Multiyear ENSO-Based Crop Yield Forecasts for Early Warning

Early warning systems provide vital decision support for effective humanitarian responses to food crises. Many of the programs implemented by government and non-governmental organizations to anticipate and respond to food crises require months to plan and execute. Examples include securing the funds needed for disaster response, procurement, and distribution of food aid, and scaling up of nutrition assistance programs. Provided the considerable logistical challenges of disaster response programs, there is a clear need for early warning information at lead times up to and beyond a year should it be possible to provide such information.

To date, preseason crop yield forecast systems have generally been evaluated for lead times of around six to nine months. Recent advances in climate forecasts, however, may provide a means of significantly extending preseason crop yield forecast lead times. In this presentation we propose that by leveraging these recent advances in climate forecasts, multi-year crop yield forecasts are not only possible but are skillful over significant portions of cropland globally.

We demonstrate that crop yield forecasts based solely on forecasts of the El Niño Southern Oscillation can make skillful preseason yield forecasts out to and beyond a year prior to harvest for both maize and wheat. Forecasts are most skillful for maize in Southeast and East Africa and Southeast Asia while for wheat they are most skillful in parts of South and Central Asia, Australia, and Southeast South America. The speed with which forecast skill degrades depends on the country, crop, lead time, and month in which the forecast was issued. Finally, we conclude with a discussion of forecast limitations and possible paths forward.