Abundance and distribution of Antarctic krill Euphausia...

Abundance and distribution of Antarctic krill(Euphausia superba) nearshore of Cape Shirreff,Livingston Island, Antarctica, during six australsummers between 2000 and 2007

Joseph D. Warren and David A. Demer

Abstract: Abundance and distribution of Antarctic krill (Euphausia superba) in the nearshore waters north of LivingstonIsland, Antarctica, were characterized from six small-boat surveys conducted in late January or early February from 2000to 2007. The first three surveys (2000, 2002, 2004) were conducted using a 120 kHz split-beam echosounder to measurewater column acoustic backscatter. The last three surveys (2005–2007) were conducted using 38 kHz and 200 kHz single-beam echosounders. A portion of the acoustic backscatter was attributed to Antarctic krill based on the results of net tows,underwater video observations, and a multiple-frequency acoustic classification algorithm. The annual mean krill biomassdensity in the survey area ranged from 11 to 84 g�m–2. Results are compared with the western Scotia Sea area of the USAntarctic Marine Living Resources (AMLR) program’s acoustic surveys of krill biomass density for the same years. Near-shore krill biomass densities were significantly larger (t test, p < 0.05), more stable, and the coefficients of variation weresmaller than the much larger AMLR surveys. Increased competition between seals, penguins, and humans for the nearshorekrill resource, especially during the austral summer months, could impact the recruitment success of these land-based krillpredators. Implications of nearshore krill biomass on small-scale management units are discussed.

Resume : Six inventaires en petit navire realises a la fin de janvier ou au debut de fevrier de 2000 a 2007 nous ont servia caracteriser l’abondance et la repartition du krill antarctique (Euphausia superba) dans les eaux cotieres au nord de l’ıleLivingston, Antarctique. Les trois premiers inventaires (2000, 2002, 2004) ont ete faits a l’aide d’un echosondeur de120 kHz a faisceau divise afin de mesurer la retrodiffusion acoustique dans la colonne d’eau. Les trois derniers inventaires(2005–2007) ont utilise des echosondeurs a faisceau unique de 38 kHz et de 200 kHz. Une partie de la retrodiffusionacoustique est attribuee au krill antarctique, d’apres les resultats des echantillonnages au filet et les observations a la videosous-marine et a l’aide d’un algorithme de classification acoustique a frequences multiples. La densite de biomasse an-nuelle moyenne du krill dans la zone d’echantillonnage varie de 11 a 84 g�m–2. Nous comparons nos resultats a ceux desinventaires acoustiques de densite de biomasse du krill realises durant les memes annees dans la region occidentale de lamer de la Scotia par le programme americain « Antarctic Marine Living Resources » (AMLR). Les densites de biomassedu krill pres de la cote sont significativement superieures (test de t, p < 0,05), plus stables et leurs coefficients de variationplus petits que dans les inventaires beaucoup plus etendus de l’AMLR. La competition accrue entre les phoques, les man-chots et les humains pour la ressource de krill cotier, surtout durant les mois de l’ete austral, pourrait affecter le succes durecrutement de ces predateurs de krill venus du continent. Nous discutons des implications de la biomasse du krill cotiersur les unites de gestion a petite echelle.

[Traduit par la Redaction]

IntroductionAntarctic krill (Euphausia superba) are the main food re-

source for many of the fish, seabirds, and marine mammalsin the Southern Ocean ecosystem. In addition to its ecologi-cal importance (Marr 1962; Croxall et al. 1999; Siegel2000), krill are also the basis of a commercial fishery(Agnew 1997; Jones and Ramm 2004). To effectively man-age this fishery, accurate estimates of the abundance and

distribution of Antarctic krill are required. Towards thisgoal, several nations, as part of the Commission for Conser-vation of Antarctic Marine Living Resources (CCAMLR),conduct surveys of the krill population using bothechosounders and net tows to measure the krill population(Hewitt and Demer 2000). The results of these surveys arethen used to set precautionary catch limits for the fishers(Hewitt et al. 2004b).

Received 15 June 2009. Accepted 5 April 2010. Published on the NRC Research Press Web site at cjfas.nrc.ca on 24 June 2010.J21254

Paper handled by Associate Editor J. Michael Jech.

J.D. Warren.1 School of Marine and Atmospheric Sciences, Stony Brook University, 239 Montauk Highway, Southampton, NY 11968,USA.D.A. Demer. Southwest Fisheries Science Center, 8604 La Jolla Shores Drive, La Jolla, CA 92037, USA.

1Corresponding author (e-mail: [email protected]).

1159

Can. J. Fish. Aquat. Sci. 67: 1159–1170 (2010) doi:10.1139/F10-042 Published by NRC Research Press

In the Scotia Sea region of the Southern Ocean (UN Foodand Agriculture Organization’s statistical area 48), the USAntarctic Marine Living Resources (AMLR) program hasconducted annual krill surveys for more than two decades(Hewitt and Demer 1991; Hewitt et al. 2003). To surveythis very large area (Fig. 1), a series of transect lines andmore than 100 sampling stations are conducted at least onceper year during the austral summer (Hewitt et al. 2003).

The AMLR survey data from 1999 and 2000 were com-bined with the results of surveys from multiple other nationsand identified three areas of consistently high krill biomassin the vicinity of the South Shetland Islands: eastern end ofElephant Island; midway between Elephant Island and KingGeorge Island; and the north side of Livingston Island (He-witt et al. 2004a). However, the large-area surveys did notsample much of the shallow, on-shelf regions of the SouthShetland Islands or Antarctic Peninsula. Because the near-shore areas have a much smaller spatial area (relative to theoffshore large-area surveys), it is unlikely that higher near-shore krill abundances would affect annual estimates of krillbiomass. However, the importance of the nearshore areas topredator foraging, fishing, and thus small-scale managementunits (CCAMLR 2006) may be disproportionately large.

Nearshore regions are defined here as areas on the conti-nental shelf that are within tens of kilometres of land. Theseregions are important areas for several land-based species(e.g., penguins (Pygoscelis spp.) and Antarctic fur seals(Arctocephalus gazella)) that have established colonies androokeries in the South Shetland Islands and, in particular,Livingston Island (Boveng et al. 1998; Croll and Tershy1998). During the austral summer months, the adult animalsregularly forage in and transit through the nearshore area ofCape Shirreff to provide food for their young on land.

