Interpreting River Health Data - Waterwatch Portal · Interpreting river health data 5 Waterwatch...

Interpreting River Health Data Waterwatch Victoria

FOREWORD

Taking a water quality sample is a complicated act. “No”, you rightly say, “surely anybody can look down a tube and estimate turbidity; it’s a bit harder to measure nutrients, but most people can be trained to do it competently. So what is complicated?”

It is what the sample means. Taking a water quality sample is to river health, what measuring a person’s temperature is to human health. Parents take a child’s temperature because it provides a surrogate for health. The parent may have only a vague idea why illness seems to lead to increased temperature, but they know that a temperature over 37oC could indicate sickness, perhaps an infection. This example provides some useful analogues for Waterwatch samplers.

By Dr. Ian Rutherfurd

Director of Integrated River Health, Department of Sustainability and Environment

Interpreting river health data

• First,therewasprobablysomereasonthattheparentmeasuredthetemperatureofthechild.Theysuspectedthattherewassomethingwrong.Theyhadaquestionthattheywantedanswered.

• Second,theywereconfidentthatthemeasuretheymadewasaccurateandthethermometerwasputinthecorrectplace.

• Third,themeasurementwascomparedagainstastandard.Inthiscasetheaveragehumancoretemperature.

• Fourth,aparentwillcontinuetomeasuretolookforchangeoratrend.Anupwardtrendisbadandwilltriggerfurtheraction.

• Fifth,thatfurtheractionmightbetotakethechildtoadoctor,anexpertininterpretingwhatiscausingthechangeinthetemperature(thegenericsurrogatemeasure).

• Sixthandfinally,aparentdoesnotbelievethattemperatureistheonlymeasureofhealth.Veryoldpeoplemightnothaveanelevatedtemperature,buttheymightalsobelievethattheyarenolongerashealthyastheywere.Human‘health’isacomplicatedidea,justasriverhealthisacomplicatedidea.

Sowhatisthismanualfor?Toextendthemedicalanalogy,thiscouldbedescribedasa‘communityhealthcentre’manualforcommunitywaterqualitymonitors.Manyparentshavegonetothecommunityhealthcentrenurseforreassurance,adviceorreferral.Thatnursehastobewelleducatedinbasicmedicine,buttheywillalsoknowwhentorefertoanexpert.

Thismanualisdesigned,inessence,tohelpinterpretthebasicsurrogatemeasuresofriverhealththatweuse.Itisalsodesignedtoavoidmakingbasicerrorsofinterpretation,inparticular,leapingtounwarrantedconclusionsandmistakingassociationforcausationwheninterpretingresults.

Followingtheexample,herearethebasicquestionsthatthismanualwillhelpyoutoaddress(althoughtheyarenotexpressedinthisorderinthemanual,youwillfindtheanswersthroughout):

1. What is the question? Waterwatchmonitorswillbeinvolvedintheprogramformanyreasons,butitisworthclarifyingthequestionsthattheyhave,sothatyoucanmakesurethattheirmonitoringwillanswerthem!Forexample,monitorsmayhaveasuspicionthatthewaterqualityintheircreekisdecliningandtheywanttoexplorewhetheritis.Willtheirroutinemonitoringtellthem?Alternatively,theymaycareabouttheconditionofareceivingwaterbody(suchasalake),inwhichcaseloadismuchmoreimportantthansampleconcentration.

2. How good is the measure? TheWaterwatchprogramisunderpinnedbybasicmeasurementprotocolsthatarebecomingmorerigorouseachyear.Thisisbecausethereisbasicallynopointhavingwaterqualitysamplesthatwecannottrust.Similarly,ourconfidenceinameasureisincreasedifthatmeasureissupportedbyothermeasuresandfollow-upsampling.

3. How does a result compare to the ‘standard’? Waterwatchmonitorsshouldknowwhatthestandardsarearoundeachclassofmeasure(somanymg/letc.),andhowtheywerearrivedat.Arethey,forexample,basedontoxicity,onpotablewaterstandardsoronaveragesfromunimpactedareas?Whataretheerrorbarsaroundthosestandards?

4. What does a change in a value mean from sample to sample? TrendisthehardestthingforaWaterwatchmonitortointerpret.Thevaluehaschangeddramaticallyoveraweek,isthatnormal?Weshouldknowthebasicprocessesthatinfluencethevariablesthatwemeasure(themostimportantofwhichisusuallysimpledischarge/rateofwaterflow).Thisalsorelatestobasicknowledgeaboutconcentrationandload.Yourloadcouldbeincreasingdramatically,butyourconcentrationcouldbegoingdown,simplybecausedischargeisrising.Keytothisissueistounderstandnaturalvariabilityasopposedtohumaninfluence.

5. When do we need to act on water quality measures? TheWaterwatchnetworkisanimportantpartofthewaterqualityreportingnetworkinVictoria.Sowhatconstitutesaproblemandwhendoyoureportor‘act’onproblems?Thiswillrelatetostandardsthatareexceededortounusualtrends.Thereareformalprotocolsforwhentocallintheregulatorsortheexperts.Weshouldknowwhentoseekhelp.

6. How do my water quality measures relate to stream health? Waterqualityisonlyoneaspectofstreamhealth,albeitacriticalone.Butweshouldallbeawareofwhereourpieceofstreamandthewaterthatflowspastusdaybyday,fitsintothelargergoalsofriverhealth.WhataretheoverallplansforthisriverandhowcanIhelp?

Aboveall,wehopethatthismanualwillhelpWaterwatchstaffandmonitorstodomorethanwatchthewater–buttoquestion,tomeasureandtoact.



1 Introduction 5

2 What is river health? 6

3 The exploration and Interpretation of data 7

3.1 Introduction 7

3.2 Preliminaries 7

3.3 The Data 8

3.4 Interpreting the Data in terms of Environmental Setting 14

4 Specific Indicators 18

4.1 Phosphorus 18

4.2 Nitrogen 20

4.3 Electrical conductivity 24

4.4 pH 26

4.5 Turbidity 28

4.6 Dissolved oxygen 30

4.7 Temperature 34

4.8 E. coli 35

5 Drawing conclusions 36

6 Presenting the data 37

7 Sources of data and expert advice 38

8 Further reading 39

9 Glossary 40

CONTENTS

Courtesy of Goulburn Broken Waterwatch.


