欢迎访问《应用生态学报》官方网站,今天是 分享到:

应用生态学报 ›› 2016, Vol. 27 ›› Issue (12): 3862-3870.doi: 10.13287/j.1001-9332.201612.030

• • 上一篇    下一篇

三种林分生物量估算方法的比较

董利虎, 李凤日*   

  1. 东北林业大学林学院, 哈尔滨 150040
  • 收稿日期:2016-05-31 出版日期:2016-12-18 发布日期:2016-12-18
  • 通讯作者: * E-mail: fengrili@126.com
  • 作者简介:董利虎,男,1986年生,博士,讲师. 主要从事林分生长与收获、森林生物量和碳储量研究. E-mail: donglihu2006@163.com
  • 基金资助:
    本文由中央高校基本科研业务费专项资金项目(2572015BX03)和黑龙江省留学归国人员科学基金项目(LC2016007)资助

Comparison of three stand-level biomass estimation methods

DONG Li-hu, LI Feng-ri*   

  1. College of Forestry, Northeast Forestry University, Harbin 150040, China
  • Received:2016-05-31 Online:2016-12-18 Published:2016-12-18
  • Contact: * E-mail: fengrili@126.com
  • Supported by:
    This work was supported by the Fundamental Research Funds for the Central Universities of China (2572015BX03), and the Scientific Research Foundation of Heilongjiang Province for the Returned Overseas Chinese Scholars (LC2016007).

摘要: 区域森林生物量的估算方法是人们目前关注的焦点,建立林分生物量模型成为一种趋势.本文以吉林省落叶松人工林固定样地为例,采用非线性似乎不相关回归法构建2种林分生物量模型,即基于林分变量的林分生物量模型(模型系统Ⅰ)和基于生物量换算系数的林分生物量模型(模型系统Ⅱ),给出落叶松人工林固定生物量换算系数值,并比较了3种林分生物量估算方法的预估精度.结果表明: 所建立的2种林分生物量模型中,总生物量和树干生物量模型拟合和预测效果较好,其Ra2>0.95,且均方根误差(RMSE)、平均预测误差(MPE)和平均绝对误差(MAE)都较小.树叶和树枝生物量模型拟合和预测效果相对较差,其模型的Ra2<0.95.模型系统Ⅰ和模型系统Ⅱ的预测精度均优于固定生物量换算系数法.基于生物量换算系数的林分生物量模型属于材积源生物量法,其本质与基于林分变量的林分生物量模型不同,但二者的预测效果相当.固定生物量换算系数的预测能力较差,将生物量与蓄积量之比假定为恒定常数是不恰当的.此外,为了使模型参数估计更有效,所建立的生物量模型应当考虑林分总生物量及各分项生物量的可加性.

Abstract: At present, the forest biomass methods of regional scale attract most of attention of the researchers, and developing the stand-level biomass model is popular. Based on the forestry inventory data of larch plantation (Larix olgensis) in Jilin Province, we used non-linear seemly unrelated regression (NSUR) to estimate the parameters in two additive system of stand-level biomass equations, i.e., stand-level biomass equations including the stand variables and stand biomass equations including the biomass expansion factor (i.e., Model system Ⅰ and Model system Ⅱ), listed the constant biomass expansion factor for larch plantation and compared the prediction accuracy of three stand-level biomass estimation methods. The results indicated that for two additive system of biomass equations, the adjusted coefficient of determination (Ra2) of the total and stem equations was more than 0.95, the root mean squared error (RMSE), the mean prediction error (MPE) and the mean absolute error (MAE) were smaller. The branch and foliage biomass equations were worse than total and stem biomass equations, and the adjusted coefficient of determination (Ra2) was less than 0.95. The prediction accuracy of a constant biomass expansion factor was relatively lower than the prediction accuracy of Model system Ⅰ and Model system Ⅱ. Overall, although stand-level biomass equation including the biomass expansion factor belonged to the volume-derived biomass estimation method, and was different from the stand biomass equations including stand variables in essence, but the obtained prediction accuracy of the two methods was similar. The constant biomass expansion factor had the lower prediction accuracy, and was inappropriate. In addition, in order to make the model parameter estimation more effective, the established stand-level biomass equations should consider the additivity in a system of all tree component biomass and total biomass equations.