Abstract Multi-scan terrestrial laser scanning (TLS) is the state-of-the-art technique for precisely characterizing the 3D structure of trees. In the acquired point clouds, wind effects such as branch duplication can significantly impact data quality and can lead to errors in derived estimates such as wood volume. As it is difficult to fully avoid wind during data acquisition, those effects are omnipresent in TLS data, widely known but still an unsolved issue. A method that can reliably correct wind effects would be of enormous value to the scientific community. We present a wind correction approach that is based on local non-rigid registration of scans. For this, we compare one deep learning method with two geometric methods. Training and evaluation of these methods requires pointwise motion vectors and windless representations of the trees. Since this reference can hardly be obtained for real-world datasets, we employ virtual laser scanning of swaying trees to generate simulated multi-scan TLS data and error-free reference data. Our comprehensive assessment using motion vectors, point cloud distances, and visualizations shows that wind correction successfully improves point cloud representation. Trained from scratch on domain-specific tree data, the deep learning method outperforms non-learning baselines in terms of non-rigid registration accuracy on simulated data and better preserves points in occluded areas. Applied to a small real-world tree dataset, both the deep learning method and the geometric methods achieve an increase in scan overlap ratio from 60% to over 79% and a decrease in chamfer distance between scans from 2.86 cm to under 1.72 cm, matching the values of a low-wind reference dataset. We believe that our approach presents a valuable preprocessing step to reduce errors in downstream applications such as quantitative structure modelling or change analysis without the need to discard wind-affected scans or to adapt algorithm settings.
Abstract Deep learning-based tree species classification models applied to terrestrial laser scanning data achieve high accuracies, yet their classification decisions remain poorly understood. Addressing this gap, we present a novel explanation framework that systematically links salient regions of Finer-CAM saliency maps—which highlight image features contributing to the classification of a target species while suppressing features shared with similar species—to structural tree features in 2D side-view images of terrestrial laser scanning (TLS) point clouds. We applied this framework using YOLOv8-based tree species classification models trained on a subset of the FOR-species20K benchmark data set. Faithfulness analysis confirmed that Finer-CAM reliably identifies the most discriminative image regions for model predictions, validating its suitability for enhancing model interpretability. Analysis of 630 saliency maps indicate that the models primarily rely on image regions associated with tree crowns across most species, particularly for Silver Birch, European Beech, English Oak, and Norway Spruce, while stem regions contribute more strongly to the differentiation of European Ash, Scots Pine, and Douglas-fir. Image perturbation experiments further demonstrated that the visibility of tree structure—such as branching patterns—in the 2D side-view images enhances classification performance, indicating YOLOv8’s capabilities to leverage detailed structural information encoded in high-density TLS point clouds. Using a cross-validation approach our trained models achieved a mean overall accuracy of 96% (SD = 0.24%) on a test data set unseen during training. Beyond assessing the contribution of image regions associated with structural tree features to tree species differentiation, our methodological framework can facilitate a better understanding of model limitations, the detection of data set biases and shortcut learning, and contributes in building confidence in deep learning-based tree species classifiers.
Abstract European beech (Fagus sylvatica L.), one of the most important deciduous timber species in Central Europe, experienced widespread crown defoliation and tree mortality during the severe drought in 2018/19, but responses varied strongly among individuals. Canopy gaps were found to be one driver of this variability. In this study, we investigated how the size, orientation, and the temporal dynamics of canopy gaps and open areas affect crown defoliation in European beech during the 2018/19 drought across 19 beech-dominated sites in Bavaria, Germany. To quantify canopy gap dynamics around individual beech target trees, we derived gap area metrics from multitemporal digital surface models derived from orthophotos across three timesteps. For each target tree, total canopy gap area was calculated within a 15 m buffer. To capture directional effects, the gap area was additionally partitioned into eight cardinal and intercardinal directions, allowing us to characterize the spatial configuration of gaps relative to each tree. Our results indicate that trees exposed to gaps on their western side (meaning the tree was located at the eastern edge of the gap) exhibited significantly higher defoliation, highlighting possible combined influences of higher exposure to wind and solar radiation. In contrast, gaps towards the north of the trees were associated with lower defoliation. Increases in gap area between 2013/14 and 2019 towards the southwest were linked to higher drought stress. While the gap-related predictors explained only a modest proportion of the total variability, site-level differences in general conditions accounted for a substantially larger portion. Our findings demonstrate the potential of digital surface models from orthophotos as a reproducible, remote sensing-based workflow for capturing detailed spatial and temporal canopy gap dynamics across forest sites. For practical management, the directional orientation of canopy openings should be considered when planning silvicultural interventions, as western and southwestern exposures may increase drought vulnerability of valuable target trees.
