Introduction
Human populations and settlements continue to grow worldwide while the number of places largely free from human disturbance continue to decline, threatening ecosystem services, biodiversity and natural ecological processes (Foley et al. Reference Foley, Defries, Asner, Barford, Bonan and Carpenter2005, Watson et al. Reference Watson, Shanahan, Di Marco, Allan, Laurance and Sanderson2016, Di Marco et al. Reference Di Marco, Ferrier, Harwood, Hoskins and Watson2019). Designated protected areas aim to preserve these ecological benefits, yet protected areas are not invulnerable to human disturbance (Hansen & Defries Reference Hansen and Defries2007, Jones et al. Reference Jones, Venter, Fuller, Allan, Maxwell, Negret and Watson2018, Belote & Wilson Reference Belote and Wilson2020). Therefore, it is important to monitor threats to protected areas in order to identify vulnerabilities and implement strategic management actions.
In the US, wilderness areas have a higher level of protection from human activity than any other protected areas (e.g., national parks, wildlife refuges), making them particularly important for preserving biodiversity and natural ecological processes (Dietz et al. Reference Dietz, Belote, Aplet and Aycrigg2015, Dietz et al. Reference Dietz, Belote and Aplet2023). Yet, the ecological integrity of wilderness areas can be impacted by surrounding land use and by daily visitors (Martinuzzi et al. Reference Martinuzzi, Radeloff, Joppa, Hamilton, Helmers, Plantinga and Lewis2015, Marion et al. Reference Marion, Leung, Eagleston and Burroughs2016, Aycrigg et al. Reference Aycrigg, McCarley, Belote and Martinuzzi2022a). Recreation in wilderness areas presents a challenge because, despite a few wilderness areas being closed to the public, the ability to visit wilderness areas is a key mandate of The Wilderness Act (US Public Law 88-577; Stankey et al. Reference Stankey, Cole, Lucas, Petersen and Frissell1985). The Wilderness Act also requires wilderness areas to be untrammelled, natural, undeveloped and have solitude or a primitive and unconfined type of recreation (Landres et al. Reference Landres, Barns, Dennis, Devine, Geissler and McCasland2008, Marion Reference Marion2016). Recreation can have direct impacts on the natural and untrammelled state of wilderness areas through loss of vegetation and soil erosion, degradation of water quality, disturbance to wildlife and proliferation of garbage and human waste (Marion Reference Marion2016, Marion et al. Reference Marion, Leung, Eagleston and Burroughs2016, Eagleston & Marion Reference Eagleston and Marion2017, Lindley et al. Reference Lindley, Blevins and Williams2018). Adding further complications to this, wilderness visitation also has notable benefits because visitors gain appreciation for wilderness preservation (Watson et al. Reference Watson, Martin, Christensen, Fauth and Williams2015, Racsh & Hahn Reference Rasch and Hahn2018, Miller et al. Reference Miller, Blahna, Morse, Leung and Rowland2022) and make significant economic contributions to local communities (Bowker et al. Reference Bowker, Cordell and Poudyal2014, Holmes et al. Reference Holmes, Bowker, Englin, Hjerpe, Loomis, Phillips and Richardson2016, Hjerpe Reference Hjerpe2018). Thus, monitoring vulnerability to overuse, wherein visitation significantly degrades the natural ecological processes being protected, is a key element in finding a balance between untrammelled wilderness and unconfined recreation (Stankey et al. Reference Stankey, Cole, Lucas, Petersen and Frissell1985, Landres et al. Reference Landres, Barns, Dennis, Devine, Geissler and McCasland2008, Marion Reference Marion2016).
Wilderness areas face unequal vulnerability to impacts from recreation because factors influencing overuse include nearby population growth and accessibility, which together increase the likelihood of visitation (Lindley et al. Reference Lindley, Blevins and Williams2018, Rasch & Hahn Reference Rasch and Hahn2018, Bowker et al. Reference Bowker, Askew, Landry, Hedges and English2022). Between 2010 and 2014, 85.6% of visits to wilderness areas managed by the US Department of Agriculture (USDA) Forest Service were <12 h in duration, 75.9% were <6 h duration and 72.1% of visitors travelled <200 miles one-way (Bowker et al. Reference Bowker, Askew, Landry, Hedges and English2022). This highlights the significance of day-use visitors in terms of both the proportional impacts associated with these visits and the value of understanding nearby populations as an indicator of wilderness area vulnerability to overuse.
Another factor influencing overuse is the appeal of the specific wilderness area, with some wilderness areas being more aesthetically pleasing because of landscape and vegetation characteristics or more popular due to their association with other popular natural attractions, such as national parks (Hanink & White Reference Hanink and White1999, Sonter et al. Reference Sonter, Watson, Wood and Ricketts2016). The appeal of a wilderness area may increase the number of visitors travelling further and spending multiple days, as well as influencing people to move closer to wilderness areas, thereby creating more potential day-use visitors (Radeloff et al. Reference Radeloff, Stewart, Hawbaker, Gimmi, Pidgeon and Flather2010, Holmes et al. Reference Holmes, Bowker, Englin, Hjerpe, Loomis, Phillips and Richardson2016, Mockrin et al. Reference Mockrin, Stewart, Matonis, Johnson, Hammer and Radeloff2018). The number of trails can also be an appeal of certain wilderness areas and has been shown to increase day-use visits (Bowker et al. Reference Bowker, Askew, Landry, Hedges and English2022).
Previous studies have quantified wilderness visitation and examined the relationships with surrounding populations (e.g., Lindley et al. Reference Lindley, Blevins and Williams2018, Rasch & Hahn Reference Rasch and Hahn2018, Bowker et al. Reference Bowker, Askew, Landry, Hedges and English2022), but these studies tend to focus on a subset of wilderness areas based on location or managing agency (e.g., Bureau of Land Management; Rice et al. Reference Rice, Armatas, Thomsen and Rushing2021). In this study, we developed the Day-Use Vulnerability Index (DUVI) for 722 wilderness areas in the contiguous US (CONUS) using eight variables influencing day-use visitation and the potential for visitation to result in overuse. We present the wilderness areas most vulnerable to impacts from day-use visitation based on our index. We also compare the mean DUVI values with visitation estimates from the USDA Forest Service to validate our results.
Materials and methods
Study description
Given data availability, we chose to focus our study on wilderness areas in the CONUS. Furthermore, our focus on wilderness visitation limited our study to wilderness areas accessible to the public. We used the US roads network to determine the accessibility of each wilderness area and we estimated travel time from any given populated census block to the wilderness area boundary. Most wilderness visits are day-uses (75.9% are visits <6 h; Bowker et al. Reference Bowker, Askew, Landry, Hedges and English2022). Therefore, we focused our analysis on human population changes within 2 h of wilderness areas (i.e., a 4-h round trip), as this reflects a reasonable amount of time for someone to travel to, recreate and return from a wilderness area in a single day.
In developing the DUVI, we evaluated variables likely to increase day-use visits or factors likely to increase the impacts of those visits, including human population in 2020 and population growth rate during 1990–2020, median travel time, wilderness size, wilderness proximity, relative accessibility, trail density and amenities (Table 1). We checked for correlation between our variables, ensuring none had a Pearson’s r greater than 0.5.
Table 1. Description of variables used to calculate the Day-Use Vulnerability Index (DUVI) and hypothesized impact of each variable on wilderness areas.
Data
We gathered spatial data for 752 wilderness areas in the CONUS from Wilderness Connect (www.wilderness.net; accessed August 2022). Thirty wilderness areas were not accessible, being islands or >1 km from the road network, bringing the total number of wilderness areas included in our analysis to 722.
To build our road network buffer around each wilderness area, we downloaded 2020 road data for the CONUS (US Census Bureau 2020). We used the ‘tigris’ R package (https://CRAN.R-project.org/package=tigris) to automate downloads for all 3110 counties in the CONUS.
