Introduction

Many urban, mountainous municipalities have abundant natural setting for “green” recreational exercise. Exercise facilities are primarily manifested in trails, which can be used for mountain biking, trail running, hiking, equestrianism, and other outdoor recreation activities. Green exercise is a physical activity done outside while in nature (eg mountainous trails, near waterways, wilderness areas, countryside, and urban green areas), and it is beneficial to physical and mental health.^1,2 Access to green exercise has most often been examined in large municipalities,^3,4 but there is little research that explores green space access in the American Intermountain West. This study sought to examine access to green exercise space (ie off-road and soft-surface trails) among urban Utah municipalities. The urban area of the Utah Wasatch Mountain Range spans roughly 65 km north and 65 km south of Salt Lake City. This setting is similar to those found in other urban, mountainous areas such as Denver and Boulder, Colorado; Santa Fe and Albuquerque, New Mexico; or Geneva, Switzerland.

Background

Any exercise, “green” or otherwise, is beneficial for reducing a host of morbidities and mortalities such as diabetes and obesity.⁵ Individuals who use their built environment (eg for walking or biking) to facilitate this activity decreased rates of obesity, diabetes, and increased levels of overall physical activity.^6,7 There appears to be an inverse relationship between distance to built-environment resources and using green spaces (ie those who live closer are more likely to use green space).⁸ Thus, our study seeks to explore the green exercise (ie off-road and soft-surface trails) access in an urban, mountainous region.

Prior research shows that access to outdoor physical activity resources, including trails, is an important predictor for participation. For example, Huston et al⁹ reported that proximal access to recreational facilities was positively associated with leisure activity engagement (odds ratio [OR] = 2.28, [1.30-4.00]). Roemmich et al¹⁰ reported that children are significantly more likely to utilize parks and be physically active with increase park density in their neighborhood. However, Franzini et al¹¹ point out that access to outdoor physical activity resources is unequal across neighborhoods and that the starkest differences are between socioeconomic and racial groups. Gladwell et al describe the positive benefits to outdoor exercise, green exercise, being increased physical activity and improved well-being. They also describe access to outdoor recreation as being an important barrier for participation.¹² This is important in the context of findings by Librett et al¹³ who reported that trail users were much more physically active than non-trail users (OR = 2.3, [1.9-2.8]) and that trail users largely rely on infrastructure within their own community to their trail activities. Thus, built environment is shown to, in part, be a determinant of green exercise.

Green exercise appears to be particularly beneficial to mental well-being. On comparing individuals who completed the same exercise routine indoors and outdoors, general results showed that outdoor exercise was more beneficial to mood, energy, and well-being.¹⁴ Similarly, Pretty et al² reported on the benefit of green exercise for improving mood and self-esteem. The effect of exercising outside can be seen across rural and urban settings. This effect is moderated by the pleasantness of the scene (eg landscapes or vistas compared with dilapidated buildings or junk yards).¹⁵ Both men and women experience benefits of using green exercise, but men seem to experience greater mood and self-esteem boosts. Young people also tend to benefit the most from green exercise, but all age groups receive positive benefits.¹

These findings can, in part, be described using health behavior theory. A foundational concept of the Social Cognitive Theory is that of reciprocal determinism; that individuals, their environment, and their behavior form an important relationship where changes to one (eg environment) can lead to changes in the other (eg behavior).^16,17 In other words, individuals are producers and products of their environment. Another important element of Social Cognitive Theory is self efficacy, and individual’s belief or confidence they are able to accomplish a task. Lack of access to outdoor physical activity resources has been shown as an inhibitor to self efficacy in context of the Social Cognitive Theory, and a reducer of individual physical activity.^18,19 Thus, examining trail access serves to explore the health behavior characteristics of environmental determinism.

Specifically, this environmental justice literature has mostly focused on urban areas, and there’s relatively little work on access to outdoor recreation in mountainous areas from this perspective (however, Floyd & Johnson outline that part of the problem is how we conceptualize and operationalize environmental justice in outdoor recreation²⁰). Studies examining environmental justice in urban areas have found differing access to parks by race and ethnicity, including park size where whites were more likely to be nearer larger parks³ and ethnic minorities being nearer more congested parks.^21,22 Wolch et al²³ argue that efforts to increase green space in urban settings can lead to gentrification and displacement of the very residents it was aimed to help. Though this urban-environment relationship has been well documented, the urban-wilderness interface has not. Therefore, our study seeks to examine soft-surface trails within a mountainous, semi-urban region.

