Assessing the potential of snow leopard tourism-related products and services in the Annapurna Conservation Area, Nepal

ABSTRACT Conservation Enterprise is increasingly promoted to support the conservation of species and landscapes through incentives, such as ecotourism, including in the Annapurna Conservation Area (ACA), Nepal. Yet the elusive behaviour of snow leopards here limits opportunities for conservation enterprise, particularly those linked to conventional ecotourism forms. Furthermore, the potential to explicitly link local snow leopard-friendly livestock production systems with the tourist market in the area, via eco-certified livestock products, has not been investigated. Therefore, this paper aims to explore the interest, from supply and demand perspectives, in introducing snow leopard ecotourism services and eco-certified products into the ACA tourist market. Questionnaire data were gathered from 406 tourists and 403 local residents. Our results, of interest to managers and researchers alike, show that there is potential to generate funds and support for both snow leopard conservation and community development, and add to the literature on utilising enterprise initiatives as conservation tools.


Introduction
Conservation Enterprise (CE) has become an increasingly popular strategy to support the conservation of iconic species and landscapes, and often complements more conventional approaches such as protected areas (PAs) and legislation (Buschke, 2015).CE draws upon various concepts, such as community-based natural resources management, ethical consumption, sustainable livelihoods and alternative livelihoods (Anand et al., 2012;Chambers & Conway, 1992;Elliott & Sumba, 2011;Scoones, 1998;Starr, 2009;Wright et al., 2016;Zollo, 2021) to address local conservation problems through innovative and cost-effective methods "in the spirit of entrepreneurship" (Buschke, 2015).CE is also influenced by the idea of market-based conservation (Roth & Dressler, 2012).Consequently, CE frequently features tourism activities (Balmford et al., 2009(Balmford et al., , 2015)), but can also involve a range of other activities including wildlife-friendly eco-labels (Treves & Jones, 2010), sustainable use approaches (Hutton & Leader-Williams, 2003) and trophy hunting (Naughton-Treves et al., 2003), among others.
CE-related tourism suffers from a plethora of often overlapping terms and definitions, including ecotourism (Weaver, 2001), nature-based tourism (Balmford et al., 2009), wildlife tourism (Reynolds & Braithwaite, 2001), sustainable tourism (Weaver, 2014) and environmentally-friendly tourism (Dolnicar et al., 2008).Based on its ubiquity and duration as a concept since the 1980s (Weaver & Lawton, 2007), we utilise the term ecotourism in this paper, it being defined as "responsible travel to natural areas that conserves the environment, sustains the well-being of the local people, and involves interpretation and education" (TIES, 2015).
A significant rationale for ecotourism is that it will provide direct benefits from conserving species and spaces, such as PAs, that otherwise can incur significant costs for local people (Kc et al., 2021;Snyman, 2012;Sekhar, 2003;Walpole & Leader-Williams, 2002).A 2015 paper estimated that terrestrial PAs received approximately 8 billion visits per year, with more than 80% of these in North America and Europe, and that these visits generated US $600 billion per annum in direct in-country expenditure (Balmford et al., 2015).Unfortunately, opportunities for direct ecotourism employment can be limited, while human-wildlife conflicts (HWC), often associated with PAs, disproportionately affect farmers and those directly dependent on natural resources for their livelihoods (Hemson et al., 2009;Munanura et al., 2016;Sandbrook & Adams, 2012).As with trophy hunting (Loveridge et al., 2010), it has been questioned how equitably or evenly ecotourism benefits or impacts are, or can be, distributed.A meta-analysis of 29 ecotourism studies, for example, found that only 25% included economic contributions to host communities as a defining feature of ecotourism (Dolnicar et al., 2008).On the contrary, homestay models, for example, have been noted to maximise benefits to local communities and households (Acharya & Halpenny, 2013).More recently, the COVID-19 pandemic has shown the risk of global stochastic events in drastically depressing tourism visits worldwide (Zenker & Kock, 2020;Zheng et al., 2021), and the concomitant risk to local communities and PAs of being overly, or even wholly, dependent on revenues from this one industry alone (Fletcher et al., 2020;Liu & Wall, 2006;Spenceley et al., 2021).
Certification, via wildlife-friendly eco-labels, attempts to unite producers and consumers of goods and services in a shared attempt to conserve wildlife (Koldewey et al., 2009).However, they vary in terms of cost and effectiveness.At one extreme of the eco-label spectrum, "Protective" labels are the most credible in the eyes of conservationists, but also the most costly to implement and monitor (Treves & Jones, 2010).At the other end of the spectrum, "Supportive" eco-labels donate funds to conservation projects and organisations but experience very limited scrutiny.In the middle, Persuasive' eco-labels theoretically certify operational procedures in order to benefit wildlife.This range of sources and standards for eco-labels can also cause confusion amongst consumers (Bhaskaran et al., 2006;Teisl et al., 1999), whose rational for such ethical consumption can range from following social norms (Starr, 2009) to non-rational motivations such as emotion (Zollo, 2021).
In the context of the Himalayas, where mountaineering tourism, if not ecotourism, is often well-established, a lack of connectivity between the tourist market and local production systems has been identified (Nepal, 2002).Here too, this has resulted in a lack of equity in revenue distribution, as well as market penetration by outside elites (Daconto & Sherpa, 2010;Nepal, 2000).Nevertheless, the Annapurna Conservation Area (ACA) of Nepal has been held-up by some authors as a successful and innovative model of community-based ecotourism and enterprise development (Aase et al., 2013;Bajracharya et al., 2006;Baral et al., 2007).
