Modeled effects of moderate and strong `Los Niños' on crop productivity in North America

Abstract

The El Niño of 1997–1998 rivaled in strength that of 1982–1983 and called attention worldwide not only for that reason but also because advances in understanding of ENSO phenomena allowed for its early forecast and preparations to deal with its impacts. We used the Erosion Productivity Impact Calculator (EPIC) to evaluate the impacts of El Niño on North American agriculture through an assemblage of 140 representative farms. We distinguish between `regular' (EN) and `strong' (SEN) El Niño defined by the extent to which sea surface temperatures in the central and eastern tropical Pacific ocean deviate from the long term normal. Each condition produced a distinct geographic distribution of temperature and precipitation anomalies with respect to Neutral (N) years (within 0.5°C of the long term normal). Using daily weather records, EPIC accounted for 87% of the total variation in historical yields in a sample of the farms. Yields simulated in EPIC with a stochastic weather generator predicted different geographic distributions of `winner' and `loser' regions for corn and wheat during EN and SEN years. Changes in water stress during EN and SEN with respect to N years was the variable that accounted for a significant proportion in the variation of simulated yield changes. Yields simulated under SEN tended to have larger variability than under EN. Further evaluation of the methodology presented here could arise from the near real-time simulation of El Niño events. Application of this type of methodology at a regional level could take advantage of interannual climatic variability for agricultural production.