5Interpreting river health data

Waterwatch Victoria is a community waterway monitoring and engagement program that connects local communities with river health and sustainable water management issues. Waterwatch coordinators and community volunteers have been collecting water quality data across Victoria for over 15 years, with a strong emphasis on data confidence and quality data since 2000.

INTRODUCTION1

Thismanualhasbeendevelopedtoimproveandsupportcoordinatorfeedbacktomonitorsandcommunitiesonwhattheirwaterqualitydatasaysabouttheirlocalsites.Themanualwillbeaninitialreferenceforcoordinatorswantingtoexploreandinterpretwaterqualitydatabeforeproducingdatareports.

ThemanualassumesusershavesomeknowledgeandskillsinMicrosoft ExcelandthatthemanualisreadinconjunctionwithrelevantstateandregionaldocumentsincludingtheWaterwatch Victoria Methods Manual (1999),Waterwatch Victoria Data Confidence GuidelinesandregionalData Confidence Plans and Equipment Manuals.Byusingthismanualonly,coordinatorscannotexpecttobecomeexpertsininterpretingwaterqualitydata.Coordinatorsareencouragedtoaccessotherresourcesforassistanceinwaterqualitydatainterpretation,someofwhicharelistedinSection7,’Sourcesofdataandexpertadvice’.

ThemanualcoverswaterqualityindicatorsregularlymeasuredintheWaterwatchprogram-theplantnutrients(phosphorusandnitrogen),dissolvedoxygen,turbidity,electricalconductivity,pH,temperatureandE. coli.MacroinvertebratesarenotcoveredinthemanualanddetailedinformationonsamplingandinterpretationcanbefoundintheEPAVictoriapublicationRapid Bioassessment Methodology for Rivers and Streams.(2003,Publicationno.604.1).


6 Interpreting river health data

An ecologically healthy river is a river that retains its major ecological features and functioning similar to that prior to European settlement and which would be able to sustain these characteristics into the future. By this definition, an ecologically healthy river need not be pristine. For example, exotic species may be present or fish passage may be provided by fishways. However, overall, the major natural features, biodiversity and functions of a river are still present and will continue into the future.

WHATISRIVERHEALTH?2

Waterqualityistraditionallymeasuredtoassessriverhealth,althoughinmorerecentyearsbiologicalindicators,suchasmacroinvertebrates,havealsobeenused.Bothwaterqualityandbiologicalassemblagesvaryfromyeartoyear,betweenseasonsandfromturbulenthighmountainstreamstolarge,meanderinglowlandrivers.Understandingthesetemporalandspatialdifferencesisimportantinassessingriverhealth.

Indicators of river health

Variousindicatorsmaybeusedtoassessriverhealth.Indicatorsarethebestpracticalrepresentationofissuesimpactingriverhealth.Forexample,phosphorusisusedasanindirectmeasure(indicator)ofpotentialexcessiveplantgrowth(theissue).Phosphorusisamajorplantnutrientandwhenthereisexcessphosphorusthereislikelytobegreaterplantgrowth.However,amoredirectmeasurewouldbetoassessplantbiomassorproductivity,butthisistimeconsuminganddifficultsoisnotgenerallyundertaken.Indicatorsareatoolandmustnotbethefocusofanassessment.Theissuemustbecentraltotheassessment.

Themostregularlyuseddirectmeasureofriverhealthisthemacroinvertebratecommunity.Themacroinvertebratecommunityistheoutcomeofenvironmentalconditions,primarilywaterqualityregime,habitatqualityandflowregime.Changesintheseconditionswillchangethecommunity.

Anothermeasuresometimesusedtoassessriverhealthispollutantload.Pollutantloadistheamountofapollutantthatpassesapointoveragivenperiodoftime,usuallyayear.Regularmonthlymonitoringisnotsufficienttoestimateloads.Toaccuratelymeasureloads,eventbasedmonitoringisneeded.Thisrequiressubstantialsamplingeffortatthetimeofhighflows,inparticularfloods.

Concentrationsprovideameasureofimmediateavailabilityoreffectonbiotaorecosystems,whereasloadsaremoreapplicabletosinkssuchaslakes,estuariesandmarineenvironmentswherepollutantsaccumulateinsedimentstobepotentiallyreleasedwhenconditionsarefavourable.LoadsarenotreportedbyWaterwatchandthereforearenotaddressedanyfurtherinthismanual.

Impacts on river health

Pollutionfromhumanactivitieshasimpactsonriverhealth.Pollutantsmayenterriverscontinuously,aswithasewagedischarge,orintermittently,aswithstormwater.Commonpollutantsenteringriversincludeplantnutrients(inparticular,phosphorusandnitrogen),sediment,saltandoxygendemandingsubstancesandtoxicantssuchasheavymetalsandbiocides.Changesinwaterqualitymaybeshortorlongtermdependingonthepollutant.Forexample,oxygendemandingsubstancessuchasrawsewagewillbreakdownwithinhoursordayswhereasheavymetalsmayremainfordecades.Whilesomepollutantsarebrokendownorwillpassdownstreamquickly,othersbindtoparticlesofsedimentinthewaterandwillsettleout,accumulatingintheriversediments.Nutrientsandmanyofthetoxicantsthataccumulateinthesedimentswillremainthereandhaveminimalimpactontheaquaticenvironment.However,undercertainenvironmentalconditionstheymaybereleasedfromthesedimenttothewatercolumnwheretheycanimpactriverhealth.LowoxygenlevelsinthesedimentorlowpHlevelscanfacilitatethisremobilisation.

Itisimportanttonotethatimpactsfrompollutantsmaynotbedetectedbyregularmonitoringbecauseofthetimingofadischarge,thefateofapollutantand/orwheremeasurementsaretaken.

Climate change

Climatechangewillhaveasubstantialimpactonaquaticecosystemsincludingincreasedwatertemperatures,decreasedstreamflows,moreextremeevents,suchasbushfiresandashiftfromperennialtointermittentstreamflows.Thesearepotentiallyprofoundchangestoaquaticecosystems.Itispossiblethattheclimaticconditionsexperiencedoverthelastdecadearearesultofclimatechangeratherthanprolongeddroughtconditions.Waterqualityandbiologicaldatagatheredpriortothisperiodarethereforelikelytorepresentdifferentenvironmentalconditionstothecurrentconditions.


