAPOR METHODS
BACKGROUND AND APPLICATIONS
Collection of Soil Vapor Samples
Soil vapor in the pore spaces, also called soil gas, soil Three methods exist for collecting soil vapor data:
air, pore gas, and interstitial gas refers to the vapor (1) active, (2) passive, and (3) flux chambers. A brief
that lies among the soil grains in the soil column lying overview of these methods follows. More comprehensive
between the ground surface and the water table (vadose descriptions of soil vapor collection methods and protocols
zone). Soil vapor data are collected in a variety of are given in documents written by Hartman (9,10),
ways, and the data can be used for various purposes. Hartman and Jacobs (5,11), EPA (1), ASTM (12), and
1
Devitt et al. (4).
Some of this material is excerpted from previous work,
The active approach is quantitative and consists of the
appearing in ‘‘Petroleum Hydrocarbon Contamination-Legal and
Technical Issues’’ and ‘‘Chlorinated Solvent Contamination- withdrawal of a volume of vapor from the subsurface,
Legal and Technical Issues’’ Copyright  1998, 1999, Argent typically using a steel sampling rod driven into the ground
Communications Group. Reprinted by permission of Argent from 0.9 to 6 meters below ground surface, and subsequent
Communication Group, P.O. Box 1425, Foresthill, CA 95631. analysis of the withdrawn vapor. Originally, surveys were
Telephone 1-800-419-2741; or sales@ argentco.com. limited to shallow depths of 3 meters because the sampling
 APPLICATIONS OF SOIL VAPOR DATA TO GROUNDWATER INVESTIGATIONS 549

rods were driven and extracted using hand techniques. Analysis of Soil Vapor Samples
Currently, active surveys can routinely achieve depths
The type of analytical instrumentation employed depends
exceeding 30.5 meters due to the use of direct-push probe
upon the goals of the project and the type of soil vapor
rigs, cone penetrometer (CPT) sampling rigs, and full-
method used. For screening-level active soil vapor surveys,
size drill rigs to collect soil vapor using newly developed
in which the primary goal is simply to find contamination
sampling tools.
qualitatively, simple handheld or portable instruments,
Sample containers for active vapor sampling include such as organic vapor analyzer (OVAs), organic vapor
syringes, gas sampling Tedlar bags, preevacuated meter (OVMs), gas-monitoring meters (e.g., IR detectors),
stainless steel canisters (e.g., SUMMA canisters), and and gas detector tubes, can be used. For applications
gastight vials (e.g., 40 mL preevacuated volatile organic where quantitative results are desired, the analytical
analysis (VOAs) vials or headspace vials). Analysis methodology employed is typically gas chromatography
can be performed on-site or off-site, usually within 48 (GC) or gas chromatography-mass spectrometry (GC-MS).
hours of collection. The quantitative results are typically These latter techniques differentiate a wide variety of
presented in concentration units (e.g., ppmv, µg/L vapor). compounds. These latter methods are also used to analyze
The active method requires actively withdrawing soil passive soil vapor samples and flux chamber samples.
vapor and is best suited to sites that have higher More comprehensive treatments of the available
vadose zone permeabilities and higher contaminant analytical options for soil vapor samples are given by
concentrations (>0.1 µg/L vapor). The active method can Hartman & Jacobs (5) and the EPA (1).
be used for investigating groundwater contamination,
and as discussed later in this chapter, the active
APPLICATIONS OF SOIL VAPOR DATA TO
method is also useful for groundwater issues relating
GROUNDWATER ISSUES
to contaminant fate and transport because it gives
concentration data.
This section discusses how to use and interpret soil vapor
The passive approach consists of emplacing a natural
data for applications specifically related to groundwater.
or synthetic absorbent into the subsurface (typically 0.15
to 0.9 meters) for an extended time, and subsequently
Assessing and Defining Groundwater Contamination
removing and analysing the absorbent. The absorbent is
typically placed in the upper end of an inverted container Soil vapor surveys have frequently been used for initial
or sleeve that has openings for migrating vapors to enter site assessment to identify groundwater contamination,
and collect. Unlike the active method, which requires one and the data are used to focus subsequent groundwater
visit to the site, the passive method requires one visit investigatory sampling programs (1,3,4). The soil vapor
to the site to deploy the absorbents and a second trip to data are typically plotted as contours or as a raised
retrieve them. surface. More sophisticated approaches, including three-
The passive method is qualitative and does not allow dimensional visualizations of the soil column, cross
for real-time data. Measured values are reported as total sections, normalization, and statistical analysis, are
absorbed mass (e.g., µg) or relative units because the sometimes used depending on the project goals.
amount of vapor meeting the absorbent is unknown. The applicability of a soil vapor survey to detect
The passive method offers more sensitivity than the potential groundwater contamination depends upon a
active method; hence it offers advantages in sites of number of factors, including the local geology, contaminant
low vadose zone permeability and lower contaminant properties, and the concentration of the contaminant in
concentrations (>0.01 µg/L vapor). Contaminants must the groundwater (11). The maximum concentration of a
have an appreciable vapor pressure to be detected by this contaminant in the soil vapor due to volatilization from
method, regardless of the deployment time or contaminant the groundwater can be estimated from Henry’s law:
concentration in the soil or water. As a rule of thumb, one
should not expect compounds whose vapor pressures are Csg = HCw (1)
less than 10−6 atmospheres to be detectable by passive soil
vapor methodology in a reasonable time frame (<50 days where Csg is the soil gas concentration in equilibrium
of exposure) for reasonable contaminant concentrations in with the groundwater,
the water. Cw is the groundwater concentration, and
Flux chambers consist of an enclosed chamber that H is the dimensionless Henry constant.
is placed on the surface for a specific period of time. As an example, consider groundwater that has a benzene
Vapor concentrations are measured in the chamber concentration of 10 µg/L. The dimensionless Henry
through time or after a period of time. This method constant for benzene is 0.22, giving an equilibrium soil
is also quantitative and yields both concentration data vapor concentration of 2.2 µg/L vapor at the groundwater
in the chamber and flux data (mass/area-time). Flux interface. Hence, a soil vapor survey conducted near
chambers are the least common soil vapor method and groundwater could at best be expected to detect only a
are seldom used for groundwater assessment. Their most maximum value of 2.2 µg/L in the vapor if groundwater
useful application is for sites where volatilization of were the only source.
contaminants from groundwater is of concern and direct Equilibrium phase partitioning is rarely achieved
fluxes are desired Hartman (9). in the subsurface, so measured soil vapor values
550 APPLICATIONS OF SOIL VAPOR DATA TO GROUNDWATER INVESTIGATIONS

from groundwater volatilization will be significantly gradient with respect to the depth below the surface.
lower (generally one to two orders of magnitude) than Diffusional transport is a three-dimensional process, so
predicted from Henry’s constants. In addition, soil vapor contaminants move away from a source in all directions,
concentrations decrease as distance increases from the similar to an expanding balloon.
groundwater due to other processes active in the vadose
zone, such as biodegradation and sorption onto soils (13). Differentiating the Source of Soil Vapor/Groundwater
For these reasons, the chances of detecting groundwater Contamination
contamination from soil vapor data increases the closer
A common problem confronted by environmental geolo-
the samples are collected to groundwater.
gists is determining the source of contamination at a site
Due to its increased sensitivity for some compounds,
that has both soil vapor and groundwater contamination.
the passive soil vapor method can offer advantages
Is the soil vapor responsible for the groundwater contami-
over the active method for this application. Similar to
nation, or is the groundwater responsible for the soil vapor
the active method, the ability of passive methods to
concentration? Contaminant transfer by gaseous diffusion
detect groundwater contamination increases the closer the
moves from areas of high concentration to areas of low con-
samples are collected to groundwater. However, regardless
centration, so vertical profiles of the soil vapor through the
of sampling depth, the capability of passive methods to
vadose zone down to groundwater can provide this answer.
detect contaminants is still limited by the vapor pressure
Figure 1 shows the expected shape of soil vapor profiles
of the compound of concern (5).
in the vadose zone for three common conditions. Processes
active in the vadose zone such as sorption, biodegrada-
Transport of Contaminant Vapors in the Subsurface
tion, and advection may alter these ideal profiles, but the
To understand the source of soil vapor and groundwater source of contamination must still be in the area of highest
contamination, it is necessary to understand how contam- concentration, as dictated by Ficks law.
inants move in the vapor phase. A condensed version of Vertical soil vapor is profiled by installing vapor
this topic is addressed in this section beginning with a dis- monitoring wells at multiple depths through the vadose
cussion of the processes by which vapors move through the zone to the groundwater. For soil vapor, vapor wells
vadose zone. A more detailed presentation can be found in are often constructed of small diameter 0.5 to 1.0 cm
Hartman (6,7,14). (1/8 in. to 1/4 in. OD) flexible tubing and ‘‘nested’’ in the
There are two types of physical processes by which same borehole (Fig. 2).
contaminants are transported in the vapor phase. The Repeated sampling of the vapor wells allows averaging
first is advection, the bulk movement of the vapor itself. any temporal effects. Thomson (19) describes two case
In advective transport, any contaminants in the vapor are histories of vertical profiling using nested vapor wells.
carried along with the moving vapor. Advective transport
processes can be important in soil vapor, for example, near Contaminant Vapor Transport Rates
the surface due to atmospheric pressure differences (15),
The distance that contaminant vapors can move by
near buildings that can create pressure gradients (16), and
gaseous diffusion can be approximated from
movement of vapor by density-induced flow (17).
The second type of transport process is gaseous
diffusion, the motion of contaminants by molecular Distance = (2 × De × t)1/2 (3)
processes through a nonmoving vapor column, which is
the primary mechanism for contaminant transport in the where De is the effective diffusivity and t is time.
vapor phase through the vadose zone (18).
Contaminant transport by gaseous diffusion is
described by Fick’s first law as

De Csg Depth
Surface source
FLUX = (2) Flux
X

where Flux is the rate of movement of a compound

per unit area,
Flux
De is the effective diffusion coefficient in the Depth Deep source
vadose zone,
Csg is the contaminant concentration gradient in
the soil vapor, and
X is depth in the vadose zone
Depth Flux Surface and deep sources
Similar to momentum transfer (e.g., water running
downhill) and heat transfer (movement from hot areas
to cold areas), contaminant transfer by gaseous diffusion
moves from areas of high concentration to areas of low Increasing concentration
concentration. The flux is always down the concentration Figure 1. Common soil vapor profiles in the vadose zone for
gradient, regardless of the orientation of the concentration several one-dimensional situations (8).
 APPLICATIONS OF SOIL VAPOR DATA TO GROUNDWATER INVESTIGATIONS 551

laterally underneath a neighboring room or building or

downward to the groundwater surface.

Into or out of Groundwater. In contrast to movement

through the vadose zone where contaminant vapors move
Surface cover
J-plug through the soil vapor, the movement of contaminant
vapors into or out of groundwater is controlled by the
Bentonite Concrete rate at which vapors partition into and move through the
liquid. Because groundwater movement is so slow, the
18" to 24" long water interface remains relatively undisturbed (laminar
2" diameter PVC flow), and vertical mixing of the water is minimal. The
pipe primary exchange process is again molecular diffusion,
Third soil gas but in this case, the exchange rate is controlled by liquid
implant #10/20 silica diffusion, not gaseous diffusion. A general value for the
sand pack liquid diffusion coefficient for compounds is approximately
0.00001 cm2 /s. Using the same factor of 10 reduction due
Bentonite to soil porosity, De for most liquids can be approximated
as 0.000001 cm2 /s.
The distance that contaminants can move into and
through groundwater in a year can be estimated as

Distance = (2 × 0.000001 cm2 /s × 31,536,000 s)1/2

Second soil #10/20 silica
gas implant sand pack ∼ 8 cm

These calculations show that although contaminant

vapors can move through the vadose zone relatively
quickly, they partition into and move through groundwater
Bentonite extremely slowly. The reverse situation is also true;
the movement and partitioning of contaminants through
groundwater and into soil vapor is also extremely slow.
First soil
gas implant #10/20 silica Estimation of the Downward Transport of Contaminant
sand pack
Vapors into Groundwater
The calculations summarized in the preceding sections
2" diameter indicate that although contaminant vapors can move
borehole quickly down to groundwater, they do not partition into
Figure 2. Schematic diagram of a multidepth, nested vapor groundwater very quickly because phase partitioning
monitoring well (19). across a liquid laminar interface is kinetically controlled.
For this reason, phase partitioning rarely reaches
equilibrium, and the use of Henry’s constants to estimate
Through the Vadose Zone. For contaminant vapor groundwater concentrations from soil vapor data (or the
transport through the vadose zone, the effective diffusion reverse) often gives estimated values one to two orders of
coefficient is the gaseous diffusion coefficient corrected magnitude from measured values.
for soil porosity. For many vapors, the gaseous diffusion Using a modification of Fick’s first law, the transfer
coefficient is approximately 0.1 cm2 /s. The effect of soil rate of a contaminant from soil vapor into groundwater
porosity varies depending upon the type of soil, and several can be estimated, and in turn, reasonable estimates of
equations are available to calculate the effect of porosity the expected contaminant concentration in groundwater
on diffusivity. A general approximation is that the porosity due to contamination in the overlying soil vapor can
reduces the gaseous diffusivity by a factor of 10. Thus, for be estimated.
vapors, De can be approximated as 0.01 cm2 /s. The flux is calculated as
The distance that contaminant vapors can move
through the vadose zone in a year can be estimated as Flux = KL (Csg /H − Cw ) (4)

Distance = (2 × 0.01 cm2 /s × 31,536,000 s)1/2 where KL is the gas exchange coefficient (length/time),
∼ 800 cm Csg is the soil gas concentration time
(mass/volume),
This calculation shows that contaminant vapors can move H is the dimensionless Henry’s law constant, and
long distances through the vadoze zone in a short period of Cw is the groundwater concentration
time. Within a few years, vapor contamination can move (mass/volume).
552 APPLICATIONS OF SOIL VAPOR DATA TO GROUNDWATER INVESTIGATIONS

The ratio Csg /H represents the groundwater concentration Importance of Downward Vapor Transport for Several
in equilibrium with the overlying soil vapor at the interface Common Contaminants
between the soil vapor and groundwater.
The potential importance of downward vapor migration for
The gas exchange coefficient has units of velocity
several of the more common compounds encountered by
and represents the distance that contaminants move
environmental consultants are given special consideration
through the groundwater per unit time. This coefficient in this section.
is highly dependent on a compound’s effective diffusivity
and advective mixing of the groundwater interface.
Leaking Vapors from Chlorinated Solvents (Dry Clean-
For groundwater velocities less than 30.5 meters per
ers and Vapor Degreasers). Many compounds, primarily
year and soil grain sizes less than 0.5 mm, diffusional
chlorinated solvents, have relatively high vapor pressures
transport dominates over dispersion, and KL can be
and vapor densities three to six times greater than air.
approximated as:
Because of these physiochemical properties, vapors may
emanate from containers or pipes holding gaseous or liq-
uid products. If air flow is restricted, such as in a closed
KL = 1.1 (De v/d)1/2 (5)
room containing a washer unit, degreaser, or tank, leak-
ing dense vapors can penetrate the floor and enter the
where De is the effective diffusivity, upper vadose zone. Density-driven advection can continue
d is the horizontal distance (downgradient) over to drive vapor downward through the vadose zone until
which the soil vapor plume and groundwater the vapor is diluted to low enough concentrations (<1%)
that density-driven advection is no longer an important
are in contact, and
factor in vapor transport and molecular diffusion takes
v is the horizontal groundwater flow velocity.
over. Vapor clouds reaching tens of feet into the upper-
most vadose zone have been documented in thousands
For the groundwater conditions described above, the
of cases. Businesses and commercial operations that are
interface remains laminar (i.e., no turbulent mixing), and
most susceptible to this situation are those that deal with
the gas transfer coefficient is approximately 8 cm/yr. Using
chlorinated solvents (dry cleaners, vapor degreasers, spray
this value for the gas exchange rate, estimates of the
facilities, etc.).
expected contaminant concentration in groundwater due
The soil vapor concentration to yield an equilibrium
to contamination in the overlying soil vapor for various soil
water concentration of 100 µg/L is 35 µg/L vapor for
vapor concentrations are summarized in Table 1. These TCA, 60 µg/L vapor for PCE, and 100 µg/L vapor for
calculations assume equilibrium partitioning at the soil vinyl chloride. In areas of groundwater deeper than 9.1
vapor/groundwater interface, transfer into groundwater meters, soil vapor concentrations for these compounds are
by molecular diffusion only, and uniform mixing of generally less than these values at groundwater, unless in
the contaminant into the groundwater over the well the presence of contaminated soil, suggesting that surface
screen length. vapor leaks are not a significant contaminant pathway
Table 1 shows that for equilibrium concentrations to groundwater in most cases. In areas of groundwater
(Csg /H) up to 100 µg/L, the resulting ground water shallower than 6 meters, soil vapor concentrations
concentration after 5 years will be low (<10 µg/L) if liquid exceeding these values are found, even in the presence
molecular diffusion is the only exchange process. of clean soil, suggesting that leaking dense vapors from
the surface may be a significant contaminant pathway.

