• 中国中文核心期刊
  • 中国科学引文数据库(CSCD)核心库来源期刊
  • 中国科技论文统计源期刊(CJCR)
  • 第二届国家期刊奖提名奖
Volume 36 Issue 2
Apr.  2023
Article Contents
Turn off MathJax

Citation:

Effects of Topography and Forest Vegetation Factors on Occurrence and Population Size of Overwintering Dendrolimus superans

  • Corresponding author: ZHANG Ji-li, xtafktj@126.com
  • Received Date: 2022-04-18
    Accepted Date: 2022-06-24
  • Objective The objective is to explore the influences of topography and forest vegetation on the occurrence probability of overwintering Dendrolimus superans and clarify the relationship between these factors and the occurrence quantity of Dendrolimus superans, which will provide scientific reference for the prevention and control of the infestations of this pest. Method Factors such as topography, forest vegetation, and the number of larvae of Dendrolimus superans were investigated in 142 plots (20 m × 20 m) of different forest types established in the main distribution area of Dendrolimus superans in the Great Xing’an Mountains. The binary logistic regression and the theory of quantificationⅠ were used to establish the models of the occurrence probability and quantity of Dendrolimus superans, respectively, and the performance of these models was evaluated. Result The discriminant accuracy of the probability model was relatively high (AUC=0.826), the slope aspect and herb cover were the key factors affecting the occurrence probability of Dendrolimus superans, the occurrence probability of which was the lowest on shady slopes and was the highest on sunny slopes. The occurrence probability increased with the increase of herb cover. There were significant differences in the number of Dendrolimus superans in the stands with different slope aspects, levels of slope gradients, slope positions, altitudes, herb coverage, and stand density (P<0.05). The model of quantification established by the five factors of slope aspect, levels of slope gradient, slope position, canopy cover, and stand density had little difference in fitting degree from the model established by the full factors. The coefficient of determination (R2) of them was 0.635 and 0.685, respectively, and their root mean square error (RMSE) was 0.964 and 0.936, respectively. The model of quantification showed that the occurrence quantity of Dendrolimus superans was the least on the shady slope and the most on the sunny slope. As the levels of slope gradient increased, its occurrence quantity would decrease. The occurrence quantity of Dendrolimus superans was the most on the ridges. Similarly, with the increase of canopy cover, or the decrease of stand density, the occurrence quantity of Dendrolimus superans would increase accordingly. Conclusion The study revealed that Dendrolimus superans were prone to occur on sunny slopes, flat slopes, and ridges, which should be the focus of prevention and control, which forest management practices such as replanting, thinning and regular weeding could help restrain the population growth.
  • 加载中
  • [1]

    SÁNCHEZ-MARTÍNEZ G, WAGNER M R. Bark beetle community structure under four ponderosa pine forest stand conditions in northern Arizona[J]. Forest Ecology and Management, 2002, 170(1): 145-160.
    [2]

    GRAY D R. The relationship between climate and outbreak characteristics of the spruce budworm in eastern Canada[J]. Climatic Change, 2008, 89(3-4): 447-449. doi: 10.1007/s10584-007-9317-5
    [3] 于 跃, 房 磊, 方国飞, 等. 气象因子对落叶松毛虫种群数量的影响[J]. 应用生态学报, 2016, 27(9):2839-2847.

    [4]

    FANG L, YU Y, FANG G, et al. Effects of meteorological factors on the defoliation dynamics of the larch caterpillar (Dendrolimus superans Butler) in the Great Xing'an boreal forests[J]. Journal of Forestry Research, 2021, 32(6): 2683-2697. doi: 10.1007/s11676-020-01277-6
    [5] 胡瑞瑞, 梁 军, 谢 宪, 等. 赤松宜林地昆嵛山腮扁叶蜂虫基指数的评估[J]. 林业科学研究, 2021, 34(1):80-87.

    [6]

    KHARUK V I, DEMIDKO D A, FEDOTOVA E V, et al. Spatial and temporal dynamics of Siberian silk moth large-scale outbreak in dark-needle coniferous tree stands in Altai[J]. Contemporary Problems of Ecology, 2016, 9(6): 711-720.
    [7] 于长义. 大兴安岭落叶松毛虫发生原因初析[J]. 森林病虫通讯, 1991, 10(3):37-38.

    [8] 于 跃, 房 磊, 王凤霞, 等. 落叶松毛虫发生的空间分布及其影响因子[J]. 生态学杂志, 2016, 35(5):1285-1293.

