Assessment of the hydrological impact of an hourly precipitation distribution correction method in the SASER modeling system

Abstract

The SASER (SAfran-Surfex-Eudyssée-Rapid) hydrological modeling system is used for simulating the continental hydrological cycle. The validation of the simulated streamflow showed that the results were generally accurate but with a tendency towards underestimating peak values. This underestimation has several causes, but in this study, we addressed the known fact that the precipitation forcing data produced by SAFRAN is good at the daily time scale, as it ingests daily precipitation data, but its performance in interpolating daily to hourly precipitation data using relative humidity as a weight is poor. Our objective was to improve the SAFRAN hourly precipitation distribution and evaluate its hydrological impact on SURFEX LSM simulations. We used a correction method to adjust the hourly precipitation distribution of SAFRAN based on a more realistic distribution obtained from a regional climate model simulation forced by ERA-Interim (CNRM-ALADIN). At the grid point scale, the correction preserves SAFRAN’s accumulated precipitation on a time window of a few days, while reflecting the RCM’s hourly variability of precipitation. We evaluated the correction method across different time windows and found that the 7-day correction showed the highest performance. Our results show that hourly precipitation patterns have a strong influence on surface runoff at the event and grid-point scales. However, at the long-term and catchment scale, the sensitivity to the use of the new precipitation forcing dataset was limited.