Mapping and Exploring the Conditions and Purpose of Using Trail Paths in a Medium-Sized Suburban Environment

Abstract

The current paper presents the results of a questionnaire survey which examines the trail path usage conditions on Peranthi Hill in Arta, a medium-sized city in northwestern Greece. The key element of the survey is the detailed mapping of the suburban trail path network of the city, and it focuses on questions about the types of activities of people who walk or bike on the trail paths and the day, time, duration and distance of their trips, as well as their equipment (bicycles, clothes, shoes, etc.). Further questions focus on maintenance and safety issues (e.g., security, guarding, cleanliness) and infrastructures which are considered necessary in order to improve the trail paths’ environment and their use. Taking into consideration that several routes are being developed on Peranthi Hill, with distinct differences in their layouts, the preferred layout is explored in relation to each activity selected by the users. The results of the questionnaire survey, and the cluster analysis specifically, aim to address these issues and assist stakeholders and policymakers in improving the trail paths in terms of their utility and usability level.

1. Introduction and Background

Citizens in urban areas look for public places for leisure and recreational activities. In cities, these areas are mainly located in urban forests, urban parks, sports fields and squares. It is widely thought that a public open space (POS), which is appropriately designed and well-structured, promotes the physical activity of residents and represents a valuable resource with the potential to enhance people’s well-being [1]. There is also a common belief that good and easy access to large attractive public open spaces is connected with higher levels of activity and walking [1]. Urban forests or municipality lands are a group of areas in which one may find recreational opportunities [2]. Similarly, urban parks are highly beneficial to communities, as they are able to provide a convenient environment for a wide range of fun, leisure and recreational activities and can also elevate a community’s image and perceived value [3]. A study by Hull and Harvey [4] investigated and assessed how individuals were affected by various suburban parks by considering the parks’ tree densities and understory vegetation densities, as well as the potential presence or absence of trail pathways [4].

Urban and suburban parks and green areas can improve citizens’ quality of life, and visitors expect satisfaction from their presence in these areas. For instance, López-Mosquera and Sánchez [5] considered two suburban natural locations in Spain and investigated the impact of various parameters related to satisfaction on the behavioral intentions of visitors. More specifically, they aimed to analyze how various psychological variables influence willingness to pay in order to improve area conservation, as well as their impacts on visitor loyalty [5]. Similarly, a study by Korpela et al. [6], in Finland, examined the main factors which are associated with restorative experiences in cherished locations. These determinants included aspects such as the ‘‘immediate’’ use of the most cherished place (including also for how long and how frequently they visit it), individual background in nature experiences (e.g., nature-oriented interests, hobbies) and also a number of personal aspects of life which are linked to stress (work-related stress, well-being, etc.) and other social aspects [6]. Another study by Tyrväinen et al. [7] demonstrates a methodological framework to define and evaluate perceived qualities of green areas within the context of strategic green area planning and suggested that the most influential predictors correlated with favorite places were found to be tranquility and the sense of naturalness.

Land-use planning and green environment services are thought to be inseparable components of trail path design in order to promote sustainable development [8]. The presence of trail paths in an area is an important asset to be used by visitors for outdoor activities. Keith et al. [9] revealed in their study that the primary motivations for visiting the examined trail paths were found to be exercising and seeking respite from the stress of urban life. Furthermore, it was observed that safety and security were found to be the most important concerns among females and among individuals from various racial and other ethnic minorities. Another study in the field by Akpinar [10] proposed that urban greenways are crucial determinants for quality of life within cities, and that it is important to realize greenways cannot be considered as mere “luxuries”. Instead, the author suggests that urban greenways play a crucial role, as they facilitate essential health, recreational and leisure activities for Turkish residents.

Other evidence shows that trail paths in urban forests are frequently chosen by many individuals who engage in a wide range of outdoor activities, such as walking/running and cycling [11]. Moreover, Wang et al. [11] found that each type of trail path in an urban forest attracts user groups with different characteristics. Meyer et al. [12] examined forest visitors focused on pre-existing trail paths in Germany and showed that the factors affecting pathway choice were found to be habit, spontaneity and proximity to visitors’ homes, while there was no effect of specific forest characteristics. Verlič et al. [13] examined how and to what degree various objectively measured recreational impacts on urban forest trail paths are perceived by visitors. Furthermore, demographic variables such as age and education were taken into account in this evaluation. The findings of the study indicated that certain impacts, such as presence of heavy mud at the trail path sections and informal trail paths, were perceived to a significantly greater degree than what assessors had objectively measured [13].

