Use of Climate Indices in Cotton Yield Risk Assessment for the Southeastern USA.

Cotton is one of the most important crops in the Southeastern USA economically, accounting for the major share of total cotton production of the United States. At the same time, cotton in the Southeast deals with the yield/production risk factors. Climate is one of the most influencing factors but sometimes its control is beyond the reach of growers. Knowing cotton yield risk in advance based on climate information would aid the growers in making alternative decisions to manage risk. The rationale was that the climate indices can be used as early indicators of cotton yield because it has lagged correlation with weather conditions during the cropping season. The main objective was to use climate indices to assess yield risk in cotton before the season in order to aid growers in managing risk. Several Pacific and Atlantic climate indices (Jan-Apr) were correlated with monthly (May-Oct) precipitation, and maximum temperature for the selected counties in Georgia, and Alabama. The same climate indices were also correlated with historic cotton yield residuals. Multiple linear regressions of significant principal components of climate indices with historic cotton yield residuals were carried out to predict cotton yield. Cross validation was used to evaluate the regression models. Predicted cotton yield residuals were re-categorized in to different groups to obtain the first yield risk prediction by the end of February. This forecast can be useful to growers when deciding about crop insurance coverage. After that the Jan through Apr climate indices were used to make another yield risk forecast by the end of April. This forecast can be useful to growers for marketing decisions. The pre-season climate indices showed statistically significant correlation with monthly weather, and historic cotton yield residuals. The principal component regression models were found to be a promising tool for the growers in managing their risk.