650 words 3 hrs

Brandtetal2019ecotourismandforestloss.pdf

Home >English homework help >650 words 3 hrs

Contributed Paper

Effects of ecotourism on forest loss in the Himalayan biodiversity hotspot based on counterfactual analyses

Jodi S. Brandt ,1 ∗ Volker Radeloff,2 Teri Allendorf,2 Van Butsic,3 and Anand Roopsind1 1Human-Environment Systems Center, Boise State University, 1910 University Drive, Boise, ID 83725, U.S.A. 2Forest and Wildlife Ecology, University of Wisconsin-Madison, 1630 Linden Drive, Madison, WI 53706, U.S.A. 3Department of Environmental Science, Policy and Management, University of California, Berkeley, Mulford Hall, 130 Hilgard Way, Berkeley, CA 94720, U.S.A.

Abstract: Ecotourism is developing rapidly in biodiversity hotspots worldwide, but there is limited and mixed empirical evidence that ecotourism achieves positive biodiversity outcomes. We assessed whether ecotourism influenced forest loss rates and trajectories from 2000 to 2017 in Himalayan temperate forests. We compared forest loss in 15 ecotourism hubs with nonecotourism areas in 4 Himalayan countries. We used matching statistics to control for local-level determinants of forest loss, for example, population density, market access, and topography. None of the ecotourism hubs was free of forest loss, and we found limited evidence that forest-loss trajectories in ecotourism hubs were different from those in nonecotourism areas. In Nepal and Bhutan, differences in forest loss rates between ecotourism hubs and matched nonecotourism areas did not differ significantly, and the magnitude of the estimated effect was small. In India, where overall forest loss rates were the lowest of any country in our analysis, forest loss rates were higher in ecotourism hubs than in matched nonecotourism areas. In contrast, in China, where overall forest loss rates were highest, forest loss rates were lower in ecotourism hubs than where there was no ecotourism. Our results suggest that the success of ecotourism as a forest conservation strategy, as it is currently practiced in the Himalaya, is context dependent. In a region with high deforestation pressures, ecotourism may be a relatively environmentally friendly form of economic development relative to other development strategies. However, ecotourism may stimulate forest loss in regions where deforestation rates are low.

Keywords: community-based forestry, environmental policy, Mahalanobis matching, quasi-experimental, sus- tainable development

Efectos del Ecoturismo sobre la Pérdida de Bosques en el Punto Caliente de Biodiversidad en el Himalaya con base en Análisis Contrafactuales

Resumen: El ecoturismo está desarrollándose rápidamente en los puntos calientes de biodiversidad en todo el mundo, pero existe evidencia emṕırica mixta y limitada de los resultados positivos que se logran con el ecoturismo. Valoramos si el ecoturismo influyó sobre las tasas de pérdida forestal y sus trayectorias entre el 2000 y el 2017 en los bosques templados del Himalaya. Comparamos la pérdida forestal en quince focos ecotuŕısticos con la pérdida forestal en las áreas sin ecoturismo de cuatro páıses del Himalaya. Utilizamos estad́ıstica correspondiente para controlar las determinantes a nivel local de la pérdida del bosque, por ejemplo, la densidad poblacional, el acceso al mercado y la topograf́ıa. Ninguno de los focos ecotuŕısticos estaba libre de pérdida forestal, además de que encontramos evidencia limitada de que las trayectorias de la pérdida forestal en los focos ecotuŕısticos eran diferentes a las trayectorias en las áreas sin ecoturismo. En Nepal y en Bután, las diferencias en la pérdida forestal entre los focos ecotuŕısticos y las áreas sin ecoturismo correspondidas no difirieron significativamente y la magnitud del efecto estimado fue menor. En la India, donde las tasas generales de pérdida forestal fueron las más bajas de cualquier páıs en nuestro análisis, las tasas de pérdida forestal fueron más altas en los focos ecotuŕısticos

∗email [email protected] Article impact statement: Whether ecotourism protects forests is context dependent, and an important factor is the deforestation pressure overall on the landscape. Paper submitted December 13, 2018; revised manuscript accepted March 13, 2019.

https://orcid.org/0000-0002-1954-5997

Brandt et al. 1319

que en las áreas sin ecoturismo correspondidas. Como contraste, en China, donde las tasas generales de pérdida forestal fueron más altas, las tasas de pérdida forestal fueron más bajas en los focos ecotuŕısticos que en donde no existe el ecoturismo. Nuestros resultados sugieren que el éxito del ecoturismo como estrategia de conservación del bosque, a como se práctica actualmente en el Himalaya, depende del contexto. En una región con presiones altas de deforestación, el ecoturismo puede ser una forma de desarrollo económico relativamente amigable con el ambiente comparado con otras estrategias de desarrollo. Sin embargo, el ecoturismo puede estimular la pérdida forestal en regiones en las que las tasas de deforestación son bajas.

