Deep-SeaResearchII49(2002)2377–2402 Distribution, annual cycle, and vertical migration of acoustically derived biomass in the Arabian Sea during 1994–1995 Carin J. Ashjiana,*, Sharon L. Smithb, Charles N. Flaggc, Nasseer Idrisib aDepartmentofBiology,WoodsHoleOceanographicInstitution,WoodsHole,MA02543,USA bRosenstielSchoolofMarineandAtmosphericSciences,UniversityofMiami,4600RickenbackerCauseway,Miami,FL33149,USA cDepartmentofAppliedScience,BrookhavenNationalLaboratory,Upton,NY11973,USA Received10April2001;receivedinrevisedform1November2001;accepted2December2001 Abstract ThedistinguishingcharacteristicthatsetstheArabianSeaapartfromotheroceanicregionsistheregularoscillation of monsoonal atmospheric conditions that produces predictable periods of upwelling or convective mixing, with associated biological response, during the Southwest and Northeast monsoons, respectively. This oscillation is also evidentincycles of standingstocksofzooplankton andmicronekton. Theverticaldistribution andspatialpattern of zooplankton and micronekton biomass were estimated using an acoustic Doppler current profiler along a 1000-km transectextendingfromthecontinentalshelfofOmantothecentralArabianSeaduringtencruisesontheR/VThomas G. Thompson (November 1994–December 1995). The influence of the Southwest Monsoon, and accompanying upwellingandenhancedacousticallyderivedbiomass,wasthedominantfeatureinthespatial-temporaldistributionsof bothzooplanktonandmicronektonneartheOmanicoast.Thedielverticalmigrationofpredators(myctophids,pelagic crabs),andtheseasonalchangesinthestrengthofthissignal,wasthemostsignificantpatternobservedinthevertical distribution of biomass and imparted a strong day–night signal to the integrated upper water-column biomass. Significant differences in the magnitude of integrated upper water-column biomass, both zooplankton (day) and migrator-zooplankton (night), were seen between inshore and offshore of the atmospheric Findlater Jet. A station locatedinthecentralArabianSeademonstratedseasonalchangesinbiomassovertheyear,despitebeingquitefarfrom the influence of the monsoonal oscillations. Predation pressure was greater offshore of the Findlater Jet than in the regioninshoreoftheJetorinthecentralArabianSea.ThepelagiccommunityoftheArabianSeamayhaveevolvedlife history strategiesto coincidewith thepredictable monsoonalcycle. r2002Elsevier Science Ltd. Allrights reserved. 1. Introduction ocean’s response to oscillating monsoon winds. The Arabian Sea is characterized by a regular, Seasonal cycles in abundance and distribution seasonal monsoon cycle consisting of the North- of zooplankton and micronekton in the northern east Monsoon (NEM: December–February; pre- Arabian Sea are influenced profoundly by the vailing winds from the NE), and the Southwest Monsoon (SWM: June–September; strong, pre- vailing winds from the SW, centered on the *Correspondingauthor. E-mailaddress:[email protected](C.J.Ashjian). troposphericFindlaterJetwhichrunsSW:NEwell 0967-0645/02/$-seefrontmatterr2002ElsevierScienceLtd.Allrightsreserved. PII:S0967-0645(02)00041-3 2378 C.J.Ashjianetal./Deep-SeaResearchII49(2002)2377–2402 offshore of the coast of Oman), punctuated by onshore–offshore declining gradient in mesozoo- Intermonsoon periods (e.g., Currie et al., 1973; plankton biomass (e.g., Qasim, 1977; Smith et al., Schott et al., 1990; Morrison et al., 1998). The 1998b; Wishner et al., 1998). The community prevailing winds have dramatic effects on the composition of the zooplankton also may be hydrographic and advective fields of the region, modified during the SWM; some copepod species with attendant influences on biological processes (e.g.,Calanoidescarinatus)increasedinabundance and distributions. During the SWM, upwelling of in greater proportions than other species (e.g., nutrient-rich water off Oman promotes primary, Smith, 1982; Smith et al., 1998b). and hence secondary, production (e.g., Banse, The influence of the monsoonal oscillation on 1987; Brock et al., 1991, 1992; Brock and hydrographic and advective structure, the atten- McClain, 1992). The offshore (100–400km) and dant changes in plankton and fish biomass, alongshore (>1000km) extent of relatively cool production, community structure, and the utiliza- sea surface temperatures (o261C; or approxi- tion and cycling of carbon, were addressed during mately 41C cooler than the adjacent sea surface the multidisciplinary Arabian Sea Expedition of not affected by upwelling) make this one of the 1994–1996underthejointauspicesoftheUSJoint