Abstract Extraction system selection is a key component of sustainable forest operations, as it directly affects productivity, environmental impact, and worker safety. While geographic information system (GIS)-based decision support systems (DSSs) have improved the transparency and consistency of extraction system selection, less is known about how stakeholder preferences affect DSS outputs. This study investigates how stakeholder-specific preferences influence extraction system allocation within a GIS-based DSS applied across three Mediterranean forest areas in central and southern Italy. The DSS uses six criteria: terrain slope, roughness, extraction distance, road density, soil bearing capacity, and timber amount. It was applied repeatedly with analytic hierarchy process weights from four stakeholder groups: researchers, forest owners, technicians, and forest workers. Differences in system allocation were analysed using Stuart–Maxwell tests and agreement statistics, while generalized linear mixed models were applied to assess the influence of stakeholder role and operational constraints. Variance partitioning was used to quantify the relative contribution of stakeholder identity and spatial factors to decision variability. Extraction system allocation differed significantly among stakeholder groups, with operational stakeholders being more likely to select ground-based systems (odds ratio = 3.18, P < .001). However, variance partitioning revealed that most of the variability in system selection was explained by parcel-level (53.9%) and study-area (44.0%) factors, whereas stakeholder identity accounted for only 2.2% of the total variance. No significant relationship was found between operational constraints and stakeholder agreement (P = .715), suggesting that, within the adopted DSS framework, increasing operational difficulty was not associated with a clear convergence among stakeholder groups. Spatial analysis showed that stakeholder disagreement was concentrated in areas where multiple extraction systems exhibited comparable suitability. These findings suggest that extraction system selection is primarily driven by environmental and logistical constraints, with stakeholder preferences playing a secondary role. GIS-based DSS can therefore provide robust and consistent recommendations, while stakeholder-specific analyses remain valuable for identifying areas of uncertainty and supporting participatory decision-making. The results contribute to a better understanding of the interaction between objective constraints and subjective preferences in forest operations planning.
Abstract Forest biometrics has evolved from a measurement-driven discipline focused on field efficiency and statistical rigor to a data-rich, technology-enabled science integrating multisensor information and advanced modeling approaches. This special issue, inspired by the Second North American Forest Mensurationists Conference held in 2022, highlights this transformation through nine studies that collectively span scales from individual branches to regional forest dynamics. Together, they emphasize a shift from identifying single optimal models to developing integrated, uncertainty-aware model systems that support operational decision-making. At the finest scale, advances in terrestrial laser scanning enable improved characterization of branch geometry under challenging conditions, yielding robust taper and form factor estimates for volume. At the tree level, extensive benchmarking of height–diameter relationships demonstrates that model form and stand origin strongly influence predictive performance, with generalized additive models often outperforming traditional approaches. Complementary work shows that calibration strategies are not universally transferable across model forms, underscoring the need for careful alignment of function choice and calibration design. Addressing biases in young stands, Bayesian model averaging offers a practical interim solution where traditional volume models trained on mature cohorts fail. At broader scales, studies demonstrate the operational potential of integrating public and low-cost remote sensing data. Freely available USGS 3DEP LiDAR supports highly accurate dominant height and site index estimation, while bias-corrected digital aerial photogrammetry provides a viable alternative in areas lacking LiDAR coverage. Landscape-level analyses using Landsat time series and permanent plots enable mapping of basal area growth, revealing spatial variability and temporal trends driven largely by stand dynamics. Collectively, these studies define a cohesive framework for modern forest biometrics: combining multiple data sources, selecting model families deliberately, applying light but effective calibration, and explicitly quantifying uncertainty. This integrated approach supports scalable, reliable predictions tailored to the needs of forest managers and policymakers. The special issue thus outlines a forward-looking research agenda that prioritizes resilient modeling systems over isolated solutions, enabling forestry to meet contemporary challenges across scales from tree components to landscapes.
