Effects of management intensity, function and vegetation on the biodiversity in urban ponds

(1)

ContentslistsavailableatScienceDirect

Urban

Forestry

&

Urban

Greening

j o u r n al ho me p a g e :w w w . e l s e v i e r . c o m / l o c a t e / u f u g

Effects

of

management

intensity,

function

and

vegetation

on

the

biodiversity

in

urban

ponds

Blicharska

Malgorzata

a,b,∗

_,

_Andersson

_Johan

c

_,

_Bergsten

_Johannes

d

_,

_Bjelke

_Ulf

e

_,

Hilding-Rydevik

Tuija

a

_,

_Johansson

_Frank

c

a_Swedish_Biodiversity_Centre,_Swedish_University_of_Agricultural_Sciences,_Box_7016,_SE-750₀₇_Uppsala,_Sweden b_Department_of_Earth_Sciences,_Uppsala_University,_Villavägen_16,₇₅₂₃₆_Uppsala,_Sweden

c_Department_of_Ecology_and_Genetics,_Uppsala_University,_Norbyvägen_18D,₇₅₂₃₆_Uppsala,_Sweden d_Department_of_Zoology,_Swedish_Museum_of_Natural_History,_Box_50007,_SE-10405_Stockholm,_Sweden, e_The_Swedish_Species_Information_Centre,_Box_7007,_SE-750₀₇_Uppsala,_Sweden,

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received26December2015

Receivedinrevisedform26August2016 Accepted26August2016

Availableonline30August2016 Keywords: Aquaticinsects Biodiversity Pondfunction Pondmanagement Pondvegetation Urban

a

b

s

t

r

a

c

t

Pondsareimportantelementsofgreenareasincitiesthathelpcounteractthenegativeconsequencesof urbanization,byprovidingimportanthabitatsforbiodiversityincitiesandbeingessentialnodesinthe overalllandscape-scalehabitatnetwork.However,thereisrelativelylittleknowledgeabouttheimpacts ofpondmanagementintensity,functionandenvironmentalvariablesonurbanpondbiodiversity.Inthis studyweaddressedthisgapbyinvestigatingwhichfactorswerecorrelatedwiththelevelofbiodiversity inurbanponds,indicatedbyspeciesrichnessofaquaticinsects,inStockholm,Sweden.Ourstudydid notconfirmanydirectlinkbetweentheperceivedintensityofmanagementorfunctionofpondsand overallbiodiversity.However,itseemsthatmanagementcaninfluenceparticulargroupsofspecies indi-rectly,sincewefoundthatTrichopterarichness(Caddisflies)washighestatintermediatemanagement intensity.Wesuggestthatthisiscausedbymanagementofvegetation,astheamountoffloatingand emergentvegetationwassignificantlycorrelatedwithboththeoverallspeciesrichnessandtherichness ofTrichoptera(Caddisflies).Thisrelationshipwasnon-linear,sincepondswithanintermediatecoverage ofvegetationhadthehighestrichness.Interestingly,theamountofvegetationinthepondwas signif-icantlyaffectedbypondfunctionandpondmanagement.Theoverallspeciesrichnessandrichnessof Trichopterawerealsopositivelycorrelatedwithpondsize.Sincewefoundthatthepatternofrelations betweenspeciesrichnessandenvironmentalvariablesdifferedbetweentheinsectgroupswesuggest thatitwillbedifficulttoprovideoveralldesignandmanagementrecommendationsforpondsinurban greenareas.Therefore,itisrecommendedthattoprovidehighaquaticdiversityofspeciesinurbanareas oneshouldaimatpromotinghighdiversityofdifferenttypesofpondswithdifferingmanagementand environmentalfactorsthatshapethem.

Theon-goingglobaltrendofurbanisationhasimportant conse-quencesforbiodiversity,leadingtotheincreasingfragmentation ofnaturalenvironmentsandhabitatloss(McDonaldetal.,2008; Miller and Hobbs 2002). Even though some plant and animal speciesareabletoinhabiturbanareas,mostaresensitivetothe effects of urbanisation (Ghert and Chelsvig, 2003; Riley et al., 2005).However, detrimentalinfluences of urbanenvironments canbealleviatedbythepresence ofgreenareasthat mayoffer

∗ Correspondingauthorat:DepartmentofEarthSciences,UppsalaUniversity, Villavägen16,75236Uppsala,Sweden.

E-mailaddress:malgorzata.blicharska@geo.uu.se(M.Blicharska).

importanthabitatsforbiodiversityincitiesandprovideessential nodesintheoveralllandscape-scalehabitatnetwork(Angoldetal., 2006;McKinney2006).Insomecases,theseareasprovidehabitats forspeciesthataredecreasingelsewhere(e.g.CarrierandBeebee, 2003).

Ponds in cities are often classified as “green-space areas”, because they are usually located within parks or other urban greenzonesandconstituteimportantcomponentsoftheseareas (Harrisonetal.,1995).Growingrecognitionoftheimportanceof pondsandothersmallwaterbodiesformaintainingbiodiversityin cities(Coldingetal.,2009;Fuyukietal.,2014;HassallandAnderson 2015)extendstoconservationprogrammesforendangeredspecies (Vermondenetal.,2009).HassallandAnderson(2015)revealed that urbanstormwater management pondscanprovidea sim-ilar level of biodiversity as urban wetlands in Ottawa, Canada. http://dx.doi.org/10.1016/j.ufug.2016.08.012

(2)

aspondcommunitiesinthesurroundinglandscape.Goertzenand Suhling(2013)showedthaturbanpondsmayhavegreatvaluefor biodiversitybutarethreatenedbyvariousurbandisturbancesand lackofsuitabledesign.

