高级检索+
文检, 宋经元, 谢彩香, 张琴, 曾凡琳, 张艺. 基于最大信息熵模型的能源物种麻疯树潜在适宜区[J]. 植物科学学报, 2016, 34(6): 849-856. DOI: 10.11913/PSJ.2095-0837.2016.60849
引用本文: 文检, 宋经元, 谢彩香, 张琴, 曾凡琳, 张艺. 基于最大信息熵模型的能源物种麻疯树潜在适宜区[J]. 植物科学学报, 2016, 34(6): 849-856. DOI: 10.11913/PSJ.2095-0837.2016.60849
WEN Jian, SONG Jing-Yuan, XIE Cai-Xiang, ZHANG Qin, ZENG Fan-Lin, ZHANG Yi. Identification of Potential Distribution Areas for Energy Plant Jatropha curcas L. Using the Maxent Entropy Model[J]. Plant Science Journal, 2016, 34(6): 849-856. DOI: 10.11913/PSJ.2095-0837.2016.60849
Citation: WEN Jian, SONG Jing-Yuan, XIE Cai-Xiang, ZHANG Qin, ZENG Fan-Lin, ZHANG Yi. Identification of Potential Distribution Areas for Energy Plant Jatropha curcas L. Using the Maxent Entropy Model[J]. Plant Science Journal, 2016, 34(6): 849-856. DOI: 10.11913/PSJ.2095-0837.2016.60849

基于最大信息熵模型的能源物种麻疯树潜在适宜区

Identification of Potential Distribution Areas for Energy Plant Jatropha curcas L. Using the Maxent Entropy Model

  • 摘要: 麻疯树(Jatropha curcas L.)为传统能源植物,是作为生物柴油最具希望的植物资源之一。本研究通过收集麻疯树分布点的经纬度数据,基于气候、土壤和地形等37个相关生态因子,采用最大信息熵模型,预测麻疯树的潜在适宜区域,分析影响其生长的主要生态因子特征。结果显示,麻疯树生长最适宜区域主要分布在我国华南地区的广东、海南、香港、台湾和西南地区的广西、云南、四川;对麻疯树分布贡献率较大的主要生态因子为:最暖季度降水量(53.5%)、温度季节性变化标准差(15.8%)、降水量变异系数(9.3%)、年均温变化范围(5.8%)、最湿季度降水量(3.6%)、最干月降水量(3.2%);Maxent模型预测的AUC值大于0.9,表明对麻疯树潜在分布的预测结果较准确。本文对麻疯树潜在分布区域以及影响其分布的主要生态条件的研究结果,可为麻疯树的种植栽培提供科学依据。

     

    Abstract: As a traditional energy species, Jatropha curcas L. is a promising biodiesel plant resource. In this study, we collected the latitude and longitude data of J. curcas distribution, and used the Maxent entropy model to predict potentially suitable areas for J. curcas based on 37 ecological factors, including climate, soil, and topography, and analyzed which of these factors most affected the growth of J. curcas. Results showed that the most suitable growth areas for J. curcas were in Guangdong, Hainan, Hong Kong, Taiwan, Guangxi, Yunnan, and Sichuan. The main environmental factors affecting the distribution of J. curcas included precipitation of warmest quarter (contribution rate of 53.5%), Standard deviation of temperature seasonality (15.8%), coefficient of variation of precipitation seasonality (9.3%), annual temperature range (5.8%), precipitation of wettest quarter (3.6%), and precipitation of driest month (3.2%). The areas under the ROC curves were all above 0.9, indicating that the predictive results with the Maxent model were highly precise. These results reveal the potential distribution areas and bioclimatic conditions for J. curcas habitat and growth, which can provide a scientific basis for its planting.

     

/

返回文章
返回