J. King Saud Univ., Vol. 3, Eng. Sei. (1), pp.

89-100 (1411H11991)

Modeling of Infiltration from An Artificial Recharge Basin

with a Decreasing Ponded Depth

Fouad F. AI-Muttair and Abdulaziz S. AI-Turbak

Civil Engineering Department, College of Enginnering, King Saud University,
P. o. Box 800, Riyadh 11421, Saudi Arabia

Abstract. The first important physical process in basin recharge is the infiltration of water from the
basin's bed to the water table below. The Green and Ampt approach was applied to model infiltration
from an artificial recharge basin with a qecreasing ponded depth. Several field experiments were carried
out. Results of one experiment were used to calibrate the model. The suggested approach was then used
to predict infiltration for the other experiments and its results were compared to field values.

Introduction

Artificial recharge in arid and semi-arid regions plays an important role in conserva-
tion of water and the avoidance of depletion of existing aquifers. An important
method of artificial recharge is basin recharge; where excess surface water is spread
over the area of a basin (or basins) and the water is allowed to infiltrate to the water
table below. A good description of spreading methods of artificial recharge (like the
basin method) can be found in Todd [1].
The first step in recharge from a basin (or any other similar spreading facility) is
infiltration from the surface of the basin after ponding of water in it. Infiltration can
be defined as "the physical process of water entry into the soil". Such a process
involves the displacement of air into the soil matrix by water.
Due to the importance of infiltration in many hydrological and engineering
problems, many researches have been devoted to its study since the beginning of the
century. Many equations, empirical as well as physically-based have been proposed
to estimate infiltration of water as a function of time. Empirical equations were
developed using infiltrometer test data and the resulting relations were applied to
describe a rate of infiltration decreasing from an initial maximum to a final minimum

89
© 199 (A.H. 141). Production and hosting by Elsevier B.V. on behalf of King Saud University. This is an open access article under
the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
90 Fouad F.AI-Muttair and Abdulaziz S.AI-Turbak

rate. The most famous empirical equations of infiltration include those proposed by
Kostiakov [2] and Horton [3,4]. Empirical equations have serious limitations. Their
parameters have little or no physical meaning and they can not usually be determined
or estimated from data available on the soil. The most important physically-based
equations of infiltration are those due to Philip [5] and Green and Ampt [6]. Both
equations, and especially the Green and Ampt equation, utilize parameters some of
which can be evaluated from physical properties on the soil. Fok [7] gives a summary
on development and limitations of different infiltration equations.
The Green and Ampt model has been the subject of considerable development
in recent years because of its simplicity and its satisfactory performance in many
hydrological problems. For instance, it has been extended to soils of non-uniform ini-
tial moisture content [8], to layered soils [9] and to crust-topped soils [10]. It has also
been used to model infiltration into homogeneous soils from constant [11] as well as
from unsteady rainfall [12].
The main objective of this work was to use the Green and Ampt model to study
the infiltration process in an artificial recharge basin with a decreasing ponded depth.
Several infiltration experiments were conducted in the basin. Using the results of one
of these experiments, the model was first calibrated. It was then used to predict infilt-
ration for the other runs and its results were compared to field values.

Data

The data utilized in this paper are obtained from five infiltration experimental
runs conducted in an artificial recharge basin. The dimensions of the basin are 24.76
X 14.47 x 0.4 m deep, with a total surface area of 358.28 m 2 • The walls of the basin
were constructed with cement blocks. The inside faces of the walls were covered by
a heavy duty plastic sheet to prevent lateral leakage of water. To measure the drop
in water level in the basin, two stilling wells equipped with pointer gages were instal-
led along the two long walls, one along each. The daily evaporation rate values in the
site were obtained from Class A pan installed near the basin. The daily values of
evaporation rate from the surface of the basin were obtained by multiplying the mea-
sured values of evaporation by a pan-to-Iake coefficient.
The groundwater levels in the vicinity of the basin were obtained from an obser-
vation well located at about 40 m downstream to the basin. Depth of water table in
the observation well was recorded before the start of each of the five runs.
Soil samples were also taken from different locations and depths in the basin.
The initial moisture content of the soil was then calculated as the average of all the
samples. Water was then pumped into the basin until a depth of 20-30 cm is ponded.
Table 1 shows the particulars of the five experimental runs.
 Modeling of Infiltration from An Artificial Recharge Basin 91

Table 1. Particulars of the Infiltration Runs

Initial Depth Initial moisture Depth of water
Run No. Duration (hr)
(cm) content table (m)
(% by volume)

1 23.24 0.504 8.06 43.00

2 22.85 2.856 3.69 53.15
3 20.76 3.912 8.98 46.13
4 21.45 2.892 13.11 45.50
5 27.04 9.468 1.60 53.52

