TheRelationshipbetweenland-coverandaquaticmacroinvertebratecommunities
offourwaterbodiesinWesternMontana,USA
AnastassiaRyan
UNDERCWest2016
Abstract
Thehealthandbiodiversityofstreamsisinvaluablefrombothlandmanagement
andwildlifepreservationperspectives.Wesampled5river,creek,andmanmade
ditchsitesinwesternMontanatoassesswaterqualityandrelatemacroinvertebrate
communitydiversitytoadjacentlandcover.Inthisstudywequantitativelyand
qualitativelydescribedthebenthicmacroinvertebratecommunitiesoffourdifferent
loticsystems.DiversityindexeswerecalculatedincludingSimpsons(asthebaseline
responsevariable),richness,andPercentModelAffinity(PMA)alongwithseveral
othersandusedinastepwiseregressiontodeterminethebestmodeltopredict
Simpson’sDiversity.AquaticmacroinvertebraterichnessandPMAwerethebest
predictorsofSimpson’sDiversitybasedonp-ValuesandtheAIC.Futurestudies
shouldinvestigatebothlandcoverandthephysicalandchemicalcharacteristicsin
thewaterbodiesstudied.
Introduction
Streams are the blood vessels of the terrestrial ecosystem. Streams play a role in
nutrient cycling and decomposition of plant material, and provide habitat to a variety of
species of fish and macroinvertebrates. Streams are an important source of biodiversity
within different landscapes, and benthic macroinvertebrates contribute to this biodiversity
(Tonkin 2014). Biodiversity of streams provides an important food source for terrestrial
animals, such as mink, which feed on freshwater mussels and crayfish, and bears, which
prey on salmonid fish. A large portion of fish diet consists of macroinvertebrates, as
different kinds of fish feed on various kinds of aquatic invertebrates. There is an evident
biological cascade that depends on the large diversity provided a healthy stream
ecosystem and good water quality. From headwaters to lower reaches, different
communities of benthic macroinvertebrates exist (e.g., shredders such as stoneflies in
headwaters, or caddisfly collector-gatherers in mid- and lower reaches). These
communities contribute to the processing and downstream transport of dissolved organic
matter (DOM) and fine particulate organic matter (FPOM) (Vannote et al. 1980)
contributing to the cleanliness and health of the stream environment.
TheMoieseValleyisalargeagriculturalcommunityinWesternMontanaon
theFlatheadReservationwherelikeeverywhereinthewesternUnitedStatesis
concernedaboutwaterqualityinthecurrentenvironmentalclimate.The
agricultureintheareaislargelyindustrialcattle,wheat,andhayallofwhichrequire
largesumsofwatertoproduce(Vorosmartyetal.2000).Theseindustriesalsohave
astresseffectonthelocalwaterqualityduetothenutrientloadingproducedfrom
runoffoffertilizersormanureintostreamsorbodiesofwater.Today’sindustrial
agricultureusesavarietyofharmfulmethodsintryingtomaximizetheiryieldsand
profits.Pesticidesareintendedtokillinsectsindiscriminately,herbicidesintended
tokillweedsmayalsodamagethediversityofthenativeflora,andfungicidesare
useddirectlyinwaterways.Alloftheseindustrialagriculturemeasuresareharmful
totheadjoinedaquaticsystems.
Agoodindicatorofecologicalhealthisthediversityofspeciesthatthe
ecosystemhouses.Whentoomanystressorsenteraripariansystemthewater
qualitygoesdownandthesensitivemacroinvertebratescannotsurviveinthe
systemanylonger(Palmeretal.2010).AccordingtotheHilsenhoffBioticIndex
(Barbouretal.)caddisfy’sandstonefly’saresensitivetopollution,right-handed
snailsandcrawlingwaterbeetlesarefacultative,andleft-handedsnailsandworms
arepollutiontolerant.Thereareseveralbiodiversityindexesorequationsmeantto
calculatewaterqualitywitharatingsystemsuchasHilsenhoff’s(HIB),thereisalso
percentEphemeroptera,Plecoptera,Trichoptera(EPT)(DeWaltetal.1999)percent
modelaffinity(PMA)(Novak&Bode1992),percentChironomidae,abundance,and
richness.