THEEXPLORATIONANDINTERPRETATIONOFDATA3

3.2 Preliminaries

The questions

Beforebeginningtheexplorationorinterpretationofwaterqualitydata,youneedtoformaclearsetofquestionsthatyouareseekinganswersto.Usuallythesequestionscomefrommanagementobjectivesorfocusonspecificwaterqualityissuesatthesiteorinthereachorcatchment.Forexample,thequestionscouldsimplybe“Whatistheconditionorhealthofthesite,reachorcatchment?”.HealthorconditioncanbeassessedbycomparingthedatacollectedagainstStateEnvironmentProtectionPolicy(WatersofVictoria(SEPWoV))waterqualityandbiologicalobjectives.Otherquestionsmayincludeexaminingtheimprovementsthathavearisenfrommanagementactiontakeninthecatchment;thechangesresultingfromdrought,floodorclimatechange;andtheassessmentofhumanimpactssuchassewagedischargesandtownandagriculturalrunoff.Beclearastowhyyouarelookingatthedata.

Conceptual models

Conceptualmodelsmayhelpdeveloptherightquestions. Aconceptualmodelisanillustrativerepresentationofanissueanditsrelationshipswithotherenvironmentalfactors.Theissueisusuallyaspecificecosystemvalueorathreattothatvalue.Forexample,ahealthynativefishcommunitymaybeavalueyouwanttomanageandprotect,whereasanalgalbloommaybeathreatyouwanttoprotectagainst.Amodelshouldrepresentallrelevantdetailsknownabouttheecosystem,includingrelationshipsbetweenthevalueorthreatandfactorsinfluencingit.Conceptualmodelscanprovidethebasisfordevelopingquestionsonpotentialcause-effectrelationshipsandforcommunicatingunderstandingoftheissue.

Figure3.1providesanexampleofaconceptualmodelforcausesofincreasedphosphorusconcentrations.Thesortofthingsthatmayexplainincreasedphosphorusatasiteincludethelossofnativevegetationinthecatchment,recentbushfiresinthecatchment,stockaccesstothestream,dischargestothestream(e.g.fromsewagetreatmentplants),intensiveagriculturalactivities(e.g.dairying)andpoorriparianzonecover.Manyofthesefactorsmaybeexacerbatedbystormeventsinthecatchmentinthehoursordayspriortosamplingasstormeventswashsedimentsintostreams,increasingphosphorusinputs.

Figure 3.1 Conceptual model for factors causing increased phosphorus concentrations.

Although‘increasedphosphorus’maybeatriggerforconcern,lowphosphorusitselfisnotavalueofthestream(i.e.astreamisnot‘valued’foritslowphosphorusconcentrations).Itistheimpactthatincreasedphosphorusconcentrationsmayhaveonthestreamecosystemthatmayaffectthestream’svalue.Forexample,increasedphosphorusconcentrationsmayincreasetheriskofanalgalbloomtherebythreateningtherecreationalandaestheticvaluesofthestream.

Conceptualmodelscanalsobeusedforproblemsolvingissuesandriskstovaluesinnaturalsystems.Figure3.2showsanexampleofanissue-basedconceptualmodelfocusingonalgalbloomsinalowlandreachofriver.Inthismodel,increasedphosphorusconcentrations,warmerwatersandreducedwaterturbulencewillincreasetheriskofanalgalbloom.Theindirectfactorsthatleadtotheseconditionsincludeclimatechange,catchmentdisturbanceandreducedriverbaseflows.Althoughitisonlynecessarytomonitorthealgallevelstoidentifythegrowthofabloom,dataneedstobecollectedforalloftheparametersintheconceptualmodeltounderstandthedriversofthebloom.Thisdataprovidesinformationthatcanthenbeusedtoassistinmanagingfutureblooms.

3.1 Introduction

This section briefly describes approaches and methods that can be used to explore patterns in data and identify what those patterns may mean. Before looking at the data it is valuable to clearly state the questions you are seeking answers to. Often these questions are best formed through the development of conceptual models. These two aspects of data exploration are examined first. Important features of the data are also discussed, including the use of summary statistics, time series graphs and spatial graphs. The section finishes with an examination of the influences of environmental setting upon water quality indicators.



Figure 3.2 Conceptual model for algal bloom risk in a lowland river reach.

ThemodelshowninFigure3.2isasimpleoneandcouldcontainmorecausalfactorsandarrows,includingfactorsandarrowsrelatedtoreducingtheriskofalgalblooms.Theimportantpointisthatconceptualmodelscanbeusedtohelpyoudeterminethequestionsthatyouareaskingofyourdataandtohelpyouvisualisetheimportanceofyourresultswithinthecontextofthewholesystem.

3.3 The Data

Tobeabletoansweryourquestionsaboutthewaterbodyyouaresampling,youneedtohaveappropriatemeasuresandsampleresults.Whenyoutakeameasurementorasamplefromariver,youobtainasinglepieceofinformationaboutthatriver.Thatpieceofinformationgivesyouanindicationoftheconditionoftheriveratthetimeofsampling.However,youdonotknowwhetheryousampledatatimewhenthewaterqualitymeasurewasunusuallyhighorunusuallylow.Inotherwords,youhavelowconfidenceinasinglemeasurebeinganaccuraterepresentationof‘typical’conditions.

Somesamplingprogramsusetheapproachoftakingone-offsamplesofasuiteofmeasures(e.g.nutrients,salinity,dissolvedoxygen,suspendedsolids)atmanysiteswithinaregiontogaina‘snapshot’ofriversinanarea.Theseprogramsareusefulforgainingsomeunderstandingofconditionsacrossabroadregionbutsufferfromhavinglowlevelsofconfidencethatanysinglesitehasbeenadequatelycharacterised.

Calculating the mean

Calculatingthemean(sometimescalledthe‘average’)requiresaddingtogetherallthedatapointsthendividingthetotalby thenumberofdatapoints.Forexample,considerthefollowing13results:

5,5,6,6,4,6,97,6,5,5,45,88,and55

Themeanvalueoftheseresultsisobtainedbyaddingthemalltogetheranddividingby13:

(5+5+6+6+4+6+97+8+5+6+45+88+5)÷13=22

Calculating the median

Themedianofasetoffiguresisobtainedbyarrangingthemallfromthelowestnumbertothehighestandtakingthemiddlevalue.Usingthesame13results,themedianisfoundtobe6:

4,5,5,5,5,6,6,6,6,8,45,88,97

Themedianisalsoknownasthe‘50THpercentile’,asitisthehalfwaypointinthedataarray(i.e.50%ofthewayfromthelowesttothehighestdatapoint).

