Spatiotemporal clustering of flash floods in a changing climate (China, 1950–2015)

Abstract. The persistence over space and time of flash flood disasters – floods that have caused either economical losses or loss life both is a diagnostic measure areas subjected to hydrological risk. concept can be assessed via clustering analyses, performed here analyze the national inventory in China occurred period 1950–2015. Specifically, we investigated spatiotemporal pattern distribution their behavior by using global local methods: first based on Ripley's K function, second scan statistics. As result, could visualize patterns aggregated events, estimate cluster duration make assumptions about evolution time, also with respect precipitation trend. Due large spatial (the whole Chinese territory) temporal (66 years) scale dataset, were able capture whether certain clusters gather specific locations times but magnitude tends increase decrease. Overall, eastern regions are much more compared rest country. Detected revealed these phenomena predominantly occur between July October, coinciding wet season China. number detected increases associated drastically decreases recent period. This may indicate change towards triggering mechanisms which typical short-duration extreme rainfall events. Finally, being directly linked realization, indirectly trigger itself has varied through enabling considerations context climatic change.