Pastoralist Decision-Making on the Tibetan Plateau


Despite a growing body of research about rangeland degradation and the effects of policies implemented to address it on the Tibetan Plateau, little in-depth research has been conducted on how pastoralists make decisions. Based on qualitative research in Gouli Township, Qinghai province, China, we analyze the context in which Tibetan herders make decisions, and their decisions about livestock and pastures. We refute three fundamental assumptions upon which current policy is premised: that pastoralists aim to increase livestock numbers without limit; that, blindly following tradition, they do not actively manage livestock and rangelands; and that they lack environmental knowledge. We demonstrate that pastoralists carefully assess limits to livestock holdings based on land and labor availability; that they increasingly manage their livestock and rangelands through contracting; and that herding knowledge is a form of embodied practical skill. We further discuss points of convergence and contradiction between herders’ observations and results of a vegetation analysis.

[The government] should try hard to change these concepts in traditional pastoralism: judging wealth by livestock numbers, perceiving rangeland as free resources, using [rangeland] without limit, and the unwillingness to slaughter or sell.

-Committee for Population, Resources, and Environment, Chinese People’s Political Consultative Conference, 2008Footnote 1

Many Tibetan herdsmen believe that the innumerable sheep and cattle crowding the range are blessings from Buddha, but many are now [because of government restrictions] realizing that less is more. -- Xinhua News, 2013.Footnote 2


A growing body of literature addresses the causes of grassland degradation on the Tibetan Plateau and the effects of various rangeland management policies that have been implemented in response. The dominant view of Chinese policy makers as well as remote sensing scientists in the Chinese academy is that degradation is caused largely by irrational management and overgrazing beyond carrying capacity, and by the burrowing and herbivory of plateau pikas (Ochotona curzoniae) (Du et al. 2004; Du et al. 2012; Pan et al. 2015; Wang et al. 2013; Wei and Chen 2001). Policy responses, including the Rangeland Household Responsibility System and tuimu huancao (“retire livestock, restore rangeland”), have been grounded in a tragedy of the commons model that presses for the privatization and division of pastoral use rights to smaller scales of management – increasingly to the household. However, studies have demonstrated negative social and environmental effects related to the reduction of flexibility and mobility, including more concentrated trampling and grazing, as well as greater susceptibility to livestock loss during snowstorms (Bauer 2005; Bauer and Nyima 2010; Cao et al. 2011a; Cao et al. 2011b; Cao et al. 2013; Gongbuzeren et al. 2015; Harris 2010; Li 2012; Yan et al. 2011; Yan and Wu 2005; Yan et al. 2005; Yeh 2009; Yeh 2013; Yeh et al. 2014).

In addition to division of pasture use rights, other policy responses have included mass pika poisoning campaigns, despite the pika’s status as a keystone species for biodiversity (Smith and Foggin 1999), and regulations to limit household livestock holdings to government-established carrying capacities. Herders are largely opposed to poisoning because of their Buddhist stance on the mass taking of life, and because they note that pika numbers bounce back quickly from poisoning. Moreover, recent studies have found that extermination campaigns negatively affect predator abundance as well as hydrological functioning (Badingquiying et al. 2016; Wilson and Smith 2015).

Regulation of livestock numbers is similarly problematic in practice. Nyima (2015) demonstrates that the purportedly scientific determination of carrying capacity on the Tibetan Plateau is not only plagued by technical problems, but also often determined more by political-economic incentives than ecological considerations. Moreover, in some areas the official carrying capacity is at or below what is considered the household poverty level. However, to date, there are few places on the Tibetan Plateau where these limits on household livestock numbers have been strictly enforced. Destocking has instead taken place most dramatically in the Sanjiangyuan area of Qinghai province (a region encompassing the sources of the Yangtze, Yellow and Mekong Rivers) through “ecological migration,” a program that moves herders completely off the grasslands to settlements in distant towns, which had led to socio-cultural dislocation without attendant evidence of grassland improvement (Bauer 2015).

In critiquing the flawed assumptions of these various policies, Harris (2010:8) notes that “most Chinese biological research has not asked, much less answered, questions regarding human motivations among the pastoralists using the rangelands of the [Tibetan plateau], but this has not kept many authors from suggesting simple reductions in livestock numbers or dramatic changes in livestock production systems.” Social science research, for the most part, has also not focused on understanding pastoralists’ decision-making about their pastures and livestock (but see Bessho (2015) on decisions to leave herding and move to town). Addressing this lacuna, this paper draws on in-depth interviews with Tibetan pastoralists to explore their practices and motivations in relation to current policies. At their core, these policies presume that transhumant pastoralism is an anachronism whose practitioners must be modernized in order to restore grassland condition. More specifically, as demonstrated in the epigraphs above, they are premised upon the assumptions that Tibetan pastoralists (1) aim to increase their livestock numbers without limit; (2) act according to tradition and thus do not actively manage or make sound decisions about their livestock and rangelands; and (3) are not knowledgeable about their environments. Following a description of our methods and study site, the paper demonstrates that these assumptions are untenable. In the first main section, on livestock numbers and management, we address the first two assumptions. The last section addresses herders’ environmental knowledge and its relationship to findings from ecological science.

Methods, Study Site, and Context

Our analysis is part of a larger interdisciplinary study conducted from 2009 to 2014 in Village Five of Gouli Township, Dulan County, Qinghai Province, China (Fig. 1), home to 175 residents in 37 households, most of whom are engaged primarily in pastoralism. Village Five occupies relatively high elevation pastures within Gouli (between 4100 and 4900 m, with vegetation sparse above 4700 m), and had been used as summer and transitional (spring/fall) pastures before collectives were dismantled in 1983. Pastoralists now use the lower areas as winter pastures, to which they move in mid-October and stay until mid-June, when they leave for higher spring/fall and summer pastures. The implementation of the Rangeland Household Responsibility System in 1996 allocated specific winter pastures to each pastoralist household on long-term leases (for details see Yeh and Gaerrang 2011).

