Nevertheless, the exercise demonstrates how data already collected for marine monitoring can be organized within the United Nations System of Environmental-Economic Accounting, known as SEEA Ecosystem Accounts. The work responds to a measurement problem in economic and environmental policy. Ocean ecosystems provide goods and services that support livelihoods, including fisheries, aquaculture, climate regulation, coastal protection and recreation.

Many of these contributions do not arise through direct extraction or use, making them difficult to represent in conventional economic accounts. The resulting omission has contributed to decisions that prioritize economic values over ecosystem degradation. International strategies for sustainable ocean development have therefore promoted integrated assessments framed in terms of natural capital and ecosystem services.

SEEA was developed as a complementary accounting standard to the System of National Accounts, using regular, objective and internationally comparable data to recognize nature’s contribution to economic and social activity. SEEA Ecosystem Accounts organize information into linked components. Extent accounts measure the area occupied by different ecosystem types, while condition accounts describe selected characteristics of those ecosystems at particular times.

Service accounts quantify flows of ecosystem services in physical or monetary terms and identify the households, businesses and government bodies that use them. Asset accounts estimate the monetary value of ecosystem stocks and their additions or reductions over time. The standard is endorsed for use in more than 90 countries, and 41 countries had compiled ecosystem accounts by 2023, but most existing work concerns terrestrial environments.

Marine accounts pose additional difficulties because habitats extend through the seabed and water column, their boundaries are not always concentrated at the surface, and resources and pressures cross national jurisdictions. The OSPAR Convention for the Protection of the Marine Environment of the North-East Atlantic provides a regional setting for testing these methods. It has 15 Contracting Parties and includes the European Union.

Its maritime area covers more than 13.5 million square kilometres, encompassing the internal waters, territorial waters and exclusive economic zones of Contracting Parties, adjacent high seas, and areas extending into the Wider Atlantic and Arctic Waters. Researchers assessed existing accounts, consulted technical experts and worked through OSPAR’s Intersessional Correspondence Group on Economic and Social Analyses. They combined guidance from SEEA and the Global Ocean Account Partnership with data from OSPAR’s Data and Information Management System, European seabed maps, European Environment Agency extent estimates, and marine natural capital accounts from the Netherlands and United Kingdom.

The resulting extent account covered more than 9.2 million square kilometres, approximately 68% of the OSPAR area. Deep-sea sedimentary habitats were the most extensive mapped class, representing 46% of the accounted area. Infralittoral and circalittoral rocky habitats together accounted for only 0.006%, or 5,281 square kilometres.

The researchers represented seabed habitats in two dimensions rather than attempting a full account of the water column and seabed as three-dimensional systems. EUSeaMap 2019 data provided the principal habitat classification through the EUNIS system, while separate European Environment Agency data supplied estimates for intertidal ecosystems and coastal estuaries. The available data were not aligned with the IUCN Global Ecosystem Typology endorsed by SEEA, illustrating a tension between international classification standards and the marine datasets that can currently be assembled.

A central limitation prevented the compilation of ecosystem condition accounts. OSPAR already reports indicators including acidity, eutrophication, species diversity, sea-surface temperature, coral condition, seagrass cover and the state of fish stocks, but many are aggregated across the entire North-East Atlantic or broad subregions. These regional measurements cannot be assigned to the specific habitat classes used in the extent account without spatially disaggregated data.

The researchers recommend shifting future collection toward ecosystem-based reporting so that condition indicators can be linked directly to mapped assets. They also note that the revised EUNIS classification, which distinguishes benthic and pelagic components, could support future work on biotic habitats such as kelp forests and oyster reefs, as well as assets and services in the water column. The physical service accounts selected for OSPAR priorities and data availability had markedly different temporal coverage.

Wild fish provisioning was estimated for 2012 through 2019 and fluctuated between about 8.1 and 9.3 million tonnes, reaching its highest reported level in 2018 and its lowest in 2019. Aquaculture provisioning was available for 2012 through 2015 and remained close to 2.12 to 2.15 million tonnes. Carbon sequestration and capture could be estimated only for 2019, while outdoor recreation was available only for 2012.

The fish figures were based on reported landings, which may omit commercially caught species and could not be assigned to the ecosystems that supplied them. Carbon estimates omitted seagrass and excluded major habitats containing significant carbon sinks, making the regional total an underestimate. Recreation used potential annual visits as a proxy and could not be extended to the whole OSPAR area.

