ISSN 00978078, Water Resources, 2010, Vol. 37, No. 5, pp. 623–637. © Pleiades Publishing, Ltd., 2010.

WATER RESOURCES AND THE REGIME

OF WATER BODIES

Use of Geoinformatics for InterBasin Water Transfer Assessment1

Niladri Guptaa, Petter Pilesjob, and Ben Maathuisc
a
School of Geography, University of Southampton,
+44(0) 7919315113 Southampton, Highfield, Southampton S0171BJ, UK
b G1S Centre, Lund University, +4646222954 Solvetrotan 12, S223, 62, Lund, Sweden
c
International Institute for GeoInformation Science and Earth Observation,
+31534874391 7500 AA, Enschede, the Netherlands
Received July 2, 2009

Abstract—Fresh water availability and demand are unevenly distributed both temporally and geographically.
Furthermore, the availability of fresh water has remained more or less constant, while the demand for clean
water is steadily increasing. With demand surpassing supply, an integrated water resource management
approach is required to ensure even distribution of potable water to all levels of society while protecting the
environment. Interbasin water transfer (IBWT) is an approach being applied in various countries around the
world, with varying environmental and social implications. The Interlinking of Rivers (ILR) scheme is an
example of such a project being planned in India. The research described in this paper was based on the ILR
project and includes an assessment of the IBWT programme in some of the tributaries of the Brahmaputra
and Ganga Rivers in the eastern part of India, covering the district of Jalpaiguri, West Bengal, India. Geoin
formation has been used in association with physical and socioeconomic factors to identify potential dam
and reservoir sites and to delineate the optimal route for canals to transfer water from the Brahmaputra basin
to the Ganga basin for further transportation to the waterdeficient regions of India.

Keywords: environment, interbasin, reservoir sites, canal.

DOI: 10.1134/S0097807810050039

1
In some areas of the world, especially regions with preventing floods in the monsoon season have always
high population density and intense economic activity, been a challenge.
the demand for fresh water has overtaken the supply.
An integrated water resource management approach is The Indian river system, which forms the lifeline of
required to balance environmental, social and eco the country, can be divided into two components: the
nomic issues, rather than the conventional technique Himalayan component, the sources of which are the
of “hydraulic mission” [9], It was suggested that such glaciers of the Himalayas and the monsoon rains, and
an approach would increase the efficiency of water use the peninsular component, consisting of water mainly
and enable society to achieve sustainability, while opti from the monsoon rains. There is surplus water in the
mising the economic return on water. They also Himalayan component and partial deficit in the pen
pointed out that, in river basins where demand exceeds insular component, especially during the dry season,
supply, new sources of water could be found either by resulting in inundation of agricultural land in areas
desalinating sea water or transferring water from a with high precipitation and water scarcity and crop
neighbouring basin where there is a surplus using failure in areas with low precipitation.
interbasin water transfer (IBWT), which is becoming Recently, the national water policy of Government
a common solution. IBWT has been undertaken in of India points to a national perspective to undertake
several countries with various environment and socio such interbasin water transfer. Water transfer is based
economic implications, for instance the USA, China on the view that the transfer of surplus water from one
and Russia. river basin to another, could permanently solve the
India is a large subcontinent and, as a result, the problems of droughts and floods, thus increasing food
challenges it faces are also on a large scale. Water grain production [1]. Thirty possible links between the
resources are unevenly distributed due to variations in major rivers in India, 14 in the Himalayan component
the geographical and temporal distribution of rainfall. and 16 in the peninsular component, have thus been
This leads to drought in some parts of the country and proposed. These were based on earlier proposals, in
floods in others. Providing water during droughts and 1972 by K.L. Rao for the National Water Grid [20],
and in 1977 by Captain Dastur for the Garland Canal
1 The article is published in the original.
Project. These earlier projects were shelved due to a

623
624 NILADRI GUPTA et al.

