conservation and policy pigs in East-Central Europe between 1940 and 2014: Consequencesfor Changing year-round habitat use of extensively grazing cattle, sheepand Agriculture, Ecosystems and Environment

Changing year-round habitat use of extensively grazing cattle, sheep and pigs in East-Central Europe between 1940 and 2014: Consequences for conservation and policy

A. Varga

*, Zs. Molnár

, M. Biró

, L. Demeter

, K. Gellény

, E. Miókovics

, Á. Molnár

, K. Molnár

, N. Ujházy

, V. Ulicsni

, D. Babai

aMTACentreforEcologicalResearch,InstituteofEcologyandBotany,H-2163Vácrátót,Alkotmányu.2–4,Hungary

bUniversityofPécs,InstituteofBiology,H-7625Pécs,Ifjúságútja6,Hungary

cDepartmentofEcology,FacultyofScienceandInformaticsoftheUniversityofSzeged,H-6726Szeged,Középfasor52,Hungary

dDepartmentofPlantScienceandBiotechnology,GeorgikonFacultyofthePannonUniversity,H-8360Keszthely,Festeticsu.7,Hungary

eFacultiesofAgriculturalandEnvironmentalSciencesoftheSzentIstvánUniversity,H-2100GödöllÅ,PáterK.u.1,Hungary

fDepartmentofEnvironmentalandLandscapeGeography,FacultyofScience,ELTE,H-1117Budapest,PázmányPétersétány1/C,Hungary

gMTAResearchCentrefortheHumanities,InstituteofEthnology,H-1014Budapest,Országházu.30,Hungary

ARTICLE INFO

Articlehistory:

Received24September2015 Receivedinrevisedform12April2016 Accepted17May2016

Availableonlinexxx

Keywords:

Agroforestrysystems Land-usehistory Landscapeecology Large-scalegrazing Post-soviet Sustainablegrazing

ABSTRACT

ManyhabitatsinEuropehavebeenmanagedbygrazingforthousandsofyears.However,extensive grazingsystemsarebecomingincreasinglyrareintheregion,andthereisalackofunderstandingofthe functioningofthesesystems.

Wecarriedout147structuredinterviewsin38landscapesthroughouttheCarpathianBasin,with3–5 informants/landscape.Thenumberofactivelygrazingcattle,sheepandpigs,theiryear-roundhabitatuse andtheproportionofherdsactivelytendedweredocumentedforfourcharacteristichistoricalperiods (before,duringandaftersocialistcooperativesandafterEUAccession).

Thenumbersofgrazing cattleandsheephad decreasedsubstantiallyby2010(by71%and49%, respectively),whilepiggrazingalmostdisappearedbythe1970s.Cattleprimarilygrazedhabitatswith tallervegetation.Sheepgrazeddrypasturesandstubbles,whilepigsweredrivenintomarshesand forests.Ingeneral,theimportanceofdryandwetgrasslandsincreased,whilethesignificanceofmarshes, stubblefields,vegetationalonglinearelements,secondgrowthonhaymeadows,wood-pasturesand forests decreased over time. Approximately half of the grazed habitats were not typical pasture grasslands,andfunctionedassupplementarypasturesduringdroughts,autumnandwinter.Thenumber ofhabitattypesgrazedpermonthpersitedropped,andherdingdecreasedsubstantially,inparticularin thecaseofcattleandpigs.

Contributingfactorsoftheeconomicandsocialchangesoftheexaminedperiodincludedthecollapse of the communist-era legal framework, the intensification of livestock husbandry, EU Common AgriculturalPolicy(CAP)regulations,andtheriseofanatureconservationethic.

Weconcludethatagriculturalpoliciesshouldtakeintoaccountthefullspectrumofhabitattypes necessaryfortheeffectiveoperationofextensivegrazingsystems.Wearguethatconservation-oriented extensivegrazingshouldusethetraditionalwisdomofherdersbutadaptedtothepresentsituations.

ã2016ElsevierB.V.Allrightsreserved.

1.Introduction

Extensive grazing systems (also called large-scale grazing, PlachterandHampicke,2010)arepredominantlybasedonnatural and semi-natural habitat types and non-intensively managed

livestockbreeds,keptusuallyatrelativelylowstockingdensities.

Thesystemisoftenfine-tunedtolocalenvironmentalandsocio- economicconditions(MeuretandProvenza,2014;Molnár,2014;

Reidetal.,2008).Extensivegrazingsystemshaveplayedandstill playanimportantroleinmaintainingbiological,andalsocultural diversity (Halada et al., 2011; Oppermann, 2014; Plachter and Hampicke, 2010; Rodríguez-Ortega et al., 2014; Vera, 2000).

Additionally, extensive grazing systems are important tools for

* Correspondingauthor.

E-mailaddress:varga.anna@gmail.com(A. Varga).

http://dx.doi.org/10.1016/j.agee.2016.05.018 0167-8809/ã2016ElsevierB.V.Allrightsreserved.

resilient and sustainable agroecosystem management, and for ensuring food security. In addition, they contribute to the maintenance ofseveral ecosystem servicesof biodiversity such as biomass production, control of pests and disease, and pollination (Baumgärtner, 2007; Bruun and Fritzbøger, 2002;

Frisonetal.,2011;Heikkinenetal.,2012;Poschlodetal.,2002;

Rodríguez-Ortegaetal.,2014).Extensivegrazingisoftenbeneficial forconservationpurposesaswell(Báldietal.,2013;Plachterand Hampicke,2010;Töröketal.,2014;WallisDeVriesetal.,2004).A highnumber ofhighnature-value habitatsrequire extensiveor transhumantgrazingmanagementinEurope(Bunceetal.,2004;

Haladaetal.,2011;Harteletal.,2013;PoschlodandWallisDeVries, 2002;vanUytvanckandVerheyen,2014;Vera,2000).