We obtained population data at the census block level for 1990 and 2020 from the University of Wisconsin–Madison SILVIS lab (Radeloff et al. Reference Radeloff, Helmers, Mockrin, Carlson, Hawbaker and Martinuzzi2022). Because census blocks change over time and due to differential privacy techniques implemented by the US Census Bureau, 1990 and 2020 census blocks cannot be compared directly; however, the error is minimized when census blocks are aggregated (D. Helmers, personal communication 2022).
We acquired line data for recreation trails across the CONUS from the US Geological Survey (USGS) National Geospatial Technical Operations Center (USGS 2023). These data were spatially intersected with the 722 wilderness areas included in our study.
To assess amenities, we used an index developed by McGranahan (Reference McGranahan1999) used in another recent study (Mockrin et al. Reference Mockrin, Helmers, Martinuzzi, Hawbaker and Radeloff2022), which combines favourable summer climate, winter climate, availability of water and topographic variation to produce county-level amenity scores. We intersected wilderness areas in our study with the amenity index using the area-weighted average if a wilderness area fell into more than one county.
To validate our results, National Visitor Use Monitoring (NVUM) wilderness visitation estimates were downloaded for each USDA Forest Service region (https://apps.fs.usda.gov/nvum/results/R01-R02-R03-R04-R05-R06-R08-R09-R10.aspx/FY2019; accessed July 2023). We used estimates collected between 2015 and 2019, which were the most recent surveys with data for all regions (Table S1). Estimates for individual wilderness areas were not available.
Data analysis
First, we estimated travel time along the road network to any given census block from each wilderness area. We converted the road lines data for each county into a 1–km raster with attributes for the major road types (i.e., primary, secondary or local) used to estimate travel resistance along the surface. While speed limits vary by state and region, we calculated an average speed of 112.7 km/h for primary roads (interstate) and 104.6 km/h for secondary roads (highway) based on 2015 state-by-state data from the US Bureau of Transportation Statistics (www.bts.gov; accessed December 2020). We assumed an average speed on local roads of 56.3 km/h, which is half the speed of primary roads. Using these average speed limits, we produced a cost raster with resistance values of 1 for primary roads, 1.08 for secondary roads and 2 for local roads. These values were calculated by scaling the speed on secondary and local roads to the speed on primary roads (i.e., primary road travel is 1.08 times faster than secondary road travel and two times faster than local road travel).
Second, we buffered each wilderness boundary by 1 km to achieve intersections with the road network, with this also representing a reasonably short walking distance for a visitor to travel on foot from the road into wilderness. We used the buffered wilderness boundaries and the cost raster as inputs for the Distance Allocation tool in ArcGIS Pro (version 2.6.0; ESRI, Inc., Redlands, CA, USA). The Distance Allocation tool generated a cost-weighted distance raster (in metres) for each wilderness area, which we converted into hours by multiplying the raster value by the primary road rate of travel in metres per hour (Fig. 1).
Figure 1. Distance (in hours) to any wilderness area within the contiguous US along roadways at a 1–km resolution. Grey areas on the map have no primary, secondary or local roads in the 2020 US Census Bureau data. The maximum distance to any wilderness area is 6.8 h.
Third, to define an area around each individual wilderness area where day-use visitors are most likely to originate, we selected pixels with a value ≤2 (i.e., within 2 h) and transformed those pixels into polygons representing 2-h buffers around each wilderness area. We intersected the population data from census blocks with the 2-h buffers to summarize population characteristics within 2 h of each wilderness area. This area was also used to determine wilderness proximity by intersecting each wilderness area with other wilderness areas within 2 h.
