ORIGINAL RESEARCH ARTICLE
Water type quanti fi cation in the Skagerrak, the Kattegat and off the Jutland west coast
Trond Kristiansen
a, Eyvind Aas
b,*
aInstituteofMarineResearch,Bergen,Norway
bDepartmentofGeosciences,UniversityofOslo,Oslo,Norway
Received1September2014;accepted17November2014 Availableonline6December2014
Oceanologia(2015)57,177—195
KEYWORDS GermanBight;
Kattegat;
Skagerrak;
Watertypes;
CDOM
Summary Anextensivedataseriesofsalinity,nutrientsandcoloureddissolvedorganicmaterial (CDOM)wascollectedintheSkagerrak,thenorthernpartoftheKattegatandofftheJutlandwest coastinAprileachyearduringtheperiod1996—2000,bytheInstituteofMarineResearchinNorway.
Inthismonth,afterthespringbloom,GermanBightWaterdiffersfromitssurroundingwatersbya highernitratecontentandhighernitrate/phosphateandnitrate/silicateratios.Thespreadingof thiswatertypeintotheSkagerrakisofspecial interestwithregardtotoxicalgalblooms.The quantificationofthespatialdistributionsofthedifferentwatertypesrequiredthedevelopmentofa newalgorithmfortheareacontainingtheNorwegianCoastalCurrent,whileanearlierDanish algorithmwasappliedfortherestofthearea.Fromtheupper50matotalof2227observationsof salinityandCDOMcontenthavebeenusedtocalculatethemeanconcentrationofwaterfromthe GermanBight,theNorthSea(Atlanticwater),theBalticSeaandNorwegianrivers.TheAtlantic Waterwasthedominantwater type,withameanconcentrationof 79%,GermanBightWater constituted 11%, Baltic Water 8%, and Norwegian River Water 2%. At the surface the mean percentagesofthesewatertypeswerefoundtobe68%,15%,15%,and3%,respectively.Within thenorthern part of theSkagerrak, closer totheNorwegiancoast, thesurfacewaters were estimatedtoconsistof 74%AtlanticWater, 20%BalticWater,and 7%NorwegianRiver Water.
TheanalysisindicatesthatthecontentofGermanBightWaterinthispartislessthan5%.
#2014InstituteofOceanologyofthePolishAcademy ofSciences.Productionandhostingby ElsevierSp.zo.o.Allrightsreserved.
PeerreviewundertheresponsibilityofInstituteofOceanologyofthePolishAcademyofSciences.
* Correspondingauthorat:DepartmentofGeosciences,UniversityofOslo,Gaustadalléen21,N-0315Oslo,Norway.
Tel.:+4722855808;fax:+4722855269.
E-mailaddress:eyvind.aas@geo.uio.no(E.Aas).
Availableonlineatwww.sciencedirect.com
ScienceDirect
jo u rn al ho m e p age :w w w. el s ev i er. co m / lo c a te /o c e an o
http://dx.doi.org/10.1016/j.oceano.2014.11.002
0078-3234/#2014InstituteofOceanologyofthePolishAcademyofSciences.ProductionandhostingbyElsevierSp.zo.o.Allrightsreserved.
1. Introduction
Skagerrakis heavily influenced byboth theBaltic andthe NorthSea.About70%ofthewaterenteringtheNorthSeais assumedtopassthroughthisarea(ICES,1983),andmanyof thehydrographicaleventstakingplaceintheNorthSeawill bereflectedintheSkagerrak.Thegeneralcirculationinthe areais cyclonic (Fig. 1),andthe distributionofthe water massesismainlyregulatedbythein-andoutflowofwaterin theNorthSea.Thesteepbottomtopographycharacterised bythedeepNorwegianTrenchisofspecialimportanceforthe steeringandmixingofthewatermasses(Danielssenetal., 1997;Rodhe,1996; Svansson,1975).In additionthere isa largefreshwatersupplytotheSkagerrakfromtheBaltic,the Kattegat,localriversandcontinentalriverdischargetothe southern North Sea (Gustavsson and Stigebrandt, 1996).
The Jutland Coastal Current appears to constitute a link between the eutrophicated waters of the southern North Sea and the waters of the Skagerrak and Kattegat, and accordingto Aarup etal. (1996a) its transport maybe in therangeof0.01—0.02Sv.Theaveragetotaltransportofthe basic cyclonic circulation in the Skagerrak has been esti- matedto0.5—1Sv(Rodhe,1987,1992,1996).Thedistribu- tionoftherelativelyfreshsurfacewatersintheSkagerrakis stronglyinfluencedbyvaryingweatherconditions,andlocal wind conditions may block as well as increase the usual patternofsurfacecirculation(AureandSætre,1981).The surfacewatersmainlyfollowthegeneralcirculation.
Eutrophicationbyanthropogenicnutrientshasbeeniden- tifiedasanissueofconcernfortheSkagerrak/Kattegatarea
(Anon.,1993;NorthSeaTaskForce,1993).InlateApril1988, justafterthespringbloom,acruisewascarriedoutbythe InstituteofMarineResearchintheSkagerrak,theKattegat andalongthewesternDanishNorthSeacoasttoinvestigate theenvironmental conditionsandanomalous nutrientcon- centrationsrelatedtothelargefreshwaterrunofffromthe GermanriversandtheBalticatthattimeoftheyear(Aure etal.,1998).Highnitrateconcentrationswerefoundinthe surface waters of the inner Skagerrak and the Kattegat, resultinginhighNO3:PO4andNO3:SiO4ratiosasbothphos- phate and silicate were near the detection limits. In the beginningofMay1988,inconnectionwithweakwinds,awell definedsurfacelayerwith hightemperaturesandlowsali- nitiesandhighnitrateconcentrationsandnitrate/phosphate ratios(Dahletal.,2005),aharmfulbloomoftheprymne- siophyteflagellateChysochromulinapolylepisoccurredina largepartoftheSkagerrakandintheentireareaofKattegat (Dahletal.,1989;LindahlandDahl,1990).Thisbloomkilled alargenumberofmarinespeciesintheupper20mofthe sublittoralzonealongmostoftheNorwegianSkagerrakcoast (Edvardsenetal.,1988;Granéli etal.,1993;Johannessen andGjøsæter,1990;Underdal etal.,1989),inadditionto 800tonnesoffarmedfish(SkjoldalandDundas,1991).Pos- siblelong-termeffectsandrecoveryintheecosystemcaused by this event have been evaluated by Gjøsæter et al.
