Abstract Oil palm is one of the main drivers of tropical deforestation, yet there is a lack of spatial evidence of species‐specific habitat loss driven by oil palm plantations. Using an unprecedented high‐resolution global oil palm map, we assessed all the vertebrate species whose habitats overlapped with oil palm plantations globally. We find that 4.5% of forest‐dependent species lost more than 1% of their potential habitats to oil palm. Indonesia exhibited the highest total biodiversity footprint for mammals, birds, and reptiles, while Malaysia ranked highest for amphibians. São Tomé and Príncipe and Sri Lanka presented high biodiversity footprints per hectare. Biodiversity footprints were led by industrial plantations that account for large habitat loss proportions and present significantly higher footprint intensity than smallholder plantations for mammals and amphibians. Our findings demonstrate the pantropical biodiversity footprint of oil palm plantations and call for measures to ensure oil palm sustainable production beyond Southeast Asia.
Abstract Humans usually try to kill wild animals that they perceive as threats. However, many scientific findings have raised doubts that killing solves wildlife damage to property. With scant, low‐quality data, uncertainty persists. We reviewed recent studies of killing gray wolves to protect domestic animals. Our inclusion criteria set a minimum threshold for robust study design and transparent data sharing. In five studies in five countries, killing wolves typically had no effect, may have had the desired effect, but more often killing wolves had the counter‐productive undesirable effect of elevating domestic animal losses. We quantified the relative rates of the three outcomes and find the balance in favor of no effect to undesirable effects. We discuss potentially confounding variables and the need for an unbiased, reproducible, transparent, randomized, and controlled trial of predator removal. We find and summarize similarities to other situations from disease transmission to crop damage.
Abstract While ivory poaching has driven elephant declines across much of Africa, South Africa's populations have expanded steadily, sparking debate over ecological impacts and population control. Yet these debates have lacked a harmonized, long‐term national evidence base. To address this gap, we analyzed 30‐year (1991–2021) population trends and current densities across 93 South African elephant populations. Owing to numerous translocations and sustained growth within reserves (mean 5.4 ± 7.9% p.a.), South Africa's national elephant population increased from ~8900 individuals across five reserves in 1991 (95% confidence interval: 5942–11,886) to ~39,400 across 93 reserves in 2021 (28,624–50,494). Most elephants (73%) occurred in Kruger National Park, while most populations (70%) occurred on private reserves/properties. Three findings carry direct management implications. First, most populations remain on a growth trajectory: over the past decade, 42% increased steadily (1%–5% p.a.) and 30% rapidly (>5% p.a.), with the largest 25 (>100 individuals) growing fastest. Second, despite this growth, most reserves support small, low‐density populations: in 2021, 63% ( n = 59) had <50 elephants and 70% ( n = 65) had <0.5 elephants/km 2 . Third, only 12 populations exceeded 1 elephant/km 2 , and these were predominantly clustered in the Greater Kruger region; they were also large (median 620 individuals) and fast growing (median 3.9% p.a.). Overall, South Africa's elephant recovery represents a conservation success. A few high‐density populations may be approaching ecological thresholds and pose management challenges due to their large size, whereas for most populations high densities are not yet a concern and may even underdeliver positive megaherbivore functions. Further work is needed to clarify the context‐specific influence of elephant density, fire, rainfall, and other drivers on ecological impact. Our study highlights the policy value of a single, standardized national dataset that integrates counts and uncertainty to support evidence‐based, spatially differentiated management of this increasingly complex national meta‐population.
