22nd ARRB Regional Symposium

DESIGN AND CONSTRUCTION OF PAVEMENTS IN HIGH RAINFALL AREAS

K.B. Wallace
22nd ARRB REGIONAL SYMPOSIUM 17

DESIGN AND CONSTRUCTION OF PAVEMENTS IN HIGH

RAINFALL AREAS

Keith Wallace, Ph.D., M.I.E.Aust., Senior Geotechnical Engineer,

Coffey & Partners Pty.Ltd., Brisbane.

ABSTRACT

This paper reviews some of the basic concepts underlying current recommend-
ations for design of pavements in high rainfall areas and refers to specific
guidelines for good design practice. The fundamental influence of pore
pressure or suction on soil strength and stiffness is emphasised, description
of moisture conditions in terms of moisture ratio is discussed and the effect
of poor compaction and plastic fines is considered. The concept of equilib-
rium moisture conditions beneath covered areas is introduced and the extent
of seasonal edge effects defined. Permeability is established as the most
sensitive and critical design parameter and requirements for low permeability
gravels are stated. Mechanisms of shoulder water entry into pavements are
also discussed briefly. The presentation concludes with an annotated list of
references which are recommended for further reading.

ACKNOWLEDGEMENT: The author acknowledges the support of Maurice Philp,

Queensland Manager of Coffey & Partners Pty.Ltd., in preparation and
presentation of this paper. However the content and opinions expressed herein
are the responsibility of the author and do not necessarily correspond with
the practice or opinions of his employer.
22nd ARRB REGIONAL SYMPOSIUM 19

1. INTRODUCTION

Nearly all of the high rainfall areas of Northern Australia are seasonally
wet with intense but variable rainfall during a 3 to 4 month long wet season
and very little rainfall during the winter months. Permanent water tables
are generally very deep except along the coastal flats. Pavements are often
constructed during the drier months when the climate is arid and water for
compaction is at a premium but there is always a risk of heavy rain during
construction and a certainty that the completed pavement will be subjected to
heavy rainfall and possibly local flooding. In this climate the subgrades
and bases of well-drained and well-maintained pavements are much drier than
the soaked condition on which earlier, more publicised, overseas pavement
design and materials selection procedures were based.

It is now 35 years since Loxton, McNicholl and Bickerstaff investigated the

moisture condition of the sandy clay subgrade at Darwin Airport (and sub-
grades at Pearce, Nowra and Essendon Airports). In one of the most signif-
icant geotechnical engineering publications they reported that the moisture
content below covered areas tended to vary according to the plastic limit of
the soil, was generally less than the plastic limit and was very much less
than the moisture content of a soaked CBR specimen. They advocated proced-
ures for subgrade strength assessment which have since stood the test of time
and have only recently been explicitly incorporated in State road authority
pavement design manuals.

Over the intervening years the basic principles governing the effect of water
on pavements have been more closely defined and practical guidelines for
design, construction and maintenance have been documented. This paper reviews
some of these basic principles and is supplemented by a reading guide. It
concludes with some presumptive subgrade CBR values for design of minor pave-
ments in high rainfall areas.

2. IMPORTANT CONCEPTS

2.1 HOW WATER AFFECTS SOIL STRENGTH AND STIFFNESS

The fundamental parameter governing the strength and stiffness of soils and
unbound pavement materials is the effective stress. With pavements applied
stresses do not vary greatly from one pavement to another so that the most
critical parameter is therefore the pore water pressure or suction. For
example, the shear strength of granular materials and normally consolidated
fine-grained soils is described by the well known effective stress equation:

T = (a — u) tan cP

Strength and stiffness is reduced if the pore pressure, u, is increased as a

result of confined seepage or as a result of external stressing of a nearly
saturated, compressible soil. Alternatively, moist to dry soils will have a
pore water "suction" which acts to increase shear strength and stiffness.

