Spatio-temporal consideration of the impact of flood event types on flood statistic

Abstract

Flood events can have very different generating processes. Floods may originate from high intensity rainfall, long-duration rainfall or snowmelt. Other factors, e.g. initial soil moisture conditions, also affect the spatial and temporal characteristics of floods. Hence, a typology for floods is often used to classify floods according to their runoff generating processes. Having in mind the changing circumstances for floods over the years, like climatic or anthropogenic changes, the question arises in how far these changes affect the occurrence of certain flood types. This does not only concern a change of frequency of flood event types over time, but also spatial changes resulting in different regional patterns. We investigate the changes of the frequency of occurrence of flood event types, their spatial patterns and the consequences for estimating extreme events. Many watershed characteristics are strongly related with specific flood types, e.g. elevation with snowmelt. Short-rain events are limited in their spatial extension and cause floods in smaller catchments, whereas in large river basins a large flood peak requires a long and spatially extended rain event. In this study, the flood typology is based on analyses of the peak-volume relationship allowing to consider long periods of observation with partially only few meteorological data, e.g. snow height. For the resulting flood types, temporal changes and spatial patterns are detected and connected with the generating processes as well as catchment characteristics. Finally, the impact of changes on the flood statistic is investigated by using a seasonal mixture model of partial duration series that is able to model event types and seasons separately and combines them in an annual model.