IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (Jan 2024)
Calibration of the SMAP Soil Moisture Retrieval Algorithm to Reduce Bias Over the Amazon Rainforest
Abstract
Soil moisture (SM) is crucial for the Earth's ecosystem, impacting climate and vegetation health. Obtaining in situ observations of SM is labor-intensive and complex, particularly in remote and densely vegetated regions like the Amazon rainforest. NASA's soil moisture active and passive (SMAP) mission, utilizing an L-band radiometer, aims to monitor global SM. While it has been validated in areas with low vegetation water content (VWC) (< 5 ${\text{kgm}}^{ - 2}$), its efficiency in the Amazon, with dense canopies and high VWC (> 10 ${\text{kgm}}^{ - 2}$), is limitedly investigated due to scarce in situ measurements. This study assessed and analyzed the SMAP SM retrievals in the Amazon, employing the single-channel algorithm and adjusting vegetation optical depth (τ) and single scattering albedo (ω), two key vegetation parameters. It incorporated in situ SM observations from three old-growth rainforest locations: Tambopata (Southwest Amazon), Manaus (Central Amazon), and Caxiuana (Eastern Amazon). The SMAP SM deviated substantially from the in situ SM. However, calibrating τ and ω values, characterized by a lower τ, resulted in better agreement with the in situ measurements. This study emphasizes the pressing need for innovative methodologies to accurately retrieve SM in high-VWC regions like the Amazon rainforest using SMAP data.
Keywords