Water Resour Manage

DOI 10.1007/s11269-011-9915-x

An Analytical Solution of Boussinesq Equation to Predict

Water Table Fluctuations Due to Time Varying Recharge
and Withdrawal from Multiple Basins, Wells
and Leakage Sites

S. N. Rai & A. Manglik

Received: 10 December 2010 / Accepted: 20 September 2011

# Springer Science+Business Media B.V. 2011

Abstract Recharging and pumping are the integral part of any scheme of ground water
resources development and both processes significantly affect the dynamic behavior of the
aquifer system. Leakage from the aquifer’s base, if present, is other process which affects
the water table variation. Therefore, an accurate estimation of water table fluctuation
induced by recharging, pumping and leakage is pre-requisite to ensure sustainability of
groundwater resources. In the present work an analytical solution of a 2-D linearized
Boussinesq equation is developed to predict water table fluctuations in the presence of time
varying recharge, pumping and leakage from any number of recharge basins, wells and
leakage sites of any dimension for any number of recharge and pumping cycles. The rate of
time varying recharge (or pumping) is approximated by using a series of linear elements of
different lengths and slopes which are dependent on the nature of variation in the recharge
(or pumping) rate. Application of the solution in the prediction of water table fluctuation in
the presence of time varying recharge, pumping and leakage is demonstrated with the help
of a numerical example. These numerical results indicate significant effect of the time
varying recharge/pumping rates and leakage on the water table variation. Such information
is useful for the proper management of groundwater.

Keywords Transient recharge . Pumping . Leakage . Analytical solution . 2D Boussinesq

equation

Abbreviation
H h2
h20
a Kh=S
h Variable water table height

S. N. Rai (*) : A. Manglik

National Geophysical Research Institute, (Council of Scientific & Industrial Research),
Hyderabad 500 007, India
e-mail: snrai@[Link]
 S.N. Rai, A. Manglik

h0 Initial water table height

h Weighted mean of the depth of saturation
K Hydraulic conductivity
S Specific yield
P Sum of recharge and withdrawal rates

1 Introduction

Mathematical models play a key role in assessing the future behavior of a groundwater
system in response to various schemes of ground water resources development such as
artificial recharging and pumping and in selection of an appropriate one out of many
proposed schemes for its sustainable development. Many mathematical models have been
developed to predict the water table fluctuations in response to recharge from basins of
different geometrical shapes (Hantush 1967; Marino 1967; Hunt 1971; Marino 1974; Rao
and Sarma 1981a, b, 1984). Most of the models are based on the assumption of constant
rate of recharge applied continuously. In some studies, intermittently applied constant
recharge has been used (e.g. Bansal and Das 2011). However, the rate of recharge largely
depends on the infiltration rate which is influenced by several factors. The infiltration rate
decreases initially mainly due to dispersion and swelling of soil particles. After some time it
increases owing to displacement of the entrapped air. After attaining a maximum value, it
again decreases owing to clogging of the soil pores (Bear 1979; Detay 1995; Dickenson and
Bachman 1995). Clogging is caused by silt and clay deposition beneath the bottom of the
basin. The rate of recharge follows almost a similar pattern of variation of infiltration rate
with comparatively less intensity and with some time lag due to the time taken by the
infiltrated water to reach to the water table. When the rate of recharge decreases to a
minimum prescribed level, the recharge operation is discontinued for some time and after
drying, cleaning, and if necessary, scrapping of the silty bottom of the basin, recharge rate is
brought back almost to its initial value and the basin is again put back to use for the next
phase of recharge operation. Zomorodi (1991) has demonstrated with the help of field
examples that the solutions based on the assumption of constant rate of recharge are unable
to predict the rise and subsequent decline of the water table which is due to decrease in the
rate of recharge. He suggested that the recharge rate should be treated as variable in time in
order to simulate actual field conditions. Several schemes have been considered to
approximate time varying recharge rate in order to develop predictive ground water flow
models. Rai et al. (1994), and Rai and Singh (1995, 1996) have used exponential function
to approximate one cycle of time varying recharge applied from a single basin and
developed 2-D analytical models to describe water table fluctuation. Yue-zan et al. (2007)
have used a scheme in which the duration of time varying recharge is divided into several
time zones according to the actual variation in the recharge rate. In each time zone, the
variation range of recharge rate should be considered so small that it can be represented by
the constant average value of recharge rate of that particular zone. He referred this recharge
rate approximation as stepped variable scheme. Manglik et al. (1997) have proposed a new
scheme for the approximation of time varying recharge rate. In this scheme time varying
recharge rate is approximated by a series of line elements of different lengths and slopes.
The number, lengths and slopes of the line elements depend on the nature of variation of
recharge rate. Advantage of this approximation scheme is that any complex nature of
recharge can be approximated with more accuracy. However, on regional scale artificial
An Analytical Solution of Boussinesq Equation