Lastly, for each variable related to day-use impacts (i.e., population in 2020, population growth rate during 1990–2020, median travel time, wilderness size, wilderness proximity, relative accessibility, trail density and amenity index) we binned the raw values into deciles such that wilderness areas were ranked 1–10 for each variable (e.g., a raw value >90th percentile for each variable was given a decile value of 10, <10th percentile was given a decile value of 1). The decile values were summed across the eight variables to determine wilderness areas with the highest overall impacts. To create the DUVI (1–10), we rescaled the summed values for the minimum value to equal 1 and maximum value to equal 10. Wilderness areas with a lower DUVI are less vulnerable to the aggregate of day-use impacts (see Table 1) compared to wilderness areas with a higher DUVI, which are more vulnerable.
Validation
To compare the DUVI with NVUM visitor count estimates, we selected wilderness areas in our study managed by the Forest Service (n = 423) and computed the mean DUVI for each Forest Service region (Table S1). We then compared the values with visitor count estimates using linear regression. There are some notable caveats to this validation, such as uncertainty in the NVUM count estimates, inclusion of all visitors (not just day-use) in the NVUM estimates, the aggregation of DUVI values to regional levels and the addition of variables in the DUVI meant to represent vulnerability to overuse (e.g., wilderness size) and not just an increase in visitor numbers. Despite these concerns, the NVUM represents the best available data for comparison with our results.
Results
Approximately 72% of the land area of the CONUS is within 2 h of any of the 722 wilderness areas we examined (Fig. 1). In 1990, 74.4% of the population (183.8 million people) in the CONUS was living within 2 h of any of the 722 wilderness areas, increasing to 75.6% (246.5 million people) in 2020. The population growth rate between 1990 and 2020 (i.e., relative to 1990) was higher within 2 h of any of the 722 wilderness areas (34.1%) than outside 2 h (26.0%).
Wilderness areas with the highest vulnerability (i.e., decile value of 10) differed by variable (Fig. 2). For instance, the amenity index was higher on average (5.5 vs 3.7) for wilderness areas in the western CONUS (i.e., west of 100° longitude; Fig. 2), whereas wilderness areas in the east were smaller on average (111 vs 362 km2).
Figure 2. Relative impacts at 722 wilderness areas within the contiguous US for each of the eight variables related to day-use visitation (see Table 1). Raw values were binned into deciles (1–10), where decile values of 10 are wilderness areas in the top 10% for vulnerability for that variable.
The combination and scaling of decile values across variables yielded the DUVI (Fig. 3). There were 30 wilderness areas with a DUVI of ≥8, representing the top 20% of the index (Table 2). The most vulnerable was the Mount Timpanogos wilderness in Utah, which had high decile values (≥7) across all eight variables. The Glacier View wilderness in Washington, the J.N. Ding Darling wilderness in Florida and the Philip Burton wilderness area in California were tied as the next most vulnerable, with a DUVI of 9.8 (Table 2). The Birkhead Mountains wilderness area in North Carolina was also in the top five (9.6). The full list of wilderness areas, raw values for each variable and DUVI values can be found in Table S2.
Figure 3. The Day-Use Vulnerability Index (DUVI) values showing potential vulnerability to impacts of day-use visitation for 722 wilderness areas in the contiguous US. Wilderness areas are represented with colour (blue to red) and circles in sizes relative to their index value.
Table 2. Wilderness areas and locations within the contiguous US with the highest Day-Use Vulnerability Index (DUVI; ≥8) and the associated values for each of the eight variables used to calculate the index. The full dataset can be found in Table S2.
a Measured within 2 h.
b Change in 1990–2020 population divided by 1990 population.
c Measured within the wilderness area.
d Roads divided by wilderness size.
e Measured within 1 km.
f Trails divided by wilderness size.
g See McGranahan (Reference McGranahan1999).
h Measured at the county level.
NVUM data indicated that the Rocky Mountain (R2), Pacific Northwest (R6) and Pacific Southwest (R5) regions had the highest numbers of visits to wilderness areas during 2015–2019, whereas the Eastern (R9) and Northern (R1) regions had the lowest (Table S1). We found a positive relationship (R2 = 0.19) between the DUVI values and NVUM visitor count estimates. R2, R6 and R5 had lower average DUVI values than expected given the high visitor estimates, whereas R1, R9 and R8 (Southern region) had higher average DUVI values than expected (Fig. 4 & Table S1).