It is therefore the pressure or suction of the water in the pavement rather

than the amount of water that affects pavement behaviour. This is sometimes
obvious at the site of a pavement blow-up due to seepage in hilly terrain.
Before the actual break-up the pavement is quite soft and spongy due to
unrelieved pore pressure but once these seepage pressures are relieved the
volume of seepage that emerges can be surprisingly small.
20 22nd ARRB REGIONAL SYMPOSIUM

Two soils of different texture may have similar strength and stiffness even
though their water contents are quite different. Therefore if we are to make
a meaningful comparison of the state of soils we should use the more basic
parameter - suction or pressure - rather than water content. Under normal
circumstances the pore pressure in our pavement materials and subgrades is
sub-atmospheric and suction measurements are appropriate. While suctions can
be measured conveniently and economically the use of suction measurement has
not found widespread acceptance in practice. Instead, moisture ratios are
widely used. In this system the water content of the soil is expressed as a
ratio of the water content which that particular soil would have at a given
strength or consistency. For fine grained soils the reference moisture
content that is used is the plastic limit corrected for the presence of
particles coarser than 425 pm.

water content x 100

moisture ratio =
plastic limit x percentage finer than 425 pm

The plastic limit is a useful reference moisture content because it is of the

same order as the optimum moisture content for compaction and it is also an
upper limit to the field moisture content of many subgrades under wet but
well-drained conditions. For coarse grained soils, the optimum moisture
content is the corresponding reference moisture content. It represents an
upper limit to the field moisture content of well-compacted and well-drained
granular pavement materials.

With pavement gravels, the critical pore pressures are those developed as the
material tends to compress under transient traffic loading. If loose granular
materials (or soft cohesive soils) are loaded quickly when they are close to
saturation the high pore pressures which are developed reduce shear strength
to a low level. Occasionally we are faced with the embarrassment of having
wet, poorly compacted gravel bases break up soon after sealing and opening to
traffic. In granular materials this is avoided by pre-loading through com-
paction at the most favourable compaction water content. Subsequent volume
change under traffic loading then tends toward dilation rather than densi-
fication. However if inclusions of clayey material are exposed to free water
they will swell to a soft state that is independent of the extent to which
they are compacted. In this swollen state they will be exceedingly weak when
subjected to traffic stresses. It is for this reason that stringent restrict-
ions are applied to the degree of swelling (plasticity index or linear shrink-
age) of the fines and the amount of fines in a pavement gravel. These two
factors may also be used to obtain a first estimate of the suitability of a
pavement gravel expressed in terms of its Californian Bearing Ratio.

2.2 EQUILIBRIUM MOISTURE CONDITIONS

If an area of ground is sealed by an impermeable layer such as a pavement the

moisture conditions beneath the central section of that area change until,
after a period varying from several months to several years, a long term
equilibrium moisture condition is attained. This long term equilibrium
moisture condition is independent of initial conditions, it does not vary
seasonally and it is in equilibrium with a groundwater table and/or average
climatic influences at the edges of the sealed area. Even in very wet areas
the equilibrium moisture content of well drained pavement layers is at or
drier than optimum moisture content. Equilibrium moisture conditions are
therefore generally quite favourable.

The time taken to change from the initial moisture conditions to long term
equilibrium conditions varies according to the texture or permeability of
the subgrade soil. With heavy clay soils it may take three or four years.
22nd ARRB REGIONAL SYMPOSIUM 21

This is of practical consequence in that the effects of soaking of the sub-

grade during construction can persist and affect pavement performance.

The presence of a shallow water table will control the equilibrium moisture
content. The maximum depth at which a water table affects the subgrade moist-
ure content varies according to soil texture. On dune sands, water tables
deeper than 900 mm have little effect whereas the water content of heavy clay
subgrades may be affected by the presence of water tables shallower than 6
metres.

