基于Maxent模型的假臭草潜在分布区预测

Prediction of Potential Distribution Area of Praxelis clematidea Based on Maxent Model

摘要:

为有效开展假臭草的防除工作，运用最大熵值模型（Maxent）和地理信息技术（GIS）对假臭草（Praxelis clematidea）在中国的潜在分布区和影响其分布的主导环境变量进行预测。运用AUC法对预测精度进行评价，达到“优秀”水平。结果表明，Maxent模型预测结果具有较高的可信度。假臭草在中国的潜在适生面积为785 985 km²，占总国土面积的8.19%，主要分布在华南及东南区域，其中福建、广东、海南和广西省为假臭草的中高适生区。福建省的假臭草分布最为广泛，61.98%的福建区域适合假臭草的生长，其中闽南沿海为假臭草的高适生区，而闽北区域假臭草的种群密度较低。2月最低温可能是限制假臭草分布的主要气候变量，贡献率为61.70%；最干季度降水量、9月最高温对假臭草的分布具有一定的影响。假臭草的分布预测为早期监测和预警提供理论依据和技术支持，防止假臭草的进一步扩散蔓延。

Abstract:

In order to prevent and remove effectively invasive plant Praxelis clematidea, the potential distribution area and dominant environment variables were predicted by using the maximum entropy model (Maxent) and the Geographic Information System (GIS) technology. The accuracy of prediction was evaluated at "excellent" level by AUC (area under curve) method. The results showed that the prediction of Maxent model had high reliability. The potential suitable growth area of P. clematidea in China was 785 985 km², accounting for 8.19% of all China land area. This species mainly distributed in southern and southeastern China, especially in Fujian, Guangdong, Hainan and Guangxi (high suitable areas). The distribution of P. clematidea in Fujian Province was the widest, which 61.98% of the area was suitable for growth. It was high suitable area of P. clematidea in coast of south Fujian, but the population density in the north of Fujian was low relatively. The minimum temperature in February might be the restrictive variable of the distribution of P. clematidea with the contribution rate of 61.7%. Precipitation of the driest quarter and maximum temperature in September have a certain influence on its distribution. These provided a theoretical basis for the prevention and control of P. clematidea and some effective measures must be taken to prevent its spread for the further.