基于MaxEnt 的云南省薇甘菊分布预测及评价

    Distribution prediction and assessment of Mikania micrantha in Yunnan province based on MaxEnt model

    • 摘要: 利用MaxEnt 生态位模型预测云南薇甘菊适生区分布,结果显示,ROC 曲线的AUC 值均在0.9 以上,预测结果具有较高的可信度。影响薇甘菊分布的主要环境因子包括最湿季降雨量(贡献率28.8%)、海拔(贡献率22.1%)、土地覆盖现状(贡献率15.3%)、最湿季平均温、年均降水量、年平均温。从因子的响应曲线分析得出,云南薇甘菊最适环境参数为海拔1 100 m,最湿季平均温为22益,最湿季降雨量为760 mm,土地覆盖为未郁闭林地、灌木林地和荒地等。从样本点数量、环境因子选择及人为影响3 个方面对MaxEnt 模型预测的云南薇甘菊分布结果进行比较和评价。

       

      Abstract: MaxEnt niche model was used to predict the suitable distribution area of Yunnan Mikania micrantha. The results showed that AUC values of the ROC curve were all above 0.9, suggesting that the predicted results had relatively high credibility. The main environmental factors affecting the distribution of M. micrantha included the rainfall in the wettest season (contribution rate of 28.8%), elevation (contribution rate of 22.1%), present land cover (contribution rate of 15.3% ), average temperature in the wettest season, annual rainfall, and annual average temperature. In the light of analysis on the response curve of factors, the optimal environmental parameters for Yunnan M. micrantha were as follows: an elevation of 1 100 m, an average temperature of 22 in the wettest season, 760 mm of rainfall in the wettest season, and land cover of free growing forest land, shrub land and wasteland, etc. Finally, from the aspects of sample point count, selection of environmental factors and man-made influences, the distribution areas of Yunnan M. micrantha predicted via MaxEnt model were compared and evaluated.