Simulation of Groundwater Levels Using Tank Model with Consideration of Mixed Hydrological Structure in Kumamoto City

Abstract

Groundwater level is a major parameter in quantifying groundwater resource. Application of simple tools to simulate and predict groundwater levels is an important issue in groundwater hydrology. Tank model is a simple and useful conceptual hydrological model which is capable to simulate different system states without considering complex hydrological structure. The Shuffled Complex Evolution-University of Arizona (SCE-UA) algorithm is an efficient automatic calibration algorithm in searching the global optimum of objective function. This paper investigated the capability of mixed-tank model for consideration of mixed hydrological structure to simulate the groundwater level fluctuation with precipitation in four observation wells in Kumamoto City. The SCE-UA algorithm was introduced and modified to minimize the root-mean-square error (RMSE) between simulation and observation data for parameter calibration. The simulation accuracy was also assessed by Nash-Sutcliffe efficiency (NSE) coefficient. The results indicate that the mixed-tank model is highly effective for the simulation of groundwater levels showing different behaviors in their fluctuation situations.