Palabras Clave: correspondencia Mahalanobis, cuasiexperimental, desarrollo sustentable, poĺıtica ambiental, silvicultura basada en la comunidad

��: ��, ��, �� 2000–2017 �� , � ��, �� , ��, �� , ��, �� , ��, �� , ��, �� , ��, �� , �� , �� , �� : ��; ��: ��

��: ��, ��, ��, ��, ��

Introduction

Ecotourism is proliferating in biodiversity hotspots, and its proponents claim it can achieve conservation and economic development goals. Ecotourism has become a major driver of economic growth and socioeconomic transformation in many areas. The amount spent on eco- tourism is estimated to be 10 times more than that spent by official aid agencies and the UN Global Environment Facility on conservation projects (Kirkby et al. 2011; Wal- dron et al. 2017). Ecotourism accounts for as much as 40% of gross domestic product (GDP) in some countries and is growing 10% per year in other countries (WTTC 2014). Despite this major investment, there is limited empirical evidence that ecotourism achieves biodiversity conservation goals in the long term and at the landscape scale.

It is difficult to ascertain whether ecotourism actually achieves biodiversity goals. In developing regions, eco- tourism and its cumulative effects on biodiversity are unclear. Ecotourism may generate the same or more income in an area than the consumption of natural re- sources (Kirkby et al. 2010, 2011). Thus, ecotourism can provide an economic incentive to protect ecosys- tems and species tourists visit. For example, governments may establish protected areas or enforce wildlife protec- tion to ensure revenue from international tourists (Buck- ley 2011). Similarly, community ecotourism projects in unprotected landscapes may dedicate a portion of ecotourism proceeds into conservation efforts to pro- tect their natural assets (Nagendra et al. 2005; Buck-

ley 2009; Wyman & Stein 2010). A major justification for ecotourism investments by developed countries is the assertion that ecotourism may lead to biodiversity conservation because it provides economic rewards for doing so.

In contrast, ecotourism may lead to biodiversity loss because it can require or encourages economic develop- ment, which often entails strong, negative environmen- tal outcomes (Mather et al. 1999). Ecotourism usually requires improved transportation networks (e.g., roads and airports), which can result in intensive natural re- source exploitation, such as logging and poaching, be- cause of increased accessibility to the area (Laurance et al. 2014; MoCTCA 2015b; Shui & Xu 2016). Increased local wealth can change residents’ consumption pat- terns, adding pressure on local forest resources (Liu et al. 2001; Brandt et al. 2012). Tourism also stimulates popu- lation growth, in the form of seasonal tourists and eco- nomic immigrants, which can raise demand for forest re- sources (Hall & Lew 2009). Tourists essentially represent a form of population growth, and they typically consume more resources per capita than local residents (Buckley 2011). Finally, tourism inherently leads to an integration of local and regional markets, another factor strongly associated with increased resource extraction (Hall & Lew 2009; Wang & Buckley 2010; Lambin & Meyfroidt 2011).

Incentives for protection and economic development thus may either result in positive or negative biodiversity

Conservation Biology Volume 33, No. 6, 2019

1320 Himalayan Ecotourism

Figure 1. Ecotourism hubs across 4 countries in the Himalayan temperate forest zone (as defined by Olson et al. 2001) and with forest-cover data from Hansen et al. (2013).

outcomes. However, even if the net result of ecotourism is negative, it may still be beneficial as long as eco- tourism leads to less biodiversity loss than would have occurred if an alternative economic development strat- egy had been implemented instead (i.e., ecotourism may not completely stop biodiversity loss, but it may be better than alternatives). Developing nations typically rely on extraction-based land uses for economic development, including the production of raw goods (e.g., mining tim- ber) or the conversion of natural ecosystems to more economically productive uses (e.g., agriculture). There- fore, even if ecotourism stimulates economic develop- ment that leads to environmental degradation, it may lead to less biodiversity loss than more extractive models, such as palm oil production. Quantifying whether or not that is the case requires the application of a counterfac- tual approach to estimate rates of negative biodiversity outcomes had there not been ecotourism (Andam et al. 2008; Butsic et al. 2017).

We sought to measure environmental outcomes of ecotourism by comparing ecotourism hubs with other areas in which other development strategies had been implemented. Specifically, we measured the association between ecotourism and forest loss in the Himalayan temperate forest zone (Fig. 1) with a counterfactual ap- proach. We identified 15 ecotourism hubs (i.e., areas where ecotourism is the primary strategy for economic development) across 4 countries with diverse tourism strategies: India, Nepal, Bhutan, and China. Our objec- tives were to characterize the type of ecotourism strat- egy implemented in each country; identify whether eco- tourism hubs have rates and trajectories of forest loss distinct from nonecotourism areas; and quantify differ- ences in forest loss rates between ecotourism hubs and nonecotourism areas.

Methods

Study Area

The Himalayan temperate forest zone extends 3000 km from southern Afghanistan to southwest China (Olson et al. 2001). It contains 2 of Earth’s biodiversity hotspots (Myers et al. 2000), an extraordinary array of ecological niches in a relatively small area, and globally the highest fractions of endemic and threatened species in the world (Grenyer et al. 2006). Himalayan temperate forests have been used for thousands of years to support subsistence- based livelihoods. Forests are the primary source of fuel for cooking, heating, and construction; are intensively used for livestock grazing, hunting, food gathering, and traditional medicines; and provide raw materials for eco- nomic development. These same forests contain highly threatened, endemic biodiversity and provide essential ecosytem services, including climate and water-cycle reg- ulation. Since the 1980s, demand for timber and fuel- wood increased, resulting in forest loss and degradation (Pandit et al. 2014). Even though forest protection is a primary conservation target across Himalayan countries, forest loss has continued (Brandt et al. 2017).