Duringthepastdecade,therehasbeenanincreaseinthe num-berofstudiesexaminingbiodiversityofurbanponds(e.g.Gledhill et al., 2005; Hamer et al., 2012; Noble and Hassal, 2014).The majorityofstudieshavefocusedonhowecologicalandland-cover variablesorwaterchemistryaffectsthebiodiversityofponds(e.g. Biggsetal.,2005;Lamy,2013;Leibold,1999;Oertlietal.,2002). Thereremains,however,alackofknowledgeonhowtheintended primaryfunctionandmanagementof pondsaffectsbiodiversity incities(Biggsetal.,2005; Hassallet al.,2011).Pondsincities canhave differentprimaryfunctions,e.g.,theirpurposecanbe treatmentofurbanrun-off,maintenanceofnaturalvaluefor bio-diversity,ordeliveryofaestheticexperiencestocitizens(Hassal, 2014).Eachfunctionmayinvolvespecifickindsofmanagement activitiesthatinfluencepondvegetation.Inoneofthefewstudies onpondmanagementandbiodiversity,NobleandHassal(2014) identifiedapossibleconflictbetweenhumanandwildlife inter-estinpondmanagementinBradfordDistrict,UK,becausemany pondsweremanagedprimarilyfortheiraestheticfunctionsuch that removal ofvegetation tokeep the pond “neat”for people potentiallydecreasedtheirecologicalvalue.However,thereislittle researchonhowenvironmentalfactorsandfunctionand manage-mentofpondsarecorrelatedandhowthisinfluencesbiodiversity ofurbanponds.

Inmanycountriesthepotentialofurbanpondstodeliverboth biodiversity and ecosystem servicessuchas water purification, floodcontrol,oraestheticexperienceshasbeenrecognized.Many pondsandsmallwaterbodieshavebeenfilledduringthepastfew decades,butthereisnowanincreasingtrendtorestoreandevento createnewwaterbodiesinmanyEuropeancities(e.g.,Åstebøletal., 2004;Segaranetal.,2014;Starkletal.,2013;GledhillandJames, 2012).InSweden,pondshavebeenincreasinglyincludedinurban landscapeplanningsince1990,bothforthepurposeofwater run-offmanagementandtoincreaseaestheticappealofnearbyhousing areas(Personalcommunicationwithrepresentativesof municipal-itiesinthecityofStockholm,2014).Recently,thefocusinSweden hasbeentoincreasethevalueofthesepondsforbiodiversity,and someprojectstorestorepondswiththeintentionofimproving theirecologicalvaluehavebeeninitiated(e.g.Ohlin,2013). How-ever,toourknowledgetherearenostudiesthathavefocusedon therelationshipsbetweentheintendedprimaryfunctionofapond, howintensivelyapondismanaged,andpondbiodiversity.Weuse theexampleofalargeSwedishcity,Stockholm,toinvestigatethis relationship.Suchknowledgeisimportantbecauseitcanhelpwith ponddesignandmanagementtoenhancetheirecologicalvalueand thusprovidebiodiversityrichhabitatsintheurbanmatrix.

Thepurposeofourstudywastoinvestigatewhichfactorsare responsiblefordeterminingbiodiversityinStockholmurbanponds witha focusonmanagement. We selectedenvironmental vari-ablesthatwereshowntoaffectbiodiversityinapilotstudyaswell asvariablesassociatedwithpondfunctionandmanagementthat wereidentifiedduringinterviews.

2. Methods

2.1. Studyareaandselectionofponds

OurstudywasconductedinthecityofStockholm,capitalof Sweden.Thecityhasca.900000residents(StockholmStad,2013)

wasthefirstcitytobeawardedtheEuropeanGreenCapitalby theEUCommissionin2010,becauseofitsholisticvisionto com-bineeconomicgrowthwithsustainabledevelopment(European Commission,2010).

Inthepresentstudy,weconsidered43pondsincentral Stock-holmcovering sevenmunicipalities(Fig.1).Wedefinepondsas naturalorman-madewaterbodieshavinganareabetween1m2

and 2ha and holding water for at least 4 months of the year (PondConservation,2002;Biggsetal.,2005).Pondswereselected frommapsandbyusinginformationfrommunicipalofficials.We focusedondenselypopulatedareasinthecity.Wedivided Stock-holminto1×1km squaresand onlyconsidered squareswhere >75%oftheareawasoccupiedbybuild-uplandusesasdefined inTerrainmap(TerrängkartanTM₎_of_the_Swedish_mapping,

cadas-tralandlandregistrationauthority(Lantmäteriet).Therefore,we excludedpondslocatedingolfcourses,mostoftensituatedoutside thepopulatedareas,eveniftheyhaveshownagreatpotentialfor fosteringbiodiversityinurbanareas(Coldingetal.,2009). 2.2. Selectionofvariablesanddatacollection

Wedefinedbiodiversityasspeciesnumber,andwemeasured itasrichnessofaquaticinsects,inthetaxadragonflies(Odonata), aquaticbeetles(Coleoptera),aquatictruebugs(Hemiptera)and cad-disflies(Trichoptera).Althoughformerstudieshaveshownthatalso othertaxonomicgroupssignificantlycontributetothebiodiversity ofurbanponds(HassallandAnderson,2015;Hilletal.,2015),the taxaselectedforthepurposeofthisstudyprovideanaccurateproxy ofgeneralbiodiversity,becausetheseinvertebratesrepresent dif-ferentfunctionalgroupsandtheirbiodiversityiscorrelatedwith biodiversityofplants,vertebratesandotherinvertebrategroups (Hassalletal.,2011;Oertlietal.,2010).Asspeciesrichnesswas significantlycorrelatedwithShannonindexforspecies(p<0.000, r=0.82) and species abundance (p=0.005, r=0.42) we did not includethesevariablesintheanalysis,becauseaddingvariables decreasesstatisticalpowerofthemodels.