It has been increasingly recognised that many large mammal species, particularly carnivores, cannot thrive in PAs alone (Woodroffe & Ginsberg, 1998).This has confirmed the need for mechanisms that allow coexistence in areas subject to more anthropogenic disturbances (di Minin et al., 2016).This includes the home range of snow leopards Panthera uncia (Johansson et al., 2016), a threatened wild felid that inhabits 12 South and Central Asian nations, including the ACA in Nepal (McCarthy et al., 2016).A number of conservation enterprise interventions, including ecotourism and handicraft production, have been attempted that use funds generated to incentivise snow leopard conservation and offset costs of mostly livestock-related impacts associated with coexistence with the species (GSLEP, 2020;Hussain, 2000;Mishra et al., 2003;Namgail et al., 2016).In the case of a snow leopard conservation handicraft project in Mongolia, a conservation contract was signed between producer communities and the product purchasers, with penalty clauses for infringement of mutually-agreed outcomes, such as poaching cessation (Mishra et al., 2003).Therefore, we define snow leopard-friendly eco-certified livestock products here as livestock products produced with minimum harm and maximum benefit to the conservation of snow leopards and their habitat.
The importance of such explicit linkages between CE programmes and concrete conservation outcomes has been stressed for snow leopards elsewhere (GSLEP, 2020; Wilman & Wilman, 2016), and for carnivores in general (Dickman et al., 2011).Conservation surcharges on tourists visiting a community-owned conservancy in Kenya, for example, funded a successful compensation scheme for livestock losses to wild lions Panthera leo (Bauer et al., 2017).Recent research from Hemis National Park in India has shown improvements in perceptions of snow leopards amongst communities who have benefited from snow leopardrelated tourism (Maheshwari & Sathyakumar, 2019;Vannelli et al., 2019).
Critics have argued that CE's should be limited to wilderness areas where conventional development institutions are lacking (Redford, 2008).Equally, if they are too complex to implement, CE's will remain ambiguous in the conservation incentives that they generate, often failing to be aligned with localised socio-ecological conditions and conservation objectives (Ferraro, 2001).Others have noted that they are often limited by Westerncentric definitions and processes (Cater, 2006).In response to some of these criticisms, it has been argued that distinguishing between development interventions and conservation payments may not be practical on the ground (Mishra et al., 2003).Indeed, snow leopard conservation often cannot be separated from sustainable rural development amidst the communities who coexist with the species (Jackson et al., 2010).
Although many studies have concurred that CE interventions can be appropriate options for achieving conservation, and sometimes development, outcomes, few studies have assessed local attitudes to these proposals, especially in emerging economies where micro-, small-and medium-sized enterprises predominate (Ghimire, 2011;Jaafar & Maideen, 2012).In fact, research had often tended to focus more on the demand-side, typically from wealthier Western consumers, rather than the supply-side, including for ecotourism (Kontogeorgopoulos & Chulikavit, 2010) and eco-certified products (Jonell et al., 2013;Teisl et al., 1999).To date, Silori's (2007) research stands out among published studies of alternative livelihoods, in having asked respondents for their opinion on a range of diversification options prior to any implementation, as well as reasons for declining the proposed options.Assessing attitudes to proposed conservation interventions is also a common failing of many conservation projects worldwide (Webber et al., 2007).
In ACA, an enhanced livestock insurance scheme, predator-proof corrals, livelihood diversification and conservation education have all been proposed as combined mitigation methods for the potential negative impacts of snow leopards on household livelihoods (Aryal et al., 2014;Hanson et al., 2019b;Hanson, 2022), as have snow leopard ecotourism services (Hanson et al., 2019a;Schutgens et al., 2019).As with general tourism in ACA, the tension between conservation and development has been noted and is ongoing (Holden, 2010).Since 2018, the Snow Leopard Conservancy (SLC) and local and international partners have worked on a Darwin Initiative-funded project to start snow leopard conservation tourism trails in the NarPhu Valley of ACA (B.Peniston, pers comm).Significant supply-and demand-side market research attached to this project has demonstrated considerable local and tourist interest in this concept, but the COVID-19 pandemic has delayed implementation.In addition, other geographical regions of ACA, as well as the concept of eco-certified livestock products, were not considered by this project.
Our review of the literature has illustrated that there is a growing body of evidence on CE, though markedly more so for ecotourism services and the demand-side of their consumption than for eco-certified products,as well as for the supply side of both products and services.In addition, both types of intervention often suffer from a lack of consultation with and engagement of stakeholders prior to implementation, especially in the Global South.Large carnivore conservation, with its need to shift the burden of coexistence from poor local communities towards the global community stands to benefit from thoughtful, reasoned engagement with these approaches.The aim of this study was therefore to offer unique, empirically-grounded perspectives on the potential of snow leopard-tourism products and services in ACA from both locals and tourists simultaneously.It contributes to the existing literature on conservation enterprise, ethical consumption, ecotourism and large carnivore conservation, illustrating, for both researchers and planners alike, that effective tourism-related CE solutions can be found at the intersection of supply and demand.Furthermore, it underscores the significance of the meaningful consultation of a range of stakeholders prior to considering, planning or developing ecotourism and CE projects.In summary, our study, of interest to managers and researchers alike, show that there is potential to generate funds and support for both snow leopard conservation and community development, and adds to the literature on utilising enterprise initiatives as conservation tools.
Accordingly, the study was structured to complement existing data and address information gaps for both the supply and demand dimensions of two potential tourism-related CE interventions in ACA: snow leopard ecotourism services, including a trail and a documentary screening, and snow leopard-friendly eco-certified livestock products.
To this end, the study sought to answer the following questions, which functioned as broad research objectives for inquiry rather than as specific hypotheses: what are tourists' attitudes to the provision of ecotourism services and which factors best explain these?; what are local residents' attitudes to the provision of ecotourism services and which factors best explain these?; what are tourists' attitudes to the provision of eco-certified products and which factors best explain these?; and, what are local residents' attitudes to the provision of eco-certified products and which factors best explain these?

