Table of Contents (All Chapters)
Table of Contents (Chapter Six)
CHAPTER SIX
Rangeland Utilization, Grassland Quality, and Biodiversity in Alpine Grasslands: A Regional Analysis
Introduction (back to top)
In recent years, China’s grasslands have undergone many changes in the way they are utilized. The total human population and population structure as well as basic resource management objectives have changed in the last few decades, with some potential negative consequences for sustainability. In the Tibetan plateau region, which comprises most of China’s alpine grasslands, many of these changes have followed directly from the region becoming increasingly integrated with the rest of the country – with its economy, its population, its politics, and its culture or predominant value system. Some of the ecological impacts of fencing (and the land use intensification, even sedentarization, that it represents) have been examined in the previous chapter.
In this chapter, the impacts on grassland quality and biodiversity of several other ecologically important variables, both natural and human, are examined. This chapter is a continuation of the previous one, simply based on a different dataset that has been compiled from official governmental sources. As with fencing, most of the variables examined here – annual rainfall, population density, presence of non-local influences, the relative importance of pastoralism, and seasonality of grassland use – are closely related to changing times, some to new development goals and objectives or to new approaches in resource management, and others simply to an increasing human population or to internal migration patterns within China. The purpose of this chapter is to examine further some of the implications of old and new (or emerging) patterns of land use (resource management) in Qinghai’s alpine grasslands by analyzing official county-level government statistics[1].
Methods and Results (back to top)
General methods (back to top)
The present dataset was compiled from four main sources. Measures of grassland areas, grassland quality, and season of use were obtained for each county in Qinghai Province Today (Jing 1986), population figures and information on main occupations in Tabulation on the 1990 Population Census of Qinghai Province (Qinghai Census Bureau 1992), rainfall in Qinghai’s Resources (Wei 1993), and information on the presence of mammal species in each county in the Annals of Qinghai’s Economic Wildlife (Li 1989).
Alpine grassland counties are here defined by the exclusion of all counties in the main agricultural area of the province, Haidong district, and of the most arid counties in Haixi prefecture. All the counties in Haibei, Hainan, Huanguan, Guoluo, and Yushu prefectures, as well as Tianjun and Wulan counties in Haixi prefecture, are included in the following analyses. The area defined by these 27 counties lies almost exclusively over 3,000 m above sea level, and the habitat ranges from mesic alpine meadow to xeric alpine steppe (Hu et al. 1992).
In this chapter, grassland quality and mammalian biodiversity are taken as two dependent variables, each affected by many other natural and anthropogenic factors. However, since many of these factors (variables) are interrelated, principal components analysis is first used to transform them into statistically independent factors, then Varimax rotation is used to increase the interpretability of each factor and thus to assign each factor to one of the original variables. Multiple regression makes it possible to determine which factors most affect (or are most closely related to) grassland quality and biodiversity.
Grassland quality (GQ) (back to top)
All of China’s grasslands have been classified according to general type and class (Su 1993). Of the 17 main types thus identified, five occur primarily on the Tibetan plateau. The total land area of each type, the proportion it comprises of the total natural grassland area in China, and the area and proportion of each grassland type that is “available” for livestock grazing are presented in Table 15. According to Huang (1992), 49 percent of Qinghai’s grassland is classified as alpine meadow, and 38 percent as steppe. All grasslands in Qinghai also have been divided into different classes according to the “proportion of the forage weight of various classes [qualities] which compose the grassland community” (Su 1993). Usable grassland thus has been classified in Qinghai as either low, medium, or high quality grassland. This latter classification was introduced in the early 1980s as part of the Household Responsibility System in order to promote a more equitable distribution of land to individual users or lease holders (Ho 1998).