    [9]

    ROOT R B. Organization of a plant-arthropod association in simple and diverse habitats: the fauna of collards (Brassica oleracea)[J]. Ecological Monographs, 1973, 43(1): 95-124. doi: 10.2307/1942161
    [10]

    PRICE P W. The plant vigor hypothesis and herbivore attack[J]. Oikos, 1991, 62(2): 244-251.
    [11] 韩瑞东, 何 忠, 戈 峰. 影响松毛虫种群动态的因素[J]. 昆虫知识, 2004, 41(6):504-511.

    [12]

    OTWAY S J, HECTOR A, LAWTON J H. Resource dilution effects on specialist insect herbivores in a grassland biodiversity experiment[J]. Journal of Animal Ecology, 2005, 74(2): 234-240. doi: 10.1111/j.1365-2656.2005.00913.x
    [13] 佟丽丽, 严善春, 王 琪, 等. 落叶松针叶内缩合单宁含量与树龄及家系的关系[J]. 生态学杂志, 2010, 29(2):221-225.

    [14] 黄咏槐, 梁 军, 马 琳, 等. 森林空间结构对昆嵛山腮扁叶蜂虫口密度的影响[J]. 林业科学, 2018, 54(7):84-90.

    [15]

    FlØ D, RAFOSS T, WENDELL M, et al. The Siberian moth (Dendrolimus sibiricus), a pest risk assessment for Norway[J]. Forest Ecosystems, 2020, 7: 48. doi: 10.1186/s40663-020-00258-9
    [16]

    MICHAEL J, CLAUDE B, DAVID C, et al. Pest categorisation of Dendrolimus sibiricus[J]. EFSA Journal, 2018, 16(6): e05301.
    [17]

    HUANG L, NING Z, ZHANG X. Impacts of caterpillar disturbance on forest net primary production estimation in China[J]. Ecological Indicators, 2010, 10: 1144-1151. doi: 10.1016/j.ecolind.2010.03.015
    [18] 张旭东. 黑龙江大兴安岭林区森林昆虫 [M]. 哈尔滨: 东北林业大学出版, 2001.

    [19] 岳书奎, 王志英, 黄玉清, 等. 落叶松毛虫生物学特性及天敌[J]. 东北林业大学学报, 1996, 24(4):1-7.

    [20] 常晓娜, 高慧璟, 陈法军, 等. 环境湿度和降雨对昆虫的影响[J]. 生态学杂志, 2008, 27(4):619-625.

    [21] 党英侨, 王小艺, 杨忠岐, 等. 中国林业害虫生物防治研究进展[J]. 中国森林病虫, 2022, 41(5):6-13.

    [22] LY/T 3030-2018, 松毛虫监测预报技术规程 [S].

    [23] 唐守正, 郎奎建, 李海奎. 统计和生物数学模型计算 [M]. 科学出版社, 2009.

    [24] 耿存娟, 夏永刚, 孙 凯, 等. 无人机喷撒白僵菌防治越冬代马尾松毛虫效果研究[J]. 湖南林业科技, 2021, 48(3):69-72.

    [25]

    HENTSCHEL R, MÖLLER K, WENNING A, et al. Importance of ecological variables in explaining population dynamics of three important pine pest insects[J]. Frontiers in Plant Science, 2018, 9: 1667. doi: 10.3389/fpls.2018.01667
    [26]

    HAYNES K J, ALLSTADT A J, KLIMETZEK D. Forest defoliator outbreaks under climate change: effects on the frequency and severity of outbreaks of five pine insect pests[J]. Global Change Biology, 2014, 20(6): 2004-2018. doi: 10.1111/gcb.12506
    [27]

    KUMAR S, NEVEN L G, ZHU H, et al. Assessing the global risk of establishment of Cydia pomonella (Lepidoptera: Tortricidae) using CLIMEX and MaxEnt niche models[J]. Journal of Economic Entomology, 2015, 108(4): 1708-1719. doi: 10.1093/jee/tov166
    [28] 吴思俊, 朱天辉, 谯天敏. 基于物种分布模型对未来气候变化下云南松毛虫在四川省适生区的预测[J]. 植物保护学报, 2021, 48(4):882-890.

    [29] 张 真, 李典谟. 马尾松毛虫暴发机制分析[J]. 林业科学, 2008, 44(1):140-150.

