432 JOURNAL

M.GEOLOGICAL
I. ADHAM AND SOCIETY
OTHERS OF INDIA
Vol.75, February 2010, pp.432-438

Study on Groundwater Recharge Potentiality of Barind Tract,

Rajshahi District, Bangladesh Using GIS and
Remote Sensing Technique

M.I. ADHAM1, C.S. JAHAN1, Q.H. MAZUMDER1, M.M.A. HOSSAIN2 and AL-MAMUNUL HAQUE1
1
Department of Geology and Mining, University of Rajshahi, Bangladesh
2
GIS Division, Center for Environmental and Geographical Information Services (CEGIS), Dhaka, Bangladesh

Abstract: The groundwater recharge potentiality in Barind Tract in Rajshahi district, Northwest Bangladesh was studied
based on Geographical Information System (GIS) and Remote Sensing technique. In this connection satellite images
(Landsat 7 ETM and SPOT) and aerial photos were subjected to several treatment processes using softwares like ERDAS
Imagine and ESRI’s Arc View. Thus various thematic maps have been prepared for drainage density, lineaments, lithology
and land cover/use that allowed deciding their interactive effect. In the present study, the degree of effect was determined
for each factor to assess the total groundwater recharge potentiality for two categories (moderate to low). The resultant
map shows that 85% of the area has low, and rest has moderate groundwater recharge potentiality. Finally only 8.6% of
the total average annual precipitated water (1685mm) percolates into subsurface and ultimately contributes to recharge
the groundwater.

Keywords: GIS, Remote sensing, Recharge potentiality, Barind tract, Bangladesh.

INTRODUCTION and undulatory dome shaped landmass of Pleistocene age,

The remote sensing (RS) study along with geographic
information system (GIS) for groundwater recharge
potentiality reflects hidden hydrogeologic characteristics
and deals with indicative elements at the surface such as,
lineaments and drainage frequency and density, lithologic
character, land cover/land use etc. It provides a better
estimate and qualitative assessment of the recharge potential.
The drainage number (frequency) has the strongest
relationship with the recharge property (Shaban, 2003);
lineaments are closely related to groundwater occurrence;
and surface and subsurface lithological characteristics
and land cover/land use play important role for recharge
process of groundwater. So the aims of the present study
are to estimate qualitatively the recharge of water into the
subsurface media using GIS and RS technique, and to
produce a map showing different zones of recharge potential
of the Barind Tract, especially in its southeastern part
covering Tanore and Godagari Upazillas (sub district) of
Rajshahi district covering 745 km 2 in the northwest
Bangladesh (Fig.1).
PHYSIOGRAPHY AND GEOLOGY
The Barind Tract - north-south elongated, but dissected
edged by river valleys with flat flood plain area running
parallel to dome, sharply demarcate the topography with an
elevation ranging from 47 m above mean sea level (AMSL)
in its central part to 11 m AMSL in the southeast in the
flood plains. The study area has agro-based fertile land, free
from flood inundation and so contributes effectively to the
ongoing drive for achievement of country’s self-sufficiency
in food. But due to harsh climatic condition and
characteristic land pattern, the area suffers from the lack of
both surface and groundwater.
The study area is bounded by the rivers Ganges (Padma)
in the south, Mohananda in the west and Sib in the east. The
Sib River flows from north to south, and their tributaries
form the drainage network of Tanore. The tributaries of Sib
River, and partly of Mohananda River mainly drain Godagari
Upazilla.
The Tanore fault zone along the eastern edge of the
Barind Tract is still active and large scale earthquake in 1952
created springs and fountains of water in the western edge
of Tanore beel (low-lying area). Small streams flowing
eastward from the central Barind Tract suggest that the fault
has a right-lateral component.
Morphostratigraphically the area is subdivided into three
geological units as: (1) Barind clay residuum - overlies and

0016-7622/2010-75-2-432/$ 1.00 © GEOL. SOC. INDIA

JOUR.GEOL.SOC.INDIA, VOL.75, FEB. 2010
 GROUNDWATER RECHARGE POTENTIALITY OF BARIND TRACT, RAJSHAHI DISTRICT, BANGLADESH
Fig.1. Location map of the study area.
developed on Pleistocene alluvium; (2) Holocene Ganges
flood-plain alluvium; (3) Active channel deposits of the
Ganges and major distributaries (modern alluvium). The
study area comprises Barind land (78.4%), Ganges flood
plain (8.5%) and Tista flood plain (4.8%), with four types
of lithology such as Alluvial sand, Alluvial silt, Barind clay
residuum and Marsh clay and peat. Some units lack distinct
sedimentological characteristics, or post-depositional
processes such as ploughing, bioturbation, and weathering
have modified the characteristics (Alam et al. 1990).
Agricultural land with homestead, pond and forest
occupy most part of the area. Here the agricultural practice
varies with the land level, soil type and water availability,
and changes gradually with invention of different varieties
of crops that is also related to the recharge processes. Another
factor that influences recharge potentiality is bund that store
water for few days in rainy season and ultimately contribute
to the water percolation.
METHODOLOGY
In the present study different methods have been applied
for different factors in various steps that influence the
recharge rate based on RS along with the topographic and
JOUR.GEOL.SOC.INDIA, VOL.75, FEB. 2010
433
geological maps (1:50000 scale), and previously collected
data. The interpreted and resultant information are analyzed
in spatial data and each factor is studied independently, thus
four major categories are demarcated to arrive at qualitative
recharge rates. Different methods are described for different
factors as:
Drainage
A combination of field data, topographic maps, aerial
photographs, and satellite images are used for analysis of
the drainage network and all algorithms use digital elevation
models (DEMs) as basic data (Chorowicz et al. 1992). Thus
a drainage map is prepared manually in order to calculate
the drainage frequencies that involve more subjectivity
than the automated ones (Jenson and Domingue 1988;
Tribe 1991; Ichoku et al. 1996; Martinez-Casasnovas and
Stuiver 1998; Abdullah et al. 2002). For morphometric
characteristics, various parameters are computed according
to laws of Horton (1945) and Strahler (1952), and for
analysis of the drainage basins mainly topographic sheet
and SPOT image have been used. The drainage system has
been indexed as 1st to 3rd orders where the main channel of
the rivers represents the highest order. The drainage
characteristics represent dendrite to sub- dendrite pattern,
434 M. I. ADHAM AND OTHERS

with minor sinuosity at places indicating homogeneous
nature of surface bedrocks and lack of structural control.
The stream order, stream number, bifurcation ratio, stream
density are the linear properties of a drainage basin which
have been analyzed in the present study. Here the higher 1st
order drainage density is related to surface topography i.e.
elevated Barind Tract, characterized by less infiltration due
to clayey and semi to impermeable Barind clay with
excessive surface runoff.
In the present study area, stream frequency values
decrease with increasing stream order, which is an indicator
of high slope and lack of facilities for the development of
higher order streams due to bed lithology and tectonic
condition (Melton, 1958). Accordingly the counting of the
tributaries followed the stream-order manner.
The methodology of “sliding windows” is used to
produce a drainage frequency map prepared by dividing the
area into grids from which a frame is produced and each
obtained value is plotted in the middle of the frame.
Therefore, for each four neighboring frames, the average
value is again calculated, and then plotted again in the
midpoint of the “conjunction” resultant frame. To calculate
the frequency densities of drainages the ‘Sliding Window’
method using a frame of specific area is used which is shown
in Fig.2. The sliding windows are moved over the area for
getting the drainage frequency values. Then the average
frequency values of the area are processed by the Arc view
GIS. For evaluating the drainage frequency density and
their distribution, a contour map is prepared with plotted
values. The drainage frequency value of the area ranges
from 1 to 7.25 and the map shown in Fig.3 indicates high-
moderate, medium and low drainage frequency.
Lineaments
The linear geological features of the study area are traced
from the satellite images of Landsat 7 ETM (30×30 m
resolution) and for identification ERDAS, Imagine software
is used. The single and multi-band enhancement is carried
out by interrelating each of the bands as one set. Thermal
band 6 (120×120 m resolution) is used for additional
interpretation as it provides information about the wet
horizons.
Major lineaments as linear signatures are identified from
the scenes of LANDSAT images (MSS) on different scales,
SPOT images, and black and white photographs. Thus the
lineament map is produced and its frequency is calculated
after Greenbaum (1985). To achieve optimum interpretation
with reference to frequency, the “sliding window” method
is applied and the number of lineaments in each cadre was
counted. The frequency value ranges from 0 to 1 evaluating
the lineament frequency and their distribution. Thus a
lineament frequency contour map is prepared (Fig. 4) and a
synoptic classification is made indicating moderate-low, low
and very low frequency value.
Lithology
The geological map (1:50000 scale) is used as base map
to bring out lithological distribution and with the application
of remote sensing, additional information is obtained from
Landsat 7 ETM satellite images. Thus, visual tracing of many
lithological patches and changes of lithilogical boundary
are considered. With the help of GIS technique (especially
Arc view program), the area is divided into grids and
assigned the percolation values of the lithological units in
the field of attribute. Thus based on percolation rates of soil
type, the study area is classified into five categories
(Table 1) (IWM, 2006). A map was then prepared to show
zones of different lithological character with respect to water
recharge (Fig.5).
Land Cover/Land Use
Based on satellite images and principal zones of similar
recharge responses, only two zones of land cover/land use
are delineated in the study area. The visual discrimination
Table 1. Estimated percolation rates for the major soil types of the
study area

