Oxygen minimum zone of the open Arabian Sea

25 Pages·2014·3.39 MB·English

Biogeosciences,11,2237–2261,2014 O
p
www.biogeosciences.net/11/2237/2014/ e
n
doi:10.5194/bg-11-2237-2014 Biogeosciences A
c
c
©Author(s)2014.CCAttribution3.0License. e
s
s
Oxygen minimum zone of the open Arabian Sea: variability of
oxygen and nitrite from daily to decadal timescales
K.Banse1,S.W.A.Naqvi2,P.V.Narvekar2,†,J.R.Postel1,andD.A.Jayakumar2,*
1UniversityofWashington,SchoolofOceanography,Box357940,Seattle,Washington98195-7940,USA
2NationalInstituteofOceanography,CouncilofScientific&IndustrialResearch,DonaPaula,Goa403004,India
*nowat:PrincetonUniversity,M45GuyotHall,DepartmentofGeosciences,Princeton,NewJersey08544,USA
†
deceased
Correspondenceto:K.Banse([email protected])
Received:15August2013–PublishedinBiogeosciencesDiscuss.:30September2013
Revised:12February2014–Accepted:17February2014–Published:23April2014
Abstract.Theoxygenminimumzone(OMZ)oftheArabian as follows: (1) an O climatology for the upper OMZ re-
2
SeaisthethickestofthethreeoceanicOMZ.Itisofglobal veals a marked seasonality at 200 to 500m depth with O
2
biogeochemicalsignificancebecauseofdenitrificationinthe levels during the northeast monsoon and spring intermon-
upperpartleadingtoN andN Oproduction.Theresidence soonseasonselevatedoverthoseduringthesouthwestmon-
2 2
timeofOMZwaterisbelievedtobelessthanadecade.The soon season (median difference, 0.08mLL−1 [∼3.5µM]).
upperfewhundredmetersofthiszonearenearlyanoxicbut The medians of the slopes of the seasonal regressions of
non-sulfidicandstillsupportanimal(metazoan)pelagiclife, O on year for each of the NE and SW monsoon seasons
2
possiblyasaresultofepisodicinjectionsofO byphysical are −0.0043 and −0.0019mLL−1a−1, respectively (−0.19
2
processes. and −0.08µMa−1; n=10 and 12, differing at p=0.01);
We report on discrete measurements of dissolved O and (2) four decades of statistically significant decreases of O
2 2
NO−,temperatureandsalinitymadebetween1959and2004 between 15 and 20◦N but an opposing trend toward an in-
2
wellbelowthetopsofthesharppycnoclineandoxyclinenear creasenear21◦Nareobserved.Themechanismsofthebal-
150,200,300,400,and500mdepth.Weassemblenearlyall ance that more or less annually maintain the O levels are
2
O determinations(originallytherewere849values,695of still uncertain. At least between 300 and 500m, the replen-
2
whichcamefromtheOMZ)bythevisualendpointdetection ishment is inferred to be due to isopycnal re-supply of O ,
2
oftheiodometricWinklerprocedure,whichinourdatabase while at 200(or 250?)m it is diapycnal, most likely by ed-
yields about 0.04mLL−1 (∼2µM) O above the endpoint dies. Similarly, recent models show large vertical advection
2
frommodernautomatedtitrationmethods.Weacknowledge ofO wellbelowthepycnoclinesandoxyclines.
2
− −
thatmuchlower(nanomolar)O valueshavebeenmeasured TheNO distribution,takenasanindicatorofactiveNO
2 2 3
recentlywiththeSTOX(SwitchableTraceamountOXygen) reduction,doesnotshowatrendintheredoxenvironmentfor
sensor in the eastern tropical South Pacific, and that similar aquarterofacenturyataGEOSECSstationnear20◦N.In
conditionsmayalsoprevailintheArabianSeaOMZ.Inspite theentireOMZ,theregressionslopesonyearwithinseasons
−
oftheerrorinO measurementsatvanishinglylowlevels,we fortherathervariableNO donotpresentaclearpatternbut
2 2
arguethatthetemporaltrendsofthehistoricdatashouldstill byothermeasurestendedtoanincreaseofNO−.
2
hold. Vertical net hauls collect resident animal (metazoan)
We find 632 values acceptable (480 from 150 stations in pelagic life in the NO− maximum of the OMZ at O lev-
2 2
the OMZ). The data are grouped in zonally paired boxes of els well below the lower limit of the Winkler titration; the
1◦ lat. and 2◦ long. centered at 8, 10, 12, 15, 18, 20, and extremely low O content is inferred from the presence of
2
21◦Nalong65and67◦E.Thelatitudesof8–12◦N,outside NO− believed to be produced through microbial NO− re-
2 3
the OMZ, are treated in passing. The principal results are duction. Instead of the difficult measurement by the STOX
PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion.
2238 K.Banseetal.:OxygenminimumzoneoftheopenArabianSea
sensor, the relation between the very low O inferred from duction rate. We are concerned, however, with the depths
2
−
presence of NO and mesozooplankton should be studied <500m. Therefore, one would divide a slightly lower in-
2
−
with100to150Lbottlesratherthannets. ventory by the substantial, measured NO reduction rate
3
The spatial (within drift stations) and temporal (daily) (Lam et al., 2011, upper Fig. 3d), which will reduce the
variabilityinhydrographyandchemistryislargealsobelow turnover=residence time greatly. The estimates of neither
the principal pycnocline. The seasonal change of hydrogra- Naqvi(1987)norOlsonetal.(1993)lendthemselvestocor-
phyisconsiderableevenat500mdepth.FutureO ornutri- rection. Further, Warren (1994) has remarked on the basic
2
entbudgetsfortheOMZmustnotbebasedonsinglecruises weakness of simple box models such as these. Incidentally,
orsectionsobtainedduringoneseasononly.Steadystatecan- note that residence time of water is not the same as the age
notbeassumedanylongerfortheintermediatelayersofthe of water estimated by, for example, freon (CFC-11 or 12).
centralArabianSea. So, accepting a short residence time, temporal O changes
2
shouldbediscernibleinhistoricaldatafromfourdecadesas
analyzedhere.BecausesmallO shiftsatthegenerallyquite
2
low concentrations in the OMZ may suddenly stop or start
1 Introduction denitrification, the temporal variability of the intensity and
geographicalextentofthisOMZareofglobalbiogeochemi-
Ourpaperaddressesvariabilityandclimatologyofdissolved calsignificance.
