Methods for quantitatively assess the impacts of natural and anthropogenic factors on vegetation changes of grassland ecosystem

定量区分导致草地生态系统植被变化的自然和人为因素，是生态系统科学管理和保育的关键。因此，综述当前应用较为广泛的定量区分方法，包括主成分分析法、层次分析法、残差趋势法和模型差值法等，比较不同方法的计算原理、优势及误差来源，进而结合典型区域或典型生态系统，对不同方法进行对比和分析。总体而言，每种方法各有其优势和缺点，当前采用同一方法在不同区域或生态系统类型应用的研究较多，但针对方法本身改进或优化的研究较少。此外，针对同一区域开展的不同区分方法间的结果也存在差异。定量区分的结果除受方法本身算法的局限外，也受数据源的时空连续性及分辨率的影响。未来定量区分方法将强调：（1）在指标的选取上，要兼顾全面、多角度、不冗余等原则；（2）进行多源数据与多时空尺度融合，在更高时空分辨率定量区分自然与人为因素，从单一因子的贡献率区分到更多综合性指标的贡献率区分；（3）对定量区分方法本身的改进，这是当前的重点与难点。以期为生态系统适应性管理与关键生态功能的针对性保育提供科学依据和政策导向。;Quantitative assessment the impact of natural factors and anthropogenic factors on grassland ecosystem changes is vital to grassland scientific protection and restoration. Therefore, this paper reviewed and summarized the quantitative assessment methods which are widely used at present, including principal component analysis (PCA), analytic hierarchy process (AHP), residual trend (RESTREND) method and ecosystem model method (difference between remote sensing model and process-based ecosystem model or climate-productivity model). We focused on comparing the algorithm, advantages and drawbacks of different methods. Furthermore, these methods were summarized and analyzed in combination with the typical region or typical ecosystems. In general, each method had its own advantages and disadvantages in data acquisition, temporal and spatial resolution. At present, many studies focused on the application of the same method in different regions or grassland types, but few studies focused on the improvement or optimization of the method itself. In addition, even in the same region and the same grassland type, the results of quantitative assessment of natural and anthropogenic factors on vegetation changes of grassland ecosystem varied in different methods. Besides the limitations of their own calculation methods, the results of quantitative assessment of natural and anthropogenic factors were also affected by the spatial and temporal resolution and continuity of the natural and anthropogenic data sources. In the future, we proposed that studies should focus on these aspects:(1) indicator selection was the first step of quantitative assessment, and it is particularly important to select indicators that can represent the changes of grassland ecosystem in different spatial and temporal scale from different perspectives. As for the selection of natural and anthropogenic factors indicators, we should also pay attention to the non-redundancy and non-correlation between indicators; (2) the data sources of natural and anthropogenic indicators were varied in different scales, it is vital to integrate multi-source of quantitative and qualitative data with different temporal and spatial scale to distinguish the contribution of single factor and the interaction between natural and anthropogenic factors in higher spatial and temporal resolution; (3) the improvement of quantitative assessment method itself, breaking through the limitations of existing methods, and developing new ideas and methods of quantitative assessment of natural and anthropogenic factors on vegetation changes of grassland ecosystem, which is also the key point and difficult point of the future research. For the purpose of providing scientific basis and policy guidance for adaptive management of ecosystem and targeted conservation of key ecological functions of ecosystems.