Ouémé River Catchment SWAT Model at Bonou Outlet: Model Performance, Predictive Uncertainty and Multi-Site Validation

Berenger Arcadius Sêgnonnan Dègan

doi:10.11648/j.hyd.20180602.13

Abstract

In a soudano-guinean climate context, the Ouémé River Basin is simulated using the semi-distributed hydrological model SWAT to understand the rainfall-runoff process on this basin and also to assess this model performance on West Africa large areas basins at daily and monthly time steps. The inputs data consist of climatic data and rain gauge discharge records. The inputs records are long-term times series for the period 1979-2010, while the considered land use is just for the year 2003. After calibration and validation of the model, spatial calibration is also performed to appreciate this other feature of the model. It gives such acceptable and disputable results. Six (06) hypotheses have been emitted to analyze this performance loss. It comes out that hypothesis H5 results perform better both in calibration and validation. This hypothesis used data for the period 1993-2010 with 1993-2004 for calibration and 2005-2010 for validation; and considered the missing data in discharge records without any completion. Considering the internal rain gauge outlet performance for this hypothesis, the best is retained and the corresponded project is realized for each individual subbasin to see how best the model could simulate discharge for the Bétérou, Kaboua and Atchérigbé individual subbasins. Hypothesis H1; an assumption which considers missing discharge with data time period of 1982-2010 with 1982-1996 for calibration and 1997-2010 for validation; is the best for Bétérou and Kaboua, whereas H5 is better for Atchérigbé subbasin. Uncertainty analysis and Global Sensitivity Analysis were performed to appreciate what are this process occurring in the basin and how these results could be validated. A last comparison effort is performed with 10km rainfall grid for climatic rainfall data at the global catchment outlet; this approach does not improve results, while at internal outlet some improvements are observed.

Highlights

World is nowadays affected by climate unmastered laws
Global Model Performance According to the results (Table 3), Soil and Water Assessment Tool’’ (SWAT) performs well to simulate discharge in calibration compared to the validation on this time period
SWAT-CUP produces output results at each station as 95PPU as well as showing the best fit, but for simplicity and clarity of presentation we only show the calibration/validation results for the Bonou basin, the individual sub basins and the final retained hypothesis simulation and report the overall statistics

Summary

Introduction

World is nowadays affected by climate unmastered laws. This situation leads to serious modification of human beings. During the Berenger Arcadius Sêgnonnan Dègan et al.: Ouémé River Catchment SWAT Model at Bonou Outlet: Model Performance, Predictive Uncertainty and Multi-Site Validation past decade, several countries have been severely affected in Asia (China, Bangladesh and Vietnam in 2002 for example), Europe (France, Germany, Hungary and the Czech Republic in 2000) and elsewhere in the world (Venezuela in 1999, Canada, 1996, United States, 2005, etc.). It is the most frequent natural hazard arising from meteorological conditions and affecting the greatest number of people on the planet [2]. These Floods sweep away agricultural crops, fields, houses, and socio-economic infrastructure, making many climate refugees

Objectives

Methods

Results

Discussion

Conclusion