Model study on horizontal variability of nutrient N/P ratio in the Baltic Sea and its impacts on ...

18 Pages·2012·5.01 MB·English

OceanSci.Discuss.,9,385–419,2012 OceanScience Dis
www.ocean-sci-discuss.net/9/385/2012/ c
Discussions u
doi:10.5194/osd-9-385-2012 s
s
©Author(s)2012.CCAttribution3.0License. io
n
P
a
Thisdiscussionpaperis/hasbeenunderreviewforthejournalOceanScience(OS). pe
r
PleaserefertothecorrespondingfinalpaperinOSifavailable.
|
D
Model study on horizontal variability of is
c
u
s
nutrient N/P ratio in the Baltic Sea and its sio
n
P
impacts on primary production, nitrogen a
p
e
r
fixation and nutrient limitation
|
D
Z.Wan1,H.Bi2,J.She1,M.Maar3,andL.Jonasson1 isc
u
s
1CentreforOceanandIce,DanishMeteorologicalInstitute,Lyngbyvej100,2100 sio
Copenhagen,Denmark n
P
2ChesapeakeBiologicalLaboratory,P.O.Box38,Solomons,MD20688,USA a
p
3AarhusUniversity,Dep.ofBioscience,Frederiksborgvej399,P.O.Box358,4000Roskilde, er
Denmark
|
Received:12January2012–Accepted:17January2012–Published:31January2012 D
is
Correspondenceto:Z.Wan([email protected]) c
u
s
PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion. sio
n
P
a
p
e
r
385 |
Abstract D
is
c
u
Theanalysisofmeasurednutrientconcentrationssuggeststhattheratioofdissolved ss
inorganicnitrogen(DIN)alterationbeforeandafterspringbloomsrelativetothealter- ion
ation of dissolved inorganic phosphorus (DIP) remains quite constant over the years P
a
5 (2000∼2009). ThisratiodiffersfromtheRedfieldratioandvariesfrom6.6:1to41.5:1 per
across basins within the Baltic Sea. If the found N/P ratios are indicators of phy-
toplankton stoichiometry, this would affect nutrient cycles in ecosystem models. We |
thereforetestedtheeffectsofusinghorizontallyvariableN/Pratioinsteadoffixedratio D
is
(N/P=10:1or16:1)onphytoplanktonuptakeandremineralizationina3-Dphysical- cu
s
10 biogeochemicalcoupledmodelERGOM.ThemodelresultsusingthevariableN/Pratio sio
show systematical improvements in model performance in comparison with the fixed n
ratios. In addition, variable N/P ratios greatly affected the model estimated primary Pa
p
production,nitrogenfixationandnutrientlimitation,whichhighlightstheimportanceof e
r
usinganaccurateN/Pratio.
|
D
is
15 1 Introduction cu
s
s
Theaveragephytoplanktonstoichiometryisoneofimportantfeaturesofecosystems, io
n
whichdeterminesmultiplenutrientcycles(Redfield,1934and1958;Falkowski,2000; P
a
Klausmeier et al., 2004a,b, 2007; Weber and Deutsch, 2010). Redfield (1934) noted pe
r
theaveragednitrogentophosphorusratioinplanktonisconsistentwithdeepoceanic
water. This averaged ratio has long been acknowledged as the “Redfield ratio” for |
20
phytoplanktonintheworld’soceans. However, laboratorystudiessuggestthatphyto- D
is
planktonareflexibleintheiroverallstoichiometry, oftenmatchingtheirnutrientsupply c
u
atlowgrowthrates(Klausmeieretal.,2004a,b). Moreover,thisclassicelementalratio ss
io
has also long been argued against in regional oceans, and recently even been chal- n
P
25 lengedforlargeareasintheSouthernOcean(WeberandDeutsch,2010). Infact,any a
p
fixed elemental ratio, Redfield or otherwise, is an oversimplification when applied to e
r
386 |
realecosystems. Manystudieshaveshownthatthephytoplanktonstoichiometrycan D
is
differamongspeciesandlifestages(MinsterandBoulahdid,1987;Arrigoetal.,1999; c
u
s
KressandHerut,2001;Wongetal.,2002;AndersonandPondaven,2003). Moreover, s
io
Klausmeieretal.(2004b)hypothesizesthatphytoplanktonstoichiometryisinprinciple n
5 a dynamical trade-off according to ecological conditions between uptake machinery Pap
andassemblymachinery. EachtypeofmachineryhasitsownN/Pratio. e
r
In the Baltic Sea, both modeling studies (Neumann, 2000; Edelvang et al., 2005;
|
Neumann and Schernewski, 2008; Savchuk et al., 2008; Eilola et al., 2009; Maar
D
etal.,2011;Meieretal.,2011)andobservation-basedstudies(OsterrohtandThomas, is
c
10 2000; Savchuk, 2005) often assume that the average phytoplankton stoichiometry is us
consistentwithRedfieldratios. OnlyafewstudieshaveconsideredNon-Redfieldratios sio
(Shaffer, 1987; Larsson et al., 2001; Kuznetsov et al., 2008). In our previous study n
P
(Wan et al., 2011), we documented that the N/P ratio of nutrient alterations before ap
e
and after spring blooms could be a good indicator to the N/P ratio of phytoplankton r
15 stoichiometryintheBalticproper. Inaddition,wedemonstratedthatRedfieldN/Pratio |
tendstobetoohighfortheentireBalticSeaandthatanotherfixedratio10:1ismore
D
appropriateformostareas. is
c
In the present study, we test with a hypothesis that in each Baltic basin, the long us
s
term mean N/P ratio of nutrient alterations before and after spring blooms can be an io
n
20 indicator for the N/P ratio of phytoplankton stoichiometry. A horizontally interpolated P
a
2-D distribution of N/P ratio based on observed data from nine monitoring stations in p
e
