rangeland responses to pastoralists’ grazing management on ......rangeland responses to...

Rangeland responses to pastoralistsrsquo grazing managementon a Tibetan steppe grassland Qinghai Province China

Richard B HarrisAGI Leah H SambergAH Emily T YehB Andrew T SmithCWang WenyingD Wang JunbangE GaerrangBF and the late Donald J BedunahA

ADepartment of Ecosystem and Conservation Sciences University of Montana MissoulaMT 59812 USA

BDepartment of Geography University of Colorado Boulder CO 80309 USACDepartment of Life Sciences Arizona State University Tempe AZ 85287 USADDepartment of Biology Qinghai Normal University Xining Qinghai ChinaEKey Laboratory of Ecosystem Network Observation and Modeling Institute of GeographicSciences and Natural Resources Research Chinese Academy of Sciences Beijing ChinaFThe Centre for Tibetan Studies of Sichuan University Chengdu Sichuan ChinaGPresent address Washington Department of Fish and Wildlife Olympia WA 98501 USAHPresent address College of Forestry and Conservation University of Montana MissoulaMT 59812 USA

ICorresponding author Email rharrismontanacom

Abstract Livestock grazing is the principal land use in arid central Asia and range degradation is considered aserious problem within much of the high-elevation region of western China termed the Qinghai-Tibetan Plateau(QTP) Rangeland degradation on the QTP is variously attributed to poor livestock management historical-culturalfactors changing land tenure arrangements or socioeconomic systems climate change and damage from smallmammals Few studies have examined currently managed pastures using detailed data capable of isolating fine-scalelivestockndashvegetation interactions The aim of the study was to understand how differences among livestock (primarilysheep) management strategies of pastoralists during winter affected subsequent rangeland condition and productivityPlant species composition annual herbage mass and indicators of erosion were quantified during four summers(2009ndash2012) on winter pastures managed by 11 different pastoralists on QTP steppe rangeland in Qinghai ProvinceChina Data came from repeated-measurements on 317 systematically located permanent plots as well as pastoralistinterviews and the use of GPS-equipped livestock Relationships between annual weather variation and herbage masswere modelled using an independent set of vegetation measurements obtained from livestock exclosures Account wastaken of inherent site differences among pastures Annual variation in herbage mass was found to be best fitted bya model containing a negative function of winter-season temperature and a positive function of spring-seasontemperature Accounting for annual and site effects significant differences among pastoralists were found for mostresponse variables suggesting that individual heterogeneity among management approaches had consequenceseven among neighbouring pastoralists Annual herbage mass of preferred plant species was positively associatedwhereas that of unpreferred species was negatively associated with mean sheep density and intensity of use Howeverthe proportion of bare soil an index of erosion and annual herbage mass of unpreferred forbs were found to havepositive relationships with sheep grazing pressure during the preceding winter whereas live vegetation cover andannual herbage mass of preferred grasses were negatively related Thus on a spatial scale pastoralists respondedadaptively to the cover of preferred plant species while not responding to total annual herbage mass Pastoralistsstocked pastures more heavily and livestock used regions within pastures more intensively where preferred specieshad a higher cover However where sheep grazing pressure was high downward temporal trends in the herbagemass of preferred species were exacerbated Pastures that were stocked at a lower density did not experience thenegative trends seen in those with a higher density

Additional keywords China grasslands livestock Qinghai-Tibetan plateau rangeland degradation steppe vegetationTibetan pastoralism

Received 18 May 2015 accepted 1 December 2015 published online 22 January 2016

Journal compilation Australian Rangeland Society 2016 wwwpublishcsiroaujournalstrj

CSIRO PUBLISHING

The Rangeland Journal 2016 38 1ndash15httpdxdoiorg101071RJ15040

IntroductionIn rangelands throughout the world livestock provide food andincome to the majority of the worldrsquos poorest people (Bedunahand Angerer 2012) yet degradation of rangelands is havingdirect effects on the livelihoods food security and way of lifeof pastoralists In addition this degradation can significantlyreduce ecosystem services provided by rangelands at regionaland global scales such as carbon sequestration and hydrologicalregulation and provision Effects of degradation such as duststorms food or commodity shortages and displaced communitiescan be felt over a large geographic scale (Yan 2001 Wang et al2005 Li et al 2013) Rangeland health also affects biodiversitydirectly and indirectly because all native flora and fauna haveadapted to the long-term evolutionary forces that have shapedthese environments Livestock grazing is the dominant form ofland use in arid biomes worldwide including on the rangelandsof the Qinghai-Tibetan Plateau (QTP) in the Peoplersquos Republicof China The QTP occupies 25million km2 ~25 of Chinarsquosarea and an estimated 70 is used by grazing livestock Thesepastures colloquially termed lsquoChinarsquos water towerrsquo are locatedupstream and upwind of upwards of an estimated 20 of theworldrsquos human population (Xu et al 2009 Immerzeel et al2010)

Awareness by Chinese scientists and policy-makers of theimpacts of degradation increased in the late 1990s as severaldisasters occurred including Yangtze River floods that killedthousands of residents downstream and cost billions of dollarsin economic losses the Yellow River running dry increasinglyoften and dust-storms and sand-storms originating in westernrangelands that affected the health and economic wellbeing ofmillions of city-dwellers in Chinarsquos east Although lacking cleardocumentation and differing in specifics scientific papers andgovernment policy statements have generally viewed the QTPas having become increasingly degraded in recent decades(Harris 2010) A frequently repeated statistic is that 90 ofChinarsquos grasslands are degraded to some extent and thatdegradation is increasing at a rate of 200 km2 yearndash1 (StateCouncil 2002)

Causes for this degradation are generally attributed to acombination of over-stocking unscientific livestockmanagementhistorical-cultural impediments to adopting modern livestockmanagement concepts global climate change and excessiveherbivory and soil disturbance from small mammals (Li 1994Chen 1996 Zhou et al 2003 Zhang et al 2004 Li et al 2013)In contrast other investigators have questioned the assumptionof wide-scale grassland degradation on the QTP and where theyagree it has occurred cast their analysis in terms of rapidchanges in socioeconomic systems and alteration of land-tenurearrangements (Miller et al 1992 Levine 1998 Holzner andKreichbaum2001Goldstein andBeall 2002Williams2002Wuand Yan 2002 Banks 2003 Banks et al 2003)

Chinese policy-makers have implemented a variety ofprograms to ameliorate negative rangeland trends (Harris 2010Li et al 2013 Shang et al 2014) some more effectively thanothers As early as the 1950s efforts were begun to exterminateplateau pikas (Ochotona curzoniae) a small-bodied burrowinglagomorph generally associated with poor vegetation cover(Smith et al 1990 Fan et al 1999) Policy during this periodblamed small mammals for causing rangeland degradation

regardless of its precise definition but this may have reflectedthe general Maoist conception of nature at the time (Shapiro2001) scientificwork on pikas began only in the 1980s Researchsubsequently began taking a more nuanced view of interactionsbetween pikas livestock and vegetation after it emerged notonly that pikas were critical parts of the natural environment(lsquokeystone speciesrsquo Smith and Foggin 1999 Lai and Smith2003 Wilson and Smith 2015) but that high density of pikaslikely resulted from rather than caused the rangeland conditionswith which they were associated (Shi 1983 Bian et al 1999Wangdwei et al 2013) Regardless programs of pika reductioncontinue to the present day (Smith et al 2006 Delibes-Mateoset al 2011 Wilson and Smith 2015)

Privatisation of livestock and dissolution of the collectivesystem occurred during the 1980s throughout the QTP and inour study area in 1983 During the 1990s following the successof quasi-privatisation policies in eastern Chinarsquos agriculturalsector similar approaches often referred to as the HouseholdResponsibility System were initiated in the pastoral sector (Yanet al 2005) Echoing widespread concerns about the lsquotragedy ofthe commonsrsquo (Hardin 1968) these programs aimed to encourageresponsible husbandry by clarifying pasture-land tenure at thehousehold level The primary tenets of the government initiativeinvolved increasing the duration of pasture-lease contracts from20 to 50 years subsidising construction of permanent winterhomes fences and livestock shelters and providing plots forgrowing supplemental winter fodder (Richard et al 2006 Wuet al 2012) Government outlays for programs related to theHousehold Responsibility System (often termed the lsquoset of fourrsquoor sipeitao in Chinese (Wu and Yan 2002)) were substantialduring 2003ndash2006 the central government reported investingsome yen71 billion (~US$1 billion at the time) for fencing alone(SEPA 2007) Whether this fundamental reform alleviated orexacerbated negative trends in rangeland condition remainscontentious

A newer set of initiatives began in the early 2000s thatemphasised land protection rather than tenure and responsibility(Yeh 2009) Increased awareness of the consequences ofupstream erosion following the devastating Yangtze River floodof 1998 led to government subsidies encouraging reforestationof cultivated lands that were unsuitable for agriculture (tuigenghuanlin lsquoretire cultivation restore forestsrsquo Grant 2003McBeathand McBeath 2010) This approach was later expanded andadapted to encompass non-forested lands that had beeninappropriately transformed from rangeland to agriculture(tuigeng huancao lsquoretire cultivation restore grasslandsrsquo) inwhich artificial seeding of forage plants took the role of thereforestation projects encouraged in more mesic climates (Shenet al 2004) A closely related program required a change of onlyone Chinese character to introduce a new approach thatrepresented a substantial change in policy This program (tuimuhuancao lsquoretire livestock restore grasslandsrsquo) was nominallyintended to conserve rangeland resources and increase long-term livestock production (Yeh 2005 Foggin 2008) Howevera central component of the lsquoretire livestockrsquo programs wasto eliminate grazing entirely for specified durations

Thus in contrast with programs that attempted to encourageresponsible husbandry through a tighter linking of pastoralistswith specific tracts of land these new programs encouraged

2 The Rangeland Journal R B Harris et al

cessation of pastoralism In some areas particularly in theSanjiangyuan area of Qinghai this approach was coupled withand often conflated with shengtai yimin lsquoecological migrationrsquo(Foggin 2008 Du 2012) which encouraged pastoralists to selltheir livestock and resettle entirely in government-constructedhousing located in often-distant towns This suggests a beliefthat physical relocation of pastoralists and reorientation of theirmeans of livelihood were necessary components of ecologicalrestoration

There has been considerable variation as well as mutabilityin local implementation of central level policy In many casestownship and village leaders have focussed more on the specificcosts and benefits (eg fencing and subsidies) rather than theunderlying ecological rationale (Bauer 2005 Bauer and YontenNyima 2010 Yonten Nyima and Yeh in press) The lsquoretirelivestockrsquo program does not appear to have entirely supplantedthe earlier responsibility-system-based programs in areasoutside of the Sanjiangyuan indeed in some places such as theTibetan Autonomous Region lsquoretire livestockrsquo has been seenas a way to further implement the Household ResponsibilitySystem (Bauer and Yonten Nyima 2010) In and aroundSanjiangyuan the two have co-existed sometimes in closegeographic proximity

