Have streamflow droughts in Europe become more severe or frequent?

Changes in the magnitude and frequency of droughts will have extensive impacts on water management, agriculture and aquatic ecosystems. With the projected global temperature increase, scientists generally agree that the global hydrological cycle will intensify and suggest that extremes will become or have already become more common. In this study, a pan-European dataset of more than 600 daily streamflow records from the European Water Archive (EWA) was analysed to detect spatial and temporal changes in streamflow droughts. Four different time periods were analysed: 1962–1990, 1962–1995, 1930–1995 and 1911–1995. The focus was on hydrological droughts derived by applying the threshold level approach, which defines droughts as periods during which the streamflow is below a certain threshold. The Annual Maximum Series (AMS) of drought severity and the frequency of droughts in Partial Duration Series (PDS) were studied. Despite several reports on recent droughts in Europe, the non-parametric Mann–Kendall test and a resampling test for trend detection showed that it is not possible to conclude that drought conditions in general have become more severe or frequent. The period analysed and the selection of stations strongly influenced the regional pattern. For most stations, no significant changes were detected. However, distinct regional differences were found. Within the period 1962–1990 examples of increasing drought deficit volumes were found in Spain, the eastern part of Eastern Europe and in large parts of the UK, whereas decreasing drought deficit volumes occurred in large parts of Central Europe and in the western part of Eastern Europe. Trends in drought deficit volumes or durations could, to a large extent, be explained through changes in precipitation or artificial influences in the catchment. Changes in the number of drought events per year were determined by the combined effect of climate and catchment characteristics such as storage capacity. The importance of the time period chosen for trend analysis is illustrated using two very long time series. Copyright © 2001 Royal Meteorological Society