Ecological Applications

Accelerating global urbanization necessitates a better understanding of how to manage cities that promote biodiversity. However, we currently lack multi-year, multi-city studies, which limits a generalizable understanding of how both within and between city differences impact the spatial and temporal dynamics of urban biodiversity. Here, we tested hypotheses about the drivers of butterfly diversity within and across urban parks by applying Bayesian occupancy models to five years of iNaturalist community science data from 2,550 parks in 22 major U.S. cities. We found that cities with bigger parks supported more species per park, including more disturbance- and edge-avoidant species. This was driven by a positive effect of park size on butterfly species colonization rates. We also found that attributes of habitat quality (plant diversity within parks and tree cover surrounding parks) contributed to butterfly species occupancy. Park connectivity increased species persistence, but the overall effects on butterfly species occupancy varied across cities. Finally, we found that the total area of tree cover throughout a city, rather than the size or connectivity of individual parks, was the primary determinant of city-wide diversity: Increasing total tree canopy cover from below-average (~6%) to above-average (~22%) increased city-wide species richness by ~10%. These findings highlight the need for cities to maintain large parks while also increasing city-wide tree cover to support biodiversity across local to regional scales. By integrating high-resolution community science data across the continental U.S., this study provides mechanistic insight into how cross-scale processes shape urban biodiversity dynamics and identifies generalizable recommendations for improving urban conservation management.

Wildlife crossing infrastructure is promoted to restore connectivity for fragmented populations, but its effectiveness at enabling natural recolonisation remains untested. We tested this using a spatially explicit agent-based model parameterised with GPS telemetry data from bobcats (Lynx rufus) in New Jersey, USA. By integrating movement behaviour, stochastic demography, habitat suitability, and traffic-dependent mortality risk, we simulated 50-year recolonisation dynamics across a highly urbanised landscape. Despite extensive unoccupied suitable habitat, natural recolonisation completely failed across all scenarios, with vehicle-induced mortality during dispersal acting as the primary limiting factor and turning the matrix into a demographic sink. Even an idealised mitigation scenario in which mortality at high-mortality crossings was reduced to zero failed to produce a self-sustaining population. Although dispersal increased, individuals at the recolonisation front remained too sparse to overcome the mate-finding Allee effect. Sensitivity analysis confirmed that the recolonisation-failure result is robust to {+/-}50% variation in per-crossing mortality and {+/-}25% variation in disperser survival. Restoring structural connectivity is not, in itself, a sufficient intervention for recovering low-density carnivore populations facing a high-mortality matrix. Instead disperser survival and local density at the recolonisation front are the rate-limiting determinants. In such systems translocation rather than crossing-structure investment is more likely to result in recolonisation success.

Urban green spaces are increasingly recognised as important refuges for biodiversity, yet their ecological value depends strongly on design and management. Here, we investigate how fine-scale structural and microhabitat components shape urban ant assemblages, using ants as indicators of broader arthropod responses to urbanisation. Ant communities were sampled in twelve urban green spaces in Cordoba (southern Spain) over a ten-year period (2004 to 2013) using pitfall traps, alongside detailed characterisation of vegetation structure and ground-layer microhabitats. In total, 38 species and 25,578 individuals were recorded. Microhabitat variables explained 58% of the variation in species occurrence. Community differences among microhabitats were driven primarily by nestedness, with dense herbaceous cover acting as a core habitat and edge-related components contributing disproportionately to beta diversity. Tree abundance showed a unimodal relationship with species richness, with maximum diversity at intermediate densities, while shrub and lawn cover had weak or inconsistent effects. Fine-scale elements such as leaf litter, stones, woody debris, and small bare-ground patches strongly influenced species occurrence by providing thermal refugia, nesting substrates, and foraging opportunities. The invasive Argentine ant (Linepithema humile) exhibited strong but spatially restricted dominance and species-specific negative effects on native ants, emphasising the role of habitat context in mediating invasion impacts. Our results demonstrate that urban biodiversity is maximised by enhancing fine-scale habitat heterogeneity rather than increasing green cover alone. We highlight practical design principles for urban green infrastructure that prioritise structural diversity and ground-layer complexity to support resilient arthropod communities.