Lithology Soil type Soil texture Percolation

Rate (mm/day)

Alluvial Sand New Ganges Silty loam 4.0

Alluvial Silt Old Ganges Silty loam 3.0
Lower Barind Silty clay 2.0
Barind Clay Middle Barind Clay loam 2.0
Upper Barind Silty clay loam 2.0
Marsh Clay Teesta Flood-
Clay 3.0
and Peat plain
Fig.2. “Sliding Windows” method. (Source: IWM, 2006)

JOUR.GEOL.SOC.INDIA, VOL.75, FEB. 2010

 GROUNDWATER RECHARGE POTENTIALITY OF BARIND TRACT, RAJSHAHI DISTRICT, BANGLADESH 435

Figs.3-6. (3) Drainage frequency map. (4) Lineament frequency map. (5) Lithological map classification based on water percolation
rate. (6) Land cover/land use classification based on water percolation rate.

of colour and resolution is established depending on
optical differentiation of satellite images of merged Landsat
7 ETM (2000) pan sharpen and IRS-1, and thus several
polygons are obtained. Within each zone, different units
are plotted, but all have relatively the same degree of effect
on the recharge potential rate (Su, 2000; Shaban, 2003).
The selection of various units in specific zones lead to a
supervised classification on the used image, thus a land
cover/land use map is produced for different water recharge
zones and is shown in Fig.6.
DATA MANIPULATION AND RESULTS
In the present study, a synoptic classification based on
factors like the lineaments, lithology, drainage and land
cover/land use of different categories has been done which
have different levels of effect, and is expressed in numerical
value ranges depending on the resultant domains, which
provide specific information about the recharge or
infiltration of water. Based on average drainage frequency
value, a three fold numerical classification is established
where segments per 25 km2 is counted. In selected domain,
three class ranges are established where each factor
influences mainly the recharging processes.
The high lineaments frequency indicates very high
recharge potential, while that of low frequency does not
indicate very low recharge potential, the average frequency
values ranging from 1.0 to 0.25 per 25 km2. The lithology
domains work from high to very low, depending upon the
type of sediments like sand, silt and clay. Similarly, the land

JOUR.GEOL.SOC.INDIA, VOL.75, FEB. 2010

436 M. I. ADHAM AND OTHERS

cover/land use map is analyzed based on human settlement Table 2. Categorization of factors affecting recharge potentials in the
study area
and vegetation cover containing different values of
water percolation, the values ranging from medium to Factor Descriptive Proposed
level weight of Domain of effect
moderate-low. Therefore, major descriptive levels are
effect
considered ranging from very high to very low including
High-moderate 6.5 4.43-6.15 (per 25km2)
some interrelated levels, and the weight rating starts from Drainage Moderate 5.0 2.72-4.43
10 points (pts) i.e. very high = 10 pts, high = 8 pts, Low 2.0 1.0-2.72
high-moderate = 6.5 pts, moderate = 5 pts, moderate- Moderate-low 3.5 0.778-1.0 (per 25km2)
low = 3.5 pts, low = 2 pts, and very low = 1 pt. All these Lineament Low 2.0 0.556-0.778
factors are integrated to obtain a recharge potential map. Very low 1.0 0-0.556