Figure 4. Comparison of National Visitor Use Monitoring (NVUM) estimated visitor counts and the Day-Use Vulnerability Index (DUVI) averaged for wilderness areas managed by the US Department of Agriculture Forest Service in each region. Forest Service regions in the contiguous US are: Northern (R1), Rocky Mountain (R2), Southwestern (R3), Intermountain (R4), Pacific Southwest (R5), Pacific Northwest (R6), Southern (R8) and Eastern (R9). The blue line indicates the line of regression, with a 95% confidence interval given in grey.
Discussion
We developed the DUVI to identify wilderness areas that might be threatened by human impacts. Validation with NVUM visitor count estimates indicated a positive relationship with the DUVI, suggesting subsequent modelling of visitor estimates is feasible with additional parameter tuning. We conducted this analysis at considerable scale and recognize that the DUVI rankings do not capture all of the possible conditions and scenarios present in individual wilderness areas. However, the index provides a novel view of day-use vulnerability across the CONUS, which could lead to the identification of threatened wilderness areas followed by more in-depth assessments based on local knowledge.
Vulnerability
Population and visitation
The primary impact of more people being within 2 h of a wilderness area is the potential for higher visitation (Lindley et al. Reference Lindley, Blevins and Williams2018, Rasch & Hahn Reference Rasch and Hahn2018, Gosal et al. Reference Gosal, Giannichi, Beckmann, Comber, Massenberg and Palliwoda2021), which in turn could result in environmental degradation, including reduction of vegetative cover, erosion of soils, disturbance to wildlife, degradation of water quality and buildup of trash and human waste (Monz et al. Reference Monz, Pickering and Hadwen2013, Marion Reference Marion2016, Marion et al. Reference Marion, Leung, Eagleston and Burroughs2016, Eagleston & Marion Reference Eagleston and Marion2017). For the Mount Timpanogos wilderness area in Utah, which had the highest DUVI, Lindley et al. (Reference Lindley, Blevins and Williams2018) reported visitation had increased by 18% between 2013 and 2017, with 82% of visitors residing in adjacent counties. More broadly, Bowker et al. (Reference Bowker, Askew, Landry, Hedges and English2022) showed that wilderness visitation to national forests had increased by >27%, which is considerably faster than US population growth (c. 8%). They also found wilderness visitation across the western US had increased, whereas visitation had decreased in the eastern US. Using population projections and historical visitor use, Rasch and Hahn (Reference Rasch and Hahn2018) estimated that the greatest increase in day-use visits to wilderness areas in national forests through 2060 would be in the south-western US.
In addition to environmental degradation, increasing visitor numbers can affect other wilderness characteristics, such as primitive and unconfined recreation or solitude. These characteristics are mandated in The Wilderness Act (US Public Law 88-577) but can be challenging to achieve with the increase in numbers of visitors (Landres et al. Reference Landres, Barns, Dennis, Devine, Geissler and McCasland2008, Cole Reference Cole2011). At wilderness areas near large urban areas there are more ‘accidental’ visitors (i.e., people who are not necessarily aware that they are entering a wilderness area). These types of visitors may or may not be interested in solitude, but they also tend to be less bothered by or cognizant of crowding (D’Antonio & Monz Reference D’Antonio and Monz2016, Lindley et al. Reference Lindley, Blevins and Williams2018).