−
oxygen (O2) and nitrite (NO2) from discrete samples col- Afterwehadlargelycompletedouranalysisofthehistor-
lectedbetween1959and2004intheupperoxygenminimum ical O data, we learned of the introduction to the field of
2
zone (OMZ) of the central Arabian Sea, which includes the theSTOXsensor(SwitchableTraceamountOXygen;Revs-
secondary nitrite maximum (SNM). The entire OMZ occu- bechetal.,2009).Itsdetectionlimitisalmosttwoordersof
pies approximately the 150–1000m depth range and is the magnitude lower than that of the automated Winkler titra-
thickest of the three major OMZs of the open ocean. Be- tion and will necessitate a re-assessment of the intensity of
tween one-fourth and one-third of the total marine denitri- theoceans’OMZs.Titrationwashithertousedworldwidefor
ficationisestimatedtooccurinthewatercolumn,ofwhich O studies,aswellasforthecalibrationoftheO probesat-
2 2
uptoone-thirdtoone-halfmaytakeplacewithintheOMZof tached to CTDs. Thamdrup et al. (2012) in February 2007
theArabianSea(Codispotietal.,2001;Bangeetal.,2005). employed the sensor along a section from 28 to 5◦S off
AccordingtoNaqvietal.(2005:Table9),thecontributionby the continental slope in the eastern tropical South Pacific
the Arabian Sea to the global marine pelagic denitrification (ETSP).IntheOMZcorebetweenabout100and300–350m
lies between 8 and 21%. The principal denitrifying zone is depth,O wasconsistentlyundetectableon10of12stations
2
locatedintheupperone-thirdoftheOMZwhereO2concen- (detection limit 0.01µmolkg−1 or less). The authors found
trations, as determined by visual endpoint detection of the depletion to <0.09µmolkg−1O (∼0.002mLL−1) to be
2
Winkler titration, fall below 0.06mLL−1 (∼2.7µM; how- normalinthecore.Further,NO− intheSNMoccurredonly
2
ever,seebelow).ThezoneisreadilyidentifiablebytheSNM, at <0.05µmolkg−1O (∼0.001mLL−1), which is lower
− 2
whichmostlyisduetoNO3 reductionandseparatefromthe by at least an order of magnitude than the best automated
−
primaryNO maximumthatismorecommonlyfoundnear endpoint detection of the Winkler titration or the colori-
2
thebottomoftheeuphoticzone(cf.SupplementFig.S.3.1; metric measurements (Codispoti et al., 2001). Thamdrup et
LomasandLipschultz,2006).OtherthanthereleaseofN2as al. (2012) concluded that the core of the OMZ off Peru is
the end product through denitrification and anammox path- broadly functionally anoxic, because the observed low O
2
ways (Dalsgaard et al., 2003, 2005; Kuypers et al., 2003; cannot sustain aerobic metabolism for any lengthy period.
Wardet al.,2009; Ward,2013), theO2 deficiencyenhances Ulloa et al. (2012) applied the term AMZ (Anoxic Marine
theproductionofnitrousoxide(N2O),agreenhousegas,an- Zone) to the region. The offshore OMZ of the ETSP, how-
other intermediate of the nitrogen redox chemistry (Naqvi ever,isnon-sulfidic,becauseNO− isbroadlypresent(Silva
3
andNoronha,1991;Devoletal.,2006;Naqvietal.,2006). et al., 2009; Thamdrup et al., 2012). Since metazoans like
The OMZ is not restricted to a particular range of tem- copepodsareobligatoryaerobes,weconsiderAMZtomean
perature,salinity,ordensityandisessentiallymaintainedby awatercolumnwithoutanimalzooplanktonona24hbasis
the balance between the supply of O2 through eddy mixing andusethetermaccordingly.
andhorizontaladvection,principallyfromthesouthwest,and Becauseofthemanysampleswithhighconcentrationsof
local consumption. The mean residence time of water ap- NO−, the OMZ of the Arabian Sea is expected to broadly
2
pearstobeaboutadecadeorless,accordingtoNaqvi(1987, presentlowoxygenconditionssimilartothoseintheETSP.
∼4years, for 100 to 1000m) and Olson et al. (1993: 679; We know of only one instance, though, when O was mea-
2
11years for 200 to 1000m). In contrast, Lam et al. (2011) suredusingSTOXintheregion(September/October2007,in
estimated a NO− turnover time of 49±20years for 100 to Jensen et al., 2011: Supplement Fig. S.2). On three stations
2
1000mdepthinthecentral–northeasternArabianSea.They overdepthintervalsofseveralhundredmeters,dissolvedO
2
− −
divided the NO inventory by the modeled NO net pro-
2 2
Biogeosciences,11,2237–2261,2014 www.biogeosciences.net/11/2237/2014/
K.Banseetal.:OxygenminimumzoneoftheopenArabianSea 2239
was≤0.09µmolkg−1 (∼0.002mLL−1),whichwasthede-
−
tection limit of the sensor on the cruise. NO has always
3
beenfoundinfairlyhighconcentrationsineverysamplean-
alyzed within the OMZ during the last five decades, how-
ever, ruling out sulfidic anoxia in the water column. Also,
metazoan plankton is present throughout, albeit in dimin-
ishednumberswithintheupperOMZ(Sect.3.2.5).
−
The OMZ and the deep NO maximum it contains were
2
first discovered in 1933/1934 by Gilson (1937). Numerous
−
NO measurementsduringmanysubsequentexpeditionsin
2
theArabianSeashowedvalues>0.5µMandupto∼5µM,
thatcommonlyoccurinourdataaswell.Theyimplythatat
least during the last 75years, the lowest O concentrations
2
inthisOMZmusthavebeenatnanomolarlevels(Thamdrup
et al., 2012), in contrast to all historical O data including
2
thoseusedherein(SupplementTableS.1.b).Theappreciable
nitrate deficit resulting from the partial denitrification and
anammox, ranging at its vertical maximum between 2 and
−
15µMNO , has been discussed since the mid-1970s (e.g.,
3
Sen Gupta et al., 1976; see also Sect. 3.2.6). The OMZ has
Fig.1.Distributionoftheboxes(leftseries,“1”,rightseries,“2”)
existed during much of the Holocene, as apparent from the withoutlinesofO2concentration(mLL−1)at200m(fromWyrtki,
sedimentary record spanning many thousands of years with
1971)andthenitritemaximum(circumscribedbythe1µMcontour)
uninterrupted series of annual varves at two sites located in in the OMZ (from Naqvi, 1991). Circle: position of drift station
the northeastern Arabian Sea (von Rad et al., 1999; Staub- shown in Fig. 2; diamond: position of GEOSECS sta. 416; aster-
wasseretal.,2002;Thambanetal.,2007). isk:positionofdriftstationdepictedinSupplementFig.S.3.1;G:
So, today, is the OMZ of the Arabian Sea, or at least Goa;Ka:Karachi;Ko:Kochi(formerCochin);Mi:Mumbai(for-
its core, functionally anoxic as suggested by Thamdrup et merBombay);Mt:Muscat.
al. (2012) and Ulloa et al. (2012) for the core of the OMZ
of the eastern tropical South Pacific? The answer is No.