largestupwellingareasknown.Onlytheupwelling Global Ocean Flux Study (JGOFS; National areasassociatedwiththeBenguelaCurrentSystem ScienceFoundation)andtheForcedUpperOcean (LutjeharmsandStockton, 1987)havedimensions Dynamics Project (Office of Naval Research). similartotheoneoffofOman;allotherupwelling During the study period, a series of ten cruises regions (Peru, 1983; Somalia, Brown et al., 1980; were conducted on the R/V Thomas G. Thompson northwestAfrica,Estrada,1974)arelessextensive. inthenorthernandcentralArabianSeaoverafull The situation is different during the NEM, when year, spanning the complete monsoon cycle. The cool, dry winds from the northeast cause con- Thompson was equipped with a 153kHz RDI vective mixing and associated primary productiv- acoustic Doppler current profiler, which operated ity (Marra et al., 1998; Madhupratap et al., 1998; continuously throughout all of the cruises. The Barber et al., 2001). The two monsoonal regimes presentstudyutilizesthebackscatterintensitydata are separated by Spring and Fall Intermonsoon collected by this instrument to analyze spatial and periods, characterized by a reduction of wind seasonalpatternsinzooplanktonandmicronekton stress over the ocean surface and consequential biomass across a 1000-km transect of the Arabian diminishing of the upper mixed-layer depth Sea. The transect was situated so that the atmo- (Gardner et al., 1999). As a result, primary spheric Findlater Jet (SWM) crossed the center production and phytoplankton standing crop are (Fig.1) so that the transect surveyed regions both minimal during the Spring Intermonsoon (Gun- inshoreandoffshoreofthelociofstrongestSWM derson et al., 1998; Marra et al., 1998; Caron and winds. The abundance and distribution patterns Dennett, 1999; Gardner et al., 1999). This annual documentedbytheADCPspananentireyearand oscillation should be evident in cycles of standing were obtained continuously across the transect, in stock of zooplankton and micronekton, especially contrast to the net-based sampling, which was in the upwelling region within 600km of limited to a smaller subset of the cruises and the Omani coast (Smith et al., 1998b; Wishner occurred at discrete locations (Smith et al., 1998b; et al., 1998). Wishner et al., 1998). Hence, the ADCP data The SWM, and its associated upwelling and describe the temporal and spatial patterns with enhanced primary production, is accompanied greater temporal and spatial resolution than by an elevation in zooplankton biomass (e.g., possible from the net-based sampling and compli- Qasim, 1977; Smith and Codispoti, 1980; Smith, ment the results of those studies. 1982, 1984; Matthew et al., 1990; Baars and Acoustic Doppler current profilers (ADCPs) Oo.sterhuis, 1997; Brink et al., 1998; Smith et al., have been used to describe the spatial and 1998a; Wishner et al., 1998). Another pattern temporal patterns in the distribution of zooplank- observed during the SWM is a pronounced ton biomass in many oceanic regions and C.J.Ashjianetal./Deep-SeaResearchII49(2002)2377–2402 2379 scattering off of the organisms collected in the net tows. In reality, organisms that are not collected quantitatively in zooplankton net tows, such as fish, fast swimming euphausiids, and siphono- phores, also may contribute substantially to the backscatter intensity (Wiebe and Greene, 1994; Holliday and Pieper, 1995; Stanton et al., 1993, 1994). However, analyses of patterns in biomass derived from ADCP backscatter intensity have demonstratedthattheinstrumentprovidesreason- able and biologically meaningful descriptions of the biomass field and its association with hydro- graphic characteristics. Acoustic Doppler current profilers are particularly adept at documenting changes in the vertical distribution of biomass in the water column associated with the diel vertical migration of zooplankton and micronekton (Plueddemann and Pinkel, 1989; Buchholz et al., 1995; Heywood, 1996; Ashjian et al., 1998; Luo et al., 2000). The diel vertical migration of acoustically derived biomass from the ADCP in the Arabian Sea is likely to be dominated by the migration of Fig.1. Location of transect in the northern Arabian Sea. myctophidfish,sincetheregionischaracterizedby Distancesalongthetransectarecalculatedasdistanceoffshore. these fish which produce a strong backscatter Standard stations (S1–S11, S15) for the US JGOFS Process intensity signal, and, seasonally, by the vertically studycruises(Smithetal.,1998a)areshown.Theapproximate locationoftheatmosphericFindlaterJetrelativetothetransect migrating pelagic crab Charybdis smithii (Currie and standard stations is demonstrated. Seasonal upwelling et al., 1973; Nafpaktitis, 1978; Gj(cid:1)saeter, 1981, occurs inshore of the Findlater Jet (S2–S4) during the SW 1984; Ropke et al., 1993; Van Couwelaar et al., Monsoon. 