On the temporal variability of widespread rain-on-snow floods

Abstract

Floods which affect several macro-scale river basins simultaneously can cause devastating damage. Future flood-risk assessment depends significantly on the knowledge about the atmospheric conditions leading to floods and external climate drivers. Nonetheless, only a few studies have investigated widespread floods, their occurrence frequency, and their underlying hydro-meteorological processes. We extended the time period analysed in previous studies to detect potential climate drivers of so-called trans-basin floods in Germany during the twentieth century. The hydro-meteorological flood evolution was examined based on climate station data and dynamically downscaled ECMWF twentieth century reanalysis (ERA‑20C). Most of the classified trans-basin floods were caused by rainfall during extensive snowmelt, so-called rain-on-snow events. We investigated the interannual to decadal variability of such rain-on-snow events by building an atmospheric proxy based on snow cover and total precipitation. As expected, the temporal variability of our proxy largely corresponds to flood-rich and flood-poor periods. Based on this, we analysed potential climate drivers of such flood events. We found a negative correlation of the Scandinavian pattern (SCA) and total precipitation, as well as a negative correlation between the North Atlantic Oscillation (NAO) index and large snow cover. The latter is supported by the fact that snowfall in advance of the floods was often associated with a cyclone track over the Mediterranean Sea which often corresponds to the negative NAO phase. Overall, spatially extended floods in Germany occurred during the 20th century predominantly, when the superposition of NAO and SCA index was negative.

Keywords

• trans-basin floods
• rain-on-snow
• twentieth century reanalysis
• north atlantic oscillation
• scandinavian pattern
• flood occurrence frequency