Global multi-resolution forest aboveground biomass reference data from field and lidar observations
2026-08-22
Forest aboveground biomass (AGB) is a fundamental variable for quantifying terrestrial carbon stocks, supporting carbon accounting, forest monitoring, and Earth system modelling. However, systematic evaluation of regional and global biomass products remains constrained by the limited availability of harmonized reference observations with documented temporal information and geographic coordinates. Here, we present a global multi-resolution reference dataset of forest AGB compiled from field observations and lidar-derived biomass datasets. Input records were standardized to common attributes, including geographic location, reference year information, source identity, and AGB expressed in Mg/ha. Records with unresolved locations, unresolved sampling years, multi-year aggregates, or harvest-derived estimates were excluded. Observations collected between 1983 and 2017 were assigned to reference years from 1985 to 2015 using symmetric 5-year temporal windows. HILDA+ annual land-use and land-cover data were used to retain observations classified as forest throughout the corresponding 5-year reference window. Outlier filtering was applied separately to field observations and each lidar-derived datasets to reduce the influence of extreme values. The retained field and lidar-derived records were independently aggregated at 0.01°, 0.05°, and 0.1° spatial resolutions. Field reference records provide mean AGB values calculated from the qualified field observations assigned to each grid cell and reference year. Lidar-derived data are released in two forms: source-specific records for individual lidar datasets and source-integrated records obtained by averaging the available source-specific aggregated estimates for the same grid cell and reference year. Source identifiers are retained in all released records. The dataset is intended for evaluating gridded biomass products and AGB models across multiple spatial resolutions and reference years.