Inconnectionwiththeresilientmaintenanceofextensiveland- use systems, several authors emphasize the importance of landscapescaleandhistoricaltimescale(Agnoletti,2014;Fischer etal.,2012;Loosetal.,2015;PlachterandHampicke,2010;Vera, 2000), aswell asoftraditionalecologicalknowledge(Agnoletti, 2014;Berkesetal.,2000;Cevascoetal.,2015;Molnáretal.,2008;

PlieningerandBieling,2013).

ExtensivegrazinginEuropeisdecreasinginmanyplaces;the associatedtraditionalecological knowledgealsoisdisappearing (Hernández-Morcilloetal.,2014;Oteros-Rozasetal.,2013;Varga and Molnár, 2014). The main role of traditional ecological knowledgeinthecaseofextensivegrazingistooptimiseextensive exploitationofthebiomassproducedinthelandscape.Thisisa kindofworkmanshipinlandscapeecology(cf.JohnsonandHunn, 2010).Itcomprisesknowledgeofwhichbiomassinthelandscape is best for grazing, where, when, how and by which livestock (Molnár,2012;Molnáretal.,2015;Veraetal.,2007).Inlandscapes where biomass production has a high inter- and intra-annual variability,e.g.steppeandalpineareas,extensivegrazingsystems must adapt tothese heterogeneities (Gugi9c,2009; Meuret and Provenza, 2014; Molnár, 2012; Plieninger et al., 2015). Beside typicalpasturegrasslandhabitattypes(dryand wetgrasslands, wood-pastures)anumberofdifferentotherhabitattypes(suchas stubble fields, second growth hay, marshlands, forests, and vegetation along linearelements) wereand are still alsotaken advantageof(Barrantesetal.,2009;Vera,2000).

Muchagricultural,ecologicalandconservationbiologicalwork hasfocusedontheroleofgrazingondedicatedpasturegrasslands, but less research is available on grazing other habitats (e.g., marshes,forests,andstubbles)(butseee.g.Andresenetal.,1990;

Middletonetal., 2006;Plachterand Hampicke,2010; Poschlod etal.,2002;RoturierandRoué,2009;Vera,2000).Verylittleis knownabouttheroleofreservepastures(e.g.grazingonarable lands) in extensive grazing systems based predominantly on (semi-)naturalhabitats(butsee,e.g.,Barrantesetal.,2009;Molnár etal.,2015;Toro-Mujicaetal.,2015;Vera etal.,2007).Current European agricultural policy also exacerbates this divide by separatingsubsidiestograsslands,forestsandcroplands(Olmeda etal.,2014).

Thecontinuouslypresentherdermanagedherdbehaviourand daily grazing circuits, often ensuring grazing optimised tothe forage offered by the landscape (Meuret and Provenza, 2014;

Molnár,2014;Oteros-Rozasetal.,2013).However,asaresultofthe socio-economic changes of the past decades, the number of knowledgeableherdersdeclined,throughoutEuropeand inthe post-Soviet countries alike (Varga and Molnár, 2014). Though pastoral grazing declined, it is still a living practice in many marginalregions(Molnár,2014;RoturierandRoué,2009;Oteros- Rozasetal.,2013).

InmostEuropeancountriesarichhistoricalandethnographic literatureisavailableonextensivegrazingsystems(seee.g.Bellon, 1996; Gunda, 1940; Jacobeit, 1961; Wealleans, 2013). These, however, rarely document landscape ecological and habitat

aspects of extensive grazing systems(but see in Hungary, e.g., Andrásfalvy,2007;Gunda,1968;Tálasi,1936).Missinglandscape ecologicalresearchissometimesimpededbytheexplicitpartialor totalprohibitionofgrazingoncertainhabitatsbysomenationalor EuropeanCommunityrulesandlegislation(forinstancetheforest law,occasionallyEUCAP;Barrantesetal.,2009;Haraszthy,2014;

VargaandMolnár,2014).

The main objectiveof this articleis to document landscape ecological features of extensive grazing and their respective changes over the past few decades in the post-communist countriesoftheCarpathianBasin.Theactualresearchquestions wereraisedasfollows:

1.How many total head of cattle, sheepor pigs wereand are grazed extensively in the areas of the communities under investigation?

2.Inwhichmonthonwhichhabitattypesdid/doesthelivestock graze throughoutthe yearand howhasthe importanceand number of habitat types grazed variedover the past 60–70 years?

3.How has the use of non-typical pasture habitat types (e.g.

stubbles,marshes,forests)variedacrosstheyearandoverthe past60–70years?

4.Howwasandisgrazingaccomplished(viaherdingorfencing)?

Researchwascarriedoutat38locationsacrosstheCarpathian Basin,studyingfourhistoricalperiodsbetween1940and2014.

2.Studyareaandmethods 2.1.Studyarea

Changesinthegrazingsystemwerestudiedin38landscapesof theCarpathianBasininCentralEurope.Researchwasconductedin sixpost-communistcountries,predominantlyinHungary,andin Croatia,Serbia,Romania,UkraineandSlovakia(Fig.1,forname, totalareaofmunicipalitiesandthecoverandchangeofgrasslands seeAppendix1).

The climateof thestudy areas is typically continental with Atlantic, sub-Mediterranean and, to a lesser extent, alpine influences. Annual average temperature varies around 11C;

annualaverageprecipitationrangesupto500–1200mm(Condé etal.,2002).Studysiteswereselectedtorepresentthemosttypical vegetationzonesandthusgrazingsystemsintheCarpathianBasin (the zone of Picea abies forestsin the higher mountains (800– 1200ma.s.l.),FagussylvaticaandQuercuspetraea,Q, cerrisforest regions in themountain and hillranges (200–800ma.s.l.), and loess,sandandsaltyforeststeppelandscapesinthelowlands(50– 200ma.s.l.)).

