• Central spikelets experienced the greatest productivity loss at low temperatures. • Low-temperature stress affected the biomass partitioning across spikelet positions. • Genetic parameter LTS was used to reflect low-temperature sensitivity of cultivars. • New model enhanced simulation of grain number and weight across spikelet positions. Preflowering low-temperature stress significantly threatens yield formation in winter wheat, yet existing crop models cannot mechanistically capture spikelet-level spatial variability in grain traits under stress. To address this critical limitation, a novel process-based prediction framework is developed. Environment-controlled phytotron experiments conducted at the booting stage revealed position-dependent responses of spikelets to low-temperature stress, with central spikelets experiencing the greatest absolute reductions in grain number and weight. These findings link low-temperature stress effects to spikelet-specific biomass partitioning through revision of the maximum partitioning index (PISP max ). A cultivar-specific and stage-dependent low-temperature sensitivity (LTS) parameter was further introduced to quantify differential stress responses. The incorporation of these mechanisms into the WheatGrow model reduced the errors in simulating grain number and grain weight by 43% and 46%, respectively, when tested against independent controlled-stress datasets. Further validation with multigenotype observations from frost-prone environments in Australia demonstrated robust model applicability under natural frost conditions in the field. A sensitivity analysis confirmed the reliable reproduction of grain distribution patterns under varying low-temperature stress scenarios. Overall, this study advances crop modeling under low-temperature stress from whole-plant resolution to within-spike precision, providing a physiologically based breeding target for developing cold-tolerant wheat cultivars.
Ji et al. (Thu,) studied this question.