Forecasting end-of-season peanut yield using mid-season crop condition rating signals

Abstract

Peanut production in the United States is concentrated in three geographically distinct regions (southeast, southwest, and Virginia-Carolinas), positioning it as a specialized commodity relative to other major crops. Due to the peanut industry’s relatively small and specialized structure, limited literature exists on yield forecasting for peanuts. Substantial research exists on yield forecasting using crop condition ratings for corn and soybeans; however, a notable gap remains for peanuts. Despite being the seventh most valuable crop in the United States, limited empirical work has focused on developing yield forecasting models for the peanut industry, underscoring the importance of this study (USDA NASS 1999). A key challenge in peanut yield forecasting is the limited availability of publicly accessible data. As a privately traded commodity, peanuts rely primarily on USDA reporting, which is consistently available only at the state level. Data were obtained from the USDA National Agricultural Statistics Service (NASS) based on survey reports of peanut crop conditions and yields for major producing states, including Alabama, Florida, Georgia, the Carolinas, and Texas. Structural break analysis indicated that observations prior to 2012 were not informative for forecasting purposes; therefore, the study period was restricted to 2012–2025. Forecast performance was evaluated weekly throughout the peanut growing season using multiple modeling approaches, including the Good + Excellent, NCI, NCI 2.0, Linear, and Stress weighting models. Results indicate that the Good + Excellent Scheme consistently outperformed alternative modeling approaches across the evaluated weeks. By week 40 of the calendar year (early October), the Good + Excellent Scheme achieved the lowest forecast error (RMSE = 0.16; MAE = 0.13), establishing it as the most accurate model among those tested. Forecasts became reliable, beginning around week 32 and remained statistically significant through the end of the growing season. Correlations between Good + Excellent condition ratings and final yield were likewise significant from week 32 onward based on historical crop-year data. Yield forecasting plays a critical role in the peanut industry by informing marketing and operational decision-making. Accurate forecasts provide value to marketing teams by enabling early assessment of crop quality and anticipated supply for upcoming sales periods. Forecast information also supports planning for storage and processing capacity at shelling and manufacturing facilities. Together, projections of crop size and quality guide marketing strategies and market allocation decisions for the subsequent production year.