基于MaxEnt模型预测气候变化下杉木在中国的潜在地理分布

为研究杉木在中国的分布特征及其对气候变化的响应模式,本研究基于现有分布记录,应用最大熵(MaxEnt)模型和地理信息系统方法,结合气候、地形等环境要素,预测杉木在当前和未来气候变化下的潜在适生区。结果表明: 影响杉木分布的最主要因素是年平均降水量,在当前气候下,杉木适生区合计面积328万km²,占全国陆地总面积的34.5%,低、中和高适生区分别占18.3%、29.7%与52.0%。在未来气候情景下,杉木生长的适宜性在我国总体上呈上升趋势,适生区面积随气候变化增大,且明显向北扩张,南方湿润亚热带地区形成集中连片高适生区。模型经受试者工作特征曲线检验,训练集平均受试者工作特征曲线下面积为0.91,可信度高。

Identification of the potential distribution area of Cunninghamia lanceolata in China under climate change based on the MaxEnt model

Based on the distribution records of Cunninghamia lanceolata, we used the maximum Entropy (MaxEnt) model and geographic information system (GIS) methods, combined with environmental factors such as climate and terrain, to predict the potential distribution areas suitable for C. lanceolata under current and future climate scenarios. The results showed that annual precipitation was the most important factor driving the distribution of C. lanceolata. Under the current climate scenario, the total area of suitable for C. lanceolata growth was about 3.28 million km², accounting for about 34.5% of the total land area of China. Among all the suitable areas, the lowly, intermediately, and highly suitable areas accounted for 18.3%, 29.7% and 52.0% of the total, respectively. Under future climate scenarios, the suitable area of C. lanceolata would increase, showing a clear trend of northward expansion in China. A concentrated and contiguous distribution region highly suitable for C. lanceolata would appear in the humid subtropical areas of southern China. The model was tested by the receiver operating characteristic curve (ROC). The average area under the curve of ROC of the training set was 0.91, showing high reliability.