【作者】 程堂仁；

【导师】 马钦彦；

【作者基本信息】 北京林业大学 ， 生态学， 2007， 博士

【摘要】 全球碳循环是全球气候变化研究的核心问题之一,森林生态系统碳循环是全球碳循环的重要组成部分,目前在全球碳循环的研究中还存在很大的不确定性,这种不确定性同样存在于森林生态系统碳循环的研究中,尤其表现在对区域森林生态系统碳库大小及其与大气CO2交换通量的估计上。因此准确估计区域森林生态系统碳库的大小及其相关碳库之间的交换通量一直是森林生态系统碳循环研究中的热点与难点。本文以甘肃小陇山地区锐齿栎,油松,栓皮栎,杨、桦,落叶松,华山松,云、冷杉,其他阔叶混交等8类林分为研究对象,重点在这些林分的生物量估算,林分乔木层生物量相容性模型建立,林分乔木层主要建群树种、灌木层、草本层各器官含碳率和枯落物含碳率测定,区域森林生态系统储碳密度、碳储量估算及现实碳库的空间分布特征等方面进行了研究,旨在探索基于森林资源调查数据来估算区域森林生态系统碳库的合适方法,期待能够为我国区域尺度的森林生态系统碳汇功能研究及中国生态地理区域森林生态系统碳循环研究提供部分基础研究数据,为在全球气候变化的条件下中国区域生态环境建设、制定区域森林生态系统碳汇管理对策以及为中国政府参与世界“碳汇贸易”谈判提供依据和参考。经过几年的大量研究,在以下方面取得一些进展。1.生物量估算(1)8类林分的单木生物量(W)与胸径(D)、与胸径一树高(D2H)之间均存在着紧密的相关关系。利用11种数学模型拟合了小陇山8类林分的单木生物量与胸径、树高之间的相关关系回归方程,结果表明均具有显著水平,其中以Power(幂回归)模型拟合的效果最好。(2)小陇山林区锐齿栎,油松,栓皮栎,杨、桦,落叶松,华山松,云、冷杉,其他阔叶混交等8类林分的乔木层生物量依次为:81.8388,60.6976,79.7980,75.3828,67.3706,67.9786,94.3733,96.5533 t hm-2;灌木层生物量依次为:1.5126,0.9884,1.2504,1.3746,1.3618,1.3691,1.3746,1.5944 t hm-2;草本层生物量依次为:0.1208,0.2266,0.1738,0.2043,0.2658,0.1968,0.2043,0.1393 t hm-2;林分生物量依次为:84.0472,62.4424,81.7747,77.4367,68.9982,70.0695,96.4865,98.7235 t hm-2。2.基于森林资源清查资料的林分乔木层生物量预测模型的建立(1)图形分析、相关分析与回归分析结果表明,甘肃小陇山8类林分各组分生物量(W)与林分活立木蓄积(V)之间存在极显著的线性相关关系。采用相关指数、相关系数、剩余标准差、平均相对误差、预估精度等5个评价指标,综合比较分析了多个回归模型的拟合结果,认为基于森林资源清查统计数据对甘肃小陇山各类林分各组分生物量估计W～V的最优模型应为W=a+bV的形式,并从理论上证明了这种建立在林分水平上的生物量线性估计模型可直接推广应用到区域尺度。(2)从估计效果来看,W～V的相容性线性模型可用于直接利用标准地资料、林分蓄积量等资料来推算林分乔木层生物量,从而避免传统生物量测定方式对森林资源的破坏。因此,这种方法上的改进具有重要的理论意义和较高的实用价值。(3)研究表明,小陇山8类林分的平均净生长量(△W总)与林分总生物量(W总)之间具有较好的相关性,并利用这种相关关系建立了甘肃小陇山8类林分△W总～W总生物量的多种数学模型。最优模型显示,不同林分类型的平均净增长量与林分总生物量之间的相关关系是不同的,锐齿栎林分,油松林分,落叶松林分,云、冷杉林分呈幂函数相关;栓皮栎林分,杨、桦林分,华山松林分呈线性相关;其他阔叶混交林分呈指数相关。与前人的研究比较,结果表明,不同地区的各类林分的平均净增长量与林分总生物量之间的相关关系不尽相同。3.含碳率测定(1)采用高精度的元素分析方法(干烧法)对甘肃小陇山8类林分不同组分(干、枝、叶、皮、根)的含碳率进行了分析测定,首次取得了甘肃小陇山地区13种乔木、14种灌木、10种草本植物的不同器官和7种林分的枯落物有机含碳率的精确测定值,并在林木个体及林分两个层次上系统地分析了这些树种各不同组分含碳率的变化特征。(2)小陇山主要林分类型的13种乔木树种的器官平均含碳率分别为锐齿栎0.4653,油松0.5049,栓皮栎0.4755,白桦0.4985,红桦0.4889,日本落叶松0.4963,华山松0.4962,云杉0.4906,秦岭冷杉0.4907,水曲柳0.4647,大叶椋子木0.4501,五角枫0.4689,辽东栎0.4632;14种灌木的器官平均含碳率为0.4446,10种草本的器官平均含碳率为0.3270,7种林分枯落物平均含碳率为0.4221;8类林分的乔木层平均含碳率分别为锐齿栎0.4676,油松0.4976,栓皮栎0.4681,杨、桦0.4837,落叶松0.4903,华山松0.4926,云、冷杉0.4902,其他阔叶混交0.4682;5种针叶树林分平均含碳率为0.4927,8种阔叶树种林分平均含碳率为0.4719。(3)在本研究对象范围内,乔木树种种内各组分的含碳率变异系数在1.55%-4.91%之间;组分含碳率的种间变异系数在1.75%-6.59%之间。阔叶树的平均含碳率值大多接近0.47,针叶树种的平均含碳率接近0.49,相对目前国内外普遍应用的两种森林植被生物量含碳率换算系数0.45与0.5,以0.48作为转换系数来估算全部森林乔木层的碳储量,估算结果可能更优。更精确的估算应该是依据不同区域不同森林类型而采用不同的含碳率转换系数。(4)不同树种各组分含碳率高低随机分布,并不呈现出某种规律性的变化,其大小完全由各树种自身的特性决定。但从树种的形态学特性比较可以看出,针叶树种各组分的含碳率普遍高于阔叶树种,所有针叶树种各组分的平均含碳率均高于阔叶树、灌木树种和草本。针叶树种组分碳的平均含碳率普遍高于阔叶树种平均高出1.47%-3.40%,相应的针叶林分的平均含碳率也高于阔叶林。各树种组分的平均含碳率值(算术平均)与林分的平均含碳率值(按生物量加权平均)非常接近,在本研究所涉及的13个乔木树种中其差值不超过1.81%。(5)与前人研究比较,小陇山地区森林植被的无论是树种还是相应林分的含碳率均普遍低于华北地区。由此可以得出结论,生长在不同区域的同一树种各组分的含碳率不尽相同,组成的林分含碳率也存在差异。因此,森林碳储量的估算精度与估算单元的区域尺度密切相关,建立在小区域尺度或区域森林生态系统尺度上的估算精度较高,树种的含碳率可能与树木的生长情况有关,即含碳率与影响林木生长的纬度、海拔、降水等气候条件相关。4.碳储量估算(1)甘肃小陇山8类林分按面积加权的森林植被层总储碳密度由大到小依次为:云、冷杉46.7342 t hm-2,其他阔叶混交46.1454 t hm-2,锐齿栎39.2335 t hm-2,栓皮栎38.1974 t hm-2,杨、桦37.3032 t hm-2,华山松34.4068 t hm-2,落叶松33.7259thm-2,油松30.9395 t hm-2。(2)小陇山8类林分森林植被层总碳储量由大到小依次分别为:锐齿栎7.0054Tg,其他阔叶混交3.5959 Tg,油松1.4466 Tg,栓皮栎0.6492 Tg,杨、桦0.2642 Tg,落叶松0.2105 Tg,华山松0.1656 Tg,云、冷杉0.0208 Tg。小陇山林区各林分森林植被层平均储碳密度为39.4254 t hm-2,总碳储量为1 3.3579 Tg。(3)研究结果表明,甘肃小陇山8类林分乔木层的平均储碳密度值与我国及世界各地森林平均储碳密度的一些估计值相比基本接近。5.创新点(1)该研究填补了我国秦岭西段森林生物量和碳储量的研究空白。(2)该研究为中国西部森林生态系统碳循环研究提供了部分基础研究数据。(3)该研究在运用森林资源清查资料估计小陇山林分各组分生物量方面取得了创新性进展,提出形如W=a+bV的区域生物量相容性线性估计模型具有很高的精度,并从理论上证明了这种建立在林分水平上的估计模型可以直接推广应用到区域尺度,对于区域生态系统生物量和碳储量研究具有重要参考价值。更多还原

