Using MODIS forest transmissivity estimates to correct passive microwave observations of snow-covered landscapes

Abstract:

Forest cover attenuation of microwave emission is a significant challenge to the estimation of snow accumulation by remote sensing microwave observations because canopy biomass can attenuate understory snowcover emission and can produce additional emission to that generated by the snowpack and sub-nivean surface. Transmissivity of radiation is the variable that describes how a tree canopy attenuates microwave emission from the ground. Although transmissivity can be measured in the field or estimated by models using field data at the in situ scale, the estimation of transmissivity at regional to global scales is a challenge. Therefore, following the work of Metsamaki et al. (2005), a transmissivity model that uses reflectance data from the Moderate Resolution Imaging Spectroradiometer (MODIS) in snow covered areas (SCA) and adjacent non-snow covered areas is applied to estimate transmissivity at regional scales. In essence, given snow covered forest and adjacent snow covered open areas, the difference in reflectance between these two landscapes will be due to forest attenuation of light which in turn is a function of forest transmissivity. In the work of Metsämaki et al. (2005) transmissivity estimates are a bi-product of the Snow Cover Area model (SCAmod) approach and are applicable to mid- to high-latitude regions where seasonal snow persists. While transmissivity estimates cannot be estimated in dominantly snow-free lower latitude regions, the MOD44B land cover data set is used to extend SCAmod transmissivity data into these lower latitude regions. The influence of vegetation on passive microwave brightness temperatures is explored by comparing the brightness temperatures (T_bs) T_bs in open areas with T_bs in forest covered areas. A regression model between the T_bs contributed by the vegetation and transmissivity was derived. In general, the brightness temperature contributed by the vegetation increases with the increase of forest vegetation density (decrease of the transmissivity). Vegetation tends to contribute more emission in a higher frequency bands or in a horizontally-polarized bands. Application of the transmissivity estimates are used to correct Advanced Microwave Scanning Radiometer brightness temperatures for snow water equivalent and snow depth estimates in forested regions.