recharging and pumping operations are carried out intermittently from more than one site in
accordance to the necessity. Rai and Manglik (2000) and Manglik et al. (2004) have used
the same scheme to approximate time dependent pumping rates by using minus sign before
the function representing time varying recharge rate. Rai et al. (2006) have used the same
scheme for approximating leakage rate along with time varying recharge and pumping rates
and developed an analytical ground water flow model to predict water table fluctuation in
response to time varying recharge, pumping and leakage from any number of recharge
basins, wells and leakage sites of different size for any number of cycle of recharging and
pumping operations. The flow system is characterized with the mixed boundary condition,
i.e. no flow condition at two adjoining boundaries and constant head condition at another
pair of boundaries.
In the present study, we have considered 2-D ground water flow in an unconfined
aquifer system characterized by the constant head boundary condition at all four
boundaries. A boundary of this kind occurs whenever the flow system is adjacent to
an open water body or the boundary is extended to such a large distance so that level
of water table at the boundary remains unaffected from recharging/pumping taking
place within the flow domain. An analytical mathematical model is developed by
solving 2-D Boussinnesq equation to predict water table fluctuation in the presence of
time varying recharge, pumping and leakage from aquifer’s base. Recharging and
pumping are considered for any number of cycles of these operations applied
intermittently from any numbers of recharge basins and/or wells of different size. The
rates of recharge and/or withdrawal are approximated by a set of line elements of
different slopes and length (Manglik et al. 1997).

2 Mathematical Formulation and Solution

The plan view of the aquifer system with a rectangular recharge basin, pumping well and a
leakage site is shown in the Fig. 1. For demonstration purposes only one recharge basin,
two wells and one leakage site is considered. But the solution will be developed for any

Fig. 1 A plan view of

unconfined aquifer with one
recharge basin, one leakage,
and two wells sites
 S.N. Rai, A. Manglik

number of recharge basins, wells and leakage sites of different dimension. The aquifer is
homogeneous and isotropic and is characterized with the constant head boundary condition
from all sides.
Ground water flow in the aquifer system with horizontal base is described by the
following 2-D linearised Boussinesq equation (Hantush 1967):

@2H @2H 2 1 @H
þ 2 þ Pðx; y; tÞ ¼ ð1Þ
@x2 @y K a @t
where H ¼ h2
h20 ;a ¼ Kh=S, and h; h0 ; h; K; S and P are the variable water table height,
initial water table height, weighted mean of the depth of saturation (Marino 1967),
hydraulic conductivity (K), specific yield, and recharge rate, respectively. The solution of
the prediction problem lies in the solution of this equation.
The initial condition is:

H ðx; y; 0Þ ¼ 0 ð2Þ

The boundary conditions are:

