Edge effects on tree architecture exacerbate biomass loss of fragmented Amazonian forests

The three-dimensional structure of trees, or tree architecture, reflects the allocation of photosynthetically fixed carbon within the plants. Tree architecture is a by-product of environmental pressures on plant growth, reproduction, and survival. Fine adjustments in aboveground architecture minimize competition, improve hydraulic conductance, limit transpiration, and maximize light capture. In Amazonian forests, tree size and architecture vary greatly across species due to evolutionary processes and across individuals due to acclimation and adaptation to changing environmental conditions (e.g., canopy gaps). Architectural traits control CO2 loss from respiration, hydraulic safety and efficiency, light capture, and mechanical stability, all modulating biomass allocation and carbon storage. Changes in tree architecture could reveal biome-wide impacts on carbon cycling with regional and global influences on vegetation feedbacks. Forest fragmentation affects Amazonian tree architecture. Fragment edges have greater light availability due to the mortality of large trees and lateral light penetration. This induces changes in tree architecture to optimize light capture, including higher vertical and horizontal crown growth. Higher temperatures and lower water availability in forest edges increase evaporative demand, potentially leading trees to shorten water and nutrient transport distances. High wind turbulence can kill asymmetrical trees, while symmetrical trees have greater stability but may be less efficient at light competition. High mortality of large individuals can damage neighbors, suppressing aboveground biomass allocation and impacting architecture and size. However, uncertainties remain regarding the effects of forest fragmentation on tree architecture because (i) tree architecture varies considerably across life stages, (ii) multiple architectural attributes interact, and (iii) responses to fragmentation vary within and between species. Edge effects on tree architecture could affect allometric models predicting aboveground biomass (AGB) as a function of diameter at breast height (DBH) and tree height. Long-term measurements show that forest fragments in Central Amazonia experience dramatic AGB loss due to large tree mortality, not offset by growth and recruitment. Terrestrial laser scanning (TLS) surveys can reduce uncertainties in tree volume estimates by considering tree geometry and shape, offering new perspectives into the 3D structure of trees, including fine-scale architectural traits and accurate allometry estimates. Understanding how forest edge trees adjust architecture and allometry helps predict how plants acclimate to environmental changes and their impacts on biogeochemical fluxes and the terrestrial carbon cycle.

The paper extensively cites existing literature on tree architecture, allometry, and the effects of fragmentation on Amazonian forests. Studies on the relationship between tree architecture and light capture, hydraulic efficiency, and mechanical stability are reviewed. The authors cite previous research showing the significant loss of aboveground biomass in fragmented Amazonian forests due to the mortality of large trees. They also acknowledge the limitations of traditional methods for estimating tree biomass and highlight the advantages of using terrestrial laser scanning (TLS) for obtaining more accurate and detailed three-dimensional measurements of tree structure. The literature review establishes a foundation for the study by demonstrating the existing knowledge gap concerning the effects of edge effects on tree architecture and the cascading consequences for biomass estimates in fragmented forests.

The study was conducted in Central Amazonian forests within the Biological Dynamics of Forest Fragments Project (BDFFP), a long-term experimental study of habitat fragmentation. The BDFFP sites consist of forest fragments isolated in 1980-1983, surrounded by a matrix regularly cleared to maintain isolation. The researchers investigated how tree architecture varied with distance from the forest edge in four fragments across two regions (Dimona and Colosso). Three-dimensional tree structure was assessed using a RIEGL VZ-400i terrestrial laser scanning (TLS) system. Data were acquired along transects near fragment edges and in the interior of a 100-ha fragment. Scans were co-registered to create a single point cloud per transect. Individual tree segmentation used an automated approach with manual corrections. Quantitative structure models (QSMs) were created to describe the topological, geometric, and volumetric properties of trees, providing estimates of branch size distribution, diameters, lengths, angles, and volumes. Branches with diameters <2 cm were trimmed from QSMs to reduce uncertainties. Tree height was used to classify trees as established before or after fragmentation (those above 20m and those below). Architectural traits (surface area per unit volume of branches and trunk, asymmetry, path fraction, and relative crown dimensions) were calculated from the QSMs. Allometric equations were developed to predict aboveground woody volume as a function of stem size and tree height or DBH only. These equations were used to estimate aboveground biomass (AGB) across larger spatial scales using data from 44 1-ha permanent plots in edge and interior forests. Linear mixed models were used to account for spatial variation and to analyze the effects of edge effects on architectural traits, allometric relationships, and AGB estimates. Edge effects extent was determined by testing the influence of distance from edges using mixed linear models. Finally, nonlinear mixed-effects models were used to assess the relationship between architectural traits and tree height. Statistical analyses used the ‘nlme’ R package.