Abstract In the United States, wildlife management policies are often developed through citizen wildlife commissions or boards that guide agency decision‐making while balancing diverse stakeholder interests. Despite their central role in wildlife governance, little information exists regarding how these commissions are structured across states or how decision‐making processes are communicated to the public. We conducted a 50‐state assessment of wildlife commission composition, stakeholder representation requirements, policy guidance regarding information sources, and public‐facing descriptions of decision‐making processes. Of the 50 states reviewed, 47 maintain active wildlife commissions, while three states manage wildlife policy directly through agency structures. We found substantial variation in commission composition, appointment mechanisms, representation requirements, regulatory authority, and commission specificity. Geographic representation requirements were common, while explicit stakeholder representation requirements were less frequent and most often focused on agricultural and sportsperson interests. We also found considerable variation in the ways that states provide formal guidance for information use and how states communicate decision‐making processes through public‐facing materials. Many states provide basic information regarding commission membership and meeting schedules but offer limited information regarding how scientific information, stakeholder input, or other considerations are incorporated into commission deliberations. This study provides the first national inventory of wildlife commission structures, policy guidance, and decision transparency across the United States. By documenting variation among states, this research establishes a baseline for future studies examining relationships between governance structures, stakeholder trust, public participation, and broader wildlife governance outcomes.
Abstract Conservation science struggles with an ongoing divide between research and action. Given its history of broad public engagement and multi‐scale management coordination, avian conservation typifies many of the challenges, successes, and opportunities across the broader field of conservation and is rich with examples that may guide solutions‐oriented conservation science. In semi‐structured focus groups composed of bird conservation scientists and practitioners from various backgrounds, we discussed the definition, characteristics, and examples of solutions‐oriented conservation science. Our discussions highlighted several examples where successful application of avian conservation science led to change in programs, policy, regulations, and social systems, including several with co‐benefits for local communities. These successes were characterized by communication, collaboration, and co‐development with local communities, decision‐makers, practitioners, and Indigenous communities. We identified remaining opportunities for improvement, including aligning incentives and priorities between research and practice, working across disciplines to mobilize solutions, dismantling barriers to inclusion, and emphasizing the importance of relationship‐building in conservation science. We outline suggestions that researchers can use to incorporate these lessons from solutions‐oriented avian conservation science into training and mentoring programs.
Abstract The conservation of endemic species requires identifying areas within their range affected by human‐modified landscapes, an unresolved challenge across Mexico, one of the world's most biodiverse countries. Although formally protected areas are the backbone of biodiversity conservation, there are often substantial gaps in their overall effectiveness. Here, we conducted a comprehensive national‐scale analysis to assess the effectiveness of conservation lands in mitigating cumulative human impacts on endemic species of plants and animals. We processed national distributional maps for 1639 endemic species, showing the overlap of conservation lands with cumulative human impacts. We quantified overlap between species distributions, conservation lands, and human‐pressure categories and evaluated differences among groups statistically. Our findings reveal that 56% and 26% of all endemic species' distribution ranges overlap ecoregions experiencing medium and high human pressure, respectively, while only 18% occur in regions with low human pressure. Biome‐level analysis shows that 87% of endemic species' ranges overlap with biomes under medium human pressure, with just 13% persisting in low‐impact biomes. Despite this, Protected Natural Areas overlap ∼10% of these species' ranges, while Voluntary Conservation Use Areas and other effective Area‐Based Conservation Measures overlap with ∼1% and 3%, respectively. Our results indicate that existing conservation lands are insufficient to safeguard Mexico's endemic species, including those most at risk of extinction, thereby foreshadowing localized extirpations and potential species loss. This underscores the urgent need for a substantial expansion of conservation networks to secure biodiversity in a rapidly human‐dominated landscape; however, achieving this goal will be politically, economically, and socially demanding without substantial national and international support. Our study advances understanding of these conservation challenges and, despite the potential difficulties, highlights the necessity for immediate, proactive measures to ensure the long‐term persistence of Mexico's endemic species.