Leaking, Hydrocarbon Fuel Vapors. In contrast to

chlorinated solvents, aboveground vapor releases are
Table 1. Expected Contaminant Concentration in
Groundwater for Various Soil Vapor Concentrations at not common for petroleum hydrocarbons because they
the Groundwater Interfacea,b rarely are stored in confined spaces (due to their
flammability). However, subsurface vapor leaks from
Water conc Water conc petroleum hydrocarbons are possible from USTs and
Csg , Csg /H, Flux, at 1 yr, at 5 yrs,
piping associated with them (vent pipes, pipe joints,
µg/L µg/L µg/cm2 -yr µg/L µg/L
and tank bungs. Gasoline vapor contains approximately
10 40 315 0.63 3.2 8000 µg/L vapor of benzene and lower, although still high
20 80 631 1.3 6.3 concentrations (>1000 µg/L vapor) of toluene and xylene.
30 120 946 1.9 9.5 Reformulated gasolines have very high concentrations of
40 160 1261 2.5 13 MtBE or ethanol, and their vapors contain approximately
50 200 1577 3.2 16 130,000 µg/L vapor of MtBE and 11,500 µg/L vapor of
100 400 3154 6.3 32
ethanol. Due to these high concentrations, leaking vapors
500 2000 15,768 31 158
1000 4000 31,536 63 315
must be considered a potential source of groundwater
contamination, as further described below.
a
Reference 5.
b
Assumptions used to compute the tabulated values: Benzene and Light Aromatics (BTEX). For light aromat-
Henry’s law constant: 0.25 (dimensionless)
Gas transfer coefficient: 8 cm/yr ics, the soil vapor concentration to yield an equilibrium
Well screen interval: 5 m water concentration of 100 µg/L is approximately 25 µg/L
 APPLICATIONS OF SOIL VAPOR DATA TO GROUNDWATER INVESTIGATIONS 553

vapor. In areas of groundwater deeper than 9.1 meters, soil sections, soil vapor values calculated from groundwater
vapor concentrations of aromatics are generally less than using Henry’s constants are likely to be overestimated
25 µg/L vapor at the groundwater, unless in the presence by amounts as large as 10 to 100 times; hence the
of contaminated soil from a leaking UST, suggesting that resulting risk will also be overestimated. The potential
vapor leaks are not a contaminant pathway in most cases. error in the calculated risk introduced by the incorrect
In areas of groundwater shallower than 3 meters, soil vapor concentration can be orders of magnitude greater
vapor concentrations exceeding 25 µg/L vapor have been than errors introduced using default values for other
found even in the presence of clean soil, suggesting that parameters, such as porosity. Thus, if a site fails the
leaking gasoline vapors from the UST are a possible upward risk calculation using calculated soil vapor values
contaminant pathway. In areas where the groundwater from groundwater data, direct measurement of actual soil
is near or above the UST, vapor leaks could be very vapor concentrations will yield more representative soil
significant in groundwater contamination. vapor values to be used in the calculation (13).
When groundwater is extremely shallow (1.5 meters)
MtBE and Ethanol. Due to extremely low Henry and representative soil vapor data are unobtainable, the
constants, the soil vapor concentration to yield an use of surface flux chambers may yield the best estimate
equilibrium water concentration of 100 µg/L is only 2 µg/L of the vapor flux (24).
vapor for MtBE and approximately 0.02 µg/L vapor for A more detailed discussion of the upward vapor migra-
ethanol. Hence, leaking gasoline vapors from a UST tion risk pathway, including a recommended protocol for
are a likely contaminant pathway to groundwater for determining the upward vapor flux, can be found in Hart-
both compounds in shallow and deep groundwater. The man (8).
likelihood is even more acute in locations where the water
table is near or above the UST and/or in areas that have
significant amounts of rainwater infiltration. BIBLIOGRAPHY

Summary. In areas of deeper groundwater (>10 m), 1. EPA. (1997). Expedited Site Assessment Tools For Under-
slow groundwater flow rates less than 30.5 meters per ground Storage Tank Sites. U.S. Government Printing Office,
year and low recharge, soil vapor concentrations for most Washington, DC.
contaminants at the groundwater surface are typically 2. Marrin, D.L. and Kerfoot, H.B. (1988). Soil gas surveying
well below equilibrium values (Csg /H) of 100 µg/L. techniques. Environ. Sci. Technol. 22 (7): 740–745.
Hence contamination of groundwater by downward vapor 3. Ullom, W.L. (1995). Soil gas sampling. In: Handbook of
transport is not likely to be significant. For compounds Vadose Zone Characterization and Monitoring. L.G. Wilson,
with very low Henry constants (MtBE, ethanol), downward L.G. Everett, and S.J. Cullen (Eds.). Lewis, Boca Raton, FL,
pp. 555–567.
vapor transport could be important, even under these
groundwater conditions. 4. Devitt, D.A., Evans, R.B., Jury, W.A., Starks, T.H., Eklund,
B., and Gholson, A. (1987). Soil gas sensing for detection
In areas where groundwater is shallower than 10
and mapping of volatiles organics, EPA/600/8-87/036. (NTIS
meters or higher groundwater flow rates and high
PB87-228516).
recharge, soil vapor concentrations for many contaminants
5. Hartman, B. and Jacobs, J. (2000). Applications and inter-
at the groundwater surface may approach high enough val-
pretation of soil vapor data to volatile organic compound
ues that contamination of the groundwater by downward contamination. In: Standard Handbook of Environmental
vapor transport may be significant. This potential will be Science, Health, and Technology. J. Lehr and J. Lehr (Eds.).
enhanced for compounds with low Henry constants and McGraw-Hill, New York.
when vapor leaks occur directly into the groundwater. 6. Hartman, B. (1999). Applications and interpretation of
Additional discussion of the downward migration of soil vapor data. In: Chlorinated Solvent Contamination:
vapors can be found in Hartman (7,14). Legal and Technical Issues. Morrison et al. (Eds.). Argent
Communications Group, Foresthill, CA.
The Upward Migration of Vapors from Groundwater 7. Hartman, B. (1998). Applications and interpretation of soil
Devolatilization vapor data. In: Petroleum Hydrocarbon Contamination:
Legal and Technical Issues. Morrison et al. (Eds.). Argent
Attention has been raised in the past several years to Communications Group, Foresthill, CA.
the risk to human health by the upward migration of
8. Hartman, B. (1997). The upward migration of vapors.
contaminant vapors into buildings and other enclosed LUSTLine Bull. 27, NEIWPCC, Wilmington, MA, December
spaces (20–22). In cases of contaminated groundwater, it 1997.
is common to assess this risk pathway before site closure
9. Hartman, B. (2003). How to collect reliable soil gas data
is obtained. for risk based applications. Part 2: Flux chamber method.
Simplified, the upward contaminant vapor flux into a LustLine # 44, NEIWPIC, August 2003.
building is computed by Fick’s first law, requiring the 10. Hartman, B. (2002). How to collect reliable soil gas data
soil vapor concentration at some depth underlying the for risk based applications. Part 1: Active soil gas method.
structure. Soil vapor data may be measured directly, LustLine # 42, NEIWPIC, October 2002.
or alternatively, in the absence of actual soil vapor 11. Hartman, B. and Jacobs, J. (2000). Soil vapor principles. In:
data, soil vapor concentrations can be calculated from Standard Handbook of Environmental Science, Health, and
groundwater data assuming equilibrium conditions using Technology. J. Lehr and J. Lehr (Eds.). McGraw-Hill, New
Henry’s constants (23). As discussed in the preceding York.
554 GROUNDWATER VELOCITIES

12. ASTM. (1993). Standard guide for soil gas monitoring in the contaminant migration and arrival time, mass flux, site
vadose zone, D5314-93. Annual Book of ASTM Standards. characterization, reactive barrier design, and calibration
Philadelphia, PA. of transport processes.
13. McNeel, P.J. and Dibley, V. (1997). Case study comparisons Hydraulic head drives flow through porous media. Two
of vapor inhalation risk estimates: ASTM RBCA model main components contributing to the hydraulic head at
predictions vs site specific soil vapor data. Proc. 1997
a given point in a groundwater flow system are pressure
Superfund XVIII Conf., Washington, DC, December 1997,
pp. 509–516. head and elevation head, as shown in Fig. 1. The pressure
head is due to the porewater pressure at that point, which
14. Hartman, B. (1998). The downward migration of vapors.
LUSTLine Bull. 28, NEIWPCC, December 1998. can be measured by a piezometer. The elevation head is
15. Massman, J. and Farrier, D.F. (1992). Effects of atmospheric
due to gravitational energy. The higher the elevation of
pressures on gas transport in the vadose zone. Water Resourc. a fluid from sea level or some selected datum, the more
Res. 28 (3): 777–791. gravitation potential energy it has. The third component,
16. Little, J.C., Daisey, J.M., and Nazaroff, W.W. (1992). Trans- kinetic energy, can be ignored, as groundwater flow
port of subsurface contaminants into buildings. Environ. Sci. velocities are usually very small in magnitude. Thus, the
Technol. 26 (11): 2058–2066. hydraulic head (h) is the sum of the pressure head () and
17. Cherry, J. and Smyth, D. (1996). Non-aqueous phase liquids the elevation head (z), expressed usually in the units of
in the subsurface environment: Assessment & remediation. length (meters).
Am. Soc. Civ. Eng., 1996 Annu. Convention & Exposition, Henry Darcy, a French engineer who studied the
Washington, DC, November 10–14, 1996. movement of water through sand in 1856, found that
18. Thorstenson, D. and Pollock, D. (1989). Gas transport in the rate of water flow through a pipe filled with sand
unsaturated porous media: The adequacy of Fick’s law. Rev. is proportional to the difference in the height of the
Geophys. 27: 61–78. water between the two ends of the tube, and inversely
19. Thomson, K.A. (1995). Soil gas sampling. In: Handbook of proportional to the length of the tube (Fig. 2). The constant
Vadose Zone Characterization and Monitoring. L.G. Wilson, of proportionality, K, is called hydraulic conductivity. This
L.G. Everett, and S.J. Cullen (Eds.). Lewis, Boca Raton, FL,
law, universally known as Darcy’s law, is given as
pp. 569–588.
20. Jury, W.A., Russo, D., Streile, G., and El Abd, H. (1990).
Evaluation of volatilization by organic chemicals residing Q = −KA[(hA − hB )/L]
below the surface. Water Resourc. Res. 26: 13–20.
21. Johnson, P.C. and Ettinger, R.A. (1991). Heuristic model for or
predicting the intrusion rate of contaminant vapors into Q = −KA(dh/dL) (1)
buildings. Environ. Sci. Technol. 25: 1445–1452.
22. Jeng, C.-Y., Kremesec, V.J., and Primack, H.S. (1996). Mod-
els of hydrocarbon vapor diffusion through soil and transport
into buildings. In: Proc. 1996 Pet. Hydrocarbons & Org. Chem. Groundwater
Ground Water: Prevention, Detection, and Remediation, Hous-
Ground surface
ton, TX, November 1996. Depth to water
23. ASTM. (1995). Standard guide for risk-based corrective action
(RBCA) applied at petroleum release sites, E-1739-95. Annual
Pressure head (y)
Book of ASTM Standards. Philadelphia, PA.
24. Smith, J.A., Tisdale, A.K., and Cho, H.J. (1996). Quantifica-
tion of natural vapor fluxes of trichoroethen in the unsatu- Total head
rated zone at Picatinny Arsenal, New Jersey. Environ. Sci. Measuring point h=y+z
Technol. 30: 2243–2250. Elevation head (z)

GROUNDWATER VELOCITIES Selected datum

Figure 1. Elevation head, pressure head, and total head
MUKAND SINGH BABEL (adapted from Reference 1).
Asian Institute of Technology
Pathumthani, Thailand

ha
Groundwater flows through interconnected pore spaces in ha − hb
subsurface soil in response to differences in fluid pressure hb
and elevation. Groundwater flow rates are extremely slow, A B
compared to surface water, and they are also highly
variable. For example, in clayey soils, groundwater moves Pipe filled with sand
as slowly as a fraction of a millimeter per day; however,
in sandy soils, it moves faster, even up to more than
a meter per day. Hydrology is concerned with both L
the speed and the direction of groundwater movement.
Groundwater velocity is very important in dealing with Figure 2. Darcy’s law.
 VISCOUS FLOW 555

or is obtained by dividing the macroscopic velocity by the

q = Q/A = −K (dh/dL) (2) effective porosity:

where Q is the quantity of flow per unit of time (m3 /day); v = −(K/ne )dh/dL (3)
K = hydraulic conductivity (m/day); A = cross-sectional
area (m2 ); hA and hB are the hydraulic head at points A or
and B, respectively; and dh/dL = hydraulic gradient. The v = Q/ne /A
term Q/A in Eq. 2 is also known as the specific discharge
(q), which is the flow per unit area [LT−1 ]. The specific or
discharge is sometimes referred to as Darcy flux, Darcy v = q/ne
velocity, or bulk velocity.
From Eq. 3, it can be seen that, for sand with an effective
The negative sign in Eq. 2 indicates that the fluid moves
porosity of 33%, the actual velocity is three times the
in the direction of decreasing hydraulic head. Darcy’s law
Darcy velocity. Therefore, to define the actual velocity, one
assumes that the flow is laminar, that is, the water follows
must consider the microstructure of the material. Usually,
distinct flow lines rather than mixing with other flow lines,
actual velocities tend to decrease with depth as porosities
which is the case for most groundwater flow in porous
and hydraulic conductivities also decrease.
media. Most groundwater movements occur in small
interstices, so that there is considerable resistance to flow
imposed by the aquifer material, making the flow laminar. BIBLIOGRAPHY
When the flow velocity increases, especially in media
1. Ward, R.C. and Robinson, M. (2000). Principles of Hydrology,
having large pores, the occurrence of turbulent eddies 4th Edn. McGraw-Hill, London.
dissipates kinetic energy, rendering the hydraulic gradient
2. Freeze, R.A. and Cherry, J.A. (1979). Groundwater. Prentice-
less effective in inducing flow. In very large interstices, Hall, Englewood Cliffs, NJ.
such as those in many limestone and volcanic areas,
groundwater flow is almost identical to the turbulent flow
of surface water. Darcy’s law is not valid for turbulent flow, VISCOUS FLOW
for unusually high velocity that may be found in fractures
or in granular aquifers. NAVRAJ S. HANSPAL
Equation 2 yields only an apparent velocity value, the VAHID NASSEHI
macroscopic velocity, through the cross-sectional area of a
Loughborough University
solid matrix and interstices, which, however, is not the true Loughborough, United
velocity of water moving through the pores. Clearly, flow Kingdom
velocities through interstices alone will be much higher
than the macroscopic value. The interstices themselves DIGANTA BHUSAN DAS
vary in shape, width, and direction, so the actual velocity Oxford University
in the soil or rock is highly variable. Furthermore, the flow Oxford, United Kingdom
path of a water particle around and between the grains
in an aquifer is tortuous, so water particles must travel
longer than the apparent distance given by the measured INTRODUCTION
length of the porous medium in the direction of flow, as
shown schematically in Fig. 3. This text describes the elementary concepts of viscous
To calculate the true linear velocity, one must take fluid flow. The principal laws that govern the motion of
into account the fact that water travels only through the fluids are also discussed. From our common experience, we
pores of the sediment. The fraction of the area covered by know that fluids deform under the action of external forces.
open space is equal to the effective porosity (ne ). Thus, the The deforming forces and rate of deformation are related
effective velocity (v), also known as the porewater velocity, to each other through a proportionality constant known
as the coefficient of viscosity. The value of this constant
depends upon the chemical identity, material properties,
Actual flow and thermodynamic state of a given fluid. To develop
paths in a theory that shows how viscosity affects fluid dynamics,
pore spaces one must understand the relationships between stress and
the rate of strain. The outlines given in this article provide
basic concepts of viscous flow. For more detailed analysis,
the reader should consult specialized texts on this topic.
Some of these texts are listed in the references at the end
of this article.
Average
linear flow
path VISCOSITY AND NEWTON’S LAW OF VISCOSITY

