Selection criteria

In this study, fifteen (15) tourist attraction sites were selected based on accessibility and location criteria, with a preference for those with easy access, located on the right and left margins of highways BR-174 and AM-240, Presidente Figueiredo municipality, State of Amazonas, Brazil (Figure 1). We selected attraction sites within and outside the "Caverna do Maroaga" Environmental Protection Area (APA, for its acronym in Portuguese), particularly those close to deforested areas, using data on forest cover loss up to 2017 from the Project for Satellite Monitoring of Deforestation in the Legal Amazon of the National Institute for Space Research (PRODES/INPE), and, above all, those with access trails to natural attributes (Table 1).

Source: elaborated by the author.

Figure 1 Location of the study area, Presidente Figueiredo, Amazonas state, Brazil. 

The tourist attraction sites evaluated are located in public and private areas. The type of use of each attraction site by community members and users depends on its geographic location and the natural access conditions (^{Reis, 2010a}). On the other hand, one of the main factors indicating natural areas managed as attraction sites was the high frequency of visits and the type of recreation developed or its potential use for tourism, spa, field research, hiking, and wildlife observation. Some attraction sites, such as "Santa Claudia", restrict public access because of a private spring fountain.

This study investigated the influence of human pressure on the biophysical environment along the trails of these attraction sites. Unsustainable tourist activities such as conventional tourism may reduce natural resources, including forest loss or degradation in public and private areas (^{Medeiros and Moraes, 2013}). We used the tourist carrying capacity (TCC), a methodology developed by Miguel Cifuentes in 1992 and improved in 1999, considered a simple and user-friendly tool compared to other methods (Limit of Acceptable Change, LAC, among others), to calculate the number of visits per day that a tourism location can support, considering physical, biological, environmental, social, and management variables. This methodology was developed to control and reduce the adverse impacts of conventional tourism, promote sustainable tourism, and provide a high degree of satisfaction to users. In this work, the tourist attraction sites evaluated are restricted to the natural attributes of the study area, according to the routes of trails within or near natural attractions (forests, waterfalls, caves, and others).

Field Data Collection

Field data were collected in fifteen (15) tourist attraction sites. The data collected were trail length and width using an analog tape measure (winch, 50 m), georeferenced data with GPS navigation (GARMIN model 76CSx), slope with an inclinometer (%), and description of the biophysical characteristics in a field notebook, to calculate the physical and actual carrying capacity of each trail. Surveys consisting of closed questions were also conducted with public officials, private officials, and landowners to determine the current status of the handling capacity of each tourist attraction, which is essential to calculate the effective carrying capacity.

Tourism Carrying Capacity (TCC)

We used the methodology by ^{Cifuentes (1992}, ¹⁹⁹⁹⁾ to determine the flow of visitors to the tourist attraction sites evaluated. This methodology establishes the maximum number of visits per day that a tourist area (natural, cultural, historical) can hold according to its physical, biological, environmental, and handling conditions. TCC includes three categories: Physical Carrying Capacity (PCC), Actual Carrying Capacity (ACC), and Effective Carrying Capacity (ECC).

Physical Carrying Capacity (PCC)

The maximum number of visits that a given attraction site can support was determined according to the area and time available (visiting hours and duration of visits). The following equation was applied (^{Cifuentes, 1992}; ^{Cifuentes et al., 1999}):

Where:

PCC = Physical Carrying Capacity (number of visits day^-1); S = overall trail length (m); sp = distance traveled per person (m); NV = number of times that a trail can be visited by the same person in one day. This variable is estimated with the following equation:

Where:

H _v = visiting hours (hours day ^-1); T _v = time needed to visit each trail (hours visit ^-1 person^-1).

Actual Carrying Capacity (ACC)

ACC was determined with the PCC and correction factors (CF) defined based on the characteristics of each location (physical, environmental, social, and management variables). The following equation was applied (^{Cifuentes, 1992}; ^{Cifuentes et al., 1999}):

Where:

CF _x = correction factor for variable “x”; Ml _x = limiting magnitude of variable “x”, and Mt _x = total magnitude of variable “x”.

