Sensitivity analysis of time-dependent generalized extreme value models for ocean climate variables

Recent advances in the description of the natural variability of extreme events associated with ocean climate variables include time-dependent variations within a certain time scale (year, season or month). These models allow incorporating smooth time variations of the parameters of the GEV distribution and also the influence of covariates (NAO, El Niño, etc.), providing more reliable results than traditional stationary models. In this paper, a sensitivity analysis of the model parameter estimates with respect to data (selected maxima, time of occurrence and covariates) is presented. The method shows how different data affects location, scale, and shape parameters as well as return level quantiles; gaining an insight into time-dependent generalized extreme value models and increasing the understanding of the models responses. The proposed method is applied to the study of the monthly maxima significant wave height at a particular location around the Brittany coast.