H ð0; y; tÞ ¼ H ðA; y; tÞ ¼ 0 0  y  B ð3Þ

H ðx; 0; tÞ ¼ H ðx; B; tÞ ¼ 0 0  x  A ð4Þ

P (x,y,t) is the sum of recharge and withdrawal rates and is represented by
(
XN
Pi ðtÞ for xi1  x  xi2 ; yi1  y  yi2
Pðx; y; tÞ ¼ i¼0 ð5Þ
0 elsewhere

where N is the total number of basins, wells and leakage sites under consideration, Pi(t) is
the time varying recharge (or pumping or leakage) rate for the ith basin (or well or leakage
site) and xi1, xi2, yi1, yi2 are the coordinates of ith basin (or well or leakage site). Pi(t) is
positive for recharge and leakage in to the aquifer and negative for the pumping and
leakage out of the aquifer. Pi(t) is approximated in the following form:
(
rij t þ cij tj  t  tjþ1 ½j ¼ 1; 2;    ; k
1
Pi ðtÞ ¼ ð6Þ
rik t þ cik t  tk ½j ¼ k

in which rij and cij are the slope and intercept of the jth line element of the recharge(or
withdrawal) rate for the ith basin (or well), and k is the number of elements. The solution of
Eq. 1 subject to the given initial and boundary condition is obtained by using the following
extended Fourier sin transform (Rao and Sarma 1981a);
ZB ZA
mpx npy
Fs ½H ðx; y; t Þ ¼ S ðm; n; tÞ ¼ H ðx; y; tÞ sin sin dxdy ð7Þ
A B
0 0
An Analytical Solution of Boussinesq Equation

After applying this transformation to Eq. 1 and using given initial and boundary
conditions, the solution of Eq. 1 in Fourier domain is given by:
" #
X
N X
k
1
S ðm; n; tÞ ¼ gi ðm; nÞ Rij þ Rik ð8Þ
i¼1 j¼1

where
2aAB h mpxi2 npxi1 i
gi ðm; nÞ ¼ 2
cos
cos
Kmnp
h A A
npyi2 npyi1 i
 cos
cos
B B

rij       
Rij ¼ tjþ1 exp
l t
tjþ1
tj exp
l t
tj
l

rij cij        

2
exp
l t
tjþ1
exp
l t
tj
l l

rik
Rik ¼ ½t
tk expf
lðt
tk Þg
l

rik cik

2
½1
expf
lðt
tk Þg
l l

ap2 m2 n2
l¼ þ
4 A2 B2

After applying the following inverse Fourier sin transform to Eq. 8

4 X1 X 1
mpx npy
H ðx; y; tÞ ¼ S ðm; n; tÞ sin sin ð9Þ
AB m¼0 n¼0 A B

gives the following desired solution for the variable water table height, h:

8a X 1 X 1
1 mpx npy
h2 ðx; y; tÞ ¼ h20 þ sin sin
Kp2 m¼1 n¼1 mn A B

N h
X mpxi2 mpxi1 ih npyi2 npyi1 i
cos
cos cos
cos
i¼1
A A B B ð10Þ
" #
X
k
1
Rij þ Rik
j¼1
 S.N. Rai, A. Manglik

2.1 Solution for Constant Recharge Rate—A Special Case

The solution for constant rate of recharge from single rectangular basin given by Rao and
Sarma (1981a) can be obtained by substituting N = 1, j = 1 and r = 0 in to Eq. 10 and is
given by

8ac X 1 X1
1 mpx npy
h2 ðx; y; tÞ ¼ h20 þ 2
sin sin
Kp m¼1 n¼1 mn A B
h mpxi2 mpxi1 ih npyi2 npyi1 i
cos
cos cos
sin ð11Þ
B B B B
½1
expð
ltÞ

This equation is the same as Eq. 13 of Rao and Sarma (1981a). In this case the constant
rate of recharge will be represented by the value of intercept, c, of a single line element with
zero slope.

3 Numerical Results and Discussion

In order to demonstrate the application of Eq. 10 in prediction of the spatio-temporal

variation water table, a synthetic numerical example is considered in which the aquifer
parameters are: A=B=5,000 m, h0 =20 m, K=5 m/d and S=0.28. For demonstration
purposes we consider one recharge basin, two wells and one leakage site. Recharge is
applied through a rectangular basin of 60 m×40 m dimension with center at (2,500 m,
2,500 m). Two cycles of recharge of 20 days duration with 20 days gap is applied (Fig. 2a).