Overthepastseveralcenturies,small-scalepeasantandallodial land-usesystemstypicallyshapedwildlifeandlandscapeinthe Carpathian Basin. Extensive livestock management played a primary role in transforming the vegetation, for instance by creatingwood-pasturesandpasturesintheplaceofformerforests, changing the species composition of primary steppes, and developinghaymeadows(Andrásfalvy,2007;BabaiandMolnár, 2014;Bellon,1996;Vargaetal.,2015).

Socialandeconomicchangesinthe20thcentury(forinstance, socialist transformation of agriculture, post-communist trans- formationsafter1989,andthentheaccessiontotheEU)brought aboutsignificantchangesintherulesofgrazing,numberofgrazing livestock and breed composition (Beaufoy and Marsden, 2010;

Bodó,2001).Correspondingly,thenumberofjobsinagriculture diminished gradually, and barely reaches 4.9% in Hungary at present(HungarianCentralStatisticalOffice,2014).

xxx–xxx

2.2.Materialsandmethods

Westudiedthemostcommonlivestockspecies:cattle,sheep andpig.Weincludedpigsinthestudybecausethisspeciesgrazed onspecialhabitatsandinadistinctway.Correspondingly,alarge quantity of historical data is available about extensive porcine grazing (e.g. Szabadfalvi, 1991). Also, documentation of now- vanished pig pasturing practices was seen as important for conservationmanagement(cf.BeinlichandPoschlod,2002).Since horsesandwaterbuffalosweremainlyusedasdraughtanimals and, as a consequence, grazed on pastures only occasionally, thereforethesespecieswerenotstudied.Grazingofgoats,geeseor ducksaffectedarelativelysmallareaonly(cf.Andrásfalvy,2007).

Breedsofcattleandpigschangedsignificantlyduringthestudy period.ThepercentagedominanceofHungarianSimmentalcattle rangedupto83%in1942,butnowstandsatonly14%(Hollóand Szabó, 2011). Its importance was overtaken by the Holstein- Friesianfrom1972onwards,andinthelastfewdecadesingrowing numberbybeefcattlebreeds(suchasCharolais,AberdeenAngus, LimousineandHereford)(OECD-FAO,2012).Whiteintensivepig breedsreplacedthetraditionalMangalitzabreedduringtheyears

betweenthe1960sandthe1970s,andasaconsequence,pasturing ofpigswasabandoned(Bodó,2001).Amongsheep,Merinotypes arepredominantthroughoutthestudyperiod(Racka,andTsigai weresignificantupuntilthe1930s).Themeatproduction(totalof cattle,pigsandsheepinHungary)was1249,460and697thousand tonsin1971,2000and2010,respectively.Productionofmilkwas 4.2, 1.7, 2.1, 1.6 million liters in 1935, 1971, 2000, and 2010, respectively.

Fourhistoricalperiodswithtypicalsocial-economicconditions affectingthepost-communistcountriesintheCarpathianBasinas awhole–irrespectiveofcountry–werestudied(Table1).

Foreachsite,structuredindoorinterviewsweremadewith3–5 respondents(147intotal,112menand35women;meanageof interviewees:70.1;theyoungestwere29,theoldest96yearsold;

the proportion of non-academic informants is 89%).Interviews wereconductedinHungarianandrecordeddigitallyornoteswere taken.

Thefirstquestionsoftheinterviewsconcernedthenumberof livestockonpasturesinthecaseofthethreespeciesstudied.Our questionswere(34questionsperinterview):“Howmanyheadof cattle/sheep/pigswerepresentonthepasturebefore/during/afterthe

Table1

Theperiodsinvestigatedandtheirmajorfeatures.

Period Years Mainfeaturesoftheperiod

I. Beforecooperatives 1940–1955 Theso-calledtraditionalperiodofagriculturebeforethesocialisttransformation, includingmainlysmallholderfarmsandlandlordestates

II. Duringcooperatives 1965–1980 Heydayofsocialistagriculture,includingforcedmechanisation,useofsynthetic fertilisersandchemicals

III. Afterthepolitical transitionin1989

1992–2000 Theperiodafterthepoliticaltransition(1989)withre-organisedcapitalist farmingpracticesbutlimitedfinancialsupportforfarmers

IV. Thelasttenyears(after EU-accession)

2005–2014 TheperiodafterEUaccession(2004);significantagriculturalsubsidies,effectsof CAP

Fig.1.Mapofthe38studysites(blackdotsonthemap)intheCarpathianBasin,CentralEurope.

Mapsource:ASTER-DEM,USGS,2009.

cooperatives/inthelast10years?”Estimatesfromrespondentswere necessary because the figures of the Central Statistical Office provideonlyatotalnumberoflivestockforeachsettlementand dataarenotadequatetodeterminetheactualnumberofgrazing livestock.Thetwofiguresareclosetoeachotheronlyinthecaseof sheep.Thechangeinthenumberoflivestockwascomparedusing Kruskal–Wallis H test, in the PAST 3.08 programme package (Hammeretal.,2001).

Inthesecondpartoftheinterviewsquestionsongrazedhabitats wereposedforthethreespeciesandfourperiods,respectively.The aimwas to document thenumber of habitattypes grazed.The respectivequestions were asfollows (3124 questions/inter- view):“Wheredidcattle/sheep/pigsstay/grazeinJanuary/February. .. etc.before/during/afterthecooperatives/inthelast10years?”To make data collectionuniform, 14broad but clearly identifiable habitatcategoriesweredetermined,anddatacollectors(11authors ofthispaperand7additionalresearchers;seeAcknowledgements) encodedprimaryinterviewdataintothosecategories.Investigated habitattypes included the following:(1) dry and mesophilous grasslands including loess, saline, rocky and sandy grasslands, mostly steppe grasslands dominated by Festuca spp.; (2) wet grasslands (such as marsh and fen meadows characterised by AlopecurionandMolinionspp.);(3)wood-pasturesandshrublands (withscattereddeciduousandconiferoustreesandbushes);(4) vegetationalonglinearfeatures(grasslandsandshrubalongroads, ditches,smallstreamsetc.);(5)haymeadows(i.e.whenthesecond growth onthesemeadowswasgrazed seasonally); (6)marshes,fens (withPhragmitesaustralis,Phalarisarundinacea,tallsedgesetc.);(7) forestsandforestedges(deciduousand coniferous);(8)stubble fields (e.g.wheat,barley,corn,sugar beetstubbles);(9)fallows (temporarilyabandonedfields);(10)cropland(e.g.cereals);(11) yards;(12)stables.Duetotheverylownumberofcases,habitatsNo.

