Journal of Applied Ecology

1. Grassland biodiversity restoration efforts have largely focused on re-establishing diverse plant communities, with less attention paid to higher trophic levels such as invertebrates. 2. We established a controlled, field-scale restoration experiment across twelve lowland grassland sites in Switzerland in close collaboration with farmers. We manipulated seed addition (hay transfer, directly harvested seeds, or multiplied seed mixtures) and soil disturbance (shallow or deep tillage), with treatments applied to meadows covering, on average, 0.5 ha. Four to six years after restoration, we quantified the components of species richness change (lost, persisting, gained from site surroundings, and gained from donors), and estimated dark diversity (species that are absent from a location but could potentially occur there) for plants, Auchenorrhyncha and Orthoptera. 3. Compared to plants, invertebrates exhibited greater species turnover, with higher proportions of species lost and gained from the surroundings, but much lower contributions from the donor meadows. However, species persistence was high across taxa. Restoration operations strongly influenced components of species richness change in plants but had weaker effects on invertebrates. In particular, multiplied seed mixtures + deep tillage had the highest proportion of plant species not transferred, suggesting that many species in the seed mixture were poorly suited to local conditions. In both plants and Auchenorrhyncha, components of species richness change were strongly mediated by site characteristics, with initial species richness in both receiver and donor emerging as the strongest predictors. 4. Within each taxonomic group, 41% of species not transferred from the donor belonged to the dark diversity, suggesting substantial unrealized restoration potential. Restoration operations impacted dark diversity for plants and Auchenorrhyncha but not for Orthoptera. In plants, dark diversity was substantially lower under multiplied seed mixtures + deep tillage than under other treatments. The probability of a species belonging to dark diversity was associated with specific leaf area in plants and initial relative abundance in Auchenorrhyncha. 5. Synthesis and applications. Components of species richness change were influenced by both restoration operations and site characteristics, with their relative importance varying among taxa, underscoring the need for multi-taxon restoration strategies.

Freshwater ecosystems in Scotland are increasingly threatened by multiple interacting stressors, including pollution, hydrological alteration, land use change, and climate-driven pressures. Scotland's rivers are amongst the most physically diverse and dynamic in the UK, contributing significantly to the country-s economy and natural heritage. Monitoring 125,000 km of waterways presents an issue, with limited funding and capacity of environmental agencies creating an environmental monitoring deficit at a time when robust data are essential for delivering national and global biodiversity targets. Citizen science offers a scalable, community driven approach to strengthening environmental evidence. This study evaluates the development, implementation, and outcomes of Guardians of our Rivers (GooR), a nationwide citizen science programme using macroinvertebrate-based river health assessments based on the Anglers' Riverfly Monitoring Initiative (RMI) established by the Riverfly Partnership. Between 2022 and 2025, GooR trained more than 850 volunteers, established 123 monitoring sites across Scotland, and generated over 860 surveys-surpassing the previous 16 years of aquatic invertebrate monitoring efforts. Standardised field training, full equipment provision, and structured support enabled high-quality data collection across diverse catchments. The establishment of nationally applied trigger and target thresholds created a consistent mechanism for detecting ecological deterioration and ensured that verified trigger breaches (n = 39 in 2025) resulted in formal investigation or follow-up. Analysis of national abundance patterns revealed ecological and biogeographic trends across the eight RMI taxa, confirming data sensitivity to habitat differences, water quality, and regional pressures. A detailed case study of the Lothian Esk catchment demonstrated the programme's capacity to detect seasonal macroinvertebrate dynamics and site level ecological trajectories over time. Overall, GooR illustrates how well-designed and well-supported citizen science can deliver high resolution spatiotemporal datasets, enhance early warning systems, empower communities, and make a meaningful contribution to promoting river health. Importantly, GooR forged a working partnership with Scottish Environmental Protection Agency (SEPA) that allowed communities to not only gather data but also transformed the way citizen science is recognized as contributing to freshwater conservation. Continued investment and long-term support will be essential to sustain these gains and realise the full potential of citizen-driven freshwater stewardship.

1. Global restoration targets, including the Kunming-Montreal Global Biodiversity Framework and the European Union's Nature Restoration Regulation, call for extending restoration beyond protected area boundaries to reconnect isolated ecosystems, yet implementing these landscape-scale interventions forces trade-offs between biophysical effectiveness and social feasibility. 2. In the Grenoble region of the French Alps, we used a gradient analysis of 12 ecosystem services across protected area interfaces to design and compare two restoration scenarios. Both were built from the same set of co-selected interventions but differed only in their spatial allocation: a 'Technical' scenario targeting areas of low ecosystem service continuity around protected areas, and a 'Participatory' scenario co-designed with local stakeholders. 3. Both scenarios physically modified about 14.5% of the landscape, mostly through a forest-based adaptation strategy, yet their effects reached up to more than 60% of the territory for the most connectivity-dependent services, showing that benefits propagate well beyond the area directly modified. 4. The two scenarios diverged mainly in how they treated agricultural land. The Technical scenario reduced abrupt drops in ecosystem service provision at the protected area edge nearly twice as often as the Participatory one (46% versus 26% of the sharp declines it addressed), and at six locations removed them entirely. However, it concentrated over 99% of its non-forest effort in agricultural diversification, such as diversifying crops and adding hedgerows, the action stakeholders ranked hardest to implement, whereas the Participatory scenario favoured more feasible actions such as river restoration, which met less resistance but missed several critical discontinuities in ecosystem service supply. 5. These contrasts show that biophysical potential and social feasibility can pull restoration in different directions, and that the more effective spatial configuration is not necessarily the more implementable one. 6. Closing the implementation gap for nature-based solutions depended less on expanding the area restored than on where interventions were placed. Gradient-based diagnostics appear most useful not as fixed prescriptions but as tools to steer existing sectoral funding in agriculture, urban planning and forestry toward high-impact locations, an approach transferable to other regions working to reconnect protected areas with their surrounding landscapes as coupled social-ecological systems.

Introduction: Despite decades of investment, stream restoration often fails to recover diverse benthic macroinvertebrate communities, as restored reaches tend to remain dominated by pollution-tolerant taxa. This shortfall in recovery is commonly attributed to either poor habitat quality or dispersal limitation of aquatic macroinvertebrate taxa, yet these hypotheses are rarely tested directly. Objectives: We conducted a manipulative field experiment in the eastern Piedmont region (Maryland, USA) within the Chesapeake Bay Watershed to evaluate whether restored streams can support pollution-sensitive taxa when dispersal barriers are removed. Methods: We translocated whole macroinvertebrate communities from three nearby reference streams into three restored streams using standardized habitat substrates and quantified macroinvertebrate survival after 28 days. We also assessed how effective leaf vs rock substrates were for accumulating diverse and abundant macroinvertebrate communities for translocation. Results: Across all sites, 77% of translocated individuals and 16 of 22 taxa sensitive and moderately sensitive to pollution persisted under restored stream conditions, demonstrating that restored reaches can support short-term survival of pollution-sensitive taxa that typically fail to recover. Substrate comparisons revealed that leaf packs accumulated higher macroinvertebrate abundances, whereas both substrates accumulated macroinvertebrate communities with similar taxa richness, EPT richness, and Shannon Diversity. Conclusions: This work suggests that dispersal limitation, rather than unsuitable habitat, may be an underappreciated constraint limiting macroinvertebrate recovery in restored streams. These findings challenge the assumption that habitat restoration alone is sufficient for biological recovery and highlight the potential for whole-community translocations to accelerate recovery and ecological uplift as part of stream restoration practices. Implication for Practice: Our study highlights the need to consider habitat connectivity and dispersal pathways during stream restoration designs. Restoring streams to meet habitat requirements of pollution-sensitive macroinvertebrate taxa may not be sufficient for their populations to recover if these taxa are unable to reach the restored stream. In scenarios where restoring connectivity to source populations is infeasible, whole community translocations may offer a novel restoration tool to overcome dispersal barriers and accelerate biological recovery in restored streams.

Agricultural intensification is a major cause of population declines in farmland birds and insects worldwide. While these declines are expected to disrupt trophic interactions through density-dependent cascading effects, examples outside of crop pest control or involving more than two trophic levels remain poorly documented. Here we quantify such cascading effects across three trophic levels involving a declining aerial insectivore, the Tree Swallow (Tachycineta bicolor), which breeding success decreases with agricultural intensity, Protocalliphora bird blowflies (Diptera: Calliphoridae) that parasitize its nestlings, and Nasonia parasitoid jewel wasps (Hymenoptera: Pteromalidae) that feed upon blowfly pupae within swallow nests. Agriculture intensity may affect each of these actors either directly via, for instance, toxicological effects from pesticide use, or indirectly through density-dependent trophic interactions, which can exacerbate or mitigate one another. Using Bayesian models within a causal inference framework, we quantified how nestling host density, meteorological conditions, and agricultural landscape context affect Nasonia parasitoid infestation of Protocalliphora pupae, and estimated Protocalliphora emergence success. Analyses were based on 3,340 Protocalliphora puparia collected from 248 Tree Swallow nests over a 16-year period (2004-2019) across a 10,200-km2 agricultural intensity gradient in Southern Quebec, Canada. Model predictions indicated that Protocalliphora emergence probability increased steeply with nestling host availability (i.e., nestling-days). Whereas Nasonia parasitoid infestation occurrence at the nest level increased with Protocalliphora host availability, the proportion of infested Protocalliphora pupae within nest decreased and led to a dilution effect at the individual level. Although Protocalliphora abundance within nests decreased with agricultural intensity, their emergence probability peaked in agrointensive landscapes. This increase was associated with reduced Nasonia infestation rates, suggesting that agricultural intensity indirectly mitigated parasitoid pressure on Protocalliphora host populations. Our results highlight how agricultural intensification can shape multitrophic interactions and life-history traits (emergence success and infestation rates) by simultaneously reducing host availability and weakening top-down control by parasitoids.

Effective management of invasive species requires surveillance methods that reliably detect populations while minimizing field effort. Environmental DNA (eDNA) offers a potentially rapid alternative to conventional monitoring, but direct comparisons with other non-invasive methods remain limited for invasive mammals. Here, we compared eDNA water sampling and camera trapping for detecting established populations of coypu (Myocastor coypus), an invasive semi-aquatic rodent. Using replicated data collected across Mediterranean river systems in southern France, we estimated detection probabilities conditional on coypu presence using Bayesian generalized linear mixed models and quantified the sampling effort required to achieve high cumulative detection probability. Detection probability was 93.0% (95% credible interval: 72.1-98.9%) for a single eDNA water sample, compared with 52.8% (22.6-91.1%) for 30 camera-trap days, with a 96.9% posterior probability that eDNA detection was higher. Two eDNA samples were sufficient to achieve a median cumulative detection probability above 95%, compared with four 30-day camera-trapping periods. Detection also varied at different spatial scales, with camera-trap detectability being heterogeneous among local camera locations and eDNA detectability varying more among catchments. These results show that eDNA provides a rapid and reliable approach for confirming the presence of established coypu populations, while camera traps provide complementary information on activity, behaviour and habitat use. Combining rapid eDNA screening with targeted camera trapping may therefore provide an efficient surveillance strategy for invasive semi-aquatic mammals.

Field ecologists often rely on observational data to understand the impact of environmental stressors and management interventions on natural systems. Natural and anthropogenic events (e.g. wildfires, protected areas, nutrient deposition) do not occur randomly in space, however, which can introduce bias into observational studies--which we refer to as causal selection bias. Field study designs that ignore the non-random occurrence of stressors may yield biased estimates of stressor effects on ecosystems. Matching methods commonly used in economics, political science and epidemiology offer a powerful framework for controlling for causal selection bias by identifying more comparable treatment and control sites. Although these methods are increasingly used in conservation, they are rarely used in ecological field-based studies. Here we review how Propensity Score Matching (PSM) can improve field sampling designs in ecology and strengthen causal identification of stressor effects. Then we apply this approach to a case study examining wildfire effects on forest recovery in California. We conclude with practical recommendations for implementing PSM to improve causal identification of ecological change, which is particularly important for developing effective management interventions.

1. This study analyzed 21 years (2000-2025) of butterfly count data from Central Alberta, integrated with intensive 5-year (2021-2025) high-resolution intra-seasonal sampling. 2. Long-term macro-scale analysis revealed a significant decline in Shannon Diversity, a change that remained obscured when relying solely on traditional metrics of species richness and evenness. 3. This diversity decline was primarily driven by the severe, long-term collapse of the native Common Ringlet (Coenonympha tullia). 4. Four other dominant species--Cabbage White (Pieris rapae), Clouded Sulphur (Colias eriphyle), European Skipper (Thymelicus lineola), and Common Wood Nymph (Cercyonis pegala)--maintained long-term population stability, though their abundances were significantly constrained by extreme winter minimum temperatures and rapid spring warming. 5. High-resolution intra-seasonal analysis (2021-2025) demonstrated that community indices and species-specific abundances were strongly limited by daily weather, particularly wind velocity and temperature. 6. These findings illustrate that while traditional metrics like richness and evenness are fundamental to community ecology, they provide incomplete insights when applied in isolation; they are most effective when utilized as part of a complementary, multi-scalar framework. 7. This study highlights the necessity of coupling multi-decadal historical datasets with high-frequency, fine-scale sampling to accurately identify the mechanisms of community turnover that simpler metrics may overlook. 8. The results underscore the critical importance of standardized citizen science monitoring in quantifying environmental impacts and establishing conservation priorities for terrestrial insect groups.

The potential of alternative agricultural systems, such as organic or conservation agriculture, for biodiversity restoration has been widely investigated. However, variability of practices within each of these two categories of agricultural system has rarely been accounted for. Furthermore, these systems have most often been studied separately, isolating either organic or conservation agriculture (itself often reduced to a soil tillage reduction), and rarely analyzing these systems side by side. This study attempts to disentangle these multiple interrelationships by sampling carabid beetles in a network of 48 commercial Belgian farms including organic agriculture (OA), conservation agriculture (CA), farms combining the two approaches (OCA) and farms applying conventional practices (CVA). In each farm, we sampled carabids using pitfall and emergence traps in one winter cereal field and collected a detailed description of crop management practices. We found that community composition was primarily structured by seasonality early in the season (March, April) followed by convergence (May, June) driven by the aggregation of the most dominant species. Based on agroecosystem classification, organic certification (i.e., the absence of synthetic pesticide use) exerted a twofold stronger effect than application of conservation agriculture principles (i.e., reduced or no soil tillage) on carabid beetle community composition. Notably, OCA supported carabid assemblages that were structurally distinct from those of CA and CVA. Furthermore, carabid communities associated with OA differed significantly in composition from those of CVA. Observed and estimated species richness were also significantly higher in OCA than in CA and CVA, while OA differed from CVA only for observed richness. Beyond differences among agricultural systems, community responses were associated with specific management practices, which explained nearly a quarter of the compositional variation with significant influence of the duration since organic conversion and the frequency of synthetic pesticide application. Carabid abundances were negatively affected by synthetic pesticides and moldboard ploughing, whereas richness was more strongly related to the time since conversion to OA, cover crops, and organic amendments. Overall, these results position synthetic inputs and time since conversion as the main filters over carabid communities. Our study highlights the importance of jointly considering system level analyses and detailed management practices to better understand the complex relationships between agricultural practices and carabid communities.

1- Biological control of mosquitoes using aquatic predators offers a sustainable alternative to chemical insecticides, yet the specific predator and prey functional traits that govern consumption efficacy remain poorly quantified at a global scale. 2- We conducted a global meta-analysis of 755 effect sizes across 59 studies to evaluate how predator identity (fish vs. odonate naiads), body size, dietary guilds, and prey characteristics influence consumption rates. Using multilevel models and robust publication-bias corrections, we quantified predation efficiency, expressed throughout as the consumption rate (CR, larvae predator -1 h -1 ), while accounting for methodological variations across experimental designs. The primary literature itself proved geographically skewed towards Asia (chiefly India), with Africa, the Americas, and Europe markedly under-represented. 3- Grouping predators solely by broad taxonomic identity concealed the central pattern in our data. Although naiads outperformed fish when compared directly within the same studies, this taxon-level difference was driven almost entirely by non-mosquitofish species, the least efficient predator group overall. Mosquitofish (\textit{Gambusia} spp.) and dragonfly naiads were statistically indistinguishable from one another, indicating that dietary specialisation, not taxonomic identity, is the stronger predictor of predation efficacy. Predator body size strongly and positively predicted consumption rates in naiads---driven primarily by dragonflies---but showed no significant or negative relationship in fish. 4- Methodological traits heavily structured the extreme heterogeneity observed across studies; notably, exposure time acted as a severe rate-suppressor, where prolonged assays drastically underestimated per-hour consumption rates due to satiety or handling constraints. Nevertheless, a combined model incorporating all significant ecological moderators simultaneously explained a substantial share of the between-study variance, confirming that predator-prey dynamics in these systems are highly predictable from functional traits. 5- Effective biological control cannot rely on broad taxonomic assumptions but requires evidence-based trait-matching. Management programs should prioritise body size when deploying insect predators, selecting the largest individuals within species known to consume mosquito larvae and favour insectivorous fish species over generalists. Crucially, because short-term laboratory assays artificially inflate efficacy, multi-duration assessments are essential to accurately scale up biocontrol predictions from experimental arenas to complex, real-world ecosystems.

Achieving global conservation and restoration targets requires knowing not only how much land to conserve and restore, but also where such actions will have the greatest ecological effect. Conservation and, in particular, restoration priorities are often inferred from habitat condition, opportunity, or broad ecological value, yet these criteria do not necessarily identify where local changes generate the largest landscape-scale consequences. Building on sensitivity-based conservation prioritization, we extend the approach to derive ecological loss and gain potential by differentiating a persistence-relevant connectivity metric with respect to local habitat condition. Sensitivity identifies areas where local degradation or improvement would most strongly affect connected habitat at the landscape scale. Combining sensitivity with current condition distinguishes high loss potential, where degradation would have large ecological consequences, from high gain potential, where ecological improvement could generate large benefits. Applying the framework to national forest-naturalness data in Norway, we show that loss and gain potential are spatially related but far from redundant, and that gain potential is not simply concentrated in the most degraded forests. Both high-loss and high-gain areas were poorly represented within current protected areas, although loss potential was consistently better represented than gain potential. Extending sensitivity-based prioritization in this way provides a general and scalable means of linking local habitat-condition change to persistence-relevant landscape outcomes, while distinguishing the ecological potentials that can inform conservation and restoration decisions.

Global biodiversity recovery demands restoration at landscape scales, where ecological outcomes are shaped not only by habitat amount but also spatial arrangement. Alternative landscape restoration strategies can increase habitat amount, improve quality, and alter configuration in fundamentally different ways and may therefore vary in their efficiency. Here, we conceptualise five landscape restoration strategies and use a spatially explicit metapopulation framework to compare their ability to increase predicted species persistence across heterogeneous landscapes and trait space defined by dispersal ability and specialisation. We find that restoration outcomes are strongly trait-dependent, with restoration of existing degraded habitats yielding the highest benefits - particularly for specialists. Improvement of the landscape matrix to create low-quality habitat primarily benefits generalists, while strategies creating new habitat patches or enlarging patches provide more consistent but modest gains. These results reveal fundamental trade-offs among restoration strategies and highlight the utility of trait-explicit approaches to inform their implementation and evaluation.

1. Agriculture is the most significant driver of global land-use change, placing increasing pressure on biodiversity. Balancing rising food demand with the need to conserve biodiversity is therefore an urgent challenge, especially as biodiversity can underpin key ecosystem functions that support agricultural systems. A greater general understanding of how biodiversity influences agroecosystem function is critical for developing agricultural practices that benefit both agricultural production and wider ecosystem processes. 2. Using a global dataset of 184 biodiversity-agroecosystem function relationships covering agricultural and silvicultural systems, we developed a classification system for different biodiversity-enhancing approaches, ranging from the planting of multiple crop species to creating tree islands or flower strips, providing the first comprehensive assessment of how these approaches affect agroecosystem function. 3. We found biodiversity enhancement generally had a positive, or at least neutral, association with multiple agroecosystem functions, indicating that it contributes to agricultural sustainability. Crop diversification had the greatest impact on key agroecosystem functions compared to other practices, likely due to the low heterogeneity already present in agricultural systems and the scale at which crop diversification is applied and measured. Overall, biodiversity's largest effect was on multifunctionality, highlighting its role in supporting the simultaneous delivery of multiple agroecosystem functions. 4. Variability in relationships among systems and agroecosystem function categories suggests that no single biodiversity-enhancing approach will universally benefit all functions simultaneously. Instead, a combination of approaches, assessed through their contribution to overall multifunctionality, is likely to be most effective, ensuring that agricultural sustainability claims reflect benefits across systems. 5. Policies should incentivise biodiversity-enhancing management approaches as effective strategies to improve agroecosystem multifunctionality, benefiting both agricultural productivity and the wider ecosystem.

The Community Conservation Index (CCI) developed by Chadd and Extence (2004) has been widely used to assess the conservation value of freshwater macroinvertebrate assemblages by integrating species rarity and community richness into a single metric. However, the original conservation scoring system was based largely on historical rarity designations and regional distribution data that no longer reflect current conservation assessments or recording coverage. This paper presents a comprehensive revision of the Conservation Scores underpinning the CCI. Species scores were updated using the latest IUCN-compliant Red List assessments for British freshwater invertebrates. For taxa classified as Least Concern, Data Deficient, or lacking a recent conservation assessment, scores were derived from national occurrence data across Great Britain for the period 1995-2024 using hectad occupancy. In addition, all non-native species were assigned a conservation score of zero to ensure they do not contribute positively to site conservation value. The revised dataset includes 1,680 taxa, compared with 1,117 in the original index, with 755 taxa receiving Conservation Scores for the first time. Thirty non-native taxa were assigned a score of zero. Among previously scored taxa, 165 received higher scores, 576 received lower scores, and 184 remained unchanged. The predominance of score reductions reflects improved knowledge of species distributions and the replacement of historical rarity classifications with contemporary assessments based on conservation status and national frequency of taxon occupancy. The revised CCI maintains the original strengths of combining rarity and richness while improving transparency, consistency and ecological relevance. By incorporating current Red List assessments, Great Britain-wide distribution data and revised treatment of non-native species, the updated index provides a more robust and nationally representative measure of freshwater macroinvertebrate conservation value. It offers an enhanced tool for biodiversity assessment, conservation prioritisation, site designation and environmental decision-making across freshwater ecosystems in Great Britain.

The reintroduction of extinct or endangered species to restore ecosystem function is an essential aspect of rewilding. The Wilder Blean Project at West Blean and Thornden Woods in Canterbury, UK, is committed to rewilding natural processes and enhancing biodiversity in one of England's oldest and largest areas of ancient woodland. The introduction of European bison (Bison bonasus) is an important part of the project. However, how the reintroduction of large herbivores influences local biodiversity and ecosystem functions during the early stages of rewilding remains poorly understood. Soil samples were collected from the same sampling sites before and two years after bison were reintroduced and profiled by metagenomic sequencing using Oxford Nanopore Technologies sequencing platforms. The results showed that the alpha diversity of soil organisms did not change significantly before and after the introduction of European bison, while beta diversity showed modest shifts in community composition. The relative abundance of some nitrogen-fixing and photosynthetic microbial genera showed declines in the 2024 Bison Area, while the mycorrhizal fungus genus Rhizophagus was significantly less abundant than in the 2024 Control Area. Despite relatively stable taxonomic diversity, functional composition differed significantly between the 2022 and 2024 Bison areas and among the 2024 rewilding treatments, revealing a decoupling between taxonomic diversity and functional composition. Amino acid synthesis pathways and carbon metabolism pathways were significantly enriched. These findings highlight the potential of long-read Oxford Nanopore metagenomics to reveal functional shifts that may not be apparent from taxonomic diversity alone. Although these early-stage responses cannot yet predict long-term rewilding trajectories, continued longitudinal monitoring integrating microbial, soil physicochemical, and ecosystem-level measurements will be essential to determine the persistence and ecological significance of these functional shifts.

Climate change is reshaping weed population dynamics, creating non-stationary management challenges for which static decision rules are insufficient. This chapter presents and evaluates an open surveillance toolkit of artificial intelligence tools for recognizing and mapping Ambrosia artemisiifolia (common ragweed) at complementary spatial scales, tested across two phenological phases (seedling, September 2024; adult, December 2024) in a lentil paddock in central Chile. At the field scale, a YOLOv11 detection model coupled with Slicing Aided Hyper Inference achieved mAP50 of 0.886. Cross-domain experiments revealed that single-site models collapse catastrophically when deployed in new geographies (mAP50 dropping to 0.108), but multi-domain training recovers performance to 0.874. At the satellite scale, PRESTO foundation model embeddings correlated strongly with ground-truth weed density (r = 0.739); geographically weighted regression explained up to 91% of local density variance. All methods are released as the ragweed-ai-toolkit (https://github.com/agroia-lab/ragweed-ai-toolkit), a modular open-source Python package enabling adaptation to new species and geographies.

Environmental DNA (eDNA) offers a non-invasive alternative to traditional, more destructive sampling methods for determining species occupancy at ecological sites of interest. Aquatic eDNA sampling entails filtering a known volume of water to capture and detect genetic material shed by organisms. While the influence of individual environmental properties on the presence of target eDNA has been widely studied, it remains unclear how variables like temperature, pH, flow rate and conductivity correlate collectively with site electrofishing counts and eDNA concentrations. Resolving this question is important for two reasons: (1) typically, only eDNA, not physical specimens, is collected and measured, and (2) when methods like electrofishing and eDNA sampling are used in tandem, results often differ. Here unsupervised association rule-based machine learning is employed to discover interesting relationships among sampled covariates within a previously published case study of native brook trout (Salvelinus fontinalis) collected from Hanlon Creek (Guelph, Ontario, Canada) in September 2019. From a dataset of only 126 observations, the mining process revealed over 12000 plausible association rules linking covariates to eDNA concentrations (low/high) and electrofishing outcomes (absence/presence of brook trout). A strict pruning strategy reduced this ruleset to a manageable size of 153 associations, some of which were corroborated by existing literature, and some of which were novel (such as those potentially relating electrical conductivity to microbial and enzymatic activity). The entire workflow is included as a new R package called RulesTools. These results highlight the promise of association rule mining as a tool for guiding eDNA metadata collection, complementing statistical modelling, and informing conservation and management decision-making.

Citizen science observations are abundant, but conservation requires turning uneven records into reliable predictions and directing new surveys to where information is missing. We developed a biodiversity platform for Japan that is updated monthly and integrates 2.32 million records to predict 8,297 species across seven taxonomic groups. Shared representation models outperformed species-specific models in four groups and extended predictions to species with few records. Five independent datasets, including structured monitoring, environmental DNA and complete forest inventories, confirmed that the models ranked observed species and occupied sites above alternatives, with median AUCs of 0.724 to 0.894 across sites and 0.650 to 0.841 across species. For any user-selected area, the platform returns candidate species, distribution predictions, a biodiversity map corrected for uneven observation effort, a conservation priority map for native species and a map recommending where to survey next. This map highlights places where species with few records are predicted to occur despite limited sampling. Independent observations showed that areas ranked highly by this predicted potential contained many such species, indicating that model predictions can help direct surveys toward knowledge gaps. New observations are incorporated into monthly updates, creating a national feedback system connecting citizen science, local conservation decisions and future surveys.

Automated monitoring camera systems offer an efficient approach for quantifying pollinator biodiversity and plant-pollinator interactions, but financial and logistical constraints limit the number of flowering plant species that can be monitored. Using a large European database of plant-pollinator networks, we evaluated whether plant subsampling can capture key metrics of interest (i.e. pollinator richness and network structure). We compared abundance-based, flower-shape-informed, phylogenetically informed and random plant sampling. Abundance-based sampling consistently outperformed random sampling, while adding flower shape provided little additional benefit and phylogenetic selection performed similarly to random sampling. Overall, monitoring 12-15 flowering species was sufficient to characterize key network properties across varying community sizes. Performance of abundance-based plant sampling declined in species-rich communities and when rare but highly attractive plants were present in the community, while a specific subset of pollinator species was consistently missed even by the best sampling strategy. Based on their relative contribution to interactions within each network, only 12.6% of pollinator species accounted for 95% of interactions, suggesting that AI classifiers could prioritize a relatively small subset of species. Our results provide practical guidance for designing efficient camera-based pollinator monitoring schemes at large spatial scales.

While plant communities that exist at the edges of agricultural fields are typically composed of 'weedy' species, they provide essential resources for insect pollinators and other wildlife. However, little is known about how sublethal exposures to herbicides via drift (off-target chemical movement by air) affects the fitness of weedy plant species across varying agro-ecological environments. If weed species vary in sensitivity to herbicide drift, then this could result in plant community shifts in favor of more tolerant species. Consequently, these changes in species composition could alter resource availability in agroecosystems. Here, we replicated a common garden experiment in three US states where agriculture is a major form of land use (Michigan, Pennsylvania, and Tennessee). We exposed nine weed species to a sublethal, drift-level rate of the herbicide dicamba and examined plant fitness. We found that fitness in the presence of dicamba drift depended on both the weed species and the environmental context. For some species, dicamba drift significantly reduced fitness. For others, it had no effect, indicating greater tolerance to the herbicide. However, more variation in tolerance to dicamba drift among species was observed in Michigan compared to Pennsylvania or Tennessee, a result that was partially explained by the severity of damage from insect herbivory that occurred at each location. In a follow-up study, we explored the relationships between plant developmental traits and fitness to determine the functional pathways that underlie tolerance to dicamba drift using structural equation modeling. This approach revealed that weedy plants that are better able to buffer the effects of dicamba drift on short-term growth are then better equipped to minimize effects on flower production and ultimately fitness (seed set). Overall, our work advances knowledge about how wild plant populations in different contexts are responding to unintentional, yet increasingly common occurrences of anthropogenic stress in the form of off-target herbicide exposures.