Abstract Manta and devil rays (collectively, mobulids) are circumglobal, planktivorous pelagic rays. Mobulids are declining globally due to targeted and non‐targeted capture. However, limited and inaccurate data have hindered species‐specific fishery trend assessments, especially in data‐limited fisheries where unregulated bycatch and low monitoring make assessments challenging. Assessing mobulid fishery trends is becoming critical due to intensifying fishing pressure, to understand population declines and guide protective measures. We used a combination of morphometric, demographic, and fishery metrics to assess how the population status of mobulids has changed temporally in India, the largest fishing nation for these species globally. Our results suggest a decline in Mobula mobular, Mobula tarapacana, Mobula birostris, and Mobula thurstoni populations between 2013 and 2023, and age‐class truncation in M. mobular populations since 2013. These population declines have occurred alongside increasing fishing pressure, suggesting overfishing is responsible. This study is the first to use multiannual data to empirically assess changes in mobulid populations in India, reinforcing the need to include all mobulid species in India's no‐take national legislation. Our research approach can be applied to other data‐limited systems, demonstrating that alternative data can effectively assess population status even when reliable total catch data are unavailable, a common challenge worldwide.
Abstract Computer vision is increasingly adopted to automate wildlife image classification, but its ecological reliability remains largely untested. We evaluated SpeciesNet, a global wildlife classifier on widespread mammal species of Florida, USA: the white‐tailed deer (Odocoileus virginianus), wild pig (Sus scrofa), coyote (Canis latrans) and bobcat (Lynx rufus). While conventional computer‐vision evaluation scores were high (e.g., F1 = 98.2% for the white‐tailed deer), they masked species‐specific false‐positive rates (FPRs) that drove bias in downstream occupancy estimates. The white‐tailed deer with FPR = 0.14% produced the largest divergence from the human‐labeled data, whereas bobcat, with a lower F1 (94.5%) but FPR = 0.01%, yielded near‐identical estimates. We show that FPRs, not conventional metrics like F1‐scores, are key factors for occupancy estimate accuracy when using standard models. Our findings highlight the limitations of conventional performance metrics when evaluating AI tools for ecological applications and the need of false‐positive‐adequate methods as AI‐classifiers gain popularity. Our analysis shows that false‐positive‐informed occupancy models quickly outperform standard models as FPRs increase. Standard models still retain an advantage at low false positive rates (<0.35% in our simulation conditions) most notably when detection probability is low. The benefits of false‐positive occupancy models are highest for species with low occupancy (ψ = 0.3) where false positives are more likely to randomly occur in unoccupied sites. Given that currently available species classifiers likely yield FPRs higher than 0.34%, our analysis indicates that false‐positive‐informed occupancy models would outperform standard occupancy models in analyses using partially human‐verified AI‐labeled data.
Abstract Modern threats to conservation are often global; however, relevant conservation decisions and actions typically occur within local jurisdictions. The socio‐ecological complexity of these problems requires collaboration across large landscapes, diverse community interests, and geopolitical boundaries, yet a lack of interoperability of planning products among decision makers, jurisdictions, and objectives impedes collaboration. Here, we provide a working model for the collaborative, iterative development of a landscape conservation design (LCD) in the Midwest region of the United States, the Midwest Conservation Blueprint, and demonstrate its applicability to the coordination of voluntary conservation actions and investments across jurisdictional boundaries. We used spatial prioritization software to synthesize over 20 datasets to develop a basemap of priority lands and waters for conservation across a diverse set of societal and ecological values (e.g., water quality, biodiversity, recreation). The Midwest Conservation Blueprint is a living plan, updated annually based on improvements in underlying data, understanding of local conditions, and public input. Landscape‐scale planning is not designed to replace local knowledge or supersede decision‐making authorities. However, by synthesizing existing data that represent diverse socio‐ecological values and planning at a broader scale, the Midwest Conservation Blueprint documents where there are shared priorities and opportunities for collaborative landscape conservation.