9,10and11wereomittedfromtheanalysis.

Thequestionsofpartthreeoftheinterviewsconcernedherding practices:“Was/isthereanyherderwiththegrazinglivestock?”,and

“Is/was livestockkeptinpermanent fencing orinportableelectric fencing?”“Whichwas/ismoretypical:onlyherding,onlyfencingor both?”

Datafromthe38studysiteswerearrangedintables.Eachand everyoccurrence ofa givenhabitatwas summarisedbymonth, livestockspeciesandperiod,andthepercentageratioofhowmany timeseachhabitattypewasmentionedwasdetermined.Takingthe numberofmentionsin thefirst periodas100%,thechangesin significanceofeachhabitattypewascalculatedforPeriodsII,IIIand IV.Thenumberofgrazedhabitattypes(habitat-usediversity)was determinedforamonthlybreakdownbysiteandperiod.Thep-value ofdifferencesbetweenperiodswasassessedbyapermutationtest.

DiversitydataweredisplayedusingtheRi3863.2.1.programme(R CoreTeam,2015).Ineachfigurepresentingtheannualcyclethe mostfrequentlymentioned initialmonth of thegrazingseason (April)isshowninthefirstcolumn.DataofpigsinPeriodsIIIandIV arenotshownonFig.2,becausethefigureswouldbemisleading duetothesmallsamplesizeofthepiggrazing.

3.Results

3.1.Numberoflivestockongrazinglandbetween1940and2014 Thenumberoflivestockongrazinglandsubstantiallydecreased for all three species from the 1940s through the mid-2000s ( Appendix2).Bytheend ofthecommunistperiod,grazing pigs vanishedfromtheCarpathianBasinalmostentirely.Thestockof sheepgrewinthecooperativeperiod,onlytodeclinelater.From PeriodIIItoPeriodIV,grazingcattleincreasedbyapproximately 10%asawholebutwerefurtherreducedinmostsites.Allinall, todayabouthalfasmanysheep(51%)andathirdasmanycattle

(approximately 29%) graze on pastures in comparison to their numbers60–70yearsago.

3.2.Annualcycleofhabitatuse

Monthlychangesofgrazedhabitatsacrossperiodsareshown onFig.2andinAppendix3.Fig.2illustratestheratioofgrazed habitats in relation to one another (as a percentage value of frequencies of all sites in the month in question). Appendix 3 showsthepercentageofsiteswherethehabitatconcernedwas usedforgrazinginthegivenmonthinthefourrespectiveperiods.

Themostcommonlygrazedhabitatsweredrygrasslands.Sheep grazedthemthroughouttheyearbutmainlyfromMarchon,while cattlegrazedsuchsitesfromAprilon.Cattleandsheepgrazeddry grasslandsuptolateautumnbutinagraduallydiminishingratio (30%and 20–28%,respectively).Theirsignificanceinthecaseof pigswaslessprominent.

Grazing onwetgrassland was alsosignificant(pigs 20–24%, cattle 18–25%, sheep10–16%). Theirimportance usually dimin- ishedfromMaythroughFebruary.Forpigs,wetgrasslandswereof greaterimportancethandrygrasslands(18%vs.12%,respectively).

Wood-pastures and shrublands were grazed by all three species.Theirsignificance increasedinPeriodIV.Grazing cattle sometimesstayedonthemthroughouttheyear.Sheepusedwood- pasturesandshrublandsacrosstheyearinallperiods,butmost intensivelyinwinterandspringtime(10–14%inwinterand10–12%

inAprilandMay).Pigsalsousedwoodpasturesandshrublands throughpracticallythewholeyear.

Grazing along linear structures (mainly roadside grazing, banksofditchesandcanals,narrowriparianzonesalongstreams) playedimportantrolesforallthreespeciesthroughouttheyear.

Inregionswherecattleweredrivenoutinwintermonths(mainly inPeriodsIandII), roadsides,banksofditchesandcanals,and riparian zones along streams represented the most important habitat(42–66%).

Haymeadowswereusedattheendofthesummer,andduring autumn.Secondgrowthwasgrazedthere.Theratioreached10– 12% for cattle and sheep (somewhat higher for cattle). Spring grazingonhaymeadowswasinfrequent;itwastypicalonlyina fewlocationsintheCarpathianBasin(suchasGyimesinRomania), mainlyinPeriodIonly.Pigsneverwenttohaymeadows.

Marshesandfenswereonceimportantmainlyaspasturesfor cattleandpigs.Theirsignificancewassmallandwasdecliningin thecaseofsheep.Theimportanceofmarshesgrewforcattlein the cooperative period. However, grazing cattle in marshes during thewinter andearlyspringperiodis now beingforced out. Pigs were driven to these habitats in Periods I and II throughouttheyear,sometimesthiswasthedominant typeof pig pasture (and still is today wherever extensive grazing practicessurvive).

Forestgrazing waswidespreadin allfourperiodsfor allthe three species in most sites. Cattle grazed in the forest mainly duringthesummermonths.Wintermonthfigureswereverylow (thoughtheyshowedagrowingtendencyrecently).Forestgrazing throughouttheannualcyclevariedbetween8and14%,declining inperiodIV tosomeextent(10%). Sheeproam (orroamed) the woodsthroughouttheyear,includingthewintermonths(9–12%), withaslightpeakinautumn.Anautumnmaximumcanalsobe observed in thecase of pigs. Survivingextensive pig pasturing happensinforests,forthemostpart.