The physical property by which the resistance to fluid

Figure 3. Concept of average linear velocity (2). flow can be characterized is known as viscosity. A more
556 VISCOUS FLOW

fundamental approach to viscosity shows that it is the water, and gasoline. In SI units, the units of viscosity are
property of a fluid that relates applied stress to a kg/m · s or Pa · s. In the cgs system, the unit of viscosity
resulting strain rate. By experience, we come across many is g/cm · s, also called the poise (after Poiseuille) and
different types of fluids. For example, consider water and equals 0.1 Pa · s. In fluid mechanics, the ratio of absolute
lubricating oil such as that used in automobiles. Water viscosity to density is also termed kinematic viscosity
can flow more freely compared to lubricating oil under which is often denoted by η. Generally, the viscosity of
isothermal conditions. Hence, we say that the viscosity of liquids decreases as temperature increases. Gases, on the
water is less than that of oil. other hand, have viscosities which generally increase with
When we deform a solid, so that it is strained, we know increasing temperature. More detailed explanations are
that the solid will exert a restoring force that opposes the based on the kinetic theory of gases and can be found
strain; for small strains, the restoring force is proportional in (1).
to the strain, the familiar Hooke’s law. Real fluids also
oppose strains; however, in a fluid it is not the amount of
NON-NEWTONIAN FLUIDS
strain that is important but the rate at which the strain
is produced. There are fluids that do not obey the simple relationship
Consider simple fluid flow between the two plates between shear stress and shear strain rate given by
shown in Fig. 1. If the force per unit area of the top Eq. 2 for a Newtonian fluid. These fluids are known as
plate is F, we find that non-Newtonian fluids. Examples are paints, solutions of
various polymers, food products such as applesauce and
F v
=µ (1) ketchup, emulsions and suspensions of various solids, and
A d coal slurries. The subject of non-Newtonian flow is actually
where d is the separation between the plates, which defines a subdivision of the science of rheology, which is also known
the viscosity µ (sometimes called the absolute or dynamic as the ‘‘science of deformation and flow.’’
viscosity). Equation 1 can be written in a more generalized We categorize non-Newtonian fluids and discuss them
form given by Eq. 2. The shear stress exerted in the x under three general classifications:
direction on a fluid surface of constant y by the fluid in the
region of lesser y is designated as τyx , and the x component 1. The first is the time independent non-Newtonian
of the fluid velocity vector is designated as vx : fluid in which the shear rate is a unique but
nonlinear function of shear stress.
dvx 2. In the second category are the time dependent non-
τyx = −µ = µγ̇ (2) Newtonian fluids that have more complicated shear
dy
stress and shear strain relationships. In these fluids,
This equation is commonly known as Newton’s law of the shearing rate is not a single valued function of
viscosity. It states that the shear force per unit area is the shear stress. The shear rate depends on shearing
proportional to the negative of the local velocity gradient. time or on the previous shear stress rate history of
The left-hand side of Eq. 2 represents the shear stress, that fluid.
and the right-hand side represents the rate of strain γ̇ · τyx 3. In the third category are viscoelastic fluids. Unlike
from Eq. 2 can also be interpreted as the viscous flux a truly viscous fluid in which all its energy of
of the x momentum in the y direction. It can be seen deformation is dissipated, some of the energy of
that the viscous momentum flux is in the direction of deformation of a viscoelastic fluid may be recoverable
the negative velocity gradient, that is, the momentum as in the deformation of an elastic fluid.
tends to go in the direction of decreasing velocity. A
velocity gradient can thus be considered a driving force for Time-Independent Fluids
momentum transport.
For time-independent non-Newtonian fluids,
Fluids in which shear stress is directly proportional to
the rate of deformation are termed Newtonian fluids. The
γ̇ = f (τ ) (3)
viscosity of such fluids is independent of the rate of shear
and rate of strain and always remains constant under
A Newtonian fluid is a special case of this category where
given conditions. Examples of Newtonian fluids are air,
the function f (τ ) is linear.
Time-independent non-Newtonian fluids can be com-
monly represented by three distinct rheograms shown in
v Fig. 2:

1. Bingham plastics
y d
2. pseudoplastics
3. dilatant fluids

x Bingham Plastics. Bingham plastics exhibit yield stress

Figure 1. Viscous drag between two plates, which the bottom at zero shear rate, followed by a straight-line relationship
plate is held fixed. between shear stress and shear rate. The characteristics
 VISCOUS FLOW 557

A B The value of n is greater than unity for pseudoplastic

D
Bingham plastic Newtonian fluid fluids.

C
Dilatant Fluids. Dilatant fluids (curve C, Fig. 2) are
Slope mp
similar to pseudoplastics in having no yield stress. They
t0 differ from pseudoplastics in a fashion such that the
Pseudoplastic apparent viscosity increases as shear rate increases. The
Slope ma power law model given by Eq. 7 is also applicable to
Shear stress

dilatant fluids, and so are Eqs. 6 and 7. The only difference

here is that the power law index n is greater than unity.
Pseudoplastic and dilatant fluids are also known as shear
thinning and shear thickening fluids.
Dilatant fluid
Time-Dependent fluids. Some fluids are more complex
Slope m than those just described, and the apparent viscosity
depends not only on the shear rate but also on the time
for which the shear has been applied. Such fluids can be
classified in two categories;
Shear strain rate
Figure 2. Shear stress–strain rate relationships for non-New- 1. thixotropic fluids
tonian fluids.
2. rheopectic fluids

of such fluids are defined by two constants. The first is the The shear stress decreases with time for a thixotropic
yield stress denoted by τ0 ; this is the stress that must be fluid, whereas the shear stress increases with time for a
exceeded for the flow to begin. The second is the plastic rheopectic fluid. Rheopexy and thixotropy are complex
viscosity µp which is the slope of the straight-line portion phenomena resulting from transient changes of the
on curve A, Fig. 2. The model for Bingham plastics can be molecular structure of time-dependent fluids under an
represented by Eq. 4: applied shear stress. The proposed models for such fluids
are hence based on many simplifying assumptions (6).
τyx = µp γ̇ + τ0 if τyx > τ0 (4)

This model closely approximates the behavior of many Thixotropic Fluids. The apparent viscosity of thixotropic
real fluids such as slurries, plastics, emulsions such as fluids depends on the length of time of shearing as well
paints, and suspensions of finely divided solids in a liquid. as on the shear rate. As the fluid is sheared from the
However, different constitutive models representing this state of rest, it breaks down (on a molecular scale), but
type of fluids have been proposed by Herschel and then the structural reformation increases with time. An
Bulkley (2), Oldyroyd (3), and Casson (4). equilibrium state is achieved when the breakdown rate is
equal to the build up rate. If allowed to rest, the fluid builds
Pseudoplastic fluids. Pseudoplastics (curve B, Fig. 2) up slowly and eventually regains its original consistency.
do not have any yield stress. A pseudoplastic fluid is As the shear is decreased, the apparent viscosity is
characterized by a progressively decreasing slope of shear less than that for increasing shear. For more detailed
stress versus strain rate. The slope has been defined as discussions, refer to (7).
apparent viscosity,
µa = τyx /γ̇ (5) Rheopectic Fluids. Molecular structure in rheopectic
fluids, is formed by shear, and the behavior is opposite
At very high shear rates, the apparent viscosity becomes to that of thixotropy. A simple example of the formation
constant, and the shear stress versus shear strain becomes of shear structure is the beating and thickening of
linear. A number of empirical relations have been used to egg whites. Many substances lose their rheopectic
describe the behavior of pseudoplastic fluids. The simplest properties at extremely high shear rates and behave as
of these is the power law model proposed by Ostwald in thixotropic fluids.
1923 [refer to (5)] and expressed in the form of Equation 6.
Viscoelastic Fluids. A viscoelastic fluid exhibits both
τyx = kγ̇
viscous and elastic properties. The simplest type of such
where n < 1 and k and n are constants for a particular a material is one which is Newtonian and obeys Hooke’s
fluid. k is known as the consistency coefficient for a given law for the elastic part. The shear rate for such fluids can
fluid, and n is the power law index. n also represents the hence be written as
deviation of a particular fluid from a Newtonian fluid. For
a Newtonian fluid, n = 1, and k = µ, the viscosity of the γ̇ = τyx /µ0 + τ̇yx /λ (8)
fluid. Combining Eqs. 5 and 6,
where λ is the rigidity modulus. Under steady-state flow,
µa = kγ̇ (n−1) (7) γ̇ = τyx /µ0 and the fluid behaves like a simple Newtonian
558 VISCOUS FLOW

fluid. However, if the shear is changed, an elastic effect is Refer to (1) to get more acquainted with shell momentum
added. Maxwell first proposed Eq. 8 in the form balance concepts, where this approach has been applied to
simple flow problems.
τyx + τ̇yx (µ0 /λ) = µ0 γ̇ (9)

Liquids that obey Eq. 9 are known as Maxwell liquids. The EULERIAN AND LAGRANGIAN COORDINATE SYSTEMS
constant (µ0 /λ)−1 is known as the relaxation time, which
is the time constant for the exponential decay of stress at Conservation equations and the constitutive equation
constant strain. If the motion is stopped, the stress relaxes are expressed in terms of relative kinematics, velocities,
at e−tλ/µ0 . Rather complex models of viscoelastic materials gradients of velocities, strains, and rates of strain, so it is
have been developed in which higher time derivatives of very important to quantify these variables.
τyx and γ̇ appear. A very readable text for these models Flow kinematics, the relative motion of fluid particles,
is (8). can be described by using either a Lagrangian or a
Eulerian description. In the Lagrangian system, the
motion of an individual particle is tracked where the
MICROSCOPIC BALANCES OF MASS, MOMENTUM,
coordinates locate the position of a particle and are
AND ENERGY
functions of time. In the Eulerian system, dependent
We define the control volume as a region in space variables such as the velocity vector and pressure are
within which our fluid flow system is present. Laws considered functions of fixed spatial coordinates and
of conservation (or change) of mass, momentum, and of time.
energy desribe the behavior of fluid flow within this The Lagrangian framework is considered a more
system. To yield the detailed descriptions of variables natural choice to represent the actual kinematics and
of interest, the control volume must be of infinitesimal stresses experienced by fluid particles. However, to solve
dimensions that can shrink to zero, yielding a point complex flow problems, this approach is limited (e.g.,
volume. This approach reduces the quantities to point due to the necessity of tracking particles along unknown
variables. The application of conservation principles to this streamlines). This system is particularly convenient in
infinitesimal system produces microscopic or differential flows of viscoelastic fluids with memory that require
conservation equations. particle tracking and calculation of deformation and
stresses along streamlines. The Eulerian formulation
Shell Momentum Balances is, in general, more convenient to use because it deals
only with local or present kinematics. In most cases,
Shell momentum balances are written to represent a fluid all variables of interest, such as strain, rate of strain,
flow problem mathematically. The momentum balance for stress, and streamlines, can be obtained from the velocity
any fluid flow system can be written as field. An additional advantage of the Eulerian system is
that it involves only time as a variable in unsteady-state
(Rate of momentum in) − (rate of momentum out)
flow, whereas the Lagrangian system uses traveling time
+ (sum of forces acting on the system) as a variable, even in steady-state flow. All equations
described in subsequent sections have been represented in
= (rate of momentum accumulation) (10)
the Eulerian coordinate system.
Momentum can enter the system by momentum transfer
according to Newtonian (or non-Newtonian) expressions EQUATIONS OF CHANGE
for momentum flux. Momentum can also enter from overall
fluid motion, bulk transfer. The forces acting on a system in The shell balance approach was used to acquaint
fluid flow problems are pressure forces (acting on surfaces) the beginner with the application of the principle of
and gravity forces (acting on the entire volume as a conservation of momentum to viscous flow problems. It is,
whole). The general procedure for solving any viscous however, not always necessary to formulate a momentum
flow problem is balance. It is quicker and easier to start with the equations
of continuity and motion in general form and then to
1. Write a momentum balance in form of Equation 10 simplify them to fit the problem in hand. These two
for a shell of finite thickness, a microscopic shell equations describe all problems of the viscous flow of a
momentum balance. pure isothermal fluid.
2. Let this thickness approach zero, and use the
mathematical definition of the first derivative
to obtain the corresponding differential equation EQUATION OF CONTINUITY
describing the momentum flux distribution.
3. Insert the appropriate Newtonian or non-Newtonian The equation of continuity is developed by applying the
expression for momentum flux to obtain a differen- law of conservation of mass to a small 3-D volume element
tial equation for the velocity distribution. within a flowing fluid, as shown in Fig. 3.
4. Integrate these differential equations which yields
(Rate of mass accumulation)
the momentum flux and velocity distributions for the
system under consideration. = (rate of mass in) − (Rate of mass out) (11)
 VISCOUS FLOW 559

We consider each of the six faces of this volume element analogously. Momentum flows in and out of the control
and evaluate the mass entering and exiting the system: volume by two mechanisms; convection, which is a result
of the bulk fluid flow and molecular transfer (i.e., viscous
Rate of mass in = yzρvx + xzρvy transfer), that is, by virtue of velocity gradients.
x y
The important forces acting on the cubic element in
+ xyρvz (12) Fig. 4 are the fluid pressure p and the gravitational force
z gx . Hence, to write Eq. 10 in mathematical form,
Rate of mass out = yzρvx + xzρvy
x+x y+y Rate of momentum in = Rate of momentum in by
+ xyρvz (13) (convection + viscous transfer)
z+z