Urbanization often reduces local biodiversity, yet cities can maintain surprisingly high total species richness at broader spatial scales. This paradox has led to the "cities-as-bioarks" hypothesis, which proposes that protected remnant habitats within cities can function as refuges for native species. However, most evidence supporting this hypothesis comes from higher latitudes, and it remains unclear whether remnant habitats in low-latitude cities can serve the same role. We surveyed ant communities across 60 sites in Atlanta, Georgia (USA), spanning streetscapes, manicured parks, and forested parks, to test whether relatively undisturbed urban forests support the highest native diversity. Contrary to the bioark prediction, native species richness was lowest in forested parks and highest in habitats with intermediate disturbance. Community composition varied with habitat structure, but surface temperature was not a significant predictor of richness or composition. Instead, native abundance and richness declined strongly with increasing abundance of the invasive Asian needle ant, Brachyponera chinensis, a forest-adapted invader capable of dominating relatively intact habitats. In contrast, two other invasive species, the red imported fire ant Solenopsis invicta and the Argentine ant Linepithema humile, were largely restricted to more disturbed habitats and had comparatively weaker associations with native diversity loss. These findings refine the bioark hypothesis by demonstrating that habitat protection alone may be insufficient to conserve insect diversity in warmer regions, and instead must be paired with active invasive species management.

Frugivorous insect larvae are dependent on fruiting plants for development, leading to complex host-parasite interactions that may be influenced by many factors at various scales. We compared the relative effects of factors at the individual, neighborhood, and landscape scales in forest patches. Our results suggest that in areas like upstate New York, where agricultural land uses are dominant, individual scale factors are the most influential. Specifically, parasitism increased with host fruit crop size, but was not associated with host species richness or proximity to forest edge. Notably, the most parasitized hosts were non-native species, including Frangula alnus Mill. (Glossy Buckthorn), indicating a potential role of invasive species to shape host-parasite interactions in our system. Our results underscore the importance of host-specific traits in structuring parasitism patterns and suggest management could consider both the ecological context of host traits and the influence of invasive species at multiple scales.

Controlling invasive species is a global conservation priority but is typically resource-limited, necessitating strategies to optimize control. Spatial prioritization methods can improve the efficiency of invader control by quantifying the benefits, costs, and risks of alternative intervention strategies. Yet prioritization designs are sensitive to the distinction between protecting native populations that are currently sympatric with invaders versus safeguarding presently allopatric native populations by preventing invader expansion. Gaining a better understanding of this distinction stands to improve our ability to prioritize the management of invaders whose distributions--and thus impacts on native species--are dynamic. We asked how prioritization designs addressing current versus future invader threats affected tradeoffs among focal species protection, biodiversity conservation, disturbance risk, and overlap with existing conservation infrastructure. As a case study, we spatially prioritized population control of invasive barred owls (Strix varia) in the northwestern US. We found that distinguishing between current versus future threats posed by barred owls to the native spotted owl (S. occidentalis) strongly mediated whether invader control stood to benefit native at-risk animal communities. Furthermore, this distinction also affected the degree to which population control would overlap with fire risk and federally protected forests, both of which plausibly affect the viability and success of conservation action. These results thus illustrate that deciding to prioritize the control of invaders based on their current versus future impacts on native species can dramatically affect the distribution and characteristics of high-priority areas for management. Our findings also directly inform control of barred owls in the northwestern US: we found that prioritizing future threats to spotted owls could protect at-risk amphibian communities from novel barred owl predation, but that high fire risk and minimal protected forest may complicate implementation. Thus, in both our system and more broadly, spatial prioritization methods are an important tool for quantitative, reproducible, and successful invader control.