It is unlikely that pastoralismwill completely disappear on theQTP other forms of agricultural land use are incompatible withenvironmental conditions found at an elevation of 3000ndash5000mThus in this paper we focus on how individual pastoralistsmight encourage or discourage sustainability of their rangelandsand herds via their own decisions regardless of the overarchingpolicy environment We were motivated by observations madeduring earlier fieldwork related to wildlife research in the studyarea (Liu et al 2007 2010Harris 2008) that rangeland conditionsappeared to vary by pasture even within a single village andeven among those with superficially similar topographicattributes We wondered if we could associate differences inapproaches to livestock husbandry with this variation andultimately if socioeconomic factors at the pastoralist level orthe ways in which pastoralists responded to the broader policyenvironment could explain the choices made by pastoralists Inthis paper we explore only the livestock-husbandryrangeland-response dynamics

Specifically our objectives were (1) accounting for siteand annual climatic variation test whether differences amongspecies herbage mass and erosion indicators were explained bythe pastoralist managing the pasture (2) accounting for site andannual variation test whether variation in herbage mass anderosion indicators was associated with variation in sheep densityandor spatially explicit measures of grazing pressure and (3)accounting for site and annual variation test whether changesin herbage mass and erosion indicators were predicted by annualchanges in sheep density

Materials and methodsStudy area

Our study was carried out in Village Five (lsquowu duirsquo) of GouliTownship Dulan County Qinghai Province China ~3558N9878E Village Five consisted of ~175 residents in 37households almost all of whom were engaged primarily in

semi-nomadic pastoralism Distance to the nearest concentrationof houses to our study area was ~6 km this village was adjacentto an historic but rejuvenated Tibetan Buddhist monastery(Harris et al 2010 Yeh and Gaerrang 2011) The landscapepart of the eastern section of the Kunlun mountain chain wascharacterised by rolling hills at elevations lt4100m rising tomoderately sloped peaks at ~4900m Vegetation was sparseabove ~4700m Vegetation formations were alpine steppedominated by Stipa purpurea Grisebach at elevations lt4300malpine meadow dominated by Kobresia spp at higherelevations and shrublands dominated by Salix spp on northerlyexposures Annual precipitation at the study site during2008ndash2013 averaged 3980mm (sd of mean 534) with ~92falling from April to September Mean annual temperature wasapproximately ndash148C with the warmest 8-day periods annuallyaveraging 1408C and the coldest averaging ndash1638C

The study area had been subject to international huntingfocussed on blue sheep (Pseudois nayar) until 2006 (Liu 1995Harris 2008) but this activity probably had little impact onpastoral practices or vegetation As had been common on theQTP during January 2007 government-sponsored workersconducted a poisoning campaign targeting plateau pikasSubsequent work showed that within a few years pikas hadrepopulated most areas

The Tibetan fauna includes several wild ungulate speciesthat could have foraged on vegetation (Schaller 1998 Harris2008) but only Tibetan gazelles (Procapra picticaudata) andblue sheep were observed in the vicinity of vegetation plots (thelatter only at the highest elevations) In addition to the commonplateau pika small mammals present in the general area includedHimalayan marmots (Marmota himalayana) the fossorialplateau zokor (Eospalax fontanierii) the Mongolian five-toedjerboa (Allactaga sibirica) mountain voles (Neodon spp) voles(Microtus spp and Lasiopodomys spp) and dwarf hamsters(Cricetulus spp) Tibetan woolly hares (Lepus oiostolus)occurred at slightly lower elevations and were rarely observedin the study pastures

The entire study area was grazed by livestock and usedprimarily as winter pastures Grazing generally occurred onlyafter livestock returned from summer and sometimes autumnpastures generally in mid-October until leaving for springndashsummer pastures in mid-June the following year (Yeh andGaerrang 2011) Village Five consisted of relatively highelevation pastures within Gouli and had been used as summerand transitional (springndashautumn) pastures before the priorcollective system was dismantled Pastoral families owned long-term leases on set pasture lands but not all grazed their ownlivestock on their own pastures Rather many pastoralists inGouli had begun sub-leasing their pastures to other grazers andor paying rental fees to graze their livestock on lands contractedto others (Yeh and Gaerrang 2011) Some winter pastureswere demarcated with boundary fences and others were notregardless most boundaries were known to pastoralists andhad traditions predating de-collectivisation Summer pasturesremained as historically in common use However somepastoralists had taken advantage of subsidised material to fencesub-sections of winter pasture in order to distribute grazingpressure reserve forage for emergencies or reduce the need forherding labour

Rangeland responses on Tibetan steppe The Rangeland Journal 3

Weather data

We measured temperature and precipitation at hourly intervalsat the research station (35834045930N 98836034710E) using asolar-powered logger (CR800-ST-SW-NC Campbell ScientificLogan UT USA) connected to a temperature probe (107-L20)and a tipping-bucket precipitation gauge with snowfall adaptor(Texas Electronics Dallas TX USA) beginning in September2009 However damage to the precipitation gauge rendered itinoperable after July 2010 thus we modelled temperature andprecipitation on the study area (see below) and used our limitedsite-specificweather data to confirm the accuracy of themodelledpredictions

To model site-specific temperature and precipitation weinterpolated meteorological data from 836 weather stations inChina using ANUSPLIN software version 42 (Hutchinson1995) This algorithm was based on the thin plate smoothingsplines of multivariates (Hutchinson 2001 Hutchinson et al2009) Observations from the weather stations were interpolatedto a 1-km spatial resolution at an 8-day time step Correlationof 8-day mean temperature of the modelled data with 71 8-daymeasurements taken at the field station from 14 September 2009to 11 June 2011was 0939 (Fig 1) After accounting for seasonalpatterns by generating residuals from regressions using Juliandate as a predictive variable correlation between modelled andsite-specific temperature measurements was 0560 (Plt 0001)

Fluctuations in herbage mass resulting from pastoralistsrsquostrategies our main variable of interest were likely to beconfounded with annual weather differences as well asdifferences in phenological stage arising from the different

dates on which each plot was sampled Thus we incorporatedboth annual weather effects and a quadratic model of Julian datein all models (see Supplementary Materials as available atjournalrsquos website) Additionally we hypothesised that herbagemass on each pasture would vary according to fixed siteconditions characterising each pasture The physical effectsof elevation slope and aspect were beyond the control ofpastoralists yet likely to affect rangeland response Thus wealso incorporated these variables in our analyses (seeSupplementary Materials)

Vegetation data and erosion indicators

Grassland vegetation and soil conditions were assessed overfour summers (2009ndash2012) using a grid of permanent plots onwinter pastures In summer 2009 we established 317 permanentsquare 05-m2 vegetation plots (measuring 071m on each side)systematically located on a 250-m grid (oriented along cardinaldirections Fig 2) across 15 winter pastures managed by 11participating pastoralists Pasture sizes varied from 46 to 1009 ha(mean = 320 ha sd of mean = 309) haWe used permanent plotsto allow for direct year-to-year comparisons and to reducesmall-scale heterogeneity In identifying the exact area toestablish each plot we walked to each pre-determined UniversalTransverse Mercator (UTM) coordinate on the sample grid (ie250-m intervals) using a handheld GPS establishing the plotcentre exactly on the intended coordinates In addition toelevation wemeasured aspect (in degrees) and slope (in per cent)of each plot using a hand-held compass To document aspectwe quantified each plot both by its absolute deviation from true

20

15

10

5

0

0 50 100 150 200

Julian date

Mea

n 8-

day

tem

p (

degC)

250 300 350

ndash5

ndash10

ndash15

ndash20

ndash25

Fig1 Measured (trianglesdashed trend line) andmodelled (circles solid trend line)meandaily temperature at8-dayperiodsatGouli field station (3558N 9878E 4033m) see text for model procedures


north and from true east in degrees (ie an aspect of 108 wasgiven a score of 10 on the north scale and 80 on the east scalean aspect of 3008 was given a score of 60 on the north scale and150 on the east scale)

Plots were designed to be relocated and identified by fieldcrews in two ways (1) via their UTM coordinates and (2) by

locating tagged wire loops connected to fixed flexible anchors(Berkshire HD Stakes with Cables Buckeye Trap SupplyAshland OR USA) that crews had inserted into the soil ~30 cmat each plotsrsquo diagonal corners when establishing the plots in2009 Each anchor left a small protruding (~3 cm diameter) loopof ~5-mm diameter steel cable to which we affixed a numbered

N

Fig 2 Detail of the study area showing the systematic grid of vegetation plots (triangles) at 250-m intervals overlaidon shaded topography Also shown are the pasture boundaries (solid polygons) and the location of the main river drainingin the area (dashed arrow)


metal tag (5 cm diameter Forestry Suppliers Jackson MSUSA) but otherwise left the surrounding area unaffected Whensampling vegetation field crews used light-weight PVC 05-m2

frames upon locating the cables they fixed the frame cornersto the two loops to re-create the precise plot location

Vegetation and site data were collected by crews of trilingual(Tibetan Chinese and English) seasonal technicians Prior toeach field season crews were trained in species identificationand field protocols Field crews assessed presence of plantspecies height cover and herbage mass for each of the 10 mainspecies in each plot as well as canopy cover estimates of totalvegetation soil rock and litter (previous yearsrsquo dry vegetation)Most plots were quantified once during the growing seasonbut some were measured twice or more depending on access

Because we used permanent plots we could not quantifyherbagemass by clipping vegetation Thuswe estimated herbagemass individually by plant species in a stepwise process Firstin the field crew members standardised their estimations ofspecies-specific fresh weight by collecting reference samplesof known fresh weight (usually 1 g) from an adjacent off-sitelocation and moving them among locations within the plotwhere that species was growing providing for close visualcomparison with unclipped vegetation Second to provide on-going calibration of fresh weight estimation a system of randomcheck-plots was set up in which crews learned whether plotswere selected for calibration only after vegetation data had hadbeen collected If selected a nearby 05-m2 location with similarvegetation to the plot was identified subjected to the fullmeasurement protocol and then clipped and sorted to speciesSpecies-specific fresh weights of the check-plot were recordedand compared with the actual (non-clipped) plot Finally toconvert fresh to dry weights clipped samples from check-plotswere placed in paper bags and either air-dried at the base-campor within a light-weight solar oven (Sport Solar Oven SolarOven Society Apple Valley MN USA) until weights stabilised(5ndash8 days) Dry-weight conversions were estimated by speciesand by month Because we were unable to dry a sufficientnumber of samples for species categorised as unpreferredforbs we analysed fresh herbage mass rather than dry herbagemass for forbs and total vegetation

