Illegal harvesting and livestock grazing threaten the endangered orchid Dactylorhiza hatagirea (D. Don) Soó in Nepalese Himalaya

Abstract Harvesting of orchids for medicine and salep production is a traditional practice, and increasing market demand is spurring illegal harvest. Ethno‐ecological studies in combination with the effect of anthropogenic disturbance are lacking for orchids. We compared population density and structure, and tuber biomass of Dactylorhiza hatagirea (D. Don) Soó for three years in two sites: Manang, where harvesting of medicinal plants was locally regulated (protected), and Darchula, where harvesting was locally unregulated (unprotected). Six populations were studied along an elevation gradient by establishing 144 temporary plots (3 × 3 m2) from 3,400 to 4,600 m elevations. Mean density of D. hatagirea was significantly higher in the locally protected (1.31 ± 0.17 plants/m2) than in the unprotected (0.72 ± 0.06 plants/m2) site. The protected site showed stable population density with high reproductive fitness and tuber biomass over the three‐year period. A significant negative effect (p < .1) of relative radiation index (RRI) on the density of the adult vegetative stage and a positive effect of herb cover on juvenile and adult vegetative stages were found using mixed zero‐inflated Poisson (mixed ZIP) models. The densities of different life stages were highly sensitive to harvesting and livestock grazing. Significant interactions between site and harvesting and grazing indicated particularly strong negative effects of these disturbances on densities of juvenile and adult reproductive stages in the unprotected site. Semi‐structured interviews were conducted with informants (n = 186) in the villages and at the ecological survey sites. Our interview results showed that at the protected site people are aware of the conservation status and maintain sustainable populations, whereas the opposite was the case at the unprotected site where the populations are threatened. Sustainability of D. hatagirea populations, therefore, largely depends on controlling illegal and premature harvesting and unregulated livestock grazing, thus indicating the need for permanent monitoring of the species.

tion, or loss of habitats are the most significant threats to the survival of orchid populations (Ehrlich, 1988;Laurance & Bierregaard, 1997;Saunder et al., 1991;Zhang et al., 2019). Also, increasing human activities may create novel environments  which limit the persistence of orchid populations. Human disturbances may result in the breakdown of ecological connections between orchids and their pollinators and mycorrhiza, changes in edaphic and microclimatic conditions, and introduction of pests and diseases (Fay et al., 2015;Light et al., 2003). Disturbances may interrupt interspecific interactions leading to reduced reproductive output and eventually altering the plant demographic dynamics (Steffan-Dewenter et al., 2006). Studies have shown that, generally, the magnitude of disturbance impacts on plant populations depends on plant life stage and features of their reproductive system (Calvo, 1990). In the case of orchid populations, the severity of disturbance impacts also depends on the level of specificity of plant-animal interactions (e.g., interactions with pollinators) and the availability of sites suitable for seedling recruitment (Schulze et al., 2019).
The distribution and abundance of orchid populations depend on a suite of biological and ecological factors including seed production and dispersal, recruitment, availability of mycorrhizal fungi, and appropriate environmental conditions (McCormick & Jacquemyn, 2014). However, in case of the smallest orchid populations the seed output may be insufficient to ensure their long-term persistence (Faast et al., 2011). The life stage dynamics of orchid populations further depends on the elevation of their habitat.
Environmental conditions and interactions associated with altitude play a significant role in the composition and distribution of orchid populations Djordjević et al., 2016Jacquemyn et al., 2005). Alpine and subalpine grasslands suffer from reduced nutrient availability and harsh environmental conditions, and plants growing in these habitats presumably develop adaptive coping strategies (Chapagain et al., 2019). Disturbance regimes, such as harvesting, grazing, trampling, and fire, also play influential positive (Chen et al., 2014;Dai et al., 2019) or negative (Chapagain et al., 2019;Kreziou et al., 2015) roles in determining the growth and persistence of plants and could be important elements of an optimal grassland management strategy for alpine meadows.
Inherently slow growth, high habitat specificity, dependency on pollinators, need of mycorrhiza for reproduction and germination, narrow range of ecological substitution options, unsustainable exploitation, and climate change are major challenges for the growth and development of orchids, such as Dactylorhiza hatagirea (Dhiman et al., 2019;Hinsley et al., 2017;Hutchings et al., 2018;Rasmussen & Rasmussen, 2018;Reiter et al., 2017;Shrestha et al., 2021;Yeung, 2017). Due to a marked decline in its natural populations, D. hatagirea has been listed as an endangered species in Nepal by Conservation Assessment and Management Plan (Bhattarai et al., 2001). According to the Forest Act of Nepal (2019), collection, use, sale, trade, and export of D. hatagirea are prohibited, and the species is strictly protected in list I of Government of Nepal (Go, 2011).