(2000). Minor blooms of Chysochromulina polylepis were since detected in 1994 and 1995. In 1998, 2000 and 2001 harmful algal blooms caused by Chattonella spp.
occurredinthearea.The bloomof1998resultedinaloss of350tonnesoffarmedsalmoninadditiontosomewildfish
Figure1 CurrentsofftheJutlandwestcoastandintheSkagerrakandtheKattegat.
Figure2 SurfacedistributionofthepercentageofGermanBightWaterinApril1996,accordingtotheDanishalgorithm.
Figure3 CDOMabsorptioncoefficientaCDOM(375)versussalinityinAprilforallyears1996—2000andalldepths(3454datapoints) fromtheentireareaofinvestigation,withtheDanishtriangleofwatertypes.
Watertypequantification 179
alongthe Danishand Norwegiancoasts, andthe bloomof 2001causedalossofabout1100tonnesoffarmedsalmon.
Occurrence of extremely high nitrate concentrations transported with the Jutland Coastal Current from the GermanBightwasthoughttoprovidefavourableconditions forharmfulalgalbloomsinthearea(Aureetal.,2001;Dahl etal.,2005).Themonitoringcruisestartedin1988bythe Institute of Marine Research has since been repeated in Aprilevery yeartodocument andassess thespringenvir- onmentalconditionsandthepossibilityforharmfulblooms.
Salinityandnutrientobservationsprovidegoodindications ofthewidthoftheJutlandCoastalCurrentalongtheDanish west coast and therefore an indication of the highly annually varying volume transports into the Skagerrak causedby thiscurrent.From1996to2000thecontentof CDOM(ChromophoricDissolvedOrganicMatterorColoured Dissolved Organic Material), originally termed Gelbstoff (yellowsubstance) by Kalle (1938)), was also sampled in the entire area to investigate whether this parameter would be a suitable tracer for the water masses of the Jutland Coastal Current and for quantification of their inter-annualvariations.
ForreferencestothepioneerworksonCDOM,leadingup tothealgorithmsappliedhere,seeHøjerslevetal.(1996) andreferencestherein.OtherrelevantstudiesoftheCDOM content in this area has been presented by Aarup et al.
(1996b), Ferrari and Dowell (1998), Højerslev and Aas
(1998, 2001), Kopelevich etal. (1989), Kowalczuk (1999), Kowalczuk andKaczmarek (1996), Kowalczuket al.(2005, 2006,2010),Lundgren(1976),Stedmonetal.(2000,2010), andWarnocketal.(1999).
2. Material and methods
2.1. DataThedatapresentedinthisarticlewerecollectedonboardr/v G.M.DanneviginthelasthalfofApriluntilMayeachyearfrom 1996 to 2000 in the Skagerrak, the northern part of the Kattegat and along the Danish North Sea coast. The area covered during the cruises is indicated by dots in Fig. 2, butwas insomeof theyearsreduced dueto badweather conditionsortechnicalproblems.Thecruisedateswere(yy.
mm.dd): 1996.04.15—1996.05.04, 1997.04.16—1997.05.05, Table1 Watertypesofthealgorithmsandtheirproperties.
Watertype Salinity aCDOM(375)[m1]
AtlanticWater 35.3 0.05
BalticWater 8.0 0.96
GermanBightWater 31.0 1.50
NorwegianRiverWater 0.0 7.00
Figure4 SurfacedistributionoftheNO3/SiO4ratioinApril1996.
1998.04.08—1998.04.27, 1999.04.15—1999.05.02 and 2000.04.15—2000.04.30. The April cruises were continued through the years, but recordings of CDOM ended in 2000.Thecontributionofthesecruisestotheenvironmental monitoringoftheseareashavebeendescribedbyMorketal.
(2012).ThehydrographicdatacanbeobtainedattheICESData Centre,theWorldOceanDatabaseortheNorwegianMarine Data Centre: http://www.imr.no/forskning/faggrupper/
norsk_marint_datasenter_nmd/en
Salinity was determined by CTD (Neil Brown), and water was sampled at standard depths (0—5—10—20—30— 50—75—100—150—200—300—400—500—600m),exceptinthe shallowerKattegatwherethesamplesweretakenatevery 5m down to a depth of 30m. Inorganic nutrients were analysedbystandardmethodsonboardusinganauto-analy- ser. The content of CDOM, expressed by aCDOM(375), the opticalabsorptioncoefficientat375nmofafiltered water samplerelativetoapurewaterreference,wasmeasuredat Figure5 SurfacedistributionofsalinityinApril1996.
Table2 InfluenceoffourwatertypesontheestimatesbytheDanishandNorwegianalgorithms.ThewatertypesareAtlantic Water(AW),BalticWater(BW),GermanBightWater(GW)andNorwegianRiverWater(NW).aCDOM,xandSxaretheresultingvaluesof theCDOMabsorptioncoefficientandsalinity,respectively.
