Application of Parity Classified Neurogenetic Models to Analyze the Impact of Climatic Uncertainty on Water Footprint

Abstract

Water footprint of an individual, community, or business is defined as the total volume of freshwater that is used to produce the goods and services consumed by the individual or community, or produced by the business. Neurogenetic models were widely used in the prediction of hydrologic variables, and outcome of such applications were found to be satisfactory. The irregular rainfall and temperature pattern, and degradation of watersheds were causing worldwide reduction of water availability (UNFCC). As water footprint is directly related to water availability and also shows the demand from industrial consumers, the present study tried to estimate the impact of climate change on water footprint between two river basins of East India with the help of neurogenetic models. The climate change scenarios were generated with the help of PRECIS climate models, and future runoff was estimated by a neurogenetic model trained with orthopareto dataset. The output from the neurogenetic model, named as PARITYCGD, was compared with a neurogenetic model trained with normal dataset (NGHYD) and conceptual hydrologic models. According to the results, the neurogenetic model trained with orthopareto dataset was selected as the better model among the five models, which shows that neural models trained with orthopareto dataset learn a problem better than a neurogenetic model trained with normal dataset. From the prediction of stream flow, water footprint of the sampling regions were calculated and according to the estimations, water footprint would be reduced in both A2 and B2 climate change scenarios where reductions would be more pronounced in A2 than in B2. Although, due to data dependency of neurogenetic models, the PARITYCGD model may not work for other basins but for the present study, it was found to have better accuracy than the conceptual hydrologic model.