Forest Age Dataset of Afforestation Areas in 2015 and 2022

This dataset focuses on the "Two Rivers and Four Rivers" (Liangjiang Sihe) watershed of Tibet — encompassing the Yarlung Tsangpo, Nujiang, Lhasa River, Nianchu River, Shiquan River, and Yapu River basins — and provides spatially distributed stand age data for afforestation areas at two time points: 2015 and 2022. The dataset aims to characterize the spatiotemporal distribution of plantation age-class structure before and after the implementation of the "Two Rivers and Four Rivers" afforestation program, thereby providing data support for vegetation carbon stock estimation, ecosystem health assessment, and refined afforestation management. Stand age is classified into five categories following the Technical Regulations for Tibet Forest Resource Inventory: young forests (< 40 years), middle-aged forests (41–60 years), near-mature forests (61–80 years), mature forests (81–120 years), and over-mature forests (> 120 years).Data production was based on the spatial distribution of evergreen broadleaf forest (EBF), deciduous broadleaf forest (DBF), and evergreen needleleaf forest (ENF) extracted from the Global 30 m Fine Land Cover Dynamic Monitoring Product (GLC_FCS30—1985–2020). Canopy height data were sourced from two independent products corresponding to different mapping years: the China Forest Canopy Height Product at 30 m spatial resolution for 2019, released by the GUO-LAB Digital Ecology Research Group, and the Global Canopy Height Dataset (2020) jointly released by Meta and the World Resources Institute (WRI). Stand age estimation models were developed using field inventory data from the 2012 Second-Class Forest Resource Survey of Tibet Autonomous Region (248 EBF plots and 34 DBF plots) and the 2016 Continuous Forest Inventory (1,844 ENF plots), with tree height as the independent variable and stand age as the dependent variable. Four regression forms — linear, logarithmic, power, and exponential — were fitted separately for each forest type, and model performance was evaluated using the coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE). The optimal model for each forest type was then applied to estimate stand age pixel by pixel from the corresponding canopy height data, generating 30 m resolution stand age thematic maps classified according to the five-category scheme. All data processing was carried out on the Google Earth Engine platform and in ArcGIS software.