Grazing of stubble fields started from mid-summer, wheat stubblesfromJulyandcornstubblesfromtheendofSeptember, respectively.Stubblefieldswerevisitedbycattleonanumberof sites,yettheywereimportantgrazinglandmainlyforsheep(5%

and 9–15% during the first three periodsfor cattle and sheep, respectively).Inwintermonthstheratioofstubblefieldsreached

xxx–xxx

10–12%and16–24%forcattleandsheep,respectively.Cattledid notgrazestubblesinwinter,whilethishappenedfrequentlyfor sheep (9–19%). Pigs also visited stubble fields often (9–14%), includingthewintermonths.Nowgrazingofjustsheepistypical onstubblefields.

Thelengthofthegrazingseasonwasshortenedbyonemonth forcattleanddidnotchangeforsheepoverthepast60–70years.

SincePeriodIIIithasbecomecommonforlivestocknottospendall seasonongrass,buttobedrivenintothestableseverynowand thenandsometimesduringthesummerperiodaswell(seegray colourinFig.2).

3.3.Changingfrequencyofhabitatuse

TakingperiodIasabaseline(100%),Fig.3providesfrequencies of mention of each habitat type in the other three periods.

Grazingof cattle onmost habitat types declined continuously duringthepast60–70years.Fewerandfeweranimalsutilisedthe biomassfromfewerandfewerhabitattypes.Grazingwasreduced mainly on more fragmented, non-pasture grasslands, such as along linear features, second growth, marshland, and mostly

stubblefields.Wood-pasturesandforestsrepresentanexception:

their proportion increased to a slight extent in the period following the political transition (Periods III and IV), after a declineduringthecooperativeperiod(PeriodII).Habitatuseby sheepwastransformedduringthecooperativeperiodtoamajor extent.Useof drygrasslands,stubblefields andlinear features increased and that of hay meadows and marshes decreased.

PeriodIIIshowedsignsofreset.InPeriodIVcertaintrendstook anotherturn.Grazingwithpigsdecreasedonallhabitatsduring the cooperativeperiod and vanished practicallycompletely by PeriodIII(notshownonFig.3).

3.4.Diversityofhabitatusebygrazinglivestockduringtheannual cycle

Habitat-usediversitybycattlewasreducedsignificantlyinthe periodbetweenAprilandOctoberduringthepast60–70years,but this decline stoppedinmostmonthsbetweenPeriodIIIand IV (Fig.4). No significantdifferences were foundacross historical periodsinthemonthsbetweenNovemberandMarch.Diversityof habitatuseinthecaseofsheepwasalsoreducedslightlyinthe Fig.2.Therespectiveratiosofuseforeachhabitattype(apercentagevalueoffrequenciesofallsitesinthemonthinquestion)duringtheannualcycleinthefourhistorical periodsstudied(1940–1955,1965–1980,1992–2000and2005–2014)in38studysitesintheCarpathianBasin(asaratioofallmentions).DataforpigsforPeriodIIIandIV haveextremelylowsamplesizes,andthusarenotshown.

past60–70yearsbetweenMayandAugust,butthesereductions werenotstatisticallysignificant.Declinewassignificantinthecase ofpigsineachmonth(Aprp<0,Mayp<0.00001,Junp<0.00001, Aug p<0.00001, Sep p<0.00001, Oct p<0.00001, Nov p

<0.00001,Decp=0.00001,Janp=0.00004,Febp=0.00002,Mar p<0.00001.Levelofsignificance:P<0.05).

3.5.Contributionofnonpasture-grasslandhabitatsthroughoutthe year

Theratioofnonpasture-grasslandhabitats(e.g.secondgrowth onhaymeadows,marshes,vegetationalonglinearfeatures,forests andstubblefields)wasslightlyreducedinthepast60–70years, butremainedsignificanttodate(Fig.5).Dominanceofpasture- grassland habitats (dry grasslands, wet grasslands, and wood- pastures)wasmosttypicalinMay–JuneandinApril–Juneinthe caseofcattleandsheep,respectively.Useofotherhabitattypes increasedfrom mid-summer, most dramatically in the case of sheep.Theirusereachedamaximuminlateautumnandduring thewinterinallfourhistoricalperiods.

3.6.Herdedvs.fencedgrazing

In PeriodI allthree species wereherdedalmost exclusively (Fig.6.).Inthecooperativeperiodfencedgrazingemerged,mainly withcattle,butithasneverbecomeadominantpractice.Inthe past ten years enclosures (mostly by electric fencing) has increased,butpasturingbyherdersisstillthedominantformof extensivegrazinginthestudiedregion.

4.Discussion

4.1.Changingnumberofextensivelygrazinglivestock

Atthe38studysitesinvestigatedthenumberofcattle,sheep andpigskeptinextensivegrazingsystemshasdecreasedfromthe 1940suptodate.Themaincauseforreducednumbershasbeen intensificationofmanagement,theextentofwhichdependedon thespeciesconcerned.Theswineindustrycouldmostreadilybe intensified,cattletoalesserextentandsheephardlyatall(Bodó, 2001;Keszeietal.,2000).Areaofgrazeablelandhasnotchanged

Fig.4. Diversityinhabitatuse(numberofdifferenthabitattypesusedaspasturesinanygivenmonthinthe38studysites)bycattleacrosstheannualcycleintheCarpathian Basin.Boxesrepresentthefourhistoricalperiods(I.1940–1955,II.1965–1980,III.1992–2000andIV.2005–2014).Cattle:Aprp=0.01447,Mayp=0.00063,Junp=0.00004,Jul p=0.00001,Augp=0.00002,Sepp=0.00001,Octp=0.00028,Novp=0.06601,Decp=0.59593,Janp=0.83531,Febp=0.77312,Marp=0.31694;sheep:Aprp=0.51592,May p=0.32481,Junp=0.25967,Julp=0.28205,Augp=0.33726,Sepp=0.67905,Octp=0.59882,Novp=0.56988,Decp=0.75851,Janp=0.72672,Febp=0.67319,Mar p=0.81177.

