Assessing the use of GPR and drone snow data for model development and runoff predictions in Northern Sweden

In cold regions, snow is a crucial component of the cryosphere, experiencing changes such as decreasing snowpack and snow cover. These changes impact the seasonal amount of snow and cause a shift in the timing of spring floods, particularly in mountainous areas. The complex and diverse snow processes and interactions in mountainous environments challenge making accurate predictions on snow and runoff. Moreover, snow is not uniformly distributed in space and time, which emphasizes the importance of monitoring mountain snowpack to enhance the understanding of hydrological processes and improve forecasting in the face of changing conditions. In the past few years, ground penetrating radar and drone acquisitions have emerged as a state-of-the-art methodology for obtaining snow data at high spatial resolution with a significant area coverage compared to traditional point observations. This study used data from ground penetrating radar and drone acquisitions to develop and evaluate a new snowfall distribution function based on wind speed, direction and topography to model wind redistribution in the semi-distributed hydrological model HYPE. We assessed the effect of the new snowfall distribution function compared to the one based on wind direction and topography on the snow distribution close to the accumulation peak in the Överuman catchment, Northern Sweden. We further assessed the impact of the two snowfall distribution functions on the catchment runoff predictions. Results show that the snowfall distribution function based on wind speed and direction better simulated the snow spatial distribution in the catchment than the snowfall function based on wind direction. Ground penetrating radar and drone acquisitions provided complementary model development and evaluation information.