Search for: All records

Forests are integral to the global land carbon sink, which has sequestered ~30% of anthropogenic carbon emissions over recent decades. The persistence of this sink depends on the balance of positive drivers that increase ecosystem carbon storage—e.g., CO₂fertilization—and negative drivers that decrease it—e.g., intensifying disturbances. The net response of forest productivity to these drivers is uncertain due to the challenge of separating their effects from background disturbance–regrowth dynamics. We fit non-linear models to US forest inventory data (113,806 plot remeasurements in non-plantation forests from ~1999 to 2020) to quantify productivity trends while accounting for stand age, tree mortality, and harvest. Productivity trends were generally positive in the eastern United States, where climate change has been mild, and negative in the western United States, where climate change has been more severe. Productivity declines in the western United States cannot be explained by increased mortality or harvest; these declines likely reflect adverse climate-change impacts on tree growth. In the eastern United States, where data were available to partition biomass change into age-dependent and age-independent components, forest maturation and increasing productivity (likely due, at least in part, to CO₂fertilization) contributed roughly equally to biomass carbon sinks. Thus, adverse effects of climate change appear to overwhelm any positive drivers in the water-limited forests of the western United States, whereas forest maturation and positive responses to age-independent drivers contribute to eastern US carbon sinks. The future land carbon balance of forests will likely depend on the geographic extent of drought and heat stress. 

Abstract When plants die, neighbours escape competition. Living conspecifics could disproportionately benefit because they are freed from negative intraspecific processes; however, if the negative effects of past conspecific neighbours persist, other species might be advantaged, and diversity might be maintained through legacy effects. We examined legacy effects in a mapped forest by modelling the survival of 37,212 trees of 23 species using four neighbourhood properties: living conspecific, living heterospecific, legacy conspecific (dead conspecifics) and legacy heterospecific densities. Legacy conspecific effects proved nearly four times stronger than living conspecific effects; changes in annual survival associated with legacy conspecific density were 1.5% greater than living conspecific effects. Over 90% of species were negatively impacted by legacy conspecific density, compared to 47% by living conspecific density. Our results emphasize that legacies of trees alter community dynamics, revealing that prior research may have underestimated the strength of density dependent interactions by not considering legacy effects. 

The ecology of forest ecosystems depends on the composition of trees. Capturing fine-grained information on individual trees at broad scales provides a unique perspective on forest ecosystems, forest restoration, and responses to disturbance. Individual tree data at wide extents promises to increase the scale of forest analysis, biogeographic research, and ecosystem monitoring without losing details on individual species composition and abundance. Computer vision using deep neural networks can convert raw sensor data into predictions of individual canopy tree species through labeled data collected by field researchers. Using over 40,000 individual tree stems as training data, we create landscape-level species predictions for over 100 million individual trees across 24 sites in the National Ecological Observatory Network (NEON). Using hierarchical multi-temporal models fine-tuned for each geographic area, we produce open-source data available as 1 km²shapefiles with individual tree species prediction, as well as crown location, crown area, and height of 81 canopy tree species. Site-specific models had an average performance of 79% accuracy covering an average of 6 species per site, ranging from 3 to 15 species per site. All predictions are openly archived and have been uploaded to Google Earth Engine to benefit the ecology community and overlay with other remote sensing assets. We outline the potential utility and limitations of these data in ecology and computer vision research, as well as strategies for improving predictions using targeted data sampling. 

Abstract Measuring forest biodiversity using terrestrial surveys is expensive and can only capture common species abundance in large heterogeneous landscapes. In contrast, combining airborne imagery with computer vision can generate individual tree data at the scales of hundreds of thousands of trees. To train computer vision models, ground‐based species labels are combined with airborne reflectance data. Due to the difficulty of finding rare species in a large landscape, many classification models only include the most abundant species, leading to biased predictions at broad scales. For example, if only common species are used to train the model, this assumes that these samples are representative across the entire landscape. Extending classification models to include rare species requires targeted data collection and algorithmic improvements to overcome large data imbalances between dominant and rare taxa. We use a targeted sampling workflow to the Ordway Swisher Biological Station within the US National Ecological Observatory Network (NEON), where traditional forestry plots had identified six canopy tree species with more than 10 individuals at the site. Combining iterative model development with rare species sampling, we extend a training dataset to include 14 species. Using a multi‐temporal hierarchical model, we demonstrate the ability to include species predicted at <1% frequency in landscape without losing performance on the dominant species. The final model has over 75% accuracy for 14 species with improved rare species classification compared to 61% accuracy of a baseline deep learning model. After filtering out dead trees, we generate landscape species maps of individual crowns for over 670 000 individual trees. We find distinct patches of forest composed of rarer species at the full‐site scale, highlighting the importance of capturing species diversity in training data. We estimate the relative abundance of 14 species within the landscape and provide three measures of uncertainty to generate a range of counts for each species. For example, we estimate that the dominant species, Pinus palustris accounts for c. 28% of predicted stems, with models predicting a range of counts between 160 000 and 210 000 individuals. These maps provide the first estimates of canopy tree diversity within a NEON site to include rare species and provide a blueprint for capturing tree diversity using airborne computer vision at broad scales. 

