基于地统计学的森林地上生物量估计

摘要

Estimation of forest above-ground biomass is great significant for evaluation of the role of forests in decelerating climate change and adaptive management.Based on bureau-level permanent plot data of forest inventory in Jincang Forest Farm,Wangqing Forest Bureau,Jilin Province,using both ordinary kriging and cokriging estimate forest above-ground biomass in the study area (above-ground biomass as a main variable and basal area as a covariate in cokriging).Meanwhile,compared these two interpolation methods,the results show that ordinary kriging is not a suitable method to estimate biomass in this study area; but contrasted to ordinary kriging,cokriging can significantly improve the prediction accuracy,generating the root mean square error (RMSE) and the average standard error (ASE) decrease by 49.0％ and 39.4％ respectively,and the correlation coefficient between predicted values and measured values increases by 68.4％.The spatial distribution pattern map of forest above-ground biomass in the forest farm can be obtained by cokriging interpolation approach.Spatially,the forest above-ground biomass varies obviously in this region from 21.48 to 261.60 t·hm-2,and the average above-ground biomass is 111.9 t·hm-2.Total above-ground biomass of this region is up to 3.578 Tg.In addition,biomass significantly varies with forest types and age groups.The results provide a method and reference for the region-scale biomass estimation based on the permanent plot and geostatistics.%基于吉林省汪清林业局金苍林场二类调查的局级固定样地数据,利用地统计学的普通克里格法和协同克里格法对研究区域内森林地上生物量进行估计(协同克里格法以地上生物量为主变量,以胸高断面积为协变量),并对2种插值方法进行比较.结果表明:在研究区域内普通克里格法预测精度较低；而协同克里格法能够显著提高预测精度,所产生的均方根误差(RMSE)减少49.0％,平均标准误差(ASE)减少39.4％,预测值和实测值的相关系数提高68.4％.通过插值得到整个林场生物量的空间分布格局图,生物量的分布存在明显的空间异质性.林场平均生物量密度为111.9 t·hm-2,总生物量达3.578 Tg;并从不同森林类型和龄组对生物量的空间格局进行分析.研究结果可为区域尺度内基于固定样地和地统计学的生物量估计提供方法和参照.