Association of Non-native Plant Species With Recreational Roads in a National Park in the Eastern Himalayas, China

Although the eastern Himalayas have high plant biodiversity, we know very little about plant invasions in the region. This study is the first to examine non-native plant distribution in a popular eastern Himalayan national park. A total of 61 non-native plant species were found in roadside plant communities, which are frequently disturbed by hikers, pack animals, and recreational vehicles. These species were annual or biennial herbs, most of which originated in America or Europe. Non-native plant richness varied with the degree of anthropogenic disturbance. Specifically, greater numbers of non-native species were found at road heads and ends, which are generally subject to intense human activity. The average number of non-native species also varied according to the type of road and road use, with more present along motor roads and horse-riding trails than along hiking trails. These results highlight the role of vehicles and pack animals as dispersal vectors and provide a foundation for future invasion management decisions. To prevent the spread of non-native plants from park roads to the adjacent landscape, we also recommend the development of educational and monitoring programs that encourage tourist participation in conservation efforts.


Introduction
Spreading of non-native plants often depends on humans (Hodkinson and Thompson 1997;Lonsdale 1999). As tourism has increased during recent decades, one might hypothesize that the number of non-native plants and the areas they occupy have also increased with recreational use (Pauchard et al 2009;Pickering and Mount 2010;Barros and Pickering 2014). The impacts of increasing tourism are of particular concern for protected areas such as national parks, heritage sites, and nature reserves Spear et al 2013;Tolvanen and Kangas 2016).
The introduction of non-native plants through tourism to protected areas can occur intentionally or unintentionally. Examples of intentional introduction are ornamental or vegetable plants in gardens and greening/ amenity planting on recreational roads and ski runs (Wilson 1992;Johnston and Pickering 2001;Hulme 2011). Examples of unintentional introduction include transport of plant parts by vehicles (Lonsdale and Lane 1994;Zwaenepoel et al 2006;von der Lippe and Kowarik 2007;Ansong and Pickering 2013a); on clothes, shoes, or equipment (Clifford 1956;Whinam et al 2005;Mount and Pickering 2009); and in the fur and dung of pack animals (Fischer et al 1996;Campbell and Gibson 2001;Ansong and Pickering 2013b).
Recreational activities also create habitats that favor non-native plant establishment and encourage further spread in protected areas. For example, native vegetation is frequently damaged by trampling or soil alteration during the construction and maintenance of tourism infrastructure (eg roads, trails, and accommodations) (T€ orn et al 2009;Wells et al 2012;Wolf and Croft 2014), in addition to direct tourist activities such as driving, hiking, biking, and horse riding (Newsome et al 2008;Pickering Mountain Research and Development (MRD) An international, peer-reviewed open access journal published by the International Mountain Society (IMS) www.mrd-journal.org

Systems knowledge
and Mount 2010;Hammitt et al 2015). Fertilizers used in gardening or site rehabilitation and the manure of pack animals also facilitate the establishment of non-native plants in many protected areas (Campbell and Gibson 2001;Gower 2008;Quinn et al 2008).
Mountain regions host one-third of the world's protected terrestrial areas (Kollmair et al 2005;Price 2007). Because of their harsh climate and limited accessibility, they have traditionally been assumed to be at low risk from plant invasions (Millennium Ecosystem Assessment 2005), but climatic change and growing human influence have altered the situation (Blyth et al 2002;Gurung et al 2012;Pauchard et al 2016). Nearly 1000 non-native plant species have now been reported in mountain areas around the world (Alexander et al 2011;McDougall et al 2011;Seipel et al 2012), and their diversity and abundance have increased dramatically in recent decades (Becker et al 2005;Py sek et al 2011;Kalwij et al 2015).
The eastern Himalayan region of China covers southeast Tibet, west Sichuan, and northwest Yunnan. The region possesses a majority of China's biological assets and is home to several large protected areas (Xu and Melick 2007;Zachos and Habel 2011). The landscape, ethnic cultures, and constantly improving infrastructure have attracted increasing numbers of domestic and international tourists since the 1990s (Fang 2002;Zhou and Grumbine 2011), elevating the likelihood of nonnative flora introduction. However, few studies have examined the distribution and diversity of non-native plants in this region, particularly as they relate to touristmediated introduction of these species.
In this study, we therefore investigated non-native plant distribution in a national park (Laojun Mountain Area) in northwest Yunnan Province (Figure 1). We hypothesized that the study site has been experiencing tourist-mediated dispersal of non-native plants for the past few decades. Using data from vegetation surveys, we aimed to determine (1) which non-native plant species occur in the park, particularly along recreational roads, (2) how non-native plant composition varies with travel distance from the park entrance, and (3) whether nonnative plant presence differs along different road types with different uses: motor roads, horse-riding trails, and hiking trails. Our results can contribute to an improved understanding of non-native plant distribution patterns in protected areas of the eastern Himalayas. This basic information can help clarify the management methods needed to mitigate the negative effects of plant invasions on eastern Himalayan ecosystems. Featuring the typical mountainous terrain of the eastern Himalayas, the park's annual average temperature  Mingyu Yang) and precipitation are around 11.68C and 900-1000 mm, respectively. Three natural or seminatural vegetation patterns can be found in the park: grassland with deciduous oak shrubs in the lowlands (,2500 m), coldtemperate coniferous forests (pine, spruce, and fir) at middle elevations (2500-3200 m), and subalpine mosaics (meadow and scree) at high elevations (.3200 m).