Fig. 1

Study site. Map by Galen Maclaurin

In addition to annual surveys of livestock numbers from 2009 to 2012, we conducted semi-structured interviews and participant observation with 17 households in 2009, 2010, and 2014. The interviews, most of which were conducted by native Tibetan researcher Gaerrang, focused on village and household rangeland and livestock management history, household socioeconomic and demographic information, herders’ understandings of the relationship between livestock and rangeland condition, snowstorms, daily herding practices, livestock sales, identity, and household aspirations and definitions of success. The broader study also included an exclosure experiment (Harris et al. 2015), and non-destructive annual vegetation sampling for ground cover and species composition on 317 permanent plots in eleven winter pastures from 2009 to 2012 (Harris et al. 2016).

Pastoralists, like all people, make everyday decisions within specific political-economic, environmental, and socio-cultural contexts. In Gouli, Tibetan herders maneuver within China’s authoritarian state capitalist institutions and forces, which have produced an increasingly marketized environment, as well as development projects that subsidize new houses for pastoralists in town, aimed at enticing herders to urbanize and give up pastoralism; compulsory education and school consolidation policies; attempts (thus far unsuccessful) to organize herders into cooperatives; and rapid construction of transportation infrastructure. In contrast to many areas of Qinghai, caterpillar fungus (Ophiocordyceps sinensis) is not found here and is thus not an important source of income for Gouli households. Limits on household herd size were announced in 2004, but have not been strictly enforced. As a way of maintaining flexibility in the face of the quasi-privatization of pastures through the Rangeland Household Responsibility System, Gouli pastoralists have, over the last 15 years, developed an increasingly complex system of sub-leasing their pastures, paying rental fees to graze their livestock on others’ pastures, providing labor for herding, and contracting livestock out to other herders. The increasing prevalence of grassland leasing has been recently documented in other Tibetan areas as well (Li 2012; Levine 2015). Whether such arrangements decrease (Li 2012) or increase (Yeh and Gaerrang 2011; Levine 2015) inequality is dependent on the leasing arrangements developed in specific locations. However, the system in Gouli appears distinct insofar as pastoralists lease not only land but also livestock.

Though herders in Gouli do not have a uniform view on trends in grassland condition, there is widespread agreement that livestock weights have decreased over time. Like Tibetans elsewhere on the eastern Tibetan Plateau (Byg and Salick 2009), herders in Gouli note significant changes in precipitation patterns, stating for example: “Many years ago, before the rain it was foggy and damp so grasses grew quickly… Now in the summer the rainfall is rare but sometimes there are sudden rainfalls and hail, which may destroy grasses and break plants.”

Culturally, an important trend since the early 2000s has been the growing influence across the eastern Tibetan plateau of the slaughter renunciation movement, led by charismatic Buddhist teachers from the Larung Gar Buddhist Academy in Serthar (Seda), Sichuan province. As part of a Buddhist ethical reform movement, Tibetan religious teachers have encouraged herders to take vows not to sell livestock to slaughterhouses, in order to prevent the severe suffering engendered by the ways the livestock are transported and slaughtered (Gaerrang 2011; Gaerrang 2015). This movement has had limited influence in Gouli, where herders have not taken such oaths. However, some are familiar with the teachings and are reluctant to sell livestock directly to slaughterhouses, preferring to sell livestock for non-slaughter purposes or, in some cases, to middlemen. Importantly, this is very much a Buddhist modernist movement, an articulation of how to be ethical, modern, and Tibetan in the twenty-first century, rather than an unchanged tradition, as state discourse tends to represent Tibetan herders’ actions and ideas (Gaerrang 2011, 2015; Gayley 2013, 2016; Kabzung and Yeh 2016).

The Simplistic View of Livestock as Wealth

Explanations about how pastoralists decide herd size, movement, and grazing strategies often fail to take seriously the complexity of their own goals and strategies (McCabe 2004). Instead, herders around the world have long been portrayed as having irrational cultural norms that give status and prestige to owners of large herds (Doran et al. 1979; Herskovits 1926). The epigraphs that begin this paper are typical of Chinese representations of Tibetan pastoralists as overly conservative and single-minded in their accumulation of too many livestock, a “backward” trait that must be overcome through development (Yundannima 2012).

Social and ecological evidence has shown these models to be flawed. In northern Kenya, McPeak (2005a, b) found that herd sizes are rational at both the household and collective levels. Household income increased with herd sizes, and wealth in livestock offered a higher rate of return than other available forms of formal savings, even with periodic herd losses taken into account. Degradation resulted from suboptimal spatial distribution of herds rather than herd sizes. Similarly, McCabe’s (2004) study refutes the long-held notion that herds are an end in and of themselves for pastoralists. Instead, among Kenya’s Turkana pastoralists, livestock herds are primarily a means to form families, a process which itself must be understood in cultural context. Moreover, herders do not sell more because they do not receive fair rates of exchange for livestock, marketing infrastructure is lacking, and there are few opportunities to make investments with greater rates of return (ibid). Furthermore, larger herd sizes have been found to be an efficient risk reduction strategy for nomadic pastoralism (Naess and Bardsen 2010; Roth 1996).

Based on research in Nagchu Prefecture, Tibet Autonomous Region, Nyima (2014) refutes the Chinese state discourse of livestock as a “symbol of wealth” for Tibetan herders on three grounds. First, he argues that outsiders may have an inflated view of livestock numbers because they do not understand the number of years needed before livestock can be used either for production or sale. Second, he argues that a larger number of livestock acts as a form of insurance against the probability of a devastating loss of the herd as a result of density-independent mortality from severe snowstorms (Nyima 2014; Yeh et al. 2014). Finally, he finds that households with larger herds have a higher standard of living, with greater access to health care, education, meat and milk for consumption, and participation in religious practices such as pilgrimage. In short he argues that the number of livestock that herders keep is economically rational.

Multi-Faceted Decision-Making about Herd Size in Gouli

Nyima’s (2015) arguments are directly relevant to Gouli, where households with larger herds have noticeably higher standards of living in the form of better housing and vehicles. Herders make decisions about livestock purchase and sales based on their needs to pay for school tuition, medical expenses, religious expenses, transportation, subsidized (but not free) housing from state development projects, and other everyday expenses in an increasingly monetized economy. In doing so, they weigh options (and labor opportunity costs) to engage in other forms of work, such as wage labor and petty sales, against the income that can be earned from animal products. These decisions are also conditioned by the extent to which they have been influenced by the slaughter renunciation campaign, and their desire (or lack thereof) to maintain a pastoralist identity by retaining a tie to the land.

Beyond the parameters of Nyima’s (2015) critique, we contribute here another argument against the assumption that herders maximize their livestock herds without limit: far from focusing exclusively on increasing livestock numbers, pastoralists in Gouli engage in complex assessments of land availability and quality as well as considering labor availability when making decisions about herd sizes. Because sheep must be followed more closely than yaks, households with less available labor power tend to keep fewer sheep. As Pastoralist LG explained, “I take labor into account when deciding whether to increase or decrease [my livestock numbers]… If there are not enough people to work then pastoralists will only herd yaks. If labor is available, they will herd both yaks and sheep. Now, my grandson only herds yaks and my son-in-law herds the sheep.” Another herder, Pastoralist L, explained, “during the winter, I divide the weaker sheep and herd them separately,” a strategy that requires either having multiple laborers within the household or hiring outside labor.

Pastoralists also consider weather and grassland conditions when assessing the number of each type of livestock that their pastures are capable of supporting. This in turn informs their decisions to sell or trade livestock, sublease additional land, or contract out part of their herds to others. Pastoralist DT explained, “I plan according to the weather. If it rains, I will plan to increase my number of livestock. If the weather isn’t good and I increase the livestock number, that won’t help.” Similarly, herder RC noted, “If the weather is good and there is rain, I will buy more livestock. If it is not, I will sell livestock.” Pastoralist LG explained, “I haven’t tried to increase or decrease my livestock numbers [in the last few years]. I know how many [sheep and yaks] I can herd based on the grassland condition and not more than that. Each year is different. I consider if there is enough grass each year. If there isn’t enough, I will sell some livestock and sheep because there isn’t enough space available on the grassland...Even if the condition is better than previously, I would not make a rushed decision. I would wait until there has been at least one month of good weather” before considering increasing herd size through purchase. These statements about the importance of temperature and precipitation for vegetation dynamics are supported by a number of studies of the Tibetan Plateau that demonstrate both temperature and precipitation are important, often interacting, drivers of vegetation dynamics (Berdanier and Klein 2011; Chu et al. 2007; Sun et al. 2013; Zhang et al. 2010), though our data in Gouli did not find spring/summer precipitation to be an important predictor of annual biomass (Harris et al. 2016).

Active Management through Contracting

Within the broad policy context of use rights privatization and division of rangeland into smaller parcels, Gouli herders try to match livestock numbers to pasture access through contracting arrangements that are multiple and fluid, often changing from year to year for each household. Most contracts are signed annually, though some have terms as long as five years. Several Gouli herders act as middlemen, brokering complicated, multi-party arrangements. Most contracting is of sheep rather than yaks, because yak herds do not grow as quickly and yaks do not need to be watched as closely as sheep. It is common for households to graze their yaks on (common) summer or fall pasture during some or all parts of the winter, saving their winter pasture, which is lower and warmer, for sheep.

Most herding households in Gouli take part in contracting of livestock and land. In general, less wealthy households prefer to take in other households’ livestock on contract, as the butter and cheese produced can be consumed or sold, and the household can also increase their own herd size. Most commonly, they return to the owner 30–35 lambs per 100 adult sheep taken in, while keeping the rest of the lambs, or alternatively return the original flock along with the monetary equivalent of 30–35 lambs, keeping the entire flock of baby sheep born in a year.

Conversely, those who are wealthier, those who have more livestock than they feel their land or labor can support – and those who make a living in town as salaried employees or petty entrepreneurs or who must move to town for other reasons (such as caring for family members in the hospital) -- contract their livestock to other households. These contracts usually stipulate that they will receive their original flock back at the end of the year, along with 30–35 lambs or their monetary equivalent. They also require that those who take in livestock must compensate owners for any livestock loss, including livestock deaths from heavy snow. Thus, these arrangements also serve as a risk mitigation strategy for wealthier herders, who essentially offload that risk to those with fewer assets (Yeh and Gaerrang 2011).

Households endowed with adequate land and labor power often decide to herd other households’ livestock along with their own, on their own pasture. If they wish to increase their own herd size, they choose the option of keeping 30–35 lambs per 100 sheep (and bearing the risk of livestock loss). Alternatively, they may choose monetary compensation for their labor. Conversely, they may mitigate risk by contracting part of their pasture to other herders to use in exchange for cash payment. Finally, those who have a reputation within the village of being particularly skillful herders can generally leverage more advantageous arrangements.

Successful Use of Contracts

Here we present two examples of successful pastoralists who have increasingly turned to contracts and rentals to manage the mismatch between their herd sizes and the amount and quality of pasture to which they have use rights. These cases further demonstrate that herders do not simply increase their livestock numbers without regard to limits.

Pastoralist T

In 2009, Pastoralist T’s household had the highest per capita number of livestock among herders in Village Five. Like many others in the village, he lost more than 100 sheep in the snowstorm of 1997–1998. In 2001, he purchased 100 sheep to increase his herd size to 200. By then he had begun to contract out some of his sheep to other herders, who kept 65–70% of the new lambs born each year, returning the original flock plus 30–35% of the new lambs to Pastoralist T. By 2005, his sheep flock had reached 500, including 100 lambs. During the winter, Pastoralist T herded them for two months on his own winter pasture and then for another four months on pasture subleased from another herder. He also sold 80 sheep that year, noting that he felt there were too many for his pasture to support. Like most other Gouli herders, Pastoralist T grazed his yaks primarily on autumn and summer pastures even during the winter.

