Analysis of 9 km quasi-global microwave land surface emissivity estimates derived from SMAP radiometer and CYGNSS reflectometer

Accurate estimation and high-resolution mapping of land surface emissivity are crucial for enhancing understanding of Earth system processes and facilitating informed decision-making across various fields. This study presents a method that combines Soil Moisture Active Passive (SMAP) radiometer and Cyclone Global Navigation Satellite System (CYGNSS) reflectometer data to derive 9 km quasi-global (within ±38° latitudes) microwave land surface emissivity estimates for the first time. In practice, the SMAP emissivity data serve as the base at 36 km scale, while CYGNSS reflectivity data are used to capture the spatial heterogeneity at 9 km scale. The combination is achieved through a spatially varying scaling factor that converts CYGNSS reflectivity to SMAP emissivity space. Global comparisons show a strong correspondence between the downscaled SMAP/CYGNSS 9 km emissivity and SMAP interpolated 9 km emissivity, yielding correlation coefficient (R) values ranging from 0.93 to 0.96 and Root-Mean-Square Error (RMSE) values from 0.018 to 0.020 for four 90-day periods. Moreover, the downscaled emissivity exhibits satisfactory spatial-temporal performance, with a median repeat period of 2.5 days for most mid-latitude regions and 4.7 days for all regions. When compared to SMAP native emissivity at 36 km scale, the upscaled SMAP/CYGNSS 36 km emissivity shows a significant increase of 66% in spatial coverage on the daily scale and a median increase of 64% in temporal coverage. These results indicate that the GNSS-R data are likely to be an effective supplement at small spatial scale with a sub-daily to daily scale for future remote sensing of microwave emissivity.