Dune heath habitats of northwest Ireland. Rachel Kempson

Transcription

1 Dune heath habitats of northwest Ireland Rachel Kempson Word count: 11,450 Thesis submitted for the degree of MSc Conservation, Dept of Geography, UCL (University College London)

2 Abstract( DuneheathhabitatsprimarilyconsistingofDecalcifiedfixedduneswithEmpetrum(nigrumandAtlantic decalcifiedfixeddunes(calluno0ulicetea)arelistedaspriorityhabitatsonthehabitatsdirective.these habitatsareofparticularpriorityduetothelimitedrangeofthecharacteristicvegetationwhichexistsin few locations in northwestern Ireland. Despite listed as a priority habitat it seems that little investigation has been undertaken to establish ecological insight into these habitats. Therefore, this study has investigated three sites along the west coast of Co. Donegal where these two dune heath habitats occur to assess their species assemblages and communities and to investigate and evaluate relationshipsbetweenvegetationcommunitiesandthewiderenvironment.itwasfoundthatvariation exists within the dune heath habitats of western Donegal between sites and within sites. It was establishedthatthereareaconfoundingnumberofenvironmentalinfluencesthatdeterminevariations in species and community composition.this complexity is attempted to be further understood to contribute to the conservation of these priority habitats that are likely to become increasingly threatenedinthefaceofclimatechange. i

3 Contents( 1. Introduction Sandduneformationanddynamics Sanddunevegetationandsuccession Speciesassemblages Edaphiccontext Duneheath Threatstodunehabitats AimsandObjectives 4 2. StudyArea Regionalsetting Climate WestDonegal LoughrosBay Keadew 7 3. Methodology Fieldsampling Analyticalprocedures DataAnalysis GISAnalysis Results Physicalcharacteristics Duneecology Vegetationassemblages Speciesandcommunities Environmentalfactors Widerduneenvironment Discussion Duneheathecology Environmentalcontrolsonduneheathdevelopment Communitycompositionandclassification Conservationofduneheathhabitats Conclusion References Appendix Ordinationoutputtables Specieslistsindicatingthosepresentatallsitesoroccurrenceatoneparticularsite Speciesabbreviations 44 ii

4 List(of(figures( Figure1Dunemorphologyasafunctionofsedimentsupply[Source:Carter,(1990)] 1 Figure2Environmentalgradientsacrossasanddunesystem[Source:Carteret(al.(1990)]. 2 Figure3Climate30yearaveragereport,MalinHead(1961_1990)[Source:IrishMeteorologicalService, (2010)]. 5 Figure 4 Regional setting of Loughros estuaries and Maghera and Magheramore sites [Source: (Burningham,2008)]. 7 Figure5CoastalMonitoringProjecthabitatmappingofsites 7 Figure6Samplepointlocations 8 Figure 7 Photographs showing a) view looking north_east across the Maghera site; b) example of characteristicvegetationfoundontheheadlandatmagheramore;c)duneheathareaatmagheramore, located central within complex; and d) area currently mapped as dune heath at Mullyvea, Magheramore. 13 Figure8ExampleofcharacteristicdunevegetationfoundatKeadew 14 Figure9Contributionofspeciestositecomposition. 15 Figure10Communitiesdefinedbyindicatorspeciesanalysis 16 Figure 11 Number of samples from each site associated with each dune grassland branch defined community CommunitynumberscorrespondwithcommunitiesdefinedinFigure10 17 Figure12Numberofsamplesfromeachsiteassociatedwitheachericaceousbranchdefinedcommunity CommunitynumberscorrespondwithcommunitiesdefinedinFigure10 18 Figure13Variationinenvironmentalparametersbetweensitesstudied 21 Figure 14 RDA ordination plot for Keadew site: Arrows represent environmental variables; Species labelledandpointsrepresentdistributionofsamplesinrelationtospeciesandenvironmentalvariables. 22 Figure 15 RDA ordination plot for Maghera site: Arrows represent environmental variables; Species labelledandpointsrepresentdistributionofsamplesinrelationtospeciesandenvironmentalvariables. 23 Figure16RDAordinationplotforMagheramoresite:Arrowsrepresentenvironmentalvariables;Species labelledandpointsrepresentdistributionofsamplesinrelationtospeciesandenvironmentalvariables. Red=Carrickahalaghheadland;Green=Mullyveaoutcrop;Blue;Duneheathpatch,centralcomplex; Yellow=Restofsamples 23 Figure 17 Variation in environmental parameters between user_defined communities derived from Figure10 25 Figure18%CoverageofA.arenariaacrosssites 26 Figure19%coverageC.vulgarisacrosssites 26 Figure20%coverofE.nigrumacrosssites 27 Figure21%coverageofJ.(communis(ssp.(nanaandS.(nigricansatMagheramore 27 Figure22Communityclassifications Toprow:C3 C7userdefinedcommunitiespersite;Middlerow: NVC communities per site; Bottom row: CORINE units per site. Left column: Keadew site different classifications; Middle column: Maghera site different classification; Right column: Magheramore site differentclassifications 28 iii

5 Acknowledgements( Thanks are due to Tim Roderick, Emer Magee and other staff from the National Parks and Wildlife ServiceIrelandforprovidingsupportandaccommodation.Thanksarealsoduetothemembersofthe geography laboratory for their help during lab work. Further thanks are due to my fellow MSc ConservationstudentsforgeneraldiscussionsandtoDrFredRumsey(botanist)attheNaturalHistory Museumforassistancewiththeidentificationofspecies. Finally,specialthanksareduetoDrHeleneBurninghamforbeingabrilliantsupervisorandinspiringme tostudyirishsanddunes iv

6 1. Introduction( 1.1. Sand(dune(formation(and(dynamics( Coastalsanddunesaredynamichabitatsandarecharacterisedbygeomorphologicaldiversitydetermined by sedimentation processes, climatic setting, both past and present, and ecological factors. The main physicalfactorsthatdeterminetheoccurrenceofsanddunesaretheavailabilityofsandsizedsediment, adequate wind to transport it and accommodation space for the deposition of this sediment within the coastalzone(labuz,2005).thesedimentthatconstitutessanddunesispredominantlyfinegrainedandis highlycalcareousduetothelargeconcentrationsofshellmaterialwhichisoftenpreferentiallytransported bywindasparticlesarelessdensethanthatofquartzsandgrains.therefore,areasexposedtowindmay have high localised concentrations of shell material (Barrett_Mold, 2007). Additionally, sediment is comprised of unaltered parent material consisting of weathered rocks from the surrounding geology (Anwar, 2009). Sand dunes are geomorphologically diverse as a result of their physical setting and a function of sediment supply (Figure 1). Geomorphological development within dune formations is dependentuponwindvelocity,sandtransportandplantdensity,distribution,heightandcover(barrett_ Mold, 2007).The morphology of dunes tends to formulate a progression landward from more mobile forms to fixed dunes that characteristically occur over a chronosequence. Embryo dunes are the initial stageofdunedevelopment.foredunesarisefromembryodunes.foredunesprogressintofixeddunesas more dunes are created on the seaward side of the system causing older dunes to migrate landwards becomingmoresheltered,causingwindvelocitiestodropandpreventingsandaccretion.therefore,the dune ceases to grow and becomes increasingly stable. The sediment of the wider littoral unit is a fundamentalfactordeterminingthepresenceofdunesandtheirmorphology(barrett_mold,2007). Figure(1(Dune(morphology(as(a(function(of(sediment(supply([Source:(Carter,((1990)] Furthermore, dune formations can occur on a variety of types of shoreline, from low, gently sloping surfacestocliffedcoastsandisoverallthefunctionofsedimentgrainsize,characteristicsoftheprofiles andthewindregime(carteret(al.1990).availableaccommodationspaceforcoastalduneformationis dependentuponthewidercoastlinegeomorphologyresultingindifferenttypesofcoastaldunes. Dunesystemsneedtobesituatedinthebroaderenvironmentalcontextastheyareintegralpartsofthe wider coastal system exchanging mass, energy, biota and information through feedback processes with adjoining ecosystems. According to Carter et( al. (1990) dunes play a strongly deterministic part in the ecosystemasmanybiologicalgradients(shelter/exposure,waterabundance/deficit,saline/fresh,nutrient 1