From 2005 to 2009 he contracted 300 sheep to other herders to graze on the other herders’ pastures. He used his own pasture for his yaks, as well as the yaks of other herders. For example, in the winter of 2008–2009, Pastoralist L herded 70 of his own yaks and all 150 of Pastoralist T’s yaks on Pastoralist T’s winter pasture from October through February, and then on common summer pasture in March and April. That year, severe snow caused 20 of Pastoralist T’s yaks to die of malnutrition and another ten were killed by wolves. In 2009–2010, Pastoralist T contracted 100 of his 140 yaks to Pastoralist PG to graze on Pastoralist T’s land. Pastoralist T also paid Pastoralist L 8000 RMBFootnote 3 to graze his remaining 40 yaks on Pastoralist L’s land for six months of that year.

In the winter of 2010–2011, Pastoralist T subleased pasture use rights from another herder in the village for three months in order to rest his own pasture. When we interviewed him about the size of his pasture, he pointedly said that his decisions have to do with the quality, rather than the area, of his rangeland. Pastoralist T was particularly attentive to the relationship between livestock trampling and the quality of his grassland, stating that he did not sublease his pasture to other herders because they might overstock it, leaving him with damaged pasture. Like others, he stated that weather, specifically precipitation, was an important factor in his annual assessments about how to modify his herd size. Through 2014, he had continued to contract his sheep out to other herders while maintaining 200 yaks. In 2014, he stated that he planned to reduce his number of livestock because, “now there are too many animals…If livestock number increases, then future grassland conditions will be poor. More livestock damages the grassland through their trampling on the grassland.”

Pastoralist H

Pastoralist H, a former village Communist Party Secretary, is respected in Village Five as a thoughtful and capable man and considered a skilled herder. He has six children: a monk, a trader living in neighboring Golog County, two high school graduates now working in salaried positions, and two herders living at home. His sheep numbers have increased slowly and steadily over time; once he assessed that his allocated land had already reached the maximum number of livestock it could support, his dominant strategy has been to contract his livestock to others. Compared to Pastoralist T, Pastoralist H was emphatic about the role of weather variables in determining grassland condition, going so far as to say, “Grassland condition does not depend on herding. It depends on the weather.” This observation about the significance of climate factors over biological ones in affecting vegetation condition does not, however, mean that he pays no attention to how many livestock he puts on the grasslands. Indeed, Pastoralist H is known for diligently and closely following his livestock, carefully managing where they graze.

The complexity of his herding strategies also reveals Pastoralist H’s considerations of livestock number and grassland condition. In 2009, he subleased the land of a monk in the village for a five-year period, for 50,000 RMB. During the winter, he herded 400 of his sheep on his own winter pastures, which were divided at the time of rangeland allocation into five different parcels, as well as that of a relative. He also contracted out his remaining 250 sheep to three other herders, each of whom grazed the sheep on their own land in return for taking 35% of the new lambs born that year. The following year, his herd reached its maximum size of 800 sheep and 170 yaks. That year, he contracted sheep to five households (20 to one household, and 50 each to four households) and contracted 140 lambs to a sixth household. He himself raised the remaining 440 sheep on his and his relative’s combined pastures. That year his family members also herded their 170 yaks on fall and summer pasture during most of the winter. Deciding that there were too many livestock for his land, he sold 90 sheep and continued to reduce his herd size the following year.

Herders’ Environmental Knowledge

Having addressed two policy assumptions about livestock numbers and pastoralists’ active management, we now turn to the third assumption, about knowledge. That herding is a practice and profession requiring skill and careful observation is not acknowledged in state policies or discourses, which represent pastoralists as having, until recent state intervention, “wandered around” in search of grass and water, implying a random and unskilled practice. However, we find that Tibetan pastoralists clearly have a great deal of environmental knowledge. This knowledge is what Ingold and Kurttila (2000) call “LTK” or “traditional knowledge as generated in the practice of locality,” as opposed to “MTK” or “traditional knowledge as generated enframed in the discourse of modernity.” The latter model conceives of knowledge as a substance, a set of discrete items that is “passed down” in order to be retrieved and applied, whereas the former understands knowledge to emerge through embodied practice. As practice, knowledge “undergoes continual generation and regeneration within the contexts of people’s practical engagement with significant components of the environment” (Ingold and Kurttila 2000:192). Ingold and Kurttila further argue that LTK can be conceptualized as skill, understood as “a property of the whole human organism-person, having emerged through the history of his or her involvement in an environment;” it is thus “refractory to codification in the programmatic form of rules and representations” (193). In short, it is that which cannot be adequately described in the two-dimensional form of a textual narrative or a set of quantitative metrics. That distilling such embodied skill into written textual, numeric, or symbolic form is difficult and awkward does not mean that herders do not have sophisticated practical knowledge about their land and livestock. This can be seen in how herders describe which livestock they graze at what times in what locations. For example, Pastoralist L explained:

I have both sunny and shadow sides in my winter pasture. When it is snowing, I use the sunny side of the pasture [which he has fenced], which is a little warmer than other parts. When the snow is not heavy, I can use both sunny and shadow parts. When it is sunny, I use the parts of the pasture that have less grass, such as the high rocky mountain area and the shadow side where the grass is shorter. This way, I can save the best parts of the pasture for harsh days when it is snowing or very cold. During cold days, I herd livestock in the relatively warm parts of the winter pasture, but not the most productive parts, which I save for days when it is snowing. On good days, I herd livestock in places where they can access just enough grass, not the best….

Herding yaks is easier than herding sheep. But these days, there are very few places where one can free yaks to graze wherever they want. Herders have to go with the yaks to graze them strategically. A herder has to see which parts of winter pasture they should graze first, and which parts later, when it is cold. If you let your yaks go freely to winter pasture and let’s say they’re not killed by wolves, they will eat the best parts of the pasture and then there won’t be any grass left during the harsh [snowy] days. A good herder will not allow this to happen.

