【作者】 孙志强；

【导师】 梁军；

【作者基本信息】 中国林业科学研究院 ， 森林保护学， 2011， 博士

【摘要】 植食性昆虫与植物多样性的关系一直是森林生态系统及其功能研究的核心问题,其种群发生机制是森林保护学家最感兴趣的话题、也是有害生物可持续控制的重要理论基础。昆嵛山腮扁叶蜂寄主在昆嵛山不同林分类型间的比例呈现从0-100%的组成梯度,这种通过无重大干扰条件下所形成的森林生态系统为本研究提供了人工条件下无法达到的研究环境。本研究利用昆嵛山森林生态系统定位研究站设立的40块永久性标准地,以昆嵛山腮扁叶蜂-赤松体系为模板,开展了两部分研究内容:(1)昆嵛山森林生态系统的群落特征及其植物多样性特点;(2)综合分析昆嵛山林分立地因子、林分特征因子对昆嵛山腮扁叶蜂种群密度效应。本研究产出的森林自调控食叶害虫的机理,可以进一步完善和补充森林自调控病虫灾害和生态调控技术的理论,对推动我国森林保护学和生态学科发展具有一定的理论意义。主要研究结果如下:(1)采用主坐标分析法将样地所代表的森林群落划分为6种林分类型,即赤松林、黑松林、日本落叶松/刺杉林、针叶树-麻栎林和针叶树-杂木林和阔叶林。六种林分类型的物种多度和物种丰富度无显著差异;阔叶林的Fisherα指数(P < 0.001)、Shannon–Wiener指数(P = 0.001),Simpson指数(P = 0.034)与其他五种林分类型相比,差异到显著性水平;其他五种林分类型的Fisherα指数和Shannon–Wiener指数无显著差异,但Simpson指数在这些林分间差异显著。物种指示值分析表明每一种林分类型有其独具代表性的指示物种。Rényi多样性曲线直观显示了其他方法无法辨别群落特征间的一些差别,如排序显示针叶树-杂木林和阔叶林林下灌草层多样性位于其他五种林分类型的之上;而另外五种林分的灌草层多样性曲线相互交错。低海拔(<200m)、下坡位、缓坡(坡度0-10o)的灌草多样性明显偏低,土壤厚度>5 cm和土壤厚度在1-3cm的灌草多样性曲线处于其他3个土壤厚度梯度的下方。昆嵛山森林自然演替13a,赤松种群密度从1996年的平均超过13000株/hm2降到了2008年的平均2377株/hm2。林分径级结构呈现典型的倒“J”形,表明这些林分已进入茎干互斥阶段。赤松纯林内灌草Shannon–Wiener指数(H)和均匀度指数(JS)分别为2.50和0.79,分别低于1996年的(H)2.69和(JS)0.85。黑松林内灌草Shannon–Wiener指数(H)和均匀度指数(JS)分别为2.27和0.79,分别高于1996年的(H)1.85和(JS)0.77。赤松和黑松林下的灌草种类的整体物种周转率T达到125.9%。昆嵛山当前的这种森林结构、林分类型的形成及其生物多样性与上世纪大规模树种引进和人工造林等人为干扰密不可分;当前的森林演替处于进展演替的初期,森林结构相对稳定。(2)2006年-2009年昆嵛山腮扁叶蜂种群动态具有显著的短期特征,不同年份间虫口密度差异极显著;其中,2006年虫口密度处于顶峰,到2009年整体种群密度已经下降到较低的水平。2006年昆嵛山立地特征中的海拔、坡度和坡向所形成的小气候促进了腮扁叶蜂集中分布在这些地方寄主资源集中的林分中,虫口密度与寄主在林间的比例极显著的正相关关系,符合“资源集中”假说。2008年昆嵛山腮扁叶蜂种群受到典型的立地所形成的小气候的直接影响。2009年虫口密度随着林龄的增加而增加,表明平均树龄越大的林分腮扁叶蜂虫口密度越高。说明在虫口密度较低时,腮扁叶蜂选择寄主的趋向性随之发生变化。以上结果说明,在景观尺度上,食叶害虫种群的发生驱动机制远比在林分尺度上要复杂。(3)采用通径分析方法,将全部通径关系的402条通径链的模型进行缩减,共去除不显著的通径链381条;缩减后的模型中共有21条通径链,占总通径链的5.22%,但这些剩余通径链能解释原模型信息的33.29%。结果表明,对昆嵛山腮扁叶蜂种群暴发有直接作用的因子包括海拔、坡度、坡位、寄主在林间的比例和Shannon–Wiener指数;并且所有这些直接作用因子的作用方式均与腮扁叶蜂虫口密度成极显著的正相关关系。所有的林分指标首先受到来自立地条件的直接作用,随后这些受到影响的林分指标又影响到其他指标,如物种丰富度和物种多度以及寄主的比例,最后这些指标的综合效应才最终影响到虫口密度;虽然Shannon–Wiener指数对虫口密度有直接作用,但Shannon–Wiener指数的变化受到物种丰富度和物种多度的直接作用,而这两个指标的变化又受到来自立地条件和林分特征的深刻影响。(4)昆嵛山森林群落的不同林分类型所遭受的腮扁叶蜂为害没有显著差异。不同林分类型中树木群体更多地表现为对昆嵛山腮扁叶蜂的“联合易感性”;只有针叶树-麻栎林在2008年虫口密度较低时,表现出较弱的“联合抗性”,但2009年这种“联合抗性”作用在虫口密度很低时未能表现出来,证明了系统发生远的树种组合能表现对食叶害虫的抑制和调控的能力;但“联合抗性”和“联合易感性”的强度、稳定性随着时间、环境条件的改变以及林内虫口密度的改变而变化。本文首次证实,在昆嵛山那些处于低海拔、阴坡的林地内腮扁叶蜂种群持续处于低密度很可能归因于由立地调控的景观水平的“联合抗性”的作用结果。(5)植物多样性对昆嵛山腮扁叶蜂的发生起到重要作用,随着植物多样性增加,整体上腮扁叶蜂的虫口密度随着增加,但因寄主种类不同而有不同的表现形式。这与大多数研究所指出的植物多样性丰富抑制寡食性食叶害虫的结论相反。(6)当昆嵛山腮扁叶蜂连续3年为害时,显著降低了赤松枝条的生长量和枝条生长率。对比2006年到2008年赤松胸径生长量,不同虫口密度危害下,生长量没有显著差异。腮扁叶蜂发生轻微的林分内的树木平均死亡率(5.5%)显著低于发生严重的林分(13.35%),树木死亡率与腮扁叶蜂虫口密度呈显著的正相关、并与林分郁闭度显著相关,但与林分密度不相关。腮扁叶蜂严重为害的林分中大树的死亡比例高于轻度发生的林分。与发生较轻的林分相比,腮扁叶蜂为害导致树木高死亡率很可能加速这些林分密度的调整,同时也是林分径级倒“J”形结构形成的促进因素之一;但另一方面,腮扁叶蜂为害在短期内未显著降低树木生长,验证了枝条“自决定理论”所认为的叶蜂对寄主整株的碳循环动态不能产生显著的影响;分析结果部分地证明和符合许多作者所证明的结论,如食叶害虫具有改变林分结构的功能和调控植物种群数量和群落动态的功能。本研究结果表明昆嵛山腮扁叶蜂的发生机制受非线性过程的制约,昆嵛山立地和林分特征对昆嵛山腮扁叶蜂近4年的时间和空间动态格局的影响具有连续的效应。因此,可以划定重点地区的林分类型,缩小昆嵛山腮扁叶蜂种群监测范围,这样可以更有效的对昆嵛山腮扁叶蜂的发生做出准确的预测。更多还原