We focused on day-use visits and potential visitors living within 2 h. However, there are other visitor types, including those recreating within wilderness areas for multiple days (i.e., multi-day visits) and day-use visitors living farther than 2 h away but staying overnight in towns or cities near to a wilderness area. While comprising a smaller proportion of visitors (c. 14.4%; Bowker et al. Reference Bowker, Askew, Landry, Hedges and English2022), multi-day recreation is more likely to impact lesser-travelled areas, causing greater ecological impacts on vegetation and soils (Monz et al. Reference Monz, Pickering and Hadwen2013, Salesa & Cerdà Reference Salesa and Cerdà2020). In this respect, we caution that some wilderness areas with a low DUVI still face threats from visitors. For instance, the Boundary Waters Canoe Area wilderness in Minnesota has a low DUVI (1.8), partly due to the small population within 2 h (306 965 people; see Table 2 for comparison). However, environmental degradation and high visitation within this wilderness area are evident (Eagleston & Marion Reference Eagleston and Marion2017, Hjepre Reference Hjerpe2018) and have resulted in this wilderness area being the first to institute a permit system (Holmes et al. Reference Holmes, Englin and Valdez-Lafarga2022). According to a survey by Hjerpe (Reference Hjerpe2018), >92% of visits to the Boundary Waters Canoe Area wilderness were multi-day trips. Additionally, >97% of visitors were self-described as not residing in the region (Hjerpe Reference Hjerpe2018). This highlights a limitation of our study for capturing visitors travelling from beyond 2 h.
Amenity migration
Migration towards wilderness areas and other public lands is expected to increase day-use visitation into the future. Radeloff et al. (Reference Radeloff, Stewart, Hawbaker, Gimmi, Pidgeon and Flather2010) demonstrated higher rates of housing growth between 1940 and 2000 within 50 km of wilderness areas than within 50 km of national parks, national forests or the CONUS average. Mockrin et al. (Reference Mockrin, Stewart, Matonis, Johnson, Hammer and Radeloff2018) showed that cities with nearby public land in the western, southern and midwestern US were more likely to see high housing growth rates between 2000 and 2010, while Hjerpe et al. (Reference Hjerpe, Hussain and Holmes2020) found that between 1980 and 2010 counties with wilderness areas and other protected lands had higher population growth rates. Similarly, we observed that population growth rates (1990–2020) were 8.1% higher within 2 h of any of the 722 wilderness areas in our study than the rest of the CONUS. The amenity index (McGranahan Reference McGranahan1999) tended to be higher on average (5.5 vs 3.7) for wilderness areas in the western US (i.e., west of 100° longitude; Fig. 2). In comparing the DUVI to NVUM visitor count estimates, the Rocky Mountain (R2), Pacific Northwest (R6) and Pacific Southwest (R5) regions each had lower DUVI values than expected based on visitor counts (Fig. 4 & Table S1). These three regions, along with the Southwestern (R3) region, also had the highest average amenity index values (5.2–6.2). This suggests that the relationship between amenities and visitation counts, such as are available through NVUM, should be explored further.
There is a growing body of literature devoted to the socioeconomic, cultural and environmental impacts of amenity migration (e.g., Gosnell & Abrams Reference Gosnell and Abrams2011, Abrams et al. Reference Abrams, Gill, Gosnell and Klepeis2012, Hjerpe et al. Reference Hjerpe, Armatas and Haefele2022). Socially constructed values regarding proximity to nature, recreational opportunities and escaping crowded urban areas have encouraged people to move closer to wilderness areas (Gosnell & Abrams Reference Gosnell and Abrams2011, Mockrin et al. Reference Mockrin, Stewart, Matonis, Johnson, Hammer and Radeloff2018). Within these wilderness-adjacent communities there can be contradicting management goals. For instance, the desirability of natural amenities encourages local policies that protect qualities such as forest cover, clean water and unobstructed views, while at the same time disrupting natural processes such as wildfires through fire suppression and aversion to prescribed fire as a management tool (Radeloff et al. Reference Radeloff, Stewart, Hawbaker, Gimmi, Pidgeon and Flather2010, Abrams et al. Reference Abrams, Gill, Gosnell and Klepeis2012, Radeloff et al. Reference Radeloff, Helmers, Kramer, Mockrin, Alexandre and Bar-Massada2018).