1997; Herring et al., 1998; Luo et al., 2000). The pelagic crab C. smithii has been observed in high abundanceintheupperwatercolumnduringboth environments (Flagg and Smith, 1989; Pluedde- the SW and NE monsoons. This species ranges in mann and Pinkel, 1989; Heywood et al., 1991; length from 6–70mm and should contribute Smithetal.,1992;FischerandVisbeck,1993;Roe substantially to the backscatter intensity when and Griffiths, 1993; Ashjian et al., 1994; Flagg present. Comparison of backscatter intensity dis- et al., 1994; Lyons et al., 1994; Zhou et al., 1994; tributions collected simultaneously using a Batchelder et al., 1995; Buchholz et al., 1995; 153kHz ADCP and 12kHz sonar in the Arabian Heywood, 1996; Ashjian et al., 1998; Luo et al., Sea suggested that the strongly migrating layers, 2000). In such studies, backscatter intensity observed by both acoustic instruments, were routinely is converted to equivalent zooplankton composed of fish whereas the non-migrating biomass using a biomass-intensity regression layers, seen only in the ADCP backscatter derived from net-collected zooplankton samples intensity, were attributed to zooplankton (Luo and spatially and temporally coincident acoustic et al., 2000). Although mesozooplankton also backscatter intensities. Such a relationship pro- migrate on a diel basis in this region (Smith et al., vides only a rough estimate of the potential 1998b; Wishner et al., 1998), the larger size of the biomass in the water column, since it assumes micronekton should produce most of the back- that the total backscatter intensity is derived from scatter intensity signal for the vertically migrating 2380 C.J.Ashjianetal./Deep-SeaResearchII49(2002)2377–2402 layers. Biomass in the upper water column should currentprofilerwasoperatingcontinuouslyonthe be dominated by non-migrating copepod species ten cruises of the Arabian Sea study, greater during day but overwhelmed by the biomass of temporal (diel and seasonal) and spatial (horizon- verticallymigratingfishandpelagiccrabsatnight. tal and vertical) resolution and coverage of the The vertically migrating myctophids and pelagic distributionofplanktonandnektonwereachieved crabs are important predators because they feed than was possible from the net-based sampling primarily during night on zooplankton of the programsalone(Smithetal.,1998b;Wishneretal., upper water column (e.g., Gj(cid:1)saeter, 1984; 1998). Dalpadado and Gj(cid:1)saeter, 1988; Kinzer and Schulz, 1991; Kinzer et al., 1993; Van Couwelaar et al., 1997). 2. Methods The goals of the present study were to (1) document the seasonal changes in integrated Tencruiseswereconductedin1994–1995across biomass, particularly those associated with the a transect extending from near the coast of Oman changing monsoonal periods, across a 1000-km (181300N, 571180E) to 1000km offshore (141270N, transect in the northern Arabian Sea, (2) contrast 651000E) on the R/V Thomas G. Thompson during the magnitude and cycles of biomass inshore and both the US JGOFS Arabian Sea Process Study offshoreof the atmospheric Findlater Jet and also and the Forced Upper Ocean Dynamics Program at a reference station located in the central (Table 1; Fig.1). Four of the cruises surveyed the Arabian Sea removed from the influence of the transect line twice, resulting in 14 total transects. monsoons, (3) describe seasonal and spatial Datawerecollectedcontinuouslyin5minaverages changes in the diel vertical migration behavior of using a hull-mounted, 153kHz RD Instruments themigratingbiomass(myctophids,pelagiccrabs), acoustic Doppler current profiler (ADCP). Most and (4) evaluate the potential predation pressure datawerecollectedin8mdepthbins,however4m of these vertically migrating myctophids in the depth bins were used over the shelf. The ADCP different regions. Because the acoustic Doppler backscatter intensity data were calibrated and Table1 Datesandcruisenumbersofthe14transectsusedinthisstudy Cruise–Transect Dates Season CentralArabianSea TN042-outgoing November30–December9,1994 NEMonsoon TN042-ingoing December9–13,1994 NEMonsoon TN043 January17–31,1995 NEMonsoon * TN044-outgoing February10–16,1995 NEMonsoon TN044-incoming February16–21,1995 NEMonsoon TN045 March23–April7,1995 