tion. Advection adds O at least occasionally to OMZs and
The OMZ of the Arabian Sea as a whole is not a non- 2
AMZsalsooffPeru,asnotedbyThamdrupetal.(2012)and
sulfidic anoxic marine zone (AMZ) as envisaged by Ulloa
off northern Chile as illustrated by Ulloa et al. (2012) for
etal.(2012)andThamdrupetal.(2012).Metazoan(animal)
June2007.
zooplankton reside in the OMZ even where the median O
2
AsshowninSupplementSect.2,thedetectionlimitofthe
content, as determined by the Winkler procedure, is lowest
titration endpoint in our data collation is ∼0.04mLL−1O
(Sect. 3.2.5). Further, we will show by the 40year clima- 2
(∼2µM)abovetheendpointofmodernautomatedmethods
tology in our wide longitudinal swath (Fig. 1) that the up-
(∼0.01mLL−1O [∼0.4µM]).However,inspiteoftheun-
per200(or250?)mareseasonallybeingstirred,presumably 2
certainties in the determination of O at vanishingly small
by eddies. In contrast, at the 300 to 500m horizons O ap- 2
2
concentrations,webelievethatthetemporaltrendsreported
parently is advected annually along isopycnals and replen-
herein remain valid. Were it not so, the ocean- and basin-
ishestheO consumedduringtheSWMperiod(Sect.4.2.1).
2
Moreover, NO− is largely observed only in the upper part scalesectionsoflowdissolvedO2 currentlyintheliterature
2 wouldnotcloselyreflectthebasin-widedistributionofnutri-
of the OMZ, generally between 150 and 400m depth, al-
ents.
though the lower boundary of the SNM could sometimes
Below, we describe first the methods and data selection
be as deep as 600–700m in the northeastern Arabian Sea
for temperature, salinity, oxygen, and nitrite in subsamples
(Naqvi,1987;Morrisonetal.,1999).Importantly,inourdata
drawn from the same water bottle. Then we review and up-
setsbetween200and500m,21%of707samplesofboxes
date the setting of the OMZ to 500m depth, including the
D1–G2 (without 500m in D1, D2, see Table 2) contained
zeroto≤0.02µMNO−(Sect.3.2.5).Thus,therewasenough small-scale spatial and temporal (days to weeks) variability
2 inordertoprovideperspectiveontheobservations.Thecore
O presenttopreventtheonsetofdenitrification.Incontrast,
2
ofthepaperaddressesseasonalandfour-decadalchangesof
this percentage was 82% of 255 samples outside the OMZ
−
O andNO .Weconcludewithasectionontheimplications
inboxesA1–C2andthedeepesthorizonintheD-boxes.We 2 2
oftheresults.
alsoobservedthataboutfour-fifthsofour707samplescom-
ing from the 200 to 500m horizons, which have significant
−
amountsofNO ,areunevenlydistributedinspace,asisthe
2
>one-fifth that contains too much O to allow denitrifica-
2
www.biogeosciences.net/11/2237/2014/ Biogeosciences,11,2237–2261,2014
2240 K.Banseetal.:OxygenminimumzoneoftheopenArabianSea
2 Materialsandmethods plement Sect. S.2). This procedure was in general use un-
tilitwaspartiallyreplacedbyautomatedtitration,bywhich
2.1 Datasources also the CTD-attached O probes were calibrated. For our
2
boxes we strove to collect all observations based on man-
TheprincipalsourcesforourdiscreteobservationsofO and
2 ualWinklertitrationswithvisualendpointdetection,buthad
−
NO togetherwithtemperatureandsalinityweretheIndian
2 to remove quite a few because of obvious bias as detailed
and US national oceanographic data centers (INODC and
below. Measurements by other methods, i.e., those involv-
NODC,respectively).Collectionsonsomeadditionalcruises
ing automated (e.g., colorimetric or photometric) endpoint
conductedbyIndia’sNationalInstituteofOceanographybut
detection of the iodometric titration, were not considered,
not yet incorporated in the INODC and NODC bases were
unless noted, because of systematic differences in the an-
also utilized. All data had been taken near 150, 200, 300,
alytical results discussed below. Excluded thus are cruises
400, and 500m, within boxes of 1◦ lat. by 2◦ long. The
118 and 159 of R/V Gaveshani, as well as the observations
boxesarecenteredat8,10,12,15,18,20,and21◦Nalong
from cruises 99 and 104 of R/V Sagar Kanya and all those
65 and 67◦E (from A1 to G1 and B2 to G2, respectively,
by R/V T. G. Thompson and other recent expeditions. To
in Fig. 1). While the text and the figures refer to rounded
our knowledge, for avoiding interference by nitrite, sodium
nominaldepths,thegreatmajorityofourdatawerecollected
azidewasnotaddedonanyexpeditionexceptthosebyR/V
within 5% and in some cases within 10% of the nominal
T.G.Thompson(Morrisonetal.,1999).
horizons.Inaddition,weincludesomeO2observationsatex- Almost all of our O data were recorded in mLL−1,
2
actnominaldepthsandtheveryfewvaluesfromdatacenters which we did not convert to µM or µmol kg−1, except for
that were already interpolated to them. The measurements
meansandmedians,becausereportingtheexactmultiplica-
used are listed in Supplement Table S.1.b together with the
tionwouldhaveaddedadecimalplaceimplyingafalsepre-
temperatureandsalinityofthesamples.
cision,whereasroundingoffmighthaveintroducederrorsin
We identify the seasons following the US Joint Global
means and medians of sets of often only five to ten values.
Ocean Flux Study (JGOFS) classification (Morrison et al., Originalobservationsexpressedasµmolkg−1 wereretained
1998) but the starting dates lagged by a half a month
butalsostatedasconvertedv/vormolarunits.
to one month: northeast monsoon (NEM), December–
During the last two to three decades O was analyzed
2
March; spring intermonsoon (SI), April–May; southwest
with automated endpoint detection in the Winkler anal-
monsoon (SWM), June–September; and fall intermonsoon
ysis (JGOFS manual; Anon., 1994). By comparing such
(FI), October–November. The lag is introduced because the
measurements with visual endpoint detection in the same
biochemicalresponseatdepthdependsalsoonthedownward
boxes during the same periods, we find an overestimate by
transmissionofsurfacesignalsbysinkingparticles. 0.04mLL−1O (∼ 2µM) (Supplement Sect. 2). Without
2
further evidence we generalize the difference as applicable
2.2 Temperatureandsalinity
toourentirematerialinSupplementTableS.1.b.Inviewof
thelowO concentrationsintheOMZ(Sect.3.1.2),thebias
AlltemperaturesandsalinitiesaccompanyingtheO and/or 2
2
− isfarfromnegligible.