Abstract The Black‐faced Spoonbill (Platalea minor), a globally endangered species, classified as Endangered on the International Union for Conservation of Nature Red List as of 2024 and reassessed as Vulnerable on June 18, 2025, is facing increasing threats in the Democratic People's Republic of Korea (DPRK) due to habitat degradation, human activities, and reclamation projects. The species relies on the DPRK's coastal wetlands and tidal mudflats for breeding, migration stopover, and over‐summering. While some areas are designated as National Migratory Bird Reserves, management measures have been limited, reducing their effectiveness. Previous data from 1997 to 2001 indicated significant Black‐faced Spoonbill populations at key sites, yet recent assessments are scarce, with only a few surveys conducted after 2016. Extensive reclamation projects have significantly reduced suitable habitats, particularly in Ryongmae Island Mudflat, South Hwanghae Province, Southwest coast of DPRK, and Honggon Islands, North Phyongan Province, Northwest coast of DPRK. The lack of up‐to‐date field data poses challenges for conservation efforts. While DPRK has established environmental protection laws and made international commitments, balancing conservation with economic development remains a challenge. This study highlighted the urgent need for comprehensive ecological surveys, stronger habitat protection policies, and international collaboration to mitigate habitat loss and ensure the long‐term survival of the Black‐faced Spoonbill in the DPRK.
Abstract Choosing management actions to protect biodiversity requires decisiveness amidst complexity. To identify barriers to effective conservation, we asked over 150 conservation practitioners working with the Canadian federal and provincial/territorial governments, non‐government organizations, and private industry about their experience with uncertainty when making decisions. Using inductive thematic coding, we identified that “data,” “resources”, and “governance” ranked as the most significant sources of uncertainty for respondents. Practitioners also proposed solutions that would reduce their uncertainty, which included improved funding processes, better collaboration, changes to governance processes, and the use of data and evidence in decision‐making. Conservation practitioners consistently mentioned elements of uncertainty that exist outside of their immediate control, indicating that substantial uncertainty in conservation decision‐making occurs due to broader dynamics in governance and society.
Abstract Conservation programs increasingly require consideration of rehabilitated individuals for translocations to reduce pressure on highly threatened wild populations. Such supplementation remains controversial due to the influence of captivity on natural behaviors. However, determining the relative conservation value of translocations requires quantifiable measures of success through post‐release monitoring over defined timescales. This case‐study evaluates the short‐term translocation success of captive‐origin cheetah ( Acinonyx jubatus ) after phased‐release reintegration into the wild in South Africa. After phased‐release training, collared cheetah ( n = 12) were released into fenced metapopulation reserves ( n = 7) and monitored intensively (2019–2021). Monthly home‐range size and overlap, as well as habitat use stabilized within 6 months for all individuals, with significant deviations representing confirmed denning events. A metapopulation‐specific cheetah body condition score was used in parallel with kill‐site cluster analysis to support direct prey consumption observations as a novel monitoring tool. While three cheetah suffered natural mortalities during the monitoring period, all females successfully mated within the first 4 months post‐release. To date, these reintegrated cheetah have contributed toward 36 first‐ and 19 second‐generation cubs, 33 of which reached independence and integrated into the wild‐born metapopulation, supplementing growth, but also providing critical demographic and genetic resilience. Rapidly evaluating such translocation success through monitoring directly informs post‐release carnivore management.
Abstract Cities offer viable habitat and are frequently located in biodiversity hotspots, but are often overlooked in conservation planning. Here, we investigated how urban greenspaces can support conservation goals within the Conservation Lands Network, a regional biodiversity conservation plan for the San Francisco (SF) Bay Area, which classifies lands that are either essential for or contributing to meeting biodiversity conservation goals using Marxan model selection. We compared vegetation (normalized difference vegetation index [NDVI]), biodiversity, and connectivity using an indicator species framework for the three most populous cities (Oakland, SF, and San Jose) in the Bay Area in three categories: greenspaces selected by Marxan as essential, greenspaces not selected by Marxan as essential, and all greenspaces. We found that greenspaces selected as essential were significantly larger in area and had a higher mean NDVI than non‐essential greenspaces. Indicator species richness was not significantly higher in greenspaces selected as essential in all three cities, while in SF, non‐essential and all greenspaces had significantly more biodiversity than essential greenspaces. Connectivity for three indicator species also significantly improved in all cities when incorporating non‐essential greenspaces. Our results demonstrate that including urban greenspaces can bolster regional conservation goals, supporting state and federal goals.
