Investigation of Sustainable Rural Tourism Activities With Different Risk: A GIS‐MCDM Case in Isfahan, Iran

Development of rural tourism activities improves the quality of life, with many associated benefits for local people. Accordingly, this study aims to prioritize tourism areas in Isfahan province using the multiple‐criteria decision‐making (MCDM) model to provide the best ecosystem services to tourists. Nine rural tourism activities (camel‐riding, hiking, mountaineering, desert tourism, wildlife observation, cycling, canyoning, camping, and horse riding) are assessed based on integrating the priorities of each suitability criteria (fuzzy membership and Analytic Hierarchical Process method) with the different risk levels. The results show that mountaineering (86.48% moderate risk; 91% high risk) and canyoning (70% moderate risk; 97% high risk) represent the best alternatives for development of tourism. According to the results of this research, tourists can make the best decisions about individual activities by determining their priorities based on associated levels of risk.

AKBARIAN RONIZI ET AL. 10.1029/2021EA002153 2 of 13 for different types of tourism (Mokarram et al., 2021;Mokarram, Pourghasemi, & Mokarram, 2022;Ronizi et al., 2020). Based on ecotourism interests and budget allocations, it determines the best tourism activities within an area spatially (Mokarram and Hojati, 2017). In fact, it is possible to provide the majority of ecosystem services to tourists by determining their specific priorities within the region, as well as taking their economic situation into account.
Considering the importance of the topic in this study, AHP and OWA methods are used to investigate tourism priorities in southern Iran's Isfahan province. Furthermore, eco-system services for tourists are determined by the economic situation of tourists. Due to its reputation as a popular tourist destination, Isfahan province was selected as a site for investigating and determining tourism situation and preparing maps for different tourism activities. The major innovation of this study is the preparation of a tourism map with different risk levels.

Characteristics of the Region
Isfahan is a province in the central part of Iran, with area of 108,317 km 2 , and is located at a longitude of 30°42΄-34°30΄ and a latitude of 49°36΄-55°25΄ (Figure 1). There is a wide range of temperature and precipitation in the region due to the transition from the Zagros Mountains in the northwest-southeast direction to lowland-flat deserts (in the eastern part). Among the regions of Isfahan province, Khor has an average annual temperature of 19.5°C, while Semirom has an average annual temperature of 10.6°C (Isfahan Meteorological Organization, 2022). Rainfall in the province averages 160 mm per year. The Zayandehrud River flows from the heights of Zardkooh Bakhtiari in the southeast of the province to the Gavkhooni wetland near Isfahan, and is one of the largest inland rivers in Iran. In terms of economy and environment, it plays a critical role in maintaining the freshness and greenery of neighboring cities. The region has a wide variety of climate (temperature, relative humidity, wind speed), topography (various landforms, different slopes), many waterways, protected areas, and roads. Consequently, different types of tourism are possible in the region.