Since the factors do not have the same degree of influence
on the recharge potentiality, a weighting approach is
incorporated and the effect of the factors on each other is
presented as schematic sketch (Fig. 7). It reveals that the
lithology was the most influential one having three major
effects, i.e. it has an effect on lineaments, drainage and
land cover/land.
Lineaments
Drainage Land cover/land
use
Lithology
Major effect Minor effect
Fig.7. Schematic sketch showing interactive influence of factors
concerning recharge potential.
Based on relative rates of each influencing factor,
recharge potentiality of the study area is categorized
(Table 2). Now to obtain a comprehensive evaluation of
each factor on recharge potentiality (Khawlie, 1986; Shaban
High 8.0 4 mm/day (Percolation rate)
Lithology Moderate 5.0 3 mm/day (Percolation rate)
Low 2.0 2 mm/day (Percolation rate)
Land cover/ Moderate 5.0 Sparse vegetation
land use Moderate-low 3.5 Human settlement
et al. 2001), the rates and weights are integrated and thus
the total weighting assessment, after rounding off values is
obtained (Table 3). The final map is derived by summation
of the factor and the weight coefficient of the factor.
Figure 8 shows calculation process to get the exact value
of the cell by weighting approach. Maps obtained for each
factor are considered as layers, and overlaying of each
layer with its own weight, in a GIS system, resulted in
different polygons of special characteristics with respect to
the overall recharge potential for the area. After considering
rate assessment, the ESRI’s Arc View software was used to
manipulate the data, through superimposing different layers
of recharge potential. Accordingly, the recharge potential
categories and their qualitative estimation in the study area
have been calculated (Table 4).
Finally, the recharge potential zonation map (Fig. 9) was
prepared with descriptive levels as moderate and low,
occupying areas of 113.34 km2 (15%) and 631.71 km2 (85%)
respectively. However, a quantitative estimation of recharged
water volume (W) to subsurface media in recharge potential

Table 3. Weight evaluation of factors influencing recharge potential capacity

Factor Descriptive Weight (a) Rate (b) Weighted Total Factors on recharge
scale (1-10) (1-3) rating (a×b) 5(a×b) potential capacity (%)
High-moderate 6.5 1.5 10
Drainage Moderate 5.0 1.5 8 21 22
Low 2.0 1.5 3
Moderate-low 3.5 2.0 7
Lineament Low 2.0 2.0 4 13 13
Very low 1.0 2.0 2
High 8.0 3.0 24
Lithology Moderate 5.0 3.0 15 45 47
Low 2.0 3.0 6
Land cover/ Moderate 5.0 2.0 10
17 18
land use Moderate-low 3.5 2.0 7
Grand Total Weight 96 100

JOUR.GEOL.SOC.INDIA, VOL.75, FEB. 2010

 GROUNDWATER RECHARGE POTENTIALITY OF BARIND TRACT, RAJSHAHI DISTRICT, BANGLADESH 437

Table 4. Recharge potential categories and their quantitative estimation

Drainage weight (Dw) ×
drainage weight coefficient (Dc) Recharge potential Estimates according Average Areal extent
category to FAO (1967) (%) (km2)

Drainage Very high 45-50% 47.5 0.0

High 30-35% 32.5 0.0
Lineament weight (Lw) ×
Moderate 10-20% 15.0 113.34
lineament weight coefficient (Lc)
Low 5-10% 7.5 631.71
Very low <5% 2.5 0.0
Lineament

Lithological weight (Liw) ×

lithological weight coefficient (Lic)

Lithology
Land cover weight (LCw) × land cover
weight coefficient (LCc)

Land cover
Final map is calculated by the sum of
these criterion, i.e., Final map = (Dw×
Dc) + (Lw × Lc) + (Liw × Lic) + (LCw ×
LCc)
Final map
Fig.8. Calculation process to get the exact value of the cell by
weighting approach.

zones of the study area is done for respective recharge

potential zones as precipitated volume (P) × recharge ratio
× % of area, where P = 1136.19×106 m3 /year. So
W = 1136.19×106 (0.475×0 + 0.325×0 + 0.15×0.15+
0.075×0.85 + 0.025×0)
= 1136.19×106 (0.15×0.15 + 0.075×0.85)
= 1136.19×10 6×0.08625
= 93.73×106 m3 /year
This means that only 8.6% of precipitated water in the
study area is percolating downward to recharge the
groundwater and the rest is lost either as evapotranspiration
or surface runs-off.
CONCLUSIONS
The present study provides an approach for qualitative
assessment of groundwater recharge potential with the
help of remote sensing and GIS in southeastern part of the
Barind Tract in Rajshahi district, Bangladesh. An integrated
groundwater potential map has been prepared and
categorized on the basis of weightages assigned to
different features such as lithology, lineaments, drainage
Fig.9. Recharge potential zones of the study area.
frequency density, and land cover/land use, which reveals
the following:
1 About 85% has low recharge potentiality and that of
rest 15% has moderate recharge potentiality.
2 Only 8.6% of the total precipitated water (1136.19×106
m3/year) is infiltrating downward to recharge the
groundwater reservoirs, while the rest is lost either as
evapo-transpiration or surface runoffs.
3 The recharge potential categories and their qualitative
estimation reveal optimistic values of recharge potential.
4 Finally it may be concluded that recharge potential map
provides first hand information on groundwater recharge
useful for planning of groundwater exploration.
So it is recommended that such study should be
applied to the whole of Bangladesh on micro-level with
instrumental verification to get more precise information
about surface signature that governs the groundwater
recharge potentiality.

