CE 101-SECTION 4CEA

CHAPTER III

PRECIPITATION – Water in the form of rainfall, snowfall, hail, frost and dew when
reaches the earth from the atmosphere.

3.1 Forms of Precipitation

a. Rainfall – when the size of water drop is larger than 0.5 mm. Rainfall may be
light, moderate and heavy if intensities of rainfall are 2.5mm/hr, 2.5mm to
7.5mm/hr. and more than 7.5mm/hr. respectively.
b. Snow – is composed of ice crystals, chiefly in complex hexagonal form and
formed into a ball of mass into snowflakes which may reach several cm in
diameter. New fresh snow has an ini tial density from 0.06 to 0.15g/cm3 and the
average value is assumed to be 0.1g/cm3.
c. Drizzle – it is a fine sprinkle of numerous water droplets of size 0.5mm and
intensity is less than 1mm/hr.
d. Glaze – It is the ice coating formed when comes in contact with cold ground at
0⁰C water drops freeze to form ice coating.
e. Sleet – they are frozen raindrops of transparent grains, which form when rain
falls through sub-freezing temperature. –combination of rain and snow.
f. Hail – It is the precipitation in the form of irregular lumps of ice of size more than
8mm. Hails occur in violent thunderstorm in which vertical currents are very
strong. Hailstones may be spherical, conical or irregular.
g. Dew – forms directly by condensation on the ground mainly during night when
the surface has been cooled by outgoing radiation.

3.2 Types of Precipitation

a. Frontal/Cyclonic Precipitation – this precipitation occurs due to conflict

between two air masses. When two air masses with contrasting temperatures and
densities clash with each other, condensation and precipitation occur at the surface of
contact. The precipitation is called frontal as the surface of the contact from which it
precipitates is called front. When two air masses, both cold and warm, are downwards
in a low-pressure area, the front is stationary. Precipitation from this condition is called
non-frontal.

b. Convective Precipitation – this type results from the upward movement of air
which is warmer than its surrounding. As the density of warm air is less, it has the
tendency to continue its rising until it reaches a level where it has the temperature of its
environment. The contrast causes convection and cools adiabatically to form the shape
of the cloud like cauliflower which finally bursts into thunderstorm. It is accompanied by
destructive winds, called tornadoes.

c. Orographic Precipitation – this precipitation caused by orographic lifting of

air over mountain barrier from a large water body. This air from water body carries a lot
of moisture and the high mountain barrier causes the air to lift up. At certain height it
condenses and precipitates in the windward slope. Generally rainfall is low.

3.3 Measurement of Precipitation

Precipitation which falls on earth’s surface for a specified time is measured in depth of
water. The instruments used for its measurement are called rain gauges. The following
are some rain gauges used to measure the rainfall:

a. Symon’s (Non-Recording Type)

b. Weighing Bucket Type
c. Tipping Bucket Raingauge
d. Siphon Raingauge or Floating Gauge
e. Storage Raingauge
f. Telemetering Raingauge
g. Radar Measurement of Rainfall
h. Automatic Radio-Reporting Raingauge
3.4 Network of Raingauges in a Catchment and their Adequacy

Thus, to have a correct data, number of rainfall stations should be within the prescribed
limit as the precipitation records are most fundamental data used in the hydrological
studies. These are essential for:

1. Carrying out storm analysis over a catchment

2. Depth-Area-Duration analysis (DAD)
3. Fixation of design food
4. Stream Flow Forecasting
5. Reservoir Regulation

Adequacy of Rainfall Stations

Number of rainfall stations, i.e., raingauge stations for an area to give necessary
average rainfall with certain percentage error can be obtained from statistical
consideration. Following are the steps in the estimation of optimum number of
raingauges:

1. Calculate Total Rainfall;

PT = P1 + P2 + P3 + ….+ Pn
2. Calculate Mean Rainfall;
Pm = PT / n
3. Calculate the sum of the squares of all the rainfall of the gauges;
∑P2 = P12 + P22 + P32 + …. + Pn2
4. Calculate the sum.
n
2
∑ ( Pi−Pm )
i=1