In this study, variables were used according to the particular characteristics of each trail. The following variables were considered: social, erodibility, accessibility, precipitation, sunlight, waterlogging, vegetation (forest degradation), and biological (wildlife disturbance). Table 2 summarizes the CFs considered (^{Cifuentes et al., 1999}):

After calculating the CFs, the ACC was determined using the following equation (^{Cifuentes et al., 1999}):

Where:

ACC = actual carrying capacity (number of visits day^-1); PCC = physical carrying capacity (number of visits day^-1); CF _xyz = correction factor _(xyz).

Effective Carrying Capacity (ECC)

ECC was calculated by multiplying ACC by handling capacity (HC) according to the following equation (^{Cifuentes et al., 1999}):

Where:

ACC = actual carrying capacity (number of visits day^-1); HC = handling capacity.

To calculate HC, infrastructure, equipment, and personnel variables were established, which are directly related to handling the flow of visits. According to ^{Cifuentes et al. (1999)}, these criteria adequately represent all the options to assess and determine HC; they can provide the necessary indicators to obtain a good approximation. They were selected based on their facility of measurement, analysis, and availability of information for each tourist attraction site. Each variable comprises four criteria: quantity, state, location, and functionality. For the "Personnel" variable, only the quantity criterion was considered.

The percent scale (Table 3) used to evaluate the criteria of each variable is an adaptation of the standard of the International Organization for Standardization (ISO 10004), which was tested and used in quality evaluation studies of services offered by public and private companies to measure the degree of user satisfaction (^{Cifuentes et al., 1999}).

The HC was calculated using the following equation (^{Cifuentes et al., 1999}):

Where:

Infr = Infrastructure; Equip = Equipment; Pers = Personnel

the number of visitors per trail was determined with the following equation (^{Cifuentes et al., 1999}):

Where:

VD = visitors per day to the trail of the tourist attraction site(visitors day ^-1); ECC = effective carrying capacity (number of visits day^-1); NV = number of times that the trail can be traveled by the same person in a single day (number of visits day^-1 person^-1).

Physical Carrying Capacity (PCC)

Considering the PCC values calculated for the fifteen trails of tourist attraction sites (Figure 2), those with the lowest PCC were "Santa Claudia" (7292.70 visits day^-1), “Quatro Quedas” (7803.3 visits day^-1), and “Santuário” (8113.6 visits day^-1). By contrast, those with the highest PCC were “Caverna do Maroaga” (13150.86 visits day^-1), “As Orquídeas” (12655.86 visits day^-1), and “Bromélias” (12569.37 visits day^-1).

Figure 2 Physical carrying capacity (PCC) of the trails evaluated. 

However, the results obtained herein for the “Caverna do Maroaga” trail differ from those reported by ^{Porto et al. (2013)}, who determined a PCC of 11,272.95 visitors per day. In that case, it was confirmed that PCC should be expressed as “number of visits day^-1” rather than “visitors day^-1” because the number of visits per day (NV) is related to the visiting hours (hours day^-1) and the estimated time spent by a person during a visit to the trail (hours visit^-1). That is to say, it not only determines the maximum number of visits that a given trail can host but also helps prevent the overload of visitors in these ecosystems and formulate a tourism management policy to improve the control of visit flow (^{Liu et al., 2020}).

For example, in the "Santuário" trail - a private protected area - the estimated PCC was 8113.7 visits day^-1. In contrast, in the "Quatro Quedas" trail - a public protected area - the PCC was 7803.3 visits day^-1. Although these two attraction sites have different “statuses” (private versus public), they yielded similar results. This similarity derives from the relationship between total trail length and the number of visits per day. Hence, the delimitation and length of each trail, the criteria of space per person, visiting hours, and the estimated time spent by each visitor during the tour in these trails define the maximum visit flow per day that a given trail is able to support. In addition, the PCC values estimated in the present study reflect the spatial and temporal conditions of these sites. Thus, the criteria related to some parameters may be modified over time (depending on the scenario) and, therefore, affect the subsequent PCC results.

