Clustering of Groundwater Wells and Spatial Variation of Groundwater Recharge in Sina Basin, India

Abstract

A spatial and temporal analysis of groundwater levels, topography, and precipitation is required to properly manage the groundwater resource. The present paper explains it in two parts: (1) spatial analysis of groundwater levels and selection of suitable clustering approach for selection of representative wells and (2) spatial and temporal variations of groundwater recharge calculated by three numerical models: Chaturvedi model, Amritsar model and ERAS model. Four clustering techniques including K-Means clustering algorithm, Hierarchical clustering technique, canopy and expectation maximisation (EM) were used for the clustering of groundwater levels. Among these, the canopy technique presents more reliable results compared to the other techniques for the spatial analysis of groundwater levels and the formation of representative wells in the Sina basin. For the groundwater recharge estimation, Chaturvedi model and ERAS model values were found very close. The recharge values show consistency with the precipitation data and found that 15% of precipitation contributes to annual groundwater recharge. Spatio-temporal variation of groundwater recharge correlated with precipitation is also carried out for the selected basin. The results show a drastic decline in the groundwater recharge from the year 1990 to 2008. An empirical expression is also developed for groundwater recharge estimation in terms of groundwater level. This provides regional scale information about the basin and helps to understand the groundwater exploitation scenario for instance.