O_LIWhen eradication is unfeasible, invasive predator control should evaluate how removal affects ecological responses by native species. Assessments often use total invasive predator abundance to evaluate prey responses, yet intraspecific variation in diet and space use means that some subgroups cause disproportionate impacts. Identifying these problem individuals, and the spatial scales over which their impacts operate, can enable targeted spatially explicit removal to maximise impact reduction. However, despite individual-level information is often already collected during trapping operations it is seldom included when analysing predator impacts, potentially biasing the conservation outcomes expected under blanket removal. C_LIO_LIWe use a novel framework and two decades of invasive predator control data to estimate how individual variation in residency status influences the distance-dependent impacts of invasive American mink Neogale vison on water vole Arvicola amphibius occupancy across two prey surveys. We also develop a sub-model to predict mink residency status for individuals with missing age data. C_LIO_LIThe probability of capturing adult mink decreased with elevation and years of control, indicating that long-term control altered the resident population and demographic composition of mink around water vole sites. C_LIO_LIDistance-dependent negative impacts of mink varied by residency status, becoming negligible at approximately 20 km from water vole sites for resident mink and 2 km for transient. The spatial scale of mink impacts was largest during the first vole survey when resident mink were more abundant, and declined rapidly for the second survey, when mink were less abundant and spatially clustered. Our results suggest that water voles have benefited mostly from reducing resident mink rather than the total population, especially in early control phases. C_LIO_LIManagers can use our framework to develop spatially explicit and impact-based strategies, not restricted to invasive species control, to construct empirically informed management buffers around populations of conservation concern. Long-term efforts will change the landscape and invasive predator contexts, and thus we recommend iteratively updating and re-evaluating management outcome evaluations. We argue that incorporating individual heterogeneity improves our understanding of ecological mechanisms influencing management success but that the suitability of targeted strategies should be evaluated for target socioecological contexts. C_LI

Understanding the relationships between soil macrofauna and decomposition is crucial for predicting how land-use change impacts ecosystem function in fragmented systems. This is because soil macrofauna affect decomposition and also respond to the changes in abiotic conditions across habitat gradients. This study investigates edge effects on the macrofauna contributions to decomposition across forest-field ecotones. We used bait lamina assay to quantify aboveground and belowground feeding activity of soil macrofauna in Autumn 2025 in three deciduous forest-old field ecotones and one coniferous forest-old field ecotone, in Southwestern Ontario, Canada. Vegetation diversity and composition, LAI and soil characteristics (i.e., soil organic matter, pH, temperature and moisture) were measured at each plot along the ecotone. Pitfall trap data collected in Summer 2025 at the same sites were used to characterize macrofauna communities. We used generalized linear mixed effects models to estimate the effect of distance to edge, site, and depth into the soil on bait lamina consumption and soil macrofauna, with transect nested within site as random effects. Consumption activity increased with distance into the forest from the field, with the edge representing an intermediate; and, decreased with increasing depth into the soil. In contrast, soil macrofauna abundance, especially isopods, decrease with distance into the forest from the field. These trends varied significantly across sites, so that consumption activity and abundance sometimes remained constant across the ecotone (i.e., site x distance interaction). The results demonstrate that macrofaunal contributions to bait consumption varied along the ecotone, shaped by interacting environmental gradients and shifts in community composition unique to each site.

1 - Automated in situ sensors - e.g., buried scanners - are transforming biodiversity monitoring by generating data at spatio-temporal resolutions unattainable through traditional sampling, including in cryptic environments such as soil that have remained largely inaccessible to existing methods. However, extracting ecologically meaningful information from these data streams requires substantial image processing effort that currently constitutes a critical bottleneck, particularly when the signal-to-noise ratio is low and annotated training data are scarce. 2 - Standard end-to-end deep learning detection pipelines offer unsatisfactory results due to the lack of training data and heterogeneity of the taxa of interest. We explore the potential of combining traditional computer vision algorithms with state-of-the-art deep learning models to build an efficient raw data processing pipelines from limited annotation effort. Specifically, based on the observation that the background barely changes, we focus on the differences between two consecutive images to turn the initial detection problem (with very low signal) into a simpler classification problem, which we solve by fine-tuning foundation models on limited annotated data. 3 - Our approach significantly reduces the annotation effort, allowing us to release an open dataset with about 600 soil scans and more than 8 000 labeled invertebrate occurrences across nine taxa. Using this dataset to train our models, we obtained population count estimates with relative errors ranging from 10% to 61% across taxa over a three-month period. Ecological validation through a land-use stability analysis showed full directional congruence between automated and expert-annotated classifications across all nine taxa examined, with effect-size discrepancies proportional to per-taxon classification accuracy. 4 - These results demonstrate that combining domain-specific heuristics with fine-tuned foundation models provides an effective and data-efficient strategy for automating ecological image processing workflows in low-signal, data-scarce contexts. The validated pipeline removes the manual annotation bottleneck that has historically limited scanner-based soil monitoring to short observational windows and restricted taxonomic scope, opening the way for continuous, large-scale tracking of soil invertebrate community dynamics at resolutions previously unachievable.

