【作者】 潘红伟；

【导师】 张星耀；

【作者基本信息】 中国林业科学研究院 ， 森林保护学， 2009， 博士

【摘要】 随着全球一体化和国际社会交往与贸易往来的与日俱增,外来有害生物随人口流动、动植物及其产品跨国扩散日益严重,给入侵地生态系统和国土安全造成巨大的灾害性损失。松材线虫(Bursaphelenchu xylophilus (Steiner & Buhrer) Nickle)已经成为世界上重要的检疫对象和我国有史以来危害最为严重的林业外来有害生物。本研究借助地理信息系统基于两种方法,分析预测了松材线虫在中国的适生范围以及在不同地区的适生程度。这对科学分析松材线虫在中国不同地区流行灾变的可能性、准确评估松材线虫在中国的危害风险具有重要意义。(1) 5种模型所预测的潜在分布区大致范围相同,松材线虫的潜在分布区范围主要分布在北纬18-40度之间,主要分布在我国的南方以及中部的一些省份,包括了实际发生的全部松材线虫分布记录。除ENFA之外四个模型的预测结果在较高纬度的山西、河北、辽宁都出现了部分低适生区,其中DOMAIN的预测结果中辽宁出现了零星的适生区。(2)5种模型预测中最佳适生区均基本形成了两大区域,第一区域主要集中在长江中下游地区,这一地区湿度和温度条件都较好,这一区域的地形主要是长江中下游平原和低山丘陵地以及四川盆地的部分地区,适合松材线虫的存在;第二区域主要集中在广东、广西和福建三省境内,这一地区的湿度和温度条件均较好,且海拔较低。两大区域都是我国南方林区的一部分,在现实情况中为松材线虫病比较严重的疫区,这与实际的发生情况是相符的。但华南、西南疫区在不同模型中的预测结果不同,最佳适生区分布面积也很小;在各模型预测的最佳适生区结果中,广西、河南、四川未见官方报道,应当作为检疫检测的重点,防止松材线虫的入侵。(3)5种模型预测中在江西和福建省交界处均存在低于周围区域的适生级别的适生区,基本上为边缘适生区和低适生区,根据该地区的地形显示为武夷山山脉的边缘地带,海拔较高,这说明5种模型预测均对该区域的地形因素反应敏感。(4)通过ROC曲线比较5种模型的模拟精度,AUC值显示5种模型都要好于随机分布模型,除ENFA与BIOCLIM差异显著外,其余模型间并无显著差异。ENFA变量模型的AUC最大,说明ENFA模型在该环境下的预测效果较其他几个模型好。(5)总的来说,ENFA,MAXENT和GARP模型都能较合理地模拟出松材线虫的潜在分布空间格局,其中,ENFA和MAXENT的结果可以用作核心分布区域的研究,而GARP结果可以作为物种潜在分布范围的边界;相对来说BIOCLIM的预测结果分布面积最大。在本研究的数据环境下,ENFA和MAXENT模型算法在结果的准确性和解释性,模型的应用性方面均优于GARP、DOMAIN、BIOCLIM模型算法。(6)通过对全国739个台站的分析,病原松材线虫最适宜站点有243个;适宜站点63个;次适宜站点72个;不适宜站点116个;极不适宜站点245个;利用ArcGIS9.0得松材线虫在全国适生性点图,再应用地统计学中的普通克里格插值方法,得到全国松材线虫适生性分布,从全国松材线虫适生性分布中看出松材线虫主要集中在我国的华东和华南地区,包括北京南部、天津、河北南部、山西南部、山东、河南、湖北、湖南、江苏、浙江、上海、江西、安徽、陕西东南部、重庆、贵州、云南中南部、广西、广东、福建、海南、新疆中西部的部分地区。(7)通过对全国739个台站的分析,传播媒介松墨天牛最适宜站点有98个;适宜站点234个;次适宜站点258个;不适宜站点11个;极不适宜站点139个;利用ArcGIS9.0得松墨天牛在全国适生性点图,再应用地统计学中的普通克里格插值方法,得到全国松墨天牛适生性分布,从中看出除了西藏、青海、四川西北部、甘肃大部分地区、黑龙江西北部、内蒙古东北和中部的部分地区外,松墨天牛在全国都可以危害。其中,广西、广东、海南、福建、云南南部和江西南部的部分地区为松墨天牛最适宜分布区。(8)通过计算病原松材线虫和媒介松墨天牛的综合适生值得到松材线虫病在不考虑寄主植物的情况下在全国潜在的分布区,包括北京南部、天津、河北南部、山西南部、山东、河南、湖北、湖南、江苏、浙江、上海、江西、安徽、陕西东南部、重庆、贵州、云南中南部、广西、广东、福建、海南、新疆中西部的部分地区。包括了当前松材线虫病在我国的实际发生地。(9)在考虑寄主分布的情况下,松材线虫的潜在分布区主要集中在我国的华东和华南地区,包括北京的南部,天津,河北南部,山西南部,山东,河南,湖北,湖南,江苏,浙江,江西,安徽,陕西南部,四川东南部,重庆,贵州,云南,广西,广东,福建,海南以及新疆的部分地区。在云南省综合病原松材线虫、传播媒介松墨天牛和寄主植物得到松材线虫病在云南省潜在的分布区。云南大部分地区都是松材线虫病的适生区,尤其是南部、东部和中部的部分地区。因此,要加强预防松材线虫病病原的传入,保护大面积云南松和思茅松的安全。(10)生态位模型分析操作较简单,但考虑因素较少,限制条件较多,而且各模型算法比较单一,因此所得预测的结果有一定的局限性;模糊数学、地统计学等方法分析需要大量的参数设定以及数据计算,分析过程繁琐。模糊数学方法的分析的我国松材线虫的潜在分布区面积要大于生态位模型的分析,其在新疆和东北地区出现了松材线虫的潜在分部区。更多还原

