基于线性分位数混合效应的辽东山区红松冠幅模型

冠幅是反映单木生长状态及构建林木生长收获模型的重要变量。本研究以辽东山区大边沟林场10～55年生红松人工林为对象,基于66块固定样地的2763株红松的每木检尺数据,选取冠幅基础模型,采用再参数化的方法引入单木竞争指标(R_d),利用哑变量的方法引入了林分密度、林层变量,构建不同分位点(0.50、0.90、0.93、0.95、0.96、0.99)的冠幅分位数回归模型,并与传统方法进行比较,选取模拟林分最大冠幅的最优分位点。为反映林分中单木冠幅在林木个体之间的差异,建立了基于样地水平的最优分位点的线性混合效应分位数回归冠幅模型,分析各变量对单木冠幅的影响。结果表明: 基于F统计检验,不同林分密度和林层的冠幅模型具有显著差异,在基础模型中引入林层、林分密度和竞争后,模型R_a²提高0.0104,均方根误差降低0.0115,均方误差降低为7.4%;与最小二乘法比较,分位数回归模型能够较好地模拟林分状态下的单木最大冠幅,并选出0.96分位点和0.93分位点作为上林层和下林层的分位数回归模型的最优分位点。引入混合效应的线性分位数回归模型的赤池信息准则、贝叶斯信息准则、HQ信息准则等评价指标优于传统分位数回归,参数标准误显著降低,混合效应的引入很好地解释了样地之间的差异。就上林层和下林层而言,林分密度越大,最大冠幅越小;相对直径越大,最大冠幅越大,其中林分密度对下林层的冠幅影响大于上林层,当林分密度足够大时,冠幅随着胸径的增大先增大后降低。本研究构建的基于混合效应的分位数回归模型能有效提高模型的拟合优度,今后可通过调控林分密度、适度抚育间伐等措施,实现对辽东山区红松人工林的科学营建和可持续发展。     

森林动态模型概论

讨论了模拟森林林分动态变化的模型,并把模型分为森林生长模型和演替模型。森林生长模型包括：全林分模型、林分级模型和单木模型;演替模型包括马尔可夫类模型和林窗模型。文中给出了演替模型的基本原理和适用性,在比较早期和最新发展的林窗模型后,叙述了林窗模型的新进展。生长和演替模型的结构和数据要求不同决定了它们的在时间和空间上的适应性,最后指出模型将向综合总体方向发展。     

森林动态模型概论

讨论了模拟森林林分动态变化的模型,并把模型分为森林生长模型和演替模型。森林生长模型包括：全林分模型、林分级模型和单木模型;演替模型包括马尔可夫类模型和林窗模型。文中给出了演替模型的基本原理和适用性,在比较早期和最新发展的林窗模型后,叙述了林窗模型的新进展。生长和演替模型的结构和数据要求不同决定了它们的在时间和空间上的适应性,最后指出模型将向综合总体方向发展     

蒙古栎红松林演替模型FOROAK的研究

以森林动态斑块理论为基础建立了蒙古栎（ＱｕｅｒｃｕｓｍｏｎｇｏｌｉｃａＦｉｓｃｈ．）红松（ＰｉｎｕｓｋｏｒａｉｅｎｓｉｓＳｉｅｂ．ｅｔＺｕｃｃ．）林动态变化的林窗模型ＦＯＲＯＡＫ,该模型包括对树木生长的生物学过程和影响树木生长的环境因子模拟两部分。用两种不同面积大小的样地分析森林动态,确定蒙古栎红松林林窗面积为０．０５ｈｍ２。以实际调查数据检验模型,证明所得模型能合理预测森林变化过程,在预测树种组成上精度很高,实际调查和预测的树种断面积组成比在６０、１００和２７０年非常接近,在森林发育后期观测样地的树种组成与预测结果吻合程度良好。对裸地上森林模拟,表明森林的动态过程规律是,蒙古栎和白桦（ＢｅｔｕｌａｐｌａｔｙｐｈｙｌａＳｕｋａｃｚ．）在林分发育开始占优势,中期形成阔叶树和针叶树混交,但以阔叶树为主的森林,后期渐被红松取代。对现实原始林预测显示,森林未来３００年变化稳定,红松株数和生物量变化很小。     