The abundance and behavior of several different types ofkrill predators (e.g., flying sea birds, penguins, and fur seals)in the nearshore area are affected by changes in nearshore

krill biomass (Miller and Trivelpiece 2008; Warren et al.2009). However, the relationship between the breeding suc-cess of these land-breeding predators and the offshore krillabundance may be complex. For example, Croll et al.(2006) found that penguin reproductive success was corre-lated with offshore krill abundance, but penguin foraging ef-fort was not. These findings suggest a possible lack ofcoherence between the nearshore and offshore krill abundan-ces. If nearshore krill biomass is sufficiently large and annu-ally stable, this could explain the long-term persistence ofpenguin and seal breeding colonies along the Antarctic Pen-insula (Croll and Tershy 1998; Hinke et al. 2007).

Alternatively, if krill abundance within the foragingranges and during the breeding season of penguins and sealsis reduced by natural variability, climate change, or fishingeffort, the apparent ecosystem balance could falter. Much ofthe commercial krill fishing takes place in these nearshoreareas of the Antarctic Peninsula (Jones and Ramm 2004)and may directly compete with foraging seals and penguins,particularly during their breeding season. As both the nativeanimals (fish, seabirds, and marine mammals) and humansare consuming the limited krill resource in the nearshoreareas, increased competition may negatively impact the re-productive success of animals confined to forage from theirland breeding sites.

To explore the hypothesis that nearshore krill biomassdensity is larger and more stable interannually than the off-shore krill biomass density, a method was developed and aseries of surveys were conducted to measure the abundanceand distribution of krill in the nearshore region north ofCape Shirreff, Livingston Island, Antarctica. Cape Shirreffis the site of several colonies of Antarctic fur seals and chin-strap (Pygoscelis antarctica) and gentoo (Pygoscelis papua)penguins. The AMLR field camp there served as a base forthe nearshore survey operations. Because of the shallowdepths and lack of accurate bathymetric charts of the area,

Fig. 1. (a) The US Antarctic Marine Living Resources (AMLR) program conducts an annual survey of krill abundance with conductivity–temperature–depth (CTD) and net tow stations (squares and circles) and multiple-frequency acoustic transects between stations. The westernarea of the AMLR survey is indicated by shaded squares. (b) An expansion of the box in (a) delineating the nearshore survey area. Thelocations of CTD and net tow stations of the nearshore survey are indicated by solid circles. The Cape Shirreff field camp (square) was theport for the nearshore survey operations.

1160 Can. J. Fish. Aquat. Sci. Vol. 67, 2010

Published by NRC Research Press

instrumented small craft were developed and used to studythe previously inaccessible nearshore region.

Materials and methodsDuring six austral summers (late January and early Febru-

ary) of 2000, 2002, and 2004–2007, an instrumented smallcraft (R/V Ernest) was used to survey the area surroundingCape Shirreff (Table 1). This area is characterized by shal-low (less than 100 m) depths except for two large submarinecanyons that exist to the northeast and northwest of theCape (Fig. 1). These canyons extend to the shelf break ap-proximately 30–40 km offshore. The bathymetry of thenon-canyon regions contains many pinnacles and other pro-nounced underwater features, some extending to within afew metres of the surface. There is also a small shelteredcove with a rocky beach that served as an anchorage forR/V Ernest. Transit to and from Cape Shirreff was onR/V Yuzhmorgeologiya during the first leg of the annualAMLR survey of the Scotia Sea krill population. The dura-tion of each nearshore survey ranged from 2 to 10 days(Table 1), depending on several factors including the dura-tion of the AMLR cruise, weather delays during the large-area survey, and weather and sea conditions during thetransfer to and from the field site and during the nearshoresurvey period.

R/V ErnestTwo people (captain, scientist) served as crew for R/V

Ernest, which had two distinct forms during this study

(Fig. 2). During the first three field seasons, an aluminumpilothouse was fitted into the 6 m inflatable (Zodiac MarkV) boat. The insert contained batteries supplying power to avariety of communication and safety equipment (radar, VHFradio, radar, EPIRB), as well as scientific instruments. Sen-sors onboard the vessel included a meteorological station(WeatherPak 2000; Coastal Environmental Systems Inc., Se-attle, Washington), which measured temperature, humidity,barometric pressure, bearing, and apparent and true windspeed and direction; multiple global positioning system(GPS) receivers; and a 120 kHz split-beam echosounder(Simrad EY500; Kongsberg Maritime AS, Horten, Norway).GPS, meteorological, and acoustic backscatter data were re-corded on a laptop computer. A liquid-crystal display,mounted behind a waterproof window, provided the scientistwith a real-time display of position and meteorological andechosounder data. A motorized downrigger was used to de-ploy a small conductivity–temperature–depth (CTD) profiler(SBE 19; Sea-Bird Electronics, Inc., Bellevue, Washington)and an underwater video-camera system (Sony Handycamwith Light and Motion Stingray housing and lights; SonyInc., Tokyo, Japan) at stations throughout the survey area.The video camera was used to confirm the classifications ofacoustically detected targets. The boat was also equippedwith survival and tool kits, manual and automatic bilgepumps, three survival suits, four fuel tanks, binoculars, andanchorage equipment. Transit operations used a 55-hp gasoutboard engine, whereas survey operations used a 9.9-hpoutboard engine with generator to allow remote steering,conserve fuel, charge batteries, and minimize noise in the

Table 1. Nearshore survey timing and duration, total length of all transects, depths sampled by acoustic echosoun-der, coverage of canyons east and west of Cape Shirreff, and timing of echosounder calibration.

Date Survey length (km) Depth recorded (m) Coverage Echosounder calibration6–10 February 2000 148 250 West, east Before17–23 February 2002 250 100 West, east Before7–09 February 2004 130 250 East During1–10 February 2005 280 400 East Before and after3–08 February 2006 234 400 East Before and after29–30 January 2007 77 400 East Before

Fig. 2. (a) R/V Ernest I was used to survey the nearshore waters of Livingston Island for krill abundance and distribution in 2000, 2002,and 2004. It contained a single-frequency (120 kHz) split-beam echosounder, a meteorological station (not shown), and an electric down-rigger for cast deployments of a CTD. (b) R/V Ernest II was used in 2005–2007 and contained a multiple-frequency (38 and 200 kHz)single-beam echosounder, a sea-surface temperature and salinity sensor, and a meteorological station.

Warren and Demer 1161


acoustic record. During the 2000 and 2002 field seasons, thetransducer was deployed on a portside gimbal mount; how-ever, transducer movement in certain sea conditions reduceddata quality so a transom-mounted transducer arm was in-stalled in 2004 and used thereafter.

