Seediscussionsstatsandauthorprofilesforthispublicationathttpwwwresearchgatenetpublication282565812

SimulatingeffectsofgrazingonsoilorganiccarbonstocksinMongoliangrasslands

ARTICLEinAGRICULTUREECOSYSTEMSampENVIRONMENTmiddotDECEMBER2015

ImpactFactor34middotDOI101016jagee201507014

READS

30

9AUTHORSINCLUDING

XiaofengChang

ChineseAcademyofSciences

18PUBLICATIONS281CITATIONS

SEEPROFILE

ShipingWang

InstituteofTibetanPlateauResearchChinhellip

106PUBLICATIONS1032CITATIONS

SEEPROFILE

BolormaaDamdinsuren

AnimalHusbandryMongolia

2PUBLICATIONS2CITATIONS

SEEPROFILE

Allin-textreferencesunderlinedinbluearelinkedtopublicationsonResearchGate

lettingyouaccessandreadthemimmediately

AvailablefromXiaofengChang

Retrievedon23December2015

Agriculture Ecosystems and Environment 212 (2015) 278ndash284

Simulating effects of grazing on soil organic carbon stocks in Mongoliangrasslands

Xiaofeng Changa Xiaoying Baod1 Shiping Wangbc Andreas WilkeseBaasandai Erdenetsetsegf Batkhishig Baivalg Danzan-osor AvaadorjhTemuujin Maisaikhani Bolormaa Damdinsureni

a State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau Institute of Soil and Water Conservation Northwest AampF UniversityYangling Shaanxi 712100 ChinabKey Laboratory of Alpine Ecology and Biodiversity of the Institute of Tibetan Plateau Research Chinese Academy of Sciences Beijing 100101 ChinacCAS Center for Excellence for Tibetan Plateau Earth Science Chinese Academy of Sciences Beijing 100101 ChinadUniversity of the Chinese Academy of Sciences 19A Yuquan Rd 100049 Beijing ChinaeValues for Development Ltd Suffolk IP33 3EQ United Kingdomf Institute of Hydrology and Meteorology Ulaanbaatar MongoliagNutag Partners LLC Ulaanbaatar MongoliahMongolian Society for Range Management Ulaanbaatar MongoliaiResearch Institute of Animal Husbandry Ulaanbaatar Mongolia

A R T I C L E I N F O

Article historyReceived 26 March 2015Received in revised form 14 July 2015Accepted 17 July 2015Available online 31 July 2015

KeywordsOvergrazingCarbon accumulationModelingMongolia

A B S T R A C T

Grassland degradation is a major issue in many parts of the world Rehabilitation of areas that have beendegraded by overgrazing can potentially accumulate soil carbon but there have been few studies in thevast grasslands of Mongolia Here we calibrated and validated Century model with 618 measurementsfrom a soil inventory covering four grassland types in a forest steppe region of Mongolia Soil organiccarbon (SOC) simulated by Century largely agreed with the observational dataset and the sign of SOCresponse to intensive grazing We employed the calibrated model to assess SOC accumulation underreduced grazing intensity scenarios and the projected accumulation rates were 220ndash369 g C m2 yr1 inthe near term (2012ndash2035) These results imply that reducing the intensity of grazing may be an effectivestrategy for restoration of degraded grasslands which can be implemented by reducing livestocknumbers andor by changing the timing and duration of grazing events Moreover the simulated SOCaccumulation was mainly determined by a conceptual slow pool which was not supported byexperimental observations in similar soils Therefore evaluating the long-term climate mitigationthrough soil carbon accumulation in degraded grasslands still warrants further attention

atilde 2015 Published by Elsevier BVThis is an open access article under the CC BY-NC-ND license (httpcreativecommonsorglicensesby-

nc-nd40)

Contents lists available at ScienceDirect

Agriculture Ecosystems and Environment

journa l homepage wwwe l sev ier com loca te agee

1 Introduction

Concerns about the steady increase in atmospheric carbondioxide (CO2) have sparked worldwide interest in the global carboncycle and the role of various potential carbon sinks (Houghton2007) Grasslands cover approximately 50 million km2 and con-tain the largest share (39) of terrestrial soil carbon stocks (Whiteet al 2000) Any change in storage in this large grassland soilcarbon reserve would have a substantial and long-lived effect onglobal carbon cycles (McSherry and Ritchie 2013) Management

Corresponding author Fax +86 10 84097102E-mail address wangspitpcasaccn (S Wang)

1 These made an equal contribution to this paper

httpdxdoiorg101016jagee2015070140167-8809atilde 2015 Published by Elsevier BV This is an open access article under the

and land use change will be key drivers expected to affect soilcarbon in the coming decades and estimates of changes ingrassland soil carbon stocks over the long term are of criticalimportance (Conant and Paustian 2002 Wang et al 2011)However despite considerable research over the past 40 yearsmuch uncertainty exists regarding the effects of grazing on soilorganic carbon (SOC) (McSherry and Ritchie 2013) Previousstudies found both strong positive and negative grazing effects onSOC (Conant and Paustian 2002 McSherry and Ritchie 2013) Insemi-arid regions intensive grazing leads to grassland degradationand SOC losses as observed in Northern China (He et al 2011)America (Neff et al 2005) and the world (Conant 2009) Curtailinggrazing intensity has been recommended to promote SOCsequestration An accumulation of SOC after reducing grazing

CC BY-NC-ND license (httpcreativecommonsorglicensesby-nc-nd40)

X Chang et al Agriculture Ecosystems and Environment 212 (2015) 278ndash284 279

has already been documented for many regions of the world(Conant et al 2001) However despite strong hints in thesescientific studies broad-scale implementation has been limited byuncertainties in the magnitude duration soil texture and grazingregime (McSherry and Ritchie 2013) For example estimates of theglobal mitigation potential of grasslands vary considerably rangingbetween 015 and 13 Gt CO2 yr1 (Lal 2004 Follett and Schuman2005 Smith et al 2007 Henderson et al 2015) The uncertainty ofglobal sequestration estimates is exacerbated by the severe lack ofstudies in developing countries (Conant et al 2001 Follett andSchuman 2005) Reports about the carbon sequestration rateswere particularly rare in the vast Eurasian steppe with theexception of China (Wang et al 2011)