The Himalayan region provides opportunity for a nat- ural experiment to investigate ecotourism impacts be- cause it contains countries in very different stages of economic and tourism development. Tourism has prolif- erated across the Himalaya as a way to balance economic development and forest conservation (Pandit et al. 2014).

We analyzed 4 Himalayan countries with active eco- tourism industries: India, Nepal, Bhutan, and China. Nepal and Bhutan are relatively small and located primar- ily in the Himalaya and its foothills. India and China are mostly outside the Himalayas and have regional adminis- trative units with distinct policies and contexts. Because of their large size, we focused on single administrative units located primarily in the Himalaya: Himachal Pradesh State, India, and Yunnan Province, China.

Forest Change Data

We defined forest loss as stand-replacing disturbance or a change from forest to nonforest. Forest cover and change data for 2000–2017 were derived from a publicly available data set of global forest dynamics (Hansen et al. 2013). The forest-cover data set contains canopy cover of each 30-m pixel in the baseline year of 2000. Each pixel is classified from 0 to 100 (0, no canopy; 100, 100% canopy cover). We considered a pixel with >50% canopy cover as forested.

Site Identification and Characterization of Ecotourism Strategies

To identify ecotourism sites and characterize ecotourism strategies and contexts, including the number, origin, and

Conservation Biology Volume 33, No. 6, 2019

Brandt et al. 1321

Figure 2. Ecotourism hubs with 15-, 25-, 35-, 45-, and 55-km circular zones surrounding each hub. The zone with the boldest outline is the 35-km zone, which was used for the counterfactual analysis. All areas within the 35-km circular zone are ecotourism hubs, and all areas outside this zone were nonecotourism areas.

purpose of tourists and economic wealth, we reviewed the peer-reviewed and gray literature on ecotourism for each country (Supporting Information). We compiled statistics when possible, but comparable and consistent statistics across sites were typically not available; thus, we relied on province-, state-, or country-level data. We obtained information about the overarching forest gover- nance strategy and forest change in each country (Brandt et al. 2017). We also used information from studies pub- lished in the literature about tourism impacts on forests (see Brandt & Buckley 2018).

Tourism areas, unlike protected areas or administra- tive units, do not have delineated boundaries. However, tourism tends to be concentrated, and tourists typically visit an ecotourism hub (i.e., a population center where tourists concentrate for accommodation, food, guides, and other amenities). To identify our ecotourism hubs, we compiled a list of the most popular general tourism hubs based on official tourism statistics and other litera- ture reviewed for each country; searched Google (search terms such as ecotourism in Bhutan and ecotourism in Yunnan) to identify hubs that advertised ecotourism; narrowed the list by asking regional experts to identify ecotourism hubs they considered the most popular; and overlaid our map of potential hubs on an ecoregion map (Olson et al. 2001) to identify sites that included forest in the Himalayan temperate zone (Fig. 1). Our selection of ecotourism sites was not designed to be a represen- tative sample of all ecotourism hubs; rather, it was an attempt to select the most important ones. We identi- fied 15 ecotourism hubs across 4 administrative units: Himachal Pradesh (n = 4), Nepal (n = 5), Bhutan (n = 3), and Yunnan (n = 3).

To determine the appropriate spatial boundary, we demarcated circular zones with 15-, 25-, 35-, 45-, and 55-km radii surrounding each ecotourism hub (Fig. 2). We summarized forest loss rates in these zones and their

respective nonecotourism areas (Fig. 3). For example, the 15-km ecotourism zone represented all forests 0–15 km from the hub, and the corresponding nonecotourism areas included all forests within the same country that were not within 0–15 km of any ecotourism hub in that country. When calculating deforestation at the hub level, forests included in >1 hub boundary were attributed to both hubs. When calculating deforestation at the country level, forests in areas of overlapping boundaries were included only once in the analysis to avoid overestimat- ing deforestation rates at the aggregate scale. In all 4 countries, forest-loss rates in ecotourism and nontourism zones came close to convergence by the 35-km bound- ary and fully converged by the 55-km boundary. Thus, we used the 35-km zone for subsequent analyses. We designated areas within the 35-km boundaries as eco- tourism hubs and areas beyond the 35-km boundaries as nonecotourism areas.

Trajectories of Forest Loss Area

We plotted trajectories of the area of forest loss for all forests in each ecotourism hub (35 km) and for all forests in the nonecotourism areas of each country. We calculated trend lines showing 2-year moving averages to smooth out errors in the annual forest loss measurements due to cloud compositing during remote-sensing analyses (Hansen et al. 2013). We also fitted a linear trend line for the entire trajectory to visualize whether each hub increased, decreased, or was stable from 2000 to 2017.

Counterfactual Analysis of Forest Loss Rate

An increasingly common counterfactual approach to de- termine the impact of conservation policies is quasi- experimental counterfactual matching analysis (Andam et al. 2008). Matching has been used, for example, to assess the effectiveness of certification policies (Miteva et al. 2015), national forest management regimes (Brandt et al. 2017), protected areas (e.g., Nolte et al. 2013), community forests (e.g., Brandt et al. 2015), and logging concessions (e.g., Brandt et al. 2016). To our knowledge, matching has not been used in the context of ecotourism.

