Abstract: Snow water equivalent (SWE) is an important part of the surface water and energy balances as well as being a significant economic resource. Studies have suggested that snow has an effect on climate locally and remotely, and also over long timescales. Typically, modelling studies are usually performed using data from field sites of a limited size for validation: validation of snow models on a large (e.g. continental or hemispheric) scale is more difficult as in situ data are sparse. This work compares the distribution of SWE at the continental scale from remotely sensed data with that obtained from various general circulation models at different resolutions, and also reanalysis data. Differences between the datasets are large, particularly over Siberia, a region where snow anomalies are thought to influence atmospheric circulation. The HadCM3 model in particular has a large circulation bias over Siberia, but comparisons with in situ observations and runoff data from major catchments show inconsistent results. In order to produce realistic climate model fields for analysis and exploit any use that snow may have as a circulation diagnostic, the differences between these datasets must be understood more fully.