Fig.3.Trendsinchangesofhabitatuseoverthepast60–70years,takinglevelsofPeriodI(1940–1955)asthereference(100%),PeriodII(1965–1980),PeriodIII(1992–2000) andPeriodIV(2005–2014).

xxx–xxx

considerablyduringthisperiod(Appendix1),sothiscannotbethe causeofthedrasticchangesinlivestocknumbers.

Thenumberofcattlekeptextensivelydeclinedgraduallyand significantly.Theunderlyingcausemightbethestatutorybreed change imposed on cooperatives and farmers in Period II in Hungary(Hungarian Simmentalhad tobereplacedbyHolstein Friesian,Bodó,2001).Withthespread ofintensivedairy farms, villageherdshavedisappearedin HungarybyPeriodIV. Atthe same time the number of grazing meat cattle, albeit growing, remainedlow.Duetoalackofstudies,noinformationisavailable onwhichimpacts variouscattlebreeds haveonpasturesof the CarpathianBasin(butcf.Rooketal.,2004).Averagebodyweightof breedsincreases,butlesssoofgrazingcattlethanofcattlekeptin intensivefarms.SheepstockgrewinPeriodII(48%),butdeclined again in Period III following the disintegration of cooperative farms.ThedecliningtrendsloweddowninPeriodIV.Thelatter might have been caused bya growing demand for lamb from abroadandtheincreasedrateoffinancialsupportfromthestate andEU(Ni_znikowskiet al.,2006).Theavoidanceof adramatic declinemighthavealsobeencausedbylandmanagers’tendency tousethisspeciestoutilizelargeandextensivedry,saline,and sandygrasslandsandabandonedfields,whichcouldnotbeputto moreintensiveuse.

Throughout the 1960s and 1970s, extensively grazed pigs vanished almost completely from the region studied (only surviving in Serbia;cf. Zingstra et al.,2009).Besides this, pigs aregrazedatonelocalityinCroatia(LonskoPoljeattheSavariver— BeinlichandPoschlod,2002;Gugic,9 2009),andonanoccasional basisinthemountainousareasofRomania.Outdoorpiggrazing disappeared in Western Europefrom the19th century onward (Beinlich and Poschlod, 2002), except the dehesa and montado systemsontheIberianPeninsula(Toro-Mujicaetal.,2015).Atthe same time, grazing pigs for conservation purposes is gaining ground(Beinlich andPoschlod, 2002;Putman,2012).Extensive grazingofpigsforconservationpurposesisbeingraisedevermore oftenintheCarpathianBasinrecently.Forinstance,it couldbe

usedtosuppressBolboschoenusmaritimusinmarshes,orthenative breedMangalitzacouldbeputtomasting,becauseofthegrowing marketforhighquality,acornfeededMangalitzapig(PéterTóth, pers.comm.).

ThereareonlylimiteddataavailableintheCarpathianBasinon how Common Agricultural Policy impacted the number of livestock grazed extensively. However,CAPhasbeen a decisive drivingforcebehindkeysectorsinagricultureforthepast10years orso,anditalsohasanimpactonnatureconservation.Thereare countriesinEuropewhereCAPreducedthenumberandratioof sheep, and others, where it has increased (e.g., in Spain, the numberofsheepdroppedtothebenefitofpigs,inGreecesheep becamemorenumerousatthecostofcattle,Toro-Mujicaetal., 2015).

4.2.Changesintheannualgrazingcycleofextensivelygrazedhabitats Cattleprimarilygrazedareaswithlongergrasses.Sheepgrazed drypasturesandstubbles,andpigsweredrivenintomarshesand forestedareas.Theportfolioofgrazedhabitatshasbeenrearranged duringthe60–70yearsstudied:theroleofdrygrasslandsandwet grasslandsincreasedsomewhat,whilethesignificanceofmarshes, stubblefields,vegetationalonglinearelements,secondgrowthon haymeadows,wood-pasturesandforestsdecreased.Theimpor- tance of stubble fields dropped most dramatically. Several underlying causes are suspected: land owners refrain from granting permission to grazing; it has become a dangerous practicebecauseoftheuseofpesticides;andalso,stubblefields are usually ploughed over after harvestbecause the allergenic Ambrosia artemisiifolia must becontrolled ona statutory basis.

Besideindustrialisationoffarmingpractices,thereduceduseof marshesmaybecausedbydrainingofwetlandhabitats.According toour datathe roleof grazing in marshes, wood-pastures and forests increasedtosomedegreein thelast decade,partlyasa result of habitat reconstruction efforts for nature conservation

Fig.6.Theratioofherdedlivestockvs.fencedlivestockinthefourhistoricalperiods(1940–1955,1965–1980,1992–2000and2005–2014)intheCarpathianBasin.

Fig.5. Proportionofnonpasture-grasslandhabitats(e.g.secondgrowthonhaymeadows,marshes,vegetationalonglinearfeatures,forests,andstubblefields)duringthe yearacrossthefourperiods(1940–1955,1965–1980,1992–2000,2005–2014)inthe38studysitesintheCarpathianBasin.

purposes(MérÅetal.,2015;Mosquera-Losadaetal.,2009;Putman, 2012;Vargaetal.,2014).

