摘要
以油松林分为研究对象,基于北京市一类连续清查数据161块油松林分2期固定样地,建立单木水平和林分水平模型,利用组合预测法把油松单木水平模型和林分水平模型组合起来,用最优加权法计算单木水平和林分水平的权重系数,最后用3项数学指标MAD、RMSE、R2评估模型的预测效果。结果表明,所建的单木水平、林分水平和组合水平模型的预测效果均较好,经过分析比较,组合预测法预测油松断面积、蓄积量生长模型的精度高于单木水平和林分水平预测的精度。组合预测法所建立的模型结合了单木水平、林分水平模型的优点,提高了油松林分生长预测模型的兼容性,保障了油松林分生长模型预测结果的一致性,也可以用来预测下一期油松林分断面积、蓄积量生长规律。
Taking Pinus tabulaeformis as the object of research, using the periodically inventory data of 161 permanent sampling plots in Beijing, growth models based on tree-and stand-levels were established, and the two models were then combined by forecast combination. Weight coefficients of different models in the forecast combination models were calculated by optimal weight method. Finally,three mathematical indicators MAD ,RMSE and R^2 were used to evaluate and predict the performance of the models. Three kinds of the models based on tree-, stand-and combination-levels all exhibited great prediction results. After analyzing and comparing, it was found that the accuracy of combination model was much better than the other two models when they were used to predict stand basal area growth and stand volume growth. The forecast combination model combined the advantages of the other two models,it improved the compatibility of P. tabulaeformis growth model from different levels, which could predict the growth of next stage in stand basal area and stand volume.
出处
《西北林学院学报》
CSCD
北大核心
2016年第6期243-249,共7页
Journal of Northwest Forestry University
基金
国家林业公益性行业科研专项(201204510)
关键词
年龄隐含的单木生长模型
油松
林分断面积
林分蓄积量
组合预测法
age implied individual tree model
Pinus tabulaeforrnis
stand basal area
stand volume
forecast combination method
