High Latitude Snow Cover from Satellite and Model-Derived Products: Quantifying Uncertainty in Current Datasets and Identifying Potential Future Opportunities
Abstract:
Monthly SCE determined from the NOAA snow chart climate data record was compared to four other snow products based on meteorological forcing from reanalysis coupled with different land surface models. While all the datasets show spring snow cover extent (SCE) in the Arctic (land surface >60N) has undergone significant reductions over the past decade, the uncertainty (standard deviation) in monthly mean Arctic SCE (April, May, June) for the 5 SCE products varies between 50 and 75% of the between-model standard deviation for 14 CMIP5 models over the 1981-2010 period. The same multi-dataset approach was utilized to determine spread in pan-Arctic SWE. There is a similar degree of variability between the SWE products relative to CMIP5 simulations, but the coupled models tend to have a higher magnitude of peak pre-melt SWE than the observational analyses, which may contribute to a slower rate of spring SCE loss in the CMIP5 simulations.
While the resolution of current SWE products (25 – 100 km) is sufficient for climate model evaluation, many applications (for example initialization of numerical weather prediction and hydrological models) are limited by these coarse resolutions. Analysis of airborne data from recent experimental field campaigns near Inuvik, Canada and Toolik Lake, Alaska have illustrated the potential for Ku-band radar measurements to provide high resolution (i.e. 500 m) SWE products sensitive to the range of SWE (0 to 150 mm) characteristic of high latitude snow cover.