Table 15. Main grassland types present on the Tibetan plateau (adapted from Su 1993)
Grassland Total Area Proportion Area of Usable Proportion of
Type in China of Total Grassland Usable Grassland
High-Cold Meadow-Steppe Type 6,870,000 1.9 % 6,010,000 87.5 %
High-Cold Desert-Steppe Type 9,570,000 2.7 % 7,750,000 81.0 %
Montane Meadow Type 16,720,000 4.7 % 14,920,000 89.2 %
Alpine Meadow Type 63,720,000 17.9 % 58,840,000 92.3 %
However, in order to relate grassland quality to average annual rainfall, rangeland population density and other variables, a unique (single) measure of grassland quality is needed instead of three separate measures of grassland area. Such a measure is obtained by weighting each class category by a different factor – high quality grassland by a factor of 3, medium quality grassland by a factor of 2, and low quality grassland by a factor of 1 – and then dividing the sum of these products by the best possible (or potential) grassland quality. Potential grassland quality is the value that would be obtained if the entire grassland area in each county were classified as high quality grassland (and hence multiplied by a factor of 3). This simple ratio is a relative measure of grassland quality, without units, and with a numeric range between 0.33 and 1.00. (Note that the lower end of this relative measure, 0.33, is contingent on the arbitrary weighting factors that were chosen). In order to increase ease of interpretation, multiplying all the values by 1.5 increases the range of relative values to one unit, and the range can then be shifted to between 0 and 1 by subtracting 0.5 units. In algebraic terms, the equation used to convert three distinct classes of grassland quality into a single relative measure of grassland quality (GQ) is
( ( ( 3H + 2M + 1L ) / 3
( H + M + L ) ) * 1.5 ) – 0.5
where H is the area of high quality grassland, M is the area of medium quality grassland, and L is the area of low quality grassland. This equation provides a measure of grassland quality that can be compared between counties with different total grassland areas. Measured this way, grassland quality varies all the way from 0.05 to 1.00 in the 27 alpine grassland counties, with an average value of 0.62 (i.e., slightly better than “medium” quality).
Biodiversity (BD) (back to top)
A basic estimate of biodiversity was derived from a Chinese publication that provides a dataset of the presence/absence of 103 mammal species for each county (Li 1989). However, given the limited resources (financial and otherwise) available to the principal researchers, it is not clear whether the presence/absence data were used to determine species’ distribution ranges, or if “known” distribution ranges were instead used to “compile” the dataset. It must be understood therefore that the present measure of biodiversity may be based only on the “likely presence” of several key species and may not be suitable for any more detailed species- or county-specific examination. However, this measure can still be taken as a general indicator of basic trends in species’ presences and, when many species are considered jointly, of species richness.
Of the 103 species in the original dataset (Li 1989), only species with the following characteristics are included in the present analyses: species that were reported in more than 25 percent of the 27 alpine grassland counties, species that were identified as “Tibetan fauna” by Hoffmann (1991) or as “resident” or “characteristic fauna” of the Tibetan plateau by Zhang (1991), and species that are known or presumed to inhabit steppe, grassland, or montane habitat. Twenty-one mammals were identified as characteristic of Qinghai’s alpine grasslands (Table 16). The measure of biodiversity used here is the percentage of these 21 species that occur in each grassland county. Biodiversity values range from 57 to 90 percent, with an average of 80 percent. It is assumed that all 21 mammal species are potential residents of each county.
Rainfall (RF) (back to top)
Rainfall is an important variable because it is likely the single most important factor affecting grassland productivity (Sala et al. 1988), and more productive grasslands are positively associated with plant biodiversity (Tilman et al. 1996). According to Smith (2000), “temperate natural grasslands develop in regions characterized by an annual rainfall between 250 and 750 mm … and [with] seasonal and annual droughts.” Annual rainfall in Qinghai is 432 mm on average, and varies between 176 and 764 mm.
Population density (PD) (back to top)
The presence of pastoralists on the grassland both impacts the grassland and its wildlife and is constrained by the availability of forage resources. The population density of pastoralists in Qinghai’s rangelands was calculated by multiplying the proportion of each county’s working population engaged in pastoralism by the county’s total population (this provides an estimate of the total population that has a predominantly pastoral livelihood), and then dividing this figure by each county’s usable rangeland area (i.e., the area “available” for livestock grazing; Su 1993). The latter is the same area that is used to calculate “season of use” and grassland quality. The average population density in the 27 counties is 2.5 people/km2, ranging from 0.3 to 5.9 people/km2. However, since pastoralists tend to have larger families than average (Qinghai Census Bureau 1994), the calculated measure of population density underestimates real densities on the grasslands.