It is also listed under appendix II in Convention on International
Trade in Endangered Species of Wild Fauna and Flora (CITES, 2020).
Nevertheless, due to its high medicinal potency, D. hatagirea is still collected illegally at all life stages and traded to, especially, India and China (Olsen & Helles, 1997;, which has pushed local populations toward extinction (Manandhar, 2002).
Given that many orchids like D. hatagirea are currently threatened or endangered, a better understanding of the factors that influence orchid population ecology and dynamics may be critical to their long-term conservation (Shefferson et al., 2020). Formulation of strategies for conservation of a species requires a sound knowledge of environmental factors, population ecology, and demographic parameters (Margules & Pressey, 2000). Disentangling factors determining successful orchid establishment and its persistence under changing conditions is a major challenge. For the long-term conservation of endangered orchids, the development of concrete conservation plans based on indigenous knowledge, long-term monitoring, genetic analysis, and scientific inputs is crucial (Dobriyal et al., 2002;Jacquemyn et al., 2007). Endangered orchid species are in need of specific conservation actions (Charitonidou et al., 2019;Mincheva & Kozuharova, 2018;Tsiftsis et al., 2011), and more attention should be paid to the management of existing sites of orchids (Stipkova & Kindlmann, 2015).
Thus, the objectives of this paper are to (a) analyze the variation in population density, structure, and tuber production in D. hatagirea over three years in two sites subjected to different levels of anthropogenic disturbances, (b) study the impact of elevation and anthropogenic disturbances on the population density of D. hatagirea, (c) examine the interaction between site and environmental factors (harvesting, grazing, and herb cover) and its effect on population density of D. hatagirea, and (d) study the socio-cultural role of D. hatagirea and assess people's perception of its status.
To meet these objectives, we identified a locally unregulated site in the western part of Nepal and a well-managed, locally regulated site in the central part of Nepal. At these sites, we established permanent plots at different elevations and carried out a survey among local people.