7 rich/poor) are interrelated with physiographic variation. However, these gradients and the processes whichformthemoperateatdifferentscalesbothwithinandbetweendunesystems(figure2). Figure(2(Environmental(gradients(across(a(sand(dune(system([Source:(Carter(etal.((1990)].( 1.2. Sand(dune(vegetation(and(succession( (Species(assemblages( Sanddunesaresemi_terrestrialecosystemsandthereforevegetationisanimportantcomponentofthese habitats. Vegetation plays a pivotal role in the development of coastal dune systems, as the degree of vegetation stabilisation is a significant factor in determining the mobility of the system (Barrett_Mold, 2007).Coastaldunevegetationcanbebroadlycategorisedintosandfixingandsandbuildingvegetation (Carter, 1991). Dune systems are considered classical examples of patterns of zonation and ecological succession.ecologicalsuccessionistheprocesswherevegetationinaparticularlocalitychangeswithtime and one unstable community progressively concedes until a stable climax community is established (Slingsby and Cook, 1986).The natural zonation of plant species that exists is typically described by communityclassificationsystemssuchasthenationalvegetationclassifications(nvc)andtheeucorine units. Speciescompositionanddiversityaredeterminedbykeyfactorssuchassaltspray,mechanicalstress(from moving sand), nutrient availability and water supply (Grunewald and Schubert, 2007; Greaver and Sternberg,2006;Moreno_Casasola,1986).Embryodunesarecolonisedbyvariouspioneerspeciesofgrass, suchasammophila(arenariaandcarex(arenaria;therootsofthesespeciesacttostabilisethesand(king, 1989).Stabilisation of the foredunes is associated with the occurrence of mosses and an increase in diversity of perennial and annual plant species. As interspecies competition becomes more intense as succession proceeds, Ammophila( arenaria becomes increasingly less successful and decreases in abundance(dowdeswell,1984).thephysiochemicalconditionsacrossthedunevaryandthisisreflectedin boththevegetationspeciesandcommunitiespresent.characteristicspeciesoffixeddunesincludegrasses suchasfestuca(rubra,luzula(campestrisandherbssuchasgalium(verum.asspeciesdiversityincreases, inter_species competition becomes the dominant factor controlling species presence or absence (GrunewaldandSchubert,2007).AccordingtoAnwar(2009) plantcommunities are demarcated bythe interactionbetweenpreferenceofplantspeciesandsandysubstrate;highwindvelocities;saltspray;sand accretionandenvironmentalheterogeneity;aconsequenceofbothbioticandabioticforces. 2

8 1.3. Edaphic(context( Soilisanimportantcomponentofecosystemsandcertainvariableseffectthecompositionofvegetation present.coastalsanddunesprovideauniqueenvironmenttoinvestigateandevaluatehowenvironmental controlsaffectmicro_environmentalfactorsessentialforplantgrowth(anwar,2009).edaphicpatternsare commonly very complex due to interaction and alternation between geomorphological and biological processes.coastalduneslocatedinmorehumidclimatesarecharacterisedbyrapidleaching,acidification duetohighpermeability,lowwaterstoragecapacityandlowacidbufferingabilityofsands(sevink,1991). Dunesoilsareinherentlycalcareousinnature;howeverthiscanvaryinaccordancetosedimentinputto thesystem.volumeofblownsandtothesystemdecreaseswithdistancefromthecoast.asfixeddunes areestablished,sandaccumulationisreducedorhalted;thereforecalcareousshellmaterialisnolonger delivered to the dune and CaCO 3 begins to leach out, resulting in a decrease in ph (Grunewald and Schubert,2007).Theprocessofsoildevelopmentcanbeviewedwithinthesoilprofilethatconsistsofan unsorted,unalteredlayerofparentmaterialtodistincthorizonsthroughtheprocessofpodsolisation.the surface horizon is characterised by leaching and litter accumulation, containing carbonates (Ca++) and magnesium(mg++)andoxidesofironandaluminium.asleachingoccurs,thesecationsareprecipitated downwardsintolowersoillayers,resultinginadecreaseinph(anwar,2009). Moisturecontentinsandysoilsisoneofthemostconstrainingfactorsonplantgrowth(Anwar,2009).The macroclimatecaninfluencesoilpedogenesisandthedevelopmentofthesoilprofileisacceleratedinareas with high precipitation because of increased decomposition of organic matter. Furthermore, evenly distributedrainfallthroughouttheyearpromotesgrowthandtheestablishmentofvegetation,increasing the overall organic input to the system. Vegetation has a major influence on the micro_environmental conditions of soil habitats as it stabilises soil surface, adds humus, develops shade, facilitates the developmentofsoilstructureandpreventsmoistureloss(anwar,2009).anincreaseinorganiccontent resultsinafurtherincreaseinsoilmoistureandincreasesnitrogenmineralisation(berendseet(al.1998). Coastal dune soils are generally nutrient poor and lack the macronutrients of nitrogen, phosphorus and potassium. Willis and Yemm (1961) illustrated that sparse growth of vegetation was primarily due to deficient resources of nitrogen and phosphorus in comparison to other environmental stresses. Plants survive in nutrient_poor conditions through symbiotic relationships with microorganisms (mycorrhiza) contained in the rhizosphere, nutrients contained in saltspray, precipitation and the accumulation of organicmatter(anwar,2009) Dune(heath( The definition of dune heath is based on the plant communities present. In coastal dune systems a gradientexistsdeterminedbysediment/soilmobility,withhighermobilityassociatedwithdynamicembryo dunesdominatedbya.( arenaria and low mobility in fixed decalcified dune heath with E.( nigrum and C.( vulgaris. C.( vulgaris is a characteristic heathland species and a strict calcifuge but can be present on calcareoussoils(leakeet(al.1990).factorsaffectingcompositionalvariabilityincludehabitatmaturity,soil stabilityandnutrient_limitedproductivity,withwater_limitedproductivityandpeataccumulationonlyof relevance to older successional stages (Frederiksen et( al. 2006). Stabilised dunes with dune heath communitiestendtobelocatedlandwardtowardstheinteriorofthecoastaldunesystem.frederiksenet al. (2006) believe that the development of dune heath communities implies a mutual action between existingvegetationandsoildevelopmentwithformationofanacidpeatysoil. 3

9 Currently,thereareonlyfoursitesrecordedwithDecalcifiedfixedduneswithE.(nigruminCo.Donegal, IrelandandsevensiteswithAtlanticdecalcifiedfixeddunes(Calluno0Ulicetea)mainlylocatedinCo.Mayo and Co. Donegal (Ryle et( al. 2009). Designated as priority habitats in the EU Habitats Directive, these habitatsareofinternationalimportanceduetothelimitedrangeofthecharacteristicvegetation(provoost et( al. 2002). Therefore, basic ecological insight of these habitats is important and demands special attention. There is still ongoing investigation and thought into deciphering the considerable number of variablesaffectingsuccessionalprogressionnotonlyindifferentecosystemsbutwithinecosystems.the discrepancy is primarily related to basic assumptions and many confounding factors (Anwar, 2009). Therefore, further investigation into the specificities of dune heath communities and the environmental situationthesehabitatsresideinisofconsiderablevalueforknowledgetoconservethesepriorityhabitats Threats(to(dune(habitats( Coastaldunehabitatsarevulnerableandthreatenedecosystems(vanderMeulenet(al.2004).Thearea coverageofcoastaldunesineuropehasdecreasedby40%since1900andbyapproximatelyonethirdsince the1970s(o Briain,2005).Duneheathisararehabitatandisseriouslyendangeredduetodunefixation, invasivealienspecies,climatechangeandtourism(frederiksenet(al.2006) Aims(and(Objectives( The aim of this study is to improve the understanding of the environmental controls and their relative importance on the vegetation communities and distribution of dune heath habitats in west Donegal. Furtheringtheunderstandingofthesehabitatsshouldcontributefundamentallytowardstheconservation ofthesecoastaldunesystems.inordertoassessthesefactorstheobjectivesofthestudyare: 1) Toexaminethehabitatcharacteristicsofthreeknownduneheathsites 2) Toassesstheinfluenceofenvironmentalvariablesonduneheathvegetation 3) Toevaluatetheplantspeciescommunitiesoccurringinthesecoastaldunesites 4) Toevaluatetheconservationofduneheath 4

