Estimation Of Belowground Biomass In Subtropical Forests Based On Ground And Remote Sensing Data

China’s subtropical forests account for about one-half of the country’s forest area.Accurately estimating the carbon budget of the subtropical region has become a major challenge for China’s ecology and earth sciences.However,at present,the estimation of subtropical forest biomass is mainly focused on aboveground biomass,and there are relatively few studies on the estimation of belowground biomass,which is one of the recognized knowledge gaps in global change research.In addition,there are large differences in the composition,structure and site conditions of subtropical forests,which leads to great uncertainty in the estimation of the spatial belowground biomass.In this study,the harvest method was first used to measure the belowground biomass of the main dominant tree species in the subtropical forest,and the spatial distribution characteristics of the belowground biomass of different dominant tree species and their relationship with the tree measurement factors(tree height,breast diameter,etc.)were analyzed.And the optimal estimation model of belowground biomass of main dominant tree species was constructed.Secondly,by integrating 613 belowground survey data of belowground biomass,climate,terrain and MODIS remote sensing data,using the Cubist machine learning method,a distribution map of subtropical forest belowground biomass with a spatial resolution of 1km was drawn to determine the impact on the main factors.The specific research results show that:(1)There are significant differences in the characteristics of the belowground biomass of the dominant tree species in subtropical forests.The belowground biomass of 11 dominant tree species accounted for 13.5%~27.0% of the total biomass.Among them,the belowground biomass of deciduous broad-leaved forest(23%)was significantly greater than that of evergreen broad-leaved(21%)forest and coniferous forest(14%);In addition,the ratio of belowground biomass to total biomass is also affected by altitude,which appears to increase with elevation.The main factors affecting the estimation of the belowground biomass of dominant tree species include forest type,tree species,longitude,above-ground biomass,breast diameter,tree height,and standing timber volume.After comparison of multiple models,it is found that the optimal model of belowground biomass suitable for the main advantages of the subtropical zone is the two-factor index model.This method can integrate factors such as latitude and longitude,altitude,breast diameter,tree height,and volume into the model,which makes it estimate accuracy is further improved.(2)The total amount of belowground biomass in subtropical forests is 149×106Mg,and the range is 0~225Mg/ha.The average belowground biomass is 50.14Mg/ha.The belowground biomass in the eastern region is significantly higher than that in the western region.Cubist model based on 17 environmental variables(nir,red,blue,green,swirl1,swirl2,thermal,NDVI,EVI,LWSI,NDWI,SAVI,precipitation,temperature,altitude,slope,and aboveground biomass)can explain 64% change of belowground biomass in subtropical forests has a root-mean-square error of 13.12Mg/ha.This result indicates that the accuracy of the belowground biomass estimated by this study is high.Among them,the most important variables for estimating belowground biomass are aboveground biomass,NDVI,SAVI,EVI,and precipitation.This study not only clarified the characteristics and influencing factors of belowground biomass of the main dominant tree species in subtropical forests,but also further integrated a large amount of ground survey data and remote sensing data of belowground biomass to accurately estimate the belowground biomass and its spatial distribution of subtropical forests.Global scale carbon stock estimation provides a reliable scientific basis and effective technical means.