    [30]

    KHARUK V I, RANSON K J, FEDOTOVA E V. Spatial pattern of Siberian silkmoth outbreak and taiga mortality[J]. Scandinavian Journal of Forest Research, 2007, 22(6): 531-536. doi: 10.1080/02827580701763656
    [31]

    ROBINET C, KEHLENBECK H, KRITICOS D J, et al. A suite of models to support the quantitative assessment of spread in pest risk analysis[J]. PLoS ONE, 2012, 7(10): e43366. doi: 10.1371/journal.pone.0043366
    [32] 徐海舟, 沈金辉, 杨晓兰, 等. 湘北地区思茅松毛虫发生与林分因子相关性研究[J]. 中国农学通报, 2018, 34(33):135-139.

    [33] 张慧霞, 吴洪渊, 黄建华, 等. 运用马尔科夫链模型预测张家界永定区松毛虫发生面积[J]. 湖南林业科技, 2019, 46(5):48-52. doi: 10.3969/j.issn.1003-5710.2019.05.009

  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Figures(3) / Tables(5)

Article views(2027) PDF downloads(67) Cited by()

Proportional views

Effects of Topography and Forest Vegetation Factors on Occurrence and Population Size of Overwintering Dendrolimus superans

    Corresponding author: ZHANG Ji-li, xtafktj@126.com
  • 1. Harbin Research Institute of Forestry Machinery, National Forestry and Grassland Administration, Harbin 150086, Heilongjiang, China
  • 2. Reseach Center of Cold Temperate Forestry, CAF, Harbin 150086, Heilongjiang, China
  • 3. Forest Pest Control and Quarantine Station of Chaoyuan Forestry Bureau, Yakeshi 022150, the Inner Mongolia Autonomous Region, China

Abstract:  Objective The objective is to explore the influences of topography and forest vegetation on the occurrence probability of overwintering Dendrolimus superans and clarify the relationship between these factors and the occurrence quantity of Dendrolimus superans, which will provide scientific reference for the prevention and control of the infestations of this pest. Method Factors such as topography, forest vegetation, and the number of larvae of Dendrolimus superans were investigated in 142 plots (20 m × 20 m) of different forest types established in the main distribution area of Dendrolimus superans in the Great Xing’an Mountains. The binary logistic regression and the theory of quantificationⅠ were used to establish the models of the occurrence probability and quantity of Dendrolimus superans, respectively, and the performance of these models was evaluated. Result The discriminant accuracy of the probability model was relatively high (AUC=0.826), the slope aspect and herb cover were the key factors affecting the occurrence probability of Dendrolimus superans, the occurrence probability of which was the lowest on shady slopes and was the highest on sunny slopes. The occurrence probability increased with the increase of herb cover. There were significant differences in the number of Dendrolimus superans in the stands with different slope aspects, levels of slope gradients, slope positions, altitudes, herb coverage, and stand density (P<0.05). The model of quantification established by the five factors of slope aspect, levels of slope gradient, slope position, canopy cover, and stand density had little difference in fitting degree from the model established by the full factors. The coefficient of determination (R2) of them was 0.635 and 0.685, respectively, and their root mean square error (RMSE) was 0.964 and 0.936, respectively. The model of quantification showed that the occurrence quantity of Dendrolimus superans was the least on the shady slope and the most on the sunny slope. As the levels of slope gradient increased, its occurrence quantity would decrease. The occurrence quantity of Dendrolimus superans was the most on the ridges. Similarly, with the increase of canopy cover, or the decrease of stand density, the occurrence quantity of Dendrolimus superans would increase accordingly. Conclusion The study revealed that Dendrolimus superans were prone to occur on sunny slopes, flat slopes, and ridges, which should be the focus of prevention and control, which forest management practices such as replanting, thinning and regular weeding could help restrain the population growth.

  • 昆虫是森林生态系统的重要组成部分和潜在内源干扰因素,当其处于动态平衡时有益于系统的整体稳定,而一旦暴发成灾则会导致树木大规模死亡,从而对系统内的物种组成、空间结构、生态功能产生极大的负面作用[1-2]

    在宏观尺度上,昆虫的繁殖和分布受温度和降水等气候因子的直接或间接影响[3];而在中观尺度上,地形是虫害暴发空间同步性主要的驱动因子之一[4]。例如,昆嵛山腮扁叶蜂 (Cephalcia kunyushanica Xiao)的危害随着海拔的升高而加重[5];西伯利亚松毛虫 (Dendrolimus sibiricus)的虫害易在西南坡暴发[6];坡位不同导致土壤养分存在差异,在土壤贫瘠条件下,寄主植物长势较弱,松毛虫 (Dendrolimus spp.)易猖獗[7]。昆虫根据生活习性而选择适合的栖居环境,导致同一群落中虫害的暴发呈现出区域化和碎片化。因此,明确影响虫害发生和数量的关键因素对于科学防治和长期监测尤为重要。