10 2. Study(Area( 2.1. Regional(setting( Coastaldunesaremoreprevalentandexpansiveonwindwardcoasts,especiallyareaswithanabundant sourceofsuitablytexturedsediment,suchastheatlanticcoastsofnorthwesteurope(carteret(al.1990). The Atlantic western coast of Ireland is highly indented with a complex of coastal estuaries, exhibiting considerablevariationinshorelineorientation.thewestdonegalcoastlinelieswithinthemixed wave_tide dominated part of the coastal energy spectrum. The physiography of which is dominated by regional geological structure and glacial modification, which determines the size and orientation of embayments (Burningham,2005).Itsupportsnumerousdunefieldsderivedlargelyfromglacigenicdepositsinthearea ofthepresentcontinentalshelf(carterandwilson,1993;burninghamandcooper,2004).by4000yrsbp, themajorityofirishduneshadbecomestabilisedsignifyingadecliningrateofsedimentsupply,tightening vegetationcontrollikelytobeaccompaniedbydevelopmentoforganic_richsoils,carbonateleachingand enhancedwaterretentionandoverallmorphodynamicchange(carterandwilson,1993).overthepast century,thedevelopmentofmanynorth_westeuropeandunesystemshavebeencharacterisedbystability (IsermannandCordes,1992;ProvoostandVanLanduyt,2001;Provoostetal.2009;Rhindetal.2001;van dermaareletal.1985).despiteshowingsignsofstability,thesedunesystemsarestilldynamicandona localscalesomesystemsappeartobeundergoingperiodicepisodesofgrowthanderosion Climate( Theclimateofthestudyareaismaritimeandconditionsaredominatedbyfrequentfrontaldepressionsto thenorthwestofirelandcombinedwiththewarmsouth_westerlygulfstream(gaynor,2006).northwest IrelandisrenownedtobeoneofthemostexposedandwindiestcoastlinesinEurope(TroenandPetersen, 1989 in Carter and Wilson, 1993).Figure 3 illustrates regional climatic conditions. There is a relatively narrowrangeinmonthlymeantemperatures,withmoderatelyhighprecipitationandwindspeeds(irish Meteorological Office, 2010). The dominant wind direction is between south and west, predominantly onshorealongthewesterncoast.regionalvariationexistswithinannualprecipitationamountswiththe westexperiencingtwiceasmuchrainfall(1400mm)astheeast(700mm)(gaynor,2006) Mean=total= precipitation=(mm) Mean=temperature=( C) Mean=wind=speed=(knots) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure(3(Climate(30(year(average(report,(Malin(Head((1961W(1990)([Source:(Irish(Meteorological(Service,((2010)].( 5

11 2.3. West(Donegal( ThesitesstudiedaresituatedonwestfacingcoastalestuarinesystemsthatarecharacteristicoftheCo. Donegalcoastline(Burningham,2005).Thesesystemsareinherentlydynamicandaretypicallyboundby duneand/orbedrockmargins(burningham,2002).metamorphicrocksunderliemostoftheseawardparts of the estuary, in particular the Loughros Group and Upper Falcarragh Pelites and Falcarragh limestone whicharecoveredbyaeoliansanddeposits.themajorityoftheinlandareaoftheestuaryisunderlainby igneous granodiorite intrusions (NPWS, 2005). These granodiorite intrusions are unique geological formationsthatonlyoccurwithinthenorthwestofireland,particularlyspecifictothedonegalarea LoughrosBay TheMagheramoresiteisincorporatedwithintheboundariesoftheWestofArdara/MaasRoadSACwhich is listed for six EU Annex I priority habitats and fifteen other Annex I habitats (Murray, 2000). The Magheramore sand dune and machair system is an expansive system located on the northern side of Loughros More which is a drowned estuary. This dune system covers 280ha and is geomorphologically diversecontainingembryonic,shiftingandfixeddune,machair,blowoutsandslackhabitats.asequenceof high fixed dune ridges flank Tramore in the west and Ballinreavy strand in the east, either side of the Carrickahalaghheadland(Murray,2000)(Figure4).TowardsMullyveathereexistsanareaofduneslocated onacentralrockyheadlanddominatedbyjuniperus(communis(ssp.(nana.thiscommunityoccursovera thinsandysubstrateunderlainbybedrock.presently,thishabitathasnotbeendefined,despitecurrently mapped as dune heath (Figure 5). Further investigation is required as this area may need to be reclassified and possibly assigned to an EU AnnexIcategoryassociatedwithJ.( communis formations on calcareousheaths/grasslands.smallpatchesofduneheath_~2haofdecalcifiedfixedduneswithe.(nigrum and~40haatlanticdecalcifiedfixeddunes(calluno0ulicetea)existinareasofexposedrockandtowardsthe Carrickahalaghheadland(Murray,2000).LocalisedpatchesofdecalcifiedduneswithE.nigrumarelocated onthecentralheadland,theseareasofheathoccuronsandsubstrateoverlyingbedrocktotheseaward edge,progressingtofixeddunesfurtherinland.atlanticdecalcifiedfixeddunes(calluno0ulicetea)occuron the Carrickahalagh headland, which is a transitional area including a mosaic from wet heath and mire vegetationtofixedduneandduneslack,includingpatchesofdecalcifiedduneheathwithe.(nigrum.the Magheramoreareaislightlygrazed,principallybycattleandrabbits,howevertheareacouldbenefitfrom moreintensivegrazinginplaces.thefixeddunelocatedaroundthecaravansiteissubjectedtotrampling anderosionpressurebyvisitorscontributingtonaturalerosion(murray,2000). MagheraislocatedonthesouthernsideoftheLoughrosBegestuary.Thissiteisanexpansivedunesystem ofabout100hadisplayingthecompletesuiteofsanddunecommunitiesfromembryonicthroughtofixed dune and dune heath (Figure 5). This site is affronted by one of the widest beaches in Ireland, approximately0.8kmfromthedunefronttothemeanhighwatermark(carterandwilson,1993). 6

12 Figure(4(Regional(setting(of(Loughros(estuaries(and(Maghera(and(Magheramore(sites([Source:((Burningham,(2008)].( Keadew TheareaofKeadewisregardedtobeafineexampleofbothdecalcifiedduneheathwithE.nigrum and Atlanticdecalcifiedfixeddunes(Calluno0Ulicetea).However,theareasurveyedwithinthisstudywasnot mappedasduneheath(figure5). Figure(5(Coastal(Monitoring(Project(habitat(mapping(of(sites( 7

13 3. Methodology( 3.1. Field(sampling( Before fieldwork was undertaken sites were investigated in a Geographic Information System (GIS) and recent aerial photographs studied to provide a general outlook on site characteristics. Subsequently, fieldworkwasundertakenthroughoutfiveweeksfrommay June2010.Arandomstratifiedapproachof samplingwasusedtolocatesamplepointsateachsite:98quadratsand27sedimentsamplesatmaghera; 121quadratsand23sedimentsamplesatMagheramoreand45quadrats,11sedimentsamplesatKeadew (Figure6). Figure(6(Sample(point(locations( A 1m 2 quadrat was used as the vegetation was predominantly composed of dune grassland/heathland species. The species present were recorded and their percentage cover estimated. Identification and nomenclatureofvascularplantswasdictatedbyrose(2005)andfitteret(al.(1984).thosethatcouldnot be identified in the field were both photographed and a sample obtained to be identified later. Some photographsandsampleswereidentifiedatthenaturalhistorymuseum,londonbyanin_housebotanist. Coverofbryophytesandlichenswasrecordedbutnotidentifiedtospecieslevel.Atotalof100species wereidentifiedandspeciesrichnessvariedbetween4and21atmaghera,5and26atmagheramoreand4 and19atkeadew.furthermore,percentagebaregroundorareasvoidofplantswererecorded. Soilsampleswereobtainedfromanumberofquadratstoensurearandomdistributionacrossthesiteand asubsetoftheoriginalsample.sedimentwassampledtoadepthof10cmandsampleswereplacedin sealed plastic bags and then remained refrigerated until analysed to prevent decay. A reading of shear strengthwastakenusingapilconshearvane at every location where a sediment sample was obtained. Throughouttheperiodoffieldworktherewasnosignificantchangewithinweathertoaffectthesediment samples taken. However, the results obtained from vegetation surveys and soils analysis need to be situated in the context of the physical site coverage. The coverage of the Maghera site was more comprehensive and takes into consideration a wider representation of the coastal dune system with a broader variety of sand dune communities. In contrast, sampling at Magheramore was focussed upon areasthathavepresentlybeenindicatedandregardedaslocationsofduneheath.inthetimegiven,this alloweddifferentscalesofcomparisontobeestablishedi.e.inter_sitevariation,withinsitevariationand variationbetweencommunitiesclassified. 8