Input Danishalgorithm Norwegianalgorithm
AW BW GW NW aCDOM,x Sx AW BW GW AW BW NW
[%] [m1] [%] [%]
84 8 8 0 0.24 32.8 84 8 8 91 7 2
84 8 7 1 0.29 32.5 80 9 11 90 7 3
84 8 6 2 0.35 32.2 76 9 15 89 7 3
84 8 4 4 0.46 31.5 68 10 22 88 7 5
84 8 2 6 0.57 30.9 60 12 29 86 8 6
84 8 1 7 0.62 30.6 56 12 32 85 8 7
84 8 0 8 0.68 30.3 52 13 35 84 8 8
Watertypequantification 181
Figure7 CDOMcontentversussalinityinAprilforallyears1996—2000andalldepths(1548datapoints)fromtheKattegat,the southernpartoftheSkagerrakandofftheJutlandwestcoast,withtheDanishtriangleofwatertypes.
Figure6 SurfacedistributionofthepercentageofGermanBightWater(southernpart)andNorwegianRiverWater(northernpart)in April1996,accordingtotheDanishandNorwegianalgorithms.
every station using a Shimadzu UV-VIS 1201 spectrophot- ometer with a 10cm long cuvette. In order to correct for possible residual particles, a baseline correction was applied,based onthevalue at750nm. Thereferencewas
asamplefromaMilliporewaterpurificationsystem.Inthe present study only the observations of salinity, CDOM, nitrate, and nitrate/silicate and nitrate/phosphate ratios willbediscussed.Fromatotalof3454observationsofsalinity
Figure9 MeanverticaldistributionofsalinityinthesectionTorungen—Hirtshals,basedontheobservationsinApril1996—2000.
Figure8 CDOMcontentversussalinityinAprilforallyears1996—2000andalldepths(1906datapoints)fromthenorthernpartofthe Skagerrak,withtheNorwegiantriangleofwatertypes.
Watertypequantification 183
Figure11 MeanverticaldistributionoftheNO3/SiO4ratiointhesectionTorungen—Hirtshals,basedontheobservationsinApril 1996—2000.
Figure10 MeanverticaldistributionoftheCDOMabsorptioncoefficientaCDOM(375)[m1]inthesectionTorungen—Hirtshals,based ontheobservationsinApril1996—2000.
Figure13 MeanverticaldistributionoftheNO3content[mmolL1]inthesectionTorungen—Hirtshals,basedontheobservationsin April1996—2000.
Figure12 MeanverticaldistributionoftheNO3/PO4ratiointhesectionTorungen—Hirtshals,basedontheobservationsinApril 1996—2000.
Watertypequantification 185
andCDOMcontentasmallernumberof2227havebeenused toquantifythespatialdistributionsofwaterfromtheGer- man Bight,theBaltic, the North Sea(Atlantic water) and Norwegianriverswithintheupper50m.
2.2. The Danishalgorithm
The water masses of the Skagerrak and the Kattegat are typicallyamixtureofthree orfourdominant watertypes.
The algorithm presentedby Højerslev etal.(1996) deter- minesthedifferentpercentagesof threewatertypes ata givenlocation anddepth, provided thesalinity and CDOM content at this depth are known. In addition the method requiressomeknowledgeofthelocalhydrographyandofthe watertypesmostlikelytobefound.
The quantification of the contents of the three water typesisbased ontheassumptionthatvolume,salinityand CDOMcontentofawatersampleareconservativeproperties.
Theconcentrationof CDOMis expressedbytheabsorption coefficientaCDOM(375)at375nm.Thesimplealgebraicrela- tionshipsbetweenrelativevolumes,salinitiesandabsorption coefficientsforawatersamplecanbewrittenas:
qBþqAþqG¼1; (1)
SAqAþSBqBþSGqG¼Sx; (2)
aCDOM;AqAþaCDOM;BqBþaCDOM;GqG¼aCDOM;x: (3) Thesymbolsaredefinedas:
qA:relativevolumeofAtlantic(NorthSea)WaterType, qB:relativevolumeofBalticWaterType,
qG:relativevolumeofGermanBightWaterType, SA:salinityoftheAtlanticWaterType,
SB:salinityoftheBalticWaterType, SG:salinityoftheGermanBightWaterType, Sx:observedsalinity,
aCDOM,A: CDOM absorption coefficient of the Atlantic WaterType,
aCDOM,B:CDOMabsorptioncoefficientoftheBalticWater Type,
aCDOM,G:CDOMabsorptioncoefficientoftheGermanBight WaterType,
aCDOM,x:observedCDOMabsorptioncoefficient.
Fromtheseequationsthethreerelativevolumescanbe expressedasfunctionsoftheobservedsalinityandabsorp- tioncoefficient.
qG¼ðSxðaCDOM;BaCDOM;AÞþSBðaCDOM;AaCDOM;xÞ þSAðaCDOM;xaCDOM;BÞÞ=ðSBðaCDOM;AaCDOM;GÞ
þSAðaCDOM;GaCDOM;BÞþSGðaCDOM;BaCDOM;AÞÞ; ð4Þ
Figure14 MeansurfacedistributionofthepercentageofAtlanticWater,accordingtotheDanishandNorwegianalgorithms,based ontheobservationsinApril1996—2000.
qA¼ SxSB
SASB
SGSB
SASB
qG; (5)
qB¼ð1qGqAÞ: (6)
The salinities and absorption coefficients of the water typesusedinthisalgorithmarepresentedinTable1.They areclosetothevaluesusedbyHøjerslevetal.(1996),who usedtheabsorptioncoefficientat380nm.Thespectralslope ofaCDOMwasnotrecordedduringthecruisesanalysedinthis paper,butitsvaluesintheBalticNorthSeatransitionzone have been discussed by Stedmon et al. (2000, 2010), Højerslev and Aas (2001) and Kowalczuk et al. (2005, 2006,2010).