Throughout we took as our response variables estimates ofherbagemass (fresh or dry see above) of the following 10 species(or groups listed here in alphabetical order not order ofimportance) (1) AstragalusOxytropis spp (forms sometimesdifficult to distinguish from one another and likely to respondsimilarly to biotic and abiotic influences) (2) Cardamine spp(3) Carex spp (4) Heteropappus altaicus Novoprokr (5)Kobresia spp (6) Leymus secalinus (Georgi) Tzelev (7) Poaspp (8) Potentilla bifurca Linnaeus (9) Stipa purpurea and10) Thermopsis lanceolata R Brown In addition we examinedaggregated life-forms as (1) grasses (which included minorspecies) (2) sedges and (3) unpreferred forbs (Liu 1986Damiran 2005) We also examined dry herbage mass of allgrasses and total herbage of all species As proxies formagnitude of the presence of or potential for erosion weexamined (1) litter canopy cover (2) proportion of bare soil (ieground unvegetated and not covered by rocks potentiallyvegetated but bare) and (3) an index of erosion that wasrecorded categorically and reflected the relative presence of

rills gulleys and pedestalling (NRC 1994) Erosion categorieswere re-coded on an ordinal scale with 1 representing theleast and 6 the most evidence of erosive forces For consistencywe have used the terminology recommended by Allen et al(2011)

Stocking rate and density

Because our intent was to examine the effects of grazingpractices as actually implemented by Tibetan pastoralistsresponding as they wished to the geophysical biologicalcultural economic and policy environments in which theyfound themselves we made no attempt to control grazing Weused direct counts to estimate the number of sheep and yaks oneach pasture in each of the three winters included in the study(2010 2011 and 2012 throughout we refer to winter grazingby the year beginning in January thus for example we use theterm lsquowinter 2010rsquo to refer to grazing that occurred during~October 2009 through early June 2010) In addition we drewfrom interviews and surveys carried out with the 11 pastoralistswhose pastures were included in the study area In 2009 and2010 in-depth interviews were carried out with each pastureowner by a member of our field crew as part of a largerinvestigation of land tenure and land management (Yeh andGaerrang 2011) Interview data included numbers of livestockof each species owned currently approximate time periodswithin each winter during which livestock grazed on thatparticular pasture and trends in livestock ownership overpast years Pastoralists also described the extent to which theyrented or contracted pastures or livestock their dependenceon herding for income and perceived changes in climate andgrassland condition Estimated livestock numbers derived fromthese two approaches were then divided by the area of thepasture (estimated by walking the periphery of each pastureswith a hand-held GPS) as well as percentage of the yearlivestock spent on the pasture to estimate mean stocking ratefor each pasture in each year

During the three winters included in our study we attachedGPS units (DG-100 Data Logger USGlobalSat Inc Chino CAUSA Qstarz BT-Q1000XT Taipei Taiwan or iBT-GPS SolarBluetooth 747 GandV Global Tech Co Taipei Taiwan) bymeans of nylon harnesses to 1ndash2 individual sheep in eachpastoralistrsquos herd for periods of 2ndash3 days for three periods eachwinter Data downloaded from the GPS units provided specificlocations of herds at 2ndash6-min intervals These data on locationwere used to demarcate levels of grazing density within eachstudy pasture in each winter All points falling within apasture in a given winter both from the ownerrsquos flock andfrom neighbouring flocks were entered into a pasture-specificanalysis We used ArcGIS 90 (ESRI Redlands CA USA)to enumerate the number of sheep locations within 250m ofeach vegetation plot To standardise metrics of relative useacross pastures that differed in their intensity of sampling werecast stocking density in terms of pasture-specific proportionaluse In addition GPS collars (Log V2 livestock collars KedzioraInnovation Group Mannsville NY USA) were attached to14 yaks belonging to four different pastoralists from autumn2011 to spring 2012 for periods of 1 week to 1 month Thesecollars provided an indication of the ways in which winter


pastures were grazed by yaks but we do not report or use thesedata further because we discovered that yaks primarily usedpastures outside the study area (ie in summer pasture evenduring winter) returning only occasionally to the study area

Statistical analysesPrior to statistical analyses we removed outlier data that wesuspected as representing data-recording errors We definedoutliers as entries gt4 sd from the mean This generally resultedin removing ~1 of data entries

To address objective 1 we asked if time-invariant measuresof species-specific herbage mass and erosion indicators (acrossall 4 years) varied by the pastoralist with user rights to pastureland during the time period We used ANOVA to test fordifferences and Tukeyrsquos HSD to differentiate levels ofdifference by pastoralists To address objective 2 we asked iftime-invariant measures of species-specific herbage mass anderosion indicators (across all 4 years) were associated with themean stocking rate of sheep (measured at the scale of the entirepasture) site-specific relative stocking density (as estimatedusing the GPS-marked sheep) or by the product of these twomeasures on the pasture during the sampling period

We developed mixed linear models examining eachhypothesis Such analyses incurred a risk of incorrectinterpretation because they lacked a before-versus-after timeelement causation could logically have run in either directionWe addressed this difficulty by means of objective 3 takingadvantage of our repeated-measures on the same plots toexamine temporal trends and to relate those to (1) time (2)differences in the stocking rate by sheep at the pasture scaleand (3) sheep density within a pasture (ie at the plot scale) inthe winter before that yearrsquos growth (as well as interactionsamong these variables) Here we used time to separate causalityreasoning that events occurring later in time could be effectsor could be unrelated but could not be causes of events thatoccurred earlier In these dynamic analyses we examineddifferences in measurements taken in 2010 2011 and 2012 fromthose taken in 2009 (the baseline year all 2009 values werefixed to zero) Standardising all measurements by their values in2009 placed data from all plots on a common basis effectivelyremoving any site effects We thus omitted elevation slopeand aspect as independent variables in these dynamic modelsBy not including site variables in regressions we implicitlyassumed that subsequent changes (ie slope coefficients) withtime sheep stocking rate and density were not themselvesfunctions of site variables However all models included theeffects of temperature (Model 2 Supplementary Materials tableS1 as available at journalrsquos website) because we consideredthese primary only trends with time or sheep stocking rate ordensity that were significant while accounting for possibleeffects of annual variation in temperature were consideredvalid We also included Julian date and (Julian date)2 in allmodels to account for differences in phenological stage arisingfrom the different dates on which each plot was sampled In allcases plot was retained in regression models as a random factor

Because the data used to address objective 3 were differencesfrom 2009 values we encountered situations in which all valueswere zero because the species was not documented in the plot

during the study Thus before conducting analyses we removedfrom consideration all plots in which a species was absentduring all sampling occasions All statistical analyses wereconducted using the software package JMP 1111 (SASInstitute Cary NC USA)

Results

Characteristics of vegetation plots

Vegetation and erosion indicators from 317 plots were measuredat least once yearly pastures contained between 8 and 55 plotsAnalysis was performed on a total of 1771 measurements(Table 1) Plots varied in elevation from 3863 to 4511m (mean(sd of mean) = 4142 (140) m) and were located on slopesranging from 0 to 708 (mean (sd of mean) = 177 (130)8) Theaspects of the plots ranged from 0 to 3598 plots on aspectswithin 608 of true south were most common (31) followedby those most nearly oriented towards the west (28) north(23) and east (18) Plots contained between 0 and 18 generaof grasses forbs and shrubs (mean (sd of mean) 65 (27) perplot) Total vegetation cover in plots ranged from 0 to985 (mean (sd of mean) 339 (203))

We obtained sufficient dry samples of the herbage massof S purpurea from off-site plots to quantify the relationshipbetween fresh- and dry weights (F= 97 df = 3 64 P lt 0001r2 = 031) We thus replaced each measurement of fresh weightof herbage of S purpurea by their dry weight In 2009 20102011 and 2012 the ratios of dry weight to fresh weight were061 049 067 and 054 Because ~65 of grass herbage wasS purpurea we applied these annual conversions to analysesof aggregated grass species

Effects of individual pastoralists on cover and herbagemass of species and erosion indicators

After accounting for the effects of site heterogeneity (tablesS2ndashS4) we observed significant differences among individualpastoralists that evidently reflected in aggregate their currentand past management practices (Table 2) Leymus secalinus

Table 1 Pastoralists with long-term leases on each (coded foranonymity) and pasture codes (numbered) in the study area Village

Five Gouli township Dulan County Qinghai 2009ndash2012Shown are pasture sizes (in ha) and number of plot readings (including

replicate within-year readings n) in each year

Pastoralist Pasture Size 2009 2010 2011 2012 Totalcodes (ha) n n n n n

B 3 46 10 8 12 25 55D1 10 ndA 10 10 20 17 57D2 7 467 49 68 80 82 279G 9 110 20 19 36 31 106H 12 15 16 253 39 40 64 51 194K 1 678 56 81 85 88 310L1 5 616 27 42 39 29 137L2 6 350 19 22 27 26 94N 11 136 16 12 21 18 67S 2 1009 48 57 73 77 255Y 13 14 17 179 29 32 51 45 157Total ndash ndash 323 391 508 489 1711

And = no data


had higher cover in pastures controlled by pastoralists B D2G and N than in pastures controlled by pastoralist K Poaspp had higher cover in pastures controlled by pastoralist B thanin those of pastoralists H and L2 The dominant and preferredperennial S purpurea had higher cover in pastures controlledby pastoralist N than all others and a lower cover in the pastureof pastoralist K than all others The disturbance-associatedlegume T lanceolata had higher cover in the pasture ofpastoralist S than others and a lower cover in pastures controlledby pastoralist H than others Due in part to the small samplessizes owing to their rarity cover of AstragalusOxytropisCardamine Carex spp H altaicus and P bifurca accountingfor topography did not differ (Tukeyrsquos HSD text Plt 005)among pastoralists (Table 2)

Total herbage mass consisting of (generally preferred) grassspecies was greater in the pastures of pastoralist N than in mostothers and less in the pasture of pastoralist K than most othersTotal herbagemass consisting of unpreferred specieswas greatestamong pastures controlled by pastoralists D1 K L1 and S thanit was among those controlled by pastoralist G The ratio ofunpreferred preferred herbage mass of species was greater inpastures controlled by pastoralist K than those controlled bypastoralists B G H N and Y Litter constituted a higherproportion of pastures controlled by pastoralists B and G thanpastoralists D1 K L2 and S a lower proportion of pasturescontrolled by pastoralist L2 than B D2 G H N and Y(Table 2) Neither the cover of bare soil nor erosion index wassignificantly affected by pastoralists

Notable pastures that appeared on the basis of theserankings to be in relatively lsquohealthyrsquo condition were thoseoperated by pastoralist B whose pastures ranked lowest inerosion indicators of over-use such as cover of bare soil anderosion index first in proportion of litter cover and of thepreferred Poa spp L secalinus and Carex spp and ranked lowin the cover of unpreferred forbs Cardamine spp andH altaicus and pastoralist N whose pastures portrayedintermediate indicators of over-use but ranked first in the coverof the preferred grass S purpurea second in the cover of total

grasses and second-to-last in the cover of unpreferred forbsincluding Oxytropis spp and H altaicus In contrast werepastures showing evidence of stress including those operatedby pastoralist K (ranked second in cover of bare soil anderosion index second-to-last in cover of litter first in the ratioof unpreferred total vegetation and last in relative cover ofboth S purpurea and L secalinus) pastoralist S (characterisedby high proportions of unpreferred species and first in the coverof the unpreferred legume T lanceolata) and pastoralist G(ranked first in cover of bare soil and erosion and last in thecover of preferred graminoids Poa and Carex spp) Otherpastoralistsrsquo areas were intermediate in these indices of stress(Table 2)

