Global patterns and climatic controls of forest structural complexity

Climate change is expected to significantly alter the structure and functioning of forest ecosystems globally, with potentially contrasting impacts on biodiversity and ecosystem services across different biomes. The response of forest biodiversity and ecosystem functions to climate change is intricately linked to changes in forest structural complexity. Therefore, understanding the impacts of climate change on these aspects requires a thorough understanding of the climatic controls on forest structural complexity. Climate directly influences forest compositional and functional diversity, which in turn are key determinants of structural complexity. However, the precise mechanisms by which climate shapes these relationships and translates into global patterns of forest structural complexity remain largely unknown. This knowledge gap hinders accurate predictions of how biodiversity and ecosystem functions will respond to future climate change. Forest structural complexity aims to quantify the three-dimensional distribution of trees and their canopies, going beyond simpler metrics like biomass, leaf area, or canopy height. Greater structural complexity at the stand level is characterized by higher diversity in tree sizes and crown morphologies, resulting in multi-layered canopies, denser packing, and greater interconnectivity of individual tree canopies. Essentially, it reflects the heterogeneity of biomass distribution in three-dimensional space and the efficiency of canopy space occupation. Measures of forest structural complexity have proven useful in understanding the relationships between three-dimensional forest structure, biodiversity, and ecosystem functions, serving as strong predictors of net primary productivity. The increased availability of LiDAR technology has facilitated the development of new methods and metrics for quantifying this complexity. Tree species composition, crown architecture complementarity, and tree size diversity (vertical stratification) all interact to determine the spatial patterns and efficiency of canopy space occupation, ultimately shaping forest structural complexity. Studies have shown a positive relationship between tree species diversity and structural complexity; higher diversity can lead to more efficient canopy space occupation due to the contrasting crown architectures of different species (niche complementarity). However, the co-existence and growth of diverse tree species, sizes, and morphologies depend on their physiological traits related to shade tolerance, crown plasticity, and resource acquisition under competitive stress. Therefore, forest structural complexity is also constrained by functional diversity and the range of plant functional strategies. Climate strongly influences forest compositional and functional diversity, with more humid and warmer climates supporting a broader range of plant functional strategies. This suggests a strong indirect link between climate and forest structural complexity via its influence on diversity. However, the global patterns and the direct relationships between climate and forest structural complexity remain largely uninvestigated. This study aims to address this gap by investigating the global variation and climatic drivers of forest structural complexity across biomes, mapping its global patterns, and estimating its potential responses to climate change. Focusing on primary forests, minimizing the effects of anthropogenic and natural disturbances, is crucial to isolate the climatic influences on forest structure.

The paper cites a number of studies which establish the link between forest structure and various ecosystem functions, including biodiversity and net primary productivity. The authors emphasize the previous lack of global-scale understanding of forest structural complexity and its relationship to climate. Several studies are referenced which highlight the individual roles of tree diversity, canopy structure, and physiological traits in determining forest complexity. The importance of LiDAR technology in measuring three-dimensional forest structure is also discussed. The authors reviewed existing literature on the relationships between climate variables (such as precipitation and temperature) and both species diversity and forest structural attributes, setting the stage for their own investigation into the direct link between climate and forest structural complexity. Previous research primarily focused on local or regional scales, and a comprehensive global analysis of primary forests was lacking, highlighting the novelty of this study's approach.