NO data were taken at face value except that five hydro-
2
graphicseries(includingtheO andNO−values)wereelim- The accuracy in historical O2 data may vary for a num-
2 2 ber of reasons. For quality control we used the accompany-
inatedasoccurringinexceptionallydeepeddies(twoseries)
−
ing NO values, all accepted at face value, and eliminated
or as clearly due to pre-trips of the entire bottle strings. A 2
O values >0.10mLL−1 when they were accompanied by
veryfewsalinityrecordsweredroppedasfalsebybeingob- 2
− −
NO values>0.2µM.Asstatedabove,highNO isanindi-
viousoutliersinT–Sdiagrams.Tofillinlargetemporalgaps, 2 2
− catorofdenitrification,andifassociatedwithsubstantialO
a few measurements without O and/or NO observations 2
2 2
itimpliesoverestimationofthelatter(seeSect.1;alsoSup-
were taken from the data centers. Data from 329 and 332
plement Sect. 2). We corrected apparent O measurements
stationswith1376temperatureand1380salinityvalues,re- 2
<0.10mLL−1accompaniedbyNO−≥0.2µMfornitritein-
spectively,wereutilized(SupplementTableS.1.b).Asinthe 2
− terference(Wong,2012)andsignifiedthesebyitalicsinSup-
caseofO andNO ,thetotalincludessomemeansofrepli-
2 2 −
plement Table S.1.b. Our cutoff of 0.2µMNO for elimi-
catecastsatroutineordriftstations.SupplementTablesS.2 2
and S.3 present the medians of temperature and salinity for nating O2 >0.10mLL−1 was rather subjective but seemed
−
allboxesanddepths. justifiable given the variability in NO2 blanks that may in-
troduce errors up to ∼0.05µM. Of course, in view of the
2.3 Oxygen recently recognized much lower O threshold for the onset
2
ofdenitrification(Sect.1),thecutoffisindeedarbitraryand
This section treats analytical bias, as well as describes our only serves to sort our data collation. The 1µM contour in
criteriaforrejectingmanyreportedvalues. ourFig.1isclosetothe0.2µMcontourinNaqvi(1991)de-
Our O2 data were generated by the iodometric Winkler marcatingtheboundariesofthesuboxiczone.
titrationtechniquewithvisualendpointdetection(seeSup-
Biogeosciences,11,2237–2261,2014 www.biogeosciences.net/11/2237/2014/
K.Banseetal.:OxygenminimumzoneoftheopenArabianSea 2241
From all boxes between about 150 and 500 m, 849 O 2.5 Statistics
2
valuesbasedonvisualendpointdetectionfrom205stations
were found (duplicate casts on the same station are aver- Seasonalanddecadalchangeswereinvestigatedbylinearre-
aged and not counted as such; also not included are re- gression analysis with the independent variable (dates, i.e.,
jecteddatafromthreecruisesmentionedbelow).IntheOMZ year and month) known, and parametric statistics without
(boxes D1-G2), 695 observations or means were recorded. testing for normality and homoscedasticity. Significance of
Ofthese,215valueswithO >0.10mLL−1accompaniedby differencesbetweenmediansorgroupsofdatawasassessed
2
NO−≥0.2µMwererejected.Theremaining480datapoints by the non-parametric rank test (Wilcoxon T test=Mann–
2
came from 150 stations. Included are 31 stations with 56 Whitney U test). Because the latter mostly addressed clear
samples with O >0.10mLL−1 not accompanied by NO− differencesbetweentwodatasets,one-tailedtestswereusu-
2 2
analysesbecausetherewasnoreasontorejectthem(ofthese, ally applied. In neither approach did the values weight the
14 came from ∼150m where higher O values can be ex- statistics that were based on means. Our statements about
2
pectedbecauseoftheproximitytotheoxycline).Theinclu- significanceof“differencesbetweenmedians”areshorthand
sionofthe56pointsdidnotalterourconclusionsexceptpos- for“testsforsignificanceofdifferencesbetweentwosetsof
siblythedecadalslopeat400minF1in1963. independentvalues.”Thep valuesreflectthedistributionof
Rejectedandnotincludedinanyoftheenumerationswere the variables tested, but also the often low number of sam-
O records from two cruises that were in their entirety far ples.Becausesomanysetsaresmall,reportingevenp=0.2
2
too high within the OMZ north of 14 to 15◦N during the seemsappropriate,butisnotintendedtoserveasproof.
four decades under consideration (R/V Priliv, May 1967;
R/VAkademikKurchatov,May1976),aswellasthosefrom
the 22nd cruise of R/V Akademik Vernadsky of 1980. Here, 3 ResultsandDiscussion
all O values >0.10mLL−1 from the OMZ were accom-
2
−
panied by NO >0.2µM, suggesting similar overestimates 3.1 Broadsetting
2
for the remaining determinations. Also dropped at the out-
setwereseveralsingleO valuesfromotherexpeditionsthat The area of study lies, strictly speaking, between 7.5 and
2
appeared unreasonably high from the context (e.g., as com- 21.5◦Nand64to68◦E(Fig.1),butthesouthernboxes(A–
pared to values from adjoining depths, or which came from C,8–12◦N),whichareoutsidethesuboxicOMZ,aretreated
transitionlayerswithstronggradients). onlyinpassing.SketchingthegeneralsettingoftheArabian
The totals are 632 accepted measurements from 196 sta- Sea,wenotethatpreviouslythetwomonsoonswerethought
tionsbetween8and21◦N(SupplementTableS.1.b).Table1 to force reversal of surface currents seasonally over the en-
presentsthenumbersandmediansoftheseO valuesforall tirebasin.Actually,muchofthewesternhalfoftheArabian
2
boxesanddepths. Sea is full of cyclonic and anticyclonic, quasi-geostrophic
mesoscale eddies and fronts with their associated meander-
2.4 Nitrite ingcurrents(FlaggandKim,1998;Shankaretal.,2002;Ar-
tamonov, 2006; Resplandy et al., 2011). The new insight is
BecausetheO valueswithintheOMZareoftenveryclose
2 illustratedbymapsofsealevelanomalies(SLAs;Kimetal.,
tothelowerlimitofdetectionandhenceperhapsnotaspre-
2001),seasurfaceheight(Fischeretal.,2002;Welleretal.,
−
cise as is desirable, we use NO as a surrogate of near-
2 2002; Resplandy et al., 2011), and geopotential anomalies
absenceofO .Onallcruisesnitritewasdeterminedfollow-
2 (Artamonov,2006).Theeddiesandfrontsmayreach500m
ingBendschneiderandRobinson(1952)orvariantsthereof.