After the pumping was stopped, the depths of water at different points in the
basin were meausred to determine the average depth of water at that time. The
installed pointer gages were then used to determine the drops in water level in the
basin. The monitoring of water levels at the gages continued for 43 to 53 hours.
Because of the high variability of infiltration rate during the beginning of each run,
readings were taken at short intervals. As time proceeded, the interval between
readings was increased. Druing each interval, the drop in water level was estimated
as an average of the measured drops at both gages.
In order to estimate the infiltration depth during each interval, it was necessary
to account for evaporation. The evaporation depth druing each time interval was
estimated by distributing the daily evaporation rate from the surface of the basin over
the time intervals based on the hourly variation of evaporation during the day which
were'taken from the data of an evaporation balance. That instrument records the
evaporation on hourly basis.
The actual infiltration depth during each time interval was then calculated by
subtracting the estimated evaporation depth for that interval from the total drop in
water level during the same interval. The resulting depth was then divided by the
length of the interval to obtain an average infiltration rate during that interval.

The Model

The Green and Ampt model can be derived by applying Darcy's law between
the soil surface and the location of the wetting front. The first assumption made in a
derivation is that at the wetting front, the water presure head remains constant. The
second assumption is that behind the wetting front, the soil is uniformally wet with a
consant hydraulic conductivity corresponding to the conductivity at natural saturation.
92 Fouad F.AI-Muttair and Abdulaziz S.AI-Turbak

The Green and Ampt infiltration rate equation can be written as [13, pp. 35-65]:

I = dw = K €Sf + W ) (1)
dt W
where

I is the infiltration rate, (LT-I)

W is the cumulative infiltration depth of water, (L)
is the time, (T)
K is the hydraulic conductivity of the wetted zone, (Lt-I)
Sf is the storage-suction factor, (L)

The storage-suction factor can be expressed by the following equation [13, pp.
35-65]:

(2)

where

Sf is the final water content behind the wetting front, fraction

Si is the initial water content for the soil profile, fraction
Hf is the effective capillary drive or suction at the wetting front, (L)
H is the ponded depth of water, (L)

There are two problems in using equations (1) and (2) to estimate the infiltration
rates from basin with variable depth of water. The first is that both I and Ware
unknowns in equation (1), while the second is that in equation (2), H is normally
assumed to be constant during the infiltration process.
Integration of equation (1) between the limits (tj _1, Wj _1) and (tj , Wj ) will result
in the following expression (14).
Sfj+Wj
K ( t·J - t·J- 1) = W.J - W·J- 1 - Sfj In ( Sfj+Wj-l ) (3)
where

tj is the time at the end ofjth period

t j _1 is the time at the end ofj-l period
Wj is the cumulative infiltration depth of water at the end of the jth period
Wj _1 is the cumulative infiltration depth ofwater at the end of the j-I period
Sfj is the variable storage- suction factor for the jth period
 Modeling of Infiltration from An Artificial Recharge Basin 93

The variable storage suction, Sfj can be estimated as:

(4)

where H j is the ponded depth of water at the beginning of the jth period.

The distance the wetting front advances to below the basin can be calculated
using the equation:

(5)

where Zfj is the distance from the bottom of the basin to the wetting front at the end
of the jth period.

The average ifiltration rate during the jth period can be calculated as:
Wj - W j_ 1
I·
J
= t j - t j _1
(6)

Computational Procedure

The procedure starts by first estimating the parameters Sf' Si' H f and K as will
be described in the next section.

The purpose of the model is to predict the infiltration rates from a basin given
the values of the above parameters and the initial ponded depth of water in the basin.

Let HI represent the initial ponded depth at the beginning of the first period. Let
E j represent the evaporation depth during the jth period.

The steps of computation for any period are as follows:

1. Calculate
ET·J- 1 = ET.J- 2 + E·J- 1

where

ET·J- 1 is the cumulative evaporation depth at the end of the j-1 period
ETj _2 is the cumulative evaporation depth at the end of the j-2 period

2. Calculate the ponded depth of water at the beginning of the period

H·J = HI - ET·J- 1 - W·J- 1

Note that during each period Wj _1 is known from the previous calculation.
94 Fouad F.AI-Muttair and Abdulaziz S.AI-Turbak

3. Calculate the variable storage-suction (Sfj) by using equation (4).

4. The cumulative infiltration depth up to the end of the jth period (Wj) is then
found by using equation (3) in a trial and error procedure.
5. Equation (5) is then used to check the position of the wetting front below the
basin. If Zfj is equal to or greater than the distance from the bottom of the basin
to the water table, the computations are stopped. Otherwise, the procedure is
continued through the next period.
6. Based on the values of Wj and "')-1' the infiltration rate through the jth period
can be calculated using equation ~ 6).