2.3. TheNorwegianalgorithm
WhenthealgorithmofHøjerslevetal.(1996)wasappliedto ourdata,itwassuccessfulindeterminingreasonablevalues for the content of water from the German Bight in the Kattegat,the southern part of the Skagerrak and offthe Jutlandwestcoast.Themodel assumesthattheobserved watermassesarecomposedofthreedominantwatertypes, butthesewatertypesarenotnecessarilythesameinthe
northern part of the Skagerrak.Closer to the Norwegian coastthe algorithmoftenproduced localmaxima ofGer- man Bight Water in the surface layer. An example from 1996 of such maxima is shown in Fig. 2. The content of German Bight Water is gradually reduced from 65—25%
alongtheDanishwest coast.IntheSkagerrakitisfurther reducedtolessthan15%,beforesomelocalmaximaof20%
appear outside the Swedish and Norwegian Skagerrak coasts. When we plotted the observations from all years 1996—2000intoanaCDOM—Sdiagram(Fig.3),wesawthat several points were lying above the triangle constituted by the three water types, indicating that another water type was present, with a higher content of CDOM and a lowersalinitythan theGermanBightWater.
Hereitshouldbenotedthatanothercharacteristicofthe GermanBightanditsriverrunoffsisthelocalnitrate/silicate ratios.Whenthe unitsofthe nitrateandsilicate contents aremmolL1,theratiostypicallyrangefrom20—400(Ryd- bergandAndersson2003,pers.comm.),whichistwoorders ofmagnitudegreaterthanintheNorwegiancoastalwaters, wheretypicalNO3/SiO4ratioslieintherange0—5.Whenwe compared the mentioned maxima of German Bight Water (Fig.2)tothecorrespondinghorizontaldistributionsofthe NO3/SiO4ratio(Fig. 4),nosimilarmaximawerefound,but the salinity distribution (Fig. 5) showed local minima.
Figure15 MeansurfacedistributionofthepercentageofBalticWater,basedontheDanishandNorwegianalgorithms,basedonthe observationsinApril1996—2000.
Watertypequantification 187
Althoughneither nitratenorsilicateconcentrations canbe regardedasconservativepropertiesofawatervolume,this still indicates thatthe apparent maxima of German Bight Water more likely were due to CDOM-rich runoffs from Norwegian rivers. It is also noteworthy that the Danish algorithm was not based on data from these parts of the Skagerrak.
In order to obtain more reliable results closer to the Norwegiancoast,anewalgorithmfortheseareaswasdevel- oped.The watermassesofthe NorwegianCoastalCurrent seemtobedominatedbywatersfromtheNorthSeaandthe Baltic Current together with fjord water and river runoff fromNorway,thelatterdefininganewwatertypethatinthis reporthasbeentermedNorwegianRiverWater.Theproper- tiesof thiswatertypehavemainlybeendetermined from measurementsintheGlomma estuary,but alsofrom mea- surementsintheDramsRiverandtheSkienRiver(Aas1995;
Korsbø1999,Dept.Geophys.,Univ.Oslo,unpublishedinter- nalreports).Allthreearemajorrivers,supplyingfreshwater totheNorthernSkagerrak.ThepropertiesoftheNorwegian River Water are presented in Table 1. The mathematical expressions for the Norwegian and Danish algorithms are similar,exceptthattheNorwegianRiverWater hassubsti- tutedtheGerman BightWater.An algorithmdistinguishing
betweenfourwatertypeswouldrequirerecordingsoffour conservative andcharacteristic properties,and thisfourth propertyremainstobefound.
Atthispointitshouldbenotedthattheconcentrations of water type, estimated by the Danish and Norwegian algorithms,aretentative, becauseerrorswilloccur when morethanthe threeassumedwatertypesarepresent,as illustratedbythesevenrowsofnumbersinTable2.Ifawater volumeinthemiddleofSkagerrakconsistsof84%Atlantic Waterand8%BalticWater,asshownbythefirstandsecond columnandthesevenrowsofTable2,andthecontentsof GermanBightWaterandNorwegianRiverWatervaryinthe ranges8—0%and0—8%,respectively,shownbythethirdand fourthcolumn,theCDOMcontentandsalinitywillvaryas shownby the fifth and sixthcolumn. The percentages of the different water types estimated by the Danish and Norwegianalgorithmsarethenpresentedincolumnsseven totwelve.
InthisexampletheDanishalgorithmclearlyproducesthe greatest deviations from the true percentages of water types.TheinfluenceofNorwegianRiverWaterontheDanish algorithm is to under- and overestimate the contents of Atlantic Water and German Bight Water by up to 30%, whiletheerrorsofBalticWateraresmaller,upto5%.The
Figure16 MeansurfacedistributionofthepercentagesofGermanBightWater(southernpart)andNorwegianRiverWater(northern part),accordingtotheDanishandNorwegianalgorithms,basedontheobservationsinApril1996—2000.
