OPTIMIZATION OF RUNOFF PARAMETERS OF A WATERSHED BY GENETIC ALGORITHM AND VISUALIZING THE RUNOFF IN 4D GIS ENVIRONMENT

Himanshu Panwar, K. Venkata Reddy

Abstract: The Genetic Algorithm has been integrated with Finite Element Method (FEM) based distributed rainfall runoff model. It optimizes the four parameters i.e. saturated hydraulic conductivity, initial water content, suction head and saturated water content which generates the simulated runoff, which is then compared with the observed runoff data of actual site. The model has been applied to Harsul watershed located in Nashik district of Maharashtra, India. The objective function used in Genetic Algorithm is Nash-Sutcliffe Model Efficiency Constant objective function, and this combination produced good results. It performed well for many rainfall events except for those which have high rainfall intensity and low runoff. The Genetic Algorithm proved to be capable and robust in terms of maximising the objective function. Within 50 generations of 10 populations each, it has provided satisfactory results. GIS has been used to prepare all the thematic maps of the study area. The simulated runoff data generated from the model has been imported to GIS environment where it is visualized dynamically in 3D format with time series animation. The methodology developed in the present study can be used to simulate flow parameters of watershed.

Keywords: FEM based distributed rainfall-runoff model, Genetic Algorithm, Nash-Sutcliffe Model Efficiency Constant, 4D GIS animation

DOI: https://doi.org/10.15623/ijret.2015.0423013