Case Study: Grazing at Squam Farm Site...

CaseStudy:GrazingatSquamFarm

SiteDescription

SquamFarmisa34-hectare(85-acre)propertyownedbyNantucketConservationFoundation(NCF),locatedontheglacialmoraineinthenortheasternsectionofNantucket.SquamFarmprovidedanideallocationforalong-termsheepgrazingprojectbecauseithadregulatedvehicleaccess,anaccessiblewatersource,andanNCFstaffmemberlivingon-siteyear-round.

Thecessationofagricultureinthe1800sinitiatedecologicalsuccessiononSquamFarm.Today,thepropertycontainsahighlevelofhabitatheterogeneity,includingcoastalshrubland,managedgrassland,deciduoushardwoodswamp,mixeddeciduousforest,andshrubswamp.PriortoNCFpurchasingtheland,someuplandareasweremowedandgrazedwithheritagebreedlivestockduringthe1980sand1990s.Asaresult,someareasremainopenandcontainadiversemixtureofgraminoids,forbs,vines,andshrubs.Afterpurchasingtheland,NCFmowedannuallytomaintaingrasslandhabitatintheseareas.

Priortothissheepgrazingproject,vegetationcommunitiescontainedanassemblageofnativesandplaingrassland-associatedspeciesincludinglittlebluestem(Schizachyriumscoparium),Pennsylvaniasedge(Carexpensylvanica)andpoverty-grass(Danthoniaspicata),aswellasnon-native,cool-seasongrassessuchasvelvet-grass(Holcuslanatus),sweetvernalgrass(Anthoxanthumodoratum),andsheep-fescue(Festucaovina).Commonforbsincludedgrass-leafflat-toppedgoldenrod(Euthamiagraminifolia),commonSt.Johns-wort(Hypericumperforatum),oxeye-daisy(Leucanthemumvulgare),commonsheepsorrel(Rumexacetosella)anddwarfcinquefoil(Potentillacanadensis).Thesegrasslandswereinterspersedwithbothnativeandnon-nativeshrubandvines,includingscruboak(Quercusilicifolia),smallbayberry(Morellacaroliniensis),blackcherry(Prunusserotina),wingedsumac(Rhuscopallinum),Japanesehoneysuckle(Lonicerajaponica),fox-grape(Vitislabrusca)andpoison-ivy(Toxicodendronradicans).Further,highpopulationdensitiesofturtlesandsnakeswereobservedinuplandareasontheproperty,includingrarespottedturtle(Clemmysguttata).Thus,sheepgrazingwasinitiatedatthissiteasanalternativetomowingduringthegrowingseasonbecauseofpotentiallylessimpactonwildlife.

Figure1.SquamFarmsheepgrazingexperiment.PhotoCredit:NantucketConservationFoundation.

ManagementGoals

1) Maintainopengrasslandstructure;

2) Increasesandplaingrasslandtargetspecies;

3) Reducewoodyshrubandvinecover;

4) Reducenon-nativeinvasivespeciescover;

5) Minimizenegativeimpactstokeywildlifespecies;

6) Testthefeasibilityofimplementingalong-termgrazingprogram.

HistoryofManagement

TargetedgrazingmanagementoccurredatSquamFarmfrom2005to2015onapproximately31hectares(75acres)ofpreviously-moweduplandsontheproperty.Atargetedresearchproject(Schlimme2006)occurredwithina0.69hectareha(1.7acre)areain2005to2008.Atthistime,grazingwastheprimarymanagementtool.

Duringaten-yearperiod,avarietyofsheepbreedsweremanagedyear-roundatSquamFarm.Duringthewinter,sheepwerekeptinpermanentfencedpasturesinthesouthernsectionofthepropertyandfedbothgrainandpurchasedorlocally-grownhay.Selectedeweswerebredinthelatefalltoproducelambsinmid-April.TheinitialsheepflockconsistedofCotswold,Romney,andRomney/Cotswoldcrossbreedsandwassupplementedin2010with15NorthCountryCheviot/ScottishBlackfacecrossbreeds.

