摘要
This research uses a quantitative methodology for directly comparing the responses of observed crop yields in the SE USA to ENSO phenomena classified using dissimilar ENSO indices. ENSO condition has been related to agricultural yields in many parts of the world. It has been generally recognized that the strongest effects on spring and summer crops occur during the boreal winter, and therefore most directly affect spring–summer field crops in the southern hemisphere. However, some ENSO effects on northern hemisphere spring and summer crops have been found, particularly when researchers have used sub-annual indicators of ENSO conditions that, unlike annual ENSO indices, distinguish between continuity and change prior to or during the crop season. To evaluate the utility of such sub-annual ENSO indicators for agriculture in the SE USA, a tercile-based scoring system was devised to compare four distinct ENSO indices: three monthly ENSO indices and the JMA annual ENSO index. Each index was scored in its ability to predict crop yields differing from the historically normal tercile for corn (Zea mays L.), cotton (Gossypiumhirsutum L.), and peanut (Arachis hypogaea L.). Annual crop yield data were used from selected counties in five Southeastern USA states. No geographic differentiation among the data was included in the analysis. This aggregation of county data increased the sample size for each crop, to address the limitation of a short time-series (47 years) distributed among up to 9 ENSO categories. Statistical significance was compared using contingency tables and the Fisher exact test. Peanut and corn yield response matched best to the Multivariate ENSO Index (MEI) and cotton, to the Oceanic Niño Index (ONI). The MEI and ONI are quantitative indices, while the lower-scoring JMA and Modified-JMA indices are categorical. Therefore, future efforts may reduce the focus on categorical (Niño, Niña, Neutral) classification, and explore the response of crop yields to quantitative ENSO data.