基于最大熵模型的云南小粒咖啡潜在适生区时空变化预估

楚雄师范学院学报 ›› 2026, Vol. 41 ›› Issue (3): 75-89.

基于最大熵模型的云南小粒咖啡潜在适生区时空变化预估

杨晓慧¹, 邓辉敏¹, 李立印¹, 段玮^2,*

1.云南省临沧市气象局,云南临沧 677000;
2.云南省气象科学研究所,云南昆明 650100

收稿日期:2026-01-16 出版日期:2026-05-20 发布日期:2026-07-22
通讯作者: ^*段玮（1979-）,男,正高级工程师,研究方向为山地气象资源和灾害。
作者简介:杨晓慧（1986-）,女,高级工程师,研究方向为气象敏感行业风险预警技术。
基金资助:
云南省气象局科技创新青年基金项目（No. QN202514）

Prediction of Spatiotemporal Changes in Potential Suitable Areas of Coffea arabica in Yunnan Province Based on Maximum Entropy Model

Yang Xiaohui¹, Deng Huimin¹, Li Liying¹, Duan Wei^2,*

1. Meteorological Bureau of Lincang Prefecture, Lincang, Yunnan Province 677000, China;
2. Yunnan Academy of Meteorological Sciences, Kunming, Yunnan Province 650100, China

Received:2026-01-16 Online:2026-05-20 Published:2026-07-22

摘要/Abstract

摘要： 利用云南小粒咖啡种植资料、历史气象资料以及CMIP6不同气候情景预估资料,结合MaxEnt模型分析了影响咖啡适生分布的关键气象与环境变量,提取了云南省小粒咖啡种植适生气候态分布,并对未来近80年（2021-2100年）的适生区时空变化进行了推演分析。结果表明：（1）训练优化后的MaxEnt模型对云南小粒咖啡适生区变化特征捕捉性能较好,模型预测效能AUC=0.939。当前气候条件下（1970-2000年）适生区与实际种植区基本吻合,适生区主要分布在滇西、滇西南和滇南,总面积约9.66×10⁴ km²。（2）研究获取了一套影响云南小粒咖啡适生分布的关键气象与环境变量,包括：年降水量（Bio_12）、最热月最高温（Bio_5）、年温差（Bio_7）、9月降雨量（Pre09）、12月最低温（Tmin12）、海拔（Elev）等9个因子。（3）在CMIP6未来不同气候情景下,到21世纪末的近期（2021-2040年）、中期（2041-2060年）和远期（2081-2100年）云南省小粒咖啡适生区呈总体缩减趋势,其中,在SSPs3-7.0情景下降最为明显,但SSPs2-4.5和SSPs5-8.5情景近期总适生区面积有所增加,中远期在减少;空间上,高适生区向滇西南的普洱、临沧、西双版纳等核心区收缩,非适生区略有扩张,低适生区波动明显。研究结果可为云南省咖啡产业布局规划和制定长期气候适应策略提供参考。

关键词: 云南小粒咖啡, 最大熵模型, 适生区, 气候变化, 时空变化

Abstract: Based on the planting data of Arabica coffee in Yunnan Province, historical meteorological data and the prediction data of different climate scenarios of CMIP6, combined with MaxEnt model, this study analyzes the key meteorological and environmental variables affecting the distribution of areas suitable for planting coffee. The distribution of climate suitable for Arabica coffee grown in Yunnan Province was extracted, and the spatio-temporal changes of suitable areas in the next 80 years （2021-2100） were deduced and analyzed. The results were as follows. Firstly, the MaxEnt model after training optimization has good performance in capturing the change characteristics of the suitable growing area of Coffea arabica in Yunnan, and the prediction efficiency of the model is AUC=0.939. Under the current climate conditions （1970-2000）, the suitable areas are basically consistent with the actual planting areas. The suitable areas are mainly distributed in western, southwestern and southern Yunnan, with a total area of about 9.66×10⁴ km². Secondly, the study obtained a set of key meteorological and environmental variables affecting the suitable distribution of Coffea arabica in Yunnan, including nine factors： annual precipitation （Bio_12）, hottest-month maximum temperature （Bio_5）, annual temperature difference （Bio_7）, September rainfall （Pre09）, December minimum temperature （Tmin12）, and altitude （Elev）. Thirdly, under different climate scenarios of CMIP6 in the future, by the end of this century, the suitable areas for Coffea arabica in Yunnan Province showed an overall reduction trend in the short term （2021-2040）, the medium term （2041-2060） and the long term （2081-2100）, of which the decline was most obvious in SSPs3-7.0. However, the total suitable areas for SSPs2-4.5 and SSPs5-8.5 would increase in the short term and decrease in the medium and long term. Spatially, the high fitness area shrinks to the core areas of Pu’er, Lincang and Xishuangbanna in Southwest Yunnan, while the non-fitness area expands slightly and the low fitness area fluctuates significantly. The results can provide reference for the layout planning of coffee industry in Yunnan Province and the formulation of long-term climate adaptation strategies.

Key words: Yunnan Arabica coffee, Maximum Entropy Model, suitable area, climate change, spatio-temporal variation

中图分类号:

杨晓慧, 邓辉敏, 李立印, 段玮. 基于最大熵模型的云南小粒咖啡潜在适生区时空变化预估[J]. 楚雄师范学院学报, 2026, 41(3): 75-89.

Yang Xiaohui, Deng Huimin, Li Liying, Duan Wei. Prediction of Spatiotemporal Changes in Potential Suitable Areas of Coffea arabica in Yunnan Province Based on Maximum Entropy Model[J]. Journal of Chuxiong Normal University, 2026, 41(3): 75-89.