Previous PageTable Of ContentsNext Page

Yak grazing and forest dynamics in Sagarmatha National Park, Nepal1

B. Brower1 and A. Dennis2

1. Geography and International Studies, Portland State University, Portland, OR 97207– 0751, USA
2. CALFLORA Database, 937 San Pablo Avenue, Albany, CA 94706, USA

1. A version of this paper appeared in 1999, entitled: 'Grazing the forest, shaping the landscape? Continuing the debate about forest dynamics in Sagarmatha National Park (SNP), Nepal'. In: Zimmerer and Young (eds), New lessons from nature's geography: Biogeographical landscapes and conservation in developing countries. University of Wisconsin Press, Madison, USA. Fulbright-Hays Faculty Research Abroad award supported the research on which the paper is based, while a Social Science Research Council provided grant for B. Brower, and University of Illinois Faculty of Research provided award for A. Dennis.


Livestock grazing in forest sets off the alarm for resource managers who understand grazing to be incompatible with trees. Particularly in protected areas, livestock are targeted for restriction and removal because of the popular perception that forest grazing damages trees and inhibits regeneration, leading to degradation. However, a small-scale study of forest dynamics in Sagarmatha (Mountain Everest) National Park, Nepal, documents active regeneration in yak-grazed plots of Abies spectabilis woodland near Sherpa villages, and supports an interpretation of warm-slope vegetation as a diverse and enduring shrubland, not degraded forest. These findings challenge much of the received wisdom about the dynamics of Himalayan forests in general and those of the Mt Everest region in particular, and suggest that yak and yak-cattle hybrids here have been unfairly targeted as agents of forest destruction. Sherpa people have managed livestock to increase their spatial and temporal distribution, reducing their impacts. Tourism has accelerated the erosion of traditional grazing management, but this is not yet manifested to damage forests.

Keywords: Forest grazing, Mountain Everest National Park/Khumbu, resource management, yak


Sagarmatha National Park (SNP), a 1110 km2 reserve on Mt Everest's south flanks, encompasses not only the world's highest mountain and its spectacular sister peaks, but also the lower valleys of Khumbu, homeland of the Sherpa people. The Sherpa, famous for exploits on Everest, continue to live within the park. They retain claim to private lands within the national park's boundaries, but find themselves subject to a variety of regulations brought about by the expectations and priorities of the park's visitors and managers.

In common with many mountain peoples, their traditional way of life depends on a variety of subsistence strategies including agriculture, trade and yak keeping. Their diversified subsistence exploits a variety of resources distributed along an extraordinarily long altitude gradient (more than 3000 metres between lowest and highest settlements), and provides what has been a sustainable way of life characterised by flexibility and resilience in a high-risk, high-altitude landscape. Yet Sherpa lives must constantly readjust to shifts in the expanding and volatile tourist economy and to ever-evolving policies for park management. The resilience of both people and landscape is being tested by the last few decades of accelerating change. The region's forest resources in particular were considered threatened by human use, and activities believed to damage woodlands, including wood cutting and livestock grazing (Blower 1971; Mather 1973; Mishra 1973; Lucas et al. 1974; Speechly 1976), were curtailed when the park was established in 1976 (Garret 1981).

Management of the park's resources in the years since has continued to be preoccupied with trees (Drew and Sharma 1977; Halkett 1981; Hardie et al. 1987; Ledgard 1989; Ledgard and Baker 1990). Rules promulgated by the park spell out where village fuel and construction wood may be gathered, restrict the kinds and movements of livestock, and promote the establishment of nurseries and plantations. This arbour-centric management persists despite a growing uncertainty about the status of the area's forests (Houston 1982; Thompson and Warburton 1985; Byers 1986; Ives and Messerli 1989; Stevens 1993), and in the face of continuing local resentment about the infringement of accustomed access to resources (Halkett 1981; Brower 1983; Brower 1987; Brower 1990; Stevens 1993). Khumbu's forests were initially assumed to be degraded remnants of formerly uninterrupted woodland covering all slopes and aspects below about 4100 metres above sea level (masl) (Naylor 1970; Stettler and MacDonald 1982 (unpublished data); Hardie et al. 1987), severely diminished in recent times by the collective assault of local people and swelling ranks of tourists (Blower 1971; Mishra 1973; Lucas et al. 1974; and Hardie et al. 1987). There are, however, questions about the extent of former forest, the time of deforestation, and the role of humans and livestock in the present patterns of forest cover (Brower 1987; Byers 1987a; Brower 1991a; Stevens 1993; Dennis and Brower 1995; Sherpa 1999).

In this paper we review the ideas and evidence that have accumulated about vegetation dynamics in SNP, and offer additional information in the form of our own study, an analysis of forest dynamics in the neighbourhood of Khumbu's main settlements. Our study refutes the conventional wisdom about a recently accelerated retreat in SNP forests. It affirms other work that argues against their declining status, challenges assumptions about the impact of grazing, suggests a new role for early human settlement in shaping the present pattern of forest and shrubland, and provokes a set of questions about forest interpretation and management. Our object is to augment the emerging picture of people–livestock–forest interactions in SNP, to warn against the use of unsupported intuition as a guide to resource management, and to argue for conservation policy that draws on both social and natural science, favouring neither and recognising the limitations of both.

The case of Sagarmatha

The history of environmental research and park planning and management in SNP reflects the increasing understanding about the interactions of culture and environment that has developed over the last twenty years. The importance of long-term human use in shaping 'natural' environments has grown, as has the awareness that successful protection of national parks and reserves requires the support and collaboration of people whose lands and resources are singled out for protection. SNP is especially intriguing for a number of reasons. This is one of several areas in the Himalaya where misgivings about conventional explanations for ecological processes first took root among researchers whose findings were at odds with expectations. Uncertainty on this Himalayan scale (Thompson et al. 1986) was of sufficiently high profile that researchers elsewhere took note, beginning to explore anew what had been considered settled questions about man's role in changing the environment. The intensity of attention that has been paid to the place, since Nepal opened its borders to the West 55 years ago, is perhaps unprecedented. When Everest was identified as earth's highest peak in 1854, it was marked as special and ensured the interest and attention of hosts of adventurers from all over the world.