differentbasinsisimplementedina3-Dphysical-biogeochemicalcoupledmodeltotest r
this hypothesis. As the primary production, nitrogen fixation and nutrient limitation is |
principally related to phytoplankton stoichiometry, this study also examines the differ- D
25 encebetweenestimatesbasedonafixedN/PratioversusthehorizontallyvariableN/P isc
u
ratio. s
s
io
n
P
a
p
e
r
387 |
2 Dataandmethods D
is
c
u
Thisstudyisanextensionofthepreviousworkofourgroup. Fordetaileddescription ss
onmodels,seeWanetal.(2011). Thefollowingprovidesbriefdescriptionondataand ion
modelsandfocusesondifferentscenarios’comparison. P
a
p
e
r
2.1 Data
5
|
The climatology of temperature and salinity configuring the initial and lateral bound- D
ary conditions for the physical model are from the World Ocean Atlas 2001 (WOA01, is
c
u
Conkright et al., 2002). The products by the operational weather model High Res- s
s
olution Limited Area Model of the Danish Meteorological Institute (DMI) are used to io
n
provide atmospheric drivers for the physical model (She et al., 2007a, 2007b). The P
10 a
dailyriverrunoffandbioloadingsarefromanoperationalhydrologicalmodelHBVrun pe
r
bytheSwedishMeteorologicalHydrologicalInstitute(SMHI)(Bergstro¨m,1976,1992)
in combination with observations from the German Bundesamt fu¨r Seeschifffahrt und |
Hydrographie (BSH) and climatology. The initial fields of biogeochemical model use D
is
15 the observed data downloaded from the website of the International Council for the cu
ExplorationoftheSea(ICES)(http://www.ices.dk/indexfla.asp)incombinationwithcli- ss
io
matologiesofWOA01. n
P
a
2.2 Physicalmodel pe
r
The physical model in this study is the HIROMB model, which was originally de- |
20 veloped by BSH (Dick et al., 2001) and further developed and maintained in DMI D
(https://hbmsvn.dmi.dk/). HIROMBisbasedontheprimitivegeophysicalfluiddynam- isc
u
icsequationsfortheconservationsofvolume,momentum,salinity,andheat. Thether- s
s
modynamicmodelcomponentcanresolvethereflectionandabsorptionofshortwave io
n
radiationbytheseabedinshallowwaterzones. Theimpactsfromwind, atmospheric P
a
25 pressure, air temperature, humidity, evaporation-precipitation, and cloud cover are all pe
r
388 |
considered and parameterized as surface boundary conditions. The tidal water level, D
is
the surge water level and the monthly climatology of temperature and salinity are im- c
u
posedastheouterlateralconditions,andtheriverrunoffsastheinnerlateralcondition. ss
The model grids cover 48◦330N to 65◦510N and 4◦050W to 30◦150E with a spatial ion
5 resolution of 60 latitude and 100 longitude, while a nested fine grid with one sixth of Pap
the coarse resolution coversthe Danish Strait in order to resolve thewater exchange e
r
throughthenarrowsillsbetweentheNorthSeaandtheBalticSea. Thecoarsegridhas
|
50verticallayersandthefinegridhas52verticallayers. Themodeldomainincludes
both the Baltic Sea and the North Sea in order to provide a sufficient transition to Dis
10 counteract the effects from open boundaries, but the focused area is only the Baltic cus
Sea(Fig.1). sio
n
P
2.3 Biogeochemicalmodel a
p
e
r
The biogeochemical model is ERGOM which was developed by Neumann (2000). It
|
has 9 state variables, including dissolved inorganic nutrients: ammonium, nitrate and
phosphate; three autotrophic functional groups: diatoms, flagellates and cyanobac- D
15 is
teria; a bulk zooplankton community; a detritus pool; and dissolved oxygen (DO). cu
s
Themodelmathematicallydescribestheprocessesofphotosynthesis,nutrientuptake, s
io
growth, grazing, digestion, respiration, excretion, mortality, remineralization, nitrogen n
P
fixation,nitrification,anddenitrification. Thismodelisnitrogen-based,andphosphorus a
p
20 was originally coupled to nitrogen via the Redfield ratio, but the N/P ratio is variable er
in this study. The parameters are same as Wan et al. (2011), which are based on
|
Neumannetal.(2002)withonlyafewadjustedvalues.
D
is
2.4 Simulationscenarios cu
s
s
io
Foursimulationscenariosareimplemented(Table1)whereeachscenariousesidenti- n
P
25 calN/Pratiosfornutrientuptakeandremineralization. Twoscenariosusethehorizon- a
p
tally variable N/P ratios, and the other two use fixed N/P ratios. In the first scenario e
r
389 |
(V1),theN/Pratioisassumedconsistentwiththemeanobservedratioin2000–2009. Dis
ThemeanobservedratioiscalculatedasthemeanvalueofsurfaceDINconcentration cu
s
alteration before and after spring blooms relative to that of DIP (Fig. 2 and Table 2). s
io
1 February is selected as a typical winter day before spring blooms, and 1 June as n
P
a typical summer day after spring blooms. The values on 1 February and 1 June are a
5 p
interpolatedfromthemonthlyobserveddataforeachyear. AsScenarioV1,thedistri- er
butionofN/Pratioisinterpolatedwiththeratiosofseasonaldeviationsbasedondata
|
from the well observed nine stations (Fig. 3a and Table 2). As Scenario V2, the N/P
D