东灵山地区落叶阔叶林长期动态的模拟

用林窗模型研究了暖温带辽东栎林的长期动态变化。用森林经营历史和实际调查数据求得模型的参数。经过与实际数据检验证明所得模型能合理地预测辽东栎林的物种组成动态变化。通过对辽东栎交生裸地的模拟,可以得到：辽东栎种群在森林动态变化过程中呈现波动的形式,波动的周期为１１０ａ左右,叶面积指数的动态变化过程与林分的竞争状况有密切的关系,生产力的变化有着明显的无规律性,变化极水稳定,在３０ａ进有一个变化的高峰期,     

暖温带落叶阔叶林动态变化的模拟研究

用森林动态林窗模型 FORET1模拟了暖温带落叶阔叶林的长期变化特征。模型参数取自暖温带地区长期森林研究和经营的历史数据 ,对过去数据中缺少的参数进行了实地测定 ,并用观测的数据对模型作了检验。结果表明模型能较好地模拟暖温带落叶阔叶林的长期动态变化特征。通过模拟可以看出 ,森林的净初级生产力没有明显变化规律且极度不稳定 ,峰值出现在30 a左右 ,相似于世界上其它地区森林动态格局变化 ,生物量格局呈循环状态变化 ,循环周期大致在 110 a左右。     

玉龙雪山自然保护区丽江云杉林林窗特征研究

研究了玉龙雪山自然保护区云杉坪典型丽江云杉林林窗干扰的基本特征．结果表明,滇西北亚高山丽江云杉林林窗占林分面积比例冠层林窗和扩展林窗分别为28．8％和42．5％,干扰频率以林窗密度计算,林窗出现的平均密度为35个·hm^-2左右．林窗大小分布为面积大于100m^2的大林窗约占25％,面积为50～100m^2的中等林窗占41％,面积小于50m。的小林窗占34％．形成林窗最重要的方式是干基和干中折断,其次是枯立滇西北亚高山丽江云杉林林窗形成木80％以上是丽江云杉,通常直径为40～50cm、高度为15～25m．每个林窗形成木数量多为1～2株,约占68．8％,5株以上很少,最多为6株．随着林窗面积由大变小,林窗中更新苗木的密度逐渐变大,小林窗中更新木密度约为大林窗的5倍．     

鼎湖山南亚热带常绿阔叶林林窗分布格局及其成因

林窗数量特征及其空间分布格局对南亚热带森林生态系统的动态变化、物种共存及生物多样性的维持等具有重要意义。本文基于鼎湖山南亚热带常绿阔叶林20 ha动态监测样地2015年的植被调查数据,结合无人机航拍图像处理技术和地理信息系统,分析了样地内林窗的几何特征和空间分布格局。结果表明:该样地的林窗空隙率为13.72%,密度为35.75个/ha,平均面积38.37 m~2。具体特征有:(1)区域内林窗数量随林窗面积的增加呈负指数分布,整体表现为小林窗多、大林窗少的规律。(2)不同成熟度林分中,过熟林林窗平均面积大于成熟林;成熟林更能体现出小林窗多而大林窗少的特点。(3)各生境林窗分布与大样地整体表现出基本一致的规律,但低谷与其他生境差异显著,林窗平均面积、林窗空隙率等都大于其他生境,而山脊林窗也在林窗空隙率与林窗密度方面低于其他生境。(4)林窗面积和地形因子显著相关:与海拔和凹凸度呈显著负相关;与坡度和坡向呈显著正相关。据此提出建立利用无人机进行森林群落林冠变化与格局的监测体系,是实现林窗与林下群落动态变化同步监测的新手段。     