Norwegianalgorithmislessinfluencedbythepresenceof GermanBightWater;AtlanticWaterisoverestimatedbyup to7%,whiletheerrorsofBalticWaterandNorwegianRiver Waterareupto1%and2%,respectively.Morespecificallythe tablealsotellsusthatintheDanishhalfoftheSkagerrak, where the content of Norwegian River water probably is muchlessthan1%,duetothecirculationpattern(Fig.1),the under- and overestimates of Atlantic Water and German BightWaterbytheDanishalgorithmarelikelytobemuch less than 4% (first and second data rowin table).In the NorwegianhalfoftheSkagerrak,ifthecontentofGerman Bight Water is less than 4%, the overestimate of Atlantic WaterbytheNorwegianalgorithmwillalsobelessthan4%
(fourthtoseventh data rowin table).We findthat these errorscanbeaccepted.
TheNorwegianalgorithmwasusedinthenorthernpartof theSkagerrak,andtheDanishalgorithminthesouthernpart, off the Jutland west coast and in the Kattegat. The line separatingthetwoareasisshowninFig.6.Fig.2presented thedistributionofwaterfromtheGermanBightaccordingto the Danish algorithm, and Fig. 6 shows the new situation when both algorithms are applied. The former aCDOM—S diagram(Fig.3)hasbeensubstitutedbytwonewones,Figs.
7and8,wheremostoftheobservationsnowfallinsidethe triangles.InSection3.1.1wewilldeterminehowdeepdown inthewatercolumnthetwoalgorithmsshouldbeapplied.
3. Results and discussion
3.1. Average conditionsMeanvaluesforsalinity,aCDOM(375),NO3/SiO4,NO3/PO4and NO3 inApril have beencalculated for thefive-yearperiod 1996—2000. The values used in Figs. 9—21 are arithmetic means for each point of observation. A vertical section Torungen—Hirtshalswiththedistributionofthemeanvalues (Figs.9—13)willbediscussedinSection3.1.1.Thehorizontal distributionsatthesurfaceofobservedandcalculatedmean properties (Figs. 14—21) will be presented in Section 3.1.2. Finally, the average conditions within the upper 50m will becommented upon in Section3.1.3. From the lasttwosetsofmeanvaluesnewintegratedmeansforthe surfaceandfor the upper50m havebeen calculated and presentedinTable3.
3.1.1. TheTorungen—Hirtshalssection
Theverticaldistributionofpropertiesin thesectionfrom theNorwegian lighthouseTorungen tothe Danishseaport Hirtshals(Fig.1)demonstratesthat mostof thevariation takesplacewithintheupper50m.Salinityhasitsminimum in the Norwegian Coastal Current (upper left corner in Fig. 9), while the CDOM content obtains its maximum in this part of the section (Fig. 10). Below 50m both
Figure17 Meansurfacedistributionofsalinity,basedontheobservationsinApril1996—2000.
Watertypequantification 189
properties show very little variation. The NO3/PO4 ratio exhibitsmaximainthesurfacelayerclosetotheDanishand Norwegiancoasts,andlittlevariationbelow50m(Fig.11).
TheNO3/SiO4 ratioonthe otherhandremainspractically constantwithdepth(Fig.12),whilethenitrateconcentra- tiondecreasesgraduallyfromabout10mmolL1at200m to8mmolL1at50m,andthendecreasesmorerapidlyto 2mmolL1atthesurface(Fig.13).Ouroverallconclusion becomes that while there seems to be clearly different watertypesintheupper50m,conditionsbelowthisdepth are more homogeneous. Although turbulent vertical diffusionmaytransportGermanBightWaterandNorwegian RiverWaterfromthesurfacetothelayerbelow50m,the validityofthetwoalgorithmsatthelowerdepthsisopento question. Accordingly we have restricted the use of the algorithmstotheupper50m.
3.1.2. Surfacedistribution
Fig. 14 presents the mean surfacedistribution of Atlantic Water.Thisisclearlythedominantwatertypeinthesurface layeroftheSkagerrak,withcontentsintherange50—90%.
ThemeancontentofBalticWaterinthesurfaceisabove45%
in the Kattegat, and its concentration is then gradually reduceddowntolessthan5%whenthewaterisdilutedin theSkagerrak(Fig.15).
TheisolinesinFig.16showsurfacecontentsupto55%of GermanBightWater,andupto5%ofNorwegianRiverWater.
Aroundthelinerepresentingtheborderbetweentheareas wherethetwodifferentalgorithmshavebeenapplied,the contentofGermanBightWaterisbetween5%and10%.North ofthisborderthecontentisthenlikelytobelessthan5%due todilution,andourconclusionbecomesthatitsmeancon- centration will probably also be smaller than 5% in the NorwegianCoastalCurrent.Hereitshouldbenotedthatthis conclusiondiffersfromtheresultofAureetal.(1998),who found from observations of salinity and nitrate that the Skagerrak surface water off Arendal contained 21%
ofGermanBightWater inMarch—April,basedon theyears 1980—1995.
Atthesurfacenorthofthealgorithmlinethearithmetic meanvaluesofthedifferentwatertypesareAtlanticWater 74%, Baltic Water 20%, and Norwegian River Water 7%
(Table 3). For the entire area of investigation the mean surfaceconcentrationsbecomeAtlanticWater68%,German Bight Water 15%, Baltic Water 15%, and Norwegian River Water 3% (Table3).Accordingto the discussionin Section 2.2 the uncertainties of the Atlantic Water and German BightWaterestimatesareprobablylessthan4%,whilethe uncertaintiesofBalticWaterandNorwegianRiverWatermay be1—2%.
Figure 18 Mean surface distribution of the CDOM absorption coefficient aCDOM(375) [m1], based on the observations in April1996—2000.
Fig.17presentsthemeansurfacesalinities,rangingfrom lessthan30intheKattegattomorethan34intheSkagerrak.
The values of ay(375) exhibit local maxima of more than 0.7m1intheKattegatandintheGermanBight(Fig.18).