Increasedconfidenceintheinformationprovidedbyadatasetisgainedthroughincreasednumbersofdatapointsforeachsite.Onceyouhaveasitewithmanydatapoints,twoquestionsthatariseare:

1. Whatdoesthedatatellmeabouttypicalconditionsofthesite?and

2. Whatdoesthedatatellmeabouttherangeofconditionsatthesite–i.e.notjustthetypicalconditions.

Whiledatainterpretationisbestundertakenwithacompleteorlargedataset,thisisnotalwayspossible.Datainterpretationshouldbeundertakenwithwhateverdataisavailable,notingthislimitationandanyassumptionsthataremade.

Interpretationofdataishelpedbyusingstandard,recognisedapproachesthatallowyoutocomparethedataagainstothersites,objectivesorguidelinesanddatafrompreviousyears(todeterminetrends).

Oneofthemorecommonwaysofdefiningthetypicalstateofawaterbodyistouseameasureofcentraltendency,suchasthemeanorthemedian.



Distribution

Agraphofthedatasetpresentedinthe‘Mean’and‘Median’boxesisshowninFigure3.3.Thistypeofgraphisknownasafrequencydistribution.Thefrequencydistributionoftheabovedatadisplaysafeaturethatiscommoninwaterqualitydatasets–a‘skew’,withmostvalueslocatedatthelower(left)endofthex-axis,andalong‘tail’totherightcreatedbyafew,muchlargervalues.AsdisplayedinFigure3.3,mostofthedatapoints(measures)areeither5or6(withafrequencyof4readingsforeach).Themedianofthedataset(6)reflectsthisdistribution,whereasthemeanvalue(22)isstronglyeffectedbythreehighvalues(45,88and97).

Figure 3.3 Frequency Distribution, Median and Mean Values of a Hypothetical Data Set.

ThedatapresentedinFigure3.3couldrepresentturbiditymeasuresfromaheadwaterstreaminanalpinearea.Themajorityofthetimethewaterisveryclear(turbiditybetween4and8NTU)butonthreeoccasions–possiblyduringorfollowingstormevents–theturbidityrangedfrom45to97NTU.Intermsofwaterqualitymanagement,itismoreusefultoknowthatthewaterisgenerallyaround6NTUthantoknowthatitsmeanis22NTU.Ifthehighestrecording(97)hadbeenmissed(forexample,duetoafaultymeteroralostsample),themedianvalueinthedatasetwouldremain6,whereasthemeanvaluewoulddropbyapproximately30%to15.75.

Inskeweddatasets,medianvaluesaregenerallymoreusefulthanmeanvaluesinreportingthetypicalwaterqualityatasite.Thisiswhywaterqualitysummariestypicallyusemedianvaluesratherthanthemean.WhenreportingwaterqualityforWaterwatch,itisrecommendedthatthemedianbeusedforsummarisingcentraltendency.

Summary statistics

Themeanandmedianareexamplesofsummarystatistics.Summarystatisticsarecharacteristicsofadatasetthatprovideinformationaboutthedatawithoutneedingtoreproducethewholeset.

Inadditiontoidentifyingcentraltendency,itisoftenimportanttosummarisetherangeofconditionsthatoccurwithinthedataset.Thehighestandlowestvaluesrecordedatasiteprovidethefullrange,althoughthesemaybeinfluencedbyextremeeventsthatarerarelyencountered.Percentilesareoftenusedtogainsomeunderstandingoftherangewithoutincludingextremes.Percentilesaresummarystatisticsthatidentifythevalueofavariablewithinthedataset,belowwhichacertainpercentofobservationsfall.Forexample,the20thpercentileisthevalue(orscore)belowwhich20percentoftheobservationsmaybefound.

EPAVictoriahassetmanyofitsSEPPobjectivesas75thpercentiles(see“Assessingagainstwaterqualityobjectives”,p.12).Forthedatasetusedinthepreviousexample,the75thpercentileis8,whichisthevaluethatis75%ofthewayfromthelowesttothehighestdatapoint:

Thisdatasetisindicativeofrelativelystablewaterqualityconditions(median6;75thpercentile8),withinfrequentdisturbances.Ifthedatasetweredifferent,forexample-

4,5,5,5,5,6,6,6,40,48,55,88,97;

thenthe25thpercentileandthemedianwouldbeunchangedbutthe75thpercentilewouldbe48,indicativeofmorefrequentdisturbances,probablyduetomorethanjuststormevents.WhenreportingagainstSEPP,youwillneedtousethe75thpercentileformostoftheWaterwatchwaterqualityindicators,aswellasthe25th percentileforpHanddissolvedoxygen.

Manysoftwarespreadsheetpackages,suchasEXCEL,canbeusedtoquicklyandeasilycalculatemediansandotherpercentiles.

50th percentile

4,5,5,5,5,6,6,6,6,8,45,88,97

75thpercentile25thpercentile

median

Median Mean

4

3

2

1

4 6 8 (22) 45 88 97

Freq

uenc

y

Indicator measurements



Using time series

Variabilityisanaturalpartoftheaquaticenvironment.Typicallywithwaterqualitydatathereisa10-20%differencebetweenconsecutivesamples.Forexample,wheresalinityismeasuredat500µS/cm,changesofupto100µS/cmfourweekslatermaybeduesimplytonaturalvariationandnotasignificantchangeintheaquaticenvironment.However,therewillbemoresubstantialchangesduetofloodsorafterlongdryperiods.Forexample,turbiditymaygouptenfoldduringafloodeventandsalinitycanmorethandoubleasflowsdecreaseandgroundwaterbecomesthemainsourceofflow.

Displayingthedatainatimeseriesprovidesapictureofchangeovertime.Timeseriesplotsareusefultoillustrateseasonalpatterns,majoreventsorwhen‘one-off’highorlowlevelsaremeasured.Figure3.4illustrateslongtermpatternsinturbidityinthelowerOvensRiver.Notetheseasonalpatternswhereturbidityrisesduringwinter/springanddropsinsummer/autumn.Alsonotetheoccasionalhighlevelsinwinter,likelytobeassociatedwithfloodevents.Droughtin2002canbeseeninlowturbiditiesduringthewinterandtheeffectsoffiresin2003canbeseenintheelevatedlevelspostfires.