JOUR.GEOL.SOC.INDIA, VOL.75, FEB. 2010

438 M. I. ADHAM AND OTHERS
References
ABDALLAH, C., KHAWLIE, M., SHABAN, A. and BOU KHEIR, R. (2002)
A method comparing between drainage networks extracted
from DEM and conventional approaches on a pilot area in
Lebanon. Proceedings of the International Symposium on
Remote Sensing, Damascus, Syria, 9-12 December.
ALAM, M.K., HOSSAIN, A.K.M.S., KHAN, M.R. and WHITNEY, J.W.
(1990) Geological Map of Bangladesh. Geol. Surv.
Bangladesh, Dhaka, (Scale 1.1,000,000).
CHOROWICZ, J., ICHOKU, V., RIAZANOFF, S., KIM, Y. and CERVELLE,
B. (1992) A combined algorithm for automated drainage
network extraction. Water Resource Res., v.28, pp.1293-
1302.
GREENBAUM, D. (1985) Review of remote sensing applications to
groundwater exploration in basement and regolith. British
Geol. Surv. Report, OD-85(8), 36p.
HORTON, R.E. (1945) Erosion development of streams and their
drainage basins. Hydrol Appl. of Morph., Bull. Geol. Soc.
Amer., v.56, pp.275-370.
ICHOKU, A., MEISEIS, A. and CHOROWICZ, J. (1996) Detection of
drainage channel networks on digital satellite images. Int. Jour.
Remote Sensing, v.17(9), pp.1659-1678.
INSTITUTE OF WATER MODELLING (IWM) (2006) Project Brief for
Deep Tubewell Installation Project in Barind Area, Ground-
water Model Study for Deep Tubewell Installation Project In
Barind Area. Final Report, IWM, Dhaka.
JENSON , S. and DOMINGUE, J. (1988) Extraction topographic
structure from digital elevation data for geographic information
(Received: 18 May 2009; Revised form accepted: 6 August 2009)
system analysis. Photogram Engg. Sens., v.54(11), pp.1593-
1600.
KHAWLIE, M. (1986) Land-use planning for the development of a
disrupted urban center: Beirut, Lebanon. Int. Jour. Dev.
Technol., v.4, pp.267-281.
MARTINEZ-CASASNOVAS, J. and STUIVER, H. (1998) Automated
delineation of drainage networks and elementary catchments
from digital elevation models, ITC J, v.3/4, pp.198-208.
MELTON, M.A. (1958) Geoetric properties of mature drainage
systems and the representation in an E4 phase space. Jour.
Geol., v.66, pp.35-45.
SHABAN, A. (2003) Studying the hydrogeology of Occidental
Lebanon: utilization of remote sensing. Etude de phydro-
geologie du Liban occidental: Utilization de la teledection.
These de doctorat, Universite Bordeaux 1, 202p.
SHABAN, A., KHAWLIE, M., BOU KHEIR, R. and ABDALLAH, C. (2001)
Assessment of road instability along a typical mountainous
road using GIS and aerial photos, Lebanon- eastern
Mediterranean. Bull. Engg. Geol. Env., v.60, pp.93-101.
S TRAHLER , A.N. (1952) Hypsometric analysis of erosional
topography. Bull. Geol. Soc. Amer., v.60, pp.117-1142.
SU, Z. (2000) Remote sensing of landuse and vegetation for
mesoscale hydrological studies. Int. Jour. Remote Sensing,
v.21.
TRIBE, A. (1991) Automated recognition of valley heads from digital
elevation models. Earth Surface Processes Landforms, v.16,
pp.33-49.
JOUR.GEOL.SOC.INDIA, VOL.75, FEB. 2010