Wilderness size
There is considerable difference in the size of wilderness areas between the eastern and western US (Aycrigg et al. Reference Aycrigg, Tricker and McCarley2022b). This can have the effect of concentrating visitor use, resulting in degradation of aquatic habitats and impacts on wildlife such as avoidance and food dependence (Monz et al. Reference Monz, Pickering and Hadwen2013, Bleich Reference Bleich2016). In validating the DUVI against visitor count estimates, the Eastern (R9) and Southern (R8) regions each had higher DUVI values than expected based on visitor counts (Fig. 4 & Table S1). This suggests that many of the highest-ranked wilderness areas in these Forest Service regions are not as likely to experience impacts from high numbers of visitors. However, wilderness areas in these regions also had the smallest average size (129 km2 for R9 and 37 km2 for R8; Forest Service-managed wilderness areas only; Table S1). Although wilderness area visitation may be decreasing in the eastern US (Bowker et al. Reference Bowker, Askew, Landry, Hedges and English2022), there could still be pronounced impacts from higher concentrations of visitors in smaller areas.
Fourteen of the wilderness areas that we found to be most vulnerable were in the eastern US (i.e., east of 100° longitude; Fig. 3 & Table 2). Among the variables we compiled, wilderness size, wilderness proximity, relative accessibility and trail density each had many wilderness areas in the eastern US in the highest decile (Fig. 2). These wilderness areas are vulnerable to environmental degradation because visitors are concentrated in smaller wilderness areas and have fewer wilderness options to visit while still having relatively high accessibility and recreational trail options. Our study did not evaluate the availability of state, county, local and other non-wilderness public land, which would affect the spatial distribution of people seeking outdoor recreation and could be an important consideration in future research.
Management implications
Our results can be a starting point to evaluate wilderness area vulnerability from day-use visits, but wilderness managers will need to consider many options to limit the impacts of increasing visitation. For instance, trailhead quota systems, which limit the number of visitors who can enter a wilderness area in a day, are a tool that can be used to limit environmental degradation (van Wagtendonk & Coho Reference Van Wagtendonk and Coho1986). However, trailhead quota systems are less popular than other options such as education and are not always fully effective due to loopholes and the unequal spatial distribution of use within a permitted area (Lindley et al. Reference Lindley, Blevins and Williams2018, Jenkins et al. Reference Jenkins, van Wagtendonk and Fincher2021, Schneller et al. Reference Schneller, Binzen, Cameron, Vogel and Bardin2021). Wilderness permits also provide a useful tool for monitoring and regulating wilderness use in areas at risk of ecological degradation, and these data can be further used to understand trends in how many people take wilderness trips, how the demographic characteristics of visitors are evolving over time and how the demand for wilderness attributes is changing (Holmes et al. Reference Holmes, Englin and Valdez-Lafarga2022). Restoration, resource modification and regulation are all available options, but these must be weighed against maintaining wilderness character (Stankey et al. Reference Stankey, Cole, Lucas, Petersen and Frissell1985, Marion Reference Marion2016, Lieberman et al. Reference Lieberman, Hahn and Landres2018). In addition, monitoring the effect of any management action is a key step in determining whether the action reduced impacts to an acceptable level, whether additional action is needed or whether a different approach is warranted (Stankey et al. Reference Stankey, Cole, Lucas, Petersen and Frissell1985).
In essence, wilderness areas are composites of ecological, societal, cultural and political goals – and not always in that order (Bleich Reference Bleich2016). In studying trends in perceptions of wilderness, Rasch (Reference Rasch2022) found that younger generations were less likely to view wilderness as untrammelled and more likely to support restoration and intervention. Other surveys have found that visitors expect wilderness areas to appear clean and natural, which may also lend support to management actions (Watson et al. Reference Watson, Martin, Christensen, Fauth and Williams2015, Lindley et al. Reference Lindley, Blevins and Williams2018). Just as wilderness-adjacent community policies can sometimes favour environmental action to preserve natural amenities (Abrams et al. Reference Abrams, Gill, Gosnell and Klepeis2012), increased recreation within wilderness areas can have positive impacts on perceptions of wilderness and support for conservation and management (Miller et al. Reference Miller, Blahna, Morse, Leung and Rowland2022). The assessment and management of individual wilderness areas should employ a holistic approach that considers both local support for conservation and ecological threats from surrounding communities.