Study site
Ranging in elevation from 1,000 m to over 8,000 m ASL, the ACA covers over 7,629 km² of Himalayan territory in the mid-western region of Nepal (Figure 1).The ACA is home to a wide variety of biodiversity including nearly 500 bird species, 100 mammal species and over 1,000 plant species (Bhuju et al., 2007).Blue sheep Pseudois nayaur are relatively common in the alpine zone and provide the primary prey species for the snow leopard, which is present at varying densities (Ale et al., 2014).In partnership with local communities' Conservation Area Management Committees (CAMCs), the National Trust for Nature Conservation (NTNC) manages the ACA through a co-management approach.
The ACA is home to more than 85,000 people (Government of Nepal, 2012), who largely derive their income from agro-pastoralist activities and tourism (Bajracharya et al., 2006;Baral et al., 2007).ACA is one of the prime destinations in the world for trekking and received an all-time high of 181,000 international tourists in 2019 (ACAP, 2021), prior to COVID-19.Tourism revenues contributed substantially to the region's economy and a large percentage of the ACA's annual operating budget (Baral & Dhungana, 2014).The COVID-19 pandemic saw visitor numbers to ACA fall sharply to 18,796 in 2020, a decrease of 89.62%.Recovery to previous volumes is likely to take several years and financial deficits will be necessary to kick-start operations in the years following the pandemic, as after previous recessions (Baral & Dhungana, 2014;ACAP, 2021).

