More precise snowmelt forecasts thanks to high-resolution satellite data

SLF researchers have used satellite data to optimize their models, which they use to predict how much snow there is and when and where it will melt. This progress is important in order to be able to warn of floods and inundations in good time.

The big question every spring: when will the snow start to melt? Is there a threat of flooding, how much snowmelt can be expected, when will the Alps be free of snow? Researchers at the SLF have used high-resolution satellite data to make their forecasting models even more precise.

“One specific challenge is ensuring that the snow cover in the models starts to melt at the right moment,” explains snow hydrologist Bertrand Cluzet from the Operational Snow Hydrology Service (OSHD) at the WSL Institute for Snow and Avalanche Research SLF. He analyzed radar satellite data from 2017 to 2021 and extracted specific information on whether there is liquid water in the snowpack or not. He then integrated this into the models. The findings are published in The Cryosphere.

Cluzet analyzed the distribution of wet snow throughout Switzerland and some border regions abroad with a total area of 98,550 square kilometers. Satellite images complement existing measurement data well, as they also record the situation from space in regions that are difficult to access in winter and from which there is therefore no information about the snow cover.

“Our results suggest that wet snow maps contain valuable real-time information for snow cover models and complement the measurements of snow depth in the flat field well, especially in complex terrain and at higher altitudes,” says Cluzet.

His analysis showed that the computer models were not as precise as previously assumed. Cluzet first compared the results of his model with the actual values of 444 measuring points in flat terrain. “The computer model and reality matched up well there,” says Cluzet.

However, the situation was different on inclined terrain. Here, the satellite data showed that the model did not always reliably calculate the snow cover processes in spring and sometimes underestimated the spread of wet snow, especially on sun-exposed slopes. This led to inaccurate forecasts in the past.

So he improved the computer model so that it now predicts the water content in the snow cover more precisely. On this basis, today’s forecasts of the amount of water available from snowmelt in spring are more reliable. “We have greatly reduced the uncertainty that previously existed,” explains the researcher.

The seasonal snow is of crucial importance for the hydrology of mountain regions, adds the scientist: “The snowmelt runoff is often decisive for downstream areas, for example in agriculture or for electricity production in hydropower plants.” Intense snowmelt combined with persistent precipitation can also contribute to devastating floods.