Soil Moisture Data Assimilation to Estimate Irrigation Water Use.

Area of Science:

• Hydrology
• Remote Sensing
• Agricultural Science

Background:

• Irrigation knowledge is vital for food security, water resource management, and understanding global water/energy cycles.
• Consistent data on irrigation water application is lacking.
• Accurate irrigation estimates are needed to address these knowledge gaps.

Purpose of the Study:

• To develop and evaluate a novel method for predicting daily to seasonal irrigation magnitude.
• To assess the impact of remotely sensed soil moisture characteristics on prediction performance.
• To compare the efficacy of a particle batch smoother against a particle filter for irrigation estimation.

Main Methods:

• A particle batch smoother data assimilation approach was employed.
• Synthetic soil moisture data assimilation experiments were conducted.
• Model configurations were varied to test different soil moisture characteristics and error sources.

Main Results:

• Near-perfect irrigation estimates were achieved with noise-free synthetic data (0.66% bias, 0.95 correlation).
• Estimates showed a median seasonal bias of <1% and 0.69 correlation with realistic satellite noise and irregular sampling.
• Systematic biases between land surface models and satellite data significantly degraded performance.

Conclusions:

• The developed framework shows potential for spatially continuous irrigation magnitude estimation.
• Broad applicability hinges on improving irrigation scheduling methods and correcting soil moisture biases.
• The particle batch smoother demonstrated superior performance over the particle filter in this application.