【作者】 朱彦鹏；

【导师】 梁军；

【作者基本信息】 中国林业科学研究院 ， 森林保护学， 2013， 博士

【摘要】 昆虫作为森林生态系统的有机组成部分，在其他生物及非生物等外界因素干扰下，或者于森林生态系统功能退化或紊乱时，由于相关控制因子失调，其种群密度、结构及空间格局等会发生变化，进而引起森林生态系统控制力丧失，并最终导致虫害暴发和流行。因此，揭示森林病虫害种群发生机理，认识森林病虫灾害形成过程和原因，可为利用天然林生物多样性和生态系统结构实现对病虫灾害自我调控功能，完善森林有害生物生态调控理论，实现森林有害生物可持续控制提供科学依据。本研究基于40块永久标准地中植物和昆虫本底数据及一块2hm2样地中乔木调查结果，以昆嵛山天然赤松林（Pinus densiflora）生态系统中昆嵛山腮扁叶蜂（Cephalciakunyushanica）为研究对象，开展了昆嵛山森林生态系统立地因子、林分结构、树种组成、寄主相邻木、植物和昆虫多样性对昆嵛山腮扁叶蜂种群影响效应等研究；构建结构方程模型，揭示了昆嵛山腮扁叶蜂种群发生机制及其影响因素之间直接和间接复杂生态学关系。主要研究结果如下：（1）立地条件中海拔、坡向和土壤厚度与昆嵛山腮扁叶蜂种群密度均无明显相关性。坡度与昆嵛山腮扁叶蜂虫口密度呈显著正相关关系，随着生境坡度增加，昆嵛山腮扁叶蜂种群密度也增加。在结构方程模型中，以坡度为观察变量的潜变量立地因子对昆嵛山腮扁叶蜂种群无显著直接效应，通过其它因子对昆嵛山腮扁叶蜂产生间接正效应和总正效应。（2）林分密度、林龄、郁闭度和草本层盖度与昆嵛山腮扁叶蜂种群密度之间均无显著相关性。赤松比例和灌层盖度与昆嵛山腮扁叶蜂显著正相关关系，即林分中寄主赤松所占比例越高，灌层盖度越大，昆嵛山腮扁叶蜂种群密度越大。林分结构对昆嵛山腮扁叶蜂种群表现为显著直接正效应。林分结构并通过植物和昆虫多样性对昆嵛山腮扁叶蜂种群产生间接负效应，总效应为正。（3）昆嵛山腮扁叶蜂虫口密度与乔木、灌木物种丰富度及多样性呈显著正相关关系，与草本植物丰富度及多样性无显著相关性。结构方程模型中以乔木、灌木和草本物种丰富度为观察指标的潜变量植物多样性，对昆嵛山腮扁叶蜂种群呈显著正效应。植物多样性受立地因子显著直接正效应，受到森林类型和林分结构直接效应很弱；植物多样性对昆虫多样性直接负效应很弱且不显著。（4）通过对40块标准地内地表层、林间层和林冠层昆虫采集，共收集昆虫标本72528号，603种（或形态种），隶属于15目，133科；昆虫物种累积曲线分析显示，一阶折刀指数(Jackknife1)获得期望物种丰富度为795（±34），实际观测物种占全部物种期望值的75.8%。地表层采集到昆虫多度占总个体数50.2%，共259种；林间层共采集昆虫28468号，物种数占总物种数91.5%；林冠层共采集到88科246种昆虫，多度占总采集昆虫数10.5%。昆嵛山昆虫群落整体物种联结性显著正联结；与昆嵛山腮扁叶蜂呈显著正联结性昆虫有四种，呈显著负联结性昆虫有七种。（5）昆嵛山昆虫群落物种丰富度、多度及多样性指数均与昆嵛山腮扁叶蜂种群密度无显著相关关系。昆嵛山腮扁叶蜂种群密度仅与地表层昆虫丰富度和多样性指数呈显著负相关性，与林间层和林冠层显著性均不明显。捕食性和寄生性天敌昆虫与昆嵛山腮扁叶蜂种群数量均无显著相关性，植食性昆虫物种多度、丰富度与昆嵛山腮扁叶蜂种群数量呈正相关。昆嵛山腮扁叶蜂种群数量受昆虫群落稳定性影响不大。结构方程模型中，昆虫多样性对昆嵛山腮扁叶蜂种群影响为不显著直接正效应。（6）基于寄主赤松最邻近木、最近四株树与关键木赤松形成的―1+4‖空间结构单元、林分三个尺度研究植物间联合作用对昆嵛山腮扁叶蜂种群影响。寄主最邻近木为赤松，对昆嵛山腮扁叶蜂影响较大，而其它树种对昆嵛山种群影响均无明显影响；―1+4‖空间结构单元中，邻近四株树中如果有赤松存在，所构成空间结构单元对关键木赤松昆嵛山腮扁叶蜂虫口密度影响均不大；但如果邻近四株全为非寄主，则会使昆嵛山腮扁叶蜂集中为害关键木赤松，而形成树种间联合易感性。在林分尺度上，赤松与系统发育关系越近（同属）树种混交成林，昆虫种群变异系数较大，易形成联合易感性；而与系统发育越远（同科、目、纲等）树种混交容易形成联合抗性。（7）本研究选择立地因子、森林类型、林分结构、植物和昆虫多样性作为潜变量构建昆嵛山腮扁叶蜂种群发生机制结构方程模型。所构建模型均达到适配标准，适配度很好，模型整体对昆嵛山腮扁叶蜂种群变化量解释均达到65%以上。结构方程模型揭示出影响昆嵛山腮扁叶蜂种群最大直接因素为植物多样性（正效应），其次为林分结构（正效应）和森林类型（负效应）；立地因子对昆嵛山腮扁叶蜂种群总影响作用为正效应。未包括潜变量昆虫多样性的结构方程模型能够很好地预测和解释昆嵛山腮扁叶蜂种群发生机制。更多还原

【Abstract】 Insects are organic components of forest ecosystem. While under biotic and abioticdisturbance or functional degradation and disorder of forest ecosystem, the related controllingfactor would lose the balance and population density, and structure and spatial pattern wouldchange, which makes forest ecosystem lose the control and eventually lead to the outbreak andprevalence of insects. Therefore, unraveling the occurrence mechanism of forest diseases andinsect pests, understanding formation process and reasons of the forest pest disasters willprovide scientific basis for utilizing biodiversity and ecosystem structure of natural forest toself-regulate its diseases and pests disaster, improving forest pest Ecological Pest Managementtheory and achieving sustainable control of harmful forest organism.This thesis identifies Cephalcia kunyushanica in Pinus densiflora ecosystem of KunyuMountains as the research subject, and bases on the statistics on plants and insects of40permanent plots as well as statistics on arbor of a2hm2plot. By studying the impacts of sitefactor, stand structure, tree species composition, neighbor tree of host, plants and insectdiversity on C. kunyushanica