Agricultural ecosystems have long been recognized as significant greenhouse gas (GHG) sources. To accurately quantify GHG emissions, researchers have developed various process models. However, there are no summary studies of comparative model simulations of GHG emissions from different crop systems. This study compared four widely used process models: APSIM, DNDC, DayCent, and STICS, analyzing their mechanisms, input variables, and simulation results from different crops in simulating GHG emissions. In this study, a total of 94 relevant peer-review literature papers were considered. The research found that these models have strengths in simulating GHG emissions from different crops, but also have certain limitations. DNDC and DayCent performed better in simulating methane (CH) emissions from rice, while APSIM was more effective in simulating nitrous oxide (NO) emissions from maize and wheat. The main factors affecting the simulation results include model mechanisms, management practices, climate, and data availability. To improve model accuracy, it is recommended that future research expands model applicability, evaluates models using standardized measured data, and enhances adaptability by optimizing algorithms and incorporating simulations of key processes.

中文翻译：

---

量化农业系统中的温室气体排放：过程模型的比较分析

农业生态系统长期以来被认为是重要的温室气体（GHG）来源。为了准确量化温室气体排放，研究人员开发了各种过程模型。然而，目前还没有对不同作物系统温室气体排放的比较模型模拟进行总结研究。本研究比较了四种广泛使用的过程模型：APSIM、DNDC、DayCent 和 STICS，分析了它们在模拟温室气体排放方面的机制、输入变量和不同作物的模拟结果。在这项研究中，总共考虑了 94 篇相关的同行评审文献论文。研究发现，这些模型在模拟不同作物的温室气体排放方面具有优势，但也存在一定的局限性。DNDC 和 DayCent 在模拟水稻甲烷 (CH) 排放方面表现更好，而 APSIM 在模拟玉米和小麦一氧化二氮 (NO) 排放方面更有效。影响模拟结果的主要因素包括模型机制、管理实践、气候和数据可用性。为了提高模型的准确性，建议未来的研究扩大模型的适用性，使用标准化的测量数据来评估模型，并通过优化算法和结合关键过程的模拟来增强适应性。