Effective conservation planning increasingly relies on species distribution models (SDMs) to guide where actions deliver the greatest biodiversity benefits through spatial conservation prioritization. However, SDMs are inherently uncertain, and this uncertainty propagates through prioritization processes, affecting the identification of priority areas and influencing conservation decisions. Here, we evaluate whether correcting SDM overprediction reduces uncertainty propagation into spatial conservation prioritization. Using two large European datasets of vertebrates and invertebrates, we compared unconstrained SDMs with models corrected for overprediction through a Bayesian integration of occurrences, expert range maps, and habitat suitability. We found that overprediction correction reduced spatial and performance uncertainty, with uncertainty strongly structured by model and algorithm choice and amplified when overprediction was not corrected. Although no single modelling adjustment fully eliminates uncertainty propagation from SDMs into prioritization, we demonstrate that overprediction correction consistently reduces it across datasets, taxa, and modelling approaches, highlighting its importance for robust conservation planning.

Land-use and land-cover (LULC) change is a dominant driver of biodiversity loss, yet its long-term role in restructuring species distribution remains poorly understood due to limited historical baselines and the scarcity of long-term analyses across multiple ecological transition zones. Here, we investigate nearly two centuries of land-use change across a biogeographically diverse region to evaluate how landscape transformation has reshaped vertebrate distribution range dynamics, and whether ecoregional context, habitat specialization, and taxonomic identity better explain these patterns than recent landscape change alone. Using Texas, a biogeographic crossroads with diverse ecoregional zones and a long history of intensive land-use, we quantify distributional changes in 100 native vertebrate species (mammals, birds, reptiles, and amphibians), using Vernon Baileys late nineteenth-century surveys as a historical benchmark. We integrated multi-temporal LULC datasets with spatial and multivariate analyses to assess how habitat change, species traits, and regional context relate to patterns of range expansion, contraction, and redistribution. We show that Texas has undergone non-linear landscape transformations characterized by early agricultural expansion followed by cropland abandonment, shrubland encroachment, and rapid urbanization. Range contractions were most pronounced among amphibians, reptiles, and habitat specialists, whereas generalist species and many birds exhibited greater range stability or expansion. Across taxa, primary habitat, vertebrate class, and ecoregional context emerged as the strongest predictors of distributional change, underscoring the importance of historical and environmental context over simple measures of habitat loss and fragmentation. These results indicate that while long-term LULC change establishes the underlying landscape template, species responses are structured by ecological context, including habitat affinity, taxonomic traits, and regional environmental gradients. By linking historical landscape trajectories to contemporary biodiversity patterns, this study provides a transferable framework for investigating how land-use legacies influence species distributions and community reassembly in heterogeneous human-dominated landscapes, highlighting the importance of adaptive capacity and connectivity for conservation planning.

Statewide tracking of urban tree canopy change is essential for evaluating progress toward policy targets, but detecting real change requires both high-resolution mapping and rigorous uncertainty estimation. We produced a four-year canopy cover time series for all California census-designated places using 60-cm NAIP aerial imagery and a U-Net deep learning model trained with semi-automated LiDAR-derived labels and manually annotated tiles. Canopy cover and change were estimated using stratified, error-adjusted area estimation, enabling comparisons across years. Statewide canopy cover showed a modest negative trend from 2016 to 2022 (Sens slope: -0.60% per year), but confidence intervals included zero across all groups and climate zones, indicating that trends were not statistically distinguishable from no change. Urban canopy cover was consistently lower than non-urban canopy by approximately six percentage points, and canopy cover was highest in the Northern California Coast and lowest in the Southwest Desert. Residential parcels accounted for 55-56% of canopy within incorporated urban areas across all years, indicating that statewide canopy increase goals will require engagement with private landowners. Error adjustment substantially altered canopy estimates relative to raw pixel-count totals, with direct implications for AB 2251 canopy tracking where baselines and targets drawn from unadjusted maps may not reflect true canopy extent. This open-source workflow is transferable to future NAIP acquisition years and other U.S. states, providing a scalable framework for long-term urban forest monitoring.