∂ρ = yzρvx vx + xzρvy vx
Rate of mass accumulation = xyz (14) x y
∂t
+ xyρvz vx + yzτxx
After inserting, Eqs. 12–14 in Eq. 11, dividing the entire z x
equation by xyz, and taking the limit of dimensions
+ xzτyx + xyτzx (17)
approaching zero, we get the equation of continuity. y z
For generality, we have expressed the equation of
Rate of momentum out = Rate of momentum out by
continuity in vector form:
(convection + viscous transfer)
∂ρ
= −(∇ · ρv) (15) = yzρvx vx + xzρvy vx
∂t x+x y+y

Here (∇ · ρv) is called the ‘‘divergence’’ of ρv. Vector ρv + xyρvz vx + yzτxx

is the mass flux and is the net rate of mass efflux per z+z x+x
unit volume. Equation 15 states that the rate of increase
+ xzτyx + xyτzx (18)
of density within a small volume element fixed in space is y+y z+z
equal to the net rate of mass influx to the element divided Sum of forces acting = body forces + pressure
by its volume. Note that the derivation can be performed  
for a volume element of any shape. = yz p − p
x x+x
A very special form of the equation of continuity is
Eq. 16, which is applicable to fluids of constant density, + ρgx xyz (19)
incompressible fluids.
Rate of accumulation of x
(∇ · v) = 0 ∂ρvx
(16) component of momentum = xyz (20)
∂t

EQUATION OF MOTION These stresses developing in viscous transfer can be

divided into normal and shear stresses. For example, τxx is
Using the shell momentum balance approach for the 3-D the normal stress, whereas τyx is the shear stress resulting
volume element shown in Fig. 5, an equation of motion from viscous forces. At each point in a continuous medium,
is developed. It should be emphasized that momentum is whether it is solid or fluid, we need six components; each
a vector quantity, and hence the equation of motion has represents a component of force per unit area to define
components in each of the three coordinate directions x, y,
and z, respectively. For simplicity, we consider only the x
component here; y and z components can then be handled
(tzx)z + ∆z

z (tyx)y + ∆y
z

∆z (txx)x + ∆x
(rnx)x (txx)x
∆z (rnx)x + ∆x

y y
∆y
∆y
∆x ( x,y,z ) ∆x
(x,y,z)
(tzx)z
(tyx)y
x x
Figure 3. Region of volume in xyz through which fluid Figure 4. Direction of transfer of x component of momentum
is flowing. through the surfaces.
560 VISCOUS FLOW


tyy ∂vz ∂vx
τzx = τxz = −µ + (27)
tnx ∂x ∂z

txy Equations 22–27 have been presented with proof in (9)

and can be applied to various complex situations.
Equations 22–24 are normal stresses and should contain
txx an additional term. For example, Eq. 22 should be
txx
written as


∂vx 2
txy τxx = −2µ + µ − κ (∇ · v) (28)
∂x 3
tnx
Similarly, Eqs. 23 and 24 are also written as


∂vx 2
tyy τxx = −2µ + µ − κ (∇ · v) (29)
∂x 3
Figure 5. Local stresses acting on a cubic element in an

∂vx 2
x–y plane. τxx = −2µ + µ − κ (∇ · v) (30)
∂x 3

the local stress completely. The local stresses acting in the κ is known as ‘‘bulk viscosity.’’ The bulk viscosity is zero for
x–y plane are shown in Fig. 5. low-density monoatomic gases and not too important for
Using Eqs. 17–20, dividing the entire equation by dense gases and liquids which are almost incompressible.
xyz, and taking the limit of the dimensions Experimental measurements to estimate the value of
approaching zero, we get the x component of the equation ‘‘bulk viscosity’’ have been discussed by Karim and
of motion. Analogously, we can find the components for Rosenhead (10). Formulas to estimate the value of bulk
the y and z components. Adding the three components density can be found in (11). When the fluid flows in the
of equation of motion, the equation of motion can be x direction between two parallel plates perpendicular to
expressed in vector notation as the y direction and vx is only a function of y, then this
set of six equations reduces to Eq. 2 described previously.
∂ Equations 22–27 inserted in Eq. 21 yield the equations
ρv = [∇ · ρvv] −[∇p]
∂t of motion for a Newtonian fluid with varying density
(21) and viscosity.
Rate of increase Rate of momentum Pressure on
of momentum gain by convection element per 1. For constant ρ and µ, the equation of motion can be
per unit volume per unit volume unit volume written as


−[∇ · τ ] +ρg ∂v
ρ + v · ∇v = −∇p + µ∇ 2 v + ρg (28)
∂t
Rate of momentum Gravitational
gain by viscous force on Equation 28 is the well-known Navier–Stokes equations,
transfer per unit element per first developed by Navier in France in 1822.
volume unit volume 2. When viscous effects are ignored, Eq. 28 reduces to


τ is known as the ‘‘stress tensor’’, [∇ · ρvv] represents the ∂v
ρ + v · ∇v = −∇p + ρg (29)
rate of loss of momentum per unit volume by fluid flow, and ∂t
(∇ · ρv) represents the rate of loss of mass per unit volume.
To use Eq. 21 to determine the velocity distribution, we Equation 29 model laminar flow when viscous effects
must insert expressions for various stresses in terms of are negligible. These flows are termed Euler, potential,
velocity gradients and fluid properties. For Newtonian or inviscid flows. Equation 29 is the well-known Euler
fluids, these expressions can be written as equation.
3. When the motion of a fluid is ‘‘very slow,’’ the flow
∂vx 2 is said to be creeping or Stokes flow. Viscous forces
τxx = −2µ + µ(∇ · v) (22)
∂x 3 dominate creeping flows and hence, the nonlinear
∂vy 2 inertia term in Eq. 28 is neglected. Then,
τyy = −2µ + µ(∇ · v) (23)
∂y 3
∂v
∂vz 2 ρ = −∇p + µ∇ 2 v + ρg (30)
τzz = −2µ + µ(∇ · v) (24) ∂t
∂z 3

 Equation 30 is the well-known Stokes equations.
∂vx ∂vy
τxy = τyx = −µ + (25)
∂y ∂x The equations of motion together with the equation of


∂vy ∂vz continuity, the equation of state p = p(ρ), viscosity mod-
τyz = τzy = −µ + (26)
∂z ∂y els illustrated in previous sections, and the intial and
 VULNERABILITY MAPPING OF GROUNDWATER RESOURCES 561

the boundary conditions, complete the problem specifica- 5. Bird, R.B., Armstrong, R.C., and Hassager, O. (1977). Dynam-
tion for any viscous flow problem. These equations, when ics of Polymeric Liquids: Fluid Mechanics. John Wiley & Sons,
solved mathematically, determine completely the density, New York.
pressure, and velocity components in a flowing isother- 6. Slibar, A. and Paslay, P.R. (1959). Retarded flow of Bingham
mal fluid. materials. Trans. ASME 26: 107–113.
7. Tanner, R.I. (1985). Engineering Rheology. Clarendon Press,
Boundary Conditions Oxford, UK.
8. Wilikinson, W.L. (1960). Non-Newtonian Fluids. Pergamon
Boundary conditions are constraints imposed on conserva- Press, New York.
tion equations to describe how flow behavior conforms to
9. Lamb, H. (1945). Hydrodynamics. Dover, New York.
the surroundings. Boundary conditions come from nature
10. Karim, S.M. and Rosenhead, L. (1952). Rev. Mod. Phys. 24:
and are mathematical descriptions of the physics at the
108–116.
boundary. These may describe conditions along the bound-
11. Papanastasiou, T.C., Georgiou, G.C., and Alexandrou, A.N.
ary dealing with motion, external stresses, rate of mass
(2000). Viscous Fluid Flow. CRC Press, New York.
and momentum flux, boundary values of field variables
12. Hirschfelder, J.O., Curtiss, C.F., and Bird, R.B. (1954). The
as well as the relations among them. When the solution
Molecular Theory of Gases and Liquids. John Wiley & Sons,
involves the time evolution of flow fields, initial condi- New York.
tions are also needed. The number of boundary conditions
required to solve a particular problem is determined by the
Nomenclature
nature of the governing partial differential equation. In
general, elliptical equations require boundary conditions A Area
on each portion of the boundary; hyperbolic equations F Force
require boundary conditions at upstream but not at down- g Gravity force vector
stream boundaries; parabolic equations require initial n Power law index
conditions and boundary conditions everywhere except p Pressure
downstream. The Navier–Stokes equations are hyperbolic v Velocity vector
at high Reynolds numbers and elliptical at low Reynolds t Time variable
numbers. The Euler equation is the upper limit of hyper-
Greek Symbols
bolicity, whereas Stokes equations are the lower limit of
γ̇ Rate of strain
ellipticity. Different types of boundary conditions in math-
κ Bulk viscosity
ematical form have been described by Papanastasiou (12)
µ Viscosity of a fluid
and are applicable to a wide range of viscous flow problems.
τ Stress
In general, there are three types of boundary conditions:
τ̇ Rate of shear
λ Elasticity modulus
1. First kind or Dirichlet boundary conditions: The
ρ Density of a fluid
value of the dependent variable along the bound-
ary is imposed as the condition. Typical Dirich- Subscripts
let boundary conditions are no-slip boundary a Stands for apparent
conditions for velocity and the specification of p Stands for plastic
inlet/outlet velocities. o Stands for yield stress
2. Second kind or Neumann boundary conditions: x Stands for x component
The normal derivative of the dependent variable y Stands for y component
is specified along the boundary. Examples are z Stands for z component
symmetry conditions and free surface and interface
stress conditions.
3. Third kind or Robin’s boundary condition: The VULNERABILITY MAPPING OF
dependent variable and its normal derivative GROUNDWATER RESOURCES
are related. Slip wall conditions are typical
Robin conditions. THOMAS HARTER
University of California
Davis, California
BIBLIOGRAPHY

1. Bird, R.B., Stewart, W.E., and Lightfoot, E.N. (1994). Trans-

port Phenomena. John Wiley & Sons, New York.
WHAT IS GROUNDWATER VULNERABILITY?
2. Herschel, W.H., and Bulkley, R. (1926). Flow of Non-
Newtonian fluids. In: Encyclopaedia of Fluid Mechanics. Gulf Groundwater vulnerability is a measure of the likelihood
Publishers, Houston, TX, Vol. 9, Chap. 4. that pollution or contamination at the land surface reaches
3. Oldyroyd, J.G. (1947). A rational formulation of the equations a production aquifer. It is a measure of the ‘‘degree of
of plastic flow for a Bingham solid. Proc. Camb. Philos. Soc. insulation’’ that natural and man-made factors provide
43: 100–105. to keep pollution away from groundwater. The term was
4. Casson, N. (1959). Rheology of Disperse Systems. Pergamon first used in Europe in the 1960s. In defining groundwater
Press, London. vulnerability, a number of issues must be examined and
562 VULNERABILITY MAPPING OF GROUNDWATER RESOURCES

clarified; all refer to these two questions: vulnerability and good scientific judgment. It is generally not intended
of what? and vulnerability to what? Exact definitions of to be a scientifically rigorous analysis of groundwater
groundwater vulnerability have therefore varied (1). resources. The National Research Council (3) identified
In principle, all groundwater is vulnerable to human four general objectives typically achieved by groundwater
activity; no groundwater is completely isolated from the vulnerability assessments: (1) to facilitate policy analysis
aboveground environment. The degree of vulnerability and development at the local and regional level, (2) to
depends on environmental conditions, on how we define provide program management, (3) to inform land use
groundwater and the part of groundwater in which we are decisions, and (4) to provide general education and
interested in (hence the question, ‘‘vulnerability of what?’’), awareness of a region’s hydrogeologic resources.
and on the timescale of interest. In addition, it depends The usefulness of vulnerability maps is limited (some
on whether or not a vulnerability measure is intended to would say, arguable). The scientific literature lists pros
account also for the presence and type of pollutants (hence, and cons to vulnerability assessments (1). The facts
‘‘vulnerability to what?’’). that speak for implementing assessments are that land
use decisions have to be made, land management
Vulnerability of What? practices need to be sensitive to the risk for groundwater
contamination, and not all anthropogenic activities can
Groundwater is produced from a wide variety of hydroge-
be carried out isolated from groundwater. Timescales
ologic environments. When defining vulnerability, there-
and distinctions based on travel time are important,
fore, we need to define the portion of groundwater to
at least with respect to some pollutants. Timescales
be assessed. Because the shallowest groundwater zone is
are also important in placing groundwater monitoring
typically the most vulnerable, vulnerability assessments
wells and in scheduling and providing for planning and
are mostly concerned with the vulnerability of the upper-
action–response time if pollution does occur. So a need
most aquifer (in a multiaquifer system) or with the water
exists to provide at least some general guidance to land
table (in an unconfined aquifer system). Deep groundwa-
use planners, decision makers, and water users. Carefully
ter is considered less vulnerable than shallow groundwater
prepared vulnerability assessments allow them to make
because of the longer travel times necessary for a pollutant
decisions that are economically sensible and at the same
to reach these deeper aquifers.
time hydrogeologically reasonable.
Some methods of assessing vulnerability therefore
On the other hand, some may argue against the use
account, directly or indirectly, for the travel time between
of vulnerability maps for land use planning because
the water table and a well by considering aquifer
groundwater flow conditions and the transport properties
permeability or travel-time zones (e.g., 2). Timescales
of the subsurface are too complex to be appropriately
are important in defining vulnerability. The practical
captured by any vulnerability tool. A second group of
question is, should we consider groundwater vulnerable
arguments against vulnerability zoning arises from the
to today’s activities only if the recharge time is a few
fact that all groundwater is vulnerable to land pollution.
years or also if recharge timescales are tens, hundreds,
The only geographically differing factor is the timescale
thousands, or tens of thousands of years? The potential
for the pollution to reach groundwater.
travel time between contaminant source and groundwater
or groundwater wells is an important consideration for
protection, monitoring, and assessment. Timescales are WHAT FACTORS DETERMINE VULNERABILITY?
also important when mapping vulnerability to specific
pollutants that become less harmful over time and with We distinguish between intrinsic factors (natural con-
travel distance due to radioactive decay, degradation, ditions) and extrinsic factors (anthropogenic activities).
inactivation (in the case of pathogens), filtration, or strong Among intrinsic factors, the thickness and hydraulic prop-
sorption to soils and aquifer materials. erties of the geologic formations above the aquifer—the
unsaturated zone and confining layers above the
Vulnerability to What? aquifer—are key factors in determining the vulnerabil-
ity of an aquifer system. They are the principal natural
An assessment of groundwater vulnerability may or may controls that determine the recharge rate and recharge
not include consideration of the presence of pollutants. time to the aquifer. The unsaturated zone also provides
Vulnerability that is independent of whether or not con- key groundwater protection by
taminants (pollutants) are present and focuses primarily
on a description of natural environmental conditions is • intercepting, sorbing, and eliminating pathogenic
often (though not always) referred to as ‘‘susceptibility,’’ viruses and bacteria;
‘‘natural vulnerability,’’ ‘‘aquifer sensitivity,’’ or ‘‘intrinsic • sorbing and degrading many synthetic organic
vulnerability.’’ Pollution-type-dependent vulnerability, or chemicals; and
vulnerability to specific land uses, is sometimes referred to
• attenuating heavy metals and other inorganic
as ‘‘specific vulnerability’’ or ‘‘integrated vulnerability’’ (1).
chemicals through sorption and complexation with
mineral surfaces within the unsaturated zone and
WHY ASSESS GROUNDWATER VULNERABILITY? through uptake into plants and crops (e.g., fertilizer).