14 3.2. Analytical(procedures( Soil is an essential part of ecosystems and plant communities can be strongly influenced by some soil factors(álvarez_rogelet(al.2007).therefore,thisanalysiscanprovideinformationontheenvironmental controlsaffectingtheexistinghabitat.soilsampleswereanalysedformoisturecontent,organiccontent, carbonate content, ph, conductivity, nitrate and phosphate. Table 1 illustrates the importance of measuredenvironmentalparametersinthecontextofduneheathenvironments. ( Table(1(Environmental(variables,(units(of(measurement(and(their(significance(to(dune(heath(environments.( Variable( Units(of(Measurement( Significance(to(dune/heath(environments( Soil(moisture(content( grams(%ofsample) Environmental gradients and species composition can be profoundly affected by soil moisture (Álvarez_Rogel et( al.( 2007). Organic(content( grams(%ofsample) Inearlystagesofsuccession,organiccontentofsoilincreases as a result of litter production and dead roots, that may simultaneouslyresultinincreasednitrogenlevels(berendse et(al.(1998) Carbonate(content( grams(%ofsample) Carbonate content can provide an indication of the provenanceofthesediment.carbonatecontentalsoaffects thephofthesoil. ph( Increased ph values strengthen nitrification and exhibits phosphoruslimitation(lammertset(al.(1999). Conductivity( microsiemens(µs) Salinityisafactoroflimitationforspeciesincoastalhabitats. Due to proximity to the coast, salt spray and/or high evaporationrateswillaffectsoilsalinitylevels. Nitrate( milligramsperlitre(mg/l) Duetoreducedamountsoforganicmatter,nitrateisoften limited in dune soils (Lammerts et( al.( 1999). Nitrate is an important environmental determinant in the dynamics of speciescomposition(berendseet(al.1998). Phosphate( milligramsperlitre(mg/l) Phosphate is also deficient in dune soils and nutrient concentrations can be further depleted by leaching consequentlylimitingplantgrowth(willisandyemm,1961). Shear(strength( kilopascal(kpa) Shearstrengthisameasureofthesoiltensionthatcanbe increasedbyrootsbindingthesubstrateanddeterminesthe stabilityofthedune.shearstrengthisalsoinfluencedbysoil moisturecontent. Initially,2gramsofeachsamplewasweighedoutintoacrucibleandplacedintheovenat105 Covernight, thesampleswerethenremovedandthedryweightrecorded.moisturecontentofthesoiliscalculated fromtheweightlossondrying.thedriedsampleswerethenplacedintothefurnaceat550 Cfor2hours andoncecooled,theremainingashweightwasrecorded.theamountoforganicmatterwascalculated fromlossonignition.subsequently,sampleswerereturnedtothefurnaceat925 Cforafurther4hours. Duringthisprocess,carbondioxideislostascarbonatesareconvertedtooxides.Onceremoved,samples werereweighed.carbonatecontentofsamplesisthedifferencebetweentheashweightandtheweight lostat925 C,thevalueofwhichismultipliedby1.36tocompensatethedifferenceinmolecularweightsof CO 2 andco 3. 9

15 Toestablishtheconductivityofsamples100mlofdeionisedwaterwasaddedto10gramsofthesample andlefttosteepfor30minutes.sampleswerethenfilteredandreadingsofconductivityweremeasured electrometricallyusingaconductivitymeter.thefilteredwatersampleswerealsousedtomeasurenitrate (Cadmiumreductionmethod)andphosphatecontent(Ascorbicacidmethod)usingaphotospectrometer. Acceptable ph readings proved problematic. Initially, the same water sample prepared for conductivity wasalsomeasuredelectrometricallywithaphmeter.however,theresultsobtainedwereunrealisticof the reality. The values returned ranged between ph 6.04 and 7.06 which were significantly neutral considering the habitat present. In the circumstances, further investigation was undertaken. ph is a problematicvariabletomeasureinitselfduetothenaturalvarianceofhydrogenionswithinsolutions.the variationinproportionsofsoil:watercanaffectthesoilbufferingcapacity,especiallyofyoungdunesoils wherethebuffercapacityissmall,whereasoldersoilsaremuchmorestronglybuffered(wilson,1960). Therefore,anumberofdifferentratiosofsoil:waterweretested.Finally,aratioof1:2.5(soil:water)was used as the concentration and 0.01 molar of dihydrate calcium chloride (CaCl 2 ) was mixed with the deionisedwatertostabilisethesolution(nieropet(al.2003).thismethodprovidedmorerealisticresults comparable to that found in other studies of similar habitat. Values of between ph 3.5 and 6.9 were discoveredfromsoilsfromadutchcoastaldunesystemincludingduneheath(nieropetal.2003).wilson (1960)alsodiscoveredlowpHvaluesatSouthHaven,Dorset,rangingbetween3.9and4.5forduneheath habitat Data(Analysis( Initially,speciesandenvironmentaldatasetswereexploredseparately.Thespeciesrichnessofeachsite wascalculatedbyaddingtogetherthespeciespresentwithineachsample.simpsonsdiversityindexwas calculatedtoevaluatespeciesrichnessandevenness.ahistogramplotwasusedtoillustratecontribution ofparticularspeciestotheoverallsitecomposition.initialexploratorydataanalysiswasundertakenusing regressiontoanalysetherelationshipbetweenenvironmentalparameters.itwasdiscoveredthatsome variablescouldbedeterminedtohaveasignificantlydependentrelationshipwithanotherenvironmental variable(townend,2002).phwasfoundtobehighlydependentupon%carbonate(r 2 =0.499;P=<0.01). Furthermore,shearstrength(r 2 =0.153;P=<0.01)andorganics(r 2 =0.731;P=<0.01)weresignificantly dependentonmoisture.alsoshearstrengthshowedasignificantrelationshipwithorganics(r 2 =0.217;P= <0.01). Species data was then analysed using WINTWINS (TWINSPAN) indicator species analysis to statistically produceuserdefinedcommunitygroupingsbasedonindicatorspecies.indicatorspeciesanalysisemploys the concept of pseudospecies, whereby the occurrence of a species at different predetermined levels is used (Kent and Coker, 1992). A key concept of this analysis is that it produces a dichotomy whereby samplescanbedividedonthepremisethatonesetofdifferentialspeciescharacteriseonegroupanda secondsetcharacterisetheother,thereforecreatingdistinctivegroupingsorcommunities(kentandcoker, 1992).Theanalysisworksintwowaysformingahierarchybytheprocessofdividingspeciesandsamples. Furthermore, TABLEFIT analysis was undertaken to evaluate the goodness of fit of the species data collected to existing NVC and EU CORINE habitat communities. The highest average of compositional satisfaction(g1)andthemeanconstancyspecies(g2)wasusedtoprovidethe bestcommunityfit with already produced classifications (Hill, 1996).Where the average was equal between two classes the highest G1 value was used to determine compositional satisfaction to that community. These analyses were undertaken to evaluate the current definition and correctness of already classified and mapped communities. 10