最大密度法则研究进展

 该文从理论推导和研究方法等方面对近几十年来关于最大密度法则的研究进展进行了综述,得出结论：1)关于最大密度法则理论主要有几何关系的3/2法则和空间填充分行支状网运输结构的WBE模型。进一步研究发现它们都是建立在一种静态的统计分析基础之上的,因而近几年研究者们开始尝试用动态的个体植物之间的竞争来建立模型。尽管如此,关于最大密度法则的模型仍然没有逃出固有的模式,如用平均植物大小代替整个植物种群。因此,关于最大密度法则理论需要进一步的研究。2)最大密度法则理论在假设条件、数学推导、 用于估计参数的原始数据选择等方面存在争议。任何模型的建立都是基于一些特定的条件和假设建立的,因而得到的关系并不是一个万能的定律。所以在分析数据时,这些模型可结合使用。3)在研究方法上,由于大家对最大密度法则的理解不同,标准不同,造成研究方法多种多样。因而建议在以后的研究中建立一个客观统一的方法。     

亚热带常绿阔叶林林窗物种丰富度的影响因素

林窗是森林群落物种多样性维持中十分重要的结构,但其本身物种多样性的维持机制尚不清楚,可能与相邻群落的物种丰富度、林窗面积以及土壤和地形因子等有关。本文选取浙江天童20 ha森林动态监测样地内的84个冠林窗作为研究对象,分析了林窗内部以及不同距离邻域内的物种组成,选取包含大部分林窗内物种(≥75%),并且对林窗内物种丰富度影响最大的邻域作为相关邻域(林窗周围成熟植株通过种子传播影响林窗物种丰富度,把距离冠林窗边缘5 m范围内的区域作为种源效应邻域;由于距离林窗较近,物种组成与林窗形成时仍存活个体的物种组成相似,故把距离冠林窗边缘4 m范围内的区域作为距离效应邻域),然后计算种源效应邻域内成熟植株的物种丰富度、距离效应邻域内所有植株的物种丰富度、冠林窗的面积、地形因子(海拔、坡度、坡向、凸度)和土壤因子(p H值、全氮、全碳、全磷),最后对11个潜在的影响因素进行随机组合,构建广义线性回归模型,筛选最优模型(AICc最小)进行分析。结果显示:种源效应邻域内成熟植株的物种丰富度和冠林窗面积显著影响林窗内物种丰富度,其他因素并未对冠林窗内物种丰富度产生显著影响。上述结果说明,冠林窗面积的大小以及外部种源的丰富程度是决定冠林窗内物种丰富度的主要因素,其他因素可能只产生间接影响。     

长白山阔叶红松林径级结构动态模拟和优化经营

以原始阔叶红松林为研究对象,采用密度依赖矩阵模型,模拟了自然生长预案下林分径级结构的动态变化,分析了一种择伐预案对林分径级结构的影响,计算了7种不同择伐强度下森林的恢复期。结果表明:原始阔叶红松林比较稳定,但也有比较缓慢的自然生长,林分株数密度呈下降趋势,符合森林的自然稀疏规律,随着时间的推移,各径阶株数的变化速度逐渐减弱,趋于稳定,验证了演替顶极理论。以生长量、收获量、保留林分结构和采伐费用为森林经营效果的评价指标,则20%的采伐强度、35年采伐周期和25%的采伐强度、45年的采伐周期的2种方案较优。     

Simulation of the Structural Development of Warm-Temperate Rain Forest Stands

The change in all-sized dbh distribution with stand age wassimulated using data from secondary stands of different agesafter clear-felling in warm-temperate rain forests in YakushimaIsland, southern Japan. The cumulative basal area [cm² m^–2]of all trees larger than a given dbh x [cm] was used as an indexof onesided density effect, which primarily regulates the growthrate and mortality of the tree at x. The relative growth ratein dbh was expressed by a negative linear function of both thecumulative basal area and logarithmic dbh, irrespective of standage. Mortality was found to be positively related to the cumulativebasal area. From these empirical relations, the change in dbhdistribution during the course not only of stand developmentafter clear-felling, but also of gap regeneration within primaryforests, was successfully simulated using the continuity equation,eliminating ‘stand age factors’ from the model.Results of the simulation satisfied the —3/2 power lawof self-thinning. Competition, continuity equation, cumulative basal area, simulation, size distribution, stand dynamics, —3/2 power law, warm-temperate rain forest, Yakushima Island     