Tourists have visited the area in tens of thousands, drawn both by mountain spectacle and challenge and by exotic culture. Development specialists have come to reforest; to tend both Sherpa and foreigners' health needs; to educate; to improve livestock; to initiate hydroelectricity at various scales; to monitor geomorphic processes; and to bolster indigenous culture and institutions. Scholars, too, have been very busy in SNP/Khumbu. Geographers, anthropologists, geologists, foresters, ecologists, biologists, and others have studied and written about the region, producing everything from magazine pieces based on a two-week trek to substantial books grounded in many years' abode. The long-term interest and deepening familiarity of a number of commentators provide insight, too, into the evolution of understanding about the workings of people and environment here (von Haimendorf 1964; Hillary 1964; von Haimendorf 1975; Ives 1979; Brower 1982; Brower 1983; Stevens 1983; von Haimendorf 1984; Brower 1987; Stevens 1989; Brower 1990; Brower 1991a; Brower 1993; Adams 1996; Brower 1996). In addition to these visitors, there is a significant voice of Khumbu-born Sherpas themselves, writing either in their role as young scholars trained in the West (Sherpa 1979a; Sherpa 1979b; Sherpa 1982; Sherpa 1987; Sherpa 1988; Sherpa 1993; Sherpa 1999) or as representatives of older Sherpa traditions (Tengboche Rinpoche and Klatzel 2000). All these eyes and voices mean a wide array of interpretations of the landscape of Khumbu/SNP.

Forests and grazing

This is perhaps nowhere better illustrated than by outsiders' perceptions about forest dynamics and the presumed role of livestock in deforestation. People who have commented about the state of Khumbu's forests include pleasure-seekers, adventurers, journalists, scientists and government agents. Some are passing through, others stop to study; each brings the preoccupations and perceptions of a particular point of view. Their understandings of what they witness, reported in casual accounts, news reports, scientific journals, and government studies, have shaped a public understanding of the interaction between local people and wildland resources that has determined national and international policy regulating the use of lands and resources in SNP.

One such perception, widely shared and popularly circulated, sees Khumbu's forest history as a piece with the supposed pattern of gradually worsening human-induced deforestation hypothesised for the Himalaya as a whole:

The problem of the Himalaya . . . can be stated very simply. In an already high energy environment, rapidly accelerating human pressures over the last 50 years are exponentially augmenting massive and rapid transfer of large volumes of material downslope overcoming, at least in part, Nature's counteracting force of slope afforestation. The counteracting forces of nature, the natural plant associations, and the hitherto reasonably well balanced agricultural associations are being disrupted; soil fertility is being reduced; and whole hillsides are being lost to productive human endeavor (Ives 1979).2

2. Examples from among the earlier writings of several of participants of this conference could also serve to illustrate how pervasive and persuasive the 'Theory of Himalayan Degradation' was: 'A number of widespread practices—such as over-collection of fuel and fodder, over-grazing, shifting agriculture, and regeneration of fodder grasses through annual burning—are well known. These ecologically unsound activities are not new; but in recent years they have caused an unprecedented amount of environmental damage and now threaten to virtually destroy the Himalaya's natural resources.' (Campbell 1979, cited in Messerschmidt 1987, p. 377).

This regional scenario, with its sense of crisis and implicit call to action, has been widely reviewed and refuted (Thompson and Warburton 1985; Ives and Messerli 1989; Metz 1991). It nevertheless persists as the guiding vision for Nepal's environment. The Khumbu variant of the scenario compounds the resource demands by the local population with the added impacts of tourists, who demand fuel wood and other products (Blower 1971; Mather 1973; Mishra 1973; Lucas et al. 1974; Bjonness 1979).

It has long been assumed that the arrival of the first Sherpa settlers, who came to Khumbu perhaps 25 generations ago, initiated a process of landscape transformation that began to accelerate in the last few decades under pressure from increasing numbers of visiting tourists. The subject of the history and environmental implications of Sherpa settlement in one part of Khumbu/Sagarmatha are explored in the brilliant doctoral dissertation of Sherpa (1999). The most frequently decried manifestation of human presence is the reduction in forest, ascribed to a spectrum of factors usually including tree felling and livestock grazing:

The forests and alpine scrub-grasslands of the park have been extensively modified and severely depleted. Apart from inaccessible country or country of colder aspect, elsewhere the forest is under pressure from human exploitation and stock-grazing, causing both a deterioration and diminution of area (Halkett 1981).

For Khumbu, as for the whole Himalaya, work in the last decade has forced a re-evaluation of the crisis scenario. Repeat photography (Houston 1982; Byers 1987a), forest analysis (Stettler and MacDonald 1982 (unpublished data); Byers 1987c; Jordan 1993; Sherpa 1999), oral history (Stevens 1993) and range analysis (Brower 1987) all suggest that the condition of SNP forests has been misinterpreted. Evidence for a significant recent reduction in forest cover is absent. Yet researchers have been quite comfortable with the still-prevailing view that the Sherpa's first arrival coincided with-and probably began-a process of deforestation, and the Sherpa's cattle are uniformly targeted as agents of forest destruction.


The Sherpas' herds of yak and other livestock enjoy pride of place in most explanations of the initial removal of forest and the subsequent suppression of regeneration. The first Sherpa warden of SNP, newly returned from New Zealand with a diploma in park management from Lincoln College, was explaining the Himalayan Trust's efforts to rid the park of goats—the quintessential destructive domestic beast—told us that he thinks there is something poisonous in goat saliva. The warden may have misinterpreted the mechanism, but it's likely he understood the essence of the message taught at Lincoln College. It's a common enough message in forestry schools and environmental studies programmes, among conservationists: livestock grazing and forests don't mix (Day 1930; Dambach 1944; Patric and Helvey 1986; de Witt 1989; Dennis in preparation).

