Data for: Distinguishing On-Year and Off-Year Improves Remote Sensing Estimation Accuracy of Aboveground Biomass in Moso Bamboo Forests.

This study proposes a biomass estimation method that integrates Moso bamboo's on-year/off-year growth cycle characteristics by combining Sentinel-2 remote sensing data with machine learning technology. The research first achieved high-precision identification of bamboo forest growth cycles using the Support Vector Machine (SVM) algorithm based on multispectral imagery. Subsequently, through PLSR variable importance analysis, the red-edge, near-infrared, and short-wave infrared bands were identified as key spectral features for biomass estimation, and a hybrid PLSR-CNN estimation model was developed. Comparative results between observed and predicted values demonstrate that incorporating the on-year/off-year stratification strategy significantly improves model accuracy compared to conventional non-stratified estimation models.