4.2.2 Observations of Snow Cover, Snow Duration and Snow Quantity

4.2.2.1 Sources of Snow Data

Daily observations of the depth of snow and of new snowfall have been made by various methods in many countries, dating to the late 1800s in a few countries (e.g., Switzerland, USA, the former Soviet Union and Finland). Measurements of snow depth and snow water equivalent (SWE) became widespread by 1950 in the mountains of western North America and Europe, and a few sites in the mountains of Australia have been monitored since 1960. In situ snow data are affected by changes in station location, observing practices and land cover, and are not uniformly distributed.

The premier data set used to evaluate large-scale snow covered area (SCA), which dates to 1966 and is the longest satellite-derived environmental data set of any kind, is the weekly visible wavelength satellite maps of NH snow cover produced by the US National Oceanic and Atmospheric Administration’s (NOAA) National Environmental Satellite Data and Information Service (NESDIS; Robinson et al., 1993). Trained meteorologists produce the weekly NESDIS snow product from visual analyses of visible satellite imagery. These maps are well validated against surface observations, although changes in mapping procedures in 1999 affected the continuity of data series at a small number of mountain and coastal grid points. For the SH, mapping of SCA began only in 2000 with the advent of Moderate Resolution Imaging Spectroradiometer (MODIS) satellite data.

Space-borne passive microwave sensors offer the potential for global monitoring since 1978 of not just snow cover, but also snow depth and SWE, unimpeded by cloud cover and winter darkness. In order to generate homogeneous depth or SWE data series, differences between Scanning Multichannel Microwave Radiometer (SMMR; 1978 to 1987) and Special Sensor Microwave/Imager (SSM/I; 1987 to present) in 1987 must be resolved (Derksen et al., 2003). Estimates of SCA from microwave satellite data compare moderately well with visible data except in autumn (when microwave estimates are too low) and over the Tibetan plateau (microwave too high; Armstrong and Brodzik, 2001). Work is ongoing to develop reliable depth and SWE retrievals from passive microwave for areas with heavy forest or deep snowpacks, and the relatively coarse spatial resolution (~10–25 km) still limits applications over mountainous regions.