生物多样性 ›› 2014, Vol. 22 ›› Issue (2): 223-230.DOI: 10.3724/SP.J.1003.2014.08178

• • 上一篇    下一篇

提高生态位模型转移能力来模拟入侵物种 的潜在分布

朱耿平1,*(), 刘强1, 高玉葆2,*()   

  1. 1 天津师范大学生命科学学院, 天津 300387
    2 南开大学生命科学学院, 天津 300071
  • 收稿日期:2013-08-03 接受日期:2013-10-15 出版日期:2014-03-20 发布日期:2014-04-03
  • 通讯作者: 朱耿平,高玉葆
  • 基金资助:
    天津师范大学人才引进基金(5RL127)

Improving ecological niche model transferability to predict the potential distribution of invasive exotic species

Gengping Zhu1,*(), Qiang Liu1, Yubao Gao2,*()   

  1. 1 College of Life Sciences, Tianjin Normal University, Tianjin 300387
    2 College of Life Sciences, Nankai University, Tianjin 300071
  • Received:2013-08-03 Accepted:2013-10-15 Online:2014-03-20 Published:2014-04-03
  • Contact: Zhu Gengping,Gao Yubao

摘要:

生态位模型利用物种分布点所关联的环境变量去推算物种的生态需求, 模拟物种的分布。在模拟入侵物种分布时, 经典生态位模型包括模型构建于物种本土分布地, 然后将其转移并投射至另一地理区域, 来模拟入侵物种的潜在分布。然而在模型运用时, 出现了模型的转移能力较低、模拟的结果与物种的实际分布不相符的情况, 由此得出了生态位漂移等不恰当的结论。提高生态位模型的转移能力, 可以准确地模拟入侵物种的潜在分布, 为入侵种的风险评估提供参考。作者以入侵种茶翅蝽(Halyomorpha halys)和互花米草(Spartina alterniflora)为例, 从模型的构建材料(即物种分布点和环境变量)入手, 全面阐述提高模型转移能力的策略。在构建模型之前, 需要充分了解入侵物种的生物学特性、种群平衡状态、本土地理分布范围及物种的生物历史地理等方面的知识。在模型构建环节上, 物种分布点不仅要充分覆盖物种的地理分布和生态空间的范围, 同时要降低物种采样点偏差; 环境变量的选择要充分考虑其对物种分布的限制作用、各环境变量之间的空间相关性, 以及不同地理种群间生态空间是否一致, 同时要降低环境变量的空间维度; 模型构建区域要真实地反映物种的地理分布范围, 并考虑种群的平衡状态。作者认为, 在生态位保守的前提下, 如果模型是构建在一个合理方案的基础上, 生态位模型的转移能力是可以保证的, 在以模型转移能力较低的现象来阐述生态位分化时需要引起注意。

关键词: 生物入侵, 生态位模型, 模型转移能力, 空间相关性, 生态位保守性

Abstract

Ecological niche modeling (ENM) seeks to characterize the ecological requirements of species using their occurrence in association with environmental variables. The classic applications of ENM to biological invasions involve the calibration of niche modeling in the native range and the subsequent transfer of the calibrated models to other regions to predict areas of potential invasion. However, low niche model transferability has been reported in certain cases, resulting in artifactual conclusions in some studies (e.g., niche shift during a species’ invasion). Improving niche model transferability would allow precise predictions of the invasion potential of species, providing valuable information for invasion risk assessment. In this review, we address model input data (i.e., occurrence records and environmental variables), using the invasive Halyomorpha halys and Spartina alterniflora, to explore protocols for improving niche model transferability. We conclude that the knowledge of the biology, population equilibrium state, geographic distribution, and biogeographic history of the invasive species is crucial prior to niche modeling. In niche model calibration, the sufficient records should not only cover the geographic extent and the ecological dimension of the species’ distribution but also reduce the sample bias and the effects of spatial autocorrelation. Selecting environmental variables should involve considerations of their biological importance in restricting the species’ distribution, the differences in occupied ecological space among geographic populations, and the dimensionality of the environmental space. Delimiting the geographic background for niche modeling should involve considerations of the species’ distributional range and population equilibrium state. We believe that, based on niche conservatism, niche model transferability can be guaranteed if niche models are built based on a reasonable approach. Caution is warranted in the case of interpretations of low niche model transferability in association with niche differentiation.

Key words: biological invasion, ecological niche modeling, transferability, spatial correlation, niche conservatism