Anotherpossibleindicatorofstreamhealthistheadjacentlandcover.Along
thesamelinesasthediversityoftaxawithinthestreamindicatinggoodquality
water,thediversityoffloraabuttingastreamwouldbeasignalofahealthyriparian
ecosystem.Wesampledfiveriver,creek,andmanmadeditchsitesinwestern
Montanatoassesswaterqualityandrelatemacroinvertebratecommunitydiversity
toadjacentland-cover(Weigeletal.1999).Theassessmentconsistedofa
quantitativeandqualitativestudydescribingthebenthicmacroinvertebrate
communitiesinfourdifferentloticsystems.Usingthequantitativedatadifferent
metricswerecalculated(suchastheHBI,EPT,richness,etc.).Alongwithourdataa
land-coverdatabaseprovidedbytheMontanaStateGovernmentwebsitewasused
toassesstherelationshipbetweenwaterqualitywithinthesystemandthe
borderingenvironment.
Methods
SamplingSites:
Fivelocationswithina15-kilometerradiusofTheNationalBisonRange
weresampledwiththekicknetmethod.Thelocationswerechosentorepresent
variouslevelsofwaterqualitywithdistinctadjacentlandcover.Threeofthesites
arelocatedintwoofthetributariestotheFlatheadRiver,theMissionRiverandthe
JockoRiver.Theothertwositesaremanmadeeitheracanaloraditchusedfor
agriculturalirrigation.
TheNinepipelocationisthenorthernmostlocationitisawetland
conservationsiteusedasfloodcontrol.ThereisacanaloriginatingfromThe
NinepipeWildlifeNationalRefugeusedforirrigationintheMoieseagricultural
valley.Thisisamanmadecanalwherethesubstrateissiltandvegetation,andthe
adjacentland-coveriscompletelygrassland.
TheMissionRiversiteselectedisrightunderwheretheMoieseWetland
Restorationsiteemptiesintotheriver.Therestorationconsistsofseveralwetlands
meanttotakewastewaterfromtheMoieseValleyindustrialagriculturalditchesand
filteroutthepollution.Thisparticularlocationisofinterestforthepeopleofthe
FlatheadReservationbecauseoftheinvestmentmadetobuildthewetland
restorationsiteandthequalityoftheirwater.TheMissionsiteislargestriver
sampledwithvarioussizedcoblesubstrateandtheadjacentland-coverishalf
grasslandandhalfriparianwoodlandandshurbland.
TheJockoRiverwassampledattwospots,theUpperJockositeisan
upstream-isolated-bradedlocationandtheLowerJockositeisdownstreaminmore
developedarea.TheregioninWesternMontanawheretheJockorunsismostly
agriculturalandthereareirrigationditchesthatenterandexitbetweenthetwo
selectedsamplingsites.Theadjacentland-covertotheUpperJockoisgrassland,
openeddeveloped,andriparianwoodlandandshrublandwhiletheland-coveron
theLowerJockoisclassifiedasmajorroads,andriparianwoodlandandshrubland.
Jerry’sDitchisasitethatrunsthougharanchingpropertyinDixion,
Montana.Theassumptionforthissiteisthatthewaterqualityisimpaireddueto
thenutrientloadingproducedbythecattleontheproperty.ThesubstrateinJerry’s
DitchissimilartoNinepipe,entirelycomposedofsiltandvegetation.Theland-cover
adjacenttotheditchisclassifiedaslowintensityresidentialandcultivatedcrops.
Alltheland-covertypeclassificationscanbeseenonthemapintheappendix(figure
1).
SamplingMethods:
Eachofthefivesitesweresampledthreetimesondifferentdates:27June
2016,5July2016,and20July2016.Everysamplingeventconsistedofthreekick
netsweepswitha150meshkicknet.TheJockolocationswheretherewasdiverse
substrateadditionalsamplesweretaken.Threekicknetsweepsweretakenina
riffle,run,andpoolontheJockoRiverateithertheupperorlowerlocationtoget
comprehensivesamplesoftheentireriversystem.Eachsamplewasstrainedand
preservedin%95ReagentAlcohol.Eachsamplewasprocessedunder
magnification;everymacroinvertebratewithinthesedimentandvegetative
materialwascollected,identifiedtoorderaccordingtoDr.J.ReeseVoshell,Jr.’s‘A
GuideToCommonFreshwaterInvertebratesofNorthAmerica’andcountedas
presentwiththestreamthesamplewastakenfrom(Table2,Figures2-6).