Species distribution models (SDMs) increasingly inform conservation and biosecurity decisions in freshwater systems, where the reliability of its uncertainty estimates matters as much as its point predictions. Ensemble SDMs derive prediction intervals from across-replicate variance, but this variance captures systematic error only when replicates disagree about it, an assumption that fails when training data are contaminated with low-accuracy records, the norm in citizen-science datasets. Whether this failure is spatially uniform or concentrates in identifiable parts of a range is unknown. Using a panel of European freshwater crayfish spanning native headwater-associated species and invasive lowland colonizers, we show that contamination-induced calibration failure is strongly spatially structured: it concentrates at stream-network headwaters, the topological tops of the network, where upstream-aggregated predictors are structurally undefined, and scales with contamination severity, replicated across four species and both dominant ensemble protocols (replicate and consensus). The failure is driven by upward prediction bias, not by intervals failing to widen: contaminated ensembles overpredict suitability in headwaters, and because the bias is shared across ensemble members, the intervals do not flag it. This is a conservation-relevant blind spot, because headwaters are both refugia for threatened native crayfish and front lines for invasion; an SDM that silently overpredicts suitability there misdirects survey and management effort toward the segments where its predictions are least trustworthy. Standard leave-one-basin-out conformal calibration, the recommended panel-wide remedy, repairs marginal coverage but leaves headwaters undercovered, because a single calibration threshold is dominated by the abundant non-headwater segments. A group-conditional (Mondrian) variant, calibrating the two populations separately, restores reliable coverage in both at no extra cost and reallocates width where it is needed. We recommend network-position-stratified calibration as a default for ensemble SDMs in dendritic freshwater systems.

Models for estimating animal density from camera traps require four parameters informing detection: movement speed, daily activity level, staying time (duration animals remain within the detection zone), and effective detection distance. These parameters traditionally come from labor-intensive manual measurements and auxiliary telemetry. Recent advances in computer vision can provide the positions of animals in camera trap images, which have been used for distance sampling. We extend this approach to extract all four parameters from imagery, providing the first AI-derived estimates of movement speed and staying time from automated coordinate tracking. We also introduce a new joint multi-species hierarchical distance function that estimates deployment-specific effective detection distances while borrowing strength across species through partial pooling. Our pipeline integrates MegaDetector for animal detection, the Segment Anything Model for segmentation, and Dense Prediction Transformers for monocular depth estimation. From frame-level coordinates, we reconstruct movement trajectories across burst sequences to estimate speed with size-biased distribution corrections, calculate staying time through bounding box interpolation, and estimate activity levels from detection timestamps. The joint hierarchical distance function decomposes the detection scale parameter into a shared deployment-level effect and species-specific offsets, so species effects represent deviations from the multi-species average, allowing data-rich species to inform detection conditions where rare species have few observations. AI-derived scene depth enters the model as a covariate on detection range, providing a vegetation openness metric from the same pipeline. To address position errors from depth estimation, we apply data quality filters. We processed 122,574 frames from 181 deployments across montane forests in Washington and Montana, generating parameter estimates for 12 species without manual annotation. Automated speed estimates produced day ranges 2.7 to 4.3 times GPS telemetry-derived daily distances, reflecting differences between encounter velocity within detection zones and landscape-scale displacement. Deployment-level variation in detectability exceeded species-level differences 3:1, with scene depth strongly predicting detection range; mean effective detection distances ranged from 4.1 to 7.6 m. Applied to a Random Encounter Model, these parameters yielded a white-tailed deer density estimate of 21.4 animals/km{superscript 2} and the Random Encounter Staying Time model yielded 11.6animals/km{superscript 2} in Montana. This pipeline enables scalable density estimation across large camera trap networks.