For the 2005–2007 field seasons, the aluminum insert wasreplaced by two waterproof equipment cases and a dodgermade of stainless steel, canvas, and vinyl (Fig. 2). Thesechanges reduced the weight of the boat, allowed the vesselto plane when transiting between survey locations, andprovided more protection from the elements for the crew.The smaller engine was removed and the larger one wasused for both transiting and survey operations. This changeallowed for faster survey speeds and an increase in surveycoverage. The single-frequency echosounder was replacedwith 38 kHz and 200 kHz echosounders (SIMRAD dual-frequency single-beam ES60; Kongsberg Maritime AS), whichused a 802.11 g wireless network to link the echosounder to alaptop computer on Ernest. In addition, a temperature andsalinity sensor (SBE 37 microCAT; Sea-Bird Electronics,Inc.) was deployed on the transom-mounted transducer arm torecord subsurface (1 m) hydrography during the survey.

Survey operationsMeteorological and position data were recorded whenever

the boat was transiting or surveying. The vessel transited athigh speeds (ca. 8 m�s–1) to the beginning of a transect line.The transducer arm was then lowered and acoustic and hy-drographic data recording commenced. Depending on thetrack lines to be run each day, transits between track lineswere made either slowly (2–5 m�s–1) while recordingechosounder data or rapidly with the transducer raised outof the water. Meteorological, date, time, and position datawere recorded every 20 s on the laptop computer. Hydro-graphic data (temperature, salinity, and pressure) from theCTD were recorded every 5 s. Occasionally, when large ag-gregations of scatterers were observed on the echosoundernear the surface (<75 m depth), the digital video camerawas lowered to record images and identify the scatterers.Weather conditions, sea state, fuel supply, and crew endur-ance determined the length of survey operations each day,but typically 5 to 8 h each day were spent collecting dataand transiting. Surveys were conducted in all types ofweather (often in the same day) and were suspended whenseas grew greater than 4 m, visibility was reduced to lessthan 10 m, or winds exceeded those for safe vessel operation(e.g., >10 m�s–1).

The survey transects varied in length and direction duringthe first few years of this study, as the operational capabil-ities and limits of Ernest were identified. Also, as measure-ments from Ernest were used to map the nearshorebathymetry, increasing portions of the nearshore area weresafely surveyed from Yuzhmorgeologiya. The survey strat-egy soon stabilized, and since 2004, nearly identical transectlegs east of Cape Shirreff were surveyed from Ernest. Tran-sits to the western side of Cape Shirreff can be difficult incertain sea-state or fog conditions, so efforts were focusedon a higher spatial coverage of the eastern side of the Capewhen survey time became limited during the 2004 field sea-son. The direction of the transects during the 2004 to 2007surveys were chosen to align with cruise tracks that Yuzh-

morgeologiya surveyed during the large-area survey so thatcomparisons could be made between the data collectedfrom the two observational platforms. However, surveytracks (direction and distance) were often adjusted depend-ing on the direction and size of the waves and wind. The2004 nearshore survey consisted of only 3 days of opera-tions due to severe weather and sea conditions during4 days of the scheduled weeklong survey period. Similarly,the 2007 nearshore survey was reduced to less than 2 daysas a result of severe weather during both the large-area andnearshore survey periods.

Acoustic data

CalibrationThe echosounders were calibrated for each survey using

the standard sphere method (Foote et al. 1987) with a38.1 mm tungsten carbide sphere (Table 1). Using a singlemonofilament tether, the sphere was lowered beneath Ernestuntil it appeared on the echogram and the target strength(TS), uncompensated for beam directivity effects (TSU; dB),was maximized. Several hundred TSU measurements of thesphere were recorded at this depth while both the sphereand transducer moved with water and vessel motion. Forthe 120 kHz split-beam measurements, the sphere positionin the beam was measured and the TSU were compensatedfor the beam directivity (dB). In the cases of the single-beam 38 kHz and 200 kHz transducers, the location of thesphere within the acoustic beam was unknown. For most ofthe single-beam calibration experiments, the distributions ofTS measurements were examined to confirm that some on-axis measurements were achieved. These measurementswere used to calculate the on-axis calibrated system gain(Foote et al. 1987).

In 2005, on-axis calibration measurements were obtainedfor the target for both frequencies of the echosounder. How-ever, in 2006, because of strong currents, TS measurementswere intermittent with the narrower beamwidth (78) of the200 kHz transducer, but continuous with the wider beam-width (138 longitudinal and 218 transverse) of the 38 kHztransducer. In this case, theoretical beampatterns were calcu-lated for both transducers, and the differences in TSU at thetwo frequencies were predicted for a range of off-axis an-gles. This information was used to constrain the possible lo-cations for the sphere within the beam. The data were thenfit to the theoretical predictions using the method of leastsquares to estimate the minimum off-axis angle achieved(i.e., 58). The TS measurements were then compensated forthe beampattern and used to estimate the calibratedechosounder gain. In 2007, a calibration was not possibledue to the short survey duration and inclement weather andsea-state conditions. Consequently, the 2006 settings wereused for the 2007 survey. For the first four field seasons,the calibration accuracy is estimated to be better than 1 dB(one-way gain accuracy £ 0.5 dB or ~12%). For the 2006–2007 field seasons, the aforementioned procedure is thoughtto result in calibration accuracy of better than 2 dB for eachfrequency (one-way gain accuracy £ 1.0 dB or ~26%).

Data collectionSurvey speeds varied depending on sea state, wind, and



transducer-mount location but were between 2 and 4 m�s–1.The transducer was located between 1 and 1.75 m beneaththe surface. Acoustic pulses of 1.024 ms duration weretransmitted every 1 s at 120 kHz (2000, 2002, and 2004)and 2 s at 38 and 200 kHz (2005–2007), and volume back-scattering strengths (Sv; dB re 1 m–1) were recorded every20 to 30 cm from the transducer to depths of 250 m (2000),100 m (2002), 250 m (2004), and 400 m (2005, 2006, and2007) (Table 1). All locations in the survey area were shal-lower than 400 m, and only the middles of the canyons weredeeper than 250 m.

Acoustic backscatter provides indirect measures of thedistribution and abundance of biological organisms in themarine environment. Validation of this metric was achievedwith a series of 2.5 m2 Isaac–Kidd midwater trawl net towsduring the 2002, 2004, and 2005–2007 field seasons. Thetows were conducted from Yuzhmorgeologiya in adjacentareas (Fig. 1). The net-tow contents were enumerated andidentified to species onboard Yuzhmorgeologiya and thenpreserved in a buffered (10%) formalin solution. During the2005–2007 surveys, adult Antarctic krill were sexed andtheir lengths were measured before preservation. At leastonce during each field season (except for 2004 and 2007),an underwater video camera was lowered into a patch ofnear-surface scatterers that were observed on theechosounder. Ernest drifted as the camera was lowered tothe depth of the scattering aggregation and then retrieved.This method was occasionally unsuccessful if the scatterersdispersed or the aggregation moved while the camera waslowered. Successful video observations indicated that aggre-gations of large euphausiids (likely E. superba) were thedominant mesozooplankton in the water column.