Table 16. Characteristic alpine grassland mammalian fauna of Qinghai province
Order Family Latin name Common Name
Carnivora Canidae Canis lupis gray wolf*
Canidae Vulpes ferrilata Tibetan sand fox*
Canidae Vulpes vulpes red fox*
Felidae Panthera uncia snow leopard
Mustelidae Meles meles Eurasian badger*
Mustelidae Mustela altaica mountain weasel*
Mustelidae Mustela eversmanni steppe polecat*
Ursidae Ursus arctos brown bear
Artiodactyla Bovidae Ovis ammon argali
Bovidae Pantholops hodgsoni Tibetan antelope*
Bovidae Poephagus mutus wild yak
Bovidae Procapra picticaudata Tibetan gazelle*
Bovidae Pseudois nayaur blue sheep*
Cervidae Cervus albirostris Thorold’s deer*
Perissodactyla Equidae Equus kiang Tibetan wild ass*
Rodentia Cricetidae Cricetulus longicaudatus long-tailed hamster
Cricetidae Myospalax baileyi Bailey’s zokor*
Sciuridae Marmota himalayana Himalayan marmot*
Lagomorpha Leporidae Lepus oiostolus Tibetan woolly hare*
Ochotonidae Ochotona cansus Gansu pika*
Ochotonidae Ochotona curzoniae plateau pika*
Note: My doctoral supervisor (Dr. A.T. Smith) or myself have observed the species marked by an asterisk.
Non-local influences (NL) (back to top)
Because so many of the changes currently taking place in the Tibetan plateau region have been introduced from the outside – both at the policy level and in terms of the arrival of new immigrants (and the various occupations, increasing demands on resources, and even the ideas that they may bring to the pastoral regions with them) – an examination of the relative influence of the “non-local” in each county could prove insightful. A simple measure of “non-local” influences in an area is the proportion of the total population that is comprised by non-local people. Such data are available from the 1990 population census (Qinghai Census Bureau 1992). The presence of non-local people varies between 0.7 and 16.1 percent over the 27 county area, with an average of 6.3 percent of the total population. The majority of local people are Tibetan pastoralists, while most non-local people are Hui and Han Chinese.
Percent pastoralism (PP) (back to top)
The proportion of each county’s population engaged in animal husbandry (pastoralism) is also examined. The relative importance of pastoralism in the 27 counties varies widely, with between 7 and 89 percent of the working population engaged in pastoralism. On average, 53 percent of people are herders or otherwise engaged in pastoral activities. In some counties, particularly those situated in the border areas of the Tibetan plateau, local people also cultivate the land, thus agriculture accounts for much of the observed variability. In county and township administrative centers, on the other hand, government posts, technicians’ jobs, and a variety of business opportunities also can be found. This statistic – the proportion of pastoralism present in each county – is calculated by dividing each county’s workforce specifically engaged in animal husbandry by the total workforce. When this statistic is examined independently of rangeland population density (PD) and the presence of non-local influences (NL), it can help determine whether the practice of pastoralism itself, intrinsically, impacts grassland quality or biodiversity.
Season of use (SU) (back to top)
Finally, the total areas of summer and winter pastures are reported for all 27 counties (Jing 1986). Although these figures may not be exact, they nonetheless provide a basic regional overview of the relative importance given to each land use category in Qinghai’s grasslands. According to Su (1993),
“1. Cold season grazing rangelands … in the frigid-temperate zone [are rangelands] where the snow depth is less than 10-20 cm in winter and spring.
2. All-year grazing rangelands are those within a radius of 10 km from the fixed inhabited spots [which] can not only be used suitably for grazing in cold seasons but also in warm seasons.
3. Warm season grazing rangelands are those of alpino-arctic ranges above the forest line with mild weather and stable water supply for man and animals. They consist of all the available ranges except those for grazing in cold seasons and those for year-round grazing.”
Thus as much grassland as possible is used for all-year (or year-round) grazing, and as much grassland as possible is used during the cold season. Given current policies in China, the area of uniquely summer pastures is likely to be decreasing, and many pastures are likely to be increasingly used in all seasons (with all of the negative ecological consequences noted in the previous chapter). The proportion of summer pastures in Qinghai’s 27 alpine grassland counties is 48 percent on average, ranging between 27 to 71 percent of the usable grassland area.
Principal components analysis (back to top)
Because the above five variables are interrelated (correlated), principal components analysis (PCA) is used to make them statistically independent. The independence of potentially explanatory variables is a basic assumption of regression analysis. The variation explained by the five independent PCA factors is reported in Table 17. Factor loadings are easy to interpret after Varimax rotation, and each factor can be assigned easily to a single original variable (Table 18).
