Study of Landuse – Groundwater relationship using an Integrated Remote Sensing and...

Study of Landuse – Groundwater relationship using an Integrated Remote Sensing and GIS appraoch


Biswajit Sarma, A. K. Saraf
Department of Earth Sciences
Indian Institute of Technology, Roorkee
Roorkee – 247667, India
[email protected]

1. Introduction
Land, air and water may be regarded as a stock of natural resource assets, which provide a range of functions to meet the increasing human need for food, fodder and fuel wood combined with industrial activities. Recent research indicates that the human induced conversions and modifications of land cover have significance for the functioning of the earth system. Though landuse is mainly controlled by, various biophysical factors like soil, climate, relief and vegetation but the human activities are mainly responsible for the change of attributes of landuse modification and conversion. Continuous exploitation of natural resources beyond threshold limit of resilience of the ecosystem accelerates various geomorphic processes on the earth surface, thereby causing imbalance in natural ecosystem resulting in large-scale disaster in present day habitation. The impact of landuse in the prevailing surface and subsurface hydrologic conditions is remarkably high. Within a basin, the dynamics of hydrologic processes are governed partially by the temporal and spatial characteristics of inputs and outputs and the landuse conditions (Shih, 1996). Keeping these factors in view a full featured spatial information about the changeable landform features and related landuse (i.e. land-water-vegetation inter-relationship) on a natural unit basis is very essential for integrating the same with the related non-spatial data (e.g. demographic, socio-economic etc.) to obtain a real world feature.

The fullest utilization of the potential of the two technologies can be realized only when an integrated approach is adopted. Blending of the these two technologies has proved to be an efficient tool in groundwater studies (Gustafsson, 1993; Saraf and Jain, 1994; Saraf et al., 1994; Krishnamurthy and Sriniwas, 1995; Krishnamurthy et al., 1996; Saraf and Choudhury, 1997; Saraf and Choudhury, 1998 and Saraf et al, 1999). The present study demonstrates the capabilities of LISS-II and LISS-III (Linear Imaging self Scanning Sensor) in understanding the landuse-groundwater relationship and changes in landuse pattern. LISS-III provides higher spatial resolution 23.5m, which makes ground features more prominent in comparison with LISS-II having spatial resolution 36.5m.

The main objectives of the study were:

  1. to develop an integrated remote sensing and GIS technique to establish and evaluate the relationship between landuse and groundwater hydrology,
  2. to identify factors influencing this relationship and their role in controlling the groundwater scenario of the study area,
  3. to evaluate the nature of changes in selected landuse categories,
  4. to have a quantitative assessment of groundwater recharge,
  5. to delineate the groundwater potential zones in the area,
  6. to suggest suitable sites for artificial recharge to augment groundwater in the study area.

2. Study Area
The study area has been chosen as it represents a typical case of landuse impact on groundwater. The Dwarkeshwar watershed with a semi-elliptical shape occupies the northeastern part of Purulia district, but the major part of the Dwarkeshwar watershed is situated in a part of Bankura district of West Bengal state, India. The Dwarkeshwar watershed is bounded by longitudes 86°37’E and 87°28’E and latitudes 23°00’N and 23°32’N. The total area of the watershed is 2270 km2 and it is in between the Damodar basin (to the north) and Kangsabati basin (to the south).

The study area comprises of Precambrian crystallines and recently deposited alluvium connected by an intervening tract (GSI, 1997). In this area, monsoon groundwater recharge cannot meet the demands for groundwater throughout the year. Artificial recharge is necessary to improve the groundwater conditions in the area. In the past, very little thought had been given to utilize artificial recharge methods at suitable locations. The present work is an attempt in this direction.

3. Data Used
The different sets of data used for the study with their sources are given in Table 1

Table 1: Details of various data sets used in the present study

Type of Data

Details of Data

Source of Data

Survey of India (SOI) toposheets 73 I, 73 M (scale 1:2,50,000) 73 I/10, I/11, I/12, I/14, I/15, I/16, 73 M/3, M/4, M/7, M/8 (scale 1:50,000) Survey of India (SOI), Dehradun
Thematic maps: Geomorphology, Soil, geology (Scale 1:2,50,000) National Bureau of Soil Survey and Land Use Planning (NBSS & LUP), Nagpur, Geological Survey of India (GSI), Calcutta
Remote Sensing digital data sets of IRS-1B-LISS-II & IRS-1C-LISS-III