Wecollectedinformationonenvironmentalvariablesthatwere previouslyfoundtobecorrelatedwithaquaticinsectdiversity.Ina pilotstudyconductedusing26pondsinnorthandcentralpartsof Stockholm(Andersson,2014)thatinvestigatedtheeffectsof envi-ronmentalvariablesontheaquaticinsectdiversity,itwasfound thatspeciesrichnessissignificantlycorrelatedwiththeamount ofaquaticvegetation(bothfloatingandemergent)inpondsand theirdistancetothenearestbuilding.Wethereforeincludedthese variables.Similareffectsofvegetationonspeciesrichnessinponds havebeenfoundinotherstudies(Biggsetal.,2005;Goertzenand Suhling, 2013;Hassall etal., 2011).However,we didnot sam-ple submerged vegetation.In addition,we added pond size as apotentialvariablethatcouldaffectbiodiversity,becauseithas beenshowntobeimportantinstudiesofruralponds(Biggsetal., 2005;Oertlietal.,2002),evenifstudiesconsideringpondsizehave yieldedconflictingresults(Hassalletal.,2011).

Vegetationcoveroftheponds(i.e.floatingplants andplants growingintheponds)wasestimatedbyeyeinAugust2013and 2014intenthsanddescribedinpercentagerangingfromatotal coverofnovegetation(0%)atalltofullcover(100%).Vegetationwas recordedintotwoseparatecategories;floatingleafvegetationand emergentvegetation.Pondsizeanddistancetonearestbuilding wasestimatedwiththesoftwareArcGIS9andtheTerrängkartanTM

mapfromLantmäteriet.

Informationontheponds’mainfunction, ageandperceived intensityofmanagementwascollectedthroughinterviews.First, an open-ended interview(Kvale, 1996)lasting about 2.5h was

(3)

M.Blicharskaetal./UrbanForestry&UrbanGreening20(2016)103–112

Fig.1.Locationofthe43pondsinthecentreofStockholmmetropolitanarea.Darkgreyshadingindicatesbuild-upareas,whilelightgreyshadingiswater.Thedarklines indicatemunicipalborders.

conductedwitharepresentativeofthecentrallylocated munici-palityinStockholm(StockholmCity)togetinitialinsightintopond managementinSwedishcitiesingeneralandtogetcontact infor-mationformanagersofparticular ponds.Then, semi-structured interviews (Kvale, 1996) were conducted with representatives ofsevenmunicipalitiesinStockholm:Danderyd,Järfalla, Sollen-tuna, Solna, StockholmCity, Sundbybergand Täby. Fourponds didnotbelongtoanymunicipality,buttoanothertype ofunit (theSwedishRoyalCourt,BergiusBotanicGarden,theStockholm WaterCompanyandaprivatehorseridingstable)andthus repre-sentativesoftheseunitswereinterviewed.Altogether,18people wereinterviewed,eitherface-to-faceorviatelephone.Additional informationwasobtainedinindividualcasesviae-mail.Each inter-viewlastedfrom15minto1.5h,dependingonhowmanyponds theparticularpersoncouldprovideinformationabout.Duringthe interviews, information was gathered onponds’ main function (run-off,amenityandnaturepond),ageclass(young,middle-aged, old)andperceivedmanagementintensity(none,moderate,strong). A descriptionofcategories for each ofthesefactors isgiven in Table1.

Toestimaterichness of aquaticinsects we sampledthemin springandearlysummer(May–June)withabottomscoopnetwith adiameterof20cmandameshsizeof1.5mm.Twentyoneponds weresampledin2013,whiletheremaining22in2014.Sixsamples weretakenineachpondatadepthof20–30cm.Thenetwasswept alongthebottominoppositedirections(lefttoright)eighttimes ona1mstretch,whichconstitutedonesample.Byusingsix sam-pleswecoveredalltypesofrepresentativemicrohabitatsalongthe shoreline,e.g.softbottom,hardbottomwithandwithout vegeta-tion.Thesamplingstrategywasderivedfromtheguidelinesbythe SEPA(2006).

Wesampledtheaquaticlifestagesi.e.larvaeinOdonataand Trichoptera and larvae and adults in Coleoptera and Hemiptera. Allinsectsweredeterminedtoorderatthepondsiteand were preserved in 70% ethanol, stored in labelled plastic tubes and broughtbacktothelaboratoryforspeciesdetermination.Allother specieswererereleasedbacktotheirrespectiveponds.Specimens thatcouldnotbedeterminedtospecieslevelwerestillincluded in thefinal analysisandsettofamilyor genus-levelandhence regardedasseparatetaxa.Inmostcasesthesespecimenswereearly instarlarvae.LarvaeofCoenagrionpuellaandC.pulchellumarenot

Table1

Definitionofthecategoricaldataclassescollectedinthestudy.

Variable

Mainfunction Urbanrun-offpond

Mainaimistoslowdowntherun-off (mayalsomeanprotectionfrom flooding)and/orcleanthewater

Naturepond

Mainaimistoenhancebiodiversity

Amenitypond

Mainaimistoincreaseattractivenessofthe areaforpeople(includingvalueforrecreation)

Ageclass Young

Upto7yearsold Middle-agedFrom8–25yearsold OldOver25yearsold

Perceivedmanagement intensity

None

Nomanagementisconducted

Moderate

Relativelylowintensitymanagement,e.g. removalofsomeportionofvegetation fromthepondeachfewyearsorregular managementaroundthepondsuchas mowingorgrazing,butnotinthepond

Strong

Frequentandintensivemanagement,e.g. removaloflargepartofvegetationeachyearor everytwoyears,orsmallremovalof vegetationeachyearandlargerinterference each4–5years.

(4)

Pondsize(m2₎_[PondSize] ₁₃ _17,219 _1905.2 _753.0

Distancetonearestbuilding(m)[DistanceBuild] 12 747 147.8 86.7

Proportioncoveredbyfloatingvegetation(%)[FloatingVeg] 0 90 19.8 10

Proportioncoveredbyemergentvegetation(%)[EmergentVeg] 0 90 36.3 30

Table3

Summarystatisticsforfunction,ageandperceivedmanagementdata(numberand%ofpondsineachcategory).

Variable Sum

Mainfunction[Function] Run-off2558% Nature 9 21% Amenity921% 43100%

Pondage[Age] Young1637% Middle-aged1944% Old 8 19% 43100%

Perceivedmanagementintensity[Management] None1637% Moderate1637% Strong1126% 43100%

distinguishableandwerethereforeregardedasthesamespeciesin theanalysis.ThesameappliestothreecasesamongtheTrichoptera where larvae could not bedistinguished between two species. Thesewerei)LimnephilusaffinisandL.incisus,ii)Limnephilusluridus andL.ignavusandiii)OligotrichastrictaandO.lapponica.

