基于数据挖掘方法的草原放牧策略研究

Research on Grassland Grazing Strategies Based on Data Mining Methods

摘要: 草原在生态系统中发挥着不可或缺的作用，合理的放牧策略对保护草原生态环境有积极的作用。本文以典型的温带草原—内蒙古锡林郭勒草原为研究对象，寻求该地区相对较优的放牧策略。首先，分析放牧强度与土壤物理性质和植被生物量的联系，以逻辑回归模型进行拟合，通过混淆矩阵和回归结果证明了模型的合理性。并采用层次分析法建立指标体系分析草原的放牧强度和放牧方式的关系。然后，采取主成分分析法对影响土壤沙漠化的因素进行分析，得到不同放牧强度对应的土壤沙漠化程度指数；接着分析影响土壤板结化的因素，得到不同放牧强度对应的土壤板结化程度指数；最后，得到最佳的放牧策略为中度放牧强度——选择划区轮牧的方式，使荒漠化板结化程度最低。本文的研究方法能为其他草原地区的放牧策略的制定提供参考。

Abstract: Grasslands play an integral role in the ecosystem and sensible grazing strategies can have a positive effect on the conservation of grassland ecosystems. This paper takes a typical temperate grassland, the Xilinguole grassland in Inner Mongolia, as the subject of this study to seek a relatively optimal grazing strategy for the region. Firstly, the link between grazing intensity and the physical proper-ties of the soil and vegetation biomass was analyzed and fitted with a logistic regression model, which was justified by the confusion matrix and regression results. A hierarchical analysis was also used to establish a system of indicators to analyze the relationship between grazing intensity and grazing practices in grasslands. Then, the factors affecting soil desertification were analyzed by principal component analysis to obtain the index of soil desertification corresponding to different grazing intensities; then the factors affecting soil denudation were analyzed to obtain the index of soil denudation corresponding to different grazing intensities; finally, the best grazing strategy was obtained as moderate grazing intensity—choosing the zoned rotational grazing method to minimize desertification denudation. The methodology of this paper can provide a reference for the develop-ment of grazing strategies in other grassland areas.