欢迎访问林业科学,今天是

林业科学 ›› 2018, Vol. 54 ›› Issue (8): 153-164.doi: 10.11707/j.1001-7488.20180817

• 问题讨论 • 上一篇    下一篇

基于优化的Maxent模型预测白栎在中国的潜在分布区

李璇, 李垚, 方炎明   

  1. 南京林业大学南方现代林业协同创新中心 南京林业大学生物与环境学院 南京 210037
  • 收稿日期:2017-04-26 修回日期:2018-06-24 出版日期:2018-08-25 发布日期:2018-08-18
  • 基金资助:
    国家自然科学基金项目(31770699,31370666,31571649);江苏省林业三新工程项目(LYSX;2016]49);河南重大科技专项(161100110400)。

Prediction of Potential Suitable Distribution Areas of Quercus fabri in China Based on an Optimized Maxent model

Li Xuan, Li Yao, Fang Yanming   

  1. Co-Innovation Center for the Sustainable Forestry in Southern China, Nanjing Forestry University College of Biology and the Environment, Nanjing Forestry University Nanjing 210037
  • Received:2017-04-26 Revised:2018-06-24 Online:2018-08-25 Published:2018-08-18

摘要: [目的]采用优化Maxent模型对白栎的适生区进行预测,了解气候因子对白栎分布的影响,同时结合植物耐寒性区域地图(PHZM)探讨白栎的栽培区划和引种区划,为白栎的引种栽培提供理论基础。[方法]采用Maxent模型,利用AICc指标对特征参数(feature)和正规化参数(β)进行筛选,建立最优模型。基于484条分布记录和10个环境变量模拟白栎在末次盛冰期、全新世中期、现代和2070年的潜在分布区。综合jackknife检验、置换重要值和百分比贡献率、限制环境因子,探讨影响白栎适生分布区的环境因子。[结果]1)最优模型的参数设置为:feature为LQP和β乘数为1.5。2)jackknife检验表明:年均温、最干月降雨量、平均日温差、温度年较差为关键因子;百分比贡献率排前3名的环境变量依次为平均日温差、温度年较差、年均温;置换重要值排前3名的为年均温、温度年较差和等温性。影响现代最适分布区的环境限制因子为年降雨量和最干月降雨量;影响未来最适分布区的环境限制因子为极端最高温和最干月降雨量。3)白栎的现代高度适生区集中分布在重庆、贵州局部地区、湖南、湖北南部、江西、安徽南部、福建北部和长江三角洲地区;末次盛冰期时白栎的高度适生区在湖南和江西零星地区,较现代分布区面积减少28.28%;全新世中期高度适生区范围与现代相似,较现代高度适生区面积增加6.44%,面积达到最大;在进行未来适生区的预测时,原本不具有适生区的辽宁出现少部分低度适生区,2070年温度可能升高,适宜分布区向北扩张,高度适生区面积减少6.44%。[结论]年均温、平均日温差、温度年较差和最干月降水量是制约白栎分布格局的重要环境因子。影响白栎的现代最适宜分布区和未来的最适分布区的环境限制因子为年降雨量、最干月降雨量和极端最高温。末次盛冰期白栎的高度适生区集中在华中地区,随气候的转暖逐渐向北移动。全新世中期时,高度适生区面积扩张达到最大。未来气温升高,白栎适生区可能发生向北扩张的趋势,高海拔地区分布的白栎更容易受到气候的影响。根据Maxent预测的白栎分布区结合中国耐寒性区域地图进行白栎的栽培区划和引种区划,新疆、北京、天津可能适合白栎的引种栽培。

关键词: 白栎, Maxent, AICc, 末次盛冰期, 适生区

Abstract: [Objective] Quercus fabri is a deciduous tree species of the genus Quercus in China, with high economic and ecological value. At present, most studies focus on the exploitation and utilization of Q. fabri resources, however, there is little research on the geographical distribution pattern of Q. fabri. In this study, we used the optimized Maxent model to predict the suitable areas of Q. fabri and understand the effect of climatic factors on the distribution of Q. fabri. As the same time, we discuss the cultivation zoning and introduction zoning of Q. fabri in combination with the PHZM, and provide some theoretical basis for the introduction and cultivation of Q. fabri.[Method] We used AICc index to screen feature and beta multipliers,and then establish the optimal model. Based on the 484 distribution records and 10 environmental variables, we simulatedthe potential distribution areas of Q. fabri during the Last Glacial Maximum, Mid-Holocene, Present and the year 2070 and investigated the environmental factors affecting the distribution of suitable habitat of Q. fabri by means of percent contribution, permutation importance, and Jackknife test, Limiting Factors.[Result] 1) The parameters of the optimal model were set as follows:feature is LQP and the β multiplier is 1.5. 2) Jackknife test showed that Annual mean temperature, Precipitation of the driest month, Temperature annual range were the key factors; the top three environmental variables of percentoge contribution rate were Mean diurnal range, Temperature annual range, and Annual mean temperature; the top three of permutation importance were Annual mean temperature, Temperature annual range, and isothermaling.Mean diurnal range the environmental limiting factors affecting the potential distribution in the present were Annual precipitation, and precipitation of the driest month; Precipitation of the driest month, and max temperature of the warmest monthwere the environmental limiting factors of potential distribution of future. 3) The highly suitable region for the present distribution covered Chongqing, part of Guizhou, Hunan, southern Hubei, Jiangxi, southern Anhui, northern Fujian, and the Yangtze River Delta region. In the Last Glacial Maximum, the highlysuitable areas of Q. fabri were scattered in Hunan and Jiangxi, which is less than the present distribution areas by 28.28%. In the middle reached Holocene, the highly suitableregion was similar to the present one, but it'sareas increased by 6.44% and the are to the maximum of the four periods. In the prediction of future suitable distribution areas, Liaoning did not have a suitable area in the present but there is a small part of the low suitable areas in the future. Temperature may rise in 2070, the suitable distribution area would expand to the north and the highly of the suitable areas of Q. fabri decreased by 6.44%.[Conclusion] Annual mean temperature, mean diurnal range, temperature annual range, and precipitation of the driest monthare important environmental factors that restrict the distribution pattern of Q. fabri. The environmental limiting factors affecting the distribution in the present and future prediction are annual precipitation, precipitation of the driest month and max temperature of the warmest month. The highly suitable areas of Q. fabri in the last glacial maximum are concentrated in central China, and then gradually moved to the north with the warming of the climate. The whole suitable areas reachto the maximum in the Mid-Holocene. The future temperature would rise, and the suitable areas of Q. fabri may be the trend of northward expansion, and Q.fabri in high altitude in is more susceptible to climate. According to the distribution pattern of Q. fabri and the PHZM in China, Xinjiang, Beijing and Tianjin may be suitable for the introduction and cultivation of Q. fabri.

Key words: Quercus fabri, Maxent, AICc, last glacial maximum, suitable distribution area

中图分类号: