Accuracy evaluation of snow water equivalent global data: the case of the Northern Dvina River basin

The article presents a comprehensive multilevel accuracy evaluation of snow water equivalent (SWE) global datasets performed for the territory of the Northern Dvina River basin where snow plays an important role in the formation of spring runoff and the increase in winter runoff during recent decades due to ongoing climate change. Six global snow water equivalent datasets, available on the National Snow and Ice Data Center website (all available global SWE data now), were used in this study. Snow surveys data of the Roshydromet observation network were used as ground-based observations. The accuracy evaluation of datasets values of SWE was carried out by comparison with the ground observations in several successive stages. At the first stage, the global SWE values were compared with the observational data for several selected weather stations, i.e. the datasets' values of SWE belong to a particular grid cell were compared with the snow surveys values of a weather station located in this cell. At the second stage, the SWE values averaged over one of the main subbasins of the Northern Dvina, the Small Northern Dvina River basin, were analyzed. The final stage was the SWE comparison for the entire Northern Dvina River basin. The methodology for ranking data depending on their accuracy based on the comparison with ground-based observations of several basic parameters of quantitative changes in SWE during periods of formation and before the start of melting (maximum SWE) was specially developed to implement the accuracy evaluation and SWEs comparison. A comprehensive comparative analysis performed based on the developed methodology revealed that the accuracy of global SWE data increases with the investigation area (i.e. the volume of used information) for which the comparison is being carried out. The closest to the observed SWE values for all datasets were obtained for the whole Northern Dvina River basin. The ranking methodology and the implemented analyze helped to determine the most promising global SWE data for future use.