Table 17. Main principal component analysis (PCA)
factors
Variable Eigenvalue Percent of Variation Cumulative Percent
Factor 1 2.0745 41.5 41.5
Factor 2 1.3205 26.4 67.9
Factor 3 0.7785 15.6 83.5
Factor 5 0.3931 7.9 100.0
Table 18. Principal
components analysis (PCA) factor matrix (after Varimax rotation)
Variables Factor 1 Factor 2 Factor 3 Factor 4 Factor 5
(new names) (Fact-SU) (Fact-PD) (Fact-NL) (Fact-PP) (Fact-RF)
NL -.00142 -.16589 .94512 -.20532 -.19256
PL .13809 -.10632 -.20987 .93713 .21763
RF .21568 .06897 -.20424 .22486 .92545
PD -.08704
.97908 -.14838 -.09207 .05782
SU .96985 -.09028 -.00184 .12552
.18834
Multiple regression analysis (back to top)
With five independent factors, each one distinctly associated with season of pasture use, rangeland population density, the presence of non-local influences, the proportion of pastoral livelihoods, or average annual rainfall, it is now possible to examine each of their separate effects on (or association with) the two independent variables of this study, grassland quality and biodiversity. The results of the two multiple regression analyses are reported in Table 19.
Table 19. Multiple regression analysis of season of use, rangeland population density, the presence of non-local people, proportion of pastoral livelihoods and average annual rainfall on grassland quality (GQ) and biodiversity (BD)
Grassland quality is associated with rainfall, population density, season
of use, and non-local influences in each county. Ecologically, rainfall is
expected to be positively associated with grassland productivity, and hence
with grassland quality. As expected, the relationship between rainfall and
grassland quality is significant (GQ by Fact-RF, p = .0072). Average annual
rainfall accounts for one-third of the total variability in grassland quality
(R2 = 0.3387; Figure 34).
Figure 34. Relative grassland quality by average annual rainfall
Grassland quality and population density also are associated with each
other (GQ by Fact-PD, p = .0072). Given the direction of the relationship,
however, it appears that this relationship is mainly the result of
pastoralists’ long-term attempts to utilize optimally all available grassland
resources (based on grassland quality as well as total grassland area).
Grassland quality accounts for one-fifth of the variability in rangeland
population density (R2 = 0.2014; Figure 35).
Even when the effect of rainfall is accounted for, as well as the association between grassland quality and population density, two other factors still have a significant impact on the quality of alpine grasslands: season of use (GQ by Fact-SU, p = .0432) and the presence of non-local people in each county (GQ by Fact-NL, p = .0513). Grassland quality increases when more grassland is used in the warm season (R2 = 0.1287; Figure 36), and decreases when more non-local people reside in the area (R2 = 0.2128; Figure 37). The proportion of pastoral livelihoods, however, is not significantly associated with grassland quality.
Figure 36. Relative grassland quality by season of use
Figure 37. Relative grassland quality by proportion of non-local
population
As noted in Table 19, rangeland population density and season of use also impact biodiversity (i.e., the proportion of characteristic Tibetan plateau mammals present in each county), as do non-local influences (as measured by the presence of non-local people). The proportion of characteristic mammals present in each county decreases as rangeland population density increases (BD by Fact-PD, p = .0011). However, this relationship is most likely due to human disturbance, as opposed to competition for forage, since the quality of grassland vegetation tends to be inversely related (insignificantly) with biodiversity. Rangeland population density explains over one-fifth of the variability of the measure of biodiversity used in this study (R2 = 0.2303; Figure 38).
Figure 38. Biodiversity by rangeland population density
The presence of non-local influences (non-local people) likewise impacts biodiversity negatively, as does season of use. Biodiversity decreases as non-local influences increase (BD by Fact-NL, p = .0382), and biodiversity is lower where more grassland is used as summer range (BD by Fact-SU, p = .0414). The magnitude of both of these relationships, however, is quite small (R2 = 0.0706 and 0.0491, respectively). Finally, biodiversity is neither impacted by rainfall nor by the relative importance of pastoral livelihoods (which is separate from population density) in Qinghai’s alpine grassland counties.