Ecotourism hubs are typically located in remote places that have relatively few people, are less accessible, and have retained more forests than the country as a whole. The goal of our matching analysis was to find areas with similar population density, accessibility, and forest cover that only differed in whether they were associated with a major ecotourism hub or not. To do so, we matched treat- ment units (e.g., forested cells influenced by ecotourism) with control units (e.g., forest cells not influenced by eco- tourism). With matched samples, it is possible to predict what outcomes would have been observed in forests with ecotourism had they not been subjected to ecotourism (Abadie & Imbens 2006). In essence, we asked: What

Conservation Biology Volume 33, No. 6, 2019

1322 Himalayan Ecotourism

Figure 3. Forest-loss rates (2000–2017) within 15, 25, 35, 45, and 55 km of ecotourism hubs and nonecotourism areas for each country.

would be the rate of forest loss in ecotourism zones of country A if ecotourism had not been adopted and if that area had been developed like other nonecotourism areas? We also performed the opposite comparison by asking: What would be the rate of forest loss in nontourism zones of country A if ecotourism had been adopted?

We performed pair-wise comparisons of the forest-loss rate in ecotourism hubs and nonecotourism areas in each country by applying Mahalanobis matching with replace- ment and bias adjustment (Sekhon 2011). We aggregated annual forest cover and forest change data into 1-km cells to achieve a sample size that was computationally feasible and consistent with similar analyses (Ferraro et al. 2013; Nolte et al. 2013; Brandt et al. 2017). We calculated an adjusted forest-loss rate, which is the total area of forest loss divided by the forested area in that cell in 2000. Cells that did not have any forest were excluded from the analyses. We matched treatment and control units based on 7 covariates (i.e., factors that influence forest loss): distance to market, population density, slope, elevation, precipitation, temperature, and percent forest cover in 2000. For each pairwise comparison, we randomly sam- pled 20% of the treatment parcels and matched them with control parcels. To determine the validity of the matches, we calculated balance statistics, which indicate

the extent to which the pool of potential controls con- tains units that are sufficiently comparable to treatment units. We dropped treatment parcels for which no com- parable control parcel could be found within 0.5 SD of each covariate. To compute the reverse estimate, we switched control and treatment group (i.e., assigned nonecotourism areas as the treatment and found matches from the ecotourism hubs). We repeated this procedure for pairwise comparisons in each country for a total of 8 different pairwise comparisons. Matched treatment and control units were always from the same country. See Supporting Information for the full results of the match- ing analysis and balance statistics.

Results

Ecotourism Strategies

The 4 countries we analyzed differed considerably in terms of the types of ecotourism they implemented (Sup- porting Information). India is one of the most populous and rapidly developing countries in the world. Himachal Pradesh, the focus of our analysis, had the second-lowest GDP/capita of any unit in our analysis (US$2200), and the

Conservation Biology Volume 33, No. 6, 2019

Brandt et al. 1323

lowest overall forest-loss rate from 2000–2017 (0.40%). Himachal Pradesh has a unique forest governance sys- tem of traditional, community-based forestry that is of- ficially recognized by the federal government (Brandt et al. 2017). Tourism in Himachal Pradesh started in the colonial era in the form of seasonal vacation centers (i.e., Hill Stations), created by the British during the 19th century (Ahluwalia & Little 1998). Since the 1970s, the state government has implemented policies to encourage the development of both corporate and leisure tourism (Pandey & Wells 1997), and the number of tourist visi- tors grew from 8 million in 2006 to 15 million in 2011 (KPMG 2012), the vast majority of which were domes- tic (ACNielsen 2012). Nature-based, adventure, and reli- gious tourism dominated (Singh 2002; Donovan 2013). We found 2 empirical articles about ecotourism and de- forestation in the Indian Himalaya, both of which link ecotourism to deforestation and forest degradation be- cause increased demand for timber and fuelwood was sourced from local forests (Singh et al. 2009; Mahapatra et al. 2012).

Nepal is among the poorest countries in the world with primarily community-based sustainable forest man- agement (Brandt et al. 2017). Nepal had the lowest GDP/capita (US$835) and the second-highest rate of tem- perate forest loss (1.0%) of our study units. Nepal has been a popular international ecotourism destination since the 1970s, and tourism has been important for economic development (Schroeder & Sproule-Jones 2012). Nepal’s tourism policy is designed to maximize the number of tourists and offers relatively inexpensive visas and few restrictions on travel and the length of time tourists may stay in the country (Schroeder & Sproule-Jones 2012). In 2014, Nepal hosted over 790,000 tourists, a 58% increase from approximately 500,000 in 2000 (MoCTCA 2015a, 2015b), and has a national goal of 2 million tourists per year by 2020 (Nepal & Karst 2017). The majority of tourists in Nepal are international and come to visit national parks and to trek (MoCTCA 2015b). Similar to Himachal Pradesh, we found 2 empirical articles for the Nepal Himalaya, both of which report more local deforestation and forest degradation due to ecotourism because of increased demand for timber and fuelwood (Stevens 2003; Garrard et al. 2016).

