Forest Carbon Density Estimation Using Tree Species Diversity and Stand Spatial Structure Indices

The forest spatial structure and diversity of tree species, as the important evaluation indicators quality, are key factors affecting carbon storage. To analyze impacts biodiversity indices stand on density, five were calculated from three aspects richness, evenness, (Reineke’s density index, Hegyi’s competition index Simple mingling degree) structure. relationships between these eight explored using Structural Equation Model (SEM). Then, used characteristic variables to predict aboveground trees (abbreviated density) in sample plots National Forest Resources Continuous Inventory (NFCI) Shaoguan City 2017. Multiple Linear Regression (MLR) four typical machine learning models Random (RF), Tree-based Piecewise (M5P), Artificial Neural Network (ANN) Support Vector (SVR) density. results show that: (1) Based analysis structural equation model (SED), species have greater (2) R2 all prediction is than 0.6, among which random highest. (3) calculation optimal RF, mean city 2017 was 43.176 tC/ha. can be accurately estimated based with algorithms. Therefore, a new method for storage explored. By analyzing different scientific reference making management measures increasing sinks reducing emissions provided.