ratiodistributionisinterpolatedwithobservedvaluesfromonlysixstationsandthead- is
c
justed values: 16:1, 18:1 and 20:1 for Stations A, G and H, respectively (Fig. 3b). u
10 s
s
Scenario NP10 uses a fixed N/P ratio of 10:1 according to Wan et al. (2011) and io
n
ScenarioNP16usesthefixedRedfieldN/Pratioof16:1. P
a
p
2.5 Efficiencyofnutrientlimitation er
|
Nitrogen and phosphorus were reported as alternative limiting nutrients in the Baltic
D
15 Sea(Grane´lietal.,1990;Kivietal.,1993;Zweifeletal.,1993;Danielssonetal.,2008). is
We define Efficiency of Nutrient Limitation (ENL) as the ratio between the duration cu
s
of one nutrient limiting and the total photosynthesis duration. The limiting nutrient is sio
determined when whose value of limitation function is the smallest among all limiting n
P
nutrients. ThelimitationfunctionisamodifiedMichaelis-Mentenformulawithsquared a
p
e
20 terms(Neumannetal.,2002), r
x2 |
Y(α,x)= ,
α2+x2 Dis
c
whereαisMichaelis-MentenconstantforspecificalgaespeciesandxiseitherDINor u
s
s
DIPconcentration. Astherearethreeautotrophicgroups,thefunctionalgroupwiththe io
n
biggestgrowthrateisselectedtocomparethevaluesoflimitationfunctions. Theratio P
a
25 of thickness between the layer where nitrogen (phosphorus) is limiting and the entire pe
photosynthesis-activelayeriscalculatedastheaveragenutrientlimitationfornitrogen r
390 |
(phosphorus) for one water column. The average nutrient limitation is considered as D
is
atwo-dimensionaldistributionwithtemporalvariation. Weconsiderawatercolumnis c
u
s
eithernitrogenlimitingorphosphoruslimiting,noothernutrientsareconsideredinthis s
io
study. ENLisnormalizedwiththepreliminary“ENL”fornitrogenandphosphorus. The n
P
5 ENL is in default for nitrogen, whose value ranges between 0 and 1. The normalized ap
“ENL”forphosphorusequals1-ENL. e
r
|
3 Results D
is
c
u
3.1 Modelvalidation s
s
io
n
Asthesamemodelwasvalidatedrobustinresolvingtheseasonalvariabilityinsurface P
a
10 andtheverticalprofilebyWanetal.(2011), inthepresentstudywearetovalidateif pe
the model results using horizontally variable N/P ratio (Scenario V2) are better than r
themodelresultsusinganoptimalfixedN/Pratio(ScenarioNP10)basedonthecon- |
sistency between the simulated results and the observation data. Three aspects of D
modelperformanceareexamined. isc
u
s
s
15 3.1.1 Improvementsofseasonalvariabilityinsurfacewatersindifferentbasins ion
P
Nine stations J-R (Fig. 1) are selected to examine the model performance of sea- ap
e
sonal variability of nutrients and chlorophyll-a in surface layer. The model results for r
chlorophyll-aareimprovedatstationsintheBalticproper(Fig.4b–e),whencomparing |
Scenario V2 to Scenario NP10. The summer bloom peaks are no longer higher than
D
thespringbloompeaks. Thisisclosertoobservations. Chlorophyll-ainScenarioV2is is
20 c
ingeneralslightlyhigherthaninScenarioNP10forstationsoutsideoftheBalticproper us
s
(Fig.4a,g,h,i). ThewinterDINconcentrationsinScenarioV2arestable,noincreasing io
n
trend like in scenario NP10 (Fig. 5b–e). The summer DIN surplus in Scenario NP10 P
a
isinconsistentwithobservationsinSkagerrakandGulfofFinland,buttheresultsfrom p
e
r
391 |
ScenarioV2areimprovedanddonotshowsuchinconsistency(Fig.5a,i). Inscenario D
is
NP10,onepartofDIPconsumptionduringspringbloomsandanotherpartduringlate cu
s
summer blooms are inconsistent with observations, but in Scenario V2, the DIP con- s
io
sumptiontendstobecontinuousforthestationsintheBalticproper, moreconsistent n
P
5 withtheobservations(Fig.6b–e). TherearesomeDIPsurplusinsummerinSkager- ap
rak, Bothnian Sea and Gulf of Finland in Scenario V2, and observations also show e
r
occasionalDIPsurplusinsummer(Fig.6a,g,i),butthemodelDIPsurplusinsummer
|
islargerthantheobservedone.
D
is
3.1.2 Improvementsintemporalvariationandverticaldistributions cu
s
s
io
10 Theresultsapplyingthecomprehensivecomparisonscheme(Wanetal.,2011)show n
theoveralleffectsofScenarioV2versusNP10(Fig.7). NotethatthevaluesinFig.7 Pa
p
arethemeansovertheobservedgridpointsinapresetmeshforstatisticpurposein- e
r
stead. If observations are roughly evenly located, the values in Fig. 7 can approach
|
to physical means. It is a feature of the comprehensive comparison scheme, aiming
toreflecttheseasonalpatternandtheverticalprofileofmodelbiases. Theobserved D
15 is
seasonal variability in Chlorophyll-a is closer to that in Scenario V2 than in Scenario cu
s
NP10 (Fig. 7a). The improvement is reflected as a correction to the excessive sum- s
io
mer peak height. The model DIN is closer to the observation in Scenario V2 than in n
P
Scenario NP10 (Fig. 7b). The model DIN in Scenario NP10 intends to increase over a
p
20 time. ThemodelDIPisalsoclosertotheobservationinScenarioV2thaninScenario er
NP10 (Fig. 7c). The vertical profile of model results has improvements in general for
|
Chlorophyll-a, DIN and DIP in Scenario V2 in contrast to Scenario NP10 (Fig. 7d–f).