Associations of mean stocking rate and density of sheepwith herbage mass of species and erosion indicators

Stocking rate at the pasture scale varied from 0 to 589 sheep handash1

annually (Table 3) We found no associations of mean stockingrate in 2009ndash2011 at the pasture scale with indicators of erosionat the scale of individual vegetation plots We found positiveassociations of mean stocking rate (during 2009ndash2011) at thepasture scale with herbage mass of litter (b = 0176 se = 0043t= 405 Plt 0001) and total herbage mass of grasses (b= 0237se = 0103 t= 231 Plt 005) at the plot scale (Table 4a) Wefound negative associations of mean stocking rate at the pasturescale with total fresh herbage mass (b = ndash0316 se = 0130t= ndash242 P lt 005) fresh herbage mass of unpreferred species(b = ndash0532 se = 0147 t= ndash363P lt 0001) and the proportionof the total herbage mass consisting of unpreferred species(b = ndash0063 se = 0017 t= ndash380 P lt 0001 Table 4b)

There were no consistent trends of any response variablewith the proportional pasture use When combining the metricsof mean stocking rate and proportional pasture use we foundnegative associations of stocking density with the per centcover of bare soil (b= ndash1713 se = 0875 t= ndash196 P= 005)herbage mass of unpreferred species (b= ndash3662 se = 1397t= ndash262 P= 001) and the ratio of the herbage mass of

Table 2 Predicted order of abundance (from most 1 to least 11) of response variables (rows) by pastoralist (columns)Values are frommodels accounting for the effects of pastoralist-specific topography (see SupplementaryMaterials table S3) with plot and the date of observation

within each year as random factors Row entries sharing letter codes were not significantly different from one another (Plt 005 Tukeyrsquos HSD test)

Pastoralist B D1 D2 G H K L1 L2 N S Y

AstragalusOxytropis 1a 5a 2a 6a 7a 8a 9a 11a 10a 4a 3aCardamine 11a 3a 7a 9a 8a 2a 5a 4a 6a 1a 10aCarex spp 1a 7a 3a 11a 8a 5a 2a 6a 10a 4a 9aHeteropappus altaicus 9a 4a 1a 3a 6a 5a 8a 7a 10a 2a 11aLeymus secalinus 1a 9ab 2a 3a 7ab 11b 5ab 10ab 4a 8ab 6abPoa spp 1a 4ab 2ab 11b 9b 8ab 3ab 10b 6ab 5ab 7abPotentilla bifurca 6a 2a 10a 11a 1a 4a 5a 7a 9a 3a 8aStipa purpurea 2abc 6bc 10abc 7abc 8bc 11c 5abc 9bc 1a 3ab 4abcThermopsis lanceolata 6ab 10ab 7ab 11ab 9b 2ab 5ab 3ab 4ab 1a 8abTotal grass herbage mass 1abc 9bcd 3abc 5abc 6abc 11d 8abcd 10cd 2a 7abc 4abUnpreferred herbage mass 11ab 2a 6ab 9b 8ab 3a 1a 5ab 10b 4a 7abUnpreferred preferred ratio 11b 2ab 6ab 9b 7b 1a 4ab 5ab 10b 3ab 8bErosion index 11a 8a 9a 1a 3a 2a 10a 6a 5a 4a 7aBare soil 11a 9a 6a 1a 3a 2a 10a 7a 5a 4a 8aLitter cover 1a 9cd 4abc 2a 6ab 10d 7abcd 11d 3ab 8bcd 5abc


unpreferred species to total herbage mass (b= ndash0392sse = 0158 t= ndash247 P lt 005 Table 5)

Effects of temporal variation in stocking rate of sheep onerosion indicators and herbage mass of species

Cover of bare soil increased with time (b= 5760 se = 0743t= 7759 Plt 00001 table S5) as well as with pasture-scalestocking rate (b= 1289 se = 0445 t= 289 Plt 001 table S6Fig 3a) The year sheep density interactionwas also significantand positive with the trend on time being greater in thosepastures with higher sheep density (interaction b = 1645se = 0321 t= 512 P lt 00001 table S7)

The erosion index did not differ with time but was positivelyrelated to the stocking rate of sheep on the pasture duringthe preceding winter (b = 0040 se = 0019 t= 208 Plt 005Fig 3b table S8) although not at the plot scale with grazingpressure When viewed together the increase in erosion withstocking rate at the pasture scale increased with time (interactionof stocking rate year b= 0037 se = 0015 t= 255 P = 001table S9)

Cover of live vegetation declined with time (b= ndash7544se = 0622 t= ndash1214 Plt 00001 table S10) and was notrelated to stocking rate in the previous winter However whentime and stocking rate in the previous winter were combined ina single model all were highly significantly negative (table S11)suggesting that vegetation cover declined more in areas withgreater than lower stocking rate (Fig 3c)

We documented no relationships among stocking rate atthe pasture scale or stocking density at the plot scale withtotal fresh herbage mass However herbage mass of grassesdeclined with time (b= ndash0956 se = 0335 t= ndash286 Plt 001table S12) The decline of herbage mass of grasses with timewas greater in pastures with higher than lower stocking rate(Fig 3d table S13)

No relationships among time stocking rate at the pasturescale or at the plot scale were observed with herbage mass ofCarex spp However among unpreferred forbs we found thatherbage mass was positively associated with density of sheepgrazing in the previous winter (b= 1496 se = 0741 t= 202P lt 005) more so towards the end than the beginning of thestudy period (Fig 3e table S14)

Herbage mass of S purpurea declined with time duringthe study (year b= ndash1412 se = 0312 t= ndash452 P lt 0001

table S15) This decline was not associated with stocking rateat the pasture scale or with stocking density at the plot scaleThe decline with time however was greater following pastureuse by higher than lower stocking rates of sheep (yearb= ndash1656 se = 0350 t= ndash472 Plt 0001 sheep densityb= ndash0567 se = 0213 t= ndash265 P lt 001 interaction ndash0441se = 0140 t= ndash314 P lt 001 Fig 3f table S16) The herbagemass of L secalinus was negatively associated with stockingrate in the previous winter the relationship strengtheningthroughout the study period (density b = ndash0326 se = 0107t= ndash305 Plt 001 year b= 0453 se = 0223 t= 204 Plt 005interaction b= ndash0208 se = 0084 t= ndash249 Plt 005 Fig 3gtable S17) When the interaction of changes over time withchanges following winter grazing by sheep were modelledtogether we observed an increasingly positive response by Poaspp with stocking rate (density b = 0402 se = 0155 t= 26P lt 001 interaction b = 0344 se = 0111 t= 310 Plt 001Fig 3h table S18) No relationships with time stocking rateat the pasture scale or stocking density at the plot scale wereobserved for other species considered individually

Discussion

Site effects on vegetation

Our focus was on responses to the management practices oflivestock by individual pastoralists but because their pasturesvaried in their inherent biological and site characteristics it wasimportant to first quantify characteristics of rangelands by siteindependently of any influences introduced by heterogeneityin management Before one can understand any effects thatlivestock management may have had on the herbage mass ofa plant species one needs to understand whether that specieswould be expected to be common or not in that pastureregardless of management For example a high herbage massof noxious legumes of the AstragalusOxytropis spp wouldgenerally be considered indicators of over-use That pasturesmanaged by pastoralists L2 and N had similarly low herbagemasses of these legumes (Table 2) might therefore suggesta similarity of management influences These legumes wereassociated with steep slopes (table S2) the pastures of pastoralistL2 were the steepest (table S3) whereas those of pastoralist Nwere relatively flat Thus we would not expect to encountermany of these legumes in the pastures of pastoralist N whereasthe paucity found in the pastures of pastoralist L2 is surprisingand potentially more informative Similarly to our analysis ofannual variation in weather accounting for site effects allowedus to isolate effects attributable to pastoralistsrsquo managementfrom those beyond their immediate control

Influence of individual pastoralists

Yeh and Gaerrang (2011) showed that despite jointmembership in a small seemingly cohesive village pastoralistsdiffered in their approach to livestock management but theseauthorsrsquo analyses did not extend to possible consequences ofthis heterogeneity as expressed by vegetation Our analysesprovided insight into the magnitude of the effects that individualvariation in management had on the rangelands over and aboveannual differences due to temperature and inherent biologicaldifferences arising from heterogeneity in the site of the pasture

Table 3 Estimated sheep densities (individuals handash1) on each pasturein the study Gouli township Dulan County Qinghai Province

2010ndash2012

Pastoralist Pasture 2010 2011 2012 Mean

K 1 003 004 007 005S 2 024 000 020 015L1 5 016 000 093 037L2 6 057 271 066 131D2 7 000 000 000 000G 9 499 195 216 303N 11 184 158 127 156Y 14 115 162 073 117H 15 147 213 127 163H 16 170 179 147 166Y 17 305 589 193 362


Table 4 Linear models relating various response variables at the plot scale to the mean density of sheep on the scaleof the individual pasture during winters 2009ndash2011

For each model coefficients (b) standard errors (se) Studentrsquos t and P-value are shown for fixed-effect explanatory variablesplot was also included in each model as a random effect a Models with mean sheep density positively associated with litterand total grass biomass b Models with mean sheep density negatively associated with total fresh biomass total biomass ofunpalatable species and the proportion of total biomass consisting of unpalatable species Response variables were square-root

transformed

a

Litter n = 1581 Adj R2 = 0696 b se t P

Julian date 0014 0013 108 02785Julian date2 ndash0001 0001 ndash129 01985Elevation 0001 0001 099 03206Slope ndash0020 0003 ndash605 lt00001North deviation 0002 0001 283 00049East deviation ndash0001 0001 ndash100 03182Spring cumulative temperature ndash1671 0059 ndash2836 lt00001January temperature 0607 0015 4061 lt00001Pasture sheep density 0176 0043 405 lt00001Herbage mass grass species n= 1575 Adj R2 = 0876 b se t P

Julian date 0086 0012 690 lt00001Julian date2 ndash0001 0001 ndash750 lt00001Elevation ndash0002 0001 ndash205 00416Slope ndash0035 0008 ndash437 lt00001North deviation 0006 0002 299 00030East deviation ndash0004 0003 ndash138 01691Spring cumulative temperature 0290 0055 525 lt00001January temperature ndash0130 0014 ndash918 lt00001Pasture sheep density 0237 0103 231 00216

b

Total herbage mass n= 1574 Adj R2 = 0694 b se t P

Julian date 0386 0027 1446 lt00000Julian date2 ndash0001 0001 ndash1506 lt00001Elevation 0002 0001 247 00139Slope ndash0036 0010 ndash358 00004North deviation 0004 0002 157 01175East deviation ndash0002 0004 ndash050 06181Spring cumulative temperature 0483 0119 406 lt00001January temperature ndash0208 0030 ndash686 lt00001Pasture sheep density ndash0316 0130 ndash242 00159