Actual Carrying Capacity (ACC)

Considering the ACC values calculated for the fifteen trails evaluated, those with the lowest ACC were "Quatro Quedas" (429.77 visits day^-1), "Bromelias" (467.51 visits day^-1), and "Neblina" (475.42 visits day^-1) (Figure 3).

Figure 3 Actual carrying capacity (ACC) of the trails evaluated. 

The social CF was not considered to estimate ACC for the trails “Perema”, “Pedra Furada”, “Santa Claudia”, “Mutum”, “Salto do Ipy”, and “Gruta do Raio” because these trails have a shorter total length (<450 m) and hold a visit flow of over twenty-five times per visitor in a single day, considering one meter as the space required by one visitor. In other words, applying this factor would further limit the carrying capacity but would not define it as a limit factor for the entire trail but as an alternative for dynamic and efficient tourism management to redistribute tourist flow (^{Chen et al., 2021}). In this context, the ACC of 541 visits day^-1 reported by ^{Porto et al. (2013)} for the “Caverna de Maroaga” trail was similar to the value calculated in the present study (581.58 visits day^-1). The CFs (Social CF, Wildlife Disturbance, Erodibility, and Accessibility) reduced the ACC values despite being carried out on the same trail; the differences between ACC values are due to the biophysical conditions and evaluation criteria established and measured at the time. These results reflect a multifaceted state because various factors influence the biophysical conditions and their performance according to the particular characteristics of each tourist attraction site.

In this context, despite the similarity with the studies conducted by ^{Porto et al. (2013)}, we consider that the ACC estimated in the present study is insufficient to propose adequate tourism management and establish guidelines to improve the control of visits. The characteristics of the “Caverna de Maroaga” trail can experience relatively significant changes in the short- and mid-term due to the lack of information, maintenance, or a more efficient control of the visit flow in this trail. In addition, if correction factors are not properly evaluated, either in their field application or due to a lack of understanding of the methodology, there will be "biases" in the data that may lead to incorrect results, especially when working with secondary data (sunlight, precipitation, and wildlife disturbance). However, it is highlighted that most of the CFs influenced the ACC values due to the physical geography and the particular conditions of each tourist attraction site. The carrying capacity should be established considering the lowest possible impacts aiming not to affect the vulnerability of forest ecosystems that provide ecosystem services (scenic beauty, maintenance of biodiversity, and regulation of natural cycles) (^{La Notte et al., 2017}; ^{Bachi et al., 2020}). Close attention should be paid when the maximum limit of visitors is approached or exceeded (tourist overload) because the optimal visit flow levels at a tourist attraction site vary according to the type of tourist, activities, and territory (^{Chen et al., 2021}). The Cifuentes method (1992, 1999) focuses only on quantitative data or figures and highlights that managers of each tourist attraction site are responsible for managing and handling tourism rather than for numbers (^{Lobo et al., 2013}).

Effective carrying capacity (ECC)

The trails that attained the lower ECC values (Figure 4) were “Bromelias” (133.51 visits day^-1), “Quatro Quedas” (183.49 visits day^-1), “Neblina” (245.63 visits day^-1), and “Caverna do Maroaga” (276.86 visits day^-1). In general, those attraction trails can receive that limited visit flow due to the poor state of the infrastructure, equipment, and personnel, which represents a handling capacity (HC) assessed as “marginally to moderately satisfactory”.

Figura 4 Effective Carrying Capacity (ECC) of the trails evaluated. 

However, they show better conditions of forest conservation, which allows visitors to enjoy ecosystem services (shade, scenic beauty, etc.) during their travel along the trail with a minimum human impact (Figure 4). It is also worth noting that trails with a good HC, such as "Complexo do Iracema Falls" and "Cachoeira da Onça", did not attain a high ECC due to the influence of CFs on the estimated values.