Uplandsformerlymaintainedbyannualmowingreceivedrepeatedgrazingeachyear,beginninginearlyspring(assoonasgreen-upoccurred)andcontinuinguntilthefirstkillingfrostoflatefall.Anecdotalobservationsandphotomonitoringconductedintheseareasshowedamarkeddecreaseinwoodyspeciesandacorrespondingincreaseingraminoidsovertimewiththismanagement.Inadditiontothesemanagedgrasslandsites,areascontainingdenseshrublandhabitatwithnon-native,invasivewoodyspeciessuchasJapanesehoneysuckle,autumn-olive(Elaeagnusumbellata),privet(Ligustrumspp.),orientalbittersweet(Celastrusorbiculatus)andAmurpeppervine(Ampelopsisglandulosa)werealsointenselygrazed.Thistreatmentwaseffectiveatremovingthedenseshrubunderstoryandfacilitatingaccessforfollow-upinvasivespeciesmanagement.

GrazingatSquamFarmwasdiscontinuedin2015primarilybecauseoflackoffunding,andbecauseofdifficultiesinmaintaininganimalsyear-round.

Research Aresearchprojecttookplacewithina0.69ha(1.7acre)sectionoftheNorthPasturessectionofSquamFarmfrom2005to2008.Theexperimentexaminedtheeffectsofrepeatedgrazingandmowingonearlysuccessionalvegetationcompositionoverafour-yearperiod(3yearsoftreatmentand1yearofrecovery),andcomparedthattreatmenttounmanagedcontrolareas(Beattieetal.2017,Schlimme2006).

TheNorthPastureresearchareawasdividedintonineresearchblocks(each33mx21m)andincluded3Graze,3Mowand3Controltreatments(theControlprovidedanexampleofcurrentlyunmanagedvegetationasinfluencedbypreviousmanagement).Vegetationcommunitycompositionwassampledwithin45permanentlymarked1m2

researchplots(5plotsperresearchblock;15plotspertreatmenttype)usinga1m2

inclinedpointquadratsamplingframe.Sheepusedfortheprojectincluded28adultCotswoldsheepduringthefirst(2005)seasonandamixed,reducedsizedflockofCotswoldsandRomneysduring2006and2007.

DuringeachGrazetreatment,thesheepremainedwithinthefirstresearchblockuntiltheyhadconsumedalmostallavailableforage,thenweremovedtothesecondandthirdblocksandallowedtograzeforthesameperiodoftime.TheMowtreatmentwasperformedwithin1to2daysfromwhenthesheepwereremovedfromtheGrazetreatmentblocks.Thenumber(2or3)andtiming(month)oftreatmentsvariedbetweeneachtreatmentyearbecauseofvariationsinvegetationresponse.Notreatmentswereconductedduring2008butpost-projectvegetationmonitoringwasdone.

Vegetationcompositionwasclassifiedbyfunctionalgroup(graminoid,forb,woodyspecies),andgroundcover(bareground,litter),andquantifiedusingcoverclasses.Priortotreatments,nosignificantdifferencewasobservedbetweencoveroffunctionalgroupsandgroundcover,allowingcomparisonoffunctionalgroupsbetweentreatmentblocks.

Figure2.ExperimentaldesignoftheresearchpastureatSquamFarm.

Afterthreeconsecutiveyearsoftreatment,shiftsinfunctionalgroupandgroundcoverdominancewereevidentintheinitialyearofrecovery(2008)inbothgrazeandMowtreatmentscomparedwiththecontroltreatmentandwithpre-treatment.Grazeandmowtreatmentssignificantlyreducedcoverofwoodyspecies.TheMowtreatmentshowedsignificantlygreaterreductioninwoodyspeciescomparedwiththeGrazetreatment.GraminoidoccurrencewassignificantlygreaterintheGrazeandMowtreatmentscomparedwiththeControl,althoughonlythemowtreatmentwassignificantlyhigherascomparedwithpre-treatment.

OverallforbspeciesdecreasedinboththeGrazeandMowtreatments,althoughtherewasnosignificancebetweenthetreatmentsorpre-treatmentforboccurrence.

