| 摘要: |
| 为构建适用于内蒙古兴安盟的文冠果 Xanthoceras sorbifolium 立木生物量定量模型, 该研究以
文冠果核心种植区扎赉特旗和突泉县为研究区域, 选取代表总体特征平均水平的人工纯林共计 9 个地块,
林龄分别为 2、 5、 9、 11、 16、 19、 33、 39 和 50 a, 共获取 28 个标准样木数据。 采用加权回归法建立基
于地径 (D) 和树高 (H) 为自变量的生物量模型, 通过引入权函数 (形式为 1 / f (x)
n
, n 取值 0. 5 ~
2. 0, 以 0. 5 为间隔) 消除异方差性, 并结合确定系数 (R
2
)、 估计值标准差 ( SEE) 和平均预估误差
(MPE) 指标筛选最优模型。 研究构建了基于地径、 树高及其组合 (D
2H) 的地上与地下生物量一元及二
元模型共 6 个; 引入权函数有效消除了异方差, 最优 n 值因模型而异; 优选模型拟合度高、 参数均显著且
稳健, 评价指标均满足 《立木生物量建模方法技术规程》 (LY / T2258—2014) 对于 MPE 精度小于 10%的
要求。 |
| 关键词: 兴安盟 文冠果 生物量模型 |
| DOI：10. 20221 / j. cnki. 2096-2053. 202504015 |
| 分类号: |
| 基金项目:中央财政林业和草原科技推广示范项目 (内林草科推 [2024] 02 号)。 |
|
| Construction of Biomass Models for Xanthoceras sorbifolium in Xing′an League |
|
CHEN Weiqiang1, XIN Weiwei2, LIU Hongye1, JI Hongfa1, ZHANG Yan2, KONG Yixin3, YU Changjiu4, WANG Yuwei4
|
|
1.Beijing Gloriam Climate Technology Co., Ltd;2.Xing ''an League Forestry Science Research Institute;3.Xing ''an League Forestry and Grassland Workstation;4.Xing′an League Forest and Grassland Bureau
|
| Abstract: |
| To establish quantitative biomass models for standing trees of Xanthoceras sorbifolium in Xing′an
League, Inner Mongolia, this study selected nine pure plantation plots representing average stand characteristics
in the core planting areas of Zhalaite Banner and Tuquan County. Sampled stands covered ages of 2, 5, 9,
11, 16, 19, 33, 39, and 50 a, yielding data from 28 standard sample trees. Weighted regression was used to
develop biomass models with ground diameter (D) and height (H) as predictor variables. Heteroscedasticity
was eliminated by applying weight functions of the form 1 / f (x)
n
, where n ranged from 0. 5 to 2. 0 in 0. 5 inter-vals. Optimal models were selected based on metrics including the coefficient of determination (R
2
), standard
error of estimate (SEE), and mean percentage error (MPE). The results showed that six univariate and bivariate models for aboveground and belowground biomass based on D, H, and D
2H were successfully developed.
Appropriate weight functions effectively eliminated heteroscedasticity, with optimal n values varying by model.
The selected models exhibited high goodness-of-fit, statistically significant and robust parameters, with all evaluation metrics meeting the accuracy requirement of MPE < 10% specified in the " Technical regulation on
methodology for tree biomass modeling (LY / T 2258—2014)". |
| Key words: Xing′an League Xanthoceras sorbifolium biomass model |