depth(e.g.,Artamonov,2006:Figs.3.15C–F,3.16B;Bobko
IncontrasttotheO2observations,thedatafromtheUSJoint and Rodionova, 2006: O at 300m in Fig. 5.6, NO− sec-
Global Ocean Flux Study (JGOFS) and World Ocean Cir- 2 3
tion in Fig. 5.8B). In addition, the salinity and temperature
culation Experiment (WOCE) programs were also included
(CTD) profiles at stations in the northern boxes down to σ
t
inourstudy.Forthefivehorizonsunderconsideration,1191 near 26kgm−3 (roughly 200m depth) show many salinity
data points from 292 stations (949 from 227 stations in the
spikesinthepycnocline(seeFigs.2and8)frominterweav-
OMZ) were utilized, all analyses having been taken at face
ing of relatively thin layers of varying salinity, presumably
value(SupplementTableS.1.b).Thesetotalscomprisesome
with varying biochemical histories. The cause probably is
averagesoftwoorthreereplicatecastsperstation,aswellas
densewaterfromwinterconvectionsubductedandadvected
meansof≥4(upto28)replicatedcastswithinthesameday
horizontally, then preserved in the pycnocline (e.g., Banse
−
orconsecutivedaysatfixedpositions.MostofthehighNO2 and Postel, 2009). Between σt of about 26–27kgm−3, sim-
valueswere<5µMexceptafewthatexceededthisconcen-
ilar layering is due to the intrusion from the Persian Gulf
tration(maximum6.2µM).Oneoutlierof10.2µMisfound
(the Persian Gulf Water, PGW, see Supplement Sect. S.3).
at200minBoxE2inJuly1970.Table2presentsthenum-
Theoverallresultof,especially,thehydrographicprocesses
bersofsamplesandthemediansforallboxesanddepths.
issubstantialvariabilityevenwithinstationsreplicateddur-
ingonetothreedays(Sect.3.2.4)andissuperimposedover
www.biogeosciences.net/11/2237/2014/ Biogeosciences,11,2237–2261,2014
2242 K.Banseetal.:OxygenminimumzoneoftheopenArabianSea
Table1.MedianO2concentrations(mLL−1)forallboxesneartheindicateddepths(numberofvaluesinparentheses;datainTableS.1.b).
One-sidedpforO2differencebetweenadjoiningmedians:*≤0.20;**≤0.10;***≤0.05);blank,non-significant.1
Box 150m 200m 300m 400m 500m
A1+A2 0.77(3) 0.99(5) 1.21(4) 0.92(1) 0.96(4)
*** *** ** **
B1 0.24(8) * 0.30(13) *** 0.72(7) 0.68(8) ** 0.48(9)
* ** ** ***
C1 0.39(2) 0.27(6) 0.36(3) 0.44(3) 0.36(6)
** *** ***
D1 0.07(9) 0.10(12) 0.04(9) ** 0.09(11) * 0.13(13)
*** ***
E1 0.22(10) ** 0.12(9) * 0.10(7) 0.09(8) ** 0.13(9)
* * **
F1 0.27(25) *** 0.14(23) 0.10(17) 0.09(19) 0.10(16)
* *
G1 0.30(19) *** 0.13(15) 0.15(12) 0.15(12) *** 0.10(7)
B2 0.16(10) ** 0.24(13) *** 0.67(11) * 0.59(9) * 0.54(12)
** ** **
C2 0.47(3) 0.37(5) * 0.25(2) 1.04(2) 0.34(4)
*** *** *** *** ***
D2 0.07(17) 0.06(23) 0.07(20) 0.11(25) 0.12(26)
***
E2 0.19(10) ** 0.04(5) 0.13(4) 0.09(5) *** 0.11(9)
F2 0.13(10) 0.09(7) 0.07(8) 0.06(8) 0.10(11)
* ** *** ***
G2 0.43(9) *** 0.08(9) *** 0.05(4) 0.04(5) 0.05(4)
1Forexample,for(A1+A2),150mcannotbedistinguishedfrom200m;at150m,however,(A1+A2)differsfromB1atp≤0.05.
marked seasonality even below the permanent thermocline rent,whichadvectsO poleward.Naqvietal.(2006)reported
2
(Sect. 4.1.2). Resplandy et al. (2011) in an eddy-resolving low salinity and elevated O near the continental slope off
2
(1/12◦)modelshowedthelargeroleofverticalnutrientsup- Goa (15◦N) even in December in 1998 after the SWM had
ply by eddies. However, as noted, for example, by Shankar lastedunusuallylongandthesurfaceflowwasstilldirected
et al. (2002), the regularity of the monsoons makes fea- towardtheequator.Intheireddy-resolvingmodel,Resplandy
tures like “monsoon currents”, which dominated the previ- et al. (2012) generate the undercurrent below about 150m
ous views of the circulation of the upper Arabian Sea, still depth and stress the importance of poleward O advection
2
stand out, including their relation to the depth of the prin- alongthecontinentalslope.
cipalthermocline.Forfurtherbackgroundinformation,Wig- ThegeneraldistributionofsalinityandO along64◦Eas
2
gertetal.(2005)reviewedbiogeochemicalpelagicprocesses. depicted by Olson et al. (1993: Fig. 2) for 1986 is proba-
RamaswamyandGaye(2006;Sta.EASTnear15◦N,65◦E) bly valid also for 65◦E. Between 200 and 500m depth the
describedtenyearsofbiogenicverticalfluxatapproximately median temperatures at each depth horizon increase with
3000mdepth. monotonousslopesby2–3◦C,withthemediansalong67◦E
Thestudyregionislargelyoutsidethestrongphysicaland (boxes B2–G2; Supplement Table S.2) tending to be cooler
biogeochemical activity offshore of the Arabian Peninsula by a few tenths of a degree than those along 65◦E (boxes
andneartheMurrayRidge.Similarlyonitseasternside,our A1–G1).Littlemorewillbesaidabouttemperature.Between
meridional band is largely beyond the influences of the sea about7and10◦N,thesalinitybelowthebottomofthepycn-
level changes, currents, and Kelvin waves near the Indian oclineatindividualstationsvarieslittlewithdepthtoatleast
Subcontinent. During the SWM period coastal upwelling 800m but increases northward from ca. 35.2 to 35.3–35.4.
reignsinthe“meso-eastern-boundary-currentregime”along To the north of 10 or 12◦N up to 21◦N and between 150
thewestcoastofIndia,withoutthelargeeddiesandoffshore- and 500m depth, the median salinities at each horizon in-
drawn filaments as seen off the western side of the basin. crease with monotonous slopes to 35.9–36.0 (35.8 at 500m
Associatedwiththeupwellingisthenorth-settingundercur- near65◦E;SupplementTableS.3).IntheOMZ,themedian
Biogeosciences,11,2237–2261,2014 www.biogeosciences.net/11/2237/2014/
K.Banseetal.:OxygenminimumzoneoftheopenArabianSea 2243
−
Table2.MedianNO concentrations(µM)forallboxesneartheindicateddepths(numberofvaluesinparentheses;datainTableS.1.b).
2
One-sidedpforNO−differencebetweenadjoiningmediansinboxesD–G:*≤0.20;**≤0.10;***≤0.05;blank,non-significant(seealso
2
footnoteinTable1).
