Multi-temporal drought rarity curves - a yearly classification of meteorological drought severity in France

Droughts are recurrent phenomena, impacting eco- and socio-systems at varied temporal and spatial scales. Their impact depends on both the severity of the antecedent meteorological conditions and the recovery dynamics of the impacted systems. The drought severity analysis proposed in this study accounts for the ”memory effect” of rainfall accumulation by considering across time the rarity of antecedent precipitation at multi-temporal scales. It applies to rainfall accumulation over a single area. In this presentation, we define the yearly curve of multi-temporal drought rarity by the non exceedance probability of the smallest rainfall accumulations observed that year over a range of accumulation durations. Each rarity curve is thus defined by as many values as the number of durations considered. We apply this concept to droughts in France from 1950 to 2022, with accumulation durations varying from 4 weeks to 260 weeks. We show that the rarity curves are easy tools to summarize how droughts build and persists across time and temporal scales. We use an automatic classification of the curves to discriminate years associated to short- to long-term droughts (basically from half a year to five years). Although the concept is here used for rainfall over a single area, France, it could be applied as well to a set of areas and/or to other drought variables such as discharge or soil moisture.