Volume backscattering coefficients (sv (m–1), where Sv =10�log10(sv)) were summed from 3 to 10 m below the seasurface to the shallower of 100 m or 1 m above the seafloor(observation range common to all years) and averaged over185 m (0.1 nautical miles (n.mi.)) trackline distances, yield-ing nautical area backscattering coefficients (sA; m2�n.mi.–2).Selected regions were excluded from the integration, includ-ing the acoustic near-field and scatter from nonbiologicalsources such as bubbles, suspended sediments (e.g., mud orsand), or the seafloor (particularly the dead zone abovehigh-relief hard substrates; see Demer et al. 2009). In the2004–2007 data, noise from the engine was also removedby coherent subtraction (Hewitt et al. 2004b) and filtered bythresholding below Sv = –80 dB re 1 m–1.

Conversion of backscatter dataThe dual-frequency echosounder data were filtered using

the DSv method (Watkins and Brierley 2002; CCAMLR2005) prior to the calculations of sA. The DSv method aimsto accept Sv from krill and reject that from other biologicalscatterers. Following the procedure of CCAMLR (2005,2009), the minimum and maximum length of krill caught inthe net samples for each year were used to calculate the DSvrange for attributing acoustic backscatter to krill. The sto-chastic distorted wave Born approximation (SDWBA) model(Demer and Conti 2005; Conti and Demer 2006; CCAMLR2009) was used to calculate the predicted difference in scat-tering (TS at 200 kHz minus TS at 38 kHz) for the smallestand largest krill lengths. These two DSv values formed theupper and lower bounds for filtering the acoustic backscatterdata (Table 2). This technique could not be applied to thesingle-frequency 120 kHz data collected during the 2000,2002, and 2004 nearshore surveys.

The sA were converted to estimates of krill biomass den-sity using the length distribution of krill caught in nets and atheoretical acoustic scattering model (Demer and Conti2005; Conti and Demer 2006; Reiss et al. 2008) that hasbeen adopted by CCAMLR as the standard method for pro-ducing estimates of krill biomass (CCAMLR 2005, 2009).Because the nearshore trackline lengths and directions werenot standardized for all survey years, the method of Jollyand Hampton (1990) was used to measure the meanweighted density and coefficient of variation (CV) for a ser-ies of transects comprising the survey for each year. Thesevalues were then comparable between the six nearshore sur-veys and the AMLR Scotia Sea surveys, which were ana-lyzed in a similar manner (method 3 from Reiss et al. 2008).

Comparisons of the nearshore survey results betweenyears are complicated by the different acoustic instrumenta-tion (dual versus single frequency), differing observationaldepths, and different amounts and locations of survey effort.To quantify the potential bias in the 120 kHz data (2000,2002, 2004) due to the possible inclusion of non-krill scat-terers, the 200 kHz data (2005–2007) were processed withand without the DSv method and compared. The sA at120 kHz were then corrected using the relationship foundusing the 200 kHz data. Although there can be significantdifferences between the TS at 120 and 200 kHz for a singleanimal (Demer and Conti 2005), krill aggregations exhibit arange of lengths (as measured from animals caught in nettows) and orientations relative to the incident sound wave.

Table 2. Conversion of acoustic backscatter data to estimates of krill biomass density used krilllength information from net tows and the SDWBA scattering model to determine the criteria toidentify backscatter from krill.

Target strength (dB)

Smallest krill Largest krill

YearRange of krilllengths (mm) 38 kHz 200 kHz 38 kHz 200 kHz

DSv range usedto identify krill

2005 40–55 –81.3 –74.4 –73.0 –71.2 2.6–7.02006 45–58 –78.7 –73.5 –73.6 –71.5 2.0–5.22007 22–60 –94.5 –79.3 –74.5 –71.9 1.8–15.2

Note: DSv is the Sv at 200 kHz minus the Sv at 38 kHz. This method was only applied to the data from2005 to 2007 as the 2000, 2002, and 2004 surveys used a single-frequency echosounder.



Thus the mean TS for a krill aggregation can be calculatedby weighting the orientation-averaged backscattering crosssection for krill at each length by the proportion of krill ofthis length caught in the net. This has the effect of dampen-ing the TS versus length oscillations in the geometric scat-tering regime (where the krill length is large relative to theacoustic wavelength).

Because Sv data were recorded to different maximumdepths (Table 1), the minimum requisite survey depth wasexplored. Most of the scattering was observed in the upper100 m of the water column, so the effect of not integratingSv deeper than 100 m is assumed to be insignificant. Thisassumption is supported by independent observations of krillthroughout the Scotia Sea during the austral summer of 2000(see fig. 9 in Demer 2004). To further validate this assump-tion, the scatterer aggregations in the 2005 and 2006 datathat were identified as krill using the DSv method were fur-ther characterized with regard to their size and depth.

ResultsNet-tow samples for all years indicate that adult and juve-

nile Euphausia superba were the largest contributors to zoo-plankton biomass in this region, although other smallereuphausiids (Thysanoessa macrura, Euphausia frigida) werealso frequently caught in smaller numbers. Other zooplank-ton and nekton in the net catches were copepods (Metridiagerlachei, Calanoides acutus, Calanus propinquus, Rhinca-lanus gigas, and Paraeuchaeta spp.), salps (Salpa thomp-soni), amphipods (Cyllopus lucasii, Primno macropa, andThemisto gaudichaudii), chaetognaths, larval fish (Lepidono-tothen larseni and Electrona spp.), and gastropods (Lima-

cina helicina, Spongiobranchaea australis, and Clionelimacina). The three taxa that accounted for the overwhelm-ing majority of biomass were Antarctic krill, copepods, andsalps. Neither copepods nor salps are considered to be strongacoustic targets at these frequencies compared with the krill(Demer 1994; Stanton et al. 1996; Stanton and Chu 2000),so these acoustic estimates of biomass should not bestrongly affected by the presence of either. Larval fish, how-ever, are relatively strong scatterers, but they were found in-frequently in the net tows, generally only at nighttime, andalways in low numerical densities. Pteropods are also strongscatterers; however, they were only found in the shelf-breaknet tows (i.e., those farthest from the nearshore survey re-gion) and were caught in very low numerical abundances,so their contributions to the nearshore backscatter are likelyto be small.