That has been the message, certainly, in nearly all attempts to come to grips with the dynamic interaction of forests and Sherpas in SNP. In almost every account promoting SNP, in plans for management of the park, and in reports by trained and lay visitors alike who comment upon the status of forests, livestock rank high as culprits in the retreat and degradation of woodland (e.g. Naylor 1970; Blower 1971; Mather 1973; Mishra 1973; Lucas et al. 1974; Speechly 1976; Bjonness 1980; Garret 1981; Jeffries 1982; von Haimendorf 1984; Byers 1987c; Hardie et al. 1987; Miehe 1987; Ledgard 1989).

The remaining forest . . . is fast disappearing and will have probably vanished altogether in a few years unless properly managed . . . The main problem . . . is that heavy grazing by yak and other livestock destroys all young tree growth through browsing and trampling and therefore prevents natural regeneration of the forest (Hardie et al. 1987).

Case study: Testing assumptions about grazing and the forest landscape

We were uneasy to lay responsibility of forest damage on Sherpa herds, in the absence of any confirming study, in light of the array of erroneous assumptions commonly made about ecological processes in the Himalaya, and in the face of conflicting preliminary work (Brower 1987). We sought to clarify the implications of livestock grazing for Khumbu's forest landscape, and set out to explore two hypotheses: (1) that forests are declining and (2) that the scrub vegetation that dominates most slopes near villages is degraded remnant forest. Evidence of declining forest would include lack of regeneration within stands, at canopy gaps, and at the edges of stands; regeneration within stands and at stand edges, and a pyramidal age structure would refute this hypothesis. Scrub vegetation composed of a subset of forest under story species plus species with 'invader' biology would be expected in shrub-grassland produced from degraded forest; floristically rich and distinctive shrubland argues against such origins. Accordingly, we designed a small-scale, exploratory study of forest plots located above the confluence of the Bhote Kosi and Imja Khola, near the Sherpa settlements of Namche Bazar, Kunde, Khumjung, and Tashinga. We conducted the study in monsoon of 1990.

The location

The triangular 'plateau' defined by these joined rivers is the area of densest population and most concentrated land use within Khumbu. It is also the place subject to most extensive speculation about the role of human occupation in eliminating forest. The slopes around these south-facing settlements are typically identified as former forest degraded to scrub through some combination of Sherpa burning, grazing, and forest felling (Halkett 1981; Hardie et al. 1987). The area is underlain by migmatitic orthogneisses and the surface topography is the manifestation of pre-glacial v-shaped troughs and platforms overprinted by glaciation and subsequent mass wasting, and relief from valley bottom to peak ranges from 2000 to 4000 masl (Vuichard 1986). Monsoon delivers yearly rainfall between June and October, some rain falls in April and May as a result of convective heating in the plains, and winter storms, though sporadic, can produce substantial snowfalls. Specific weather data are scarce, and conditions vary widely across Khumbu, but a weather station maintained at Namche Bazar (3440 masl) reported a January mean of –0.4°C (the coldest month), a July mean of 12°C, and an average annual precipitation of 1048 mm with a range from 708 to 1710 mm (Joshi 1982).

The study area is subject to grazing by a total population of about 1200 animals, maintained by 300 households and a government yak-breeding facility. Traditionally, Sherpa cattle made use of ranges extending from below main settlements to the snowline above 6000 masl. Sherpas managed their animals in such a way as to distribute the impacts of grazing over space and time, using individually-owned huts and hay-fields as a base from which to graze successively higher pastures, and discouraged by such cultural mechanisms as the di from leaving livestock too long near settlements (Brower 1987; Sherpa 1999). Tourism's need for load-carrying packstock and competition for herder labour have undermined such traditional range-management mechanisms, and today livestock remain concentrated nearer the villages most of the year. Although the villagers of Kunde and Khumjung pride themselves on maintaining their di, Namche's livestock population has grown, and depends on our study area's grazing resources. Yak, 'hilly cattle,' and yak-cattle hybrids (female 'zum'; male 'zopkio') using the area are stabled at night in the villages and driven out each morning to forage for themselves. In winter, limited feeding of hay and household scraps may supplement the diet of particularly valuable animals, but most depend primarily on grazing and browsing. According to previous studies (Bjonness 1979; Bjonness 1980; Brower 1987) and our observations, grazing pressure in the vicinity of these villages varies from moderate to severe. Six of the seven sample sites (Namche North and Komuche) are subject to grazing year-round; the others—at Tashinga and Kunde—are protected, at least in principle, by restrictions on grazing during part of the summer (Brower 1990). Each sample plot is well within the daily foraging of livestock maintained by local Sherpas. Pure cattle are a larger proportion of the stock having access to some sites (Namche North), crossbreeds dominate others (Komuche), and pure yak are a larger percentage of the total stock likely to be grazing those plots near Tashinga. These differences may be reflected in grazing effects, for the feeding behaviours and preferences of Sherpa cattle vary according to breed (Bonnemaire and Jest 1976; Brower 1996).3

3. In addition to impacts of wandering livestock, sample sites are also subject to collection of forest litter and deadfall, occasional wood-cutting, and to browsing by musk deer and tahr (a wild goat).