5. Calculate the square of standard deviation.

n
σ = s = ∑ ( Pi−Pm )2 / n-1
2 2

i=1

6. Calculate the coefficient of variation.

Cv = 100 √ s2 / Pm
7. Optimum number of raingauge N that would be necessary to estimate the
average rainfall within percentage error P ;
N= (Cv/P) 2
8. For error less than or equal to 10%,
N= (Cv/10)2
9. Additional number required = (N-n);

The additional numbers should be distributed in different zones in proportion to their

area.
Examples:

1. The average rainfall of 5 raingauges in the base stations are 89,54,45,41 and 55
cm. If the error in the estimation of rainfall should not exceed, how many
additional gauges may be required?
2. In certain watershed (River Basin), there are four raingauges stations. The
normal precipitation amounts respectively to 800, 520, 440 and 420 mm.
Determine the optimum of raingauges in the watershed if it is required to limit the
error in average rainfall to 6%.

3.5 Methods of Computing Average Rainfall

The average rainfall over a catchment or watershed is obtained by the following four
different methods from available rainfall data of different raingauges:

a. Arithmetic Mean Method: It is the simplest method. If there are n numbers of

raingauges stations in a watershed, and P1, P2, P3,…., Pn are amount values of
rainfall in those n stations in a given period (say annual precipitation), then
arithmetic mean of those rainfall records give the average value of precipitation
to the watershed or catchment, i.e.,
n
Pm = ( P1 + P2 + P3 + …. + Pn ) / n = 1/n ∑ ( Pi )
i=1

where ;
Pm = is the average rainfall in the catchment.
If the catchment area is more or less flat, distribution of raingauge stations is uniform,
and then average P gives good result. This method is quite rapid and easy.
b. The Thiessen Polygon Method: A method of assigning a real significance to
point rainfall values. Perpendicular bisectors are constructed to the lines joining
each measuring station with those immediately surrounding it. These bisectors
form a series of polygons, each polygon containing one station
P = (P1A1 + P2A2 + P3A3 + P4A4 + P5A5 + P6A6 ) / (A1 + A2 + A3 + A4 + A5 + A6)
Thus, in general for N stations;
N
P = ∑ ( PiAi ) / A
i=1

where A is the total Area of the catchment.

c. Isohyetal Map Method: The most accurate method of finding average

precipitation over an area provided it is used by an experienced analyst. Location
 of stations and amount of precipitation are plotted on a suitable map and
contours of equal precipitations called Isohyets.

Let the isohyets represent the rainfall P1, P2, P3,…., Pn and areas between the
successive isohyets are A1, A2, . . . ., An-1.

P 1+ P 2 P 2+ P 3 Pn−1+ Pn
P = [A1 ( ¿ + A2 ( ¿ +…. + An-1( ¿] / A
2 2 2

d. Weighted Length Method: Considering linear variation of rainfall between two

raingauge stations, mean rainfall of a catchment is equal to the summation of
product of the two lengths of the line joining the two stations divided by total
length of the line joining the raingauge stations.

Examples:

1. Using Thiessen polygon method, compute the depth of average precipitation

over a basin using the following data:

Raingauge Area of Polygon Observed Depth of Precipitation in mm

1 480 50
2 2860 65
3 2200 85
4 1600 90
5 710 60
6 270 70
2. The isohyets for annual rainfall over a catchment basin were drawn. The areas of
strips between the isohyets are given. Determine the average precipitation over
the basin.

Isohyets Area in sq.km.

75-85 1600
85-95 3000
95-105 2800
105-115 1000
115-135 900
 135-155 700

3.6 Interpolation and Adjustment of Missing Data

Sometimes rainfall data in one or two stations may be missed. Inconsistency of average
rainfall may occur in catchment due to various reasons like missing precipitation data,
exposure of station may be changed with growth of trees and buildings, etc. in such
situations, interpolations in the estimates of average rainfall is required. These
interpolations are done by the following methods depending upon causes of
inconsistency.

a. Arithmetic Mean Method: Sometimes, some station has a break in record due
to absence of observer or failure of the instrument. It is then necessary to
estimate that missing data. To estimate at least three stations close to this station
are selected. It is necessary that these three stations are evenly distributed
around the station. If normal precipitation at each of these selected stations is
within 10% of that for station with missing data, then simple arithmetical mean of
precipitation of those three stations will give the value of the missing station. If
PA, PB and PC are precipitation of the nearby stations and P x is the estimated
precipitation of the missing station, then