Box 150m 200m 300m 400m 500m
A1+A2 0.01(10) 0.01(10) 0.01(7) 0.01(7) 0.01(7)
B1 0.02(14) 0.01(18) 0.00(15) 0.00(14) 0.00(14)
C1 0.03(12) 0.04(18) 0.00(11) 0.00(13) 0.00(12)
D1 0.13(30) 0.81(34) * 1.17(32) *** 0.00(29) *** 0.00(29)
*** *** *** ***
E1 0.02(15) *** 0.40(16) ** 2.86(15) *** 1.32(16) *** 0.04(11)
* *** ***
F1 0.04(36) *** 0.14(45) ** 1.36(41) ** 1.37(38) *** 0.43(37)
* *** **
G1 0.03(30) *** 0.23(34) 0.53(31) 0.75(32) ** 0.54(28)
B2 0.02(9) 0.00(9) 0.00(8) 0.00(4) 0.00(5)
C2 0.00(4) 0.00(6) 0.00(5) 0.01(4) 0.00(6)
D2 2.00(35) *** 2.83(40) *** 1.22(38) *** 0.01(36) *** 0.00(35)
*** *** *** ***
E2 0.08(16) *** 2.53(18) 2.74(15) *** 1.28(17) *** 0.06(17)
** ** **
F2 0.07(6) *** 1.74(9) *** 2.82(9) *** 1.44(7) ** 0.95(5)
* *
G2 0.04(12) *** 0.46(14) *** 2.67(14) *** 1.98(14) *** 1.26(13)
salinities tend to be slightly lower along 67◦E than in the 3.2 SettingoftheOMZ
western(along65◦E)boxes.
Overtherangeofabout7to10◦N,nearandsomewhatbe- 3.2.1 Generalhydrography
low200m,O asmeasuredpriortotheadventoftheSTOX
2
probe declines from about 2 to about 0.4mLL−1 (∼90 to We focus on the OMZ poleward of 12◦N. Our 150m, up-
18µM; cf. Wyrtki, 1971: Tables 441 and 502; see also the permosthorizonisbelowthe salinity maximumofthenon-
numerous significant differences among median O2 values seasonal pycnocline near a σt of 24kgm−3 and below the
between the C and D boxes at 10 and 12◦N, respectively, sharp oxycline, which is the upper border of the OMZ.
in Table 1). Naqvi et al. (1993) noted that the transitional Thedepthsofthesediscontinuitiesvaryseasonally(Colborn,
zone is observed between these two latitudes, which during 1975;Molinarietal.,1986),beingdeeperintheSWMand,
1995 was only one degree wide. The position seems to be inthenorth,alsoduringtheNEM.Theclimatologicaldepths
fairly stable in time, probably owing to the distribution of of the bottom end of the steep vertical gradients of temper-
wind stress and the resulting quasi-zonal circulation (War- ature,salinity,andoxygenoccursomewhatabovethe200m
ren,1994;butseethesectionfortheSWMindeSousaetal., horizon (see Fig. 2 as an example). Much of the OMZ has
1996, and Supplement Sect. S.1). The very steep horizontal onlyweakverticalgradientsofO2,butzooplanktonspecies
O gradients at the southern edge of the upper half of the may be layered, perhaps responding to these gradients, as
2
OMZ in our meridional band, not described previously, are found with copepods (Böttger-Schnack, 1996) and Wishner
tobecoveredbyBanseandPostel(2015). etal.(1998).
3.2.2 Oxygen
The intense (suboxic) OMZ extends from north of the C-
boxes (12◦N) poleward, although in September 1994 the
depthsatandslightlyabove200minourBoxC1werepart
www.biogeosciences.net/11/2237/2014/ Biogeosciences,11,2237–2261,2014
2244 K.Banseetal.:OxygenminimumzoneoftheopenArabianSea
ofthiszone(R/VT.G.ThompsoncruiseTN039;seealsothe Oxygen (mL L-1)
variablelatitudeofthedenitrificationzoneinNaqvi,1991). 0.0 2.0 4.0 6.0
0
In the OMZ, the median visual Winkler O levels are often
2
below0.1oreven0.05mLL−1 (Table1).Themanysignifi-
cantdifferencesinthetablebetweenthe150and200mhori-
100
zonssuggestrestrictingourtreatmentoftheOMZtothe200 0
to 500m levels but adding the 150m level in the D-boxes Oxygen
(15◦N).Themediansforthefourdeeperhorizonsrangeonly
200
between 0.04 and 0.15mLL−1 O2. Within the upper OMZ m) 100 De
aosftsetnudthieadnhneortet,htehdeivffeerrteicnaclegsrbaedtiwenetesntednedptthohboeriszmonasll;armeoinre- epth ( Salinity pth (m
D300 )
significant,asarethenorth–southdifferenceswiththeexcep- 200
tion of the G-boxes (Table 1). Regarding zonal differences, 26
the means of the annual medians of the depths in the four 400 C) 24
westernboxesareallhigherthanthoseoftheeasternboxes, ure (° 2220
but the ranking of the medians in Table 1 shows significant at 18 Time
differencesonlyat200and400m(p=0.01and0.2,respec- 500 mper 16 0079::1050
tively). Te 14 1125::4050
12
Resplandyetal.(2012),withtheireddy-resolvingmodel, 35.4 35.6 35.8 36.0 36.2 36.4
Salinity
studied the factors that control the O balance in the OMZ.
2 500
For a depth range of 250–300m their computed median
35.6 36.0 36.4 36.8
O2 utilization rate was 2.8µmolL−1a−1 (range: 2.1–6.8 Salinity
[0.06mLL−1a−1 (0.05–0.15)] L. Resplandy, 2013, in litt.,
◦ ◦
Fig.2.Sta.4010ofR/VSagarSampadanear19 N,67 E(circle
cf.theirFig.6aCAS).TheyalsoreportedfortheNEMand
SWM that the amplitude of seasonal O change from verti- inFig.1)inJanuary1998.(a)Salinityand(b)O2recordsfromthe
2
CTDcaststakenover25h,mostofthemhourlyuntil16:45LT.The
cal eddy-driven advection was several times that from bio-
shipdriftedbyabout41kmbetweenthefirstandlastcasts;(c)T–S
logical consumption. Similarly, McCreary et al. (2013) in a
diagramforfourofthecastsduringthefirst8h(drift,about13km).
non-eddyresolving(0.5◦)modeldeterminedthelargeroleof
lateralO advectiontothecentralArabianSea.Neitherpaper
2
addressedshiftsontheclimatologicalscaleaswedohere.