The direction of the
relationships between biodiversity and both population density and non-local
influences is in the one expected (i.e., both of these factors impact
biodiversity negatively), but the relationship between biodiversity and season
of use is not as straightforward. However, since human influences (population
density and non-local influences) negatively impact biodiversity, and this is
likely the result of various forms of physical disturbance, then it is expected
that wildlife would survive better in areas characterized by less human
disturbance. Wildlife generally will not inhabit the low-lying winter pastures
since these tend to be occupied by pastoralists throughout most of the year. It
is expected that their preferred habitat is the more remote, mountainous,
summer pastures, or uninhabited grassland. It is therefore normal that wildlife
would tend to avoid areas where pastoralists are more common in their (the
wildlife’s) preferred habitat (i.e., the remote alpine pastures). This is the
relationship that has already been noted above (BD by Fact-SU, p = .0414). If
human physical disturbance is one of the main factors that affects
biodiversity, as clearly seems to be the case, then it is also expected that
wildlife will be most abundant (biodiversity will be higher) in areas where
there is more rangeland left unused by pastoralists (even though such rangeland
also may be sparser or of lower quality). Data are available to examine this
relationship, and it is found that unused grassland (classified as “unusable”
grassland by most herders and government planners) is a strong predictor of
biodiversity. Both the proportion of the grassland area considered to be
“unusable” by pastoralists (R2 = 0.3183, p = .0027) and the total area
(absolute area) of “unusable” grassland (R2 = 0.4352, p = .0002; Figure 39) in
each county are closely related to biodiversity in Qinghai’s alpine grassland
region. This confirms the strong influence of human physical disturbance on
biodiversity, whether such disturbance is because of hunting by non-local
people or the geographic proximity of herders in preferred wildlife habitat.
Figure 39. Biodiversity by area of “non-usable” grassland
General Discussion (back to top)
Rainfall (back to top)
Rainfall is the single most important determinant of grassland productivity in many temperate ecosystems (Sala et al. 1988). In the Tibetan plateau region, both water and temperature (short growing season) affect primary productivity because of the very high altitude (Ekvall 1974, Holzner and Kriechbaum 1999). In the present study, it was confirmed that annual rainfall significantly impacts overall grassland quality. Furthermore, the proportion of unusable rangelands (because of insufficient water supplies to support livestock; Su 1993) also is inversely related with the quality of adjacent usable grasslands in each county (Pearson’s product moment r = -0.6275, p = 0.0005, unpublished analysis). Thus, average rainfall is clearly a very important determinant of grassland productivity and quality on the Tibetan plateau.
Percent pastoralism[2] (back to top)
The most obvious anthropogenic factor on the Tibetan plateau grasslands is the presence of pastoralists. Pastoralism has been practiced on the plateau for between 2,200 and 3,000 years (Zhao 1992, Hare 1998). As noted in this study, the practice of pastoralism itself (as distinct from population density, different aspects or forms of resource management, or a variety of non-local influences; see below) does not affect grassland quality or biodiversity. This finding is consistent with many other researchers’ assertion that, given the long history of pastoralism on the Tibetan plateau, sustainable husbandry practices have long been the norm for the region’s indigenous pastoralists (Ekvall 1968, Clarke 1987, Goldstein and Beall 1990, Miller 1995, Wu N. 1997). In contrast to this view, however, many government leaders and development workers assume that traditional forms of pastoralism are irrational, a view that has led to a plethora of misguided policies, usually related to the intensification of land use and promoting more sedentary lifestyles for pastoralists (Aronson 1980, Bennett 1988, Coughenour 1991, Barnett 1993, ‘Old ways…’ 1994, Environment Science and Technology 1998a, 1998c). Since pastoralism does not appear to impact the alpine grassland ecosystem, broadly defined, biodiversity protection and pastoralism should no longer be considered intrinsically at odds with each other. Instead the needs of each should be addressed together in an integrated way that synergistically builds on the contributions that each area can make to the other. In other words, pastoralism (and pastoralists) should be included in the formulation of broad regional plans for environment protection, not blamed and excluded from the process (Ghai and Vivian 1992, Noss and Cooperrider 1994, Stevens 1997a) or, worse still, moved to other regions altogether (a recurring suggestion made by several government bureaus in Qinghai).