Figure 3.4 Turbidity time series 2000 to 2008 for Ovens River at Peechelba (VWQMN site no. 403241). Source: VWQMN Data Warehouse. Graph generated using Microsoft Excel.

Trendlinescanalsobeaddedtodetermineiftherehavebeensubstantialincreasesordecreasesovertime.Fortheclearestpicture,alongtimeperiodisrequiredtoshowtemporalchanges.Usuallylessthanfiveyearswillnotprovideagoodassessment.Figures3.5and3.6displayfiveand14yearsofmonthlydatarespectively.Thetrendlinessuggestsubstantialchangesatbothsites.PhosphoruslevelsinBennisonCreek(Figure3.5)appeartohavedecreased.Notethelackofveryhighlevelsinthelateryearscomparedtoearlieryears,suggestingthatdroughtmayhavebeenthecauseofthisresult. InBrokenCreek,turbiditylevelshaveincreasedbetween1993and2007(Figure3.6).Levelsgraduallyincreasedovertimeandthereisnoobviousinflectionpoint,suggestinggradualincreasesfromthecatchmentratherthanachangeinactivityinthecatchment. By2002,turbiditylevelswereconsistentlyabove50NTUandoftenabove100NTU.Suchlevelsarelikelytobeharmfultoaquaticlife.



Figure 3.5 Total phosphorus time series and trend line for Bennison Creek at south Gippsland Highway (site no. BNN020). Source: West Gippsland CMA. Graph generated using Microsoft Excel.

Figure 3.6 Turbidity time series and trend line for Broken River at Gownagardie (site no. BAR010). Source: Goulburn-Broken Waterwatch. Graph generated using Microsoft Excel.



Assessing water quality against objectives

ThemajorityofassessmentsusingwaterqualitydatacollectedforWaterwatchwillbeagainsttheobjectivesprovidedintheSEPP(WoV).TherearesomeimportantpointstonoteabouttheSEPP(WoV)objectives:

1. Theyareambientobjectivesandarenotdesignedforuseinsingleorshort-termsamplingprograms.Theobjectivesaredesignedtoindicatelow-riskconditionswithinaregionorsegmentoveranextendedperiodoftime(monthstoyears).However,singlesamplingorafewsamplesfromasitecanstillbeusedtoprovidesomeindicationofconditionandmaysuggestthatfurthersamplingiswarranted

2. Theyaretypicallyprovidedas75thpercentiles(e.g.totalphosphorus,totalnitrogen,turbidity,electricalconductivity,pH-high)or25th percentilesforindicatorsthatshouldnotfallbelowparticularvalues(e.g.pH-low,dissolvedoxygen-low).Theuseof75thand25thpercentilesrequiresaminimumof11samplesforadequateconfidence.Therefore,anannualmonthlymonitoringprogramshouldbesufficientforassessmentagainstSEPPobjectives

3. Theyare‘trigger’values.Itisimportanttorecognisethedifferencebetweentriggervaluesandpass/failobjectives.Ifapass/failobjectiveisnotmet,thesiteisdeemedtohavefailedanassessmentofitsecologicalcondition.Althoughthisisusefulforauditingorcatchmentconditionreporting,itdoesnotnecessarilytriggerfollow-upaction.Incontrast,ifatriggervalueobjectiveisnotmet,thenthisshould‘trigger’someaction.Theactionthatistriggeredmayrangefromspendingafewhourstoidentifypossiblereasons(suchasnaturaldisturbanceslikestorms,droughtsorbushfires),throughtoundertakingecologicalriskassessmentsortarget-settingprograms.

4. Theobjectiveshavebeenderivedforregionsbasedonecologicalcondition(Figure3.7).

AdatasetcollectedfrommonthlysamplingatasitecanbecomparedtotheSEPP(WoV)objectivesforthesegment(waterwaysection)todeterminewhetheranyresultstriggerfurtheraction.TheSEPP(WoV)anditsbackgrounddocumentsprovideimportantinformationfortheassessmentofdataagainstobjectives.

0 100 20050 Kms

LEGENDSEPP (WoV) Segments

Highlands

Forests A

Forests B

Cleared Hills and Coastal Plains

Murray and Western Plains

Mallee

Figure 3.7 State Environment Protection Policy (Waters of Victoria) segments.



Table 3.1 Water quality data and annual percentiles for 2007 for Bennison Creek at South Gippsland Highway. SEPP (WoV) Cleared Hills and Coastal Plain segment. (site no. BNN020) Source: West Gippsland CMA

Date Electrical pH Total Turbidity Conductivity (pH Units) Phosphorus (NTU) (µS/cm) (mg/LP)

2 Mar 777 6.99 0.151 8 6 Apr 591 6.72 0.035 3 4 May 625 6.9 0.06 11 25 May 680 7.16 0.021 6 15 Jun 492 6.73 0.029 10 6 Jul 290 6.68 0.053 25 27 Jul 310 6.85 0.046 8 17 Aug 253 6.9 0.038 12 7 Sep 325 7.02 0.04 5 28 Sep 304 7.06 0.049 8 19 Oct 263 7.15 0.049 9 2 Nov 328 7.1 0.057 5 23 Nov 325 7.12 0.092 10

Annual 75th percentile 591 7.1 0.057 10 Annual 25th percentile - 6.85 - -

WoV Objective 75th ≤ 7.7 (annual percentile) 75th ≤ 500 25th ≥ 6.4 75th ≤ 0.045 75th ≤ 10

Asingleindicatortriggeringbyasmallamountforthefirsttime(i.e.ithasn’ttriggeredpreviously),maybeconsideredaresultworthyof‘watch’status.Incontrast,asuiteofindicators(e.g.totalphosphorus,totalnitrogenanddissolvedoxygen)thatcontinuallytriggerbyasubstantialamountand/orareincreasingintheirmagnitudeoftriggeringmaywarrantafullriskassessment.Betweenthesetwoscenarios,furtherassessmentbywaterwaymanagersoranaquaticecologistmaybewarranted.FurtherdetailsontheSEPP(WoV)approachcanbefoundinRisk-based assessment of ecosystem protection in ambient waters(EPAVictoria2004).