Underwater receiver networks (passive acoustic telemetry systems) are deployed to track animals in aquatic habitats all over the world, but remarkably limited attention has been given over to how we can strengthen the value of these networks through statistical and computational advances. Here, we upscale state-of-the-art methods of Bayesian inference to big animal-tracking datasets from acoustic telemetry, with the largest geolocation analysis in a sparse passive acoustic telemetry system (with non-overlapping receivers) to date. Using four years of data from 93 lake trout (Salvelinus namaycush) in North Americas Lake Champlain (657,360 timesteps per individual), we formulate and fit state-space models to reconstruct animal movements through time. Uniquely, we directly embed biological expertise and detailed complementary datasets from fine-scale positioning systems, accelerometry, swim-tunnel experiments and field range tests in our analysis. Using simulated and real-world datasets, we map movement patterns and estimate residency in distinct management zones. We quantify array precision and deliver maps and residency estimates with a median error and precision (standard error) below 1 %. These results strengthen the evidence base for management. This work takes us a step towards robust inference of movement patterns at scale in acoustic telemetry systems across the world. We can, and should, build on prior scientific progress and extend the value of hard-earned data beyond individual studies to refine inferences for ecology and management. Significance statementAcoustic receivers are deployed across the globe to track aquatic animals, but reconstructing detailed movement patterns from detections at receivers remains a considerable challenge. Here, we upscale state-of-the-art methods of Bayesian inference by two orders of magnitude to analyse big, real-world datasets, using an extensive case study of lake trout (Salvelinus namaycush) in Lake Champlain. By directly integrating diverse complementary datasets from animal-borne tags, swim-tunnel experiments, field studies and close-kin mark-recapture in our analysis, we resolve detailed movement patterns over a four-year period, with broad implications for ecology and management. This work provides a powerful framework for acoustic telemetry studies that strives to meet the challenges of big, real-world datasets from telemetry networks across the world.

AimHuman land-use change contributes to biodiversity declines, but also creates new niches that facilitate novel biotic interactions. These interactions can reshape ecological communities and ecosystem function, yet remain poorly understood. Swiftlets and swallows in Southeast Asia present a classic example: coexistence is facilitated by fine-scale diet partitioning, with population sizes historically limited by available nesting substrates. However, several species now nest on manmade structures, particularly "nest farms" built to harvest edible swiftlet nests. We evaluated whether land-use change, especially the spread of nest farms, is leading to breakdowns in niche partitioning and increased competition among six sympatric swiftlets and swallows. LocationNorthern Borneo MethodsWe calculated geographic niche overlap using species distribution models (SDMs) with different environmental predictors, hypothesizing greater overlap when land-use variables were included. We then implemented joint species distribution models (JSDMs) to partition shared environmental responses from potential biotic interactions, predicting that competition would emerge as negative residual correlations. We used sightings from citizen-science datasets and structured surveys to evaluate the influence of climate, land-use, nest farms, morphology, and foraging behavior on species occurrences. ResultsSDMs that included land-use variables showed high niche overlap, suggesting that human activity homogenizes niches. The optimal JSDM, based on structured survey data, identified distance to nest farms as the strongest predictor of occurrence for all species, with species showing both positive and negative responses. Morphology and behavior had small effects, and residual correlations were weak, indicating limited unexplained biotic interactions. Main conclusionsHuman activity, through the creation of artificial nesting sites, broadly drives co-occurrence of swallows and swiftlets across our study region. These effects appear to operate primarily through environmental filtering rather than direct competition. Our findings reveal substantial and complex impacts of land-use change and anthropogenic nest sites on the distribution and composition of aerial insectivore communities.

Landscape connectivity is a key regulator of dispersal, which is an important process in community assembly. Theory predicts that connectivity may influence spatial and temporal patterns of community assembly; however, empirically evaluating the role of connectivity is nearly impossible due to the need to isolate its influence over long time frames and large spatial extents. We overcome these challenges through a large-scale, long-term connectivity experiment to test how connectivity affects plant community turnover and directionality of change over 24 years of assembly. Plant communities within connected patches had lower temporal variability in composition compared to plant communities within unconnected patches. Differences in composition between patches and the directionality of compositional changes were driven more by the amount of edge habitat in a patch and the time since the start of assembly. All community responses to connectivity were stronger for species with wind or unassisted dispersal compared to those with seeds dispersed by animals. Connectivitys role in regulating local community dynamics is critical for understanding community assembly and increasingly relevant in an era of anthropogenic land-use change. Significance StatementConnectivity between habitat patches facilitates dispersal to localities, yet the impact of connectivity on local species assemblages is exceptionally challenging to isolate from other spatial changes over time. In a 24-year experiment, we found that connectivity stabilized local community composition as a higher number of species persisted across years within patches connected by corridors. Independent of connectivity, edge effects were more important for driving compositional differences between patches. Importantly, these patterns would not have been captured with short-term data or without controlling for confounding spatial changes. Our findings have broad conservation relevance. Anthropogenic landscape changes that result in a loss of connectivity or increased edge effects may disrupt local community assembly over time.

