新疆植被NDVI时空变化及定量归因

为揭示自然与非自然因素以及因素之间的交互作用对新疆植被覆盖变化的影响,以归一化植被指数(Normalized Difference Vegetation Index,NDVI)为植被覆盖状况表征指标,结合气候、地形、水文、土壤以及人为干扰等15种因子,采用线性趋势分析、随机森林和地理探测器模型,分析了2000—2018年新疆地区植被NDVI的时空演变特征,并对其空间分异的驱动力进行了定量探测。结果表明:新疆NDVI分布呈现北高南低、西高东低、山区高平原低的特点; 研究期间植被覆盖状况有明显改善,60.91%的植被区域NDVI呈增加趋势。土壤湿度的解释力最高(0.394),与耕地面积和潜在蒸散发共同主导植被NDVI的时空格局; 双因子的交互作用能以非线性方式增强对NDVI空间分布的解释力,其中,尤以土壤湿度和潜在蒸散发的交互作用影响力最强; 各因子存在促进植被生长的最适宜范围或特征,其中,NDVI空间变化基本与水分因素呈正相关,与辐射量、海拔呈负相关。总体上,土壤水分条件对新疆植被变化的影响更为直接,此外,适度的人类活动对植被恢复已有明显的促进作用。

Spatiotemporal Dynamics of NDVI in Xinjiang and Quantitative Attribution Based on Geodetector

The spatial and temporal changes of vegetation in Xinjiang were firstly identified based on the Normalized Difference Vegetation Index(NDVI)data from 2000 to 2018 by using linear trend analysis.In combination with 15 other factors including climate,topography,hydrology,soil,and human disturbance etc.,the spatial stratification of the impacts of natural and non-natural factors on vegetation changes was then explored by applying the geographical detector model with the Random Forest.The results showed that:the vegetation coverage of the whole region was characterized by high in the north and west,low in the south and east,while the vegetation coverage of the mountains were higher than that of the plains;during 2000—2018,the vegetation had improved significantly,with 60.91%of the vegetated areas showing an increasing trend in NDVI;root-zone soil moisture had the highest explanatory power(0.394),and dominated the spatio-temporal patterns of NDVI together with arable area and potential evapotranspiration;the interaction of each two factors could enhance the explanatory power on the spatial distribution of NDVI in a non-linear manner,among which the interaction of root-zone soil moisture and potential evapotranspiration had the strongest influence in particular;there existed optimum ranges or characteristics for each factor to promote vegetation growth,among which,NDVI spatial variation was basically positively correlated with moisture factors and negatively correlated with radiation and elevation.In conclusion,vegetation change in Xinjiang is more directly influenced by soil moisture conditions,and in addition,moderate human activities have had a significant contribution to vegetation recovery.