The Mongolian steppe spanning an area of about 115 millionkm2 (ALAGC 2012) is part of the Eurasian steppe These grasslandshave been grazed by wild and then domestic livestock for millennia(Hilker et al 2014) In the period of Soviet influence the nationallivestock population was relatively constant at about 23 millionhead (Olonbayar 2010) and Mongolian steppe was sustainablygrazed Since the transition to a market economy and privatizationof formerly state-owned livestock in 1992 the number of livestockhas steeply increased to 45 million by the end of 2013 (NSO 2014)During this time long-term trends of reduced grassland biomass orproduction were documented over much of Mongolia (Liu et al2013) Official data suggest that about 22 of the grassland area isdegraded (ALAGC 2011) Another recent national assessmentusing a new methodology suggests that about 90 is in an lsquoalteredstatersquo much of which can be rehabilitated through improvedmanagement (NAMEM 2015) Several studies have attempted touse analysis of remote sensing imagery to differentiate the effectsof climate change from the effects of grazing suggesting thatovergrazing is the main driver of grassland degradation in much ofthe central and northern parts of the country where averageprecipitation is higher (Hilker et al 2014 Liu et al 2013) This isconsistent with other research suggesting that equilibrium grass-lands are more vulnerable to degradation due to inappropriatemanagement (von Wehrden et al 2012) The presence ofdegradation in equilibrium grasslands may also indicate potentialfor increasing SOC stocks through changes in management practice

Fig 1 Distribution of sampling site

(Booker et al 2013) Considering it large area and potentialimportance to the regional and global carbon cycle it is importantto assess the soil carbon pools in Mongolian steppe in response tochanges in grazing pressures

This study assessed the impact of grazing on the SOC ofMongolian grasslands using process based modeling approachescombined with inventory data Conducting soil inventory oncecannot reflect changes in soil carbon stocks in response to specificland use and management To the contrary the process-basedmodeling approach can predict detailed effects of land use andmanagement changes but requires measured data for modelcalibration and validation (Bortolon et al 2011) In the past theCentury model has been used in Mongolia to evaluate the effects ofclimate change on NPP (Dagvadorj et al 2009) but there havebeen limited applications to investigating SOC stock changes underdifferent grassland management practices Here we developed aSOC inventory for four major grassland types with varying degreesof degradation in a forest steppe area of central Mongolia We thenapplied the Century model using historic grazing regimes based onknowledge derived from local grassland managers Estimationsbased on the inventory method were compared to the modelsimulation results Our objective was to characterize the responsesof SOC stocks to changing grazing pressures and assess the size ofthe accumulated soil carbon in degraded Mongolian steppe

2 Materials and methods

21 Soil and biomass survey

We conducted a field survey in Tariat soum Arkhangai aimag incentral Mongolia during the summers (July and August) of 2011ndash2012 (Fig 1) The climate in Tariat soum has a mean annualtemperature range of 57 to 19 C with mean annualprecipitation ranging from 820 to 3503 mm (Fig 2) The soiland biomass survey covered more than 47000 ha of summergrassland in Tariat which is about 30 of the total grassland area inthe soum The vegetation is dominated by four grassland typesmountain meadow (MM) riparian meadow (M) riparian meadowand marsh (MWM) and mountain steppe (MS) Each vegetation

s and study area in Mongolia

Fig 2 Annual temperature and precipitation (a) and N deposition (b) at Tariat from1972 to 2011 The median and error bars in box-and-whisker plot indicate datarange within 5th and 95th percentiles The N deposition was modeled based on therelationship between wet deposition and precipitation across China (Jia et al 2014)

280 X Chang et al Agriculture Ecosystems and Environment 212 (2015) 278ndash284

type was further categorized as non-degraded or lightlymoderately or heavily degraded based on the characteristics ofground cover plant community composition and biomass produc-tion We sampled 618 sites across the study area including 285 inMM 86 in MWM 208 in M and 39 in MS At each site a typicalquadrat of 05 05 m2was randomly selected Within the selectedquadrat all plants were harvested to measure abovegroundbiomass Then we collected three soil cores using a 5-cm corerfrom 0 to 20 cm depth Bulk density samples were obtained using astandard container of 100 cm3 in volume and weighed after beingoven dried at 105 C for 24 h Soil samples were sieved (2 mm) air-dried and ground to a powder SOC of ground soil samples weredetermined by combustion in a TOC analyzer (Shimadzu TOC-5000 Kyoto Japan) after phosphoric acid treatment to removecarbonates Soil texture was determined by a particle size analyzer(MasterSizer 2000 Malvern Worcestershire UK) Abovegroundbiomass samples were oven dried at 65 C for 48 h and weighed

22 Century model description calibration and validation

Century is an ecosystem model that simulates biogeochemicalfluxes on a monthly time step In the model soil organic matter isdivided into three pools (active slow and passive) depending onturnover rates In the plant submodel the maximum plant growthis calculated according to the genetic potential of the planttemperature and availability of moisture and nutrients The plant

Table 1Mean particle size distribution bulk density plant biomass and SOC stocks (0ndash20 cm)

Grassland type Status Clay () Silt ()

MM Non-degraded 254 500

Light 323 420

Moderate 235 421

Heavy 102 287

MWM Light 296 448

Moderate 215 417

Heavy 89 247

M Moderate 189 421

Heavy 103 287

MS Moderate 213 383

Heavy 80 318

MM mountain meadow MWM riparian meadow with marsh M riparian meadow

module also determines the timing of net primary production andits proportional allocation to roots or foliage The grazing event canbe specified by defining the fractions of aboveground biomassconsumption and excrement export to folds (urine and dung)Grazing effects were determined based on relationships of grazingto biomass production changes as grazing intensity increases(Evans et al 2010) This approach has been used to simulategrazing in other studies (Wang et al 2008 Chang et al 2014)

To represent the long history of grazing on Mongolian grass-lands we simulated an equilibrium period for 5000 years with aperennial grass system under moderate grazing (50 of live shootsremoved by grazing event per month) in summer (JunendashOctober)From 1990 to 2012 the model was run with higher removal of plantproduction by heavy grazing The modeled biomass removal ratesfor degraded grassland strata during this period were 80 whichwas determined on the basis of an inventory of livestock herds andgrazing patterns among households using grasslands in the studyarea For non-degraded mountain meadow sites moderate grazingwas kept constant during entire period The climatic data used inthe Century model consisted of monthly precipitation monthlymaximum and minimum air temperature provided by Tariatmeteorology station Because the geospatial N deposition datawere unavailable we set the two N deposition parameters (EPNFA(1) and EPNFA(2)) based on the relationship between wetdeposition and precipitation over the neighboring country China(Jia et al 2014) The basic site-specific parameters such as soiltexture bulk density and rock fragments were obtained fromanalysis of soil samples Drainage conditions were also predefinedaccording to local topography Potential aboveground production(PRDX) parameter was adjusted so that the modeled carbon stockaccurately matched the measured values from 309 sampling sitestaken randomly from our sampling campaign (Fig S1) Once themodel had been calibrated the modeled SOC was validated againstanother dataset comprised of data from the field campaign thathad not been used in calibration

