Earlier emergence of high-flow changes at higher elevations

The magnitude, frequency and spatial extent of hydrological extremes are changing because of climate change. However, the sign and magnitude of projected future changes in high flows remain uncertain in many regions, including Switzerland, because of internal climate variability, which causes streamflow fluctuations on annual to decadal timescales. To disentangle the changes in high flows that can be attributed to climate change from those related to internal variability, one needs to quantify this uncertainty. For this task, one can use single-model initial-condition large ensembles (SMILEs), which are climate models composed of different members representing equally plausible realisations of the climate. In this study, we use the climate variables projected by a 50-member SMILE at the daily time step as inputs to a hydrological model to project future changes in high-flows for a large set of catchments in Switzerland. We calculate changes in high streamflow quantiles and changes in the seasonality of maximum annual streamflows to investigate the changes in high flows between current climate and future conditions until the end of the century. We then calculate the signal-to-noise ratio and the time-of-emergence to determine where and when these changes are significant. We show that under the RCP8.5 scenario (1) high flows are likely to decrease at high elevations and increase at low elevations; (2) annual streamflow maxima are projected to occur earlier at high elevations and later at low elevations; and (3) the high flow change signal emerges from internal climate variability before 2050 at high elevations and after 2050 at low elevations. Our findings will likely have implications on flood frequency estimation in the alpine region.