Within this area we chose a dispersed set of sample sites in fir (Abies spectabilis) woodlands within a few minutes walk of settlements. Sites are similar in elevation but represent a range in both grazing use conditions and levels of stand closure. All but one of our study sites lie between 3500–3680 masl in three locations, representing different combinations of slope, aspect, and level of grazing use; we designate them Tashinga, Komuche, and Namche North. We chose an additional site within a stand above the village of Kunde that had been previously identified as manifesting stand decline.4

At these elevations within Khumbu, more southerly aspects support shrub/grassland vegetation dominated by low-growing rhododendron (Rhododendron lepidotum) and cotoneaster (Cotoneaster microphyllus). Tall shrub vegetation with birch (Betula utilis), fir, and tree rhododendron (Rhododendron campylocarpum, R. campanulatum) dominate the more northerly aspects. East and southeast aspects support fir forests of various densities mixed with shrub vegetation including tall species of Rhododendron and Cotoneaster, Salix, and Rosa and many species of the rhododendron-cotoneaster shrub/grassland assemblage. We make no statistical inferences regarding differences, similarities or trends among plots; rather we treat them as an aggregate of separate case studies.


Our findings, summarised here, suggest that interplay of site factors shapes the present configuration of the forest/shrub landscapes of Sagarmatha National Park (SNP). Though each of the four study locations tells a somewhat different story, together the sites suggest that altitude, aspect, and slope steepness are the most important elements determining whether forest becomes established and subsequently thrives. Our study sites show that contemporary grazing has at most a limited effect on the regeneration of forest, significant—if at all—only when acting in conjunction with other type of stresses.


Tashinga is a gunsa (winter settlement) of small houses and associated fields, occupying an area of moderate slopes facing dominantly east-southeast. All four of these sites showed light to moderate grazing; the only evidence of wood-cutting were a few cut stumps of small trees on forest sites. The two plots on east-southeast aspects support fir forest, the oldest trees established since the 1930s. Plot-1 trees are larger, the peak period of establishment was between 1960 and 1964, and it is still a fairly open stand (Table 1). Plot 2 is denser, its trees smaller and younger, with high densities of seedlings and saplings mostly established since 1970. Height of large saplings on both sites exceeds the rates on all other sites we examined. Saplings less than 3 m tall, about 5% of Plot 2, indicate damage from browsing, but leader growth is equal in both browsed and unbrowsed trees, suggesting minimum impact. These can be characterised as actively growing Abies stands, within which human activities are having small effect.

Tashinga's north-north-east aspect supports tree cover comparable to these sites, but one-third of the trees over 1 m tall are birch rather than fir. Abies seedlings and small saplings are fairly abundant, but growth apparently stagnates, for trees up to 60 years old have diameters less than 10 cm; the oldest fir tree on the plot was established in about 1926 but is only 17 cm diameter at breast height (DBH). The more mesic conditions of this site, as of all north-oriented slopes in this elevation zone, appear to favour birch. Such sites permit the establishment of fir seedlings but forestall their continued growth.

The Tashinga plot with the most southerly aspect is a forest-free zone on the south-southeast slope, and supports rhododendron-cotoneaster shrub/grassland vegetation; a thorough search revealed a small number of small juniper (Juniperus recurva) plants but no fir or birch.


At Komuche we sampled one forest plot (Plot 5) and two shrub/grassland plots (Plots 6 and 7). Komuche lies about 1 km northeast of the village of Namche Bazar, 5 km south of the Tashinga site. The general aspect here is more southerly and the slope steeper than at Tashinga. The area shows high current grazing use (a majority of grass clumps had been grazed short) by livestock—particularly hybrids—musk deer, and an expanding herd of tahr. Grazing pressures are concentrated on reduced range resources in the area near Namche Bazar: approximately 35 of the national park's 45 hectares of fenced plantation (Ledgard and Baker 1990) are located here, the yak farm has claimed a substantial area of former hayfield and range, and accessible forage has been significantly reduced. Forest stands in this area occur in isolated positions above rocky outcrops or in ravines diagonal to the generally steep slope of the hillside. This is the setting for Plot 5, which has the same east-southeast aspect and approximate elevation as the fir forest plots at Tashinga. The oldest trees here were established in about 1945—more recently than those at Tashinga. Establishment has apparently been intermittent since, with clear peaks between 1960–65 and between 1970–75. Long-term growth rates are less than at other sites in this elevation zone, and there are few seedlings or small saplings; sapling growth is much slower than on the plot with similar canopy closure at Tashinga (Table 1). Although about 45% of the saplings and small trees show browsing damage, as at Tashinga, browsed saplings do not show a reduced growth rate relative to unbrowsed saplings. There was no evidence of woodcutting–no surprise; perhaps, since the site is close to SNP headquarters and the patrolling park rangers are housed there. Conditions on this rocky, steep and exposed site appear more hostile for seedlings and less nurturing for growing fir than Tashinga's fir plots.

One shrub/grassland plot here is very like that at Tashinga: predominantly Rhododendron lepidotum and Cotoneaster microphyllus less than 0.5 m tall. The other occupies a more northerly aspect and more closely resembles the Tashinga fir/birch site, though with only a few birches, most with multiple stems less than 3 m tall, occur on the site. These show signs of having been repeatedly browsed, although there is little evidence of browsing on current growth. Willows apparently have not been browsed and neither willow nor birch appears to have been cut.

Namche North

The sites occupying the southwest-facing slope of the Bhote Kosi valley about 0.5 km northwest of Namche, contain more well-developed forest cover than the Tashinga and Komuche sites. The site nearer Namche (Plot 8) shows somewhat heavy grazing use, and the one farther on (Plot 9) shows light-moderate use. We saw crossbreeds at these sites but found no sign of musk deer or tahr.

Plot 8 had the largest and oldest trees of all the plots in this elevation zone—up to 77 rings and 36 cm DBH, implying trees established around 1914 (Table 1). This site showed a peak of establishment in 1956–58. Plot 9 had more complete canopy closure and more trees in larger size classes. Maximum age and size were, however, somewhat less than at Plot 8. The oldest tree on Plot 9 was established in about 1936 and measured 29 cm DBH. Long-term growth rates were high in terms of the sites considered here—similar to those at the Tashinga fir sites and considerably higher than at Komuche. Sapling growth, however, was quite slow. Both plots had low numbers of small saplings but very high numbers of seedlings. Low recruitment of seedlings into the sapling class and slow current sapling growth may be a response to shading by the fairly closed canopy of these sites.

