Comparison of Statistical and Multifractal Properties of Soil Moisture and Brightness Temperature From ESTAR and PSR During SGP99

摘要

Global soil moisture products are based on passive microwave sensors of brightness temperature at different frequencies, including L- and C-bands. Two airborne sensors used to develop and test retrieval algorithms for soil moisture estimation are the L-band Electronically Scanned Thinned Array Radiometer (ESTAR) and the C-band Polarimetric Scanning Radiometer (PSR). In this letter, we compare the statistical and multifractal properties of soil moisture and brightness temperature from ESTAR and PSR during the Southern Great Plains Experiment in 1999. We show that differences between products are minimized at a support scale close to the satellite resolution. Then, we compare the subfootprint variabilities of each product in eight coarse domains of $25.6 times 25.6 hbox{km}^{2}$. Scale-invariance and multifractal properties were found in all the three available soil moisture products, but multifractality was negligible in brightness temperature fields, implying that the nonlinear transformations in retrieval algorithms introduce the multifractal behavior in soil moisture. Subfootprint variability and the multifractal behavior for diverse wetness conditions are also different for the three products. Since discrepancies between the subfootprint variabilities of PSR and ESTAR brightness temperatures are small, the factors leading to the differences among the soil moisture products are mainly due to the level of detail of the retrieval algorithm and to the spatial variability of forcing data and parameters related to surface roughness and vegetation.