    森林植被因子是影响同一植物群落内昆虫数量的限制因素[8]。资源集中假说[9]认为寡食性昆虫更易停留在寄主集中分布的纯林。除生物多样性影响外,与植物生长过程密切相关的植食性昆虫更倾向在生长势旺盛的寄主上产卵[10]。对松毛虫的研究发现,树冠体积与松毛虫数量呈线性正相关关系[11]。还有研究认为,寡食性昆虫发生概率随寄主植物密度增加而增加,但达到某一密度后,随着密度增加其发生概率增速下降,即资源稀释效应[12]。森林中虫害的发生还与林龄具有很高的相关性,幼龄林内可抗虫的缩合单宁含量显著高于其它林龄组[13]。此外,树木的空间分布对虫口密度也有显著影响[14]。例如,避免树木聚集分布能降低虫口密度,从而提高林分抵御虫害的能力。

    松毛虫属鳞翅目枯叶蛾科,主要以云杉 (Picea spp.)、冷杉 (Abies spp.)、松树 (Pinus spp.)和落叶松 (Larix spp.)等针叶树种为食,是严重威胁北方针叶林的主要虫害,并且分布范围仍在扩展[15-16],在我国东北林区分布的主要为落叶松毛虫 (Dendrolimus superans)[17]。落叶松毛虫每年对我国森林的危害面积约为2.5 × 106 hm2,损失的木材生长量约为3.2 × 106 m3[18]。低龄幼虫期为影响落叶松毛虫数量的关键阶段,其昼夜取食,初孵幼虫主要取食针叶的尖端,此时期的死亡率极高,这与该时期落叶松毛虫对不良环境抗性较弱有关[19]。此外,当长期处于湿度较高的环境时,落叶松毛虫的幼虫体内水分失衡,发育迟缓[20]。目前,针对该虫害的主要防治手段为生物防治,根据其生活习性和栖居环境,从森林经营的角度进行防治还未广泛应用[21],而在种群的定量研究方面仍有待开展。本研究旨在探索不同地形和森林植被因子对落叶松毛虫发生和数量的影响,定量分析松毛虫的发生量与地形和森林植被条件的关系,可为预测落叶松毛虫的种群发展和有效防治提供科学依据。

    • 研究地点位于内蒙古大兴安岭绰源林业局 (120°17′52″~121°40′19″ E,47°35′21″~48°37′50″ N),总面积约为3.13 × 106 hm2,森林覆盖率83.7%,海拔750~1 450 m。该地区属大兴安岭山脉南麓,寒温带大陆性气候,夏季气候凉爽湿润,冬季冰雪覆盖期长,年平均气温-2.1℃,年均降水量461.9 mm,无霜期87 d。该区域常见植物60科214属390种,具有寒温带针叶林生态系统的典型特征,主要建群种为兴安落叶松 (Larix gmelinii (Rupr.) Kuzen)和白桦 (Betula platyphylla Suk.),伴生乔木主要包括蒙古栎 (Quercus mongolica Fisch.ex Ledeb.)、黑桦 (Betula dahurica Pall.)、山杨 (Populus davidiana Dode)等,下层灌木主要包括杜香 (Ledum palustre L.)、越桔 (Vaccinium vitis-idaea L.)等。

    • 2019年在绰源林业局所辖林场 (翠岭、梨子山、青岭、苏格河和育林林场)进行样地的选取,依据绰源林业局森林资源二类调查数据提供的各小班林组信息,筛选林组为近熟林的小班,随机设立142个20 m × 20 m不同林型的临时调查样地。调查每个样地的坡向、坡度、坡位、海拔。坡向分为平地、阴坡 (北坡和东北坡,337.5°~22.5°和22.5°~67.5°) 、半阴坡 (东坡和西北坡,67.5°~112.5°和292.5°~337.5°) 、半阳坡 (西坡和东南坡,112.5°~157.5°和247.5°~292.5°) 和阳坡 (南坡和西南坡,157.5°~247.5°)。坡度分为< 6° (平坡)、6°~15° (缓坡)、16°~25° (斜坡)和 >25° (陡坡)。坡位分为谷地、坡下、坡上和山脊。

