Univariate modelling of multivariate precipitation extremes using the block quantiles method

Abstract This paper presents the approach of Team Moscow State to solving the tasks of the EVA (2025) Conference Data Challenge. Although the challenge tasks were originally formulated in a multivariate setting, they can be reformulated as problems of univariate statistics of extremes, which largely determined our approach to their solution. It relies on a unified framework that combines classical techniques of statistics of extremes with the recently proposed block quantiles method.