Bhutan is a small Buddhist kingdom, known for its gross domestic happiness (GDH) policy, where environmen- tal protection and economic growth are equally priori- tized, a sustainable development approach that is unique among developing nations (Brooks, 2010, 2013). Bhutan had the second highest GDP/capita of any unit in our study (US$3110) and the second lowest forest-loss rate (0.9%). Bhutan’s national forest policy emphasizes forest conservation (Brandt et al. 2017), and tourism in Bhutan is a relatively recent phenomenon compared with India and Nepal. Tourism was introduced as a means of attaining foreign currency to help achieve economic development

and autonomy from donor aid (Nepal & Karst 2017), but, in contrast to Nepal, Bhutan has pursued a controlled approach. In 1974, Bhutan implemented a policy known as ‘‘high value, low volume’’ (Schroeder & Sproule-Jones 2012). Tourist visas are expensive and short, and travel permissions are tightly controlled, which limits both tourism numbers and the activities tourists can engage in. Although there is great potential for adventure-based tourism, such as climbing and trekking, it is limited due to the tight control (Gurung & Seeland 2008), and the primary activity of most tourists is “cultural sight-seeing” on designated tours to specific sites (Bhutan 2014; TCB 2015). In 2014, Bhutan hosted 133,480 tourists, about one-fifth as many as Nepal (Bhutan 2014). We found no empirical studies about ecotourism and forests in Bhutan.

China has had the fastest growing economy in the world in recent decades. Strong economic development policies for western China have stimulated high rates of economic growth in Yunnan (Xu et al. 2006). Yun- nan’s forest governance policy emphasizes for-profit use of forests (Brandt et al. 2017). Yunnan had the highest GDP/capita (US$5117) and the highest forest-loss rate (2.9%) of all of our study units. The most common eco- nomic development strategies include extractive-based activities, including cash crops, mining, and hydropower, except for specific areas where ecotourism has been im- plemented (Li & Han 2000; Donaldson 2007; Wang & Buckley 2010). The Himalayan region of Yunnan is des- ignated as the premiere ecotourism destination in China and aggressively marketed to the growing middle class in eastern China (Nyaupane et al. 2006). Ecotourism has grown exponentially since 1990 and the vast majority of tourists are domestic (Jenkins 2009; Brandt et al. 2012; HKTDC 2017). For example, tourist visitors in Diqing Prefecture grew from 40,000 tourists in 1995 to 5.3 mil- lion visitors in 2009. We found 2 empirical studies from the Chinese Himalaya about the impacts of tourism on forests, both of which reported that ecotourism led to accelerated deforestation due to rapid economic devel- opment and population growth (Liu et al. 2001; Brandt et al. 2012).

Rates and Trajectories of Forest Loss in Ecotourism Hubs and Nonecotourism Areas

In simple comparisons, we did not find that forest loss dif- fered clearly between ecotourism hubs and nontourism areas, but there were differences among countries (Fig. 4). India had the lowest forest-loss rates, ranging from 0.4% in Shimla and nonecotourism areas to 1.1% in Manali. The highest hub-level forest-loss rates occurred in China, ranging from 2.0% in Tacheng to 4.3% in Lijiang; nonecotourism areas had an intermediate forest-loss rate of 2.9%. Bhutan and Nepal’s forest-loss rates were be- tween those in India and China, and forest-loss rates in

Conservation Biology Volume 33, No. 6, 2019

1324 Himalayan Ecotourism

Figure 4. Forest-loss rates in each ecotourism hub and nonecotourism areas, and the average of all ecotourism hubs in each country.

nonecotourism areas were within the range of individual ecotourism hubs in the respective countries.

In terms of annual forest-loss trajectories, nontourism areas and ecotourism hubs in India all had a decreasing forest-loss trend from 2000 to 2017 (Fig. 5). Dalhousie had the highest forest loss of all sites and years (1.5 km2

in 2002) but had no forest loss in the most recent years (2014–2017). Other Indian hubs and the nontourism ar- eas also showed high interannual variability; there was no specific temporal pattern other than low or 0 forest loss after 2013. In Nepal, nontourism areas showed a stable to slightly increasing trend of forest loss, whereas 4 of 5 ecotourism hubs had a decreasing forest-loss tra- jectory. Bhutan displayed a unique temporal pattern that was consistent across the country; a spike in forest loss occurred in all 3 hubs and the nontourism areas in 2010 and rates increased steadily from 2012 to 2017. China’s nonecotourism areas, similar to Nepal’s, showed a slight increase in forest loss over the entire period, and the ecotourism hubs varied greatly: 1 (Lijiang) increased, 1 (Shangrila) decreased, and 1 (Tacheng) was stable.

Differences in Forest-Loss Rates in Ecotourism Hubs and Nonecotourism Areas Based on Counterfactuals

According to our counterfactual analysis, forest-loss rates in Nepal and Bhutan in ecotourism hubs were not signif- icantly different (p < 0.05 threshold) relative to noneco- tourism areas (Fig. 6 & Supporting Information). The difference in loss rates between ecotourism hubs and nonecotourism areas was +0.08% in Nepal and +0.12% in Bhutan. In India, ecotourism hubs had higher forest-loss

rates than matched nonecotourism areas; effect size was small but significant (+0.40%, p < 0.001). Similarly, when we matched cells in nonecotourism areas with those in ecotourism hubs, nonecotourism areas had less forest loss than ecotourism hubs; effect size was −0.74% (p < 0.001). China was the only country where ecotourism hubs had lower forest-loss rates than matched noneco- tourism areas (effect size of −1.73%, p < 0.001). The inverse comparison was also significant; cells in noneco- tourism areas matched with those in ecotourism hubs had +0.77% more forest loss (p < 0.001).