A basic question in ecological research and biodiversity monitoring concerns the estimation of species abundances from trap catches. As a case in point, a Malaise trap can yield thousands of arthropod individuals, but how this count should be converted to numbers of individuals per unit area has remained an open question. Here, we supplement observational data with an experimental approach targeted at quantifying catchability. We released marked insects in a boreal forest and examined their capture rate by a grid of Malaise traps around the release location. We estimated insect movement rates, mortality rates, and Malaise trapping capture rates by fitting a joint species movement model to these data. As a methodological novelty, we show how to convert the movement model parameters to the expected number of captured individuals, given their actual population density. Our results show that multiplying the sample content by 30 000 yields a rough estimate of the number of individuals per hectare. This conversion factor depends on the species, generally decreasing with increasing body size. We apply the estimated conversion factors to conclude that typical boreal forest contains some four million insect individuals per hectare, out of which around half belong to Diptera. SIGNIFICANCE STATEMENTTraditionally, the Malaise trap method has been used for assessing the state of the local communities and to estimate population abundances. However, a topical question is: how does the number of individuals observed in a sample relate to the true density of individuals in the surrounding community? To answer this question, we implement a movement model parametrized by a carefully designed mark-recapture experiment, in which we are able to obtain taxon-specific conversion factors for different groups of insects. We found that different insect groups come with different conversion factors, causing a mismatch between sample contents and true community composition. Thus, treating the sample contents as a direct representation of the reference community will be misleading.

Reducing pesticide risks while maintaining food production remains a central challenge for sustainable agriculture. Although pesticide reduction is pursued through centralized regulation and farm-level Integrated Pest Management, how these governance pathways translate into pollinator recovery in agroecological systems remains poorly understood. Existing ecological network models often treat pesticide pressure asm external forcing and management actions as fixed parameters, limiting their ability to capture feedbacks among governance decisions, network structure, and population dynamics. Here, we develop a dynamical framework that embeds pesticide management within tripartite pollinator-plant-pest networks using a policy variable and a farm-level adoption variable. Across empirical and synthetic networks, we show that recovery is not determined by pesticide reduction alone, but by how management acts through ecological interaction structure. More modular networks require stronger intervention, and pollinators with similar degrees show different recovery outcomes, indicating that degree alone does not determine recovery potential. Further, increasing policy strength generally expands the persistence domain more than increasing farmer adoption alone. These results show that pesticide reduction does not automatically yield ecological recovery, and effective strategies must match governance scale to ecological condition and network structure.

O_LIEffective pest control requires a better understanding of natural enemy ecology, particularly how their distribution responds to landscape structure and local management. Agricultural intensification has simplified landscapes, reducing biodiversity and constraining pest control. Landscape-scale surveys are therefore needed to identify strategies that support natural enemies in agricultural systems. C_LIO_LIFrom 2021 to 2022, over three seasons, we surveyed 25 clementine orchards across a gradient of landscape complexity in Corsica. We modelled the tree-level occurrence of four natural enemy species in relation to ecological context across multiple spatial and temporal scales. Site-level evenness, reflecting differences in species occurrence frequencies, was also estimated to assess how natural enemy assemblages vary across local management and landscape gradients. C_LIO_LIClimate and local management were the primary drivers of species occurrence, with additional contributions from landscape factors. Species displayed distinct responses, while model performance increased with the number of sampled trees, plateauing at around half of the sampling effort. C_LIO_LISite-level evenness was higher in organic orchards than in conventional orchards. In conventional orchards, it responded to landscape structure in a scale-dependent manner, increasing at the larger scale with the heterogeneity and spatial continuity of semi-natural habitats. C_LIO_LISynthesis and applications. Organic farming enhances natural enemy presence and diversity in Corsican citrus orchards, while the naturalness and complexity of surrounding landscape can mitigate species scarcity in conventional orchards. We recommend continued long-term monitoring of arthropod communities in these systems, prioritizing the number of sampling sites and years while reducing the number of trees sampled per site to optimize effort. C_LI