    • 2020年4月下旬,在落叶松毛虫幼虫上树为害前,采用五点取样法,在每个样地的4角和中心附近各选取2株兴安落叶松,共10株,清除目标树木胸径以下的侧枝和与其树干相连的灌木枝条,在每株树的胸径处环刮去树干的粗皮,用塑料胶带在此处粘闭合的胶带环,再往胶带上均匀涂抹粘虫胶,形成一个宽度15 cm的闭合隔离环。幼虫开始上树后,每隔3天记录闭合环上的幼虫数量,并清除杀死,直到连续3天未见有新虫上树,观察结束,统计每株的幼虫上树数量。测量每株树的胸径、树高,记录各样地林分密度、郁闭度、草本盖度等林地环境因子。其中,林分密度通过计数样地内胸径≥5 cm的林木株数换算得到。

    • 林分密度指单位面积上的林木株数,可利用下式得到:

      式 (1)中,S为林分密度 /(株·hm−2),N为样地内林木株数,A为样地面积,本研究中A为0.04 hm2

      调查142个样地的松毛虫数量,统计发生落叶松毛虫样地中10株树木的落叶松毛虫幼虫总数,计算平均值作为该样地落叶松毛虫的发生量,依据《松毛虫监测预报技术规程》[22],利用落叶松毛虫的发生量将各个样地的松毛虫危害程度进行分级,分别为极轻度 (0~20 条·株−1)、轻度 (20~40 条·株−1)、中度 (40~70 条·株−1)、重度 (>70 条·株−1)危害。

    • 利用二元Logistic回归模型,将样地中是否存在落叶松毛虫作为因变量,以地形和森林植被因子为自变量,其中坡向、坡度等级、坡位和林分类型为定性变量称为项目,需将定性变量进行划分称为类目(表1),并以指数函数(0, 1)进行定量处理,只要定性变量中有一个类目的影响效果显著(P<0.05),则把该变量代入到模型中。通过判别预测结果,得到混淆矩阵,然后计算敏感性和特异性,做ROC曲线图,并计算AUC作为模型精度的预测指标。二元Logistic回归模型的基本公式为:

      项目
      Item
      代号
      Code
      类目等级 Category hierarchy
      12345
      坡向 Slope aspectx1平地 Flat阴坡 Shady slope半阴坡 Semi-shady slope半阳坡 Semi-sunny slope阳坡 Sunny slope
      坡度 Slope gradientx2<6°6°~15°15°~25°>25°
      坡位 Slope positionx3谷地 Valley下坡 Lower slope上坡 Upper slope山脊 Ridge
      森林类型 Forest typesx4
      注:海拔 /m:x5;树高/m :x6;胸径/cm:x7;郁闭度 :x8;草本盖度:x9;林分密度/ (株·hm−2):x10。Ⅰ:落叶松纯林,Ⅱ:以落叶松为优势种的落叶松-白桦混交林,Ⅲ:以白桦为优势种的白桦-落叶松混交林,Ⅳ:有少量落叶松树种的混交林。以下模型中的自变量均参考表1
        Notes: Elevation /m:x5, Tree height/m:x6, DBH/cm:x7, Canopy cover:x8, Herb coverage:x9, Stand density:x10. I: Pure larch forest, II: Larch-white birch mixed forest with larch as the dominant species, III: White birch-larch mixed forest with white birch as the dominant species, IV: Mixed forest with a few larch species.The independent variables in the following models refer to Table 1

      Table 1.  Each topography and forest vegetation factor of classification criteria

      对式 (2)进行转化得出发生概率的函数与自变量之间的线性回归模型公式为:

      式 (2)和 (3)中,Pi为落叶松毛虫发生的概率;y为因变量,表示该样地中落叶松毛虫是否发生,发生=1,未发生=0;xi为自变量,表示第i种影响因素;βi为自变量的回归系数。

    • 在落叶松毛虫发生的126个样地中,对各定量变量进行分级处理,利用单因素方差分析(one-way ANOVA)探讨地形和森林植被对落叶松毛虫发生量的影响,并利用 LSD 检验进行多重比较。采用对数转化以保证其发生量的正态分布和方差齐性。