The purpose of vulnerability assessments is to provide The potential of the soil and the unsaturated zone to sorb,
a decision-making tool based on the best available data degrade, or eliminate substances depends on the type of
 VULNERABILITY MAPPING OF GROUNDWATER RESOURCES 563

pollutant and therefore is considered only in vulnerability States. Put together by a group of experts and the EPA
studies that are specific to certain land uses. in the mid-1980s, it has been applied to a number of
The amount of recharge at the land surface is groundwater basins, regions, even states.
another important factor that determines vulnerability. The name stands for depth to groundwater, recharge
The higher the recharge, the higher the vulnerability rate, aquifer media, soil media, topography, impact of
of the aquifer. If climatic conditions are such that little vadose zone media, and hydraulic conductivity of the
or no recharge occurs at the land surface, downward aquifer. In DRASTIC, scores of 0 to 10 are assigned
movement of moisture through the unsaturated zone is to each parameter, 0 means low risk of groundwater
very limited, regardless of the hydraulic properties of the contamination; 10 means high risk of contamination. The
unsaturated zone. scores are multiplied by a parameter-specific weight. The
In vulnerability assessments that encompass more than weighted scores of all seven parameters are then added
just aquifer susceptibility (see above), the presence of for the final DRASTIC score. Two sets of weights can be
contamination sources or possible contaminating activities used: one for general (intrinsic) vulnerability analysis, one
is considered an important factor. The amount, spatial for specific vulnerability to pesticides.
distribution, and chemical properties of the contaminants
(possibly) released into the subsurface are highly specific GOD (1). GOD is a vulnerability assessment method
to individual sources and regions. developed in Great Britain, where most groundwater
resources are in hardrock aquifers, primarily sandstone
CHARACTERIZATION OF GROUNDWATER and limestone aquifers. Unconsolidated overburden or
VULNERABILITY soil layers cover the fractured hardrock in many places.
Like DRASTIC, GOD is an index-and-overlay method
Numerous schemes have been developed for assessing and designed to map groundwater vulnerability across large
mapping vulnerability. These methods can be grouped into regions based on a few important parameters; GOD
three major categories: stands for groundwater occurrence, overall lithology of
the unsaturated zone or overlying aquitard, and depth
• index-and-overlay methods to the groundwater table. Scores are assigned to each of
• process-based computer simulations the three categories and then multiplied to yield a final
score. In developing GOD, the method’s authors have given
• statistical analyses
particular consideration to the likelihood of fractures or
fracture systems to develop in the soils, overburden, or
Index-and-Overlay Methods
overlying geologic units of the aquifer.
Index-and-overlay methods are based on assembling infor-
mation on the most relevant factors affecting aquifer Process-Based Computer Simulations
vulnerability (soil type, geologic formation type, recharge,
etc.), which then is interpreted by scoring, integrating, or Process-based computer simulations afford a great amount
classifying the information to produce an index, rank, or of realistic complexity and detail to be built into the
class of ‘‘vulnerability.’’ The scoring, ranking, and inte- vulnerability assessment. Computer models can account
gration methods are based on expert opinion rather than for complex physical and chemical processes on a very
processes and are inherently subjective to some degree. detailed scale. Unlike the two-dimensional maps and
In the United States, the most prominent vulnerability map layers used with other methods, computer modeling
assessment method in this category is ‘‘DRASTIC’’ (2). allows for complete three-dimensional resolution. Geologic
Others include the vulnerability analysis specified by Cal- and hydrogeologic variations with depth can therefore
ifornia’s Drinking Water Source Assessment Program (4) be reproduced to evaluate their effect on vulnerability.
and the British ‘‘GOD’’ method (1). A brief overview of Process-based computer models focus on re-creating the
DRASTIC and GOD is given below. flow and transport patterns within the unsaturated zone
The advantage of index-and-overlay methods is that or in an actual aquifer and can be used to compute
they provide relatively simple algorithms or decision trees travel times or concentrations of a contaminant in the
to integrate a large amount of spatial information into unsaturated zone or in the aquifer. Computer models do
maps of simple vulnerability classes or indexes. The not compute vulnerability directly. Rather, vulnerability
methods are designed to rely on data that are readily is defined as a function of what the computer models
available from local, state, or federal agencies, such as simulate. For example, high vulnerability may be defined
information on soils, water level depth, precipitation, as any region in the aquifer for which the computer model
and geology. These methods are particularly suitable for shows a travel time of less than 5 years.
use with computerized geographic information systems The most advanced computer models also allow the
(GIS), which is a digital form of mapmaking, because they analyst to compute the uncertainty that is inevitably
usually involve overlaying and aggregating multiple maps associated with the computer model predictions due to
showing soil properties, depth to water table, recharge, shortcomings in the database fed into the computer
and so on. and due to our limited knowledge of the ‘‘underground
world’’. These methods use statistical methods to quantify
DRASTIC Vulnerability Mapping (2). DRASTIC is the uncertainty. However, even simple computer models can
most popular vulnerability mapping tool in the United include a sensitivity analysis, that is, an investigation
564 VULNERABILITY MAPPING OF GROUNDWATER RESOURCES

of the sensitivity of the model predictions to the Statistical Methods

input provided.
Computer models are not commonly used for vulnera- Statistical methods are used to quantify the risk of ground-
bility assessment due to their considerable data require- water pollution by determining the statistical dependence
ments and the expertise required to implement models. or relationship between observed contamination, observed
In other words, computer simulation models are rarely an environmental conditions that may or may not charac-
economical alternative for vulnerability mapping. How- terize vulnerability (e.g., unsaturated zone properties,
ever, computer modeling is an excellent and economical recharge), and observed land uses that are potential
tool for vulnerability mapping if sources of contamination (e.g., fertilizer applications, sep-
tic tank occurrence). Once a model of this dependence
• a more localized analysis of specific vulnerability or relationship has been developed from the statistical
to particular land uses (particular contaminants) is analysis, it can be used to predict—in a similar area
required and sufficient data are available or can be elsewhere—the chance or risk of contamination. Such an
collected to prepare the computer model, and application requires, of course, knowledge of significant
• a number of ‘‘what-if’’ scenarios involving complex environmental conditions for that area. When statistical
processes need to be evaluated for making important methods are used, the risk of contamination is essentially
land use planning decisions. a quantitative measure of ‘‘vulnerability.’’ The higher the
contamination risk, the higher the vulnerability.
Four computer model codes are particularly popular In principle, statistical methods are not much dif-
for computing the fate and transport of contaminants ferent from index-and-overlay methods. Both establish
as those contaminants travel downward through the a relationship between inherent natural conditions and
unsaturated zone to the water table: VLEACH (EPA), groundwater vulnerability (which the statisticians refer
PRZM (EPA), LEACH (Cornell University), and HYDRUS to as groundwater contamination risk). In overlay meth-
(U.S. Department of Agriculture/Agricultural Research ods, the relationship is established by a team of experts. In
Service Salinity Laboratory, Riverside, California) (for up- statistical methods, the relationship is established by sta-
to-date information on these codes, see the web pages of tistical analysis. The advantage of the statistical method is
these agencies). All four codes (listed here roughly in order that the statistical significance can be explicitly calculated,
of complexity and user expertise required; VLEACH is which provides a measure of uncertainty or certainty of the
least complex and HYDRUS the most complex) compute model. The disadvantage is that statistical methods are
the downward movement of water (flow) and contaminants difficult to develop and, once established, can be applied
(transport). These codes are designed specifically to only to regions that have environmental conditions simi-
determine the amounts of sorption and degradation that lar to those of the region for which the statistical model
occur in soils and in the unsaturated zone and are typically was developed.
used to compute the residual, nondegraded mass of a Few statistical methods have been developed for
contaminant that arrives, over time, at the water table. For vulnerability assessment, primarily because they rely on
vulnerability assessments, subregions or smaller sections large data sets from regions already contaminated. The
of land with similar unsaturated zone properties are key to the development of a statistical method is a large,
selected and grouped. One simulation is performed for high quality data set of a specific contaminant or several
each group or section, and the results are mapped in specific contaminants commonly found in a region. That
various ways; a map of the travel time to the water table data set is then correlated in one fashion or another to
(perhaps specific to a contaminant), a map of the percent specific properties of the subsurface (depth to water, soil
removal of a contaminant within the vadose zone, and permeability, slope, hydraulic conductivity, etc.) to create
so on.
a statistical predictive model that can be used to make a
The most common computer code used for groundwater
quantitative statement about the contamination risk, such
modeling (in all three dimensions) is MODFLOW (U.S.
as, ‘‘if this kind of unsaturated zone, aquifer, and land use
Geological Survey). This model and other 3-D computer
properties exist in a location, then the risk for groundwater
models can be used for highly specific site studies of
contamination is X%, with a confidence interval of plus or
groundwater contamination that are far beyond the
minus Y%.’’ Some examples follow.
purpose of a typical vulnerability assessment. The data
requirements for running these models are tremendous
and require careful data preparation and data processing, CALVUL. In California, a statistical method has been
particularly in a fully three-dimensional simulation. developed by the California Department of Pesticide Reg-
Sometimes, groundwater models are used to compute ulations (DPR) to determine the specific vulnerability of
explicitly the complex source area of the water reaching groundwater to pesticide residues. The method is nick-
a well. In that case, the same computer model can named CALVUL (California Vulnerability approach; (5)).
also be used to compute specific vulnerability for DPR uses the vulnerability analysis to determine, for each
particular contaminants of concern within the source section of agricultural land in California, whether ground-
area of the well. If a well-calibrated and well-documented water in that section of land is vulnerable or not. If it is
groundwater model exists, it may be preferrable to use vulnerable, the land section (square mile) becomes part
the groundwater model to address specific vulnerability of a groundwater protection area with restricted pesticide
issues rather than using a basic index or a scoring-based use. The DPR method distinguishes two types of vulnera-
vulnerability assessment. bility: (1) vulnerability to pesticide leaching in coarse soils
 VULNERABILITY MAPPING OF GROUNDWATER RESOURCES 565

with a shallow water table and (2) vulnerability to pesti-

cide runoff over hardpan soils (and subsequent leaching
in dry wells). Pesticide management practices depend on
whether pesticide leaching or pesticide runoff is of concern.
A pesticide database for the San Joaquin Valley was
used to determine the statistical relationship between pes-
ticide residue in groundwater and a large number of soil
(unsaturated zone) properties, including water holding
capacity, texture, organic matter content, permeability,
shrink-swell potential, slope, and infiltration. By deter-
mining which soil and geographic properties are good
predictors of pesticide residue in groundwater, the statis-
Probability that nitrate
tical model can be used to identify sections of land that exceeds 4 mg/L
have similarly ‘‘vulnerable’’ soil and geographic conditions 0–.17
but no past or current pesticide residue detections (5). >.17–.33
>.33–.50
Presumably, the most vulnerable areas are identified and >.50–.67
can be targeted for implementation of mitigation manage- >.67–.83
>.83–1
ment practices.
Figure 1. Groundwater vulnerability map for the United States,
where vulnerability is here defined as the probability that nitrate
Texas Case Study. In Texas, a statistical analysis
concentration in groundwater exceeds 4 mg/l. (reprint from (7),
similar to DPR’s San Joaquin study, albeit based on
also published at http://water.usgs.gov/nawqa/nutrients/pubs/est
a different set of statistical tools, was developed by v36 no10/fig3.html).
Evans and Maidment (6). Instead of pesticide residues
in groundwater, this method uses an analysis of nitrate
in groundwater as the basis for delineating vulnerable
The regression model was validated (tested) on a large set
groundwater areas. As in CALVUL, the assumption
of well records not used in developing the regression model
made by the developers of the method was that
and then applied to predict the probability of elevated
where high contamination exists, the aquifer is more
nitrate (a measure of nitrate vulnerability) in groundwa-
vulnerable than elsewhere. Nitrate was used as the target
ter throughout the United States at a resolution of 1 km2
variable because a large amount of data was available
(Fig. 1, from Reference 7). The study demonstrates the
throughout Texas and because nitrate is neither sorbed
advantage of the statistical method over index-and-overlay
nor significantly degraded in groundwater (except under
methods. The statistical assessment model uses objective,
anaerobic conditions). The analysis was done by using data
quantifiable measures to assess vulnerability. Its predic-
from the entire state. From a 30-year statewide database,
tive capability has been quantified as well (through the
the probability (risk) that a well would have nitrate levels
validation test).
in excess of some threshold value (1, 2, 5, or 10 mg/L NO3 -
N) was computed separately for each analysis unit. The
exceedance probability is used as a quantifiable measure of WHAT ARE THE LIMITATIONS OF VULNERABILITY
vulnerability; the higher the chances that nitrate is found ASSESSMENTS?
in wells within an analysis unit (area), the higher the
vulnerability. The study illustrates that overall results of Because of the implied imprecise nature of vulnerabil-
the statistical analysis follow the general patterns laid out ity assessments and the inevitable subjectiveness of the
by the index-and-overlay methods, albeit at considerably underlying interpretive scheme, the National Research
higher expense. Council (3) issued ‘‘three rules of groundwater vulnera-
bility.’’ These rules, or limitations, should be spelled out
National Groundwater Nitrate Vulnerability Assess- explicitly with every vulnerability assessment:
ment. The U.S. Geological Survey recently published a
national assessment of the probability of nitrate contami- • All groundwater is vulnerable to some degree.
nation in recently recharged groundwater (7). The assess-
ment relied on a specialized statistical regression analysis • Uncertainty is inherent in all vulnerability assess-
among major factors affecting the specific vulnerability ments.
of groundwater to nitrate and the measured occurrence • There is a risk that the obvious may be obscured and
of elevated nitrate levels in wells. The data set used to the subtle may become indistinguishable.
develop the regression model included wells from 20 major
groundwater basins across the United States; each con- The last refers to the danger, especially when using
tained 20–30 test wells. Nitrogen fertilizer loading and complex vulnerability assessment tools, that in light of
percent cropland were significant extrinsic (nitrogen use the final vulnerability index or ranking, one may lose sight
related) factors. The presence of well-drained soils, depth of the data used for the analysis and of the assumptions
to a seasonally high water table, and the presence of uncon- underlying vulnerability assessment schemes. The results
solidated sand and gravel aquifers were the key intrinsic of vulnerability assessments may mask real differences
factors that determine groundwater nitrate vulnerability. between specific sites or aquifer conditions.
566 WATER/ROCKS INTERACTION

