Analysis of Forest Biomass Dynamics Based on Local Regression Model

This study based on the 4-period Landsat remote sensing images and the meteorological station data in Fenglin County, the global regression model (multiple linear model) and 2 local regression models (geographically weighted regression model and geographically and time weighted regression model) were used to establish the relationship between above-ground biomass of trees and remote sensing factors in the study area. The optimal model was selected to study the spatial and temporal variation of above-ground biomass in Fenglin County. The results showed that the simulation results of the 3 models were better than the global model, and the geographically temporal weighted regression model with the addition of temporal characteristics had the best fitting effect, and the model evaluation indexes were better compared with the geo-weighted regression model. The total above-ground biomass of trees in the study area was 1.63 × 10⁷, 2.05 × 10⁷, 2.32 × 10⁷, 3.37 × 10⁷ t. The average above-ground biomass of trees in the 4 periods was 54.82, 68.98, 77.87, 113.46 t/hm².The total above-ground biomass of trees in the study area showed a trend of increasing from period to period. The use of remote sensing factors to estimate the above-ground biomass in the rich forest area provides a basis for estimating the future biomass distribution in the area.