Abstract-Soil moisture retrieval algorithms based on passive microwave remote sensing observations need to account for vegetation attenuation and emission, which is generally parameterized as vegetation optical depth (VOD). This multisensor study tests a new method to retrieve VOD from cross-polarized radar backscattering coefficients. Three years of Aquarius/SAC-D data were used to establish a relationship between the cross-polarized backscattering coefficient σ H V and VOD derived from a multitemporal passive dual-channel algorithm (VOD M T ). The dependence of the correspondence is analyzed for different land use classes. There are no systematic differences in the slope for woody versus nonwoody vegetation, resulting in a strong correlation (80% explainedvariance) and a global linear relationship when all classes are combined. The relationship is stable over the years of observations. The comparison of the Aquarius-derived VOD M T to Soil Moisture and Ocean Salinity's multi-angular VOD estimates shows similar spatial patterns and temporal behavior, evident in high correlations. However, VOD M T has considerably higher mean values, but lower dynamic range globally. Most of the differences can be attributed to differences in instrument sampling. The main result of this study, a relationship between backscatter and VOD, will permit high-resolution mapping of VOD with synthetic aperture radar measurements. These maps allow future studies of scaling and heterogeneity effects of vegetation on soil moisture retrieval at the coarser scales of land microwave radiometry. The study shows that VOD based on passive measurements and predicted by active measurements are comparable globally and that the breakdown by land cover classification does not affect the relationship appreciably.
Manuscript