In general, evidence shows that investment in trail paths creates many benefits for an area. For example, Lukoseviciute et al. [14] assessed how trail-related tourism in a coastal area of Portugal adds to income and overall wealth and disclosed numerous benefits to local economy and communities [14]. In a similar study by Lukoseviciute et al. [15], the most famous hiking trail path in Portugal was studied in order to define and evaluate recreation opportunities of the trail’s management, development, preferred trail path attributes and factors affecting visitor loyalty.

Without doubt, trail paths play a vital role in the creation of tourist space [16]. Trail paths for tourism constitute a vital component of infrastructure in the natural landscape [17] and are considered to be significant to tourists as they frequently offer several benefits, including safety, education, nature protection and more [17]. Besides, tourist trail paths can perform different roles (i.e., recreational, ecological, economic), as stated in a study by [18]. Moreover, tourism trail paths could contribute to cross-border integration [19]. According to the United Nations World Tourism Organization (UNWTO), tourism focusing on walking is widely embraced as one of the most popular and desired ways to experience a place one is visiting. In addition, if properly developed and effectively managed, trail paths can result in social and economic advantages and benefits for communities and local residents [20]. During the actual development phase, some of the core elements included route characteristics (attractiveness, safety, difficulty level, accessibility) and essential facilities (appropriate signage, provision of toilets, benches), appropriate maintenance, economic opportunities (accommodation, retail, cultural programs linked to gastronomy and wine tourism), as well as commercialization and marketing [20].

In addition, hiking trails (HTs) have gained great attention and are a global trend in modern leisure activities [21]. Hiking trails are thought to be a potential option for local development in rural areas, although they are not limited exclusively to them [22]. Mountaineering hiking trails serve forest recreation, mountain tourism and ecotourism, as well as an approach to inaccessible areas or points in the countryside and traditional settlements and, at the same time, they contribute to the promotion of natural and cultural heritage, as well as to raising awareness in order to enhance the protection of these areas [23]. Furthermore, hiking paths relate to the quest for cultural roots [24]. McNamara and Prideaux [25] revealed that hiking has important effects on the ecosystem. Törn et al. [26], carried out a comparative study in Finland and analyzed how hiking, cross-country skiing and horse riding might affect trail characteristics and vegetation. There was evidence that trail characteristics and vegetation had a strong correlation with recreational activity, research site and forest type [26].

To contextualize the scale of trail path networks, it is worth noting that the European long-distance trail paths (E-paths) form an extensive network of footpaths spanning across Europe, consisting of 12 trail paths (E1 to E12) covering more than 65,000 km [27,28]. In Canada, the Trans Canada Trail (TCT) has earned the distinction of being the world’s longest trail, currently stretching over 27,000 km, and it is famous for providing a platform for millions of users to engage in a diverse array of outdoor pursuits [29]. Particularly, in Greece, many kilometers of trail paths and routes were unfortunately lost, expropriated or totally abandoned [23,30]. However, by securing funds from councils, hiking clubs and associations, as well as various individuals, the maintenance and revival of large sections of trail paths and routes were finally managed. This ultimately resulted in a creation of a large network of trail paths and routes throughout Greece, having a total length of about 3500 km [23,30].

A recent and relevant study in Greece by Kantartzis et al. [31] investigated the attitudes and perceptions of citizens at the Evros region, regarding the contribution of trail paths to protecting, promoting and enhancing cultural and natural environments. It was observed that the main contributory factor was age, which directly influences attitudes and perceptions of local citizens towards trail paths, and it was revealed that younger users had more positive views than older users [31]. However, the efficient management of a trail path is necessary for it to be seamlessly used by visitors. For instance, a study in Greece by Mertzanis et al. [32] proposes a number of interventions in order to manage trail paths (Hercules’ Trail to be specific) in an efficient manner, such as critical trail path infrastructure (construction of trail structures, wooden steps and boardwalks, litter bins), appropriate signage (e.g., destination signs, etc.), improvement of the trail path (e.g., widening/clearing, tree and shrub trimming), maintenance (accessibility, cleanliness, etc.), visitor monitoring and management and promotion and marketing.

Hence, within this context, the objective of the present study is to add to the current knowledge of this increasing timely topic, by analyzing user profiles according to their preferences and attitudes, as well as trail path choices (please see Figure 1 for an overview of the successive step sequence of the present study). More specifically, a questionnaire survey was used to analyze the attitudes and perceptions of respondents regarding the trail path usage conditions on Peranthi Hill [33] in Arta [34], a medium-sized city in northwestern Greece. Arta is the capital of the greater Arta region, which is part of the Epirus region. The population of the municipality was 43,166 citizens in the year 2011 [34]. Peranthi Hill is a low geological formation (elevation 218 m), on the outskirts of which is built the city of Arta. Today, Peranthi Hill is the suburban forest of the city of Arta, and it is an ideal place for walking, running, mountain biking and exploring [35].

