Surface Soil Moisture in Challenging Environments: Methodological Advances and Drought Monitoring Applications

Surface soil moisture (SSM) is a key environmental variable controlling water and energy fluxes at the land-atmosphere interface. As part of the water cycle, SSM influences infiltration, runoff, evapotranspiration and determines plant water availability. Its accurate monitoring is essential for early agricultural drought detection, hydrological modelling and climate change adaptation. Over the last decades, satellite-based approaches have become prominent to monitoring SSM, providing temporally and spatially continuous observations in near-real-time. In particular, the Sentinel-1 satellites, carrying C-band synthetic aperture radar, provide SSM retrievals at an unprecedented sub-kilometer resolution, distributed through the Copernicus Land Monitoring Service. Despite the potential of the operational Sentinel-1 SSM, the product shows limitations over dense vegetation and complex topography, limiting its applicability for environmental monitoring in these landscapes. This thesis first describes methodological advances to address the dense vegetation and complex topography limitations, then demonstrates the potential of Sentinel-1 for agricultural drought monitoring in subtropical context.The first study investigates the impact of vegetation on the Sentinel-1 SSM retrievals across Europe. Three new products are developed by applying vegetation masks derived from optical and microwave datasets to selectively discard densely vegetated backscatter pixels during the resampling from 20 m to the kilometer scale. The validation with both ERA5-Land and in-situ observations highlights that selective masking for non-sensitive pixels significantly improves retrieval accuracy, particularly over mixed pixels combining agricultural and forested land cover.The second study extends SSM retrievals over the Austrian Alps, where approximately 80% of the CLMS product is masked due to topography masking. A novel resampling approach is developed, aggregating terrain-flattened backscatter (γ0) by sub-basins, elevation bands and aspect instead of fixed kilometer-scale grid. The resulting product is compared to ERA5-Land and 264 in-situ precipitation stations, showing that the Sentinel-1 product is sensitive to SSM and precipitation changes over 80% of the Austrian Alps.The third study tests the Sentinel-1 SSM for agricultural drought monitoring over Mozambique, where climate and dominant rainfed agriculture make rural livelihoods particularly vulnerable to drought. First, the Sentinel-1 SSM product is compared to state-of-the-art SSM products (SMAP, ASCAT, ERA5-Land), then used as input to develop two drought indicators: the Soil Water Deficit Index (SWDI), comparing the Sentinel-1 SSM with soil properties, and the Z-score anomalies combining ASCAT long-term climatology with Sentinel-1 high-resolution. The two drought indicators are compared to vegetation and precipitation anomalies, highlighting the potential of surface soil moisture (SSM) to detect the early onset of agricultural droughts at a sub-kilometer resolution.These studies demonstrate that adapting the Sentinel-1 resampling strategies extends the spatial coverage and accuracy of the Sentinel-1 SSM product and suggests its potential for agricultural drought monitoring, hydrological modelling and climate change adaptation.The methodological advances of this thesis contribute to the development of robust and physically informed high resolution SSM products for operational use.