Inconsistencies in global soil moisture products and discrepancies in their relationship with vegetation productivity

Abstract

Soil moisture is one of the critical environmental variables influencing ecosystem function and plays a vital role in regulating vegetation dynamics. In recent years, various soil moisture datasets have been developed using different methodologies, including land surface modeling, remote sensing-based retrievals, and data assimilation techniques. These datasets have been widely applied to study vegetation responses to water availability. However, their consistency has not been thoroughly evaluated, which introduces biases and inconsistencies in vegetation response analyses. Such inconsistencies may lead to biases when interpreting vegetation response patterns and long-term environmental trends.

To address this issue, this study explores the differences among multiple soil moisture products and assesses their consistency in evaluating vegetation water stress responses. We focus on five widely used soil moisture products European Centre for Medium-Range Weather Forecasts Fifth-Generation Land Reanalysis Dataset (ERA5-Land), Global Land Evaporation Amsterdam Model (GLEAM) soil moisture dataset, the second Modern-Era Retrospective analysis for Research and Applications (MERRA-2), Global Land Data Assimilation System (GLDAS) soil moisture dataset, and a globally gap-filled surface soil moisture dataset and in-situ soil moisture observations were collected for this analysis. These products were chosen to represent different soil moisture estimation approaches, ensuring a comprehensive assessment of their consistency.

To evaluate discrepancies among these datasets, we applied statistical correlation analysis, trend comparisons, and spatial pattern assessments using satellite-derived vegetation index and solar-induced fluorescence were used as proxies for vegetation activity. The results indicate that correlations between each product and observed data varied seasonally, with stronger performance during the growing season compared to the non-growing season. These products showed conflicting long-term trends at global scale. Additionally, there were significant discrepancies in the relationships between different moisture products and vegetation indices, particularly in spatial patterns. In about half of the global regions, conflicting correlations emerged between different products and vegetation indices. These inconsistencies highlight the challenges of using a single soil moisture dataset for ecological studies, emphasizing the necessity for cross-product comparisons to improve data reliability and integration. The findings of this paper provide new perspectives for future research on soil moisture and atmospheric dryness and help improve the effectiveness of data integration strategies.