Using soil moisture forecasts for sub-seasonal summer temperature predictions in Europe

Soil moisture exhibits outstanding memory characteristics and plays a key role within the climate system. Especially through its impacts on the evapotranspiration of soils and plants, it may influence the land energy balance and therefore surface temperature. These attributes make soil moisture an important variable in the context of weather- and climate forecasting.In this study we investigate the value of (initial) soil moisture information for sub-seasonal temperature forecasts using a mostly observation-driven approach. For this purpose we employ a simple water balance model to infer soil moisture from streamflow observations in 150 catchments across Europe. Running this model with forecasted atmospheric forcing, we derive soil moisture forecasts, which we then translate into temperature forecasts using simple exponential relationships. The resulting temperature forecasts show skill beyond climatology up to 2 weeks in most of the considered catchments. Even if there is significant skill only in some catchments at the longer lead times of 3 and 4 weeks, this simple approach shows local improvements compared to the monthly ECMWF temperature forecasts at these lead times. In addition, the average forecast skill of all catchments at 4 weeks lead time is slightly higher than in the ECMWF product. For both products, we find better forecast performance in the case of extreme events, especially at long lead times. The forecast skills are mainly controlled by the strengths of (i) the soil moisture-temperature coupling and (ii) the soil moisture memory. The negative relationship between these controls weakens the forecast skills, nevertheless there is a middle ground between both controls in several catchments, as shown by our results.