摘要
连翘(Forsythia suspense (Thunb.) Vahl.)为大宗药材,也是常见观赏植物和水土保持植物。本文基于连翘的119个地理分布点数据和气候、土壤、地形共12个指标,利用BioMod2程序包提供的广义线性模型(GLM)、随机森林模型(RF)、推进式回归树模型(GBM/BRT)及人工神经网络模型(ANN)对连翘的潜在地理分布区进行预测。结果表明:4种模型预测均成功,效果最好的是RF模型,AUC(0.95)、TSS(0.81)、KAPPA(0.68)值均较高,温度对连翘生长的敏感程度高于降水,连翘的潜在地理分布区位于陕西、山西、湖北、河南、辽宁的山地与河流沿岸区域。综合4种模型得出,连翘适宜生境为海拔250~2200 m,年均温12.1~17.3 ℃、年降水量400~1000 mm。研究结果为连翘的人工栽培提供指导。
Forsythia suspense (Thunb.) Vahl. is a kind of bulk medicinal material. It is also a common species for ornamental use and for water and soil conservation. Based on 119 current geographical distribution records of F. suspense and 12 related factors, including climate, soil and topography, the potential geographic distribution of F. suspense was assessed by generalized linear models (GLM), random forest (RF), generalized boosting model or usually called boosted regression trees (GBM/BRT), and artificial neural network (ANN) in the BioMod2 package. The results showed that the four models had good performance, among which the RF model had the best efficiency, with high values of evaluation indexes, including AUC (0.95), TSS (0.81), and KAPPA (0.68). For the growth of F. Suspense, the sensitivity of temperature was higher than that of precipitation. The potential geographic distribution of F. suspense situated in the mountains and along rivers in Shaanxi, Shanxi, Hubei, Henan, Liaoning. The suitable habitats of F. suspense are characterized by an altitude of 250-2200 m, annual mean temperature of 12.1-17.3 ℃ and annual precipitation of 400-1000 mm. This study provides guidance for F. suspense’s planting.
出处
《生态学杂志》
CAS
CSCD
北大核心
2016年第9期2562-2568,共7页
Chinese Journal of Ecology
基金
国家自然科学基金项目(31070293)
国家"十一五"科技支撑计划项目(2006BAI06A13-06)
陕西省科学技术研究发展计划项目(2014K-01-02)资助
关键词
连翘
物种分布模型
BioMod2
潜在地理分布
Forsythia suspense ( Thunb. ) Vahl.
species distribution model
BioMod2
potential geographical distribution.