Introduction

Continuous interactions between people and natural systems can result in the formation and development of coupled human and natural systems (Liu et al. 2007). Traditional approaches of managing resources and ecosystems have mostly relied on instrumental or technical end-of-pipe solutions. However, these approaches cannot adequately deal with different stakeholder perspectives in a complex ecosystem (Pahl-Wostl et al. 2008). In most cases, resource management is related to policy science and social learning. The introduction of the term “governance” signaled a change in thinking that places a strong emphasis on social learning as an essential element of policy development and implementation. Social learning can have a critical regulatory role in the effectiveness of ecosystem management (Folke et al. 2005; Liu et al. 2007; Pahl-Wostl 2007). People generally develop social learning through their historical and cultural heritages. This is especially true in the grassland ecosystem of the Tibetan Plateau where changing forms of social organization can affect the sustainable development of a coupled system.

In the late 1970s, Chinese agrarian policy instituted major reforms in the agricultural and animal husbandry sectors. This involved decollectivization and a return to a market-oriented economy. Communes were dissolved and livestock was divided among former nomads. Central planning was replaced with management at the household level, and once again families became responsible for livestock and the marketing of their products. Pasture tenure remained state property but was gradually contracted out to households on a leasehold basis (Manderscheid 2001). The assignment of rangeland use rights to individual households and the establishment of exclusive boundaries in rangelands via fencing and stocking rates for individual household pastures were mandated, monitored, and enforced by Grassland Supervisor Stations (Ranks 2004). About 70% of China’s usable rangeland was leased through long-term contracts by the end of 2003; 68% being contracted to individual households and the rest to groups of households or villages (Yao 2003). The rationale for this form of privatizing rangeland in China was the assumption that open access of privately owned livestock to common rangeland has led in the past to severe rangeland degradation. To eliminate open-access, property rights needed to be defined. The assumption has indeed been widely accepted in China (Gadgil et al. 1993; Yan et al. 2005) and is based on the “Tragedy of the Commons” theory proposed by Hardin (Ashenafi and Williams 2005; Wang 2009a, b). In Hardin’s paper (1968), it is in the individual’s interest to maximize their utilization of common land even if this result in its degradation, because the individual gains any short-term benefits and the costs are spread over all those with right of access. A suggested resolution to this issue is to make sure all land is retained as private property, which encourages the owner to utilize resources sustainably and to improve the productive motivation of herders.

At the beginning of the grassland contract system on the Qinghai-Tibetan Plateau, contract rules were clearly defined by local governments. Namely, winter grassland was to be contracted to single-households, and summer grassland could be contracted to multi-households, but the combined number of households could not exceed three. Nevertheless, many multi-households have been operating on winter grasslands in defiance of the grassland contract system policy. The objective of this study is to compare social, economic, and ecological costs and benefits between multi-household and single-household grassland management systems, and to propose a potential alternative management model for grassland areas in the Qinghai-Tibetan Plateau and similar areas.

Study Area

Maqu County

Maqu County (101°–102°E, 33°–34°N) is located in the eastern Tibetan Plateau at the junction of Qinghai and Sichuan provinces (Fig. 1). The altitude ranges from 2,900 to 4,000 m with an annual rainfall of 450–780 mm. The annual average temperature is 1.8°C with a low of −10.7°C in January and a high of 11.7°C in July. The highest atmospheric temperature during the growth season can reach 28.9°C, and there are, on average, 270 frost days annually. The grassland area covers about 87 × 104 ha, and 59% of it is alpine meadow. Grassland of Maqu acts as a natural ecological barrier for the Yellow River.

Fig. 1
figure 1

Geographic location of Maqu County in China

Full size image

The Yellow River originates in Qinghai Province and has a 38.9 × 108 m3/s water flux. It flows around Maqu for about 433 km, and again enters Qinghai with a 147 × 108 m3/s water flux. Runoff from Maqu to the Yellow River is 10.81 × 109 m3, accounting for 58.7% of the total runoff from the source area of the Chief Yellow River, or 18.413 × 109 m3, and for 1/6 of the total runoff of the Yellow River (Wang 2009a, b). This area is well known as the ‘water tower plateau.’ A number of Tibetan ethnic minorities lives in Maqu. There are seven villages and one town (a sub-administrative unit). In 2004, the county’s total population was 42,900, including 33,313 herdsmen, accounting for 77.7% of the total population. The production value related to animal husbandry is 119.79 million RMB which constitutes 93.9% of the agricultural production value of this area.