3.2.4 Day-to-dayandweeklyvariability
3.2.3 Nitrite
Using four examples, the section demonstrates that a num-
WithintheOMZandincontrasttoO ,Table2showsmany
2 ber of relatively small water masses with apparently differ-
−
significant depth differences for NO in spite of the great
2 ent biochemical history are frequently present and often on
variability of concentrations. The values are large at and
thesamedensitysurfaces.Thehighlydynamicnatureofthe
poleward of 15◦N except near 150m (Table 2). At 15◦N
hydrography creates temporal as well as spatial variability
−
(D-boxes),however,thehighNO valuesatthe150mhori-
2 eveninthecentralArabianSeaatthesamestationswithre-
zonindicatefairlyintensedenitrificationinconjunctionwith
peated casts and is especially noticeable on week-long drift
low oxygen. In contrast, there is little denitrification at 400
stations.Fortheareatreatedherein,previouspapersempha-
and500mat15◦NandlittleintheE-boxesat500mdepth,
sizedtemperature(isopleths),twiceforfourdaysinBoxE2
reflecting the trends to slightly higher O (medians, 0.10–
2 (Ramesh Babu et al., 1981), as well as one and two weeks
0.13mLL−1).Forthesamereason,theNO−mediansat300
2 of mean isopleths of four stations centered in Box C1, both
and 400m in Box G1 decline relative to the adjoining Box
to200m(Rao,1987),ortemperatureandsalinityfor13days
F1. The low medians at 150m in boxes E1 and F2 and at
(PrasannaKumaretal.,2001;Naqvietal.,2002).
400m in the D-boxes, though, hide the fact that several to
InTable3thefirsttwoexamplespresentamong-daysvari-
many high values were present along with zero concentra- −
ability for temperature, salinity, O , and NO on stations
tions(seeSupplementTableS.1.b).Themapofhighandlow 2 2
nearBoxD1andinBoxE1,eachmaintainedwithin1–2km.
−
valuesofNO atthe250mhorizoninBobkoandRodionova
2 The standard deviations (SD) for O2 are absolutely small,
(2006)fromthecentralandnorthernArabianSeaillustrates
but may be as high as the low means. On both stations the
the role of downwelling from eddies to appreciable depths −
variabilityofNO at200–400misverylarge(thevariance
(cf.Artamonov,2006). 2
being larger than the mean), in part because of zero values
but in part so even among the high concentrations, as will
alsobeapparentinlatersections.Measurementswithfewer
repeats on seven other locations in our data sets, including
Biogeosciences,11,2237–2261,2014 www.biogeosciences.net/11/2237/2014/
K.Banseetal.:OxygenminimumzoneoftheopenArabianSea 2245
Table3.MeanvalueswithSDondriftstations(numberofsamplesinparentheses).
150m 200m 300m 400m 500m
SK99:17,19–21February1995,westofD1*(O2automatedendpoint)
T (◦C) 18.95±0.50(7) 16.63±0.27(7) 13.96±0.09(8) 12.86±0.07(6) 12.18±0.09(5)
S 35.69±0.04(7) 35.73±0.01(7) 35.64±0.01(8) 35.61±0.01(6) 35.60±0.01(5)
O2(mLL−1) 0.30±0.065(6) 0.05±0.03(4) 0.05±0.03(6) 0.08±0.08(7) 0.09±0.09(7)
NO−(µM) 0.06±0.11(7) 0.90±0.85(6) 2.36±0.47(7) 0.05±0.08(5) 0(5)
2
ME32/3:243-252,10–18May1995,E1**(O2automatedendpoint);T andSincludesta.240of9May
T (◦C) 20.94±0.45(9) 17.54±0.13(11) 15.10±0.27(12) 13.57±0.10(11) 11.49±0.10(11)
S 35.94±0.13(9) 35.62±0.02(11) 35.82±0.06(12) 35.76±0.02(11) 35.57±0.01(11)
O2(mLL−1) 0.31±0.31(11) 0.08±0.015(10) 0.09±0.02(5) 0.08±0.02(7) –
NO−(µM) 0.04±0.02(13) 3.54±0.77(10) 2.86±0.82(4) 2.36±0.58(7) –
2
SK121:8,10–22February1997,westernedgeofG1***(O2visualendpoint)
T (◦C) 20.23±0.57(24) 18.08±0.25(20) 15.78±0.38(23) 14.05±0.26(23) 13.06±0.19(23)
S 35.99±0.06(22) 35.89±0.06(18 36.03±0.07(23) 35.85±0.05(23) 35.77±0.03(23)
O2(mLL−1) 0.25±0.20(27) 0.04±0.04(24) 0.06±0.04(26) 0.05±0.04(25) 0.06±0.07(27)
NO−(µM) 0.03±0.06(27) 0.57±0.79(24) 0.38±0.86(27) 0.85±0.99(27) 0.34±0.60(27)
2
*Driftingoverarangeofabout1km.
**Driftingoverarangeofabout2km.
***Driftingoverarangeofabout45km,seeNaqvietal.(2002:Fig.1)andSupplementFig.S.3.1.
U.S.JGOFScruises,otherthanthefollowingonesshowsim- exclude each other spatially below the O threshold for the
2
ilar variability. The third set of data in Table 3 come from onset of denitrification of <0.002mLL−1 (<0.09µM). As
a drift station near 21◦N, where temporal and spatial vari- stated,21%of707samplescontained≤0.02µMNO−;the
2
ability change cannot be separated, but the spatial range of majorityshowedzeroor0.01µMNO−;theaveragesinclude
2
almost half a degree is relevant to comparisons of stations NO−datanotaccompaniedbyO (SupplementTableS.1.b).
2 2
on sections with one-degree spacing (see also Supplement Inviewoftheaforementionedsalinityspikesreflectingstrat-
Fig.S.3.1). ification (also Fig. 2), we visualize the dimensions of such
The fourth example (Fig. 2) illustrates within-station and patcheshorizontallytobemuchlargerthanvertically.Onthe
within-dayvariabilityofhydrographysouthofBoxD2by16 average the patches must last many months, if not a year,
castsduringa28hdriftofabout41kmextentwherespatial such that planktonic animals (e.g., copepods) live and per-
and temporal inhomogeneity came into play. The T–S dia- sist in an otherwise nearly anoxic milieu of a few tens of
gram for the first 8h (Fig. 2c) shows massive replacement nanomolesofO (seealsothelarge“patches”foundin1960
2
of water between about 100 and 300m, below the principal asfreeofNO−attheendofSect.3.2.4).
2
pycnocline.
Last,aninklingoflarge-scaledistributionsisprovidedby
3.2.6 Animallife
−
the NO at 250m depth on a one-half to two-thirds degree
2
grid during the spring intermonsoon of 1990, coupled with
ThissectionandSupplementSect.S.4showthateventheup-
the associated O2 at the horizon and vertical sections indi- perOMZoftheArabianSeawiththeNO− maximumisbi-
catingdeep-reachingeddiesinBobkoandRodionova(2006: 2
ologicallynotfunctionallyanoxic,incontrasttothesugges-
Figs.5.6and5.11B).
tionbyThamdrupetal.(2012)formuchoftheupperOMZ
The geographical heterogeneity is also to be expected
in the eastern South Pacific. Metazoan zooplankton, which
in regional surveys or in interannual studies at fixed loca-
byitsnaturerequiresdissolvedO ,isfoundyearround.As
2
tions.Therefore,datafromafewstationsmustnotbeover- mentionedabove,NO−≤0.02µM(i.e.,zerowithinthepre-
interpreted. 2
cisionoftheanalysis)wasmeasuredinone-fifthofoursam-
plesfromthe200to400mhorizonsoftheupperOMZwith
3.2.5 Oxygenandnitriteco-occurtemporally itsSNM.ThusenoughO ispresentinsuchsamplestopre-
2
vent denitrification and is sufficient to maintain a reduced
Theobservationsabouthydrographicvariabilitysupportour numberofanimalspeciesandspecimens.