23 Grazing management scenarios

Once the model calculations are optimized we assessed carbondynamics according to the different grazing scenarios Grazingmanagement scenarios for the degraded grasslands were projectedbased on lsquorule of thumbrsquo sustainable biomass removal rates appliedto each degradation stratum within each vegetation type 50biomass removal for non-degraded grasslands 45 for lightlydegraded 40 for moderately degraded and 35 for heavilydegraded grasslands The timing of grazing events simulated in thesummer grasslands followed local common practice with grazingbeginning on June 1 and ending on October 31 of each year

for Mongolian grasslands at different degrees of degradation

Sand () Bulk density(g cm3)

Above-groundbiomass (g m2)

SOC stock(kg C m2)

246 09 1353 560257 09 1258 473345 11 1262 347611 11 1065 156

257 09 1884 450368 11 1189 384663 09 504 138

390 10 1046 345610 10 1068 170

405 11 1122 310602 10 878 109

MS mountain steppe

Fig 3 Comparison observed and modeled soil organic carbon (SOC) for non-degraded light moderate and heavy degraded Mongolian grassland soils on theTariat soum MM mountain meadow MWM riparian meadow with mash Mriparian meadow MS mountain steppe

X Chang et al Agriculture Ecosystems and Environment 212 (2015) 278ndash284 281

modeled A 23 years (2012ndash2035) spin-up was applied on the basisof average historical climate data

24 Model performance and uncertainty assessment

Performance of the Century model was assessed using themethodology described by Cheng et al (2014) According toMethodology for Sustainable Grassland Management (SGM)depicted in CDM EB- approved General Guidelines For SamplingAnd Surveys For Small-Scale CDM Project Activities we assessedthe parametric uncertainty using the key parameters with a givenerror The range of estimated changes in SOC stock demonstratesthe model parameter uncertainty which is expressed as follows

UncertaintyethTHORN frac14 jSmax Sminj2 S

100 (1)

Smax frac14 ModelethPmax Temperaturemin Precipitationmin ClaymaxTHORN(2)

Smin frac14 ModelethPmin Temperaturemax Precipitationmax ClayminTHORN(3)

where Smax and Smin denote the upper and lower values of soilcarbon accumulation corresponding to the parameters at the 95confidence level and S is the modeled SOC accumulation with themean of parameters Pmin and Pmax are the minimum andmaximum value of parameter at the 95 confidence levelrespectively This is clearly a very limited way to assess manypossible uncertainties but was the best we could do given thelimitations in data availability

3 Results

31 Effects of degradation

Our field inventory revealed that variations of plant and soilvariables were significantly across a degradation gradient (Table 1)Grassland degradation coincided with an increase in sand contentand decreases in clay and silt content The differences in bulkdensity among degradation classes were negligible Above-groundbiomass generally decreased as degradation increased SOC stockdecreased markedly with degradation SOC in lightly degraded MMgrasslands were 187 lower than that in non-degraded grasslandsThe SOC further declined by about 20 and 50 in moderately andheavily degraded grasslands respectively

32 SOC response to grazing

The model predicted a marked decrease in SOC stock in theperiod of high grazing intensity (1990ndash2012) The modeled SOCstocks corresponded well with measured values (Fig 3 Table S1)Although grazing scenarios were set for each grassland stratum onthe basis of rule of thumb biomass removal rates reducing grazingintensity on the Mongolian grasslands showed substantial carbonaccumulation potential except for a marginal gain in non-degraded MM (Fig 4) In addition lower SOC accumulation waspredicted in heavily degraded compared to moderately degradedgrasslands The annual accumulation rates were 332ndash369 and220ndash295 g C m2 yr1 for moderately and heavily degradedgrassland strata respectively The lightly degraded grasslandstrata had moderate SOC accumulation rates ranging from273 to 318 g C m2 yr1 Simulations to 2035 indicated thatreducing grazing intensity would allow SOC stocks to recoverfrom 312 to 414 of the intact stocks in heavily degraded grassland

strata and from 734 to 903 in moderately degraded strataHowever the grasslands did not reach a steady-state SOC level inthe near term as shown by the continuing accumulation trends

4 Discussion

41 SOC loss from grassland degradation

The Century simulations revealed an evident negative impact ofgrassland degradation on SOC stocks This patter is consistent withour geospatial soil inventory and several findings of experimentalstudies from other grassland ecosystems with similar soils andclimates (Huang et al 2010 Zhang et al 2011) SOC depletioncould partly be attributed to the reduction in grass biomassproduction Studies have shown that overgrazing can often lead toan amount of plant photosynthetic tissue consumption andcommunity composition shifts that lower plant production andthereby limit potential litter inputs to the soil (Bagchi and Ritchie2010 Gao et al 2008 Wang et al 2012) A gradient decrease inaboveground biomass and soil carbon stocks from intact soilstowards the heavily degraded soils has also observed in other semi-arid grasslands confirming a close connection between soil carbonand grass biomass production and suggesting that overgrazingcould restrict soil carbon sequestration (He et al 2011 Wen et al2013 Zhang et al 2011) In addition the SOC loss may be likelydriven by increasing erosion due to a decrease in vegetation coverassociated with continuous heavy grazing (Hoffmann et al 2008)Erosion can amplify the negative effects of heavy grazing on SOCand potentially even further coarsening the soil and reducing itscapacity to hold SOC (McSherry and Ritchie 2013) Anotherpossible reason for SOC depletion may be the impact of animaltrampling Animal trampling generally leads to a deterioration ofsoil structure particularly to destruction of soil aggregatesfollowed by a release of physically protected soil organic matterwhich then mineralized (Steffens et al 2008) In Century soilerosion is treated in a simplified form Erosion events areuncoupled from precipitation and only flat monthly soil loss ratescan be entered as input variables (Tornquist et al 2009)Furthermore information regarding soil erosion was unavailablein our study area Thus we have not explored soil erosion hereAlthough it was in fact posited as a potential mechanism for SOCdepletion in several studies (Wiesmeier et al 2012 Steffens et al

Fig 4 Trend of SOC change in degraded Mongolian grassland soils in relation to the different simulated grazing managements Black lines are the SOC equilibrium levlesbefore overgrazing The shading is the uncertainty range MM mountain meadow MWM riparian meadow with mash M riparian meadow MS mountain steppe

