Sentinel-1 Soil moisture validation with Multiple sources of observations in the mountainous area in Korea

Soil moisture is a key factor in analyzing the water cycle on the land surface. Active microwave sensor has been widely used for spatially representative soil moisture content regardless of weather conditions. Especially C-band microwave sensors (Scatterometer and Synthetic Aperture Radar) loaded on a satellite were adopted for capturing soil moisture over the vegetated area. However, in heterogeneously or thickly vegetated areas, it is difficult to get accurate soil moisture content with SAR sensor for high spatial resolution (< 1 km). In this study, high-resolution soil moisture content in mountainous areas is estimated and evaluated with in-situ soil moisture observation and a Cosmic-Ray Neutron probe (CRNP) sensor. To evaluate the satellite-based soil moisture product, the SMC Soil Moisture observation site, designed for monitoring soil moisture content, was used. The site has 16 FDR sensors for 10 cm, 20 cm, and 30 cm. At the center of the site, CRNP is operated for measuring spatial soil moisture content. Firstly, the Sentinel-1 backscattering signal, strongly affected by land surface conditions in the mountainous areas, was analyzed. Then, canopy attenuation and the relation between the backscattering signal and the local incidence angle on the mountain were evaluated. Secondly, Sentinel-1 images on the observation site were resampled to 10 m, 50 m, 100 m, and 150 m. Water Cloud Model and change detection method were applied to estimate soil moisture content for the 4 scales. Lastly, estimated soil moisture content was compared with CRNP soil moisture data and ASCAT data on observation sites. Error analysis for each pixel included in ASCAT pixels was conducted to figure out the main obstacles of soil moisture estimation on mountains. With the result of this study, high-resolution soil moisture estimation on the Korean peninsula which mainly consists of the mountainous area would be suggested.

Acknowledgment: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (NRF-2022R1A2C2010266). This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education(2021R1A6A3A01087645)