A separate case: Kunde

We sampled a 10th plot on the hill above Kunde at 4020 masl on a southwest-facing slope (Table 1). This area is a gently sloping, smooth herbaceous sward, in contrast to the steep terraced topography of all other sites in the study. This site had the largest and oldest trees (up to 43 cm DBH, established as long ago as 1890) but the stand was very sparse with only 16 percent canopy closure (Table 1). There were few seedlings. Over half of the individuals in the small sapling class were dead, and there was abundant evidence of severe browsing damage. Sapling growth estimates have limited meaning since they are based only on the few live individuals in this size class, but are certainly very low. Long-term growth rates of the larger trees have also been very slow. In contrast to other sites, cut stumps were obvious. Most of these were less than 5 cm in diameter, but there were some stumps of larger trees and holes where large stumps had been removed in the vicinity of the plot. There was little shrub cover at this site, and species composition had little resemblance to other sites.

Table 1. Number of individual trees within size classes found in each of the sample sites.


Number of individuals/1000 m2


(Fir sites)

Tashinga (Birch)



North Sites



Year established              

Since 1976




































Before 1931








Oldest tree
















50 (+43 dead)


Saplings (to 5 cm DBH)







28 (+60 dead)


Small trees (to 20 cm)









Large trees (>20 cm)









Largest tree (cm DBH)








(Plots within sites are averaged).


Kunde clearly shows evidence of forest decline: few seedlings, high mortality of saplings, and few trees in intermediate size and age classes. Low diameter growth rates over the lives of the few large trees show that conditions for growth at this site are poor even for individuals in size classes that are little affected by current browsing and trampling.

However, browsing and tree cutting have every manifestation of being major factors in mortality of small trees. Clearly, unless conditions improve for survival and growth of small trees, the large trees on this site are unlikely to be replaced and the stand will thin even further or disappear. In spite of apparently heavy grazing, vegetation cover at this site is complete and there is clearly no accelerated soil loss-taking place. This, together with observations by Byers (1986) at similar sites, suggests that it is unlikely that livestock impacts on soil productivity are a major factor in the decline of the forest at this site.

The under-story community here, though equally diverse, shares relatively few species with any of the sites at lower elevations. All of the shrub species that form major components in both the lower elevation forest understory and shrub/grassland sites are absent here, as are all the grasses and herbs that are major vegetation components on those sites. Clearly, many plant species typical of fir forests and associated communities reach their environmental limits below the elevation of this site, and this site is clearly at the uppermost limit for fir. Timberlines fluctuate, and the older trees on this site may have been established under environmental conditions more favourable than those that exist today. Damage from grazing and woodcutting may not be the only limiting factor operating here.

Fir forests at Tashinga and Namche North, in contrast, do not have characteristics we expect in declining stands. On the contrary, what we see at these sites is consistent with young stands at various stages of filling in. While at Kunde cut stumps and holes where stumps have been dug out were obvious, at these sites we did not find any evidence to suggest that tree cutting in recent years has reduced the number of large trees. Rather, the oldest trees here appear to have been pioneers on a previously unforested site. Composition and floristics of the under-story vegetation support the interpretation of these as sites that were shrub/grassland with scattered trees rather than more dense forest vegetation in the recent past. On all these sites, the understories bear strong resemblance to nearby shrub/grassland. The herbaceous understory differs from the shrub/grassland by absence of individual common species, a different set in each case, rather than by presence or absence of a common set of species or, alternatively, presence of species characteristic of more fully developed fir forests elsewhere in these valleys (see Byers 1986 for lists of these species). The under-story shrub component does show signs of acquiring a distinctive set of species not characteristic of shrub/grassland—the tall shrub/small tree species of Rhododendron, Salix, Cotoneaster and Rosa. At present, these species are minor components of cover (less than 5%) on these sites.

The Komuche site presents a somewhat less certain picture. Looking only at trees over 1 meter tall, we see a young, expanding stand, the pioneers of which became established in the late 1940's. However, populations of seedlings and small saplings are relatively small. Sapling numbers and growth rates resemble the rates in closed canopy at Namche North rather than the rapid rates of open stands at Tashinga. Growth rate of large trees is lower than at either Tashinga or Namche North. Direct or indirect effects of livestock and tahr, both of which use this area heavily, could well be major factors in low seedling numbers. However, contrary to the Kunde case, it is not clear that browsing is directly causing either reduced growth or mortality for saplings.

As with Tashinga and Namche North, understory characteristics support an interpretation of this as a young stand, not a depleted remnant of a formerly more dense forest. Intermittent establishment appears to have been the pattern over the history of the existing stand. Animal impacts may interact with environmental factors to limit regeneration, except in years with favourable combinations of both the animal and environment. It is certainly possible that recent changes in livestock use in this area could reduce the frequency at which those favourable circumstances occur. Trees now in the sapling stage can reasonably be expected to survive and produce some filling in of the stand, but it is not clear if further stand development is likely to occur. Site 3, the north-facing stand near Tashinga, shows that poor conditions for growth, not grazing, are probably the limiting factor in expansion of fir forest onto northerly aspects. Although the mesic environment of these sites is apparently favourable for seedling establishment and initial growth, it is clearly not favourable for continued growth. It is highly unlikely that livestock trampling or browsing would be limiting growth in this size class without also affecting smaller trees. The shift in understory species composition between these sites and adjacent more southerly aspects shows that other species also reach environmental limits on these shady slopes. It is striking that no fir seedlings were found in shrub/grassland plots. Plots had vegetation similar in structure and composition to the fir stand understories and there were no apparent differences in level of livestock use, density or condition of trails and terracettes, or level of herbage removal between adjacent scrub/grassland and forested sites. We do not have specific information on seedling physiology at such sites for Abies spectabilis, but we do for a number of other Abies species. All the North American Abies species are shade tolerant and sensitive to heat and desiccation at seedling stage. For these Abies species, seedling survival is greatly enhanced by shading, although established trees tolerate and grow more rapidly in full sun (Burns and Honkala 1990). If Abies spectabilis is similar, high seedling mortality in full sunlight may largely account for the scarcity of seedlings on the scrub/grassland sites. This would fit with several other patterns we have observed: the delay between establishment of the first cohort on a site and subsequent population expansion; abundance of seedlings on the Tashinga birch/fir site; the Namche North sites where conditions for subsequent growth are poor at the present time; and the scarcity of seedlings on the steep, exposed Komuche site.