【Abstract】 Relationships between herbivorous insect and plant diversity are a central topic in biodiversity-functioning research in forest ecosystem, and mechanism under this relationship is the most important theoretical foundation for sustainable management of forest pests. The near-natural conditions, in which the host of Kunyushan web-spinning sawfly (Cephalcia kunyushanica) ranges a proportion from 0% to 100% in Kunyushan region naturally formatted away from anthropogenic interference, have the advantages of providing a herbivorous insect-host system with a multitude of interactions that might not have developed in the same way in artificial experiments.This study was conducted in the 40 permanent plots set by Kunyushan Ecosystem Station and to use Kunyushan web-spinning sawfly-host system as a model. Fisrt, forest community and plant diversity of Kunyushan ecosystem was analyzed. Based on this, the interactions between site conditions, plant community factors and larval density of C. kunyushanica were analyzed to explain factors and structures of forest community mediating C. kunyushanica population fluctuations. Furthermore, the functions of herbivory by C. kunyushanica on growth and mortality of Pinus densiflora was detected. The study could further improve and supply knowledge of forest system mediating pests and ecologically based pest management technologies. It also has important therotical significance for promoting development of forest protection and forest ecology. The main results are as follows:1. Principal coordinate analysis revealed that plant communities at Kunyushan Forest Reserve could be grouped into six stand types. Species indicator values (IndVals) analysis suggested the each stand type had its own unique indicators. No statistically significant difference was found among the six stand types in their abundance and species richness, while the diversity indices of Fisherα, Shannon–Wiener and Simpson were significantly different among these stand types. The highest Fisherαindex and Shannon–Wiener index were found in broadleaved stand types, whose Simpson index was the lowest. Fisherαand Shannon–Wiener indexes showed no statistically significant differences among the other stand types, i.e. P. thunbergii stands, P. densiflora stands, Larix kaempfer/Cunninghamia lanceolata stands, conifer-Quercus acutissima mixed stands and conifer-broadleaved mixture. Rényi diversity profiles were more intuitive to display some detailed differences of species diversity in communities than that which can not be distinguished by other analysis approaches. ANOVA analysis suggested that shrub and herb diversity indices under various elevations, slope positions, slope gradients and soil depths were not significantly different, except that Shannon-Winner index (F=4.665, P=0.002)and Simpson index(F=2.465, P=0.044) were significantly different in various stand types. However, Rényi diversity profiles indicated that, by having consistently higher profiles than other stand types, broadleaved stands and conifer-broadleaved mixed stands appeared to have higher species diversity, while the other five stand types could not be ordered in diversity with one another as they had corresponding diversity profiles that intersected. On the other hand, site conditions had profound effect on diversity. Plots at lower elevation, lower slope position