lack of technical feasibility, desirability and economic layan trend, and transverse faults trending NS, NW
viability [3]. SE and NESW running across the Himalayan trend
The Ganga and Brahmaputra basins (the two major (see Fig. 1). A few of these faults have been shown to
river basins in India) constitute a major part of the be active in geologically recent times, while others are
Himalayan component. There is a surplus of water in only thought to be active. These networks of faults/lin
these two river basins and floods are therefore a recur eaments are considered to be a major contributor to
ring phenomenon. A significant reduction in regional the instability of this region [5]. Any movement or
imbalance could be achieved by storing these water adjustment along these structural elements will have
resources in reservoirs, and transferring the excess an effect on river geomorphology in the area, e.g. shift
water to other parts of the country via canals. However, ing river courses, which can be observed in the rivers
it is not known whether such a project would actually included in the ILR project. The aims of this study
prevent drought, or merely move the problem from were
one place to another. Many other questions also to investigate the geomorphology, geology, land use
remain. and drainage characteristics of the area using multi
Will it be necessary to relocate people, and if so, temporal and multisensor satellite data;
how many? to identify suitable reservoir sites and to determine
Are there any smallscale alternatives to engineer the storage capacity of the major tributaries traversing
ing on this grand scale? the study area using a numerical method and available
Is the terrain suitable for building canals to transfer elevation data from a Digital Elevation Model (DEM)
the water? of NASA’s Shuttle Radar Topographic Mission
(SRTM);
What types of engineering structures are required,
and where can they be located? to determine possible routes of the link canal for
IBWT;
What will the effect of the project be on the envi
ronment? to evaluate the possible benefits of the transported
water in terms of usage assuming a specific transfer
How much water will actually be made available? rate, and taking losses during transportation and
The project thus requires a multidisciplinary decrease in reservoir capacity into account.
approach.
Presently, IBWT accounts for about 540 × 109 m3
water per year, i.e. 14% of all global withdrawal, and STUDY AREA
there are proposals for an additional 940 × 109 m3 per The area studied is located in the Jalpaiguri district
year [11]. These volumes together form only a quarter in the state of West Bengal, in the eastern part of India
of the projected need for water withdrawal by 2025, (Fig. 1).
and are thus not considered significant in terms of The area is located between 26°16′ and 27°0′N and
water resource management. Interbasin water trans 88°4 and 89°53′E, covering an area of about 6190 km2.
fer has effects on both the donor and recipient basins. It is bounded by the foothills of the Himalayas in the
Such projects are also limited by the geoenvironmen north and northwest and by the alluvial plains of the
tal and socioeconomic conditions in the area. Expe Ganga and Brahmaputra rivers in the south and east.
riences of water transfer projects such as the Central The climate varies from sub tropical in the northern
Valley Project in the USA and the Aral Sea Project in part to humid sub tropical in the southern part. The
the former USSR have revealed some of the limita annual rainfall is unevenly distributed. The mean
tions and consequences of such projects [14, 18, 4, 8]. annual rainfall varies from 3160 mm in the monsoon
Experts in the field have acknowledged that there season (May–October) to 201 mm in the dry period
may not be any real surplus of water despite the fact (November–April). The Mahananda River is the only
that there are real shortages [12]. Thus, an integrated major river of the Ganga basin and the Teesta,
water resources management plan is required to Jaldhaka, Torsa, Raidak and Sankosh are the major
ensure that both water and profits are properly distrib rivers of the Brahmaputra basin within the area stud
uted. The sustainability of IBWT programmes must be ied.
assessed in terms of design, operation and manage Geomorphologically, the area is part of the Teesta,
ment. Jaldhaka and Torsa interfluve belt of North Bengal.
According to the ILR project, the excess water Locally, the northern part of the area is called “Duars”
from the Ganga and Brahmaputra rivers and their trib and is part of the Piedmont plains at the foothills of the
utaries must be transported through the Himalayan Himalayas. These are gradually transformed into allu
foothill region, which is structurally characterized by vial plains further south. The Piedmont region is dis
EW trending faults/lineaments parallel to the Hima sected by the major rivers and the tributaries consid

WATER RESOURCES Vol. 37 No. 5 2010

 N
g

WATER RESOURCES
27°0′00″N W E n B
i h
S l
e u

Vol. 37
t a
e n

No. 5
r
R.Murti
26°45′0″N

2010
a
R.Mahananda
R.Kaljani R.Sankos

D
R.Jaldhaka R.Torsa
A s s

R.Raidak
a m

B
26°30′0″N o c h
a R.Teesta b
n K
e
g Index h
a r
Rock type
International

l
Boundary Baikunthapur Formation Canal
h State Boundary MitialiChalsa Formation River
District Boundary

a
SamsingThalijora

s
Source: GSI Geology Map Sandy Area
Railways Formation Scale
SOI Topo Map 1975 Roads Shagaon Formation
26°15′0″N IRS P6 LISS III, Nov. 2005 d e 0 5 10 20 30 40
Faults/Lineaments SubHimalayan Siwaliks
SRTM Data kilometres

88°30′00″E 88°45′0″E 89°0′0″E 89°15′0″E 89°30′0″E 89°45′0″E

USE OF GEOINFORMATICS FOR INTERBASIN WATER TRANSFER ASSESSMENT

Fig. 1. The area studied, showing the lineaments identified from IRS P6 satellite sensor LISS III (15 Nov., 2005) and DEM generated from SRTM data (90 m resolution) super
imposed on the geology.
625
626 NILADRI GUPTA et al.

