The project deals with the development and improvement of statistical methods for the analysis of extreme weather events. The focus is on the topics of “serial clustering”, “homogeneity analysis” and “improved inference for block maxima”. Serial clustering refers to the accumulated occurrence of several extreme events within short periods of time. Methods from the first funding phase will be expanded, for example by taking seasonalities and temporal trends into account or by using multivariate observations. Homogeneity analysis refers to the application of statistical methods to check whether observations from several stations or from different time periods can be regarded as statistically homogeneous. Homogeneous observations can then be pooled in further analyses to increase statistical accuracy by enlarging the database; an approach that is frequently used in event attribution studies. The project is working on the development, improvement and application of such methods. Motivated by promising results in the first funding phase, work is also being carried out on the methodological development and application of the sliding block maxima method for the statistical analysis of extreme events.

Website: SCAHA
Institutions:  Department of Statistics, TU Dortmund University¹; Faculty of Mathematics, Ruhr University Bochum²
Contact: Prof. Dr. Roland Fried¹, Prof. Dr. Axel Bücher²