AnexampleofassessingsitedataagainstSEPPobjectivesisprovidedforBennisonCreek(Table3.1).Inthisexample,onlypHandturbiditymeettheSEPPobjectivesfortheClearedHillsandCoastalPlainsegment.Althoughnotlikelytobearisktotheecosystem,theresultsareanalertthattheremaybeanissueatthesite.Electrical

conductivityisovertheobjective,althoughnotsubstantiallyandnosinglemeasurementisabovepotentialtoxiclevels(greaterthan1,000µS/cm).Whiletotalphosphorusisalsojustovertheobjective,someofthemeasurementswereparticularlyhighandifsustainedcouldresultinexcessiveplantgrowth.Furtherinvestigationoftotalphosphorusislikelytobewarranted.Withthisprocessthereisaneedtolookatpastresultsandseeifapatternisemergingorwhethertheeventsareoneoffandareunlikelytooccuragainoroccurinfrequently.

TheSEPP(WoV)objectivesdonotprovideascaleofoutcomes(e.g.verygoodcondition,goodcondition,poorcondition,etc.).However,theydoindicateifasiteisgood(i.e.thesitemeetsthetriggervalue),orpossiblybad(thesitedoesnotmeetthetriggervalue).Thedistinctionisnotalwaysclearwhenresultsareclosetothetriggervalues,asdemonstratedintheBennisonCreekexample.



3.4 Interpreting the Data in terms of Environmental Setting

Relationships between indicators and other ecological components

Theresultsobtainedfromanysamplingeventorprogramwillreflectavarietyofphysicalandbiologicalfeaturesthatmaybeaffectingwaterqualityateachsite.Theseinclude:

• thenatureofthesurroundingcatchment(steepness,soiltype,vegetationcover);

• sitelocationwithinthecatchment(headwater/upland,lowlandplains);

• weatherconditionsduringorpriortosampling(storms,droughts);

• landusewithinthecatchment,particularlyclosetothesamplingsite(urban/industrial,agricultural,naturereserve);and

• in-streaminputsandactions(stockaccess,wastedischarges,waterremoval,storages).

Similarly,thewaterqualityatasiteaffectsothercomponentsofthewaterbody,includingotheraspectsofwaterqualityandthebiotathatthewaterbodysupports.Afewofthemorecommonrelationshipsbetweenwaterqualityindicatorsandotherecologicalcomponentsarepresentedinthissection.Greaterdiscussionofthemostsampledindicatorsisprovidedinsectionfive.

The surrounding catchment

Oneofthemostimportantnaturalfeaturesofthesurroundingcatchmentissoilerodibility(whichitselfisafunctionofvegetationcover,slopesteepnessandlengthandsoiltype).Inparticular,catchmentsthathavesteepslopesand/orsparsevegetationcoverareparticularlysusceptibletosoilerosion,withpotentiallyseriousimplicationsforwaterquality.

Soilthatiserodedfromthecatchmentandsubsequentlydepositedinstreamscandirectlyandindirectlyimpactontheconditionofthereceivingwaterbodies.Onedirectimpactisphysical,withcoarsergrainedsediments(sand,finegravel)fillingindeeppoolsrequiredbyfishandfillingininterstices(gaps)betweenrocksandstonesthatarevitalhabitatforsomemacroinvertebratespecies.Thedepositionoffinersedimentsderivedfromsoil(suchassiltsandclays)cansmotherhabitatandclogthegillsoffishandinvertebrates.

Ameasureoftheextenttowhichfinerparticlesarepresentinthewatercolumnisprovidedbyananalysisofsuspendedparticulatematter(SPM)concentrationinthewater(oftenalsocalledsuspendedsolids(SS),totalsuspendedsolids(TSS),ornonfilterableresidue(NFR)).Finerparticlescanalsoacttoreducelighttransmissionthroughthewatercolumn,inhibitinggrowthpotentialforaquaticmacrophytesandalgaeandimpactingonfishspeciesthatarevisualpredators.Turbidityisasurrogatemeasureoflighttransmissionthroughthewater.Highturbiditiesareoftenanindicationofpoor

landusemanagementwithinthecatchment,includingtheriparianzone.ElevatedturbidityandSPMareoftencorrelatedwithincreasedtotalnutrientconcentrations(totalnitrogenandtotalphosphorus),assoilparticlesoftenhavenutrientsattachedtothem.

Site location within the catchment

Sitesthatarehighwithincatchmentsareoftenmorepronetomarkedswingsinwaterquality,astheyhavelittlebufferingfromrainfallevents.Thisisbecauseheadwaterstreamsareoftensmallanddrainasmallcatchmentarea,soanyraineventwithinthatcatchmenthasthepotentialtocausesubstantialfluctuationsinwaterquality. Incontrast,largeriversinthelowerreachestypicallyhavemanytributariesfeedingintothemandatanytimemaybereceivingwatersfromsometributariesthathaveexperiencedstorms,andothersthathavenot.Therefore,waterqualitychangesareoftenonlysmall.Ofcourse,changesinwaterqualityfromamajorcatchment-widestormeventwillbereadilyobservableevenatlowlandsites.

Weather conditions prior to sampling

Undertypicalflowconditions,aperennialstreamwithinagivencatchmentwillbeexpectedtohaveareasonablypredictablesetofwaterqualitymeasurementsandvariability.However,unusualweatherconditionspriortosamplingcanleadtolargechangesinwaterquality.Forexample,duringdroughtsthegradualreductioninflowsleadstoanincreasedproportionofgroundwatercontributingtostreamflow.LargepartsofVictoriahavegroundwaterwithsubstantiallyhighersaltandnitrateconcentrationsthantypicallyfoundinflowingsurfacewaters.Concurrently,thereducedinputsofwaterfromoverlandflowandthroughflowresultinlesssedimentbeingwashedintothestreamandthereforelessassociatednutrients.

Asaresult,prolongeddroughtconditionsareoftenreflectedbyreducedstreamflowswithgreaterelectricalconductivities(EC),increasednitrateconcentrationsandreducedturbidities,SPMandtotalnutrientconcentrations.

Incontrast,highflowsfollowingstormeventsorsubstantialrainfallsaretypicallyassociatedwithgreaterdilutionofgroundwaterandincreasedsedimentloadsbeingwashedintothestream(Figure3.8).TypicalimpactsonwaterqualityarethereforereducedECandnitrateconcentration,accompaniedbyhigherturbiditiesandassociatedincreasedconcentrationsofSPM,totalnitrogen(TN)andtotalphosphorus(TP)(Figure3.8).Increasesinnutrientconcentrationscanleadtofurtherproblemsassociatedwitheutrophication(seesectionsonnutrientsandoxygen).


