Abstract:
Within the context of climate change, the hypothesis of the stationarity of observed datasets for performing classical flood frequency analysis is no longer valid. We explore the use of see surface temperature (SST) and see level pressure (SLP) as covariates for modelling the annual maximal discharges (AM) at 5 gauging station of the Ouémé basin. Significant correlations (at 5% level) were found between the AM and the SST/SLP of the Gulf of Guinea. Non-stationarity was introduced to the generalized extreme value (GEV) distribution using a linear function of the location and scale parameters. Different combinations of the model parameters were explored with the stationary model based on three criteria of goodness of fit. The non-stationary model superior to others and explains a substantial amount of variation in the data. The good correlations found provide a possibility of using these climate indexes as predictors for flood early warning system implementation.