Abstract In multilateral environmental agreements, procedural fairness depends not only on a Party's right to attend but on its capacity to participate. We examined which Party‐level traits predict delegation size at the Convention on International Trade in Endangered Species (CITES) Conferences of the Parties (CoPs) and how delegation size shapes voting and working group participation when Committees meet concurrently. Using Bayesian multilevel models, we analyzed attendance, voting records, and working‐group rosters for CoP16–20. On average, 10% of Parties sent no delegates and 16% sent just one, while Committees I and II overlapped for 80% of session time. Small island developing Parties more frequently sent one‐person delegations. CITES working‐language match, OECD membership, and stronger national CITES legislation predicted larger delegations. Voting and working group participation rose with delegation size. One‐person delegations were more often absent, or ineligible to vote due to credential issues, with Committee II most affected. Under present‐and‐voting rules, coordinated blocs and sparse presence can shape outcomes. Because the most constrained Parties are often lower‐middle‐income range states expected to implement listings, these gaps weaken input legitimacy and voluntary compliance. Options include credential pre‐checks, retargeted sponsorship ensuring at least two delegates per Party, and scheduling or hybrid adjustments.
Abstract Evaluating the impact of human‐caused removals on wildlife is challenging, especially when population data are limited, future monitoring is uncertain, and habitat conditions are changing. The East Greenland (EG) polar bear subpopulation contains 2275 animals (95% CI = 1360–3807) and is a subsistence resource for Indigenous Peoples. We developed a state‐dependent modeling and management framework to evaluate different harvest strategies for EG bears. Our approach considered sea‐ice loss and followed international consensus principles for the sustainable harvest of polar bears. Over the next 34 years (approximately three polar bear generations) and assuming a 2:1 male‐to‐female harvest sex ratio, the starting sustainable harvest ranged from 63 bears/year in the absence of scientific monitoring, to 93 bears/year if updated demographic parameters are estimated after 15 years and used to adjust the harvest, if necessary. We also identified an intermediate harvest strategy of up to 78 bears/year that requires a moderate level of monitoring and uses a threshold harvest rule. Our findings demonstrate how monitoring can make it possible to safely implement higher sustainable harvest levels (i.e., approaching maximum sustainable yield). Our methods are relevant to wildlife managers seeking to balance demographic risk, investment in monitoring, and opportunities for subsistence use.
Abstract Appendix III of The Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) allows any Party to unilaterally extend its domestic species protections internationally without requiring consensus at the Conference of the Parties. Despite this advantage, Appendix III remains marginal, comprising approximately 1.3% of CITES‐listed species. We present the first global synthesis of Appendix III use since 1975, drawing on the complete historical record of listings, delistings, and uplistings. Appendix III is most effective for endemic species where origin is unambiguous and it frequently functions as a staging mechanism toward Appendix II or I, with 96% of uplisting proposals succeeding. Longstanding perceptions of ineffectiveness largely reflect early over‐listing, which has since been corrected. We argue that Appendix III is underused not because it is inherently weak, but because many Parties may be unaware of how effectively it functions when applied under the right conditions, and we outline a framework for strategic deployment.
Abstract In 2016, Ecuador established a no‐take marine sanctuary in the northern portion of the Galapagos Marine Reserve (GMR), ostensibly to protect sharks. Arguably because of this, experimental longlining was permitted in the rest of the reserve as part of a compensation package to local fishers. Considering this agreement and very low bycatch rates inside the GMR, we question the efficacy of creating a large no‐take area in the existing reserve. To examine the potential efficacy of alternative management strategies to protect sharks, we constructed a stage‐structured demographic model for three shark species (scalloped hammerhead [Sphyrna lewini], blacktip [Carcharhinus limbatus], and silky [Carcharhinus falciformis]). We used the model to evaluate the efficacy of mortality reduction based on a spatial management approach and a hypothetical alternative bycatch mitigation scenario. We found that when adult dispersal rates were high, the sanctuary scenario generated lower absolute changes in the population growth rate than the bycatch mitigation scenario. Our results show bycatch reduction strategies likely have a higher conservation impact compared to the establishment of a sanctuary inside a reserve where sharks are already legally protected. With this case study, our work underscores the promising role of demographic models in informing management decisions for marine conservation.