| Study area
This study was carried out in two sites: (i) Lolu-Pilkanda (N29°60.095′ and E080°56.754′ to N29°57.719′ and E080°57.672′) within Api Nampa Conservation Area (ANCA) in Darchula District, north- of the Government of Nepal, and due to poor implementation of regulatory mechanisms, the human exploitation of natural resources at this site is very heavy . Further, the local community has not taken any particular initiatives to conserve the area.
Commercial and illegal trade of MAPs from the area has increased drastically in the last few decades . Hence, the Lolu site is hereafter referred as the locally "unprotected site." ANCA is the youngest and the most remote conservation area of Nepal. The climate is temperate to nival with an annual average precipitation of 2,100 mm and annual mean minimum and maximum temperatures of 4. 7°C and 27°C, respectively (DNPWC, 2015). Important livelihood activities are collection and trade of highvalue MAPs, notably Ophiocordyceps sinensis, Dactylorhiza hatagirea, Fritillaria cirrhosa, Neopicrorhiza scrophulariiflora (Pouliot et al., 2018;Pyakurel et al., 2018), and traditional mountain farming systems tightly integrated with transhumance and other livestock systems (DNPWC, 2015). The area has legal permission for commercial Therefore, the Bhimthang site is hereafter referred to as the locally "protected site." At the protected site in Manang, the climate varies from temperate to nival and is influenced by the summer monsoon. The annual mean minimum and maximum temperatures are 4.7°C and 16.8°C, respectively, and the average annual precipitation is 972 mm (DNPWC, 2015). The basic livelihood activities include and combine traditional mountain farming systems, transhumance and animal husbandry, small-scale trade at lower altitude during winter, tourism, and the collection and trade of highly valued MAPs (Chhetri, 2014;Subedi & Chapagain, 2013).
The study sites differ with respect to edaphic, topographic, and substrate conditions. The unprotected site has a silty-loamy soil and is rich in herbs and grass and also has some bare ground cover, while the protected site has a sandy-loamy soil and is rich in moss, lichen, litter, rock, and scree cover. D. hatagirea is found on slopes ranging between 1 and 66° at the unprotected site and 4-65° at the protected site. The sites do not vary much in terms of relative radiation index (RRI) (Table S1: Appendix S1). The unprotected site is subjected to higher anthropogenic pressure as revealed by higher disturbance scores (harvesting, grazing, trampling, and animal droppings) than observed at the protected site.
D. hatagirea is an erect perennial herb (30-90 cm tall) that grows in moist alpine meadows and forest gaps between 2,800 and 4,600 m asl (Ghimire et al., 1999) favoring high soil Ca content (Thakur et al., 2020). It bears palmate tubers, 5-7 lanceolate or oblong leaves, which are progressively smaller toward the top, and has a robust stem (Figure 2, left). The inflorescence is up to 15 cm long, with a large number of densely packed flowers. Flowers are resupinate and purple to light pink, arranged around the rachis, resembling a hyacinth. Capsules bear thousands of dust-like seeds.
Seeds are minute and have no endosperm and therefore lack storage reserves, and the orchid partly depends on the mycorrhizal fungi Rhizoctonia for nutrition (Giri & Tamta, 2012;Kalimuthu et al., 2007;Warghat et al., 2014). The rate of vegetative propagation is very slow and seed germination in nature is very poor, that is, 0.2%-0.3% (Vij, 2001).
The tubers of D. hatagirea (Figure 2, right) yield a high quality salep (a beverage made from the powder of the orchid tuber), which is used as an aphrodisiac or a nutritive and restorative tonic, and are also eaten raw as a farinaceous food (Baral & Kurmi, 2006;Sood et al., 2005;Thakur & Dixit, 2007;Vij, 1995;Watanabe et al., 2005). It is also used in the treatment of diabetes, chronic diarrhea, dysentery, coughs, hoarseness of voice, paralysis, fractures, during convalescence and to correct malnutrition (Das, 2004;Singh & Duggal, 2009). The whole plant possesses antibacterial properties and is used in curing various bacterial diseases (Ranpal, 2009). Tubers of D. hatagirea contain a wide range of chemical compounds including dactylorhins, dactyloses, glucosides, starch, and albumin (Kizu et al., 1999;Lama et al., 2001). Recent findings indicate that D. hatagirea also has anticancerous properties (Popli, 2017 Table S1: Appendix S1). In each population, we established four transects at a minimum vertical distance of approximately 100 m. In each transect, we established six (3 m × 3 m) plots at a minimum horizontal plot to plot distance of 10 m. Each plot was divided into nine 1 m × 1 m subplots, and the four corner subplots were systematically sampled and measured.

| Sampling design used in the vegetation survey
For each plot, the geographical location (latitude and longitude) and topographical characteristics (elevation, slope, and aspect) were recorded and used to calculate the relative radiation index (RRI) (Oke, 1987;Vetaas, 1992aVetaas, , 1992b. In each subplot, the ground cover (%) for vascular plants (grasses, herbs, and shrubs), nonvascular plants (lichens and bryophytes), litter, bare ground, rock, and scree cover were estimated using standard methods (Pauli et al., 2015).
Disturbance (harvesting, trampling, grazing, and animal droppings) scores ranging from 0 (none) to 4 (very high) were recorded for each subplot after careful observation of the evidence, for example, large holes caused by excavation, wilted or fresh uprooted aerial parts, tuber fragments, browsed plant parts, defoliated aerial parts, and animal droppings.
We classified the individual plants into four life stage classes based on the number and size of leaves and presence of reproductive structures. The four stages are as follows: seedlings (Sd; leaf breadth ≤1 cm, leaf number = 1-2), juveniles (Jv; leaf breadth ≥1 cm, ≤2 cm, leaf number = 2-3), vegetative adults (Adv; leaf breadth ≥2 cm, leaf number >2, nonflowering), and reproductive adults (Adr; flowering or fruiting individuals). Individuals at different stages were counted in each subplot to calculate the density. The population structure was described as the proportion of each life stage within the studied population.
To estimate reproductive traits, we selected fifteen mature individuals from each population at the two sites and recorded the reproductive traits (number of flowers and fruits). For the estimation of dry biomass of tubers, we recorded the weight of fifteen dried tubers of D.hatagirea from the local MAPs collectors.