      利用数量化理论Ⅰ将样地内落叶松毛虫的发生量作为因变量,以地形和森林植被因子为自变量纳入数量化模型中,其中,定性变量需以指数函数 (0, 1)进行定量处理,例如,将谷地定量为 (1, 0, 0)、上坡定量为 (0, 1, 0)、下坡定量为 (0, 0, 1)、那么山脊则定量为 (0, 0, 0),并带入到线性模型中,因此,每个定性变量的最后一个类目 (山脊)是多余的参数,得分为0,并筛选出显著性影响因子 (P<0.05),筛选方法与二元Logistic回归模型相同。数量化模型公式为:

      式 (4)中:yi为因变量,是第i块样地的落叶松毛虫发生量;$ {\widehat{b}}_{0} $为常数;$ {\delta }_{i\left(i,k\right)} $为类目的反应值;$ {b}_{jk} $j项目k类目的得分值,(j=1,2,…,m;k=1,2,…,rj);xj是定量变量的值。

      方程模型由随机抽取的96个样本数据拟合而成,剩余30个未参加模型建立的数据用于该模型精度的检验和评价。模型精度检验采用决定系数 (R2)、均方根误差 (RMSE)对模型进行评价,以确定模型的可靠性和精度。本研究中二元Logistic回归模型、ROC曲线、单因素方差分析和多重比较在SPSS Statistics 26.0中完成,数量化理论Ⅰ是在自变量中包含定性变量的多元线性回归,通过Forstat 2.1完成[23],随机抽样和绘图采用Excel 2010软件。

    2.   结果与分析
    • 坡向和草本盖度对落叶松毛虫的发生影响显著 (P<0.05)。将表2中常量和回归系数代入公式 (3)得到模型:

      项目
      Item
      类目
      Category
      回归系数
      B
      标准误差
      S.E.
      WaldsPExp (B)
      常量 Constant −4.344 2.184 3.956 0.047 0.013
      坡向 Slope aspect (x1) 1 −1.521 1.561 0.950 0.330 0.218
      2 −3.165 1.208 6.861 0.009 0.042
      3 −2.733 1.303 4.399 0.036 0.065
      4 −2.630 1.223 4.625 0.032 0.072
      5 0
      草本盖度Herb coverage (x9) 0.110 0.029 14.329 <0.001 1.116
        注:类目参照表1
        Note: Category refer to Table 1

      Table 2.  Estimated results of the Dendrolimus superans occurrence model

      模型 (6)表明,落叶松毛虫发生在阴坡的概率最低,发生在阳坡的概率最高,且随着草本盖度的增加,发生概率增加。AUC为0.826,范围为0.717~0.939,说明模型的预测结果较好,精度较高(图1)。

      Figure 1.  ROC curve of the occurrence probability of Dendrolimus superans

    • 在落叶松毛虫发生的126个样地中,落叶松毛虫的发生数量受到地形和森林植被因子的显著影响 (P<0.05)。图2A表明,落叶松毛虫在阳坡的幼虫数量显著最多,在阴坡的幼虫数量显著最少 (P<0.05)。图2B表明,落叶松毛虫数量随坡度等级的增大而显著减少 (P<0.05)。图2C图2D显示,落叶松毛虫在山脊的数量显著多于其他相应坡位,且在海拔<900 m的样地发生数量显著多于海拔>1 100 m的样地 (P<0.05)。图2E表明,草本盖度在60%~70%的样地,虫数显著少于另外3个草本盖度等级的样地。图2F表明,在林分密度大于1200株·hm−2的样地虫数显著少于另外3个林分密度等级的样地。

      Figure 2.  Influences of topography and forest vegetation on the occurrence quantity of Dendrolimus superans

    • 表3中各变量的得分代入到公式 (4)中,得到落叶松毛虫发生数量与地形和森林植被因子的方程,模型为:

      项目
      Item
      类目
      Category
      得分
      Score
      P
      坡向 Slope aspect (x1) 1 −1.132 0.205
      2 −2.558 <0.001
      3 −0.548 0.093
      4 −1.125 <0.001
      5 0
      坡度 Slope gradient (x2) 1 1.699 0.028
      2 1.401 0.025
      3 0.945 0.148
      4 0
      坡位 Slope position (x3) 1 −0.083 0.625
      2 −1.105 <0.001
      3 −0.537 0.083
      4 0
      森林类型 Forest types (x4) 1 −0.863 0.155
      2 0.068 0.895
      3 0.119 0.827
      4 0
      海拔Elevation (x5) −0.001 0.187
      树高Tree height (x6) −0.059 0.458
      胸径 DBH (x7) −0.050 0.331
      郁闭度 Canopy cover (x8) 4.506 0.003
      草本盖度Herb coverage (x9) 0.382 0.812
      林分密度Stand density (x10) −0.001 0.030
      截距 Intercept 4.717 0.033
      注:类目参照表1
        Note: Category refer to Table 1