Historically, herders in Maqu have been engaged in nomadic collective pastoral activities and are referred to as the “the nationality on horseback.” Under long-term natural and artificial selection, the yak, Tibetan sheep, and Hequ horse are the major livestock species well-adapted to the harsh local environment. In recent years there have been significant changes as a result of the grasslands policy of “increasing sheep, stabling cattle, and decreasing horses,” Following road improvements and increased farmers’ incomes, the number of horses has gradually declined. The primary modes of transportation have changed to motorcycles and farm vehicles. Prior to implementing the reform program and more open policies, this area was dominated by a small-scale peasant economy. With reform, most production has focused on livestock raised for market sale.

Currently, 90% of herders’ income is derived from the trade of livestock and their byproducts. The remaining 10% of income comes from the sale of medicinal materials and other goods. Previously, sheepdogs were only used to prevent wild animals from attacking livestock, but now the use of sheepdogs has expanded to include pasture management by keeping unauthorized livestock out.

Evolution of Two Different Grassland Management Patterns

In 1995, a grassland contract system was developed in Maqu. By 2003, 96% of all available pasture in the county had been contracted to households; giving 30 years of use rights for fixed-term pasture. Although all winter grasslands were contracted to single-households as required by the contract articles, many herders were unwilling to separate from others because of their traditions of kin-based nomadic movement as well as their more recent dependence on the collective lifestyle (Cao et al. 2009).

Given the circumstances, the local government made some necessary modifications to cater to local conditions during the implementation of the policy. Those herders who were unwilling to voluntarily operate as a single-family formed a community to implement joint household operations, although pasture tenure was nominally assigned to a household. In some circumstances, this got the acquiescence of the Government. However, because of the difference in policy enforcement across the county, the MMP situation was also different. Generally, in suburban areas implementation of the policy was comparatively thorough. There were two reasons for this. One was that the location was geographically adjacent to the administrative headquarters. This area was heavily targeted for fulfilling the policy of the grassland contracts. The other reason was that some herd owning households who settled in the suburban zone chose a single-household business pattern for the convenience in renting grassland. According to our survey, around 50% of herd owning households in the suburban zone were engaged in single-household business both in the winter and summer pastures. In remote areas, the implementation of the official policy was relatively lax and flexible; thus permitting more than 80% joint business in winter pastures. In summer pastures the rate of joint-business households appeared to be over 90%.

Thus two grassland management strategies developed: multi-household form of management with grassland jointly held by two or more households without fencing (MMP) and single-household management with fences (SMP) (Fig. 2). The MMP reflects the traditional pastoral livestock production model. By combining the cultural inheritance of herders with the flexible implementation of policy, an integrated management model emerged that matched the local culture and natural conditions. This grassland co-management model combines indigenous and scientific strategies, allowing for more flexible policy interpretations and locally appropriate adjustments (Richard et al. 2006)

Fig. 2
figure 2

Photos showing grazing pressure, grassland fencing and material inputs, labor input, and a guard sheepdog. a Grazing pressure in SMP system. b Residence and fence in SMP. c Fencing material for SMP. d Mutual aids in MMP system. e Sheepdog guard

Full size image

Pre- and Post-Grassland Contract Grazing Patterns

Traditionally, what was termed “grazing along with water and grass” typified local animal husbandry production systems. Pastures for each of the four seasons were distinct, with frequent movement among them. However, with the implementation of grazing tenure laws, the scope and space of available grassland was reduced. In addition, with barrier fences separating each household’s allotment, herders lost the ability to transit seasonally. Currently, with the exception of a few single-households without a winter or a summer pasture, most households have summer and winter pastures, with the majority of the time spent at the winter pasture (usually the main residential area). Compared with previous nomadic patterns, the frequency with which herders migrate between pastures has markedly decreased. Previously, migration activities usually started in the spring, and herders moved without long halts among the four seasonal pastures. Currently, however, migration often is simply between summer and winter grazing areas. During the growing season in June, herders transfer animals from winter grasslands to summer grasslands and return to winter grasslands in August. In September, they go to summer grasslands and then return to winter grasslands in October. When the era of pasture contracts began, the contracted winter grasslands of both multi-households and single-households were enclosed by government subsidized fences. Few summer pastures were fenced because of difficulties in transporting materials. In recent years the government no longer provides subsidies. To the extent that the MMP households are more mobile than the SMPs, then MMPs can be thought of as approximating the older pastoralist heritage.