Abstract Effective forest health monitoring requires timely, accurate detection of crown dieback and dead trees. Artificial intelligence and particularly deep learning (DL) approaches integrated with high-resolution RGB aerial imagery overcome past remote sensing limitations, related to model transferability and spatial scale by enabling automated, tree-level monitoring across large forest landscapes. This study evaluated four DL models (YOLO11s-SAM, YOLO11s-seg, DeepLabV3+, and SegFormer) for mapping dieback and dead tree crowns using imagery from five ecologically distinct forest sites in Iran, Germany, Luxembourg, and the USA. Two training strategies were evaluated: models trained separately on individual sites (site-specific training) and models trained on combined data from multiple sites (integrated training), with one site held out as an independent test set to assess transferability. In site-specific experiments, YOLO11s-SAM achieved the highest point-based (verifying the correct localization of dead and dieback trees) accuracy (94.81%–98.51%) and the best polygon-based performance (verifying the correct delineation of the affected areas using the metric segments shape similarity) in three of four sites, while SegFormer led in the Luxembourg Forest. Under integrated training, SegFormer achieved higher point-based accuracy (95.55%–97.77%) and polygon-based accuracy (90.50%–94.54%) in most sites, whereas YOLO11s-SAM remained superior only in the Alang Darreh Forest. Both models generalized well to the independent test site, with YOLO11s-SAM and SegFormer achieving point-based accuracies of 94.66% and 93.33%, and polygon-based accuracies of 87.38% and 86.46%, respectively. YOLO11s-SAM produced the most segments shape similarity but tended to underestimate crown extent, whereas SegFormer consistently overestimated it. These findings highlight the robustness and generalizability of YOLO11s-SAM and SegFormer DL models for forest health monitoring. They enable a fully automated, spatially transferable workflow using widely available RGB aerial imagery, demonstrating strong potential for developing a global foundational model to accurately detect canopy dieback and dead tree at scale.
Accurate and efficient assessment of forest structure is crucial for both ecological research and effective forest management. This paper introduces DendRobot, an innovative software pipeline developed to automate the inventory of forest sample plots or entire forest stands using terrestrial LiDAR scans or ground-based photogrammetric point clouds. DendRobot incorporates a novel 2D density-based tree-detection algorithm (Detection Rate = 93%) alongside a new vertical clustering approach for estimating tree height. Both methods are implemented together with established and widely trusted methods to process three-dimensional data into GIS layers. By leveraging these algorithms, DendRobot derives key forest inventory metrics of individual trees, including diameter at breast height (Mean Absolute Error = 3.4 cm), tree height (Mean Absolute Error = 0.7 m), tree locations, and crown projection areas at a fine spatial scale with the resolution of individual trees. Additionally, it produces Digital Terrain Models (DTMs), Digital Surface Models, and Canopy Height Models (CHMs) with user-defined resolution, supporting advanced spatial analyses of forest environments and providing information for forest management planning. Optionally, these data can be enriched with individual-tree point clouds, segmented by a novel approach. Designed as a comprehensive tool for forest researchers, managers, and students, DendRobot supports efficient, data-driven decision-making with minimal manual intervention. Initial tests conducted in complex forest environments demonstrate its capacity to streamline workflows and generate forest-stand-scale inventory data with accuracy comparable to state-of-the-art methods and software. DendRobot (available at https://www.dendrobot.czu.cz/) is a user-friendly, free and open-source solution for the practical application of terrestrial LiDAR scanning in real-world forestry challenges.
Reliable forest biomass assessments are becoming increasingly important, as Parties to the Climate Convention are required to report changes in multiple carbon pools, including both above- and belowground biomass. In some regions, use of remote sensing is the only viable option for obtaining such estimates, whereas in other regions it bears potential to improve the accuracy of ground inventory-based biomass estimates. However, statistically rigorous estimation through remote sensing poses several challenges. This study systematically and comprehensively reviews the methodological quality of large-area biomass assessment studies from 1992 to 2022, based on core survey elements for successful biomass surveying assisted by remote sensing. For each element, we reviewed the studies in relation to “ideal standards” derived from the literature, which served as evaluation criteria. Our review revealed an increasing trend in use of remote sensing for biomass surveys, coupled with gradual improvements in methodological quality for all survey elements evaluated. For example, advances in remote sensing techniques, particularly the increased use of Light Detection and Ranging, Radio Detection and Ranging, and digital aerial photogrammetry, all technologies able to capture information on forest structure, have enhanced the reliability of biomass estimates. However, several problems remain, such as field data scarcity for model calibration, signal saturation in high-biomass regions, and misconceptions about the use of statistical methods. We identified five remaining key challenges for improving remote sensing assisted large-area biomass assessments. These include (i) obtaining sensor data that correlate stronger with biomass, (ii) acquiring larger sets of harmonized field data at the level of trees and plots for calibrating models, (iii) adequate use of statistical principles, (iv) developing methods for domain estimation, and (v) improved quality assurance and quality control. While upcoming new airborne technologies and satellite missions may mitigate some challenges, continued methodological innovation and further enhancement of the rigor of statistical and other procedures will remain essential for advancing remote sensing-based biomass assessments.