16 The variation in environmental variables between sites and communities was analysed using boxplots. These are useful graphical presentations using measures of the median and quartiles to illustrate the dispersion and range from the central value within the data (Kent and Coker, 1992). User defined communities (C7 C14, Figure 10) generated from indicator species analysis were used to analyse the environmentalsubtletiesbetweencommunities. Multivariate statistical analysis was used to assess and compare the importance of environmental parametersonspeciesatdifferentsites.44specieswereselectedintotaltoformulateordinationplotsin conjunctionwithenvironmentalvariables.speciesthatwerelowinabundancewereremovedfromthe analysistoassesstheeffectsandvariationofenvironmentalparametersonthespeciesmoreabundantat all sites. The cover of lichens and bryophytes were removed from this analysis. Detrended Correspondance Analysis (DCA) of the species data established that two sites have an axis 1 length of greaterthan4standarddeviationunitsandthereforecomprisehighbetadiversityvalues,whereaskeadew hasalengthofjustover2standarddeviations(appendix8.1).despite,this,afterproductionofcanonical CorrespondanceAnalysisplots,whichareunimodalformsofanalysissuitedtolonggradientlength,itwas decidedthatgradientlengthsweren tsubstantialenoughtocreateeffectiveordination(kentandcoker, 1992).Therefore, Redundancy Analysis (RDA) was performed on individual site data to understand the variationbetweenspeciesandenvironmentalvariables,inclusiveofsamples.toimprovethecomparability ofdataforanalysissomeenvironmentalvariableswerelogtransformedfornormality(log 10 (x+1) GIS(Analysis( ThedataatallsiteswasexaminedinESRIArcGIS9.2usinggeo_rectifiedaerialphotographsfrom2005and shapefiledatafromthecoastalmonitoringprojectofireland(2009)providedbythenationalparksand WildlifeService(NPWS).Samplesandvegetationcommunities(establishedfromindicatorspeciesanalysis andtablefitanalysis)wereanalysedatthespatialscaleusingagis.spatialpatternsofthedatawere statisticallytestedusingspatialautocorrelation(moransi)totesthowsignificantlyclusteredorrandomthe assignedcharacteristicisinnaturebasedonfeaturelocationsandvalues.valuesofmoransindexcloserto 1.0indicatesignificantclustering(ESRI,2010).Furthermore,thisallowedthephysicalandregionalcontext ofthedatatobeviewed. 11

17 4. (Results( 4.1. Physical(characteristics( Considerablevariationexistswithinthephysicalcharacteristicsofthesitesstudied.Figure7apresentsthe siteatmagherawhichisamosaicoffixeddunehabitatsaffrontedbyembryonicdunes.visibledistinctions exist between the communities present within the site. A prominent dune ridge exists towards the seawardedgeofthesite,distinctlycharacterisedbyareasofsandandadominantcoverofa.(arenaria. Heathlandvegetationappearsdistinctlydarkerincolourandislocatedtowardsthelandwardareaofthe site, situated to the back of the dune system in more sheltered areas. In general, the presence of A.( arenariadecreasesinabundancetowardsthelandwardedgeofthedunesystem. The site at Magheramore has a more varied topography making it comparatively different to that of Maghera.Thesanddunesystemflankseithersideofacentralheadlandandisofrelativelyhigheraltitude (4.21_52.76m),thanthatofMaghera(1.57m 18.15m)orKeadew( m).Thissystemcouldbe classifiedasaclifftopduneformation.figures4canddillustratehowclosethebedrockistothesurfaceas rockoutcropsthethinlayerofsubstrate.thereisavarietyofvegetationexhibitedwithindifferentareasof the site. Figure 7b shows an example of vegetation located on the headland. A myriad of heathland speciesarepresentincludingc.(vulgaris,p.(erecta,e.(angustifolium,s.repens,c.(dissectum andamixof Carex species mainly C.( flacca and C.( panicea. Overall, the vegetation exists in a low sward formation. Figure4cportraysadifferentvariationofduneheathhabitatatMagheramore.Aspreviouslyindicated, there is a greater expanse of bedrock outcropping this area. Furthermore, there are indications of the areasnaturalhistoryasthislocationisdistributedwithglacialerratics,bouldersdepositedduringthelast periodofglacialaction.theareainfigure7cisboundedbymorecharacteristicfixedduneformationsand this patch of dune heathland is delimited by the occurrence of A.( arenaria. This area comprises a predominantly heathland community with an abundance of Carex species, C.( vulgaris and E.( tetralix. Furthermore,distinctivetallertuftformationsofS.(nigricansfeaturewithinthisareaofhabitat.Figure7d portraysadifferentvegetationcompositionincontrasttootherareasatmagheramore.similarityexists betweenalllocationsatmagheramoreasbedrockcanbeobservednearoroutcroppingthewiderdune formation.however,therearedistinctivedifferencesindominantvegetationspecies.j.(communis(ssp.( nana carpets an expansive area of the inland headland formation at Mullyvea. Other species that contribute to the habitat include S.repens, C.( vulgaris, A.( diocia, E.( cinerea and E.( tetralix and other heathland species. As can be seen from Figures 7b, c and d the distribution of A.( arenaria occurs in restricted,localisedoccurrencesthroughouttheareassurveyed. 12

18 a) b) c) d) Figure7Photographsshowinga)viewlookingnorth6eastacrosstheMagherasite;b)exampleofcharacteristicvegetationfoundontheheadlandatMagheramore;c)duneheathareaat Magheramore,locatedcentralwithincomplex;andd)areacurrentlymappedasduneheathatMullyvea,Magheramore. 13

19 TheKeadewsiteisamorecharacteristicsanddunesystem,comprisingofanabundantswardofA.arenaria andgrassessuchasf.rubraandl.campestris(figure8).thissiteislessexpansiveinareaandcouldbe postulatedtobeyoungerinnature.thissiteislocatedonaheadlandandthereisasignificantdropontoa rocky shoreline. There are some outcrops of the bedrock throughout the site. Localised occurrences of heathland communities tend to be distributed in areas where bedrock outcrops the dune formation. SpeciesthatoccurinthismanneratKeadewincludeC.vulgarisandE.nigrum. Figure8ExampleofcharacteristicdunevegetationfoundatKeadew 4.2. Duneecology Vegetationassemblages Speciesandcommunities Atotalof100vascularspecieswereidentified(abbreviationslistedAppendix8.3);41specieswerepresent onallsites;whereas7speciesmadealocalisedoccurrenceatmaghera;23speciesatmagheramoreand5 speciesatkeadew(appendix8.2).figure9illustratesthecontributionofabundantspeciestooverallsite vegetationcomposition.a.arenaria,c.vulgaris,f.rubraands.repensmakeasignificantcontributionto the overall species composition at Maghera, C. vulgaris and F. rubra exhibiting the greatest overall coverage.atmagheramoreitisclearthatc.vulgarisisthedominantspecies,withasignificantcoverageof F.rubra,S.pratensisandS.repens.AtKeadew,grasslandspecies,A.arenariaandF.rubradominatesite vegetation. Species such as L. corniculatus, A. vulneraria, C. arenaria and L. campestris have a greater contribution to vegetation communities at Keadew than at the other sites. Comparison across sites indicates that E. nigrum has the greatest overall contribution to vegetation composition at Keadew, R. spinossissma at Maghera and J. communis ssp. nana at Magheramore. Furthermore, E. tetralix and E. cinerea comprise small amounts to vegetation composition at Maghera and Magheramore but are not present at Keadew. Overall, evaluation of species composition on a general scale, heathland species appeartoconstituteagreatercontributiontotheoverallvegetationcommunitiespresentatmagheraand Magheramore.Incontrast,speciescharacteristicofdunegrasslandpredominatevegetationcomposition atkeadew. 14

