基于海拔梯度的华北落叶松林地上碳储量模型

打开文本图片集

中图分类号：S718.56 文献标志码：A 文章编号：1673-923X（2025）09-0159-11

Abstract：Objective】Elevatioalgradintisacriticalenvironmentalfactorafectingforestcabostoragedistrbutioninmountaious ecosystems，yetthenonlinearmechanismsandoptimalelevationZonesremainunclear.Thisstudyaimstorevealthedistribution paternsofforestcabostoragealonglevatioalgadientsiLrixpncipis-rupchtiiforestsmparetepredictivepfce ofdifferentmodelingapproaches，andtoexplorethapplicationpotentialofLiDARtechnologyinforestcarbostorageasessnt. 【Method】Basedonfieldsurveydatafrom35plotsandLiDARdata，thisstudyappliedlinearregression，polynomialgsion mixed-efectsmodels，andgeneralizedaditivemodels（GAM）tomodelandcomparetherelationshipbtweencarbonstorageand elevation.Leave-one-outcros-validation wasused toevaluate modelpredictiveperformance，andstandardevaluationmetrics such as the coefficient of determination （R2） ，standard error of estimation （SEE），and mean percentage error （MPE）were calculated. 【Result】ForestcbonstorageofLrixprincipis-rupprechtiiehitedasgnificantinvertedU-sapeddistrbutionpatealogthe elevational gradients（ 1900-2500m ），withanoptimal elevation zone of 2 000-2 200m and a theoretical optimal elevation of 2165m Elevation was the dominant factor explaining carbon storage variation，accounting for 89.5% ofthevariance （η2=0.895 P<0.001 ） Among the four modeling approaches，GAM performed best （fitted R2=0.907 ，cross-validation R2=0.888 ，prediction standard error= （2号 10.15t/hm2 with only 1.86% overfiting.LiDAR-derivedcanopydensity （Intensitycv） emerged asan important predictor，significantly improvingmodelpredictionaccuracy.【Conclusion】This studyconfirmedthenonlineareffectsof elevational gradienton forest carbonstorage，withGAMdemonstratingclearadvantagesinmodelingcomplexecologicalrelationships.LiDARtechnologyhows greatpotentialinforestcabonstorageassessment，providingthicalsupprtforstablsingmoutainfoestcarbonsnkoitoring systems.Theresearchfndings willprovidescientificbasis foradaptive forestmanagement under climate changeand precision forest management based on elevational gradients.

Keywords：forestcarbonstorage;elevationalgradient;generalizedaditivemodel;LiDAR;Larixprincipis-rupprechti

森林是全球碳循环的重要组成部分，面积达40.6亿 hm2 ，约占陆域的 1/3[1] 。（剩余21031字）