Research Methods

Preliminary Interviews

Our 1999 survey indicated that the MMP strategy was more frequently adopted by herders under the stringent grassland management system. In early 2005, we interviewed a limited number of herders and found that those herders who had worked in the MMP were very satisfied with this form of collective management. Several SMP households also indicated a preference for multiple household arrangements.

Formal Interviews

To determine which management pattern the herders preferred to adopt overall, between May and August 2005, 559 questionnaires were completed using randomized cluster sampling. Before the formal investigation, six interviewers were trained; these included three graduates from Lanzhou University and three local translators from local elementary schools. All interviews were face-to-face as this method generally leads to the highest number of survey responses and usually provides more control over the order of questions in the questionnaire. It also can offer practical advantages in maintaining respondent motivation and allows the use of graphic supplements (Arrow et al. 1993; Carson et al. 2001; Nunes et al. 2004). From the 559 questionnaires, 32 were eliminated because of incomplete answers, one was lost, and one interviewee was underage. Therefore, 525 questionnaires were treated as valid, accounting for 94% of the total.

Initially, we planned to estimate income by number of livestock per household as well as household size, grassland area, the carrying capacity of grassland as approved by the local government, and also the average projected marketable livestock. However, we found that many herders were evasive or deliberately distorted the number of their livestock. Historically, the number of livestock served as a benchmark by which herders were judged to be “rich” or “poor” sometimes leading to social consequences. Even today, herders still hold to the convention that a family in possession of at least 300 sheep and 100 cattle is considered to be relatively rich. Respecting local customs, we aimed to establish a trusting relationship with interviewees, and adopted an indirect method to estimate their income. Herders’ income was calculated based on the sale of livestock and their byproducts, the sale of medicinal herbs, and their other sources of income. At the same time, herders’ expenditures were estimated based on tuition fees, medical care expenses for family and livestock, as well the expense of purchasing forage for livestock during the winter. Although this method was indirect, it proved to be effective and feasible (Cao et al. 2008).Table 1 lists the social characteristics of the interviewees, and Tables 2 includes the income and expenditures of the interviewees.

Table 1 Social characteristics of the interviewee
Full size table
Table 2 Annual income and expenditure of the investigated families
Full size table

To compare the differences between the effects of the two grassland management systems, an integrative survey was conducted from May 2008 to July 2009, using anthropology and field experiment methods. In this manner the comprehensive benefits assessed include social, economic, and ecological benefits.

Data Analysis

In our ecological benefits analysis, we chose biomass, vegetation cover, richness of plant species and the heights of plants as a means to evaluate grassland quality because these factors are sensitive markers grassland degradation. Reed et al. (2007) used these variables to assess the change of grassland quality. The results showed that these factors were useful to help pastoralists make an informed short-list that could be used sensitively and reliably to detect long-term grassland degradation. Biomass data consisted of functional plant groups including sedge, grass, poisonous weed, and legume biomass. The fresh samples were oven dried at 80°C, and each biomass type was expressed as its dry weight. Samples were dried to a constant weight (48 h) and weighed to a thousandth of a gram. During weighing, biomass was recorded as zero when the biomass of a functional group was less than 0.01 g. Vegetation cover was determined by on-site visual estimates. Species diversity was determined by artificial recognition. Plant height was measured manually. Species richness was evaluated with the number of species in a quadrat.

The data were analyzed using SPSS 15.0 statistical software. ANOVA was used to determine differences in the parameters involved in the study at a level of P < 0.05.

Results

Of the 525 interviewees, 489 were male (Table 1), a reflection of local history where the social status of woman has been low and the concept of male superiority prevalent. The mean age of the interviewees was 38.5 years, ranging from 18 to 75. Age groupings of old, middle-aged, and young might reflect differences in respondents’ thoughts about the two grassland management patterns. Generally, the older generations have extensive experience in grassland system changes and tend to have strong views on the different grassland management patterns. Of the 525 interviewees, 299 were illiterate, 188 had primary school level education (many having only 2 or 3 years of education), 30 had junior middle school education and only seven had received a high school level education or above. Statistical analysis indicated that income and expenses were positively correlated (r = 0.272, P < 0.01). Herders usually did not have money saved unless their annual income was over 35,000 RMB.

