Abstract
Improving the global monitoring of above-ground biomass (AGB) is crucial for forest management to be effective in climate mitigation. In the last decade, methods have been developed for estimating AGB from terrestrial laser scanning (TLS) data. TLS-derived AGB estimates can address current uncertainties in allometric and Earth observation (EO) methods that quantify AGB. We assembled a global dataset of TLS scanned and consecutively destructively measured trees from a variety of forest conditions and reconstruction pipelines. The dataset comprised 391 trees from 111 species with stem diameter ranging 8.5 to 180.3 cm and AGB ranging 13.5–43,950 kg. TLS-derived AGB closely agreed with destructive values (bias <1%, concordance correlation coefficient of 98%). However, we identified below-average performances for smaller trees (<1,000 kg) and conifers. In every individual study, TLS estimates of AGB were less biased and more accurate than those from allometric scaling models (ASMs), especially for larger trees (>1,000 kg). More effort should go to further understanding and constraining several TLS error sources. We currently lack an objective method of evaluating point cloud quality for tree volume reconstruction, hindering the development of reconstruction algorithms and presenting a bottleneck for tracking down the error sources identified in our synthesis. Since quantifying AGB with TLS requires only a fraction of the efforts as compared to destructive harvesting, TLS-calibrated ASMs can become a powerful tool in AGB upscaling. TLS will be critical for calibrating/validating scheduled and launched remote sensing initiatives aiming at global AGB mapping.