Tourist questionnaire
Between March and May 2014 we administered a written questionnaire in simple English to 406 international visitors in ACA, walking the popular tourist circuit in a clockwise direction (Figure 1).We employed cluster sampling of hotels to ensure a random composition of visitors were included in the sample.Furthermore, we surveyed a total of 33 hotels after securing permission from the hotel owner.We gave all visitors a short overview of the research project and then requested their participation.Consenting participants took 23 min on average (SD = 9.5; range = 12-55 min; n = 58) to complete the self-administered questionnaire.A team member remained nearby to provide clarification if necessary.Completed questionnaires were photographed and catalogued prior to data entry on an Excel spreadsheet.In total we approached 572 international visitors and 406 agreed to participate yielding a 71.0%response rate.Of those that provided a reason for not participating, more than 50.0%stated a lack of interest, followed by language difficulties (14.9%) and fatigue (10.6%).For further information on the variables used and their measurement, see Supplementary Material 1.

Local questionnaire
These data were also collected between March and May 2014 with the surveys administered orally in Nepali by two Nepali research assistants.Systematic sampling was used due to the informal nature of many of the settlements and the absence of a reliable sampling frame.Census data provided the number of households in each Village Development Committee (VDC; redesignated as Gaunpalikas since 2017) in the Manang and Lower Mustang districts that harboured snow leopards.The region of Upper Mustang could not be accessed for logistical and budgetary reasons.
Approximately one quarter of households was considered a realistic and achievable goal, giving a target sample of 440.Every fourth house in each settlement was sampled, using a similar approach to the sampling of every third house in a study in the Middle Hills of Nepal (Paudel & Thapa, 2001).To prevent any overlap in sampling by the two research assistants, each had responsibility for data collection in one half of each settlement only.In addition, random back-checking of 10% of surveys carried out by research assistants was implemented to promote and ensure inter-observer reliability (Newing et al., 2011;White et al., 2005).The process did not uncover any cases of error in data collection.Overall, we approached 467 local individuals, of whom 445 agreed to participate, with a response rate of 95.3%.See Supplementary Material 1 for further information on the variables used and their measurement.

Tourist questionnaire
Three ordinal response variables were included in the questionnaire, which required conversion into binary variables.This conversion resulted in the inevitable loss of some information, because some levels had few observations, which in turn interfered with the minimum cases required per variable, in turn compelling us to build logistic regression models.With logistic regressions, the ratio of the number of events in lower categories to the number of predictors was around 10:1, suggesting sufficient events per variable for the analysis.All the positive events (yes) were coded as "1" and negative events (no) as "0" for the three derived binary response variables: (1) Expectation of seeing snow leopards influencing visit; (2) Interest in participating in snow leopard ecotourism services; and (3) Willingness to pay extra for eco-certified yak products.We built three logistic regression models, taking these variables as the response variables and 10 socio-demographic and attitudinal variables as predictors (see Supplementary Material 1 for definitions).We selected predictors based on earlier research and bivariate exploratory analyses, and made a priori decision to enter them collectively in the model (Bagley et al., 2001) because our aim was to explain the relationships of these predictors with the response variables in a comparative framework.
Interactions should be pre-specified, so they were not considered in the models because of any prior knowledge to suggest potential interactions.Multicollinearity among the predictors did not appear to be an issue in the models as indicated by the mean variance inflation factor (VIF) of 1.16 and tolerance values greater than 0.70.The final models were selected based on the goodness-of-fit measures and post-estimation regression diagnostics.The models with predictors had a better fit with the data than the null models (χ 2 10 >26, P < .05;Table 1), and the models correctly classified more than 72.0%cases.Logistic regression assumes a linear relationship between the log odds of the response variable and the predictor variables.This assumption of linearity was tested following the Box-Tidwell test to determine the need for transformations in the variables and the Linktest to explore a model specification error.The residual plots revealed no influential and leverage points.These results suggest that the final models were adequate to make valid inferences.

Local questionnaire
The multi-variable scales were treated as continuous data for the purposes of analysis.This follows the trend in recent attitudinal analyses of large felids (Bhatia et al., 2016;Carter et al., 2014;Cavalcanti et al., 2010;Suryawanshi et al., 2014).To test for reliability, independent t-tests were used to test for inter-observer consistency between the two research assistants (Field, 2013).Multiple regression models were used to determine the variables that best explained attitudes to ecotourism services and eco-certified products scales, focusing on explanation rather than prediction (Mac Nally, 2000).Prior to constructing the models, the data were checked to ascertain whether they met the necessary assumptions: linearity, reliability, homoscedasticity and normality (Osborne & Waters, 2002).Furthermore, multicollinearity between variables was not found to exceed the recommended limit of 0.70 (Green et al., 2013).Hierarchical entry based on theoretical suitability was used and model selection to determine the most suitable, parsimonious model used the r² change results to determine goodness-of-fit (Mac Nally, 2002).Bootstrapping was used for all models as P-P plots to test for normality in multiple regression models indicated some evidence of non-normality (Field, 2013).

