Content uploaded by Santosh Dhungana
Author content
All content in this area was uploaded by Santosh Dhungana on Nov 30, 2017
Content may be subject to copyright.
www.cafetinnova.org
Indexed in
Scopus Compendex and Geobase Elsevier,
Geo-Ref Information Services-USA, List B of Scientific
Journals, Poland, Directory of Research Journals
ISSN 0974-5904, Volume 09, No. 04
August 2016, P.P.1376-1382
#02090404 Copyright ©2016 CAFET-INNOVA TECHNICAL SOCIETY. All rights reserved.
Assessment of Groundwater in Ghataprabha Sub-Basin Using
Visual MODFLOW Flex
NAGRAJ S PATIL1, SANTOSH DHUNGANA1 AND B K PURANDARA2
1Department of Water and Land Management, VTU, Belagavi, Karnataka, India
2National Institute of Hydrology, Hard Rock Regional Centre, Belagavi, Karnataka, India
Email: nagrajspatil@yahoo.com, nspatil@vtu.ac.in
Abstract: In the present study, an attempt is made to simulate the current groundwater level in the Ghataprabha
sub-basin (Krishna basin) in southern India. For this purpose, a single layered aquifer was conceptualized using
Visual MODFLOW Flex ver.14.2. Model calibration was carried out using Parameter Estimation (PEST), with
R2, RMSE and NRMSE as model evaluation criteria. Model behaved well with the value of 0.99, 6 m and
2.13% for R2, RMSE and NRMSE respectively. Groundwater modelling results showed that there is an increase
of 1.46 m of groundwater level in 5 years. Incorporating the increasing trend of groundwater level in planning
water resource projects would be fruitful.
Keywords: Groundwater, Ghataprabha, PEST, Visual MODFLOW FLEX
1. Introduction
Groundwater is finite yet most valuable natural
resources for human survival, economic development
and ecological diversity. Due to its several inherent
qualities such as consistent temperature, widespread
and continuous availability, excellent natural quality,
limited vulnerability, low development cost and
drought reliability, it has become an important and
dependable source of water supplies in all climatic
regions including both urban and rural areas of
developed and developing countries [1] [2].
However, with the ever growing population,
industrialization, urbanization, and technological
advancement in agriculture, most of the hard rock
aquifers in India may have already been stressed [3].
Ghataprabha, a sub basin of mighty Krishna River
basin in peninsular India, has been facing a severe
water shortage problem for both irrigation and
domestic purposes over the past few years [4]. Every
year in summer most surface water sources dry up,
causing serious water shortages for both domestic and
irrigation purposes. In addition, because of the
capricious nature of the south-west monsoon in India,
the availability of surface water cannot be ensured in
the right quantity at the required time. Hence, the
majority of the irrigated area in the Ghataprabha basin
is being cultivated with the help of groundwater
acquired from dug wells and tube wells. However, the
unrestricted excessive pumping of groundwater has
resulted in groundwater lowering in some parts of the
study area [2].
As we cannot see into the subsurface formation, we
need a tool that could provide insight into the
complex system behavior; this is where groundwater
models come into play. The findings of modeling
would be very helpful to develop safe exploitation and
management policies. However, modeling in hard
rock aquifer is difficult as it is constrained to its
prevailing heterogeneity [5] and data availability.
In this study, Visual MODFLOW Flex 14.2
developed by Schlumberger Water Services, was as
an interface to MODFLOW, a three dimensional
numerical engine. MODFLOW uses finite difference
method to describe the movement of groundwater.
2. Study Area
2.1 Introduction
Ghataprabha sub-basin lies approximately between
latitude 15° 45' and 16° 25' N and longitude
74° 00' and 75° 55' E. Total catchment area of the
sub-basin is 8829 km2, out of which 6815.988 km2
(77.2%) lies in Karnataka and rest 2013.012 km2
(22.8%) falls under Maharashtra. Most of the sub-
basin area is flat to gently undulating except for
isolated hillocks and valleys [6].Geographical location
and talukas associated with the study area has been
displayed in figure 1.