Intermonsoon * TN048-outgoing June26–July4,1995 SWMonsoon TN048-incoming July4–9,1995 SWMonsoon TN049 July30–August11,1995 SWMonsoon * TN050 August30–September12,1995 SWMonsoon * TN051-outgoing September21–October2,1995 Intermonsoon TN051-incoming October2–5,1995 Intermonsoon TN053 November6–18,1995 NEMonsoon * TN054 December12–26,1995 NEMonsoon * Cruises TN043, TN045, TN049, TN050, TN053, and TN054 were conducted as part of the US JGOFS study (National Science Foundation).CruisesTN042,TN044,TN048,andTN051wereconductedaspartoftheForcedUpperOceanDynamicsProgram (OfficeofNavalResearch).Theseasonsofthemonsoonsampledbyeachcruisearenotedand,withtheexceptionofTN044-incoming, areidenticaltothosedefinedbyMorrisonetal.,1998.CruisesduringwhichthecentralArabianSeastationwassampledalsoare noted(*). C.J.Ashjianetal./Deep-SeaResearchII49(2002)2377–2402 2381 range-correctedaccordingtothemethodsofFlagg using the functional regressions of Wiebe et al. andSmith(1989)andRDInstruments(1990).The (1975) and Wiebe (1988). The biomass-intensity ADCPdatawereenteredintoandaccessedfroma relationship was derived by comparing zooplank- Common Oceanographic Data Access System ton dry weight from discrete depth ranges with (CODAS) database, developed by E. Firing at spatially and temporally averaged coincident the University of Hawaii. The ADCP transducer backscatter intensity data from the ADCP. The ontheThompsonismountedonapodthatextends overall correlation between backscattered inten- below the hull, producing a remarkably low-noise sity, in dB, and log of the dry weight, in mg/m3, environment and one free of bubble interference was 0.68. A Model 2 regression was used to frombubble-sweep-down.TherangeoftheADCP convert all backscatter intensities (dB) in the excluded the upper B20m of the water column database to approximate equivalent dry weight, because of the depth of the instrument on the hull using the following relationship: Log (dry 10 of the ship and the blanking interval of the weight)=5.996+0.0662(cid:2)intensity. Data from instrument. The deck unit of the ADCP was both day and night were utilized in deriving this located in a climate-controlled laboratory that relationship. It is likely that the 1m2 MOCNESS experienced constant temperature during the undersampled the night biomass because the net period of data collection. may not effectively capture fast swimming micro- Duringsixofthecruises,alocationfaroffshore nekton. Unfortunately, sampling with a net in the Arabian Sea (101000N, 651000E), removed appropriate to capture fast swimming micronek- from the influence of the monsoons and the ton, such as a 10m2 MOCNESS, was not oxygen minimum zone, was sampled to serve as conducted from the Thompson during these a reference point (station S15) (Fig.1; Table 1). cruises, so estimates of micronekton abundance The ADCP datafrom this location were extracted were not available. Hence the biomass estimates and used to verify the consistency of the ADCP derived from the regression, particularly for the data over the course of the study period, and to datacollectedduringthenight,mayunderestimate establishabaselineofminimal seasonalityintotal the actual biomass present. biomass and diel vertical migration behavior to All data between the beginning and end points compare with those characteristics observed along of the transect were extracted from the database the 1000-km transect line (Fig.1). Examination of without vertical or temporal averaging. The the backscatter intensity from this location re- various cruise tracks frequently did not lie vealed little variation in minimum intensity preciselyalongthetransect.Excursionsbothalong (noise), but a degradation in the penetration of and off the transect occurred during stations and theacousticsignalanddepthrangeovertheperiod during the execution of grid surveys by towed of the study. instruments (cruises TN042, TN044, TN048, Vertically discrete zooplankton samples were TN051) (Lee et al., 2000). Furthermore, station collected using a 1m2 Multiple Opening/Closing keeping resulted in multiple time periods being Net and Environmental Sensing System (MOC- surveyed at a single location. Hence, a secondary NESS; Wiebe et al., 1976), equipped with 150mm data set consisting of data along the transect line meshnets,duringthreeofthecruises(Smithetal., was derived from the complete data set by 1998b). These samples were used to derive an removing observations collected during station empirical relationship between backscatter inten- keeping