StatisticalAnalysis:
GPScoordinatestakenateachsamplinglocationwereusedinArcMap(Esri
2013)alongwithland-coverclassificationdataobtainedfromtheMontana
Governmentwebsite.Thebuffertoolwasusedtocreatea15-meterradiusbuffer
aroundthesamplingsites.Thentheland-coverclassificationlayerwasclippedto
justthe15-meterradiustoobtainthepercentageofland-covertypedirectly
adjacenttoeachsamplingsite.Thereweresevenland-covertypesforallfiveofthe
samplinglocations:Type1isNorthernRockyMountainlowermontaneriparian
woolandandshrubland,Type2isRockyMountainlowermontanefoothillandvalley
grassland,Type3iscultivatedcrops,Type4isopenwater,Type5ismajorroads,
Type6lowintensityresidential,andType7isopendeveloped.
Severalmetricswerecalculatedpriortostatisticalanalysisfromtheraw
macroinvertebratecountdata.Allthemetricsarediversityindexesintendedto
relatethequalityofthewaterorthestreamhealthtothediversityandabundance
ofpresentmacroinvertebrates.ThebaselinemetricusedwasSimpsonsindexbut
alsocalculatedforcomparison:HBI,EPT,PMA,percentChironomidae,abundance,
andrichness.
Thetypeland-coverpercentagesandeachofthemetricsareusedtopreform
aforwardstepwiseregressionanalysisinSystatVersion13toobtaintheAkaike
InformationCriterion(AIC)valuesgiveninTable1.UsingSimpsonsdiversityindex
asourstandardforgoodqualitywaterandahealthystreamhabitat,wecompared
boththeland-covertypesandthecalculatedmetricsagainstitasourinitialanalysis.
Asfollowupanalysiswetookland-covertypeandcalculatedmetricsasindividual
categoriesandpreformedtheforwardstepwiseregressiononthemindividually.A
principlecomponentsanalysiswasalsopreformedinSystatVersion13tovisualize
therelationshipsamongallofthevariables(figure8).InRpackage“mvabund”
(Wangetal.2016)wepreformedanordinationandageneralizedlinearmodelon
therawmacroinvertebratecountdatatovisualizethesimilaritiesbygroupings
(figure9).Becausethisisabundancedata,whichistypicallynon-normal“mvabund”
takesthatintoaccountandisabletosortthroughthenon-normalcyaccordingly.
Results
TheordinationraninRshowscleargroupingsbasedonsimilaritiesinthe
rawcountdata.Eachsitehaddistincttaxanomicgroupsandvaryingabundance.The
dominantfamilyintheorderDipteraintheLowerJocko,UpperJocko,andinJerry’s
DitchwastheChironomidae.Jerry’sDitchhadtheleasttotalabundancewithonly
70specimenscollected,theUpperJockohadthemostabundancewith1067
specimenscollected.TheMissionsitewasthesecondmostabundantwith735
collectedspecimenswiththedominantorderbeingTrichoptera.
Thestepwiseregressionusingallofthefactors(land-covertypeandthe
calculatedmetrics)gavePMAandrichnessastheleadingpredictorvariablesin
determiningwaterquality(Simpsonsdiversity)(Table1).
Thenwewantedtoseewhattheleadingpredictorvariablewouldbeifwe
lookedateachcategoryindividually,land-covertypeandcalculatedmetrics.When
takingonlythecalculatedmetricsforstepwiseregressionagainsttheSimpsons
responsevariablebothrichnesswithap-Valueof0.00andaPMAp-Valueof0.048
gaveusthelowestAICvalueof-71.177.Whentakingland-covertypeastheonly
predictorvariables,Type1isthestrongestpredictorvariablewithap-Valueof
0.004andanAICvalueof-49.771.ThePCAshowsaclearrelationshipbetween
Simpsons,richness,abundance,Type1,andType7.
Discussion
ThereisaclearpositiverelationshipbetweenSimpsons,richness,andType1
land-cover.ThismeansthattheNorthernRockyMountainlowermontaneriparian
woodlandandshrublandland-covertypeispositivelycorrelatedwithhigher
diversityofmacroinvertebratepopulation.ThehighertheSimpsonsdiversityvalue
orrichnessofspeciesinastreamsystemwouldindicategoodwaterqualityanda
healthyaquaticecology.
ItisnaturalthatwhenanalyzingallofthefactorsagainsttheSimpsons
responsevariabletherichnessandPMAaretheleadingpredictorvariables.When
analyzingonlythecalculatedmetricsagainstSimpsonsitmakessensethatrichness
andPMAwouldalsobethestrongestpredictorvariables.Thefactthatrichnessgot
ap-Valueof0.00whencomparedwithSimpsonsdiversityindexmeansthatbecause
richnessisafactorincalculatingSimpsonsindexyouwouldexpecttheretobea
directrelationshipbetweenthetwometrics.Sincethetwo,Simpsonsandrichness
aresosimilaronecoulduserichnessinplaceofSimpsonsasaquickestimateofthe
healthofastreamandinturnwaterqualityinthestreamsystem(McGuire2006).