AFTER THE ASSESSMENT: VERIFICATION AND POSTAUDIT Snyder, D.T., Wilkinson, J.M., and Orzol, L.L. (1998). Use of a
Ground-Water Flow Model with Particle Tracking to Evaluate
Verification refers to some independent procedure that can Ground-Water Vulnerability, Clark County, Washington,
Water-Supply Paper 2488. U.S. Geological Survey, Denver, CO.
verify the results of the vulnerability analysis. Postaudit
is essentially the same as verification but typically occurs Vrba, J. and Zoporozec, A. (Eds.). (1994). Guidebook on Map-
ping Groundwater Vulnerability. International Association of
years later, when additional data have been sampled
Hydrogeologists: International Contributions to Hydrogeology
that can be compared with the predictions made by the Vol. 16, Verlag Heinz Heise, Hannover, Germany.
vulnerability analysis.
Verification and postaudits of vulnerability assess-
ments can be done in many different ways. The most WATER/ROCKS INTERACTION
common approach, particularly for verifying assessments
done with index-and-overlay methods, is to compare the HENRY TENG
vulnerability map with the actual occurrence of some com- The George Washington
mon pollutant in groundwater. Typical pollutants used are University
nitrate and pesticides. However, such verification works Washington, DC
well only where the appropriate pollution source is actually
present and has been present for some time.
Water–rock interaction is a process whereby solid earth
materials and aqueous phases in the Earth system
BIBLIOGRAPHY
exchange matter and, ultimately, energy through chemical
and biochemical reactions. The extent of reactions is
1. Robins, N.S. (Ed.). (1998). Groundwater Pollution, Aquifer
determined by how far the system is away from
Recharge and Vulnerability. Geological Society Special Pub-
lication No. 130, Geological Society, London. equilibrium, and the rate of reactions is controlled largely
by the kinetics of mineral-water interfacial processes.
2. Aller, L., Bennett, T., Lehr, J.H., and Petty, R.J. (1985).
DRASTIC—A standardized System for Evaluating Ground In a broad sense, these reactions can be classified into
Water Pollution Potential Using Hydrogeologic Settings. U.S. two categories based on the reactants involved: (1) those
EPA, Robert S. Kerr Environmental Research laboratory, between H2 O and the mineralogical components of rocks
Office of Research and Development, EPA/600/2-85-018. and (2) those between solutes dissolved in natural waters
3. National Research Council. (1993). Ground Water Vulnerabil- and the minerals. Water–rock interactions participate
ity Assessment, Contamination Potential under Conditions of in many geological processes such as ore deposit
Uncertainty. National Academy Press, Washington, DC. formation, weathering, and pedogenesis. The outcome of
4. California Department of Health Services. (2000). Drinking these interactions plays an important role in regulating
Water Source Assessment and Protection Program. Available: geochemical cycling of elements, the atmospheric CO2
http://www.dhs.ca.gov/ps/ddwem/dwsap/ level, and the chemistry and nutrient availability of
DWSAP document.pdf. surface/near-surface waters at both local and global
5. Troiano, J., Spurlock, F., and Marade, J. (1999). Update of the scales.
California Vulnerability Soil Analysis for Movement of Pesti- Water–rock interactions are heterogeneous reactions
cides to Ground Water. Department of Pesticide Regulation, because of the involvement of multiple phases (that is,
Sacramento, CA 95814-3510, Document EH 00-05. Available: solid and liquid phases in most cases). Despite the com-
http://www.cdpr.ca.gov/docs/empm/pubs/ehapreps/eh0005.pdf.
plexity associated with the heterogeneity and variations in
6. Evans, T.A. and Maidment, D.R. (1995). A Spatial and Statis- mineral assemblages, water chemistry, and environmen-
tical Assessment of the Vulnerability of Texas Groundwater
tal conditions, reactions between aqueous phase and rocks
to Nitrate Contamination. Technical Report CRWR 260, Cen-
at the fundamental level can be regarded as consisting
ter for Research in Water Resources, Bureau of Engineering
Research, The University of Texas at Austin, Austin, TX. of only two types of mechanistic movement (1), chemical
species attachment to and detachment from mineral sur-
7. Nolan, B.T., Hitt, K.J., and Ruddy, B.C. (2002). Probability of
nitrate contamination of recently recharged groundwaters in faces. Individual or combined development of the two basic
the conterminous United States. Environ. Sci. Technol. 36(10): actions over time yields three primary types of water–rock
2138–2145. interactions: adsorption (including ion exchange), dissolu-
tion, and precipitation. Following is a detailed discussion
of the three types of reactions.
READING LIST

Blodgett, R. (1993). Drinking Water Vulnerability Assessments ADSORPTION

in Texas. Proc., 1993 Ann. Conf., American Water Works
Association, Denver, CO. Chemical sorption results directly from the occurrence
Holtschlag, D.J. and Luukkonen, C.L. (1998). Assessment of of undercoordinated sites and associated charges on
Ground-Water Vulnerability to Atrazine Leaching in Kent mineral surfaces. Development of surface charge in
County, Michigan: Review, Comparison of Results of Other aqueous environments can be understood as a consequence
Studies, and Verification, Water-Resources Investigations of surface complexation, by which water molecules
Report 98-4006. U.S. Geological Survey, Lansing, MI. form chemical bonds with the undercoordinated surface
Rupert, M.G. (2001). Calibration of the DRASTIC ground water atoms (2). These surface-bound water molecules are then
vulnerability mapping method. Ground Water 39(4):625–630. subject to a proton transfer process that shifts hydrogen
 WATER/ROCKS INTERACTION 567

ions onto neighboring surface anions. This dissociative Because the total amount of surface sites is [>S]tot =
sorption of water molecules results in one or more [>S] + [>SA], Eq. 1d can be rewritten as
types of surface species, depending on the chemical
composition of the minerals in question. For (hydr)oxide Kad [A]
[>SA] = [>S]tot (1e)
materials, hydroxylated surface cation centers >X(OH)0 , Kad [A] + 1
where > symbolizes lattice bonds and X represents
the cationic component, seem to be the only concern. High concentrations of [A] lead to [>SA] ∼ = [>S]tot ,
For salt-type minerals Xm An , an additional, second-type whereas low values of [A] reduces Eq. 1e to a linear form
nonhydroxylated surface anion site >AH0 has to be taken of [>SA] ∼
= Kd [A]. Another common empirical adsorption
into consideration. model, the Freundlich isotherm, uses a logarithmic
When present in solution environments, the surface relationship between [>SA] and [A] so that
hydroxyl groups can undergo protonation or deprotonation
to form charged surface species. The relative abundance [>SA] = kf [A]n (1f)
of protonated and deprotonated species thus governs the
surface electric properties. Consider an oxide mineral XO. where both kf and n are constants. The Freundlich model
The sign and quantity of surface charges incurred in pure can be thought of as the Langmuir isotherm integrated
water will be controlled by the following reactions over many types of surface sites (4).
In the surface complexation model, adsorption is
treated as a chemical reaction
≡ XOH2 + ⇔≡ X(OH)0 + H+
app
Ka1 (1a)
>X(OH)0 + M (n+1)+ ⇔ >XOM n+ + H+ (1g)
and
≡ X(OH)0 ⇔≡ XO− + H+
app
Ka2 (1b) The equilibrium constant K app

app app
where Ka1 and Ka2 are the apparent equilibrium [>XOM n+ ][H+ ]
K app = = K int K coul (1h)
constants of reactions 1a and 1b. The participation of [>XOH0 ][M (n+1)+ ]
hydrogen ions in the above reactions indicates that the
relative concentrations of the protonated and deprotonated now has contributions from chemical (intrinsic) as well
surface species, and hence the amount of surface charges, as electrostatic (coulombic) interactions between the
are controlled by the pH of the aqueous environment. reactants. Quantitatively, K coul = exp(FZ), where Z
The pH at which there are equal amounts of >XOH2 + is the change of charges in surface species because of
and >XO− (i.e., an uncharged surface) is defined as the adsorption. The surface potential  is a function of surface
pH of zero net proton charge, below which the surface is charge σ , and the charge-potential relationship can be
positively charged for >XOH2 + is more abundant than derived from electrochemistry theories. A common one
>XO− and above which the surface is negatively charged
because of the excess of >XO− . σ = (8RTεε0 I)1/2 sinh(zF/2RT) (1i)
In the presence of other sorbing ions, say M (n+1)+ and
L(n+1)− , in solutions, competitive complexation becomes is given by the Gouy–Chapmen theory (diffuse double-
operative as M can exchange for H and L for OH layer model). Here ε is the relative dielectric constant
at the >X(OH)0 sites to result in the formation of water, ε0 is the permittivity of vacuum, I is the ionic
of surface complexes such as >XOM n+ and >XLn− . strength of the bulk solution, and z is the valence of
Surface complexation under this condition becomes the symmetrical background electrolyte. The quantity of
adsorption where both chemical bonding and electrostatic surface charge can be computed from the concentration of
interaction take place between solute particles and charged surface species. For Eq. 1g,
surface atoms (3). Empirical approaches to adsorption
employ simple equations to describe the adsorption σ = F[≡ XOM (n−1)+ ]/AS (1j)
isotherm (relationship between the solution concentration
or activity and the surface concentration of an adsorbate), where A and S are the specific area (m2 /g) and solid
without concerning the specific interfacial interactions concentration (g/l) of the adsorbent material, respectively.
between adsorbate and adsorbent. For example, the Overall, assuming the total amount of surface sites
Langmuir isotherm, one of the most common empirical in Eq. 1g is approximated by [>XO]tot = [>X(OH)0 ] +
models, simply treats adsorption as a free surface site >S [>XOM n+ ], one can rearrange Eq. 1h into
occupied by an adsorbate A to form a sorbed site >SA:  
exp(nF)K int [M (n+1)+ ]
[>XOM n+ ] = [>XO]tot
>S + A ⇔ >SA (1c) exp(nF)K int [M (n+1)+ ] + [H+ ]
(1k)
It becomes apparent that, after inspecting Eqs. 1e, 1f,
The adsorption constant Kad given by the mass law is and 1h, the surface complexation approach is advanta-
geous over the empirical isotherms in that the formal pro-
[>SA] vides additional insight into the dependence of adsorption
Kad = (1d) upon pH and ionic strength (through Eq. 1i). Moreover, the
[>S][A]
568 WATER/ROCKS INTERACTION

use of K int in surface complexation models makes it possi- a system containing the crystal in contact with an
ble to compare results from different experiments because, undersaturated solution can be described by
unlike empirical parameters such as Kad (Eq. 1d) and Kf  
(Eq. 1f), the intrinsic surface complex formation constant dG = − νi µi dn + γj dAj (3a)
is not experimental condition dependent. i j
It is noteworthy to point out that although surface
complexation reactions can occur on any mineral crystals, where µi is the chemical potential of the ith species in
effective adsorption requires small grain sizes. Because the liquid phase; νi is the stoichiometric coefficient of the
of the inverse relationship between particle size and ith species in the solid phase; n is the number of moles of
specific surface area (defined as the ratio of surface area solute that are involved in the dissolution reaction; and γj
to volume of a particles), small grains (<1 µm) tend to and Aj are the surface free energy and the surface area of
have a significant percentage of surface-bound atoms. the jth type of faces, respectively. Using the definition of
Consequently, small particles have large numbers of active chemical potential,
surface sites available to undergo complexation reactions.
Micron-scale suspended particles and sub-micron-scale µi = µoi + RT ln αi (3b)
colloids of minerals are a common occurrence in surface
waters and groundwaters (Ref. 5). It is estimated that one can rewrite Eq. 3a as
80% of the materials transported by rivers and discharged
dG  
into the oceans is in the form of suspended particles. =− νi µoi − RT ln αiνi + γj (dAj /dn) (3c)
Although the exogenic cycle of the major cations and dn
i i j
anions is controlled by solution process, in most cases
the geochemistry of trace elements in natural waters is where µoi and αi are the standard chemical potential
governed by adsorption on separate mineral grains (6). and the activity of the ith species, ! respectively. At
equilibrium conditions, dG/dn = 0, i νi µoi = −RT
" ln Ksp ,
it then follows that the crystal solubility Kb = i αiνi can
DISSOLUTION be expressed as

Minerals dissolve because the chemical potential of solute 1 

ln Kb = ln Ksp + γj (dAj /dn) (3d)
in solid and liquid phases has not reached equality, RT
j
which indicates that the driving force of any dissolution
reaction may be estimated by the free energy difference, Equation 3d states that the solubility of a crystal is
G, between reactants and products. As the reaction always greater than the standard solubility product of the
approaches equilibrium conditions where G becomes solute. Rewrite Eq. 3d using the volume v and density ρ
zero, dissolution should cease to proceed. of the crystal and the molecular weight W of the solute,
Consider the dissolution of a solid compound ML, one gets

ML ⇔ M n+ + Ln− (2a) W/ρ 

ln Kb = ln Ksp + γj (dAj /dv) (3e)
RT
j
The free energy change in this system is described by
For isotropic crystals with any characteristic dimen-
G = Go + RT ln(aM aL ) (2b) sion l (i.e., length, width, or height), γj (dAj /dv) can
be approximated by (2γ )/(3l) and Eq. 3e can be simpli-
where Go is the standard free energy change, R is the fied to
2W/ρ
gas constant, and T is the temperature; aM and aL are the ln Kb = ln Ksp + (γ /l) (3f)
activity of aqueous species M n+ and Ln− , respectively. 3RT
At equilibrium conditions, G = 0 and Eq. 2b can be Equation 3f is the statement of the classic Gibbs–Thomson
rewritten as relationship, which predicts that the difference between
aM aL = exp(−Go /RT) (2c) crystal solubility and Ksp is size-dependent, and the
solubility of any crystals can only be approximated by
Because Go can be computed as the difference between the standard solubility product of the solute when the
the standard free energy of formation for the dissolved crystal is infinitely large. In a practical sense, the ‘‘infinite’’
species and the solid phase, Eq. 2c yields a numerical size is usually no larger than a few micrometers. For
value defined as the standard equilibrium constant or anisotropic minerals, each crystal face has to be considered
solubility constant, Ksp , of the compound ML. individually because the surface energy is face-specific
When the solid–fluid interface is concerned, surface and Eq. 3e cannot be easily simplified. Nonetheless,
energies and grain sizes have to be taken into account the conclusions drawn from Eq. 3f are just as well
to define the equilibrium conditions. Consider a mineral applicable.
crystal of any geometry with any number of faces that Dissolution not only takes place in the form of
are not related by any symmetry operations. At constant dissociation reactions (Eq. 2a), but it can also result
temperature and pressure, the free energy change in from reduction–oxidation reactions where the number
 WATER/ROCKS INTERACTION 569