Most prior efforts to address access and environmental justice have examined the issue from a neighborhood, tract, or block group perspective. Prior literature has shown that residents are more likely to recreate using facilities in their own neighborhoods.^3,4,22 These studies make observations in large, urban settings (eg Baltimore or Los Angeles), where each neighborhood or zip code would have greater residential density and population density than many municipalities in the Intermountain West. With this fact, it is worth considering the rural-urban municipality as comparable to the urban neighborhoods of these other cities. At least geospatially and culturally, there are similarities. Similarly, some studies have used zip code as its geographic unit,^24,25 but in the Intermountain West, many municipalities, even those classified as urban, only have a single zip code. Rigolon et al²⁶ note the micro-level analysis as a limitation of park access in the United States because it limits the broad context view municipal-wide analysis allows. Others noted the value of city-level is also where many policy, funding, and facilities management decisions are made.^26,27

We specifically sought to address the following research questions: (1) What is the prevalence of municipal trail head access?; (2) Is there significant spatial clustering of trailhead access?; and (3) What are municipal-level predictors of trail density and trailhead access?

Methods

Setting

The Western face of the Wasatch Front mountain range is the most populated region in Utah, where about 80% of the state population resides (Figure 1).²⁸ This narrow strip of land is roughly 140-km long and 5- to 30-km wide. The Salt Lake metropolitan area and Provo-Orem have population-weighted densities exceeding 4250 people per square mile (1642 per square km) which places them among the top 50 most densely populated cities in the United States.²⁹ The natural environment in this area affords residents the ability to find green exercise spaces (eg hiking, skiing, and mountain biking). If predictions are correct, Utah’s population is expected to double by the year 2050.³⁰ Many other mountainous states, such as Colorado and Arizona, have experienced greater than 6% population growth in the past 15 years.³¹ This region was selected because it has a long (140 km) region of urban municipalities that directly borders mountain land; it can provide insights into other similar regions; and, data were widely available to answer research questions posed in this article. The purpose of this article is to characterize trail access in an urban, mountainous area.

Figure 1.

Population by municipality in Utah.

The map shows the highly populated area between Utah and Weber Counties. Counties with municipalities selected for this study are indicated in gray.

Data and analysis

Socioeconomic and commuting data for each municipality were obtained from the US Census.³³ The state-wide trail system geographic information systems (GISs) shapefile and municipal boundaries were obtained from the Utah State mapping portal website.³⁴ A 1-km buffer was added to the mountain-facing side of each municipality (East) to capture trailheads that may not have been within city limits (roughly 90% of land within this buffer is private, city-owned land³⁵). The trails shape file was clipped to the buffered municipalities to measure trailhead access. Similarly, household point data were obtained from the same portal website for activity space analysis.³⁶ R statistical software was used for data management analysis³⁷ and QGIS was used for geospatial visualization.³⁸

Trailhead access

A trailhead is the access points to a trail or network of trails. Access to trailheads within municipalities was determined, first by computing each municipality’s activity space, and second, by determining if the activity space overlapped with trailheads. Activity space can be thought of as a way to explain individual’s movements and how they are likely to interact with their environment, including how accessible community resources are.³⁹ In our case, we used a two-standard deviation ellipse (2SDE) around the municipality residential points. This provides a two-dimensional ellipse or activity space. If a point of interest were to appear within that ellipse it is very likely the resident population has access to it. In our case, we observed if trailheads were within each municipality’s activity space. For example, Figure 2 shows the city of Holladay (population 26 400), the corresponding activity space, and one trailhead present within that space. Though there are other methods to measure access, our focus on municipal-wide analysis leads us to this larger scoped analysis. The activity space provides insight into “coverage” or “cumulative opportunities” within the core area of the municipality.⁴

Figure 2.

Activity space for Holladay, UT.

