Validate satellite remote sensing soil moisture with ground-based methods in dryland

Soil moisture is a key component of the Earth system and hydrological cycle. Accurate soil moisture estimates are critical for many applications. Global soil moisture measurements are primarily derived from microwave remote sensing (RS); however, their spatial resolution is typically coarse, often on the order of kilometers, and is impacted by various factors. Therefore, in situ ground measurements should be used to improve the spatial and temporal representation of soil moisture in RS. The current study presents a comparative analysis of soil moisture data retrieved from Time Domain Reflectometry (TDR), Electromagnetic Induction (EMI), Cosmic-Ray Soil Moisture Observation System (COSMOS), and satellite remote sensing soil moisture derived using the OPTical TRApezoid Model (OPTRAM). The study site is located in a semi-arid environment, with a mean annual rainfall of 150 mm that falls between October and May. EMI measurements were conducted manually during the dry summer and wet winter seasons. Concurrently, TDR at depths of 10 and 20 cm and COSMOS continuously monitored and collected soil moisture data, respectively. Satellite information for the dates of the EMI surveys was retrieved from Sentinel-2 images.

Various correlation analyses were performed. The spatial and seasonal relationships between apparent electrical conductivity (ECa) and remote sensing soil moisture (RSSM) were also tested. At the beginning of the winter season, after a long dry spell, the ECa values correlated negatively with the RSSM. The best positive correlation occurred only after a long period of water percolation. The correlation between TDR and RSSM was the strongest among the methods. Meanwhile, COSMOS soil moisture also showed a strong positive correlation with RSSM, stronger than with ECa.

Concerning EMI measurements, soil moisture variability was minimal after five months of a dry, hot summer. Following several rain events, the ECa values exhibited high variability, which was related to increases in soil moisture. The RSSM showed a corresponding phenomenon: during the dry period, a narrow distribution of values was observed, and after a number of rain events, the distribution expanded. Thus, the ground-based EMI method and RSSM indicated the same spatiotemporal dynamics of soil moisture in the subsurface of dryland.

It is concluded that the RSSM represents the spatiotemporal conditions of the top-soil moisture conditions, but only after sufficient time for water percolation and distribution. TDR and COSMOS provide reliable soil moisture data to correct RSSM across time and space, whereas EMI is seasonally dependent (positive correlation during very wet periods and negative correlation after long dry spells).