Unpreferred herbage mass n = 1570 Adj R2 = 0794 b se t P

Julian date 0284 0023 1238 lt00000Julian date2 ndash0001 0000 ndash1279 lt00001Elevation 0005 0001 403 lt00001Slope ndash0003 0011 ndash023 08195North deviation ndash0001 0003 ndash029 07756East deviation 0004 0004 101 03153Spring cumulative temperature 0056 0102 055 05816January temperature ndash0039 0026 ndash150 01335Pasture sheep density ndash0532 0147 ndash363 00003

Unpreferred total herbage mass n = 1526 Adj R2 = 0860 b se t P

Julian date 0010 0002 459 lt00001Julian date2 0001 0001 ndash467 lt00001Elevation 0000 0000 316 00018Slope 0003 0001 218 00303North deviation 0000 0000 ndash150 01348East deviation 0001 0000 156 01204Spring cumulative temperature ndash0022 0009 ndash235 00188January temperature 0008 0002 323 00013Pasture sheep density ndash0063 0017 ndash380 00002


We observed patterns suggesting that some pastures were lessliable to change than others even taking site differences intoaccount For example pastures such as those allocated topastoralists B and N were characterised both by a higher coverof preferred species (eg S purpurea L secalinus and Poaspp) as well as less bare soil and erosion In contrast cover ofT lanceolata P bifurca and Cardamine spp on pastures suchas those of pastoralists K and H was generally associated withgreater bare soil and erosion Analysis of factors that mighthave caused pastoralists who shared cultural practices andwere exposed to a similar policy environment to differ in theirmanagement approaches is beyond the scope of this paper

Spatial associations among mean grazing levelsand rangeland response

Our analyses relating stocking rate and density to changesin vegetation and erosion indicators (Tables 4 5) could beinterpreted as suggesting that stocking level was negativelyassociated with cover of bare soil and unpreferred specieswhile being positively associated with litter and herbage massof grass species Might grazing have had such unexpected andbeneficial effects Analyses conducted by Harris et al (2015)

using livestock exclosures suggested that most QTP steppevegetation appears to bewell adapted to grazing inwinter at leastat low stocking rates In most plant species increased intra-or inter-specific competition for resources in the absence ofherbivory evidently counteracted whatever benefits plantsenjoyed from a respite from herbivory Thus the possibilitythat grazing actually increased herbage production (Hilbertet al 1981 McNaughton 1983) in this system should not bedismissed Harris et al (2015) however also found that grazingincreased erosion over conditions prevailing within livestockexclosures

A simpler interpretation however is that pastoralists returningto winter pasture in October adjusted their stocking levels totheir (generally correct) perceptions of the herbage mass of thepreferred vegetation Despite pastoralists occasionally statingduring interviews that they lacked ability to distinguish onespecies from another (sometimes referring to all vegetationsimply as lsquograssrsquo) our analyses suggest that pastoralists rationallyadjusted herd size in winter to reflect the herbage mass ofpreferred grasses and actually reduced grazing pressure onspecific pastures if herbage consisted disproportionately ofunpreferred species Within pastures the proportion of timespent near each plot by GPS-monitored sheep was lower where

Table 5 Linear models relating various response variables at the plot scale to the mean grazing pressure (meansheep densitypasturepasture-specific proportional use) during winters 2009ndash2011

For each model coefficients (b) their standard errors (se) Studentrsquos t and P-value are shown for fixed-effect explanatoryvariables plot was also included in each model as a random effect Models shown indicate that mean grazing pressure wasnegatively associated with percent bare soil biomass of unpalatable biomass and the proportion of total biomass consisting

of unpalatable species Response variables were square-root transformed

Bare soil n= 1584 Adj R2 = 0658 b se t PElevation ndash0001 0001 ndash172 00871Slope ndash0009 0007 ndash119 02354North deviation ndash0002 0002 ndash131 01912East deviation ndash0001 0003 ndash033 07397Julian date ndash0077 0018 ndash4147 lt00000Julian date2 0001 0001 413 lt00001Spring cumulative temperature 0979 0083 1173 lt00001January temperature ndash0135 0021 ndash635 lt00001Stocking rate proportional use ndash1713 0875 ndash196 00511

Unpreferred herbage mass n= 1570 Adj R2 = 0794 b se t P

Elevation 0005 0001 463 lt00001Slope ndash0002 0012 ndash013 08949North deviation ndash0002 0003 ndash073 04680East deviation 0004 0004 096 03388Julian date 0283 0023 1232 lt00000Julian date2 ndash0001 0001 ndash1273 lt00001Spring cumulative temperature 0055 0102 053 05933January temperature ndash0039 0026 ndash149 01358Stocking rate proportional use ndash3662 1397 ndash262 00092

Unpreferred total herbage mass n= 1562 Adj R2 = 0861 b se t P

Elevation 0001 0001 385 00001Slope 0003 0001 324 00309North deviation ndash0001 0001 ndash202 00446East deviation 0001 0001 149 01386Julian date 0009 0002 454 lt00000Julian date2 0001 0001 ndash462 lt00001Spring cumulative temperature ndash0022 0009 ndash237 00181January temperature 0008 0002 324 00012Stocking rate proportional use ndash0392 0158 ndash247 00139


unpreferred species were higher in herbage mass than elsewhereThus both stocking rate comparing among pastures andstocking density viewed within each pasture respondedadaptively increasing with vegetation selectively eaten bysheep (Harris et al 2015 see also Cincotta et al 1991)

Rangeland response to annual variation in stocking rate

Importantly however our dynamic analyses indicate thatgrazing exerted a negative effect on preferred vegetation and

soils displaying patterns generally consistent with the hypothesisthat at least at sufficiently intense levels herbivory of senescentvegetation and winter-time compaction reduced subsequentherbage masses of many preferred species The modelssummarised in Fig 3 used as explanatory variables pasture-specific measurements of stocking rate during the winterpreceding the observed rangeland response Thus unlike thestatic analyses summarised in Tables 4 and 5 the two types ofvariables were not contemporaneous but rather followed one

40 05

04

03

02

01

0

0

0

5

4

3

2

1

0

ndash1

ndash01

ndash02

ndash2

ndash2ndash3ndash4ndash5ndash6ndash7ndash8ndash9

ndash10

ndash1

ndash4

ndash6

ndash8

ndash10

ndash12

ndash14

ndash03

35

(a) (b)

(c) (d)

(e) (f )

(g) (h)

30

25

20

15Bar

e so

il

Ero

sion

inde

x

Veg

etat

ion

cove

rU

npre

ferr

ed fo

rb fr

esh

herb

age

mas

sL

sec

alin

us fr

esh

herb

age

mas

s

S p

urpu

rea

dry

herb

age

mas

sP

oa s

pp f

resh

herb

age

mas

sG

rass

fres

h he

rbag

e m

ass

10

15

13

11

9

7

5

3

1

ndash1

15

10

05

0

ndash05

ndash10

ndash15

ndash20

ndash3

5

0

0

0 1 2 3 4 5

Sheep handash1

0 1 2 3 4 5

ndash5

ndash5

ndash10

ndash15

ndash20

ndash25

ndash30

ndash35

ndash40

Fig 3 Differences from summer 2009 levels (summer 2010 solid summer 2011 dash and summer 2012 dot-dash) as a function ofthe density of sheep on the pasture during the preceding winter (a) Proportion of plots devoid of vegetation (ie bare soil) (b) Indexof erosion (c) Proportion of plots covered by live vegetation (d) Fresh herbage mass of all grass species (e) Fresh herbage mass ofunpreferred forbs (f) Dry biomass of Stipa purpurea (g) Fresh biomass of L secalinus (h) Fresh biomass of Poa spp


another in sequence Because our data were observationalrather than experimental it is possible that relationship shownin Fig 3 were not causal but rather that causes resided in someother unmeasured variable However it could not have beenthe case that the response variables depicted in Fig 3 (egherbage mass of S purpurea) produced the stocking ratesdocumented because the stocking events occurred and endedbefore the response variables (eg vegetation growth) werequantified

Accounting for fluctuations in annual temperature and sitevariability rangeland condition as measured appeared toworsen during the 4 years of our study independently of annualstocking rate of sheep The cover of bare soil increased andvegetation cover decreased during 2009ndash2012 Although totalfresh herbage mass did not change herbage mass of S purpureaas well as total grasses declined We found no evidence oftemporal changes in the herbage mass of L secalinus and Poaand Carex spp However the herbage mass of unpreferredspecies as well as the proportion of unpreferred species of thetotal herbage mass increased with time during 2009ndash2012 Tworeasonable hypotheses to consider to explain these dynamicsare long-term climate changes associated with changes in soilmoisture content or temperature or alternatively a laggedresponse to herbivory levels before our study

Quantifying the size of the negative effects of winter-timestocking rates of sheep on the subsequent herbage mass ofthe preferred grasses is not straightforward Most effects wefound were complicated by the general downward trend inherbage mass observed among most species through timeOur dynamic models showing trends in herbage mass resultingfrom changes in the stocking rates of sheep considered all plotsbut the starting cover of each species varied by pasture andpastoralist The size of effects associated with the higheststocking rates documented in this study were substantial ForS purpurea we estimated the mean herbage mass in 2009 overall plots as ~180 kgDMhandash1 Our models suggested that by2012 at a stocking rate of 2 sheep handash1 would have reducedherbage mass by ~100 kgDMhandash1 or more than half

Studies in experimental situations have suggested thatliveweight gains by sheep were negatively correlated withstocking rates (Zhou et al 1995) and among yaks werepositively correlated with ratios of preferred to unpreferredplants (Dong et al 2003) Additional studies of the relationshipsamong stocking rates herbage mass of preferred species andliveweights of sheep in working pastures would be useful

Conclusions

Our analysis shows that pastoralists stock their pastures withsheep on the basis of the herbage mass of preferred grasses noton the basis of total herbage mass We found that the stockingrate of sheep at the pasture scale was negatively associatedwith total herbage mass and of the herbage mass of unpreferredforbs but was positively related to the herbage mass of grassesStocking rate at the pasture scale did not appear to be related toindicators of erosion (cover of bare soil total vegetation coverand erosion index) but within pastures sheep tended to avoidareas with a relatively large cover of bare soil and relativelylarge proportions of unpreferred vegetation

We detected annual responses in most indicators of erosionas well as in the herbage mass of most (albeit not all) preferredspecies to sheep grazing pressure during the preceding dormantseason Accounting for annual weather fluctuations and sitevariability both cover of bare soil and erosion varied positivelywith stocking rate of sheep Although total herbage mass wasnot related to stocking rate of sheep total vegetation coverherbage mass of S purpurea and of all grass species variedinversely with stocking rate of sheep in the preceding winterwhereas herbage mass of unpreferred forbs varied directlywith stocking rate Within pastures the increase in cover ofbare soil was more pronounced where stocking rate was higherand per cent live vegetation cover and herbage mass ofS purpurea declined more strongly where stocking rate washigher than lower