282 X Chang et al Agriculture Ecosystems and Environment 212 (2015) 278ndash284

2008) effects of animal trampling on decomposition was notsimulated in Century where decomposition are primarily affectedby molecular properties of the soil organic matter and are modifiedby temperature and moisture (Dib et al 2014) Taken as a wholeCentury appears to provide an accurate depiction of SOC responsesto grazing events though still possibly potential for improving themodel structure and parameterization to further reduce uncer-tainties (Tornquist et al 2009)

42 SOC accumulation in response to reduced grazing intensity

With open access or collectively managed grasslands through-out the country (Olonbayar 2010) it is difficult to implementexperiments on the effect of land use and management on SOC

Fig 5 Predicted changes in active slow passive and total SOC in light degra

stocks in grasslands in Mongolia However it is important to fillknowledge gaps to support policy decisions In this situation theprocess-based modeling approach may be capable of providinguseful projections of SOC response to a variety of grasslandmanagements (Lugato et al 2014) In our study the Centurymodel predicted effective SOC accumulation in degradedMongolian grassland under reduced grazing intensity with SOCaccumulation rates in the near term (ie by 2035) of 2196ndash3691 g C m2 yr1 If these accrual rates are applied to 90 ofMongoliarsquos total forest steppe grassland area (332 million ha)annual accumulation to 2035 would be 73ndash123 Mt C yr1 underfuture grazing recommendations assuming no climate change Bycomparison in 2006 Mongoliarsquos total net GHG emissions were425 Mt C (MNET 2012)

ded mountain meadow (other scenarios give similar trends not shown)

X Chang et al Agriculture Ecosystems and Environment 212 (2015) 278ndash284 283

Our estimated accumulation rates were comparable to thereported values (170ndash370 g C m2 yr1 for 20 cm) in alpinemeadow environments (Fan et al 2012) with similar vegetationand climate conditions to Tariat but lower than or within the lowerend of the ranges of accumulation rates for grasslands in northernChina reported by Wang et al (2011) (23ndash330 g C m2 yr1 0-20 cm) and Guo et al (2008) (213ndash731 g C m2 yr1) However thehigher results in that study were obtained through exclusion fromgrazing which is not feasible in the Mongolian context asextensive livestock production remains the source of livelihoodsfor a considerable proportion of the population (Sankey et al2009) In Tariat surveys of land users suggested that the averagehousehold moves their grazing camp just over two times duringthe summer grazing season However since we cannot be sure thateach specific plot of grassland is not grazed more than once bydifferent herds in the season our scenario settings retained theconservative assumption that grazing is continuous in the summergrazing season If the duration of each grazing event is in factshorter than we simulated the accumulation rate would be higherGiven the strong dependence of rural households in Mongolia onlivestock for their income it is likely that the most feasible route toreducing grazing intensity is through changes in the timing andduration of grazing for example through increased rotation andresting of pastures which may be supported by strengtheningcommunity-based rangeland management institutions (Fernaacuten-dez-Gimeacutenez et al 2015) Efforts to address livestock numbers arelikely to require changes in livestock management and marketingto maintain herdersrsquo incomes (Kemp et al 2013) which woulddepend on longer-term and more systemic changes in theMongolian livestock economy

43 Carbon quality and SOC recovery

Soil carbon is heterogeneous in nature and consists of severalSOC pools (Allen et al 2010) Labile pool accounts for a smallfraction of SOC and has a fast turnover rate while recalcitrant SOCis a large pool and has a slower turnover rate (Davidson andJanssens 2006) Therefore sequestration of atmospheric carbon insoils in a stable form is considered to have potential for global CO2

mitigation The Century simulations for grazing scenarios showedthat the shape of the total SOC time-series is mainly dictated by thechanges occurring within the slow pool (Fig 5) During this periodchanges in the active and passive pools are relatively smallHowever a previous field soil survey in steppe soils in NorthernChina founded that labile SOC is preferentially lost in the course ofsoil degradation under intensive grazing (Wiesmeier et al 2015)According to observations of experiments with reduced grazingintensity or grazing exclusion SOC increases mainly resulted froman accumulation of labile SOC which was sensitive to land use andclimate change (Steffens et al 2011) Therefore the observed SOCincreases after improved grassland management cannot be viewedas contribution to long-term carbon sequestration The differencesbetween the predicted and measured trajectories of carbonfractions may be attributed to particular features of the Centurymodel Century has a rapid transfer rate of SOC from active poolsinto pools with lower decomposition rates (Dib et al 2014) Somechanges in the slow SOC pools are needed to improve modelpredictions in grassland soils even if quantitative SOC pools inCentury often do not easily correspond to measurable entitiesWithout such improvements the magnitude of the predicted SOCsequestration simulated by Century may be amplified

Even after 23 years of improved grazing management the SOCstorage of the intact grassland cannot be restored completely(Fig 4) Wind erosion related to intensive grazing should be takenin to account Evidence gathered in several studies indicated thatwind erosion resulted in a loss of silt and clay particles (Steffens

et al 2008 Wiesmeier et al 2009) Our study also clearly showsthe smaller silt and clay contents were found in more degradedgrassland soils The loss of fine soil particles consequentlyundermines the ability to stabilize SOC which makes the recoveryof the fine mineral fraction SOC largely irreversible (Wiesmeieret al 2015) Such a wind erosion effect is a contributory reasonwhy the SOC did not recover completely However the dynamicsand stabilization mechanisms differed between the soil carbonfractions The physically protected pool may be slow to rebuild pastreserves because reassembly of macro- and micro-aggregatestructures after disturbance might resume over a long timespan(OrsquoBrien and Jastrow 2013) This suggests that SOC restoration inthe aggregate-protected fraction may require a longer time thanour simulations For active pools rebuilding past reserves wasfaster than other fractions Its recovery dynamics depend oncarbon input level and decomposition rate (OrsquoBrien and Jastrow2013) Hence if improved managements fuel long-term grasslandproductivity active carbon pools of a new equilibrium may overthe past values The objective was to simulate effects of grazing ontotal SOC however there is also need to further investigate thedynamics of different SOC fractions

5 Conclusions

Overgrazing has caused widespread degradation in Mongoliansteppes Reducing grazing intensity is a common strategypromoted by government sources to combat grassland degrada-tion However few experimental data of the soil carbon recoveryare available This study assessed the impacts of grasslandmanagement on SOC stocks using a modeling approach combinedwith a geospatial inventory dataset Although there are significantpossibilities for future improvement validation against SOC stocksfrom inventories confirmed the robustness and accuracy of themodified Century model when simulating the effect of grazingmanagement practices on forest steppe grasslands in MongoliaThe prediction of future SOC accumulation potential was in therange of 2196ndash3691 g C m2 yr1 in near term (by 2035) underdecreased grazing intensity Our results demonstrate that reduc-tion in grazing intensity may provide a near term solution forMongolian steppe that have experienced soil carbon loss fromintensive grazing managements The reduction in grazing pres-sures offers a profitable and sustainable solution to our needs forpairing livestock production with SOC recovery

