生态系统服务时空演化的非线性影响因子——以钱塘江流域为例

为深入探究不同尺度、不同类别影响因素对不同类型生态系统服务的影响程度和复杂关系，本文测定了钱塘江流域7项生态系统服务的时空变化，并通过随机森林模型识别了各单项生态系统服务不同尺度自然环境因素和社会经济因素的影响程度和相对重要性，并评估了主导影响因素的边际效应.结果表明：2010~2017年，各单项生态系统服务在时空变化上具有明显的空间分异特征；不同类型生态系统服务受不同尺度自然环境因素和社会经济因素的共同影响，但其影响程度存在明显差异，格网尺度、自然环境因素的影响相对较大，建设用地、耕地及与其直接相关的社会经济因素的变化改变和塑造着生态系统服务的演化过程；主导影响因素与各单项生态系统服务之间存在复杂的非线性关系，各影响因素指标的变化对各年份生态系统服务具有非平稳的影响，如建设用地增加初期对多项生态系统服务的影响更为复杂和剧烈.

Spatiotemporal dynamics of ecosystem services and their nonlinear influencing factors-A case study in the Qiantang River Basin

In order to unravel the impacts of different categories of influencing factors across scales on different types of ecosystem services, this paper measured 7 ecosystem services and untangled their spatiotemporal changes from 2010 to 2017 within the Qiantang River Basin, China. The magnitude and relative importance of influence for natural factors and socio-economic factors across different scales of each ecosystem service type are identified using the random forest algorithm, and the marginal effects of the dominant influencing factors are quantified. The results showed that ecosystem services presented obvious spatially heterogenous characteristics. Different types of ecosystem services were jointly affected by varying natural factors and socio-economic factors across scales, but their influence degrees varied with types. The natural environmental factors at grid scale had relatively larger influence. The changes of construction land, cultivated land and directly related socio-economic factors shaped the evolution process of ecosystem services. There existed a complex nonlinear relationship between the leading influencing factors and individual ecosystem services. The change of the indicators of each influencing factor had a non-stationary impact on ecosystem services in each year. For example, the impact of the initial increase of construction land was much more complex and intense for many ecosystem services.