Figure 1: Ghataprabha Sub-basin
Assessment of Groundwater in Ghataprabha Sub-Basin Using Visual
MODFLOW Flex
International Journal of Earth Sciences and Engineering
ISSN 0974-5904, Vol. 09, No. 04, August, 2016, pp. 1376-1382
1377
2.2 Climate
The climate of sub basin is marked by a hot summer
and a mild winter. Summer usually commences in
mid of February and ends on May, April being the
hottest month with an average daily maximum
temperature of 37.5⁰ C and minimum of 19.5⁰ C.
Winter records the mean daily maximum and
minimum temperature of 29.3⁰C and 19.3⁰C
respectively. Average annual rainfall in the study is
650 mm. Southwest monsoon contributes about 65%
of total precipitation, remaining 35% being
contributed from northeast monsoon. Rainfall shows
spatio-temporal variation. The humidity is high during
the monsoon period accounting for 85% and low
during non-monsoon accounting for about 30%.
2.3 Land Use, Soil and Agriculture
According to National Water Development Agency
(NWDA) report, 1991 [7], in Ghataprabha sub-basin,
agriculture which covers 63.7% of area is the largest
consumer of water in Ghataprabha basin. The
dominating soil types are black soil, blend of black,
red soil and lateritic soil. Black soil occurs in shallow
depth (25-30 cm) with moderately well drain and high
permeability properties. Coarse black soil occurs very
deep (100-150 cm) and covers low lying areas, with
poor drainage and low permeability. Major crops
grown in the basin under rain fed condition are jowar,
wheat, cotton, groundnuts, tobacco, chilies, wheat,
pulses, etc. and under irrigation schemes are
sugarcane, paddy, wheat, maize, tobacco, turmeric,
vegetables etc.
2.4 Hydrogeology
Geologically, the area is underlain by rocks of
Archaean crystallines to recent
alluvium. Groundwater in the study area occurs in
weathered to semi weathered and fractured hard rock,
under unconfined to semi-confined state.
Groundwater, in this region, is usually found to be
tapped by dug wells, dug cum bore wells and bore
wells.
A hard rock aquifer usually extends the first tens of
meters from ground level [8]. In this case, a shallow
unconfined aquifer extends 30 m from the top. The
fractured aquifer below the shallow zone and extend
down to 80 m and beyond [9]. In the canal command
area, the depth to water level varied between 2 to 8
mbgl during post -monsoon period. The entire area
indicates a rise in water level during the period of
canal operation [10].
3. MODFLOW
Any device that represents an approximation of a field
condition is a model [11] [12]. Groundwater flow
model solves the distribution of head under certain
assumptions and provided boundary conditions.
MODFLOW can simulate confined, leaky
confined and unconfined aquifers and only
simulates saturated flow in a porous medium with
uniform temperature and density [13]. Visual
MODFLOW Flex is a graphical user interface (GUI)
with MODFLOW as a 3-D numerical model, for
which the governing equation is
KXX, Kyy and Kzz are the hydraulic conductivity along
the x, y, and z coordinate axes, (Lt-l); h is the
potentiometric head (L); W is a volumetric flux per
unit volume and represents sources and/or sinks of
water (t- l); Ss, is the specific storage of the porous
material (L-l); and t is time (t). In general, SS, KXX, Kyy,
and KZZ may be functions of space (SS= Ss(x,y,z), Kxx=
KXX(x,y,z), etc.) and W may be a function of space
and time (W = W(x,y,z,t)). Equation 1 describes
ground-water flow under non-equilibrium conditions
in a heterogeneous and anisotropic medium and
provided the principal axes of hydraulic conductivity
are aligned with the coordinate directions.