      Table 3.  Quantitative scores and significance of each topography and forest vegetation factor

      坡向、坡度等级、坡位、郁闭度和林分密度的影响效果显著 (P<0.05),为影响落叶松毛虫发生数量的关键因子 (表3)。

      利用这5项关键因子重新建立其与落叶松毛虫发生数量的方程,模型为:

      落叶松毛虫的发生数量在阴坡最少,阳坡最多。随着坡度等级的增加,其数量随之减少。在山脊的落叶松毛虫发生量最多。森林郁闭度增加,或林分密度降低,落叶松毛虫的发生数量随之增多 (表4)。

      项目
      Item
      类目
      Category
      得分
      Score
      P
      坡向 Slope aspect (x1) 1 −0.656 0.428
      2 −2.266 <0.001
      3 −0.336 0.286
      4 −0.944 0.001
      5 0
      坡度 Slope gradient (x2) 1 1.862 0.010
      2 1.301 0.033
      3 0.948 0.134
      4 0
      坡位 Slope position (x3) 1 −0.794 0.225
      2 −1.179 <0.001
      3 −0.681 0.028
      4 0
      郁闭度 Canopy cover (x8) 2.021 0.008
      林分密度Stand density (x10) −0.001 0.020
      截距 Intercept 2.633 0.008
      注:类目参照表1
      Note: Category refer to Table 1

      Table 4.  Quantitative scores and significance of key factors

    • 模型 (7)和 (8)的R2分别为0.685和0.635,RMSE分别为0.936和0.964,且回归方程极显著 (P<0.001),表明2个模型均较可靠,用5个关键因子可以达到全部地形和森林植被因子的拟合效果 (表5)。

      模型
      Model
      评价指标
      Evaluation index
      R2RMSEP
      模型70.6850.936<0.001
      模型80.6350.964<0.001

      Table 5.  Evaluation index of the occurrence quantity model of Dendrolimus superans

      利用模型 (8)计算得到未参加方程建立的剩余30个样地的落叶松毛虫发生数量预测值,与其实测值作比较,有6个样地的落叶松毛虫预测与实测危害等级不同,且多为危害等级为中度的预测。预测危害等级的准确率为80.0% (图3)。

      Figure 3.  Predicted and measured occurrence quantity of Dendrolimus superans

    3.   讨论
    • 坡向和草本盖度是决定大兴安岭南部落叶松毛虫是否发生的关键因子 (表2),随着从阴坡到阳坡的光照增强和水分减少,落叶松毛虫的发生概率逐渐增加。这主要是因为温度的升高有利于虫卵的发育和幼虫摄食强度提高;而湿度的增加会打破落叶松毛虫体内水分平衡,同时使白僵菌等微生物的繁殖加快,提升其死亡率[20, 24]。对西伯利亚松毛虫 (Dendrolimus sibiricus)和欧洲松毛虫 (Dendrolimus pini)的研究也发现,夏季高温和少雨造成的干旱生境会缩短这两类松毛虫的生命周期,是造成它们大规模暴发的主要诱因[15, 25-26]。由此可以推测,随着气候变暖,本研究区域的落叶松毛虫暴发风险将持续提升。另外,草本盖度对该虫害发生概率的作用途径很可能是通过增加土壤呼吸速率,土壤温度和肥力的增加更有助于落叶松毛虫产卵,并且幼虫在枯枝落叶层生存以林内灌草为掩护,不易被天敌捕食和人工防治。本研究中发生概率模型的AUC值为0.826 (图1),与于跃等基于大兴安岭南部落叶松毛虫建立模型的AUC值相近[8]。然而,相对于Kumar等[27]建立的苹果蠹蛾 (Laspeyresia pomonella Linnaeus)全球分布预测模型 (AUC=0.94)以及吴思俊等[28]对云南松毛虫 (Dendrolimus houi)适生区预测模型 (AUC=0.93)略低,这可能由于本研究选择小班为单位,空间尺度较小、为避免多重共线性采用的变量也较少,因而预测精度会随之降低,如能综合考虑气候、天敌等更多维度的因素,模型精度可能会进一步提高。