Fig. 2 (a) Time varying recharge

rate from basin, and (b) time
varying pumping rate from the
well
An Analytical Solution of Boussinesq Equation

In each cycle the rate of recharge decreases from 0.8 m/d to 0.7 m/d during first 2 days and
again reaches to its maximum value of 0.9 m/d during the next 2 day period. Thereafter, it
continuously decreases to 0.2 m/d during next 14 days period. After that recharge operation
is discontinued. As a result, the rate of recharge decreases to zero in the next 2 days
(Fig. 2a). Two wells, namely well-1 and well-2 of 10 cm×10 cm dimensions with centers at
(2,400 m, 2,500 m) and (2,500 m, 2,400 m), respectively are considered for pumping. Two
cycles of pumping each of 10 days duration with a gap of 20 days is carried out from both
wells at a rate of 80 m3/d. The first cycle of pumping begins after 10th day from the beginning
of the first cycle of recharge and continues till 20th days, i.e. the last day of the first cycle of
recharging. After a gap of 20 days, the second cycle of pumping starts at 40th day and continues
until 50th day (Fig. 2b). Leakage from the aquifer’s base is considered through an area of
200 m×100 m dimension with center at (2,700 m, 2,740 m). The leakage rate is considered
constant equal to 0.02 m/d. The weighted mean of the depth of saturation, h, is computed by
using successive method of approximations suggested by Hantush (1967).
In order to see the effect of variation in the recharge and pumping rates on the water
table variation in the presence of leakage from the aquifer base, water table heights are
computed using Eq. 10 for t=10, 20 and 40 d along a profile parallel to x-axis at Y=2,500 m
which passes through the centres of well-1 and the recharge basin and are plotted in Fig. 3.
Water table heights are measured from the initial water table height. This figure shows that the
water table attains its maximum height equal to 4.18 m at t=10 d at the centre of the recharge
basin. At this time growth of the water table along the profile is confined between 2,300 m
and 2,650 m. Decline of the water table due to leakage occurs between 2,675 m and 2,900 m
with the maximum value of 0.09 m between 2,725 m and 2,760 m. After 10th day, first cycle
of pumping starts and continues up to 20th d. As a result, water table heights below recharge
basin and its adjoining area decreases with time. At t=20 d the maximum growth below the
centre of recharge basin reduces to 1.65 m and at t=40 days, it further reduces to 0.17 m. The
maximum decline of the water table height below the well-1 at t=20 d is 2.37 m. Till 20 d,
there is no signature of the decline of water table due to leakage. But at t=40 d, the maximum
decline of 0.22 m due to leakage is noticed between 2,650 m and 2,700 m.
Figure 4 illustrates nature of water table variation along a profile at Y=2,400 m which
passes through the centre of the well-2 and the centre of the leakage area. Since this profile
is 80 m away from the outer boundary of the recharge basin, the maximum growth of water
table due to recharge is 0.54 m which is 3.65 m less than the maximum growth of the water
table along the profile at Y=2,500 m. The maximum decline of the water table at well-2 for

Fig. 3 The nature of water

table variation at different times
along a profile at Y=2,500 m
which passes through a well
and recharge basin
 S.N. Rai, A. Manglik

Fig. 4 Nature of water table

variation at different times along
a profile at Y=2,400 m which
passes through a well and the
centre of leakage area

t=20 d is 2.45 m. A continuous decline of the water table is observed above the leakage
site. The maximum magnitude of water table decline above the centre of leakage area at
t=10 d is 0.41 m which increases to 0.70 at t=40 d. A 3-D view of water table variation
consisting of ground water mound below recharge basin, decline of water table in conical
shape at wells and almost flattened nature of water table decline above the leakage area for

Fig. 5 Three dimensional view of growth and decline of water table at t=20th day (top) and a surfacial view
of the same (bottom)
An Analytical Solution of Boussinesq Equation

t=20 d is shown in top of Fig. 5. A plan view of the nature of water table variation is
presented below in Fig. 5.