In addition, it is expected that the lower the estimated ECC value, the lower the concern regarding the number and management of visitors to the trails, considering the suggested maximum number of visitors per day that the local ecosystem is able to support. Given this scenario, the carrying capacity should be managed comprehensively and closely related to the basic services offered in any category, such as interpretation, accommodation, and food services. As regards the latter, the contractors in charge of providing these services should be monitored to prevent impairing ecotourism management. Human resources should be incorporated to implement strategies that favor the conservation of tourist resources, preventing their deterioration; they should always be present for the recreation of visitors, which contributes to the sustainability of ecotourism as a conservation strategy. All the above informs managers to make short-, mid-, and long-term decisions once the tool provides early warnings that identify environmental, biological, and social imbalances. This strategy will translate into proper considerations to support ecotourism activities that are feasible and consistent with the management plan of each protected area.

The ECC of the trails evaluated is the recommended maximum visit flow per day. The number of visits should be respected as indicated and proposed in this present study to promote sustainable tourism quality in terms of visits and low impact. Accordingly, the deterioration of the biophysical environment that could compromise the conditions (correction factors) and management of these tourist attraction sites is avoided or reduced. In this sense, based on the recommended visit flow, the trails “Complexo do Iracema Falls” (C. Sucurijú, Iracema, and de Araras), “Cachoeira da Onça” and “C. Neblina” showed the highest number of visitors per day and year, while the trails “Quatro Quedas”, “Santa Claudia”, and “Bromélias” obtained lower ECC values (Figure 5). It is worth noting that this carrying capacity varies over time; therefore, action strategies and conservation measures should follow this dynamic behavior (^{López-Dóriga et al., 2019}).

Figure 5 Recommended number of visitors (per year and day) to each trail evaluated. 

Analysis of Forest Degradation in Tourist Attraction Sites

The state of forest degradation (in terms of forest degradation by the impact of human activities, specifically on trails) and deforestation in the tourist attraction sites evaluated showed that the vegetation CF is directly related to forest degradation, driven by human activities and pressures (e.g., deforestation, forest clearing/burning, areas undergoing forest regeneration, pollution), causing impacts and changes in the local landscape. Figure 6 shows that the trails with the most significant impact on vegetation were “Perema” (40.3%), “Santuário” (26.9%), MNP “Galo da Serra” (26.9%), and “Pedra Furada” (26.1%). In addition, the percentage of areas of attraction sites that suffered the heaviest deforestation as of 2017, according to PRODES/INPE land-cover data, were “Santuário” (39.9%), “Santa Claudia” (38.6%), “Mutum” (33.8%), and “Salto do Ipy” (24.3%).

Figure 6 Percent loss of forest cover as of 2017. A) Total trail length; B) area per attraction site. 

Most of these attraction sites showed a "good" environmental conservation status because their natural attributes (e.g., forest, waterfall, grotto) were in favorable conditions to receive visits from users. However, some of them (except for "C. da Onça" and "Complexo do Iracema Falls") lack suboptimal tourist infrastructure to provide a high degree of satisfaction to visitors. For example, “Santuário”, the attraction site with the best tourist infrastructure (along with “C. da Onça” and “Complexo do Iracema Falls”), was the site with the most significant loss of forest cover (39.9%) according to land-cover data from ^{PRODES/INPE (2018)}. In addition, this site has an environmental license for aquaculture activities issued by the Amazon Environmental Protection Institute (IPAAM) regarding fish farming of Colossoma macropomum in a semi-intensive farming system in nine nurseries with a total area of 4.24 ha, classified by the IPAAM as a “medium” potential polluter/degrader.

The percent loss of forest cover as of 2017, specifically in tourist attraction sites, is related to the previous deforestation due to the construction and expansion of tourist infrastructure (hotels, restaurants, swimming pools) as tourist attractions, such as “Santuário” (39.9%) and “Santa Claudia” (38.6%), which showed high values of forest cover loss. In addition, in the case of “Mutum” (33.8%) and “Salto do Ipy” (24.3%), the previous and current deforestation are related to slash-and-burn agriculture, pastures, and livestock raising. ^{Reis et al. (2013)} also observed that the “Porteira” and “Sanctuary” sites already had a significant impact due to deforestation along the Mutum River, considered the watercourse that sustains these attractions. Besides, the local managers lack controls or monitoring of visits, management plans, public use plans, route circuits, or on-site signaling, so that the intervention of environmental agencies in the ecological licensing and monitoring of these attraction sites became a critical aspect.

