Work

Are locally trained allometric functions of forest aboveground biomass universal across spatial scales and forest disturbance scenarios?

Year: 2025

Type: article

Abstract: Large scale AGB estimation remains highly uncertain. Given the need for multi-sensoral, multi-scale and multi-temporal analyses to grasp natural and anthropogenic changes and heterogeneity in forests, considering the spatial scales of the input datasets is of imminent importance. Our study analyses, whether locally trained forest height to above-ground biomass allometries are scale dependent and how much regional forest disturbances impacts these estimations. For this, simulations of the individual-based forest gap model FORMIND, initialized on forest inventories of tropical lowland forest close to Manaus (Brazil), were rasterized under different forest height metrics (e.g. H100, Lorey’s height). Through regression analysis, allometric parameter sets for resolution scales ranging 20 m to 200 m were acquired. Under different application scenarios, the usage of rigid off-scale allometric parameters, local forest density adaptation and off-scenario cases was tested. As expected, the application of off-scale parameter sets introduces a considerable additional error. Generally, the error increases with differing spatial scale and is also dependent on the type of forest degradation scenario. By dynamically adapting the allometric parameter sets to local conditions, the mentioned additional error was compensated. In addition, the effect of forest disturbance was tested, by applying the locally trained allometric parameter sets to differing degradation scenarios, from which vast errors arise. Concluding, applying off-scale allometric parameter sets is possible but introduces additional challenges. Our study offers a pathway for a possible mitigation of these errors, but also highlights the importance of developing local forest structure products.

Source: Ecological Modelling

DOI: 10.1016/j.ecolmodel.2025.111339

Authors: Benedikt Hartweg, Leonard Schulz, Andreas Huth, Konstantinos Papathanassiou, Lukas W. Lehnert

Institutions: Ludwig-Maximilians-Universität München, Helmholtz Centre for Environmental Research, Deutsches Zentrum für Luft- und Raumfahrt e. V. (DLR)

Language: English

Topic: Forest ecology and management

Subfield: Nature and Landscape Conservation

Field: Environmental Science

Domain: Physical Sciences

SDG: Life on land

Open Access status: hybrid

Estimated APC paid: $2,810