of valence electrons changes on one or more reactant This expression depicts a linear relationship between the
elements. Redox reactions particularly affect water–rock driving force and the kinetics of dissolution reactions
interactions involving minerals that contain multivalence at near-equilibrium conditions. When applying Eq. 5b
state elements such as S and Fe. However, because of to water–rock interactions, however, caution has to
the relatively few minerals concerned, the contribution be used because the dissolution of many minerals in
of redox chemistry to the exogenic cycle of elements is aqueous environments is initiated by the formation of
considered auxiliary (7). Most natural redox reactions etch pits on crystal surfaces. Etch pit nucleation usually
are mediated by microbial communities or induced by requires one to overcome an energy barrier that decreases
solar radiation. In the formal scenario, oxidized mineral with descending saturation state (9,10). At very close to
species are often used as electron acceptors for metabolism, equilibrium conditions, the saturation state may be too
whereas in the latter situation, electrons in the reactants’ high to bring down the barrier to the extent that allows pit
molecular orbital are excited photochemically to pair up formation, which indicates that dissolution kinetics may
with more electrons or to be released to lower the energy change discretely, rather than in a linear fashion, with
state of the system. decreasing saturation state (11,12).
Traditionally, geochemists treat redox reactions as A unique characteristic of mineral dissolution is
reactants’ gaining or losing oxygens. The more modern that some (alimino)silicate minerals may dissolve in an
approaches, on the other hand, choose to use electron as incongruent or nonstoichiometric manner. This type of
a basic component. In this convention, reductants are reaction is best exemplified by the dissolution process
electron donors that release e− and oxidants electron of Na- and K-feldspars in acidic environments. It is
acceptors that receive e− . The activity of electron can generally agreed that the dissolution starts with an
be expressed as a potential (Eh, in volts) or as an activity initial fast ion exchange of hydrogen ions for alkali
unit (pe = − log ae− ). Quantitatively, Eh and pe can be and alkaline earth metals, followed by a slow process
converted to each other through that disrupt the Si-O-Al-O framework. The disruption
seems to proceed by preferential release of aluminum (i.e.,
F nonstoichiometric surface leaching) so SiO2 accumulates
pe = Eh (4a) in situ. Observations by numerous workers indeed reveal
2.303RT
the presence of a Si-rich surface coating on feldspar crystal
where F is the Faraday constant. In addition, Gibbs free faces when the minerals are exposed to solutions with pH
energy and the activity of electrons are related by <3–5 (Ref. 13). What researchers are not in complete
agreement on is to what extent the Si-rich layer is
G = −nFEh (4b) originated from surface alteration. Although results from
investigations up to the end of the last century imply
where n is the number of electrons in the reaction. If a that most coating (up to several thousand angstroms
redox dissolution reaction can be written as in depth) forms through nonstoichiometric leaching of
the underlying minerals (Ref. 13), recent studies using
more modern experimental techniques suggest otherwise.
Ox (solid) + ne− ⇔ Rd (aqueous) (4c)
Synchrotron radiation and high-resolution transmission
electron microscopy show that the incongruent dissolution
the standard free energy expression (Eq. 2b) becomes may be operative only for, at most, the first few unit
cells at the mineral-water interface, and the thick
RT
Eh = Eo − ln aRd (4d) surface alteration thus forms largely by interfacial
nF dissolution–reprecipitation (14,15).
or
PRECIPITATION
1 1 1
pe = pe − log aRd = log K o − log aRd
o
(4e)
n n n Much of the discussion on reaction driving forces in
o o dissolution section is directly transferable to precipitation
where E and pe are the activity of electrons under
reactions. However, precipitation is not a pure mirror
standard conditions. Notice that the equilibrium constant
image of dissolution in that the former is a combination
of redox reaction K o can still be calculated from the of two separate but sequentially occurring processes,
difference between the standard free energy of formation nucleation and growth. Nucleation refers to the atomic
for the dissolved species and the solid phase (i.e., Eq. 2c). process in which solute components congregate to
In general, the rate R of dissolution reactions can be form a thermodynamically stable cluster, whereas the
related to G by (8) following accretion of the cluster is termed growth.
Both nucleation and growth require the occurrence of
R = k(1 − eG/RT ) (5a) supersaturation, but each of the two processes has its
unique physiochemical controls.
where k is the rate constant. When the reaction is Homogeneous nucleation occurs when a thermodynam-
sufficiently close to equilibrium (i.e., |G|  RT), Eq. 5a ically stable 3-D cluster is born directly from solution. On
reduces to the other hand, solute particles can assemble onto an exist-
R = −(k/RT)G (5b) ing surface, and hence through heterogeneous nucleation,
570 WATER/ROCKS INTERACTION

to form a stable 2-D cluster. No matter how nucleation frequency (20) that can be approximated by kB T/hp where
takes place, the classic nucleation theory requires the hp is the Planck constants; µL and µS are the chemical
sizes of the clusters to satisfy the following energetics for potentials of solute in liquid (solution) and solid (crystal),
the atomic assemblage to become stable nuclei: A surface respectively, and E is the free energy of an intermediate
energy increase associated with the emergence of the solid activated complex.
phase precisely offsets the chemical potential decrease in Equation 6e can be further rearranged into
the solution because of losing solute. Mathematically, free
energy change upon formation of a spherical cluster of vs = (d2 /λo )υ exp[(µL − E)/kB T]{exp[(µS − µL )/kB T] − 1}
radius r can be expressed as (6f)
Considering that the term exp(µL − E) has an entropy
G = 4π r2 γ − (4/3π r3 /v)µ (6a) factor, exp(S/kB ), and an energy factor, exp(−E/kB T),
and that (µS − µL ) is essentially the driving force for
where γ is the average surface free energy and v is the growth and is quantitatively related to supersaturation,
specific molecular volume of the solid. The minimal size of σ , by
a stable cluster µ = kB T ln(a/ae ) = kB Tσ (6g)
rc = 2vγ /µ (6b)

can be found by solving the equation d(G)/dr = 0. Any one can rewrite Eq. 6f as
cluster with r < rc is unstable and will be dissolved,
whereas the ones with r > rc will move on to the growth vs = {(d2 /λo )υ exp(−S/kB ) exp(−E/kB T)}/
stage. Inserting Eq. 6b into 6a, one has an expression that kB T]}{(a − ae )/ae } (6h)
estimates the energy barrier, G∗ , for nucleation:
Notice that step velocity has a linear dependence on
16π γ 3 v2
G∗ = (6c) the difference between solute activity and its equilibrium
3µ value (a − ae ). By convention, the quantity between the
The nucleation rate Rn , according to the conventional rate first two curly parentheses in Eq. 6h is termed the kinetic
theory, is then available through coefficient, β, of step growth. This parameter, as shown
here, is characterized by the activation energy barriers,
−G∗ E, and the corresponding entropic changes for the
Rn = A exp( ) (6d) formation of activated intermediate complexes.
kB T
Complementary to the classic approach of atom-
where kB is the Boltzmann constant and A is a con- by-atom growth, recent findings suggest that crystals
stant related to the efficiency of particle collision (16). can also grow by oriented aggregation of nano-sized
All implications from Eqs. 6a–6d apply to heterogeneous clusters (21,22). In this growth mode, adjacent nano-
nucleation, except that, quantitatively, G∗ is usually crystallites coagulate after spontaneous self-organization
smaller when nucleation occurs on a substrate surface. that situates the two parties to a common crystallographic
This result leads to a general conclusion that heteroge- orientation. The new, larger particles gain thermodynamic
neous nucleation often occurs at a lower supersaturation stability by forming chemical bonding between the original
than does homogeneous nucleation. clusters to eliminate the interface between them. It
Precipitation enters the growth stage with the onset is entirely possible that different growth mechanisms
of enlargement of the minimal sized clusters. Nuclei operate at different conditions. When nucleation rates
increase their size through attachment of individual are low and nuclei are far apart in the media, growth
solute particles onto the cluster surfaces. During growth, may only be accomplished by atom-by-atom attachment
crystallographic control becomes strongly operative. That on growing surfaces. When sufficient nuclei are present,
is, faces that are more closely packed will prevail, whereas they can coalesce, through orientated coagulation, to form
the ones with smaller packing density will be phased larger crystals, as long as the system is shifting toward a
out. The more closely packed faces usually are the flat lower energy state.
face (F-face) on the ideal Kossel crystals (17); the growth It is important to be informed that the kinetics of
of flat surfaces is commonly recognized as following a interfacial reactions such as dissolution and growth can
layer-by-layer mechanism where solute particles attach be controlled either by the pace of the actual chemical
to kink sites along steps (i.e., the edges of individual reactions or the speed at which the reactants and products
layers) generated either by crystal imperfection or surface are delivered to and removed from the reaction sites,
nucleation. The net rate of step advance, vs , is determined or both. To date, it is generally agreed that surface
by the difference between solute fluxes attaching to and reactions, rather than transport, play a predominant
detaching from the step kinks. This relationship can be role in controlling the rate of mineral-water interactions.
described mathematically by (18,19) This point of view is supported by the nearly ubiquitous
occurrence of etching features on both laboratory dissolved
vs = (d2 /λo )υ{exp[(µS − E)/kB T] − exp[(µL − E)/kB T]} and naturally weathered grains of many minerals. The
(6e) significance of this understanding is directly reflected by
Here, d is the size of the attaching species in the lattice, the intensive interests in mineral surface geochemistry
λo is the interlink distance, and υ is the fundamental that has been growing for more than two decades.
 GROUND WATER: WELLS 571

BIBLIOGRAPHY

1. Hochella, M.F., Jr. and White, A.F. (1990). Rew. Mineral. 23:
1–16.
2. Schindler, P.W. and Stumm, W.W. (1987). The surface
chemistry of oxides, hydroxides, and oxide minerals. In:
Aquatic Surface Chemistry. W. Stumm (Ed.). John Wiley &
Sons, New York, pp. 83–110.
3. Stumm, W. and Morgan, J.J. (1981). Aquatic Chemistry. 2nd
Edn. John Wiley & Sons, New York, pp. 600–601.
4. Stumm, W. (1992). Chemistry of the Solid-Water Interface.
John Wiley & Sons, New York, p. 95.
5. Langmuir, D. (1997). Aqueous Environmental Geochemistry.
Prentice-Hall, Englewood, Cliffs, NJ, p. 344.
6. Morel, F.M.M. and Hering, J.G. (1993). Principles and
Applications of Aquatic Chemistry. John Wiley & Sons, New
York, p. 509.
7. Morel, F.M.M. and Hering, J.G. (1993). Principles and
Applications of Aquatic Chemistry. John Wiley & Sons, New
York, p. 421.
8. Lasaga, A.C. (1998). Kinetic Theory in the Earth Sciences.
Princeton Univ. Press, Princeton, NJ, p. 188.
9. Frank, F.C. (1951). Acta Cryst. 4: 497–501.
10. Cabrera, N. and Levine, M.M. (1956). Phil. Mag. 1: 450–458. the bottom of the hole. Some wells are still dug by hand
11. Nagy, K.L. and Lasaga, A.C. (1992). Geochim. Cosmochim. today, but more modern methods are available. It’s still a
Acta 56: 3093–3111. dirty job, though!
12. Lasaga, A.C. and Luttge, A. (2001). Science 291: 2400–2404. Wells are extremely important to all societies. In many
13. Blum, A.E. and Stillings, L. (1995). Rev. Mineral. 31: places wells provide a reliable and ample supply of water
291–351. for home uses, irrigation, and industries. Where surface
14. Teng, H.H., Chen, L., Fenter, P.P., and Sturchio, N.C. (2001). water is scarce, such as in deserts, people couldn’t survive
Geochimica et Cosmochimica Acta. 65: 3459–3474. and thrive without ground water.
15. Hellmann, R. et al. (2003). Phys. Chem. Minerals 30:
192–197. TYPES OF WELLS
16. Stumm, W. (1992). Chemistry of the Solid-Water Interface.
John Wiley & Sons, New York, p. 216. Digging a well by hand is becoming outdated today (would
17. Kossel, W. (1927). Nach. Ges. Wiss., Gottingen, p. 135. YOU want to do it?). Modern wells are more often drilled
18. Chernov, A.A. (1961). Soviet Phys. 4: 116–148. by a truck-mounted drill rig. Still, there are many ways to
19. Chernov, A.A. and Komatsu, H. (1995). Topics in crystal put in a well—here are some of the common methods.
growth kinetics. In: Science and Technology of Crystal
Growth. J.P. van Eerden and O.S.L. Bruinsma (Eds.). Kluwer Dug Wells
Academic, Amsterdam, pp. 67–80.
20. Lasaga, A.C. (1998). Kinetic Theory in the Earth Sciences.
Hacking at the ground with a pick and shovel is one way
Princeton Univ. Press, Princeton, NJ, p. 178. to dig a well. If the ground is soft and the water table is
21. Penn, R.L. and Banfield, J.F. (1998). Science 281: 967–971.
shallow, then dug wells can work. They are often lined
with stones to prevent them from collapsing. They cannot
22. Penn, R.L. and Banfield, J.F. (1999). Geochimica et Cos-
mochimica Acta. 63: 1549–1557.
be dug much deeper than the water table—just as you

GROUND WATER: WELLS

ROGER M. WALLER
U.S. Geological Survey

There’s a good chance that the average Joe who had to

dig a well in ancient Egypt probably did the work with
his hands, a shovel, and a bucket. He would have kept
digging until he reached the water table and water filled

This article is a US Government work and, as such, is in the

public domain in the United States of America.
572 WELL SCREENS

cannot dig a hole very deep when you are at the beach. . . 1. Adequate open area (typically greater than 3–5%)
it keeps filling up with water! to maximize well efficiency1
2. Strong pipe-based well screen (e.g., horizontal
Driven Wells louver shutter screen) to provide maximum col-
lapse strength.
Driven wells are still common today. They are built by
driving a small-diameter pipe into soft earth, such as sand 3. Constant diameter throughout the entire length of
or gravel. A screen is usually attached to the bottom of the well screen.
the pipe to filter out sand and other particles. Problems? 4. Single material type (e.g., all copper-bearing or all
They can only tap shallow water, and because the source stainless steel) to avoid galvanic corrosion from two
of the water is so close to the surface, contamination from different metals coming in contact with one another.
surface pollutants can occur.
Well screens may be constructed with a variety of
Drilled Wells different materials and designs. The screen material
can be different steels, such as carbon, copper-bearing,
Most modern wells are drilled, which requires a fairly high strength low-alloy, and stainless, or PVC materials.
complicated and expensive drill rig. Drill rigs are often The type of material used is dependent on the purpose
mounted on big trucks. They use rotary drill bits that chew of the well, strength requirements, groundwater quality
away at the rock, percussion bits that smash the rock, or, (corrosion potential), the expected well life, and cost factor.
if the ground is soft, large auger bits. Drilled wells can be There are many different types of well screens; however,
drilled more than 1,000 feet deep. Often a pump is placed the most commonly used are discussed below.
at the bottom to push water up to the surface.
HORIZONTAL LOUVER SCREENS
WATER LEVELS IN WELLS
Horizontal louver screens are a pipe-based well screen
Ground-water users would find life easier if the water level manufactured by punch-formed downward facing horizon-
tal louver-shaped openings (Fig. 1). This type of screen is
in the aquifer that supplied their well always stayed the
particularly suited for use in large diameter, deep, gravel
same. Seasonal variations in rainfall and the occasional
envelope wells. The horizontal louver screen can be manu-
drought affect the ‘‘height’’ of the underground water level.
factured in a variety of patterns, with various percentages
If a well is pumped at a faster rate than the aquifer around
of open areas. Percentage open areas range from approxi-
it is recharged by precipitation or other underground flow,
mately 3.4% to 7.4% for Ful Flo to Super Flo patterns (2).
then water levels around the well can be lowered. The
The smooth, unobstructed interior of the horizontal louver
water level in a well can also be lowered if other wells near
screen allows for efficient use of highly effective devel-
it are withdrawing too much water. When water levels
opment techniques such as swabbing and also facilitates
drop below the levels of the pump intakes, then wells will
employment of efficient maintenance and repair meth-
begin to pump air—they will ‘‘go dry.’’
ods. Disadvantages of horizontal louver screen include

WELL SCREENS
DENNIS E. WILLIAMS
Geoscience Support Services
Claremont, California

A well screen is a fairly generic term that denotes

the component of a well that provides (1) an opening
for water to enter the well casing from the aquifer;
and (2) stabilizes material in the ‘‘near-well zone’’ (i.e.,
materials immediately adjacent to the well) and prevents
it from entering the well.
Proper design of well screens therefore is paramount in
successful water wells. Most modern wells are constructed
using a filter pack placed between the well screen and
the aquifer. In a proper well screen/filter pack design (i.e.,
gravel envelope well), the fundamental principle is: The
Figure 1. Horizontal louver screen.
purpose of the filter pack is to stabilize the aquifer; the
purpose of the well screen is to stabilize the filter pack. 1
[It has been demonstrated (1,2) that if the percentage open area
A properly designed well screen therefore allows sand- is above 3–5%, well efficiencies do not significantly increase. It
free water to flow into the well with minimum drawdown. has also been demonstrated that entrance velocity is not a critical
Features of a properly designed well screen include: design factor and only needs to be below 0.5 m/s (1.5 ft/s).]
 WELL SCREENS 573

unavailability of slot size smaller than 1 mm (0.040 in.), slot screen is usually installed in gravel envelope wells.
and difficulty in perforating though thick walled pipe [i.e., Its chief advantages are reasonably high area of opening
greater than 9.5 mm (0.375 in.)]. However, the superior (approximately 4.8%) and minimum frictional head losses
strength capabilities and long-life characteristics have at relatively low cost. One important disadvantage is its
made horizontal louver screens an industry standard in low collapse strength due to the large number of vertically
deep high capacity water supply wells. oriented slots. The manufacturing process is restricted
with respect to wall thickness, which is limited to a
maximum of 6.35 mm (0.250 in.).
CONTINUOUS WIRE WRAP SCREEN

Wire wrap screen is manufactured by wrapping a shaped MACHINE-SLOTTED (MILLED) SCREENS

wire (steel or plastic) around an internal array of
longitudinal rods (Fig. 2). The wire and rod are joined Horizontal and vertical machine-slotted screens are
by resistance welding, which produces a cage-shaped manufactured from casing in which openings are milled
cylindrical configuration. Continuous slot screens are very with axially oriented precision cutters (Fig. 4). Vertical
effective in naturally developed wells in uniform fine- slotted screens are also manufactured by prepunching flat
grained sands in relatively shallow, thin aquifers since sheets of steel prior to fabrication into cylinders. Machine-
they are able to be manufactured with very small slot slotted screens have disadvantages that include clogging
sizes and yet still maintain the necessary open area due to the parallel surfaces within the opening. As slot
to minimize frictional head loss. Continuous wire wrap
screens typically have open areas greater than 25%.
Additionally, the design allows small particles to pass
completely through the screen without becoming wedged
in and causing clogging of the screen. Disadvantages of
the continuous slot screen are associated with its relatively
low strength, which necessitates special handling during
installation and development, and also restricts the
extent of future well maintenance and repair efforts.
As the interior is obstructed by rods, the use of highly
effective well development techniques such as swabbing
is precluded. Should the screen be damaged, repair by
swedging is impractical.