National Remote Sensing Agency (NRSA), Hyderabad
Scene Date Scene Date
19/51/A2 03.11.88 106/55 09.11.96
19/52/A1 03.11.88 106/56 04.11.97
20/51/B2 26.11.88 107/55 14.11.96
20/52/B1 26.11.88 107/56 14.11.96
Groundwater data: Depth to water level of 34 wells The depth of water level for the month of January, April, August, and November for years from 1984 to 1997 State Water Investigation Directorate (SWID), Calcutta

4. Methodology
The methodology adopted in the present study is represented schematically (Figure 1) and described in the following steps:

Figure 1 Flow chart showing data flow and different GIS analysis operations followed in the present study

  1. In the initial stage of GIS spatial database development various analogue maps, which were in different scales obtained from different organizations, were converted into digital format by using manual digitization method in ILWIS software.
  2. In the second stage, digital image processing of the satellite data were carried out for extraction of pertinent information. The IRS-LISS-II and IRS-1C-LISS-III data were classified using supervised classification technique. The landuse maps of the year 1988 and 1996 were prepared and the original extent of the landuse in 1988 is compared with the changes that have occurred in 1996 to compute an overall change patterns in each category (Figure 2).

    Figure 2 Landuse change map derived from the analysis of IRS-1A-LISS-II data of 1988 and IRS-1C-LISS-III data of 1996

  3. In the third stage, all the above themes were brought into Arc View 3.1 for further processing and analysis.
  4. The fourth stage involves the integrated analysis of multi-disciplinary data sets to construct composite information set to explain various queries in the spatial context. GIS and landuse are natural partners as both of them deal with spatial data. The landuse change evaluation with respect to groundwater changes is provided.

5. Weighted Index Overlay Model for Groundwater Prospects
Weighted overlay analysis is a simple and straightforward method for a combined analysis of multi-class maps. The efficacy of this method lies in that human judgement can be incorporated in the analysis. A weight represents the relative importance of a parameter vis-a-vis the objective. Weighted index overlay method takes into consideration the relative importance of the parameters and the classes belonging to each parameter. There is no standard scale for a simple weighted overlay method. For this purpose, criteria for the analysis should be defined and each parameter should be assigned importance (Saraf and Choudhury, 1997 and Saraf and Choudhury, 1998).

Determination of weightage of each class is the most crucial in integrated analysis, as the output is largely dependent on the assignment of appropriate weightage. Consideration of relative importance leads to a better representation of the actual ground situation (Choudhury, 1999). Considering the hydro-geomorphic conditions of the area weighted indexing has been adopted (Table 2) to delineate groundwater prospective zones (Figure 3) considering five parameters namely geology, geomorphology, soils, slope and lineaments.

Figure 3 Groundwater prospective zones analysed on the basis of geological, geomorphological, soils, lineaments and slope information and their analysis using GIS

Table 2: Weightage of different parameters for Groundwater prospects & Artificial Recharge sites

Serial No. Criteria Classes Weights For Groundwater Prospects Weights For Artificial Recharge
1 Geology Alluvial deposits (sand, silt, clay) 3 3
Ferruginous gritty sandstone & shale 2 2
Pyroxenite 1 1
Pink granite 1 1
Quartzite/ Quartz schist 1 1
Laterite 2 2
2   Geomorphology Lower alluvial plain 5 3
Flood plains and alluvial fill 5 3
Upper undulating alluvial plain 4 3
Gently to moderately sloping land interspersed with mounds and valleys 3 2
Moderate to strongly sloping land interspersed with isolated hills 2 1
Rock outcrops 1 1
Hillocks and mounds 1 1
Residual hills 1 1
3 Soil Loam 4 4
Sand 3 3
Silt 2 2
Clay loam 1 1
4 Slope 0 – 5° 4 4
5 – 10° 3 3
10° – 20° 2 2
>20° 1 1
5 Lineament <1 km 3 3
1 – 2 km 2 2
>2 km 1 1
6 Recharge 25 – 35 Not Applicable 3
25 – 15 Not Applicable 2
0 – 15 Not Applicable 1

6. Selection of artificial recharge sites
A remote sensing and GIS based method is found to be very useful in suitability analysis for artificial recharge sites in the hard rock terrain (Saraf and Choudhury, 1998). For such analysis the first task was to identify the factors facilitating recharge to take place. The existing artificial recharge system in the area has been studied with respect to its hydrogeology, topography and response in the water level of the wells. Based on these observations, a set of rules has been designed to demarcate the most suitable zones and also to find out the exact sites for artificial recharge. The thematic information layers used in this suitability analysis and weighted indexed overlay model are: (a) Geology, (b) Geomorphology, (c) Soils, (d) Slope, (e) Distance to lineament,(f) Recharge map.