2.3. Dataanalysis

WeusedmultivariateGeneralLinearModels(GLM)toexplain the relationships between explanatory variables and overall speciesrichness(allinvertebrategroupstogether)andrichnessof eachoftheseparategroups.Intheinitialanalysesweincludedall explanatoryvariablesandthenweusedabackwardsmodel selec-tionforselectingthemodelsthatbestexplainedtherelationships betweentheresponsevariable(overallspeciesrichness or rich-nessofparticularspeciesgroups)andtheexplanatoryvariables. Categoricalexplanatoryvariables(Function,AgeandManagement) wereusedasfixedfactorsin themodels,whilethecontinuous explanatoryvariables(PondSize,DistanceBuild,EmergentVegand FloatingVeg)werecovariates.Wedidnotaccountforthe interac-tionsbetweentheexplanatoryvariablesduetosmallsizeofsome categoriesofthecategoricalvariables.ForvariablesEmergentVeg andFloatingVegwealsoincludedaquadraticterm,inadditionto thelinearfunction,astherelationshipbetweenspeciesrichness andvegetationcoverseemedtobenon-linear(seeResultssection). Wesetathresholdvalueofp=0.15forremovingthevariablesfrom themodelinfindingthebestmodel(seeBendelandAfifi,1977for explanationofthechoiceofthisthreshold).Inshort,usingap-value

of0.15reducestheprobabilityofperformingatype Istatistical error.Visualinspectionofourresponsevariableshowedthatitdid notdeviatefromnormality.

WhenthemultivariateGLMmodelsproducedsignificanteffects weexploredtherelationshipsbetweenindividualvariablesfurther usingeitherlinearorquadraticregressionmodels.

Asvegetationhadthestrongesteffectonspeciesrichnesswe alsoinvestigatediftherewasaneffectofmanagementandpond functionontheamountofvegetation.ThiswasdoneusingGLM modelsthathadpondmanagementorpondfunctionascategorical variablesandvegetationasresponsevariables.

3. Results

3.1. Overallspeciesrichness

Overallspeciesrichnessinthepondsvariedfrom1to22,and themeannumberofspeciesperpondwas9.74.Themostfrequent specieswereColeopteraandOdonataspecies(Onlineresource1). Themostspeciesrichwerepondsdesignedtoenhancebiodiversity, followedbyrun-offandamenityponds(Onlineresource2).Sizeof pondsrangedfrom13toslightlyover17000m2 _and_they_were

locatedwithinthedistanceof12–747mfromthenearest build-ing,whilevegetationcoverrangedfrom0to90%(Table2).Mostof thepondswereeithermiddle-agedoryoungwithnoneor moder-atemanagementandasmuchas58%ofthemwererun-offponds (Table3).

(5)

M.Blicharskaetal./UrbanForestry&UrbanGreening20(2016)103–112 Table4

ResultofGLMmodelfortheoverallspeciesrichness.

Variablesintheselectedmodel F P Models’R2

Function 2.411 0.104 0.270 FloatingVeg 5.213 0.029 FloatingVeg2 _4.237 _0.047 EmergentVeg 3.223 0.081 EmergentVeg2 _2.338 _0.135 PondSize 5.198 0.029

Duringthebackwardselectionthevariablesageandperceived management intensitywere excludedfrom themodelsas they didnotshowanysignificanteffectonthespeciesrichness.Inthe selectedmodelthat bestexplainedoverallpond richness, float-ingvegetationandpondsizehadmostsignificanteffectonoverall richness(Table4).Visualinspectionontheinfluenceoffloating vegetationonspeciesrichnesssuggestedahumpedshaped rela-tionshipwithrichness,withthehighestrichnessatintermediate vegetationcover(Fig.2a).Aregressionmodelusingquadratic float-ingvegetationcoverasa quadratic responsevariableshoweda significanteffectoffloatingvegetationcover(P=0.013;R2₌_0.195; Fig.2a).Therewasatrendforasignificantrelationshipbetween emergentvegetationandspeciesrichness(Table4),andregression modelusingquadraticemergentvegetationcoverasaresponse variableshowedasignificanteffectoffloatingvegetationcoveron speciesrichness(P=0.041;R2₌_0.148;_Fig.₂_b)._In_contrast_when

inspectingtherelationshipbetweenrichnessandpondsizeusing alinearregression,thelineareffectofsizeonrichnessdisappeared (P=0.137;R2₌_0.053;_Fig.₃_)._Pond_function_had_no_significant_effect

onoverallspeciesrichness(Table4).

Table5

ResultofGLMmodelfortheTrichopterarichness.

Management 4.007 0.028 0.547 Age 2.627 0.087 FloatingVeg 10.489 0.003 FloatingVeg2 _7.406 _0.010 EmergentVeg 9.299 0.004 EmergentVeg2 _7.054 _0.012 PondSize 4.709 0.037

Fig.3.Therelationshipbetweenspeciesrichnessandpondsize.Thelineshows thebestfitfromalinearregressionmodel,butitshouldbenotedthatitwas non-significant.Thelinewasaddedtofacilitatevisualinterpretationofthefullmodel, whichshowedasignificantrelationship.

3.2. Trichopterarichness

Trichopterarichnesswasaffectedsignificantlybyfloatingand emergentvegetation(Table5).Visualinspectionoftheindividual influenceofthesevariablesonrichnesssuggestsahumpedshaped relationshipbetweenrichnessandvegetation(Fig.4).Wetherefore ran two quadratic regressions using quadratic floating vegeta-tioncoverandquadraticemergentvegetationcoverasresponse variables respectively. These two models showed a significant relationship between vegetation cover and richness (P<0.001; R2₌_0.374_for_floating_and_P₌_0.008;_R2₌_0.216_for_emergent

veg-etation;Fig.4aandb).TherichnessofTrichopterawasalsoaffected significantlybymanagementandpondsize(Table5).Pondswith a moderatemanagement intensityshowedthehighestrichness (Fig.5).Wheninspectingtherelationshipbetweenrichnessand pondsizeusingalinearregression,thelineareffectofsizeon rich-nessdisappeared(P=0.148;R2₌_0.050;_Fig.₆_).