Abstract Community‐based social marketing (CBSM) is increasingly used in conservation to influence human behavior, yet its biological effectiveness is rarely evaluated. We assessed relationships between CBSM campaigns and both behavioral and demographic responses of six shorebird species across five U.S. Atlantic coast states during the 2024 breeding season. Using nearly 1000 point counts, >850 behavioral observations, and nest monitoring data analyzed with generalized additive models, we evaluated patterns across three CBSM campaigns (dog‐leashing, dog‐free beaches, and walk‐around‐flocks) alongside field‐based stewardship presence. Walk‐around‐flocks campaigns coincided with significantly lower shorebird vigilance, reduced variability in vigilance, and evidence consistent with improved fledging success for the federally listed Atlantic coast Piping Plover (Charadrius melodus). In contrast, dog‐leashing campaigns showed mixed outcomes: leashed dog presence was negatively related to shorebird abundance, and apparent campaign benefits diminished at sites with high numbers of unleashed dogs. Steward presence also corresponded with reduced disturbance and lower bird vigilance. Although not all interventions produced consistent responses, our findings suggest that CBSM strategies promoting bird‐friendly, conflict‐avoidance behaviors, particularly when paired with stewardship, may help reduce human disturbance and support improved conditions for shorebirds during critical breeding periods.
Abstract Monitoring wildlife abundance is important for understanding species population status and trends and, thus, for informing conservation planning and wildlife management. Yet, abundance monitoring can be challenging and costly to implement, particularly for wildlife of conservation concern occurring in remote and rugged terrain. Remotely‐sensed proxies, such as the relative abundance index (RAI) based on camera‐trap data, can be a useful and cost‐efficient approach in such situations. However, RAI is typically assumed to be linearly, continuously, and positively related to actual abundance—the assumptions that are rarely tested. In this study, we explored how well RAI from camera‐traps corresponds to direct field estimates of the abundance of the bezoar goat (Capra aegagrus) in the rugged mountains of the South Caucasus. Specifically, we summarized the data of RAI and count estimates obtained during the same one‐month periods (all in winter) and 5 × 5 km2 grid cells in 2018, 2019, and 2022, and used linear regression to determine the relationship between both measures. We found that the RAI‐abundance relationship was reliable, with a positive, statistically significant, accurate and precise effect, and a large effect size. We conclude that camera‐trap based RAI can be a reliable metric for winter monitoring of bezoar goats in remote and rugged mountain regions such as the South Caucasus. This offers a widely applicable, relatively low‐cost pathway to upscale mountain ungulate monitoring to larger areas to inform conservation planning and wildlife management.
Abstract Nocturnal seabirds, particularly Procellariiforms, are vulnerable to being negatively impacted by artificial lights as they are prone to encounter various forms of light pollution in coastal and pelagic habitats, leading to strandings or other changes to their normal behavior. While adopting blackout periods can be effective to mitigate the impacts at brightly lit sites, this is not always an option when human health and safety requirements during night‐time operations are a concern. We exposed adult Leach's Storm‐Petrels ( Hydrobates leucorhous ) to four light (blue, bright white, dim white, and red) and a dark (no light) treatment to investigate whether we could influence their frequency of occurrence, quantified using a thermal imaging camera, while birds were naturally active at a colony. We used a generalized linear regression to examine the relationship between the number of birds detected per minute and light treatments. The treatments of blue and bright white versus dark were significant covariates with the mean number of birds per minute detected highest during the dark treatment (1.52, LCL = 1.13, UCL = 1.91) and lowest during the blue treatment (0.93, LCL = 0.69, UCL = 1.16), suggesting avoidance to these lights. Our study adds to the growing scientific support that seabirds show a stronger sensitivity to lights with shorter wavelengths and demonstrates the potential of using blue lights as a mitigative approach to reduce adult strandings at industrial sites without compromising the safety of night‐time operators.