Our data diverge from the opinions of most of the pastoralistswho although recognising their pastures had finite capacitiestended to view possible negative effects of heavy stocking onlyin terms of potential livestock mortality and not in rangelandproductivity itself although they did note a downward trendin the liveweight of sheep without attributing it to grazing Inour study area pastoralists had the ability to encourage thesummer-time herbage mass of preferred species and discourageexpansion of unpreferred species through their stocking levelsin winter

Acknowledgements

Principal funding for this work was from the USA National ScienceFoundation Dynamics of Coupled Natural and Human Systems ProgramAward 0815441 Supplementary funding was provided by the TraceFoundation and the Bridge Fund Field data were ably collected byChungjyid Dorjiejyal Drubgyal Gurudorjie Hulchendorjie LamojiaPagmostso Pemabum Puhuadongzhi Rinchentso Sonam SonamtsoTseringdorjie and Wanmananqing Pemabum ably ran the field stationFor field assistance we thank Zhou J K Ma L L Shi Y H and QiS F Statistical assistance was provided by B Steele For administrativesupportwe thankLArendsTBaerwaldHBjornWBleisch JBurchfieldM Garry Kunchok Gelek and E Yang The map was produced byG Maclaurin

References

Allen V G Batello C Berretta E J Hodgson J Kothmann M Li XMcIvor J Milne J Morris C Peeters A and Sanderson M (2011)An international terminology for grazing lands and grazing animalsGrass and Forage Science 66 2ndash28 doi101111j1365-2494201000780x

Banks T J (2003) Property rights reform in rangeland China dilemmason the road to the household ranchWorld Development 31 2129ndash2142doi101016jworlddev200306010

Banks T J Richard C Li P and Yan Z L (2003) Governing thegrasslands of Western China Mountain Research and Development23 132ndash140 doi1016590276-4741(2003)023[0132CGMIWC]20CO2

Bauer K (2005) Development and the enclosure movement in pastoralTibet since the 1980s Nomadic Peoples 9 53ndash81 doi103167082279405781826119

Bauer K and Yonten Nyima (2010) Laws and regulations impacting theenclosure movement on the Tibetan Plateau of China Himalaya 3023ndash38


Bedunah D J and Angerer J P (2012) Rangeland degradation povertyand conflict how can rangeland scientists contribute to effectiveresponses and solutions Rangeland Ecology and Management 65606ndash612 doi102111REM-D-11-001551

Bian J H Jing Z C and Fan N C (1999) The effect of grasslandfencing on the population density of plateau pikas Acta BiologicaPlateau Sinica 14 110ndash115 [in Chinese]

Chen S (1996) Inner Asian grassland degradation and plant transformationIn lsquoCultural and Environment in Inner Asia Volume 1 The PastoralEconomy and the Environmentrsquo (Eds C Humphrey and D Sneath)pp 111ndash123 (The White Horse Press Cambridge UK)

Cincotta R P van Soest P J Robertson J B Beall CM and GoldsteinM C (1991) Foraging ecology of livestock on the Tibetan Changtanga comparison of three adjacent grazing areas Arctic and AlpineResearch 23 149ndash161 doi1023071551379

Damiran D (2005) lsquoPalatability of Mongolian Rangeland Plantsrsquo (EasternOregon Agricultural Research Center Union Station OR)

Delibes-Mateos M Smith A T Slobodchikoff C N and Swenson J E(2011) The paradox of keystone species persecuted as pests a call forthe conservation of abundant small mammals in their native rangeBiological Conservation 144 1335ndash1346 doi101016jbiocon201102012

Dong Q M Zhao X Q Ma Y S Li Q Y Wang Q J and Shi J J(2003) Studies on the relationship between grazing intensity for yaksand plant groups inKobresia parva alpine meadow Acta Agrestia Sinica13 334ndash343 [in Chinese English abstract]

Du F C (2012) Ecological resettlement of Tibetan herders in theSanjiangyuan a case study in Madoi County of Qinghai NomadicPeoples 16 116ndash133 doi103167np2012160109

Fan N C Zhou W Y Wei W H Wang Q Y and Jiang Y J (1999)Rodent pest management in the Qinghai-Tibet alpine meadowecosystem In lsquoEcologically-based Management of Rodent Pestsrsquo (EdsG R Singleton L A Hinds H Leirs and Z B Zhang) pp 285ndash304(Australian Centre for International Agricultural Research CanberraACT)

Foggin J M (2008) Depopulating the Tibetan Grasslands national policiesand perspectives for the future of Tibetan herds in Qinghai ProvinceChina Mountain Research and Development 28 26ndash31 doi101659mrd0972

Goldstein M C and Beall C M (2002) Changing patterns of Tibetannomadic pastoralism In lsquoHuman Biology of Pastoral Populationsrsquo (EdsWRLeonardandMHCrawford) pp 131ndash150 (CambridgeUniversityPress Cambridge UK)

Grant A (2003) A study of the implementation of Chinarsquos sloping landconversion policy lsquotui geng huan linrsquo a case study ndash Hanyuan CountySichuan Province Forests Trees and Livelihoods 13 331ndash343doi1010801472802820039752469

Hardin G (1968) The tragedy of the commons Science 162 1243ndash1248doi101126science16238591243

Harris R B (2008) lsquoWildlife Conservation in China Preserving theHabitat of Chinarsquos Wild Westrsquo (ME Sharpe Inc Armonk NY)

Harris R B (2010) Rangeland degradation on the Qinghai-Tibetanplateau a review of the evidence of its magnitude and causes Journal ofArid Environments 74 1ndash12 doi101016jjaridenv200906014

Harris R B Bedunah D J Yeh E T Smith A T and Anderies J M(2010) Determinants of rangeland dynamics on the Qinghai-Tibetplateau China livestock wildlife and pastoralism Pastoralism 1325ndash326

Harris R B WangW Y Badinqiuying Smith A T and Bedunah D J(2015) Herbivory and competition of Tibetan steppe vegetation inwinter pasture effects of livestock exclosure and plateau pika reductionPLoS One 10 e0132897 doi101371journalpone0132897

Hilbert D W Swift D M Detling J K and Dyer M I (1981) Relativegrowth rates and the grazing optimization hypothesis Oecologia 5114ndash18 doi101007BF00344645

Holzner W and Kreichbaum M (2001) Pastures in south and centralTibet (China) probable causes of pasture degradation Die Bodenkultur52 37ndash44

Hutchinson M F (1995) Interpolating mean rainfall using thin platesmoothing splines International Journal of Geographical InformationScience 9 385ndash403 doi10108002693799508902045

Hutchinson M F (2001) lsquoANUSPLIN version 42 User Guidersquo(Centre for Resource and Environmental Studies Australian NationalUniversity Canberra ACT)

Hutchinson M F McKenney D W Lawrence K Pedlar J HHopkinson R F Milewska E and Papadopol P (2009) Developmentand testing of Canada-wide interpolated spatial models of dailyminimum-maximum temperature and precipitation for 1961ndash2003Journal of Applied Meteorology and Climatology 48 725ndash741doi1011752008JAMC19791

Immerzeel W W van Beek L P H and Bierkens M F P (2010)Climate change will affect the Asian water towers Science 3281382ndash1385 doi101126science1183188

Lai C H and Smith A T (2003) Keystone status of plateau pikas(Ochotona curzoniae) effect of control on biodiversity of native birdsBiodiversity and Conservation 12 1901ndash1912 doi101023A1024161409110

Levine N E (1998) From nomads to ranchers managing pasture amongethnic Tibetans in Sichuan In lsquoDevelopment Society and Environmentin Tibet Proceedings of the Seventh Seminar of the InternationalAssociation for Tibetan Studiesrsquo (Eds G E Graz and G E Clarke)pp 69ndash119 (Verlag der Oumlsterreichischen Akademie derWissenchasftenVienna Austria)

Li M S (1994) Characteristics and rational exploitation of Tibetrsquos landresources Journal of Natural Resources 9 51ndash57 [in Chinese]

LiXLGao JBrierleyGQiaoYM Zhang J andYangYW (2013)Rangeland degradation on the Qinghai-Tibet plateau implicationsfor rehabilitation Land Degradation amp Development 24 72ndash80doi101002ldr1108

Liu R T (1986) lsquoAksai Kazak Autonomous County Grassland Resourcesand Planningrsquo (Gansu Province Aksai Kazak Autonomous CountyAgricultural Planning Office Gansu China) [in Chinese]

Liu Y S (1995) International hunting and the involvement of localpeople Dulan Qinghai Peoplersquos Republic of China In lsquoIntegratingPeople and Wildlife for a Sustainable Futurersquo (Eds J A Bissonetteand P R Krausman) pp 63ndash67 (The Wildlife Society Bethesda MD)

Liu Q X Harris R B Wang X M andWang Z H (2007) Home rangesize and overlap of Tibetan foxes (Vulpes ferrilata) in Dulan CountyQinghai Province Acta Theriologica Sinica 27 370ndash375 [in Chinese]

Liu Q X Harris R B and Wang X M (2010) Food habits of Tibetanfox (Vulpes ferrilata) in the KunlunMountains Qinghai Province ChinaMammalian Biology 75 283ndash286 doi101016jmambio200902002

McBeath J H and McBeath J (2010) lsquoEnvironmental Change and FoodSecurity in Chinarsquo (Springer Dordrecht Heidelberg Germany)

McNaughton S J (1983) Compensatory plant growth as a response toherbivory Oikos 40 329ndash336 doi1023073544305

Miller D J Bedunah D J Pletscher D H and Jackson R M (1992)From open range to fences changes in the range-livestock industryon the Tibetan Plateau and implications for development planningand wildlife conservation In lsquoProceedings of the 1992 InternationalRangelandDevelopmentSymposiumrsquo (EdsGK Perrier andCWGay)pp 95ndash109 (Society for Range Management Littleton CO)

NRC (1994) lsquoRangeland Health New Methods to Classify Inventory andMonitor Rangelandsrsquo (National Academies Press Washington DC)


Richard C Yan Z L and Du G Z (2006) lsquoThe Paradox of the IndividualHousehold Responsibility System in the Grasslands of the TibetanPlateau ChinarsquoUSDA Forest Service Proceedings RMPS-P-39 (USDAForest Service Fort Collins CO)

Schaller G B (1998) lsquoWildlife of the Tibetan Steppersquo (University ofChicago Press Chicago IL)

SEPA (State Environmental Protection Agency Nature Reserve Bureau)(2007) Report on environmental 2006 Grasslands Available at wwwmepgovcn (accessed 18 June 2007)[in Chinese]