(6)

Fig.5.Trichopterarichnessinpondswithdifferentmanagementintensity.Thebars representthe95%confidenceinterval.

Table6

ResultofGLMmodelfortheOdonatarichness.

FloatingVeg 3.129 0.085 0.208 FloatingVeg2 _2.785 _0.104 EmergentVeg 3.193 0.082 EmergentVeg2 _2.539 _0.120 DistanceBuild 2.916 0.096 PondSize 2.380 0.132 3.3. Odonatarichness

TherewerenostrongeffectsofthevariablesonOdonatarichness (Table6).Howevertherewasatrendforasignificanteffectof veg-etationanddistancetobuildings.Visualinspectionoftheeffectof thesevariablestakenindividually,suggeststhatbothfloatingand emergentvegetationshowedahumpedshapedrelationshipwith richnessshowingthehighestrichnessatintermediatevegetation cover(P=0.019;R2₌_0.179_and_P₌_0.064;_R2₌_0.128,_{respectively;}

quadraticregressionmodels;Fig.7aandb).Pondswithagreater

Fig.6.TherelationshipbetweenTrichopterarichnessandpondsize.Thelineshows thebestfitfromalinearregressionmodel,butitshouldbenotedthatitwas non-significant.Thelinewasaddedtofacilitatevisualinterpretationofthefullmodel, whichshowedasignificantrelationship.

distancetobuildingstendedtohavehigherrichness,howeverthis relationshipwasnotsignificant(P=0.149;R2₌_0.050_linear

regres-sion;Fig.8).

3.4. ColeopteraandHemipterarichness

NeitherColeopteranorHemipterarichnesswasaffected signifi-cantlybyanyoftheexplanatoryvariablesweincluded,andthefull modelshadaverylowR2_values₍_Tables₇_and₈_).

3.5. Vegetationintheponds

Theamountofemergent,butnotfloatingvegetationwas signif-icantlyaffectedbymanagementintensity(P=0.005andP=0.254, Fig.9aandb),withthelowestamountofvegetationatthestrongest management.Theamountofemergentvegetationwasalso signif-icantlyaffectedbythepondmainfunction(P=0.005),withrun-off pondshavingmostvegetation,followed bynatureandamenity ponds(Fig.9c).Therelationshipbetweenpondfunctionandthe

(7)

M.Blicharskaetal./UrbanForestry&UrbanGreening20(2016)103–112 Table7

ResultofGLMmodelfortheColeopterarichness.

Management 0.056 0.946 0.161 Function 1.943 0.161 Age 0.161 0.852 FloatingVeg 0.025 0.877 FloatingVeg2 _0.007 _0.932 EmergentVeg 0.569 0.457 EmergentVeg2 _0.595 _0.446 DistanceBuild 0.003 0.954 PondSize 0.540 0.568

amountoffloatingvegetationwasclosetosignificant(P=0.068), withamenity pondshavinglargest cover offloating vegetation (Fig.9d).

4. Discussion

Wedidnotfindstrongevidenceoftheimportanceofperceived managementintensityforthebiodiversity,butwefoundthatone ofourstudiedinsectorderswasaffected.However,the interpreta-tionofour“managementintensity”variable,whichwasobtained throughageneralizationofinformationfrominterviews,needsto beinterpretedwithcaution.Themeasureweusedcanbeseenasa humbleindicationofmanagementintensity,butdoesnotaccount forconcretemanagementmeasuresandapproaches.Therefore,the future researchneedstoaddress urbanpond managementin a morecomprehensiveandquantitativeway,measuringtheactual management, includingtheamountofvegetationremoved, the amountof workinghoursspent,etc., over longertime periods, and comparingparticularmanagement strategies. Nevertheless, ourresultssuggestthatmanagementcanaffectrichnessindirectly sincemanagementoftenincludestheremovalofvegetation.The GLMmodelsinourstudyhaverevealedthatoverallrichnessand richness ofTrichopterawasassociated withtheamountof veg-etation(especiallyfloating vegetation),and that pondswithan intermediatecoverageofvegetationhadthehighestrichness.This wasnotshownintheGLMforothertaxonomicgroups,however, forOdonata,thelinearregressionshowedsignificanteffectofthe floatingvegetation.

Fig.8. TherelationshipbetweenOdonatarichnessanddistancetobuildings.The lineshowsthebestfitfromalinearregressionmodel,butitshouldbenotedthatit wasnon-significant.Thelinewasaddedtofacilitatevisualinterpretationofthefull model,whichshowedasignificantrelationship.

Table8

ResultofGLMmodelfortheHemipterarichness.

Management 1.853 0.174 −0.051 Function 1.307 2.86 Age 0.304 0.740 FloatingVeg 0.835 0.368 FloatingVeg2 _0.586 _0.450 EmergentVeg 0.172 0.681 EmergentVeg2 _0.089 _0.768 DistanceBuild 0.022 0.884 PondSize 0.667 0.421

In the case of general species richness, there was more pronounced effect of vegetation on richness when considering individualvariables(vegetationcover)inlinearmodel,compared to GLM. The differences between results of linear regression and GLM can be linked to the differences in complexity of thesetwo typesof models.In theGLMmodelmulticollinearity is present whereas that is not the case in the simple regres-sion model. Nevertheless, the relationshipbetween richness of at least some of the investigated groups and vegetation is in linewithfindingsofotherauthors.Forexample,inthestudyby Goertzen andSuhling (2013), bothterrestrialand aquatic vege-tationwerethemajordeterminantsofdiversityofdragonfliesof urban ponds.The same pattern with respectto vegetationhas alsobeenfoundinruralwaters withregard todragonflies(e.g. Hinden et al., 2005; Remsburg and Turner, 2009)and for sev-eral different groups of macro-invertebrates (Hill et al., 2015). Thegreatestrichnessofmacro-invertebratesinwaterswith inter-mediate level of vegetation was explained by previous studies investigatingpondmicrohabitatsavailablefordifferenttaxonomic groups.Forexample,Bazzantietal.(2010)haveshownthatlower macro-invertebrate richness (i.e.,reduced faunal diversification andabundance)inunvegetatedzoneswasexplainedbythe pres-enceoffinesediments,higherlevelsofnutrients,andrelativelylow oxygencontent.Ontheotherhand,theauthorsemphasizedthe importanceofvegetationinprovidingstabilityofsediments, bet-teroxygenationanddiversityofhabitatsandavailabilityoffood resourcesthatincreasedspeciesrichness.