D
The improvements happen in upper 20m for Chlorophyll-a, below 40m for DIN and is
c
below 80m for DIP. Even though the improvements are not very large, the improving u
s
25 patternissystematic. sio
n
P
a
p
e
r
392 |
3.1.3 Improvementsinstatistics D
is
c
u
The statistic results clearly show the model performs better in Scenario V2 than in ss
Scenario NP10 (Table 3). If we define the bias as the difference between the model ion
results and the observed data, the percentage biases decrease from scenario NP10 Pa
p
5 to V2: 32% to 28% for Chlorophyll-a, 20% to 11% for DIN, and −10% to −6% for er
DIP. The correlation coefficients increase from 0.50 to 0.52 for DIN and from 0.89 to
|
0.91forDIP,butnochangeforChlorophyll-a. ThestandarddeviationofmodelDIPfor
scenarioV2largerthanthatforNP10,i.e.,themodelDIPforscenarioV2isclosertothe D
is
observeddata,butnochangesforChlorophyll-aandDIN.Althoughtheimprovements cu
s
in statistics are not very large, but the signal is clear that a horizontally variable N/P s
10 io
ratioisbetteranoptimizedfixedN/Pratio(Wanetal.,2011). n
P
a
p
3.2 Primaryproduction e
r
The seasonal mean of primary production in Scenario V2 mostly ranges from |
20gCm−2yr−1 to 140gCm−2yr−1 (Fig. 8a). The extreme values can be up to D
15 200gCm−2yr−1 insomecoastalareasandestuaries. Thegeneralpatternofprimary iscu
production is higher in the south than in the north, in coastal regions than in offshore ss
io
regions, in estuaries than outside of estuaries. The total model primary production n
for the entire Baltic Sea varies seasonally, from 10ktonCday−1 to 200ktonCday−1 Pa
p
(Fig.8d). Theseasonalpatternisfeaturedwiththespringpeakandthesummerpeak. e
r
Thespringpeakvaluearound200ktonCday−1 ismuchhigherthanthesummerpeak
20
valuearound110ktonCday−1. |
In contrast to Scenario V2, the excessive primary production in Scenario NP10 in- D
is
creases from 60◦N northward up to 10gCm−2yr−1 at the south end of the Bothnian cu
Bayandfurtherupto30gCm−2yr−1 atthenortherntipoftheBothnianBay(Fig.8b). ss
io
Italsoincreaseseastwarduptoaround30gCm−2yr−1intheGulfofFinland. Thepri- n
25 mary production in Scenario NP10 is mostly 10∼20gCm−2yr−1 higher than in the Pa
p
Scenario V2 in the transition zone between the Baltic Sea and the North Sea. It er
393 |
is only 0∼10gCm−2yr−1 lower in Scenario NP10 than in Scenario V2 in the Baltic Dis
proper. The primary production in Scenario NP16 is mostly 0∼20gCm−2yr−1 lower cu
s
than Scenario V2, but can be up to 40∼50gCm−2yr−1 lower in the Swedish coast sio
of the Gotland Sea, the Gulf of Riga and other eastern coast of the Gotland Sea n
P
5 (Fig. 8c). The primary production in Scenario NP16 is seldom higher than in Sce- ap
nario V2, only 0∼10gCm−2yr−1 higher in the Russian part of Gulf of Finland and er
10∼20gCm−2yr−1 higherwestof10◦E. |
Theseasonalpatternofmeanprimaryproductionaveragedovertheentiredomain
D
in Scenario NP10 is quite similar to Scenario V2, only around 20gCm−2yr−1 higher is
c
10 during summer blooms in Scenario NP10 than in Scenario V2 (Fig. 8d). The mean us
primaryproductioninScenarioNP16is10∼60gCm−2yr−1 smallerthaninScenario sio
V2duringspringbloomsandinearlysummer,butaround20gCm−2yr−1higherduring nP
a
latesummerblooms. p
e
r
3.3 Nitrogenfixation |
D
15 In2007∼2008,thenitrogenfixationinScenarioV2occursmainlyinthesouthernBaltic is
c
SeaandattheentranceoftheGulfofFinland(Fig.9a). ThenitrogenfixationinSce- u
s
narioV2ismostlylessthan40mmolNm−2yr−1,andtheintegratednitrogenfixationis sio
nomorethan1.0ktonNday−1(Fig.9d). ThenitrogenfixationinScenarioNP10occurs n
P
inthewholeSouthernBalticSeaandmostoftheBalticproper,anditismostlyhigher ap
than40mmolNm−2yr−1 intheSouthernBalticSea(Fig.9b). Thenitrogenfixationin er
20
ScenarioNP16occursinthewholesouthernBalticSea,theBalticproper,mostofthe |
GulfofFinlandandtheeasternBothnianSea(Fig.9c). Itshorizontaldistributiontends
D
increasing southeastwards and southwards. The highest nitrogen fixation appears in is
c
the southern coast of the Baltic Proper, up to 200mmolNm−2yr−1. The total nitro- us
s
25 genfixationisseveraltimeshigherinScenariosNP16andNP10thaninScenarioV2 ion
(Fig.9d). P
a
p
e
r
394 |
3.4 Nutrientlimitation D
is
c
u
ThemodelresultsshowthattheprimaryproductionintheBalticSeaismostlyN-limited ss
(ENL>0.5,ENLisdefinedinSect.2.5),buttheextentoflimitationvarieshorizontally ion
anddiffersacrossscenarios. InScenarioV2,itisheavilyN-limited(mostlyENL>0.8) Pa
p
5 in the Gotland Sea and in the Skagerrak and the Kattegat (Fig. 10a). The P-limited er
(ENL<0.5)areaisquitelimited,thesevereP-limited(ENL<0.2)appearsinthewest-
|
ernBothnianSeaandintheGulfofRiga,otherslightP-limited(0.3<ENL<0.5)areas
include the southern coast of the Baltic proper, the eastern tip of the Gulf of Finland D
is
and the northern tip of the Bothnian Bay. In Scenario NP10, the Baltic proper, the cu
s
Southern Baltic Sea, the entrance of the Gulf of Finland and the entrance of the Gulf s
10 io
ofRigaareallN-limited(ENL>0.7), buttheBothnianSea, theBothnianBay, theSk- n
P
agerrak and the Russian part of Gulf of Finland are P-limited (ENL<0.1) (Fig. 10b). a
p
InScenarioNP16,theBalticSeaisN-limitedingeneral,onlyafewofcoastalregions e
r
are P-limited (Fig. 10c). The Skagerrak, the Kattegat and the connection between
|
the Bothnian Sea and the Bothnian Bay are heavily N-limited (ENL>0.9). The Baltic
15
D
proper is media N-limited (0.7<ENL<0.8), and the southern Baltic Sea is weak N- is
c
limited(0.5<ENL<0.7). AstheentireBalticSeaisconcerned,thenutrientcondition u
s