BothGrazeandMowtreatmentssignificantlyincreasedthepresenceofbaregroundinthesampledplots,ascomparedtotheControlandpre-treatmentsampling.IntheGrazetreatment,baregroundwassignificantlyhigherattheendofthestudycomparedwiththemowtreatment.Treatmentshadlittleimpactontheoccurrenceoflitter,althoughtheGrazetreatmentshowedanon-significanttrendofdecreasedlitter.

Individualspeciesofmanagementinterest(includingthreesandplaingrassland-indicativespecies,threeweedy/agriculturalspecies,andtwowoodyspecies)wereanalyzedseparatelyfortrendsinresponsetotreatment.Significantchangesinthreeoftheeightspeciesweredetectedafterthreeconsecutiveyearsofmanagement.Grassleafflat-toppedgoldenrod(Euthamiagraminifolia),asandplaingrassland-indicativespecies,decreasedsignificantlyinboththeMowandGrazetreatmentscomparedwiththecontrolandpre-treatment.Theoccurrenceofsweetvernalgrass,anon-nativepasturegrass—significantlyincreasedintheMow

Figure3.Pre-andpost-treatmentresultsofgroundcoverandcoverofvegetationbyfunctionalgroup.

treatmentcomparedwithboththeGrazeandControltreatments.Velvet-grass(anon-native,invasivegrass)occurrencesignificantlyincreasedinboththemowandgrazetreatmentscomparedtothecontrolandpretreatment.Theotherspeciesexaminedwereox-eyedaisy(Leucanthemumvulgare),Pennsylvaniasedge(Carexpensylvanica),littlebluestem(Schizachyriumscoparium),scrub-oak(Quercusilicifolia),andJapanesehoneysuckle(Lonicerajaponica),andallshowednosignificantchangesbetweentreatments.

Threeconsecutiveyearsofsheepgrazingormowingtreatmentssignificantlydecreasedthewoodyspeciesandincreasedgraminoidspecies.Grazingalonesignificantlyincreasedpatchesofbaregroundwithinthelandscape,providingareasforseedgerminationandspeciesrecruitmentinthegrasslands.

Althoughmowingandgrazinghadsimilareffectsonreducingwoodyspeciesoccurrenceandincreasinggraminoidcover,theecologicalimpactsofeachtreatmentvaried.Mowingreducedallvegetationtoauniformheight,includingnewleafandshootgrowthaswellasolderwoodystems.Incontrast,grazingwasselectiveandsheeptendedtoprefernewergrowthandprunetheyoungerfoliageandstems.Inthisstudy,theGrazetreatmentcreatedapatchymosaiceffect,andtheamountoftimespentineachgrazingtreatmentblockinfluencedtheefficiencyoftreatment.

Ultimately,disturbancebybothgrazingandmowingtreatmentsincreasedweedyspeciesinthetreatmentarea.Mowingincreasingsweetvernalgrass(Anthoxanthumodoratum)significantlymorethangrazing.Managementtreatmentsimplementedinareasthathavenotbeenrecentlymowedorgrazedshouldbemonitoredfortheestablishmentandspreadoftheseweedyspecies.Mowingandgrazingtreatmentssignificantlyreducedtheoccurrenceofgrass-leafflat-toppedgoldenrod,anativeperennialforb.Managementresultinginamosaicoftreatmenttimeandintensitymaybemoreeffectiveatmaintainingthesuiteofnativeforbspecies.

TheoverallconclusionsfromthisstudyandexperiencesatSquamFarmwerethat:(1)sheepgrazingreducedclonalshrubandvinecover,(2)grazingalsointroducedseedsofweedyagriculturalplantspecies,andthatgrazingalonewillnotlikelyresultinhabitatrestorationoverarelativelyshortperiodoftime.Constraintsofcostsandthedifficultyofmanaginganimalsyear-roundledNCFtoendthesheepgrazingprogram.

References

Beattie,K.C.,Karberg,J.M.,Omand,K.A.,&O'Dell,D.I.,2017.Sheepgrazingasagrasslandmanagementtool:LessonslearnedonNantucketIsland,Massachusetts.NortheasternNaturalist24:45-64.

Schlimme,K.S.2006.ReturningsheeptotheislandofNantucket:aconservationexperimentatSquamFarm.Unpublishedthesis.DukeUniversity,Durham,NC.