Themeansurfacecontentofnitrateamountstomorethan 15mmolL1intheGermanBight(Fig.19),andthisareaalso exhibits the highest values of the NO3/SiO4 and NO3/PO4
ratios, being up to 35 and 50, respectively (Figs. 20 and
21).Themeansurfacevaluesoftheseratiosfortheentire areaare7and51,respectively(Table3).
Stedmonetal.(2010) usedobservations ofsalinity and CDOM content (absorption coefficient at 300nm) with a three-componentwatertypemodeltoestimatetherelative volumesofAtlanticWaters,GermanBightWatersandBaltic WatersintheBaltic—North Seatransition region,between 68Eand148E.IntheSkagerraktheirobservationsextendedto Figure19 MeansurfacedistributionoftheNO3content[mmolL1],basedontheobservationsinApril1996—2000.
Table3 Meanvaluesandrmsdeviationsoftheinvestigatedquantities,nisthenumberofobservations.
Northernareasurface n=185
Totalareasurface n=443
Totalareaupper50m n=2227
Mean rms Mean rms Mean rms
AtlanticWater[%] 74 16 68 18 79 16
BalticWater[%] 20 16 15 13 8 10
GermanBightWater[%] 15 8 11 7
NorwegianRiverWater[%] 7 9 3 4 1.7 1.8
Salinity 27.2 4.6 29.4 3.6 31.9 3.1
aCDOM(375)[m1] 0.69 0.57 0.59 0.42 0.42 0.27
NO3[mmolL1] 2.4 3.0 4.5 5.3 5.3 5.4
NO3/SiO4 1.2 1.9 7.0 10.5 6.4 8.8
NO3/PO4 55 128 51 91 37 63
Watertypequantification 191
alittlemorethan halfwaybetweenDenmarkandNorway.
TheirdistributionsofAtlanticandBalticWatersresembleour resultsinFigs.14and15,whilethecontentofGermanBight WaterismorelikeourdistributioninFig.2,withvaluesupto 20%inthenorthernpartoftheSkagerrak.
3.1.3. Upper50m
Figs.9—13showedthattherewereverticalgradientsabove 50mdepth,incontrasttothemorehomogeneousconditions below.Accordinglythosepropertiesthathavetheirvertical maximaatthesurface,willobtainsmallermeanvalueswhen alldepthswithintheupper50maretakentogether.These propertiesareBalticWater(8%),GermanBightWater(11%), Norwegian River Water (2%), ay(375) (0.42m1), and the NO3/SiO4(6.4)andNO3/PO4(37)ratios.Thepropertieswith minimaat thesurfaceobtain greatermean values forthe upper 50m. These properties are Atlantic Water (79%), salinity(31.9)andnitrate(5.3mmolL1).
3.2. Annualvariations
Weshallnotdiscussallthedetailsoftheannualvariations inourdataset,butonlymentiontwoofitsmanifestations.
During the years of investigation the salinity distribution showedthat theamount ofwaterfromthe German Bight brought by the Jutland Coastal Current along the Danish westcoastandintotheSkagerrakvariedtoalargeextent.
In both 1996 and 1997 the salinity along the DanishSka- gerrakcoastwashigherthanintheyearsafter.In1996and 1997theamountof nitrateinthe GermanBight waterat thesouthernmostsectionontheDanishNorthSeacoastwas extremely low, while in the years 1998 and 1999 high amounts of nitratewere spreadingfromthis areaand up along the Danish west coast and into the Skagerrak.
The great difference in the nitrate concentrations from yeartoyearatthesouthernmostsectionisassumedtobe due to variation in runoff and flooding on the continent, brought from the German Bight with the Jutland Coastal Current.
Theconsiderableannualvariationofthesurfacepercen- tagesofGermanBightWaterandNorwegianRiverWateris illustrated in Table 3 by the root-mean-square deviations fromthemeanvalues,beingofthesameorderofmagnitude asthemeanvaluesthemselves.Inthetablesomeoftherms deviationsaregreaterthanthecorrespondingmeanvalues, implyingaveryunsymmetricaldistribution ofobservations aroundthemeanvalue.
Figure20 MeansurfacedistributionoftheNO3/SiO4ratio,basedontheobservationsinApril1996—2000.
4. Summary and conclusions
Initiatedbyanumberoftoxicalgalbloomsoccurringalong the Norwegian Skagerrak coast a series of cruises were carried out in the 5-year period 1996—2000. The central issuewastodetermineifnutrient-richwaterfromtheGer- manBightcouldbethecauseoftheseblooms.Byusingan algorithm for water massanalysis developed by Højerslev etal.(1996),reasonableresultswereobtainedintheKatte- gat,offtheJutlandwestcoastandinthesouthernpartofthe Skagerrak. However, close to the Norwegian coast in the SkagerraktheapparentdistributionofGermanBightWater became improbable, yielding local maxima outside the fjords.Consequentlyanalternativealgorithm,withNorwe- gianRiverWatertypesubstitutedfortheGermanBightWater type,wasconstructedfortheseparts.Theborderbetween the areas for the two algorithms (the line in Fig. 6) is approximatelyconsistentwiththelimitsoftheDanisharea ofinvestigation.
TheresultofouranalysisinTable3isthatnorthofthe borderline the mean surface contents of Atlantic Water, Baltic Waterand NorwegianRiverWater are74%, 20%and 7%,respectively.Sincethisresultisbasedontheassumption
thatGermanBightWaterisnotoneofthethreewatertypes oftheappliedalgorithm, thesenumbers mayof coursebe questioned, but our analysis also indicates that the mean contentofGermanBightWaterinthiscurrent,regardlessof theappliedalgorithm, willbelessthan5%.For theentire area of investigation the mean surface contents become AtlanticWater68%,GermanBightWater 15%,BalticWater 15%,andNorwegianRiverWater3%.Foralldepthswithinthe upper50m takentogether the Atlantic Water constitutes 79%,GermanBightWater11%,BalticWater8%,andNorwe- gianRiverWater 2%. The values inTable 3 are arithmetic means,andtheiruncertaintieshavebeen estimatedtobe lessthan4%fortheAtlanticandGermanBightWaters,and 1—2%orlessfortheBalticandNorwegianRiverWaters.