For deep water tables equilibrium moisture conditions are primarily dependent
on the long term average climate at the edge of the sealed area. (The pres-
ence of permeable pavement layers or upper soil horizons will also tend to
increase equilibrium moisture contents.) One approach to estimating equil-
ibrium moisture content of a subgrade is to select an existing pavement or
similar construction and subject to the same climatic and/or water table
influences. The moisture ratio of the subgrade beneath the central area of
this pavement is then measured and it is assumed that under similar circum-
stances the new pavement subgrade will have the same moisture ratio.

It must be noted that all discussions of the concept of equilibrium moisture

conditions assume that the pavement is well drained and maintained so as to
preclude the effects of water entry through the seal or seepage in cuttings.

2.3 SEASONAL EDGE EFFECTS

Between the central zone of moisture equilibria and the outer edges of a
covered area there will be zones within which moisture content varies accord-
ing to changes in climate. Rainfall at the edge will infiltrate and cause
progressive wetting-up of this edge zone. During dry periods there will be
a progressive drying-out due to evapo-transpiration and drainage.

The best way of detecting these seasonal changes in moisture conditions is

to use the Benkelman beam to measure surface deflection under a heavy single
axle loading. Changes in surface deflection show the integrated effect of
changes in moisture content on pavement and subgrade stiffness. With well
constructed pavements the resilient deformation of the pavement layers is
relatively small and therefore seasonal changes in pavement surface deflect-
ion are primarily due to seasonal changes in subgrade strength. Seasonal
measurements of pavement deflection show that, while wheel inner wheel-path
deflections may not change, deflections in the outer wheel path may be up to
20 percent or more higher towards the end of the wet season. Measurements of
deflections at closely spaced intervals across the pavement have shown that
generally seasonal weakening is confined to the outer 0.5 m to 1 m.

It seems that the most practical and economical method of counteracting the
potential damage due to seasonal weakening at the pavement edges is to extend
the sealed surface one metre beyond the trafficked pavement. However this
does not occur naturally when a pavement is widened because the heavier
vehicles tend to travel at the same distance from the outer edge of the seal.
Heavy vehicles must be positively deterred from travelling along the edge of
wider seals. This is especially important and difficult on steep climbing
gradients.

2.4 CRITICAL FACTORS AFFECTING SHOULDER WATER ENTRY

If shoulder water entry is not controlled the lateral extent of edge effects
will be increased and equilibrium moisture conditions will be wetter. The
rate of water entry depends both on the potential gradients that are tending
22 22nd ARRB REGIONAL SYMPOSIUM

to move water in and the permeability of the materials. Of these two sets of
factors, permeabilities are by far the most sensitive and the most important
to control. The permeability of a pavement material may be reduced one
hundred-fold by barely perceptible changes in texture while the potential
gradient is relatively unaffected.

The grading of a material may be described in terms of a grading index, n

where:
D
Percentage finer than particle size D = ( ) x 100.
D
max

For a continuously graded granular material with grading index n < 0.5 the
permeability is sensitive to the amount and nature of the finer components.
As a first approximation the permeability varies as the square of the 10
percent size i.e. the size for which 10 percent of the material is finer.
Permeability is further reduced if the finer components have some degree of
plasticity.

Segregation of gravel can increase permeability by several orders of magnit-

ude. Gravels with high grading indices (n = 0.5) and non-plastic fines are
most liable to segregation.

By good compaction the construction engineer can effect a ten-fold reduction

in permeability.

The question often arises as to whether cement modification of a pavement

material will reduce permeability. The answer is not straightforward. If
the cement modified material is well compacted and remains intact without
shrinkage cracking then its permeability will be less than if it were not
cement modified. (This is the situation which is demonstrated in laboratory
tests.) Perhaps the worst situation that can occur is for the cement to
set-up before compaction thus inhibiting compaction and leaving a highly
permeable granular material. Shrinkage cracks will be highly permeable but
reduced permeability of the matrix between cracks could provide added
protection of moisture-susceptible gravels.

