Estimation Of Aboveground Biomass Of Natural Secondary Forest Based On Optical-ALS Variable Combination And Non-parametric Model

Forest Biomass is an important parameter to evaluate forest ecological value and carbon sequestration capacity.Accurate estimation of Aboveground Biomass(AGB)is of great significance to the study of global carbon cycle and climate change.In this study,Feature variables extracted through a combination of airborne laser scanning(ALS)and Sentinel-2A data,the best variable combination mode and estimation meth od were explored for estimating forest aboveground biomass of natural secondary forest.Based on ALS data of 2015,Sentinel-2A data,the fixed sample plots of secondary forest resources inventory of Maoershan Forest Farm of 2016,this study extracted height features from ALS data(All the LiDAR variables,AL),several vegetation indices from Sentinel-2A(All the Optical variables,AO),first,filter the variables,and then combined the two kinds of variables into new variables(COLI1 and COLI2),finally,constructed five models,including stepwise multiple linear regression(SMLR),k-nearest neighbor(K-NN),support vector regression(S VR),random forest(RF)and stack sparse auto-encoder(SSAE),of AGB for natural secondary forests using six feature combinations(AO+AL,COLI1,COLI2,COLI1+AO+AL,COLI2+AO+AL,COLI1+COLI2+AO+AL).The influence of COLIs variable and different models on the accuracy of AGB was investigated.The results show that the accuracy of each model is improved by the selection of variables.COLIs variable could efficiently improve the accuracy of AGB estimates;Comparing to the other four models,SSAE had the highest accuracy regardless of variables;The SS AE model with the combination of optical and ALS features(COLI1+COLI2+AO+AL)had the best model performance of R2=0.83;RMSE=11.06 t/ha;rRMSE=8.23%.The combined variable COLIs can effectively improve the estimation accuracy of the natural secondary forest AGB,and the deep learning model(SSAE)is superior to other prediction models in estimating the natural secondary forest AGB.This conclusion will help to further apply the deep learning model to draw a large area AGB spatial distribution map and estimate other forest parameters.In general,the SSAE model with the combination of ALS and Sentinel-2A data could estimate AGB more accurately than other models,which provides technical support for AGB estimation and carbon evaluation of natural secondary forests.