population of Kunyu Mountains, and building up structuralequation model, it reveals occurrence mechanism of C. kunyushanica population and complexdirect and indirect ecological relationship between its influencing factors. The results reveal:(1) Elevation, aspect and soil depth has no obvious correlation with C. kunyushanicapopulation. Slope has a significant positive correlation with density of C. kunyushanica(P<0.001), and it increases with a bigger gradient. In structural equation model, taking slope aslatent variable of observed variable, site factors have positive direct effects on plant diversityand have no significant effects on C. kunyushanica population. Site factors have indirect effectsand positive total effects on C. kunyushanica population through the co-effect of other factors.(2) C. kunyushanica population has no significant correlation with stand density, standsage, canopy, and herbaceous coverage and is significantly positive correlated with P. densiflora percentage and shrub coverage, which suggests C. kunyushanica population has a biggerdensity with a higher P. densiflora percentage and shrub coverage. Stand structure, latentvariable of structural equation model, is observed by P. densiflora percentage and shrubcoverage as the observed variable. The result shows stand structure has a relatively significantpositive correlation with both P. densiflora percentage and shrub coverage. Stand structure hasdirect effects on C. kunyushanica population, and has no significant direct effects on plantsdiversity and insect diversity, through the co-effective of which, stand structure has negativeindirect effects on C. kunyushanica population diversity.(3) Larval density of C. kunyushanica is significantly positive correlated with speciesrichness and diversity of trees and shrub, and has no significant correlation with speciesrichness and diversity of herbaceous plants. Plants diversity, as latent variable observed bytrees, shrub and herbaceous richness in the structural equation model, has significant positiveeffects on C. kunyushanica population. Species richness of trees and shrub is significantlypositive correlated with plant diversity, and with less correlated with herb richness. Plantdiversity is significantly positive direct effected by stand factors and less directly affected byforest types and stand structure, and has a weak and insignificant negative direct effects oninsect diversity.(4)72328insect specimen, in603species or morphospecies,133families and15orders,was collected from litter, understory and crown canopy layer of40plots. Species accumulationcurve analysis shows Jackknife1acquires a expecting species density of795(±34), expectingspecies density of actual observed species accounting for75.8%of the whole species.50.2%ofthe total insects were collected from Litter layer, with259species.28468insects werecollected from understory layer, taking up91.5%of the total species.88families and246species of insects were collected from crown canopy layer occupying10.5%of the total insectamount. C. kunyushanica community is significantly net associated with interspecificassociation, and C. kunyushanica population is significantly positive associated with