Research ethics
The Ethics Review Group at the University of Cambridge's Department of Geography provided ethical approval for the study.The National Trust for Nature Conservation, Nepal, and the Department of National Parks and Wildlife Conservation, Nepal, provided fieldwork approval and research permits.At the time of study's design and implementation, none of the authors were involved in the Snow Leopard Conservancy's ecotourism and conservation enterprise project in the NarPhu valley of the ACA.Given increasing scrutiny of the reflexivity of conservation researchers and practitioners (for example, Boyce et al., 2022), it is worth noting that the research team exhibited both disciplinary and cultural diversity, enhancing their capabilities to critically reflect on the broad range of issues pertaining to this study, as well as their approach to, and involvement in, it.

Results
3.1 Tourist interest in snow leopard ecotourism 3.1.1Tourist expectations of seeing snow leopards influenced their choice of visiting ACA Respondents were asked: "To what extent did the expectation of seeing snow leopards in the wild influence your decision to visit Annapurna?"Their responses were recorded with the following five ordinal options: (i) not at all (77.6%), (ii); to a limited extent (9.8%); (iii) to a moderate extent (8.5%); (iv) to a large extent (2.0%) and (v) to a very great extent (2.0%).
Because more than 66% of responses belonged to one category, a dichotomous variable was created to assess whether the expectation of seeing snow leopards influenced tourists' decision to visit the area.Responses 2-5 suggest a wish to see snow leopards, so these responses were merged together and coded as 1 (22.4%) to compare with negative responses coded as 0 (77.6%).

Tourist interest in snow leopard ecotourism services
Respondents were asked, "In which of the following activities are you likely to take part, if available?"We gave tourists four options from which to choose: (i) Snow Leopard Trail-a local guide escorts you to the high pastures in search of snow leopards (42.7%); (ii) Paid Documentary-watching a documentary film on snow leopards (13.4%); (iii) Both the activities (11.1%); and (iv) None of these activities (32.8%).After carefully examining the distribution of responses, a binary variable with two categories was constructed: 0 = participation in none of the activities (32.8%) and 1 = participation in one or both of the activities (67.2%).

Tourist interest in eco-certified products
Figures 2 and 3 detail tourists' actual and intentional purchase or consumption of yak products when visiting the ACA, including meat, cheese, butter, wool clothes and handicrafts.
Respondents were asked whether they would be willing to pay extra for eco-certified yak products consumed in the ACA, with the following responses: Yes (27.4%),Maybe (49.0%), and No (23.6%).We then created a binary variable recoding definite yes responses as "1" (27.4%) and compared them with no or maybe responses coded as "0" (72.6%) in the model.

Tourist snow leopard identification
When asked to identify the snow leopard from the accompanying photos of six species of large carnivores, 65.8% respondents correctly identified snow leopards and their response is coded as positive (1).However, 34.2% respondents could not identify correctly and their response was coded as negative (0).

Tourist support for snow leopard conservation
Respondents were asked to indicate how important they believe it is to implement the Snow Leopard Conservation Action Plan to protect snow leopards in the ACA.Their responses were initially recorded in the following five categories: (i) Not at all important (0.5%); (ii) A little important (3.3%), (iii) Somewhat important (20.7%),(iv) Very important (54.3%); and (v) Extremely important (21.2%).Because there were only two responses in the first category, it was merged with the next category to yield a 4-point ordinal scale.

Factors explaining tourist interest in snow leopard ecotourism
Various factors best explained tourist attitudes to aspects of snow leopard ecotourism in our logistic regression models (Table 1), the explanatory power of which correctly predicted more than 72.0%cases.Firstly, North Americans, Australians and those who supported snow leopard conservation were significantly more likely to have had their visit to ACA influenced by the expectation of seeing snow leopards.Secondly, respondents who were older and who supported snow leopard conservation were significantly more likely to be interested in participating in snow leopard ecotourism services.Thirdly, North Americans, Australians, Europeans, those who supported snow leopard conservation and those with more years of education were significantly more willing to pay extra for eco-certified yak products.