ABSTRACT Conspecific density dependence (CDD) in plant populations is widespread, most likely caused by local‐scale biotic interactions, and has potentially important implications for biodiversity, community composition, and ecosystem processes. However, progress in this important area of ecology has been hindered by differing viewpoints on CDD across subfields in ecology, lack of synthesis across CDD‐related frameworks, and misunderstandings about how empirical measurements of local CDD fit within the context of broader ecological theories on community assembly and diversity maintenance. Here, we propose a conceptual synthesis of local‐scale CDD and its causes, including species‐specific antagonistic and mutualistic interactions. First, we compare and clarify different uses of CDD and related concepts across subfields within ecology. We suggest the use of local stabilizing/destabilizing CDD to refer to the scenario where local conspecific density effects are more negative/positive than heterospecific effects. Second, we discuss different mechanisms for local stabilizing and destabilizing CDD, how those mechanisms are interrelated, and how they cut across several fields of study within ecology. Third, we place local stabilizing/destabilizing CDD within the context of broader ecological theories and discuss implications and challenges related to scaling up the effects of local CDD on populations, communities, and metacommunities. The ultimate goal of this synthesis is to provide a conceptual roadmap for researchers studying local CDD and its implications for population and community dynamics. 

Abstract The duration of tree seedling persistence in the understory varies greatly between forests and across environmental conditions within a forest ecosystem. To examine species‐level variation in understory persistence and passage to the sapling life stage, we followed 5236 seedlings in single‐tree canopy gaps and closed canopy conditions over three years and simulated seedling passage times and the number of seedlings required to produce one 1.5‐m tall sapling of five common tree species in a hemlock–hardwood forest of Massachusetts, USA. Averaged across species, it took 26 years in gaps and 31 years under closed canopies to go from a first‐year seedling to a 1.5‐m sapling. Across species, the average number of seedlings needed for one sapling was 294 in gaps and 2674 in closed canopy environments. We observed high interspecific variation in passage times and number required for one sapling.Betulacongeners andPinus strobustook less time and significantly fewer individuals thanAcer rubrumandTsuga canadensis, which are generally regarded as more tolerant of understory conditions. The largest intraspecific difference in gaps versus closed canopy environments was forQuercus rubra, where we estimated the number of seedlings required to produce one sapling in closed canopies to be 172 times higher than in gaps. Stem breakage also increased the number of seedlings needed per sapling, especially in closed canopy environments. We evaluated our estimates in the lab by aging cross‐sections obtained from seedlings in gap and closed canopy conditions. Compared to our empirical age‐to‐height relationships, most simulations tended to underpredict seedling age for a given height, suggesting that passage times may be even longer than our simulations indicated. Our study shows that trees can persist for decades in the seedling life stage, highlighting a need for better‐parameterized recruitment processes in demographic forecasting. 

ABSTRACT The Janzen‐Connell Hypothesis posits that plant species diversity is maintained by a reduction in seedling survival near living conspecific trees relative to heterospecifics–known as negative conspecific density dependence (CDD). CDD facilitates coexistence if stronger than heterospecific density dependence (HDD). However, whether and how long CDD persists after trees die is unknown. In a three‐year study across three forests, we monitored seedling survival near living and dead trees, both conspecific and heterospecific, across a seven‐year chrono‐sequence since tree death. CDD persisted for at least 5 years after tree death (‘legacy CDD’), and most species showed stronger CDD relative to HDD through time. We used our empirical findings to parametrize a theoretical community dynamics model. Our model suggests that both stabilising niche differences and fitness differences persist after tree death. While legacy CDD can facilitate coexistence, fitness differences often overwhelmed niche differences, making competitive exclusion the most likely outcome. 

Protecting and enhancing forest carbon sinks is considered a natural solution for mitigating climate change. However, the increasing frequency, intensity, and duration of droughts due to climate change can threaten the stability and growth of existing forest carbon sinks. Extreme droughts weaken plant hydraulic systems, can lead to tree mortality events, and may reduce forest diversity, making forests more vulnerable to subsequent forest disturbances, such as forest fires or pest infestations. Although early warning metrics (EWMs) derived using satellite remote sensing data are now being tested for predicting post-drought plant physiological stress and mortality, applications of unmanned aerial vehicles (UAVs) are yet to be explored extensively. Herein, we provide twenty-four prospective approaches classified into five categories: (i) physiological complexities, (ii) site-specific and confounding (abiotic) factors, (iii) interactions with biotic agents, (iv) forest carbon monitoring and optimization, and (v) technological and infrastructural developments, for adoption, future operationalization, and upscaling of UAV-based frameworks for EWM applications. These UAV considerations are paramount as they hold the potential to bridge the gap between field inventory and satellite remote sensing for assessing forest characteristics and their responses to drought conditions, identifying and prioritizing conservation needs of vulnerable and/or high-carbon-efficient tree species for efficient allocation of resources, and optimizing forest carbon management with climate change adaptation and mitigation practices in a timely and cost-effective manner.