ered in this study. The northern part of the area is char mate aim of joining the new canal to the existing canal
acterized by faninfan morphology and the ancient system between the Teesta and the Mahananda basins.
deposits are fluvioglacial in origin, as evidenced by
huge boulders. Later fluvial activity can be observed in
the form of terraces, where coarse pebbles to clay size Terrain Analysis and Reservoir Location
materials are found. Rill and gully erosion over a long The method used to identify potential sites for res
period of time has produced deposits with a dissected ervoirs takes into account both the physical suitability
undulating surface. In the alluvial plains levees, back and the socioeconomic suitability of the location.
swamps and oxbow lakes are dominant, and all consist Available IRS 1D LISS III, IRS P6 LISS III, and
mainly of recent sediments. The area is distinctly dif LANDSAT TM data were processed together with
ferent from the rest of the frontal Himalayas and is Survey of India (SOI) topographical maps and other
characterized by the absence of large EW flowing riv ancillary data, using visual or digital classification
ers like the Ganga to the west and the Brahmaputra to techniques to generate prefield thematic maps per
the east. Instead, huge fans and flood plains have taining to geomorphology, land use/land cover and
evolved from southerly flowing rivers such as the drainage of the area on a scale of 1 : 50000. The geol
Teesta and the Mahananda [5]. ogy map was generated by scaling up the Geological
Geologically, the Piedmont zone, which forms the Survey of India (GSI) geological map on the scale of
northern part of the study area, can be divided into 5 1 : 250000 to 1 : 50000 and updating it using satellite
morphostratigraphic units based on their oxidation data. The prefield thematic maps thus generated were
index, the colour of the sediments in the weathering then validated by ground truthing to generate the final
zone and the colour of the topsoil [5]. The study area geomorphology, geology, drainage and land use/land
has been influenced by tectonic activity that took cover maps. These thematic maps were used to asses
place in the eastern Himalayas from the Pliocene to the physical suitability of potential reservoir sites and
recent times [10]. The northern part of the area is were used as input to the multicriteria analysis. The
characterized by alluvial fan deposits which have coa canal transport network and the village maps were
lesced to form the Piedmont zone. derived from the SOI topographical maps updated
The land use and land cover of the study area also with satellite data. These were used as socioeconomic
reflect the characteristics of the terrain. The northern suitability inputs for the multicriteria analysis
part is mainly covered by dense forest with occasional (MCA). The detailed methodology is illustrated in
degraded forest, encroachments and tea plantations in Fig. 3.
the hills and the fan lobe areas, while the southern part Two GISbased approaches can be adopted for ter
is used for agricultural purposes, with one or two crops rain analysis and multicriteria analysis (TA and
a year, and occasional scrubland. MCA) to evaluate spatial features and suitable loca
tions for reservoirs [16]: one, a subjective, knowledge
driven approach that emphasizes expert knowledge in
METHODOLOGY the matter, and the other a datadriven approach
The methodology was based on the problems, employing objective criteria and weighting. In the
aims, the data available, and technological knowhow present case a datadriven approach was adopted as it
and its implementation. Various methods can be used provides adequate information to identify the section
to identify possible locations for dams and reservoirs, of the river suitable for the location of the dam and the
and to determine the best route of canals. The choice reservoir. Table 1 lists the criteria applied in the MCA
of method depends on the drainage characteristics and to identify suitable locations for dams and reservoirs
the terrain, as well as the technology appropriate for a based on certain considerations.
particular environment. Details of the approach
adopted are shown in the form of flow diagram in
Fig. 2. Calculation of the Reservoir Volume
The procedure was divided into 3 stages. Stage I A dam was created at each potential reservoir loca
deals with the terrain analysis (TA) required for reser tion identified in Stage I of the procedure, taking into
voir location. Thematic maps were generated from consideration dam height, freeboard and designed
remote sensing data and a multicriteria analysis was water level. The dams at each potential reservoir loca
undertaken using the thematic information to identify tion were converted into a raster and then integrated
possible dam and reservoir sites. Stage II deals with the with the available DEM data (SRTM Version 3) [21].
estimation of reservoir storage capacity in a GIS The SRTM data had already been preprocessed to fill
framework in which SRTM data were used as the input the voids using an interpolative technique, i.e. the
for an iterative numerical propagation method to TOPOGRID algorithm of Arc/Info, providing a
determine the spatial extent of the reservoir and fac seamless elevation dataset of 5 by 5 degree tiles. To
tors related to reservoir capacity. The final stage is con determine the areal extent of the reservoir, a neigh
cerned with identifying the best route for canals to bourhood connectivity function of ILWIS software
transport the water from the reservoirs, with the ulti [13] was used to help identify all neighbouring pixels in

WATER RESOURCES Vol. 37 No. 5 2010

 USE OF GEOINFORMATICS FOR INTERBASIN WATER TRANSFER ASSESSMENT 627

Satellite data
+
Ancillary data

Digital/Visual
interpretation
STAGE–I

Generation of
thematic maps

Optimal canal route

STAGE–III
Multicriteria Potential
analysis reservoil
sites

DEM
(SRTM Data)

Neighbouhood
function
STAGE–II

(Raster analysis)

Flooded area
Preferred

map
locations
dam

Reservoir
capacity
atatistics

Fig. 2. Overall research methodology.

the DEM with an elevation lower than or equal to the then determined by integrating the reservoir depth
specified water level of the reservoir behind the dam (the value for each pixel) over the total inundated
[7, 2, 15]. Figure 4 illustrates the process. area. The most suitable dam and reservoir locations
The reservoir depth was calculated by subtract were selected based on the output of the neighbour
ing the elevation contained in the DEM from the hood function. The detailed methodology is illus
reservoir water level. The reservoir capacity was trated in Fig. 5.

WATER RESOURCES Vol. 37 No. 5 2010

628 NILADRI GUPTA et al.

RS Data: Geocoded
IRS P6 LISS III (2006)
IRS 1D LISS III (2001)
LANDSAT TM (1990) Preparation of prefield
Visual interpreatation thematic maps in 1:50000
and/digital scale viz/ geomorphology,
classification geology, drainage, landuse
SOI toposheet (1:50000 scale)
Ancillary data
Ground verification
and/
data collection
Preparation of
administrative and
transport network Final geomorphology map

Final geology map

Updation
Final drainage map
Final land use/land cover
map
Final village map

Final transport
map Terrarn analysis &
classification

Socio – Muiti criteria analysis Physical

economic
suitabiloty suitability

Potential sites of
reservoirs

Fig. 3. Method used for terrain analysis and the identification of sites suitable for the location of dams and reservoirs.