BRIDGE SLOT SCREENS

Bridge slot screens are manufactured on a press from

flat sheets or plates. The slot opening is usually vertical
and provides two orifices, longitudinally aligned to the
axis. The perforated steel sheets or plates are then rolled Figure 3. Bridge slot screen.
into cylinders and the seam is welded (Fig. 3). A bridge

Figure 2. Continuous wire wrap screen. Figure 4. Horizontal machine-slotted screen.

574 WELL TEST

clogging is directly related to wall thickness, thicker from a pumping test involves a carefully planned test
material encourages greater plugging. Some vertical program, including consideration of continuous operation
machine-slotted casings are machined with an undercut and execution of the necessary steps.
to reduce this tendency. A second drawback is the low Step-drawdown tests are generally the first type of
area of openings (less than 1%), although this can be test performed on a newly completed production well
overcome at higher cost by increasing the number of slots. following final well development. A step-drawdown test
Collapse strength of vertically machine-slotted casings, involves pumping the well at several discharge rates and
however, is substantially reduced when the number of measuring the change in depth to pumping level in the
openings is increased. The main positive characteristics of well as time progresses (Fig. 1). At least three ‘‘discharge
machine-slotted screens are low cost, uniform openings, steps’’ are required for the test with subsequent discharge
and availability in a fairly wide range of apertures rates increasing over the previous steps. Data obtained
and patterns. from step-drawdown tests are used to determine well
production and drawdown characteristics, from which
WELL CASING PERFORATED IN PLACE the permanent production pump can be designed. The
discharge step rates are determined based on results from
Well casing perforated in place has historically been the final development pumping. For example, if during
used with cable tool wells drilled in unconsolidated final development, the maximum discharge rate reached
alluvium. Once the casing has been driven or jacked to 2500 gpm with good groundwater level stabilization, then
its total depth, a mills knife (vertical slots) or hydraulic the three rates for the step-drawdown test might be 750,
louver-type perforator (horizontal slots) is used down the 1500, and 2200 gpm. The discharge rates for a step-
hole to perforate the casing at specific intervals. The drawdown test should be reasonably spread out over the
hydraulic louver-type perforator is superior to the mills maximum range available for the well (as determined
knife in producing slots due to better physical casing during development).
characteristics and more control over the aperture size; Following the step-drawdown test, a constant-rate
in addition, the louver-shaped openings provide greater pumping test is usually performed. The purpose of
stabilization of particles in relative fine-grained sand the constant-rate test is to verify the design discharge
formations. The hydraulic louver-type perforator is limited rate estimated from the step-drawdown test and also
to a slot size of 3.2–4.0 mm (0.125–0.156 in.). Typically, to measure longer-term (usually greater than 24 h)
a well casing perforated in place would have a percentage drawdown effects on the pumping well and any nearby
open area of less than 1%. wells. Data obtained from both the pumping well and
The design of the screen slot or aperture size for gravel nearby wells can be used to calculate aquifer parameters,
envelope wells requires the filter pack to be designed first. which can then be used to design spacing between wells,
The screen slot size must result in 90% retention of filter calculate groundwater storage volumes, or determine the
pack material outside the screen. rate of movement of groundwater. The discharge rate
during the constant-rate test should be kept constant for
the entire duration of the test and should be at the design
BIBLIOGRAPHY
discharge rate for the well (Fig. 2). The design discharge
rate may be based on both development pumping and
1. Williams, D.E. (1985). Modern techniques in well design. J.
AWWA 77: 9. results from the step-drawdown test. The step-drawdown
test is seldom performed for a duration sufficient to
2. Moss, R. (1990). Handbook of Ground Water Development. John
Wiley & Sons, Hoboken, NJ. design the permanent pump and should always be verified
with a longer-term constant-rate test. However, the final
design pumping rate is still subject to ‘‘sound engineering
WELL TEST
DENNIS E. WILLIAMS
Geoscience Support Services Time since pumping began, minutes
Claremont, California 1 10 100 1,000
0
Drawdown, ft

A well test is a test conducted in a completed and developed

well from which data can be collected to calculate certain
hydraulic parameters of the well and surrounding aquifer.
Hydraulic parameters may then be used, for example,
to design the maximum safe pumping rate of the well,
determine well field spacing, calculate aquifer storage,
or determine the rate of movement of contaminants
in the groundwater. Well tests include pumping tests,
recovery tests, and slug tests. Essential measurements
50
and data collected during well tests include time, depth
to water, and discharge rate. Obtaining accurate data Figure 1. Step-drawdown test data.
 SAFE YIELD OF AN AQUIFER 575

Time since pumping began, minutes SAFE YIELD OF AN AQUIFER

1 10 100 1,000
0 ERIC S. WILSON
E. L. Montgomery & Associates, Inc.
Drawdown, ft

Tucson, Arizona

Safe yield is a controversial term and concept that

may be defined for common usage as the amount of
water an aquifer system can yield over a long period
Constant drawdown Recovery without producing unacceptable effects. Unacceptable
effects may include an excessive decline in groundwater
level, sometimes described as groundwater mining;
50 reduction or elimination of riparian vegetation; reduction
or elimination of water in streams, lakes, springs,
Figure 2. Constant-rate test with recovery data.
and wetlands; aquifer compaction and land subsidence;
reduction in groundwater underflow to downgradient
aquifer systems; a change in groundwater temperature;
judgment’’ and should consider both demand requirements an undesirable change in the chemical quality of
and pumping test results. groundwater; and salt water intrusion into aquifer
Recovery tests are carried out after pumping tests. The systems adjacent to oceans or saline lakes.
test involves measuring the rate at which the groundwater The definition of safe yield has evolved for nearly a
level recovers to its initial level after pumping has stopped. century; initially, it was defined by Lee (1) as ‘‘. . .the
In general, water level data from recovery tests are more limit and quantity of water which can be withdrawn
reliable due to the lack of water level fluctuations caused regularly and permanently without dangerous depletion
by discharge variations. of the storage reserve.’’ The definition was modified and
A slug test is a popular and inexpensive method of expanded by Meinzer (2), Conkling (3), and Todd (4). Due
estimating the hydraulic properties of aquifers in small to ambiguity and lack of agreement on the definition, many
diameter wells when the hydraulic conductivity of the authors have suggested abandoning the term (5–7) for an
aquifer materials is too low to conduct a pumping test. alternate term such as ‘‘sustainable yield.’’
Incorrect concepts, regulations, or laws may indicate
Slug testing involves either rapidly adding or removing
that safe yield is equal to recharge. This incorrect concept
a ‘‘slug of water’’ from a well and measuring the rate
has been described by Bredehoeft et al. (8) as the ‘‘water-
at which the water level recovers (either falling-head or
budget myth’’ because the concept oversimplifies the
rising-head). Disadvantages with this method are that
necessary information needed to understand the effects
it can only be used to determine the aquifer’s hydraulic
of groundwater development on an aquifer system.
conductivity in the immediate vicinity of the well, and
It is a great deal more difficult to define safe yield of
results can be influenced by filter pack material adjacent
an aquifer based on analyzing the unacceptable effects
to the well screen.
of overpumping than to assume simply that safe yield is
The type of test to be performed depends on the test
equal to recharge. When water is pumped from an aquifer,
purpose, available resources, and site-specific limitations.
even in amounts equal to recharge, natural groundwater
For example, in an area containing only a single well, it
discharge continues. The initial result of pumping is the
is not possible to calculate aquifer storage parameters
decline of groundwater levels near the pumping well,
from a pumping test without an observation well
and with time, across larger areas of the aquifer. This
(i.e., nonpumping well). Similarly, observation wells not
water level decline increases the hydraulic gradient and
screened in the same aquifer as the pumping well cannot
assists in drawing groundwater from marginal parts of the
provide reliable interference data.
aquifer system into the central part of the aquifer system,
Once groundwater level data from the well tests have but eventually water levels drop sufficiently that natural
been collected, compiled, and graphically plotted, the groundwater discharge becomes smaller. As a result, even
data are evaluated using various methods to determine by pumping groundwater at rates equal to recharge, a
hydraulic parameters of the aquifer and well. In many decline in groundwater discharge may occur. This decline
cases, the calculation of parameters requires only solving in discharge could possibly be defined as unacceptable,
simple equations or a set of equations. For each method and the aquifer may not be in a state of safe yield.
there are basic underlying assumptions from which the When considering the pragmatic aspects of safe yield,
method equation was derived. These assumptions need it is important to address the principal elements of the
to be carefully considered with respect to the test data water budget expressed by the law of conservation of mass
and specific aquifer conditions when a method is used. In and the water budget equation:
addition to the basic assumptions, each of the analysis
methods has other assumptions specific to the method. Inflow = outflow ± change in storage
Many of the analytical solutions are based on the basic
differential equation of groundwater flow, which can be Pumping from an aquifer system over time changes
derived from fundamental laws of physics. the amount of water in storage and subsequently changes
 Next Page
576 SPECIFIC YIELD STORAGE EQUATION

values of inflow and outflow, thus changing values in the BIBLIOGRAPHY

water budget equation. Theis (9) stated that groundwater
withdrawals are balanced, over time, by a loss of water 1. Lee, C.H. (1915). The determination of safe yield of
underground reservoirs of the closed basin type. Trans. Am.
somewhere. Groundwater withdrawals by pumping result
Soc. Civ. Eng. 78: 148–251.
in one or more of the following: (1) a decrease in
2. Meinzer, O.E. (1923). Outline of Ground-Water Hydrology
the amount of groundwater in storage; (2) capture of
with Definitions. U.S. Geological Survey Water-Supply Paper
previously rejected recharge, if available; (3) an increase 494.
in water entering the aquifer system through increased
3. Conkling, H. (1946). Utilization of ground-water storage in
underflow, if available; (4) decreased outflows by way stream system development. Trans. Am. Soc. Civ. Eng. 3:
of underflow to downgradient aquifers; or (5) decreased 275–305.
outflow by reduction of discharge to streamflow or 4. Todd, D.K. (1959). Ground Water Hydrology. John Wiley &
to evapotranspiration. The hydraulics of an aquifer Sons, New York.
system evolve toward, and eventually may approach, 5. Thomas, H.E. (1951). The Conservation of Ground Water.
a new dynamic equilibrium under increased pumping. McGraw-Hill, New York.
An exception may occur in hydraulically isolated, arid 6. Kazmann, R.G. (1956). Safe yield in ground-water develop-
environments where little recharge is available. Under ment, reality or illusion? J. Irrig. Drainage Div. Am. Soc. Civ.
these conditions, if pumping exceeds the meager natural Eng. 82(IR3): 12.
recharge, no subsequent equilibrium will be reached and 7. DeWiest, R.J.M. (1965). Geohydrology. John Wiley & Sons,
groundwater mining is expected. New York.
Groundwater withdrawal or development of water 8. Bredehoeft, J.D., Papadopulos, S.S., and Cooper, H.H., Jr.
resources of an aquifer system occurs in stages. These (1982). Groundwater: The water-budget myth. In: Scientific
stages include the predevelopment stage, partially devel- Basis of Water-Resource Management. National Academy
oped stage, and fully developed stage. Under natural Press, Washington, DC, pp. 51–57.
or predevelopment conditions, an aquifer system is in 9. Theis, C.V. (1940). The source of water derived from
wells—essential factors controlling the response of an aquifer
dynamic equilibrium among change in aquifer storage,
to development. Civ. Eng. 10: 277–280.
groundwater underflow, and interaction between ground-
10. Wood, W. (2001). Water sustainability—science or manage-
water and surface water. As additional development
ment. Ground Water 39(5): 641.
occurs, water budget values change as a result of an
extended period of groundwater pumping. A partially
developed aquifer system will see moderate declines in SPECIFIC YIELD STORAGE EQUATION
groundwater level and decreases in groundwater in stor-
age; an increase in recharge may occur, and a decrease THOMAS HARTER
in natural discharge is expected. Full development of an University of California
aquifer system results in a large magnitude of ground- Davis, California
water level decline, additional depletion of groundwater
in storage, maximization of recharge, elimination of dis-
charge to surface waters, and underflow to downgradient The specific yield, Sy , is the ratio of the volume of water
aquifers may cease. that can be drained by gravity from a mass of porous
Safe yield cannot be defined solely on the basis of hydro- rock (unconsolidated or consolidated) and the volume of
logic conditions. Hydrologists may describe and quantify that mass:
VW
the effect of groundwater withdrawals on an aquifer sys- Sy = ,
tem, but safe yield is properly a determination by a team VT
of natural resource managers, and the results must be where VW is the volume of drainable water and VT is the
acceptable to concerned citizens. It is often difficult to total volume of porous rock or sediments (Fig. 1). Specific
reach consensus on the true safe yield of an aquifer sys- yield, Sy , is typically expressed as a percentage.
tem. Although difficult, the term and concept are the basis Specific yield is part of the total porosity of a porous
of many water management programs and groundwater rock or sediment. The total porosity (total pore space) of a
laws and therefore must be given careful attention. rock or sediment formation is much larger than the specific
Wood (10) states, ‘‘. . .scientists and engineers must yield. Total porosity includes the fraction of pore space that
avoid the trap of blurring the boundary between scientific is interconnected (called ‘‘effective porosity’’) and porosity
investigations and management decisions. Neither I nor due to isolated pores. The effective porosity itself includes
any other scientist or engineer can say whether any given specific retention and specific yield. Specific retention is
volume of water or delivery rate meets the needs of anyone that volume fraction of water that is held back by adhesion
beyond the minimum personal requirement of two or three and capillary forces, when an aquifer is drained. Specific
liters per day necessary for long-term survival. ‘Meet- yield is the amount of water that is actually available
ing the needs’ requires many value judgments, and their for groundwater pumping, when sediments or rocks are
assessment lies outside the realm of science or engineer- drained due to lowering of the water table.
ing. In many ways, sustainability appears to be analogous Specific yield depends on the distribution of pores, their
to the management concept of aquifer ‘safe yield’ that has shape, and grain size (Table 1). The larger the surface
been extensively used and misused in the last 85 years.’’ area of the rocks or sediment particles per unit volume