and then projecting the position of each sity from the ADCP and zooplankton biomass. off-track ADCP profile orthogonally back onto Zooplankton biomass was determined from the the transect line. The primary and secondary data net plankton samples by first determining the sets were plotted as vertical sections as a function displacement volume of each preserved sample of time (primary data set) or distance offshore using methods modified from Ahlstrom and (secondary data set) for the 14 transects. Thrailkill (1963) and then converting the displace- Integrated water-column biomass was calcu- ment volume to equivalent dry weight (mg/m3) lated for both the 1000-km transect (edited data) 2382 C.J.Ashjianetal./Deep-SeaResearchII49(2002)2377–2402 and the offshore station over a 20–120m depth fromthedatafield)basedonfieldvariance,sensor interval. Because the ADCP is not located at the error, and the uncertainty in the estimation of the sea surface, but mounted on the hull of the ship, mean. datawerecollectedonlybelow20m.Thedepthsto Diel vertical migration (DVM) in the data was which reliable estimates of backscatter intensity identified byexamining thevertical distributionof (% good >92; e.g., Flagg and Smith, 1989) were biomass relative to the timing of day and night obtained shoaled dramatically during night, pre- overthesampledwatercolumn.Themediandepth sumably because of the absence of scatterers in (depth at which 50% of the biomass is located deeperlayersduetoupwarddielverticalmigration above and below) was selected to represent the (FlaggandKim,1998;similarpatternobservedby verticaldistributionofbiomassinthewatercolumn Herring et al., 1998). Hence, only those profiles and to identify periods of diel vertical migration that extended to 120m were considered in this (e.g., Ashjian et al., 1998). This statistic best analysis to ensure that backscatter intensity represented the observed vertical distribution of measurements were present for all depths consid- biomass of the several statistics considered (depth ered when integrating and comparing biomass. ofmaximumbiomass,meandepth).Useofasingle Because of the strong diel signal, day and night statistic to represent vertical distribution could be integrated biomasses were considered separately. misleading if multiple migrating scattering layers Day and night integrated biomasses and chlor- were observed, such as seen in Luo et al. (2000). ophyll a concentrations were further processed by For the present data, only single migrating layers objectiveanalysis(OA)accordingtothemethodof were observed. The median depth was calculated Mariano and Brown (1992) to calculate predation foralldatacollectedduringatransect(noremoval pressureindicesandtoregressdaytimebiomasson of data points because of station keeping or chlorophyll a concentration. Objective analysis is excursions from the transect line). Potential useful in projecting data collected asynoptically along-track variation in migration behavior was onto temporally and spatially uniform fields. A examinedbycomparingvertical distributions from modified format was adopted to analyze the different along-track locations for each cruise. averaged data over distance (1000-km along Vertical migration velocities were calculated transect) and time (180 days) for all cruises. The from the median depth of biomass. The median observation variable (i.e. daytime biomass, night- depth first was smoothed over time using a time biomass, chlorophyll a concentration) was 10-point running mean (B50min) to reduce decomposedintoameanfield,anaturalvariability small-scale variation. Vertical migration velocities (mesoscale) field, and a subgrid-scale noise error then were calculated between each successive data field (for equations, see Mariano and Brown, point as the change in depth (m) per time period 1992). The mean field was estimated by a two- (s). Most observations were separated by 5min dimensional bicubic spline with adjustable intervals except when a profile was missing from smoothness and tension parameters. The mean the data because of poor quality. distributions were removed from the data before Periodswhendielverticalmigration(DVM)was OAcalculations.Aninterpolationgridwasdefined strong were identified using a DVM Index. This based on the number of points and resolution in index reflects the adherence of the vertical space and time for the output field estimate distribution of the median depth to a sinusoid (101(cid:2)101 grid, 0.011 resolution). At each grid curve (e.g., Pearre, 