for the primary production seasons (except for during the cyanobacteria blooms) is sio
n
slightly, medially and heavily N-limited in Scenario NP10, V2 and NP16, respectively
P
20 (Fig. 10d). In June–September, ENL is generally low, as it is the season of likely ap
e
cyanobacteriablooms,whichcanonlybeP-limited. r
|
4 Discussion D
is
c
u
4.1 HorizontalvariabilityofN/Pratio ss
io
n
ThehorizontalvariabilityofN/Pratio(Fig.3b)isfeaturedwithaminimumaround7:1 P
a
25 intheGotlandSeaandnomorethan9:1intheBalticproper, andwithanorthwards pe
r
395 |
increasing trend up to 18:1 in the Bothnian Sea and the Bothnian Bay and an east- D
is
wards increasing trend up to 22:1 in the Gulf of Finland. This pattern is synthesized cu
s
through comparing model results among Scenarios V1, NP10, NP16. The relatively s
io
lowN/PratiointheBalticproperiswellsupportedbythebetterconsistencybetween n
P
5 observationsandmodelresultsofScenarioV2thanthatofScenarioNP10. Theout- ap
performance of Scenario V2 over Scenario V1 (see Fig. 3a,b for N/P ratios) is very e
r
clear in the Bothnian Sea, the Bothnian Bay and the Gulf of Finland for the resultant
|
DIP(Fig.11d–h). DIPshowsexcessivelysurplusatthestationsinthoseareasinSce-
D
narioV1. TheadjustmentinScenarioV2relativetoScenarioV1significantlyimproves is
c
10 theconsumptionofDIP,butnotmuchchangesforDINandchlorophyll-a,becausethe us
adjustmentfromScenarioV1toV2doesnotchangethepropertyofnutrientlimitation sio
inthoseareas. Therefore,thehypothesisthatthelongterm(years2000∼2009)mean n
P
ratioofseasonaldeviationofDINrelativetothatofDIPindicatesthestoichiometryN/P ap
ratioisgenerallyvalidforoffshorestations,exceptthreeestuarinestationsA,HandI. er
15 Wan et al. (2011) suggested that the N/P ratio in the Baltic proper is smaller than |
10:1. Wanetal.(2011)reachedthisaccordingtonotonlycomparingsimulationsce-
D
nariosbutalsoestimatingthebiologicalnutrientremovalwithobserveddata. Wenotice is
c
that there are several arguments on the N/P ratio in the Baltic proper. Osterroht and us
s
Thomas (2000) noticed that the N/P ratio of nutrient alteration before and after the io
n
20 growingseasonwasmuchdifferentfromtheRedfieldratio,howevertheyexplainedthe P
a
elementalratioofnutrientuptakeintheBalticproperwereconsistentwiththeRedfield p
e
ratio,butthenutrientsremineralizedfromfreshlyproducedorganicmaterialhadanon- r
Redfieldratio. OncetheconceptoftheN/Pratioofnutrientuptakeisextendedtothe |
N/Pratioofnetbiologicalremovalofnutrients,itcouldalsobealternativelyexplained D
25 that the elemental ratio of nutrient uptake were inconsistent with the Redfield ratio. isc
Shaffer (1987) suggested the nutrient N/P ratio was smaller than the Redfield ratio us
s
basedonsedimentmeasurements. Accordingtomeasurements,Larssonetal.(2001) io
n
concludedthattheN/Pratioofelementalcompositionofoceanicsestonisonlyslightly P
a
smallerthantheRedfieldratiobeforeJunebutmuchlargerthantheRedfieldratiolater pe
r
396 |
on until the winter recovery. We think that the N/P ratio of elemental composition of Dis
oceanic seston probably differs much from the N/P ratio nutrient uptake. Otherwise, cu
s
the conclusion of Larsson et al. (2001) conflicts with two facts: that the N/P ratio of s
io
nutrient alteration before and after spring blooms is much smaller than the Redfield n
P
ratio16:1intheBalticproper(Table2),andthatnomuchDIPdeclinationisobserved a
5 p
aftermid-June. AstotheN/PratioofnutrientuptakeoutsideoftheBalticproper,there er
arefewrelevantinvestigationstocompareto.
|
D
4.2 EcologicalimpactsofhorizontallyvariableN/Pratio is
c
u
Although the horizontal variability of N/P ratio in the Baltic Sea suggested by the im- ss
io
10 plemented scenario simulations needs to be verified with other investigations in the n
P
future, thosesimulationsallowustoestimatethelikelyecologicalimpactsfromusing a
p
horizontallyvariableN/Pratioinsteadofafixedone. Wefocusonthelikelyecological e
r
impactsonprimaryproduction,nitrogenfixationandnutrientlimitation.
|
4.2.1 Primaryproduction D
is
c
u
15 The direct measurement of primary production can be made through the classic 14C ss
method. The measurements by the classic 14C method are generally dynamic and ion
have only narrow representation. As pointed by Wasmund et al. (2005), data on pri- P
a
p
mary production are actually limited and hard to map the seasonal and the spatial e
r
features, although there are some direct measurements (Elmgren, 1984; Wasmund
|
et al., 2001). We compare our model results with the calculated primary production
20
based on relevant measurements (Osterroht and Thomas, 2000; Struck et al., 2004; D
is
Wasmundetal., 2005; SavchukandWulff, 2007)andtherecentmodelresults(Maar cu
s
etal.,2011). s
io
Wasmundetal.(2005)estimatedthenetprimaryproductionintheeasternGotland n
25 Sea was 6454mmolCm−2 during the most productive period (28.3.2001∼13.7.2001) Pap
whichtheybelievedwereclosetotheannualnetprimaryproduction. Thisvalueequals e
r
397 |
77gCm−2yr−1. Strucketal.(2004)calculatedtheannualnewproductionintheeast- Dis
ern Gotland Sea ranging from 247 to 588mmolNm−2yr−1, it equals ranging from 41 cu
s
to97gCm−2yr−1 iftheRedfieldratioandthef-ratio0.48(Wasmundetal.,2005)are sio
applied. Ourmodeledvalueinthisregionvariesbetween40to80gCm−2yr−1,which n
P
5 roughlyfitstheestimatesbasedonmeasurements. Thedistributionpatternofprimary ap
e
production(Fig.8a)isquitesimilarwithMaaretal.(2011). BothMaar’sandourmodel r
resultsshowanorthwardsdecreasingtrend,whichisconsistentwiththeestimatesof |
SavchukandWulff(2007).
D
The estimates of primary production depend on N/P ratio of nutrients, and even is
c
10 ratios of nutrients to carbon if the model does not explicitly include a state variable us
s
carbon. This recovers the uncertainty involved in primary production estimates from io
n
models. Eventhoughtheabsolutevalueofmodelestimatesinvolvesuncertaintymore P
a
or less, model can capture the horizontal distribution pattern and seasonal pattern p
e
reasonable well in the present study (Fig. 8a,d). As one of the three usual methods r
15 estimating primary production, the “nutrient concept” method uses the phytoplankton |
stoichiometry(Wasmundetal.,2005). Itmeanssomeofprimaryproductionestimates D
even based on measurements can also include uncertainty from the selected phyto- isc
u
planktonstoichiometry. s
s
ThephytoplanktonstoichiometryintheBalticSeaistraditionallyregardedasconsis- io
n
tentwithRedfieldratios(Neumann,2000;Edelvangetal.,2005;NeumannandScher- P
20 a
newski,2008;Savchuketal.,2008;Eilolaetal.,2009;Maaretal.,2011;Meieretal., pe
r
2011). If this is not true, the estimates of primary production might be overestimated
20∼40%,accordingtotheimpactofdifferentN/Pratios(Fig.8a,c). |
D
4.2.2 Nitrogenfixation isc
u
s
s
25 As methods to measure nitrogen fixation directly are not available yet, there are no ion
precisemeasurementsfornitrogenfixation. Wecompareourmodelresultswiththees- P
a
timatesbasedrelevantmeasurements(Niemisto¨ etal.,1989;Hu¨belandHu¨bel,1995; p
e
r
398 |
Rahm et al., 2000; Larsson et al., 2001; Rolff et al., 2007) and other model studies D
is
(NeumannandSchernewski,2008;Neumannetal.,2010). Weestimatethetotalan- c
u
nualnitrogenfixationfortheBalticSeais41,111and344ktonNyr−1forScenariosV2, ss
io
NP10andNP16,respectively,accordingtothemodelresults(Fig.9d). Niemisto¨ etal. n
5 (1989) estimated the nitrogen fixation in early 1980s around 100ktonNyr−1. Hu¨bel Pap
andHu¨bel(1995)estimatedthenitrogenfixationfortheBalticproperincludingtheBelt e
r
Sea but not the northern Baltic proper ranging 20∼190ktonNyr−1. The estimate by
|
Rahm et al. (2000) was 30∼260ktonNyr−1. The estimate by Larsson et al. (2001)
for years 1994∼1998 was 180∼430ktonNyr−1. The estimate by Rolff et al. (2007) Dis
10 reached the highest 1000ktonNyr−1 since ever. The model result by Neumann and cus
Schernewski(2008)showedthenitrogenfixationvarying200∼800ktonNyr−1 forthe sio
period 1960∼2000. The result from similar model by Neumann (2010) showed the n
P
nitrogen fixation varying 80∼280ktonNyr−1 for the period 1960∼2100, with a value ap
around160ktonNyr−1foryears2000s. Inbrief,ourmodelresultsareintherightorder, er
15 butrelativelylow. |
TheamountofnitrogenfixationintheBalticSeadependsontheDIPavailabilityand
D
the N/P ratio of nutrient uptake of cyanobacteria. The DIP available for cyanobacte- is
c
ria mainly depends on the winter DIP level minus the DIP consumption during spring us
s
bloomswhichdependsontheN/Pratioofnutrientuptakeagain. Thustheamountof io
n
20 nitrogen fixation significantly depends on the N/P ratio. The involvement of the N/P P
a
ratiocanpartiallyexplainwhytheestimatesofnitrogenfixationcanvaryoneorderof p
e
magnitude (Niemisto¨ et al., 1989; Hu¨bel and Hu¨bel, 1995; Rahm et al., 2000; Lars- r
sonetal.,2001;Rolffetal.,2007;NeumannandSchernewski,2008;Neumannetal., |
2010). D
25 Our model estimates in Scenarios V2, NP10 and NP16 are 41, 111 and isc
344ktonNyr−1,respectively. ThisreflectsthesignificantimpactofN/Pratio. Wethink us
s
that the Redfield ratio or even a further bigger N/P ratio for all seasons, which most io
n
ofotherestimatesused,tendtooverestimatethenitrogenfixation. IftheN/Pratioof P
a
nitrogenfixationisnotsmallerthantheRedfieldratio, inadditionthattheratioofDIN pe
r
399 |
relativetoDIPinriverloadingsismuchlargerthantheRedfieldratio(Sta˚lnackeetal., D
is
1998),thismightconflictwiththefactthattheratioofDINrelativetoDIPintotalnutrient cu
s
stocksismuchsmallerthan16:1(Danielssonetal.,2008). s
io
n
P
4.2.3 Nutrientlimitation a
p
e
r
While ocean waters are usually N-limited (Nixon et al, 1996) and freshwaters are P-
5
limited(Schindler,1974),theBalticSea,asanadjacentsea,isrelativelycomplicated. |
TheBalticSeawasregardedN-limitedintheBalticproperandtheGulfofFinland,but D
is
P-limited in the Bothnian Sea and the Bothnian Bay (Grane´li et al., 1990; Kivi et al., c
u
1993;Zweifeletal.,1993;Danielssonetal.,2008). Thepropertyofnutrientlimitation ss
io
10 to primary production can vary across basins and seasons, depending on both the n
P
nutrientsuppliesandphytoplanktonstoichiometry. RegionsP-limitedcanbeswitched a
p
to N-limited when the N/P ratio is increased. For instance, the Bothnian Bay, the e
r
BothnianSea,theSkagerrakandtheeasternendoftheGulfofFinlandareP-limited
|
inScenarioNP10,butturntobeN-limitedinScenarioV2(Fig.10a,b).
