Non-stationary Pointwise and Spatial Approaches by Peaks-Over-Thresholds. Applications to climate warming in Burkina Faso

In this study, trends in temperature extremes in Burkina Faso are evaluated using the statistical theory of extreme values, in particular the Peaks-Over-Thresholds in a non-stationary context. The first issue addressed in this study is the temporal evolution of extremes handled pointwise by the approach of threshold exceedances. The trend in parameters of the distribution of excesses and the intensity of the occurrences of extreme events for temperature variables are described as linear or quadratic functions of time. Then the spatial dimension is integrated using a geostatistical approach. From the extrapolation of the identified trends, non-stationary return levels and their confidence intervals are also calculated and non-stationary return level maps are produced. Finally, our approach is applied to temperature data in Burkina Faso.