and gradient, soil thickness > 5cm and ranging 1-3 cm, had lower species diversity. The mean tree number density decreased from 13000 trees/hm2 in 1996 to 2377 trees/hm2 in 2008, which indicated that stands in this region have entered the stem exclusion stage. Shannon diversity index (H) and evenness index (JS) of shrub and herbage plant communities under P. densiflora pure forest was 2.50 and 0.79 in 2008, lower than (H) 2.69 and (JS) 0.85 in 1996, respectively. In P. thunbergii mixed forest, Shannon diversity index (H) and evenness index (JS) of shrub and herbage plant communities was 2.27 and 0.79 in 2008, higher than (H) 1.85 and (JS) 0.77 in 1996, respectively. Species turnover rate (T) was 125.9%, which implied the change of species over 13 years in Kunyushan forest communities. The formation of diverse forest stand types mainly resulted from anthropogenic disturbance about 30 years ago, when new species were introduced and many trees were planted. 2. Larval density of C. kunyushanica differed significantly among year 2006, 2008 and 2009, which it peaked in 2006 and maintained at the lowest level in 2009. In 2006, larval density was significantly positive related to elevation, slope, slope aspect, the host proportions, and Shannon-winner index. The significant positive relation between larval density and host proportion corresponded with“resource concentration hypothesis”. Microclimate triggered by site conditions directly affected the occurrence of C. kunyushanica in 2008. However, larval density in 2009 positively related to stand age, which suggested that the trend of host selection changed along with larval density. Thses results revealed that mechanism for herbivorous insect population outbreak is far more complex in landscape scale than in stand level.3. Path analysis permitted insight into the causal mechanism by making a variety of predictions about the strength of direct and indirect interactions causing web-spinning sawfly occurrences in Kunyushan region. 21 significant paths were extracted from total 402 paths, which 381 paths were not significant. The residual paths with 5.22% in proportion of total paths explained 33.29% information of original model. The results indicated that direct interaction to affect web-spinning sawfly outbreak in 2006 included elevation, slope, slope position, host proportion, and Shannon-Winner index. All these factors had a significant positive effect on sawfly population outbreak. All stand features were directly affected by site conditions, and these stands features further affect other factors and all these combining factors together affected sawfly population. Shannon-Winner index had a direct effect on sawfly population, while it was affected by both site conditions and stand features.4. The study indicated that no difference of herbivory to host trees was found among different forest types. Almost all forest types performed strong“associational susceptibility”to Kunyushan web-spinning sawfly, except that weak“associational resistance”was observed in conifer-Quercus acutissima mixed stands in 2008, while this