Of the 525 interviewees, 417 preferred MMP, and 108 preferred SMP. The reasons for choosing vary. After excluding reasons that were chosen less than 10 times, results indicated that people preferred MMP because they can help each other when they have difficulties, livestock can graze larger areas, and people can access grassland resources including water more equitably. Out of eight choices, about 40% of the interviewees selected these three choices together. Interviewees who preferred SMP did so because this system involved less conflict over grazing and single family grassland management was personally convenient. Of the 108 interviewees, nearly 77% selected these two choices together. From these results, we conclude that interviewees who selected MMP were more concerned with fairness and sustainable use of grassland resources, while those who selected SMP were more concerned with their immediate family. Using chi-square statistical tests (Wu and Wang 1996; Xue 1997), we found that no social or economic variable significantly affected MMP or SMP selection.

Economic Benefits

Generally, herders involved with MMP assist each other in livestock production and management. This mutual assistance can enhance the capacity to cope with risks and reduce production costs by sharing labor and tools and by assisting each other with sheep shearing (Fig. 2). Timely sheep shearing is extremely important for maintaining the health of sheep. Although these benefits are somewhat difficult to quantify, the costs of fencing and sheepdog breeding saved under the MMP can be measured.

Fence Cost Savings

From our investigations, we found that a single household would spend 20,000 RMB (average length of the fence for each household was about 2,000 m, and the average cost of fence per meter is 10 RMB) on constructing fences to establish boundaries between the grasslands of other households. In Maqu, fences last about 10 years (Fig. 3). On average, fences depreciate at approximately 2,000 RMB per year for each household. Therefore, 2,500 RMB (500 RMB used for fence maintenance) would be saved each year per household if there was no fence construction.

Fig. 3
figure 3

Different fencing patterns of MMP and SMP

Full size image

Figure 3 shows the difference in expenses among different household union sizes. One boundary line is set up between two households, and two fence lines are needed to separate three households that are adjacent to each other. This pattern is extended with increasing numbers of households (Fig. 3). Therefore, N-1 enclosure lines would be needed when there are N households settled in the same location. Here, we simply regard the cost of one enclosure line to be 2,500 RMB. Therefore, N united households would save (N-1) × 2,500 RMB per year for not fencing. There are 7,406 households in Maqu, and 4,575 households were investigated in our study. Eight-five percent of the investigated households were in the MMP, with the average size of a multi-household community being 5.42 households. From this, we can infer that there are 1,161 (7,406 × 0.85/5.42) households that belong to MMP and 13millon RMB [1,161× (5.42−1) × 2,500] would be saved each year as a result of the implementation of this management method in Maqu.

Sheepdog Breeding Cost Savings

Local herders often raised and bred sheepdogs to ensure the security of their livestock (Fig. 2). Sheepdogs appear to be more critical to herders in SMP. Generally, single-households lived in scattered places and the risk of their livestock being stolen was higher than that of multi-households. This was especially true in summer pastures where herders faced more threats from wild animals and cross-border grazing of livestock belonging to other herders. Thus, single-households had to raise more sheepdogs to ensure stable livestock production. We made a systematic survey of sheepdog population and expenditures in MMP and SMP (Table 3).

Table 3 Number of Sheepdogs in MMP and SMP
Full size table

The average number of sheepdogs per household in MMP was 2.14, while the average number of sheepdogs in SMP was 4, 1.87 times greater than MMP (Table 3). Based on data from interviews, we found that each sheepdog eats 12 bags of dog food per year. At the time the market price of one bag of dog food was 50 RMB. Therefore, it would cost at least 600 RMB per year to keep one sheepdog. Projecting this number forward, single-households would spend an additional 1,116 RMB [(4−2.14) × 600] more than multi-households on sheepdogs each year. As described above, 85% of households belonged to multi-household in Maqu, for a total of 6,295 households in the MMP system. Therefore, the reduced number of sheepdogs in MMP would save 7 million RMB per year.

By comparing income and product cost per household between the two management patterns, we found that herder households in MMP had higher incomes and lower production costs than SMP households. Average annual income per household in MMP was 28,000 RMB, while annual production costs were 7,000 RMB. For SMP, the average annual income per household was 25,000 RMB but annual production costs reached 10,000 RMB. The disposable income of an MMP household each year was 21,000 RMB, higher than the 15,000 RMB of an SMP household. The greater economic benefit was mostly responsible for the collective mutual aid and high grassland productivity in the MMP.