【Abstract】 Global carbon cycle is one of the core issues in the researches of global climate changes and the carbon cycle in forest ecosystem is an important part of the global carbon cycle. In present studies, there are still a lot of uncertainties in the studies of global carbon cycle and these uncertainties also exist in the estimats of the size of the regional forest ecosystem carbon storage and the exchange flux between the forest ecosystem and CO₂. Therefore to estimate the gross of carbon storage in regional forest ecosystem and its exchange flux with other carbon storages, all the while, is the hot and difficult problem.This paper does investigations on eight forest stands of Xiaolong Mountains, Gansu Province: Quercus aliena var. acuteserrata, Pinus tabulaeformis, Quercus variabilis, Populus sp. and Betula sp., Larix sp., Pinus armandii, Picea sp. and Abies sp., and other broad leaved mixed forest. It focuses on stand biomass estimate, foundation of tree layer biomass model with compatibility, organic carbon ratio measures of tree layer, shrub layer and herb layer, litter fall carbon ratio measures, the evaluation of the density and amount about regional forest ecosystem carbon storage and the allocation. It is aimed to find an appropriate way to exactly estimate the size of carbon storage of the regional forest ecosystem basing on the forest resource inventory data in the hope of supplying some data for researches on carbon sequestration functions of the regional forest ecosystem and carbon circle of the ecological geographic regional forest ecosystem in China, and providing the foundation and reference for the regional environment construction, the regional forest ecosystem carbon storage management policy and the negotiation of the carbon storage trade that China government participated in. After years of extensive studies,we reached the following conclusings:Part 1 Biomass estimates（1） In eight stands in Xiaolong Mountains, there are close relationship among single wood biomass （W） and trunk diameter （D）, and diameter-height （D²H）. With eleven mathematical models, we simulate regression equations about W-D-D²H correlations and results show that the correlations are prominent, and the simulative effect of Power equation is best.（2） In the sequence of the eight stands: Quercus aliena var. acuteserrata, Pinus tabulaeformis, Quercus variabilis, Populus sp. and Betula sp., Larix sp., Pinus armandii, Picea sp. and Abies sp., and other broad leaved mixed forest:Their tree layer biomass are: 81.8388, 60.6976, 79.7980, 75.3828, 67.3706, 67.9786, 94.3733, 96.5533 t hm^-2 separately.Their shrub layer biomass are: 1.5126, 0.9884, 1.2504, 1.3746, 1.3618, 1.3691, 1.3746, 1.59441 hm^-2 separately.Their herb layer biomass are: 0.1208, 0.2266, 0.1738, 0.2043, 0.2658, 0.196 8, 0.2043, 0.1393 t hm^-2 separately.Their stand biomass are: 84.0472,62.4424, 81.7747,77.4367,68.9982,70.0695, 96.4865, 98.7235 t hm^-2 separately.Part 2 Foundation of Tree Layer Biomass Prediction Model Based on Forest Resource Inventory Data（1） Graphic Analysis, Correlative Analysis and Regression Analysis about eight stands in Xiaolong Mountains show that there exists a prominent linear relationship between species wood biomass （W） and living forest volume （V）. With five evaluation indexes, i.e. Correlation index, Correlation coefficient, Residual standard deviation, Average fractional error and Prediction accuracy, by the comprehensive comparsion of the simulative results about regression equations, and basing on forest resourceinventory statistical data for forest biomass estimates （W～V）, we build that the optimal model in the form of W=a+bV, and we prove theoretically that the forest biomass linear estimation model can be directly applied to the regional scale.