林窗模型的几个基本问题的研究 Ⅰ.模拟样地面积的效应

以中国东北长白山阔叶红松林为例,应用林窗模型NEWCOP探索了不同模拟样地面积对林窗模型输出结果的影响,结果表明,模拟样地面积大小变化可影响模拟出的森林群落的树种组成和模拟样地的林窗出现周期,通过应用这一特点确定了阔叶红松林的林窗面积为400－800m^2。     

林窗模型的几个基本问题的研究:I.模拟样地面积的效应

以中国东北长白山阔叶红松林为例,应用林窗模型NEWCOP探索了不同模拟样地面积对林窗模型输出结果的影响.结果表明,模拟样地面积大小变化可影响模拟出的森林群落的树种组成和模拟样地的林窗出现周期,通过应用这一特点确定了阔叶红松林的林窗面积为400～800m².     

林窗模型BKPF模拟红松针阔叶混交林群落对气候变化的潜在反应

用森林动态模型ＢＫ荭松针阔叶混交林群落在气候变化和ＣＯ２倍增后的反应得出,在气候变化５０ａ后现存的天然红松林林木总株数减少２０％以上,地上部分生物量将减少９０％以上,蒙古栎的生物量占锯林分的５７％以上,林分生产力与现在气候条件下比较接近,略高４％,但主要以蒙古栎、山杨、白桦为主,林分叶面积指数大幅度减少。     

江西中南部红壤丘陵区主要造林树种碳固定估算

本文根据江西第6次森林清查小班数据,通过基于实地调查数据拟合的森林植被生物量与蓄积量的关系,估算了2003年江西中南部红壤丘陵区泰和县和兴国县主要人工造林树种马尾松、湿地松、杉木林的生物量和碳储量,并采用空间替代时间的方法,利用Logistic方程拟合了三个树种林龄与碳密度的曲线关系,估算了研究区1985-2002年的森林植被生物量和碳储量,分析了时空动态特征。结果表明：(1) 2003年研究区主要造林树种林分面积31.04?10⁴hm²,总生物量22.20Tg,总碳储量13.07TgC,平均碳密度42.36tC/hm²。(2) 1985、1994、2003年三个树种植被碳储量分别为4.91、11.41和13.07TgC,年均固碳量0.45 TgC.a^-1。(3) 海拔位于700-900m之间的树种平均碳密度最大,坡度位于25～35?之间的树种平均碳密度最大。森林植被碳密度总体上呈现随海拔高度的增加而增加,随坡度的增大而增大的分布。人工造林工程使江西中南部红壤丘陵区森林植被碳储量明显增加,合理的森林经营管理可以提高森林生态系统的固碳能力。     

A computer model of succession and fire response of the high-altitude Eucalyptus forest of the Brindabella Range,Australian Capital Territory

The BRIND model, a computer model of the high altitude forests in the Brindabella Range near Canberra (Australian Capital Territory), is documented and the results of a series of tests on the model are provided. The BRIND model simulates a 1/12 ha forest stand by computing the growth of each individual tree in the stand. It considers establishment and death of trees on a tree-by-tree basis using stochastic functions. The model also simulates the effects of prescribed fire and wildfire on the forests. The model presently is restricted to southeasterly facing slopes (moist, sheltered situations) above 850 m in altitude. The BRIND model is tested in four ways: (1) A single example simulation (for 500years) is inspected for agreement with stand dynamics in wet sclerophyll forests. (2) By varying wildfire frequency, the model is used to develop a succession diagram for forests in the alpine ash (E. delegatensis) zone of the Brindabella Range. This diagram is considered in terms of the successional patterns described for this ecological zone. (3) By subjecting the model to different climatic conditions and wildfire frequencies, a simulated altitude zonation is developed. This simulated pattern of forest types is compared with the extant forest types in the Brindabella Range. (4) The model is tested on its ability to duplicate basal area, stocking density, and average diameters for different age stands found in an independent data set. The model was found to simulate patterns of vegetation that resemble those of the forests of the Brindabella Range in both space and time. The successional pattern was found to be complex and to differ from classic theories of succession originating with Clements. Potential model applications are discussed     