2.5 PATTERNS OF SHOULDER WATER ENTRY

In steady rain, water will be entering the shoulder surface under a vertical
hydraulic gradient slightly in excess of the unit gravity gradient. The
rate of entry is primarily dependent on the permeability of the shoulder
surface. Compared with permeability, moisture suction effects are quite
secondary. The foolproof strategy for reducing shoulder water entry is to
provide a low permeability shoulder surfacing. The leading question is ...
"how low should the permeability of the shoulder material be so as to effect-
ively cancel out shoulder water entry" ... The writer believes that a
permeability of 5 mm/day or less is desirable. To achieve this the material
should be continuously graded and the grading index should not exceed 0.45,
at least 10 percent should be finer than the 75 pm sieve size and the fines
should have a linear shrinkage of at least 3. It is thought that these are
necessary but not necessarily sufficient conditions for low permeability of
a pavement material.

Once the water has entered the shoulder it will tend to percolate downward
until it encounters a layer of lower permeability. The ideal situation is
to have permeability constant or increasing continuously with depth. This
situation is encountered sometimes on very sandy subgrades but it only
requires the slightest reduction of the underlying permeability for develop-
22nd ARRB REGIONAL SYMPOSIUM 23

ment of a perched water table. For example, if the underlying permeability

is one-third that of the surface layer a perched water table will form to
two-thirds of the depth of the surface layer. This must be a common occurr-
ence because more than three-fold variations of permeability occur throughout
apparently homogeneous materials.

Once a perched water table develops the infiltration tends to be diverted

laterally into the pavement. From what is seen of water running off pavement
surfaces due to crossfall one might expect that the crossfall of the under-
lying layer might counteract lateral diversion of water into the pavement.
Further consideration of the relative hydraulic gradients shows that this
is not so. Water will be entering the shoulder under at least unit gradient
but the controlling gradient for lateral run-off is of the same order as the
4% to 5% crossfall. Therefore, for water to be able to drain rapidly enough
through the shoulder the permeability of the shoulder underlayer would have
to be at least 25 times greater than that of the shoulder surfacing and
adjacent pavement material.

Situations in which the pavement material is relatively permeable and is

contained between low permeability shoulders and subgrade are referred to
as boxed construction and are generally undesirable. With respect to shoulder
water entry, any infiltration will be trapped at the pavement edge where it
will soften the subgrade. Even where the permeability of the pavement gravel
appears to be quite low, segregation or poor compaction of the edge of the
pavement can produce the same result as boxing in of a permeable material.

3. ADDITIONAL READING

3.1 Main Roads Dept., Queensland (1981). "Interim Manual for Pavements".
(This manual outlines a set of practical guidelines for selection of
design moisture conditions, material specification and cross section
design in a State that contains some exceedingly wet regions.)

3.2 Murphy, H.S. (1978). "Controlling Sub-soil Water and its Effects".
Rid. M.R.D. Form No. 11AX2. (A more extensive guide covering: Cuttings;
Embankments; Pavement Configuration and Materials; Sub-soil Drainage
and Drainage Layers.)

3.3 Lambe, T.W. and Whitman, R.V. (1969). "Soil Mechanics". Wiley.
(Chs. 16-19, 26, 28 and 29 provide an essential background knowledge of
pore pressure development, permeability and undrained shear strength.)

3.4 Aitchison, G.D. and Richards, B.G. (1965). "A broad scale study of
moisture conditions in pavement subgrades throughout Australia". Proc.
Symp. on 'Moisture Equilibria and Moisture Changes in Soils beneath
Covered Areas' pp. 184-232, Butterworths : London. (Describes equilib-
rium moisture conditions throughout Australia and establishes basic
principles.)

3.5 Richards, B.G. (1967). "Moisture flow and equilibria foundations".

Permeability and Capillarity of Soils, A.S.T.M., S.T.P. No. 417, pp.4-34.
(An excellent account of the basic principles.)

