Satellite and modelling based snow season time series for Svalbard: Inter-comparisons and assessment of accuracy (SATMODSNOW)

• Main Authors: Malnes, Eirik, Vickers, Hannah, Karlsen, Stein Rune, Saloranta, Tuomo, Killie, Mari Anne, Van Pelt, Ward, Zhang, Jie
• Format: Report
• Language: English
• Published: Svalbard Integrated Arctic Earth Observing System 2021
• Subjects: Snow; remote sensing; modelling; snow cover; snow water equivalent; Svalbard
• Online Access: https://zenodo.org/record/4294072; https://doi.org/10.5281/zenodo.4294072

This is chapter 8 of the State of Environmental Science in Svalbard (SESS) report 2020 (https://sios-svalbard.org/SESS_Issue3). We document differences and similarities between three satellite-based and three model-based snow cover datasets, showing the geographical distribution and amount of snow across Svalbard for several periods from 1957 to 2020. The study shows that the datasets have many differences and that work needs to be done to accurately represent the snow cover in Svalbard. Low resolution datasets tend to predict longer winters than higher resolution datasets. We studied differences between the datasets and suggest methods to improve each dataset. Satellite data have been available since 1978, but early sensors had low resolution, and can only provide correct information over larger areas. Current sensors, available since 2016, have high resolution. Older low-resolution data may be improved by utilising overlapping time-series of high- and low-resolution data since local snow distribution patterns recur annually with a time-shift depending on average temperature and precipitation during the winter. The snow models predict in general the amount of snow (Snow Water Equivalent or SWE), but the timing of snow disappearance predicted by the models can be compared with estimates from satellite snow cover observations. Since the snow models depend on uncertain models of precipitation and temperature to estimate SWE there is potential to integrate satellite data to improve the models for snow in the future.