A Radiation‐Modulated Frost Index Improves the Representation of Frost Stress and Yield Response in Winter Wheat

Late spring frost is a major abiotic stress affecting winter wheat ( Triticum aestivum L.) production, yet its impact on yield remains difficult to quantify because crop responses depend not only on the intensity of low‐temperature exposure but also on environmental conditions following frost events. Conventional frost indices are primarily temperature‐based and therefore do not fully capture variability in yield responses under comparable thermal stress conditions. This study aims to improve the representation of frost impacts by incorporating radiation‐related information as an additional explanatory factor. A radiation‐modulated accumulated frost degree day index (R_AFDD) was developed to integrate temperature deficits with radiation conditions. Using daily meteorological data and county‐level yield records from 2015 to 2024 across Shandong Province, China, the performance of this index in explaining yield variability was evaluated in comparison with conventional frost indices. The results indicate that yield responses to frost are nonlinear and strongly condition‐dependent. Compared with temperature‐based indices, the proposed approach improved the ability to explain yield variability and reduced prediction errors. This improvement is associated with a better representation of variability under similar levels of thermal stress, suggesting that incorporating radiation‐related information helps account for differences in environmental conditions linked to yield outcomes. These findings indicate that frost impacts on winter wheat are influenced by both stress intensity and associated environmental conditions. Although the underlying physiological processes were not directly measured, the results are consistent with a role of radiation in shaping conditions that may affect post‐frost responses. Overall, this study provides an empirically improved framework for assessing frost impacts and offers practical value for regional‐scale risk assessment under increasing climate variability.