【作者】 蒋春先；

【导师】 程登发；

【作者基本信息】 中国农业科学院 ， 农业昆虫与害虫防治， 2012， 博士

【摘要】 稻纵卷叶螟和稻飞虱是我国水稻上的重要害虫，近年来给我国水稻生产造成严重的损失。兴安地区位于广西东北部是每年稻纵卷叶螟和稻飞虱在我国南北往返的必经之路，近年来兴安地区水稻栽培制度发生变化，逐渐形成了单双混作模式，对广西兴安地区当前单双季水稻混作制度下“两迁”害虫的发生动态、虫源性质及迁飞轨迹进行系统研究和分析，明确以兴安为代表的南岭以北双季稻区单季稻大面积种植对“两迁”害虫发生规律及迁飞轨迹的影响，对我国北方稻区乃至全国“两迁”害虫监测预警和防治工作意义重大。本研究采用灯光诱捕、田间系统调查、雌蛾卵巢解剖的方法研究了广西兴安地区稻纵卷叶螟和稻飞虱的发生动态和世代虫源性质。运用大气质点轨迹分析平台HYSPLIT和气象图形处理软件GrADS对广西兴安地区稻纵卷叶螟早期迁入虫源和2011年褐飞虱迁飞高峰期进行了迁飞轨迹和气象背景分析。采用传统空间分析方法和地理信息系统对中稻田褐飞虱种群空间分布进行了研究。对2007年毫米波扫描雷达监测到的稻纵卷叶螟的一次迁出过程进行了再分析。采用逐步回归法和bp神经网络模型对广西兴安地区稻纵卷叶螟、白背飞虱和褐飞虱发生期和发生量进行了预测研究。获得了以下主要结果：1．探照灯诱虫器能有效监测稻纵卷叶螟种群动态，其数量变化规律与佳多测报灯灯下虫量变化规律基本一致，监测峰次多于佳多灯。探照灯诱虫器对稻纵卷叶螟的诱蛾量显著高于佳多灯，且对起飞迁出、过境、迁入降落虫群均具有较强诱捕作用。本章提出了以探照灯诱虫器和佳多测报灯诱虫突增情况和雌蛾卵巢解剖情况为标准的探照灯诱虫器诱蛾虫源性质判断方法。2．广西兴安地区稻纵卷叶螟从4月中下旬开始迁入，共发生6代，即第二-七代。其中第三-五代为主害代，发生量最大。在双季早稻、单季中稻和双季晚稻田均能发生3代，且不同类型田间重叠发生。双季早稻田发生量明显高于单季中稻和双季晚稻田。第二、三代主害双季早稻，均为迁入虫源；第四代数量最多，主害双季早稻，属于迁入虫源在本地繁殖世代，也有部分迁入虫源；第五代主害单季中稻，虫源性质复杂，主要为本地虫源本地繁殖型，既有迁入，又有迁出；第六代前期主害单季中稻，后期主害双季晚稻，第七代为害双季晚稻，虫源性质均为本地虫源大部迁出型。单季中稻田的大面积种植，在双季早、晚稻田之间起到重要的“桥梁田”作用，可使稻纵卷叶螟滞留本地为害，为双季晚稻田积累了虫源。3．广西兴安稻纵卷叶螟早期迁入量和迁入峰次在年度间存在差异，最大迁入峰逐年推迟，800hPa高空气温较低以及偏北气流时次较多，是造成稻纵卷叶螟早期迁入量和迁入峰次年度间差异的主要原因。广西兴安5月份迁入虫源主要来自广西东南稻区和广东西南稻区，6月份迁入虫源集中在广西东南部、广东西南部和海南北部稻区。西南和偏南低空激流是稻纵卷叶螟北迁的运载气流，降雨是引起稻纵卷叶螟集中降落的主要原因。4．稻纵卷叶螟在华南地区的迁飞行为参数与华东地区有所不同。稻纵卷叶螟在日落前30min开始起飞，20:00左右达到起飞高峰，起飞可持续1-1.5h；迁飞高度主要在1000m以下，具有聚集成层现象，可形成2-3层，400-500m成层密度最大。成层现象与局部风速极值有关，与风向关系不大，但最大密度并不总是出现在最大风速处。5．广西兴安单双季稻混作的栽培模式，对褐飞虱种群积累增值有很大作用。兴安褐飞虱发生高峰期与中稻生长期吻合，中稻受害重，并且也为晚稻积累了虫源。2011年广西兴安褐飞虱迁入虫源主要来自广西东南部、广东西部及海南沿海稻区，回迁虫源来自湖南西南部和江西西部，10月份从本地迁出的虫群主要迁向广西中部，部分迁往贵州南部的虫群由于缺少食物来源而死亡。偏南方向的低空急流是广西兴安褐飞虱的大规模迁入提供了运载气流，降雨是迫使稻飞虱降落的原因。6．除前期由于田间虫口密度过小外，呈均匀或随机分布外，中稻田褐飞虱成、若虫混合种群、褐飞虱长翅型成虫以及褐飞虱若虫在田间均呈聚集分布，聚集强度是随着种群密度升高而增加。聚集中心分布范围及聚集程度随时间有所不同，田间有明显的1至多个聚集中心，总体呈中部聚集——东西向扩展——中部聚集——东西向扩展的趋势。7．采用逐步回归方法和bp神经网络模拟法对广西兴安地区稻纵卷叶螟和稻飞虱发生期和发生量进行了预报研究。逐步回归法回报准确率为80-87.5%，预报准确率为75-80%；神经网络模拟法回报准确率为100%，预报准确率为75-100%。采用逐步回归法筛选预报因子，再用神经网络模型进行预测，是一种较好的预测方法。更多还原

