ASSESSMENT OF GROUNDWATER REGIME USING GEOSPATIAL AND FLOW MODELING TECHNIQUES

Abstract

Groundwater is a natural and most widely distributed precious resource upon which all-social and economic activities and ecosystem depends. In spite of its importance, it is often misused and poorly managed. The main objective of this research is to understand the groundwater system and its behavior in response to stresses using geospatial techniques, particularly remote sensing and GIS, and 3-D groundwater flow modeling towards providing scientific input for groundwater management. The study area in the present research work covers Unnao district, Uttar Pradesh, India, which is characterized by an alluvial aquifer system. The groundwater resource is poorly managed in the study area. Many places suffer from the problem of declining groundwater levels, while water-logging and soil salinity problems also exist due to seepage from canals and excessive canal irrigation. The research is divided into six parts: (1) understanding hydrogeological framework; (2) spatio-temporal analysis of groundwater levels; (3) groundwater resource assessment; (4) delineation of groundwater potential zones; (5) groundwater flow modeling; and (6) designing a web-based geoportal for visualization and query of the groundwater and ancillary information. To understand the hydrogeological framework, lithologs of 10 exploratory wells drilled by Central Ground Water Board (CGWB), Uttar Pradesh, and water levels in the observation wells (45 no.) of CGWB were analysed. The maximum depth of these exploratory wells is 400 meter below ground level (mbgl). Based on the disposition of aquifers and aquitards/ aquicludes, it is found that the area is mainly characterized by the presence of three aquifers. The first (top) aquifer extends upto 90 mbgl; the second (middle) aquifer occurs from100 m to 250 mbgl; and the third aquifer extends from 200 m to >400 mbgl. These three main aquifers are separated by impermeable clay beds. While the first aquifer is mainly unconfined, the second and third aquifers are confined in nature. The water levels in the shallow wells suggest that a subtle groundwater divide running NW-SE exists in the area dividing the groundwater flow system into two sub-flow systems, one draining towards Ganga River (i.e. towards west) and the other towards Sai River i.e. (towards east). The long-term trend analysis of groundwater levels has been carried out using non-parametric Mann Kendall’s test, remote sensing and GIS. A Network of 45 observation wells has been used to explore the well-wise trend in the study area. The Mann Kendall’s test indicates downward trend in groundwater levels at many of the wells, while upward trend is observed at few locations. In the pre-monsoon season, about 49%, 53% and 58% wells showed iii downward trends and about 11%, 13% and 16% wells showed upward trends at 1%, 5% and 10% significance level, respectively. In the post-monsoon season, about 49%, 56% and 58% wells showed downward trend and about 11%, 13% and 13% wells showed upward trend at 1%, 5% and 10% significance level, respectively. The rate of the trend has been also estimated using Sen Slope estimator. The results are further verified with the satellite images of 1986 and 2009, indicating that the salt-affected and water-logged areas have decreased over time, which may be attributed to increased usage of groundwater. These results provide useful information to ensure sustainable groundwater development. Groundwater resource assessment has been carried out using the methodology and norms suggested by Groundwater Estimation Committee (1997). Water table fluctuation (WTF) and rainfall infiltration factor (RIF) methods are used to estimate monsoon recharge, whereas non-monsoon recharge is calculated using RIF method. Groundwater recharge for the entire district is estimated as 1.64 BCM (billion cubic meters), about 40% of this (i.e. 0.67 BCM) is contributed by monsoon rainfall recharge. The net annual groundwater availability is 1.48 BCM, while annual groundwater draft is 1.28 BCM. Thus, the stage of groundwater development is 86%. On the basis of long-term trend analysis and stage of groundwater development, it is found that the Sikandarpur Karan and Purwa Blocks (administrative units) fall under over-exploited category. Groundwater potential zones have been delineated using remote sensing, GIS and MCDM techniques. Of the many techniques available in literature, Analytical Network Process (ANP) has been used. ANP provides a method for input judgment and measurement to derive ratio scale priorities for the distribution of influence between the different factors. Five categories of groundwater potential zones have been delineated namely, 'poor', 'moderate', 'good', 'very good' and ‘excellent’. Verification of groundwater potential zone map using the well yield data of 37 pumping wells provided satisfactory results. Using the results obtained from the above research components, a groundwater flow model has been developed using PMWIN (Processing Modflow for Windows) software. Initially, the model was calibrated for steady state. The results were fed as input to transient state and then the model calibration has been done for a period of 8 years (1995-2002) and validation for a period of 5 years (2003-2007). Hydraulic conductivity and specific yield were essentially calibrated. The results obtained from the calibrated model were used to predict future changes in water levels in different scenarios. In the first scenario, under steady state, increasing the groundwater abstraction by 25% indicated decrease in groundwater level in the iv central and central-western parts of the study area (average decline estimated to be ~1.45 m); however, it is observed that such an increase in abstraction does not lead to any change in the groundwater flow system. The second scenario has been generated for the year 2025 under transient state wherein groundwater abstraction rate has been assumed to be same as that at present. It is observed that a maximum drawdown of 4.5 m will take place in central-western part of the study area with an average decline of 1.62 m without affecting the existing overall groundwater flow system. Another important outcome of the groundwater flow modeling was to obtain calibrated, spatially distributed maps of hydraulic conductivity and specific yield of the aquifer. On the basis of sensitivity analysis, it has been found that hydraulic conductivity is more sensitive as compared to specific yield. The final part of the research-involved development of a geoportal, which is a webbased application. Geoportal allows the users of the system to share and explore the information related to administrative boundaries, geography, demography, groundwater scenario, groundwater recharge, hydrological parameters etc. The developed application is hosted in internet domain. The geoportal can be used simultaneously by a large number of users, as the knowledge of basic software is not necessary to handle web-GIS based modules. This research has provided a holistic understanding of the groundwater system and its behavior of a typical alluvial aquifer system. The calibrated groundwater flow model and the hydrogeological information generated through this research can be used to simulate impact of human activities on groundwater levels and flow regime for formulating sustainable groundwater resources development scenarios and disseminating them to the public and decision-makers. Presently over-exploited and the potential areas where decline in groundwater levels is modeled need attention of the water resource managers to regulate the groundwater pumping and undertaking artificial recharge measures in critical zones based on detailed ground-based hydrogeological investigations. Similarly, in the water-logged and saltaffected areas, conjunctive use of canal and groundwater irrigation is recommended. A better network of observation wells (including automated piezometers in critical areas) and more pumping tests are required to better constrain the aquifer parameters and calibration datasets so as to help improve the model.