（2） As far as estimate effects as concerned, the linear models with compatibility of W^V can be used to calculate tree layer biomass by using the sample plots data, stand volume and other information directly, so as to avoid the disadvantage about the traditional methods which cause the destruction of forest resources. Therefore this improved approach is of higher theoretical and practical value.（3） Researches of the eight stands in Xiaolong Mountains show that there is a better correlation between average net growth (△W_total) and overall biomass (W_total).We make use of this relationship to build several mathematical models about eight stands for △W_total ～ W_total , which best model shows that correlations between averagenet increase in the total amount of forest biomass in different stand are different. Those of Quercus aliena var. acuteserrata, Pinus tabulaefonnis, Larix sp., Picea sp. and Abies sp., are Power model related. Those of Quercus variabilis, Populus sp. and Betula sp., Pinus armandii, are linear model related. And other broad leaved mixed forest are Exponential model related.Compared with previous researches, results show that, the correlativity between average stand net growth and volume of stand biomass in different area is different.Part 3 Measures of Carbon content Rate（1） Using high-precision element analytical method （Dry Combustion method） to measure the carbon content rate of different components （stems, branches, leaves, barks, roots） of eight stands of Xiaolong Mountains, we first acquire accurate value of organic carbon content rate of 13 tree species, 14 shrub species, 10 herbaceous plants, and the forest litters of 7 stand types. We also systematically analyze the changing characteristics of carbon content rate of individual tree and stand.（2） The average organic carbon content rate of 13 tree species, the main standtypes of Xiaolong Mountains, separately, Quercus aliena var. acuteserrata, 0.4653; Pinus tabulaeformis, 0.5049; Quercus variabilis, 0.4755; Betula platyphylla, 0.4985; Betula albo-sinensis, 0.4889; Larix leptolepis, 0.4963; Pinus armandii, 0.4962; Picea asperata, 0.4906; Abies chensiensis, 0.4907; Fraxinus mandschurica, 0.4647; Cornus macrophylla, 0.4501, Acer mono, 0.4689; Quercus liaotungensis, 0.4632. The average organic carbon content rate of 14 shrubs is 0.4446. The average organic carbon content rate of 10 herb plants is 0.3270. The average carbon content rate of litter fall of 7 stands is 0.4221. The average carbon content rate of tree layer of 8 stands separately, Quercus aliena var. acuteserrata, 0.4676; Pinus tabulaeformis, 0.4976; Quercus variabilis, 0.4681; Populus sp. and Betula sp., 0.4837; Larix sp., 0.4903; Pinus armandii, 0.4926, Picea sp. and Abies sp., 0.4902; and other broad leaved mixed forest, 0.4682; The average carbon content rate of 5 coniferous forests is 0.4927. The average carbon content rate of broad leaved forests of 8 stands is 0.4719.（3） For the objects studied in this article, the variation coefficient of carbon content rate of components is within 1.55%～4.91%. The variation coefficient between species is within 1.75%～6.59%. The average carbon content rate of broad leaved forest is close to 0.47. The average carbon content rate of coniferous forest is close to 0.49. Compared to domestic and universal application of the two types forest biomass conversion factor of carbon content rate, 0.45 and 0.5, using 0.48 as conversion coefficients to estimate the total forest tree layer of carbon storage may make estimate result in better effects. for more accurate, we should adopt corresponding conversion coefficients of carbon content rate according to different region and different forest style.