Consideringbothourfindingsandfindingsoftheprevious stud-ies,weagreewithGoertzenandSuhling(2013),whoconcluded that“themostprominentdeterminantofdiversityaturbanponds isthesameasinnaturalenvironments”.Ourpilotstudy(Andersson, 2014)showedarelationshipbetweenvegetationcoverandspecies richness.However,ourresultsshouldbeconsideredwithcaution, becausesubmergedvegetationwasnotsampled.Lowto intermedi-atedensitiesofsubmergedmacrophytesprobablyincreasespecies richnessbecausemorestructureandthushabitatisavailablefor aquaticinvertebrates(StPierreandKovalenko,2014).Buthigh den-sitiesofsubmergedvegetationmightresultinnegativeeffectssuch asananaerobicconditionsasacauseofthedecompositionofthis organicmaterial.

Wefoundthattherelationshipbetweenrichnessandvegetation cover wasnon-linearand showedthehighestrichness at inter-mediatevegetationcover. Wesuggestthatthis humpedshaped relationshipmaybecausedbyacombination ofhabitat hetero-geneityandproductivity(Huston,2014;Rosenzweig,1995).When productivity increases, more plant species can colonize, which causesanincreaseinspatialheterogeneity,availabilityof micro-habitats,andresourceavailabilityforplanteatinginsectsand,in turn,theirinsectpredators(Bazzantietal.,2010).However,with evenhigherproductivityplantrichnessgoesdown(Huston,2014; Rosenzweig,1995), asthecommunitybecomes dominatedbya few competitiveplant speciesthat provide less habitat hetero-geneityandalessdiversefoodresource.Thisfindinghaspotential

(8)

Fig.9.Relationbetweenmanagementintensityandamountofemergentvegetation(a);managementintensityandfloatingvegetation(b);pondfunctionandemergent vegetation(c);pondfunctionandfloatingvegetation(d).Thebarsrepresentthe95%confidenceinterval.

implicationforponddesignandmanagement,asitsuggeststhat maintenanceofintermediatelevelsofvegetationhaspotentialto improveoverallspeciesdiversity.Whendesigning urbanponds, however,oneshouldconsidertheprimarypurposeofanindividual pond(Hilletal.,2015).Asdifferenttaxonomicgroupshavevarious requirements,“onefitforall”managementstrategywouldbenefit onlysomegroups.Basedonourresults,keepingintermediatelevels ofvegetationwouldsupportTrichopteraandoverallspecies rich-ness,whileOdonata,HemipteraandColeopteramayrequireother managementapproaches.Moreresearchisneededtoinvestigate howparticularmanagementsolutionsinfluenceparticularspecies, especiallywhenoneconsiderstheimportanceofmicrohabitatson biodiversity(Bazzantietal.,2010).Gledhilletal.(2005)suggested thatmanagementofurbanpondscanbeineffectiveinmaintaining highbiodiversity,particularlyincaseofaquaticplantsrequiring morecomplexmanagementstrategies.Supportforthisfromour studyisthelinkbetweenspeciesrichnessinurbanpondsandtheir vegetation,which can indirectlybe influenced by management intensity.NobleandHassal(2014)claimedthatproper manage-mentaimingatincreasingaquaticplantsincreasestheecological qualityofponds.Incontrast,ourstudyshowsthatthe vegetation-richnessrelationshipisnotnecessarilylinear,asrichnessincreases uptoacertainpointofvegetationcoverandthendecreases.Thereis

needforresearchthatwouldinvestigatethisrelationshipinmore detailwithregardtoconcretemanagementmeasurestoexplore itsparticularnuancesthatcouldbeusefulinplanningtheproper management.

Becausetheprimaryfunctionofapondaffectedtheamountof vegetation,withrun-offpondscharacterizedbythehighestamount ofemergentandlowestamountoffloatingvegetation,onecould alsoexpecta significantrelationship betweenprimaryfunction andspeciesrichness(influencedbyvegetation).NobleandHassal (2014)suggestedthatpondsmaintainedforamenityvaluemaybe lessspeciesrich,duetodecreasedamountofvegetation,because thepond managerstend tokeep theponds “neat”for esthetic purposes.Forexample,Nassauer(2004)foundthatclean,mowed wetlandsareperceivedasmoreattractivebypeople.Tomitigate oravoidpotentialconflictbetweenhumanandwildlifeinterests, NobleandHassal(2014)promotedinformingthepublicaboutthe importanceofaquaticvegetationinpondstofacilitatechangesin perceptionsconcerningpondattractiveness.Theyalsosuggested thatmanagersworkingwithpondsshouldbetrainedaboutthe ecologicalrequirementsofdifferentspecies(e.g.abouttheneed forplantingaquaticvegetationthathaspoordispersalrate). How-ever,inourstudywe foundnoevidenceforadirect significant linkbetweenpondfunctionandrichness,onlyanindirecteffect,

(9)

M.Blicharskaetal./UrbanForestry&UrbanGreening20(2016)103–112

whereasmanagementaffectedvegetationandvegetationaffected richness.