Material and methods
Once isolated and hard to access, the area was populated only by small, local ethnic groups (Lisu, Yi, Naxi, and Tibatan) who practiced subsistence farming and herding (Buntaine et al 2007;Grumbine 2012). Since the 1990s, however, tourism has become a viable option for local communities to alleviate poverty and support nature conservation (Zhou and Grumbine 2011;Zinda 2014). The park has become an important site for outdoor recreation in southwest China; in 2015 more than 150,000 tourists visited it to hike, mountain climb, and view the wildlife and landscape.

Plant sampling
Definition of non-native plants and nomenclature: Following definitions by Richardson et al (2000) and Py sek et al (2004), we determined non-native plants in the Laojun Mountain Area to be those not endemic to southwest China. Plant nomenclature, origin, growth form, dispersal mode, and reproduction mode were recorded based on The Plant List (URL: www.theplantlist.org), Flora of China (URL: http://frps.eflora.cn), and other publications (Ding et al 2006;Xu and Lu 2006;He 2012;Ma et al 2013;Mo et al 2013). To ensure that we did not over-report the number of non-native species, plants that could be identified only to the genus level were considered natives.
Field sampling: During 2014-2016, plant surveys were conducted in the growing season (August to October), when plant species composition is relatively stable and at its peak. With the assistance of park managers, 125.5 km of roads were identified as frequently used by tourists and were sampled.
To achieve systematic data collection, recreational roads were segmented into 2 km bands, each containing 2 plots, 1 m 3 2 m each, with the long axis parallel to the road. Plot size was determined based on pilot surveys in 2013 that revealed that most non-native plants in the area are herbs (rather than shrubs or trees). Sites with recent natural disturbances (eg landslide or fire) or non-tourismrelated anthropogenic disturbances (eg settlement or farming) were avoided. A total of 122 plots were sampled.
Within each plot, native and non-native species were identified and their percentage covers were estimated according to the Brann-Blanquet scale (Peet and Roberts 2013; also see footnote to Table 1). The location and elevation of each plot were recorded using a global positioning system (GPS, Garmin eTrex Vista HCx). Plot distance from the park entrance was measured using a distance-measuring wheel (Topmeasure DM60).

Data analysis
Pearson's correlation was applied to detect the relationship between non-native species richness (the number of non-native species per plot) and distance from the park entrance. To explore the difference between non-native and native species richness, the corresponding pattern of native species was analyzed with the same method. A Brown-Forsythe multiple comparison test was used to compare the number of non-native plants along 3 road types (motor roads, horse-riding trails, and hiking trails). This method was preferred to Dunnett's T3 test, because it does not assume equal variances among the types. All statistical computations were performed using SPSS 21.0 software (SPSS Inc., Chicago, IL, USA). Significance was determined at an alpha level of 0.05 unless otherwise stated.

Species richness and composition by travel distance
Species richness and distance from the park entrance were significantly correlated (r ¼ À0.24; P ¼ 0.01). Most non-native plants were found closer to park entrances (,5 km) (Figure 2). The fewest non-natives (less than one species in plot on average) were found at distances of 15-20 km. Near road ends, the number of non-natives gradually increased, with .2 species per plot. In contrast, native plant richness showed a hump-shaped relationship with distance from park entrances (Figure 2), being the highest in middle-distance plots and less abundant near road heads or ends. Almost half of the species were distributed only near park entrances (,5 km, Figure 3). Only 17 species (B. bipinnata, C. serotinum, Galinsoga parviflora, G. quadriradiata, N. physalodes, O. corniculata, P. asiatica, Cosmos bipinnatus,