4. Methodology
Figure 2 shows the flow diagram form groundwater
modeling using MODFLOW.
Figure 2: Work flow diagram form groundwater
modeling using MODFLOW
4.1 MODFLOW Setup
Setting up of a model has been carried out in two
steps, data processing and data assembly.
4.2 Data Requirement and Processing
Various data are required to carry out groundwater
modeling. The data such as Precipitation and
Recharge, Hydraulic Conductivity, Observation wells
and River stage and Discharge aren’t available in a
NAGRAJ S PATIL, SANTOSH DHUNGANA AND B K PURANDARA
International Journal of Earth Sciences and Engineering
ISSN 0974-5904, Vol. 09, No. 04, August, 2016, pp. 1376-1382
1378
single department and were collected from various
sources. Data that have been used to carry out the
modeling are described briefly.
4.2.1 Precipitation and Recharge
Precipitation data for the Belgaum district was
obtained from statistical department, Bangalore
whereas, for regions coming under Maharashtra state
and for Bagalkot district precipitation value was
directly adopted from the literature [6]. In hard rock
area of southern India, approximately 12- 37% of
average annual precipitation contributes to
groundwater recharge [3] and 3 -13 % in basaltic
terrain. According to NIH report [6], the attempted to
model groundwater level in Ghataprabha sub-basin
whereby they have considered 15% of average annual
precipitation as groundwater recharge, same was
adopted for this study.
Figure 3 represents the monthly variation in
precipitation and groundwater recharge in the study
area from year 2008 to 2012.
Figure 3: Average monthly precipitation and
groundwater recharge in the study area from the year
2008 -2012
4.2.2 Hydraulic Conductivity
Data on hydraulic conductivity has been derived from
the CGWB report on Karnataka state. Taluka wise
bore wells havebeen drilled by CGWB to study the
hydrogeological formations. CGWB carried out
various pumping test to calculate the transmissivity of
the geological formations underlying the area.
Transmissivity ranges between 0.22 m2/day to 2220
m2/day in the basin. Visual MODFLOW Flex
requires the conductivity values as an input parameter
which was calculated dividing transmissivity with the
depth of drill. However, specific yield and storage
coefficient has not been registered in the Karnataka
CGWB report. Specific yield and storage coefficient
was adopted from NIH report [6]. Figure 4 displays
the location details of bore wells in the study area.
4.2.3 Observation wells
CGWB, Bangalore has been collecting taluka wise
groundwater level data at different period for the
entire Karnataka state. Groundwater level is recorded
in mbgl (meters below ground level). For the present
study, groundwater level data from 2008 to 2013 were
taken for the talukas coming under Ghatapraha sub-
basin. Observation wells falling under the study area
were clipped using ArcGIS ver. 10.1. Data on
observation wells then was transformed into the
required format by Visual MODFLOW Flex. Figure 5
displays the distribution of observation wells in the
study area.
Figure 4: Bore well details
Figure 5: Location of Observation wells in the study
area
4.2.4 River stage and Discharge
Visual MODFLOW Flex software, used in this study,
has the specific requirement of River or stream
properties to assess the river aquifer interaction. River
stage, bottom, bed thickness and its hydraulic
conductivities are the major requirements. River
discharge and stage data were obtained from Water
Assessment of Groundwater in Ghataprabha Sub-Basin Using Visual
MODFLOW Flex
International Journal of Earth Sciences and Engineering
ISSN 0974-5904, Vol. 09, No. 04, August, 2016, pp. 1376-1382
1379
Resource Information System (WRIS ver. 4.0). The
obtained river stage data were analyzed and on an
average of 3 m river stage was taken. River width was
obtained from DEM processing under ArcGIS
environment. River bed thickness of 0.6 m has been
adopted.