In summary, we concluded that about 20 million RMB were saved per year from implementation of the MMP in Maqu. However, many economic benefits of the MMP system are hard to quantify. For example, in 2008, 18 cattle were stolen from one family at their summer grasslands, although they were found several days later thus recovering nearly 50,000 RMB in potential losses. According to one herder, the stolen cattle found so quickly because an agreement exists among herders in the MMP system in which all households must search for lost livestock. Generally, each herder has unique interpersonal relationships. With the help of this network of relationships, stolen livestock can usually be found quickly. It is much more difficult to locate livestock stolen in the SMP system because herders lack the social networks present in the MMP system. Usually single-family residents only had a single relationship as a result of the scattered residences. When they experience problems such as theft of livestock, they cannot obtain timely on-site aid.

Social Benefits

The MMP system is important for preserving the indigenous heritage. Local herders have a long history of nomadic grazing and livestock management. Such knowledge provides invaluable insights for sustainability and to conserve biodiversity (Gadgil et al. 1993) which cannot be replaced by modern scientific knowledge. Furthermore, many traditional activities, such as singing, dancing, painting, and horse racing are important for local herders as these activities strengthen their network of interpersonal contacts. Such knowledge and activities seem threatened under the SMP system.

As previously noted, the MMP system is more focused on equality and sustainability. To balance the relationship between livestock numbers and grassland capacity, strict regulations were enacted to limit the stocking rates within the MMPs to which each household has to adhere. Inspections of livestock numbers are made once or twice each year according to the household units. Besides the effects of policy, strategic modification of herders’ behavior acts as a social indicator. The herders choose to form a community for production and business on their own. When there is no conflict with the assigned number of livestock approved by the local government, herders can make adjustments according to the availability of grass. For example, in years when grass is plentiful, the number of livestock can be increased, and when grass production is limited, livestock numbers can be reduced. Through these dynamic adjustments, the overall carrying capacity remains basically in line with the livestock capacity authorized by the government. For single-households supervision and restraint is largely lacking except for the regulation of the government. Therefore, single-households rarely make such dynamic adjustments.

Further evidence of social benefits can be found in the MMP system, such as contractual transfer of grassland rights. For example, if household A has fewer people but more grassland than does household B, the amount of livestock of household A may not reach the contracted number. Therefore, grassland from household A could be rented to household B. By this system of mutual aid, household B can raise more livestock. By sharing grassland within the MMP, a better balance between available forage and livestock can be achieved. In this internal system, each household in the MMP can make reasonable use grassland resources and can achieve sustainable development by balancing livestock and forage. However, for a single household, it would frequently be hard to achieve the same degree of grassland land circulation because of the lack of good information about availability, fencing barriers, the great distances and other factors that affected fair use of grassland resources. On the one hand, those with smaller households but sufficient grass would not adequately utilize their grassland resource because labor shortages can limit the number of livestock. On the other hand, those households with ample labor but insufficient grass resources are similarly limited. During our survey, we observed that some pastures were underutilized while others had an excess of animals grazed.

The MMP supports development of infrastructure. Generally, the MMP residential sites are concentrated which facilitates the provision of education, medical care, communications, and transportation. During our investigation, we found that drinking water, large-scale utilization of solar energy, and recreational facilities projects were more complete in the MMP than in the SMP. The ‘yi te si hua’ project to be launched in 2010 sets limits on the size of multi-household units. Only MMP systems that include over 10 households are qualified to engage in the “professional layout, industrial management, standardized production and skill training that focus on unique yak and Tibetan sheep.”

The MMP system helps maintain good relationships among neighbors by requiring fewer fences. In general, with more fences, conflicts increase between neighborhoods. There are three main reasons for this: First, pasture conditions can vary on either side of a fence and livestock often cross for better forage, causing disputes. Second, fences result in privatization of public water sources resulting in difficulties among households in providing drinking water to livestock. Neighborhood conflicts result from competition for water. Third, fences clearly define grassland boundaries and cross-border grazing and traffic are likely to cause disputes. As one herder stated, “The border, particularly where fences are set up, not only cuts off interactions among livestock, but also interactions between people are cut off.” During the investigation, some of these phenomena were witnessed and some reported by informants. But inside the MMP, where there are no fences, water resources are shared resulting in less conflict and more harmonious relationships among neighborhoods.

