Mapping The Forest Canopy Height And Aboveground Biomass By Fusion Of ICESat-2 And Multi-Source Remote Sensing Data

Forest canopy height is defined as the distance between the highest point of the tree canopy and the ground,which is considered to be a key factor in calculating above-ground biomass,leaf area index,and carbon stock.The Ice,Cloud,and Land Elevation Satellite-2(ICESat-2)was launched in 2018,with the Advanced Topographic Laser Altimeter System(ATLAS)instrument taking on the task of mapping and transmitting data as a photon-counting Li DAR,which offers an opportunity to obtain global forest canopy height and other forest structure parameters.Jiangxi Province is rich in forest resources,with a stable forest cover of over 63.1%.Forest canopy height monitoring of Jiangxi Province can provide information on dynamic changes in forest resources to policymakers.In this study,the relationship between canopy height and 33 primary predictors such as forest age was first analyzed at the sample scale to evaluate the potential of forest age-years as a new predictor in predicting forest height and the Spearman correlation coefficient(r_s)was used to screen the predictors for developing the canopy height extrapolation models.Subsequently,we integrated ICESat-2 and multi-source remote sensing data,including Sentinel-1,Sentinel-2,the Shuttle Radar Topography Mission,and forest age.Meanwhile,four canopy height extrapolation models was developed by Random Forest(RF),Support Vector Machine(SVM),K-nearest neighbor(KNN),Gradient Boosting Decision Tree(GBDT)to link canopy height in ICESat-2,and spatial feature information in multi-source remote sensing data.After that,we extrapolated the canopy height from the sample point scale to the regional scale by using the preferred model and the remote sensing images corresponding to the screened predictors to achieve the 30 m resolution forest canopy height mapping in Jiangxi Province.Based on that,the above-ground biomass estimation models was developed by linear,exponential,logarithmic and power functions,combined with the measured above-ground biomass sample data of forest survey and forest canopy height.Finally,the optimal model was applied to map the spatial distribution of above-ground biomass at 30 m resolution in Jiangxi Province.Based on the above study,the following main conclusions were obtained:(1)Of the 33 initial predictors,21 predictors with r_s greater than 0.2 were used to develop the canopy height extrapolation models.Among them:elevation(r_s=0.62)and slope(r_s=0.55)were significantly correlated with canopy height;forest age and forest canopy height were moderately correlated at the sample scale(r_s=0.42)which implies that it contributes to prediction accuracy as a predictor;and the predictors extracted by Sentinel-1 were the least correlated(r_s from 0.13 to 0.32)(2)The coefficient of determination(R~2)of four canopy height extrapolation models ranged from 0.47 to 0.61,and the root mean square error(RMSE)ranged from5.29 to 6.18 m.The best-fitting model was RF base on the R~2=0.61 and RMSE=5.29m.In addition,RF significantly depended on elevation,slope,S2＿BAND5,forest age was a moderately RF-dependent variable,and RF was the least sensitive to texture features extracted from Sentinel-1.(3)The heights predicted by RF in the Jiangxi Province ranged from 4.88 to35.55 m,with an average height of 23.8 m.In general,the forest canopy height was higher in the mountains and lower in the plains and hills.The validation reveals that the canopy height extrapolation model we created could effectively map the forest canopy height in the research area(R~2=0.69;RMSE=4.02 m).(4)Among the four above-ground biomass estimation models for Jiangxi Province,the linear function estimation model with the highest accuracy(R~2=0.616)was used to estimate above-ground biomass in Jiangxi Province.The results indicated that the above-ground biomass in Jiangxi Province ranged from 30.59 to 149.08 t/ha.Finally,comparison of above-ground biomass obtained from forest surveys with predicted above-ground biomass shows that above-ground biomass information can be effectively estimated by forest canopy height(R2=0.6;RMSE=17.15 t/ha).