To our knowledge, the relevant studies using questionnaire surveys in the field are relatively limited [9,10,11,31,36] and do not employ clustering techniques for grouping users. Hence, we attempt to add to the current knowledge in the field, as our ultimate goal was to understand and create trail user profiles by applying clustering methods.

A key element of this survey is the detailed mapping of the suburban trail path network of the city. Specific attention is given to questions about the types of activities people who walk or cycle on the trail paths engage in and the day, time, duration and distance of their trips, as well as their equipment (bicycles, clothes, shoes, etc.). Additionally, this survey has a specific focus on questions about maintenance and safety issues (e.g., security, guarding, cleaning) and infrastructure, as these issues are considered necessary in order to improve the trail paths’ environment and their usage. Considering that several routes are developed on Peranthi Hill with distinct layout differences, the preferred layout is explored in relation to each activity selected by users.

This survey aims to address these issues and assist stakeholders and policymakers in improving the utility and usability level of the trail paths, as sustainable trail path planning and management can create a better trail path network for visitors [37]; hence, the present study could further contribute to this field. Finally, the scope of the present research is to be a case study for future improvements on trail path environments in relative areas.

2. Materials and Methods

2.1. Case Study’s Trail Paths Description

Three distinct round routes constitute this case study’s trail paths, all of which are located on Peranthi Hill in the Arta region in western Greece (Figure 2). They mainly serve recreational purposes, although many amateur athletes carry out their workouts on them.

Peranthi Hill is the periurban forest of the city of Arta and is an ideal place for walking, running, mountain biking and exploring. In recent years, the Arta Runners Association, in cooperation with the Mountaineering Club of Arta, cleaned and signaled the trail paths on the hill. Because of the great need for outdoor activities during the COVID-19 pandemic, (e.g., sports and contact with nature), people initially voluntarily cleared the forgotten trail paths that were built about 40 years ago on Peranthi Hill. This triggered the opening of new trail paths which connected with the old paths, and a network of about 9 km of trail paths was created. All trail paths are marked and mapped so that their use is safe.

From a technical point of view, the trail paths are characterized as either easy, moderate or advanced (see

Table 1. Layout characteristics of the considered routes (trail paths).

	Route 1	Route 2	Route 3
Length (km)	1.57	1.74	2.36
Max grade (%)	12	16	18
Total ascent (m)	48	46	115
Classification	Easy	Moderate	Advanced

). The surfaces of the trail paths are either earthy, rocky or mixed (Figure 3). Along the routes, there are archaeological locations of interest from ancient Greece but also more recent monuments from World War II, such as pillboxes (Figure 4). Visitors of the hill can enjoy the magnificent view of the panoramic landscape of the lowland area of Arta to the Amvrakikos Gulf, Tzoumerka, the Pournari Dam, the artificial lakes of the Arachthos River and the lakeside park (Figure 5). Each user of the trail paths can select the route and the length they wish to travel, as well as their preferred landscape. Moreover, they can choose between the trail path that lies inside the forest (with cypress and pine trees) or, alternatively, the clear path with the panoramic view. Additionally, in the year 2021, the 1st Arta City Trail was held by the Arta Runners Association.

2.2. Questionnaire Survey

The present paper describes the findings of a survey of trail users that was conducted during spring 2023, after ethical approval for the survey was obtained. The Arta Runners Association and the Mountaineering Club of Arta (both located in the study area) were the two entities that circulated the questionnaire via email, social media and other channels (hard copy handout) to trail users. Additionally, it was made available through a Google survey app.

The questions in the anonymous questionnaire survey consist of various aspects of urban trail parameters, and they are based on a workbook which provides guidelines on how to conduct a trail-user survey [38]. In the questions related to the qualitative assessment of the existing characteristics of the trail paths, as well as the evaluation of the need for future improvements, a 5-point Likert scale was used. The qualitative characteristics of the trail paths could be evaluated by the respondents as “Very poor”, “Poor”, “Adequate”, “Good” or “Excellent”, while the eventual improvements and additional facilities along the trail paths could be assessed as “Not necessary at all”, “Rather unnecessary”, “Indifferent”, “Necessary” or “Absolutely necessary”.

There were 128 completed questionnaires in total, the majority of which were local respondents. More details regarding the demographics and other main questions are presented in Section 3.1 below, which presents the preliminary analysis of the sample. Ethical approval was acquired before the launch of the survey by the responsible Committee at the Civil Engineering Department of University of Thessaly, Greece.

