1 km L-VOD retrieval from the synergy of Sentinel-1 SAR and SMOS radiometer data for subtropical evergreen vegetation dynamics in China

The L-band vegetation optical depth (L-VOD) with high spatial resolution is essential for monitoring vegetation traits. However, the spatial resolution of current VOD products is relatively coarse (∼ 9–36 km). This study investigates a methodological framework for generating 1 km L-VOD (L-VOD 1km ) by coupling active/passive spatial downscaling methods with a radiative transfer model and extended Q-H model. The framework was tested in red soil hilly region of southern China. First, we derived L-VOD 1km collaboratively using Sentinel-1backscatter, SMOS brightness temperature, and soil properties. Second, the performance of L-VOD 1km was evaluated for parameter sensitivity, spatial details, and responsiveness to disturbance using Normalized Difference Vegetation Index (NDVI), aboveground biomass (AGB), SMOS INRA-CESBIO L-VOD product (L-VOD SMOS-IC ) as well as climate/human activity data. The results are as follows: i) the mean temporal Distance Correlation (Dcor) values between L-VOD 1km and NDVI/AGB were 0.61 and 0.72, respectively. Across different regions, topographic conditions, and vegetation types, the spatial Dcor between L-VOD 1km and NDVI consistently remained high. ii) L-VOD 1km can distinguish more vegetation types and spatial details than L-VOD SMOS-IC . iii) The pronounced decline in L-VOD 1km during 2018–2020 corresponds to the preceding rainfall deficits (2017–2019), with an approximate one-year lag relative to the precipitation anomalies. Our results indicated that the proposed framework could generate high-resolution L-VOD with good performance in mapping evergreen vegetation properties in RSHR. This study presents new perspectives for advancing finer-scale L-VOD with evident physical interpretability, offering great potential to assist potential studies such as dynamic monitoring and health assessment of vegetation ecosystems. • 1 km L-VOD (L-VOD 1km ) generated by fusing physical and semi-empirical models. • L-VOD 1km excels in parameter sensitivity, spatial detail, and disturbance response. • L-VOD 1km can serve as foundational data for depicting finer-scale vegetation traits. • L-VOD 1km effectively captures vegetation dynamics in red soil hilly region of China.