Given the evidence for stand dynamics at these sites, which among them represent a range in altitude, aspect, slope and livestock-use variables, we suggest the following process:

  1. Establishment of fir seedlings in shrub/grassland is a rare event because seedlings are vulnerable to desiccation. Establishment events are likely to be more frequent on shadier aspects, on sites protected from wind, and near seed sources.
  2. Past the critical establishment phase, shrub grassland sites are suitable for fir growth.
  3. Because of the relative unpalatibility of fir and its rapid ability to replace damaged leaders, moderate browsing by livestock and other herbivores does not result in major foliage losses or retard growth of trees that are otherwise healthy.5
  4. Once pioneer trees grow big enough to produce a shaded, protected area beneath, seedling establishment increases rapidly.

5. This may not be true under conditions of heavy browsing on plantation-grown Abies by native browsers such as tahr, as evidenced by substantial reduction in fir shoots relative to pine in plantations above Namche recorded this summer.

Closing remarks

Our results show that concern about livestock impacts on fir regeneration in Khumbu is justified in some situations, but on the whole fir forests in the Namche-Kunde-Khumjung region are expanding and livestock grazing has little indirect adverse effect on tree regeneration and growth. Regeneration problems do occur where environmental limits to growth interact with browsing pressure near timberlines. However, we found that there was ample regeneration on sites representing a range of grazing conditions, from relatively moderate, controlled use near Tashinga to fairly heavy and less controlled use within half a kilometres of Namche Bazar. Observations of vigorous fir regeneration on very heavily grazed sites at 3800 masl just south of Kunde (in an area of much recent tree cutting) further support our conclusion that grazing alone would rarely prevent fir regeneration. Our findings on age structure, vegetation composition, and floristic characteristics all support this conclusion. We see the relationships between human impacts and forest stand dynamics hinging on the particular biology of Abies spectabilis in this location, especially its sensitivity to desiccation at the seedling stage, its relative unpalatiblity to large herbivores, and its tolerance of moderate browsing when otherwise healthy. We would not necessarily expect to see similar relationships with species that differ from Abies spectabilis in these characteristics.

Our conclusions agree with those drawn by Houston (1982) and Byers (1987a, b, c): forest cover is increasing in the 3400–3800 masl zone in the Namche triangle. Sparse stands and lone trees in this zone are pioneers, not remnants of formerly more dense forests, as researchers in the field speculated (Hardie 1974; Speechly 1976; Bjonness 1979; Bjonness 1980; Halkett 1981; Bjonness 1983; Miehe 1987).

We would like to make a further point regarding the shrub/grassland vegetation, characterised as 'degraded' by these authors. Our results show that Rhododendron lepidotum-Cotoneaster microphyllus shrub/grassland is a rich and diverse plant community in its own right. It is in no way a collection of forest species that have clung to life in spite of overstory removal and grazing, nor is it composed of pioneering species. It is clear that boundaries between shrub/grassland and forest have shifted over time on the sunny slopes of Khumbu. In the long term, each has repeatedly been 'secondary' to the other. The fact that human activities have probably had a major role in some of these shifts does not inherently make the community type favoured by human activity 'degraded'. We can in fact recognise degraded conditions in either shrub/grassland or forest and we can find examples of both in Khumbu (the steep eroding trails and heavily trampled areas above Namche, for example). We recognise degraded conditions as ones where accelerated soil loss is occurring, or where species characteristic of the community are missing. In contrast, intact shrub/grasslands like the ones we examined in this study and observed elsewhere in Khumbu are not undergoing soil loss and provide habitats for a wide range of native species that do not occur in forests. Attempts to 'restore' intact shrub/grasslands to forest cover are counterproductive to maintaining the diversity of habitats of Khumbu and the full array of native biota these habitats support.

Implications for conservation

Our study suggests that ecology, to some extent, is in the eye of the beholder. Foresters, park planners and others convinced of the dangers of Himalayan deforestation, and unkindly inclined toward livestock, saw decline in the patterning of Khumbu forests, and agents of degradation in Sherpa herds. They somehow overlooked what would seem to be clear evidence of forest vitality (distracted, perhaps, by abundant evidence, such as that provided by ubiquitous stock-terrecettes, that in this landscape livestock have clearly been agents of other landscape change).

More partial to yak, leery of conventional explanations for ecological processes in the Himalaya, and with experience of unexamined prejudices against forest grazing, ours was a different perspective: we set our questions to test it, and our study offers a different result. It is, clearly, an exploratory study, and raises no end of further questions about environmental processes (e.g. what conditions and processes impede stand establishment? What factors explain the varying dates for the apparent pioneering of the stands we sampled? What is the role of natural and human-initiated fire in this area?). Cultural practices (e.g. Do Sherpa livestock management practices influence the degree of grazing impact? How closely do contemporary grazing patterns reflect past practices? How do long term residents account for the patterns of forest establishment and maintenance that we observe?). And park policy objectives (e.g. is the vision of a forested Khumbu drawn from nature or from outsiders' expectations? Is plantation forestry appropriate given successful natural regeneration? Do we know enough to meddle with the environment and local culture? How can ecological understanding be separated from prejudicial training and values?). We await with interest the results of future work on some of these and other questions, and brace ourselves for the critique we hope will follow on our own attempt to make sense of the landscape of Khumbu. The outcome we would hope for is a conservation strategy for SNP that reflects a thorough and thoroughly neutral understanding of both ecological processes and the changing role of local people through time, one that recognises the mutability of scientific explanations and the legitimacy of local needs and values. Within the incredibly complex geopolitical-economic framework within which decisions about SNP and the world's other protected areas are made, let science play a central role-but always in the awareness of its limits.

