Variations in phenology are regarded as a dynamic bio-indicator of global climate change. Rising incomes and the shift toward a diverse diet have been increasing the cash crop demand. However, the response of cash crop phenology to climate change and adaptive management practice remains largely unknown. In this study, using phenology records from 1991 to 2010, we separate the effects of climatic factors and management on phenological changes in the cash crops of sorghum, peanut, and canola based on the statistical and machine learning models. Our results show: (1) The sowing, emergence, three-leaf, and milk ripening date of sorghum showed a significantly (p < 0.05) advanced trend. The sowing, emergence, and five-leaf for canola likewise exhibited a significantly advanced trend (0.55 to 0.91 days a⁻¹). The phenophases of the peanut were generally delayed (0.12 to 0.86 days a⁻¹). (2) For sorghum, canola, and peanut, there is a delayed effect of increasing sunshine hours on heading/anthesis dates. The sowing date for sorghum and peanut delayed with the increased temperature. The three-leaf and milk ripening date of sorghum were sensitive to the wind speed. (3) Sunshine hours contributed to the extension of the whole growth period for sorghum, peanut, and canola, by 68.1, 60.7, and 40.0 %, respectively. The wind speed and temperature had comparable contributions to the maturity date, the former even dominating the heading date of sorghum. (4) Adaptive management practice partially offsets the effect of climate change and supports the length of the whole growth period for sorghum and peanut. Climate change exerts a positive effect on the vegetative growth period for three crops. Our results identified the main climatic factors regulating cash crop phenology. Wind speed must be incorporated in the process-based model to better account for the phenological variations associated with climate change.

物候变化被认为是全球气候变化的动态生物指标。收入的增加和向多样化饮食的转变一直在增加经济作物的需求。然而，经济作物物候学对气候变化和适应性管理实践的反应仍然很大程度上未知。在本研究中，我们利用 1991 年至 2010 年的物候记录，基于统计和机器学习模型，分离了气候因素和管理对高粱、花生和油菜经济作物物候变化的影响。结果表明：（1）高粱的播种期、出苗期、三叶期、乳熟期均呈现显着（p<0.05）提前趋势。油菜的播种、出苗和五叶同样表现出显着的提前趋势（0.55 到 0.91 天 a ^-1）。花生的物候期普遍延迟（0.12 至 0.86 天 a ^-1）。(2) 对于高粱、油菜和花生，增加日照时数对抽穗/开花日期有延迟作用。高粱和花生的播期随气温升高而推迟。高粱的三叶期和乳熟期对风速敏感。(3)日照时间对高粱、花生和油菜整个生育期延长的贡献分别为68.1%、60.7%和40.0%。风速和温度对成熟期的贡献相当，前者甚至主导了高粱的抽穗期。(4) 适应性管理实践部分抵消了气候变化的影响，支持了高粱和花生整个生长期的长度。气候变化对三种作物的营养生长期产生积极影响。我们的研究结果确定了调节经济作物物候的主要气候因素。风速必须纳入基于过程的模型中，以更好地解释与气候变化相关的物候变化。