Harnessing Ocean Salinity as a Measure of Oceanic Moisture Export and Terrestrial Rainfall Variability for Improved Predictions of Wheat Yields in Australia

Crop failures in Australia, one of the world’s largest wheat producers, can have ripple effects on global food security and prices. These failures are often associated with climate variability, which may subject crops to floods, droughts, and heat stress. To help minimize societal impacts to climate-related variations in crop yields or develop well-informed strategies in anticipation of favorable growing conditions, it would be useful to identify predictors of crop yield in advance of the growing season. Ocean salinity is increasingly used as a metric that tracks moisture export from the ocean surface and has been used as a skillful predictor for terrestrial hydroclimate in Australia at seasonal or longer lead times. Hence, we assess whether sea surface salinity (SSS) could also be a useful predictor directly for wheat yield in Australia. We find that years with low wheat yields are often associated with characteristic SSS anomalies in the tropical Indo-Pacific prior to the growing season. The salinity patterns identified capture the effect of anomalous atmospheric moisture export from the Indian and Pacific Oceans and variations in moisture transfer towards Australia. Relevant SSS indices are linked to key climate modes, such as the El Niño-Southern Oscillation and Indian Ocean Dipole, and used for predicting New South Wales wheat yields. To assess the predictive power of the SSS patterns, separate wheat yield linear prediction models were created using cropping season rainfall (May – October) and the SSS variations in preceding months. SSS demonstrated a stronger predictive power for New South Wales annual wheat production compared to cropping season rainfall in this region. Improvements of the prediction model are evaluated by integrating additional oceanic, atmospheric, and terrestrial variables at a range of forecast lead times. Comparison of the optimal predictive models for wheat yields in different states are linked to understanding ocean moisture transport near Australia. These predictions may be useful for farmers, enabling them to adjust their strategies in anticipation of the cropping season, especially at longer lead times offered by SSS.