Modeling Extreme Floods Susceptibility Using The Generalized Extreme Value Distribution: Case Study Of Gonse And Wayen, Burkina Faso.

Abstract

Over recentes decades, Burkina Faso has experienced extremes events such as droughts and floods. In this study, flood frequency has been ascertained based on Generalized Extreme Value (GEV). To this end, discharge data from Gonse and Wayen stations are collected from the National Center for Water resources. The period of analysis goes from 1980 to 2022. The Kolmogorov-Smirnov test is applied to check the distribution of the time series. Then, the Maximum Likelihood Estimation (MLE) method is implemented to estimate the location, the scale and the shape parameters of the GEV distribution. The goodness-of-fit between the empirical data and the theorical distribution is then evaluated based on Akaike criterion (AIC) and Bayenan criterion (BIC). The results revealed that across Gonse station, the probability that the annual maximun discharge will be less than 30m3/s is 0.7 and the 50-year return period discharge is 37.33 m3/s. In Wayen station, the probability that the annual maximun discharge will be less than 200m3/s is 0.5 and the 50-year return period discharge is 226.38m3/s. The AIC is 308.10 and 484.61 respectively for Gonse and Wayen station. The BIC is 313.65 and 490.16 respectively for Gonse and Wayen station. The findings may provide a scientific base for managing the risks of floods to advance climate change adaptation over the Nakambe watershed.