Acknowledgements

This paper is based on work carried out as the National NaturalScience Foundation of China (41230750) the National BasicResearch Program (2013CB956000) the lsquoLinking Herders to CarbonMarketsrsquo project mandated by Swiss Development Cooperation toUnique Forestry and Land Use GmbH (Project 7F-07809) and wassupported by the National Science Foundation for Young Scientists(41303062) and the Fundamental Research Funds for the CentralUniversities (NWAFU-2014Y13057) We would like to thank CindyKeough for providing the Century model version 45

Appendix A Supplementary data

Supplementary data associated with this article can be found inthe online version at httpdxdoiorg101016jagee201507014

References

Administration for Land Affairs 2011 Geodesy and Cartography (ALAGC) AnnualReport of ALAGC ALAGC Ulaanbaatar

Fig 4 Trend of SOC change in degraded Mongolian grassland soils in relation to the different simulated grazing managements Black lines are the SOC equilibrium levlesbefore overgrazing The shading is the uncertainty range MM mountain meadow MWM riparian meadow with mash M riparian meadow MS mountain steppe

282 X Chang et al Agriculture Ecosystems and Environment 212 (2015) 278ndash284

2008) effects of animal trampling on decomposition was notsimulated in Century where decomposition are primarily affectedby molecular properties of the soil organic matter and are modifiedby temperature and moisture (Dib et al 2014) Taken as a wholeCentury appears to provide an accurate depiction of SOC responsesto grazing events though still possibly potential for improving themodel structure and parameterization to further reduce uncer-tainties (Tornquist et al 2009)

42 SOC accumulation in response to reduced grazing intensity

With open access or collectively managed grasslands through-out the country (Olonbayar 2010) it is difficult to implementexperiments on the effect of land use and management on SOC

Fig 5 Predicted changes in active slow passive and total SOC in light degra

stocks in grasslands in Mongolia However it is important to fillknowledge gaps to support policy decisions In this situation theprocess-based modeling approach may be capable of providinguseful projections of SOC response to a variety of grasslandmanagements (Lugato et al 2014) In our study the Centurymodel predicted effective SOC accumulation in degradedMongolian grassland under reduced grazing intensity with SOCaccumulation rates in the near term (ie by 2035) of 2196ndash3691 g C m2 yr1 If these accrual rates are applied to 90 ofMongoliarsquos total forest steppe grassland area (332 million ha)annual accumulation to 2035 would be 73ndash123 Mt C yr1 underfuture grazing recommendations assuming no climate change Bycomparison in 2006 Mongoliarsquos total net GHG emissions were425 Mt C (MNET 2012)

ded mountain meadow (other scenarios give similar trends not shown)

X Chang et al Agriculture Ecosystems and Environment 212 (2015) 278ndash284 283

Our estimated accumulation rates were comparable to thereported values (170ndash370 g C m2 yr1 for 20 cm) in alpinemeadow environments (Fan et al 2012) with similar vegetationand climate conditions to Tariat but lower than or within the lowerend of the ranges of accumulation rates for grasslands in northernChina reported by Wang et al (2011) (23ndash330 g C m2 yr1 0-20 cm) and Guo et al (2008) (213ndash731 g C m2 yr1) However thehigher results in that study were obtained through exclusion fromgrazing which is not feasible in the Mongolian context asextensive livestock production remains the source of livelihoodsfor a considerable proportion of the population (Sankey et al2009) In Tariat surveys of land users suggested that the averagehousehold moves their grazing camp just over two times duringthe summer grazing season However since we cannot be sure thateach specific plot of grassland is not grazed more than once bydifferent herds in the season our scenario settings retained theconservative assumption that grazing is continuous in the summergrazing season If the duration of each grazing event is in factshorter than we simulated the accumulation rate would be higherGiven the strong dependence of rural households in Mongolia onlivestock for their income it is likely that the most feasible route toreducing grazing intensity is through changes in the timing andduration of grazing for example through increased rotation andresting of pastures which may be supported by strengtheningcommunity-based rangeland management institutions (Fernaacuten-dez-Gimeacutenez et al 2015) Efforts to address livestock numbers arelikely to require changes in livestock management and marketingto maintain herdersrsquo incomes (Kemp et al 2013) which woulddepend on longer-term and more systemic changes in theMongolian livestock economy

43 Carbon quality and SOC recovery

Soil carbon is heterogeneous in nature and consists of severalSOC pools (Allen et al 2010) Labile pool accounts for a smallfraction of SOC and has a fast turnover rate while recalcitrant SOCis a large pool and has a slower turnover rate (Davidson andJanssens 2006) Therefore sequestration of atmospheric carbon insoils in a stable form is considered to have potential for global CO2

mitigation The Century simulations for grazing scenarios showedthat the shape of the total SOC time-series is mainly dictated by thechanges occurring within the slow pool (Fig 5) During this periodchanges in the active and passive pools are relatively smallHowever a previous field soil survey in steppe soils in NorthernChina founded that labile SOC is preferentially lost in the course ofsoil degradation under intensive grazing (Wiesmeier et al 2015)According to observations of experiments with reduced grazingintensity or grazing exclusion SOC increases mainly resulted froman accumulation of labile SOC which was sensitive to land use andclimate change (Steffens et al 2011) Therefore the observed SOCincreases after improved grassland management cannot be viewedas contribution to long-term carbon sequestration The differencesbetween the predicted and measured trajectories of carbonfractions may be attributed to particular features of the Centurymodel Century has a rapid transfer rate of SOC from active poolsinto pools with lower decomposition rates (Dib et al 2014) Somechanges in the slow SOC pools are needed to improve modelpredictions in grassland soils even if quantitative SOC pools inCentury often do not easily correspond to measurable entitiesWithout such improvements the magnitude of the predicted SOCsequestration simulated by Century may be amplified