| Interview survey
Semi-structured interviews were conducted during 2015-2017 among 117 persons in the unprotected site (Darchula) and 69 persons in the protected site (Manang). The informants were local MAPs users, leaders, teachers, and students in the villages, and hotel owners, local tourist guides, and cattle herders working in the sites where we carried out our ecological survey. We explained the goal of our research and obtained the informants' consent before starting the interview. The informants were also informed about their right to withdraw their consent at any stage of the interview. Informants' responses were documented by written notes and in possible cases were also supplemented by voice recording, with the permission of the informants.
We asked the informants to comment on the abundance of D.
hatagirea, collectors (villagers or outsiders), reasons for collecting the plant (local use or commercial purpose), local uses, the collection (areas and time of collection), when they started collecting, indigenous knowledge transfer practices, special tools, prices, qualities, markets (to whom they sold), processing after collection, and their other sources of income. We also asked questions concerning nature conservation; changes observed in the biotope, perceived trends in populations of the species, causes of population change, threats, conservation status, protection measurements, and practices that could ensure survival and sustainable management.

| Data analysis
A relative radiation index (RRI), which is a relative measure of the exposure to solar radiation at noon at a specific location (Oke, 1987;Vetaas, 1992aVetaas, , 1992b, was calculated for each plot as a function of aspect, latitude, and slope: where Ω is aspect (slope azimuth in degrees), Φ is latitude (degrees), and β is slope inclination (degrees).
The densities of different stages (seedling, juvenile, vegetative, and reproductive adults) were compared by Kruskal-Wallis tests, and the reproductive traits were compared among the three populations in each of the two sites using one-way ANOVA.
Direct field observations confirmed the plant as rare in the study site, and we therefore expected that the data collected would exhibit a large number of zeros. We tried with different model alternatives but based on the Akaike information criterion (AIC) the best fit was obtained using mixed zero-inflated Poisson (ZIP) models. The mixed ZIP model allowed us to analyze relationships between density of D. hatagirea plants at different stages and a set of independent variables including population, cover of shrubs or herbs, relative radiation index (RRI), and anthropogenic disturbance indicators such as harvesting, trampling, grazing, and animal droppings.
We prepared ten sets of candidate models using the glmmTMB package (See Appendix S2) and finally prepared an average model based on the set of five best candidate models (selected on the basis of delta AIC) using the MuMIn package (Barton, 2018). More specifically, the final average models were prepared using five models with delta AIC values ≤402 for the seedling stage, ≤748 for the juvenile stage, ≤645 for the adult vegetative stage, and ≤763 for the adult reproductive stage. The full models were in all cases expressed as: where a (intercept), b (population), and c 1 …c 6 are fixed model parameters. i = 1…144 is the plot (included as a random effect); j = 1…4 is the subplot. The population variable had six categories, three at each site (see the section Study area). Confounded variables were excluded from the analysis. All the tests were conducted using the R version 3.5.3 (R Development Core team, 2017).

| Variation in population density and structure
The population density of D. hatagirea at the unprotected site ranged from 0.60 to 0.79 individuals/m 2 while the density ranged from 0.70 to 2.16 individuals/m 2 at the protected site. Population densities were highest in mid-elevation populations at both sites (Table 1). At the unprotected site, all populations showed highest densities for the juvenile and adult vegetative stages, whereas at the protected site, the density of the adult reproductive stage was mostly higher than for the juvenile and adult vegetative stages (Table 1). At the protected site, the variation of density among the populations was significant (Kruskal-Wallis test, p < .05), overall, and for all stages except the adult vegetative stage, but at the unprotected site, there was no significant variation among populations (p >.05).
The population structure varied between the sites (Figure 3).

