Terrestrial laser scanning-based allometric models for East-African tropical rainforest trees

An in-depth understanding of tree allometry is a crucial prerequisite to quantify the role of tropical forests in global carbon cycling. Here, we present site-specific, multi-species allometric models to quantify height, crown length, crown projection area, volume and biomass for East-African tropical rainforest trees. We used terrestrial laser scanning data to derive individual tree structures and developed quantitative structure models (QSMs) to estimate tree volume and biomass (n = 85). To account for the irregular structure of buttressed trees, we combined QSM modelling with a meshing approach for accurate three-dimensional representation of the buttressed lower stem sections. Our results show that tree volume (adjusted R² = 0.90), tree height (adjusted R 2 pseudo = 0.72), and crown projection area (adjusted R² = 0.80) can be estimated with high accuracy from tree diameters only, while crown length (adjusted R 2 pseudo = 0.35) only showed weak correlations with diameter. The presented allometric models can be readily applied to estimate tree structure and biomass for large forest inventory datasets. A comparison with other published allometric models shows that pan-tropical models tend to overestimate tree height and underestimate tree biomass for East-African tropical trees, highlighting the relevance of site-specific allometric functions for unbiased tree structure and biomass estimation. We recommended for future studies to include species-specific TLS data in building allometric models to broaden their applicability.