Pastoralist DT explains his strategies for herding different types of livestock as follows:

On the winter pasture, there are sunny and shadow parts, but there is not enough for the livestock because parts are bare soil with no grasses left. I have flat grassland that I use in the winter for livestock that are old and weak, and also sheep and yaks that already have newborn lambs and baby yaks. When it is windy and chilly, we herd on the sunny part of the land and the valley… Sometimes it is necessary to herd [different kinds of livestock] separately. At other times, it is necessary to herd them together when there is only one part of the grassland that can be grazed. If there is snow, then there should be more grass on the shadow side of the winter pasture that is covered with snow until February. If there isn’t much snow, then the sunny side should be better. In years when there isn’t much snow, we do not allow the livestock to eat grass on the sunny side [because we save it for later]. When we want to herd the livestock on the higher parts of the pasture, we have to be very careful not to trample the grass when we drive the herd to the destination. We drive them through a very small valley where there is a path so that the grass will not be affected by trampling.

Our vegetation study corroborates the fact that herders direct their attention to pastures and livestock. Harris et al. (2016) found that overall, pastoralists stocked their pasture in response to the relative abundance of palatable forage, and stocked more lightly where indicators of erosion were higher. Though Gouli herders did not verbally articulate their strategies in relation to palatable vs. unpalatable species, their practical actions consisted of avoiding putting more sheep on pastures that had high biomass but significant weedy species. That is, their embodied skills included a differentiation between palatable and unpalatable species, even when, in interviews, they did not articulate these distinctions (see results from Harris et al. 2016).

Of course, like other groups of people, individual Gouli herders vary in their level of knowledge and skill in herding. Some are more thoughtful, meticulous, diligent, and observant in their management and embodied skill than others. By demonstrating that herders are knowledgeable about their environments through continual engagement with place, and do not just “wander” aimlessly in search of water and grass, our intention is not to paper over differences in the degree to which this is the case. After accounting for both annual and site specific differences in forage quality and quantity, Harris et al. (2016: Table 2) found that pastures displayed heterogeneity in their responses to pastoralist decision-making, and that pastoralists evidently differed in their ability to move pastures toward palatable, away from unpalatable, and away from eroded conditions.

The “Scent” of the Soil

In discussing qualities of different pastures, herders frequently deploy the concept of the soil’s “scent” (sa dri). For example, Pastoralist W stated, “the best pasture has the qualities of dense grass, soil that has a good scent, and a combination of sunny and shadowy slopes, wetlands, and flat areas.” Gouli pastoralists state that because their soil has a particularly good scent, their livestock are relatively plentiful, large, and healthy compared to those in other pastoral areas on the Tibetan Plateau, even though grass cover is sparser. As one put it, “Even though there is less grass than other places, the livestock here can better survive. Around Qinghai Lake, the grass is much denser, but their livestock are not as good as ours. It’s said this is due to the different scent of the soil.” Another explained, “in a place with abundant grass but without good soil scent, the livestock will starve in the fall even if there is grass to eat.” Several herders attributed this to a greater salt content in the soil. Importantly, herders emphasize that grass is not everything; tall grass growing on soil with poor sa dri will result in poor livestock, whereas good sa dri will cause livestock to be fat and healthy even if the land looks degraded to outsiders because of sparse grass. This insight speaks to frequent clashes between outsiders’ landscape aesthetics and local understandings of vegetation history, in pastoral areas and beyond (Williams 2002; Yundannima 2012).

Despite their conviction that Gouli had relatively good soil, no herders stated that its scent was improving over time. Instead, in explaining why livestock weights have decreased, Pastoralist K stated that it was the scent of the soil that has deteriorated over time. A number asserted that mining and the digging of medicinal herbs undermines the good “scent” of the soil. This concept of the soil’s scent is found in other Tibetan pastoral areas as well, and is related to the concept of the soil’s “nutrition” or “essence” (sa bcud). The latter is widely deployed to explain why grassland condition suffers as a result of mining (see Yeh 2014). Sa bcud is understood as a regional condition, whereas sa dri is a more localized condition that can vary from kilometer to kilometer. Thus, though there is limited mining in Gouli itself, confined to the land of one pastoralist-businessman who willingly rented out his land for mineral extraction, herders asserted that mining locally and more broadly on the plateau affected both the scent and the essence of the soil and thus rangeland conditions and livestock weight.

Grazing, Climate and Grassland Condition

As is the case with much ecological research on the Tibetan plateau and elsewhere, Gouli herders’ analyses of vegetation conditions sometimes converge with and sometimes conflict with those of ecological science (Klein et al. 2014; Nightingale 2016). The majority of Gouli herders we interviewed stated that grassland conditions at the present time are worse than in the past, a general trend that is reinforced by their observations of declining livestock weights, and that is supported by Harris et al.’s (2016) vegetation study, which found declines in most rangeland indicators over the 2009–2012 study period, even accounting for annual weather fluctuations. However, there were some dissenting voices. For example, Pastoralist RC stated in 2014 that though conditions worsened for a number of years after the division of winter pastures, “since 2005, grassland condition has been improving. The grass is taller and thicker. There is more grass cover and less bare ground.” Herder L also stated that the years immediately following winter pasture allocation were characterized by more bare ground, but that this changed after 2007–2008. He attributed his observations of less bare ground and higher grasses to his own skillful management of the pastures.

Gouli herders also have complicated and at times self-contradictory understandings of the extent to which grazing density - as opposed to weather conditions - affects vegetation. As discussed above, herders emphasize the importance of highly variable weather conditions as a key driver of annual variation in vegetation growth. For example herder LG states, “When there is more rain, vegetation grows better, particularly those with flowers, which are good for yaks. If there is more grass mixed with flowering plants yaks will become more productive in milk, which is an important indicator that yaks have good nutrition.”

Some herders’ statements about weather events appear to represent a radically non-equilibrium view of the ecosystem, in which weather events make livestock numbers inconsequential for grassland dynamics. These herders explain that their pastures have limits in terms of how many livestock they can support, but that the consequences of exceeding these limits would be the death of livestock, rather than grass growth the following year (as long as precipitation is plentiful). Several claimed that “in places where grass grows well, it will continue to grow regardless of livestock number; in places where livestock do not grow well, then the grass does not grow even where no livestock are put on the land,” a proposition contradicted by our vegetation study (Harris et al. 2016).

