Relationships between ENSO and crop yield across the southeast US

Atmospheric conditions play an integral role in agricultural production across the Southeast US region. El Niño-Southern Oscillation (ENSO) has been linked to climate variability across the Southeast US, including shifts in temperature and precipitation. Large scale climate patterns, such as ENSO, can also influence crop growth. As a result, specific relationships can be identified and are particularly meaningful since ENSO is currently the only predictable seasonal climate oscillation. This research explores the historical relationship between seasonal ENSO and crop yields of peanuts, cotton, and corn across the southeastern US from 1950 to 2009.

To establish if these relationships are sensitive to the ENSO index utilized, three different ENSO indices are evaluated: Japan Meteorological Agency (JMA), Oceanic Niño Index (ONI), and Multivariate ENSO Index (MEI). Monthly ENSO values are averaged over five seasonal periods of DJF (December-January-February), MAM (March-April-May), JJA (June-July-August), SON (September-October-November), and growing season (May-June-July-August-September).

To determine if these relationships are sensitive to the spatial patterning of climate, three different spatial aggregations are examined: counties, National Climatic Data Center (NCDC) climate divisions, and National Drought Mitigation Center (NDMC) experimental clusters.

Since seasonal ENSO data are non-normal and the sample size is relatively small (N<100), correlations are computed using Spearman's method to determine the relationship between seasonal ENSO values and annual crop yield anomaly percentages. The five seasonal ENSO periods, three crops, three spatial aggregations, and three ENSO indices are analyzed to identify regional patterns during different ENSO seasons and among different ENSO indices.

Only 27 out of 14,220 total correlations are stronger than |0.5| when considering the relationship between crop yield anomaly percentages and seasonal ENSO values. Most correlations are considered weak between growing season ENSO and crop yield anomalies for all crops, spatial aggregation types, and ENSO indices, with values less than |0.4|. Higher correlations are observed during pre-growing season periods of DJF and MAM, especially for corn. Overall, corn has the strongest relationship with ENSO yet most correlations are less than |0.5|.

Spatial aggregations with larger area (NCDC, NDMC) emphasize regional signal as compared with county-level scale. The spatial pattern is usually consistent across the three spatial aggregation types. However, differences in spatial pattern are observed between the three ENSO indices although sometimes two or more are similar. MEI values have the highest spatial correlation with crop yield for all spatial aggregations, all crops, and ENSO periods of DJF, MAM, and growing season.

Overall, MEI is recommended as the best ENSO index for the relationship between crop yield anomaly percentages and ENSO. NDMC experimental clusters are recommended as the best spatial aggregation for this particular relationship. However, ENSO is only a weak indicator of summer crop yield across the Southeast US.