In the case of the Mount Timpanogos wilderness area, which ranked highest on our DUVI, surveys in 2013 indicated that people valued the natural characteristics of the area and recognize the impacts of overuse, even though the idea of a quota system was unpopular (Lindley et al. Reference Lindley, Blevins and Williams2018). In 2023, the national forest instituted a quota parking permit system at popular trails for the busiest periods (https://www.fs.usda.gov/recarea/uwcnf/recarea/?recid=15110; accessed July 2023). While some wilderness visitors will likely not approve of any restrictions (Lindley et al. Reference Lindley, Blevins and Williams2018), this type of system balances the need for limitation while still allowing visitors flexibility to choose less busy trails and to recreate during off-peak periods. For the wilderness areas we identified as most vulnerable to day-use impacts, there are probably additional opportunities to engage with visitors and nearby communities to develop mutually beneficial strategies allowing for recreation and limiting impacts.
Conclusion
We developed the DUVI for 722 wilderness areas in the CONUS using eight variables influencing day-use visitation and the potential for visitation to result in overuse, highlighting individual wilderness areas that are most vulnerable (Table 2). A positive correlation between the DUVI and NVUM visitor counts demonstrated the potential for future development of models for estimating visitation. We focused on vulnerability to impacts from day-use visits in wilderness areas of the US, but protected places worldwide face many of the same impacts to wildlife and the environment from recreation (Salesa & Cerdà Reference Salesa and Cerdà2020, Salvatori et al. Reference Salvatori, Oberosler, Rinaldi, Franceschini, Truschi, Pedrini and Rovero2023). Population proximity has been shown to be an indicator of visitation in Brazil, Canada, Europe and Sub-Saharan Africa (Gosal et al. Reference Gosal, Giannichi, Beckmann, Comber, Massenberg and Palliwoda2021, Hausmann et al. Reference Hausmann, Toivonen, Heikinheimo, Tenkanen, Slotow and di Minin2017, Nabout et al. Reference Nabout, Tessarolo, Pinheiro, Marquez and de Carvalho2022). Furthermore, populations have been growing faster near the edges of protected areas than other rural areas (Wittemyer et al. Reference Wittemyer, Elsen, Bean, Burton and Brashares2008). Visitor education, habitat restoration, resource modification, visitor quotas and wilderness permits are useful tools for managing the impacts of increasing populations near protected areas (Watson et al. Reference Watson, Martin, Christensen, Fauth and Williams2015, Marion Reference Marion2016, Lindley et al. Reference Lindley, Blevins and Williams2018). Finally, it is worth noting that there are stark inequities between protected areas globally; strategies used to preserve wilderness character in the US might not apply in places facing different financial and cultural challenges (Jones et al. Reference Jones, Venter, Fuller, Allan, Maxwell, Negret and Watson2018). Engaged local communities are the best resource for incorporating actions most likely to succeed in a particular location with unique risks.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S0376892923000279.
Author contributions
TRM, JLA, SM, RTB and TPH conceived the research idea and design. TRM and JLA carried out the research. TRM conducted the analyses. TRM, JLA, SM, RTB and TPH wrote, revised and reviewed the manuscript.
Acknowledgements
We thank Susan Fox, Beth Hahn, Sean Parks, Jason Taylor and Dave Helmers for their valuable input throughout this project.
Financial support
This work was supported by the Aldo Leopold Wilderness Research Institute through US Department of Agriculture Forest Service Agreement No. 17JV11221639050.
Competing interests
The authors declare none.
Ethical standard
None.
Open access
