Evaluating the Utility of Passive Microwave-Derived Soil Moisture Estimates for Forecasting Canola Yields across the Canadian Prairies

Abstract

Soil moisture is a key variable in the determination of crop yields in arid regions around the world. While in-situ soil moisture measurements are sparsely-distributed, passive microwave remote sensing offers an efficient and accurate means for acquiring large-scale observations of surface soil moisture. This research evaluated the utility of passive microwave-derived soil moisture for forecasting canola yields across the Canadian Prairies using soil moisture observations obtained by the Soil Moisture Ocean Salinity Mission (SMOS) satellite. Initial work explored the relationship between soil moisture and canola yields and determined that canola yields are strongly associated (p < 0.01, df = 1) with excess soil moisture conditions throughout the growing season, and in particular, during the stand establishment stage (SM ≥ 26.6%), in low-yielding years. The Integrated Canadian Crop Yield Forecaster (ICCYF) was then employed to assess the added-value of utilizing SMOS soil moisture observations for forecasting canola yields. Improved model fit (R2diff > 0) was observed across most of the Canadian Prairies when SMOS soil moisture indices were included in the ICCYF, however R2 values revealed that model performance overall was relatively low. These findings suggest that while passive microwave-derived soil moisture observations provide an effective indicator of canola yields, forecast skill is limited by the short temporal record of these datasets.