Population density (back to top)
In contrast to the simple presence or practice of pastoralism, population density is closely related to both grassland quality and biodiversity. Population density is positively associated with grassland quality, which would indicate that pastoralists tend to distribute themselves on the rangeland proportionately with grassland quality which, in conjunction with total grassland area, is a good measure of an area’s potential to meet livestock’s nutritional requirements. The fact that increasingly more people live in areas with higher grassland quality is also indicative of the fact that this relationship is likely due to human behavioral responses to grassland conditions (see Figure 35; while a linear curve explains 20 percent of the variation in rangeland population density, an exponential curve explains even more of the observed variation, R2 = 0.3491; unpublished analysis). It seems that population density does not so much impact grassland quality as the reverse: grassland quality is a limiting factor that impacts (or constrains) pastoralists’ general distribution on the rangelands. This finding does not mean that various other factors usually related with population density do not affect the grasslands – but in the present analysis, such factors (e.g., the demand for livestock products, and hence livestock populations and overall grazing pressure) are more likely to be included in the following, independent variable termed non-local influences.
In contrast to grassland quality, biodiversity is greatly impacted by population density on the Tibetan grasslands. From wildlife’s point of view, population density is a straightforward measure of human disturbance. This disturbance can be of two varieties, the simple presence (geographic proximity) of pastoralists, or the practice of various non-pastoral activities such as hunting (it has already been shown that pastoralism itself does not directly impact grassland quality or biodiversity). Based on the literature, it is safe to assume that both forms of disturbance affect wildlife on the Tibetan plateau (Schaller et al. 1988, Cai et al. 1990, Schaller and Liu 1996, Schaller 1998, World Wide Fund for Nature 1998, Harris et al. 1999).
Non-local influences (back to top)
Although Tibetans do occasionally hunt wild animals (Hedin 1925, Ekvall 1968, Vigoda 1989, Geoffrey 1993), it generally has been non-local people that poach the most wildlife. In the 1950s, for example, army units were ordered to meet large hunting quotas, and wildlife was decimated in most areas of the province (Qinghai People’s Government 1951, ‘Open up…’ 1956, Cai et al. 1990, Becker 1996; also personal interviews with several pastoralists). Even today, poaching poses a huge threat to wildlife across the entire plateau (Miller and Schaller 1996, 1997, Schaller 1998, World Wide Fund for Nature 1998). It is this and other activities undertaken by outsiders that are included in the indicator of non-local influences examined in this chapter. Not only are these activities generally detrimental to wildlife populations, they also negatively impact the grassland vegetation, probably due to increasing demands for livestock products made by the non-local population. Such market demands provide an incentive to increase overall livestock production, and hence grazing pressure on the grassland. The presence of non-local people in an area equally is representative of a more general increase of external influences in the area, influences that usually include a strong impetus for change in many different areas, from economic development and the adoption of more modern production methods (and, conversely, the abandonment of traditional practices) to education and literacy, health care, and the rapid development of basic infrastructure. The merits and drawbacks of each of these areas of change are extremely varied, and any evaluation also depends on the observer’s own development perspective. Each area would therefore require an extensive study in its own right. However, at least from an ecological point of view, it already is clear that non-local influences (as measured by the proportion of each county’s total population comprised by non-local people) have to date tended to be more negative than positive. Fortunately, though, there are many indications of change at higher (national) government levels (Drake 1997, Yan 1997, Environment Science and Technology 1998d), and new perspectives on the environment and development will eventually trickle down to local leaders and decision-makers as well. In the meantime, similar to the development history of many other developing countries in the world (Bennett 1988, Galaty and Johnson 1990, Germeraad and Enebisch 1996, Sheehy 1996), the current direction of change introduced by external forces is to limit pastoralists’ mobility and to increase overall ease of governance by promoting more sedentary ways of life on the Tibetan plateau. But as Fernández-Giménez and Erdenebaatar (1995) explain, “to maintain sustainability in the pastoral system, national land policy and its local implementations must allow for and encourage movement in response to spatial and temporal variability in resources. In other words, herders must not be excluded from extensive pasture resources.”