4 Conclusion

An analytical mathematical model (Eq. 10) is developed to predict the water table
fluctuation in response to time varying recharge, pumping and leakage from multiple
basins, wells and leakage sites of different dimension and for any number of recharge and
pumping operation. The analytical solution for constant recharge from a single basin is
shown as a special case. In this model recharge, pumping and leakage can also be
considered either separately or any combination of these three. Numerical results indicate
that only recharge will raise the water table height to a larger value. But a combined
operation of recharging and pumping from a nearby well will reduce the growth of the
water table. The scheme of approximating time varying recharge/pumping rate by a set of
linear elements of different lengths and slope has advantage of approximating any complex
nature of variation in recharge/pumping rate with more accuracy than any other scheme such as
exponential function and stepped variable. This model can be used for sensitivity analysis of the
effects of variation in the controlling parameters such as duration and intensity of recharge and
pumping, dimension and distribution pattern of the recharge basins and wells etc. Such analysis
is useful for selection of an appropriate scheme of ground water resources development out of
many proposed scheme to achieve the preset objective of sustainable development. This model
can also be used for testing the validity of the numerical models under developing stages before
their applications to the real field problems.

Acknowledgements Thanks are due to Dr. S. Thiagarajan for his help in preparation of the manuscript and
Director, N.G.R.I for according permission to publish this work.

References

Bansal RK, Das SK (2011) Response of an unconfined sloping aquifer to constant recharge and seepage from
the stream of varying water level. Water Resour Manage 25:893–911
Bear J (1979) Hydraulics of groundwater. McGraw-Hill, New York
Detay M (1995) Rational groundwater reservoir management, the role of artificial recharge. In: Johnson AI,
Pyne RDG (eds) Artificial recharge of groundwater II. ACSE, New York, pp 231–240
Dickenson JM, Bachman SB (1995) The optimization of spreading ground operations. In: Johnson AI, Pyne
RDG (eds) Artificial recharge of groundwater II. ACSE, New York, pp 630–639
Hantush MS (1967) Growth and decay of ground water mounds in response to uniform percolation. Water
Resour Res 3:227–234
Hunt BW (1971) Vertical recharge of unconfined aquifer. J Hydraul Div ASCE 96(HY7):1017–1030
Manglik A, Rai SN, Singh RN (1997) Response of an unconfined aquifer induced by time varying recharge
from a rectangular basin. Water Resour Manage 11:185–196
Manglik A, Rai SN, Singh VS (2004) Modelling of aquifer response to time varying recharge and pumping
from multiple basins and wells. J Hydrol 292:23–29
Marino MA (1967) Hele-Shaw model study of the growth and decay of groundwater ridges. J Geophys Res
72:1195–1205
Marino MA (1974) Rise and decline of water table induced by vertical recharge. J Hydrol 23:289–298
Rai SN, Singh RN (1995) Two-dimensional modelling of water table fluctuation in response to localized
transient recharge. J Hydrol 167:167–174
Rai SN, Singh RN (1996) On the prediction of groundwater mound formation due to transient recharge from
a rectangular area. Water Resour Manage 10:189–198
 S.N. Rai, A. Manglik

Rai SN, Manglik A (2000) Water table variation due to timevarying recharge and withdrawal. In: Sililo et al
(eds) Groundwater—past achievement and future challenges. Balkema, Rotterdam, pp 259–262
Rai SN, Manglik A, Singh RN (1994) Water table fluctuation in response to transient recharge from a
rectangular basin. Water Resour Manage 8:1–10
Rai SN, Manglik A, Singh VS (2006) Water table fluctuation owing to time varying recharge, pumping and
leakage. J Hydrol 324:350–358
Rao NH, Sarma PBS (1981a) Recharge from rectangular areas to finite aquifers. J Hydrol 53:269–275
Rao NH, Sarma PBS (1981b) Groundwater recharge from rectangular areas. Ground Water 19:271–274
Rao NH, Sarma PBS (1984) Recharge to aquifers with mixed boundaries. J Hydrol 74:43–51
Yue-zan T, Mei Y, Bing-feng Z (2007) Solution and its application of transient strean/groundwater model
subjected to time-dependent vertical seepage. Appl Math Mech 28(9):1173–1180
Zomorodi K (1991) Evaluation of the response of a water table to a variable recharge. Hydrol Sci J 36:67–78