1979), which approximates point, a weighted local average of the influential the diel changes in depth of isolumes through the detrended data points was calculated and sub- water column. Before calculating the index, the tractedfromeachinfluentialdetrendeddatapoint. vertical distribution of the median depth for each The output for each variable was an OA field of day was time-standardized so that sunrise and identical spatial and temporal extent calculated sunsetoccurredat0600and1800,respectively,and from the sum of the OA estimate, the local depth-normalized to a range of 0–1 so that the average,andthetrend,andatrendfield(deviation maximum median depth was set to 1 and the C.J.Ashjianetal./Deep-SeaResearchII49(2002)2377–2402 2383 minimum to 0 (e.g., Ashjian et al., 1998). An both of along-transect distance (Fig.2a) and of ‘‘ideal’’ distribution then was calculated for each time (Fig.2b) from the March–April, 1995 cruise day by converting the observed times of each are shown as examples of typical vertical sections profile to equivalent radians and calculating the observedforthe14transects.Astrongdielvertical cosine function of the sinusoid curve. The sum of migration signal was the prevalent feature of the squares (SS) of the deviation then was calculated verticaldistributionforallofthecruises,withonly for each day and used as the DVM index: a single migrating layer observed. Greatest bio- mass was observed near the surface during night 1 XN dev:SS¼ ðobserved median depth and at depth (>250m) during day. Strong and N(cid:4)1 i i rapidverticalredistributionsofbiomassasasingle (cid:4)predictedmediandepthÞ2; coherent mass were associated with periods of i sunrise and sunset. Most profiles collected during fori¼1toN;whereN isthenumberofprofilesin dayextendedquitefarintothewatercolumn.The that day. Therefore, a lower value of the DVM penetration of the acoustic signal to depth was indexindicatesgreateradherencetoanidealDVM limited during night because of the aforemen- distribution, as simulated by a sinusoid. tioned absence of sound scatterers (zooplankton The vertical range of the acoustic data was and micronekton). The along-transect vertical limited during both day and night for the last two distribution (Fig.2a) demonstrates the temporal cruises (November 1995 (TN053); December 1995 discontinuities resulting from station-keeping and (TN054)). Because of an apparent degradation of the editing of the data. the ADCP signal, the DVM index was not The vertical distribution of the median depth calculated for December 1995 (TN054). Examina- clearly showed the effect of DVM on the distribu- tion of the vertical sections of intensity from the tion of biomass, following a sinusoid pattern reference station located in the central Arabian throughout the 24h, and was an effective means Sea over the period of the study demonstrated of representing the vertical distribution of the that, although the depth range of the instrument biomass (Fig.2b, upper panel). Occasionally, the clearlywasreduced,themagnitudeoftheintensity median depth of biomass deviated from this did not appear to be affected. Comparisons pattern when the biomass signal was reduced involving the magnitude of biomass from Novem- (e.g., Fig.2b, day 28). Vertical velocities (Fig.2b, berandDecember1995wereconsideredwithcare. lower panel) showed prominent peaks of upward Statistical analyses were conducted using MA- or downward movement (migration) associated TLAB (Mathworks, Inc.) or the Statistical Analy- with times of sunset or sunrise, respectively. sis System (SAS; SAS Institute, Inc., 1985). For multiple comparisons of means, analysis of var- 3.2. Spatial and temporal distribution of ADCP iance was used to identify whether significant biomass differences existed and the Student-Newman- Keuls post-hoc test was used to identify which The ADCP biomass present in the upper 120m means or groups of means were significantly during day was considered to be zooplankton, different (Zar, 1984). while that present during night was considered to be zooplankton plus those organisms that migrate to the upper 120m during night, primarily 3. Results myctophidfish,swimming crabs,andeuphausiids. Because of the different compositions of the 3.1. Vertical distribution of acoustically-derived biomass, day and night distributions and patterns biomass of biomass were considered separately. The influence of the SWM (June–September, 1995) Theverticaldistributionsofacoustically-derived was clear in the along-transect distribution