Discussion

We found little evidence that ecotourism reduces rates of forest loss, but also little evidence that ecotourism leads to higher forest-loss rates due to more rapid devel- opment. At first glance, our results seemed to suggest that ecotourism spurs forest loss because forest-loss rates themselves were in most cases higher in ecotourism hubs than in nonecotourism areas (i.e., Fig. 3). However, sim- ple comparisons of forest-loss rates are only valid when ensuring that sites with similar deforestation pressure are compared. Indeed, when we controlled for deforestation pressure, we found that the effects of ecotourism varied among countries. We were surprised by our results be- cause previous empirical studies in both India (Singh et al. 2009; Mahapatra et al. 2012) and Nepal (Stevens 2003; Garrard et al. 2016) found more deforestation in eco- tourism hubs. However, none of these prior studies used counterfactual approaches to control for deforestation

Conservation Biology Volume 33, No. 6, 2019

Brandt et al. 1325

Figure 5. Annual forest loss, 2-year moving averages, and linear trend-lines from 2000–2017 in each ecotourism hub and nonecotourism areas in each country.

pressure, and their findings may reflect that ecotourism hubs are often located in places that inherently have higher deforestation pressure. Similarly, we expected that Bhutan’s strategy of a tightly controlled ecotourism industry would result in better forest conservation out- comes compared with Nepal, which hosts many more tourists and exerts less control on where and how they travel (Schroeder & Sproule-Jones 2012). However, we found no evidence that ecotourism in Bhutan reduced deforestation pressures or differed in its impact from ecotourism in Nepal.

Yunnan was the only study unit where ecotourism hubs had lower forest-loss rates than nontourism areas, and even when comparing areas with similar deforesta- tion pressure, the effect size was considerable (1.73%).

This result surprised us because prior case studies re- ported empirical evidence of accelerated deforestation after ecotourism started in the Chinese Himalaya (Liu et al. 2001; Brandt et al. 2012). However, our matching results suggest that in the context of rapid development in China as a whole, and in Yunnan in particular, eco- tourism led to less forest loss than areas where tourism was not prominent. It is important to note that while forest-loss rates in Yunnan’s ecotourism zones were less than that in China’s nonecotourism areas, they were still 2 to 3 times higher than forest-loss rates in eco- tourism hubs of Bhutan, Nepal, and India. Overall, forest- loss rates in Yunnan were very high due to China’s national-level forest management policy that encourages for-profit use of forests instead of the sustainable use or

Conservation Biology Volume 33, No. 6, 2019

1326 Himalayan Ecotourism

Figure 6. Matched comparisons of forest-loss rates between ecotourism hubs and nonecotourism areas in each country (black, ecotourism hubs with higher forest-loss rates than nonecotourism areas; gray, ecotourism hubs with lower forest-loss rates than nonecotourism areas; no shading, no significant difference; ∗estimates significant at p < 0.001). See Supporting Information for full results.

conservation-oriented policies implemented in the other 3 countries (Brandt et al. 2017). China’s background forest-loss rates were 3–5 times higher than in the other 3 countries, suggesting intense forest loss pressures due to other economic development strategies in southwest China, including timber extraction from forests, mining, and hydropower development (Buckley 2010). When compared with these other development strategies, eco- tourism may be a relatively environmentally friendly form of economic development in China.

Our empirical analysis highlights an urgent need for more rigorous, empirical, and multiscale analysis of the effects of ecotourism in biodiversity hotspots. To our knowledge, this analysis is the only multinational and the only counterfactual analysis that evaluates ecotourism outcomes. Because economic development is also a goal of ecotourism, there is an urgent need to analyze eco- nomic benefits concurrently with forest change, for ex- ample, by calculating a forest change per unit of eco- nomic growth among different economic development strategies. It is likely that, similar to other environmental governance interventions, the effects of ecotourism vary in space and time (Ostrom et al. 2007). Thus, case stud- ies in diverse social-ecological contexts, and at different spatial and temporal scales, followed by rigorous meta- analyses will be essential to build a stronger knowledge base.

Our findings have important implications for policy makers because they highlight that forces of economic

development, even when stimulated by a nonextractive development strategy like ecotourism, can lead to en- vironmental degradation. Specifically, our results suggest the rates of forest loss resulting from ecotourism are com- parable to those resulting from other, more conventional, development strategies. The exception to this rule ap- pears to be areas where deforestation pressures are very high. In these high deforestation areas, ecotourism may slow forest loss. More research, at finer spatial scales, and in other biodiversity hotspots, is necessary to build the evidence base about under what conditions ecotourism generates sustainable forest conservation outcomes.

Acknowledgments

J.S.B. acknowledges support from the National Science Foundation (IIA-1301792). V.R. acknowledges support from the NASA Land Cover and Land Use Change Pro- gram.

Supporting Information

Covariates used in the analysis (Appendix S1), Full match- ing results and balance statistics (Appendix S2), articles reviewed for each country (Appendix S3) and summary of ecotourism types and context in each country (Ap- pendix S4) are available online. The authors are solely

Conservation Biology Volume 33, No. 6, 2019

Brandt et al. 1327

responsible for the content and functionality of these materials. Queries (other than the absence of material) should be directed to the corresponding author.

Literature Cited

Abadie A, Imbens GW. 2006. Large sample properties of matching esti- mators for average treatment effects. Econometrica 74:235–267.