2.3. k-Means Cluster Analysis

In order to categorize respondents and create meaningful profiles on the basis of a number of related variables (as discussed later in the paper), a k-means cluster analysis was carried out. Our k-means cluster analysis, the most commonly used unsupervised machine learning partitioning algorithm, attempts to identify relatively homogeneous groups of respondents and contributes to the identification of patterns and inherent relationships in the respondents of the questionnaire survey by splitting the trail users into k groups (clusters) with similar characteristics according to their statements concerning the evaluation of trail paths [39].

3. Results

3.1. Preliminary Analysis

The questionnaire sample consists of 128 responses. A total of 81.2% of respondents were male, while 18.8% were female, and none of them declared other or nonbinary gender. The dominance of males in the questionnaire survey was expected and is considered inevitable, as it reflects the ratio of males and females participating in the two entities (the Arta Runners Association and the Mountaineering Club of Arta) that circulated the questionnaire. A similar unequal distribution of males and females has also been observed in relevant past literature in the field [36].

Regarding age distribution, more than one out of two of respondents were between 36 and 55 years old, while around 28% of the sample were younger than 35 years old, and about 17.2% were older than 55 years old.

Table 2. Descriptive statistics of main variables considered.

Variable (Question of the Survey)	Percentage Distribution
Gender:	Man	Women	Other
	81.2%	18.8%	-
Age:	≤35	36 to 55	≥56
	28.1%	54.7%	17.2%
Usage frequency of trail paths:	≥3 times/week	1–2 times/ week	2–3 times/ month	≤once a month
	19.5%	17.2%	14.8%	48.5%
Personal main activity:	Walking alone or with human company	Walking with pet	Jogging	Bicycling
	53.1%	3.9%	40.6%	2.4%
Main days of visit:	Only weekends	Only weekdays	Any day of the week
	32.8%	6.3%	60.9%
Part of day:	until 11:00	11:00 to 15:00	15:00 to 19:00	after 19:00
	15.6%	17.2%	65.6%	1.6%
Activity duration on each visit:	≤1 h	1 to 2 h	≥2 h
	35.9%	52.4%	11.7%
Appropriate use of trail paths:	Recreation	Health, exercise	Movement, commuting	Training
	5.5%	60.1%	5.5%	28.9%
Level of trail path maintenance:	Very poor	Poor	Adequate	Good	Excellent
	-	7.9%	45.3%	35.9%	10.9%
Level of trail path cleanliness:	Very poor	Poor	Adequate	Good	Excellent
	3.1%	7.9%	31.2%	44.5%	13.3%
Level of trail path safety and security:	Very poor	Poor	Adequate	Good	Excellent
	4.7%	7.1%	27.3%	44.5%	16.4%
Need for parking spaces near trail paths:	Not necessary at all	Rather unnecessary	Indifferent	Necessary	Absolutely necessary
	28.1%	7.8%	15.6%	24.2%	24.2%
Need for additional viewing areas near the trail paths:	Not necessary at all	Rather unnecessary	Indifferent	Necessary	Absolutely necessary
	15.6%	8.7%	24.2%	24.2%	27.3%
Need for rest benches along the trail paths:	Not necessary at all	Rather unnecessary	Indifferent	Necessary	Absolutely necessary
	7.8%	7.1%	31.2%	25.8%	28.1%
Need for drinking water along the trail paths:	Not necessary at all	Rather unnecessary	Indifferent	Necessary	Absolutely necessary
	4.7%	10.2%	16.4%	35.1%	33.6%
Need for directional signs alongside trail paths:	Not necessary at all	Rather unnecessary	Indifferent	Necessary	Absolutely necessary
	-	2.3%	9.4%	34.4%	53.9%
Need for children recreation areas:	Not necessary at all	Rather unnecessary	Indifferent	Necessary	Absolutely necessary
	2.3%	5.5%	21.1%	38.3%	32.8%
Need for improving accessibility fordisabled users:	Not necessary at all	Rather unnecessary	Indifferent	Necessary	Absolutely necessary
	2.3%	5.5%	14.1%	28.1%	50.0%
Need for guarding the trail paths:	Not necessary at all	Rather unnecessary	Indifferent	Necessary	Absolutely necessary
	23.4%	23.4%	9.4%	18.8%	25.0%
Volunteering on trail path tasks:	No	Yes
	22.7%	77.3%
Willingness to pay an annual fee for trail path improvement:	No	Yes, until EUR 10	Yes, until EUR 20
	48.4%	42.2%	9.4%
Preferable trail path for recreation:	Trail 1	Trail 2	Trail 3	All	None
	21.6%	19.7%	20.6%	33.3%	4.8%
Preferable trail path for health, exercise:	Trail 1	Trail 2	Trail 3	All	None
	25.8%	34.1%	37.6%	2.5%	-
Preferable trail path for training:	Trail 1	Trail 2	Trail 3	All	None
	25.0%	26.6%	32.8%	4.7%	10.9%

below provides the descriptive statistics of the main variables (questions of the survey) considered in our analysis.

