Comparing global passive microwave freeze/thaw records: Investigating differences between Ka- and L-band products

Abstract

The NASA L-Band Soil Moisture Active Passive (SMAP) satellite mission launched in 2015 has produced soil moisture and freeze thaw (FT) products at a global scale. While the use of L-band (1.41 GHz) passive microwave radiometry (P-MW) has proven useful in detecting changes in the surface FT state, these classifications have not been comprehensively assessed against similar existing FT products, such as the global FT record from the Special Sensor Microwave/Imager (SSM/I, Ka-band, 37.0 GHz) as part of the FT Earth System Data Record (FT-ESDR). In order to fill in this gap, this study investigates regions in which FT classifications diverge and identifies potential sources of classification variability. The SMAP and SSM/I FT records are compared over an extended period covering multiple seasonal cycles from April 2015 through December 2017. The spatially and temporally varying relationship between these products is examined in relation to climate (Köppen-Geiger climate classes and air temperature), MODIS (MoDerate Resolution Imaging Spectrometer) land cover, and topography (using Global Multi-resolution Terrain Elevation Data). SMAP and SSM/I FT product agreement proportion (Ap) was corrected for seasonality and then separated by land cover classes and compared to the global Ap mean. The agreement between these products vary most notably during freeze and thaw onset and in areas near abundant surface water, snow and ice, and wetlands. Relative to other vegetation types, reduced agreement between FT products is also observed over grasslands, sparsely vegetated lands, as well as mixed and evergreen forests. Distinct seasonal differences in FT classification agreement were also detected between products over cold arid regions and between continental and temperate classes. Similarly, as topographic complexity increases, a decreasing trend in agreement between L- and Ka-band FT products is observed. While reiterating challenges in FT classifications identified by prior studies, this work also contributes new insights by providing detailed geospatial and seasonal analyses into the factors contributing to FT product divergence.