Finally, the MMP system promotes the appropriate use of rural resources and reduces risk. As Manderscheid (2001) stated: “Reduction of pasture tenures in favor of industrial projects, exploration of natural resources, nature reserves, and sedentarization projects are among the causes which force pastoralists worldwide to give up a nomadic lifestyle.” Only a well-organized community can successfully resist unreasonable exploitation and use of resources. For example rules regarding no long hair, no stealing, and no gambling greatly influence members within the MMP system. Joint-production household groups can act with enhanced bargaining power; therefore they can effectively resist incursions with negative impacts.. On the other hand, a single household operating independently is weak in terms of its bargaining position. In Maqu, natural disasters (e.g., snow storms, freezes, hail) occur frequently. In serious disasters, some households may lose all their livestock and capital needed for later development. If this happens in the MMP system, those who have lost their livestock are often helped by others and the social benefits of MMP are available. In addition, sound welfare services were common within some MMP systems. For example, in an MMP community we investigated, there was a certain amount of pasture land reserved as commons. Generally, these commons were used only for sheep or old, weak cattle in winter. Those who used the commons must reserve a certain number of livestock for collective compensation and these were used as welfare funds. With these welfare funds, poorer households within the MMP could be assisted or even households outside the MMP could be helped.

Ecological Benefits

In this study, ecological benefits were defined as the quality of the grassland. To explore ecological benefits of the two management systems, a field experiment was conducted using conventional ecological analytical methods. Two criteria were used to select households to be included in the sample. One criterion was that the MMP and the SMP had to use a similar grassland management model. For instance, they had to transfer livestock from winter grassland to summer grassland on the same date, and they had to have the same timeframe for doing management tasks. The other criterion was that all samples had to belong to the same grassland community type. Generally, summer grassland is often remote and difficult to reach by vehicle. Our test was conducted in winter grassland after the herders transferred to summer grassland. As previously mentioned, the effectiveness of enforcing pasture contract policies in Maqu varied among locations, which resulted in uneven distribution of MMP and SMP strategies. Few places had both systems.

In this study, 30 MMP and 30 SMP samples were selected in 2009. Within each sampling site, grassland plant communities were investigated within three 50 × 50 cm quadrats with three sampling replicates statistically (Table 4).

Table 4 Grassland quality of MMP and SMP
Full size table

In the pasture investigated, sedge plants included Kobresia capillifolia, Scirpus distigmaticus, Carex pseudo-supina, and Kobresia humilis, and grasses included Poa pratensis, Elymus nutans, Festuca sinensis, and Stipa aliena. The data indicated that the biomass of the sedge group was significantly higher in the MMP than in the SMP (P < 0.05). In addition, in sedge plants, the height of Scirpus distigmaticus and Kobresia humilis was significantly (P < 0.001) different between the two patterns. For grasses, Festuca sinensis was significantly higher in MMP than in SMP (P < 0.001) (Table 5). Although there were no significant differences in average biomass, average vegetation cover, average vegetation height, and average plant biodiversity of each quadrat between the two patterns, sedge biomass and the height of some important grass were significantly different between the two management patterns. Since sedge and grass are the main edible forage and dominant graminoids in the grasslands of Maqu, some indicators from our analysis inferred that SMP pasture was likely more degraded than was MMP pasture (Table 4). This was mainly because edible forage had been grazed more intensively in the SMP during the seasonal transition of grazing activity. Remaining stubble of edible grass was low, which was not conducive to self-recovery of grassland vegetation.

Table 5 The height of plants under the MMP and the SMP
Full size table

Discussion and Conclusions

The MMP appears to be a mixed grazing pattern, somewhere between nomadic and fixed grazing. It encompasses many of the advantages of nomadism such as grazing on a large scale, resilience to natural disasters, and cooperative production. By grazing on a large scale, livestock can use abundant low-quality food and typically create frequent but small disturbances across a landscape. There may also be efficient seed dispersal effects by transporting soil and undamaged seeds over larger distances. Spatially-heterogeneous urine deposition could also increase regeneration sites and soil heterogeneity (Olff and Ritchie 1998). Concurrently, mobility on a large scale allows flexibility in management enabling sound utilization of resources during the annual production cycle and exploitation of patchy and temporally-variable environments. The practice of grazing on a large scale also encompasses a strong conservation ethic that recognizes the idea of sustainable use and high biodiversity of grasslands. This can be essential to favoring multiple use and conservation of different ecological environments (Adriansen and Nielsen 2002; Gimenez 2002; Ladio and Lozada 2004; Spiegelberger et al. 2006). With collective action and cooperative production, social relationships are strong within the MMP, and the community is better equipped to deal with external entities, be they governmental, commercial enterprises, or the market. The MMP system is thus better able to deal with emergency situations than is the SMP system.