However, these assessments are tempered and somewhat contradicted by other statements, sometimes by the same herders, that do suggest the importance of livestock density, usually couched in terms of a problem of trampling rather than grazing per se. Several stated that the soil might be “killed” by the trampling action of too many livestock. Others suggest that both trampling and grazing affect grassland condition, a proposition supported by Harris et al.’s (2016) study. Pastoralist GW asserts, “if you herded your livestock in a way that does not allow for some grasses to be leftover, this will affect next year’s grass growth.” Pastoralist GK agreed that, “if you herd fewer livestock on the pasture, this is good for the grassland condition.” More importantly, even those who stated that livestock numbers made no difference for grassland condition had clear notions of how many livestock are appropriate for their pastures, and engaged in contracting of land and livestock to maintain these levels, with varying degrees of success. That is, though they appeared to articulate a radically non-equilibrium view of the ecosystem, their embodied practices were more consistent with a much more nuanced view (and one more in keeping with current ecological understandings) in which biotic and density-dependent factors of herbivory and trampling are consequential at many times, while being made irrelevant at other points in time.

These disparate views can be interpreted together to suggest that short-term weather variations may outweigh stocking rate in determining annual variation in range condition, but management of grazing density and timing also matter, particularly over longer periods. Our vegetation study found evidence that, normalizing for the effect of annual weather variations, some ecological parameters (e.g., proportion of bare soil, erosion index, vegetation cover and grass herbage mass) responded to livestock density over the 2009–2012 study period. Moreover, Harris et al.’s (2015) exclosure experiment suggested that the predominant forage species, Stipa purpurea, is adapted to moderate levels of herbivory and competes with itself in the absence of herbivory, so that grazing exclusion did not have strong effects on annual biomass production, though it did improve bare soil and other erosion indicators.

Finally, it should be noted that interpreting apparent discrepancies between herders’ articulated environmental knowledge and the results of ecological studies of vegetation is complicated not only by the gap that can be produced in the translation from LTK to MTK, but also by different time scales of observation. Our vegetation data, collected annually between the growing seasons of 2009 through 2012, cannot account, for example, for Herder L’s recollection that rangeland conditions worsened in 1998–1999 because of a lack of rain and a sudden increase in pika numbers, or pastoralist DT’s statement that there is more bare ground on his pasture now than in the past decade, but that 30 years ago there was more bare ground than today. These issues should be further explored with longer scientific data sets along with a recognition of the integrative character of local knowledge (Berkes and Berkes 2009; Klein et al. 2014).


Gouli herders are in agreement about their observations of decreased livestock weights over time. They interpret these observations through a set of causal explanations that partially overlap with and partially diverge from ecological understandings of rangeland degradation. Many point as an ultimate cause to their understanding that the contemporary world is in the midst of an “age of degeneration” or “Dharma-ending age,” a historical period following the life of Shakyamuni Buddha when sentient beings become greedy and filled with hatred, the dharma cannot be transmitted properly, and there is a state of general world decline. At least one herder in Gouli also attributed what he observed as increased grass cover on his pasture over the past six years to efficacious religious activities, including inviting monks to chant prayers over barley grains he provided as an offering. Simultaneously, herders also attribute the rangeland conditions they observe to their own management practices, to mining and quarrying both locally and regionally, to an overabundance of pikas, and to the household division of winter pasture and its attendant concentration of grazing and trampling.

While neither ecologists nor officials of an atheist state are likely to find much use for Buddhist explanations of the ultimate causes of grassland degradation, we have suggested here that there is certainly value for them in herders’ environmental observations and the proximate theories of degradation that they deduce from them. These are not blindly inherited from the timeless past, but rather constitute knowledge that is continually produced through embodied practice. Neither their observations nor our vegetation studies (Harris et al. 2015, 2016) can definitively determine whether and to what extent lag effects of overgrazing, household division of pasture and decreased mobility, current overstocking by some pastoralists or in some years, and climate change, alone and in combination, are leading to changes in grassland characteristics. However, the points of convergence and divergence between different forms and sources of knowledge highlighted by this study should be a particularly productive entry point for further research (Gearheard et al. 2010; Klein et al. 2014; Nightingale 2016; Popke 2016; Yeh 2016).

More broadly, we have demonstrated that Tibetan pastoralists assess limits to livestock holdings based on land and labor availability, and make active decisions about grazing practices based on their observations of vegetation, the weather, and their livestock, and in the context of multi-scalar and contemporary political-economic and cultural-political forces. Within a policy context that pushes privatization, they have increasingly turned to contracting of livestock and rangelands to maintain flexibility in management. Like all people, pastoralists differ in their levels of skill and knowledge, but our study suggests that they do have significant knowledge in the form of embodied, practical skill in stocking and managing their pastures. Policy makers who wish to achieve policy outcomes that benefit both herders and grasslands would do well to consider how to engage in dialogue with herders’ existing knowledge, management practices, cultural context, and goals rather than continue to act as if these do not exist.


  1. 1.

    Research report on establishment of long-term rangeland ecological protection compensation mechanism in the TAR (guanyu jianli Xizang caoyuan shengtai buchang changxiao jizhi de diaoyan baogao) in Proposals on Sustainable Development 2007 (in Chinese), Forestry Publishing House, Beijing, 366–72, in Nyima (2014:186).

  2. 2.

  3. 3.

    The exchange rate is roughly 6.5 RMB = 1 USD.


  1. Badingquiying,  Smith, A. T., Senko J., and Siladan M. (2016). Plateau pika Ochotona curzoniae poisoning campaign reduces carnivore abundance in southern Qinghai, China. Mammal Study 41(1): 1–8.

  2. Bauer K. (2005). Development and the enclosure movement in pastoral Tibet since the 1980s. Nomadic Peoples 9(1): 53–83.

    Article  Google Scholar 

  3. Bauer K. (2015). New lives, new homes - the social and economic effects of resettlement on Tibetan nomads (Yushu prefecture, Qinghai province, PRC). Nomadic Peoples 19(2): 209–220.

    Article  Google Scholar 

  4. Bauer K., and Nyima Y. (2010). Laws and regulations impacting the enclosure movement on the Tibetan plateau of China. Himalaya 30(1–2): 23–38.

