Stress-testing the buffering role of glaciers in the Rhine basin: How much worse could summer low flows get under future glacier retreat?

Past drought years, characterized by large scale precipitation deficits and high summer temperatures, such as 2018 in Europe, have resulted in extreme low flow situations with negative ecological, economic and social consequences. In large river basins that originate in high mountain ranges, such as the Rhine river originating in the Swiss Alps, glaciers and snowpack alleviate drought situations by continuously providing meltwater to downstream reaches during summer. However, this meltwater contribution is under threat due to ongoing climate warming and retreating glaciers. Here, we designed a stresstest model experiment to answer the question ‘what if a historical drought year reoccurs in future conditions with retreated glaciers?’ A model framework was used combining the HBV model for the glacierized headwater catchments of the Rhine and the LARSIM model for the rest of the basin. Three historical drought and low flow years, 1976, 2003 and 2018, were selected and their meteorological conditions were transferred and used as stresstest model input to three future conditions in time, namely now (2018), near future (2031) and far future (2070). These three model states were obtained by transient simulations up to the respective moment in time using meteorological observations or an ensemble of bias-corrected climate model output from the RCP 8.5 scenario, using coupled glacio-hydrological model runs. The results show an aggravation of downstream low flows, especially when drought years happen under conditions in the far future. From the three years, 2003 has the strongest effect in the future, because the ice melt contribution was highest in that year in the past. During August, flows reduce up to 80% upstream for highly glacierized catchments, compared to the streamflow of the original drought years. At downstream gauges, where flows were already critically low in the past, streamflow reduces by 5%-20%. This model experiment shows a glimpse in future low flow events and emphasizes the importance of upstream cryospheric changes for downstream streamflow dynamics during drought.