Materials
In order to determine suitable places for nine types of rural tourism activities (camel-riding, hiking, mountaineering, desert tourism, wildlife observation, cycling, canyoning, camping, horse riding) in the study area, data on the distance from the road, map of landform types in the area, distance from the waterway, slope, temperature, wind speed, protected areas, and relative humidity were collected (Mokarram, 2022). Topography, distance from roads and water resources are the major criterion for selected tourist activities such as mountaineering and desert climbing. In addition, research shows that factors such as distance from roads and water resources can influence tourism (Wang & Hung, 2015) However, it is the climate that has the greatest impact on the types of tourism activity (Mahdavi et al., 2015), and it is also one of the most important factors for the tourism industry as a whole.
In order to prepare the map of distance from the road and water resources (rivers), the topographic map of the region with a scale of 1: 250,000 prepared by the Geographical Organization of the country in 2018 was used. In addition, the landform and slope data were derived from the USGS digital elevation model with a spatial resolution of 90 m. Climatic data (temperature, wind speed, and relative humidity) from the Fars Meteorological Bureau (FMB) (https:// www.farsmet.ir/) for 20 stations (2000-2020) was used in this study. The protected-area map utilized the data from the Department of Environmental Protection (DEP) (https://www.doe.ir/portal/home/). All data was collated during the pre-processing (clip, transformation, and geo-referencing) before the spatial analyzes were performed using the inverse distance weighted (IDW) method (Bartier & Keller, 1996) and ArcGIS v.10.8 software (in Figure 2).
To determine the weights of the criteria in the Expertchoice software, different questions were prepared based on the opinions of experts. The questionnaire questions for each of the 9 tourism activities (camel-riding, hiking, mountaineering, desert tourism, wildlife observation, cycling, canyoning, camping, horse riding) are as follows.
1. What is the best slope for tourism? 2. Which is the most important climate parameter among temperature, humidity, and wind speed? 3. What effect does the distance from the river have on tourism? 4. What effect does increasing the distance from the road have on tourism? 5. Which type of landform is more suitable for the desired tourism? 6. How are protected areas and their impact on tourism activities? 7. Among the climatic parameters, geomorphology, protected areas, distance from the river and road, and topography, which are the most important and which are the least important parameters in tourism activity?
After averaging each answer in each questionnaire, they were converted from qualitative to quantitative in Expert Choice v.11 software.

Fuzzy Method
In 1965, Zadeh introduced fuzzy logic as an approach to scientific reasoning based on real human reasoning. Fuzzy sets are prepared using the membership function. This function's range changes from {0,1} for definite sets to closed intervals [0,1] for fuzzy sets. In the fuzzy set, the value of μA (x) indicates the degree of membership of element x in the fuzzy set. The membership of an element in the set is zero if it is completely out of the set, and one if it is completely in the set. When an element's membership is between zero and one, it indicates gradual membership (Saaty, 2008). Using the linear membership function, a fuzzy map was prepared for each criterion based on expert opinions and previous studies (Jeong et al., 2016). Each tourism activity has a different minimum and maximum value for individual input criteria. In this study, two types of membership functions are used that are defined in Equations 1 and 2, and shown in Figure 3 (Mokarram & Mirsoleimani, 2018).

AHP Method
In the 1970s, Saaty invented hierarchical analysis, one of the most popular multi-criteria decision-making techniques. This method is useful when

Earth and Space Science
AKBARIAN RONIZI ET AL.
10.1029/2021EA002153 5 of 13 a decision action faces multiple options indicators. It is possible to use quantitative or qualitative indicators. Pair-wise comparisons are the basis of this method. This is the stage where the problem is defined and factors and elements that make up the decision are arranged hierarchically (Lee, 2010;Lin & Yeh, 2012). Hierarchical analysis involves breaking the decision problem into several levels. Four levels of a decision tree are used for this purpose. The first level includes the decision's overall purpose. Decisions are made based on general criteria at the second level. At the third level, the sub-criteria are included. The last level contains the decision options, which are land suitability classes for each tourism activity. In the AHP method, the input data were compared pairwise, and one to nine values were considered for each input data (Saaty, 1986).
It is possible to calculate the relative weight (W i ) of each criteria after forming the pair-wise comparison matrix. In order to normalize the pair-wise comparison matrix, each column's values must be added together and their sum divided by the column's value for each element (Equations 3 and 4). Next, the weight vector of the criterion is calculated by averaging the elements in each row of the normalized matrix.
where p equals the number of columns, o equals the number of rows, s ab equals normalized matrix elements for options a and b, l ab represents pair-wise comparison matrix arrays, and f ab equals the weight of option a.