3.2. Cluster Analysis for Identifying Groups of Trail Users

The selection of the appropriate number of clusters was based on the well-known Elbow method, a graphical technique in which initially the total within-cluster sum of squares is calculated as a function of the number k of clusters. As the number of clusters increases, the total within-cluster sum of squares starts to decrease. When adding more clusters does not significantly improve the total within-cluster sum of squares, we can choose the corresponding value of k as the optimal number of clusters, avoiding overfitting by creating too many clusters [40].

The cluster analysis that was carried out revealed two groups (clusters) of trail path users. According to our analysis, the optimal number of clusters was found to be two (Figure 6). Convergence was achieved due to no or small change in cluster centers. Moreover, based on the cluster analysis and the reduction in the sum of squared distances when adding more clusters (as shown in Figure 6 below), the selection of two clusters reduces the sum of squared distances by 41.47%. When attempting to increase the number of clusters from two to three (a 50% increase), it resulted in only an additional 4.8% reduction in the sum of squared distances. Therefore, the optimal number of clusters was two, as an attempt to add more clusters would potentially increase the risk of overfitting without significantly contributing to a substantial reduction in the sum of squared distances or to the interpretability of results.

The maximum absolute coordinate change for any center is 0.000 after five iterations. The minimum distance between initial centers was found to be 9.110.

Table 3. Analysis of variance and F-test for each clustering variable—Cluster centers.

	Cluster		Error		Univariate F Test		Cluster Centers
	Mean Square	df	Mean Square	df	F	Sig.	1	2
Level of trail path maintenance	8.127	1	0.570	126	14.247	0.000	3.214 *	3.722 *
Level of trail path cleanliness	10.214	1	0.787	126	12.980	0.000	3.250 *	3.819 *
Level of trail path safety and security	36.969	1	0.710	126	52.045	0.000	3.000 *	4.083 *
Need for parking place near the trail paths	79.961	1	1.810	126	44.171	0.000	3.982 **	2.389 **
Need for additional viewing areas near the trail paths	64.643	1	1.411	126	45.804	0.000	4.196 **	2.764 **
Need for rest benches along the trail paths	75.446	1	0.837	126	90.168	0.000	4.464 **	2.917 **
Need for drinking water along the trail paths	44.941	1	0.963	126	46.691	0.000	4.500 **	3.306 **
Need for directional signs alongside trail paths	17.813	1	0.435	126	40.906	0.000	4.821 **	4.069 **
Need for children’s recreation areas	37.786	1	0.680	126	55.545	0.000	4.554 **	3.458 **
Need for improving accessibility for disabled users	26.583	1	0.844	126	31.515	0.000	4.696 **	3.778 **
Need for guarding the trail paths	102.691	1	1.582	126	64.930	0.000	4.000 **	2.194 **

* 1: Very poor, 2: Poor, 3: Adequate, 4: Good, 5: Excellent. ** 1: Not necessary at all, 2: Rather unnecessary, 3: Indifferent, 4: Necessary, 5: Absolutely necessary.

shows the variables considered in the cluster analysis and presents the final cluster centers.

The distribution of responses across the two clusters is as follows: 43.75% of respondents are assigned to Cluster 1, while 56.25% belong to Cluster 2. Figure 7 illustrates the distribution of the trail path users in the two clusters per gender. It can be observed that 64.3% of walkers in Cluster 1 are male, and none of those in Cluster 1 declared “other gender/nonbinary”. However, Cluster 2 consists mainly of male respondents (more than 90%). As for age, Cluster 1 consists mainly of people aged between 36 and 55 years old (Figure 8). On the other hand, age distribution in Cluster 2 is more balanced, since it involves a large proportion of younger as well as middle-aged walkers (41.7% and 38.9%, respectively).

The cluster analysis also shows the distribution of the variables selected in

Table 4. Final cluster distribution.