ThePrincipalComponentsAnalysisalsoshowsthattherearesimilaritiesin
Simpsons,richness,PMA,andType1butsuggeststhatType7,abundance,EPT,and
Chiromayalsobegoodindicatorsofdiversityandwaterquality(figure8).While
theordinationconfirmsthatthemacroinvertebratepopulationsareconsistent
withineachsamplinglocationduringthethreeseparatesamplingevents.
UpperJockoprovedtobethemostabundantsiteandthisislikelybecause
thereweredistinctsubstrateareassampledseparatelybutwechosetoincludethe
samplesinourstatisticalmodelstoenhancethestrengthoftheanalysisbyutilizing
allofthedatawecollected.Thustherewasavarietyofmicrohabitatsthatcouldbe
utilizedbyadiverseassemblageofbenthicinvertebrates.TheMissionhadalarge
numberofTrichoptera,whichareasensitivetaxa.TheUpperJockoandMissionhad
thehighestSimpsonsscoresrespectively(Table2).TheUpperJockohadtheleast
alteredland-coverandwasinanareaclassifiedasNorthernRockyMountainlower
montaneriparianwoodlandandshurblandwhichwastheland-covertypefoundto
havethemostinfluenceonthemacroinvertebratediversity.TheMissionsitewasin
anareathathadbeendevelopedasawetlandrestorationsitesothefactthatithas
thesecondbestSimpsonsscoresandahighnumberofthesensitivetaxa
Trichopterasuggeststhatthewaterqualitythereisbetterthanateitheragricultural
site(Jerry’sDitchandNinepipe).
Clearlythereisarelationshipbetweenland-covertypeandthehealthofa
loticsystem(Wangetal.2011).Whenchoosingoursiteswewentoutphysicallyand
lookedfordifferenttypesofloticsystemstocompare.Infuturestudiesitwouldbe
interestingtofindmultiplelocationswithacertainland-coverclassificationanddo
replicatesandthencomparewithotherland-covertypereplicates.Alsowewerenot
ablecollectothermetricssuchasvelocity,temperature,dissolvedoxygen,etc.that
wouldbeimportantintellingthewholestoryofwhatisgoingoninariversystem.
Acknowledgements
IwouldliketothankmyMentorKathrineBarrettforteachingmeeverything
Iknowaboutmacroinvertebratesandforherinfectiousenthusiasm.Iwouldliketo
thankDr.GaryBelovskyandJenifferBelovskyforgivingmethiswonderful
opportunityandexperience.DavidFlagelformakingallthearrangementstogetthis
projectdone.TotheUniversityofNotreDameforhavingsuchagreatprogramfor
studentstoliveandlearninthefield.
References
Barbour,M.T.,Graves,C.G.,Plafkin,J.L.,Wisseman,R.W.andBradley,B.P.1992.
EvaluationofEPA'srapidbioassessmentbenthicmetrics:Metricredundancyand
variabilityamongreferencestreamsites.EnvironmentalToxicologyandChemistry
11:437–449.
DeWalt,R.E.,Webb,D.W.Webb,Harris,M.A.1999.SummerEphemeroptera,
Plecoptera,andTrichoptera(EPT)speciesrichnessandcommunitystructureinthe
lowerIllinoisRiverbasinofIllinois.GreatLakesEntomologist3:115-132.
McGuireD.L.2007.ClarkForkRiverBiomonitoringMacroinvertebrateCommunity
Assessments.UnitedStatesEnvironmentalProtectionAgencyRegion8.
Novak,M.A.,Bode,R.W.1992.Percentmodelaffinity:anewmeasureof
macroinvertebratecommunitycomposition.JournaloftheNorthAmerican
BenthologicalSociety11:80-85.
Palmer,M.A.,Menninger,H.L.,Bernhardt,E.2010.RiverRestoration,habitat
heterogeneityandbiodiversity:afailureoftheoryorpractice.FreshwaterBiology
55:205-222.
Tonkin,J.D.2014.Driversofmacroinvertebratecommunitystructureinunmodified
streams.PeerJ2:e465;DOI10.7717/peerj.465.
Vannote,R.L.,G.W.Minshall,K.W.Cummins,J.R.Sedell,ANDC.E.Gushing.1980.