The map shows residents of Holladay, Utah, the corresponding activity space, and one trailhead appearing within the activity space.

Deciphering spatial patterns of municipal trail densities

Determining if municipalities exhibit spatial patterns with respect to trail density was done so using two-step process. First, the global pattern across the entire space was determined using Moran’s I for spatial autocorrelation (ie I < 0 is a uniform pattern, I = 0 is a random pattern, and I > 0 represents a clustered pattern).⁴⁰ Although these results determine what type of pattern exists, the nature of the pattern remains unknown. A second step is needed to visualize local clusters. Second, determining the nature of the pattern was determined using Local Indicators of Spatial Autocorrelation (LISA). This procedure works by comparing individual municipality variables with neighboring municipality variables.⁴¹ In essence, this is a comparison of the standardized municipal observation, the standardized spatial lag (or average of the municipality neighbors), and the local Moran’s I statistic. Results from this analysis typically produce a combination of “high” or “low” in a two-word sequence for each municipality (eg high-high or high-low). The first word indicates the individual municipality’s variable, and the second word indicates what the individual municipality’s neighbors (ie spatial lag) variable is in comparison to the individual. For example, for trail density, a high-low result demonstrates an individual municipality that has high level of trail density and its neighbors are generally low.

Identifying predictors of tail density and trailhead access

Municipal-level predictors were identified using multiple linear regression with trail density as the dependent variable and municipal-level predictors as the independent variables. Assumptions for linear regression, namely normal distribution and equal variance, were determined using Normal QQ and equal variance plots. Municipal-level variables used were population density;^24,27 median household income, housing value, poverty rate, educational attainment, and property tax;^26,424344-45 businesses, land area, elevation;⁴⁶ percent women;⁴⁷ and bicycle and walking commuters.⁴⁸ Backward elimination was used to find a best-fit model.⁴⁹

Results

A total of 33 urban municipalities were surveyed along the western face of the Wasatch Mountains. On average, municipalities had 19.64 km (standard deviation [SD] = 30.63, range = [0, 153.93]) of off-road, soft-surface trails. On average, municipalities had 2.10 (SD =2.45) trailheads and an average trail density of 0.54 km/km² (SD = 0.60). Excluding municipalities that had no trailhead, municipalities with a trailhead had an average trail density of 0.78 km/km² (SD =0.64) and an average of 3.63 trailheads (SD = 3.92). Roughly, two in five municipalities had no trailhead access within its activity space (42.4%), and 57% of these had no trailheads within 2 km of municipal boundaries.

Trail density was observed in a significant clustering pattern (I = 1.54, P = .06). Since this initial step is used to justify proceeding with a more spatially local analysis, the slightly elevated P value is noted but considered low enough to proceed with LISA analysis. Number of trail heads was not significantly clustered (I = 0.06, P = .48) and appeared to be a random pattern. Trail density and number of trailheads were significantly correlated (r = 0.49, P = .004), indicating that increased number of trailheads is related to greater trail density. Figure 3 shows a large, positive cluster of trail density around the Alpine/Pleasant Grove area, indicating that these areas have significantly high trail density, as do their neighbors.

Figure 3.

Localized clusters of trail density urban Utah municipalities.

Analysis shows the results of LISA analysis, where the first word indicates the individual municipality and the second word represents the aggregated measure, or lag, of its neighbors (eg Highland [center, blue] is low for trail density and its neighbors are high for trail density).

Trail density and trailheads were regressed on the set of municipal-level predictors using multiple linear regression. Backward elimination was used to identify variables from among this set that were significant predictors of trail length. Table 1 presents the full and reduced models for both model sets. Both trail density and trailheads were significant predictors for the opposite models, due to their correlated nature. In the reduced model, median household income was a significant predictor for trail density; and elevation was a significant predictor for trailheads.

Table 1.

Predictors of trail density and trailheads among Utah urban, mountainous municipalities.

Discussion

We sought to characterize urban, mountainous green exercise trail access among municipalities bordering the Wasatch Mountains in Utah. In general, trail density correlated with trailhead access points. In addition to this, elevation was a significant predictor for trailheads and home value was a significant predictor of trail density. There was some clustering effect for trail density. Roughly, two in five municipalities had no trailhead in their activity space.