15 The model result of Scenario V2 shows the Bothnian Bay is N-limited, as the N/P Dis
ratio is assumed higher 20:1. Zweifel et al. (1993) regarded the Bothnian Bay is P- cu
s
limited. However, nitrogen was found depleted in summer (Andersson et al., 1996; s
io
Humborg et al., 2003). It means the Bothnian Bay can also be N-limited. The model n
P
resultofScenarioV2showsalsotheBothnianSeaismostlyN-limited. Thisisconsis- a
p
20 tentwithAnderssonetal.(1996)andDanielssonetal.(2008). er
|
5 Summary D
is
c
u
Thisinvestigationstartsfromahypothesisthatthelongterm(years2000∼2009)mean s
s
ratioofseasonaldeviationofDINrelativetothatofDIPindicatestheN/Pratioineach ion
basin in the Baltic Sea. A 3-D physical-biogeochemical coupled model is applied to P
a
p
25 testthishypothesis. ThemonthlyseriesofobservedDINandDIPfromninestationsin er
400 |
differentbasinsareusedtogeneratethehypothesizedN/Pratiosandtheninterpolated Dis
for the whole Baltic Sea. The monthly series of observed surface chlorophyll-a, DIN cu
s
and DIP from another nine stations and other various observations for chlorophyll-a, s
io
DIN and DIP are used to verify that the hypothesized horizontally variable N/P ratio n
P
distributionisbetterthananoptimalfixedN/Pratio. Thehypothesisisverifiedgenerally a
5 p
validforobservationsfromoffshorestations,butnotforthreeestuarinestations. er
SimulationsforScenariosV2(withtheN/Pratiodistributionbasedonthehypothe-
|
sized N/P ratio but manipulated for three estuarine stations), NP10 and NP16 (fixed
D
N/Pratios10:1and16:1)areanalyzedforprimaryproduction,nitrogenfixationand is
c
nutrientlimitation. ScenariosV2andNP10havealmostsameprimaryproduction,and u
10 s
they have 10∼60gCm−2yr−1 higher primary production during spring blooms and sio
in early summer than Scenario NP16 does. Different N/P ratio scenarios can cause nP
significantnitrogenfixationdifference,forinstance,ScenarioV2,NP10andNP16es- ap
timate nitrogen fixation 41, 111 and 344ktonNyr−1, respectively. Different N/P ratio er
15 scenarioscanalsocausesignificantchangeforthenutrientlimitationproperty. |
D
Acknowledgement. WewouldliketothankPerBergfortechnicalassistancewiththeHIROMB is
model code and setups, and to thank the Swedish Meteorological and Hydrological Institute cu
and the Bundesamt fu¨r Seeschifffahrt und Hydrographie in Hamburg, Germany for providing ss
io
theriver data. Thiswork wassupported byEuropean Commission FP6projects MYCOEAN n
20 (ContractNo.218812)andMEECE(ContractNo.DK18159104). Pa
p
e
r
References |
D
Andersson, L., Hajdu, S., Haecky, P., Kuparinen, J., and Wikner, J.: Succession and growth is
c
limitationof phytoplanktonin theGulfof Bothnia(Baltic Sea), Marine Biol., 126, 791–801, u
s
1996. s
io
25 Anderson,T.R.andPondaven,P.: Non-redfieldcarbonandnitrogencyclingintheSargasso n
Sea:pelagicimbalancesandexportflux,Deep-SeaRes.Pt.I,50,573–591,2003. Pa
Arrigo,K.R.,Robinson,D.H.,Worthen,D.L.,Dunbar,R.B.,DiTullio,G.R.,VanWoertand,M., pe
r
401 |
and Lizotte, M. P.: Phytoplankton community structure and the drawdown of nutrients and D
is
CO intheSouthernOcean,Science,283,365–367,1999. c
2 u
Bergstro¨m, S.: Development and application of a conceptual runoff model for Scandinavian s
s
catchments,Ph.D.thesis.,SMHIReportsRHO,No.7,Norrko¨ping,1976. io
n
5 Bergstro¨m, S.: The HBV model – its structure and applications, SMHI Reports RH, No. 4, P
a
Norrko¨ping,1992. p
e
Conkright, M. E., Locarnini, R., Garcia, H., O’Brien, T., Boyer, T. P., Stephens, C., and r
Antonov, J.: World ocean atlas 2001, objective analyses, data statistics and figures, CD-
|
ROMdocumentation,NationalOceanographicDataCenter,SilverSpring,MD,2002.
10 Danielsson, A˚., Papush, L., and Rahm, L.: Alterations in nutrient limitations – Scenarios of Dis
achangingBalticSea,J.MarineSys.,73,263–283,2008. cu
s
Dick,S.,Kleine,E.,Mueller-Navarra,S.,Kleine,H.,andKomo,H.: Theoperationalcirculation s
modelofBSH(BSHcmod)–modeldescriptionandvalidation, BerichtedesBSH29/2001, ion
Bundesamtfu¨rSeeschifffartundHydrographie,1–48,2001. P
a
15 Edelvang, K., Kaas, H., Erichsen, A. C., Alvarez-Berastegui1, D., Bundgaard, K., and pe
Jørgensen, P. V.: Numerical modeling of phytoplankton biomass in coastal waters, J. Ma- r
rineSys.,57,13–29,2005. |
Eilola, K., Meier, H. E. M., and Almroth, E.: On the dynamics of oxygen, phosphorus and
D
cyanobacteriaintheBalticSea:amodelstudy,J.MarineSys.,75,163–184,2009. is
c
20 Elmgren,R.: Trophicdynamicsintheenclosed,brackishBalticSea.Rapp.P.-v.Re´un.Cons. us
Int.Explor.Mer.,183,152–169,1984. s
io
Falkowski,P.G.:Rationalizingelementalratiosinunicellularalgae,J.Phycol.,36,3–6,2000. n
Grane´li, E., Wallstro¨m, K., Larssson, U., Grane´li, W., and Elmgren, R.: Nutrient limitation of Pa
p
primaryproductionintheBalticSeaarea,Ambio,19,142–151,1990. e
25 Humborg,C,Danielsson,A˚.,Sjo¨berg,B.,andGren,M.:Nutrientland-seafluxesinoligotrophic r
and pristine estuaries of the Gulf of Bothnia, Baltic Sea, Est. Coast. Shelf Sci., 56, 1–13, |
2003. D
Hu¨bel, H. and Hu¨bel, M.: Blaualgen-Wasserbluten in der Ostsee: Ursachen – Ausmass – is
c
Folgen,Dtsch.Hydrogr.Z.(Suppl.),2,151–158,1995. u
s
30 Kivi, K., Kaitala, S., Kuosa, H., Kuparinen, J., Leskinen, E., Lignell, R., Marcussen, B., and sio
Tamminen,T.:NutrientlimitationandgrazingcontroloftheBalticplanktoncommunityduring n
P
annualsuccession,Limnol.Oceanogr.,38(5),893–905,1993. a
p
Klausmeier, C. A., Litchman, E., and Levin, S. A.: Phytoplankton growth and stoichiometry e
r
402 |
undermultiplenutrientlimitation,Limnol.Oceanogr.,49,1463–1470,2004a. D
is
Klausmeier, C. A., Litchman, E., Daufresne, T., and Levin, S. A.: Optimal nitrogen-to- c
u
phosphorusstoichiometryofphytoplankton,Nature429,171–174,2004b. s
s
Klausmeier,C.A.,Litchman,E.,andLevin,S.A.: Amodelofflexibleuptakeoftwoessential io
n
5 resources,J.Theor.Biol.,246,278–289,2007. P
a
Kress,N.andHerut,B.: Horizontalandseasonalevolutionofdissolvedoxygenandnutrients p
e
intheSouthernLevantineBasin(EasternMediterraneanSea): chemicalcharacterizationof r
thewatermassesandinferencesontheN:Pratios,Deep-SeaRes.Pt.I,48,2347–2372,
|
2001.