Monetary values were derived mainly through benefit transfer, a method that applies estimates from a study site to another location with similar characteristics. United Kingdom resource-rent values informed the wild-fish estimate, a Norwegian aquaculture study supplied values for aquaculture products, an efficient carbon price from a low-reduction scenario was applied to sequestration, and the INCA project provided recreation values for European Union members that were also OSPAR Contracting Parties. Fish provisioning was valued between €1.7 billion and €2.8 billion, with no clear relationship between changes in physical and monetary values.

Aquaculture’s monetary value increased from €1.3 billion to €3.7 billion between 2012 and 2015 despite a nearly stable physical flow. Carbon sequestration and capture was valued at €1.6 million in 2019, while outdoor recreation in OSPAR countries with available data was estimated at €253 million in 2012. Using the net present value of expected future service income, the researchers estimated the value of ecosystem assets across the OSPAR region at €125.75 billion.

Carbon sequestration represented €48.15 billion of this preliminary total, and outdoor recreation represented €7.56 billion; together, they contributed more than 40%. These are modelled asset values, not market prices for the North-East Atlantic, and the incomplete carbon and recreation estimates imply that the asset totals are also underestimated. The study further cautions that absolute monetary values depend strongly on the valuation method and may be less informative than changes over time.

For example, lower landings in 2016 received a higher resource-rent value than higher landings in 2015 because the method is connected to factors such as fuel prices rather than directly representing fish abundance or ecosystem health. Beyond the numerical estimates, the exercise exposed a set of practical requirements for regional ocean accounts. Future data should align more closely with international ecosystem and service classifications; condition indicators should be spatially linked to named habitats; and OSPAR Contracting Parties should use common methods, tools and reporting practices.

The convention’s existing data platform could support more systematic collection and collation, while repeated and coherent reporting would permit future accounts to track changes rather than provide isolated snapshots. Higher-resolution ecosystem mapping and remote sensing could improve estimates of extent change, and three-dimensional accounting could clarify flows involving dynamic stocks such as fisheries. The study’s principal contribution is therefore a tested blueprint for assembling SEEA-aligned accounts for a regional sea, accompanied by a clear inventory of the data, harmonization and valuation work still required.

Subject of Research: Preliminary SEEA ecosystem accounting for marine and coastal ecosystems in the OSPAR North-East Atlantic region

Article Title: Compiling preliminary SEEA Ecosystem Accounts for the OSPAR regional sea: experimental findings and lessons learned

Article References: Alarcon Blazquez, M., van der Veeren, R., Gacutan, J., & James, P. (2023). Compiling preliminary SEEA Ecosystem Accounts for the OSPAR regional sea: experimental findings and lessons learned. One Ecosystem, 8, Article e108030. https://doi.org/10.3897/oneeco.8.e108030

Image Credits: AI Generated

DOI: 10.3897/oneeco.8.e108030

Keywords: SEEA Ecosystem Accounts, OSPAR, North-East Atlantic, marine natural capital, ecosystem services, environmental accounting, fisheries, carbon sequestration, outdoor recreation, Compiling, preliminary, SEEA

The study, led by Dr. Mirjam Barrueto as part of her doctoral research at the University of Calgary, spanned western Alberta and eastern British Columbia, encompassing the Rocky Mountains and the Columbia Mountains. Co-authors include her supervisor, Dr. Marco Musiani, an adjunct professor in the Faculty of Science, along with scientists from Parks Canada, the Nature Conservancy of Canada and Montana State University. Together, the team set out to answer a deceptively simple question: how is the wolverine population actually doing, and what kinds of human pressures are shaping its numbers? The answer, drawn from one of the most extensive wolverine surveys ever conducted, is that the animals are far scarcer than the seemingly wild landscape would suggest, and that two distinct forms of human disturbance are leaving measurable fingerprints on the population.

The headline numbers are stark. Across 81,800 square kilometres of mountain terrain, the researchers estimate that fewer than 300 wolverines currently make their home. That figure, low as it is, may actually understate the loss: according to Barrueto, the population could be roughly 30 per cent higher if current levels of recreation and road development were absent. In other words, in the absence of the trails, ski operations, highways and resource roads that now thread through these mountains, the region could plausibly support dozens more of these solitary predators. The finding is particularly striking because it emerges at the scale of entire landscapes rather than individual sites, meaning the effect is not simply that wolverines shift their movements to avoid a busy trailhead or a noisy road corridor on a given day.