Parameter Estimation
To use the infiltration approach just described, it is necessary to estimate the
parameters Si' Sf' H f and K.
The initial moisture content Si was estimated by taking four soil samples (one
from each quarter of the basin) just before the beginning of each run, determining
the moisture content for each sample in the laboratory, and taking the average value
of the four moisture contents. The values of the initial moisture contents for each run
shown in Table 1 represent these average Si values.
The parameter Sf represents the maximum water content (as a fraction) behind
the wetting front. It was approximated by the average porosity of the soil in the basin.
That average porosity was calculated by taking the average of the porosities of four
soil samples (one from each quarter of the basin). The value of Sf used in the model
was 0.3184. The approximation of Sf by the average porosity is reasonable approxima-
tion becuase the model is applied under continuous application of water from above.
This will prevent any drying in the soil profile. Had the model bee.n applied under
conditions where drying and we~ting of the profile coccur, this approximation may be
questionable.
The effective capillary drive H f is a measure of the soil capillary pull, expressed
as an equivalent depth of water. Almost all reported values in the literature for this
parameter range between 5 and 40 cm [15,16]. This parameter can be neglected after
few minutes in the infiltration process and its value becomes insignificant even at early
times if the ponded head is in the order of 30 cm [17]. Since the durations of all the
experiments used in this study were between two and three days and the ponded
heads were between 20 and 30 cm, the choice of the value for H f will not affect the
results. The value of H f which was used in both the calibration and the validation of
the model was 35 cm.
The last parameter to be estimated is K; the hydraulic conductivity of the wetted
zone (below the basin). The value that should be used is the average hydraulic con-
 Modeling of Infiltration from An Artificial Recharge Basin 95

ductivity for the soil layers below the basin. Due to the difficulty in establishing an
average K value for the profile during infiltration, this parameter will be determined
by calibration.

Calibration and Application of the Model

As stated previously the value of the hydraulic conductivity of the wetted zone
K will be determined through calibration. Data of Run No. 1 were used for this pur-
pose. To avoid the fluctuations in the values of the actual infiltration rate that
resulted from field measurements, the total depth of water that infiltrated through
the soil from the time of ponding until the end of the run was used as a criterion in
the calibration. Different values of K were used in the model and the optimum value
that gave the measured cumulative infiltrated depth (3.657 cm/day) was selected.
Fig. 1 shows the actual and predicted cumulative infiltrated depths versus time for
Run No. 1.

20 Actual
18 _ _ _ _ _ Predicted
A
~
16
e ~

...
,/
~
..c: 1 4 /'
Q. ./
QJ /
"C
"C
12 /.
QJ
~ /
:: 1 0 /
is /
.5 /
QJ
.~ 8 /
~ /
"3 6
E /
u= /
4 /
/
2
0
4 8 12 16 20 24 28 32 36 40 4it 48 52
Time (hr)

Fig. 1. Cumulative infiltrated depth versus time Run No. 1

96 Fouad F.AI-Muttair and Abdulaziz S.AI-Turbak

To validate the model, the model ~ith its estimated parameters was then applied
to determine the variation of cumulative infiltrated depth with time for Runs No.
2,3,4, and 5. Figs. 2 to 5 show the results of the application. Table 2 shows the actual
and predicted cumulative infiltrated depths by the end of the run for all the runs.

20 Actual
18 Predicted

I 16
a
~
14

~ 12
~
b 10
is
.$ 8
6
4

O ............_,..,........_"""'-......._ ........._""--....._ ...........--.I~~

4 8 12 16 20 24 28 32 36 40 44 48 52

Time (hr)

Fig. 2. Cumulative infiltrated depth versus time Run No. 2

,..,.
c:: 20
Actual
Predicted

4 a 12 16 20 24 28 32 36 40 44 4& 52
Time (hr)

Fig. 3. Cumulative infiltrated depth versus time Run No. 3

 Modeling of Infiltration from An Artificial Recharge Basin 97

20
Actual
18 _ _ _ _ .... Predicted

E
~
16
.c
a
~
14
"C
"C
~
12
b
e
.5
10
~
~ 8
~
-=e 6
u= 4
2
0
4 a 12 16 20 24 28 32 36 40 ~4 Ita 52
Time (hr)

Fig. 4. Cumulative infiltrated depth versus time Run No. 4

Actual
20 ----- Predicted

E 18
~
...
..c:
Q..
16
~
"C
14
......
"C
~

12
...=
is
.5 10
~
.~ 8
~
-=e 6
u= 4
2
0
4 8 12 16 20 24 28 32 36 40 44 48 52
Time (hr)