In weighted index overlay, the individual thematic layers and also their classes are assigned weightage (Table 2) on the basis of their relative contribution towards the output. In the present study, weighted indexing method has been used to demarcate the suitability zones for artificial recharge sites (Figure 4). The classes with higher values indicate the most favourable sites for artificial recharge structures.

Figure 4 Potential zones for future artificial recharge sites to provide better groundwater recharge conditions. This map is prepared using the slope map derived from DEM together with the geology, geomorphology, groundwater recharge and lineament maps
7. Results and Discussion
In order to make an accurate prediction of hydrologic response due to landuse changes, it is required (i) to be able to track changes as they actually occur and (ii) to quantitatively understand the effect caused by such changes. Landuse change studies with respect to groundwater hydrological changes have become extremely necessary for planning of any development activity on the earth’s surface. However, in order to understand the landuse changes various influencing factors and interdependent intricacies, a GIS based modeling approach is required.

The land surface is subjected to continuous change due to natural and man-made causes. As the landuse in a watershed is altered in space and time, the factors that influence the hydrologic response of the watershed also changes. Evaluation of the relationship between landuse changes and such factors is one of the goals of the study of landuse-groundwater. The following observations are made after the landuse change analysis:

  1. A mixture of forest cover, agricultural activities and wasteland characterized the landuse of the study area. The major part of the study area was predominantly wasteland especially in the western part of the watershed, which decreased significantly within last 8 years. The next predominant category of landuse was fallow land. The total forest cover inclusive of the plantation increased significantly. Irregular patches of arable land occur in a random fashion in the study area.
  2. The landuse maps of 1988 and 1996 generated from satellite images show that the wasteland is the most dominant class followed by fallow land in the year 1988. However, on comparing with the landuse map of 1996 reveals a decline in the wasteland from 48.17 percent to 8.65 percent, an increase in fallow land from 33.63 percent to 58.8 percent, dense forest increased from 2.9 percent to 7.1 percent. Water bodies had decreased marginally from 0.8 percent to 0.61 percent. While arable land increased from 13.88 percent to 24.1 percent, settlement increased from 0.21 percent to 0.25 percent and river sediment increased from 0.41 percent to 0.49 percent (Figure 2).
  3. Nature of landuse change: The maps pinpointed the spatial locations where landuse changes have taken place. These landuse maps show the areal extent of landuse change (1988 to 1996) of a single particular category into other categories (Figure 2).

In order to establish the impact of future landuse change on the groundwater resources of a watershed, it is necessary to understand the existing groundwater conditions of the watershed and to quantify the extents to which the water resources will be modified. The hydrological factors influencing the landuse change are rainfall and groundwater conditions. The rainfall is considered to be one of the major factors in causing landuse change. Despite sufficiently high rainfall, droughts are a major problem owing to the uneven distribution of rainfall. As it is clear from the rainfall map, both the eastern and western parts of the watershed receive relatively a good amount of rainfall. However, comparing with the recharge map, groundwater recharge is relatively high on the eastern part due to the presence of dissected pediment, flood plain, alluvial fill and alluvial plain, which facilitate better recharge conditions. In the hard rock terrain in the western part, the suitability is poor. This can be attributed to the changes in the landuse along the eastern part where there is a sufficient conversion of wasteland into other categories of landuse. Whereas, in the western part, mostly dominated by wasteland due to the lithology being impermeable, there is little chance of recharge and most of the water flows as runoff.

The rainfall volume calculated for the years 1994-1996 shows a minor variation in rainfall volume, which in comparison with recharge volume, it was found that runoff has been more in the year 1994 than subsequent years, when the recharge increased sufficiently causing changes in landuse. This is supported by the fact that the fluctuation of water level along the eastern part of the watershed is due to the extraction of water for irrigation purposes.