（4） Carbon content rates of different tree species, at random distribution, do not show regular changing orderliness and are determined by the characteristics of the trees characters. Judged from the morphological characteristics of the species, component carbon content rate of coniferous forest is higher than that of broad leaved forest, shrub forest and herbaceous plants. The average carbon content rate of conifer species composition is generally in the range between 1.47% and 3.40%, higher than that of broad leaved forest. Coniferous forest corresponding rate is also higher thanthe average carbon content rate of broad leaved forest. The average carbon content rate of species （arithmetic mean） and that of stands, （by weighted average of biomass volume） is very close. In the 13 tree species in this research, the margin does not exceed 1.81%.（5） Compared with previous researches, in forest vegetation of Xiaolong Mountains, the carbon content rate of species and stands both are lower than those in North China. We may conclude that there exists different carbon content rate of the same species in different areas. There are also differences in the carbon content rate within the same stand. The estimation accuracy of forest carbon storage is closely related to estimation of the regional scale module. Estimation accuracy based on Regional scale or regional forest ecosystems scale is better. Carbon content rate of tree species may be related with the growth state, i.e. carbon content rate may be related with latitude, altitude, rainfall and other climate-related conditions, all of which affect tree growth.Part 4 Carbon Storage Estimates（1） The total carbon density of forest vegetation about the eight types in Xiaolong mountains weighted by the area in sequence: Picea sp. and Abies sp., 46.7342 t hm^-2 ; other broad leaved mixed forest, 46.1454 t hm^-2 . Quercus aliena var. acuteserrata, 39.2335 t hm^-2; Quercus variabilis, 38.1974 t hm^-2; Populus sp. and Betula sp., 37.3032 t hm^-2. Pinus armandii, 34.4068 t hm^-2. Larix sp., 33.7259 t hm^-2; Pinus tabulaeformis, 30.9395 t hm^-2.（2） Carbon storage of forest vegetation in Xiaolong mountains weighted by the the area in sequence: Quercus aliena var. acuteserrata, 7.0054 Tg, other broad leaved mixed forest, 3.5959 Tg; Pinus tabulaeformis, 1.4466 Tg; Quercus variabilis, 0.6492 Tg; Populus sp. and Betula sp., 0.2642 Tg. Larix sp., 0.2105 Tg; Pinus armandii, 0.1656 Tg; Picea sp. and Abies sp., 0.0208 Tg. The density of carbon storage of forest vegetation in each stand is 39.4254 t hm^-2. The total carbon storage is 13.3579 Tg.（3） Research results show that, the average density of carbon storage in tree layerof 8 stands in Xiaolong Mountains is close to the estimated carbon density of other forests in China and in the world.Part 5 Highlights（1） This dissertation filled the research vacancy on forest biomass and carbon storage of western Qinling Mountain in China.（2） This dissertation provided some basic data for research on carbon cycle of forest ecosystem in west of China.（3） This dissertation made an innovative progress in estimating the forest stands biomass of Xiaolong Mountains with forest resource inventory data, found linear model with regional biomass compatibility in the form of W = a + bV with high accuracy, and we has proved from theory that the model W = a + bV can be directly applied to the regional scale and it has important reference value for the biomass and carbon storage study about regional ecosystem.更多还原