Criteria for Optimal Canal Routing calculation was carried out assuming normal flow
After determining the suitable locations for reser depth in a rectangular channel where the slope of the
voirs based on estimates of the dam height and reser water surface and channel bottom is same, depth
voir storage capacity, possible canal routes were iden remaining constant and no acceleration of flow
tified based on the gradient of the terrain for the trans occurs. A MS Excel based application which uses an
port of water by gravity. The ultimate aim was to join iterative process and Secant’s method was used for this
the proposed canal to the existing TeestaMahananda purpose [17].
canal. The fan lobes of the region were also taken into
consideration as it is not always possible to build canals RESULTS AND DISCUSSION
through the lobes. A transfer rate from each reservoir
was estimated based on the storage capacity of each Terrain Analysis and MultiCriteria Analysis
reservoir and the number of days the flow rate could be Both the physical and the socioeconomic suitabil
maintained to accommodate the dry season’s lower ity were considered in the terrain analysis and the
flow. The dimensions of the proposed main link canal, multicriteria analysis. The information on which
divided into 5 sections, were based on the water trans physical suitability was determined included the land
fer rate from each reservoir to the main link canal. The use map, the geomorphology map, the geology map,

WATER RESOURCES Vol. 37 No. 5 2010

 USE OF GEOINFORMATICS FOR INTERBASIN WATER TRANSFER ASSESSMENT 629

Table 1. Key Criteria for dam site selection and the underlying condition considered for different thematic information
used for MCA
Thematic Layer Criteria for dam site selection Underlying condition
Settlement Not to be located near or within 1 Km settlement area Safety (to avoid flooding)
Transport Not to be located within 1 km of major roads or railroad Safety (to avoid flooding and disruption of commu
Network nication)
DEM To be located above 110 m elevation New canal alignment is to be connected to the exist
ing Teesta – Mahananda link canal.
Slope To be located in a slope <12% Hydraulic condition/Stability
Structure To be located at a distance >500 m Stability
of an existing lineament/fault scarp
Drainage Preferably on river bed To have sufficient reservoir capacity
Geomorphology To be located avoiding low altitude fan/alluvial plain Hydraulic condition/Stability

the structural map, the slope map and the DEM. The outcome was presented in the form of a composite
socioeconomic suitability was assessed based on the suitability map showing the potential dam and reser
settlement map and the transport network map. The voir locations as can be seen in Fig. 6.
multicriteria analysis involved raster analysis in which
all the input maps were converted into rasters. Each TA and MCA were used to identify potential dam
geomorphic unit was weighted with a value from 0 to 5: and reservoir sites in the study area. Five sites were
a value of 0 being assigned to alluvial plains (least suit found. The existing barrage on the river Teesta (Dam 1)
able) and a value of 3 to midaltitude intermediate fans was used to test the iterative propagation method and
(reasonably suitable). Of the 15 land use classes iden the capacity of the existing reservoir was estimated.
tified, 5 major classes were considered to be likely to be Other potential reservoir locations were identified on the
affected by the reservoir, and agricultural land was
assigned a weight of 0 (least suitable), and river beds a tributaries of the Jaldhaka (Dam 2), Murti (Dam 3),
weight of 5 (most suitable). A Boolean operation was Torsa (Dam 4) and Kaljani (Dam 5) rivers.
applied to the elevation and slope data in which it was According to the results of the TA and MCA, one
required that the reservoirs be located above an alti of the five identified sites was not suitable (Dam 3) as
tude of 110 m and the slope <12%. The pixel values
that satisfied the criteria were given assigned the value the land cover on both sides of the river was forest,
1 (suitable) and those that did not were assigned a which has a high weighting factor. However, the site
value of 0 (unsuitable) to identify the physically suit was suitable with regard to the slope and drainage
able locations for reservoirs and dams. Similarly, to characteristics, so a dam was proposed here and the
identify sites that were socioeconomically suitable a reservoir volume was estimated.
buffer zone was created around roads and railways and
settlements (as listed in Table 1). Areas within the Although Dam A was suitable according to the
buffer zone were considered unsuitable (and assigned multicriteria analysis, the reservoir capacity was not
a value of 0) while those outside the buffer zone were sufficiently high and this location was therefore not
considered suitable (and assigned a value of 1). The considered.

Lower but not

connected

Freeboard
Dam height

Design water
Reservoir

level
level

Dam
Terrain

Fig. 4. Determination of the areal extent of a reservoir using topography and a neighbourhood connectivity operator [7].

WATER RESOURCES Vol. 37 No. 5 2010

630 NILADRI GUPTA et al.

Dam site

Designed
Dam height Free board water
level

Mark cells with

location of dam

Combine with DEM

DEM (SRTM Data V3)

Map with marked Compare with

DEM (= future reservoir elevation of terrain
+ area) cell upstream (neighbourhood
Dam site
of dam site calculation)

Mark cell < = water

level of marked cells

NO

YES Elevation of
Flooded neighbouring cell >
area map water level of marked
cell (s)

Storage
capacity/
Reservoir
volume

Fig. 5. Methodology for reservoir map creation and volume calculation from the DEM using the iterative propagation method [7].

Estimation of Reservoir Capacity the labelled neighbouring cells. The process was
repeated until no more connected cells had an eleva
A cell was defined behind the dam, and an algo tion lower than or equal to the required level. All the
rithm was used to identify all the neighbouring cells identified cells were labelled as being flooded.
with an elevation lower than or equal to the specified
reservoir water level. The cells that satisfied this condi The extent and capacity of the reservoirs were
tion were labelled. The algorithm was then applied to determined for dams with three different levels, and

WATER RESOURCES Vol. 37 No. 5 2010

 N
INDIA

W E

WATER RESOURCES
S
N G
27°0′0″N I B
L
E

Vol. 37
E H
U T
Dam 2 WEST BENGAL
J A

No. 5
Dam A

R
N

A
Dam 3

2010
Dam 1 Dam 4
26°45′0″N Dam 5
A

D
Teesta Jaldhaka
S

Torsa
S

Raidak
B
A M

A
26°30′0″N
N O C
G H
K B R

L
E H A

A
Legend

D
E Not Suitable for Reservoir
26°15′0″N H Scale Location
S Suitable for Reservoir Location
Source: SRTM Version 3 Data 0 5 10 20 30 40
Thematic Maps used in MCA Kilomets Dam Location
SOI Topo Map, 1975

88°15′0″E 88°30′0″E 88°45′0″E 89°0′0″E 89°15′0″E 89°30′0″E 89°45′0″E

USE OF GEOINFORMATICS FOR INTERBASIN WATER TRANSFER ASSESSMENT

Fig. 6. Possible locations of reservoirs based on multicriteria analysis and the hydrological character of the study area.
631
632 NILADRI GUPTA et al.