That distinction matters enormously for how the results should be interpreted. As Barrueto explains, the analysis shows that in areas with more disturbance, whether from road density or recreation, there were simply fewer individual wolverines living there than in areas with less disturbance. The animals are not merely redistributing themselves within the landscape, crowding into quiet corners while the overall total stays the same. Instead, disturbed areas support genuinely smaller populations, a pattern that suggests the disturbance is functionally removing habitat or imposing costs, such as stress, energetic expense or reduced reproductive success, that translate into fewer animals able to persist. On a population scale, that is a fundamentally different and more consequential outcome than localized avoidance behavior.

Measuring recreation across such an enormous and remote area presented a methodological challenge, and the research team turned to an increasingly common tool in wildlife science: publicly available heat maps generated by recreational apps. These digital traces capture the intensity of human activity in the backcountry, including backcountry skiing, heliskiing and cat skiing, cross-country skiing and hiking in both winter and summer. Notably, the recreation data does not include snowmobiling, but Barrueto points out that snowmobile activity is likely captured indirectly in the road density variable, since snowmobiles depend on roads and trails for access. The approach demonstrates how crowdsourced recreational data, originally designed to help people navigate and share their outings, can be repurposed to quantify one of the most diffuse and hard-to-measure pressures on wildlife across entire mountain ranges.

While the recreation finding came as a surprise even to the researchers, the role of roads was more expected. Barrueto notes that road density emerged as the other main predictor of wolverine population density, and that it was also a key predictor for females specifically, a result consistent with previous studies. Roads have long been recognized as a powerful driver of ecological change, fragmenting habitat, facilitating human access into remote areas, and increasing mortality risk. What the new study adds is a clearer picture of how these two disturbance types, one ancient in origin and one distinctly modern, combine to shape the distribution and abundance of a carnivore that evolved in landscapes without either.

One of the most important new insights from the research is that the impacts of disturbance are sex specific. Road density, in particular, affects female wolverines the most, and that matters far beyond simple demographics. Females are the engine of population recovery, and wolverines are among the slowest-reproducing carnivores in North America: females give birth to only one to three kits every few years, meaning that any reduction in female survival or occupancy ripples through the population over long time horizons. A landscape that loses its females is a landscape that loses its capacity to rebound, even if males continue to move through the area. For wildlife managers, this makes the protection of low-disturbance habitat for females a particularly high-leverage conservation strategy.

The biological realities of the species compound the concern. Wolverines are, as Barrueto describes them, pretty small carnivores, and it generally seems stressful for most of these animals to interact with humans. They are built for remoteness: solitary, wide-ranging, dependent on deep snow and cold conditions, and naturally occurring at low densities even in pristine habitat. That combination makes them inherently vulnerable to cumulative human pressure. An animal that reproduces slowly, occupies enormous home ranges and avoids areas of human activity has little capacity to absorb additional disturbance without its population density declining. The study’s findings suggest that the wolverines of the Canadian Rockies are already living close to that edge, with the Upper Columbia region of southeastern British Columbia standing out as an area with very few animals.

The research does not stop at diagnosis. It makes five recommendations aimed at giving managers and planners concrete tools, including maintaining roadless areas and areas of low road density, protecting low-disturbance, high-quality wolverine habitat, and limiting recreation in reproductive areas. Barrueto frames the work as a call to awareness for anyone who works or plays in wolverine habitat, herself included, and poses the question at the heart of the matter: do we want a landscape with just deer and coyotes, or do we want a landscape that has space for wild things? Now that the impacts are understood, she argues, management and planning decisions can be made deliberately so that wolverines can thrive in the places people love.

That actionable quality is what resonates most with co-author Dr. Aerin Jacob, director of science and research and Weston Family senior scientist at the Nature Conservancy of Canada. Describing the findings as stark, Jacob emphasizes that the study provides something concrete to act on at every level, whether that means decisions about building trails, protecting and restoring habitat, or working to prevent and reduce impacts. In her view, that kind of collaboration among scientists, land managers and conservation organizations is what effective conservation actually looks like. For a species that most visitors to the Rockies will never glimpse, the study offers both a warning and a roadmap: the quiet presence of wolverines in these mountains is not guaranteed, but with deliberate choices about roads, recreation and habitat protection, it can endure.

Subject of Research: Effects of outdoor recreation and road density on wolverine population density in the Canadian Rocky and Columbia Mountains

Article Title: Study shows wolverine populations affected by recreation, roads

Article References: Study shows wolverine populations affected by recreation, roads. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: wolverine, outdoor recreation, road density, carnivore conservation, Rocky Mountains, Columbia Mountains, population density, Biological Conservation, University of Calgary, habitat disturbance, wildlife management, backcountry skiing

The study was conducted by Jonathan Hughes, an economist at the University of Colorado Boulder and a fellow at the Renewable and Sustainable Energy Institute. Hughes, who has spent years navigating these very reservation systems himself, examined eleven reservation systems spanning more than five million reservations across seven national parks, including Rocky Mountain, Glacier, and Yosemite. His central finding is striking: among systems that offered advance reservations, people who secured permits weeks or months ahead of their visits tended to come from ZIP codes with notably higher median household incomes than those who obtained permits on the same day or just one day in advance.

The income gap was most pronounced for the scarcest reservations. For highly competitive permits, where the majority of available tickets sell out within five minutes of release, successful reservation holders came from ZIP codes with median household incomes roughly three to six percent higher than those associated with less competitive reservations. That may sound like a modest difference, but across millions of transactions it represents a systematic skew in who gets access to parks that belong, by law and by founding ideal, to everyone. Hughes points to the organic act that established Yellowstone National Park, the nation’s first, which set the land apart as a public park for the benefit and enjoyment of the people, not just rich people.

The timing of the study is significant. Rocky Mountain National Park, one of the busiest in the system, is preparing to end its timed-entry permit system for the season on October 18. Under that system, during peak hours from mid-May to mid-October, visitors must obtain a permit costing two dollars, in addition to an entrance pass, to enter the park during a specific time window. The park releases some permits on the first day of each month for the remainder of that month, while another batch becomes available the evening before each entry date. This two-tier release structure is common across the National Park Service and is precisely where Hughes identified the equity gap.

The pressure on these systems has grown dramatically over the past decade. National parks in the United States have experienced steady increases in visitation since around 2013, and the COVID-19 pandemic pushed attendance even higher as Americans sought open-air recreation. In 2024, more than 331 million people visited national parks. That surge has placed real strain on park infrastructure: crowded roads, trampled ecosystems, overwhelmed restrooms, and emergency response operations complicated by gridlock. In response, many parks have turned to online reservation systems for parking, popular trails, and entry during peak periods, effectively rationing access through a combination of technology and advance planning.

Hughes’s interest in the question was personal as well as professional. He recalls arriving on campus early to take advantage of faster WiFi connections and running two computers simultaneously to maximize his odds of securing a permit. Those experiences prompted him to think about who is actually able to participate in such a process. Not everyone, he notes, can plan trips months in advance, is comfortable navigating online booking platforms, or has access to a fast and reliable internet connection. The reservation race, in other words, rewards resources that are unevenly distributed: flexible schedules, digital literacy, and the financial security to commit to a trip long before it happens.

The new findings build on Hughes’s earlier research into river permit systems in the American West, which govern access to popular boating and fishing stretches through a combination of lotteries and advance reservation systems. In that work, he found that people who secured permits through advance reservation tended to have higher incomes than those who obtained them through lotteries. The parallel is telling. Advance reservation systems, by design, favor those with the foresight, flexibility, and technical skill to act the moment bookings open. Lotteries, by contrast, randomize the outcome and appear to distribute access more evenly across income groups.

The mechanism behind the bias is likely a combination of factors, Hughes suggests. Higher-income visitors tend to have more flexibility to plan trips months ahead, whether because their jobs allow predictable vacation scheduling or because they can absorb the risk of a nonrefundable booking. They are also more familiar with the intricacies of reservation systems, knowing exactly when permits are released and how to move quickly when they are. And they tend to be more tech-savvy, with better devices, faster connections, and greater comfort competing in a five-minute digital sprint. Each of these advantages compounds the others, producing the measurable income gradient Hughes observed across the five million reservations in his dataset.

The implications for park managers are considerable. As the National Park Service continues to adjust its reservation systems, Hughes argues that the findings could help administrators think differently about how permits are distributed. One option is to experiment with the balance between permits released months in advance and those released closer to the visit date, potentially giving spontaneous or less well-connected visitors a fairer shot. In some cases, he suggests, random lotteries may provide a more equitable alternative to first-come, first-served advance booking, echoing the pattern he documented in Western river systems. The goal is not to eliminate reservations, which have proven effective at managing crowds and protecting fragile resources, but to design them so that the benefits of access do not accrue disproportionately to the wealthy.