The forest products supply chain (FPSC) is a complex distributed network that transforms raw forest resources into finished goods. It faces inherent complexities because of factors like divergent processes, coordination of independent business units, volatile markets, logistical challenges, and resource constraints. As supply chains across industries become more data driven, artificial intelligence (AI) has emerged as a powerful tool for optimizing supply chain operations. However, there has been limited research that systematically investigates the usage of such technologies in the FPSC. Here, we used a combination of a systematic literature review and a hermeneutic approach to examine the existing implementations and recent advancements of AI applications in the FPSC, and discuss key research challenges and future opportunities for AI adoption. It was found that a wide range of AI-based applications and algorithms were developed for specific purposes along the FPSC. For example, reinforcement learning was found to be especially suitable for spatial planning while convolutional neural networks were favoured for species classification and quality assurance from image data. Using a framework developed for this review, we highlight underexplored domains and open challenges which relate to fibre supply, forest operations, log storage, and transportation. AI methodologies are still rarely applied for tasks like harvest block allocation, inventory policy, and forest road layout design. For these underexplored domains, we suggest methodological solutions adopted from broader supply chain research which we assume to have high transferability potential to the FPSC. With this review, we aim on guiding stakeholders in leveraging AI for enhanced operational efficiency and informed decision-making.
Remote sensing-based forest inventories create statistical relationships between 2D or 3D remote sensing data and field plots to predict forest attributes for grid-based population units that cover the entire area. In addition to conventional airborne laser scanning (ALS) feature-based models, deep learning alternatives have been extensively explored. However, most of the studies to date have only investigated performance in experimental setups where the training and testing data units have the same geometry (typically a circle). In practical applications, predictions are required for grid-based population units, which are typically square grid cells. While the conventional approaches that use lidar metrics appear to be largely unaffected by the geometric mismatch between the training and population units, this may not hold for the deep learning alternatives. Therefore, this study investigated the sensitivity of three-dimensional convolutional neural networks (3D CNNs) to discrepancies in shape and size between training and population units. We compared the predictions errors associated with 3D CNN architectures [fully convolutional (FCN) and Inception-V3] and Gaussian Process Regression (GPR), serving as a reference approach from conventional area-based approach, under different training (n = 376) and population unit configurations (n = 555 000). We used ALS data alone and in combination with aerial images to predict growing stock volume. Based on our results, both 3D CNN architectures were more sensitive than GPR to the discrepancy in shape and size between training and population units. For example, predictions to square population units using the FCN model trained with circular plots resulted in considerable underestimations (%RMSE: 49.8, %MD: -40.9). The utilization of either hexagonal or circular areas of remote sensing data in both training and population units resulted in the smallest prediction biases. Furthermore, we found that additional information on aerial images was beneficial with all experimented prediction methods. These results indicate that future work applying deep learning to model forest attributes with ALS data must consider shape discrepancies between training and population units.
Large-scale outbreaks of the European spruce bark beetle (SBB, Ips typographus L.) have severely disturbed Norway spruce-dominated forests across Northern Europe in recent decades. In Finland, SBB outbreaks and the subsequent mortality of Norway spruce have increased over the last two decades, particularly raising concerns after the outbreaks in 2012-2013. In this study, we developed generalized additive models to assess how various environmental drivers (stand structure, landscape, topography, weather and their interactions) and spatio-temporal drivers contribute to the variability in the SBB-induced mortality of Norway spruce during an SBB outbreak in south-eastern Finland. The study data covered managed areas where salvage-sanitation loggings and clear-cuttings were implemented, as well as conserved areas where wind-damaged trees were left unsalvaged. Among the grouped environmental drivers, the stand variables had the highest impact on SBB-induced spruce mortality, followed by the landscape and topography variables. The SBB-induced spruce mortality was higher in forest stands, which had a high proportion (> 65%) of spruce with a diameter larger than 20 cm, the mean stand diameter larger than 25 cm, were growing on east-facing slopes, were situated at less than 200 m from the managed or unsalvaged gaps, and had growing season temperature sum of more than 1300 degree days. The SBB-induced spruce mortality was spatially clustered in the study locations and forest compartments. Our findings improve understanding of the contribution of environmental and spatio-temporal drivers to the SBB-induced spruce mortality. They provide support for identifying high-risk areas for SBB outbreaks in the boreal region of northern Europe and highlight timely preventive risk management measures to mitigate SBB outbreaks after severe wind damage.
