欢迎访问林业科学,今天是

林业科学 ›› 2019, Vol. 55 ›› Issue (9): 103-110.doi: 10.11707/j.1001-7488.20190911

• 论文与研究报告 • 上一篇    下一篇

基于节子分析技术构建落叶松人工林树冠基部高动态模型

陈东升1, 李凤日2, 孙晓梅1, 张守攻1   

  1. 1. 中国林业科学研究院林业研究所 国家林业和草原局林木培育重点实验室 北京 100091;
    2. 东北林业大学林学院 哈尔滨 150040
  • 收稿日期:2017-07-28 修回日期:2018-05-19 发布日期:2019-10-28
  • 基金资助:
    国家自然科学基金重点项目(31430017);"十三五"国家重点研发计划课题(2017YFD0600401)。

Reconstruction Dynamic Models of Height to Crown Base of Larch(Larix olgensis)Plantation Applying Knot Analysis Technique

Chen Dongsheng1, Li Fengri2, Sun Xiaomei1, Zhang Shougong1   

  1. 1. Key Laboratory of Tree Breeding and Cultivation, National Forestry and Grassland Administration Research Institute of Forestry, CAF Beijing 100091;
    2. School of Forestry, Northeast Forestry University Harbin 150040
  • Received:2017-07-28 Revised:2018-05-19 Published:2019-10-28

摘要: [目的]基于节子分析技术构建落叶松人工林树冠基部高动态预测模型,分析落叶松树冠衰退规律及其影响因素,为制定合理的经营措施提供理论依据。[方法]以2007年设立的8块落叶松人工林标准地获取的40株解析木数据为基础数据,采用节子分析技术,得到树冠基部高随年龄的动态变化数据,应用传统线性模型、理查德和逻辑斯蒂非线性模型构建落叶松树冠基部高动态模型。[结果]传统线性模型、理查德和逻辑斯蒂非线性模型可较好拟合树冠基部高动态变化过程,模型参数均具有统计意义(P<0.01),以理查德方程为基础模型构建的树冠基部高模型拟合效果最好,加入权重因子可消除异方差,降低估计参数标准误,提高预测精度,模型的确定系数(R2)为0.904,绝对误差(Bias)和均方根误差(RMSE)分别为0.002和1.251,最优落叶松树冠基部高模型形式为HCB=(3.146+0.036CCF+0.225Bas+0.788HT-0.481CL)(1-e-0.086 t4.278。[结论]树冠基部高动态变化过程与林分发育规律一致,符合"S"形生长曲线,可通过树冠竞争因子(CCF)、林分断面积(Bas)、调查当年的树高(HT)和冠长(CL)解释,解释率达90.4%。树高、树冠竞争因子和林分断面积增大会导致树冠基部高上升,加速落叶松树冠衰退。竞争对树冠的影响较敏感,落叶松人工林10~41年间,树冠竞争因子大(187.33)的林分冠长率从75%下降到36%,而树冠竞争因子小(105.82)的林分冠长率从75%下降到40%;落叶松人工林树冠基部高平均每年上升0.66 m。本研究构建的树冠基部高动态模型可较好模拟落叶松人工林树冠基部高动态变化过程,利用单木和林分变量能够解释落叶松人工林树冠衰退趋势。通过检验验证,基于节子分析技术获取的树冠基部高数据构建的动态模型精度较高,可作为一种获取长期树冠动态变化数据的有效手段。

关键词: 落叶松人工林, 树冠基部高, 树冠衰退, 线性和非线性模型, 节子分析技术

Abstract: [Objective] The dynamic model of height to crown base was built to show crown recession in larch plantation by knot analysis technique, and to evaluate the influence of individual tree and stand variables on crown recession.[Method] Based on falling 40 sample trees in 8 plots of larch plantation in 2007, the dynamic data of height to crown base were obtained by applying knot analysis technique. The traditional linear model, Richard and Logistic non-linear model were used to construct the dynamic model of height to crown base.[Result] Linear model, Richard and Logistic nonlinear model could well fit the dynamic change process of height to crown base, model parameters were statistically significant (P<0.01). One of the highest fitting precision was based on Richard equation model, but the model had heteroscedasticity, by adding the weight to eliminate heteroscedasticity, eventually model R2 was 0.904, Bias and RSME were 0.002 and 1.251 respectively. The model form is HCB=(3.146+0.036CCF+0.225Bas+0.788HT-0.481CL)(1-e-0.086 t)4.278.[Conclusion] Dynamic process of height to crown base was consistent with the stand growth rule which it was accordance with the "S" type growth curve. It can be explained through crown competition factor (CCF), stand basal area (Bas), tree height (HT) and crown length (CL). The height to crown base increased with the increase of HT, CCF and Bas, accelerating larch crown recession. From 10 years to 41 years, crown length of larch plantation fell from 75% to 36% in biggest CCF(187.33), however, fell from 75% to 40% in smallest CCF(105.82). We also found that a mean annual increment of height to crown was 0.66 m in larch plantation. The dynamic process of height to crown base of larch plantation was well simulated by model, crown recession could be explained by tree, stand variables. Consequently, knot analysis technology was considered as an effective method to acquire a long-term dynamic data of crown.

Key words: larch plantation, height to crown base, crown recession, linear and nonlinear model, knot analysis technique

中图分类号: