Abstract: As a key economic crop in China，cotton plays a vital role in the agricultural economy of Xinjiang. Against the background of intensifying climate change，the precise assessment of suitable cotton planting areas is critical for optimizing agricultural layouts. Based on the MaxEnt model and ArcGIS technology，this study integrated 1 402 cotton occurrence records and 47 environmental factors to construct a prediction model for the potential distribution of cotton in Xinjiang.The results indicated that slope（55.7%），≥10 ℃ accumulated temperature（25.8%），and water vapor pressure（10.0%） had a cumulative contribution rate of 91. 5% in the screening of dominant environmental factors. In addition，annual mean temperature showed the highest permutation importance value（39.6%），and together these factors constituted the core constraint system affecting the spatial distribution of cotton.Model validation showed that the average AUC value of the training set reached 0. 918，indicating that the MaxEnt model had high reliability in predicting the potential distribution of cotton in Xinjiang. Environmental response curves further revealed that cotton suitability significantly declined with increasing slope，while gentle terrain conditions with a slope of less than 2° and sufficient heat accumulation，with ≥10 ℃ accumulated temperature ranging from 1 600 ℃·d to 3 200 ℃·d were necessary prerequisites for cotton growth and development. In terms of spatial zoning characteristics，the total potential suitable area for cotton in Xinjiang was 509 737 km^²，accounting for 30.61% of the total land area of Xinjiang. High⁃，moderate⁃，and low⁃suitability zones accounted for 20.43%，18.52%，and 61.05% of the total suitable area，respectively. High⁃suitability zones were concentrated at the northern foot of the Tianshan Mountains，the Ili River Valley，and the northern marginal oases of the Tarim Basin，while low⁃suitability zones were mainly located at oasis margins and other secondary habitats. This study clarifies the spatial suitability pattern and dominant factors of cotton in Xinjiang，providing a scientific basis for the optimization of local agricultural planting structures and the rational allocation of land and water resources.

Key words: Cotton, MaxEnt model, Geographic information system, Suitable zoning, Environmental factors

摘要： 棉花作为我国重要的经济作物，对新疆的农业经济具有重要意义。随着气候变化的加剧，精准评估棉花适宜种植区域对优化农业布局至关重要。基于MaxEnt模型和ArcGIS技术，整合1 402个棉花分布点数据及47个环境因子，构建新疆棉花潜在分布预测模型。结果表明，在主导环境因子筛选中，坡度（55.7%）、≥10 ℃积温（25.8%）以及水蒸气压（10.0%）的累计贡献率达91.5%，且年平均温度的置换重要度值最高（39.6%），共同构成影响棉花空间分布的核心约束体系。模型精度验证显示，训练集的平均AUC值达到0.918，证明MaxEnt模型在预测新疆棉花潜在分布方面具有极高的可靠性。环境因子响应曲线进一步揭示，棉花适宜性随坡度增加而显著下降，平缓的地形条件（坡度<2°）及充足的热量积累（≥10 ℃积温1 600~3 200 ℃·d）是棉花生长发育的必要前提。在空间区划特征上，新疆棉花潜在适生区总面积为50.973 7万km²，占全疆土地总面积的30.61%，其中，高、中、低适宜区分别占适生区总面积的20.43%、18.52%和61.05%。高适宜区集中分布于天山北麓、伊犁河谷及塔里木盆地北缘绿洲，而低适宜区则主要位于绿洲外缘等次生生境。明确新疆棉花的空间适宜性格局及其主控因子，可为当地农业种植结构优化及水土资源合理配置提供科学依据。

关键词: 棉花, MaxEnt模型, 地理信息系统, 适宜性区划, 环境因子