Catchment land use

Landusewithinacatchmentcanhaveamajorinfluenceonthewaterqualityofstreamsystems.Catchmentsthathavealargeproportionofnativevegetation,naturereservesandnationalparkswillhavelowerratesoferosionandthereforelessinputsofsedimentintostreams.Inpristineornearpristinecatchments,alargepercentageofwaterfromrainfalleventstendstosoakintotheground,recharginggroundwaterandgraduallycontributingtostreambaseflowoveranextendedperiodoftime.

Incontrast,catchmentswithalargedegreeofdisturbance(suchasurbanisation,largescalevegetationclearanceorpooragriculturalpractices)tendtohavehighersurfacerunoffduringrainfallevents.Thesurfacerunoffwillpickuppollutants(fromurbanareas)andsoilparticles(particularlyinagriculturalareas)anddepositthemintothereceivingwaterways.Anotherfeatureofincreasedoverlandflowindisturbedcatchmentsisthattherainwaterisdeliveredmorequicklytothestreams,resultinginahigherfloodpeakthatpassesquicklythroughthesystem(Figure3.8).Thehigherfloodpeakresultsinincreasedwashingawayerosionofstreambedsandbanks,inturncontributingmoresedimenttothestreamsystem.Therapiddrainingofcatchmentsalsoleadstoreducedgroundwaterrechargeandoncethefloodpeakhaspassed,thereislessgroundwatertocontributetobaseflow,resultinginagreaterpotentialforstreamstodryduringlowrainfallperiods.


Therefore,streamsthatdraincatchmentsdisturbedbylandusessuchasurbanisation,intensiveagricultureandlarge-scaleclearingcanbeexpectedtohavehigherconcentrationsofcontaminants,sedimentsandassociatedtotalnutrients,particularlyduringandsoonafterlargerainfallevents.

Closertothestreamchannel,removalofriparianvegetationcanintensifytheseimpactsandalsoaddtotemperaturefluctuations,particularlythroughreducedstreamshadinginsummer.Thewarmerwaterscandirectlystressin-streambiota,aswellasleadtoreduceddissolvedoxygenconcentrations(seeoxygensection),furtherstressingthebiota.

In-stream inputs and actions

Othercommoninfluencesonwaterqualitythatmaybereflectedinroutinesamplingincludedirectdischargestostreams(e.g.fromsewagetreatmentplants,dairyeffluentorstormwater),stockaccesstostreamsandflowregulationorextraction.Theeffectsofdischargeswillvaryaccordingtothedischargeanditsmake-up.

Sewagedischargeswilloftenhavehighnutrients(includingreactivephosphorus,ammonia/ammoniumandnitrates),highoxygen-demandingsubstances(whichmayleadtolowdissolvedoxygenfurtherdownstream),highsalts(reflectedinECreadings)andmiscellaneousorganicandmetalcontaminants(dependingonthesourcestothesewagetreatmentplant).

Figure 3.8 Typical flow hydrograph after rainfall in a forested and a disturbed catchment.

Natural vegetation CatchmentDisturbed Catchment

Flow

s (e

.g. M

L/da

y)

Rain event Time

Flood peak higher and shorter duration in disturbed catchment

Rapid increase in �ows after rain event

EC, nitrates diluted by substantial surface �ows

Surface runo� washes in lots of soil causing very high TSS, turbidity, TP and TN during high �ow event

High �ow from surface runo� will increase TSS Turbidity, TP and TN but not as much as in disturbed catchment

Flood peak lower in forest catchment

More rainfall reaching groundwater, resulting in higher base�ow (with lower EC and nitrates) between rain events

Gradual increase in �ows Less rainfall reaching groundwater, resulting in lower base�ow (with higher EC and nitrates) between rain events



Stormwatermayalsocontaincontaminantssuchasmetalsfromindustryandfromroads,aswellasparticulateswithnutrientsandhighoxygendemand.Dairyeffluentdischargingtostreamswilloftenhaveveryhighoxygendemandfromtheorganicmatter,aswellashighnutrients(includingreactivephosphorus,ammonia/ammoniumandnitrates).

Directstockaccessintowaterwaystypicallyresultsinincreasedbedandbankerosion,whichcontributestoincreasedturbidity,SPMandassociatedTNandTP.Faecalmatterfromstockalsocontributestooxygendemandandhighnutrients(includingreactivephosphorus,ammonia/ammoniumandnitrates).

Streamflowregulationcanleadtogreatertemperaturefluctuations.Thiscanbeattributedtocoldwaterdischargesfromdeepstoragesandreducedflowhavinglessthermalmassandthereforeheatingupandcoolingdownmorerapidlywiththeambientairtemperatures.Rapidfluctuationsinflowsassociatedwithsomeregulationcanalsocreatebankinstability,leadingtobankfailureandincreasedsedimentinput.

Howqua River upstream of Running Creek Campground (looking downstream).

Looking for spatial patterns

Waterqualitychangesasyougodownstream.Small,shallowheadwaterstreamschangetobigger,deeperlowlandrivers.Waterclaritydecreasesnaturallyandnutrientandsalinitylevelsincreasenaturally.Themagnitudeofchangeswilldependonthecatchment.Forestedcatchmentswithlittlelanddisturbancechangegradually,whilemoredisturbedcatchmentswillhaverapidlychangingwaterquality.Humanactivitiesincludingagricultureandurbanisationcanhavesubstantialeffectsonwaterquality.

Assessingdownstreamwaterqualityisavaluabletooltodeterminepointsourcedischarges,suchassewageorstormwaterdischarges,anddiffusesources,suchasfromagriculture.Togiveagoodspatialrepresentationofchangesandgooddifferentiationoflikelysources,sitesneedtoberelativelyclosetogether.Inaddition,usinglongertermdata(atleasttwelvemonths)willgiveamoreeffectiveandaccurateassessment.

