Extreme water temperatures in mountain rivers – changes and driving processes

Abstract

Water temperature is one of the most important indicators of water quality as it regulates physical, chemical and biological processes in rivers. When water temperatures reach extreme values, this can have potentially severe consequences for the survival of aquatic ecosystems. Extreme water temperatures can be caused by extreme weather phenomena such as heat waves and prolonged droughts. In mountain regions, the complexity of water temperature dynamics is greater than in lowland regions due to changes in the hydrological regime caused by glacier retreat and changes in the contribution of snowmelt to streams. Despite the potential impacts of water temperature extremes, knowledge of the occurrence and driving processes of water temperature extremes in mountain rivers remains limited. Here, we aim to improve our understanding of the spatial and temporal variability and long-term changes in the occurrence of extremes. In addition we aim to identify the main processes influencing the occurrence of water temperature extremes in mountain rivers in Europe.  First, we compare 30 years of water temperature data in 18 catchments in the Alps to gain insight into the temporal variability of water temperature extremes. We examine the seasonality of these extremes and use trend tests to assess long-term trends. Second, we compare 177 catchments across four different mountain regions in Europe to understand the frequency, severity and variability of water temperature extremes at a regional scale. Finally, we use random forest models to investigate the importance of different processes contributing to water temperature extremes and how the main driving processes vary in both time and space. The results of the trend analysis in the Alps show that extreme water temperatures, i.e. water temperatures exceeding a locally varying threshold, have increased faster than mean water temperatures during the summer period of 1991-2021. The most severe extreme events are mainly found in low elevation catchments. The number of extreme events has increased over time at all elevations, with the strongest increase for high elevation catchments. Furthermore, the analysis of the driving processes shows that air temperature is the main driver of non-extreme water temperature. However, to predict water temperature extremes, other hydroclimatic variables such as soil moisture, snowmelt, and baseflow should also be considered. This suggests that current water temperature models, which use only air temperature and discharge as input variables, may not be suitable for predicting water temperature extremes at high elevations. These insights into the behaviour of water temperature extremes are valuable for predicting future changes in extremes in mountain rivers.