4.3 Groundwater Modeling
Groundwater models are, by definition, a
simplification of complex reality which have been
extensively used to address the groundwater problems
and to support the decision making [14]. Hence, it is
used to make predictions about the subsurface
system’s response to stresses. Modeling further helps
to increase our understanding on the hydrological
system of the catchment. In this case, modeling has
been carried as a single layered unconfined model
assumed based on the literature, that the hard rock
aquifer extends first few tens of meters and
groundwater occurs in unconfined to semi confined
condition [15].
To carry out this study, a single layered unconfined
aquifer was conceptualized with appropriate boundary
conditions.
4.3.1 Boundary Condition
Boundary conditions are the key components in
conceptualizing the groundwater flow system or
model [16] [17]. Various boundary conditions used in
this study are described briefly.
River Boundary Condition
Visual MODFLOW Flex uses the river package to
incorporate streams and rivers into the groundwater
model. Surface water bodies such as stream and rivers
may act as recharging and discharging boundary
condition based on the gradient between the surface
water and groundwater. Interaction between surface
water and groundwater was simulated by assigning
stage of water in the stream, stream bed thickness and
conductivity values of streambed. Figure 6 describes
the stream boundary condition in MODFLOW. Blue
region in the figure 6 indicates the active cells where
river aquifer interaction takes place.
Figure 6: Stream boundary condition
Recharge
MODFLOW recharge package was used to
incorporate recharge into groundwater model.
Recharge values for the single layered is tabulated in
table 1. Recharge in this case was assigned taluka
wise as 15% of the precipitation.
Table 1: Taluka wise precipitation and corresponding
recharge for the year 2008
Taluka
Year
Precipitation
Recharge
Belgaum
2008
1297.88
194.68
Gokak
2008
520.04
78.00
Hukkeri
2008
687.23
103.08
Sayandatti
2008
525.88
78.88
Bailhongal
2008
828.43
124.26
Chikkodi
2008
732.38
109.85
Raibag
2008
437.66
65.64
Ramdurga
2008
608.06
91.20
Chandgad
2008
2814.00
422.10
Gadinglaz
2008
644.00
96.60
Bagalkot
2008
456.00
68.40
Bigli
2008
395.00
59.25
Mudhol
2008
334.00
50.10
4.4 Groundwater Calibration and Validation
Calibration refers to the fine tuning of the model
parameters such that model simulated results properly
matches the field conditions. So, for this purpose, it is
necessary that field conditions should be properly
characterized. Improper characterization may results
in a model, which is not representative of real field
conditions [18]. Parameter ESTimation (PEST) was to
calibrate the model in the current study.
Doherty and Johnston in 2003 [19] developed the
PEST, which is a non-linear parameter estimation and
optimization package. It offers model independent
optimization routines. PEST uses Levenberg-
Marquardt algorithm (i.e. gradient-based
methodology) to search for the optimal solution. The
objective in case of groundwater model calibration is
to minimize the sum of squared residuals. Residual
are differences between observed and simulated
groundwater heads.
4.5 Sensitivity Analysis
Prior to calibration, it is necessary that sensitivity
analysis is performed. It helps to understand the effect
of parameters on the model performance. In this
study, PEST was used to perform sensitivity analysis.
Figure 7 describes the sensitivity of parameters. It is
noticed that the specific yield is highly sensitive
parameter.
After the sensitivity analysis, model calibration was
performed. Initial value of the objective function was
656286.00. The objective function is the sum of
square of the residuals (i.e. sum of squares of
simulated- observed head). Based on the parameter
alteration the value of the objective function was
lowered to 6823.8. Values of objective functions over
iterations are tabulated in table 2 and displayed in
graphical form as shown in figure 8.
NAGRAJ S PATIL, SANTOSH DHUNGANA AND B K PURANDARA
International Journal of Earth Sciences and Engineering
ISSN 0974-5904, Vol. 09, No. 04, August, 2016, pp. 1376-1382
1380
Figure 7: Results of Sensitivity analysis
Table 2: Showing Iteration vs. Phi
Figure 8: Plot of objective function values vs.
iteration
4.6 After Calibration
After optimization of parameters, model was once
again run with the optimized parameters which
yielded a very good result. Observed and simulated
head shows a near perfect correlation with R2 of 0.99.