Species richness and composition by road type
Non-native plants occurred in 44 of the 53 plots (83%) along motor roads, 32 of the 36 plots (89%) along horse trails, and 19 of the 33 plots (58%) along hiking trails.
Multiple comparison testing showed that the average number of non-native plant species per plot differed among the 3 road types (F 2,119 ¼ 9.163, P , 0.001) and was significantly lower along hiking trails than along motor roads and horse trails (Table 2). In contrast, the number of native plant species did not differ across the 3 road types (F ¼ 2.60, P ¼ 0.078). Moreover, plots along hiking trails had a significantly lower non-native to native ratio than plots along the other 2 road types (Table 2). In total, 55 non-native species were identified along motor roads and 31 species along horse trails. Only 15 non-native species were recorded along hiking trails. The 3 most frequent non-native species (G. quadriradiata, O. corniculata, and P. asiatica) occurred along all 3 road types. B. pilosa, E. sumatrensis, B. bipinnata, and B. alba were also relatively common along motor roads and horse trails, while T. mongolicum and C. serotinum were more frequently encountered on hiking trails.

Discussion
This study is the first to examine plant invasion in a popular national park of the eastern Himalayas. The region has seen a considerable rise in tourism to its protected areas but no corresponding rise in research on how the influx of human activity has contributed to nonnative flora occurrence. Such paucity of knowledge poses a real challenge for local biodiversity conservation. In this study, we inventoried 61 non-native vascular plants and described their distribution along recreational roads in the park. Our results highlight potential threats of tourism to local ecosystems.
Most identified non-native plants were from America, Europe, and other temperate regions, which may suggest an overlap in climatic conditions (Khuroo et al 2007;Kosaka et al 2010;Sekar 2012), while also partially reflecting travel or trade patterns between the eastern Himalayas and source regions (Weber and Li 2008).
Non-native plants in the study area were predominantly herbaceous and annual or biennial, rather than woody or perennial. Non-natives depended more on  seed output, and few had the capacity to spread vegetatively. These findings indicate that non-native plants in the Laojun Mountain Area are mainly early succession species that are widely considered to benefit from human disturbance and can take advantage of roadside environments (Weber 2003;Pickering and Hill 2007;Wells et al 2012).
Non-native plants that exhibited high frequency and coverage were from the families Compositae (eg G. quadriradiata, B. pilosa, T. patula, E. sumatrensis), Amaranthaceae (eg C. serotinum), and Solanaceae (eg D. stramonium), consistent with previous research on mountain regions worldwide (McDougall et al 2011;Seipel et al 2012;Kueffer et al 2013). These species possess shared traits that contribute to their successful dispersal and establishment in the study area, such as prolific seed production, easily spread seeds, and rapid initial growth (Figure 4). Because of these traits, continued introduction or spread of these species by tourism activities may lead to their proliferation in the park, with substantial ecological and economic impacts.
Non-native plants were recorded more frequently in plots close to entrances. This effect might be associated with higher tourist presence and greater disturbance in these areas. Disturbance may result in an increase of resource availability (eg light and soil nitrogen) and provide new niches for colonization by new plants (Hobbs and Huenneke 1992;Py sek et al 2010). Simultaneously, hikers, pack animals, and vehicle traffic can also bring high quantities of seeds to sites near the entrances. This propagule pressure is likely a major reason that nonnative richness dropped with increasing distance from park entrances (von der Lippe and Kowarik 2012; Lockwood et al 2013).
In middle-distance plots, species richness of nonnative plants was lower than that of native plants ( Figure  2). In addition to having lower levels of disturbance and propagule pressure, roads in this portion of the park are lined with dense forests. Lack of sufficient sunshine and high competition from native species thus hamper germination and colonization of non-native plants. This suggests that the types of native vegetation that are present in an area influence the extent and pattern of non-native plant invasions (Davis et al 2000;Davies et al 2011).
Plots close to road ends contained a higher diversity of non-native species than roads in the middle distances. Road-end plots were all in or close to subalpine meadows, where park tourists commonly visit to enjoy the scenic views and often camp overnight. Visitors frequently bring pack animals into the meadows, resulting in intense grazing and manure accumulation, which combine with the meadows' high light intensity and wet soils to facilitate the establishment of non-native species ( Non-native plants appeared to prefer plots along motor roads and horse trails to those along hiking trails. Likewise, the ratio of non-native to native species was higher along the former 2 road types than along hiking trails. These patterns confirm the importance of vehicle traffic and pack animals in spreading non-native plants in the park (Campbell and Gibson 2001;Quinn et al 2008). The effectiveness of these vectors is attributable to the presence of multiple dispersal modes including vehicle surfaces, tire mud, horse dung, fur, and hooves, as well as traffic-related disturbances such as to road width or ground hardness (Landsberg et al 2001;Quinn et al 2010).
