欢迎访问《生态学杂志》官方网站,今天是

生态学杂志 ›› 2026, Vol. 45 ›› Issue (9): 3031-3040.doi: 10.13292/j.1000-4890.202609.017

• 研究报告 • 上一篇    下一篇

基于优化MaxEnt模型的金铁锁潜在适生区预测

余淏1,2,苏旭1,2,3,刘玉萍1,2,3*,吕扬1,2,张羽新1,2,才让扎西1,2,郑莹晖1,2

  

  1. 1青海师范大学生命科学学院, 西宁 810008; 2青海师范大学青海省青藏高原生物多样性形成机制与综合利用重点实验室, 西宁 810008; 3青海师范大学高原科学与可持续发展研究院, 西宁 810016)
  • 出版日期:2026-09-10 发布日期:2026-09-08

Prediction of the potential suitable habitat of Psammosilene tunicoides (Caryophyllaceae) based on the optimized MaxEnt model.

YU Hao1,2, SU Xu1,2,3, LIU Yuping1,2,3*, LYU Yang1,2, ZHANG Yuxin1,2, CAIRANG Zhaxi1,2, ZHENG Yinghui1,2   

  1. (1School of Life Sciences, Qinghai Normal University, Xining 810008, China; 2Key Laboratory of Biodiversity Formation Mechanism and Comprehensive Utilization of the QinghaiXizang Plateau in Qinghai Province, Qinghai Normal University, Xining 810008, China; 3Academy of Plateau Science and Sustainability, Qinghai Normal University, Xining 810016, China).

  • Online:2026-09-10 Published:2026-09-08

摘要: 为科学预测珍稀濒危药用植物金铁锁(Psammosilene tunicoides)的潜在适生区,本研究利用野外分布数据和多源环境因子,采用交叉验证与网格搜索法优化MaxEnt模型参数,模拟当前及未来不同气候变化情景下金铁锁的适生区,分析主导环境因子及适生区质心迁移规律。结果表明:(1)优化后的MaxEnt模型预测精度较高,平均训练数据AUC值达0.979。(2)等温性、海拔、年降水量、气温季节性变动系数、最冷月最低温度和气温年较差是影响金铁锁分布的主导环境因子,累计贡献率和置换重要性分别达93%和78.8%。(3)当前气候背景下,金铁锁的潜在适生区总面积为167.47×104 km2,主要分布于西藏东部、云南、四川、贵州等地,其中高适生区集中于横断山脉中南段,面积29.65×104 km2,占比17.7%。(4)未来不同SSP情景下,金铁锁的适生区呈“收缩—扩张—收缩”的阶段性变化,2030s以收缩为主,2050s时SSP5-8.5情景下川西—藏东高海拔地区适生区收缩,2070s时SSP3-7.0和SSP5-8.5两种高排放情景下显著扩张,低排放情景不同程度收缩,至2090s时扩张潜力趋于饱和;适生区质心总体沿西北向迁移,迁移率随排放强度增大而加快。本研究可为金铁锁就地保护、迁地保育、人工栽培以及资源可持续利用提供重要科学依据。


关键词: 金铁锁, 最大熵模型, 环境因子, 模型优化, 质心迁移

Abstract: To predict the potential suitable habitats of the rare and endangered medicinal plant Psammosilene tunicoides, we used occurrence records combined with multi-source environmental variables to construct a species distribution model, simulating the distribution of suitable habitats under current and future climate change scenarios. We further analyzed the dominant environmental factors as well as the spatiotemporal dynamics and migration patterns of the centroid of suitable habitats, by adopting optimized MaxEnt model parameters obtained via cross-validation and grid search. The results showed that: (1) The optimized MaxEnt model exhibited high predictive performance, with an AUC value of 0.979 based on the average training dataset. (2) Isothermality, altitude, annual precipitation, temperature seasonality, minimum temperature of the coldest month, and temperature annual range were the dominant factors affecting the distribution of P. tunicoides, with a cumulative contribution rate of 93% and a cumulative permutation importance of 78.8%. (3) Under current climatic conditions, the total area of potential suitable habitats for P. tunicoides was 167.47×104 km2, mainly distributed in eastern Xizang, Yunnan, Sichuan, and Guizhou. Among them, highly suitable habitats were concentrated in the middle and southern sections of the Hengduan Mountains, with an area of 29.65×104 km2, accounting for 17.7% of the total suitable area. (4) Under future SSP scenarios, the suitable habitat for P. tunicoides exhibits a phased transformation characterized by a “contraction-expansion-contraction” pattern. In the 2030s, habitat contraction is the dominant trend. After the 2050s, high-emission scenarios show expansion into the high-altitude regions of Western Sichuan and Eastern Tibet. By the 2070s, only the SSP5-8.5 scenario shows expansion, while all other scenarios experience varying degrees of contraction. By 2090s, the potential for expansion tends to become saturated. In addition, the centroid of suitable habitats generally migrated toward the northwest, and the migration rate increased with higher emission intensity. These findings provide an important scientific basis for in situ and ex situ conservation, artificial cultivation, and the sustainable utilization of germplasm resources of P. tunicoides.


Key words: Psammosilene tunicoides, MaxEnt, environmental factor, model optimization, centroid migration