Assessment of Rainwater Harvesting Potential on Groundwater Recharge in Vellore City, India.

The study suggests the importance of inclusion of rainwater harvesting in Groundwater level in the state of Tamil Nadu, India. To serve the purpose, groundwater levels of two observation wells in the Katpadi and Vellore City blocks of Vellore district, Tamil Nadu, India were modeled using two different approaches of Multiple Linear Regression (MLR) and Neural Network Analysis (ANN). Different input parameters were selected on the basis of previous studies and cross correlation analysis (CC). Volume of rainwater harvested was calculated by employing image classification technique on different images of the study area obtained from Google Earth’s historical database. During CC volume of rainwater harvested emerged as an important parameter. Results of MLR model suggested inclusion antecedent rainfall conditions produced the best results. ANN modeling suggested that present condition and absence of antecedent conditions produced the best model. In comparison, on the basis of coefficient of determination ANN produced better model than MLR analysis. Considerably better results were obtained for more recent years suggesting that rainwater harvesting may have reduced fluctuations in groundwater levels. The study presented itself as the motivation for other states to follow the successful scheme of Tamil Nadu government of mandating rainwater harvesting in all the structures.