Accurate maps of tree species distribution are essential for forest science and management but remain difficult to generate over large regions. New algorithms from the field of deep learning may have better abilities to extract species-specific signatures from complex time-series signals of satellite imagery. Here, we propose a transformer-encoder model, derived from SITS-BERT that integrates Sentinel-2 time-series with forest inventory data to classify dominant tree species in Baden-W & uuml;rttemberg, Germany. The model distinguishes eight classes of major tree species with an overall accuracy of 77%, based on a conservative validation approach that relies on completely independent test samples. In direct comparison, our classification clearly outperforms alternative map products. We further analyse the influence of training data quality and quantity, and particularly examine whether the canopy cover at the location of the reference sample points has an effect on the obtained results. We demonstrate that transformer architectures remain relatively robust even with limited or noisy references but at the same we show that the model improves if only samples with a high canopy cover are used during training. Our results highlight the potential of deep learning for large-scale forest species mapping.
The ecological viability of using logging residues as a postharvest silvicultural treatment remains poorly understood in tropical forests. This study evaluated the effects of residue removal on floristic composition and natural regeneration structure at different successional stages. The study was conducted in the Tapaj & oacute;s National Forest, a dense ombrophilous forest located in the Brazilian Amazon, across three timber production units selectively logged in 2021, 2017, and 2013, representing situations 2, 6, and 10 years after logging. In the initial harvesting process, commercial tree species with a diameter at breast height >= 50 cm were extracted including for example Hymenaea courbaril, Handroanthus serratifolius, Handroanthus impetiginosus, Dipteryx odorata, and Manilkara elata. The extraction followed reduced-impact logging techniques, which involved directional felling, log skidding, and log measurement. The utilization of forest logging residues consisted of removing branches and forks from these same species, using the same skidding and measurement system applied to logs. Areas with and without logging residue removal were compared using sixty-six 10 & times; 10 m plots. In these plots, natural regeneration was measured considering structural attributes, diversity, and floristic composition, categorizing individuals into three size classes (seedlings, saplings, and small trees). Multivariate analysis of variance and discriminant analysis were applied as statistical methods to detect differences between treatments. The results indicated significant variations among the successional stages for the size classes but no significant differences between areas with and without logging residue removal. Therefore, under the conditions of this study, logging residue removal did not compromise natural regeneration, supporting its ecological viability as a postharvest silvicultural treatment in managed forests in the Amazon.
Thinning alters forest structure and functioning, yet its effects on canopy biochemistry, growth, carbon uptake and hydraulic dynamics remain poorly quantified across spatial and temporal scales. We combined Sentinel-2 PROSAIL inversion with explicit uncertainty propagation to derive monthly canopy traits in paired thinned and control stands of Pinus sylvestris and P. nigra, integrating these with high-frequency eddy-covariance GPP, maximum daily stem-shrinkage (MDS), and decadal basal area increment (BAI) from tree-ring records. Thinning reduced canopy density, pigment content, and albedo, indicating a shift in stand optical properties toward a more open but less reflective canopy structure. Functionally, thinning increased long-term basal-area increment by similar to 90% in P. sylvestris and similar to 35% in P. nigra, but reduced spring GPP by up to similar to 9 mu mol CO2 m(-2) s(-1) and intensified summer hydraulic drawdown (Delta MDS approximate to -60 mu m). Trait-function models explained 61% of BAI, 85% of GPP, and 30% of MDS variance, indicating distinct biophysical controls across response variables: leaf structural traits (especially LMA) was the strongest predictor of long-term growth, canopy architecture primarily explained seasonal productivity, and pigment-structure interactions contributed most to stem-water dynamics. This multi-scale, uncertainty-aware framework shows how integrated satellite, flux-tower and dendrochronological measurements can robustly detect and interpret thinning impacts in Mediterranean pine forests.
