Knowledge Resource Center for Ecological Environment in Arid Area
| 非线性优化方法在陆地生态系统模拟及其不确定性研究中的应用 | |
| 其他题名 | Application of nonlinear optimization method to simulation of the terrestrial ecosystem and its uncertainty |
| 孙国栋 | |
| 出版年 | 2009 |
| 学位类型 | 博士 |
| 导师 | 穆穆 |
| 学位授予单位 | 中国科学院大气物理研究所 |
| 中文摘要 | 本文利用五变量草原生态系统模式和Lund－Potsdam－Jena（LPJDGVM）模式探讨了条件非线性最优扰动（CNOP）方法和条件非线性最优参数扰动（CNOPP）方法在陆地生态系统模拟及其不确定性研究中的应用，分别考察了人类活动与气候变化对草原（沙漠）生态系统和陆地生态系统的影响。主要内容和结论如下：\n（1）利用五变量草原生态系统模式探讨了草原（沙漠）生态系统平衡态的稳定性和平衡态突变的非线性特征。结果表明，当初始扰动足够大时，线性稳定的草原平衡态是非线性不稳定的。将线性奇异向量（LSV）和Lyapunov向量（LV）的结果与CNOP的结果进行比较，我们发现：对于某些约束条件参数，CNOP能够导致平衡态的突变，而LSV和LV不能导致平衡态的突变。这说明了草原（沙漠）平衡态的突变具有非线性特征。将CNOP类型与非CNOP类型的初始扰动的结果进行比较，我们发现枯草量的遮荫作用在平衡态突变过程中有着重要的作用。对于其它湿润度指数（μ=0.325），草原（沙漠）平衡态的稳定性及其突变的非线性特征的结果与上述结果是类似的。湿润度指数越大，草原平衡态越稳定，沙漠平衡态越不稳定。对于前者较大的约束条件参数才能导致草原平衡态的突变，对于后者较小的约束条件参数就能导致沙漠平衡态的突变。\n（2）利用五变量草原生态系统模式考察了气候变化对草原（沙漠）生态系统平衡态的影响。结果表明，当湿润度指数为0.31时，不同生态系统的CNOPP具有非线性特征。在其影响下的草原（沙漠）平衡态各个分量在优化时间内减少（增加）。 在约束条件参数较小情况下，草原（沙漠）平衡态不会发生突变，而约束条件参数足够大时，草原（沙漠）平衡态将会发生突变。这里约束条件参数的大小代表气候变化的大小。具有和某些约束条件参数大小相同的、线性特征变化的气候条件，不会导致草原（沙漠）平衡态的突变，这说明了气候条件的非线性变化(CNOPP)对生态系统的影响更大。不同湿润度指数敏感性（μ=0.32）的研究结果表明湿润度指数越大草原平衡态越稳定，而沙漠平衡态越不稳定。\n（3）利用Lund－Potsdam－Jena模式模拟了中国区域近20年植被分布以及碳通量的变化。研究结果发现模拟的植被功能类型的分布与观测基本相符。模拟的中国区域净初级生产力（NPP）总量以及其年际变化与其它模式的模拟结果类似，模拟的净生态系统生产力（NEP）在可接受的范围内。利用条件非线性最优扰动方法研究了土地利用对中国区域植被碳贮量和净初级生产力（NPP）的影响。结果表明北方常绿针叶林带(BoNE)的相对变化大于温带夏绿阔叶林带(TeBS)的相对变化，而温带常绿阔叶林带(TeBE)是三种植被功能类型中相对变化最小的。进一步研究结果发现土地利用程度越高，中国区域NPP总量增加的幅度就越大。我们还探讨了三种植被功能类型相对变化与气候条件的关系。此外，在土地利用情况下，三种植被功能类型的分布基本保持不变，只是在东北的部分区域稍有变化。在此区域北方常绿针叶林被北方夏绿针叶林或北方夏绿阔叶林所代替。\n（4）利用Lund－Potsdam－Jena模式评估了温度和降水的变化对中国区域陆地生态系统的影响。CNOPP叠加在原始温度序列后，不仅使得温度平均值发生变化（增加2o），而且温度变率也发生变化。东北区域和南方区域NPP增加，研究区域的中部地区（即干旱与半干旱区域）NPP减少。NPP相对变化较大的区域主要分布在中部地区和南方区域。而对于温度序列仅平均值增加2o，NPP变化的区域与CNOPP的结果基本一致，但在数值上偏小。NPP相对变化较大的区域分布在中部区域，这个结果与CNOPP的结果一致。但是在南方区域，NPP的相对变化很小，这个结果与CNOPP的结果是不同的。对于降水，CNOPP导致降水平均值增加20%，变率也有所变化，但是变化幅度较小。NPP在整个研究区域增加，在北方区域NPP增加的幅度较大。对于降水平均值增加20%，而变率不发生变化的情形下，NPP的变化与CNOPP的结果基本一致，但是变化幅度没有CNOPP的结果大。 |
| 英文摘要 | Using a five variable grassland ecosystem model and aLund-Potsdam-Jena dynamical global vegetation model (LPJ DGVM), the conditional nonlinear optimal perturbation (CNOP) and the conditional nonlinear optimal parameter perturbation (CNOPP) are employed to study simulation of the terrestrial ecosystem and its uncertainty, and to consider the response of the grassland (desert) ecosystem and the terrestrial ecosystem to human activity and climate change. The main conclusions are as follows.\n(1)Based on a five-variable theoretical ecosystem model, the stability of equilibrium state and the nonlinear feature of the transition between a grassland state and a desert state are investigated. The numerical results indicate that the linearly stable grassland and desert states are nonlinearly unstable to large enough initial perturbations. The results obtained by CNOP, LSV and Lyapunov vector (LV) are compared to analyze the nonlinear feature of the transition between the grassland state and the desert state. It is shown that the CNOP could lead to the transition and the LSV and the LV fail for some initial perturbations. Besides, it is shown that there are similar results about the stability of equilibrium state and the nonlinear feature of the transition for the other moisture index(μ=0.325). The larger the moisture index is, the stabler the grassland ecosystem is and the unstabler the desert ecosystem is.