Search for: All records

Finding where the trees are in a city and monitoring any changes are essential for sustainable urban management. Historically, urban forests are mainly inventoried via manual processes often limited to public lands. Leveraging advances in computing, we present a novel generative artificial intelligence (AI) method along with a first-ever national-scale dataset, to automatically localize trees in cities across the nation using satellite imagery. Our monitoring approach is fully automated and can be completed for 330 U.S. cities within less than a day of computing, enabling actionable knowledge of changes in urban trees and supporting sustainable development decisions. We successfully localized and counted over 278 million trees, achieving an average tree count accuracy of 92.5% and spatial accuracy of 1.5m for 2024–2025. Our computational approach allows for novel nationwide analysis to be performed. For example, we can localize approximately 117 million trees on private lands and 161 million on public lands. Further, we show and quantify that urban tree distribution exhibits strong spatial disparity, with low-income communities having substantially fewer trees and less canopy cover than others. In addition, we compare tree count and layouts before and after multiple major events (e.g., major fires and destructive weather phenomena). Overall, our approach enhances computational urban planning, including weather and extreme event forecasting, for the development of sustainable cities. 

Accurate tree species identification through bark characteristics is essential for effective forest management, but traditionally requires extensive expertise. This study leverages artificial intelligence (AI), specifically the EfficientNet-B3 convolutional neural network, to enhance AI-based tree bark identification, focusing on northern red oak (Quercus rubra), hackberry (Celtis occidentalis), and bitternut hickory (Carya cordiformis) using the CentralBark dataset. We investigated three environmental variables—time of day (lighting conditions), bark moisture content (wet or dry), and cardinal direction of observation—to identify sources of classification inaccuracies. Results revealed that bark moisture significantly reduced accuracy by 8.19% in wet conditions (89.32% dry vs. 81.13% wet). In comparison, the time of day had a significant impact on hackberry (95.56% evening) and northern red oak (80.80% afternoon), with notable chi-squared associations (p < 0.05). Cardinal direction had minimal effect (4.72% variation). Bitternut hickory detection consistently underperformed (26.76%), highlighting morphological challenges. These findings underscore the need for targeted dataset augmentation with wet and afternoon images, alongside preprocessing techniques like illumination normalization, to improve model robustness. Enhanced AI tools will streamline forest inventories, support biodiversity monitoring, and bolster conservation in dynamic forest ecosystems. 

Abstract Structural diversity—the volume and physical arrangement of vegetation within the three‐dimensional (3D) space of ecosystems—is a predictor of ecosystem function that can be measured at large scales with remote sensing. However, the landscape composition and configuration of structural diversity across macrosystems have not been well described. Using a relatively recently developed method to quantify landscape composition and configuration of continuous habitat or terrain, we propose the application of gradient surface metrics (GSMs) to quantify landscape patterns of structural diversity and provide insights into how its spatial pattern relates to ecosystem function. We first applied an example set of GSMs that represent landscape heterogeneity, dominance, and edge density to Lidar‐derived structural diversity within 28 forested landscapes at National Ecological Observatory Network (NEON) sites. Second, we tested for forest type, geographic location, and climate drivers of macroscale variation in GSMs of structural diversity (GSM‐SD). Third, we demonstrated the utility of these metrics for understanding spatial patterns of ecosystem function in a case study with NDVI, a proxy of productivity. We found that GSM‐SD varied in landscapes within macrosystems, with forest type, geographic location, and climate being significantly related to some but not all metrics. We also found that dominance of high peaks of height and vertical complexity of canopy vegetation and the heterogeneity of the vertical complexity and coefficient of variation of canopy vegetation height within 120‐m patches were negatively correlated with NDVI across the 28 NEON sites. However, forest type always had a significant interaction term between these GSM‐SD and NDVI relationships. Our study demonstrates that GSMs are useful to describe the landscape composition and configuration of structural diversity and its relationship with productivity that warrants further consideration for spatially motivated management decisions. 

Abstract Forest canopy complexity (i.e., the three‐dimensional structure of the canopy) is often associated with increased species diversity as well as high primary productivity across natural forests. However, canopy complexity, tree diversity, and productivity are often confounded in natural forests, and the mechanisms of these relationships remain unclear. Here, we used two large tree diversity experiments in North America to assess three hypotheses: (1) increasing tree diversity leads to increased canopy complexity, (2) canopy complexity is positively related to tree productivity, and (3) the relationship between tree diversity and tree productivity is indirect and driven by the positive effects of canopy complexity. We found that increasing tree diversity from monocultures to mixtures of 12 species increases canopy complexity and productivity by up to 71% and 73%, respectively. Moreover, structural equation modeling indicates that the effects of tree diversity on productivity are indirect and mediated primarily by changes in internal canopy complexity. Ultimately, we suggest that increasing canopy complexity can be a major mechanism by which tree diversity enhances productivity in young forests. 

Abstract Invasive forest pests can affect the composition and physical structure of forest canopies that may facilitate invasion by non‐native plants. However, it remains unclear whether this process is generalizable across invasive plant species at broad spatial scales, and how other landscape characteristics may simultaneously facilitate non‐native plant invasion. Here, we assembled a dataset of over 3000 repeatedly measured forest plots and quantified the impact of emerald ash borer (EAB,Agrilus planipennis) residence time, land cover, and forest structure on the accumulation and coverage of invasive plants. We show plots in counties with longer EAB residences tended to accumulate more invasive plants than plots with shorter EAB residences. On average, nearly half of the plots with ash (Fraxinusspp.) in counties with EAB accumulated an additional 0.48 invasive plant species over the 5‐ to 6‐year resample interval compared to plots with ash in counties without EAB at the time of sampling. Increases in invasive species coverage were also evident in counties with EAB—although residence time did not have a strong effect, while forest gap fraction and vertical complexity were each negatively associated with increased coverage. This work has implications for understanding how invasive forest pests can facilitate the spread of non‐native plants. 

Abstract Decades of theory and empirical studies have demonstrated links between biodiversity and ecosystem functioning, yet the putative processes that underlie these patterns remain elusive. This is especially true for forest ecosystems, where the functional traits of plant species are challenging to quantify. We analyzed 74,563 forest inventory plots that span 35 ecoregions in the contiguous USA and found that in ~77% of the ecoregions mixed mycorrhizal plots were more productive than plots where either arbuscular or ectomycorrhizal fungal-associated tree species were dominant. Moreover, the positive effects of mixing mycorrhizal strategies on forest productivity were more pronounced at low than high tree species richness. We conclude that at low richness different mycorrhizal strategies may allow tree species to partition nutrient uptake and thus can increase community productivity, whereas at high richness other dimensions of functional diversity can enhance resource partitioning and community productivity. Our findings highlight the importance of mixed mycorrhizal strategies, in addition to that of taxonomic diversity in general, for maintaining ecosystem functioning in forests.