A key factor in the conversion of sA to estimates of bio-mass is the estimated probability density function (pdf) ofkrill lengths. This length pdf is used to calculate both theweighted scattering cross-section and the biomass per krill.For 2000, 2002, and 2004, the length pdfs for the nearshorearea were assumed equivalent to those measured in the west-ern area of the AMLR broad-scale survey, which includesthe South Shetland Islands. For the 2005–2007 surveys, thelength pdfs were calculated from net samples from the near-shore area (Fig. 3). In 2005, the nearshore and AMLR west-ern area distributions were nearly identical (coefficient ofdetermination, R2 = 0.96), whereas the 2006 (R2 = 0.52)and 2007 (R2 = 0.81) distributions showed slightly (approxi-mately 5 mm) larger krill in the nearshore waters. Addition-ally, in 2007, smaller (juvenile) krill were caught in bothregions, although the nearshore region had fewer of these

Fig. 3. Length distributions of Euphausia superba collected from net tows from (a) 2005, (b) 2006, and (c) 2007 during the nearshoresurvey (shaded squares) and AMLR western area (solid circles).



younger krill. Given the general agreement between thenearshore and western area distributions in the 2005–2007surveys, errors resulting from using the western area lengthpdfs in 2000, 2002, and 2004 are likely small.

Using the 38 kHz and 200 kHz data from 2006, biomassdensities calculated from the 38 or 200 kHz data, withoutthe DSv method, were, as expected, higher than those de-rived with the dual-frequency algorithm (Fig. 4). Therefore,the biomass densities estimated from the single-frequencydata collected in 2000, 2002, and 2004 are probably overes-timated. To account for this bias, a correction factor was de-veloped and applied. First, biomass densities were calculatedfor each of the 2005–2007 surveys using only the 200 kHzdata (i.e., replicating a single-frequency echosounder) andusing both frequencies (i.e., the DSv method). The differen-ces between the biomass densities calculated using the sin-gle-frequency (200 kHz) and DSv methods were fairlyuniform across all values of measured backscatter (Fig. 4b)and for each survey. The 2006 data included the largestrange of biomass density values and were therefore used toestimate the correction factor, but the 2005 and 2007 datahad the same pattern where they overlapped. This suggeststhat the DSv method, tuned by the krill length pdfs, is notaffected by the overall level of backscatter and that scatterfrom small (e.g., salps and smaller euphausiids) and largeorganisms (fish) was rejected, as desired, using the DSvmethod.

The length-weighted TS for krill at 120 kHz for each yearof the survey using the SDWBA model (Conti and Demer2006) were 0.45 ± 0.15 dB larger than that at 200 kHz. Toaccount for this small, expected difference, the Sv at120 kHz were reduced by 0.45 dB before calculating the sA

(Demer 2004). Next, a correction formula was found for the2006 data by fitting a straight line (in log–log space) to the200 kHz biomass density data and then transforming thesedata to match the biomass density using the DSv method(Fig. 4). The correction equation, based on the 200 kHz bio-mass density data, was applied to the adjusted 120 kHz bio-mass density (BD120 kHz) from 2000, 2002, and 2004 data,resulting in a revised biomass density estimate (BD120 new)that was used in all subsequent analyses (Fig. 4): BD120 new =10–1.551�BD120 kHz

1.084.During each nearshore survey, krill biomass density was

observed to be high throughout the shallow regions proxi-mate to Cape Shirreff (Fig. 5). High biomass densities weremapped both along the edges and within the canyon east ofCape Shirreff. Areas to the southeast and south of the east-ern canyon had relatively lower biomass. Coupling the 2000and 2002 Ernest surveys of the canyon west of Cape Shir-reff with observations from Yuzhmorgeologiya (Warren etal. 2009), there is an indication that krill are equally asabundant there as in the eastern canyon.

In addition to examining horizontal distributions, the verti-cal distributions of krill aggregations were examined in both2005 and 2006. Unlike the echo-integration analyses, whichwere limited to the upper 100 m of the water column, krillaggregations were examined throughout the water column(as the maximum observational depth of 400 m exceededthe maximum bottom depth in these surveys). Aggregationsof krill were identified from acoustic-backscatter data withinthe DSv range corresponding to the range of krill lengths ob-served each year. Data from both years showed that krill ag-gregations had similar depth distributions with mean (±1standard deviation, SD) depths of 62 (±26) m in 2005 and

Fig. 4. Comparison between values of integrated krill biomass density (BD) (g�m–2) (circles) estimated from single-frequency backscatterdata at (a) 38 kHz and (b) 200 kHz (BD38 kHz and BD200 kHz, respectively) and from the two-frequency DSv method (BDDSv ). The value ofthe single-frequency estimates is on the horizontal axis, while the two-frequency DSv estimates are on the vertical axis. The shaded brokenline in both plots represents where the single-frequency and DSv method estimates of biomass density are equal. To correct single-frequencysurvey data (which cannot discriminate between krill and non-krill scatterers), an equation (shown in (b)) was found that related BD200 kHz

and BDDSv , which was used to produce corrected 200 kHz biomass estimates (BDnew, shaded circles). The black broken line in (b) is alinear regression fit to the data that was used to calculate the correction function.



53 (±27) m in 2006. The center of mass of krill aggregations(e.g., depth of krill aggregations weighted by the biomass ofthe swarm) was 55 m in 2005 and 48 m in 2006. During both2005 and 2006, the overwhelming majority of krill aggrega-tions (92% and 95%, respectively) and their biomasses (97%and 99%, respectively) were located at depths less than orequal to 100 m (Fig. 6).

To compare estimates of biomass from each survey, a ser-ies of roughly parallel transects were selected from eachnearshore survey. The transect directions were not the sameeach year because of the different survey designs; however,for the last three years of the survey, the transects wereroughly oriented northwest to southeast. Following Jollyand Hampton (1990), the mean biomass density for each

Fig. 6. Distributions of the depth of the center of individual krill aggregations identified during the (a) 2005 and (b) 2006 nearshore surveysand the depth of the center of mass (depth weighted by biomass density) of each krill aggregation sampled during the (c) 2005 and (d) 2006nearshore surveys. There were 867 (in 2005) and 793 (in 2006) krill aggregations identified and analyzed.