Variable (Question of the Survey)	Cluster 1					Cluster 2
Usage frequency of trail paths:	≥3 times/week	1–2 times/week	2–3 times/month	≤once a month		≥3 times/week	1–2 times/week	2–3 times/month	≤once a month
	7.1%	16.1%	10.7%	66.1%		29.2%	18.1%	18.1%	34.6%
Personal main activity:	Walking (with or without pet)	Jogging, cycling				Walking (with or without pet)	Jogging, cycling
	73.2%	26.8%				44.4%	55.6%
Main days of visit:	Only weekends	Only weekdays	Any day of the week			Only weekends	Only weekdays	Any day of the week
	50.0%	10.7%	39.3%			19.4%	2.8%	77.8%
Part of day:	Until 11:00	11:00 to 15:00	15:00 to 19:00	after 19:00		until 11:00	11:00 to 15:00	15:00 to 19:00	after 19:00
	28.6%	7.1%	64.3%	-		5.6%	25.0%	66.7%	2.8%
Activity duration on each visit:	≤1 h	1 to 2 h	≥2 h			≤1 h	1 to 2 h	≥2 h
	57.1%	39.3%	3.6%			19.4%	62.5%	18.1%
Appropriate use of trail paths:	Recreation	Health, exercise	Movement, commuting	Training		Recreation	Health, exercise	Movement, commuting	Training
	10.7%	80.4%	1.8%	7.1%		1.4%	44.4%	8.3%	45.9%
Level of trail path maintenance:	Very poor or poor	Adequate	Good or excellent			Very poor or poor	Adequate	Good or excellent
	14.3%	53.6%	32.1%			2.8%	38.9%	58.3%
Level of trail path cleanliness:	Very poor or poor	Adequate	Good or excellent			Very poor or poor	Adequate	Good or excellent
	7.1%	60.8%	32.1%			13.9%	8.3%	77.8%
Level of trail path safety and security:	Very poor or poor	Adequate	Good or excellent			Very poor or poor	Adequate	Good or excellent
	25.0%	39.3%	35.7%			1.4%	18.1%	80.5%
Need for parking spaces near trail paths:	More or less unnecessary	Indifferent	More or less necessary			More or less unnecessary	Indifferent	More or less necessary
	14.3%	17.9%	67.9%			52.8%	13.9%	33.3%
Need for additional viewing areas near the trail path:	More or less unnecessary	Indifferent	More or less necessary			More or less unnecessary	Indifferent	More or less necessary
	3.6%	17.9	78.5%			40.3%	29.2%	30.5%
Need for rest benches along the trail paths:	More or less unnecessary	Indifferent	More or less necessary			More or less unnecessary	Indifferent	More or less necessary
	-	10.7%	89.3%			26.4%	47.2%	26.4%
Need for drinking water along the trail paths:	More or less unnecessary	Indifferent	More or less necessary			More or less unnecessary	Indifferent	More or less necessary
	7.1%	0	92.9%			20.8%	29.2%	50.0%
Need for directional signs alongside trail paths:	More or less unnecessary	Indifferent	More or less necessary			More or less unnecessary	Indifferent	More or less necessary
	-	-	100%			4.2%	16.7%	79.1%
Need for children recreation areas:	More or less unnecessary	Indifferent	More or less necessary			More or less unnecessary	Indifferent	More or less necessary
	-	3.6%	96.4%			13.9%	34.7%	51.4%
Need for improving accessibility fordisabled users:	More or less unnecessary	Indifferent	More or less necessary			More or less unnecessary	Indifferent	More or less necessary
	-	3.6%	96.4%			7.8%	21.1%	71.1%
Need for guarding the trail paths:	More or less unnecessary	Indifferent	More or less necessary			More or less unnecessary	Indifferent	More or less necessary
	14.3%	7.1%	78.6%			72.2%	11.1%	16.7%
Volunteering on trail path tasks	No	Yes				No	Yes
	17.9%	82.1%				26.4%	73.6%
Willingness to pay an annual fee for trail path improvement:	No	Yes, until EUR 10	Yes, until EUR 20			No	Yes, until EUR 10	Yes, until EUR 20
	46.4%	42.9%	10.7%			50.0%	41.7%	8.3%
Preferable trail path for recreation:	Trail path 1	Trail path 2	Trail path 3	All	None	Trail path 1	Trail path 2	Trail path 3	All	None
	22.9%	20.1%	19.9%	29.0%	8.1%	20.5%	19.4%	21.2%	36.5%	2.4%
Preferable trail path for health, exercise:	Trail path 1	Trail path 2	Trail path 3	All	None	Trail path 1	Trail path 2	Trail path 3	All	None
	32.2%	35.7%	32.1%	-	-	20.8%	32.9%	41.8%	4.5%	-
Preferable trail path for training:	Trail path 1	Trail path 2	Trail path 3	All	None	Trail path 1	Trail path 2	Trail path 3	All	None
	32.2%	21.5%	35.7%	3.5%	7.1%	19.4%	30.6%	30.6%	5.5%	13.9