20 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Maghera Magheramore Keadew Thymus;polytrichus Succisa;pratensis Salix;repens Rosa;spinosissima Potentilla;erecta Luzula;campestris Lotus;corniculatus Juniperus;communis;ssp.;nana Hypochaeris;radicata Festuca;rubra Erica;tetralix Erica;cinerea Empetrum;nigrum Carex;arenaria Calluna;vulgaris Anthyllis;vulneraria Ammophila;arenaria Figure9Contributionofspeciestositecomposition. Figure 10 shows the divisions and communities defined by indicator species analysis (Wintwins, TWINSPAN).Thefirstdivisioncreatedisacleardistinctionbetweenericaceouscommunities(C2)andthose morecharacteristicofdunegrassland(c1).thedunegrasslandisfurtherclassifiedintoascrubbierdune community(c3)consistingofspeciessuchasp.aquilinum,r.spinossissmaandv.chamaedrys;andtruer dunegrassland(c4)includingspeciessuchasa.vulneraria,p.officinarum,l.corniculatusandt.polytrichus. Afurtherdivisionofthescrubbydunecommunitydefinestheadditionofamoreericaceouscommunity (C8)withindicatorspeciesincludingC.vulgaris,E.angustifolium,PerectaandH.pulchrum.Incontrastto themoreericaceousgrasslandcommunityisacommunity(c7)definedbypresenceofp.aquilinumandg. verum,thatwhendividedfurtherresultsinasubvcommunitysolelydefinedbyincidenceofs.repens(c15). SubVcommunitiescreatedbydivisionofthetruerdunegrasslandresultinacommunitywithmarkerspecies ofs.repens,l.corniculatus,j.communisssp.nanaandp.maritima(c10)andanalternativecommunity withindicatorspeciesl.campestrisandt.repens(c9).speciessuchase.nigrumonlyappearasindicator specieswithinthesubtlydefinedcommunities(c20).incontrast,theericaceousbranchisdividedintoa community characterised by species such as E. tetralix, C. flacca, C. dissectum, C. nigra and P. maritima (C6); and a community classified on the presence of F. rubra and E. angustifolium (C5). The more ericaceouscommunityissubvclassifiedintoacommunitywheree.tetralixisthesoleindicatorspecies(c14) andanalternativecommunitycharacterisedbythepresenceofs.nigricansandp.officinarum(c13).final divisioncommunitiesincludeacommunitycharacterisedbye.tetralix,s.repensandc.nigra(c27)anda communitydefinedbyhigherpresenceofgrassesincludingf.rubraandh.lanatus(c28). 15

21 C1&Dune&grassland N=168 A.arenaria; G.verum; L.campestris;T.polytrichus C3&Scrubby&dune N=61 P.aquilinum; R.spinossissma; V.chamaedrys; E.angustifolium C8&CallunaC Ericaceous N=28 C.vulgaris; E.angustifolium; P.erecta; H.pulchrum C17& Eriophorum N& =&10 E. angustifolium C18&Carex Trifolium N& =&18 C.arenaria; T.repens C7&GaliumC Pteridium N=33 G.verum; P.aquilinum C15&Salix&N&=& 15 S.repens C16&VeronicaC TrifoliumC Ranunculus N&=& 18 V.chamaedrys; T.repens; R.bulbosus Total&Data&N&=&264 C2&Ericaceous N=96 P.erecta; E. tetralix; C.vulgaris C4&Dune&grassland N=107 T.polytrichus; A.vulneraria; P. officinarum;poly.vulgaris; L.corniculatus;C.flacca C5 FestucaCE.& angustifolium N=28 F.rubra; E.angustifolium C6&EricaCCarex N=68 E.tetralix; C.flacca; C.dissectum; C.nigra; P.maritima C9&LuzulaC Trifolium N=73 L.campestris; T.repens C10&SalixC LotusCJuniper N=34 S.repens; L.corniculatus; J.communis; P.maritima C11&TrifoliumC Calluna N&=&25 T.repens; C.vulgaris C12&AgrostisC JuniperusC Ranunculus N&=&3 A.capillaris; J.communis; R.bulbosus C19&N&=&45 G.verum; H.pulchrum; T.polytrichus; Prun.vulgaris; S.pratensis C20&N&=&28 P.lanceolata; E.nigrum; C21&N&=&12 A.arenaria C22&N&=&22 P.maritima; Poly.vulgaris; A.vulneraria; T.polytrichus C23&N&=&4 P.lanceolata C24&N&=&21 EricaCSalixC Rosa E.tetralix; H.pulchrum;S. repens;r. spinossissma C25&N&=&5& P.anserina C13&SchoenusC Pilosella N&=&12 S.nigricans; P.officinarum C26&N&=&7 CallunaCJuniperusC EmpetrumCSalix A.diocia;J. communis;e. nigrum;c.vulgaris; S.repens C14&Erica&N& =&56 E.tetralix C27&Salix& EricaC Carex N& =&44 S.repens; E.tetralix; C.nigra C28&FestucaC HolcusCPlantago N&=&12 F.rubra; H.lanatus; P.lanceolata Figure10Communitiesdefinedbyindicatorspeciesanalysis 16

22 Communitiesbecomesubtlerindefinitionwithfurthersub5division.Thenumberofsamplesfromeachsite whichfallintoeachcommunityareshowninfigure11forthedunegrasslandbranchandfigure12forthe ericaceousbranch.samplesfrommagheraareprincipallyclassifiedunderthedunegrassland(c1)branch withahighnumberofsamplesclassifiedintothescrubbydunecommunity(c3).additionallysamplesfrom Maghera are classified into some of the increasingly ericaceous (C2+) communities, although there occurrenceislessfrequent.however,apredominanceofmagherasamplesisclassifiedaserica&salix&rosa- (C24)- community. In contrast, samples from Magheramore are classified within a wide number of communities.however,somecommunitiesdefinedwithintheericaceousbranch(c2+)arepredominantly samplesfrommagheramore.ahighnumberofsamplesfrommagheramorehavebeenclassifiedastruer dune grassland (C4) and the Erica&Carex&Cirsium- (C6)- community. There are however, particular understated communities that only occur at Magheramore including; Schoenus&Pilosella- (C13);- Festuca& Holcus-(C28);-Calluna&Juniperus&Empetrum&Salix-(C26)communities.Also,incomparisontotheothersites there are a higher proportion of samples classified within the Salix&Lotus&Juniper- (C10)- community. SamplesofKeadewarepredominantlyclassifiedwithincommunitieslocatedonthedunegrasslandbranch (C1+).AhighnumberofsamplesfromthissiteareclassifiedwithintheLuzula&Trifolium-(C9)-andPlantago& Empetrum-(C20)communities C1 43 C C C C C9 23 C C C C C20 10 C17 15 C18 7 C21 16 C Figure11Numberofsamplesfromeachsiteassociatedwitheachdunegrasslandbranchdefinedcommunity Community numberscorrespondwithcommunitiesdefinedinfigure10 17

23 65 C C5 59 C C C12 C13 C C27 12 C28 4 C23 20 C C25 7 C Figure12Numberofsamplesfromeachsiteassociatedwitheachericaceousbranchdefinedcommunity Communitynumbers correspondwithcommunitiesdefinedinfigure10 ClassificationofcommunitieswasinvestigatedfurtherusingTABLEFITwhichenabledagoodnessoffitto evaluatehowwellspeciescompositionandcoverfitwiththecurrentclassificationofthenvc(rodwell, 2000) and the EU CORINE units. Table 2 shows the variation of comparative classification communities found present at each site. There are differences in the communities present varying from dry dune communitiestowetterdunecommunitiesthroughtowetheathandmirevegetation.thecommunitiesat eachsitevaryinaccordancetositeconditions. 18