Other studies on deforestation highlight the influence of road infrastructure on deforestation in the study area. For example, ^{Reis and Pinheiro (2010)} reported that deforestation in the “Caverna do Maroaga” EPA was concentrated along highways BR-174 and AM-240, rural settlements, and the margins of Permanent Preservation Areas (APP, for its acronym in Portuguese). This observation was confirmed in the present study, which found the same deforestation trend in attraction sites located along highways (BR-174 and AM-240) and in rural settlements, such as “Santuário” (39.9%), “Santa Claudia” (38.6%), “Mutum” (33.8%), and “Salto do Ipy” (24.3%). In addition, these results agree with the study carried out by ^{Massoca (2010)} within the framework of the Settlement Project - PA “Uatumã” of the National Institute of Colonization and Agrarian Reform - INCRA of Brazil, indicating that the opening and expansion of secondary roads and forest clearing and burning for agriculture are more common in rural settlements. Therefore, it affects the areas of tourist attraction sites near or within the protected area, which is precisely the case of the "Bromelias" attraction site that is under anthropic pressure.

Although the use of tourist carrying capacity based on the methodology developed by ^{Cifuentes (1992}, ¹⁹⁹⁹⁾ is adequate for the present study, this approach involves some limitations. In this sense, ^{Delgado (2007)} considers it an incomplete methodology abundant in subjective and qualitative aspects related to tourism that yield a "magic" figure on the maximum number of visits, arguing that it cannot be decisive when stating that a certain number of visitors will have no impact on the environment. Although this method essentially provides a quantitative result, it does not necessarily represent a final estimate of the carrying capacity and can also be used as a criterion to estimate the impact of other tourism management strategies (^{Pires, 2005}). In this regard, the correct application of the methodology should be conducted along with impact monitoring, given the need to set indicators of the limit or threshold, whose value determines the acceptability of the measured change. These indicators inform managers about the deterioration of natural values and the satisfaction level of visitors or the local community before these become unacceptable. In addition, depending on the results of monitoring activities, the carrying capacity of a given site may be modified (increased or decreased) according to the range of indicators (understood as the assessment of the impact and performance of tourism management in those places). In this way, monitoring and tourist carrying capacity set the grounds for negotiation and agreement with site managers, prioritizing the conservation of natural resources, environmental education, and the quality of visit care and satisfaction.

Despite these limitations, the Cifuentes methodology (1992, 1999) is most commonly used (62%) in tourism research, followed by the Limit of Acceptable Change, LAC (18%), due to its facility of application relative to other methods (^{Limberger and Pires, 2014}). This method proposes a maximum level of users that a tourist area is able to support with minimum environmental impact and the highest degree of user satisfaction possible, oriented to sustainable environmental conservation (^{Ruschmann, 1993}; ^{Teixeira and Oliveira, 2015}; ^{Perruolo Laneti and Camargo Roa, 2017}; ^{Han et al., 2018}). Its application is recommended as it is an effective and relevant approach for specific areas, particularly highly vulnerable ecosystems and protected areas, to avoid the deterioration or, in extreme cases, loss of natural attributes (^{Cupul-Magaña and Rodríguez-Troncoso, 2017}). in addition to guiding managers to alleviate tourist overload (^{He y Xie, 2019}; ^{Chen et al., 2021}).

In summary, our results on the carrying capacity and forest degradation in the trails of the tourist attraction sites evaluated reflect the need to promote the application of this tool in public and restricted areas. This is particularly important in protected areas that include vulnerable ecosystems to avoid the degradation or loss of their unique characteristics and promote the lowest possible impact from the flow of visits. There is an urgent need to encourage further detailed studies as part of permanent control, monitoring, and management activities, considering other tourist attraction sites not evaluated in the study area.