Characterizing rainfall-groundwater dynamics in a hard-rock aquifer system using time series, GIS and geostatistical modeling

Abstract

The aim of this study was to investigate rainfall-groundwater dynamics over space and annual time scales in a hard-rock aquifer system of India by employing time series, GIS and geostatistical modeling techniques. Trends in 43-year (1965-2007) annual rainfall time series of ten rainfall stations and 16-year (1991-2006) pre- and post-monsoon groundwater levels of 140 sites were identified by using Mann-Kendall, Spearman rank order correlation, and Kendall rank correlation tests. Trends were quantified by Kendall slope method. Furthermore, the study involves novelty of examining homogeneity of pre- and post-monsoon groundwater levels, for the first time, by applying seven tests. Regression analysis between rainfall and post-monsoon groundwater levels was performed. The pre- and post-monsoon groundwater levels for four periods: (a) 1991-1994, (b) 1995-1998, (c) 1999-2002, and (d) 2003-2006 were subjected to GIS-based geostatistical modeling. The rainfall showed considerable saptio-temporal variations, with a declining trend at the Mavli rainfall station (p-value<0.05). The Levene’s tests revealed spatial homogeneity of rainfall at α = 0.05. Regression analyses indicated significant relationships (r2>0.5) between groundwater level and rainfall for eight rainfall stations. Non-homogeneity and declining trends in the groundwater level, attributed to anthropogenic and hydrologic factors, were found at 5 to 61 more sites in pre-monsoon compared to post-monsoon season. The groundwater declining rates in phyllite-schist, gneiss, schist and granite formations were found to be 0.18, 0.26, 0.21 and 0.14 m year-1 and 0.13, 0.19, 0.16 and 0.02 m year-1 during the pre- and post-monsoon seasons, respectively. The geostatistical analyses for four time periods revealed linkages between the rainfall and groundwater levels.