Biomass inversion of typical mangrove forests in Qinzhou Bay based on LiDAR
The total mangrove biomass in the test area is critical to the conservation of the ecosystem in the study area.In combination with the measured sample plot data and LiDAR point cloud data,random forest method was employed to estimate the mangrove biomass in the study area.A biomass estimation model was then built based on the subsequent accuracy test to invert the entire mangrove biomass in the test area.The results showed that:(1)The biomass of Sonneratia apetala in the sample plots of the study area ranged from 0.55 kg·m-2 to 13.57 kg·m-2,averaging 5.40 kg·m-2.(2)The training set obtained by calculating mangrove biomass through the random forest model was R2=0.9516,RMSE=0.8142,and rRMSE=0.1486;and the test set was R2=0.6598,RMSE=2.0276,and rRMSE=0.3983.These indicated that the biomass calculated by the surface biomass estimation model was basically consistent with that calculated based on the measured sample plot data,thereby verifying the accuracy of the random forest algorithm.(3)In the study area,the total mangroves biomass was 459.18 Mg,averaging 4.15 kg·m-2.The single-tree mangroves with higher heights and larger DBHs were mainly distributed near the tidal creek as well as the central and southern parts of the study area.
国家科技期刊平台
NSTL
万方数据