Fig. 5. Spatial distributions of krill biomass density (g�m–2) estimated from acoustic backscatter sampled during (a) 2000, (b) 2002,(c) 2004, (d) 2005, (e) 2006, and (f) 2007 of the Livingston Island nearshore survey. The 200 m isobath (thin black line) marks the edge ofthe submarine canyon where large krill aggregations were regularly found.



segment and the transect lengths were calculated. The meanbiomass density for each transect, weighted by transectlength, were then calculated for each survey along with theircoefficients of variation (CV values; Table 3; Fig. 7). TheCV values were generally similar between years, except for2000 and 2007, which were higher. In 2000, this is likely theresult of the longest transect in that survey having a meanbiomass density six times larger than the other transects. In2007, the large CV is due to the short duration of the surveyand, consequently, only three transects available for analysis.

Annual mean krill biomass densities ranged from 11 to84 g�m–2 and were significantly larger than krill biomassdensity estimates from the AMLR western area survey(t test, p < 0.05; Table 3; Fig. 7). The estimates of biomassdensity in 2004 and 2007 (57.9 and 84.3 g�m–2, respectively)were substantially larger than those for the other years, mo-

tivating a thorough re-examination of the echograms. Thehigh biomass estimates were the result of krill aggregationsin both years being both more numerous and dense. In everyyear except 2002, the nearshore biomass density was greaterin the nearshore region than in the western area of theAMLR survey (AMLR 2008), and in some years, nearshorebiomass densities were substantially greater than those off-shore (Table 3; Fig. 7). The mean nearshore biomass densityestimates were larger (mean, 38.6 g�m–2; SD, 28.0 g�m–2)and always greater than 11 g�m–2, whereas the western areaestimates were smaller (mean, 14.1 g�m–2; SD, 13.1 g�m–2),exhibited a much larger variance from year to year, and insome years were very small (i.e., less than 1 g�m–2). Basedon these six years of survey data, krill biomass densities re-mained high in the nearshore regions even when the off-shore krill biomass densities were low.

Table 3. Comparison of mean krill biomass densities (BD) and coefficient of variation(CV) for the transects from the nearshore survey and the AMLR Western Area (AMLR2008), and the number of nearshore transects and total transect distances used to calcu-late BD.

Nearshore survey AMLR Western area

YearTransects(number)

Distance(km)

BD(g�m–2) CV (%)

BD(g�m–2) CV (%)

2000 9 93 14.9 46 4.51 (D) 32.2 (D)2002 11 103 11.2 25.6 21.02 (A) 44.6 (A)

0.41 (D) 46.4 (D)2004 7 84 57.9 13.2 34.37 (A) 8.9 (A)

18.87 (D) 44.0 (D)2005 24 222 26.9 17.1 17.11 (A) 26.6 (A)

0.37 (D) 85.2 (D)2006 17 162 36.2 17.5 0.81 (A) 45.9 (A)2007 3 32 84.3 36.1 29.23 (A) 19.7 (A)

Note: Analyses of both the AMLR Western Area and nearshore data use the same SDWBA scat-tering model and procedure for converting acoustic backscatter to krill biomass. AMLR surveys oc-curred once or more each year in January (A) or February (D). Biomass densities from 2000 to 2004(single-frequency data) were corrected using the relationship shown in Fig. 4.

Fig. 7. Annual mean krill biomass density (g�m–2) and coefficient of variation (whiskers) for the nearshore survey (solid circles) and thewestern area of the AMLR Scotia Sea January (open squares) and February (open diamonds) surveys. For points where the uncertainty isnot visible, the CV is smaller than the marker.



DiscussionThis multiyear study has demonstrated that a small boat

equipped with a scientific echosounder and either video ornets to verify the acoustic scatterers as krill can successfullysurvey shallow, nearshore waters of Antarctica. The esti-mates of nearshore krill biomass density are significantlylarger, exhibit smaller interannual changes, and have smallerCV values within each survey year than the acoustic surveysof the much larger offshore areas of the Scotia Sea. For thearea north of Cape Shirreff, during the austral summers ofthe years of this study, krill were consistently found in sim-ilar areas, had a minimum biomass density of at least11 g�m–2, and had mean biomass densities that were muchlarger than the adjacent offshore waters in all but one year.The western area of the AMLR survey often had biomassdensities that were one order of magnitude smaller than thenearshore areas, although in some years, biomass differenceswere even greater. Physical (advection) and biological (up-welling, increased phytoplankton biomass) factors in thenearshore area are possible reasons for the consistentlyhigher krill biomass in this region.

Note that the AMLR survey data collected at three acous-tic frequencies (38, 120, and 200 kHz) are currently ana-lyzed using two DSv ranges to identify acoustic scatteringfrom krill (e.g., Reiss et al. 2008), following CCAMLR(2005, 2009). These differences in survey data and analysistechniques may partially explain their smaller estimates ofkrill biomass density. However, because of the absence ofsmall krill in the nearshore net tows, the nearshore surveyhad a much smaller DSv range (except in 2007 when theranges are comparable) than that used by the AMLR survey.Consequently, the data presented here are conservative esti-mates of the nearshore krill biomass density (CCAMLR2009). In addition to the nearshore waters having larger andmore uniform krill densities compared with the nearby off-shore region, the nearshore waters generally contain morelarge adult krill and fewer juvenile krill. Thus, the nearshoreregion may avail the land-based predators with higher-quality prey (i.e., more biomass per krill). Why larger krillpreferentially aggregate nearshore is yet to be discovered.

Although the sampling methods and instrumentationchanged somewhat between the nearshore surveys, the dif-ferences probably do not affect the overall results. Themean krill biomass densities from 2000, 2002, and 2004 aresimilar (and in general smaller) than those from the yearswhen the dual-frequency discrimination was done, suggest-ing that our method of correcting these data did not producepositively biased results. The different integration depthsused for different years do not contribute appreciable biasesbecause the majority of the krill biomass was in the upper100 m of the water column and all the integration depthswere greater than or equal to 100 m. These distributions aresimilar to that observed for krill surveyed over the entireScotia Sea during 2000 (Demer 2004) and in other studiesof Antarctic krill (Miller et al. 1993; Siegel and Kalinowski1994). Moreover, Miller and Trivelpiece (2008) found thatchinstrap penguins did not exhibit foraging dives deeperthan 93 m (2005) and 88 m (2006) during the day, whichsuggests that sufficient krill prey resources exist shallowerthan 100 m.