assumption set forth toward the end of Sect. 1 that O and Supplement Sect. S.4 reviews results of zooplankton col-
2
−
NO may co-exist temporally in the OMZ. Of course, they lectionsfortheplanktonminimumintheupperOMZofthe
2
www.biogeosciences.net/11/2237/2014/ Biogeosciences,11,2237–2261,2014
2246 K.Banseetal.:OxygenminimumzoneoftheopenArabianSea
−
ArabianSea.Onlynighttimedataareconsidered,whichex- accompaniedbyNO ,isapttobeduetomixingdownward
2
clude the diel migrators among plankton and nekton enter- fromthedepthofmaximum.InthenortheasternArabianSea,
−
ing the top of the OMZ during daytime but which during Naqvi et al. (1990) observed lower NO deficits during the
3
the night in the mixed layer pay off the O debt incurred SWM period of 1987 than during the NEM period of 1988
2
at depth. The lowest specimen numbers of a few species of (2–3µMversus8–9µM,respectively).
copepodswere10–20m−3(Böttger-Schnack,1996)andsev-
eraltensper1000m3ofindividualspeciesofcalanoidcope-
pods(Wishneretal.,1998).Wetweightsseemtorangefrom 4 SeasonalanddecadalvariabilityintheOMZ
1 to 5mgm−3. The first author remarked on the large frac-
Besides a T–S diagram of seasonal medians for the boxes
tionofcopepodcarcassesandexoskeletons.Ignatyev(2006)
(Fig.3),theprincipalmethodofstudyislinearregressionof
reported that on the average one-fifth to two-fifths of wet
thevariableofinterestonyear(decimalfractionsaccording
weight caught by mid-water trawls of 0.5mm mesh size
to the month of observation). Decadal changes of tempera-
consistedofunidentifiableremaindersofgelatinousanimals.
ture are not considered because linear regressions of sam-
Thesewereinadditiontotheone-thirdofwetweightofcoe-
pling depth on year unexpectedly showed positive or nega-
lenteratesinanticycloniceddiesandthree-fifthsoftunicates
tivetrends.Weneglectthissourceofpossiblebiasforsalin-
aloneincycloniceddies.
ity,oxygen,andnitritebecauseofthemuchsmallervertical
Below the minimum in slightly better oxygenated wa-
gradientsatthehorizonsofconcern.
ter, copepod numbers tended to increase tenfold, as did the
biomass, before declining toward depth at >1km. Species
4.1 Temperatureandsalinity
occurrenceinthesecondarybiomassmaximumwaslayered,
presumablyinresponsetoO (Wishneretal.,1998).
2 4.1.1 Temperature
TheanimaldistributionintheupperOMZoftheArabian
Seaisnotunique.Similarfeatures,includingtheminimaof
By the climatology of temperature of the upper 500m, the
biomassandnumberofspecies,arefoundintheOMZofthe 20◦C isotherm in the middle of the thermocline in our re-
Costa Rica Dome of the eastern tropical Pacific, including
gionmovesseasonallybyabout50m(Molinarietal.,1986).
thesecondarymaximumofbiomassandspeciesoccurrence
Ramesh and Krishnan (2005) studied the seasonal down-
inthelowerpartoftheOMZ(SaltzmanandWishner,1997). ward flux of heat to ≥200m as zonal averages between 50
Inthesameregion,Vinogradovetal.(1991),usingplankton and 70◦E, for 0–20◦N. Presumably, O is also transported
2
haulsandvisualobservationsfromasubmersible,confirmed
(cf.Sect.1).Theycalculatedthatdownwellingfromconver-
the similarity of vertical patterns between the Arabian Sea
genceandhorizontaladvection,andthedeepeningandcool-
andtheCostaRicaDomebutaddedthelargecontributionof
ingoftheupperlayercoupledwithstrongverticaldiffusivity
biomassbygelatinousanimals,whichareapartoftheOMZ
fromwind-causedturbulence,warmedthethermoclinenear
fauna.Jelly-likeanimalsaccountedfor92%ofwetmassand 150,200,and250mdepthssouthof12–13◦Nbyabout1.2,
16%ofcarbontothemeso-andmacroplanktonintheupper 0.9, and 0.4◦C, respectively. Their Figs. 6, 8, and 9 also il-
500m, omitting very large (>15cm) but rare medusae and
lustratethattheOMZareaofourstudyislargelyoutsidethe
ctenophores (see also Hamner et al., 1975, and Supplement
centersofthegreatesthydrographicandassociatedchanges
Sect.S.4).
inthewesternandsouthernArabianSea.Finally,the2-year
time series by an Argo float, launched at 12◦N, 65◦E and
3.2.7 Nitratedeficit slowly drifting toward the southeast (Prakash et al., 2012),
illustrates an apparently seasonal vertical shift of the top of
−
The maximum of the NO deficit resulting from denitrifi-
3 theoxycline,theextremesrangingbetween35–40and130m
cation is also observed in the core of the OMZ, although
depth(duringtheearlyNEMandtheSI,respectively).
−
tending to be a few tens of meters deeper than the NO
2
maximum (Morrison et al., 1999: Fig. 8); the deficit was 4.1.2 Temperature–salinityrelations
calculated by the NO approach, which below about 250m
depthmightyieldslightoverestimates(NaqviandSenGupta, OurT–S diagram(Fig.3;cf.Fig.1)isbasedonmediansof
1985: Fig. 2). Other authors are Mantoura et al. (1993) and discretedataandpresentstheclimatologyfortheOMZinour
Chang et al. (2012). The maximal values in a season or a meridional swath. Obvious are the strikingly differing pat-
cruise range between about 2 and 15µM. As illustrated by terns at the 150 and 200m horizons down to the isopycnals
Morrison et al. (1999) for 1995, the deficit declines toward (σ) of 26.3–26.4kgm−3 versus those at the 300 to 500m
t
−
depth and vanishes by 600–900m, while NO is not ob- levels.Theformeraredominatedbyseasonal,seeminglydi-
2
servedbelow350mdepth(below600matoneoftheirfour apycnal warming to >200m depth (but <300m) in boxes
−
stations).However,notethatthepresenceofanNO deficit B–G, extending well below the permanent pycnocline. In
3
doesnotreflectongoinganoxia.Rather,thedeficitregularly contrast, the three lower horizons exhibit clear isopycnal,
observed in the OMZ well below the deficit maximum, not seasonally periodic change at and poleward of the D-boxes
Biogeosciences,11,2237–2261,2014 www.biogeosciences.net/11/2237/2014/