Inconclusion,withinourinvestigatedareaAtlanticWater isthedominantwatertype(70—80%),whilethecontentsof GermanBightWaterandBalticWaterareanorderofmagni- tude smaller.The content of Norwegian RiverWater is on averagesmallerthanthecontentofBalticWaterbyafactor of3—5(Figs.14—16,Table3).
Wehopethepresentanalysismayserveasacontribution tofurtherinvestigationsofthemixingofwatertypesinthis area.
Figure21 MeansurfacedistributionoftheNO3/PO4ratio,basedontheobservationsinApril1996—2000.
Watertypequantification 193
Acknowledgements
We are thankful to D.S. Danielssen and E. Svendsen for initiatingtheanalysispresentedinthispaper,andforprovid- ingdataandfunding.Wealsowishtothankourtworeviewers forconstructivecomments.
References
Aarup,T.,Holt,N.,Højerslev,N.K.,1996a.Opticalmeasurements in the North Sea—Baltic Sea transition zone. II. Water massclassificationalongtheJutlandwestcoast fromsalinity andspectral irradiance measurements. Cont.Shelf Res. 16, 1343—1353.
Aarup,T.,Holt,N.,Højerslev,N.K.,1996b.Opticalmeasurementsin theNorthSea—BalticSeatransitionzone.III.Statisticalanalysis ofbio-opticaldatafromtheEasternNorthSea,theSkagerrakand theKattegat.Cont.ShelfRes.16,1355—1377.
Anon.,1993.NorthSeaSubregion8.AssessmentReport1993.North SeaTask Force. StatePollutionControl Authority(SFT),Oslo, Norway,77.
Aure,J., Danielssen,D.S.,Skagen,M.,Svendsen,E.,Søiland,H., Pettersson,L.,2001.EnvironmentalconditionsduringtheChat- tonellabloomintheNorthSeaandSkagerrakinMay1998.In:
Hallegraeff,G.,etal.(Eds.),Proc.NinthInt.Conf.HarmfulAlgal Blooms2000,Hobart,Australia,7—11Feb.2000.IOCUNESCO, 82—85.
Aure,J.,Danielssen,D.,Svendsen,E.,1998.TheoriginofSkagerrak coastalwateroffArendalin relationtovariations innutrient concentrations.ICESJ.Mar.Sci.55,610—619.
Aure,J.,Sætre,R.,1981.WindeffectsontheSkagerrakoutflow.In:
Sætre,R.,Mork,M.(Eds.),Proc.Univ.Bergen.TheNorwegian CoastalCurrent,263—293.
Dahl,E., Bagøien, E., Edvardsen, B., Stenseth, N.C., 2005. The dynamicsofChrysochromulinaspeciesintheSkagerrakinrelation toenvironmentalconditions.J.SeaRes.54,15—24.
Dahl,E.,Lindahl,O.,Paasche,E.,Throndsen,J.,1989.TheChry- sochromulina polylepis bloom in Scandinavian waters during spring 1988. In: Cosper, E.M., Bricelj, V.M., Carpenter, E.J.
(Eds.),NovelPhytoplanktonBlooms.Springer-Verlag,NewYork, 383—405.
Danielssen,D.S.,Edler,L.,Fonselius,S.,Hernroth,L.,Ostrowski,M., Svendsen,E.,Talpsepp,L.,1997.Oceanographicvariabilityinthe SkagerrakandnorthernKattegat,May—June,1990.ICESJ.Mar.
Sci.54,753—773.
Edvardsen,B.,Anstensrud,M.,Christie,H.,Fredriksen,S.,Gray,J.
S., Leinaas,H.P., Schram,T., Saanum,I., Winther-Larsen, T., 1988. Rapport fra undersøkelse om effekter på bunnlevende organismerogstrandlevendefiskpåkyststrekningenLangesund -TvedestrandetteroppblomstringenavChrysochromulinapoly- lepis.Univ.Oslo,1—48(inNorwegian).
Ferrari,G.M.,Dowell,M.D.,1998.CDOMabsorptioncharacteristics withrelationtofluorescenceandsalinityincoastalareasofthe SouthernBalticSea.Est.Coast.ShelfSci.47,91—105.
Gjøsæter,J.,Lekve,K., Stenseth,N.C., Leinaas,H.P.,Christie, H.,Dahl,E.,Danielssen,D.S.,Edvardsen,B.,Olsgard,F.,Oug, E.,Paasche, E., 2000.A long-term perspectiveon theChry- sochromulinabloomontheNorwegianSkagerrakcoast1988:a catastrophe or an innocent incident? Mar. Ecol. Prog. 207, 201—218.
Granéli,E.,Paasche,E.,Maestrini,S.Y.,1993.Threeyearsafter theChrysochromulinapolylepisbloominScandinavianwaters in1988:someconclusionsofrecent researchmonitoring. In:
Smayda,T.J., Shimizu,Y.(Eds.),Toxicphytoplantkonblooms in the sea. Proc. 5th Int. Conf. Toxic Phyt., Newport, R.I., 28October—1November1991.Elsevier,Amsterdam,25—32.