Sagarmatha, of course, is not unique in requiring understanding of both natural and cultural processes as the foundation for conservation planning and management. The landscape valued here, a high-altitude, peopled foreground to the world's biggest mountains, has its own distinctive history, and reflects responses to a very particular set of conditions. The inherent instability of high mountains presents one set: an ever-changing interaction of climate, slopes, and biota. People and their yak introduce another set of influences on landscape processes that change along with the lifeways of Khumbu's residents, and interventions of Sagarmatha's managers. The processes and patterns that characterise this particular Himalayan landscape are not, after all, part of some regional scheme. Yet for all its distinctiveness, Sagarmatha/Khumbu's forest landscape is a product of the same broad dynamics of establishment and decline that shape landscapes everywhere. Understanding the particular expression of those dynamics, here as anywhere, requires a suspension of presumptions, a careful analysis of a complex spectrum of factors, and a willingness to re-examine interpretations of landscape-making derived from inadequate understanding of dynamism, complexity, and untested assumptions.


Adams V. 1996. Tigers of the snow and other virtual Sherpas: An Ethnography of Himalayan Encounters. Princeton University Press, Princeton, USA.

Bjonness I-M. 1979. Impacts on a high-mountain ecosystem: Recommendations for Action in Sagarmatha (Mt. Everest) National Park. Unpublished report for Sagarmatha National Park.

Bjonness I-M. 1980. Animal husbandry and grazing, a conservation and management problem in Sagarmatha (Mt. Everest) National Park, Nepal. Norsk Geografisk Tiddscrift 34:59–76.

Bjonness I-M. 1983. External economic dependency and changing human adjustment to marginal environment in the High Himalaya, Nepal. Mountain Research and Development 3(3):263–272.

Blower J.H. 1971. Proposed National Park in Khumbu District. Memorandum, Wildlife Conservation Officer, FAO, to His Majesty's Government's Secretary of Forests, March 21, 1971.

Bonnemaire J. and Jest C. 1976. Le yak; son rôle dans la vie matérielle et culturelle des éleveurs d'Asie Centrale. Paris, Société d'Ethonozootechnie. Ethnozootechnie, no. 15, Paris, France, 170 pp.

Brower B. 1982. The problem of people in natural parks: Chitwan and Sagarmatha National Parks, Nepal. University of California, Berkeley, USA (Unpublished paper).

Brower B. 1983. Mountain hazards and the people of Khumbu-Pharag. Report, Nepal Mountain Hazards Mapping Project.

Brower B. 1987. Livestock and landscape: The Sherpa pastoral system in Sagarmatha (Mt. Everest) National Park, Nepal. PhD thesis, Department of Geography, University of California, Berkeley, USA.

Brower B. 1990. Range conservation and Sherpa livestock management in Khumbu, Nepal. Mountain Research and Development 10(1):34–42.

Brower B. 1991a. Sherpa of Khumbu: People, livestock, and landscape. Oxford University Press, Delhi, India.

Brower B. 1991b. Crisis and conservation in Sagarmatha National Park. Society and Natural Resources 4:2.

Brower B. 1993. Co-management vs. co-option: Reconciling scientific management with local needs, values and expertise. Proceedings, conservation and development: Bottom-up strategies for the roof of the World. Yale School of Forestry and Environmental Studies, New Haven, USA.

Brower B. 1996. Geography and history in the Solukhumbu landscape. Mountain Research and Development 16:3.

Burns R.M. and Honkala B.H. 1990. Silvics of North America. Volume 1, conifers. Agriculture Handbook 271. USDA (United States Department of Agriculture), Washington DC, USA. 675 pp.

Byers A. 1986. A geomorphic study of man-induced soil erosion in Sagarmatha (Mt. Everest) National Park, Nepal. Mountain Research and Development 6(1):83–87.

Byers A. 1987a. An assessment of landscape change in the Khumbu Region of Nepal using repeat photography. Mountain Research and Development 7(1):77–81.

Byers A. 1987b. Landscape change and man-accelerated soil loss: The case of Sagarmatha (Mt. Everest) National Park, Nepal. Mountain Research and Development 7(3):209–216.

Byers A. 1987c. A geoecological study of landscape change and man-accelerated soil loss: The case of Sagarmatha (Mt. Everest) National Park, Nepal. PhD thesis, Department of Geography, University of Colorado, Boulder, USA.

Dambach C.A. 1944. A ten-year ecological study of adjoining grazed and ungrazed woodlots in north-eastern Ohio. Ecological Monographs 14:257–270.

Day R.K. 1930. Grazing out the Birds. American Forests 36:555–557, 594.

Dennis A. Fir reproduction in subalpine forests of Sagarmatha National Park (in preparation).

Dennis A. and Brower B. 1995. Forest reproduction in subalpine forests of Khumbu. Paper presented to 24th Annual Conference on South Asia, Madison, Wisconsin, USA.

Drew I.K. and Sharma U. 1977. A proposal for a reforestation programme in the Sagarmatha (Mt. Everest) National Park. Report, Sagarmatha National Park, Kathmandu, Nepal.

Garret K. 1981 Sagarmatha National Park Management Plan. Department of Lands and Surveys, Wellington, New Zealand.

von Haimendorf C.F. 1964. The Sherpas of Nepal. University of California Press, Berkeley, USA.

von Haimendorf C.F. 1975. Himalayan traders: Life in highland Nepal. John Murray, London, UK.

von Haimendorf C.F. 1984. The Sherpas transformed: Social change in a Buddhist society of Nepal. Sterling Publishers, New Delhi, India.