Even after 23 years of improved grazing management the SOCstorage of the intact grassland cannot be restored completely(Fig 4) Wind erosion related to intensive grazing should be takenin to account Evidence gathered in several studies indicated thatwind erosion resulted in a loss of silt and clay particles (Steffens

et al 2008 Wiesmeier et al 2009) Our study also clearly showsthe smaller silt and clay contents were found in more degradedgrassland soils The loss of fine soil particles consequentlyundermines the ability to stabilize SOC which makes the recoveryof the fine mineral fraction SOC largely irreversible (Wiesmeieret al 2015) Such a wind erosion effect is a contributory reasonwhy the SOC did not recover completely However the dynamicsand stabilization mechanisms differed between the soil carbonfractions The physically protected pool may be slow to rebuild pastreserves because reassembly of macro- and micro-aggregatestructures after disturbance might resume over a long timespan(OrsquoBrien and Jastrow 2013) This suggests that SOC restoration inthe aggregate-protected fraction may require a longer time thanour simulations For active pools rebuilding past reserves wasfaster than other fractions Its recovery dynamics depend oncarbon input level and decomposition rate (OrsquoBrien and Jastrow2013) Hence if improved managements fuel long-term grasslandproductivity active carbon pools of a new equilibrium may overthe past values The objective was to simulate effects of grazing ontotal SOC however there is also need to further investigate thedynamics of different SOC fractions

5 Conclusions

Overgrazing has caused widespread degradation in Mongoliansteppes Reducing grazing intensity is a common strategypromoted by government sources to combat grassland degrada-tion However few experimental data of the soil carbon recoveryare available This study assessed the impacts of grasslandmanagement on SOC stocks using a modeling approach combinedwith a geospatial inventory dataset Although there are significantpossibilities for future improvement validation against SOC stocksfrom inventories confirmed the robustness and accuracy of themodified Century model when simulating the effect of grazingmanagement practices on forest steppe grasslands in MongoliaThe prediction of future SOC accumulation potential was in therange of 2196ndash3691 g C m2 yr1 in near term (by 2035) underdecreased grazing intensity Our results demonstrate that reduc-tion in grazing intensity may provide a near term solution forMongolian steppe that have experienced soil carbon loss fromintensive grazing managements The reduction in grazing pres-sures offers a profitable and sustainable solution to our needs forpairing livestock production with SOC recovery

Acknowledgements

This paper is based on work carried out as the National NaturalScience Foundation of China (41230750) the National BasicResearch Program (2013CB956000) the lsquoLinking Herders to CarbonMarketsrsquo project mandated by Swiss Development Cooperation toUnique Forestry and Land Use GmbH (Project 7F-07809) and wassupported by the National Science Foundation for Young Scientists(41303062) and the Fundamental Research Funds for the CentralUniversities (NWAFU-2014Y13057) We would like to thank CindyKeough for providing the Century model version 45

Appendix A Supplementary data

Supplementary data associated with this article can be found inthe online version at httpdxdoiorg101016jagee201507014

References

Administration for Land Affairs 2011 Geodesy and Cartography (ALAGC) AnnualReport of ALAGC ALAGC Ulaanbaatar

X Chang et al Agriculture Ecosystems and Environment 212 (2015) 278ndash284 283

Our estimated accumulation rates were comparable to thereported values (170ndash370 g C m2 yr1 for 20 cm) in alpinemeadow environments (Fan et al 2012) with similar vegetationand climate conditions to Tariat but lower than or within the lowerend of the ranges of accumulation rates for grasslands in northernChina reported by Wang et al (2011) (23ndash330 g C m2 yr1 0-20 cm) and Guo et al (2008) (213ndash731 g C m2 yr1) However thehigher results in that study were obtained through exclusion fromgrazing which is not feasible in the Mongolian context asextensive livestock production remains the source of livelihoodsfor a considerable proportion of the population (Sankey et al2009) In Tariat surveys of land users suggested that the averagehousehold moves their grazing camp just over two times duringthe summer grazing season However since we cannot be sure thateach specific plot of grassland is not grazed more than once bydifferent herds in the season our scenario settings retained theconservative assumption that grazing is continuous in the summergrazing season If the duration of each grazing event is in factshorter than we simulated the accumulation rate would be higherGiven the strong dependence of rural households in Mongolia onlivestock for their income it is likely that the most feasible route toreducing grazing intensity is through changes in the timing andduration of grazing for example through increased rotation andresting of pastures which may be supported by strengtheningcommunity-based rangeland management institutions (Fernaacuten-dez-Gimeacutenez et al 2015) Efforts to address livestock numbers arelikely to require changes in livestock management and marketingto maintain herdersrsquo incomes (Kemp et al 2013) which woulddepend on longer-term and more systemic changes in theMongolian livestock economy

43 Carbon quality and SOC recovery

Soil carbon is heterogeneous in nature and consists of severalSOC pools (Allen et al 2010) Labile pool accounts for a smallfraction of SOC and has a fast turnover rate while recalcitrant SOCis a large pool and has a slower turnover rate (Davidson andJanssens 2006) Therefore sequestration of atmospheric carbon insoils in a stable form is considered to have potential for global CO2

mitigation The Century simulations for grazing scenarios showedthat the shape of the total SOC time-series is mainly dictated by thechanges occurring within the slow pool (Fig 5) During this periodchanges in the active and passive pools are relatively smallHowever a previous field soil survey in steppe soils in NorthernChina founded that labile SOC is preferentially lost in the course ofsoil degradation under intensive grazing (Wiesmeier et al 2015)According to observations of experiments with reduced grazingintensity or grazing exclusion SOC increases mainly resulted froman accumulation of labile SOC which was sensitive to land use andclimate change (Steffens et al 2011) Therefore the observed SOCincreases after improved grassland management cannot be viewedas contribution to long-term carbon sequestration The differencesbetween the predicted and measured trajectories of carbonfractions may be attributed to particular features of the Centurymodel Century has a rapid transfer rate of SOC from active poolsinto pools with lower decomposition rates (Dib et al 2014) Somechanges in the slow SOC pools are needed to improve modelpredictions in grassland soils even if quantitative SOC pools inCentury often do not easily correspond to measurable entitiesWithout such improvements the magnitude of the predicted SOCsequestration simulated by Century may be amplified

Even after 23 years of improved grazing management the SOCstorage of the intact grassland cannot be restored completely(Fig 4) Wind erosion related to intensive grazing should be takenin to account Evidence gathered in several studies indicated thatwind erosion resulted in a loss of silt and clay particles (Steffens

et al 2008 Wiesmeier et al 2009) Our study also clearly showsthe smaller silt and clay contents were found in more degradedgrassland soils The loss of fine soil particles consequentlyundermines the ability to stabilize SOC which makes the recoveryof the fine mineral fraction SOC largely irreversible (Wiesmeieret al 2015) Such a wind erosion effect is a contributory reasonwhy the SOC did not recover completely However the dynamicsand stabilization mechanisms differed between the soil carbonfractions The physically protected pool may be slow to rebuild pastreserves because reassembly of macro- and micro-aggregatestructures after disturbance might resume over a long timespan(OrsquoBrien and Jastrow 2013) This suggests that SOC restoration inthe aggregate-protected fraction may require a longer time thanour simulations For active pools rebuilding past reserves wasfaster than other fractions Its recovery dynamics depend oncarbon input level and decomposition rate (OrsquoBrien and Jastrow2013) Hence if improved managements fuel long-term grasslandproductivity active carbon pools of a new equilibrium may overthe past values The objective was to simulate effects of grazing ontotal SOC however there is also need to further investigate thedynamics of different SOC fractions