Forestsplayedanimportantroleinallfourperiodsandallthree species. Grazing in forests was typical mainly in the summer (cattle)orfromspringuntilautumn(sheepandpigs).Grazingin forestsusedtobeanimportantnon-timberforestproductonthe continentuptothe periodbeforeWorldWar II(Johann, 2007;

Harteletal.,2015).Itisstillanimportantsourceofbiomass,even though grazing in forests is an activity which is forbidden or subject to permission throughout Europe (it is completely forbiddeninSlovakiaandHungaryandsubjecttopermissionin Romania,SerbiaandUkraine—(Cirellietal.,2001).

Accordingtoourdata,thefrequencyofgrazingwasreducedin the case of several habitat types over the past 60–70 years.

Abandonmentof grazing may be harmful from a conservation point of viewin many cases,since it mayresult in decreasing species diversity both in pasture grasslands, and non pasture- grasslandhabitats, for instancedue tolight-demanding species beingsuppressedbytheclosingcanopy(Lasantaetal.,2015;Paltto etal.,2011;Vera,2000).

Weshowedthatapproximatelyhalfofthegrazedhabitatsin theseextensivegrazing systemswerenottypicalpasturegrass- lands,butsomeothertypeofhabitat.Anumberofsituationswere documentedasextremes(e.g.insandregions),wheregrazingwas restrictedalmost exclusively tonon pasture-grassland habitats, beingbasedmainlyonstubblefields,fallowland,vegetationalong linearelementsand forestplantations.Grazingonthese ‘other’ typesofhabitatshasbecomemorefrequentmainlyafterJuly(cf.

Etienne, 2005; Kerven et al., 2006), when they functioned as supplementarypasturesduringthelatesummerandearlyautumn drought periods. The particularities of the continental climate (suchas thelow precipitationin theJuly-August period)cause systematic shortages in forage biomass on classical pasture grasslandsbetweenJulyandSeptember(Molnár,2014).Thus,in spiteofadeclineinlivestockthebiomassoftheseotherhabitatsis required.Theirgrazinghasconservationbenefitsaswell,because grazingmay–among othereffects– suppressinvasive species, increase dispersal of propagules or regeneration of secondary habitats,suchasold-fields(BruunandFritzbøger,2002;Haraszthy, 2014;Mosquera-Losadaetal.,2009;vanUytvanckandVerheyen, 2014).

4.3.Decreasingdiversityofgrazedhabitats

Extensive grazing is a culturally and ecologically relatively conservativeland-usesystemadaptedtotheneedsoftheanimals andtheconditionsofthelandscape(MeuretandProvenza,2014).

Thismightbethemainreasonwhyinspiteofdramaticsocialand economicchanges—thenumberofhabitattypesgrazedpermonth per site often did not drop drastically over the four periods investigated.Inthecaseofcattleandthepigsthereductionwas significantfromApriltoOctoberandduringthefirsttwoperiods, respectively,whilenumberofhabitatsandsitesusedbygrazing sheepdidnotchangeasmuch,andthedeclinewasnotsignificant.

In other words, grazing of sheep was found to have themost conservativeattitude,hardlychangingactivity.

Thereductioninthenumbersofhabitattypesgrazedcanbe explained by reduced micromobility in the landscape. That is, livestockstaysonanygivenpastureland longer.Theunderlying causesmayincludetheongoingslowintensificationofmanage- mentpracticesontheonehandandthebusyroadnetworkonthe other,whichmakesmobilisationofanimalsincreasinglydifficult (cf. Luick, 2008; Oteros-Rozas et al., 2013). One of the most importantecologicalconsequencesofreducedmobilityisthata key vector of seed dispersal is lost, which may have serious

economic,ecologicalandconservationimplications(cf.Agnoletti, 2014;BruunandFritzbøger,2002;Poschlodetal.,2002).

Unfortunately,nosimilardataareavailableonthediversityof habitatusewithinextensivegrazingsystemsofotherEuropean countries,althoughtheywouldbeessentialforthemaintenance, set-upandsupportofsustainablegrazingsystems.CAPdoesnot directlyaffectthediversityofgrazedhabitats.However,iffarmers usereedbedsforgrazing,theydonotgetthesamepaymentfor these areasas for grasslands. Arable land is notallowed tobe grazedinareasunderagri-environmentalschemes.Forestgrazing –subjecttonationalregulation–isnotaffectedbytheCAP.

4.4.Decreaseofherding

Herdingdecreasedsubstantiallyoverthepast60–70years,in particularthatofcattleandpigs.Nonetheless,herdingisstillquite commoninthesepost-communistEast-Central Europeancoun- tries,especiallytheherdingofsheep.Theaverageageofherders, however,is increasing,whiletheirnumber declinesasin other Europeanregions(Bernuésetal.,2011).Oneconsequenceisthat there are not enough people with the appropriate skills and competence to herd the livestock (Oteros-Rozas et al., 2013;

Molnáretal.,2016).Infact,theknowledgeofherdingisbeinglost atanevenmorerapidratethanthenumberofherders,because therearemany‘compelled’herderswhoarenotqualifiedforthe job.However,preservationofherdingisanimportantsocialand conservationvaluebecauseconsciousherdingofgrazinganimals contributestothemaintenanceofmanyecosystemservices(e.g.

plantdiversity,agriculturalproductivity,mitigateclimatechange impacts)andtotheimplementationofanumberofconservation measures(cf.Heikkinenetal.,2012;MeuretandProvenza,2014;

Molnár,2014;Rodríguez-Ortegaetal.,2014).CAPshouldmotivate farmerstomaintainherdingbyprovidingthemfinancialreward andrecognitionfortheirwork.

5.Conclusions

Extensive grazing systems based on traditional (landscape-) ecological knowledge are culturally and ecologically relatively conservativeland-usesystems.Ratiosofutilizationamongvarious habitatsaresetbyharmonisingtheinterestsoflivestockandthe biophysical conditions of thelandscape. Asa consequence the impactsof socio-economicandpolitical changesaresometimes lessexplicitthaninthecaseofintensivesystems.

