Data driven parameter estimation and Application of two Storage Based Runoff models on Oshin Catchment using Mod16 Evapotranspiration datasets

Abstract

This research aims to apply two storage-based rainfall-runoff models which have been proven and used in other catchments/watersheds, such as the  Australian Water Balance Model and the Sacramento storage-based Runoff Model, to the Oshin catchment area of Kwara State, Nigera. This is to  determine the applicability of the storage-based runoff models which were developed in watersheds of other countries on a Nigerian watershed.  Previous researches focused more on physically-based models, but none has focused on an optimization of the parameter datasets making a  comparative evaluation of the runoff parameters lacking. This paper presents for the first time, the parameter optimization of Australian Water Balance  Model and SACRAMENTO runoff parameters and a comparative evaluation of the efficiencies of the European-developed storage-based runoff models in  runoff estimation on the Oshin Watershed and it sensitivity analysis. The SACRAMENTO model outperformed the AWBM model with a NASH Sutcliffe  Criterion of 0.753 during calibration and a Nash Sutcliffe of 0.742 during validation, while the AWBM had a Nash Sutcliffe of 0.517 during calibration and  0.423 during validation. The AWBM model had two parameters that were sensitive during optimization using pattern search algorithms (BIF and Kbase).  During optimization trials, the Sacramento had none of its parameters sensitive. Therefore, the applicability of both the rainfall-runoff models were  confirmed, with the SACRAMENTO as the most suitable for the catchment.