Headwater forests function as key hydrological regulators, enhancing soil water conservation and mitigating erosion in hilly agroforestry systems. Although the role of multi-layered forest structures in regulating the hydrological cycle is widely recognized, the specific mechanisms by which headwater forests govern soil water infiltration and replenishment in terraced landscapes remain unclear. Based on in-situ hydro-meteorological data from 2020-2021 collected in fir, bamboo, and mixed forests in the Longji Terraces of southern China, the replenishment of soil water and its seasonal dynamics under different rainfall intensities were examined. The results showed that: (i) The mixed forest maintained significantly higher soil water content (SWC) and exhibited lower annual variability compared to the bamboo and fir forests. (ii) The replenishment of soil water decreased as the antecedent SWC increased across all soil layers in the bamboo and fir forests, and the mixed forest displayed minimal interlayer differences in replenishment and the weakest sensitivity to rainfall intensity. (iii) The replenishment efficiency of soil water decreased with rainfall intensity in the fir and mixed forests, while the bamboo forest demonstrated deep activation potential during torrential rain events. These findings can help to better understand the dynamic response of soil water to rainfall in complex agroforestry ecosystems, and provide a reference for the creation of regional ecological landscapes.
Understanding the roles of trait-environment interactions in shaping subtropical seasonal tree communities is crucial for elucidating the demographic trade-off strategies employed by seedlings. So far, limited research has examined whether seedling survival is influenced by such interactions, and whether these effects are season-dependent. We assessed the effects of trait-environment interactions on seedling survival across dry and rainy seasons in a 4 ha subtropical mid-montane moist evergreen broadleaved forest. From 2020 to 2022, we monitored woody seedling survival and measured nine leaf traits for 936 individuals representing 56 species. We also characterized 15 environmental variables related to light availability, topography, soil properties, and seasonal rainfall. Using generalized linear mixed models, we modeled seedling survival as a function of initial seedling height, leaf traits, environmental factors, and their interactions and quantified variation in these effects between the dry and rainy seasons. We observed significantly positive effects of seedling height on seedling survival in both seasons. In the dry season, only leaf dry-matter content showed a strong positive association with survival, whereas in the rainy season, survival was negatively associated with both leaf nitrogen and phosphorus content. Seedling survival was positively influenced by sufficient light in the rainy season but negatively affected by higher rainfall. Topographic and soil variables showed no significant effects in either season. Trait-environment interactions influenced seedling survival in both seasons, with stronger explanatory power in rainy season models. The effects of interactions involving leaf thickness, leaf area, and specific leaf area with environmental factors (topography, soil, and rainfall) were detected across both seasons, but their strength varied seasonally. Notably, the interaction between light availability and leaf chlorophyll content affected survival only in the dry season, while interactions involving leaf nitrogen content were significant only in the rainy season. These findings demonstrate that the relationship between trait-environment interactions and seedling survival is strongly season-dependent in our study system. Our results highlight the potential of trait-based approaches to inform species selection and planting strategies in seasonal subtropical forests, with implications for improving the efficiency of forest restoration and plantation management.
Cryphalus morivorus (Coleoptera: Curculionidae: Scolytinae), a bark beetle species previously known only from East Asia, is reported here for the first time in Europe. The species was found in association with Morus alba and M. alba 'Pendula' at four municipality-level localities in the Danubian Lowland and the Lu & ccaron;enec Basin, western and southern Slovakia. Infested twigs and small branches showed wilting and dieback associated with subcortical galleries, although no cases of whole-tree mortality were observed. Identification was based on morphological traits and supported by molecular analyses of the 28S rDNA and COI genes. In addition to recently collected field material from May 2025, previously collected herbarium specimens were examined, allowing the earliest occurrence of the species in Slovakia to be dated to 2022. The pathway of introduction of C. morivorus into Europe remains uncertain but was most likely linked to the international trade of ornamental Morus plants. The presence of characteristic galleries and adult beetles across multiple sites and years indicates that the species is established and reproducing locally. This discovery highlights the increasing risk of exotic bark beetle introductions associated with ornamental and landscape tree plantings. It underscores the importance of early detection and monitoring of mulberry trees to mitigate potential long-term ecological, socio-cultural, and management-related impacts.