N
W E
S
N
26°52′30′′ Reservoir extent on R. Teesta

26°47′0′′

Dam hidht 13 m
Dam hidht 18 m
26°41′30′′ Dam hidht 22 m
0 5 10 km

88°29′30′′ 88°35′0′′ 88°40′30″ E

Fig. 7. Areal extent of the reservoir behind Dam 1 on the River Teesta for different dam heights superimposed on the IRS P6 LISS
III data.

three reservoir extent maps were created for each Similar reservoir extent maps were generated for
potential site. The catchment area of the rivers and the three different dam heights at the other four potential
tributaries from the dam locations were determined locations. The runoff was assumed to be 50% of the
from the drainage map generated from the DEM. The mean annual precipitation. The mean annual precipi
“DEM Hydroparameterization” module of ILWIS tation was assumed to be the same for all the catch
ment areas and was determined as the mean of the
was used [13]. The dam location was used to delineate mean annual precipitation recorded at four weather
the catchment area upstream of the dam. Figure 7 stations in the study area (3517 mm per year). The
shows the extent of the reservoir when Dam 1 on the present Teesta barrage has a transfer rate of about
River Teesta has three different levels. The calculated 30 m3 s–1, which is equivalent to about 946 million m3
volumes of the reservoir at different dam height on the per year. The total runoff available (50% of the mean
river Teesta (Dam 1) are shown in Fig. 8. Similar plots annual precipitation in the study area) is about
were made for the reservoirs on the other rivers. In all 15360 million m3. Thus, the total volume of water
cases an almost linear relationship was found between transferred is about 6% of the total runoff. The volume
the dam height and the reservoir capacity, and there of water available through runoff, the maximum stor
age capacity of the reservoir and the percentage of run
fore no intermediate levels were calculated, as shown off utilized for storage was estimated for each of the
by the regression equation in Fig. 8. five reservoirs. Assuming a flow rate of 20 m3 s–1 from
the other 4 reservoirs ~ 5 m3 s–1 from each (that from
the River Teesta being 30 m3 s–1 ) the amount of water
Dam 1 – River Teesta available each day would be about 50 m3 s–1. With a
Capacity (milion m3)

1000
800 total available reservoir capacity of 972 million m3,
Reservoir

water would be available for only 225 days in a year.

600
400 Since this scenario did not provide sufficient water
y = 76.903x – 831.46
200 for a whole year, another configuration was investi
R2 = 0.9324 gated. Two of the existing reservoirs, one on a tributary
0
12 14 16 18 20 22 24 of the River Jaldhaka (Dam 2) and the other on River
Dam Height (m) Torsa (Dam 4), were relocated upstream at the north
ern edge of the study area, and one of the reservoirs on
the River Kaljani (Dam 5) was rejected. Correspond
Fig. 8. Reservoir storage capacity curve for different dam ing estimates were made of the volume of water avail
heights on R. Teesta as derived from digital elevation able through runoff, the maximum storage capacity of
model. the reservoirs and the percentage of runoff being uti

