A comparison of SMAP and SMOS L-band brightness temperature observations over the global landmass

Abstract

ABSTRACT This study compared the consistency of brightness temperature (Tb) measurements over the global landmass as observed by the Soil Moisture and Ocean Salinity (SMOS) and Soil Moisture Active Passive (SMAP) radiometers. We evaluated the consistency of horizontally and vertically polarized (H-Pol and V-Pol) SMAP and SMOS Tb data via the Pearson correlation coefficient (r), the bias (B, SMOS minus SMAP), and the Hovmöller diagram analysis. The r and B results were displayed and analyzed through the global land covers and Koppen climates. The results indicate a strong consistency of SMAP and SMOS Tb data for the r (mostly > 0.8) and B (mostly varying from −5 to 5 K) metrics. For different land covers and climates, relatively low r values (< 0.8) were found for the continental climates, deciduous needleleaf forests, mixed forests, and barren/sparsely vegetated land covers. The SMOS Tb gave warmer values with respect to the SMAP Tb over the global landmass at the dual polarizations. But the B values are larger at the V-Pol (global average is 0.70 K) than at the H-Pol (global average is 0.16 K). Only the croplands yielded significantly positive B values at the H-Pol. The largest discrepancies of SMAP and SMOS Tb data were observed in Eurasia and area of eastern North Africa and we speculate that the RFI (Radio Frequency Interference) contamination from the SMOS is a reason that may explain this phenomenon. We suggest users pay careful attention to the RFI contamination when using SMOS Tb in applications. As revealed by the Hovmöller diagrams, SMAP and SMOS Tb data all resolved the same seasonally dynamic changes for both dual polarizations. This study is helpful to construct a long-term and reliable Tb product by combining SMAP and SMOS Tb data to retrieve soil moisture information.