Satellite and Reanalysis Soil Moisture Products for Irrigation Monitoring in Northern Algeria

ABSTRACT Accurate estimation of soil moisture (SM) is a critical factor for sustainable irrigation management and food security, especially in semi‐arid agricultural regions where water availability is highly variable. Although reanalysis and satellite‐based products provide extensive spatial coverage, their reliability must be systematically validated against in situ measurements. This study establishes a multi‐product validation benchmark by evaluating and comparing several satellite and model‐based SM products over northern Algeria, including the Soil Water Index from the Advanced Scatterometer (ASCAT), Sentinel‐1 (S1), the SM Active Passive (SMAP) mission, European Centre for Medium‐Range Weather Forecasts (ECMWF) Reanalysis v5‐Land (ERA5‐Land), and the Global Land Data Assimilation System (GLDAS). The analysis focuses on surface‐to‐near‐surface SM dynamics while accounting for site heterogeneity, product resolution, and sensing‐depth differences. The results reveal pronounced spatial variability linked to local pedoclimatic conditions, including soil properties, vegetation cover, and land‐use characteristics. Across the evaluated sites, ASCAT shows the highest average temporal consistency ( r ≈ 0.84; maximum 0.96 at Chiffa), whereas SMAP shows more balanced performance and Sentinel‐1 exhibits strongly site‐dependent behavior, with local failures discussed in the text. ERA5‐Land effectively reproduces temporal variability ( r up to 0.91), whereas the low unbiased root mean square error (ubRMSE) values obtained for GLDAS indicate a low random error component after bias removal rather than absolute accuracy. The bias profiles show that temporal agreement does not guarantee the absence of systematic error: At Chiffa, biases reach +12.75 percentage points in volumetric SM for ASCAT and +15.40 for ERA5‐Land, whereas the ASCAT bias reaches +14.00 at Oued Smar. The performance observed for ASCAT is specific to the evaluated sites and validation period; it does not constitute a universal ranking of the products nor does it validate irrigation‐event detection or direct irrigation scheduling, which require independent rainfall and irrigation records.