ACNielsen. 2012. Final report for Himachal Pradesh—tourism survey. Ministry of Tourism, Government of India, Himachal Pradesh, India.

Ahluwalia I, Little I, editors. 1998. India’s economic reforms and de- velopment, essays for Manmohan Singh. Oxford University Press, Delhi, India.

Andam KS, Ferraro PJ, Pfaff A, Sanchez-Azofeifa GA, Robalino JA. 2008. Measuring the effectiveness of protected area networks in reducing deforestation. Proceedings of the National Academy of Sciences of the United States of America 105:16089–16094.

Bhutan. 2014. Bhutan tourism monitor annual report. Tourism Council of Bhutan, Thimpu, Bhutan.

Brandt JS, Allendorf T, Radeloff V, Brooks J. 2017. Effects of national- level forest management regimes on non-protected forests of the Himalaya. Conservation Biology 31:1271–1282.

Brandt JS, Buckley RC. 2018. A global systematic review of empirical evidence of ecotourism impacts on forests in biodiversity hotspots. Current Opinion in Environmental Sustainability 32:112–118.

Brandt JS, Butsic V, Schwab B, Kuemmerle T, Radeloff VC. 2015. The relative effectiveness of protected areas, a logging ban, and sacred areas for old-growth forest protection in southwest China. Biological Conservation 181:1–8.

Brandt JS, Kuemmerle T, Li H, Ren G, Zhu J, Radeloff VC. 2012. Us- ing Landsat imagery to map forest change in southwest China in response to the national logging ban and ecotourism development. Remote Sensing of Environment 121:358–369.

Brandt JS, Nolte C, Agrawal A. 2016. Deforestation and timber produc- tion in Congo after implementation of sustainable forest manage- ment policy. Land Use Policy 52:15–22.

Brooks JS. 2010. Economic and social dimensions of environmental behavior: balancing conservation and development in Bhutan. Con- servation Biology 24:1499–1509.

Brooks JS. 2013. Avoiding the limits to growth: gross national happiness in Bhutan as a model for sustainable development. Sustainability 5:3640–3664.

Buckley R. 2009. Ecotourism principles and practices. CAB Interna- tional, Oxfordshire, United Kingdom.

Buckley R. 2010. Conservation tourism. CAB International, Oxfordshire, United Kingdom.

Buckley R. 2011. Tourism and environment. Pages 397–416 in Gadgil A, Liverman DM, editors. Annual review of environment and resources. Volume 36. Annual Reviews, Palo Alto, California.

Butsic V, Lewis DJ, Radeloff VC, Baumann M, Kuemmerle T. 2017. Quasi-experimental methods enable stronger inferences from ob- servational data in ecology. Basic and Applied Ecology 19:1–10.

Donaldson JA. 2007. Tourism, development and poverty reduction in Guizhou and Yunnan. The China Quarterly 190:333–351.

Donovan R. 2013. What have been the ecological and cul- tural impacts due to tourism in the Himalayan region of Dharamshala in India? Journal of Experiential Education and Learn- ing https://doi.org/10.1111/cobi.13341.

Ferraro PJ, Merlin MH, Daniela AM, Gustavo Javier C-B, Subhrendu KP, Katharine RES. 2013. More strictly protected areas are not necessar- ily more protective: evidence from Bolivia, Costa Rica, Indonesia, and Thailand. Environmental Research Letters 8:025011.

Garrard R, Kohler T, Price MF, Byers AC, Sherpa AR, Mahar GR. 2016. Land use and land cover change in Sagarmatha National Park, a world heritage site in the Himalayas of Eastern Nepal. Mountain Research and Development 36:299–310.

Grenyer R, et al. 2006. Global distribution and conservation of rare and threatened vertebrates. Nature 444:93–96.

Gurung DB, Seeland K. 2008. Ecotourism in Bhutan: extending its bene- fits to rural communities. Annals of Tourism Research 35:489–508.

Hall C, Lew A. 2009. Understanding and managing tourism impacts. Routledge, London, United Kingdom.

Hansen MC, et al. 2013. High-resolution global maps of 21st-century forest cover change. Science 342:850–853.

HKTDC. 2017. Yunnan: market profile. HKTDC Research, Wanchai, Hong Kong.

Jenkins M. 2009. Searching for Shangri-La. National Geographic, Wash- ington, D.C.

Kirkby CA, Giudice-Granados R, Day B, Turner K, Velarde-Andrade LM, Duenas-Duenas A, Lara-Rivas JC, Yu DW. 2010. The market triumph of ecotourism: an economic investigation of the private and social benefits of competing land uses in the Peruvian Amazon. PLOS ONE 5 (e13015). https://doi.org/10.1371/journal.pone.0013015.

Kirkby CA, Giudice R, Day B, Turner K, Soares-Filho BS, Oliveira- Rodrigues H, Yu DW. 2011. Closing the ecotourism-conservation loop in the Peruvian Amazon. Environmental Conservation 38: 6–17.

KPMG. 2012. Tourism in Himachal Pradesh and the way ahead. KPMG, PHD Chamber of Commerce and Industry, Delhi, India.

Lambin EF, Meyfroidt P. 2011. Global land use change, economic global- ization, and the looming land scarcity. Proceedings of the National Academy of Sciences of the United States of America 108:3465– 3472.