【Abstract】 The rice leaf roller, Cnaphalocrocis medinalis Guenée, and rice planthopper are the major crop peststhat cause serious damage to rice production. Xing’an, in northeast Guangxi Municipality, is the keyannual migratory route for the rice leaf roller and the rice planthopper migrating back and forthbetween southern and northern China. Recent years, the rice cultivation system changed to single anddouble mixed rice system. A better understanding of the population dynamics, populationcharacteristics and migration trajectory, explicit the effect of the a scale planting of middle seasonrice on the population dynamics and the migration trajectory of the two migration pests will bebeneficial for the monitoring, warning and long-term control of these pest in Xingan and in China. Inthis study, light traps, systematic field surveys, and dissection of female ovaries were used to studythe population population dynamics and population characteristics of these pests. The spatio temporaldistribution of source areas of C. medinalis early migration peaks and source areas and landing areasof BPH miagration peaks in2011were analyzed using HYSPLIT (Hybrid Single-Particle LagrangianIntegrated Trajectory), a software package that simulates migration trajectories, and GrADS (GridAnalysis and Display System), a program for analyzing and displaying the synoptic meteorologybackground during the migration periods. The spatial distribution of BPH in middle season rice wasanalysised by traditional spatial analysis method and Geographic Information System. A emmigrateprogress of C. medinalis monitored by the millmetircs scanning entomological radar in2007wasreanalysised in this paper. Forecast of the occurrence period and occurrence quantity were studied bystepwise regression method and artificial neural network. The results were as follows:1. The vertical pointing searchlight-trap could monitor the population dynamics of C. medinaliseffectively.It could catch much more moths than Jiaduo light trap, including the immigration, passingand emigration airborne populations of C. medinalis. So that the peaks of catches on the two kinds oftraps and the ovarian developmental level of the female in the vertical pointing searchlight-trap wouldbe able to determine the characteristics of C. medinalis. According to this method, three case studiesof immigration, passing and emigration airborne populations of C. medinalis were conducted in thispaper.2. The population dynamics of C. medinalis varied seasonally in Xing’an. C. medinalis had sixgenerations a year, namely the2nd-7thgeneration. The3rd-5thwere the major pestilent generation.Each of the three annual rice crops (early double-crop, middle-season single-crop, and latedouble-crop) experienced three generations of C. medinalis, with some overlap of leaf rollergenerations onto more than one crop. The population density of C. medinalis in the early double-croprice field was significantly higher than in other two rice fields. The ratio of mated females and theratio of females with matured ovaries indicated that the second and third generations which mainlyinfested early rice, were immigrants. The fourth generation, which also mainly infested early rice, emerged locally and partly from immigration population. The fifth generation mainly infestedmiddle-season rice and were mostly locally emergent, but with some immigration and someemigration. The sixth generation was mostly emigrants. The early sixth generation infestedmiddle-season rice, while the late sixth generation infested late rice. The seventh generation, mainlyinfested late rice and were basically emigrants. The middle-season rice which largely planntedbecame the trasitional brige between the early and late rice. It caused C. medinalis settled down andinfested middle-season rice.3. The early immigration amont and immigration peaks were different among years. The biggestimmigration peak postponed with years. The occurance time of partial northern air stream on800hPaisobaric was the main reason caused these differences. Most of the immigrants in May came fromsoutheast of GuangXi and southwest of Guangdong, and the immigrants in June came from southeastof GuangXi, southwest of Guangdong and north of Hainan.Analysis of weather conditions during themigration peaks showed that strong southwest and partial southlow-level jets carried massivenumbers of C. medinalis into Xing’an, while heavy rainfall caused mass descent.4. The flight parameters of C. medinalis in sounthern China were different form that in eastern China.C. medinalis started take-off behind30min of sunset, reached the take-off peak at about20:00,take-off can last1-1.5h. C.medinalis generally flew below1000m It accumulate and form2-3layersdense layers when flew. The biggest density was at400-500m. The insect layer was coincide closelywith partial maximum wind speed and was not conincide with wind direction. However, the biggestdensity was not always at the height of biggest wind speed.5. The single and double mixed rice system in Xing’an was beneficial for the accumulation andincrement of BPH. The BPH whose peak period of occurance was coincidence with themiddle-season rice growth period, not only caused serious damage to middle-season rice, but alsoaccumulated pest source for late rice. The immigrants to Xing’an in2011mainly came from southeastof GuangXi, southwest of Guangdong and coastal area of Hainan. The autumn immigrants camefrome the southwest of Hunan and west of Jiangxi. In October local emigrants mainly flew to themiddle of Guangxi, some population which flew to the south of Guizhou can not survive because oflacking food. Partial southlow-level jets carried massive numbers of BPH into Xing’an, while heavyrainfall caused mass descent.6. The spatial distribution of the adult and larva mixed population, the macropterous adult populationand the larva population were mainly aggregation distribution in middle-season rice, except ofuniform distribution and random distribution in early stage, because of less density of BPH.Aggregation intensity increased as the population density increased. The aggregation center changedwith time. There were1or sevel aggregation center in field, generally showed middleaggregation—west and east extending—middle aggregation—west and east extending.7. Forecast of the occurrence period and occurrence quantity were studied by stepwise regressionmethod and artificial neural network. The historical coincidence of stepwise regression method was80-87.5%, while forecast accuracy was75-80%. The historical coincidence of artificial neural network was100%, while forecast accuracy was75-100%. So, screening forcast factories by stepwiseregression method, then forecast by artificial neural network was a good forecast methed.更多还原