fourspecies and significantly negative associated with seven species. (5) Richness, abundance and diversity index of insects population in Kunyu Mountains isnot significantly correlated with population density of C. kunyushanica, which is onlysignificantly negative correlated with litter layer insects species and diversity index (Pn=0.023;Ps=0.035) and is not correlated with understory and crown canopy layer. The amount of C.kunyushanica population has no significant correlation with richness, abundance andShannon-Weiner diversity index of predates and parasitism natural enemy insects, and hassignificant positive correlation with richness and abundance of phytophagous insects. Twocommunity stability index, richness and abundance ratio and abundance ratio of natural enemyand phytophagous insects, have no significant correlation with the amount of C. kunyushanicapopulation. It suggests the amount of C. kunyushanica population is not quite affected byinsects community stability. In structural equation model, insect diversity has no significantpositive direct effect on C. kunyushanica population.(6) Plant associational effect on C. kunyushanica population was observed on three scales,nearest neighboring tree of host,―1+4‖spatial structural unit of the nearest four tree and thehost and the stand. Nearest neighboring tree of host has a bigger effects if it is P. densiflora,while has on obvious influence as other species. One or more P. densiflora in―1+4‖spatialstructural unit makes no great change in C. kunyushanica population density, while no P.densiflora in―1+4‖spatial structural unit would push C. kunyushanica to focus on the host, P.densiflora, which may easily cause associational susceptibility. P. densiflora in nearestneighboring is positively related with density of C. kunyushanica population. In the stand scale,the more the C. kunyushanica nearest to phylogenetic relatedness, where trees tend to form theforest, the bigger the coefficient of variation is, and it‘s likely it form a associationalsusceptibility. The mixed forests tend to be associational resistance while it is farther fromgrowth system (Family, Order, and Class). In structural equation model forest type, as latentvariable, directly negative associated with C. kunyushanica population and co-effected throughsome other factor. (7) In this thesis, structural equation model was built up based on latent variable standfactor, forest type, and stand structure as well as plants and insects diversity. It has thegoodness of fit and can explain more than65%of the variable of C. kunyushanica population.Structural equation model reveals that plant diversity (positive effect) is the most importantdirect factor on the amount of C. kunyushanica population, with stand structure (positive effect)and forest type (negative effect) ranked the second. Stand factor positively influence C.kunyushanica population on the whole. Structural equation model without insects diversity aslatent variable, AIC (Akaike Information Criteria) and CAIC (Consistent Akaike InformationCriteria), has an average smaller than the model with insects diversity, could excellentlypredict and explain occurrence mechanism of C. kunyushanica.更多还原