【作者】 国红；

【导师】 陆元昌；

【作者基本信息】 中国林业科学研究院 ， 森林经理， 2010， 博士

【摘要】 结构-功能模型,是指能明确表达由生理过程和环境因子调控的植物三维结构生长和变化的一类模型。林木结构-功能模型整合林木形态结构和生理生态功能特征,并考虑环境因子和树木生长发育的交互作用,输出树木生长和发育的三维动态信息,从而取得传统的经验模型、过程模型和形态结构模型难以达到的效果,可以作为林木生长预测、可视化和经营决策的重要支持工具。GreenLab模型作为通用性植物结构-功能模型已广泛应用于农作物,如玉米、番茄、黄瓜、向日葵等,在林业中上有一些应用,如油松、樟子松和山毛榉等,但是对于不同年龄阶段树木尤其是成熟树木的研究还较少,缺少功能与结构之间反馈机制的检验和林分层次的应用研究。因此,本研究的目的是应用GreenLab模型,构建油松的结构-功能模型,分析油松的生长和生理过程,检验GreenLab模型的假设,并将其扩展到树木群体层次,模拟采伐对生长和结构的影响。为GreenLab模型在森林生长和经营中的应用、揭示林木生长规律和生物学驱动的林木三维可视化提供方法。研究地点位于北京市昌平区十三陵林场苗圃、园艺奇苗圃和北京市西山试验林场,数据为1-5年生、10年生、13年生、18年生和41年生不同年龄阶段油松植株的生物量和形态结构、拓扑结构测量数据。模型参数求解方面,GreenLab模型的直接参数通过实验数据获得;隐含参数通过最小二乘法反求获得。利用实测数据和经验模型对模型进行了校准和验证。通过拟合生物量需求满足率与树木器官个数的反馈机制,标定了GreenLab反馈机制模型。通过建立生存面积竞争指数与模型隐含参数的关系,将GreenLab结构-功能模型的应用从单木层次扩展到林分层次。研究的主要内容和结论如下:1)异速生长模型是结构-功能模型的重要组成部分,是分析生物量分配和树木器官几何属性的工具。本文通过幂指数方程建立了10年生和13年生油松一级枝、二级枝、三级枝的当年生小枝的针叶单叶长和单叶生物量、节间生物量和节间长、节间生物量和针叶生物量、节间截面积和节间生物量及节间截面积和针叶生物量等的异速生长关系,并通过独立数据进行检验。结果表明观测值和预估值之间的相关系数除了节间截面积和针叶生物量外均大于0.85,说明上述异速生长关系的相关关系非常显著。小枝生物量和总针叶生物量之间存在显著的正相关关系,其权度关系明显小于1,表明较大的枝具有较低的叶生物量分配比例和较高的枝生物量分配比率。小枝的针叶总生物量和小枝的截面积呈等速生长关系,符合管道模型理论。2)根据植物学原理,运用植物构筑型中生理年龄的概念,对树木的拓扑结构进行了纯数字编码,解决了结构-功能模型应用中由于树木组件众多而使树木测量和数据处理、编程分析变得复杂和困难的系列问题。3)建立了油松幼树和大树的结构-功能模型。在构建油松大树结构-功能模型时,采用分层抽样的方法,将树枝级别和轮作为层,减少了油松形态和生物量测量的工作量;利用子结构方法大大减少了模型拟合所需要的基本单元个数,缩短了模型的计算时间。运用λ值将Pressler模型和全局分配模式结合起来,解决了年轮生长中分配模式的选择问题,发现油松的年轮生物量分配模式受年龄的影响较大。另外,本文参考发表的经验模型对GreenLab模型进行了验证,结果表明GreenLab模拟油松的总生量和经验模型的结果相关系数达0.98,总节间生物量相关系数达0.95,总叶生物量的相关系数为0.75,说明GreenLab模型模拟树木总生物量好于叶生物量。4)从研究林木个体之间的光竞争入手,分析光竞争对树木个体、器官的形态和生物量特征及模型隐含参数的影响。运用异速生长模型建立了隐含参数植株投影面积Sp、环境因子E与生存面积竞争指数Sd之间的关系,通过采伐引起的竞争指数的变化,模拟了采伐对于树木结构和生长的影响。结果表明:GreenLab结构-功能模型可以合理的表达油松的结构与功能之间的互反馈关系,从而输出具有生物学机理的树木的三维结构信息;在林分层次描述光竞争和模拟采伐效应方面有着巨大的潜力,在森林生长模拟和经营决策中有广阔的应用前景。下一步要加大样本量,扩展结构环境因子反馈机制关系的研究,深化树木生长规律的生物学解释,加强模型验证,并和林分层次的森林经营措施相结合,为森林经营提供决策支持工具。更多还原