Inourmodelspondsizeshowedapositiverelationshipwith overallspeciesrichnessandrichnessofTrichoptera.Wenote, how-ever, that this relationship disappeared when linear regression modelswereusedtoinvestigaterelationshipbetweenrichnessand pondsize.Themainreasonforthesedifferentresultsisthatthe multivariateGLMstakeintoaccountall variables.Nevertheless, theresultfromthefullmodelisinaccordancewithsomeother pondstudies(e.g.Kadoyaetal.,2004;Oertlietal.,2002).We sug-gestthat largerpondshave morehabitats,whichsupportmore species(Rosenzweig,1995).Alargereffectiveareaalsoresultsina higherpopulationsizewhichreducesextinctionrates(Melbourne andHastings,2008).AlthoughGoertzenandSuhling(2013)found anegativerelationshipbetweenpondsizeandrichnessofOdonata inurbanponds,whileOertlietal.,(2010)foundapositive relation-shipforthesamegroup.Furthermore,Oertlietal.(2010)showed thattherelationshipspeciesrichnessandpondsizewasveryweak forothertaxonomicgroups.Inourstudy,therelationshipbetween pondsizeandspeciesrichnessalsoheldonlyforsomegroupsof species,andwefoundhighandlowspeciesrichnessamongthe smallpondswesampled.Takingintoaccountthiswidespanof values,weagreewithOertlietal.(2002)thatconservationefforts shouldbeappliedtopondsofdifferentsizesinordertocoverthe fulltaxonomicbreadthoftheinsectfauna.

Wedidnotfindsignificanteffectsofourselectedvariableson Odonata,ColeopteraandHemiptera,butweacknowledgethatthey mightbeinfluencedbyotherfactorsthatwerenotmeasuredinour study.Suchdifferencesmakeitdifficulttoprovideoverall recom-mendationsforponddesignandmanagementalthoughourresults suggestmaintenance ofintermediate vegetationcover inponds forsometaxonomicgroups.In addition,whilevegetationcover isanimportantvariabletoimproveoverallspeciesrichness,the resultsofthisstudyshowthatdifferentenvironmentalvariables areimportantdriversofrichnessindifferentmacro-invertebrate orders.Becauseofthat,individualmanagementstrategiesneedto bedevelopedtotargeteitherparticulartaxaoroverallrichness. Hassalletal.(2011)recommendedinsuchcasestopromotehigh diversityofdifferenttypeofponds(andthuswithdiffering man-agementandenvironmentalfactorsthatshapethem)topromote highspeciesdiversityatlandscapescale.Ourresultssuggestthat whenplanningpondsingreenareasincities,oneshouldconsider creatingpondsofdifferentsizesandbothpondswith intermedi-atevegetationcoverandpondswithotherlevelsofvegetationto supportspecieswithotherrequirements.

Acknowledgements

WewouldliketothankPer-OlaHedwallforpreparingthemap andthehelpwithstatistics,ClaudiavonBrömssenforthehelpwith statisticsandKevinHolstonforlanguageediting.Wealsothank thetwoanonymousreviewersforcommentsthathavehelpedto greatlyimprovethemanuscript.

References

Åstebøl,S.U.,Hvidved-Jacobsen,T.,Simonsen,Ø.,2004.Sustainablestormwater managementatFornebu—fromanairporttoanindustrialandresidentialarea ofthecityofOslo,Norway.Sci.TotalEnviron.334–335,239–249.

Andersson,E.,etal.,2009.Patternsandscalerelationsamongurbanization measuresinStockholm,Sweden.Landsc.Ecol.24,1331–1339.

Andersson,J.,2014.AquaticInsectCommunityStructureinUrbanPonds:Effects ofEnvironmentalVariables.DegreeProjectinBiology,MasterofScience. BiologyEducationCentreandDepartmentofEcologyandGenetics.Uppsala University,Sweden.

Angold,P.G.,Sadler,J.P.,Hill,M.O.,Pullin,A.,Rushton,S.,Austin,K.,Small,E.,Wood, B.,Wadsworth,R.,Sanderson,R.,Thompson,K.,2006.Biodiversityinurban habitatpatches.Sci.TotalEnviron.360,196–204.

Bazzanti,M.,Coccia,C.,Dowgiallo,M.G.,2010.Microdistributionof macroinvertebratesinatemporarypondofCentralItaly:taxonomicand functionalanalyses.Limnologica40,291–299.

Bendel,R.B.,Afifi,A.A.,1977.Comparisonofstoppingrulesinforward‘Stepwise’ regression.J.Am.Stat.Assoc.72,46–53.

Biggs,J.,Williams,P.,Whitfield,M.,Nicolet,P.,Weatherby,A.,2005.15yearsof pondassessmentinBritain:resultsandlessonslearnedfromtheworkofpond conservation.Aquat.Conserv.Mar.Freshw.Ecosyst.15,671–693.

Carrier,J.-A.,Beebee,T.J.C.,2003.Recent,substantial,andunexplaineddeclinesof thecommontoadBufobufoinlowlandEngland.Biol.Conserv.111,395–399.

Colding,etal.,2009.Golfcoursesandwetlandfauna.Ecol.Appl.19,1481–1491.

EuropeanCommission,2010.http://ec.europa.eu/environment/

europeangreencapital/winning-cities/2010-stockholm(accessedon13.03.15). Fuyuki,A.,Yamaura,Y.,Nakajima,Y.,Ishiyama,N.,Akasaka,T.,Nakamura,F.,2014.

Pondareaanddistancefromcontinuousforestsaffectamphibianegg distributionsinurbangreenspaces:acasestudyinSapporo,Japan.UrbanFor. UrbanGreen.13,397–402.

Ghert,S.D.,Chelsvig,J.E.,2003.Batactivityinanurbanlandscape:patternsatthe landscapeandmicrohabitatscale.Ecol.Appl.13,939–950.

Gledhill,D.G.,James,P.,2012.Socio-economicvariablesasindicatorsofpond conservationinanurbanlandscape.UrbanEcosyst.15,849–861.

Gledhill,D.G.,James,P.,Davies,D.H.,2005.Urbanpond:alandscapeofmultiple meanings.In:PaperPresentedatthe5thInternationalPostgraduateResearch ConferenceinTheBuiltandHumanEnvironment,UniversityofSalford.