Comparison between observed and simulated head is
displayed in figure 9 by line graph.
Figure 9: Line graph for observed vs. simulated
heads after calibration
4.7 Evaluation of Calibration
Anderson and Woessner [18] suggest evaluating the
results of calibration both quantitatively and
qualitatively. In this study, three methods of error
quantification techniques (mean error, root mean
square error, and normalized root mean square) and
coefficient of linear regression were used to evaluate
the results of calibration. Table 3 describes the values
of model evaluation criterion before and after
calibration. A drastic change in model performance
can be seen before and after calibration.
Table 3: Model evaluation criteria and their
respective values before and after calibration
Evaluation
Criterion
Before
Calibration
After
Calibration
Minimum residual
(m)
0.16
-0.0083
Maximum Residual
(m)
156.38
-15.4
Residual mean (m)
22.88
0.43
RMSE (m)
45.8
6
NRMSE (%)
16.17
2.13
Regression
Coefficient (R2)
0.65
0.99
5. Results and Discussions
5.1 Groundwater Modeling Results
Figure 10 - 12 represents the spatial distribution of
simulated hydraulic head over different period.
Figure 10: Distribution of Hydraulic head (Time of
simulation: 5 years)
Assessment of Groundwater in Ghataprabha Sub-Basin Using Visual
MODFLOW Flex
International Journal of Earth Sciences and Engineering
ISSN 0974-5904, Vol. 09, No. 04, August, 2016, pp. 1376-1382
1381
Figure 11: Distribution of Hydraulic head (Time of
simulation: 1 year)
Figure 12: Distribution of Hydraulic head (Time of
simulation: 0 days)
Spatial groundwater head distribution profile indicates
the movement of groundwater along the topographic
profile from higher elevation to lower elevation. In
the present, study only the effect of recharge on the
groundwater level with stream as the only source of
groundwater discharge is highlighted. Groundwater
level profile for different period of simulation has
been shown in figure 10-12. The simulated head
indicates the increasing water level, with a head
increase of 0.46m for the first 1 year of simulation
and approximately 1.46 m for the simulation period of
5 years. Model simulation complies with the observed
groundwater level trend. Table 4 shows the
groundwater balance rate under transient case from
2008 to 2014, where recharge exceeds the discharge
rate indicating increase in groundwater level.
Table 4: Daily groundwater balance rate under
transient condition for the month of October
Year
Rates
[m^3/day](IN)
Rates
[m^3/day](OUT)
In- out
1/10/20088
1066443.88
1066523
78.75
1/10/2009
1061585
1061862
276.75
1/10/2010
1057806.38
1057997
190.37
1/10/2011
1054671.25
1054771
100
1/10/2012
1051778.63
1051866
87.62
1/10/2013
1049150.25
1049172
21.75
1/10/2014
1046766.13
1046849
83.31
Figure 13 shows the daily groundwater balance rate
from year 2008 to 2012.
Figure 13: Daily groundwater balance in the study
region from year 2008 – 2014
6. Conclusion
Groundwater modeling showed an increase in
groundwater level in the study area. Groundwater
model was calibrated for the period of 2008-2013. R2,
NRMSE and RMSE were used to evaluate the model
performance. Model behaved well with R2, RMS and
NRMS for 0.99, 2.13 % and 6m respectively.
Increases in groundwater level possess a severe threat
of creating waterlogging condition and groundwater
flooding in low lying areas. Incorporating these
results in planning the water resource in the study area
would be advantageous.
References
[1] Todd, D.K. and Mays, L.W, Groundwater
Hydrology, 3rd edn, New York: John Wiley &
Sons (2005).