WATER RESOURCES Vol. 37 No. 5 2010

 USE OF GEOINFORMATICS FOR INTERBASIN WATER TRANSFER ASSESSMENT 633

lized for storage in each of the four reservoirs. The Table 2. Water availability from each reservoir in 3rd Scenario
assumptions were the same as used above. In this case, considering different transfer rates
it was found that, assuming a water transfer of 50 m3 s–1,
Dam Q out per day, No. of
it would be possible to supply water for a maximum of no.
River
(m3 s–1) days
276 days.
A third alternative was then investigated in which 1 Teesta 30 355
two more reservoirs were located at two new locations, Jaldhaka Tributary 2 296
one on the River Raidak (New Dam 5) instead of River 2 (Upstream new location)
Kaljani (Dam 5) and the other on the River Sankosh
(Dam 6). Both these rivers flow through the study 3 197
area, but no suitable reservoir site was found within the 5 118
study area using TA and MCA. The new reservoirs 3 Murti 2 124
were thus located on these two rivers further upstream, 3 82
outside the study area. Besides being outside the study
area, the reservoir extended over the international 5 49
boundary into Bhutan. In this case, it was found that 4 Torsa (Upstream new location) 5 453
the stored water would be able to supply water for a 5 Raidak (Outside area) 3 416
maximum of 345 days, which is almost the whole year. 5 250
Estimates of the water availability from each of the
6 Sankosh (Outside area) 5 446
six proposed reservoirs in the third alternative can be
found in Table 2. The calculations indicate that the
reservoirs behind Dam 1 on the River Teesta could
supply water for almost a year with the present flow was divided into five sections of different dimensions
rate of the Teesta canal, while the reservoirs at Dam 4 according to the amount of water being transferred
and Dam 6 could supply water at a flow rate of more from each reservoir.
than 5 m3 s–1. The reservoir at Dam 5 could supply As the total transfer rate was assumed to be 50 m3 s–1
water on a continuous basis if the transfer rate were and the existing TMLC had a capacity of about
3 m3 s–1, while those at Dam 2 and Dam 3 could supply 30 m3 s–1, the rest of the reservoirs would contribute
water almost all the year with a combined flow rate of about 20 m3 s–1. Taking this into consideration a trans
5 m3 s–1. fer rate of 6 m3 s–1 was assumed for the first section of
It was found in all the alternatives that the percent the canal with input from the River Sankosh (Dam 6)
age of runoff stored for transfer from each reservoir did to the proposed main link canal. An additional flow of
not exceed 8% of the total amount of water available 4 m3 s–1 was added to the second section of the canal
from runoff, which was about 2% more than the value from the River Raidak (New Dam 5), while a flow of
assumed (6% runoff being transferred from the present
5 m3 s–1 was added to each of the third and fourth sec
Teesta reservoir at a rate of 30 m3 s–1). This indicates
that the natural flow of the river downstream would tions of the canal, from the River Torsa (Dam 4) and
not be hampered to any great extent. The only excep combined transfer from the River Murti (Dam 3) and
tion was the reservoir behind Dam 2 on the tributary of the tributary of the River Jaldhaka (Dam 2), respec
the River Jadhaka, which had about 22.5% of runoff tively. Section 5 of the canal was assigned a flow rate of
transferred. In this case as the reservoir and the dam 50 m3 s–1 considering the input of 30 m3 s–1 from the
were located on one of the tributaries of the main River River Teesta (Dam 1). A schematic diagram of the link
Jaldhaka, the water available downstream in the river canal and the contributions from each reservoir on the
would not be affected to any great extent as the other rivers is shown in Fig. 10. Based on the flow rate from
tributaries would still feed the main river. each reservoir the dimensions of the canal were calcu
lated using an iterative process by considering normal
flow depth in a rectangular channel using Secant’s
Optimal Route for the Link Canal method [17]. The estimated dimensions of the pro
The ultimate aim is to merge the proposed canal posed link canal are given in Table 3. An average flow
with the existing TeestaMahananda link canal velocity of 1 m s–1 was assumed in the proposed link
(TMLC). An optimal route was found, taking into canal.
consideration the slope, roughness and character of
the terrain, especially the midaltitude intermediate
fan, which was avoided as far as possible. Although the Assessment of IBWT
land use in the area should be taken into consider
ation, especially forested areas, tea plantations and An assessment of the possible utilization of the
agricultural land, it was not possible to avoid all the water available in the three alternatives described
above land use categories due to the location of the above was made, based on the following two assump
reservoirs (see Fig. 9). The proposed main link canal tions:

WATER RESOURCES Vol. 37 No. 5 2010

634 NILADRI GUPTA et al.

W
Dam 2 New E
270′0″N G B S
I N
L H
E Dam 4 New
U T
E A
J
N
Dam 1
D A R

Section 4 Section 3 Dam 6

Dam 5
26°45′0″N Section 5
Section 1
Section 2

A S S A M
Torsa
Mahananda
Jaldhaka Raidak
B

Teesta
Sankosh
A

26°30′0″N
N
G
L
A

Legend
Link Abandoned Channel Land wih scrub INDIA
Teesta (DAM1) Urban Area Open Spaces
D

Jaldhaka Tributary River Island

(DAM2 NEW) Canal
Murti (DAM3) Kharif River
H

26°15′0″N E Torsa (DAM4 NEW) Dense Forest Scrub Forest

S Raidak (DAM5) Forest Plantation Villages
Scale Kharif + Rabi
Sankosh (DAM6) (double Cropping) Tea Gardens
0 3.75 7.5 15 22.5 30 Distriet Boundary Ponds/Tanks Source: IRS P6 LISS III 2005, Jan 2006
Kilometrs WEST BENGAL
SOI Topo Map 1975 (1: 50000 scale)

88°30′0″E 88°45′0″E 89°0′0″E 89°15′0″E 89°30′0″ 89°45′0″E

Fig. 9. Map showing proposed canal route superimposed on the land use and land cover map.

the amount of water required for irrigation was 2 mm carry a great deal of sediments [5, 6]. As the reservoirs
per day (based on a requirement of 2 cm over a 10day are located in the Piedmont zone heavy silting can be
period); expected. Figure 11a shows the area that can be irri
70–135 litres of water are consumed per person per gated if 40% of the available water is used for irrigation
day for domestic use, the lower value in rural areas and and Fig. 11b – the number of days per year water will
the higher value in urban areas [22]. be available if 20% of the available water is used for
The loss of water due to evaporation and during human consumption.
conveyance was also considered, and was assumed to The results indicate that if water were available at a
be 10%, although the actual loss would vary with discharge rate of 50 m3s–1 then a total of 4.32 million m3
changing seasons, climatic conditions and the aging of water would be available daily. If 40% of the available
the canal. Reductions in reservoir capacity over time water were used for irrigation a total of 2160 km2 could
were assumed to vary from 10 to 40% as it has been be irrigated daily and if 20% were used for human con
found in earlier studies that the rivers in the study area sumption, water could be provided for 12.34 million

Table 3. Estimated dimensions of the proposed link canal using reiterative method
Canal Sections Section 1 Section 2 Section 3 Section 4 Section 5
Factors
Transfer Rate, m3 s–1 6 10 15 20 50
Manning Roughness, n 0.04 0.04 0.04 0.04 0.04
Slope m m–1 0.002 0.0017 0.00125 0.001 0.00055
Width of Canal m 3 4 5 6 9
Flow depth m 1.99 2.31 2.81 3.16 5.25
Flow Velocity m s–1 1.01 1.08 1.07 1.05 1.06