of biomass(hereinafterADCPbiomass)asafunction vertically integrated (20–120m) zooplankton 2384 C.J.Ashjianetal./Deep-SeaResearchII49(2002)2377–2402 Fig.2. Vertical distributions of biomass from March–April, 1995 (TN045) as a function of along-track distance (n=585 profiles) (upperpanel),andtime(n=3447profiles)(middlepanel).Periodsofday(white)andnight(dark)areshowninthebarsacrossthetop ofeachpanel.Themediandepthofbiomassasafunctionoftimealsoisshown(blackline).Thelowerpanelshowstherateofvertical displacementofthemediandepthofbiomass,calculatedafterapplyinga10-pointrunningmean.Verticalprofilesweregriddedtoa densityof1point/20mintheverticaland1point/5kmfordistanceor1point/hfortimeinthehorizontal.Thealong-tracklocationsof standardUSJGOFSprocessstudystations(S2–S10)areshownasreference(upperpanel). biomass throughout the year (Fig.3a). Elevated also was observed inshore of the Findlater Jet zooplankton biomass was observed inshore (sta- during the SWM (July); this peak corresponds tions S2–S5) of the Findlater Jet (Jet located with the peak in zooplankton biomass. between stationsS6 and S7; seeFig.1)duringlate When the mean, vertically integrated ADCP June–early October, extending offshore to the biomass for regions inshore (stations S2–S4; Findlater Jet region during a single cruise in 104–403km along-transect) and offshore (stations August. Peak zooplankton biomass was observed S7–S11; 584–1064km along-transect) of the inJulyatstation S3,with the greatest areal extent Findlater Jet for each sampling period are of enhanced biomass occurring during the late compared, seasonal patterns emerge (Fig.4). SWM(September).ElevatednightADCPbiomass Biomass was significantly greater during night (the sum of zooplankton and migrator biomasses) (migrators plus upper-ocean zooplankton) than was observed offshore during the NEM and day (upper-ocean zooplankton only) for each Spring Intermonsoon (SI) periods from December sampling period within each region (ANOVA, to April (Fig.3b). Elevated night ADCP biomass po0:0001). Similar temporal patterns were C.J.Ashjianetal./Deep-SeaResearchII49(2002)2377–2402 2385 Fig.3. AnnualdistributionofintegratedADCPbiomassintheupperwatercolumn(20–120m)forday(upper)andnight(lower)asa functionofalongtrackdistanceandtime.Thealong-tracklocationsofthestandardUSJGOFSprocessstudystations(S1–S11)are shownasreference.Dataweregriddedtoadensityof1point/15daysintheverticaland1point/25kminthehorizontal.Gapsinthe distributionresultedwhenalong-transectlocationswerenotsampledduringthatperiod(dayornight).Notethescalechangebetween dayandnight.NocruisestookplaceinMayandmostofJune1995. observed for both zooplankton and zooplankton intermediate biomass during the 1994 NEM, and plus migrators within each region. Greatest lower biomass during the fall 1995 NEM (ANO- biomass inshore was observed during the SWM, VA, po0:001; Student Newman Keuls (SNK), 2386 C.J.Ashjianetal./Deep-SeaResearchII49(2002)2377–2402 Fig.4. Average biomass in the upper water column (20–120m) for regions inshore (S2–S4; 104–403km from shore) and offshore (S7–S11; 584–1064km from shore) of the atmospheric Findlater Jet along the transect line during day (zooplankton) and night (zooplanktonandmigrators).Onlythoseprofilesextendingdownto120mwereused.Standarddeviationsareshownwiththeerror bars.Eachmeanwascalculatedutilizing392–659profiles.Separatemeansarecalculatedforeachlegforcruiseswithbothanincoming and outgoing leg. The periods of the Northeast Monsoon (NE), Spring Intermonsoon (SI), Southwest Monsoon (SW), and Fall Intermonsoon(FI)areseparatedbydashedlines. po0:05). The patterns of both types of biomass December 1995 for all four region/diel period offshore were the reverse, with greatest biomass groupings (with the exception of September– generally observed during the fall 1994 NEM and October 1995, but see large variability in mean). Intermonsoon periods, lower biomass during the This may have resulted from interannual varia- SWM, and lowest biomass during the early NEM bilityin biomass between the two years or may be in 1995 (ANOVA, po0:001; SNK, po0:05). For a consequence of the degradation of the ADCP repeat transects within a cruise, biomasses usually signal at the end of 1995 (see methods section). were similar, suggesting that spatial and temporal Averages of integrated biomass (20–120m) at patchiness had little effect for those periods. thecentralArabianSeaoffshorestation(S15)were Biomasses were lowest during November and compared to averages from the regions inshore
Description: