Movement Ecology

O_LIClassifying animal movement strategies from GPS tracking data is essential for understanding space use, population dynamics and conservation planning. However, existing approaches either require strong parametric assumptions about trajectory shape, large labelled datasets (i.e. expert-annotated) for machine learning, or lack formal uncertainty quantification. These limitations create barriers for researchers working with novel species or limited sample sizes. C_LIO_LIWe present a profile-based classification framework consisting of three steps. First, trajectories are segmented using breakpoint detection applied to Net Squared Displacement (NSD) time series. Movement metrics are then extracted from each segment and classified by comparing them to empirically derived behavioural profiles via Z-score distances transformed to softmax probabilities. Bootstrap resampling quantifies uncertainty in the resulting classifications from both training and test data. We validated the framework through simulation experiments and applied it to GPS tracking data from two ecologically contrasting species: gray wolf (Canis lupus;43 individuals) and northern lapwing (Vanellus vanellus;15 individuals). C_LIO_LISimulations showed that 5-10 training segments per movement strategy suffice for reliable classification, with overall accuracy of 91.1%across residential, floating and dispersal strategies. Segment duration of 30-60 days was required for confident discrimination of residential and floating behaviour. For wolves, the framework clearly distinguished residency, floating or dispersal (91.2%of segments classified with >50%probability). For lapwings, migration was identified with high confidence, while residential-floating discrimination reflected genuine ecological ambiguity confirmed by domain experts, with bootstrap confidence intervals transparently flagging uncertain cases. C_LIO_LIThe profile-based framework provides an accessible, interpretable alternative to parametric NSD fitting and machine learning approach, requiring modest training data while delivering probabilistic classifications with honest uncertainty estimates. An R package (moveprofile) implementing the complete workflow is freely available. The framework is applicable to any tracked species where distinct movement strategies can be identified by experts knowledge. C_LI

BackgroundThe Andean bear (Tremarctos ornatus), South Americas only ursid, is one of the worlds most elusive large mammals, making movement data collection exceptionally rare. Addressing this gap, we present the largest telemetry dataset ever assembled, spanning 19 individuals tracked across three Ecuadorian National Parks over two decades, paired with a novel analytical approach. MethodsWe integrated Continuous-Time Movement Models (CTMM), Auto-correlated Kernel Density Estimators (AKDEs), Hidden Markov Models (HMM) and a diel niche theoretical framework to mitigate biases previously unaccounted for the species in telemetry studies. Fine-scale AKDEs and non-linear movement metrics were calculated to understand seasonal space use and movement behaviors. Speed and diffusion from CTMM and behavioral states from HMM were modelled with environmental covariates to investigate which conditions shape diel and seasonal activity. ResultsPopulation mean home range was 138.2 km2 (95% Confidence Intervals 78.7-225.5), with males (239.8 km2; 182.8-307.5), significantly exceeding females (58.5 km2; 35.5-90.3). Notably, three females exhibited ranges comparable to some males. Weekly and monthly AKDEs uncovered cyclic home range dynamics potentially driven by resource availability, with contractions around corn harvests, mortino and achupalla fruiting, and expansions during paramo transitions. Decoupling speed from diffusion rates showed region-specific behaviors: intensive patch exploitation in Llanganates, broad exploratory ranging in Cayambe-Coca, and suppressed female locomotion in Cotacachi-Cayapas. Statistical analyses identified temperature as a key diel modulator and precipitation as the seasonal driver. Foraging probability increased between 2:00-6:00, large displacements between 7:00-14:00, and nocturnal movement rose significantly under colder conditions. Across diel hypothesis frameworks, bears were classified as cathemeral rather than strictly diurnal, corroborated by camera-trap records from Colombia, Ecuador, and Peru. ConclusionsWe propose a cathemeral diel phenotype that responds to thermal fluctuations and situates Andean bears within a broader ursid context of thermoregulatory niche plasticity. This dataset reveals unprecedented resolution of regional and sex specific behaviors that will facilitate and accelerate comparative studies in rapidly changing Andean landscapes. By releasing this long-term dataset as an open resource, we provide a foundation for climate-resilient conservation strategies. More broadly, we advocate for data democratization and invite collaboration.

Winter can affect animal population dynamics by limiting resource availability and increasing energetic costs of movement caused by deep snow. Given the rapid alteration of snowpack properties due to climate change, quantifying how snow characteristics influence reproduction and physical condition is critical. We evaluated how snow cover duration, depth, and density affect spring body mass, reproduction probability, and subsequent autumn body mass of bighorn sheep (Ovis canadensis) using 45 years of individual-based data at Ram Mountain, Alberta, Canada, along with historical snow records reconstructed via the SNOWPACK model. Using Bayesian structural equation modeling, we quantified the direct and indirect effects of snow across different sex and age classes. Long and deep snow covers reduced spring body mass across all demographic groups, with yearlings, especially males, losing up to 0.12 kg per additional cm of snow depth. Harsh snow conditions reduced the probability of reproduction for adult females and generated a compensatory indirect effect on mass by avoiding the energetic costs of reproduction. In contrast, yearlings showed no compensatory responses and entered the following autumn in poor condition (up to 14% lighter for males and 8% for females following the deepest snow years). The impact of snow density on autumn mass of adult males was density-dependent, shifting from beneficial at low density (+0.09 kg per kg/m3) to detrimental at high density (-0.04 kg per kg/m3). The effects of snow conditions generate persistent, context-dependent carry-over effects across seasons. Our study suggests that distinct demographic groups rely on different mechanisms to cope with environmental constraints, highlighting complex, time-lagged consequences of changing winter climate on alpine herbivore populations.

BackgroundWintering migratory birds must balance energetic requirements, resource availability, and disturbance in increasingly human-modified landscapes. However, individual-level variability in daily movement and winter space use remains poorly understood in South Asian populations of the common crane. We investigated how seasonal dynamics, landscape composition, and individual differences structure winter movement ecology in a semi-arid agro-wetland system in western India. MethodsWe analysed high-resolution GPS telemetry data from multiple tagged cranes tracked across three consecutive winters. Daily movement distances were modelled using mixed-effects approaches to partition variance within and among individuals and among winters. Daily movement trajectories were evaluated using non-linear temporal terms. Landscape predictors, including cropland proportion, built-up area, and habitat heterogeneity, were incorporated to assess environmental drivers. Winter range distributions were estimated using autocorrelation-informed kernel density estimation within a continuous-time movement modelling framework. ResultsMost variation in daily movement occurred within individuals rather than among them, indicating strong behavioural flexibility. Interannual differences explained substantial variance, suggesting sensitivity to changing environmental conditions. Daily movement distance followed a non-linear seasonal pattern consistent with shifts in the profitability of agricultural resources over winter. Cropland proportion and landscape evenness were negatively associated with movement distance, whereas a high proportion of built-up areas increased daily movement distance, reflecting a trade-off between resource concentration and anthropogenic disturbance. Winter range distribution size varied markedly both within individuals across years and among individuals within seasons. ConclusionWinter movement and space use in common cranes are predominantly context-dependent and environmentally driven. Seasonal dynamics, agricultural landscapes, and human disturbance jointly structure movement patterns, with limited but consistent individual differences. Multi-year, individual-based telemetry provides a comprehensive understanding of winter spatial strategies in dynamic semi-arid agro-wetland systems.