For Hughes, the issue ultimately goes to the heart of what national parks are meant to be. For many years, he notes, the people responsible for stewarding America’s public lands have worked to give everyone equal access, and any indication that some people have better access than others runs counter to that mission. Income, he cautions, is just one indicator; education, age, and race could also play a role in who succeeds in the reservation race, and those dimensions deserve further study. As parks grapple with record visitation and finite resources, the challenge will be to manage crowds without quietly converting shared national treasures into amenities reserved for those with the time, money, and technology to outcompete everyone else. The study, published in Nature Communications with the DOI 10.1038/s41467-026-76030-4, offers park managers a data-driven starting point for redesigning systems so that the benefit and enjoyment of the people extends to all of the people.

Subject of Research: Income-based inequality in national park reservation and permit systems

Article Title: Richer people have an edge when it comes to getting national park permits

Article References: Richer people have an edge when it comes to getting national park permits. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: national parks, reservation systems, income inequality, Rocky Mountain National Park, timed-entry permits, Nature Communications, recreation equity, National Park Service, lotteries, outdoor recreation, Jonathan Hughes, public lands

The research tackles a stubborn bottleneck in outdoor recreation management: the concept of carrying capacity has evolved well beyond the search for a single maximum number of visitors. Modern practice, codified in the Interagency Visitor Use Management Framework, asks managers to define desired conditions, select measurable indicators, set thresholds, and link monitoring to adaptive actions. Yet implementation routinely stalls at the measurement step. Manual counts are expensive, automated counters suffer from group undercounting and false triggers, and neither approach tells managers where vehicles can plausibly park, when parking fills, or how use concentrates along a linear corridor that may stretch for many kilometers. The Payette River, with its fragmented jurisdiction spanning the USDA Forest Service, the Bureau of Land Management, state transportation agencies, and local law enforcement, exemplifies the problem.

The authors built two complementary workflows. The first quantifies what they call parking opportunity, the spatial extent of surfaces where a vehicle could realistically be parked, using high-resolution aerial imagery from the National Agriculture Imagery Program at 0.6-meter resolution. The analysis area was defined by intersecting 500-meter buffers around major highways and the river itself, then constrained to terrain with slopes of five percent or less using a one-meter digital elevation model from the USGS 3DEP program. This slope filter was applied twice, first to screen candidate training labels and again to mask out implausible predictions on steep hillsides during inference, ensuring the model never reports parkable surface on a cliff face.

Label development for the segmentation workflow relied on the Segment Anything Model, or SAM, in its largest ViT-H configuration. SAM generated 12,942 candidate object proposals from the tiled aerial imagery, but manual review revealed that most of these were trees, a striking illustration of why general-purpose foundation models still require human curation when applied to infrastructure mapping. After screening, 817 polygons survived with meaningful labels, of which only 16 represented paved parking and 50 represented dirt or gravel parking. This severe scarcity of positive examples shaped the modeling strategy: the team trained a DeepLabV3+ network with a ResNet-34 encoder using a weighted multi-class cross-entropy loss that penalized parking misclassification heavily, and applied on-the-fly augmentations including flips, rotations, affine transforms, and brightness perturbations to squeeze more effective diversity from the sparse labels.

The final segmentation model achieved a mean intersection-over-union of 0.64 across classes, with 0.60 for paved parking and 0.56 for dirt and gravel parking, substantially better than an unweighted, non-augmented baseline. Post-hoc checks found that fewer than five percent of model-derived paved parking areas fell within four meters of roadway centerlines, suggesting that confusion between roads and parking lots, the error mode with the greatest practical consequence, was marginal in this setting. The resulting maps revealed that parking opportunity is highly uneven along the corridor, giving managers a replicable, spatially explicit supply-side baseline for the first time. The authors are careful to note that mapped opportunity is not the same as authorized or safe parking, and that field verification of ownership, legality, and stall geometry remains essential before any investment decisions.

The second workflow addresses the demand side, using YOLO26 object detection applied to a large corpus of images from motion-triggered cameras positioned at parking areas and river viewpoints. Rather than labeling images at random, the team employed an active learning strategy: a COCO-pretrained baseline detector produced preliminary detections, and an uncertainty- and diversity-aware selector chose an initial batch of 500 images for human annotation, prioritizing ambiguous predictions, hard negatives likely to cause false positives, and predicted-empty images to stabilize specificity. A second loop of 1,000 river-only images specialized the detector for the notoriously difficult visual conditions of water scenes, where glare, specular highlights, partial occlusion, and clustered watercraft confound naive models. Two annotators achieved inter-annotator agreement above 0.80 on Krippendorff’s alpha, following strict rules such as labeling only the visible portion of occluded vehicles and refusing to label reflections or wake patterns as boats.