AnexampleofspatialpatternsisgiveninFigure3.9fortheBrokenRiver/BrokenCreekcatchment.Inthiscatchment,turbiditylevelsarelowintheforestedreachesofthecatchmentupstreamofLakeNillahcootie(aroundtheSEPPWoVobjectiveof5NTU).Downstreamofthelake,turbiditylevelsincreasedtomorethanfivetimestheupstreamlevels.However,theincreaseatCaseysWeirisfargreater,tomorethantwicetheSEPPWoVobjectiveof30NTUforthislowlandsegment.Theincreaseisalmostcertainlyduetotheveryturbidwatersenteringfromanoff-linestorage(LakeMokoan).Surprisingly,evengreaterincreasesweremeasuredfurtherdownstreamofCaseysWeir,eventhoughobvioussourceshavenotbeenidentified.Thecontinuedincreaseislikelytocomefromoneormoresources,mostprobablyfromthebedandbanks,irrigationreturnwatersorsurfacerunoff.

Inthisexample,sitesarerelativelyclosetogetherwhichgivesagoodspatialrepresentationofchangesandgooddifferentiationofreachchanges,howeverithasnotidentifiedsources.Sourcescouldbeinvestigatedbyundertakingsiteinspections,discussionswithlocallandholdersand/orfurthersamplingdesignedspecificallytoidentifypotentialsources.


Benalla

Goorambat

Devenish

Yundool

Youarang

Katamatite

L a k e M o k o a n

Swanpool

Lima South

L a k e N i l l a c o o t i e

Bridge Creek

Barwite

B R O K E N R I V E R

Mans�eld

Midland Hwy

H U M E F W Y

Midland H

wy

Midlan d

Hw

y

BR

OK

EN

RIV

ER

B R O K E N R I V E R

Midland Hwy

Benalla-Tocumw

al Rd

BR

OK

E N C

RE

EK

St James Rd

Tungamah Main Rd

Yarrawonga-Katamatite Rd

Waggarandall

M o u n t S a m a r i a P a r k

0 3 6

K I L O M E T R E S

N

Melbourne 200km

0 50 100 150 200 250 300

Katamatite

Cobram Rd

Youarang Rd

St James Rd

Devenish

Goorambat

Caseys Weir

Downstream Benalla

Lake Benalla

Benalla (Footbridge)

Evans Bridge

Downstream Nillahcootie

O’Hallorans Rd

Whit�eld Rd

Byrnes Rd Ford

Barragunda

Turbidity (NTU)


Figure 3.9 Turbidity in Broken River and Broken Creek. Long term 75th percentiles (1996-2008). Data Source: Goulburn-Broken Waterwatch.


SPECIFICINDICATORS4

Althoughphosphorusisoneofmanynutrientsrequiredbyplantsandanimals,infreshwatersystemsitisoftenthenutrientlimitingplantgrowth(i.e.itisthenutrientinshortestsupply).Thisplacesaspecialimportanceonthemonitoringandmanagementofphosphorus.

Phosphorusisrarelyfoundinitselementalform(P).Itusuallyoccursinwaterwaysasaformofphosphate(PO4

3-).Differentformsofphosphatearecategorisedaseitherchemicallybasedformsoranalyticalmethods-basedforms.Chemicallybasedformsofphosphateinclude:

• OrthophosphatesThesearetheformsmostreadilyavailabletoplantsandincludethesimpleinorganicformsofphosphates,suchasPO4

3-,HPO42-,H2PO4-,andH3PO4.

• CondensedPhosphatesThesearecomplex,tightly-bound,inorganicphosphatecompoundssometimesreferredtoas‘polyphosphates’.

• OrganicphosphorusThisreferstoaphosphatemoleculeassociatedwithacarbon-basedmolecule,asinplantoranimaltissue.

Organicandinorganicphosphatecanbedissolvedinthewaterorsuspended(attachedtoparticlesinthewatercolumn).

Thethreemostcommonlyusedanalyticalmethodsformeasuringphosphorusconcentrationsare:

• ReactivePhosphorus(RP)Thismeasuresanyformofphosphorusthatreactswithreagentsinacolorimetrictestwithoutpriorfilteringordigestion(acidandheating).Thismethodisalsoknownas‘totalreactivephosphorus’andislargelyameasureoforthophosphates.However,theprocedureisn’t100percentaccurateandsoresultswillincludeasmallfractionofsomeotherformsoforganicandinorganicphosphorusthatareeasilybrokendowninwater.

• FilterableReactivePhosphorus(FRP)Thisisthefractionofthetotalreactivephosphorusthatisinsolutioninthewater(asopposedtobeingattachedtosuspendedparticles).Itisdeterminedbyfirstfilteringthesample,thenanalysingthefilteredsampleforreactivephosphorus.

• TotalPhosphorus(TP)Thisisthesumoforganicandinorganicformsofphosphorusinunfilteredwatersamples.Theprocedureusessampledigestion(acidandheating)andmeasuresbothdissolvedandsuspendedorthophosphate.

AsimpleconceptualmodelofthedifferentmeasuredformsofphosphorusispresentedinFigure4.1.Ameasurementoffilterablereactivephosphorus(FRP)isacomponentofmeasuringreactivephosphorus(TRPor,moreusually,RP),whichitselfisacomponentoftotalphosphorus(TP).

Figure 4.1 Conceptual model of common forms of phosphorus in water.

How is it measured?

Asdescribedabove,thethreemostcommonwaystomeasurephosphorusareRP,FRPandTP.WithinVictoria,theSEPP(WoV)objectivesforphosphorususeTP.AlthoughTPislikelytobeanoverestimateofthebiologicallyavailablephosphorusinawatersample,biochemicalprocessessuchasremineralisationoforganicphosphorusandconversionsbetweenthevariousforms,meanthatmeasuresofTPgiveareasonableindicationoftheamountofphosphorusultimatelyavailable.

TheSEPP(WoV)objectivesarepresentedas75thpercentilesandvarybetweensegmentsofeachwaterway.

Thedifferentmeasuresofphosphoruscannotbesubstituted.However,ifameasureofFRPorRPwasapproachingorgreaterthantheSEPP(WoV)objectiveforTP,thiswouldbeastrongindicationofexcessivephosphorusconcentrations.

4.1 PhosphorusWhat is it?

Phosphorus is a naturally occurring element originating from minerals in rocks and is essential for animal and plant life. In natural circumstances, phosphorus usually enters waterways from the weathering of rocks (inorganic phosphorus) and the decomposition of plant and animal material (organic phosphorus).

Interpreting River Health Data - Waterwatch Portal · Interpreting river health data 5 Waterwatch...

Documents

Transcript of Interpreting River Health Data - Waterwatch Portal · Interpreting river health data 5 Waterwatch...