Anthropogenic and natural disturbances change forest structure and subsequently habitat for wildlife in boreal ecosystems. Timber harvesting and wildfire can constrain availability of forage for species that often use older seral forest, like threatened woodland caribou (Rangifer tarandus caribou), or enhance forage availability for generalists like moose (Alces americanus) and omnivorous predators like bears (black bears: Ursus americanus, grizzly bears: Ursus arctos). However, few studies have assessed long-term changes in the availability of forage following natural and anthropogenic disturbance. We used forest growth models to project future stand conditions and characterized the effects of timber harvest and wildfire on corresponding availability of forage for caribou, moose, and bears. We initialized models with field data collected from 250 harvested and 259 burned stands (0-40 years since disturbance) across boreal and montane forests in Alberta, Canada. We compared the projected amount of forage to the availability of forage observed in 256 older stands (> 40 years since disturbance) that were used by caribou. The development trajectory of forage varied among disturbance type and ecosystem subtype. Within the first 50 to 60 years post-disturbance, forage for caribou, moose, and bears experienced the greatest rate of change. In some ecosystem subtypes, the availability of terrestrial lichens, winter forage for caribou, was similar to undisturbed stands 40 years following timber harvest or wildfire. Forage for moose and bears, comprised of large deciduous shrubs, was projected to reach a greater abundance in most cutblock and wildfire sites when compared to the undisturbed sites used by caribou. Our results suggest that for the first few decades following timber harvest and wildfire, there will be more forage for moose and bears and less for caribou. Those post-disturbance increases in forage could directly or indirectly (i.e. via apparent predation) result in greater risk for caribou.
As the climate changes, coastal temperate forests in western North America are facing more rapid and extreme temperature fluctuations. This increases drought and heat stress, but also exposes trees to unexpected cold spells, potentially leading to frost damage and reduced growth and wood quality. Coastal Douglas-fir (Pseudotsuga menziesii var. menziesii) is a keystone species in western North America and is of great economical importance. Genetic selection programs have produced fast-growing seed that is widely used in reforestation. However, it is currently unknown if these fast-growing trees are more susceptible to frost, especially under different levels of competition. Frost rings, damaged rows of cells in the cambium, are a direct marker of a tree's frost susceptibility. This study uses frost ring data from five replicated realized gain trial sites in coastal British Columbia, each consisting of three levels of genetic gain for volume, which are tested at four planting densities. By characterizing the number of frost rings across all factorial combinations, we evaluated the effects of genetic selection and competition on frost ring occurrence during the first 10 years of growth. We observed large differences in frost ring occurrence across sites and planting densities, with microtopography playing a significant role at the subsite level. Moreover, we found a negative correlation between height, diameter, and frost ring occurrence, and a positive correlation between mortality and frost ring occurrence. Our results suggest that genetic selection for volume gain did not increase coastal Douglas-fir susceptibility to frost, but higher competition may, depending on site conditions and microtopography. Understanding site-specific climatic drivers of frost ring formation is critical in the site selection process prior to reforestation, to ensure optimal growth, quality, and survival of the stand.
Understanding forest development following the cessation of management remains a key challenge in forest ecology. Gap dynamics, a driver of structural and compositional change, might serve as an important indicator of progression toward old-growth conditions. This study investigates and compares the long-term gap dynamics of two Central European forest reserves (Kekes-old-growth and OserdO-long untouched) dominated by European beech (Fagus sylvatica L.) over a 42-45-year period, using aerial imagery. Our aim was to assess how historical management legacies influence gap size, formation, persistence, and closure, and to evaluate these processes relative to the forests' naturalness levels. Gap fractions ranged between 4.9%-9.9% in Kekes and 2.3%-10.7% in OserdO. These values are mostly consistent with previous reports for primary European beech forests. In Kekes, smaller gaps (<200 m(2)) predominated, while in OserdO, following a sequence of recent disturbances, gaps of 200-499 m(2) became dominant. Statistical analyses revealed a significantly steeper increase in total gap area in OserdO, reflecting greater sensitivity to exogenous disturbances, likely due to its more homogeneous stand structure and residual management effects. Site conditions, browsing pressure, and topography might have further contributed to the observed divergence. Our results indicate that the fine-scale endogenous dynamics, typical of old-growth forests (characterized by small, recurring gaps of endogenous processes), develop only several decades after the last human intervention. The study highlights that gap dynamics might serve as an indicator of forest naturalness and can guide restoration planning. Promoting structural heterogeneity and mimicking natural gap processes may accelerate recovery toward old-growth characteristics in formerly managed forests.