Changing climates are reshaping animal populations, but our understanding of how demographic trends are shaped by individual responses to local environmental conditions is often limited to long-term studies with restricted spatial scales. Increasing evidence suggests that climatic extremes exert differential selection pressures across environments, often leading to nonstationary biological responses among populations. Participatory science (i.e. citizen science) observations can detect this variation at large geographic scales, but analyses of these data often lack insight into the individual-level responses that are required to explain the origins of such variation. Here we present a new research framework that uses long-term data to validate, then inform analyses of participatory science data to measure reproductive responses to environmental variation across large geographic scales. We use this approach to investigate how reproduction in a montane-adapted songbird, the mountain chickadee (Poecile gambeli), varies across elevations and latitudes in response to extreme scarcity and extreme accumulation of snow throughout the Sierra Nevada Mountains in North America. Chickadee reproduction in lower and higher elevation populations was often differentially impacted by drought and deluge snowfall extremes, but these relationships varied across latitudes. Reproductive performance in the northern Sierra Nevada was negatively affected by snow deluge conditions at high elevations, whereas snow drought conditions reduced reproductive output at low elevations. These relationships changed in the central Sierras where drought conditions negatively impacted both elevations, but deluge conditions improved reproductive performance at both low and high elevations. Reproduction in the southern Sierra Nevada was less affected by spring snow levels, likely due to the lower snow accumulation and earlier snowmelt in this region. These results emphasize the power of long-term studies to inform and interpret participatory science data in order to better understand how animal responses to environmental extremes vary across large geographic scales.

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.

AimHome range size is a fundamental aspect of animal spatial ecology, and understanding the factors that shape it is important for conservation purposes. Several hypotheses, based on energy needs or competition, assume that home range size negatively correlates with population density. However, this pattern has been little investigated on a global scale, and it remains unclear whether it would stand at both intra- and interspecific levels. To fill this gap, we conducted a global exploration of this relationship at the level of an animal family. Location: Global. Time period: Contemporary. Major taxa studied: Wild Felidae. MethodsIndividual home range size records (n = 1022) and population density estimates (n = 1061) were retrieved from the literature for 23 felid species across the world. We first investigated the interspecific relationship by modelling the median home range size of a species as a function of its median population density. To study the intraspecific relationship, we spatially merged data points based on their spatial or temporal proximity. We then applied a mixed-effects linear model using species as a random factor. ResultsWe found that home range size was negatively associated with population density, at both interspecific (-1.323 {+/-} 0.180, p < 0.001) and intraspecific levels (-0.569 {+/-} 0.201 to - 0.537 {+/-} 0.201 depending on the merging approach, p < 0.01). Landscape features were also predictors of home range size, without confounding the effect of population density. Main conclusionsSeveral processes likely govern the relationship between home range size and population density: differences in body mass between species may drive the interspecific relationship, whereas the intraspecific pattern is probably explained by conspecific competition. Although more research is needed to quantify their relative contribution, our study highlights a worldwide ecological pattern that exists at multiple biological levels in the wild.

Prolonged association between mothers and their offspring is common in ungulates, with the level of maternal investment likely to play a central role in shaping this trait. Here we examined patterns of association between mothers and offspring over time, the apparent benefits of association to offspring, and costs to mothers. We analyzed 40 years worth of census data from an individually-monitored, food-limited population of red deer (Cervus elaphus) on the Isle of Rum, Scotland. Starting from birth, female calves associated more frequently with their mothers than male calves in their first year. Calves also associated less with their mothers if the mother did not conceive a new calf. Association frequency decreased with mothers age and population density, and survival over the first year was not related to mother-calf association. Yearlings, now in their second year, were more often associated with their mothers if they were female, if there was no subsequent calf (or the subsequent calf died as a neonate), and if they were still being suckled. Increased association between mothers and yearlings was associated with increased survival to adulthood at 28 months, but suckling a yearling did not improve its probability of survival. For individuals that reached maturity, increased association in the yearling year was associated with slightly shorter adult life spans. The level of association between a calf and mother was not associated with the mothers immediate survival or fecundity. Our findings suggest that juveniles born to poor-condition mothers benefit from prolonged association through improved yearling survival.