Season of use (back to top)
An important characteristic of traditional nomadic pastoralism in most areas of the Tibetan plateau is the use of distinct seasonal pastures (Lattimore 1940, Ekvall 1968, Spooner 1973). Although pastoralists may spend over two-thirds of the year in the winter pastures, it is the summer pastures that are most crucial for livestock to recover from the previous winter and to prepare for the next. Summer pastures are also the most easily degraded (Lang et al. 1997). However, most development activities in Qinghai almost exclusively target the winter pastures, such as seeding plots to grow additional winter forage and building livestock shelters, and these activities are often to the exclusion of any regard for the condition of summer pastures (Cincotta et al. 1992). Furthermore, any grassland that can be used in winter will be used in winter (Su 1993), mainly because of the closer-to-home focus on livestock, rather than on grassland conditions, adopted by most pastoralists and by many government leaders alike. This focus on livestock has translated into a disproportionate concern for winter pastures because livestock are most vulnerable in winter. Most development activities have therefore aimed to improve the over-winter survival of livestock, including the classification of as much grassland as possible as favorable for winter grazing. In practice, this means that more modernized areas are likely to have a larger proportion of winter pastures or pastures grazed in both summer and winter (i.e., former winter-only pastures, equivalent to the unfenced pastures examined in the previous chapter), while less developed grassland areas are likely to have a larger proportion of summer pastures. The proportion of summer grasslands can thus be used as a measure of relative grazing seasonality because as much grassland as possible is now grazed in the winter season (in order to increase the over-winter survival of livestock), and any land that is grazed in winter, even if also grazed in the summer, is generally classified as winter grassland (Su 1993). The larger the area of summer pastures in a county, therefore, the more seasonal the character of pastoralism in that county, and to a lesser degree, the greater the overall pastoral mobility and flexibility.
In the present study, overall grassland quality was found to be significantly higher where more grassland is grazed in summer only. This result has direct implications for current development priorities and plans in the province, and suggests in particular that more attention should be given to the conservation of summer pastures as well as winter pastures. Biodiversity, on the other hand, is lower in counties where there is more summer pasturage, but this was found to be the result of the preference of wildlife for undisturbed grassland habitat, as opposed to summer pastures that are used more extensively by pastoralists. Clearly, most mammals are common only in remote, relatively inaccessible mountain areas (Harris et al. 1999), yet these areas also tend to comprise the pastures most suitable for summer grazing by livestock (i.e., pastures in the high mountains, in contrast to the lower plains or valley floors grazed during the long cold season). This suggests that, while the overall area of grassland used for summer grazing should be increased where possible (in order to reduce the amount of overgrazing by overstocking small summer pastures, and thus to improve overall grassland quality), specific measures must also be taken to ensure the protection of wildlife in these areas. Protective measures should include the establishment of some core zones in high mountain areas where wildlife cannot be disturbed.
These present findings justify seasonal use of the grassland and other aspects of spatial mobility, not only as long-term effective strategies for pastoralists, but also as judicious techniques for managing and conserving alpine grassland quality and biodiversity in one of the world’s harshest environments, the Tibetan plateau.
The way forward (back to top)
The alpine grasslands of the Tibetan plateau are characterized by a semi-arid climate and variable primary productivity. Precipitation in Qinghai’s grassland areas ranges from around 200 to 700 mm per year (Jing 1986; Carey 1996). In the Qinghai Lake area, between 1961 and 1975, average grassland productivity ranged from 607 kg/ha in dry years to 1,449 kg/ha in wet years (Western Resources and Environment Research Center 1994). Local pastoralists have traditionally capitalized on the variable climate by adopting a flexible production strategy, namely nomadic pastoralism (Spooner 1973, Galaty and Johnson 1990, Wu N. 1997). The key feature of pastoral livelihoods is their inherent mobility – instead of working fields of land, pastoralists work “fields” of livestock, or “fields on the hoof” (Ekvall 1968). Livestock can be moved between pasture as climate and annual productivity warrant, to capitalize on regular seasonal pastures in good years and on more distant pastures in poor years. Mobility thus allows pastoralists to “maintain within a wide geographical region a total livestock population far greater than that which could be sustained … by independent herds operated separately on small plots of land” (Thompson and Wilson 1994). Many bureaus and agencies in China, however, consider traditional practices as backward, including the use of distant seasonal pastures, and herders are encouraged to “turn to modern production methods” (Xie 1997) and to abandon all aspects of nomadism as rapidly as possible.
This analysis of sustainability (via grassland quality and biodiversity) in Qinghai’s alpine grasslands has shown that in areas with relatively large winter pastures, poor grassland quality may result from the increased concentration of livestock on the proportionally smaller summer pastures. A positive feedback loop can develop if livestock enter the winter in poor condition because of insufficient grazing during summer, and pastoralists perceive a need to further invest in and enlarge their winter pastures to avoid starvation of livestock in late winter; the next year, livestock are grazed on even smaller summer ranges and overall intensity of grazing on the summer range increases, causing even more grassland degradation. Thus, livestock in higher numbers on increasingly degraded land put on less weight during summer, the only growing season for grassland resources, and concomitantly enter winter in ever poorer condition. This completes the vicious cycle. However, even though the poor condition of summer pasture is an underlying cause of livestock starvation in winter, it is still ignored by most decision-makers.
