Climate change impact on the hydrological processes over an alpine basin: the Adige River

Hydrological extremes (drought and floods) have undeniable financial implications and are predicted to grow in the next years. Yet to understand their future local impacts it is necessary to understand the evolution of the governing hydrological processes. Such is the case for the Adige basin, an important basin in Italy, where understanding changing patterns of hydrological processes is crucial to optimally plan competing water uses, such as hydroelectric production and agricultural water allocation.

Euro-CORDEX models provide future climate projections throughout the region, for different emissions scenarios (RCP 2.6, 4.5, 8.5) and climate models (13). Upon the application of a downscaling and bias correction methodology against observed climate variables (i.e. air temperature and precipitation), a process-based semi-distributed digital twin of the Adige River basin is implemented. Hydrological process variables (snow, actual evapotranspiration, soil moisture and discharge) are obtained for the entire basin and the timespans of the different Euro-CORDEX models (1980-2005 for the historical baselines, 2005-2100 for the projections) at daily temporal scale and 5 km² spatial resolution. The temporal and spatial patterns for discharges are evaluated through the average monthly values for 6 sub-catchments. Other process variables such as snow, actual-evapotranspiration (AET) and soil moisture (SM) are assessed against remote sensing datasets. The resulting climate and hydrological end of the century projections (2075-2100) are compared to historical baselines (1980-2005), to assess projected changes.

The digital-twin model is found to reproduce discharge patterns accurately, with an average KGE of 0.8, and provides a good fit for snow and AET, with average correlations of 0.95 and 0.96 respectively. A reasonable fit is found for SM, with an average correlation of 0.5. Careful assessment of the digital twin model through these variables ensures that it reproduces accurately historical local hydrological processes and increases confidence in the quantification of these variables under future projections.

The results of our study give regional policymakers insights into possible future scenarios and how these affect water resources and their potential impacts and adaptations on several economic sectors.