Laurance WF, et al. 2014. A global strategy for road building (vol 513, pg 229, 2014). Nature 514:262–262.

Li W, Han N. 2000. Ecotourism management in China’s nature reserves. Ambio 30:62–63.

Liu J, Linderman M, Ouyang Z, An L, Yang J, Zhang H. 2001. Ecological degradation in protected areas: the case of Wolong Nature Reserve for Giant Pandas. Science 292:98–101.

Mahapatra PS, Pandey R, Pradhan S. 2012. River rafting in moun- tainous regions of Uttarakhand: impacts, suggested mitigation measures and sustainability. Journal of Mountain Science 9:511– 522.

Mather AS, Needle CL, Fairbairn J. 1999. Environmental kuznets curves and forest trends. Geography 84:55–65.

Miteva DA, Loucks CJ, Pattanayak SK. 2015. Social and environmental impacts of forest management certification in Indonesia. PLoS ONE 10 (e0129675). https://doi.org/10.1371/journal.pone.0129675.

MoCTCA (Ministry of Culture, Tourism and Civil Aviation). 2015a. Nepal tourism statistics 2014. Ministry of Culture, Tourism and Civil Aviation, Kathmandu, Nepal.

MoCTCA (Ministry of Culture, Tourism and Civil Aviation). 2015b. Nepal Tourism statistics 2015. MoCTCA, Kathmandu, Nepal.

Myers N, Mittermeier R, Mittermeier C, da Fonseca GAB, Kent J. 2000. Biodiversity hotspots for conservation priorities. Nature 403:853– 858.

Nagendra H, Karmacharya M, Karna B. 2005. Evaluating forest manage- ment in Nepal: views across space and time. Ecology and Society 10:24.

Nepal SK, Karst H. 2017. Tourism in Bhutan and Nepal. Pages 187–197 in Hall CM, Page SJ, editors. The Routledge handbook of tourism in Asia. Routledge, Oxon, United Kingdom.

Nolte C, Agrawal A, Silvius KM, Soares-Filho BS. 2013. Governance regime and location influence avoided deforestation success of pro- tected areas in the Brazilian Amazon. Proceedings of the National Academy of Sciences 110:4956–4961.

Nyaupane GP, Morais DB, Dowler L. 2006. The role of community involvement and number/type of visitors on tourism impacts: a con- trolled comparison of Annapurna, Nepal and Northwest Yunnan, China. Tourism Management 27:1373–1385.

Olson DM, et al. 2001. Terrestrial ecoregions of the worlds: a new map of life on Earth. BioScience 51:933–938.

Conservation Biology Volume 33, No. 6, 2019

https://doi.org/10.1111/cobi.13341

https://doi.org/10.1371/journal.pone.0013015

https://doi.org/10.1371/journal.pone.0129675

1328 Himalayan Ecotourism

Ostrom E, Janssen MA, Anderies JM. 2007. Going beyond panaceas. Proceedings of the National Academy of Sciences of the United States of America 104:15176–15178.

Pandey S, Wells M. 1997. Ecodevelopment planning at India’s Great Himalayan National Park for biodiversity conservation and participa- tory rural development. Biodiversity & Conservation 6:1277–1292.

Pandit MK, Manish K, Koh LP. 2014. Dancing on the roof of the world: ecological transformation of the Himalayan landscape. BioScience 64:980–992.

Schroeder K, Sproule-Jones M. 2012. Culture and policies for sustainable tourism: a South Asian comparison. Journal of Comparative Policy Analysis 14:330–351.

Sekhon JS. 2011. Multivariate and propensity score matching software with automated balance optimization: the matching package for R. Journal of Statistical Software 42:1–52.

Shui W, Xu GW. 2016. Analysis of the influential factors for changes to land use in China’s Xingwen Global Geopark against a tourism development background. Geocarto International 31:22–41.

Singh RB, Mal S, Kala CP. 2009. Community responses to mountain tourism: a case in Bhyundar Valley, Indian Himalaya. Journal of Mountain Science 6:394–404.

Singh S. 2002. Managing the impacts of tourist and pilgrim mobility in the Indian Himalayas. Revue de Geographie Alpine 90:25–36.

Stevens S. 2003. Tourism and deforestation in the Mt Everest region of Nepal. Geographical Journal 169:255–277.

TCB (Tourism Council of Bhutan). 2015. Bhutan tourism monitor annual report 2015. TCB, Thimpu, Bhutan.

Waldron A, Miller DC, Redding D, Mooers A, Kuhn TS, Nibbelink N, Roberts JT, Tobias JA, Gittleman JL. 2017. Reductions in global biodi- versity loss predicted from conservation spending. Nature 551:364– 367.

Wang CH, Buckley R. 2010. Shengtai Anquan: managing tourism and environment in China’s forest parks. Ambio 39:451– 453.

WTTC (World Travel and Tourism Council). 2014. The economic im- pact of travel and tourism. WTTC, London, United Kingdom.

Wyman MS, Stein TV. 2010. Modeling social and land-use/land-cover change data to assess drivers of smallholder deforestation in Belize. Applied Geography 30:329–342.

Xu JT, Yin RS, Li Z, Liu C. 2006. China’s ecological rehabilitation: unprecedented efforts, dramatic impacts, and requisite policies. Ecological Economics 57:595–607.

Conservation Biology Volume 33, No. 6, 2019