Is it Possible to Quantify Irrigation Water‐Use by Assimilating a High‐Resolution Satellite Soil Moisture Product?

Ehsan Jalilvand,Ronnie Abolafia‐Rosenzweig,Masoud Tajrishy,Mohammad Reza Mohammadi,Sujay V Kumar,Narendra N Das

DOI: https://doi.org/10.1029/2022WR033342

IF: 5.4

2023-03-27

Water Resources Research

Abstract:Irrigation is the largest human intervention in the water cycle that can modulate climate extremes, yet irrigation water use (IWU) remains largely unknown in most regions. Microwave remote sensing offers a practical way to quantify IWU by monitoring changes in soil moisture caused by irrigation. However, high‐resolution satellite soil moisture data is typically infrequent (e.g., 6 ‐12 days) and thus may miss irrigation events. This study evaluates the ability to quantify IWU by assimilating high‐resolution (1km) SMAP‐Sentinel 1 (SMAP‐S1) remotely sensed soil moisture with a physically‐based land surface model (LSM) using a particle batch smoother (PBS). A suite of synthetic experiments is devised to evaluate different error sources. Results from the synthetic experiments show that unbiased simulations with known irrigation timing can produce an accurate irrigation estimate with a mean annual bias of 0.45% and a mean R2 of 0.97, relative to observed IWU. Unknown irrigation timing can significantly deteriorate the model performance, resulting in an increased mean annual bias to 23% and decreased mean R2 to 0.36. Adding random noise to synthetic observations does not significantly decrease model performance except for the experiments with low observation frequency (>12 days). In real‐world experiments, the PBS data assimilation approach underestimates observed IWU by 18.6% when the timing of IWU is known. IWU estimates are consistently significantly higher over irrigated pixels compared to the non‐irrigated pixels, indicating data assimilation skillfully conveys irrigation signals to the LSM.

environmental sciences,water resources,limnology