TheRiverContinuumConcept.Can.J.Fish.Aquat.Sci.37:130-137.
Vorosmarty,C.J.,Green,P.,Salisbury,S.,Lammers,R.B.2000.GlobalWater
Resouces:VulnerabilityfromClimateChangeandPopulationGrowth.Science
5477:284-288.
Voshell,J.R.2002.AGuideToCommonFreshwaterInvertebratesofNorthAmerica.
DepartmentofEntomologyCollegeofAgricultureandLiveSciencesVirginiaTech.
Wang,L.,Lyons,J.,Kanehl,P.,Gatti,R.2011.InfluencesofWatershedLandUseon
HabitatQualityandBioticIntegrityinWisconsinStreams.Fisheries22:2-12.
Wang,Y.,Naumann,U.,Wright,S.,Eddelbuettel,D.,Warton,D.2016.Statistical
MethodsforAnalysingMultivariateAbundanceData.3.11.9.
Weigel,B.M.,Lyons,J.,Palne,L.K.,Dodson,S.I.,Undersander,D.J.1999.Using
StreamMacroinvertebratestoCompareRiparianLandUsePracticesonCattle
FarmsinSouthwesternWisconsin.JournalofFreshwaterEcology15:93-106.
Tables
Table1:ThistablegivestheresultsofeachforwardstepwiseregressiondonewiththeSimpson’s
DiversityIndexastheresponsevariableandLand-coverTypes[Type1-NorthernRockyMountain
lowermontaneriparianwoolandandshrubland,Type2-RockyMountainlowermontanefoothilland
valleygrassland,Type3-cultivatedcrops,Type4-openwater,Type5-majorroads,Type6-low
intensityresidential,Type7-opendeveloped]andCalculatedMetrics[EPT,PMA,HBI,percent
Chironomidae,abundance,andrichness]asthepredictorvariables.
Table2:Showsthenumberofeachmajorfamilyofmacroinvertebratecountedateachlocation.
Thesenumberswereusedtomakethepiecharts(figures1-5).TherangeofSimpsonsvaluesare
given,theSimpsonsDiversityIndexwascalculatedindividuallyforeachofthethreesamplingevents
ateachlocation.
Figures
Figure1:Mapofsamplingsiteswithland-covertypesateachsite.
LowerJocko
Chironomidae
Coleoptera
Diptera
Ephemeroptera
Plecoptera
Trichoptera
Worm
UpperJocko
Chironomidae
Coleoptera
Diptera
Ephemeroptera
Plecoptera
Trichoptera
Worm
Jerry'sDitch
Chironomidae
Coleoptera
Diptera
Ephemeroptera
Plecoptera
Trichoptera
Ninepipe
Chironomidae
Coleoptera
Diptera
Ephemeroptera
Plecoptera
Trichoptera
Worm
Scud
Mission
Chironomidae
Coleoptera
Diptera
Ephemeroptera
Plecoptera
Trichoptera
Worm
Scud
Figures2-6:Anindividualpiechartforthemostabundantfamiliesfoundateachsite.TheJocko
RiverandJerry’sditcharesimilarinthatthemostabundantmacroinvertebratefoundatthosesites
aretheChironomidaewhicharemostlyfacultative.AtNinepipethedominatingmacroinvertebrate
areScudswhicharealsomostlyfacultative.InMissionTrichopteraarethedominanttaxawhichare
sensitivetopollution.
Figure7:Theresultoftheforwardstepwiseregressionwithinputsofallthefactorsi.e.land-cover
typeandcalculatedmetrics,showingthatrichnessandPMAarethestrongestpredictorvariablesin
determiningSimpsonsdiversity.
Figure8:PrincipleComponentsAnalysispreformedinSystat.Thisgraphshowsthattherearestrong
similartiesbetweenSimpsons,Type1land-cover,andrichnessatoursitesbutthereseemstobea
weakeryetstillpresentrelationshipbetweenthepreviousmentionedmetricsandEPTandPMA.
Figure9:InRpackage“mvabund”wepreformedanordinationandageneralizedlinearmodelon
therawmacroinvertebratecountdatatovisualizethesimilaritiesbygrouping.Thesitesarecolor
coded:Mission-grey,Ninepipe-red,JockoUpper-green,Jerry’sDitch-purple,JockoLower-blue.
Therearespatialdistinctionsfromeachsamplecollectedateachsite.Themacroinvertebrate
communitiesremainedconstantthroughoutthesamplingperiod(lateJunetolateJuly2016).