Local interest in provision of snow leopard ecotourism services
Two questions were asked.For attitudes to raising money for local snow leopard conservation activities (n = 402), 24.9% were very positive, 20.1% were positive, 45.8% were neutral, 8.2% were negative and 1.0% were very negative.For attitudes to raising awareness of snow leopards amongst tourists (n = 401), 35.4% were very positive, 20.9% were positive, 41.4% were neutral, 2.0% were negative and 0.2% were very negative.

Local interest in providing eco-certified products
Three questions were asked.Firstly, attitudes to higher prices for local yak products paid to herders (n = 403), with 8.2% of respondents very positive, 33.5% positive, 31.0%neutral, 8.9% negative and 18.4% very negative.Secondly, attitudes to higher prices for other local livestock products paid to herders; (n = 402).Here, 11.4% were very positive, 34.3% were positive, 43.0% were neutral, 8.2% were negative and 3.0% were very negative.Thirdly, with attitudes to higher prices paid by tourists for local livestock products (n = 402), 18.4% were very positive, 15.2% were positive, 46.3% were neutral, 18.7% were negative and 1.5% were very negative.

Local interest in snow leopard ecotourism services provision
A summated rating scale was developed from the two questions, utilising 5-point Likert-type scales, ranging from 1 (very positive) to 5 (very negative) (Spector, 1992).The data in the .50p = .001Note.* 0 = no; 1 = yes.Note.See Supplementary Information for additional and diagnostic information.
scale were then reverse-scored.Its Cronbach's alpha score was .73,above the commonly accepted threshold of .70 for internal reliability (Nunnally, 1978).Table 2 shows a linear model explaining local individuals' attitudes to snow leopard ecotourism services, with moderate explanatory power (R 2 = 0.35).Local interest in service provision was best explained by attitudes to snow leopard conservation, years of education and livelihood scores.

Local interest in providing eco-certified products
Following the methodology for the services scale construction, The Cronbach's alpha score for the scale was .68.In practice, scores of >.60 are reported as significant in the medical (Bosma et al., 1997;McKinley et al., 1997) and conservation literatures (Nepal & Spiteri, 2011).Therefore, the scale's internal consistency was considered acceptable, and Table 3 shows the linear model that explains local attitudes to eco-certified products.However, the model had low explanatory power (R 2 = 0.04) and the number of livestock killed by snow leopards best explained local interest in product provision.

Tourist sample
The mean age of respondents was 32.2 ± 11.1 years, with a range from 18 to 73 years.Males comprised 47.8% and females 42.2% of the sample.For education, 22.5% were educated to high school level or below, 12.8% to associate degree level, 29.3% to bachelor's degree level, 27.8% to master's degree level and 7.6% to doctoral degree level.More than half (55.4%) had annual household income of less than $40,000 from all sources and before taxes, 28.1% had in between $40-80,000 and 16.4% had more than $80,000.A minority, 40.7%, were active in the labour force, while the remainder, 59.3%, were inactive.In terms of nationality, 22.2% were North American, 52.0% were European, 10.1% were Australasian and 15.7% were Asian.A total of 17.1% of respondents held membership of any environmental, conservation or wildlife organization.

Local sample
Years of education varied between 0 and 18, with a mean of 3.76 ± 4.58 years.Males equalled 49.2% of those surveyed, and females 50.8%.The majority of respondents were local (92.8%)rather than Nepali non-locals (7.2%).Five categories were used to declare religion, but the numbers in all except Buddhist (91.2%) meant that the other four-none, Bon, Hindu and other-were aggregated into a single result (8.8%).Likewise, the five-point Likert-type scale used to measure religiosity was collapsed into two categories due to a lack of data for categories three to five.Very religious respondents comprised 64.0% of those surveyed, while 36.0%were less than very religious.Finally, 60.0% of respondents were able to positively identify a snow leopard, compared to 40.0% who failed to do so.The mean attitudes to snow leopards score was 3.42 ± 1.04, while the mean attitudes to snow leopard conservation score was 3.98 ± 0.42.The mean household Sustainable Livelihoods Index score was 0.55 ± 0.10.With livestock owned per household, the mean was 21.35 ± 36.56, with a range between 0 and 312.For self-reported losses of livestock to snow leopards in the previous 12 months, mean losses per household were declared as 0.82 ± 2.46, with a minimum of 0 and a maximum of 21 animals.