Social indicators can be studied over time and used to formulate policy and management decisions (Force and Machlis 1997). Applications of social indicators are valuable for optimizing future management initiatives in Maqu. These include 1) monitoring local social conditions; 2) conducting comparative studies between the MMP and SMP systems over time; 3) evaluating human ecosystem responses to grassland management actions; and 4) providing managers and herders with information for collaborative decision making. In this study, two important indicators are presented as examples of the 525 households in the study area. One is the spontaneous modification by individual herders from SMP to MMP management strategies. This can be considered to be a social learning process that addresses the increasing uncertainty of the grassland ecosystem, and can be regarded as demonstrating that herders have successfully learned from past experiences. The other is the flexibility of policy as implemented by the local government. Based on our results, it can be argued that social learning played an important role in the sustainable management of grassland. Theory shows that social learning is an essential element of policy development and implementation for resource management (Folke et al. 2005; Pahl-Wostl 2007). Joint tenants within the MMP system were better able to utilize social learning from previous experience with livestock production. As a result, herders within the MMP were better able to make up for deficiencies in current policy and were able to achieve higher levels of business effectiveness.

In the study area there no longer is traditional nomadic grazing; hence we analyzed the difference between two grassland use patterns, the MMP and the SMP. To understand these different grazing patterns, some representative herders were interviewed. The results showed that majority of sampled respondents thought nomadic and the MMP patterns were better than the SMP. Far fewer favored the SMP system. Those who preferred nomadism thought that in a nomadic system livestock can graze on a large scale, feed on different forage, and move among areas based on the season. But in most MMP systems, only two types of pastures remain, summer and winter, while in the SMP systems there is sometimes no distinction between summer and winter pasture. Those who preferred the MMP system based their view mainly on the fact that grassland has been contracted to households and the experiences of joint business of multiple households. They believed that it was a wise strategy to choose a grazing system as similar as possible to the nomadic grazing system under the conditions where they cannot return to the past nomadic grazing style. This conclusion came mainly from comparative understanding and practical experience about the MMP and the SMP systems. The preference about the appropriate size of a MMP group was from 8 to 15 households. If it increased over 15, production efficiency would decrease and difficulties in management would increase. Those who liked the SMP system believed that degradation of grassland was caused by natural factors, and not associated with management patterns.

As for the evaluation in this study of the ecological effectiveness, only the aboveground part of the grassland was tested. Aspects such as soil, organic carbon storage (SOC), seed bank, and microbes must be studied in the future to better understand these different grassland management systems. SOC is the most important factor for grassland management as it modifies carbon inputs to the soil, affecting net primary production, root turnover, and carbon allocation between roots and shoots (Ogle et al. 2004). SOC is often related to climate change, so what choices humans make about future land use could have a significant impact on regional and seasonal climate (Feddema et al. 2005).

Until now, it has been widely assumed in China that open access for privately-owned livestock to common rangeland could lead to severe rangeland degradation. Indeed, the Chinese government still regards grassland contracts as an important tool in reversing trends in grassland degradation (Yan et al. 2005). However, there is evidence that suggests the theory is not universally supported (Crépin and Lindahl 2009) and that common ownership in the right social setting may be environmentally beneficial. Grassland contracts have universally caused grassland degradation (Yang 2007). Following implementation of grassland contracts, 90% of grassland has become degraded to varying degrees, regardless of the number of people, amount of precipitation, degree of development, or differences in climate.

To conclude, our results suggested that the MMP system of grassland use is one under which local natural ecosystems and human-cultural characteristics can achieve synchronous coordination and harmonious development. The MMP system not only maximized use of the natural landscape and succession of grassland ecosystems but also was user friendly to the nomadic cultural heritage and its development. The MMP can be regarded as a highly efficient collective grassland management system (Yan et al., 2005). Local communities and governments need flexibility to create tenure regimes that match local cultural and ecological characteristics (Richard et al. 2006), here and in similar areas.