Fig.S. Cumulative infiltrated depth versus time Run No. S

98 Fouad F.AI-Muttair and Abdulaziz S.AI-Turbak

Table 2. Actual and predicted cumulative infiltrated depths

Cumulative infiltrated Depth (cm) Actual-predicted

RunNo. ------x100
Actual Predicted Actual

1* 18.597 18.597 o
2 20.177 20.545 1.82
3 17.009 18.444 8.44
4 18.770 18.496 1.46
5 19.540 19.517 0.12

* Used in the Calibration

As can be seen from Table 2 the predicted cumulative infiltrated depths by the
end of Runs No. 2,3,4 and 5 are in good agreement with the measured values. Figs. 1
to 5 show that the model resulted in an over-prediction of the cumulative infiltrated
depths at the early stages of each run. This indicates that, the model is over-predict-
ing the infiltration rates at the early times of each run. This over-prediction is attri-
buted to the fact that, the model assumes that the infiltration process starts at the
starting time of application of the model. In reality, the infiltration process in the
basin started by the time when the pumping of water started in the basin. However,
these figures show that as time passes, tpe predicted values of cumulative infiltrated
depth become close to the actual values.

Concluding Remarks

The paper presents a method to model and predict the infiltration process
through a recharge basin with decreasing ponded depth. The predicted values of
cumulative infiltrated depths were compared with the actual values for four runs.
The model overpredicted the cumulative infiltrated depths at the early stages of each
run because the starting time of its application is different than the actual starting
time of infiltration. However, the cumulative infiltrated depth by the end of each run
is close to the actual value.

The technique used in this paper can be very helpful in designing recharge
basins. One of the essential requirements of the design of a recharge basin (or basins)
is to estimate infiltration rate from the bottom of the basin. The approach used in this
work can be easily applied for that purpose.

References

[1] Todd, D.K. Groundwater Hydrology. New York: John Wiley and Sons, 1980.
[2] Kostiakov, A. N. "On the Dynamics of the Coefficients of Water Percolation in Soils and the Neces-
sity of Studying it from a Dynamic Point of View for Purposes of Amelioration." Trans. Corn. Int.
Soc. Soil. Sci., 6th Moscow, Part A (1932),17-22.
 Modeling of Infiltration from An Artificial Recharge Basin 99

[3] Horton, R.E. "The Role of Infiltration in the Hydrologic Cycle." Trans. Am. Geophysical Union,
14 (1933), 446-460.
[4] Horton, R.E. "An Approach Toward Physical Interpretation of Infiltration Capacity." Proc. Soil
Sciences Soc. Am., 5 (1939), 399-417.
[5] Philip, J.R. "The Theory of Infiltration. 1. The Infiltration Equation and Its Solution." Soil Sci-
ence, 83, No. 5 (1957),345-357.
[6] Green, W.H. and Ampt, G.A. "Studies on Soil Physics. 1. The Flow of Air and Water Through
Soils." Journal of Agricultural Science, 4 (1911),1-24.
[7] Fok, Y.S. "Evolution of Algebraic Infiltration Equations." Proc. of the Int. Conf. on Infiltration
Development and Application, Univ. of Hawaii, Jan. 6-9 (1987).
[8] Bouwer, H. "Infiltration of Water into Non Uniform Soils." ASCE, 95, IR4 (1969), 451-456.
[9] Childs, E.C. and Bybordi M. "The Vertical Movement of Water in a Stratified Porous Material, 1.
Infiltration." Water Resources Research, 5, No. 2 (1969), 446-451.
[10] Hillel, D. and Gardner, W. R. "Transient Infiltration into Crust-Topped Profiles." Soil Science, 109
(1970), 410-416.
[11] Swartzendruber, D.S. "Infiltration of Constant-Flux Rainfall into Soil As Analyzed by the
Approach of Green and Ampt." Soil Science, 117 (1974), 272-281.
[12] Chu, S.T. "Infiltration During Unsteady Rain." Water Resources Research, 14, No. 3 (1978), 461-
466.
[13] Morel-Seytoux, H.J. Conjunctive Use of Surface and Ground Waters, in Asano, T. (ed.),Artificial
Recharge of Groundwater, Stonham, U .S.A.: Butterworth Publishers, 1985.
[14] Morel-Seytoux, H.J. Personal Communications. April (1987).
[15] Morel-Seytoux, H.J. and Khanji, J. "Derivation of An Equation of Infiltration." Water Resources
Research. 10, No. 4 (1974), 795-800.
[16] Brakensiek, D.L. "Estimating the Effective Capillary Pressure in the Green and Ampt Infiltration
Equation." Water Resources Research, 13, No. 3 (1977), 680-682.