Crop growth models, such as the WOrld FOod STudies (WOFOST) model, mimic the mechanistic processes involved in crop development, growth, and yield production. The accuracy of simulation is decreased in unfavorable low‐temperature settings because these models do not accurately represent crop response processes in low‐temperature stress. Enhancing the WOFOST crop growth model's accuracy in simulating crops' responses to cold temperatures is the aim of this work. Given its vulnerability to low temperatures, the inquiry uses winter wheat in Henan Province as a focal point. It integrates the WHEATGROW wheat phenology model with the Frost model of Lethal Temperature 50 (FROSTOL) inside the framework of the crop growth model. This link aims to improve simulation accuracy and supplement the model's mechanisms, particularly when it comes to the impact of low temperatures on crop development. The study uses Long Short‐Term Memory networks to build a yield model that integrates remote sensing data with information from simulated crop models. Under low temperatures, the leaf area index, total above ground biomass, and total weight of storage organs of the model WWF—which combines FROSTOL and WHEATGROW with WOFOST—show a considerable decline. It was discovered that there is a greater improvement in simulation accuracy of the linked model WWF relative to the WOFOST model in frost years than in normal years, based on a comparison analysis between typical frost years and normal years. To be more precise, the improvement is 8.03% in frost years and 1.98% in regular years. When all is said and done, the coupled model advances our knowledge of how winter wheat is impacted by low temperatures.

中文翻译：

---

基于多模型耦合的低温胁迫对冬小麦的影响研究

作物生长模型，例如世界粮食研究 (WOFOST) 模型，模拟作物发育、生长和产量产生的机械过程。在不利的低温环境下，模拟的准确性会降低，因为这些模型不能准确地代表低温胁迫下作物的响应过程。这项工作的目的是提高 WOFOST 作物生长模型模拟作物对低温反应的准确性。鉴于其易受低温影响，本次调查以河南省冬小麦为重点。它将 WHEATGROW 小麦物候模型与致死温度 50 (FROSTOL) 的霜冻模型集成在作物生长模型的框架内。该链接旨在提高模拟精度并补充模型的机制，特别是在低温对作物发育的影响方面。该研究使用长短期记忆网络构建一个产量模型，将遥感数据与模拟作物模型的信息相结合。在低温下，将 FROSTOL 和 WHEATGROW 与 WOFOST 相结合的 WWF 模型的叶面积指数、地上总生物量和储存器官总重量均出现相当大的下降。通过典型霜冻年与正常年的对比分析发现，霜冻年联动模型WWF相对于WOFOST模型的模拟精度比正常年有较大提高。准确地说，霜冻年改善8.03%，平年改善1.98%。总而言之，耦合模型增进了我们对冬小麦如何受到低温影响的了解。