Ordered Weighted Averaging (OWA) Method
For each activity, the OWA method was used to determine maps with different risk levels (DRLs). Each effective criterion in the relevant tourism activity (Q) was ranked and weighed using the opinion of experts. Equation 5 is utilized to calculate the weight assigned to each criterion which is associated with a weight vector W = [w 1 ,w 2 ,… ,w n ] satisfying w j ∈ [0,1] and ∑ =1 [0 1] (Liu, 2008): where x 1 , x 2 ,…,x n are the input values to be aggregated and v j denotes the jth of the x i .
The score of each option is calculated based on Equation 6: V α is calculated based on Equation 7: where Z iq is ranked q by the second-ranking of option values at the q target level and u q is the q target by the weight ranking. α g is the criterion that connects the divine quantifiers with the ultimate goal of spatial decision-making based on the hierarchical structure.
Based on the final option chosen, an individuals' risk aversion and risk-taking are calculated using a weighted method. To describe and evaluate the OWA operator, ORness and trade-off operators are used. ORness measures how much emphasis is placed on improving or deteriorating values, or how much emphasis is placed on taking or avoiding risks by decision-makers; it is calculated by Equation 8: According to Equation 9, the higher the ORness value in the decision-making process, the higher the risk. The OWA method also uses a trade-off indicator to show the impact of each criterion on the decision (Equation 9) (Mokarram & Aminzadeh, 2010).
The stages of the research method are summarized in Figure 4.
Finally, to provide tourists with the best ecosystem services, maps of different tourism activities with varying levels of risk were prepared considering the amount of budget allocated to each activity. As a result, once the tourist has determined the level of risk they are willing to accept, and how much money they plan to spend, suitable activities can be found. The study only focuses on tourist transportation expenses but does not consider accommodation, food, or any other travel expenses.

Results of Fuzzy Approach to Determine Fuzzy Maps for Each Criterion
To determine maps with different risk levels, fuzzy membership functions were used for each of the nine types of rural tourism activities. The information from 30 questionnaires completed by experts and professors were used to determine the membership function for each input data (distance from the road, map of landform types, distance from the waterway, slope, temperature, wind speed, protected areas, and relative humidity). Table S1 shows a sample questionnaire. In this study, nine different fuzzy maps have been prepared for each activity (Table 2). For each of the criteria, the data were set between 0 and 1 based on the relative importance of the activity (Eastman, 2010).
The results of AHP method is shown in Figure 5. According to Figure 5, not all criteria are equally important for different tourism activities. In horse riding, desert tourism, canyoning, and camel riding, the climate is more important. Furthermore, the most important factor for watching wildlife is protected areas, and the distance from the river has a lesser importance for wildlife observation than for other tourist activities.

Results of OWA-AHP Approach to Prepare Tourism Maps With Different Levels of Risk
To prepare maps with DRL for each tourism activity, the OWA method was used. Figure 6 illustrates that each of the six factors examined for the tourism activities was plotted by using DRL for seven different value ranges. The levels of risk were calculated using α = 0.2, α = 0.4, α = 0.8 and α = 10, α = 8, α = 4, α = 1, α = 1 α = 20. For each of the criteria, different weights were considered based on the level of risk. As shown in Figure 6, as the level of risk-taking increases, more areas are suitable for tourism activities and vice versa. Horse riding: Based on Figure S1a in Supporting Information S1, the south and center have more potential for this tourism activity. Moreover, the higher the risk levels, the more areas are suitable for this type of activity.
Camping: Figure S1b in Supporting Information S1 describes the maps of different risk levels for tourism activities related to camping. The results show that the center and south of the region are more important for this type of tourism activity.
Canyoning activities: According to Figure S1c in Supporting Information S1, the study area is better for canyoning activities than other activities. Most areas except limited parts of the north, south and center have high values (over 0.8) for this activity. The priority of tourists in the study area is canyoning.
Cycling: Figure S2a in Supporting Information S1 illustrates with an increase in risk levels, more areas in the north of the region can be considered suitable for this activity. As shown in Figure S2a in Supporting Information S1, Humidity between 20 and 60 Note. A, Maximum; B, Minimum; C, Medium. one of the suitable areas for cycling tourism is Shahrereza, in the southern study area. This area has been identified at different levels of risk-taking suitable for this activity, which indicates the high accuracy of the model.
Wildlife: According to Figure S2b in Supporting Information S1, most parts of the area are suitable for wildlife observation except for protected areas. Figure S2b in Supporting Information S1 shows that Abbasabad in Nain in the eastern part of the study area and Kolah Ghazi in the central part of the study area have interesting wildlife for tourists, which are also identified as suitable areas in this model. Figure S2c in Supporting Information S1, the northern, central and southeastern parts of the study area are suitable for this activity due to desert conditions, which increases with higher levels of risk. Studies show that Maranjab desert, which is one of the best desert climbing areas, is very suitable for this type of activity, which confirms the accuracy of the model. Maranjab desert is located in the northernmost point of Isfahan province (34° 7′ and 51° 48′) and 50 km from Aran and Bidgol cities.

