高级检索
    王冬至, 张冬燕, 张志东, 黄选瑞. 塞罕坝华北落叶松人工林断面积预测模型[J]. 北京林业大学学报, 2017, 39(7): 10-17. DOI: 10.13332/j.1000-1522.20170072
    引用本文: 王冬至, 张冬燕, 张志东, 黄选瑞. 塞罕坝华北落叶松人工林断面积预测模型[J]. 北京林业大学学报, 2017, 39(7): 10-17. DOI: 10.13332/j.1000-1522.20170072
    WANG Dong-zhi, ZHANG Dong-yan, ZHANG Zhi-dong, HUANG Xuan-rui. Prediction model for basal area of Larix principis-rupprechtii plantation in Saihanba of Hebei Province, northern China[J]. Journal of Beijing Forestry University, 2017, 39(7): 10-17. DOI: 10.13332/j.1000-1522.20170072
    Citation: WANG Dong-zhi, ZHANG Dong-yan, ZHANG Zhi-dong, HUANG Xuan-rui. Prediction model for basal area of Larix principis-rupprechtii plantation in Saihanba of Hebei Province, northern China[J]. Journal of Beijing Forestry University, 2017, 39(7): 10-17. DOI: 10.13332/j.1000-1522.20170072

    塞罕坝华北落叶松人工林断面积预测模型

    Prediction model for basal area of Larix principis-rupprechtii plantation in Saihanba of Hebei Province, northern China

    • 摘要: 如何实现林分水平和单木水平预测林分断面积的兼容性,并提高不同水平断面积模型的预估精度,在森林经营过程中是一个亟待解决的科学问题。本文以华北暖温带华北落叶松人工林为研究对象,运用105块连续观测的固定样地数据,首先采用Gauss-Newton算法建立林分水平和单木水平断面积生长模型;其次分别采用不同形式的逻辑斯蒂方程对单木生存概率方程进行拟合;最后将不同水平最优预测模型进行组合并建立组合方程,采用最小二乘法估计组合方程参数,以提高对不同水平断面积的预测精度。结果表明:在约束参数法中,林分密度模型、林分断面积预测模型和单木断面积模型均具有较好的预测效果,并均能解释90%以上的变异;在分解法中采用逻辑斯蒂方程来预测单木生存概率和林分密度,经检验获得的ROC曲线下面积为0.906,表明该方程可以较好地预测林木生存概率;在组合预测法中,采用不同水平的最优模型进行组合后的预测效果最佳。在预测林分密度和断面积时,组合预测方程预测精度最高,林分水平模型预测精度次之,单木水平模型预测精度最低。组合预测法能够预测不同水平下的林分密度、立木生存概率、林分断面积及单木断面积,提高了模型预测精度,为预测林分生长动态、空间结构变化及经营效果评价等提供参考依据。

       

      Abstract: It is a scientific problem to be solved urgently in the process of forest management that how to realize area compatibility between whole-stand growth model and an individual-tree model and improve the prediction accuracy. In this paper, data from 105 permanent sample plots of Larix principis-rupprechtii plantation were used to develop both whole-stand growth model and individual-tree model. In a first step, using Gauss-Newton algorithm, a whole-stand growth model and an individual-tree model were established. In a second step, the single tree survival probability equation was fitted based on the logistic equation in different forms. Finally, the best combinations obtained in each step were compared. Regarding the disaggregation of predicted stand density, the approach based on considering the intercept of the logistic function for tree survival as a specific parameter of each sample plot and optimizing its value produced the best results. The results showed that the prediction model of stand density, stand basal area and individual basal area had a good predictive effect, and can explain more than 90% of the variance in the constraint parameter methods. In the decomposition method, survival probability of single trees and stand density were predicted based on the logistic equation. The area under the ROC curve obtained by the test was 0.906, which indicated that the equation could predict the survival probability of forest trees. In combination forecasting methods, the combination forecasting method had the best effect using different levels of optimal model. When predicting the stand density and basal area, the combination forecasting equation had the highest accuracy, and the stand level model was the second, and the accuracy of the single tree level model was the lowest. The combined forecasting method can predict the stand density, tree survival, stand basal area and tree basal area. The method improves the prediction accuracy of the model, and provides a reference for the prediction of stand growth, dynamic change of spatial structure and management effect evaluation.

       

    /

    返回文章
    返回