利用最大熵模型预测药用植物海南蒟的潜在地理布局

Predicting Potential Geographical Distributions of Medicinal Plant Piper hainanense Using Maximum Entropy

利用获得的24个种群位点和27个环境图层,选择最大熵模型(Maxent)在地理和环境空间上模拟了我国药用植物海南蒟(Piper hainanense)的潜在地理布局。结果表明,Maxent模型的预测准确率高达87.5%,p值检验显示具有统计显著性。海南蒟的潜在区主要集中在海南、广东南部和广西部分地区,与已知的实际分布密切相关,此外在越南北部和老挝中部也有潜在分布的可能。目前已有部分分布区位于政府设定的自然保护区内,但仍有超过80%的分布区未被有效保护,常遭受人为干扰破坏。根据留一法检验和环境因子分析,部分与温度和降水有关的环境因子决定了海南蒟的潜在地理布局。

The potential geographical distributions of Piper hainanense were predicted with Maxent software based on twenty four locations and twenty seven environmental layers. The results showed that the Maxent model had 87.5% success rate and was statistically significant. The potential distributions of P. hainanense with the highest suitability were mainly located in the Hainan Island, southern Guangdong and Guangxi. Northern Vietnam and Central Laos where the species had never been documented were also predicted as the suitable habitat. Only small area of the species＂ potential distribution had been under protection. According to the jackknife test, some precipitation and temperature parameters were identified as highly effective factors on the distribution of the species and could possibly account for its narrow distribution scale. This approach presented here appeared to be quite promising in predicting suitable habitat for threatened and endangered species with small sample records and could be an effective tool for biodiversity conservation planning, monitoring and management.