WATER RESOURCES Vol. 37 No. 5 2010

 USE OF GEOINFORMATICS FOR INTERBASIN WATER TRANSFER ASSESSMENT 635

DAM 6
DAM 2 NEW
(R. Jaldhaka (R. Sankosh)
Tributary)
DAM 5
DAM 4 NEW (R. Raidak)
(R. Torsa)
6 m3/s
DAM 1 (R. Teesta)
4 m3/s
5 m3/s
3 m3/s
DAM 3 (R. Murti)

30 m3/s
2 m3/s

LINK
L
CANA
Section 1

Section 2

Section 3

Section 4

Section 5

Fig. 10. Estimated dimensions of the proposed link canal.

people if they used 70 litres per day, or 6.4 million peo considering specific potential reservoir locations,
ple if they used 135 litres per day. highresolution elevation data are essential as the
In reality, the water would be used for many other SRTM data may underestimate slope steepness [19]
purposes apart from irrigation and human consump resulting in over or underestimation of the reservoir
tion, such as power generation and in industry, but this capacity. The uncertainties related to elevation infor
preliminary assessment gives an idea of the benefits of mation and the effects of radarreflective surfaces were
the planned ILR project. not considered in the present study.
Neither the slope nor the condition of the terrain
was considered in detail when determining the most
Limitations of the Methodology suitable route of the link canal. The alternative pro
The present methodology has the following main posed was the result of using a methodology with lim
limitations. ited ground data.
The terrain analysis and multicriteria analysis The assessment of the utilization of the water made
included only some of the factors affecting site selec available was based on certain assumptions regarding
tion such as slope, drainage, geomorphology, geology, water transfer and reservoir capacity. The actual avail
structure and land use regarding physical suitability ability in the study area was not determined. The
and village location and transport networks regarding transfer volume was also assumed and only irrigation
socioeconomic suitability. In reality, other factors such and human consumption were considered. In reality,
as the kind of drainage, drainage discharge, soil type the water will also be used for other purposes.
and soil character would be required in the multicri
teria assessment.
CONCLUSIONS AND RECOMMENDATIONS
The SRTM data used as the input for elevation had
a resolution of 90 m and were taken from a global Interbasin water transfer is being considered a
dataset. This type of dataset is useful for preliminary dominant solution to cope with the deficit of water in
assessments of possible reservoir locations, but when some major river basins of the world. The ILR project

WATER RESOURCES Vol. 37 No. 5 2010

636 NILADRI GUPTA et al.

No. of Days of Water Avalability

10% + 10%
Irrigated Area (million km2) 10% + 10% 20% + 10%
0.30 300
20% + 10% 30% + 10%
0.25 30% + 10% 250 40% + 10%
40% + 10%
0.20 200
0.15 150
0.10 100
0.05 50
0 0
1 2 3 1 2 3
Scenario Scenario

Fig. 11. a  Estimates of the area that can be irrigated in 3 scenarios assuming 40% of the available water is used for irrigation, and
b  estimated number of days of water availability in 3 scenarios for a rural population of 12.34 million consuming 70 litres/day or
an urban population of 6.4 million consuming 135 litres per day, taking into consideration a reduction in reservoir capacity
(10⎯40%) due to silting and a 10% loss due to evaporation and conveyance losses.

is an example of this. In this study, an approach was tioned factors. The iterative process used indicates that
developed to make a preliminary assessment of IBWT to maintain a flow of about 1 m3 s–1, the channel slope
using geoinformatics. Terrain analysis and multicrite should not exceed 2 m km–1. Detailed topographic
ria analysis were used to identify potential reservoir information is therefore required to determine slopes.
locations. SRTM data were used to provide elevation No hydrological information on the river discharge
information to determine reservoir capacity. The or silting was available for the study area. River dis
required canal dimensions were estimated, and the charge information will enable more accurate esti
optimal route for the canal was proposed. Utilization mates of the water availability in the rivers. Silting data
of the water in terms of irrigation and human con will be useful to calculate the reduction in reservoir
sumption was also assessed. capacity over time. The latter is important as the suit
The study showed that multitemporal and multi able sites are located in active flood plains.
sensor satellite data can be used to map the relevant An assessment of the utilization of the available
geomorphology, as well as the geology, land use and water for irrigation and human consumption was
drainage characteristics of a study area, and these can made, taking into consideration losses due to evapora
be used as input for terrain analysis and multicriteria tion, conveyance and reduction in reservoir capacity.
analysis to identify potential reservoir sites. A combi It can be concluded that this type of assessment
nation of elevation and slope data and the abovemen enables water utilization to be estimated in the differ
tioned thematic information improves the identifica ent scenarios of water availability prevailing in the real
tion procedure of the potential reservoir sites. The world. Information on actual water requirements in
method presented here could be used for preliminary various sectors is essential for better feasibility assess
terrain analysis and multicriteria analysis in large ments of the use of water stored in reservoirs.
river basins. However, more data must be collected for The information gathered and the results obtained
more precise identification of potential reservoir sites. can be used in a preliminary costbenefit analysis, to
We used SRTM Version 3 data [21] as the input for obtain an indication of the cost of the ILR project pro
the characterization and estimation of the reservoir posed by the Indian Government.
capacity in an iterative propagation model. This In conclusion, many issues are involved in inter
method has previously been used to model flood inun basin water transfer programmes apart form reservoir
dation due to dyke failure [7] and storm water surges in site identification, capacity estimation, canal routing
urban areas [15]. SRTM data are thus useful for pre and the possible utilization of the available water. The
liminary reservoir characterization and volume calcu research described above provides a stepping stone for
lation. The use of field data collected by GPS survey more detailed studies.
ing, or other local elevation information of much
higher resolution, for example, GDEM (30 m resolu
tion) from ASTER Stereo, would provide better char ACKNOWLEDGMENTS
acterization of potential reservoir sites. Niladri Gupta would like to express his thanks to
The optimal route of the canal linking the reser the European Union and the Erasmus Mundus con
voirs was determined by taking into account the eleva sortium (University of Southampton, UK, Lund Uni
tion, slope and the characteristics of the terrain, versity, Sweden, Warsaw University, Poland and ITC,
including the geomorphology and land use. The canal the Netherlands). This paper evolved from a disserta
route proposed here was based only on the above men tion by Niladri Gupta, who would like to express his