    Google Scholar 

  5. Berdanier A. B., and Klein J. A. (2011). Growing season length and moisture interactively constrain high elevation above ground net primary production. Ecosystems 14: 963–974.

    Article  Google Scholar 

  6. Berkes F., and Berkes M. K. (2009). Ecological complexity, fuzzy logic, and holism in indigenous knowledge. Futures 41: 6–12.

  7. Bessho Y. (2015). Migration for ecological preservation? Tibetan herders’ decision making process in the eco-migration policy of Golok Tibetan autonomous prefecture (Qinghai province, PRC). Nomadic Peoples 19(2): 189–208.

  8. Byg A., and Salick J. (2009). Local perspectives on a global phenomenon - climate change in eastern Tibetan villages. Global Environmental Change 19(2): 156–166.

    Article  Google Scholar 

  9. Cao J.J., Holden N. M., Lü X.T., and Du G. (2011a). The effect of grazing on plant species richness on the Qinghai-Tibetan plateau. Grass and Forage Science 66(3): 333–336.

  10. Cao J.J., Xiong, Y.-C., Sun J., Xiong W.-F., and Du G.-Z. (2011b). Differential benefits of multi- and single-household grassland management patterns in the Qinghai-Tibetan plateau of China. Human Ecology 39(2): 217–227.

  11. Cao J.J., Yeh E. T., Holden N. M., Yang Y.Y., and Du, G.Z. (2013). The effects of enclosures and land-use rights contracts on grassland degradation on the Qinghai-Tibetan plateau. Journal of Arid Environments 97: 3–8.

  12. Chu D., Lu L., and Zhang T. (2007). Sensitivity of normalized difference vegetation index (NDVI) to seasonal and interannual climate conditions in the Lhasa area, Tibetan plateau, China. Arctic, Antarctic and Alpine Research 39(4): 635–641.

    Article  Google Scholar 

  13. Doran L., Low A., and Kemp R. (1979). Cattle as a store of wealth in Swaziland: implications for livestock development and overgrazing in Eastern and Southern Africa. American Journal of Agricultural Economics 61: 41–47.

  14. Du M., Kawashima S., Yonemura S., Zhang X., and Chen S. (2004). Mutual influence between human activities and climate change in the Tibetan plateau during recent years. Global and Planetary Change 41(3–4): 241–249.

    Article  Google Scholar 

  15. Du M., Yonemura S., Zhang X., He Y., Liu J., and Kawashima S. (2012). Climatic warming due to overgrazing on the Tibetan plateau. Journal of Arid Land Studies (Shamuo yanjiu) 22(1): 119–122.

    Google Scholar 

  16. Gaerrang (2011). The alternative to development on the Tibetan plateau: preliminary research on the anti-slaughter movement. Revue d'Etudes Tibetaines 21: 31–43.

    Google Scholar 

  17. Gaerrang (2015). Development as entangled knot: the case of the slaughter renunciation movement in Tibet, China. Journal of Asian Studies 74(4): 927–951.

    Article  Google Scholar 

  18. Gayley H. (2013). Reimagining Buddhist ethics on the Tibetan plateau. Journal of Buddhist Ethics 20: 247–286.

    Google Scholar 

  19. Gayley H. (2016). Controversy over Buddhist ethical reform: a secular critique of clerical authority in the Tibetan blogosphere. Himalaya: the Journal of the Association for Nepal and Himalayan Studies 36(1): 22–43.

    Google Scholar 

  20. Gearheard S., Pocernich M., Stewart R., Sanguya J., and Huntington H. (2010). Linking Inuit knowledge and meteorological station observations to understand changing wind patterns at Clyde River, Nunavut. Climatic Change 100: 267–294.

    Article  Google Scholar 

  21. Gongbuzeren,  Li Y., and Li W. (2015). China's rangeland management policy debates: what have we learned? Rangeland Ecology & Management 68: 305–314.

  22. Harris R. B. (2010). Rangeland degradation on the Qinghai-Tibetan plateau. Journal of Arid Environments 74: 1–12.

    Article  Google Scholar 

  23. Harris R. B., Wang W., Badinqiuying, Smith, A. T., and Bedunah D. J. (2015). Herbivory and competition of Tibetan steppe vegetation in winter pasture: effects of livestock exclosure and plateau pika reduction. PLoS One 10. doi:10.1371/journal.pone.0132897.

  24. Harris R. B., Samberg L. H., Yeh E. T., Smith A. T., Wang W., Wang J.,  Gaerrang, and Bedunah D. (2016). Rangeland responses to pastoralists’ grazing management on a Tibetan steppe grassland, Qinghai Province, China, The Rangeland Journal. 38(1): 1-15. doi:10.1071/RJ150410.

  25. Herskovits M. (1926). The cattle complex in East Africa. American Anthropologist 28: 361–388.

    Article  Google Scholar 

  26. Ingold T., and Kurttila T. (2000). Perceiving the environment in Finnish Lapland. Body & Society 6(3–4): 183–196.

    Article  Google Scholar 

  27. Kabzung, and Yeh E. T. (2016). Slaughter renunciation in Tibetan pastoral areas: Buddhism, neoliberalism and the ironies of alternative development. In Rojas C., and Litzinger R. (eds.), Ghost Protocol: Development and Displacement in Global China, Duke University Press, Durham, N.C., pp. 109–130.

  28. Klein J. A., Hopping K., Yeh E. T., Nyima Y., Boone R., and Galvin K. A. (2014). Unexpected climate impacts on the Tibetan plateau: local and scientific knowledge in findings of delayed summer. Global Environmental Change 28: 141–152.

    Article  Google Scholar 

  29. Levine N. (2015). Transforming inequality: Eastern Tibetan pastoralists from 1955 to the present. Nomadic Peoples 19(2): 164–188.

  30. Li J. (2012). Land tenure change and sustainable management of alpine grasslands on the Tibetan plateau: a case study from Hongyuan County, Sichuan Province, China. Nomadic Peoples 16(1): 36–49.

    Article  Google Scholar 

  31. McCabe T. (2004). Cattle bring us to our enemies: Turkana ecology, politics and raiding in a disequilibrium system. University of Michigan Press, Ann Arbor.