Goertzen,D.,Suhling,F.,2013.Promotingdragonflydiversityincities:major determinantsandimplicationsforurbanponddesign.J.InsectConserv.17, 299–399.

Hamer,A.J.,Smith,P.J.,McDonnel,M.J.,2012.Theimportanceofhabitatdesignand aquaticconnectivityinamphibianuseofurbanstormwaterretentionponds. UrbanEcosyst.15,451–471.

Harrison,C.,Burgess,J.,Millward,A.,Dawe,G.,1995.AccessibleNatural GreenspaceinTownsandCities:AReviewofAppropriateSizeandDistance CriteriaEnglishNatureResearchReportNo.153.EnglishNature,Peterborough.

Hassal,C.,2014.Theecologyandbiodiversityofurbanponds.WileyInterdiscip. Rev.:Water1,187–206.

Hassall,C.,Anderson,S.,2015.Stormwaterpondscancontaincomparable biodiversitytounmanagedwetlandsinurbanareas.Hydrobiologia745, 137–149.

Hassall,C.,Hollinshead,J.,Hull,A.,2011.Environmentalcorrelatesofplantand invertebratespeciesrichnessinponds.Biodivers.Conserv.20,3189–3222.

Hill,M.J.,Mathers,K.L.,Wood,P.J.,2015.Theaquaticmacroinvertebrate biodiversityofurbanpondsinamedium-sizedEuropeantown (Loughborough,UK).Hydrobiologia766,225–238.

Hinden,H.,Oertli,B.,Menetrey,N.,Sager,L.,Lachavanne,J.-B.,2005.Alpinepond biodiversity:whataretherelatedenvironmentalvariables?Aquat.Conserv. 15,613–624.

Huston,M.A.,2014.Disturbance,productivity,andspeciesdiversity:empiricismvs logicinecologicaltheory.Ecology95,2382–2396.

Kadoya,T.,Suda,S.,Washitani,I.,2004.Dragonflyspeciesrichnessonman-made ponds:effectsofpondsizeandpondageonnewlyestablishedassemblages. Ecol.Res.19,461–467.

Kvale,S.,1996.InterViews−AnIntroductiontoQualitativeResearchInterviewing ThousandOaks.SAGEPublications,California.

Lamy,T.,2013.Variationinhabitatconnectivitygeneratespositivecorrelations betweenspeciesandgeneticdiversityinametacommunity.Mol.Ecol.22, 4445–4456.

LeViol,I.,Mocq,J.,Julliard,R.,Kerbiriou,C.,2009.Thecontributionofmotorway stormwaterretentionpondstothebiodiversityofaquaticmacroinvertebrates. Biol.Conserv.142,3163–3171.

Leibold,M.A.,1999.Biodiverityandnutrientenrichmentinpondplankton communities.Evol.Ecol.Res.1,73–95.

McDonald,R.I.,Kareiva,P.,Forman,R.T.T.,2008.Theimplicationsofcurrentand futureurbanizationforglobalprotectedareasandbiodiversityconservation. Biol.Conserv.141,1695–1703.

McKinney,M.L.,2006.Effectsofurbanisationonspeciesrichness:areviewof plantsandanimals.UrbanEcosyst.11,161–176.

Melbourne,B.A.,Hastings,A.,2008.Extinctionriskdependsstronglyonfactors contributingtostochasticity.Nature454,100–103.

Miller,J.R.,Hobbs,R.J.,2002.Conservationwherepeopleliveandwork.Conserv. Biol.16,330–337.

Nassauer,J.,2004.Monitoringthesuccessofmetropolitanwetlandrestorations: culturalsustainabilityandecologicalfunction.Wetlands24,756–765.

Noble,A.,Hassal,C.,2014.Poorecologicalqualityofurbanpondsinnorthern England:causesandconsequences.UrbanEcosyst.18,649–662.

Oertli,B.,Joye,D.A.,Castella,E.,Juge,R.,Cambin,D.,Lachavanne,J.B.,2002.Does sizematter?Therelationshipbetweenpondareaandbiodiversity.Biol. Conserv.104,59–70.

Oertli,B.,Céréghino,R.,Biggs,J.,Declerck,S.,Hull,A.P.,Miracle,M.R.,2010.Pond ConservationinEurope.DevelopmentsinHydrobiology,vol.210.Spinger Verlag.

Ohlin,V.,2013.GroddjuriStockholm,projektrapportering.Lokala Naturvårdsatsningen.StockholmStad,Stockholm.

PondConservation,2002.AGuidetoMonitortheEcologicalQualityofPondsand CanalsUsingPSYM.PondConservationTrust.OxfordBrookesUniversity.

(10)

J.L.,Fisher,R.N.,Sauvajot,R.M.,2005.Effectsofurbanizationonthedistribution andabundanceofamphibiansandinvasivespeciesinSouthernCalifornia streams.Conserv.Biol.19,1894–1907.

Rosenzweig,M.L.,1995.SpeciesDiversityinSpaceandTime.CambridgeUniversity Press,Cambridge.

SEPA(SwedishEnvironmentalProtectionAgency),2006.BottenfaunaIsjöars litoralochvattendrag.SS–EN27828.

Segaran,R.R.,Lewis,M.,Ostendorf,B.,2014.Stormwaterqualityimprovement potentialofanurbanisedcatchmentusingwatersensitiveretrofitsintopublic parks.UrbanFor.UrbanGreen.13,315–324.

India.J.WaterSanit.Hyg.Dev.3,500–511.

StockholmStad,2013.http://www.stockholm.se/OmStockholm/Fakta-och-kartor

(accesses02.06.15).

Vermonden,K.,Leuven,R.S.E.W.,vanderVelde,G.,vanKatwijk,M.M.,Roelofs, J.G.M.,JanHendriks,A.,2009.Urbandrainagesystems:anundervaluedhabitat foraquaticmacroinvertebrates.Biol.Conserv.142,1105–1115.