Tourists and snow leopard ecotourism services
Our study showed that 67.2% of visitors were interested in one or both of: (i) a "snow leopard trail", in which a local guide escorts visitors into high pastures looking for snow leopards or signs of their existence, and, (ii) a "paid documentary", showcasing information about snow leopards, their prey and their habitat in the local area.Taken together with more recent unpublished market research undertaken by the Snow Leopard Conservancy and its partners among potential international tourists (B.Peniston, pers comm), there appears to be considerable demand for snow leopard ecotourism services in ACA, notably some sort of "snow leopard trail".This is in accordance with wider trends towards snow leopard ecotourism activities in North-West India especially (Vannelli et al., 2019) and across snow leopard range states more generally (GSLEP, 2020).
The inferential analysis showed that older respondents and those who supported snow leopard conservation were more likely to be interested in taking part in snow leopardrelated tourism services.This is consistent with other recent work on snow leopard conservation tourism in Nepal (Schutgens et al., 2019;Hanson et al., 2019a).Given these findings, tourism managers active in ACA could explicitly market snow leopard ecotourism services to older tourists, who may be more likely to visit in an organised group-making the provision of services to them more efficient and effective.Group dynamics may also help to reinforce such ethical consumption activities as a social norm (Starr, 2009).Meanwhile conservation managers could ensure snow leopard conservation messages are effectively and explicitly communicated to visiting tourists, further facilitating their interest in ecotourism services.

Locals and snow leopard ecotourism services
From the supply side of the ACA tourism equation, locals expressed interest in snow leopard ecotourism services.Forty-five percent of respondents were positive or very positive about money being raised for local snow leopard conservation activities from a proposed scheme, while 56.3% were positive or very positive about it raising awareness of snow leopard among tourists.This is consistent with other recent studies from Hemis National Park in India, where snow leopard conservation tourism was popular locally (Vannelli et al., 2019;Maheshwari & Sathyakumar, 2019).
Similar work in Africa has also linked employment in ecotourism with more positive attitudes to conservation (Snyman, 2012).
The linear model showed that local interest in service provision was best explained by attitudes to snow leopard conservation.As with tourists in ACA and in other areas (Schutgens et al., 2019;Hanson et al., 2019a), attitudes to snow leopard conservation can have a strong influence on conservation outcomes (Hanson et al., 2019b).In addition, more years of education positively influenced attitudes to the provision of ecotourism services, a finding common from similar studies in this PA (Hanson et al., 2019a).Furthermore, livelihood scores were inversely correlated with support for snow leopard ecotourism services, with those scoring lower on the Sustainable Livelihoods Index significantly more supportive of any proposed scheme.This may be because the appeal of ecotourism schemes to residents of PAs is strongest among those most likely to bear the costs of coexisting with wildlife, a trend noted at other sites (Snyman, 2012;Sekhar, 2003;Walpole & Leader-Williams, 2002).
Conservation and tourism managers could continue to develop the potential for snow leopard conservation tourism in ACA, capitalising on considerable local support.In particular, managers could proactively address the concerns of those less supportive of snow leopard conservation, those with less education, and wealthier individuals with less exposure to losses from snow leopards and other predators.

Tourists and snow leopard-friendly eco-certified products
Consumption of local livestock products by tourists is considerable in ACA (Figures 2 and 3).Furthermore, 76.4% of respondents stated they would, or would consider, purchasing ecocertified local yak products even if they were more expensive than non-certified products.This suggests that there is potential to supply a tourist market in ACA with snow leopardfriendly products and handicrafts, as has been trialled in Mongolia and other snow leopard range states (Mishra et al., 2003).However, the challenges inherent in eco-certification schemes, from increased costs and verification effort with more "Protective" labels through to limited trust and verification with more "Supportive" approaches (Treves & Jones, 2010), mean that any such scheme would be challenging to design and implement to ensure positive conservation and development outcomes for all stakeholders.
Our inferential analysis of the origin of snow leopard tourists showed that: North Americans, Australians and Europeans; support for snow leopard conservation; and more years of education all influenced willingness to pay extra for eco-certified yak products.These findings were similar to other work on tourist support, and willingness to pay, for snow leopard conservation in ACA (Schutgens et al., 2019;Hanson et al., 2019a).Therefore, conservation and tourism managers in ACA and other similar sites could consider the importance of Western tourists, their levels of education and their support for snow leopard conservation in marketing snow leopard-friendly eco-certified livestock products.Social norms (Starr, 2009) and non-rational factors, such as emotion (Zollo, 2021), are likely to be important in influencing consumer's choices here.

