Sub-daily downscaling of discharge in glacierized Alpine catchments

Alpine glaciated catchments exhibit complex hydrological streamflow dynamics influenced by temperature effects on snow and ice melt as well as precipitation, resulting in seasonally varying diel streamflow cycles. These cycles shift and become more intense during the summer melt season due to reduced buffering by the declining snow cover and the associated progressive development of more efficient subglacial drainage systems. This variation is of importance, especially for sediment transport, which is commonly a non-linear function of instantaneous discharge above a critical threshold. However, these diel streamflow cycles remain challenging to simulate due to a lack of high-quality meteorological data for remote areas and a general lack of observed streamflow data in highly glaciated catchments for model calibration. Consequently, many classically used hydro-glaciological models, such as those that use a degree-day approach for melt simulation, cannot capture sub-daily streamflow dynamics well, unless they are combined with temporal downscaling to sub-daily timescales. This work aims to develop an innovative downscaling approach that captures the specific features of streamflow patterns in Alpine glacierized catchments. 

The work benefits from an exceptionally high-resolution dataset that comprises 15-minute discharge records for 45 years from 7 small, highly-glacierized catchments in the South-Western Swiss Alps (relative glacial cover ranging from 5 to 70%). It adopts a maximum entropy (POME) approach more commonly used to downscale non-glacial discharge records available at the monthly scale. We couple this approach with a semi-distributed hydrological model that predicts mean daily discharge using modeled hydrological characteristics (e.g., snow depth, ice melt rates) to drive the downscaling. 

Results show that a simple sigmoid equation can be used to fit the daily flow duration curves of glacierized catchments. Furthermore, the progressive evolution of the sigmoid parameters over the last 45 years shows the influence of rapid climate warming on the dynamics of sub-daily flows. The downscaling method based on daily simulated discharge and informed by simulated hydrological and glacial characteristics offers a promising and transferable solution for reconstructing sub-daily discharge in data-scarce regions, as well as for improving hydrological modeling at high temporal resolutions.