Improving runoff modelling using satellite-derived snow covered area?

Abstract

Operational use of satellite-observed snow covered area (SCA) in the HBV model was studied in order to improve the spring flood prediction. The study included calibration of HBV models against both discharge and SCA, against discharge only, and updating of the HBV models based on satellite observed SCA. Ten test catchments were selected for the study. The results confirmed that HBV models calibrated against SCA in addition to discharge simulate discharge nearly as well as models calibrated against discharge only. The simulated SCA was markedly improved when SCA was included in the calibration. Deviations between observed and simulated SCA was used to trigger model updates. In such cases, the model input (precipitation and temperature) was adjusted until a reasonable fit in the SCA was obtained. The results of the updates were ambiguous. Only 28% of the updates improved the runoff simulations. Improved simulations were most frequent at large SCA values. SCA derived from radar and optical satellite sensors differed considerably. Harmonisation is required before both types of data can be integrated in the same model simulations.