Locals and snow leopard-friendly eco-certified products
Snow leopard-friendly eco-certified livestock products attracted less local support than did ecotourism services.Overall, 41.7%, 45.7% and 33.6% of respondents, respectively, were positive or very positive about paying higher prices for local yak products to the herder; paying higher prices for other local livestock products paid to herders; and by tourists paying higher prices for local livestock products.Elsewhere, creating snow leopard-friendly eco-certified products in particular (Mishra et al., 2003), and ecocertified products in general (Treves & Jones, 2010;Koldewey et al., 2009), that demonstrate tangible and verifiable links between the product's pledge and the product's conservation impact have proved challenging.As elsewhere, there would be a high likelihood of such products in ACA being subject to questions of consumer knowledge and trust on one hand (Dolnicar et al., 2008), and supplier willingness to participate in a degree of verification effort on the other (Treves & Jones, 2010), especially given limited literacy and educational standards among many livestock keepers in ACA.
Our second inferential analysis suggested that the number of livestock reported as killed by snow leopards best explained local interest in the provision of eco-certified products.However, the linear model had limited explanatory power and the results should be interpreted and extrapolated with caution.Nevertheless, given that livestock owners negatively affected by predation are a key constituent group for snow leopard (Hanson et al., 2019b;Suryawanshi et al., 2014) and other large carnivore (Dickman et al., 2011) conservation, there is still merit in conservation and tourism managers in ACA working explicitly to meet the needs and concerns of this group, in order to assess the potential for a snow leopard-friendly eco-certification scheme.

Conclusion
There is potential to generate funds and support for both snow leopard conservation and community development through snow leopard CE in the ACA, underscoring its potential as a conservation tool more generally.However, there is likely to be greater potential for snow leopard ecotourism services than for snow leopard-friendly eco-certified livestock products.This may be due to a clearer conceptual link between support for snow leopard conservation and snow leopard ecotourism services, as well as there being fewer logistical constraints with establishing and maintaining snow leopard tourist trails and documentaries, compared with eco-certified products.There are likely to be new and continued challenges and tensions reconciling cost/benefit sharing and conservation/development with any such proposals in ACA, as elsewhere (Acharya & Halpenny, 2013;Holden, 2010;Kc et al., 2021).
In theoretical terms, the findings contribute to the diverse bodies of knowledge in ecotourism, livelihoods, ethical consumption and others that underpin CE (Buschke, 2015).The study also contributes to all three tenets of the TIES (2015) ecotourism definitioninterpretation, conservation and communities-while supporting Treves and Jones (2010) assertion, in relation to eco-labels and certification, that a balance needs to be struck between consumers and producers to ensure the success of any schemes.Lastly, the study's simultaneous focus on local and tourist perspectives on these issues in relation to snow leopard conservation in ACA contributes to correcting a historical imbalance often tilted against the meaningful consultation of local communities in both conservation and development interventions (Nepal, 2002;Silori, 2007).
Although this paper provides valuable preliminary data and analysis, it should be verified by further research and consultation to inform ongoing and future planning and implementation in ACA.Here and elsewhere, conservation would benefit from more and better consultation with local communities prior to any project design and implementation.In addition, future research could address the dearth of data on the linkages between ecotourism and eco-label schemes and their intended and unintended conservation and development outcomes.Although the data are somewhat dated, the pause on tourism caused by the COVID-19 pandemic makes for a timely opportunity to reconsider tourism-related CE planning and development in ACA and beyond.

Figure 1 .
Figure 1.Annapurna Conservation Area showing sampling locations and time-frame in 2014.

Figure 2 .
Figure 2. How frequently visitors consumed or purchased local yak products during their visits in the ACA.

Figure 3 .
Figure 3. Respondents' intention to purchase or consume local yak products while in the ACA.

Table 1 .
Logistic regression models explaining tourist attitudes to aspects of snow leopard conservation tourism.

Table 2 .
Linear model explaining local individuals' attitudes to aspects of snow leopard conservation tourism.

Table 3 .
Linear model explaining local individuals' attitudes to snow leopard-friendly eco-certified livestock products.