Due tothe prevailing continental climatein the Carpathian Basin,thequantityandqualityofbiomassfitforgrazingissubject tosubstantialintra-andinterannualvariability.Adaptedtothese conditions,extensivegrazingsystemsutilisedandareutilisingstill anumberofnonpasture-grasslandhabitattypesevenatthecostof violatinglawsandregulationspresentlyineffect.Asopposedto this, themajorityofagriculturaland conservationresearchand policiesconcerningthesupportoffarmingonhighnature-value habitatsfocusonclassicalpasturegrasslands(Changetal.,2015;

Haraszthy,2014;Olmedaetal.,2014).EU-level(CAP)andnational agriculturalpolicies should take intoaccount not only pasture grasslandsbutalsothefullspectrumofhabitattypes(e.g.forests, marshes)necessaryfortheeffectiveoperationofextensivegrazing systems.CAPshouldpromotegrazinginforestsandmarshes,on stubblefields,thesecondgrowthonhaymeadowsandvegetation alongside linear landscape features, ensuring the necessary legislativeandfinancialframework.Thesehabitatsareallrequired in order to successfully maintain sustainableextensive grazing systems(cf.MeuretandProvenza,2014).

We argue however that conservation-oriented extensive grazingcannotbebuiltonlyontraditionalherderwisdom;italso needstobeadaptedtothecurrentsocio-ecologicalenvironment.

xxx–xxx

habitatswithstrictlyregulatedbutannuallyadaptedspatialand temporalpatterns(Molnáretal.,2015;Vera,2000).Unfortunately, littleis knowncurrentlyabouttheecologicalaspectsof former extensive grazing systems: for instance, the spatiotemporal patternof grazingintensity, theimpactof extensivegrazing on variousvegetationtypesifgrazedina network,and theherder knowledgeusedinvarioushabitatsandvariousseasons.Allthese would be highlynecessary in order to make extensive grazing systems built on the mosaics and networks of semi-natural habitatstrulysustainable.

Acknowledgements

Firstofallwegratefullyacknowledgealltheinterviewees,who patientlyandgenerouslysharedtheirknowledgeandtimewithus, especiallyIstvánÁgoston,ImreBálint,IstvánBende,JánosDobcsák andhisfamily,JózsefKis,JánosMáté,AntalNémeth,LászlóPapp, PálPethes,LászlóSáfián,LídiaSomogyi,MiklósStencinger,Károly Vándory,AttilaVinczeandLászlóZenics.

WealsothankSándorBéres,DénesHorváth,MirjamHorváth, MelindaJuhász,IldikóJuhászné Türke,KatalinMargóczi,Zoltán Tamás Samuand Gergely Szácsforselflessly helpingwithdata collection,Peter Fehérvári and Roland Hollósi for theirhelp in statisticalanalysisandBélaBorsos,LauraSutcliffe,andSusannahR.

McCandlessforEnglishtranslationandrevision.

Thisresearchwaspartlysupportedbytheproject“Sustainable ConservationonHungarianNatura2000Sites(SH/4/8)”withinthe framework of the Swiss Contribution Program. Furthermore, Daniel Babai was supported by the János Bolyai Research Scholarshipof theHungarian Academy of Sciences and by the Human Capacities Grant Management Office (EMET) (National Talent Program), as a background institution of the Hungarian MinistryofHumanCapacities(EMMI).

Appendix1.

Basic data of the 38 study sites: Name, total area of municipalities, proportional cover of grasslands in the 1970s, and change of the areaof grasslands between 1935 and 2010.

Officialstatistical(KSH)datawereused.Legend:didnotchange

Appendix2.

SeeFig.A2.

Dánszentmiklós 3801 8 ##

Dörgicse 1913 17 #

Drégelypalánk 2221 23 ##

Dunasziget 3589 6 #

ErdÅbénye 4579 14

FelsÅszentmárton 1945 22 ##

FertÅrákos 3969 5 #

Furta 4285 35 "

Fülöpszállás 9127 39 ""

GömörszÅlÅs 855 32 "

Hajdúsámson 6952 7

KisgyÅr 7116 16

Kopa9cevo(HR) 450 50 #

Kunadacs 8997 30

LuncadeJos(RO) 5906 60

MezÅszilas 6497 4 ##

Mórahalom 8314 21

Muraszemenye 1550 14 #

Nádudvar 22591 40 #

NagykÅrös 22786 12 #

Nuşfalau(RO) 5131 8 "

Olaszfalu 2184 16 "

Ordacsehi 2262 30 ##

PlaieşiideJos,P.

Sus,Imper(RO)

30250 35 ""

Pusztakovácsi 4339 11 #

Somotor(SK) 1631 10 "

Stana(RO) 1400 15 ""

Székkutas 12400 21

Tahitótfalu 3921 11 #

Tiszasüly 9182 5 #

Vácrátót 1815 11 ##

Vámosatya 2429 26

Višnjicevo(SRB) 6800 7

Zalaszántó 3773 20 #

Zerind(RO) 5088 23 #

%e:4ki

#^epe(4—Veliki Berehi(UKR)

550 25 "

Fig.A2.Numberoflivestockturnedouttopastureinthe38studysitesinvestigatedduringthefourperiods(1940–1955,1965–1980,1992–2000,and2005–2014).Cattle p=0.000000109,H:35.11;sheepp=0.07851,H:6.744;pigsp=2,07E–17,H:64.3.

Appendix3.

SeeFig.A3.

Fig.A3.Thepercentageofdifferenthabitattypes mentionedinthe38 study sitesintheCarpathian Basin,duringtheannualcycle inthe fourhistoricalperiods studied(1940–1955, 1965–1980,1992–2000and2005–2014).Fortheannualcyclethemostfrequentlymentionedinitialmonthofthegrazingseason(April)wasincludedinthefirstcolumn.

xxx–xxx