Remote Sensing Estimation of Bamboo Forest Aboveground Biomass Based on Geographically Weighted Regression

Bamboo forests are widespread in subtropical areas and well known for their rapid growth great carbon sequestration ability. To recognize the potential roles functions of bamboo regional ecosystems, forest aboveground biomass (AGB)—which is closely related to productivity, cycle, and, particular, sinks ecosystems—is calculated applied as an indicator. Among existing studies considering AGB estimation, linear or nonlinear regression models most frequently used; however, these methods do not take influence spatial heterogeneity into consideration. A geographically weighted (GWR) model, a local can solve this problem certain extent. Based on Landsat 8 OLI images, we use Random Forest (RF) method screen six variables, including TM457, TM543, B7, NDWI, NDVI, W7B6VAR. Then, build GWR model estimate AGB, results compared with those cokriging (COK) orthogonal least squares (OLS) models. The show following: (1) had high precision strong prediction accuracy (R2) was 0.74, 9%, 16% higher than COK OLS models, respectively, while error (RMSE) 7% 12% lower errors respectively. (2) estimated by Zhejiang Province relatively dense distribution northwestern, southwestern, northeastern areas. This line actual Province, indicating practical value our study. (3) optimal bandwidth 156 m. By calculating variable parameters at different positions bandwidth, close attention given variation law estimation order reduce error.