Modeling Ground Water Dynamics Using Artificial Neural Network ANN ground water flow model for Kanpur Districts Uttar Pradesh

Abstract

Ground-water system structures are dynamic and changes frequently in short and long newlinedurations due to deviations in climate, ground-water extraction patterns, and land use of any newlinespecific region. Ground water level measurements, monitoring and analysis are the principal newlinesource of information that how hydrologic stresses acting on aquifers and howit is affecting newlinegroundwater inflow, outflow and change in storage. Long-term monitoring of ground water newlinelevels provides essential data needed to evaluate temporal changes in the resource, to develop newlineground-water models and to forecast trends which helps to design, implement, and newlinemonitorthe effectiveness of ground-water management and protection strategies and newlinedatabase. Modeling of groundwater dynamics helps to assess prevailing status of ground newlinewater resources of any area and the impact of the on- going ground water management newlinepractices on the groundwater resources. newlineThis study had been planned to develop, calibrate and validate MLP based ANN models for newlinepredicting future trends of ground water level fluctuations at 10 blocks of Kanpur District in newlineorder to utilize groundwater resources optimally and to suggest best groundwater newlinemanagement practices in the study area. Ground water Level data for the duration 1998-2016 newlinefrom 50 wells of different blocks (Kakwan, Bilhaur, Ghatampur, Shivrajpur, Chaubeypur, newlineKalyanpur, Vidhnu, Sarsaul, Bhitargaon and Patara) of Kanpur district were collected and newlineused for the modeling purpose. Out of total 18 years data, Fifteen years were used for testing newlinewhile three years data were used for validation of the developed models. Based on the newlineGamma Test results and standard errors, various models of different architectures were newlinedeveloped considering all input combinations affecting the ground water fluctuation (GWF). newlineThe best input combination found was the current session rainfall (RF), current session newlineeffective rainfall (ERF), and one and two session lag GWF as inputs variables to predict newlineground water fluctuations (GWFs)as output va