Tourists themselves are only modest vectors of plant dispersal.
The above findings allow us to recommend several strategies for managing plant invasions in protected areas of the eastern Himalayas. These appear most timely given the growing number of tourists visiting the area and the effects of rapid climate change.
First, education programs should be developed that focus on changing public perception and behavior related to non-native plants. Specifically, educators must emphasize the negative impact of non-native plants and detail appropriate actions tourists can take to avoid disturbances and introductions (Burt et al 2007). Encouraging tourists to identify and report the occurrence of non-native plants during their stay will contribute to improved stewardship and might be an effective instrument for managing local invasions. The increasing popularity of mobile apps provides a powerful opportunity for developing a tool that allows tourists to participate with ease.
Second, management of non-native species in protected areas of the eastern Himalayas should focus on dispersal corridors and other tourist-disturbed habitats (eg campsites and viewing spots). These habitats are especially vulnerable to colonization by non-native flora but are also far easier to control than adjacent rugged environments, where management efforts become extremely unwieldy and costly (Kueffer et al 2013).
Possible actions that can prevent invasion into difficult terrain include minimizing seed production of non-native plants, especially those with ecological significance and high dispersal potential. In addition, native species can be planted in these habitats to reduce opportunities for nonnative establishment. Third, park managers should carefully reevaluate the use of vehicles and pack animals, given their role as the primary vectors of non-native plants. Specifically, we recommend the establishment of effective quarantine procedures, including wash units to clean vehicles arriving and leaving protected areas (Pickering and Mount 2010) and a certified weed-free forage program for pack animals, to discourage the introduction of hay infested with non-native plants (McDougall et al 2011).
Finally, the results of our study indicate 3 possible directions for future research. First, systematic surveys of non-native species must be prioritized in protected areas of the eastern Himalayas. Most such areas are currently assumed to be free of biological invasions. But this study has clearly shown that non-native species have already invaded. Second, future studies should investigate factors that encourage human-mediated dispersal and establishment of non-native species. We still do not have basic information such as the maximum travel distance of a seed carried by human agents, the characteristics of plant communities particularly susceptible to invasion, and how environmental factors (eg climate, soil, or even elevation) determine plant growth and success. Third, researchers should aim to build an information-sharing network across the region that includes neighboring nations (eg Bhutan, India, and Nepal). The increased FIGURE 4 Examples of biological characteristics that enable non-native species to become established in Laojun Mountain National Park. (A) Galinsoga quadriradiata can produce 3000 flower heads and up to 7500 seeds per plant. (B) Chenopodium serotinum produces an abundant amount of very small seeds (1.2-1.8 mm in diameter) that are easily spread and can germinate immediately upon contact with moist soil. (C) Datura stramonium has a seed capsule that is 3-8 cm in diameter; at maturity, the capsule splits into 4 chambers, each with dozens of seeds. Seedlings establish quickly to become large thickets with big leaves that shade out all surrounding vegetation. (D) Tagetes patula is a fast-growing and light-demanding pioneer species, establishing readily and densely in disturbed sites. The dense growth makes it highly competitive to native plant species. (E) Bidens pilosa has seeds with 2 to 4 barbed spines that increase the likelihood of attaching to animal fur or human clothing (Global Invasive Species Database, URL: www.iucngisd.org; European Network on Invasive Alien Species, URL: www.nobanis.org; Weber 2003). (Photos by authors) availability of relevant data will both support local management capacities and contribute to understanding plant-invasion mechanisms across the Himalayan region.

Conclusion
In contrast to the prevailing perception, a wide range of non-native plant species were identified in protected areas of the eastern Himalayas. Repeated disturbances and propagule introduction by tourists and their animals and vehicles have increased the establishment of nonnative species in this region, although different vectors vary in the degree of their contribution. Education, monitoring, and quarantines are the recommended defense strategies for invasion management in these areas because of their rugged terrain and limited management resources. Mechanisms by which successful dispersers invade local ecosystems should be further explored, particularly in response to increased tourism and climate change.

ACKNOWLEDGMENTS
We would like to thank undergraduate students Chaoli Yu, Juping Zhang, Lishuang Peng, and Feiyan You of Yunnan University for their assistance with field data collection. We would also like to thank the staff at the Laojun Mountain National Park, particularly Shihong Wang and Hua Li, for their support with fieldwork. We are also grateful to the anonymous reviewers for their insightful comments. This work was supported by the National Natural Science Foundation of China (grant number 41401641).