• Some grid boxes (around Greenland and Antarctica) snow continually accumulates during the entire simulations, reaching very large values. I think this is because the model does not represent calving of glaciers. These grid boxes should probably just be ignored, and certainly should be removed from global-mean snow cover for instance. (They have not been removed from the area-average series included on the SO&P website, which is why those series show a continual upward trend in mean snow cover.)
• The headers in the full fields of data indicate that values are stored in m, but it looks like they are actually in mm!
• Be careful when using the full fields of data, because snow can (of course) occur on sea ice, and thus there are snow depth values over the (time-varying) region of the oceans that has sea ice.