The watershed shows a low and high fluctuation (0.5-1.9 and 6.3 m) along the central part and it gradually increases towards the southwest and southeast. The maximum fluctuation in the southeast is 4.9 m b.g.l (CSME, 1993). This fluctuation could be attributed to the presence of agricultural tracts in these areas where sizable extraction of water takes place. Further, it can also be justified by the landuse change image, where the eastern part of the watershed is mainly covered by fallow and arable land. Further, it has been observed that high groundwater recharge areas are associated with maximum landuse change areas (eastern part of the watershed), however, the western part of the study area though showing rapid landuse change but less recharge to groundwater regime. The reason for this could be the presence of highly impermeable zone mainly consisting of pink granite.

8. Conclusion
This study has successfully demonstrated an integrated remote sensing and GIS based methodology for evaluation of landuse-groundwater relationship in a hard rock area. This study shows the capability of integrated GIS in suitability analysis of artificial recharge sites in the area as well as establishing a relationship of landuse with consequent changes in groundwater regime. It further demonstrates the utility of IRS-1A-LISS-II and IRS-1C-LISS-III data in studying the landuse and the groundwater regime. Based on the above the following conclusions can be made:

  1. In this study, an integrated remote sensing and GIS technique has been developed for evaluating landuse-groundwater relationship.
  2. The present study has demonstrated that the change in landuse is mainly due to the hydrological factors.
  3. Weighted index overlay model has been found very useful in delineating groundwater prospective zones.

The integrated remote sensing and GIS based approach demonstrated by (Saraf and Choudhury, 1998) on basaltic terrain has been found completely applicable with minor modifications, however, geologically the present study area is mainly granitic terrain (GSI, 1997). Further, it may be concluded that the above technique may be applied to similar terrain conditions, with some local considerations and modifications.


  • Choudhury, P. R. 1999, integrated remote sensing and GIS techniques for groundwater studies in part of Betwa basin, Ph.D. Thesis (unpublished), Department of Earth Sciences, University of Roorkee, INDIA.
  • CSME, 1993, Groundwater Resources Assessment and Management of the Bankura District, West Bengal, A report by Center for Study of Man and Environment, Calcutta, 26p.
  • GSI, 1997, Georesource Map of Bankura District, West Bengal, Geological Survey of India, Calcutta, 1997.
  • Gustafsson, P., 1993, High resolution satellite data and GIS as a tool for assessment of groundwater potential of a semi-arid area, IXth Thematic conference on Geologic Remote Sensing, Pasadena, California, USA, 8-11 Feb. 1993, 609-619.
  • Krishnamurthy, J. and Srinivas, G., 1995, Role of geological and geomorphological factors in ground water exploration: A study using IRS-LISS-II data, International Journal of Remote Sensing, 16 (14): 2595-2618.
  • Krishnamurthy, J., Venkatesesa Kumar, N., Jayraman, V. and Manivel, M., 1996, An approach to demarcate ground water potential zones through remote sensing and a geographical information system, International Journal of Remote Sensing, 7 (12): 1867-1884.
  • Saraf, A. K., Gupta, R. P., Jain, S. K. and Srivastava, N. K., 1994, GIS based processing and interpretation of groundwater quality data, Proceedings of the Regional Workshop on Environmental Aspects of Groundwater Development, Oct. 17-19, 1994, Kurukshetra, India, II-9 – II-4.
  • Saraf, A.K. and Jain, S.K., 1994, Integrated use of remote sensing and GIS methods for groundwater exploration in parts of Lalitpur District, U.P., International conference on Hydrology and Water Resources, New Delhi, Dec 20-22, 1993, 251-259.
  • Saraf, A. K. and Choudhury, P. R., (1997), Integrated Application of Remote Sensing and GIS Groundwater exploration in hard rock terrain, Proceedings. Int. Symp. on Emerging trends in Hydrology, Department of Hydrology, Roorkee, September 25-27, 1997, Vol. I, 435-442.
  • Saraf, A. K. and Choudhury, P. R., (1998), Integrated Remote Sensing and GIS for Groundwater Exploration and Identification of artificial recharge sites, International Journal of Remote Sensing. 19(10), 1825-1841.
  • Saraf, A. K., A. Srivastava, B. Roy, B. S. Pradhan and P. R. Choudhary, (1999), Landuse Modelling in Geographic Information System for Dwarkeshwar watershed, Bankura district, Proceedings of National Symposium on Remote Sensing Applications for Natural Resources: Retrospective and Perspective organised by Indian Society of Remote Sensing, January 19-21, 1999,pp 99-100.
  • Shih, S. F., 1996, Integration of remote sensing and GIS for hydrologic studies, geographical information system in hydrology, pp. 15-42, Kluwar Academic Publishers, Netherlands.