5 Conclusions

Overgrazing has caused widespread degradation in Mongoliansteppes Reducing grazing intensity is a common strategypromoted by government sources to combat grassland degrada-tion However few experimental data of the soil carbon recoveryare available This study assessed the impacts of grasslandmanagement on SOC stocks using a modeling approach combinedwith a geospatial inventory dataset Although there are significantpossibilities for future improvement validation against SOC stocksfrom inventories confirmed the robustness and accuracy of themodified Century model when simulating the effect of grazingmanagement practices on forest steppe grasslands in MongoliaThe prediction of future SOC accumulation potential was in therange of 2196ndash3691 g C m2 yr1 in near term (by 2035) underdecreased grazing intensity Our results demonstrate that reduc-tion in grazing intensity may provide a near term solution forMongolian steppe that have experienced soil carbon loss fromintensive grazing managements The reduction in grazing pres-sures offers a profitable and sustainable solution to our needs forpairing livestock production with SOC recovery

Acknowledgements

This paper is based on work carried out as the National NaturalScience Foundation of China (41230750) the National BasicResearch Program (2013CB956000) the lsquoLinking Herders to CarbonMarketsrsquo project mandated by Swiss Development Cooperation toUnique Forestry and Land Use GmbH (Project 7F-07809) and wassupported by the National Science Foundation for Young Scientists(41303062) and the Fundamental Research Funds for the CentralUniversities (NWAFU-2014Y13057) We would like to thank CindyKeough for providing the Century model version 45

Appendix A Supplementary data

Supplementary data associated with this article can be found inthe online version at httpdxdoiorg101016jagee201507014

References

Administration for Land Affairs 2011 Geodesy and Cartography (ALAGC) AnnualReport of ALAGC ALAGC Ulaanbaatar

284 X Chang et al Agriculture Ecosystems and Environment 212 (2015) 278ndash284

Administration for Land Affairs 2012 Geodesy and Cartography (ALAGC) UnifiedLand Territory Report ALAGC Ulaanbaatar

Allen DE Pringle MJ Page KL Dalal RC 2010 A review of sampling designs forthe measurement of soil organic carbon in Australian grazing lands Rangeland J32 227ndash246

Bagchi S Ritchie ME 2010 Introduced grazers can restrict potential soil carbonsequestration through impacts on plant community composition Ecol Lett 13959ndash968

Booker K Huntsinger L Bartolome JW Sayre NF Stewart W 2013 What canecological science tell us about opportunities for carbon sequestration on aridrangelands in the United States Global Environ Change 23 (1) 240ndash251

Bortolon ESO Mielniczuk J Tornquist CG Lopes F Bergamaschi H 2011Validation of the Century model to estimate the impact of agriculture on soilorganic carbon in Southern Brazil Geoderma 167ndash168 156ndash166

Chang X Wang S Zhu X Cui S Luo C Zhang Z Wilkes A 2014 Impacts ofmanagement practices on soil organic carbon in degraded alpine meadows onthe Tibetan Plateau Biogeosciences 11 3495ndash3503

Cheng K Ogle SM Parton WJ Pan G 2014 Simulating greenhouse gasmitigation potentials for Chinese croplands using the DAYCENT ecosystemmodel Global Change Biol 20 (3) 948ndash962

Conant R 2009 Challenges and Opportunities for Carbon Sequestration inGrassland Systems FAO technical report Rome

Conant RT Paustian K 2002 Potential soil carbon sequestration in overgrazedgrassland ecosystems Global Biogeochem Cycles 16 (4) 1143

Conant RT Paustian K Elliott ET 2001 Grassland management and conversioninto grassland effects on soil carbon Ecol Appl 11 343ndash355

Dib AE Johnson CE Driscoll CT Fahey TJ Hayhoe K 2014 Simulating effectsof changing climate and CO2 emissions on soil carbon pools at the HubbardBrook experimental forest Global Change Biol 20 1643ndash1656

Dagvadorj D Natsagdorj L Dorjpurev J Namkhainyam B 2009 MARCC 2009Mongolia Assessment Report on Climate Change 2009 Ministry of NatureEnvironment and Tourism Mongolia

Davidson EA Janssens IA 2006 Temperature sensitivity of soil carbondecomposition and feedbacks to climate change Nature 440 165ndash173

Evans S Burke I Lauenroth W 2010 Controls on soil organic carbon and nitrogenin Inner Mongolia China a cross-continental comparison of temperategrasslands Global Biogeochem Cycles 25 (3) doihttpdxdoiorg1010292010GB003945 GB3006

Fan YJ Hou XY Shi HX Shi SL 2012 The response of carbon reserves of plantsand soils to different grassland managements on alpine meadow of threeheadwater source regions Grassland Turf 32 (5) 41ndash46

Fernaacutendez-Gimeacutenez ME Batkhishig B Batbuyan B Ulambayar T 2015 Lessonsfrom the Dzud community-based rangeland management increases theadaptive capacity of mongolian herders to winter disasters World Dev 68 48ndash65

Follett RF Schuman GE 2005 Grazing land contributions to carbonsequestration In McGilloway DA (Ed) Grassland a Global ResourceWageningen Academic Publishers Wageningen the Netherland pp 265ndash277

Gao YH Luo P Wu N Chen H Wang GX 2008 Impacts of grazing intensity onnitrogen pools and nitrogen cycle in an alpine meadow on the eastern TibetanPlateau Appl Ecol Env Res 6 67ndash77

Guo R Wang XK Lu F Duan XN Ouyang ZY 2008 Soil carbon sequestrationand its potential by grassland ecosystems in China Acta Ecol Sin 28 862ndash867

He NP Zhang YH Yu Q Chen QS Pan QM Zhang GM Han XG 2011Grazing intensity impacts soil carbon and nitrogen storage of continentalsteppe Ecosphere 2 doihttpdxdoiorg101890ES10-000171 art8

Henderson B Gerber P Hilinski T Falcucci A Ojima D Salvatore M Conant R2015 Greenhouse gas mitigation potential of the worldrsquos grazing landsmodelling soil carbon and nitrogen fluxes of mitigation practices Agric EcosystEnviron 207 91ndash100

