Work

Early-season crop yield prediction using machine learning and drought conditions in the southeastern United States

Year: 2026

Type: article

Abstract: Drought is a major hydroclimatic hazard with significant impacts on agriculture. However, drought is very unpredictable and the tandem between drought conditions and crop yield is still poorly understood. Particularly, in a humid region such as the southeastern coastal plain of the United States (US) where rainfed agriculture is dominant, a better understanding of crop responses to drought conditions is essential for anticipating yield fluctuations. Hence, this study addresses drought implications on crops by proposing a modeling framework for an early-season prediction of corn, cotton, peanuts, and soybeans yields in a region of the southeastern US encompassing the states of Georgia, North Carolina, and South Carolina. The framework uses county-level standardized precipitation and evapotranspiration index (SPEI) and detrended crop yields time series. Explicitly, a z-score transformation was used to develop a classification scheme that distinguishes gradual levels of crop yields. Supervised Machine Learning models including the random forests (RF) and the k-nearest neighbors (kNN) were distinctly trained and tested for yield prediction based on SPEI. Results showed a disparity of performance depending on the model, the crop, and the prediction scales. For instance, the accuracy of predicting crop yields at regional and state scales was up to 74% with RF and 77% with kNN models. This slight outperformance of the kNN models was consistent at all scales for each of the four crops, but the kNN structure is relatively complex compared to RF. However, macro-averaged F1score values sustained variable predictability among yield levels. Nevertheless, both RF and kNN models showed reasonable performances for predicting three or two levels of yield. Interestingly, the models’ inputs are mid-season SPEI values, allowing yield prediction months ahead of the harvesting periods. Hence, the framework's prediction can be useful for anticipating crop yield anomalies at the regional scale.

Source: Farming System

DOI: 10.1016/j.farsys.2026.100224

Authors: Clement D.D. Sohoulande, C. Prakash Khedun

Institutions: Agricultural Research Service, South Carolina Water Resources Center, Clemson University

Language: English

Topic: Hydrology and Drought Analysis

Subfield: Global and Planetary Change

Field: Environmental Science

Domain: Physical Sciences

SDG: Zero hunger

Open Access status: gold

Estimated APC paid: $1,800