The study revealed that edge effects significantly impacted tree architecture and allometry, exacerbating biomass loss in fragmented Amazonian forests. **Edge effects on architectural traits:** Edge effects influenced architectural traits differently depending on when trees were established. Surviving tall trees ( >20 m) at the edges had thinner trunks, were more symmetrical, and had a reduced path fraction compared to interior trees. Colonizing short trees (<20 m) at the edges exhibited thicker branches and trunks, were more asymmetrical, and had a higher path fraction, indicating adaptation for maximizing light capture. Most trait variability was explained by within-plot variability reflecting local edge effects, species-specific, and ontogenetic influences. **Edge effects on tree allometry:** Allometric models for predicting woody volume were affected by edge effects. For a given DBH and height, colonizing short trees at the edges had 50% more woody volume than interior trees. However, surviving tall trees at the edges, while maintaining similar woody volume to interior trees when considering both DBH and height, showed a 30% decline in woody volume when considering DBH only, primarily due to reduced height. **Edge effects on AGB estimates:** Analysis of 44 1-ha plots revealed a significant reduction in AGB of 24.7 Mg ha−1 due to edge effects (10% reduction relative to intact forests). This reduction comprised of an 18.7 Mg ha−1 decline due to edge effects on forest structure and 6.0 Mg ha−1 due to edge effects on tree allometry. Altered tree allometry alone accounted for one-third of all AGB losses.

The findings highlight the previously overlooked role of edge effects on tree allometry in exacerbating carbon losses in fragmented Amazonian forests. The increased woody volume in colonizing trees reflects their adaptation to high-light conditions but the disproportionately lower height of surviving tall trees near edges results in a net AGB loss. The study demonstrates the importance of using advanced techniques like TLS to accurately assess the three-dimensional structure of trees and to reveal subtle changes in architecture that traditional methods may miss. The contrasting responses of colonizing and surviving trees to edge effects suggests a complex interplay of factors influencing tree growth and survival in fragmented landscapes. Future research should focus on understanding the mechanisms driving architectural changes and how these changes affect long-term carbon dynamics in fragmented forests. This includes considering species-specific responses and the influence of ongoing disturbances.

This study demonstrates that edge effects on tree architecture and allometry significantly contribute to biomass loss in fragmented Amazonian forests. The 30% reduction in woody volume of large trees near edges, coupled with the structural changes in forest composition, represents an important factor impacting forest carbon storage. The use of terrestrial laser scanning revealed these previously unquantified impacts. Future studies should integrate this knowledge with ecological and biophysical models to better predict the impact of fragmentation on carbon cycling in tropical regions. Long-term monitoring of tree architecture in these fragments is crucial to understand the long-term consequences of these changes and to help guide conservation strategies.

The study focused on a specific region in Central Amazonia and may not be generalizable to all Amazonian forests. The within-plot variability in architectural traits was substantial, suggesting that factors beyond edge effects influence tree architecture. The study utilized a specific TLS system and methodology, the results may be influenced by the choice of parameters and methods used for data processing. While the BDFFP provides unique insights, it minimizes some anthropogenic influences such as illegal logging; therefore, extrapolation to more heavily impacted forest fragments may need caution. More research is necessary to understand species-specific responses and fully resolve the underlying mechanisms responsible for architectural changes.