Hilker T Natsagdorj E Waring RH Lyapustin A Wang Y 2014 Satelliteobserved widespread decline in Mongolian grasslands largely due toovergrazing Global Change Biol 20 418ndash428

Hoffmann C Funk R Wieland R Li Y Sommer M 2008 Effects of grazing andtopography on dust flux and deposition in the Xilingele grassland InnerMongolia J Arid Environ 72 792ndash807

Houghton RA 2007 Balancing the global carbon budget Annu Rev Earth PlanetSci 35 313ndash347

Huang Y Sun W Zhang W Yu Y 2010 Changes in soil organic carbon ofterrestrial ecosystems in China a mini-review Sci China Life Sci 53 766ndash775

Jia Y Yu G He N Zhan X Fang H Sheng W Zuo Y Zhang D Wang Q 2014Spatial and decadal variations in inorganic nitrogen wet deposition in Chinainduced by human activity Sci Rep 4 3763 doihttpdxdoiorg101038srep03763

Kemp DR Han GD Hou XY Michalk DL Hou FJ Wu JP Zhang YJ 2013Innovative grassland management systems for environmental and livelihoodbenefits P Natl Acad Sci U S A 110 8369ndash8374

Lal R 2004 Soil carbon sequestration impacts on global climate change and foodsecurity Science 304 (5677) 1623ndash1627

Liu YY Evans JP McCabe MF de Jeu RA van Dijk AI Dolman AJ Saizen I2013 Changing climate and overgrazing are decimating Mongolian steppesPloS One 8 (2) e57599 doihttpdxdoiorg101371journalpone0057599

Lugato E Panagos P Bampa F Jones A Montanarella L 2014 A new baseline oforganic carbon stock in European agricultural soils using a modelling approachGlobal Change Biol 20 313ndash326

McSherry ME Ritchie ME 2013 Effects of grazing on grassland soil carbon aglobal review Global Change Biol 19 1347ndash1357

Ministry of Nature Environment and Tourism (MNET) 2012 Mongolia SecondNational Communication under the United Nations Framework Convention onClimate Change Ulaanbaatar

National Agency for Meteorology and Environmental Monitoring of Mongolia(NAMEM) 2015 Mongolian State of Rangeland Report Ulaanbaatar

National Statistical Office of Mongolia (NSO) 2014 Available at httpwww1212mnencontentsstatscontents_stat_fld_tree_htmljsp (accessed09115)

Neff JC Reynolds RL Belnap J Lamothe P 2005 Multi-decadal impacts ofgrazing on soil physical and biogeochemical properties in southeast Utah EcolAppl 15 87ndash95

OrsquoBrien SL Jastrow JD 2013 Physical and chemical protection in hierarchical soilaggregates regulates soil carbon and nitrogen recovery in restored perennialgrasslands Soil Biol Biochem 61 1ndash13

Livelihood Study of Herders in Mongolia In Olonbayar M (Ed) Mongolian Societyfor Range Management Ulaanbaatar

Sankey TT Sankey JB Weber KT Montagne C 2009 Geospatial assessment ofgrazing regime shifts and sociopolitical changes in a Mongolian rangelandRangeland Ecol Manag 62 522ndash530

Smith P Martino D Cai Z Gwary D Janzen H Kumar P McCarl B Ogle SOrsquoMara F Rice C Scholes B Sirotenko O 2007 Agriculture In Metz BDavidson OR (Eds) Climate Change 2007 Mitigation Working Group IIIContribution to the Fourth Assessment Report of the Intergovernmental Panelon Climate Change Cambridge University Press Cambridge United Kingdomand New York NY USA

Steffens M Koumllbl A Totsche KU Koumlgel-Knabner I 2008 Grazing effects on soilchemical and physical properties in a semiarid steppe of Inner Mongolia (PRChina) Geoderma 143 63ndash72

Steffens M Koumllbl A Schoumlrk E Gschrey B Koumlgel-Knabner I 2011 Distribution ofsoil organic matter between fractions and aggregate size classes in grazedsemiarid steppe soil profiles Plant Soil 338 63ndash81

Tornquist CG Mielniczuk J Cerri CEP 2009 Modeling soil organic carbondynamics in Oxisols of Ibirubaacute (Brazil) with the Century Model Soil Till Res105 33ndash43

von Wehrden H Hanspach J Kaczensky P Fischer J Wesche K 2012 Globalassessment of the non-equilibrium concept in rangelands Ecol Appl 22 (2)393ndash399

Wang Y Zhou G Jia B 2008 Modeling SOC and NPP responses of meadow steppeto different grazing intensities in Northeast China Ecol Model 217 (1ndash2) 72ndash78

Wang S Wilkes A Zhang Z Chang X Lang R Wang Y Niu H 2011Management and land use change effects on soil carbon in northern Chinarsquosgrasslands a synthesis Agric Ecosyst Environ 42 329ndash340

Wang S Duan J Xu G Wang Y Zhang Z Rui Y Luo C Xu B Zhu X Chang XCui X Niu H Zhao X Wang W 2012 Effects of warming and grazing on soilN availability species composition and ANPP in an alpine meadow Ecology 932365ndash2376

Wen L Dong SK Li YY Li XY Shi JJ Wang YL Liu DM Ma YS 2013 Effectof degradation intensity on grassland ecosystem services in the alpine region ofQinghai-Tibetan Plateau China PloS One 8 (3) e58432 doihttpdxdoiorg101371journalpone0058432

White RP Murray S Rohweder M Prince SD Thompson KM 2000 GrasslandEcosystems World Resources Institute Washington DC USA

Wiesmeier M Steffens M Koumllbl A Koumlgel-Knabner I 2009 Degradation andsmall-scale spatial homogenization of topsoils in intensively-grazed steppes ofNorthern China Soil Till Res 104 299ndash310

Wiesmeier M Steffens M Mueller C Koumllbl A Reszkowska A Peth S Hor R nKoumlgel-Knabner I 2012 Aggregate stability and physical protection of soilorganic carbon in semioumlarid steppe soils Eur J Soil Sci 63 22ndash31

Wiesmeier M Munro S Barthold F Steffens M Schad P Koumlgel-Knabner I 2015Carbon storage capacity of semioumlarid grassland soils and sequestrationpotentials in Northern China Global Change Biol doihttpdxdoiorg101111gcb12957 in press

Zhang G Kang Y Han G Mei H Sakurai K 2011 Grassland degradation reducesthe carbon sequestration capacity of the vegetation and enhances the soilcarbon and nitrogen loss Acta Agric Scand B-S P 61 356ndash364