基于随机森林模型的国家重点保护陆生脊椎动物物种优先保护区的识别

准确可靠地识别国家重点保护陆生脊椎动物物种的优先保护区,是生物多样性保护的热点问题之一。采用随机森林(random forests)模型,基于12个环境变量,对中国263种国家重点保护陆生脊椎动物建模,并预测各个物种在背景点的适生概率,迭加计算得到国家重点保护陆生脊椎动物物种的生境适宜性指数。此外,基于对生境适宜性指数的空间自相关分析,识别和确定国家重点保护陆生脊椎动物物种优先保护区,并对优先保护区目前的被保护情况进行分析。结果表明,国家重点保护陆生脊椎动物物种的优先保护区的面积为103.16万km~2,约占我国国土面积的10.90%。优先保护区主要分布在我国的西部地区,包括西南地区的秦岭-大巴山山区、云南省与印度及缅甸的交界地区、武陵山山区、喜马拉雅山-横断山脉山区、阿尔泰山脉山区、天山山脉山区、昆仑山山脉山区;东北的大、小兴安岭、东北-华南沿海地区及长江中下游地区有少量分布。优先保护区中被保护的面积为50.40万km~2,占优先保护区总面积的48.86%,保护率偏低,未被充分保护。利用系统聚类分析,将未被保护的优先保护区划分成3种优先保护顺序,以期为相关部门的决策提供科学依据,更好地保护生物多样性。

Identifying the priority conservation areas for key national protected terrestrial vertebrate species based on a random forest model in China

The accurate and reliable identification of key national protected terrestrial vertebrate species in China is vital to biodiversity conservation. In this research, 12 environmental variables were used to model the 263 key national protected terrestrial vertebrate species with a random forest model, to predict the probability of occurrence of each species in background points. The habitat suitability index of the key national protected terrestrial vertebrate species was calculated by superimposing the data. Furthermore, spatial autocorrelation analysis performed on the habitat suitability index to identify priority conservation areas for these key species, with the protection status of these priority conservation areas analyzed using gap analysis. The results show that priority conservation areas account for 10.90% of China''s land area (1031600 km²). These priority conservation areas are distributed primarily in western China, including in the Qinling-Daba mountain area in southwestern China, the Yunnan and India-Burma border region, the Wuling mountain area, the Himalayas-Hengduan mountain area, the Altai mountain area, the Tian Shan mountain area, and the Kunlun mountain area. Some distribution is also found in the Daxing''anling and Xiaoxing''anling mountain areas in northeast China, the northeast-south China coastal areas, and the middle and lower reaches of the Yangtze River. The protected areas found within the priority conservation areas account for 48.86% of the priority conservation areas (504 000 km²). The rate of protection is relatively low, meaning that the priority conservation areas are not sufficiently protected. Using a hierarchical cluster analysis, a priority protection order for the unprotected priority conservation areas was presented, to serve as a scientific basis for decision-making for relevant departments and provide better protection of biodiversity in this region