A role-type model (rope) and its application in assessing climate change impacts on forest landscapes

Gap-phase replacement is a general phenomenon found in forest ecosystems, worldwide. Different tree species can be expected to produce different sizes of gaps when they die. Species also vary in their regeneration success in gaps of different sizes. In this paper, the gap-phase interactions among tree species in a forest stand are simulated by a role-type stand model called ROPE. By incorporation of environmental effects on tree height, ROPE can simulate forest composition and stand leaf area under different climate conditions. The model was developed for forest ecosystems in northeastern China and was used to simulate the forest landscape structures under current climate conditions and under four climate change scenarios for greenhouse gas related warming. These scenarios were obtained from general circulation models developed by different atmospheric research centers. Korean pinebroadleaf mixed forest and larch forest are the major stand types in the study area under present conditions. Under the four climate change scenarios, Korean pine-broadleaf mixed forest would be expected to occur only on the higher parts of large mountains. Larch forest only would be found north of the study area. Broadleaf forest would become the dominant vegetation over the study area. Use of the Kappa statistic to test for similarity in spatial maps, indicates that each climate change scenario would result in a significant change of forest distributions.Supported by The United States National Science Foundation Grant BSR-8702333 to University of Virginia.     

EVALUATING THE IMPACT OF INSECT COMMUNITY ON PINE CATERPILLAR DENSITY IN DIFFERENT STAND CONDITIONS

Abstract  Natural enemies play an important role in inhibiting population growth of pine caterpillar, but they are difficult to be used in systematic management models. This paper attempts to use some forest stand factors to substitute the effect of natural enemies, which could be used in management models. Through the studies from more than 200 stand plots with different stand conditions in Qu County, Changshan County, and Longyou County of Zhejiang Province, 16 predominant species groups from insect community were selected, and the canonical correlation coefficient between the diversity index, species number, individual number of predominant species and the stand factors was analyzed. The results indicated that the canopy density and ground vegetation were the key stand factors for characterizing the function of insect community. The forest area was divided into four types according to the variation of canopy and vegetation conditions by using cluster analysis. The regression models between the rate of variation of pine caterpillar density and the forest factors including canopy density and vegetation covers were developed in each type of forests. There were significant differences in diversity index, species number, individual number of insect community and population density of pine caterpillar among four types of forests, and correlation coefficients of the four regression models were rather high ( r =0. 924 — 0. 964). This means that the classification of the forest stands is reasonable and the stand factors can characterize the effect of natural enemies in practice.     

Scale and resolution of forest structural pattern

An individual tree-based forest succession model was modified to simulate a forest stand as a grid of contiguous 0.01-ha cells. We simulated a 9 ha stand for 750 years and sampled the stand at 50 yr intervals, outputting structural variables for each grid cell. Principal components analysis was used to depict temporal patterns in forest structure as observed in 0.01 ha samples (individual grid cells). We then resampled the grid using square aggregates of 4 to 100 grid cells as quadrats. Principal component scores recalculated for the aggregates, using the original (0.01 ha scale) scoring matrix, depict the effects of obervational scale on perceived patterns in forest structure. Larger quadrats reduce the apparent variation in forest structure and decrease the apparent rate of structural dynamics. Results support a scale-dependent conceptualization of forest systems by illustrating the qualitative difference in forest dynamics as viewed at the scale of individual gap elements as compared to the larger scale steady state mosaic. The aggregation exercise emphasizes the relationship between these two observational scales and serves as a general framework for understanding scaling relationships in ecological phenomena.