【Abstract】 With the development of the global integration and international exchanging and trading, the invasive alien species dispersal causes a large damage to the ecology system of the country invaded. Pine wood nematode (Bursaphelenchus xylophilus), as an agent of pine wilt disease and the important quarantine target of many countries around the world, has been the most serious invasive pest in China. Using tow methods based on GIS, the potential geographic distribution of Bursaphelenchus xylophilus in China is analysed in this paper.It is impotant for the analysis of the posibility of disease epidemic of Bursaphelenchus xylophilus in the different part of China and the assessment of the risk Bursaphelenchus xylophilus in china.(1)The potential distribution of Bursaphelenchu xylophilus is similar in 5 models.The distribution is mainly between 18°-40°N,including some provinces in the middle and southern part of China.Except ENFA,some low suitable distribution appears in Shanxi,Hebei,Liaoning in other four models,and suitable distribution appears in Liaoning of DOMAIN analysis.(2)There are two parts of optimal suitable distribution in 5 models.The first area is the Middle and Lower Yangtze regions with gently rolling topography ,good tempreture and humidity suitable to Bursaphelenchu xylophilus .The other area is mainly in Guangdong, Guangxi and Fujian with low altitude, good tempreture and humidity.These tow parts are really serious pine wood nematode epidemic area. Guangxi,Henan nnd Sichuan should be the quarantine focus to prevent the invading of Bursaphelenchu xylophilus ,because they are optimal suitable distribution and not reported.(3)The suitable rank of the area between Jianxi and Fujian is lower than nearby area and they are marginal and low suitable distribution in 5 models.This area is the marginal zone of the Wuyi mountain with high elevation.The result shows the 5 models analysis are sensitive to the topography.(4)The AUC values of 5models are all better than the random models.Except the significant differences between ENFA and BIOCLIM, there are no significant differences between others.The AUC values of ENFA the biggest in five models.This shows that the analysis result of ENFA is bettee than the other four models.(5)In this paper ENFA,MAXENT and GARP can give us reasonable potential geographic distribution of Bursaphelenchus xylophilus in China.The result of ENFA and MAXENT can be the study on core distribution area and the result of GARP can be the ege of the distribution;the area of BIOCLIM anlysis is relatively big. The accuracy, interpretability and applicability of ENFA和MAXENT are better than GARP、DOMAIN、BIOCLIM.(6)Through the analysis of the 739 Climatic sites, there are 243 suitablest sites, 63 suitabler sites,72suitable sites,116unsuitable sites and 245 unsuitablest sites for Bursaphelenchu xylophilus.The figure of Suitable distribution of Bursphelenchus xylophilus in China shows that the suitable sites mainly exit in southern and easten part of China,including the southern Beijing,Tianjing, the southern Hebei, the southern Shanxi, Shandong, Henan, Hubei, Hunan, Jiangsu, Zhejiang, Shanghai, Jiangxi, Anhui, south east Shanxi, Chongqing, Guizhou, Yunnan, Guangxi, Guangdong, Fujiang, Hainan, Xinjiang.(7)Through the analysis of the 739 Climatic sites, there are 98 suitablest sites, 234 suitabler sites,258suitable sites,11unsuitable sites and 139 unsuitablest sites for Monochamus alternatus.The figure of suitable distribution of Monochamus alternatus in China shows that Monochamus alternatus exit in most of China except Xizang,Qinghai,north west of Sichuan,most of Gansu, north west of Hei Long-jiang,north east of Neimeng.And Guangxi, Guangdong, Hainan, Fujiang,southern Jiangxi, southernYunnan are suitablest distribution of Monochamus alternatus.(8)Through calculating the comprehensive suitable value,not tinking host,we get the the potential distribution of Bursaphelenchu xylophilus in China including the southern Beijing,Tianjing, the southern Hebei, the southern Shanxi, Shandong, Henan, Hubei, Hunan, Jiangsu, Zhejiang, Shanghai, Jiangxi, Anhui, south east Shanxi, Chongqing, Guizhou, Yunnan, Guangxi, Guangdong, Fujiang, Hainan, Xinjiang.(9) Considering the host distribution,the potential distribution of Bursaphelenchu xylophilus include southern Beijing,Tianjing, the southern Hebei, the southern Shanxi, Shandong, Henan, Hubei, Hunan, Jiangsu, Zhejiang, Shanghai, Jiangxi, Anhui, south east Shanxi, Chongqing, Guizhou, Yunnan, Guangxi, Guangdong, Fujiang, Hainan, part of Xinjiang.Most of Yunnan is the potential distribution of Bursaphelenchus xylophilus.(10) The operation of niche modles are simple but the little considering factors and single algorithm make the limitation of the results.The analysis of the fuzzy mathematics need a lot of parameters setting and data calculation which is complex.The area of the potential distribution of Bursaphelenchu xylophilus in the fuzzy mathematics analysis is larger than that in niche modles analysis.更多还原