Multi-Source Soil Moisture Data Fusion Based on Spherical Cap Harmonic Analysis and Helmert Variance Component...

Area of Science:

• Earth Science
• Climate Science
• Remote Sensing

Background:

• Soil moisture (SM) is a critical climate variable influencing land-atmosphere interactions.
• Existing global SM products suffer from discontinuous observations and coarse resolution, limiting fine-scale applications.

Purpose of the Study:

• To develop accurate, high-resolution (1 km) spatio-temporally continuous soil moisture products.
• To address limitations of current satellite and model-based SM data.

Main Methods:

• Integration of in situ, satellite, and model datasets using Spherical Cap Harmonic Analysis (SCHA) and Helmert Variance Component Estimation (HVCE).
• Bias correction and resampling of diverse datasets prior to data fusion.
• Comprehensive evaluation against representative and validation sites, and comparison with other SM products.

Main Results:

• Fused SM products achieved high accuracy, with minimal correlation coefficient (R) > 0.85 and largest root mean square error (RMSE) < 0.040 m³ m⁻³ at representative sites.
• Overall validation showed R = 0.889 and RMSE = 0.036 m³ m⁻³ for fused products, significantly outperforming others (R < 0.75, RMSE > 0.06 m³ m⁻³).
• The fused products demonstrated superior alignment with in situ measurements and better capture of SM spatial and temporal dynamics.

Conclusions:

• The developed data fusion methodology effectively overcomes limitations of individual SM datasets.
• The resulting 1 km SM products offer enhanced accuracy and reliability for various applications.
• This approach provides a valuable tool for understanding and monitoring soil moisture at fine scales.