EURO-FRIEND, Flow Regimes From International Experimental and Network Data

Contact |  Imprint |  Sitemap |  Print |  Send

Assimilating satellite derived snow covered area (SCA) in hydrological models

Recent research

A major cause of flooding in Norway is the combination of intense snowmelt and precipitation. In order to be able to forecast these flooding events, we need reliable forecast of precipitation and temperature, and a good estimate of the snow reservoir and its coverage in the catchment at the time of the forecast.

Successful updating for the catchment Atnasjø year 2002. Note how the snow reservoir in the top panel is sharply reduced in the middle of the melting period due to the assimilation of satellite derived SCA

Modelling the SCA correctly is considered a prerequisite for the applied rainfall-runoff model in order to capture the dynamics of the snowmelt induced spring flood, and through this study, an analytical link between SCA and the parameters of the spatial distribution of snow water equivalent (SWE) is developed. The spatial distribution of snow water equivalent (SWE) is modelled as a two parameter gamma distribution with parameters dependent on the number of accumulations and ablations. The strict analytical control of the spatial distribution of SWE at all times allows for the development of algorithms which relates accumulated or ablated snow to changes in snow covered area (SCA) of a catchment. The algorithms are further developed so that remotely sensed information of SCA can be used in order to update the snow reservoir and the spatial distribution of SWE. The snow distribution model and the updating algorithms are implemented in the Nordic HBV model and have been tested for ten Norwegian catchments. The overall improvements on the prediction of discharge by updating for several satellite derived SCA scenes are modest, but significant improvements for some scenes are observed (Fig. 1). Errors and temporal inconsistencies in the quantification of SCA from the satellite scenes are found, which may lead to serious errors in the predicted discharge. The methodology is totally dependent on the quality of the SCA data and special care in quantifying SCA has to be taken for operational use of the updating algorithms.

For more information contact: Thomas Skaugen.

Top top      

River Trent, Nottingham, UK
River Trent, Nottingham, UK

Heavy rain, Penang, Malaysia
Heavy rain, Penang, Malaysia