The relative importance of nearshore krill in terms of thetotal krill stock in the Scotia Sea (e.g., Hewitt et al. 2004b)is unknown; however, the spatial area of this nearshore sur-vey is quite small (~200 km2) when compared with that ofthe Scotia Sea. The relatively high and stable densities ofkrill biomass in the nearshore regions of the South ShetlandIslands may be important to include in some estimates ofkrill biomass, particularly in subregions considered forsmall-scale management units. Although the three identified‘‘hot spots’’ of krill biomass from the CCAMLR 2000 multi-national, Scotia Sea survey (Hewitt et al. 2004a; CCAMLR2006) are all proximate to nearshore areas, the surveys usedto identify these hot spots did not extensively sample theshallower nearshore waters, as was done in this study. Thelarge nearshore krill biomass is generally most accessibleand attractive to the land-breeding predators, as well as tohuman fishers competing for this valuable resource. Animalsand fishers alike are drawn to the nearshore hot spots duringthe austral summer months (Jones and Ramm 2004). Giventhe different objectives of research vessels (standardized sur-vey effort to collect high quality data) and fishing vessels(maximize profits by catching more animals), fishing vesselsmay be willing to work closer to land than research vessels,which may increase competition between land-based preda-tors and fishers for the krill resource.

This study focused on a relatively small nearshore regionof Livingston Island, and it is difficult to say whether theseresults are applicable to other shallow, nearshore areas alongthe South Shetland Islands, Antarctic Peninsula, or else-where in the world. Acoustic surveys of krill backscatterfurther south along the Antarctic peninsula (MargueriteBay) have also observed higher levels of backscatter in theshallower, nearshore regions (Lawson et al. 2004) that havebeen related to the distribution or abundance of at least onekrill predator (Friedlaender et al. 2006), so the results of thisstudy corroborate surveys in other areas conducted by largervessels (but not in as shallow a region as this study).

The nearshore surveys were conducted over a very smallarea (hundreds of square kilometres) and very short timescale (days) over six of eight consecutive years. Because ofthese spatial and temporal sampling constraints, these dataare susceptible to biases associated with short-term orsmall-scale phenomena that would be ‘‘smoothed out’’ by alarger area or longer survey period (Warren et al. 2009). Po-tential disturbances that may affect the nearshore survey re-sults include mesoscale meteorological events such as low-pressure systems or storms passing through the region orpossibly even the difference that a spring or neap tide mighthave on the ecosystem close to shore. Some of these events(particularly storms) are frequent occurrences in the studyarea, so these data may reflect some of these biases. How-ever, the data from the nearshore survey during 2005 didnot exhibit the same reduction in krill biomass that Warrenet al. (2009) observed in the waters immediately adjacent tothe nearshore survey area during the same time period, sug-gesting that the nearshore areas may be resistant to the ef-fects of these submesoscale perturbations.

This study demonstrates that acoustic surveys conductedfrom instrumented small craft can augment larger-vessel sur-veys and extend the multidisciplinary investigations intowaters commonly inaccessible, but ecologically important.



Results show that krill biomass can be as high or higher insome nearshore areas, possibly influencing the sites and suc-cess of penguin and seal rookeries, and should be consideredin krill-fishery management schemes utilizing small-scalemanagement units (Hewitt et al. 2004b). During the australsummer months, many of these highly productive nearshoreareas are preferred foraging areas for large populations ofpenguins and Antarctic fur seals (Croll and Tershy 1998;Croll et al. 2006) and commercial fishing activities (Jonesand Ramm 2004).

When rearing their chicks and pups on land, adult pen-guins and fur seals need a reliable food resource in the near-shore areas to assure reproductive success. Without adequateprey nearshore, the land-breeding krill predators must forageon other species, or further offshore, consequently jeopardiz-ing the health of their offspring and themselves. Also, itshould be noted that even in times of large nearshore krillabundance, the land-breeding predators might switch pre-ferred prey. For example, Miller and Trivelpiece (2008)found that in 2004, the year with the second highest ob-served nearshore krill biomass density, the chinstrap pen-guins at Livingston Island preferentially foraged on fishinstead of krill. The reason for this switch is unknown butmay be due to increased accessibility of more nutritiousprey. Increased competition between seals, penguins, andman for the nearshore krill resource, especially during theaustral summer months, could have a large impact on the re-cruitment success of these land-based krill predators. Conse-quently, the biomass of krill within the foraging ranges ofbreeding penguins and seals should be routinely surveyedand considered when developing small-scale units, both inspace and time, for managing the krill resource.

AcknowledgementsThe financial and logistical support of the US AMLR Pro-

gram has made this work possible, and Rennie Holt, its Di-rector, is thanked heartily. During 2001, J.D.W. joined thisresearch as an Office of Naval Research postdoctoral fellow(grant no. N00014-01-1-0166) with D.A.D. at NOAA’sSouthwest Fisheries Science Center (SWFSC). The 2005–2007 field seasons were supported by the National ScienceFoundation’s Office of Polar Programs (grants OPP-0338196 and OPP-0633939) and the Advanced SurveyTechnologies group (AST) at SWFSC. The captains of R/VErnest, Adam Jenkins (2000, 2002, 2004) and Steve Ses-sions (2005, 2006, and 2007), allowed these data to besafely collected in a wide variety of conditions. D.A.D.’sconcept for R/V Ernest I was cleverly designed and skill-fully built by Leif Knutsen (Big Foot Marine). Derek Need-ham and Steve Sessions assisted with the design andconstruction of R/V Ernest II. The Captain and crew of R/VYuzhmorgeologiya provided a vigilant guard and reliablesupport for the small-boat operations. The scientists aboardthe R/V Yuzhmorgeologiya provided great company andhigh-quality acoustic (A. Cossio, J. Emery, R. Hewitt, andC. Reiss), hydrographic and meteorologic (D. Needham, M.Prowse, M. Soule, and M. van den Berg), and zooplankton(V. Loeb and the zooplankton sampling teams) data. B.Cobb, B. Parker, D. Krause, and R. Haner of AMLR and J.Evans of NSF/Raytheon Polar Services provided logisticalsupport for our land-based operations. Finally, this work

was facilitated by the AMLR Cape Shirreff field camp per-sonnel (led by R. Holt, M. Goebel, and W. Trivelpiece),who graciously welcomed, fed, and sheltered us during thesmall-boat operations.

ReferencesAgnew, D.J. 1997. The CCAMLR ecosystem monitoring pro-

gramme. Antarct. Sci. 9(3): 235–242. doi:10.1017/S095410209700031X.

Antarctic Marine Living Resources Program. 2008. Antarctic Mar-ine Living Resources Program 2007/2008 field season report.Report No. NOAA-TM-NMFS-SWFSC-427, National MarineFisheries Service, La Jolla, California.