Shang Z H Gibb M J Leiber F Ismail M Ding L M Guo X S andLong R J (2014) The sustainable development of grassland-livestocksystems on the Tibetan plateau problems strategies and prospects TheRangeland Journal 36 267ndash296 doi101071RJ14008

Shapiro J (2001) lsquoMaorsquos War against Nature Politics and theEnvironment in Revolutionary Chinarsquo (Cambridge University PressCambridge UK)

Shen Y Y Ma Y S and Li Q Y (2004) Grassland restoration inDari County Qinghai Province In lsquoImplementing the Natural ForestProtection Program and the Sloping Lands Conversion ProgramsLessons and Policy Implicationsrsquo (Eds E Katsigris J Xu andT A White) pp 303ndash40 (Beijing Forestry Publishing House BeijingChina)

Shi Y Z (1983) On the influences of rangeland vegetation on the densityof plateau pika (Ochotona curzoniae) Acta Theriologica Sinica 3181ndash187 [in Chinese]

Smith A T and Foggin J M (1999) The plateau pika (Ochotonacurzoniae) is a keystone species for biodiversity on the Tibetan PlateauAnimal Conservation 2 235ndash240 doi101111j1469-17951999tb00069x

Smith A T Formozov N A Hoffmann R S Zheng C L and ErbajevaM A (1990) The pikas In lsquoRabbits Hares and Pikas Status Surveyand Conservation Action Planrsquo (Eds J A Chapman and J E C Flux)pp 14ndash60 (IUCN Gland Switzerland)

Smith A T Zahler P and Hinds L A (2006) Ineffective andunsustainable poisoning of native small mammals in temperate Asiaa classic case of the science-policy divide In lsquoConservation Biology inAsiarsquo (Eds J A McNeely T M McCarthy A T Smith L Olsvig-Whittaker and E D Wikramanayake) pp 285ndash293 (Society forConservation Biology Asia Section and Resources HimalayaFoundation Kathmandu Nepal)

State Council (2002) Some suggestions regarding strengtheninggrassland protection and construction State Council Circular 2002 19[Beijing China]

WangYBWangGX ShengYP andWangWL (2005)Degradationof the eco-environmental system in alpine meadow on the Tibetanplateau Journal of Glaciology and Geocyrology 27 634ndash640 [inChinese]

Wangdwei M Steele B and Harris R B (2013) Demographic responsesof plateau pikas to vegetation cover and land use in the TibetanAutonomous Region China Journal of Mammalogy 94 1077ndash1086doi10164412-MAMM-A-2531

Williams D M (2002) lsquoBeyond Great Walls Environment Identity andDevelopment on the Chinese Grasslands of Inner Mongoliarsquo (StanfordUniversity Press Stanford CA)

Wilson M C and Smith A T (2015) The pika and the watershedthe impact of small mammal poisoning on the ecohydrology of theQinghai-Tibetan plateau Ambio 44 16ndash22 doi101007s13280-014-0568-x

Wu N and Yan Z L (2002) Climate variability and social vulnerabilityon the Tibetan Plateau dilemmas on the road to pastoral reformErdkunde 56 2ndash14 doi103112erdkunde20020101

Wu N Yan Z L and Lu T (2012) Enclosure and resettlement in theeastern Tibetan Plateau dilemma of pastoral development duringthe last three decades In lsquoPastoral Practices in High Asia Agencyof lsquoDevelopmentrsquo Effect by Modernization Resettlement andTransformationrsquo (Ed H Kreutzmann) pp 291ndash306 (SpringerDordrecht Heidelberg Germany)

Xu J Grumbine R E Shrestha A Eriksson M Yang X Wang Y andWilkes A (2009) The melting Himalayas cascading effects of climatechange on water biodiversity and livelihoods Conservation Biology23 520ndash530 doi101111j1523-1739200901237x

Yan J P (2001) lsquoStrategies and Countermeasures of Chinarsquos GreatWestern Development Strategyrsquo (Science Press Beijing China) [inChinese]

Yan Z L Wu N Dorji Y and Ru J (2005) A review of rangelandprivatization and its implications in the Tibetan Plateau ChinaNomadic Peoples 9 31ndash51 doi103167082279405781826155

Yeh E T (2005) Green governmentality and pastoralism in westernChina lsquoConverting pastures to grasslandsrsquo Nomadic Peoples 9 9ndash30doi103167082279405781826164

Yeh E T (2009) Greening western China a critical view Geoforum 40884ndash894 doi101016jgeoforum200906004

Yeh E T and Gaerrang (2011) Tibetan pastoralism in neo-liberalisingChina continuity and change in Gouli Area 43 165ndash172 doi101111j1475-4762201000976x

Yonten Nyima and Yeh E (in press) Environmental issues and conflictin Tibet In lsquoEthnic Conflict in Western Chinarsquo (Eds B Hillman andG Tuttle) (Columbia University Press New York)

Zhang H F Liu F G Zhou Q and Duo H R (2004) Degradationmechanism of the grass in Qinghai Plateau and its prevention and controlcountermeasures Ziran Zaihai Xuebao 13 115ndash120 [in Chinese]

Zhou L Wang Q J Zhao J and Wang Q (1995) Studies on optimumstocking intensity in pasturelands of alpine meadow I stocking intensityto maximize production of Tibetan sheep In lsquoAlpine MeadowEcosystem 4rsquo pp 365ndash375 (Science Press Beijing China) [in Chinese]

Zhou H K Zhou L Zhao X Q Liu W Yan Z L and Shi Y (2003)Degradation process and integrated treatment of lsquoblack soil beachrsquograsslands in the source regions of the Yangtze and Yellow RiversChinese Journal of Ecology 22 51ndash55 [in Chinese]


wwwpublishcsiroaujournalstrj

IntroductionIn rangelands throughout the world livestock provide food andincome to the majority of the worldrsquos poorest people (Bedunahand Angerer 2012) yet degradation of rangelands is havingdirect effects on the livelihoods food security and way of lifeof pastoralists In addition this degradation can significantlyreduce ecosystem services provided by rangelands at regionaland global scales such as carbon sequestration and hydrologicalregulation and provision Effects of degradation such as duststorms food or commodity shortages and displaced communitiescan be felt over a large geographic scale (Yan 2001 Wang et al2005 Li et al 2013) Rangeland health also affects biodiversitydirectly and indirectly because all native flora and fauna haveadapted to the long-term evolutionary forces that have shapedthese environments Livestock grazing is the dominant form ofland use in arid biomes worldwide including on the rangelandsof the Qinghai-Tibetan Plateau (QTP) in the Peoplersquos Republicof China The QTP occupies 25million km2 ~25 of Chinarsquosarea and an estimated 70 is used by grazing livestock Thesepastures colloquially termed lsquoChinarsquos water towerrsquo are locatedupstream and upwind of upwards of an estimated 20 of theworldrsquos human population (Xu et al 2009 Immerzeel et al2010)

Awareness by Chinese scientists and policy-makers of theimpacts of degradation increased in the late 1990s as severaldisasters occurred including Yangtze River floods that killedthousands of residents downstream and cost billions of dollarsin economic losses the Yellow River running dry increasinglyoften and dust-storms and sand-storms originating in westernrangelands that affected the health and economic wellbeing ofmillions of city-dwellers in Chinarsquos east Although lacking cleardocumentation and differing in specifics scientific papers andgovernment policy statements have generally viewed the QTPas having become increasingly degraded in recent decades(Harris 2010) A frequently repeated statistic is that 90 ofChinarsquos grasslands are degraded to some extent and thatdegradation is increasing at a rate of 200 km2 yearndash1 (StateCouncil 2002)

Causes for this degradation are generally attributed to acombination of over-stocking unscientific livestockmanagementhistorical-cultural impediments to adopting modern livestockmanagement concepts global climate change and excessiveherbivory and soil disturbance from small mammals (Li 1994Chen 1996 Zhou et al 2003 Zhang et al 2004 Li et al 2013)In contrast other investigators have questioned the assumptionof wide-scale grassland degradation on the QTP and where theyagree it has occurred cast their analysis in terms of rapidchanges in socioeconomic systems and alteration of land-tenurearrangements (Miller et al 1992 Levine 1998 Holzner andKreichbaum2001Goldstein andBeall 2002Williams2002Wuand Yan 2002 Banks 2003 Banks et al 2003)

Chinese policy-makers have implemented a variety ofprograms to ameliorate negative rangeland trends (Harris 2010Li et al 2013 Shang et al 2014) some more effectively thanothers As early as the 1950s efforts were begun to exterminateplateau pikas (Ochotona curzoniae) a small-bodied burrowinglagomorph generally associated with poor vegetation cover(Smith et al 1990 Fan et al 1999) Policy during this periodblamed small mammals for causing rangeland degradation

regardless of its precise definition but this may have reflectedthe general Maoist conception of nature at the time (Shapiro2001) scientificwork on pikas began only in the 1980s Researchsubsequently began taking a more nuanced view of interactionsbetween pikas livestock and vegetation after it emerged notonly that pikas were critical parts of the natural environment(lsquokeystone speciesrsquo Smith and Foggin 1999 Lai and Smith2003 Wilson and Smith 2015) but that high density of pikaslikely resulted from rather than caused the rangeland conditionswith which they were associated (Shi 1983 Bian et al 1999Wangdwei et al 2013) Regardless programs of pika reductioncontinue to the present day (Smith et al 2006 Delibes-Mateoset al 2011 Wilson and Smith 2015)

Privatisation of livestock and dissolution of the collectivesystem occurred during the 1980s throughout the QTP and inour study area in 1983 During the 1990s following the successof quasi-privatisation policies in eastern Chinarsquos agriculturalsector similar approaches often referred to as the HouseholdResponsibility System were initiated in the pastoral sector (Yanet al 2005) Echoing widespread concerns about the lsquotragedy ofthe commonsrsquo (Hardin 1968) these programs aimed to encourageresponsible husbandry by clarifying pasture-land tenure at thehousehold level The primary tenets of the government initiativeinvolved increasing the duration of pasture-lease contracts from20 to 50 years subsidising construction of permanent winterhomes fences and livestock shelters and providing plots forgrowing supplemental winter fodder (Richard et al 2006 Wuet al 2012) Government outlays for programs related to theHousehold Responsibility System (often termed the lsquoset of fourrsquoor sipeitao in Chinese (Wu and Yan 2002)) were substantialduring 2003ndash2006 the central government reported investingsome yen71 billion (~US$1 billion at the time) for fencing alone(SEPA 2007) Whether this fundamental reform alleviated orexacerbated negative trends in rangeland condition remainscontentious

