Abstract

Interpolated gridded datasets for temperature and precipitation are often used in large-scale hydrological modeling. Usually, the quality of such datasets is only assessed by cross-validation of the used station data. This raises the question of how good their representativity is in high-altitude areas where the observation station density is low. The representativity of the meteorological datasets for the modeling of the snow cover is of particular relevance here. The present study compares the interpolated temperature and precipitation datasets HYRAS-TAS/HYRAS-PRE (a product of the German Weather Service and of the German Federal Institute of Hydrology, further called HYRAS-dataset) with TabsD/RhiresD (a product of the Federal Office of Meteorology and Climatology, further called MS-dataset) for the period 1971-2006 for the Swiss part of the River Rhine basin. Both Products contain uncorrected average daily values for temperature and precipitation taken from similar measurement stations, but which have been interpolated using different approaches. The direct comparison of the datasets showed systematic deviations, especially for the precipitation. Overall the HYRAS-product showed lower mean daily precipitation sums in higher elevation zones than the MS-product. In a further step, the suitability of both datasets for the modeling of the snow water equivalent (SWE) was tested. To this aim, an expanded version of the Temperature-Degree-approach (as used, for instance, in the hydrologic runoff model HBV) was used to calculate the mean daily SWE of 766 sub-basins with an average area of 35 km(2). The modeled SWE values of each sub-basin were then compared with an independent dataset consisting of daily SWE maps of Switzerland which are based on daily snow depth measurements and specific snow models (a product of the operational snow hydrological service of the Swiss Institute for Snow and Avalanche Research SLF, further called SLFmap-dataset). The comparison with the SLFmap-dataset showed that the overall snow amount in high elevations (above 2000 m.a.s.l.) was somewhat underestimated by the SWE modeled on the HYRAS-product and highly overestimated by the SWE modeled on the MS-product. The overestimation of the SWEMS increased in the course of the winter and was especially noticeable in April and Mai. The overall difference between the SWEHYR and SLFmap was smaller than the difference between the SWEMS and SLFmap. The estimation of the overall amount of SWE improved when the HYRAS-dataset was used, but the MS-product showed a more detailed spatial structure in the precipitation, which was more consistent with the SLFmap-dataset. These differences should be taken into consideration when selecting the meteorological datasets for hydrological modeling in regions where snow plays an important role.

Keywords

Snow water equivalent; temperature; precipitation; hydrological Modeling; Snow model; interpolated meteorological dataset

Published in

Hydrologie Und Wasserbewirtschaftung
2016, Volume: 60, number: 6, pages: 353-367
Publisher: BUNDESANSTALT GEWASSERKUNDE-BFG

UKÄ Subject classification

Oceanography, Hydrology, Water Resources


Publication identifier

DOI: https://doi.org/10.5675/HyWa_2016,6_1

Permanent link to this page (URI)

https://res.slu.se/id/publ/98781