This study involved an extensive global field campaign across 20 primary forest sites spanning five biomes (boreal, temperate broadleaf, temperate conifer, subtropical savannas and woodlands, and tropical moist broadleaf). The selection of primary forests aimed to minimize the influence of disturbances on forest structure, allowing for a clearer assessment of the climatic influences. At each site, multiple 1-hectare plots were established, and terrestrial laser scanning (TLS) was used to acquire detailed three-dimensional data on forest structure. Five scans per plot were conducted using FARO Focus 120 or FARO M70 scanners, employing a systematic ‘five on a dice’ approach for scan placement. Data processing involved filtering the point clouds to remove erroneous points and exporting them in .xyz format for subsequent analysis. The Stand Structural Complexity Index (SSCI), a well-established metric based on the fractal dimension of cross-sectional polygons derived from the 3D point clouds, was used to quantify forest structural complexity. Canopy height, canopy openness, and basal area (as a proxy for aboveground biomass) were also measured. Climate data for the period 1971-2000 were obtained from the WorldClim2 database at a 30-arcsecond resolution, including variables such as mean annual precipitation (MAP), precipitation seasonality (coefficient of variation), mean annual temperature (MAT), solar radiation, and mean growing season temperature. Soil data, including soil water holding capacity, soil nitrogen content, and cation exchange capacity, were obtained from SoilGrids and the Regridded Harmonized World Soil Database. Linear regression and linear mixed-effects models were employed to analyze the relationships between forest structural metrics (SSCI, canopy height, canopy openness, basal area) and climatic and edaphic variables. Model selection was guided by AIC values, and model robustness was assessed through leave-one-out cross-validation and biome exclusion tests. Spatial autocorrelation in model residuals was also evaluated. Finally, using the best-performing climate-structure model, a global map of potential structural complexity (SSCIpot) was created using globally available climate data. This map represents the structural complexity that could potentially develop at a given site without anthropogenic disturbance. A systematic global sampling grid (50 km spacing) was used to predict and map SSCIpot for forest and woodland ecoregions, excluding areas outside the climatic range of the study sites. Hotspots of high potential structural complexity were identified and compared with hotspots of plant diversity based on existing global biodiversity datasets. Generalized additive models were used to evaluate latitudinal patterns in forest structural complexity.

The analysis revealed a strong correlation between forest structural complexity (quantified by SSCI) and water availability across all biomes. The best-performing model, using mean annual precipitation (MAP) and precipitation seasonality as predictors, explained 89.4% of the variation in SSCI (R² = 0.89). Other climate variables, such as temperature and solar radiation, showed weaker or less consistent relationships with SSCI. Soil water holding capacity also showed a significant positive relationship with SSCI. The model demonstrated robustness in leave-one-out cross-validation (R² = 0.86, RMSE = 0.71) and maintained explanatory power even when entire biomes were excluded (except for subtropical savannas and woodlands). There was no spatial autocorrelation detected in the model residuals. The global map of potential structural complexity (SSCIpot) showed a clear latitudinal pattern, peaking at the equator and in the mid-latitudes, and decreasing towards the poles. Hotspots of very high potential structural complexity (SSCIpot ≥ 9) were identified in several tropical and subtropical moist broadleaf forest ecoregions, including the Amazon, Borneo, Sumatra, and New Guinea. High SSCIpot values were also found in temperate rainforest ecoregions in South America, North America, and Australia. Lower values were observed in subtropical savannas and woodlands. A strong correspondence was observed between hotspots of high potential structural complexity and hotspots of plant diversity.

The strong correlation between forest structural complexity and water availability suggests that climate's influence on structural complexity is largely mediated by its effects on functional traits and structural attributes related to efficient canopy space occupation. Greater water availability appears to support higher functional diversity, leading to greater complementarity in crown architectures and more efficient packing of canopies. Additionally, water availability influences maximum tree height, contributing to vertical stratification and complexity. The model's robustness suggests that the relationship between precipitation and structural complexity is relatively consistent across different forest types and geographic regions. The global map of potential structural complexity provides a valuable benchmark for forest management, restoration, and conservation. It highlights regions with high potential for structural complexity, which often correspond with biodiversity hotspots. However, the actual structural complexity of forests is also influenced by disturbances, which were minimized in this study's sample selection. Future research should incorporate disturbance regimes and small-scale variations in soil conditions to improve model predictions and understand the dynamic interactions between climate, disturbance, and forest structure. The concordance between the patterns of potential structural complexity and plant diversity underscores the importance of considering structural complexity in conservation planning and ecosystem management.

This study demonstrates a strong relationship between forest structural complexity and water availability, quantified through a globally applicable model. The resulting global map of potential structural complexity offers a valuable benchmark for sustainable forest management, restoration efforts, and conservation strategies. Future research should focus on integrating disturbance regimes and finer-scale soil data to refine the model and further elucidate the intricate interplay between climate, disturbance, and forest structural complexity. This integrated approach is crucial for accurately predicting the responses of biodiversity and ecosystem functions to ongoing climate change.

The study's focus on primary forests, while crucial for isolating climatic effects, limits the generalizability of the findings to forests impacted by disturbances. The global map of potential structural complexity relies on climate and ecoregion data, which have limitations in spatial resolution and accuracy. Incorporating high-resolution data on disturbances and soil conditions could further improve the model's predictive power and reduce uncertainty in areas outside the studied climatic range. The study did not explicitly address the intricate feedback mechanisms between changing climatic conditions, disturbance regimes, ecosystem resilience, and forest structural complexity; further investigation is necessary in this aspect.