Gustavsson,B.,Stigebrandt,A.,1996.Dynamicsofthefreshwater- influencedlayersintheSkagerrak.J.SeaRes.35,39—53.
Højerslev,N.K.,Aas,E.,1998.SpectrallightabsorptionbyGelbstoff incoastalwatersdisplayinghighlydifferentconcentrations.In:
Conferencepapers,OceanOpticsXIV,Kailua-Kona,Hawaii,No- vember10—13,vol.3,10—13.
Højerslev,N.K.,Aas,E.,2001.Spectrallightabsorptionbyyellow substance in theKattegat— Skagerrakarea. Oceanologia43, 39—60.
Højerslev,N.K.,Holt,N.,Aarup,T.,1996.Opticalmeasurementsin theNorthSea-BalticSeatransitionzone.I.Ontheoriginofdeep waterintheKattegat.Cont.ShelfRes.16,1329—1342.
ICES,1983.FlushingtimesoftheNorthSea.Coop.Res.Rep123,159.
Johannessen,T.,Gjøsæter,J.,1990.AlgeoppblomstringeniSkager- rak i mai 1988 —ettervirkninger på fiskog bunnfauna langs Sørlandskysten.FlødevigenMeldinger6,1—68,(inNorwegian).
Kalle,K.,1938.ZumProblemderMeereswasserfarbe.Ann.Hydrol.
Mar.Mitt.66,1—13.
Kopelevich, O.V., Lyutsarev, S.V., Rodionov, V.V., 1989. Spectral light absorptionby yellow substance inseawater. Oceanology 29,305—309.
Kowalczuk,P.,1999.Seasonalvariabilityofyellowsubstanceabsorp- tioninthesurfacelayeroftheBalticSea.J.Geophys.Res.104, 30047—30058.
Kowalczuk,P.,Kaczmarek,S.,1996.Analysisoftemporalandspatial variabilityof`yellowsubstance0absorptioninthesouthernBal- tic.Oceanologia38,3—32.
Kowalczuk, P., Olszewski, J., Darecki, M., Kaczmarek, S., 2005.
Empiricalrelationshipsbetweencoloureddissolvedorganicmat- ter(CDOM)absorptionandapparentopticalpropertiesinBaltic Seawaters.Int.J.RemoteSens.26,345—370.
Kowalczuk,P.,Stedmon,C.A.,Markager,S.,2006.Modelingabsorp- tionbyCDOMintheBalticSeafromseason,salinityandchloro- phyll.Mar.Chem.101,1—11.
Kowalczuk,P.,Zablocka,M.,Sagan,S.,Kulinski,K.,2010.Fluores- cencemeasuredinsituasaproxyofCDOMconcentrationinthe BalticSea.Oceanologia52,431—471.
Lindahl,O.,Dahl,E.,1990.OnthedevelopmentoftheChrysochro- mulinapolylepisbloomintheSkagerrakinMay—June,1988.In:
Granéli,E.,Sundström,B.,Edler,L.,Anderson,D.M.(Eds.),Toxic MarinePhytoplankton.Elsevier,NewYork,189—194.
Lundgren, B., 1976. Spectraltransmittancemeasurements inthe Baltic Rep.Dept.Phys. Oceanogr.,vol.30.Univ.Copenhagen, Copenhagen,38.
Mork, K.A., Naustvoll, L.J., Albretsen, J., Ingvaldsen, R., 2012.
Utredningavprogramforovervåkingavklimaendringseffekteri norskekyst-oghavområder.Fiskenoghavet5,76.
NorthSeaTaskForce,1993.NorthSeaQualityStatusReport1993.
OsloandParisCommissions,London,Olsen&Olsen,Fredensborg, Denmark,1—132.
Rodhe,J.,1987.Thelarge-scalecirculationintheSkagerrak;inter- pretationofsomeobservations.Tellus39A,245—253.
Rodhe,J.,1992.StudiesofcurrentsandmixingintheSkagerrak.
Doctoralthesis.Dept. Oceanogr., Univ.Gothenburg,Gothen- burg,127.
Rodhe,J.,1996.Onthedynamicsofthelarge-scalecirculationofthe Skagerrak.J.SeaRes.35,9—21.
Skjoldal,H.R.,Dundas, I., 1991.The Chrysochromulinapolylepis bloom inthe Skagerrakand theKattegat in May—June 1988:
environmental conditions, possible causes, and effects. ICES Coop.Res.Rep.1—59.
Stedmon,C.A.,Markager,S.,Kaas,H.,2000.Opticalpropertiesand signaturesofchromophoricdissolvedorganicmatter(CDOM)in Danishcoastalwaters.Est.Coast.ShelfSci.51,267—278.
Stedmon,C.A.,Osburn,C.L.,Kragh,T.,2010.Tracingwatermass mixingintheBaltic—NorthSeatransitionzoneusingtheoptical propertiesofcoloureddissolvedorganicmatter.Est.Coast.Shelf Sci.87,156—162.
Svansson, A., 1975. Physical and chemical oceanography of the SkagerrakandtheKattegat.I.OpenSeaConditions.Fish.Board Swed.Inst.Mar.Res.1,1—88.
Underdal,B.,Skulberg,O.M.,Dahl,E.,Aune,T.,1989.Disastrous bloom of Chrysochromulina polylepis (Prymnesiophyceae) in
Norwegian coastal waters 1988 — mortality in marine biota.
Ambio18,265—270.
Warnock,R.E.,Gieskes,W.W.C.,vanLaar,S.,1999.Regionaland seasonaldifferencesinlightabsorptionbyyellowsubstanceinthe SouthernBightoftheNorthSea.J.SeaRes.42,169—178.
Watertypequantification 195
