Shifting hydrological process importance in alpine catchments: Combined effects of climate and land cover change on alpine evapotranspiration

In alpine catchments, evapotranspiration (ET) is regularly considered a minor component of the hydrological system and is therefore only rudimentarily regarded in modelling studies and climate change projections. The focus is usually on snow and glacier related processes, projecting a rapid retreat of glaciers in central Europe within the next few years and decreasing snow volumes at lower elevations due to climate warming. This leads to a reduction in the dominance of snow and glacier related processes. Changes in vegetation characteristics due to climate- and/or land use change will presumably lead to a relative increase in the importance of other processes such as ET, interception and soil moisture, which will affect spatial and temporal variations in discharge generation.

To identify spatial and temporal patterns of changing process importance, a case study of the Fundusbach catchment in Tyrol, Austria is conducted using the fully distributed and physically based model WaSiM-ETH. As the process representation within all hydrological models is affected by parameter equifinality, substantially different parameter combinations and therefore process representations can lead to similar values of performance criteria when looking at discharge only. To address this issue, elevation dependent, spatial and temporal parameter sensitivity analysis is coupled with a multi-objective calibration. This approach aims to improve the spatial and elevation dependent process representation within the model domain. Instead of calibrating on different output variables, additional field and remote sensing data are used to constrain the parameter space of individual submodels and thus the process behaviour. In a final step, the constrained model is calibrated against discharge. Based on this reference model, scenario analyses are carried out to investigate individual process responses to changes in climate and land use.

The results show, that integrating additional field data and constraining the calibration parameter space considerably improves the process representation within the model. Furthermore, first results show, that changes of the land cover do not influence the overall discharge regime, but ET and soil moisture. Increasing amounts of shrub cover limit infiltration and evaporation, while interception and soil moisture increase. Most process responses intensify with increasing elevation and reflect the spatial patterns of land cover.