森林土壤有机碳深度分布模型的构建与应用

Abstract

PDF HTML阅读 XML下载 导出引用 引用提醒 森林土壤有机碳深度分布模型的构建与应用 DOI: 10.5846/stxb201403250556 作者: 作者单位: 河南科技大学林学院,中国林业科学研究院森林生态环境与保护研究所,中国林业科学研究院森林生态环境与保护研究所,秭归县林业局 作者简介: 通讯作者: 中图分类号: Q149 基金项目: 中央级公益性科研院所基本科研业务费专项资金(CAFRIFEEP201101) Construction and application of a depth distribution model for soil organic carbon in forest areas Author: Affiliation: College of Forestry,Henan University of Science and Technology,Luoyang,State Forestry Administration Key Laboratory of Forest Ecology and Environment,Research Institute of Forest Ecology,Environment and Protection,Chinese Academy of Forestry,State Forestry Administration Key Laboratory of Forest Ecology and Environment,Research Institute of Forest Ecology,Environment and Protection,Chinese Academy of Forestry,Forestry Bureau of Zigui County,Yichang Fund Project: 摘要 | 图/表 | 访问统计 | 参考文献 | 相似文献 | 引证文献 | 资源附件 | 文章评论 摘要:了解森林土壤有机碳 (SOC)的深度分布模式对正确估算森林碳储量,充分发挥森林碳汇功能,减缓全球气候变化有着重要意义.选取寒温带针叶林、温带落叶林、亚热带针阔混交林、热带常绿阔叶林等4类森林生物群系,建立SOC深度分布数据库,构建SOC质量密度的深度分布模型;使用Nash-Sutcliffe效率系数 (E)、误差百分比 (PE)、决定系数 (R2) 等统计参量评定模型的模拟效果;利用构建的深度分布模型外推更深层SOC密度.研究结果表明:(1) 本文所构建的森林SOC深度分布模型模拟值与观测值较为吻合,Nash-Sutcliffe效率E、误差百分比PE和决定系数R2平均为0.74、6.95%、0.88 (P < 0.05),模型模拟能力较高 (E > 0.6),模拟误差值低于可接受的临界值 (PE < ±15%),说明构建的模型可以对该地区森林SOC密度值进行估算;(2) 寒温带针叶林0-20 cm层SOC质量密度较高,热带常绿阔叶林较低;20 cm以下则是寒温带针叶林较低,热带常绿阔叶林较高,热带常绿阔叶林具有更深层的SOC分布;用0-100 cm深度的SOC来表征区域SOC储量时结果偏低.若考虑0-200 cm深度,0-100 cm深度SOC值平均偏低约21.8%,在热带地区这种偏低趋势可能更加突出,误差可能更大.(3) 模型对表层SOC密度有偏低预测趋势,对深层SOC密度预测值可能偏高;作为一个森林SOC深度分布模拟工具,模型可以在有限区域条件下估算不同深度SOC密度值. Abstract:The pool of soil organic carbon (SOC) in a forest forms an important component in the global carbon (C) cycle. SOC plays an important role in enhancing forest productivity and mitigating the net rate of global greenhouse gas emissions. The risk of global warming has caught the attention of the scientific community as it relates to SOC stocks in forest ecosystems. The precise measurement of SOC stocks and verification of the amount of C sequestered in the soil are critical factors for the implementation of C trading programs. SOC in mineral soils generally decreases with depth; however, this decrease is non-linear and has been frequently modeled as an exponential function. We selected four forest types (boreal forest, temperate deciduous forest, subtropical mixed forest, and tropical evergreen broadleaved forest) and analyzed the exponential function for SOC mass density. We established an SOC database for layers of the soil profile by measuring the SOC in typical areas in the four forest biomes. The depth distribution models for the mass density of SOC were established by a typical sampling method. The model was calibrated using 60% of the data of the profiles, and 40% of the data was used for validation purposes. The entire evaluation for the results of model simulation consisted of determining the coefficient of determination (R2), Nash-Sutcliffe coefficient of efficiency (E), and the percentage error (PE). Next, the depth distribution models evaluated here were used to simulate the distribution of SOC deeper into the soil profile. The results showed that the simulation values for the depth distribution models of the four forest biomes and the observed values were relatively consistent. The average values of R2, E, and absolute PE were 0.88, 0.74, and 6.95%, respectively. The model simulations had a relatively high capacity (E > 0.6), and the PE of the model was simulated within a range with acceptable accuracy (PE < ±15%). The model could be used to simulate the depth distribution of forest soil organic carbon. Second, the boreal coniferous forest had a much higher density of SOC in the 0-20 cm layer than those of the tropical deciduous forest and the two other forest types. In contrast, the SOC densities in other layers of boreal coniferous forest were lower, while those of the tropical deciduous forest were higher. The regional SOC densities were lower when SOC densities in the 0-100 cm soil layers were used to characterize the regional SOC density. When compared with the SOC densities in the range of 0-200 cm in the soil profile, the SOC densities in the 0-100 cm soil layer were about 21.8% lower than the overall density. Any error in this calculation may be greater and more prominent in regions with high temperatures and precipitation rates. For rainfall events of a small magnitude, the model generally over-estimated mass density at the bottom of the soil profile, while the opposite was true; that is, for regions with large amounts of rainfall, the model generally under-estimated the surface SOC density. In general, the model performs well at simulating the depth distribution of SOC, and it can be used as a forest SOC management tool to simulate the depth distribution of SOC in some regions. 参考文献 相似文献 引证文献