基于机载激光雷达数据估计主要森林参数联立模型  

Joint Modeling for Estimating Major Forest Parameters Based on Airborne Laser Scanning Data

作　　者：邹文涛 曾伟生 孙乡楠 ZOU Wentao;ZENG Weisheng;SUN Xiangnan(Research Institute of Forestry Policy and Information,Chinese Academy of Forestry,Beijing 100091,China;Academy of Inventory and Planning of Forest and Grassland of NFGA,Beijing 100714,China)

机构地区：[1]中国林业科学研究院林业科技信息研究所,北京100091 [2]国家林业和草原局林草调查规划院,北京100714

出　　处：《中南林业调查规划》2024年第4期32-38,共7页Central South Forest Inventory and Planning

基　　金：国家自然科学基金项目(31971653);中央及公益性院所基本科研业务费专项资金项目(500103-6018)。

摘　　要：为规范主要森林参数建模方法与评价指标,利用吉林长白山林区和黑龙江伊春林区200个红松林样地的机载激光雷达数据和地面实测数据,根据8项林分因子之间的关联性,以及平均树高、断面积和蓄积量与主要激光雷达变量之间的相关性,应用误差变量联立方程组方法,建立了基于相同激光雷达变量的红松林不同林分因子联立预估模型。结果表明:所建立的红松林8项林分因子估测模型,其平均预估误差(MPE)大都在5%以内,仅每公顷株数的MPE超出了5%;平均胸径、平均树高、平均优势高预估模型的平均百分标准误差(MPSE)在15%以内,每公顷断面积、蓄积量、生物量、碳储量预估模型的MPSE在20%以内,仅每公顷株数的MPSE超出了允许误差范围;最核心的三储量(蓄积量、生物量、碳储量)模型的确定系数(R^(2))都在0.7以上。据此可以得出结论:在技术上基于机载激光雷达数据估计主要林分因子是完全可行的;所建立的红松林8项林分因子联立模型的预估误差满足森林资源调查的精度要求(除每公顷株数外),可以在实践中推广应用,且能为其他松科树种林分因子估算提供参考。In order to standardize the methods and evaluation index of major forest parameters,the airborne laser scanning(ALS)data and field-measured data from 200 sample plots of Pinus koraiensis forest stands in Changbai Mountain Area of Jilin Province and Yichun Forest Area of Heilongjiang Province were utilized.Based on the correlations among eight quantitative stand factors and the relationships between mean tree height,basal area,volume stock per hectare with primary variables of ALS data,a system of simultaneous equations with error-in-variables was employed to develop simultaneous prediction models for different stand factors of Pinus koraiensis forests using the same ALS variables.The results showed that the mean prediction errors(MPE)of eight models were mostly within 5%,except for the model for estimating stems per hectare.The mean percent standard errors(MPSE)for the prediction models of mean diameter at breast height,mean tree height,and mean dominant height were within 15%;while the MPSE for the models estimating basal area,volume stock,biomass,carbon storage per hectare were within 20%,with only the MPSE for the model of estimating stems per hectare exceeding the allowable error range.The coefficients of determination(R^(2))for the most critical three storage parameters,volume stock,biomass,carbon storage per hectare,were all above 0.7.Therefore,it can be concluded that estimating major stand factors based on ALS data is technically feasible.The simultaneous models establishedin this study for eight forest parameters of Pinus koraiensis forest stands can meet the precision requirements for forest resources inventories(except for stems per hectare)and can be promoted and applied in practice,providing a reference for the forest parameters estimation inother pine species.

关 键 词：机载激光雷达 森林参数 误差变量 联立模型 红松林 

分 类 号：S758[农业科学—森林经理学] S771.8[农业科学—林学]