    Google Scholar 

  32. McPeak J. (2005a). Confronting the risk of asset loss: what role do livestock transfers in northern Kenya play? Journal of Development Economics 81: 415–437.

    Article  Google Scholar 

  33. McPeak J. (2005b). Individual and collective rationality in pastoral production: evidence from northern Kenya. Human Ecology 33(2): 171–197.

    Article  Google Scholar 

  34. Naess M. W., and Bardsen B.-J. (2010). Environmental stochasticity and long-term livestock viability: herd-accumulation as a risk reducing strategy. Human Ecology 2010(38): 3–17.

    Article  Google Scholar 

  35. Nightingale A. J. (2016). Adaptive scholarship and situated knowledges? Hybrid methodologies and plural epistemologies in climate change adaptation research. Area 48(1): 41–47.

    Article  Google Scholar 

  36. Nyima Y. (2014). A large herd size as a symbol of wealth? The fallacy of the cattle complex theory in Tibetan pastoralim. Area 46(2): 186–193.

    Article  Google Scholar 

  37. Nyima Y. (2015). What factors determine carrying capacity? Area 47(1): 73–80.

    Article  Google Scholar 

  38. Pan T., Wu S., and Liu Y. (2015). Relative contributions of land use and climate change to water supply variations over Yellow River source area in Tibetan plateau during the past three decades. PLoS One. doi:10.1371/journal.pone.0123793.

    Google Scholar 

  39. Popke J. (2016). Researching the hybrid geographies of climate change: reflections from the field. Area 48(1): 2–6.

  40. Roth E. A. (1996). Traditional pastoral strategies in a modern world: an example from northern Kenya. Human Organization 55(2): 219–224.

    Article  Google Scholar 

  41. Smith A. T., and Foggin J. M. (1999). The plateau pika (Ochotona curzoniae) is a keystone species for biodiversity on the Tibetan plateau. Animal Conservation 2: 235–240.

    Article  Google Scholar 

  42. Sun J., Cheng G. W., and Li W. P. (2013). Meta-analysis of relationships between environmental factors and aboveground biomass in the alpine grassland on the Tibetan plateau. Biogeosciences 10: 1707–1715.

  43. Wang J., Zhang X., Chen B., Shi P., Zhang J., Shen Z., Tao J., and Wu J. (2013). Causes and restoration of degraded alpine grassland in Northern Tibet Journal of Resources and Ecology 4(1): 43–49.

  44. Wei Y., and Chen Q. (2001). Grassland classification and the evaluation of grazing capacity in Naqu Prefecture, Tibet Autonomous Region, China. New Zealand Journal of Agricultural Research 44: 253–258.

  45. Williams D. M. (2002). Beyond great walls: environment, identity and development on the Chinese grasslands, Stanford University Press, Stanford.

    Google Scholar 

  46. Wilson M. C., and Smith A. T. (2015). The pika and the watershed: the impact of small mammal poisoning on the ecohydrology of the Qinghai-Tibetan plateau. Ambio 44: 16–22.

    Article  Google Scholar 

  47. Yan Z., and Wu N. (2005). Rangeland privatization and its impacts on the Zoige wetland on the eastern Tibetan plateau. Journal of Mountain Science 2(2): 105–115.

    Article  Google Scholar 

  48. Yan Z., Wu N., Dorji Y., and Jia R. (2005). A review of rangeland privatization and its implications on the Tibetan plateau, China. Nomadic Peoples 9(1): 31–52.

    Article  Google Scholar 

  49. Yan J., Wu Y., and Zhang Y. (2011). Adaptation strategies to pasture degradation: gap between government and local nomads in the eastern Tibetan plateau. Journal of Geographic Science 21(6): 1112–1122.

    Article  Google Scholar 

  50. Yeh E. T. (2009). Greening western China: a critical view. Geoforum 40: 884–894.

    Article  Google Scholar 

  51. Yeh E. T. (2013). The politics of conservation in contemporary rural China. Journal of Peasant Studies 40(6): 1165–1188.

    Article  Google Scholar 

  52. Yeh E. T. (2014). Reverse environmentalism: contemporary articulations of Tibetan Buddhism, culture, and environmental protection. In: Van der Veer P, Miller J, Yu D.S., (eds.), Religious Diversity and Ecological Sustainability in China. Routledge, New York, pp. 194–218.

  53. Yeh E. T. (2016). How can experience of lcoal residents be 'knowledge'? challenges in interdisciplinary climate change research. Area 48(1): 34–40.

    Article  Google Scholar 

  54. Yeh E. T., and Gaerrang. (2011). Tibetan pastoralism in neoliberalising China: continuity and change in Gouli. Area 43(2): 165–172.

  55. Yeh E. T., Nyima Y., Hopping K., and Klein J. A. (2014). Tibetan pastoralists' vulnerability to climate change: a political ecology analysis of snowstorm coping capacity. Human Ecology 42(1): 61–74.

    Article  Google Scholar 

  56. Yundannima (2012). From 'Retire livestock, restore rangeland' to the compensation for ecological services: state interventions into rangeland ecosystems and pastoralism in Tibet. PhD dissertation. University of Colorado Boulder.

  57. Zhang Y., Wang G., and Wang Y. (2010). Response of biomass spatial pattern of alpine vegetation to climate change in permafrost region of the Qinghai-Tibet plateau, China. Journal of Mountain Science 7(4): 301–314.

    Article  Google Scholar 

Download references


We thank Pemabum for field assistance and the many herders in Gouli who patiently answered our questions over a number of years.

Author information



Corresponding author

Correspondence to Emily T. Yeh.

Ethics declarations


This study was funded by the US National Science Foundation, Dynamics of Coupled Natural and Human Systems Program, Award 0815441.

Conflict of Interest

The authors declare that they have no conflict of interest.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Yeh, E.T., Samberg, L., Gaerrang et al. Pastoralist Decision-Making on the Tibetan Plateau. Hum Ecol 45, 333–343 (2017).

Download citation


  • Tibet
  • Pastoralism
  • Rangeland condition
  • Livestock management
  • Environmental knowledge