Halkett L.M. 1981. Forest management in Sagarmatha National Park. Report by New Zealand Project Forester.

Hardie N. 1974. The Sherpa people. Manuscript (New Zealand Mission Report).

Hardie N., Benecke U., Gorman P. and Gorman P. 1987. Nepal–New Zealand Project on Forest Management in Khumbu–Pharak. Forestry Research Center, Christchurch, New Zealand.

Hillary E. 1964. Schoolhouse in the clouds. Hodder & Stoughton, London, UK.

Houston C. 1982. Return to Everest: A sentimental journey. Summit 28:14–17.

Ives J. 1979. Applied mountain geoecology: Can the scientist assist in the preservation of mountains? In: Webber P. (ed), High Mountain Geoecology. Westview press, Boulder, Colorado, USA.

Ives J. and Messerli B. 1989. The Himalayan dilemma: Reconciling development and conservation. Routledge, London, UK.

Jeffries B.E. 1982. Sagarmatha National Park: The impact of tourism in the Himalayas. Ambio 11(5):274–281.

Jordan G. 1993. GIS modeling and model variation of erosion and deforestation risks, Nepal. MSc thesis, School of Agricultural and Forest Sciences, University of Wales, UK.

Joshi D.P. 1982. The climate of Namche Bazar. Mountain Research and Development 2:4.

Ledgard N. 1989. Nepal (Khumbu/Pharak) visit-April 1989: Forestry notes for the Himalayan Trust. Forest Research Institute, Mountainland Forestry Unit, Christchurch, New Zealand.

Ledgard N. and Baker G. 1990. Nepal (Solu-Khumbu) Visit-June/July 1990: Forestry Notes for the Himalayan Trust. Forest Research Institute, Mountain Land Forestry Unit (Contract Report FWE 90/OSI), Christchurch, New Zealand

Lucas P.H.C., Hardie N. and Hodder R.A.C. 1974. Report of New Zealand Mission on Sagarmatha (Mt. Everest) National Park, Nepal. Report prepared for Ministry of Foreign Affairs, Wellington, New Zealand.

Mather A.D. 1973. Reforestation in the proposed Mt. Everest National Park. Report, Sagarmatha National Park, Kathmandu, Nepal.

Messerschmidt D. 1987. Conservation and society in Nepal: Traditional forest management and innovative development. In: Little P.D. and Horowitz M.M. (ed), Lands at risk in the third world: Local level perspectives. Westview Press, Boulder, Colorado, USA. pp. 373–397.

Metz J. 1991. A reassessment of the causes and severity of Nepal's environmental crisis. World Development 19(7):805–820.

Miehe G. 1987. An annotated list of vascular plants collected in the valleys south of Mt Everest. Bulletin British Museum Natural History (Botany) 16(3):225–268.

Mishra H. 1973. Conservation in Khumbu: The proposed Mt. Everest National Park. Confidential Report, Department of Wildlife and National Parks, Kathmandu, Nepal.

Naylor R. 1970. Colombo Plan Assignment in Nepal. Report, New Zealand Forest Service, Wellington, New Zealand.

Patric J.H. and Helvey J.D. 1986. Some effects of grazing on soil and water in the Eastern Forest. USDA (United States Department of Agriculture) Forest Service General Technical Report, NE-115, Washington, DC, USA. 24 pp.

Sherpa L.N. 1979a. Considerations for management planning of Sagarmatha National Park. Dissertation, Diploma in Parks and Recreation, Lincoln College, University of Canterbury, New Zealand.

Sherpa L.N. 1987. Social functions of Rara National Park: A case study of park and people interaction at Rara National Park, Nepal (Unpublished manuscript).

Sherpa L.N. 1988. Conserving and managing biological resources in Sagarmatha (Mt. Everest) National Park, Nepal. East-West Centre Environment and Policy Institute, Working Paper 8. Honolulu, Hawaii, USA.

Sherpa L.N. 1999. Effects of forest/land use on structure and distribution patterns of high-altitude forests: A case study of Nangpa and Hinku valley, Northern Solukhumbu. PhD thesis, Forestry, University of Washington, Seattle, USA.

Sherpa M.N. 1982. Preserving a Mountain Heritage. National Geographic 161(6):696–702.

Sherpa M.N. 1993. Grass Roots in a Himalayan kingdom. Indigenous peoples and protected areas. Sierra Club, San Francisco, USA.

Sherpa N.W. 1979b. A report on firewood use in Sagarmatha National Park, Khumbu Region, Nepal. Report, HMG Department of Wildlife and National Parks, Kathmandu, Nepal.

Speechly H. 1976. Proposal for forest management in Sagarmatha National Park. Report, HMG Department of National Parks and Wildlife Conservation, Kathmandu, Nepal.

Stevens S.F. 1983. Tourism and change in Khumbu. Bachelor's thesis, Geography Department, University of California, Berkeley, USA.

Stevens S.F. 1989. Sherpa settlement and subsistence: Cultural ecology and history in Highland Nepal. PhD thesis, University of California, Berkeley, USA.

Stevens S.F. 1993. Claiming the high ground. University of California Press, Berkeley, USA.

Tengboche Rinpoche and Klatzel F. 2000. The stories and customs of the Sherpas. Mera Publications, Kathmandu, Nepal.

Thompson M. and Warburton M. 1985. Decision-making under contradictory certainties: How to save the Himalayas when you can't find out what's wrong with them. Journal of Applied Systems Analysis 12:3–34.

Thompson M., Warburton M. and Hatley T. 1986. Uncertainty on a Himalayan scale. Milton Ash Editions Ethnographica, London, UK.

Vuichard D. 1986. Geological and petrographical investigations for the Mountain Hazards Mapping Project, Khumbu Himal, Nepal. Mountain Research and Development 6:(1):41–51.

de Witt B. 1989. Forest grazing hurts. Missouri Conservationist 50:18–19.

Previous PageTop Of PageNext Page