| 摘要: |
| 【目的】研究蒙古栎天然林的进界模型，为其合理经营提供依据。【方法】以蒙古栎天然林为研究对象，基于吉林省汪清林业局195块固定样地的2期复测数据，采用两阶段条件法建立蒙古栎天然林分的进界生长模型：首先使用二分类的Logistic回归方法建立进界概率模型，其次使用普通线性逐步回归建立蒙古栎林的进界株数模型，2个模型的自变量均包括海拔、地位级指数、林分每公顷断面积、林分每公顷株数。【结果】最终确定的蒙古栎天然林的进界概率模型χ2检验结果显示，样地实测进界数和预测进界数无显著差异；ROC（Receiver operating characteristic）曲线显示，AUC（Area under curve）值为0.755。进界株数模型的残差大部分为-1.5～1.5，预测株数与实测株数无显著差异，残差未表现出显著的异质性。【结论】检验结果及ROC曲线均显示，所建立的蒙古栎天然林进界概率模型具有较好的解释性和预测性；进界株数模型的检验结果也表明该模型比较合理。因此，所建立模型可用于吉林省汪清地区蒙古栎天然林的进界生长预测。 |
| 关键词: 蒙古栎天然林 两阶段条件法 进界概率模型 进界株数模型 |
| DOI： |
| 分类号: |
| 基金项目:国家“十二五”科技支撑计划项目“长白山过伐林可持续经营技术研究与示范”（2012BAD22B02） |
|
| Growth models for natural Mongolian oak forests:Ⅳ.Recruitment model |
|
MA Wu,LEI Xiang-dong,XU Guang,et al
|
| Abstract: |
| 【Objective】This study established the recruitment model to provide foundation for reasonable management of Mongolian oak natural forests.【Method】The data used to develop individual-tree diameter growth model for the natural Mongolian oak forests were collected from 195 re-measured permanent sample plots in Wangqing Forest Bureau,Jilin province.Two-stage estimating method was used for logistic regression for recruitment probability model and ordinary linear stepwise regression for recruitment trees model.The independent varibles for both models covered elevation,site class index, stand basal area per hectare,and number of trees per hectare.【Result】Chisquare-Test of final recruitment probability model shows that there was no significant difference between observed and predicted plots.Receiver operating characteristic (ROC) curve shows that the value of area under curve (AUC) was 0.755.The major residuals of recruitment trees model were -1.5-1.5,and there was no significant difference between the oberved and predicted recruitment trees either.The residuals showed no heterogeneity.【Conclusion】Chisquare-Test and ROC curve show that models developed in this study had good explanation and predictability.The model could be easily applied for recruitment projection of Mongolian oak forests in the Wangqing region of Jilin province. |
| Key words: natural Mongolian oak forest two-stage estimate recruitment probability model recruitment trees model |