\n(2)By using the five-variable theoretical ecosystem model and the CNOPP method, the stability of the grassland (desert) ecosystem equilibrium state is investigated. The CNOPPs of different ecosystems have a nonlinear feature. The values of four variables decrease (increase) gradually for the grassland (desert) ecosystem during the optimization time. When the parameter constraint condition is small, the ecosystem influenced by the CNOPP is stable. With the increasing of the parameter constraint condition, the abrupt change from the grassland (desert) to the desert (grassland) will occur. The parameter perturbation represents climate change. However, the abrupt change would not occur for a linear parameter perturbation of the same amplitude with the CNOPP. It is shown that the nonlinear change of climate condition play an important role in the ecosystem. When the moisture index is large, the grassland ecosystem is stable and the desert ecosystem is unstable.\n(3)With the LPJ-DGVM (Lund-Potsdam-Jena Dynamic Global Vegetation Model), the distribution of vegetation and the interannual variation of carbon flux are simulated in China. The vegetation distribution is in general agreement with the observation. For total and annual NPP, the simulation results by LPJ model is similar to those by other models. The simulated NEP is within its accepted range.\nThe approach of CNOP is applied to investigate the response of the vegetation carbon and the net primary production (NPP) of these three PFTs to land use. The relative change of the vegetation carbon for the boreal needleleaved evergreen tree (BoNE) is greater than that for the temperate broadleaved summergreen tree (TeBS). And the relative change of the vegetation carbon for the temperate broadleaved evergreen tree (TeBE) is the least among three PFTs. Furthermore, it is demonstrated that the larger the perturbations related to land use are, the more significantly the total amount of NPP in China increases. The relationship between the relative change of the vegetation carbon and climate is investigated. The spatial distribution of vegetation keeps invariable except northeastern China. The BoNE is substituted by boreal needleleaved summergreen tree or boreal broadleaved summergreen tree for.\n(4)The impact of the terrestrial ecosystem on climate change is explored. Not only the average temperature increases by 2o, but also its variability increases as the temperature series plus the CNOPP. The NPP decreases in arid and semiarid region, while the NPP increases in northeastern and southern China. If the average temperature increases by 2o , the spatial variation is similar to that of the CNOPP. However, the extent of the variation is smaller than that influenced by the CNOPP. There is a distinct difference about the relative change of the NPP between them. In the southern China, the relative change influenced by the CNOPP is larger than that by the other method. The precipitation series plus CNOPP change its average value, its variability is less variational. In the northern China, the NPP increases distinctly. There are similar results for the precipitation average increasing by 20%. However, the extent of the variation is smaller than that influenced by the CNOPP. |
| 中文关键词 | 草原生态系统 ; 陆地生态系统 ; 非线性优化方法 ; 条件非线性最优扰动 ; 条件非线性最优参数扰动 ; 不确定性 |
| 英文关键词 | grassland ecosystem terrestrial ecosystem nonlinear optimization method conditional nonlinear optima |
| 语种 | 中文 |
| 国家 | 中国 |
| 来源学科分类 | 地球流体力学 |
| 来源机构 | 中国科学院大气物理研究所 |
| 资源类型 | 学位论文 |
| 条目标识符 | http://119.78.100.177/qdio/handle/2XILL650/286721 |
| 推荐引用方式 GB/T 7714 | 孙国栋. 非线性优化方法在陆地生态系统模拟及其不确定性研究中的应用[D]. 中国科学院大气物理研究所,2009. |
| 条目包含的文件 | 条目无相关文件。 | |||||
| 个性服务 |
| 推荐该条目 |
| 保存到收藏夹 |
| 导出为Endnote文件 |
| 谷歌学术 |
| 谷歌学术中相似的文章 |
| [孙国栋]的文章 |
| 百度学术 |
| 百度学术中相似的文章 |
| [孙国栋]的文章 |
| 必应学术 |
| 必应学术中相似的文章 |
| [孙国栋]的文章 |
| 相关权益政策 |
| 暂无数据 |
| 收藏/分享 |
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