below. As mentioned earlier, the cluster analysis was conducted on the basis of the selected variables, as illustrated in Table 3. In Table 4, and in order to make the characteristics of the participants in each of the two clusters clearer and more distinguishable, certain responses from the respondents are grouped together. For example, the choices “Not necessary at all” and “Rather unnecessary” are summed together and are presented (and calculated) in Table 4 as “More or less unnecessary”. Similarly, activities along the trail paths that require similar physical exertion are also grouped together (e.g., jogging or cycling as opposed to walking with or without pet).

4. Discussion

4.1. Analysis and Discussion of Clusters

The preliminary analysis as well as the cluster analysis that was subsequently carried out, revealed insightful trends and results regarding walkers’ attitudes and preferences regarding the trail paths at Peranthi Hill, Greece. Similar circular trail paths have also been examined in past literature [32]. A detailed description of the two clusters is presented in this section.

Regarding the usage frequency of the trail paths, it was found that more than two out of three respondents that belong to Cluster 1 are very occasional users (fewer than one visit per month), while the other visit frequencies generally vary from 7% to 16%. On the other hand, Cluster 2 is a more balanced cluster, as it varies from 18.1% to 34.6% in all usage frequencies. Cluster 1 also includes respondents whose main activity is walking (three out of four), while the more active users (jogging, cycling) belong to Cluster 2. Another interesting remark is that half of Cluster 1 walkers visit the hill mainly on weekends and almost 40% at any day of the week. However, the majority of walkers in Cluster 2 stated that they visit Peranthi Hill on any day of the week (77.8%). Thus, they seem to focus on regular exercise lasting from 1 to 2 h (62.5%). Additionally, they exercise mainly from 15:00 to 19:00 (two out of three) and from 11:00 to 15:00 (one out of four), while in Cluster 1, they visit the hill from 15:00 to 19:00 (two out of three) and until 11:00 (fewer than one out of three). Respondents in both clusters, however, state that the main reasons are health and exercise; however, almost one out of two respondents in Cluster 2 also mention that they visit for training.

Moreover, both clusters involve people who are mainly interested in participating in volunteering actions. Similarly, the majority of respondents in both clusters (around 90% in each cluster) would either disagree or be willing to pay EUR 10 maximum for improving the trail paths.

Another important group of questions involved the attitudes and perceptions towards various aspects of the trail paths, such as maintenance, cleanliness, safety, security, availability of parking spaces near the trail paths, sight points, rest points, drinking water, guide signs and the availability of children’s recreation areas, as well as potential need for improved accessibility for disabled users. These questions indicate a serious difference in beliefs and perceptions among the two groups/clusters. Generally, the majority of respondents in Cluster 1 feel that maintenance, cleanliness and safety/security are adequate (53.6%, 60.8% and 39.3%, respectively). It seems, however, that Cluster 2 mainly evaluates these aspects as excellent, with high percentages (58.3%, 77.8% and 80.5%, respectively). A similar major difference in perceptions is also observed when asked about the necessity of parking spaces near trail paths: 67.9% of Cluster 1 state that this is necessary infrastructure, while two out of three walkers in Cluster 2 evaluate this as unnecessary or feel indifferent.

Similar differences in perceptions among the two clusters are also observed concerning the necessity of adding sight points and rest points alongside trail paths, as well as sufficient guarding of the trail paths. Nonetheless, it is interesting to note that the two clusters generally agree that drinking water, guide signs and children’s recreation areas should be present alongside the trail paths, showing high percentages overall and less variation. It is also notable that walkers in both clusters generally show similar trends when answering the questions regarding trail path choice, with very slight differences. In summary, all trail path choices for recreation are equally preferred by both clusters (around one out of 4 in each choice in each cluster). When health/exercise and training are considered, no major differences are observed either.

4.2. Identification of Types of Trail Path Users

Based on the cluster analysis results, the next step is a detailed description of the two clusters, which is as follows:

• Cluster 1 “Indifferent occasional users”: Cluster 1 shows no particular preference for any trail path regardless of use. Even the increased percentage in the question of preferred trail path for training reasons in trail paths 1 and 3 is insignificant, since this cluster is mainly characterized by occasional use of the trail paths (fewer than once a month), mainly for short walks.
• Cluster 2 “Fitness users”: On the contrary, Cluster 2 shows a trend of use for all questions on the 3rd trail path which, according to Table 1, is also the most demanding due to its longer length, steep gradients and overall positive elevation. When users state that the reason for selecting the trail paths is health, exercise and training, the superiority of the 2nd and 3rd path is observed, which means that Cluster 2 (which is also the most fitness-oriented cluster) prefers paths of increased technical difficulty.