24 Table2ResultsofTablefitanalysis CORINECode CORINEName NVCCode NVCName SitePresent C BritXerobrcalcgrass CG1f Festuca-ovina- -Carlina-vulgaris-Festuca-rubra&Sci-ver- K C ShortMesobrcalcgrass CG7b Festuca-ovina- -Hier-pil- -Thymus-Cladonia-spp.- MM C Low5AltDryasheath CG13 Dryas-oct- -Carex-flacca- MM C Coastalscilvernheath H7 Calluna-&-Scil-vern-heath- M,MM,K C Coastalscilvernheath H7b Calluna- -Scil-vern-heath-Viola-riviniana- MM C Coastalscilvernheath H7c Calluna- -Scil-vern-heath-Erica-tetralix- M,MM C Coastalscilvernheath H7d Calluna&Scil-vern-heath-Empetrum-nigrum- MM C Coastalscilvernheath H7e Calluna- -Scil-vern-heath-Calluna-vulgaris- M,MM C Ulex5galliheath H8 Calluna- -Ulex-gall-heath- M C Ulex5galliheath H8c Calluna- -Ulex-gall-heath-Sanguis-minor- M C Submontaneheathermoor H10a Calluna- -Erica-cinerea-heath-typical- M C Submontaneheathermoor H10c Calluna- - Erica- cinerea- heath- Festuca- ovina- - M Anthoxanthum-odoratum- C16.24 DuneheathwithCalluna H11 Calluna-&-Carex-arenaria-heath- M C16.24 DuneheathwithCalluna H11c Calluna-&-Carex-arenaria-heath-species&poor- M C51.12 Boghollows M3 Eriophorum-angustifolium-bog-pool- K C31.11 Northernwetheath M16 Erica-tetralix&Sphagnum-compactum- MM C31.11 Northernwetheath M16b Erica-tetralix-&-Sphagnum-compactum-Succisa-pratensis- - M,MM Carex-panicea-- C52.22 Cottongrassbog M20 Eriophorum-vaginatum-blanket/raised- M C AcidMoliniagrassland M25a Molinia- -Potentilla-erecta-mire-Erica-tetralix- M,MM C18.21 Atlanticseacliff MC8b Festuca-rubra&-Armeria-maritime-Crithm-martimum- K C18.2 Vegetatedseacliffs MC9 Festuca-rubra- -Holcus-lanatus- M,MM,K C18.2 Vegetatedseacliffs MC9a Festuca-rubra- -Holcus-lanatus- -Plantago-maritime- MM,K C18.2 Vegetatedseacliffs MC9c Festuca-rubra- -Holcus-lanatus-Achillea-millefolium- MM C18.2 Vegetatedseacliffs MC12 Festuca-rubra- -Hyacinthoides-non&scripta- MM C81 Fertandreseededgrass MG7f Lolium-perenne-recreational-grass-Lolium-perenne- -Poa- MM pratensis- C Grassyfloodswards MG11 Festuca-rubra- -Agrostis-stolinifera- -Potentilla-anserina- MM C16.22 Fixed dune and herbaceous OV27e Epil/Cham-angustifolium-tall-herb-Ammophila-arenaria- M vegetation C AtlanticAmmophiladune SD6 Ammophila-arenaria-mobiledunecommunity M C AtlanticAmmophiladune SD6e Ammophila-arenaria-subc.Festuca-rubra- M,MM,K C AtlanticAmmophiladune SD6f Ammophilaarenaria Poapratensis M,K C AtlanticAmmophiladune SD7 Ammophila- arenaria&festuca- rubra- semi5fixed dune M,MM,K community C AtlanticAmmophiladune SD7d Ammophila-arenaria-subc.Elymus-pycnanthus- M C Calcareousnorthernfixeddune SD8 Festuca-rubra&Galium-verum-fixeddunegrassland MM,K C Calcareousnorthernfixeddune SD8a Subc.Typical(Festuca&Galium)- M,MM,K C Calcareousnorthernfixeddune SD8b Festuca-rubra&Galium-verum-subc.Luzula-campestris- M,MM,K C Calcareousnorthernfixeddune SD9 Ammophila- arenaria&arrhenatherum- elatius- dune M grassland C Thermophdunefringe SD9b Ammophila arenaria Arrhenatherum elatius Ger M sangineum C Northernfixeddune SD10a Carexarenaria Festucarubra M,K C Northernfixeddune SD12 Carex-arenaria&Agrostis-capillaris-dunegrassland-- K C Northernfixeddune SD12a Carex- arenaria&agrostis- capillaris- subc. Anthoxanthum- odoratum- K C16.33 Calcareousfenspp.inslack SD15 Salix- repens&calliergon- cuspidatum- dune5slack MM community C16.26 Creepingwillowdune SD16 Salix-repens&Holcus-lanatus-dune5slackcommunity M,K C16.26 Creepingwillowdune SD16a Salixrepens HolcuslanatusOnonisrepens M,MM,K C16.26 Creepingwillowdune SD16b Salix-repens&Holcus-lanatus-subc.Rubus-caesius- M,MM,K C16.34 Moistslackgrass/rush SD17b Potentillaanserina CarexnigraCarexflacca M,MM C Juncusgerardiiuppersaltmarsh SM16d JuncusgerardiiFestucarubra MM C35.12 Agrostis FestucaGrassland U4 Festucaovina Agrostiscapillaris Galiumsaxatile MM C35.12 Agrostis FestucaGrassland U4b Festuca ovina Agrostis capillaris Galium saxatile Holcuslanatus Trifoliumrepens MM 19

25 Environmentalfactors The variation in environmental parameters between sites is shown in Figure 13. The overall moisture contentofdunesoilsiswide,rangingfrom0.87%to64.72%.however,generallythemoisturecontentis fairlydrywithhalfofthedatahavingvaluesbetween5.63%and13.50%.soilsatmagheramorearehigher in moisture content, ranging between 1.86% and 64.72%, with an average of 16.30%. The average moisturecontentofsoilsatmagherais8.69%and11.39%atkeadew.similarly,theoverallorganiccontent ofsoilsislowwiththemajorityofvaluesoccurringbetween1.41%and2.91%.however,thereissignificant rangewithlowestvalueof0.24%occurringatmagheraandthehighestvalueof23.91%atmagheramore. Theoverallcarbonatecontentofsoilsamplestakenaresignificantlylowinvalueandvariabilitywithan averageof0.599%.however,thereissignificantvariabilitywithincarbonatevaluesofthesoilsatkeadew, with an average of 2.32% which is notably higher than that of Maghera (0.257%) and Magheramore (0.178%).ThelargestrangeinconductivityoccursatMagheramorewithvaluesrangingbetween57.1µS and 155.6µS. Keadew has considerably higher values overall with the lowest value being 79.20µS. Conductivity at Maghera has the lowest range of values, ranging between 42.4µS and 89.8µS with an averageof60.2µs,whichisappreciablylowertothatofmagheramore(88.5µs)andkeadew(99.45µs).ph has a moderately wide range with values as low as 4.07 at Magheramore and values as high as 7.25 at Keadew.ThemajorityofpHvaluesofKeadewareoverallhigherthanthoseoftheothersites.Maghera hasmoreuniformphvalueswithhalfofthedatarangingbetween5.01and5.57.overall,keadewhasthe highest average of 6.59 in comparison to Maghera (5.34) and Magheramore (5.32). Concentrations of nitratesandphosphatesaregenerallyverylowandprincipallyuniform,withhighestvaluesofphosphate occurringatkeadew(0.5mg/l)andhighestvaluesofnitrateatmagheramore(0.4mg/l).overall,thereisa considerablerangeinvaluesofshearstrengththroughoutthedunesiteswithvaluesrangingbetween6kpa atkeadewand70kpaatmagheramore.thegreatestvariationinshearstrengthoccursatmagheramore, which is characterised by the largest range and deviation. In contrast, lowest values and range are associated with Keadew. Generally there is considerable variability within species richness at all sites. SpeciesrichnesswasslightlyloweratKeadewincomparisonwiththeothersites;thelowestvaluebeing4. Highest values of species richness were present at Magheramore; the highest value being 23 species. However,bothMagheraandMagheramorehadequalaveragesof14speciespresentperquadrat. 20

26 Figure13Variationinenvironmentalparametersbetweensitesstudied Figures14,15and16illustratetherelationshipbetweenparticularspeciesandenvironmentalvariablesat individual sites through use of ordination. The RDA for Keadew (Figure 14) illustrates some patterns of associationbetweenspeciesandenvironmentalvariables.aclusteringofspeciesincludingp.-officinarum, L.-campestris,L.-corniculatus,A.-vulneraria,C.-vulgaris,E.-angustifolium,T.-polytrichus,Poly.-vulgarisandP.- maritimaexists.thisclusterofspeciesexhibitspreferencesofbeingmoreericaceousinnature,associated withlowerproportionsofcarbonates,averageshearstrength,nitratesandphosphatesandpreferencefor alowerphandslightlyloweraltitude.c.-arenariadisplayspreferenceforlowershearstrength.species suchash.-lanatus,v.rivinianaandr.-bulbosuspreferconditionswithahigherph. TheRDAforMagheraalsoillustratesclusteringofspecies.Aclusterofspeciesexistsatthetopofaxis2 (Figure 15). The nature of clustering indicates that these species have similar environmental characteristics; species such as G.- verum; T.- polytrichus; S.- pratensis; L.- campestris; C.- fontanum exhibit similar habitat conditions of low carbonate values, higher moisture content and increased levels of 21