The correlated nature of trail density and trailheads may be explained in two ways. In reciprocal determinism, more trails contribute to more trailheads, vice versa and so on.¹⁶ Another way would be the “herd effect” described among bicyclists. Jacobsen and colleague^48,50 described the additive effect having bicyclists in a community. The more bicyclists there are in an area, the safer they tend to be on the road. It is thought that more bicyclists on the road create a protective, “herd effect” on the collective bicycling population. It may be a similar mechanism with trail density, trailheads, and participant usage, in that more participants lead to more trails, and so on. It is possible that municipalities with more trails tend to have a more developed trail use society and culture, thus driving the capital infrastructure investment.

We expected some level of geographic variation for trail density among municipalities, but the location of those “highs” and “lows” were previously unknown. Geographic variation has been noted in a variety of other settings. For example, bicycle crashes²⁴ and access to physical activity facilities and exercise performance.⁵¹ Because of this, trail density and access was not expected to be uniform across the space studied. There may be topological differences among these locations that make trail building more conducive. More likely, these patterns correspond to municipal-level differences in socioeconomic levels, and/or trail use. Some of these same socioeconomic factors were significant predictors of municipal trail length, namely higher home income (higher home income correlated with more trails, or increases in US$10 000 income results in roughly 0.12 more km/km²). Over the space of a municipal boundary interface with the mountain area, this can result in many more kilometers of trails available. One study notes that bicycling infrastructure often attracts and caters to higher socioeconomic groups.⁵² This higher home value could be related to greater municipal-driven trail development. For example, Draper City (population = 45 285, trail density = 1.98 km/km² [highest among study area], and home value = US$371 000 [third highest in study area]) have significant municipal trail development investment.⁵³

Elevation being a significant predictor within the trailhead model may be a function of convenience. As noted by Jones et al,⁸ proximity and green exercise are inversely related. Thus, higher elevation is more distant from residential areas. This may have developed naturally as a result of the reciprocal determinism: participant is a producer and product of their environment.¹⁶ That is, users do not want to travel far to participate in green exercise.

This study has inherent limitations. As this was an exploratory research, we only examined one metropolitan, mountainous region at a single point in time. Although there are other ways to classify a municipality as urban or not, we chose to measure urban using the USDA Business & Industry classification. In the Intermountain American West, there are much fewer large metropolitan areas than the Eastern United States or European countries. This classification allowed us to include municipalities along the Wasatch Front. We measured municipalities as being mountainous as those with boundaries adjacent to the mountains. We chose not to include municipalities not directly bordering the mountains because the study purpose was to examine trailhead access. In addition, we did not have access to bicycling data from these areas. For most of them, these data does not exist. For the few who do collect these data, it is infrequently collected and the measures are generally not comparable. This study only analyzed data for 33 municipalities across one metropolitan region in the United States. Despite this, there are important lessons worth considering in other comparable areas, such as New Mexico, California, Colorado, the Swiss Alps (eg Geneva), or Southern British Columbia, Canada.

Our research demonstrates patterns of access to green exercise trails that differ by location and with respect to home value and elevation. The discovery that 42% of municipalities do not have trailhead access may be a restriction on actual green exercise. These unpaved, mountainous trails may be a way to provide infrastructure for individuals to get adequate exercise.¹ In municipal investment, Dill and Carr point out that built infrastructure gets used.⁵⁴ These mountainous trails have also been shown to improve mental health and act as a way to improve mood and self-esteem.^2,15 It is important for parks and land managers, as well as public health practitioners interested in promoting exercise to consider ways to promote “green” exercise via off-road, soft-surface trail development. Future research needs to examine the relationship between trail infrastructure and actual participant behavior. It would be beneficial to examine other urban, mountainous regions as well (eg municipalities in Colorado, California, British Columbia, and other European regions) Urban, mountainous municipalities have ample opportunity to engage participants in green exercise. Some appear to be making this a reality for residents and others are not. It is recommended for city planners, recreation managers, and public health officials to consider a variety of ways to promote green exercise. This may include trail building programs^5556-57 or transit to trail programs; in either case, residences have easier access to green exercise space.