Decomposing environmental, spatial, and spatiotemporal components of species distributions
Decomposing environmental, spatial, and spatiotemporal components of species distributions
复制标题
DOI：
10.1890/10-0602.1
复制
发表时间：
2011-05-01
影响因子：
6.1
通讯作者：
Brandl, Roland
中科院分区：
文献类型：
--
作者：
Hothorn, Torsten;Mueller, Joerg;Brandl, Roland
物种分布模型是预测全球变化对物种分布范围和群落组合影响的重要工具。虽然统计建模在过去十年中取得了相当大的进展，但许多方法仍然忽略了物种分布的重要特征，如非线性和预测因子之间的相互作用、空间自相关和非平稳性，或者最多只包含其中的一些特征。然而，生态学家需要一个同时灵活和一致地处理所有这些特征的建模框架。在这里，我们描述了这样一种方法，除了处理时空自相关和非平稳效应的局部成分外，还可以估计环境变量的全局效应。局部分量可以用来推断未知的时空过程；全局组件描述了物种如何受到环境的影响，并可用于预测，允许拟合许多众所周知的回归关系，从简单的线性模型到复杂的决策树，或从添加模型到受机器学习程序启发的模型。时空预测的可靠性可以通过分别评估局部和全球影响的重要性来进行定性预测。我们通过对主要出现在西欧的猛禽红鸢（Milvus Milvus）的繁殖分布进行建模，基于巴伐利亚州40平方公里网格的两次测绘活动的存在/缺失数据，展示了新方法的潜力。该模型的全球分量选择了从CORINE和WorldClim数据库中提取的7个环境变量来预测红鸢的繁殖。海拔高度的影响在空间上是非平稳的，而且数据具有空间自相关，这表明不考虑空间变化效应和空间自相关的物种分布模型可能忽略了决定红鸢繁殖在巴伐利亚分布的重要过程。因此，不考虑现实世界复杂性的标准物种分布模型的预测可能是相当错误的。我们对红鸢育种的分析体现了物种分布模型创新方法的潜力。该方法同样适用于计数数据的建模。
Species distribution models are an important tool to predict the impact of global change on species distributional ranges and community assemblages. Although considerable progress has been made in the statistical modeling during the last decade, many approaches still ignore important features of species distributions, such as nonlinearity and interactions between predictors, spatial autocorrelation, and nonstationarity, or at most incorporate only some of these features. Ecologists, however, require a modeling framework that simultaneously addresses all these features flexibly and consistently. Here we describe such an approach that allows the estimation of the global effects of environmental variables in addition to local components dealing with spatiotemporal autocorrelation as well as nonstationary effects. The local components can be used to infer unknown spatiotemporal processes; the global component describes how the species is influenced by the environment and can be used for predictions, allowing the fitting of many well-known regression relationships, ranging from simple linear models to complex decision trees or from additive models to models inspired by machine learning procedures. The reliability of spatiotemporal predictions can be qualitatively predicted by separately evaluating the importance of local and global effects. We demonstrate the potential of the new approach by modeling the breeding distribution of the Red Kite (Milvus milvus), a bird of prey occurring predominantly in Western Europe, based on presence/absence data from two mapping campaigns using grids of 40 km 2 in Bavaria. The global component of the model selected seven environmental variables extracted from the CORINE and WorldClim databases to predict Red Kite breeding. The effect of altitude was found to be nonstationary in space, and in addition, the data were spatially autocorrelated, which suggests that a species distribution model that does not allow for spatially varying effects and spatial autocorrelation would have ignored important processes determining the distribution of Red Kite breeding across Bavaria. Thus, predictions from standard species distribution models that do not allow for real-world complexities may be considerably erroneous. Our analysis of Red Kite breeding exemplifies the potential of the innovative approach for species distribution models. The method is also applicable to modeling count data.