resistance only existed other than in 2008. The result confirmed that taxonomically dissimilar plants may be more likely to interact in ways that lead to“associational resistance”, while the strength, consistency, and relative impact of“associational susceptibility”or“associational resistance”on plant fitness also are likely to vary temporally and spatially as environmental conditions and herbivore and plant abundance change. Further, this was probably the first study to prove that site conditions might be contributors for landscape level“associational resistance”which led to continious low larval density in sites at low elevation and shady slope.5. Plant diversity played a main role for Kunyushan web-spinning sawfly occurrence. Sawfly larval density increased positively with increasing of plant diversity index values, which the diversity performance differed with the change of forest types. This conclusion is in contrast to many studies reporting a decrease in herbivory or herbivorous insect abundance with increasing plant diversity both in forests and in other systems.6. Current year shoot length under different feeding modes was significantly different (F1,7=2.422,p=0.02) and so did the growth rate (F1,7=2.439,p=0.019). Growth of current year shoot was not affected by current year feeding while was affected by last year’s feeding. Current year shoot suffered significant length reduction when it had previously experienced high intensities of two successive herbivore attacks last year and the year before. However, increment of DBH in various stands under different sawfly severities showed no significant difference. This indicated that infestations of sawfly may have obvious impact on shoot growth while it might have little affection on persistence of whole pine stands. We found a significant higher mortality rate (average 13.35%) in severely infested stands whose sawfly larval density was > 25 larvae/tree than that in slightly infested stands (p=0.003). Meanwhile, a positive correlation was found between tree’s mortality and sawfly larval density and canopy cover. Growth of diameter at breast height (DBH) of severely infested stands was slightly higher than that of slightly infested stands while there was no significant difference between them. Sawfly disturbance was most likely the promoter for reverse“J”shape of diameter distribution of stands. This indicated that higher tree mortality caused by sawfly attack could accelerate adjustment of stand density in severely infested stands compared to the same course by population self-thinning in slightly infested stands, however, the attack of sawfly did not inhibited tree growth promisingly. C. kunyushanica population dynamics is governed by nonlinear processes, while the impact of site and stand features on its population has a continious effect. The results suggested that future monitoring of sawfly population could be limited in key sites and forest types, which would be effective for accurate prediction of future sawfly occurrence. Keywords: Kunyushan Forest Ecosystem Station, forest ecosystem, Cephalcia kunyushanica,更多还原