| Variation in reproductive traits and tuber production
At the protected site, the reproductive output and tuber biomass were about three times higher than at the unprotected site. In comparisons between populations, the reproductive traits (number of flowers, number of fruits, and total reproductive output per individual) were found to decrease from the lowest to the highest elevation at both sites ( Table 2). The dry biomass of daughter tubers showed similar trends at both study sites. A comparison of tuber production across three consecutive years (2015, 2016, and 2017) showed reduced tuber production in 2017 at the unprotected site and increasing tuber production at the protected site ( Figure 5).  (Table 3).  (Table 4).

| Interview survey
About forty-three percent of the MAP users interviewed at the unprotected site (n = 117) were aware that D. hatagirea is strictly protected, seventy-one percent were aware of its use value, five percent were aware of its population ecology, and forty-two percent harvested D. hatagirea for local uses to treat cuts and wounds, boils, fractures, and to use it as a tonic. Ninety-two percent of the informants claimed that illegal harvesting of D. hatagirea was a common practice and that the population had decreased drastically over the last few decades. They further disclosed that there are illegal traders in the district headquarters of Darchula who motivate MAP users and cattle herders to harvest D. hatagirea by promising to buy the dried tubers. Further, perceived difficulties to collect Ophicordyceps sinensis and Fritillaria cirrhosa increased the temptation to carry out illegal and premature harvest of D. hatagirea. The informants mentioned that for every 1 kg of dried tubers sold, approximately 500-1,000 mature plants are harvested. Eighty-one percent of the TA B L E 2 Variation in reproductive output of Dactylorhiza hatagirea in populations in the locally unprotected (Darchula) and locally protected (Manang) sites

Dhauli up
15 .  informants are of the view that the major source of their livelihood is MAP collection.
At the protected site, about ninety-three percent of the MAP users interviewed (n = 69) were aware of the harvesting ban, ninetyeight percent were aware of the species' use value, nineteen percent were aware of its population ecology, and twenty-one percent occasionally harvested a few (2-5) individuals of D. hatagirea for home use to treat cuts and wounds, burns, and boils and for religious purposes by Buddhist Lamas. About twenty-four percent were of the opinion that the population is decreasing, while seventy-three percent thought that the population has been almost constant in the last few decades as the local community is involved in patrolling the area to control illegal collection during the maturation period. Our interview results also revealed that none of the families totally rely on MAP collection for their livelihood as they had access to income from the flourishing tourism. We also observed that awareness programs were run at community level, emphasizing the sustainable use of available MAPs. Further, they also used a specific route to protect sensitive plants from damage caused by grazing and trampling.

| D ISCUSS I ON
The study is the first of its kind in this region, but similar research has been done in Greece (Charitonidou et al., 2019). Our work provides results based on a long-term study from the Nepal Himalayas and should therefore be of significance to the conservation management of the endangered orchid D. hatagirea. Environmental variables and human-mediated disturbances such as harvesting and livestock grazing had significant effects on the population structure, density, reproductive traits, and tuber production of D. hatagirea.

| Variation in population density and structure
We recorded a maximum population-level mean density of At both sites, we recorded a low proportion of seedlings. This could partly be explained as a consequence of orchids being habitat specific and seedling establishment depending on a suite of environmental factors, which are rarely present at the same time and place (Shefferson et al., 2020). Our results compare well with previous research as small populations have substantially lower viability compared to larger populations, and seedling recruitment rates can be considerably lower in small populations, which results in significantly lower population growth rate and density (Hens et al., 2017;Pellegrino & Bellusci, 2014). The low proportion of reproductive plants observed at the unprotected site compares well with observations made by Pellegrino and Bellusci (2014)

