| 摘要: |
| 森林碳储量估测可以为评估碳源 / 碳汇及其空间格局提供定量化信息。青藏高原由于其独特的地理、气候特征，了解其森林在陆地生态系统碳汇中的作用非常重要。文章提出了一种直观的、基于GIS 数据来估测青藏高原地上森林碳储量的方法，这种方法结合了覆盖 3 个实验区（西藏、云南和四川）250 m 空间分辨率的 MODIS 数据、气候数据、1 km 的 DEM、1:2 500 000 林相图和 1 086 个森林资源清查小班数据（NFI）。采用相关分析和线性诊断方法对关键变量进行选择，线性回归模型和对数模型用于森林碳储量模型建立。结果表明，采用三个不同模型分别对地理特征不同的子区域内森林碳储量的测算结果，比使用一般模型对没有划分子区域的整个青藏高原森林碳储量的测算更准确。虚拟变量 ( 如层、从DEM 提取的植被类型、NFI 和青藏高原森林地图 ) 的融合提高了碳估算模型的性能。同时，线性模型和对数模型可决系数（R2）也分别随之增大。青藏高原地上森林碳储量不同子区域间变化较大，从 50 000到 250 000 kg/hm2，平均值 19 000 kg/hm2，而灌木碳储量则不及 10 000 kg/hm2。 |
| 关键词: 森林碳储量 虚拟变量 对数模型 MODIS数据 青藏高原 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金(31800529) |
|
| Quantification of forest carbon storage in Tibetan Plateau |
|
Wang Xiyang1, KE Biying1, HUANG Zhiqing1, YANG Guang2, HUANG Guihua2, HU Qipeng3, SUN Lingling4
|
|
1.Guangdong Eco-engineering Polytechnic;2.Key Laboratory of National Forestry and Grassland;3.Sino-Forest china Investments Corp LTD,Guangzhou;4.The Pearl River Hydraulic Research Institute,Guangzhou
|
| Abstract: |
| Estimating forest carbon storage can provide quantitative information to assess carbon source/sink size and its spatial patterns. Tibetan Plateau is prone to climate change due to its unique geographic feature, and thus it is important to understand the role of Tibetan forests in China terrestrial ecosystem carbon sink. This paper presented a spatially explicit GIS based method, which incorporated MODIS data with 250 m spatial resolution, climate data, 1 km DEM, 1:2,500,000 forest map and
1 086 plot data of National Forest Inventory (NFI) covering the three test regions (Tibet, Yunnan and Sichuan), to estimate the above-ground forest carbon storage in Tibetan Plateau. Correlation analysis and linearity diagnosis were used to choose key variables, and linear regression model and logarithmic model were applied to develop the forest carbon storage models. The results showed that forest carbon storage estimation by using the three separated models for different respective sub-regions according to its physical geographic characteristics was more accurate than by using the general model for the whole Plateau without dividing sub-regions. Incorporating dummy variables, such as aspect and vegetation types derived from DEM, NFI and forest map of Tibetan Plateau, improved the carbon estimation model performance. The determining coefficients (R2) of the linear models and logarithmic models were also increased, respectively. The above-ground forest carbon storage in Tibetan Plateau varied greatly with the subregions, ranging spatially from 50 000 kg/hm2 to 250 000 kg/hm2, with the mean of 19 000 kg/hm2, while the carbon storage of shrub was less than 10 000 kg/hm2. |
| Key words: forest carbon storage dummy variable logarithmic model MODIS data Tibetan Plateau |