ContextAnimals balance resource acquisition with risk mitigation. These trade-offs are rarely uniform, being mediated by spatial scale, demographic traits, and environmental constraints. Understanding these divergent spatial behaviors is critical for management across human-dominated landscapes. ObjectivesWe investigated how sexual dimorphism and ontogeny interact with landscape structure to influence scale-dependent resource selection. Specifically, we sought to determine how these demographic factors mediate spatial trade-offs between optimal foraging habitats, top-down intraguild predation risk, and bottom-up severe winter weather. MethodsWe examined the spatial ecology of a solitary carnivore, the bobcat (Lynx rufus), across a heterogeneous, human-modified landscape in northern Minnesota, USA. Using spatial data derived from harvested adult and juvenile individuals, we evaluated multi-scale selection relative to land cover, structural ecotones, intraguild predator activity, and winter severity. ResultsHabitat selection was scale-dependent and partitioned demographically. Whereas bobcats universally selected for ecotones and avoided homogeneous open habitats at fine scales, responses to other features diverged by sex and age. Females actively avoided areas with high coyote activity and freezing temperatures; males exhibited high risk tolerance, apparently indifferent to coyote activity and tolerant of freezing temperatures. We identified a distinct ontogenetic spatial shift among females. Subordinate juveniles were competitively excluded from optimal natural ecotones, forcing them into riskier, anthropogenic agricultural edges. In contrast, adult females optimized foraging opportunities by selecting productive ecotones at the intersection of woody vegetation and semi-natural grasslands. ConclusionsOur findings demonstrate that habitat selection is not a static species-level trait, but instead a dynamic process resulting from the interaction between ontogeny, sex, and landscape heterogeneity. The reliance of vulnerable demographic groups on marginal or anthropogenic habitats highlights how human land-use changes can inadvertently produce ecological winners and losers within the same species. Consequently, landscape management and conservation planning for solitary carnivores must shift from broad, population-wide habitat prescriptions to strategies that explicitly accommodate the divergent spatial requirements of specific demographic cohorts.

Understanding how migratory birds balance energetic costs of movement with wintering benefits requires quantifying their energy expenditure across the full annual cycle. Here, we present a novel approach to reconstruct annual energy budget from multi-sensor geolocator data on atmospheric pressure and activity. We apply this method to tracking data of Eurasian Wrynecks from a European breeding population exhibiting striking variation in migration distance: short-distance migration to southern Europe, medium-distance to northern Africa, and long-distance to sub-Saharan Africa. Long-distance migrants had 21-26% lower annual energy expenditure than shorter-distance migrants, primarily due to reduced thermoregulation costs during boreal winter. However, they faced extreme physiological demands during migration, with daily energy expenditure exceeding 9.5 times basal metabolic rate. Since 1950, climate warming has progressively reduced winter thermoregulation demands disproportionately benefiting and potentially promoting shorter-distance strategies. These results reveal shifting energetic trade-offs under climate change, potentially driving evolution of migration patterns.

Elucidating how habitat degradation facilitates extinction is critical for effective conservation efforts. Here, we propose integrating physiologically-structured population models into stochastic population viability analyses to assess how differing consequences of habitat degradation interact to drive extinction dynamics in a focal population. Using the isolated spectacled caiman Caiman crocodilus population/ecomorph from the Apaporis River as a case study, we find that threatening the resource base, which individuals increasingly rely upon, to outgrow vulnerable size ranges and mature accelerates extinction. We also found that when habitat degradation impacts both the primary adult and juvenile resource bases, this can have marked synergistic effects on threatening population viability. By contrast, destroying nesting sites has only a small effect on accelerating the impact of deteriorating prey availability. Through integrating community-level feedback between habitat degradation/change and population dynamics/structure, our approach provides a comparative framework for assessing the relative importance of distinct mechanisms through which habitat degradation ultimately drives extinction risk.

Mark-recapture analyses on the delineation of natural populations between areas often assume random sampling, with a between/within (B/W) area resighting ratio that declines towards zero as the population components of two areas become more-and-more isolated from one another, with fewer-and-fewer individuals mixing between areas. I use an individual based population model split in two areas to simulate this result, analysing also for the potential effects of the space-time fidelity of the mark-recapture sampling in the areas. I find that small B/W resighting ratios--that traditionally is taken as evidence of population isolation--can easily be observed within a completely mixing population if a random sampling scheme is restricted in space and/or time. Random sampling within restricted areas and time windows is not sufficient to estimate mixing rates and population isolation between areas, unless the resighting rates are analysed by a method that accounts both for the space-time fidelity of the scientific sampling scheme and the space-time fidelity of the distributional behaviour of the individuals in the population.