| Variation in the reproductive traits
We found reduced reproductive fitness in D. hatagirea at the unprotected site. Alterations due to anthropogenic disturbances in natural habitats often reduce the size and density of populations (Aguilar et al., 2006;McKinney, 2002). Anthropogenic disturbances increase the spatial distance between the plant populations as well as between individuals within a population, thereby disrupting insect movement between plants (Öckinger et al., 2009), decreasing pollinator abundance (Liu & Koptur, 2003), and altering their behavior and the frequency of flower visits (Aguilar et al., 2006). This process ultimately decreases the reproductive fitness of the plants (Peterson et al., 2008). The level of inbreeding may be higher in small, isolated populations (Miao et al., 2014) because of the higher rate of selfing and more frequent mating between close relatives. The unprotected populations had a lower population size and a lower fruit set than did protected populations, suggesting that the latter populations are better buffered. This could be a consequence of inadequate pollinator visitation in small populations, resulting in insufficient pollen transfer, poor pollination, and lower seed set (Smithson, 2006;Tremblay et al., 2005;Xia et al., 2013). By contrast, larger populations of plants are likely to be more attractive to pollinators, resulting in higher visitation rates and therefore higher pollination success (Mustajärvi et al., 2001). ably has a huge negative impact on its growth and development by breaking the aerial parts before fruit maturation and seed dispersal.
Besides, collectors who are unable to collect sufficient amounts of O. sinensis and F. cirrhossa are tempted to illegally harvest the tubers of D. hatagirea irrespective of its degree of maturity. Such practices are also common in other parts of the world (Ghorbani et al., 2014;Kreziou et al., 2015). Further, the harvesting of the orchids involves destructive uprooting of the daughter tubers, which kills the plants.
When ascending from lowland to alpine environments in the Himalayas, plant species experience a large variation in abiotic conditions over an extremely short distance (Korner, 2003). With increasing elevation, changes in pressure, temperature, wind speed, UV exposure, and soil properties have been shown to affect different phenological and morphological properties of plants Hodkinson, 2005), thus also influencing growth and reproductive performance. The decreasing number of reproductive parts observed along the altitudinal gradient could be further attributed to the time of flowering, which is influenced by the ambient temperature and the timing of the snow melt (Kudo & Hirao, 2006).

| Effects of different environmental variables on the density of different stages
The occurrence and distribution of orchid species are influenced by environmental and topographical factors such as latitude, altitude, slope, and aspect (Bulafu et al., 2007;Djordjević et al., 2016. In this study, we observed a very weak negative effect of the Relative Radiation Index (RRI) on the adult vegetative density ( (Pouliot et al., 2018). This is also in agreement with Poudeyal et al. (2019) who observed that intense human disturbances, especially harvest, played a crucial role as determinants of the density and structure of Neopicrorhiza scrophulariiflora populations.

| Interview survey
We Annapurna Conservation Area Project (ACAP), which has worked in this area for three decades, has also contributed to increase conservation awareness and promote sustainable use of natural resources (Baral & Heinen, 2007).
In contrast, in the unprotected site there is no strict local protection system and people from outside ANCA are also allowed to harvest MAPs. The local MAP users were also found with to have a low level of conservation awareness (in relation to D. hatagirea) as also reported for northwestern Greece by Kreziou et al., (2015).
Moreover, due to the lack of protection measures and awareness, the MAP users were found to engage in intensive harvesting of D.
hatagirea, and selling it to local traders

| CON CLUS I ON AND CON S ERVATI ON IMPLI C ATI ON S FOR D. HATAG IRE A
Harvesting Note: Parameter estimates with standard errors in brackets. Significance levels are stated as: *p < .1, **p < .05, ***p < .01, ****p < .001.  during fieldwork. The local research assistants who helped us to establish the plots are highly acknowledged. We express our sincere thanks to the locals for their coordination and warm hospitality during the fieldwork and for sharing their knowledge. We thank Naresh Subedi for his constructive suggestions on an earlier version of the manuscript. We also thank Pratikshya Chalise and Yagya Raj Paneru for helping us during the fieldwork in Manang.
The Central Department of Botany, Tribhuvan University, facilitated the fieldwork. Finally, we thank two anonymous reviewers and editors for critical comments which helped to improve the manuscript.

CO N FLI C T O F I NTE R E S T
The authors declare no competing interests exist.

AUTH O R ' S CO NTR I B UTI O N
DJC and SKG designed the experiment; DJC and SBM collected the data; DJC, HM, CBB, and SKG analyzed the data and wrote the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data used for the ecological analysis are presented in the paper and the Supplementary material. Data that identify human participants in the study will not be made publicly available. However such data can be made available upon request to the authors.