27 nitrogen.particularspecieshavestrongassociationwithindividualenvironmentalvariables.e.-cinereais particularly associated with moisture, whereas E.- angustifolium and P.- erecta indicate preference for conditions associated with higher organics. C.- arenaria exhibits a similar environmental preference for decreased values of shear strength at both Maghera and Keadew. C.- vulgaris indicates preference for lowervaluesofsimpsonsreciprocal;thisimpliesthatthespeciestendstobemoreprominentinareasof low species richness. R.- spinossissma specifies increased values of phosphate; although values of phosphate are low throughout the site. S.- repens- indicates preference for conditions with a higher ph, howeverahigherphvalueatmagheraisquiteacidicinthebroaderenvironmentalcontext. TheRDAforMagheramore(Figure16)illustratesagreaterclusteringofspeciesthanMaghera.Toillustrate the variation within habitat across Magheramore, samples from different subsites were highlighted. An overlapofbothspeciespresentandenvironmentalconditionsdoesexist,althoughitisevidentthatthere arethreevariationsofduneheathcommunitypresentatmagheramore.speciessuchasa.-arenaria,c.- arenaria,a.-vulneraria,l.-campestris,t.-repens,g.-verum,l.-catharticum andc.-fontanum areclustered together.thesespeciesexhibitequalpreferenceofaveragephandphosphate.h.-pulchrumexhibitsa preference for higher organics, whereas species such as E.- nigrum, C.- dissectum and S.- repens exhibit preferences for higher proportions of carbonate with lower moisture and organic content. P.- maritimaappearstobestronglyinfluencedbyaltitude.s.-nigricansissituatedfairlycentralintheordinationplot indicatingthatthespeciesshowspreferenceforfairlyaverageenvironmentalconditionsatthissite,the relativeimportanceofindividualenvironmentalparametersisambiguous. Figure14RDAordinationplotforKeadewsite:Arrowsrepresentenvironmentalvariables;Specieslabelledandpointsrepresent distributionofsamplesinrelationtospeciesandenvironmentalvariables. 22

28 Figure15RDAordinationplotforMagherasite:Arrowsrepresentenvironmentalvariables;Specieslabelledandpointsrepresent distributionofsamplesinrelationtospeciesandenvironmentalvariables. Figure16RDAordinationplotforMagheramoresite:Arrowsrepresentenvironmentalvariables;Specieslabelledandpoints representdistributionofsamplesinrelationtospeciesandenvironmentalvariables.red=carrickahalaghheadland;green= Mullyveaoutcrop;Blue;Duneheathpatch,centralcomplex;Yellow=Restofsamples 23

29 Furthermore,therelationshipbetweencommunitiesdefinedandenvironmentalvariableswasevaluated (Figure17).Thethirddivisioncommunitieswereselected(C7 C14fromFigure10)asthisenabledthe finer details of relationships to be analysed. Lower numbered communities associated with the dune grassland branch are clearly determined by different environmental preferences to that of the higher numbered communities belonging to the ericaceous branch. C14 has been defined as an ericaceous community based on the presence of E.- tetralix and environmental conditions associated with this communityincludesoilswithahighermoistureandorganiccontent,verylowlevelsofcarbonates,lowph values ranging between 4.5 and 5.09, high values of shear strength and overall low and fairly uniform concentrationsofnutrients.c13isadifferentvarietyofericaceouscommunity,exhibitingapreferencefor a moisture content ranging between 12.34% and 16.64%, low organic and carbonate content, with an averagephof5.83andapreferenceforslightlyhigherconcentrationofphosphatesthannitrates.c10and C12arebothcommunitiesincludingJ.-communis-ssp.-nanaasapreferentialindicatorspecies.C10hasa widerangeinmoisture,organicandcarbonatecontent,withhighestvaluesofconductivity(155.6µs)and ph(7.25).c12hasmoderatelyhighmoisturecontent(38.95%),ph(6.54)andanotablyhighervalueof phosphate in comparisontoothercommunities (0.5). C8 is defined as the more ericaceous community variant within the dune grassland branch. This community has a fairly high moisture content (max = 13.03%)andlowerpHvalues(min=4.26)incomparisontootherdunegrasslandcommunities. 24

30 Figure17VariationinenvironmentalparametersbetweenuserUdefinedcommunitiesderivedfromFigure Widerduneenvironment ThewidercoastaldunesystemwasanalysedandevaluatedusingaGIS.Distributionoftheabundanceof particularcharacteristicspecieswasanalysedonaspatialscale(figures18521).figure18illustratesthe localoccurrenceofa.arenaria.thehighestvaluesofthisspeciesoccurpredominantlyinareaspresently 25

31 mappedasfixeddune.thereisaparticularlowabundanceofa.-arenariaatareasmappedasduneheath atbothmagheraandmagheramore.thespatialdistributionofa.arenariaishighlysignificantatallsites; Keadew(MoransI=0.03;P=<0.05);Maghera(MoransI=0.07;P=<0.01)andMagheramore(MoransI= 0.08;P=<0.01).C.-vulgarisshowstheoppositepatterntoA.-arenaria(Figure19),wherebyhighervaluesof abundance occurring in areas predominantly mapped as dune heath. However, there is an expansive presenceofthisspeciesonthecarrickahalaghheathlandwhichispresentlymappedas Other(undefined). TheoccurrenceofC.-vulgarisexhibitshighlysignificantclusteringatMaghera(MoransI=0.16;P=<0.01) and Magheramore (Morans I = 0.26; P = <0.01), whereas the presence at Keadew is considered to be neitherclusterednordispersed.figure20showsthelocalisedoccurrenceofe.-nigrum.thisspeciesinnot at all present in the current area mapped as decalcified dune heath with E.- nigrum at Maghera and thereforethemapshouldbemodified.thisspeciesclearlyexhibitsgreaterabundanceatkeadewandis moderately clustered (Morans I = 0.02; P=<0.10). Figure 21 shows the abundance and occurrence of J.- communis-ssp.-nanaands.-nigricansatmagheramore.itisclearthedistributionofj.-communis-ssp.-nana ishighlyclustered(moransi=0.26;p=<0.01)andrestrictedtotheinlandrockoutcropatmullyveawhereit exhibitsmoderatelyhighlevelsofabundance.s.-nigricansexhibitshighestamountofoccurrenceinthe dunepatchlocatedinthecentreofthedunecomplex.however,thespeciesdoeslocateitselfinclustered patches(moransi=0.04;p=<0.01)withinotherduneheathareasacrossmagheramore. Figure18%CoverageofA.arenariaacrosssites Figure19%coverageC.vulgarisacrosssites 26

32 Figure20%coverofE.nigrumacrosssites Figure21%coverageofJ.1communis1ssp.1nanaandS.1nigricansatMagheramore Figure22showsthedifferentclassificationschemesobtainedfromTABLEFITandindicatorspeciesanalysis. Ascanbeseen,particulargeomorphologicalfeaturesarehighlightedbydifferentcommunitiesandthese featuresareemphasisedregardlessoftheclassificationschemeused.forexampleatmaghera,thefixed duneridgeisparticularlyhighlighted,whereasatmagheramore,thecarrickalahaghheadlandandtherocky inlandheadlandatmullyveaareemphasisedthroughcommunityclassification.furthermorecommunities exhibithighlysignificantpatternsofclustering,themajorityofwhichexhibitalessthan1%likelihoodthat theseclusteredpatternsoccurredbychance.overallmagheramoreexhibitsthehighestvaluesofmorans I,varyingbetween0.1and

33 Figure22Communityclassifications Toprow:C3 C7userdefinedcommunitiespersite;Middlerow:NVCcommunitiesper site;bottomrow:corineunitspersite.leftcolumn:keadewsitedifferentclassifications;middlecolumn:magherasite differentclassification;rightcolumn:magheramoresitedifferentclassifications 28