Desert tourism: As shown in
Mountaineering: As shown in Figure S3a in Supporting Information S1, with the increase in risk levels, wider areas of higher elevations will be suitable for this activity. Sofeh Mountain, located in the south of the study area, is one of the attractive areas for mountaineering in this area, which has been identified by the model suitable for this activity. It is worth mentioning that the topographic criterion is of special importance in this type of activity (Sahani, 2019).
Hiking: The map with different levels of risk for hiking is illustrated in Figure S3b in Supporting Information S1. It shows the central areas are of greater importance for this type of activity due to access to urban areas with more facilities.
Camel riding: Finally, Figure S3c in Supporting Information S1 shows the maps with different levels of risk for camel riding. It shows that protected areas and elevations at all levels are not suitable for this activity. With increasing levels of risk, large parts of the center of the region are suitable for this activity. In this study, land suitability for each risk level of the tourism activities is divided into 5 classes: very low (0-0.2), low (0.2-0.4), medium (0.4-0.6), high (0.6-0.8), very high (0.8-1). In Figure S4 in Supporting Information S1, it can be seen that the high classes are more suitable for tourism activities. As a final step, individual tourist activity maps were merged to create the final tourist map. The purpose of preparing these maps is to determine the best type of tourism at different levels of risk. Thus, tourists can choose the best type of tourism from these 9 types of activities based on the desired levels of risk. Using this map, tourists can make the right choice. In the first stage, the tourist selects a map with risk levels. Based on Table 3, it selects the type of tourism activity it wants in the region. Therefore, the tourist can choose several tourism activities in the desired area. For example, as shown in Figure 7, each level of risk was classified into 15 classes, each of which represents the best type of tourism activity for the region. Therefore, after choosing the type of tourism activity by the tourist, the amount of cost can be determined by considering the distance from the airport, costs such as transportation, accommodation,  etc. Based on Figure 8, a tourist seeking a low-cost mountaineering excursion in the region can select the shortest route from the best possible areas in one of the risk levels. It is important to emphasize that only travel costs in tourist areas are considered in this research, and not accommodation or food costs.
Lastly, the costs for rural tourism activities in Isfahan province were determined by assessing the risk to tourists. At lower levels of risk, the tourist that can pay more will choose more restricted but better quality locations. On the other hand, tourists are more likely to choose lower-quality areas with higher levels of risk if he or she can pay less. Based on the assumption that Shahid Beheshti International Airport is the primary point of entry for tourists, the best place for tourism can be determined after selecting a type of activity, risk level, and distance from the airport. For example, if mountaineering is the chosen tourism activity, tourists can select one of five locations based on the levels of risk and costs presented in Table 4 ( Figure 8). Figure 8 and Table 4 shows the locations of sample points for mountaineering tourism activities with DRLs.