A newer set of initiatives began in the early 2000s thatemphasised land protection rather than tenure and responsibility(Yeh 2009) Increased awareness of the consequences ofupstream erosion following the devastating Yangtze River floodof 1998 led to government subsidies encouraging reforestationof cultivated lands that were unsuitable for agriculture (tuigenghuanlin lsquoretire cultivation restore forestsrsquo Grant 2003McBeathand McBeath 2010) This approach was later expanded andadapted to encompass non-forested lands that had beeninappropriately transformed from rangeland to agriculture(tuigeng huancao lsquoretire cultivation restore grasslandsrsquo) inwhich artificial seeding of forage plants took the role of thereforestation projects encouraged in more mesic climates (Shenet al 2004) A closely related program required a change of onlyone Chinese character to introduce a new approach thatrepresented a substantial change in policy This program (tuimuhuancao lsquoretire livestock restore grasslandsrsquo) was nominallyintended to conserve rangeland resources and increase long-term livestock production (Yeh 2005 Foggin 2008) Howevera central component of the lsquoretire livestockrsquo programs wasto eliminate grazing entirely for specified durations

Thus in contrast with programs that attempted to encourageresponsible husbandry through a tighter linking of pastoralistswith specific tracts of land these new programs encouraged













Weather data







20

15

10

5

0

0 50 100 150 200

Julian date

Mea

n 8-

day

tem

p (

degC)

250 300 350

ndash5

ndash10

ndash15

ndash20

ndash25






N



















Results











And = no data
























Discussion






2010ndash2012


K 1 003 004 007 005S 2 024 000 020 015L1 5 016 000 093 037L2 6 057 271 066 131D2 7 000 000 000 000G 9 499 195 216 303N 11 184 158 127 156Y 14 115 162 073 117H 15 147 213 127 163H 16 170 179 147 166Y 17 305 589 193 362




transformed

a




b


























40 05

04

03

02

01

0

0

0

5

4

3

2

1

0

ndash1

ndash01

ndash02

ndash2


ndash10

ndash1

ndash4

ndash6

ndash8

ndash10

ndash12

ndash14

ndash03

35

(a) (b)

(c) (d)

(e) (f )

(g) (h)

30

25

20

15Bar

e so

il

Ero

sion

inde

x

Veg

etat

ion

cove

rU

npre

ferr

ed fo

rb fr

esh

herb

age

mas

sL

sec

alin

us fr

esh

herb

age

mas

s

S p

urpu

rea

dry

herb

age

mas

sP

oa s

pp f

resh

herb

age

mas

sG

rass

fres

h he

rbag

e m

ass

10

15

13

11

9

7

5

3

1

ndash1

15

10

05

0

ndash05

ndash10

ndash15

ndash20

ndash3

5

0

0

0 1 2 3 4 5

Sheep handash1

0 1 2 3 4 5

ndash5

ndash5

ndash10

ndash15

ndash20

ndash25

ndash30

ndash35

ndash40







Conclusions




Acknowledgements


References



















































































Weather data







20

15

10

5

0

0 50 100 150 200

Julian date

Mea

n 8-

day

tem

p (

degC)

250 300 350

ndash5

ndash10

ndash15

ndash20

ndash25






N



















Results











And = no data
























Discussion






2010ndash2012


K 1 003 004 007 005S 2 024 000 020 015L1 5 016 000 093 037L2 6 057 271 066 131D2 7 000 000 000 000G 9 499 195 216 303N 11 184 158 127 156Y 14 115 162 073 117H 15 147 213 127 163H 16 170 179 147 166Y 17 305 589 193 362




transformed

a




b


























40 05

04

03

02

01

0

0

0

5

4

3

2

1

0

ndash1

ndash01

ndash02

ndash2


ndash10

ndash1

ndash4

ndash6

ndash8

ndash10

ndash12

ndash14

ndash03

35

(a) (b)

(c) (d)

(e) (f )

(g) (h)

30

25

20

15Bar

e so

il

Ero

sion

inde

x

Veg

etat

ion

cove

rU

npre

ferr

ed fo

rb fr

esh

herb

age

mas

sL

sec

alin

us fr

esh

herb

age

mas

s

S p

urpu

rea

dry

herb

age

mas

sP

oa s

pp f

resh

herb

age

mas

sG

rass

fres

h he

rbag

e m

ass

10

15

13

11

9

7

5

3

1

ndash1

15

10

05

0

ndash05

ndash10

ndash15

ndash20

ndash3

5

0

0

0 1 2 3 4 5

Sheep handash1

0 1 2 3 4 5

ndash5

ndash5

ndash10

ndash15

ndash20

ndash25

ndash30

ndash35

ndash40







Conclusions




Acknowledgements


References











































































N



















Results











And = no data
























Discussion






2010ndash2012


K 1 003 004 007 005S 2 024 000 020 015L1 5 016 000 093 037L2 6 057 271 066 131D2 7 000 000 000 000G 9 499 195 216 303N 11 184 158 127 156Y 14 115 162 073 117H 15 147 213 127 163H 16 170 179 147 166Y 17 305 589 193 362




transformed

a




b


























40 05

04

03

02

01

0

0

0

5

4

3

2

1

0

ndash1

ndash01

ndash02

ndash2


ndash10

ndash1

ndash4

ndash6

ndash8

ndash10

ndash12

ndash14

ndash03

35

(a) (b)

(c) (d)

(e) (f )

(g) (h)

30

25

20

15Bar

e so

il

Ero

sion

inde

x

Veg

etat

ion

cove

rU

npre

ferr

ed fo

rb fr

esh

herb

age

mas

sL

sec

alin

us fr

esh

herb

age

mas

s

S p

urpu

rea

dry

herb

age

mas

sP

oa s

pp f

resh

herb

age

mas

sG

rass

fres

h he

rbag

e m

ass

10

15

13

11

9

7

5

3

1

ndash1

15

10

05

0

ndash05

ndash10

ndash15

ndash20

ndash3

5

0

0

0 1 2 3 4 5

Sheep handash1

0 1 2 3 4 5

ndash5

ndash5

ndash10

ndash15

ndash20

ndash25

ndash30

ndash35

ndash40







Conclusions




Acknowledgements


References
























































































Results











And = no data
























Discussion






2010ndash2012


K 1 003 004 007 005S 2 024 000 020 015L1 5 016 000 093 037L2 6 057 271 066 131D2 7 000 000 000 000G 9 499 195 216 303N 11 184 158 127 156Y 14 115 162 073 117H 15 147 213 127 163H 16 170 179 147 166Y 17 305 589 193 362




transformed

a




b


























40 05

04

03

02

01

0

0

0

5

4

3

2

1

0

ndash1

ndash01

ndash02

ndash2


ndash10

ndash1

ndash4

ndash6

ndash8

ndash10

ndash12

ndash14

ndash03

35

(a) (b)

(c) (d)

(e) (f )

(g) (h)

30

25

20

15Bar

e so

il

Ero

sion

inde

x

Veg

etat

ion

cove

rU

npre

ferr

ed fo

rb fr

esh

herb

age

mas

sL

sec

alin

us fr

esh

herb

age

mas

s

S p

urpu

rea

dry

herb

age

mas

sP

oa s

pp f

resh

herb

age

mas

sG

rass

fres

h he

rbag

e m

ass

10

15

13

11

9

7

5

3

1

ndash1

15

10

05

0

ndash05

ndash10

ndash15

ndash20

ndash3

5

0

0

0 1 2 3 4 5

Sheep handash1

0 1 2 3 4 5

ndash5

ndash5

ndash10

ndash15

ndash20

ndash25

ndash30

ndash35

ndash40







Conclusions




Acknowledgements


References






























































































Discussion






2010ndash2012


K 1 003 004 007 005S 2 024 000 020 015L1 5 016 000 093 037L2 6 057 271 066 131D2 7 000 000 000 000G 9 499 195 216 303N 11 184 158 127 156Y 14 115 162 073 117H 15 147 213 127 163H 16 170 179 147 166Y 17 305 589 193 362




transformed

a




b


























40 05

04

03

02

01

0

0

0

5

4

3

2

1

0

ndash1

ndash01

ndash02

ndash2


ndash10

ndash1

ndash4

ndash6

ndash8

ndash10

ndash12

ndash14

ndash03

35

(a) (b)

(c) (d)

(e) (f )

(g) (h)

30

25

20

15Bar

e so

il

Ero

sion

inde

x

Veg

etat

ion

cove

rU

npre

ferr

ed fo

rb fr

esh

herb

age

mas

sL

sec

alin

us fr

esh

herb

age

mas

s

S p

urpu

rea

dry

herb

age

mas

sP

oa s

pp f

resh

herb

age

mas

sG

rass

fres

h he

rbag

e m

ass

10

15

13

11

9

7

5

3

1

ndash1

15

10

05

0

ndash05

ndash10

ndash15

ndash20

ndash3

5

0

0

0 1 2 3 4 5

Sheep handash1

0 1 2 3 4 5

ndash5

ndash5

ndash10

ndash15

ndash20

ndash25

ndash30

ndash35

ndash40







Conclusions




Acknowledgements


References










































































transformed

a




b


























40 05

04

03

02

01

0

0

0

5

4

3

2

1

0

ndash1

ndash01

ndash02

ndash2


ndash10

ndash1

ndash4

ndash6

ndash8

ndash10

ndash12

ndash14

ndash03

35

(a) (b)

(c) (d)

(e) (f )

(g) (h)

30

25

20

15Bar

e so

il

Ero

sion

inde

x

Veg

etat

ion

cove

rU

npre

ferr

ed fo

rb fr

esh

herb

age

mas

sL

sec

alin

us fr

esh

herb

age

mas

s

S p

urpu

rea

dry

herb

age

mas

sP

oa s

pp f

resh

herb

age

mas

sG

rass

fres

h he

rbag

e m

ass

10

15

13

11

9

7

5

3

1

ndash1

15

10

05

0

ndash05

ndash10

ndash15

ndash20

ndash3

5

0

0

0 1 2 3 4 5

Sheep handash1

0 1 2 3 4 5

ndash5

ndash5

ndash10

ndash15

ndash20

ndash25

ndash30

ndash35

ndash40







Conclusions




Acknowledgements


References


























































































40 05

04

03

02

01

0

0

0

5

4

3

2

1

0

ndash1

ndash01

ndash02

ndash2


ndash10

ndash1

ndash4

ndash6

ndash8

ndash10

ndash12

ndash14

ndash03

35

(a) (b)

(c) (d)

(e) (f )

(g) (h)

30

25

20

15Bar

e so

il

Ero

sion

inde

x

Veg

etat

ion

cove

rU

npre

ferr

ed fo

rb fr

esh

herb

age

mas

sL

sec

alin

us fr

esh

herb

age

mas

s

S p

urpu

rea

dry

herb

age

mas

sP

oa s

pp f

resh

herb

age

mas

sG

rass

fres

h he

rbag

e m

ass

10

15

13

11

9

7

5

3

1

ndash1

15

10

05

0

ndash05

ndash10

ndash15

ndash20

ndash3

5

0

0

0 1 2 3 4 5

Sheep handash1

0 1 2 3 4 5

ndash5

ndash5

ndash10

ndash15

ndash20

ndash25

ndash30

ndash35

ndash40







Conclusions




Acknowledgements


References












































































Conclusions




Acknowledgements


References








































































rangeland responses to pastoralists’ grazing management on ......rangeland responses to...

Documents