[2] Hutti, B., and Nijagunappa, R, Identification of
Groundwater Potential Zone using Geo
informatics in Ghataprabha basin, north
Karnataka, India. Universal Journal of
Environmental Research and Technology, 2011,
pp. 500-514.
[3] Varalakshmi, V., Rao, B., Surinaidu, L.,
Tejaswini, M., 2014. Groundwater flow modeling
of a Hard Rock Aquifer – a case study. J.
Hydrolo.Eng. 10.1061 (ASCE) HE. 1943-
5584.000627. v.19 (5), 2014, pp: 877–886.
[4] (GOK, 2008). Government of Karnataka (2008),
Revised Water planning of Ghataprabha
Subbasin.
[5] Nico, G., and David, D. Methods of Karst
hydrogeology, Taylor and Francis, London, 2007,
246.
[6] Groundwater Modelling in Ghataprabha sub
basin of Krishna river basin. Roorkee: National
Institute of Hydrology, (1996-97).
[7] NWDA, Technical Study NO.17, Water Balance
Study of the Ghataprabha Sub-basin of the
Krishna Basin. , 1991.
[8] Detay, M., Poyet, P., Emsellem, Y., Bernardi, A.,
and Aubrac, G. Development of the Saprolite
reservoir and its state of saturation: Influence on
NAGRAJ S PATIL, SANTOSH DHUNGANA AND B K PURANDARA
International Journal of Earth Sciences and Engineering
ISSN 0974-5904, Vol. 09, No. 04, August, 2016, pp. 1376-1382
1382
the hydrodynamic characteristics of drillings in
crystalline basement. C. R. Acad. Sci. Paris II,
309, 1989, pp. 429–436.
[9] Central Groundwater Board. Groundwater
information Booklet, Belgaum, District,
Karnataka, 2004.
[10] Central Groundwater Board, India, 2014,
<http://cgwb.gov.in/conjuctive_use.html>
accessed on: 07/10/2014.
[11] Reilly, T. E. System and boundary
conceptualization in groundwater flow
simulation.Chapter B8, Techniques of water-
resources investigations, Book 3, U.S. Geological
Survey, Denver, CO, 26, 2001
[12] Anderson, M. P., & Woessner, W. W. Applied
groundwater modeling: simulation of flow and
advective transport. San Diego Academic
Press.,1992
[13] Fetter, C. W. Applied hydrogeology Visual
Modflow, Flownet and Aqtesolv student version
software on CD – ROM (Fourth edition Ed.).
Upper Saddle River: PrenticeHall.,2001.
[14] Kumar, C. P, Numerical modeling of
groundwater flow using MODFLOW.
International Journal of Science 2(4), 2013, pp.
86-92.
[15] Detay, M., Poyet, P., Emsellem, Y., Bernardi, A.,
and Aubrac, G. Development of the Saprolite
reservoir and its state of saturation: Influence on
the hydrodynamic characteristics of drillings in
crystalline basement. C. R. Acad. Sci. Paris II,
309, 1989, pp. 429–436.
[16] Franke, O. L., Reilly, T. E., and Bennett, G. D.
Definition of boundary and initial conditions in
the analysis of saturated ground-water flow
systems—an introduction. Techniques of Water-
Resources Investigations 3-B5, U.S. Geological
Survey, Denver, CO, 15, 1987.
[17] Reilly, T. E. System and boundary
conceptualization in groundwater flow
simulation. Chapter B8, Techniques of water-
resources investigations, Book 3, U.S. Geological
Survey, Denver, CO, 26, 2001.
[18] Anderson, M. P., & Woessner, W. W. Applied
groundwater modeling: simulation of flow and
advective transport. San Diego Academic Press,
1992.
[19] Doherty, J., and Johnston, J. M. Methodologies
for calibration and predictive analysis of a
watershed model, J. Am. Water Resour. Assoc.,
39, 2003, pp. 251–265.