The release of tens of millions of common pheasants (Phasianus colchicus) across the UK for shooting may pose an ecological risk to native species and sensitive habitats, particularly if the birds move into protected areas (PAs) such as Special Areas of Conservation (SAC), Special Protection Areas (SPA), and Sites of Special Scientific Interest (SSSI). The extent of this ecological risk depends on the abundance of pheasants in these sensitive sites, especially if they are attracted there after the shooting season when game management efforts to retain the birds cease. We used relative pheasant abundance measures derived from British Trust for Ornithology bird atlas data from 3793 2km tetrads across four English counties (Berkshire, Cornwall, Devon, and Hertfordshire) to determine if pheasants preferentially disperse into or reside in areas with greater PA coverage. We analysed relative abundance in both the winter shooting season and the breeding season using a Bayesian occupancy-abundance model, controlling for habitat type and diversity. Our results showed a strong influence of habitat on pheasant abundance, consistent with known habitat preferences. However, we found no evidence of a relationship between relative pheasant abundance and the proportion of ecologically relevant PA coverage in a tetrad. This lack of a relationship was consistent across all four counties and across both the winter and breeding seasons. Our finding suggests that common pheasants do not preferentially disperse into or reside in protected areas compared to surrounding, unprotected land, suggesting that the ecological impacts caused by released pheasants are no more likely to occur in protected areas than in non-protected areas.

Guinea worm disease eradication efforts may benefit from environmental surveillance methods capable of detecting infected copepod intermediate hosts in aquatic habitats. We developed a three-dimensional, spatially explicit agent-based model to examine how ecological processes influence detection probability for a hypothetical water sampling method. The results show that surveillance sensitivity is shaped by the combined effects of larval diffusion, copepod density, and pond size, with interactions among these factors producing nonlinear relationships. Detection, in our model, was concentrated within a relatively restricted period after larvae matured to the infective stage and before dispersal and mortality reduced presence, indicating a limited spatiotemporal window for effective sampling. Surveillance performance peaked under intermediate dispersal regimes that generated sufficient spatial overlap between larvae and intermediate hosts, while both limited dispersal and excessive diffusion reduced detection by constraining encounters or diluting larval concentrations. Increasing habitat size reduced detection by diluting larval concentrations, but the magnitude of this effect depended on copepod density and dispersal dynamics, producing nonlinear and threshold responses rather than simple scaling with pond volume. Spatial and temporal patterns of detection shifted as larvae dispersed, with the most favorable detection periods occurring when both larval abundance and intermediate host encounters were elevated. These findings indicate that surveillance can be guided by local ecological conditions. When the timing of larval introduction is uncertain, effective surveillance requires repeated sampling over time to capture transient windows of detectability and the sampling will be less effective in very stagnant and highly mixed waterbodies. Overall, this study demonstrates how mechanistic modeling can support the design and interpretation of environmental surveillance strategies for Guinea worm eradication programs. Author summaryGuinea worm disease is close to eradication but confirming that transmission has fully stopped remains difficult because detecting infectious larvae in water is challenging. Transmission depends on freshwater copepods that become infected after ingesting Guinea worm larvae. These copepods are short-lived and unevenly distributed within ponds, and infected individuals may die before larvae reach the infective stage. As a result, environmental detection is inherently uncertain. We developed a three-dimensional agent-based model to simulate larval dispersal, copepod infection, and water sampling in a pond environment. The model shows that detection is constrained to a brief period when mature larvae and copepods overlap in space and time, and that this window depends strongly on local ecological conditions such as larval dispersal, copepod density, and pond size. Because infected copepods can be present outside these narrow detection windows, negative water samples do not necessarily indicate absence of transmission, highlighting the need for repeated, spatially targeted surveillance during the final stages of eradication.