【Abstract】 Functional-structural plant models are models explicitly describing the development over time of the 3D architecture of plants as governed by physiological processes which, in turn, are driven by environmental factors. Functional-structural tree models (FSTMs) which combine trees’morphological and functional parts depict the true 3D presentation of trees for analyzing tree growth and interactions between trees and the environment. FSTMs can be as an important means of forest management and decision-making with more outputs expected than traditional empirical models, process-based models and morphological models. As a general plant functional-structural model, GreenLab model has been applied to crop simulation such as maize, tomato, cucumber and sunflower. There are some applications in forestry including Fagus Sylvatica, Pinus sylvestris and Salicaceae. But it is seldom applied to trees with different age class especially to adult trees. There it is never applied to trees at stand or tree population level with feedback mechanism. The goal of this paper was: 1) to construct functional-structural model of Chinese pine based on GreenLab model to analyze its growth process and development with different age class; 2) to test biological hypothesis behind GreenLab models; and 3) to extend the application of GreenLab model to stands or tree population level with light environmental effects. It is expected to provide the method for applying GreenLab model to forest growth and management and implementing biology-driven tree visualization.Experimental sites were located in nursery of Shisanling forest farm, Yuanyiqi nursery and Xishan forest farm in Beijing. Detailed data including tree geometry, tree topology and biomass measurements were collected for 1-5 years old, 10 years old, 13 years old, 18 years old and 41 years old trees. With experiment data, direct parameters of the model were calculated including organ sink parameters and allometric parameters and target files were created. Hidden parameters were acquired by the nonlinear least square method inversely. At the same time, GL3 model which is called feedback model were calibrated by fitting demand satisfaction rate. By the relationship between tree Voronoi area which is called area potentially available index(APA) and hidden parameters, we extended GreenLab model to stand or tree population level and simulated the variation of trees biomass and topological structure with different competition index APA after cutting. Main contents and results are listed as follows:1) Organ allometry is one of important components in functional-structural models. It is essential to analyze the biomass allocation and organ geometry attributes. In this thesis allometric rules between internode biomass and internode length, between internode biomass and needle biomass, between internode sectional area and internode biomass, and between internode sectional area and needle biomass of current year twigs on 1 level branches, 2 level branches and 3 level branches of 4 trees with 10 years old and 13 years old were analyzed by power exponent equation, Independent data were used to test the allometric models. The results showed that the correlation coefficient between predicted and observed values is more than 0.85 except the case between sectional area and needle biomass, and the correlation is statistically significant. The size scale between internode biomass and needle biomass is less than 1 which means that biomass allocation rate is higher for internode biomass than for needle biomass. The size scale between sectional area and needle biomass is close to 1 which means that it is symmetry between these two attributes and accord with pipe model.2)Based on botany rules and the concept of physiological age of architecture, we developed topological code by numerical way. It overcomed the difficulties in dealing with hundreds and thousands components of trees. It is convenient to be used in measurement, data process and programming.3) Functional-structural model of young and adult Chinese pine trees are constructed based on GreenLab methodology. Stratified random sampling was used to measure adult Chinese pine trees. Stratifies were whorl and branching orders. Substructure algorithm was used to fill target file to overcome the difficulties in dealing with thousands and millions of units and saved computation time. The thesis introducedλto combine the Pressler pattern and common pool patter for analyzing the ring growth. It is effective and flexible to apply to trees with different ages and different environment factors. By referring to published empirical models, we tested the accumulated biomass of whole trees, internodes and needles. The results showed that the correlation coefficient is 0.98 for total biomass, 0.95 for total internode biomass and 0.75 for total needle biomass. More samples applied to GreenLab model will be helpful to improve the estimation accuracy.4) The thesis introduced Voronoi area as density of tree group. By comparing the attributes two patterns with different Voronoi areas, the effects on organ biomass and organ dimension of light competition was analyzed. Hidden parameters were also compared after fitting models. By allometric models, the relationship between Voronoi area and hidden parameters which indicate the light competition were established. According to this relationship, we simulated the tree topology and biomass after cutting.The study proved that functional-structural GreenLab model could reasonably describe the feedback between tree structure and functions of Chinese pine, and produced 3D architecture information with biological mechanism. The model has potentials in explaining light competition and simulating cutting effects at stand level, and expected to be broadly applied in forest growth simulation and management decision-making in future. The next research will include linking FSTMs to mechanism of botany and ecology, extending the feedback between tree structure and environmental factors, strengthening model validation, integrating the model with management practice, and optimization of parameter fitting. We hope the model could provide more supports for forest management and decision making.更多还原