3.6 Morris, P.O. et al (1968). "Moisture and density conditions in base and
subgrade materials in roads in Tasmania and Victoria". A.R.R.B. Internal
Report. (Extensive data from this small area of South-Eastern Australia.)
24 22nd ARRB REGIONAL SYMPOSIUM

3.7 Wallace, K.B. and Leonardi, F. (1978). "Seasonal Changes in the Strength
of Rural Road Pavements in Coastal North aueensland". A.R.R.B. Special
Report No. 16. (An extensive study which leads to a series of recommend-
ations for reduction of seasonal weakening.)

3.8 Wallace, K.B. (1977). "Moisture transients at the pavement edge:

analytical studies of the influence of materials and cross-section
design". (Basic principles of shoulder water entry and drainage.)

3.9 Hilf, J.W. (1975). "Compacted Fill". Ch. 7 of 'Foundation Engineering

Handbook' Ed. Winterkorn and Fang, Van Nostrand Reinhold, PP. 244 to 311.
(Compaction processes, compaction control and the properties of compact-
ed soils.)

4. PRESUMPTIVE CBR VALUES FOR HIGH RAINFALL AREAS

(Seasonally Wet with Average Annual Rainfall >600 mm)

The following subgrade CBR values are suggested for design of minor pavements
where extensive evaluation of subgrade moisture and strength conditions is not
economical.

SUBGRADE CBR FOR MOISTURE CONDITION

SUBGRADE TYPE
A* B* C*

. SAND (non to slightly plastic

with generally t 12% finer than
0.075 mm.)
(a) DENSE 10 20 20
(b) MODERATELY DENSE 7 10 7
(often poorly graded)
2. SILT (non to slightly plastic) 1 to 2 test test
3. SOFT TO VERY SOFT CLAYEY SILTS test test test
AND CLAYS in-situ in-situ in-situ
4. CLAYS
(a) P.I. = 10 5 8 6
(b) P.I. = 20 4 7 6
(c) P.I. = 30 3 5 6
(d) P.I. > 40 2 3 5
5. LATERITIC CLAYS
(a) in cut Treat as for P.I. = 20
(b) in fill Treat as for P.I. = 30
6. EXTREMELY WEATHERED AND EXTREMELY Treat as for soil
TO HIGHLY WEATHERED ROCK
7. SLAKING MUDSTONE 2 10 5
(Undesirable as earthfill)

_....._... . .
22nd ARRB REGIONAL SYMPOSIUM 25

• Moisture Condition A refers to: shallow water tables (< lm below formation
level); floodways, causeways and other pavements which could be subjected
to prolonged flooding; pavements with permeable bases or subgrades which
could be subject to temporary flooding of the pavement or to prolonged
ponding adjacent to the formation; and cuttings below the water table or
where seepage is likely.

• Moisture Condition B refers to well-drained pavements of low permeability

in which surface water drains away from the pavement without prolonged
pondage on or near the shoulders but in which subgrade moisture conditions
are controlled by a water table at 2m depth below formation level.

• Moisture Condition C refers to well-drained and well-maintained pavements

similar to those for condition B but in which the water table is deep and
subgrade moisture conditions are determined by the climate. Where surface
and subsurface drainage conditions are particularly favourable and can be
guaranteed the CBR values for clays will be conservative.

5. CONCLUSION

Over the last 35 years, developments in our understanding of the mechanics of

soils and of water movement through unsaturated soils have provided a more
rational background for assessment of problems of design pavements for wet
areas. Unfortunately these developments have not provided us with accurate
tools for prediction of pavement behaviour and therefore the most important
and difficult development of all is the combination of accumulated experience
and established principles to provide guidelines for good design practice.
This paper has reviewed some of the important principles and referred to one
set of current guidelines. The most fruitful area for further development
lies in documentation of field experience in terms of established principles
so that existing guidelines will evolve into even better rules for future
practice.