PA + PB+ PC
Px =
3

b. Normal Ratio Method: This method is used for interpolation when the normal
annual precipitation of, say, three nearby stations A, Band C NA, NB and NC,
differs from that of the Nx of station X with missing data from previous record by
more than 10%. If PA, PB ,PC are average storm precipitation of A, B, and C in the
year when average precipitation at X, i.e., Px is missing, then

Nx Nx Nx
Px = [ PA + PB + PC ]
NA NB Nc

c. Station Year Method: In this method, the records of two or more are combined
into one long record provided station’s record are dependent and areas of
stations are climatologically the same. The missing record at a certain station in a
particular year may be found out by the ratio of the average or by graphical
comparison. For example, in a certain year total rainfall of station A is 78cm and
for the neighbouring station B, no record is available. But if average annual
precipitation at A and B in previous year record are 72 and 82 cm respectively,
then missing year’s rainfall at B, i.e., PB may be obtained by simple proportion as:
 78 PB
= ; PB = 88.833 cm
72 82

Note: This result must be checked by another station C for better results. If the new P B
is slightly different, average of the two will provide a bit accurate P B.

d. Double Mass Curve: Kohler3 presented analysis, testing and adjustment of

double mass curve. A double mass curve shown in Figure 3.9 is drawn for a
period of 29 years of a catchment to check the inconsistency of rainfall record of
a station (say X) and accordingly to adjust the incorrect results.

The cause of inconstancy may be:

1. Shifting of rainfall station to a new position

2. When ecosystem in the vicinity is changed due to land slide, forest fire, flood and
earthquake.
3. Error in observation from a certain year.
4. Change in the vicinity of the station due to growth of trees, buildings, fencing,
cutting of forest nearby changing the wind pattern, etc.
5. Raingauge may be faulty from a certain period.
6. Site of instrument may be changed or replaced without record.

To draw this curve, a group of stations (say 10) is taken as base station in the neighbourhood of
the problem station X. the accumulated rainfall of station X, (∑P x ) and accumulated values of
average of group of base stations (∑P av) are calculated starting from the latest record, i.e., from
1999. In figure 3.9, the values of (∑Px ) as ordinate and ∑Pav, as abscisa to a scale is plotted for
available data of rainfall from the year 1999 to 1971. In the plot, a break in the slope has been
seen from the year 1983. It indicates a change in precipitation regime of station X. The values at
X beyond the period of change of regime, i.e., the point 1983 in Figure 3.9 (double mass curve)
has been corrected. The original slope of the mass curve of rainfall from 1999 to 1983 is
extended by a dotted lins as shown in the figure. The correction to the data of rainfall at X from
1982 to 1971 is slope of the dotted line divided by the slope of the second mass curve of 1983
to 1971, i.e.,
c /b c
Correction factor = =
s/b s
The corrected data at X is:

c
PcX = PX ( )
s

Thus, al the inconsistent data of X from 1982 to 1979 are corrected by multiplying by c/s, the
value of which is obtained from the plot measuring c and s as per scale.
The procedure of testing and correcting the inconsistency of data of a point by double mass
curve has been presented first by Max A. Kohler in 1949.

Examples:
1. The normal annual rainfall of stations A, B, C and D in a catchment is 80.97, 67.59,
76.28, 92.01 cm respectively. In the year 2006, the station D was inoperative when
station A, B, C recorded annual rainfall of 9.11, 72.23, 79.89 cm respectively. Estimate
the missing rainfall at D in the year 2006 by normal ratio method.
2. Rainfall data for station X as well as the average annual rainfall measured at a group of
10 neighbouring stations located in a meteorologically homogeneous region for 22 years
is given as follows. Test the inconsistency of data at X by double mass curve method.
Find the year in change of regime and the correction factor to be applied.

Year Station X 10 station Year Station X 10 station

(cm) average (cm) (cm) average (cm)
1999 163 161 1988 141 156
1998 130 146 1987 196 193
1997 137 130 1986 160 128
1996 130 143 1985 144 117
1995 140 135 1984 196 152
1994 142 163 1983 168 155
1993 148 135 1982 194 161
1992 95 115 1981 162 147
1991 132 145 1980 178 146
1990 145 155 1979 144 132
1989 158 164 1978 177 143