Discussion
Recently, multi-criteria decision-making methods, such as fuzzy models, AHP, and OWA, along have been successfully applied to tourism use planning (Büyüközkan & Çifçi, 2011). Fuzzy-AHP models can be applied to a variety of real-world scenarios (Zabihi et al., 2020). High-quality land suitability maps can be prepared using multiple-criteria decision-making (MCDM) methods. For more accurate results, MCE and GIS can be used simultaneously. Ashouri and Faryadi (2010), Mahdavi et al. (2015), for example, used the MCE method to prepare land suitability maps, demonstrating the model's accuracy. Furthermore, Fuzzy, WLC, and ANP methods were used to determine suitable ecotourism spots in Alborz province in Iran (Aliani et al., 2017) and to develop the tourism industry in Zabol, Iran (Zabihi et al., 2020). A combination of multi-criteria and GIS evaluation methods were used to assess rural tourism potential. In addition to integrating layers and evaluating land use potential, WLC uses a fuzzy approach to standardize criteria. This method offers a high degree of flexibility in management and is easy to implement (Hajehforooshnia et al., 2011). Furthermore, this method has many advantages, including the ability to simultaneously evaluate land potential alongside other selected environmental, social, and economic criteria. The OWA approach assigns weights to criteria values at the individual pixel level. By applying appropriate weights to spatial decision making, a variety of maps can be generated. In addition, this method is flexible enough to meet the needs and priorities of decision makers. Modeling allows a better understand ing of land use, and can make a significant contribution to the solutions for many multi-criteria decision problems. According to results of Mokarram, Pham, and Khooban (2022), the OWA model is capable of modeling complex real-world decision problems. This model is not only designed to find one option, but also to emphasize other aspects of it. As a convenient feature of the OWA method, equivalence and compensability between criteria can be controlled or adjusted, facilitating quick interpretation and assessment. AHP, AHP-OWA, and Fuzzy AHP-OWA can be used in many spatial decisions.  Using fuzzy methods and OWA criteria, Tavakkolinia et al. (2018) assessed spatial relationships in the Rudbar region of Qasran, Iran, for the development of ecotourism. According to the results of the different scenarios, a high proportion of the region is suitable for ecotourism development. Using the OWA method, managers and planners can examine different scenarios in an area, and the results can be used to guide tourists to the most appropriate tourism activity. Tourists can be recommended different places based on their capacity for risk-taking (quality of tourism at various levels of risk-taking). Additionally, a suitable place can be selected based on their financial situation (the amount of costs they are willing to take on), which is a step forward in tourism studies.

Conclusion
Tourism has been viewed as an important tool in growing and developing economic activities in various regions and a way to boost gross national income in all countries. Tourism is a very important part of sustainable development in rural areas, and the growth, implementation and development of this industry can contribute to that goal. Even though rural tourism represents a small share of this industry, it can contribute significantly to economic growth in rural areas and play a vital role in preserving a region's natural and historical heritage. In order to determine the appropriate places for various rural tourism activities, this study analyzed nine types of activities with varying levels of risk. Multi-Criterion Assessment (MCE) and GIS techniques have been used to assess the potential and risk levels of natural ecosystems for the purposes of developing rural tourism activities in the province of Isfahan in Iran. OWA method allows considering multiple criteria simultaneously in determining land suitability among the different methods. To create maps with varying levels of risk, OWA was used to overlap map layers. For activities with low risk levels, the OWA-AHP method indicates that the areas located in the south and center of the region are optimal for horse riding, wildlife, desert tourism, those in the south are most suitable for camping, cycling, and mountaineering tourism activities, and the center of the region is best for hiking and camel riding. One of the most important innovations of this study is that tourists can use any of the risk maps based on their intended spending on tourism activities. Therefore, at high risk levels, less attractive areas are also considered suitable for tourism, and vice versa.

Data Availability Statement
All the datasets used in the study, including topography, climate, river, road, geomorphology, protection layers, are available in the in-text data citation reference (Mokarram, 2022) and available at https://figshare.com/articles/ figure/Untitled_Item/20453457. Figures were made with ArcMap software version 10.8. For more data details, please contact us via email. One additional table containing an example of a questionnaire form is included in Supporting Information S1 file and available at https://doi.org/10.6084/m9.figshare.22155245.