Estimating snowmelt using remote sensing information

The degree-day method has been an important tool in estimating snowmelt in the past, but a modern technique using remote sensing images is becoming more routinely used in areas where snow accumulation and melt is important to water resources. Remote sensing can most easily reveal the area of snow-covered ground and the surface temperature of the snow pack; it cannot so easily determine the total snow water equivalent in a pack. Passive microwave sensors have been used to estimate snow water equivalent (deeper packs have lower microwave brightness temperatures) but available satellite systems have too coarse a resolution to be useful other that on large areas of relatively flat terrain (Schmugge et al., 2002). Estimates of snow-covered area can still be operationally useful, however, during the melt season. This is because the changes in area can be superimposed over a digital terrain map of an area and extrapolated patterns of air temperature to estimate how much water equivalent has melted between successive images. The snowmelt runoff model (SRM) combines the changes in area with the degree-day method to provide operational forecasts of snowmelt runoff (Seidel and Martinec, 2003).

The MODIS sensors on the AQUA and TERRA satellites of the Earth Observation Program provide both visible wavelength and thermal infra-red images and allow repeat imaging every 1-2 days at a resolution of 500 m (e.g. Fig. 10.16). Snow can be differentiated from cloud cover by using both visible and infra-red images,

Fig. 10.16 (a) Moderate-resolution imaging spectroradiometertrue-colour image acquired on 12 April 2005, showing the result of a new snowfall in Colorado; (b) the corresponding swath snow map, with pink areas indicating a classification as possibly either snow or cloud (after Hall and Riggs, 2007, with kind permission of John Wiley & Sons).

Fig. 10.16 (a) Moderate-resolution imaging spectroradiometertrue-colour image acquired on 12 April 2005, showing the result of a new snowfall in Colorado; (b) the corresponding swath snow map, with pink areas indicating a classification as possibly either snow or cloud (after Hall and Riggs, 2007, with kind permission of John Wiley & Sons).

although cloud cover can prevent full coverage of an area being retrieved for all images. Operational snowmelt prediction systems based on satellite data are now being used in several countries including the USA, Canada and Spain.

0 0

Post a comment