10 Kuznetsov,I.,Neumann,T.,andBurchard,H.:Modelstudyontheecosystemeffectofavariable Dis
C:N/PratioforcyanobacteriaintheBalticProper,Ecol.Model.,219,107–114,2008. cu
s
Larsson,U., Hajdu,S., Walve,J., andElgren,R.: BalticSeanitrogenfixationestimatedfrom s
thesummerincreaseinuppermixedlayertotalnitrogen, Limnol.Oceanogr., 46, 811–820, ion
2001. P
a
15 Maar, M., Møller, E. F., Larsen, J., Kristine, S. M., Wan. Z., She, J., Jonasson, L., and Neu- pe
mann, T.: Ecosystem modeling across a salinity gradient from the North Sea to the Baltic r
Sea,Ecol.Model.,222,1696–1711,2011. |
Meier,H.E.M., Eilola,K., andAlmrouth,E.: Climate-relatedchangesinmarineecosystems
D
simulated with a 3-dimensional coupled physical-biogeochemical model of the Baltic Sea, is
c
20 Clim.Res.,48,31–55,2011. us
Minster,J.-F.andBoulahdid,M.: Redfieldratiosalongisopycnalsurfaces–acomplementary s
io
study,Deep-SeaRes.,34,1981–2003,1987. n
P
Neumann,T.: Towardsa3D-ecosystemmodeloftheBalticSea,J.MarineSys.,25,405–419, a
p
2000. e
25 Neumann,T.: Climate-changeeffectsontheBalticSeaecosystem: amodelstudy,J.Marine r
Sys.,21,213–224,2010. |
Neumann,T.andSchernewski,G.: EutrophicationintheBalticSeaandshiftsinnitrogenfixa- D
tionanalyzedwitha3-Decosystemmodel,J.MarineSys.,74,592–602,2008. is
c
Neumann,T.,Fennel,W.,andKremp,C.: Experimentalsimulationswithanecosystemmodel u
s
30 oftheBalticSea: anutrientloadreductionexperiment,GlobalBiogeochem.Cy.,16,7(1)– sio
7(19),2002. n
P
Niemisto¨,L.,Rinne,I.,Melvasalo,T.,andNiemi,A˚.:Bluegreenalgaeandtheirnitrogenfixation a
p
intheBalticSeain1980,1982and1984,Meri,17,3–59,1989. e
r
403 |
Nixon,S.W.,Ammerman,J.W.,Atkinson,L.P.,Berounsky,V.M.,Billen,G.,Boicourt,W.C., D
is
Boynton, W. R., Church, T. M., Ditoro, D. M., Elmgren, R., Garber, J. H., Giblin, A. E., c
u
Jahnke,R.A.,Owens,N.J.P.,Pilson,M.E.W.Q.,andSeitzinger,S.P.:Thefateofnitrogen s
s
andphosphorusattheland-seamarginoftheNorthAtlanticOcean, Biogeochemistry, 35, io
n
5 141–180,1996. P
a
Osterroht,C.andThomas,H.:Newproductionenhancedbynutrientsupplyfromnon-Redfield p
e
mineralisationoffreshlyproducedorganicmaterial,J.MarineSys.,25,33–46,2000. r
Rahm,L.,Jo¨nsson,A.,andWulff,F.: NitrogenfixationintheBalticProper: anempiricalstudy,
|
J.MarineSys.,25,239–24,2000.
D
10 Redfield, A. C.: On the proportions of organic derivations in sea water and their relation to is
thecompositionofplankton,in: JamesJohnsonMemorialVolume,editedby: Daniel,R.J., cu
s
UniversityPressofLiverpool,Liverpool,177–192,1934. s
Redfield, A. C.: The biological control of chemical factors in the environment, Am. Sci., 46, ion
205–221,1958. P
a
15 Rolff, C., Almesjo¨, L., and Elmgren, R.: Nitrogen fixation and abundance of the diazotrophic pe
cyanobacteriumAphanizomenonsp.intheBalticproper, MarineEcol.ProgressSer., 332, r
107–118,2007. |
Savchuk, O. P.: Resolving the Baltic Sea into seven subbasins: N and P budgets for 1991–
D
1999,J.MarineSys.,56,1–15,2005. is
20 Savchuk, O. P. and Wulff, F.: Modeling the Baltic Sea eutrophication in a decision support cus
system,Ambio,36,141–148,2007. s
io
Savchuk,O.P.,Wulff,F.,Hille,S.,Humborg,C.,andPollehne,F.:TheBalticSeaacenturyago n
P
–areconstructionfrommodelsimulations,verifiedbyobservations,J.MarineSys.,74,485– a
p
494,2008. e
r
25 Schindler,D.W.:Eutrophicationandrecoveryinexperimentallakes:implicationsforlakeman-
agement,Science,184,897–899,1974. |
Shaffer, G.: Redfield ratios, primary production, and organic carbon burial in the Baltic Sea, D
Deep-SeaRes.Pt.I,34,769–784,1987. is
c
She, J., Berg, P., and Berg, J.: Bathymetry effects on water exchange modeling through the u
s
30 DanishStraits,J.MarineSys.,65,450–459,2007a. sio
She,J.,Høyer,J.,andLarsen,J.: Assessmentofseasurfacetemperatureobservationalnet- n
P
worksintheBalticSeaandNorthSea,J.MarineSys.,65,314–335,2007b. a
p
Sta˚lnacke, P., Grimvall, A., Sundblad, K., and Tonderski, A.: Estimation of riverine loads of e
r
404 |