Policies in the 1960s and 1970s aimed at raising the standard of living of pastoralists in Mongolia (Mearns 1995) and Inner Mongolia (Li et al. 1993; Williams 1996) closely resemble those of Qinghai today. In both instances land use was intensified, herd mobility was reduced, and entire grassland ecosystems were degraded. According to Ho (1998), attempts to fence grasslands are especially ill-suited to the ecosystem’s highly variable productivity (e.g., with periodic snow disasters, such as the snowstorms that covered Zhiduo in October 1999; Disaster Information Center 2000), and the grasslands would “benefit more from flexible arrangements than from rigid ones.” Ho continues: “It should be no surprise that the experiments for dividing the rangeland into delimited plots have failed.” Several development practices current in Qinghai today (e.g., building livestock shelters, increasing winter forage production) also aim to maintain local livestock populations artificially at otherwise unsustainable levels (Cincotta et al. 1992, Fernández-Giménez and Erdenebaatar 1995), and privatization in general “can be less meaningful for good resource management than other factors, such as secure tenure, equitable access to community resources, and meaningful institutional supports in the form of credit, production services, and legal protection” (Williams 1996).
Promoting land use intensification and reducing herd mobility are the result of a short-term livestock management perspective. For example, capital investments in “grassland construction” projects in one county in southern Qinghai have focused almost entirely on livestock, not on the rangeland, in an attempt to increase livestock over-winter survival. Yet, over a decade of construction has failed to achieve an increase in total livestock numbers while the grasslands continue to deteriorate (Shu 1993; Dari county government, personal communication). In many counties in Qinghai, grassland degradation has become so severe that many pastoralists have been forced to leave their family pastures to search for better land or to beg for a living (Guoluo and Yushu prefecture governments, personal communications), and many pastoralists have even been termed “ecological refugees” (Lang et al. 1997).
In contrast to a livestock management perspective, one should instead focus more attention on the grassland biome as a whole and on its long-term sustainable utilization. Adopting this perspective, the underlying resource, the grassland ecosystem, comes first, and utilization patterns should be based on local ecological conditions instead of perceived development goals by non-local forces. In practice, this means that there are limits to development on the Tibetan plateau and that grassland quality (and the integrity of the ecosystem as a whole, including the plateau’s biodiversity) must be maintained. The grasslands themselves must therefore be managed as well as the livestock. Such a perspective also calls for a long-term approach to be adopted in development planning.
Although such a (partial) shift from livestock to grassland management does not ensure that all needs will be met – resources are still finite – it may be the only way for development to be sustainable and to protect the native biodiversity of Qinghai’s alpine grasslands for future generations. Furthermore, expertise and financial resources that are currently being directed at counter-productive activities could be re-directed to work in concert, not at odds, with local ecological conditions. Finding new ways to use natural resources, or to diversify the economy, is already an explicit goal in China. The present analysis, based on local government bureaus’ own findings, simply highlights the ecological need to find such alternate development mechanisms. Instead of investing effort and resources primarily on winter pasture areas and attending almost exclusively to livestock, it would be best to focus more attention on the underlying rangelands. Traditional land use patterns should also be examined further to determine what aspects might be incorporated beneficially into future pastoral development plans and activities.
_________________________
Chapter Six
Rangeland Utilization, Grassland Quality and Biodiversity in Alpine
Grasslands: A Regional Analysis
6.1. Introduction
6.2. Methods and Results
6.2.1. General methods
6.2.2. Grassland quality (GQ)
6.2.3. Biodiversity (BD)
6.2.4. Rainfall (RF)
6.2.5. Population density (PD)
6.2.6. Non-local influences (NL)
6.2.7. Percent pastoralism (PP)
6.2.8. Season of use (SU)
6.2.9. Principal components analysis
6.2.10. Multiple regression analysis
6.3. General Discussion
6.3.1. Rainfall
6.3.2. Percent pastoralism
6.3.3. Population density
6.3.4. Non-local influences
6.3.5. Season of use
6.3.6. The way forward