4.3. Summary of Research Findings

In conclusion and according to what was described in the previous section, the indifferent occasional users in Cluster 1 appear to be more critical in their evaluations of the existing qualitative characteristics of the trail paths (maintenance, cleanliness, safety and security) and more demanding regarding eventual improvements of the trail paths (parking spaces, viewing areas, rest benches, drinking water, directional signs, children’s recreation areas, accessibility for disabled users, trail path guarding). On the contrary, the fitness users in Cluster 2 are more tolerant of the existing shortcomings of the trail paths and less demanding regarding eventual future improvements. Hence, it is clear that the fitness users in Cluster 2 mainly use the trail paths for specific personal needs related to improving their physical condition and physical strength, giving less importance to the qualitative elements of the trail paths, in contrast to the indifferent occasional users in Cluster 1, who, although having shorter and more sporadic contact with the trail paths, are nonetheless more demanding.

Overall, our analysis showed a number of interesting findings which could contribute to sustainable trail path planning and management in order to create a better trail path network for visitors, as suggested by [37]. However, we feel that direct comparisons of our findings with past literature should be made with care ass to our knowledge, we employed different analysis methods from similar studies in the past, as for instance in [31]. In that specific study [31], the authors state that the gender, age and income of respondents may affect the experience of using the trail paths. In our study, there is some evidence that gender might affect user experience, because Cluster 2 (the fitness users) mainly consist of males, who are less demanding than the occasional users mentioned above. On the other hand, there was no information about trail path users’ income, while age distribution did not clearly show any solid trend (Figure 8).

Some of our preliminary findings (descriptive statistics) were found to be consistent with the findings of past reports in the US [41] showing, for example, that the top motivations for using a trail path include exercise, fun and overall health/wellness reasons. Similar results were found in Turkey [10]. However, in our study, recreation/fun was ranked very low. Another potential difference with other relevant studies might be the fact that a non-negligible percentage of our sample used the trail paths for less than 1 h (around 36%), while some evidence from the international literature shows that only 17% of users [41] used trail paths for less than 1 h. On the other hand, in a recent study [9], it was observed that about 46% of users spent less than 1 h on trail paths. Lastly, it is notable that the need for water was among the top priorities of users in our study, as well as other relevant studies in the field [10].

4.4. Limitations and Directions for Further Research

The present research paper does not come without limitations. Our study focuses on a specific area of Greece (Arta) and, as such, it utilizes a relatively limited sample size of locals who use the considered trail paths. Also, the impact of topographic alignments on user perceptions was not considered in this study, such as, for example, a sustainability assessment of trail paths on the basis of topographic alignments [37,42,43].

Future research would benefit from utilizing a larger nationwide sample or more case study areas, for example, in large urban areas such as Athens or Thessaloniki. The inclusion of more genders, locals and tourists, as well as users from various running associations and entities should also be attempted so that they are more widely represented in surveys. Moreover, further research would benefit from carrying out more advanced statistical analyses and machine learning methods to further explain or predict trail path choice according to a number of influential factors. Lastly, the transferability of our results further adds to the knowledge on the topic.

5. Conclusions

The current paper aimed to carry out an analysis of a questionnaire survey which examined the trail path usage conditions on Peranthi Hill in Arta, a medium-sized city in northwestern Greece. For that reason, three round routes (trail paths) constituted the case study’s trail paths, all located on Peranthi Hill. The ultimate aim of the study was to add to current knowledge in the field by identifying meaningful profiles of trail path users. For that reason, a k-means cluster analysis was carried out.

A key element of the survey was the detailed mapping of the suburban trail path network of the city. The survey also encompasses questions about the types of activities people who walk or bike on the trail paths engage in and the day, time, duration and distance of their trips, as well as their equipment. Further questions examining various maintenance and safety issues, such as security, guarding, cleanliness and necessary infrastructure to improve the trail paths’ environment and use were also considered. The results of the k-means cluster analysis reveal two distinct clusters (i.e., groups of trail path users); the first mainly consists of indifferent occasional users, while the second consists regular fitness-oriented users. Furthermore, additional attitudes and perceptions of the users in each cluster were identified and generally showed that each group has different priorities.

Overall, the findings of the study aimed to identify related issues and assist stakeholders and policymakers in improving the utility and usability level of the trail paths in this area. Further studies should attempt to carry out additional analyses and utilize larger nationwide samples.