Watercraft remained so rare in the imagery that the workflow added a targeted mining phase. The team calibrated a confidence threshold by sweeping candidate values against validation data, then ran large-scale inference across the full river corpus, ranking images by a mining score that favored confident, meaningful detections. High-ranked candidates were annotated and oversampled during a final fine-tuning stage, with mined examples strictly restricted to the training pool to avoid evaluation leakage. The payoff was dramatic: watercraft recall increased substantially in the final iteration alongside precision of 0.843, lifting mean average precision at the 0.50 threshold to 0.554. Vehicle detection, by contrast, was strong throughout, reaching precision of 0.856, recall of 0.733, and mAP50 of 0.865 in the final model, supporting the use of vehicle counts as a reliable comparative indicator of access pressure.

Aggregated results paint a vivid picture of corridor dynamics. Parking occupancy proxies showed sharp within-day pulses aligned with arrival and departure cycles, with the highest demand at Banks and Beehive Bend concentrated on weekends and midday, while sites such as Alder Creek, Danskin, and Deadwood showed much smaller peaks. On the river, watercraft detections clustered into two pronounced windows, mid-morning and mid-afternoon, with the Banks-to-Beehive stretch showing the most intense activity, consistent with typical launch-to-takeout travel times. Calibration against 67 hours of manual hand-tally counts yielded only modest absolute agreement, with a correlation of 0.35 and an R-squared of 0.12, but crucially the temporal ordering of peak-use windows and the relative ranking of sites remained stable between calibrated and uncalibrated series. The authors therefore recommend treating the on-river metric as a relative activity indicator rather than an absolute census of boats.

The study’s broader significance lies in its honesty about what machine learning can and cannot deliver for public land management. The authors explicitly frame their outputs as physical infrastructure and demand indicators, necessary inputs to capacity frameworks but not capacity thresholds themselves, since acceptable levels of crowding and ecological impact require social and ecological standards that no camera can supply. They also confront the ethics of camera deployment in public spaces, recommending placement that avoids capturing faces, resolution limited to detection needs, masking of sensitive regions, signage, and aggressive data retention policies that discard raw imagery once detections are extracted. Post-hoc benchmarks showed DeepLabV3+ outperforming U-Net and FPN architectures for segmentation, while the compact YOLO26n variant, at just 2.5 million parameters versus 32.1 million for an FCOS comparison, delivered comparable vehicle detection, an operational advantage for resource-constrained agencies.

For a field where monitoring capacity has long been the weakest link in adaptive management, the Payette River case offers a template that could extend to trail networks, scenic byways, and other linear recreation corridors wherever aerial imagery and fixed cameras are feasible. The workflow’s active learning strategy, in which a modest amount of targeted annotation informed by uncertainty and hard-example mining yields outsized gains on rare but management-critical targets, transfers naturally to other sparse-event monitoring problems. The authors caution that generalization beyond a single corridor remains a working hypothesis, and that seasonal domain shift, changing vegetation, and camera mounting constraints will demand local recalibration. But the core message is clear and timely: with careful human-in-the-loop design, transparent error accounting, and explicit acknowledgment of uncertainty, computer vision can finally give recreation managers the continuous, spatially explicit evidence they need to diagnose bottlenecks before they become conflicts.

Subject of Research: Computer vision workflows for measuring parking capacity and recreation demand along a linear river recreation corridor

Article Title: Computer vision workflows for physical carrying capacity analysis and monitoring in linear recreation corridors

Article References: Smith, J. W., & Lamborn, C. C. (2026). Computer vision workflows for physical carrying capacity analysis and monitoring in linear recreation corridors. Machine Learning with Applications, 26, Article 101012. https://doi.org/10.1016/j.mlwa.2026.101012

Image Credits: AI Generated

DOI: 10.1016/j.mlwa.2026.101012

Keywords: computer vision, carrying capacity, outdoor recreation, semantic segmentation, object detection, YOLO, Segment Anything Model, DeepLabV3+, active learning, Payette River, visitor monitoring, machine learning