O_LIFear of humans can drive persistent changes in wildlife behavioural and life-history traits, with cascading effects on entire ecosystems. Human multimodal cues and pet cues may influence impact of such fear, yet no study has tested how wildlife fear responses change when human acoustic cues and pet visual and acoustic cues are added to human visual cues. Filling this gap is important for managing human and pet outdoor activities while conserving wildlife. C_LIO_LIHere, with dogs representing the pet, we tested the effects of human and dog cues on fear responses of wild sika deer (Cervus nippon yesoensis) in approximately 800 km2 area, northern Japan, using alert distance (AD) and flight initiation distance (FID). First, we measured AD/FID with an approaching surveyor alone and with additional cues. Then, we fitted linear mixed-effects models while controlling for key covariates. C_LIO_LIFrom analyses with 266 observations, AD was estimated at 80.0 m with the human visual cue alone, and dog barking increased AD by 18.4m. FID was estimated at 57.1 m with the human visual cue alone, and human voice and the dog decoy increased FID by 11.3m and by 8.5 m, respectively. C_LIO_LIThese results demonstrate that human multimodal cues and pet cues can increase prey fear responses. Our findings also suggest that dog walking may expose wildlife to simultaneous human and pet cues more consistently than predator co-occurrence typically does in nature. The increase in FID with human acoustic cues, in contrast to previous studies, suggests that animals may shift cue weighting depending on predator species, potentially using human voices to help identify the threat as human. C_LIO_LIPrevious studies show that multimodal predator cues increase prey fear responses, and our findings extend this flamework to fear responses towards humans. Our findings can inform more tolerant management of human recreation and pet walking in sensitive areas. Reducing human and pet cues through signage, guidance, and zoning may prevent flight and associated energy expenditure, whereas mitigating vigilance may require behavioural guidance and spacing between pet-walking visitors. Overall, shaping how humans and pets behave may be more practical than blanket restriction. C_LI

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.

There is an urgent need to gather data on harvest rates of waterbirds in Europe to assess the sustainability of hunting. Estimates of total waterbird harvest in the United Kingdom (UK) and the relative harvest of different huntable species come from two separate surveys, the Value of Shooting (PACEC 2014) and National Gamebag Census (NGC, Aebischer 2019), and these have been recently used to explore the likelihood of unsustainable harvests of wild waterbirds by UK hunters (Ellis and Cameron 2022; Madden et al., 2025). The reliability of these sustainability estimates depends on how representative the original surveys are of hunter behaviour and success. There are also 1-3 million released game-farm mallard (Anas platyrhynchos) that takes up considerable and unquantified proportions of the UK waterbird harvest. Here we explore uncertainties in the UK winter harvest of wild waterfowl by comparing estimates from the NGC dataset with those from the Crown Estate coastal hunting clubs, and a novel approach using analysis of social-media images (2019/20 to 2023/24). We explore the difference in species-specific harvest with and without the uncertainties in the number of released mallard and the total number of duck harvested in the UK. Waterbird harvest estimates differ markedly depending on the input dataset and whether released mallard are included in the analysis. Confidence intervals of each estimate are inflated by uncertainties in the number of released game-farm mallard contributing to, and the size of that national bag. Estimates extrapolated from social media suggest the national harvest of several species may be considerably larger than the corresponding NGC estimates (e.g. Teal *2.07 and gadwall *11.2), while mallard harvests away from formal shoots represented by NGC are significantly lower (*0.71). Excluding released mallard reduces the statistical estimate of total wild duck harvest by 56-63%, which would have biologically significant effects if realised.

