GIS-BASED STUDIES FOR ARTIFICIAL GROUNDWATER RECHARGE IN WESTERN GANGA PLAINS

Abstract

The Indo-Gangetic Plains, the land where our civilization has been nurtured, is a land of fertile soil, moderate climate and abundant water. These factors have combined to make this aregion ofplenty for human settlement. Groundwater is amajor source ofwater available for consumption in this area. However, over the years due to swelling population, increasing industrialization and expanding agriculture, the demand of water has increased manifold. Simultaneously, the available per-capita water resource has been reduced due to generally declining groundwater table. Hence, there is a tremendous need to take up augmentation measures. In the present work, asystematic study has been taken up for developing astrategy to replenish groundwater artificially in the study area, lying in the western part of the Indo- Gangetic Plains. The main objectives of the research work are: (a) application of Remote Sensing—GIS techniques to map spatial distribution of porous and permeable stretches, which happen to be parts of paleochannels of the Ganga river, (b) evaluation of hydrogeological characteristics of the paleochannel-aquifers and also the adjacent alluvial plains, (c) estimation of allowable recharge volume of the rechargeable aquifer, and (d) estimation of source water availability and planning for artificial recharge. The study area (between latitudes 29°10'N to 29°50'N and longitudes 77°30'E to 78°10'E) exhibits the characteristics of a river flood plain. The area slopes gently from north to south, at an average gradient of less than 0.38 mkm"'. Hydrogeologically it comprises extensive, multiple alluvial aquifer systems, composed of unconsolidated to semi-consolidated deposits of sand, clay and calcium carbonate concretions that constitute a good groundwater reservoir system. The following datasets have been used for the study: (a) Remote Sensing data: IRS- 1B LISS-II multi-spectral; (b) Survey ofIndia toposheets at 1:50,000 scale; (c) Soil map from National Bureau of Soil Survey and Land Use Planning, (d) other data such as specific yield, storage coefficient etc. collected from various sources, and (e) field data. The remote sensing data has been processed by using ERDAS Imagine-8.7 software. The GIS analysis has been carried out using ILWIS-3.3, ARCVIEW-3.2 and R2V software. Litholog data analysis is carried out by using ROCKWORKS-2006 software. A base map has been prepared from the Survey of India topographic maps by scanning, geo-referencing, mosaicking and digitizing. All the data layers have been coregistered with the base map. Point data obtained from field and laboratory experiment are properly placed on the base map and finally various information have been obtained using GIS tools. The IRS-IB LISS-II multispectral data have been co-registered with the base map, and corrected for atmospheric path radiance and striping, and enhanced for improving interpretability. The IRS-1B-LISS-II sensor data has been used as the primary data source -ft* to implement the supervised classification for generating landuse/landcover (LULC) map. Six LULC classes - agricultural land, paleochannel, dry streams, water body, built-up area, and marshy land, have been chosen and a LULC map has been generated with an overall accuracy of 87.9% using Maximum Likelihood Classifier (MLC). Finally, integrating information from colour infra-red composites and LULC map, a paleochannel map has been generated. The existence of paleochannels has also been cross-checked from litholog data and field observations. In the study area, three major paleochannels characterized by serpentine and meandering pattern, have been deciphered. The paleochannels are located to the west of the present day course ofthe river Ganga. Most ofthe paleochannels are very wide (2-5 km) suggesting their formation by a large river. Thus, it can be inferred that the Ganga river has gradually shifted from the west to the east. Field observations have revealed that the soils in the paleochannels are generally coarse sand. Rather sparse vegetation and low surface moisture over the paleochannel areas are indicative of highly permeable, porous, coarse grained materials possessing high infiltration rate. This is amply indicated by the spectral characteristics such as, very light tone in NIR band, and yellowish-white colour in CIR composites. The litholog data have been analysed to determine aquifer depth and lithological details. It has been observed that the paleochannel aquifer mainly consists ofcoarse sand occasionally mixed with pebbles, and boulders of varying sizes. On the other hand, the aquifers of adjacent alluvial plains are mainly composed of medium to fine grained sand along with clay and kankar beds. Construction ofsubsurface lithological cross-section, construction ofpaleochannel aquifer geometry and its inter-connectivity with the adjacent alluvial plains aquifer has been done by aggregating and synthesizing all the information, such as the lithological information, the base map, the CIR composite image, paleochannel map, well location map, and the DEM. The first aquifer (- 25-30 mthick) in the alluvial plains is unconfined and consists offine to medium sand with several lenses of clay and kankar. The second aquifer is confined in nature and mainly consists of fine to medium grained sand along with some lenses of kankar. The paleochannel aquifer is unconfined and is mainly composed of coarse sandy material along with boulder and pebbles beds. This paleochannel aquifer extends upto a depth of about 65 m and is well inter-connected with the adjacent alluvial aquifers. A series of 17 observation wells systematically sited on the paleochannel and its either flanks have been drilled and sampling has been carried out for collecting lithological information at different depths. Grain size analysis has been conducted for 82 samples. Based on grain-size analyses and the use of the Hazen (1911) equation, the bulk hydraulic conductivity for selected core samples at different depths has been estimated. It is found that the value ofhydraulic conductivity ranges from 30 to 75.3 m/day for samples falling in the paleochannel, and that between 13.5 to 22.3 m/day for the alluvial plain aquifers. The natural groundwater recharge rate due to precipitation has been estimated using tritium tagging technique. Comparison of recharge rates and hydrogeologic characteristics in different landforms indicates that: (a) paleochannel area have coarse grained soil (sandy loam) and high recharge rate of 18.9 to 28.7%, and (b) the alluvial plains have medium to fine grained soil (silty loam)and relatively lower recharge rate (6.3 to 8.9 %). Stable isotopes of groundwater samples from the first unconfined aquifer of the study area have been analysed using Dual Inlet Mass Spectrometer. The study indicates that the alluvial plains aquifers get recharged dominantly by canal and/or rainfall. The data also indicates that the influence of canal water to groundwater recharge decreases away from the canal, where rainfall recharge component relatively increases. It is also inferred that rainfall/precipitation constitute the dominant source for groundwater recharge in paleochannel aquifers. Groundwater levels have been monitored at 37 locations (12 on the paleochannel aquifers and 25 on the adjacent alluvial plains aquifer) over 2 years (2005-2006) for both pre- and post-monsoon period. The precise locations (x, y, z co-ordinates) have been determined through differential GPS. Interpretation have been made by combining information of paleochannel map, reduced water level contour map and flow direction map. It has been observed that groundwater flows away from the paleochannel in both preand post-monsoon period, which is related to the high hydraulic conductivity and porosity of the paleochannel aquifer. This further indicates that recharging of groundwater through paleochannels would lead to gradually recharging ofthe aquifers in the alluvial plains. For estimating the rainfall runoff in the three watersheds in the study area, the Soil Conservation Service Curve Number (SCS-CN) method has been used. Further, for planning artificial recharge of the paleochannel aquifer, its storage potential has been estimated, and the value is found to be 89.5 x 106 m3. On the other hand, the volume of water required for arresting the decline ofgroundwater table over the three watersheds is estimated as 34.6 x 106 m3. Sources of water considered for artificial recharge are rainfall runoffand the canal water. A flow accumulation has been generated from DEM in GIS. Considering the various aspects, an integrated three-stage planning using rainfall runoff water and barely 1% ofcanal discharge in the lean water demand period (July-September), has been suggested - that would be sufficient to meet the requirement ofartificial recharge and arrest the declining groundwater table in the area.