WATER RESOURCES Vol. 37 No. 5 2010

 USE OF GEOINFORMATICS FOR INTERBASIN WATER TRANSFER ASSESSMENT 637

special thanks to Mr. Sanjoy Nag, Senior Scientist, Engineering, Science and Politics Interlock, Journal of
and Dr. Parthasarathi Chakrabarti, Chief Scientist, at Physics and Chemistry of the Earth, 2007, 33, pp. 28–40.
the Department of Science & Technology, Govt. of 10. Haller, J., Himalayan Orogenesis in Perspective Meta
West Bengal, India for their invaluable support and morphic Rock Sequence in Eastern India, Geohydrol
suggestions during the field work and primary database ogy and Geotectonics of the lower Ganga Basin III, 1980.
generation. 11. ICID. Experiences in InterBasin Water Transfers for
Irrigation, Drainage or Flood Management (3rd Draft 15
August 2005). Unpublished Report, New Delhi: Interna
REFERENCES tional Commission on Irrigation and Drainage ICID
CIID, 2005.
1. Bandyopadhyay, J. and Perveen, S., The Interlinking of
Indian Rivers: Some Questions on the Scientific, Eco 12. International Joint Commission. Protection of the Waters
nomic and Environmental Dimensions of the Proposal, of the Great Lakes: Final Reports. Govt. of Canada and
Seminar on Interlinking Indian Rivers: Bane or Boon? United States. 2000.
Kolkata: IISWBM, 2002. 13. ITC. ILWIS. The Integrated Land and Water Informa
2. Brouwer, D.H., Flood study in the MeghnaDhona tion System. User’s Manual First Edition. 2001. I. D.
goda Polder, Bangladesh, 15th Asian Conference on ITC, P.O. Box 6, 7500 AA. Enschede, The Nether
Remote Sensing, Bangalore, 1994. lands.
3. Chakrabarti, P., River Interlinking Proposal—Water 14. Liu, C., L. J. C. Ma, Interbasin Water Transfer in
shed Development Programme; The Reality, in Water China, Geographical Review, 1983, pp. 253–270.
Security and Management of Water Resources, Saha, G.N., 15. Maathuis, B.H.P.B., et al., Evaluating Urban Stormwa
Ed., New Delhi: Vedams eBooks (P) Ltd, 2004, pp. 97– ter Drainage Using GIS and RS Techniques: A Case
104. Study in Dhaka, Bangladesh, GeoCarto International,
4. Changming, L., Environmental Issues and the South 1999, 14, pp. 19–30.
North Water Transfer Scheme, The China Quarterly, 16. Malipa, J.M.Sc., Dissertation, Site Selection for Dam
1998, 156, pp. 899–910. Construction Using RS and GIS To Select Dam Site for
5. Das, A., and Chattopadhyay, G.S., Neotectonics in the Irrigation along Mpiji River in Bagamoyo District Tanza
Tista, Jaldhaka, and Torsa Interfluve Belt of North nia. Enschede, ITC, the Netherlands, 2005.
Bengal, Geological Survey of India Record, 1993, 101, 17. Mannaerts, C., Water Erosion Prediction and Control
pp. 101–109. Software, Water Resource Division, ITC, the Nether
6. Dasgupta, A., Water Potential, Fluvial, Arteries A Brief lands, 1992.
Scenario, Sechpatro, Irrigation and Waterways Director 18. Micklin, P.P., Data Acquisition and Data Capture
ate, Govt. of West Bengal, 1993, pp. 26⎯32. through Terrain Mapping Units, ITC Journal. Bulletin
7. Fazlur Rahman, A.K.M, Flood Study in the Meghna de l’ ITC, 1988, pp. 23–44.
Dhonagoda Polder, Bangladesh in Introduction to the 19. Wu, S., et al., Characterization and Evaluation of Ele
Use of GIS in Practical Hydrology, Meijerink, A.M.J., vation Data Uncertainty in Water Resources Modelling
Brouwer, H.A.M.d., Mannerts, C.M., Valenzuela, C.R., with GIS, Water Resources Management, 2007, 22,
Eds., The Netherlands: ITC and Ministry of Education pp. 959–972.
and Science; France: UNESCO, Division of Water 20. Interlinking of Rivers, http://www.nwda.gov.in/index
Resources, 1994, pp. 205–212. main.asp?linkid=84&langid=1 (accessed 28 March
8. Garone, P., The History and Ecology of the Wetlands of 2007).
California’S Great Central Valley, EPA STAR Graduate 21. SRTM 90m Digital Elevation Data. Available
Fellowship Conference. University of California. Davis. http://srtm.csi.cgiar.org (accessed 11 September 2007).
CA, 2004, (poster presentation). 21. Water Supply and Sanitation Sector Status: Rural and
9. Gupta, J. and Zaag, P.V.D., Interbasin Water Transfers Urban, http://mahawssd.gov.in/HomeImage/Urban%20
and Integrated Water Resources Management: Where sectorstatus.pdf (accessed 20 December 2007).

WATER RESOURCES Vol. 37 No. 5 2010