Explainable machine learning (ML) methods are gaining increasing attention in environmental and ecological research for their ability to reveal relationships between environmental drivers and population dynamics. However, there remain questions on the reliability of these tools, especially given recent research shows that these explanations can be highly sensitive to model architecture. In ecology, it is typical to use a single ML model, and a comparative evaluation of sensitivity of explainability for different ML approaches is overlooked. In this paper, we develop a novel framework that quantifies explanation consistency between multiple ML model architectures. This framework provides a discrepancy measure for each model prediction, with high discrepancy indicating substantive explanation disagreement across models and low discrepancy indicating strong consensus in explanations across models. We then demonstrate that low explanation discrepancy aligns well with ground truth mechanism. Furthermore, high explanation discrepancy provide a mechanism to identify areas for model refinement and further investigation by domain experts. We do this by using a simulation study based on synthetic coral cover data that incorporate spatio-temporal variability driven by known disturbance effects. Our method provides a quantitative approach to assess the sensitivity of explainable ML in the absence of ground truth. As a result, this enhances the utility of ML approaches in conservation and ecological management. While we focus primarily on ecological modelling for coral reefs, our methods are generally applicable to other ecological and environmental modelling settings.

Large-scale annual releases of pheasants Phasianus colchicus and their subsequent management for recreational shooting create various ecological impacts in the UK. While effects at release sites are fairly well understood, dispersing birds may influence areas farther away. If they enter ecologically important but sensitive protected areas (PAs), any negative impacts could be especially harmful. Using tracking data, from 766 birds across 10 sites, we estimated survival and dispersal of released pheasants and applied these patterns to gamebird release records near English PAs to gauge intrusion risk. Of 2,885 registered release sites, just over half lay within 2 km of a PA. A large number of shoots release relatively few birds while a small number release many birds. Thus, numbers expected to enter a particular PA likely depend both on the size of releases and proximity to the PA. We estimate that, at a national level, a maximum of between 525,000 and 784,000 pheasants might be found within PAs very soon after release, representing around 1.7% of all the pheasants released annually. This number declines over the months after release until in February, we estimate that there are between 131,000 and 196,000 pheasants (0.4% of the total release) might be found within PAs. The critical metric by which ecological damage might occur is their density within PAs. Mean densities soon after release averaged 12.0 birds/ha in PAs within 250 m of release sites. This density declined markedly both in time (as birds died) and space (as they moved further from the pen as potential areas increased). By November, densities in PAs 500-1000m from release sites peaked at 0.5 birds/ha, falling to 0.16 birds/ha in February. These estimated densities are around two orders of magnitude lower than those known to cause strong, lasting impacts within release pens. The results are subject to assumptions about movement behaviour, game management and bias in registration. Despite these constraints, considerable local variation exists, with a minority of high-volume release sites very near PAs posing the greatest potential ecological risk.

Despite substantial global commitments to expand protected-area networks, the strategic allocation of limited resources remains challenging. Spatial conservation planning helps identify priority regions that maximise conservation benefits per unit area. Yet, they also tend to neglect two fundamental aspects of conservation: climate-driven range shifts and the representation of environmentally distinct populations within species. Here, we propose a continental-scale conservation planning framework that explicitly accounts for both processes through novel routines implemented in the conservation planning software CAPTAIN. We apply this framework to European crop wild relatives (CWR), for which niche coverage is a focal priority, as it underpins their potential to support agricultural adaptation to future environmental stressors through breeding programs. Comparative analyses on a subset of 186 CWR associated with five focal crops show that accounting for range shifts and niche coverage leads to substantially different conservation priorities from those obtained with a baseline model based on current distributions only. These additions reduced the number of non-protected species by 64%, increased the average protected distribution range by 43%, increased mean niche coverage from 75.8% to 84.5% and reduced the number of species with less than half of their niche protected from 35 to 10. Applied to a more comprehensive checklist of 1,140 European CWRs, the final framework identifies continental-scale priority areas representing 93.5% of these taxa and includes 94.4% of its critically endangered species. Our results highlight the importance of incorporating both temporal dynamics and within-species environmental representation when designing conservation strategies under climate change. RepositoryThe repository will be made publicly accessible after publication at doi: https://10.5281/zenodo.19855597