      除坡向和草本盖度外,坡度等级、坡位、海拔和林分密度也是影响落叶松毛虫发生数量的关键因子。其中,坡度越大落叶松毛虫的发生数量越少,这与对马尾松毛虫 (Dendrolimus punctatus)的研究发现类似,坡度较大时,林内小气候和微环境复杂,虫害在环境和天敌的作用下不易暴发[29]。从山谷到山脊土壤含水量不断下降,因此落叶松毛虫更适合在干燥的山脊处生存繁衍 (图2C)。落叶松毛虫的发生数量在海拔小于900 m的样地显著多于海拔大于1 100 m的样地 (图2D),这间接反映了落叶松毛虫对温度的适应性趋势[3]。需要注意的是,虫害的数量多不完全等同于危害大。对更高纬度的西伯利亚松毛虫的研究发现,其对森林危害 (主要指虫害导致的树木死亡)最严重的地形条件主要为海拔400 m左右,坡度处于11°~13°的缓坡地带[6, 30]。当林分密度大于1 200株·hm−2时 (图2F),落叶松毛虫发生数量显著减少,而其数量在落叶松纯林和混交林之间无显著差异,此时资源稀释效应更加突显[12],也可能是由于降低的林内风速会阻碍该虫的远距离扩散和定向转移[31]。落叶松是落叶松毛虫的优先寄主,其长势与抗虫能力一致[10],然而,本研究中落叶松胸径和树高并没有显著影响该昆虫的发生数量。这可能受选取样地时落叶松龄组均为近熟林所致,落叶松代谢出对害虫生长起抑制作用的植物单宁含量在不同林龄阶段存在显著差异[13]

      在数量化模型 (公式7)中,筛选出坡向、坡度等级、坡位、郁闭度和林分密度这5个关键因子。其中,落叶松毛虫的发生数量随郁闭度的增加而显著增多,高的郁闭度使林内通风不畅,该虫尤其是幼虫多生存在背风区域,这可能由于成年雌蛾飞行能力受腹中卵粒数量影响,导致背风区域其飞行距离有限而繁殖在此环境下[31]。徐海舟等在湘北地区的研究发现,思茅松毛虫 (Dendrolimus kikuchii)发生量随郁闭度的增加而减少[32],本研究与其不一致的原因可能归因于气候条件的差异性。

      在预测和防治方面,应用马尔科夫链模型对虫害进行预测较为广泛,然而,这种方法只能预报虫情的大致趋势。有关松毛虫的研究表明,该模型精度为76.9%,且划分的虫害发生量等级越多,精度越低,预测条件至少需10年的虫情资料[33]。而本研究运用数量化理论Ⅰ建立的模型,精度较高且减少了数据收集的工作难度。但实测与预测危害等级的样地出现了偏差,尤其是对虫害较高样地的预测。胡瑞瑞等[5]运用数量化模型建立了立地因子与虫基指数的关系,其对预测昆嵛山腮扁叶蜂危害等级的准确率达到87.5%,本研究为80.0%。在模型中加入土壤因子,即考虑土壤质地对落叶松毛虫发生量的影响,有可能提高预测危害等级的准确率。虽然气候变暖在时间和空间上对松毛虫生活史和主要寄主的影响仍存在很大的不确定性[15, 25],但目前的防治重点应放在干旱时期海拔较低的阳坡和半阳坡区域,通过增加乔灌混交度降低林下草本盖度,从而最大程度的降低该类虫害暴发的风险。此外,本文仅着重分析了地形和森林植被对落叶松毛虫发生和数量的影响,而有关于落叶松毛虫幼虫不同生活史阶段对地形和森林植被的响应有待于长期的监测和研究。

    4.   结论
    • 本研究分析了地形和森林植被因子与落叶松毛虫发生概率和数量的关系,证实了立地条件和森林植被特征均对其发生概率和数量具有显著影响,并发现坡向和草本盖度为影响其发生概率的关键因子。根据坡向和草本盖度,做好发生概率的预测可以准确的为落叶松毛虫的防治节省大量资源。不同坡向、坡度等级、坡位、海拔、草本盖度和林分密度之间的落叶松毛虫发生数量存在显著差异。利用坡向、坡度等级、坡位、郁闭度和林分密度可以有效预测该区域的松毛虫发生数量。为防止落叶松毛虫暴发成灾,应对处于阳坡、平坡和山脊的森林区域重点防治,通过合理的补植、疏伐和定期除草等森林经营措施将虫害控制在较低水平。

Reference (33)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return