Boveng, P.L., Hiruki, L.M., Schwartz, M.K., and Bengtson, J.L.1998. Population growth of Antarctic fur seals: limitation by atop predator, the leopard seal? Ecology, 79(8): 2863–2877.doi:10.1890/0012-9658(1998)079[2863:PGOAFS]2.0.CO;2.

Commission for the Conservation of Antarctic Marine Living Re-sources. 2005. Report of the first meeting of the subgroup onacoustic survey and analysis methods. Report No. SC-CAMLR-XXIV/BG/3, Commission for the Conservation of AntarcticMarine Living Resources, Hobart, Tasmania, Australia.

Commission for the Conservation of Antarctic Marine Living Re-sources. 2006. Report of the second workshop on managementprocedures for allocating catch among small-scale managementunits. Report No. SC-CAMLR-XXV/3, Commission for theConservation of Antarctic Marine Living Resources, Hobart,Tasmania, Australia.

Commission for the Conservation of Antarctic Marine Living Re-sources. 2009. Report of the fourth meeting of the subgroup onacoustic survey and analysis methods. Report No. SC-CAMLR-XXVIII/8, Commission for the Conservation of Antarctic Mar-ine Living Resources, Hobart, Tasmania, Australia.

Conti, S.G., and Demer, D.A. 2006. Improved parameterization ofthe SDWBA for estimating krill target strength. ICES J. Mar.Sci. 63(5): 928–935. doi:10.1016/j.icesjms.2006.02.007.

Croll, D.A., and Tershy, B.R. 1998. Penguins, fur seals, and fish-ing: prey requirements and potential competition in the SouthShetland Islands, Antarctica. Polar Biol. 19(6): 365–374. doi:10.1007/s003000050261.

Croll, D.A., Demer, D.A., Hewitt, R.P., Jansen, J.K., Goebel, M.E.,and Tershy, B.R. 2006. Effects of variability in prey abundance onreproduction and foraging in chinstrap penguins. J. Zool. (Lond.),269(4): 506–513. doi:10.1111/j.1469-7998.2006.00090.x.

Croxall, J.P., Reid, K., and Prince, P.A. 1999. Diet, provisioningand productivity responses of marine predators to differences inavailability of Antarctic krill. Mar. Ecol. Prog. Ser. 177: 115–131. doi:10.3354/meps177115.

Demer, D.A. 1994. Accuracy and precision of acoustic surveys ofAntarctic krill. Ph.D. thesis, University of California San Diego,San Diego, Calif.

Demer, D.A. 2004. An estimate of error for the CCAMLR 2000survey estimate of krill biomass. Deep Sea Res. Part II Top.Stud. Oceanogr. 51: 1237–1251.

Demer, D.A., and Conti, S.G. 2005. New target-strength model in-dicates more krill in the Southern Ocean. ICES J. Mar. Sci.62(1): 25–32. doi:10.1016/j.icesjms.2004.07.027.

Demer, D.A., Cutter, G.R., Renfree, J.S., and Butler, J.L. 2009. Astatistical–spectral method for echo classification. ICES J. Mar.Sci. 66(6): 1081–1090. doi:10.1093/icesjms/fsp054.

Foote, K.G., Knudsen, H.P., Vestnes, G., MacLennan, D.N., andSimmonds, E.J. 1987. Calibration of acoustic instruments forfish density estimation: a practical guide. International Council



for the Exploration of the Sea (ICES) Cooperative Research Re-port No. 144.

Friedlaender, A.S., Halpin, P.N., Qian, S.S., Lawson, G.L., Wiebe,P.H., Thiele, D., and Read, A.J. 2006. Whale distribution in re-lation to prey abundance and oceanographic processes in shelfwaters of the Western Antarctic Peninsula. Mar. Ecol. Prog.Ser. 317: 297–310. doi:10.3354/meps317297.

Hewitt, R.P., and Demer, D.A. 1991. Krill abundance. Nature(London), 353(6342): 310. doi:10.1038/353310b0.

Hewitt, R.P., and Demer, D.A. 2000. The use of acoustic samplingto estimate the dispersion and abundance of euphausiids, with anemphasis on Antarctic krill, Euphausia superba. Fish. Res.47(2–3): 215–229. doi:10.1016/S0165-7836(00)00171-5.

Hewitt, R.P., Demer, D.A., and Emery, J.H. 2003. An 8-year cyclein krill biomass density inferred from acoustic surveys con-ducted in the vicinity of the South Shetland Islands during theaustral summers of 1991–1992 through 2001–2002. Aquat. Liv-ing Resour. 16(3): 205–213. doi:10.1016/S0990-7440(03)00019-6.

Hewitt, R.P., Kim, S., Naganobu, M., Gutierrez, M., Kang, D., Ta-kao, Y., Quinones, J., Lee, Y.-H., Shin, H.-C., Kawaguchi, S.,Emery, J.H., Demer, D.A., and Loeb, V.J. 2004a. Variation inthe biomass density and demography of Antarctic krill in the vi-cinity of the South Shetland Islands during the 1999/2000 aus-tral summer. Deep Sea Res. Part II Top. Stud. Oceanogr. 51:1411–1419.

Hewitt, R.P., Watkins, J., Naganobu, M., Sushin, V., Brierley, A.S.,Demer, D., Kasatkina, S., Takao, Y., Goss, C., Malyshko, A.,Brandon, M., Kawaguchi, S., Siegel, V., Trathan, P., Emery, J.,Everson, I., and Miller, D. 2004b. Biomass of Antarctic krill inthe Scotia Sea in January/February 2000 and its use in revisingan estimate of precautionary yield. Deep Sea Res. Part II Top.Stud. Oceanogr. 51: 1215–1236.

Hinke, J.T., Salwicka, K., Trivelpiece, S.G., Watters, G.M., andTrivelpiece, W.Z. 2007. Divergent responses of Pygoscelis pen-guins reveal a common environmental driver. Oecologia (Berl.),153(4): 845–855. doi:10.1007/s00442-007-0781-4.

Jolly, G.M., and Hampton, I. 1990. A stratified random transect de-sign for acoustic surveys of fish stocks. Can. J. Fish. Aquat. Sci.47(7): 1282–1291. doi:10.1139/f90-147.

Jones, C.D., and Ramm, D.C. 2004. The commercial harvest ofkrill in the southwest Atlantic before and during the CCAMLR

2000 survey. Deep Sea Res. Part II Top. Stud. Oceanogr. 51:1421–1434.

Lawson, G.L., Wiebe, P.H., Ashjian, C.J., Gallager, S.M., Davis,C.S., and Warren, J.D. 2004. Acoustically-inferred zooplanktondistribution in relation to hydrography west of the Antarctic Pe-ninsula. Deep Sea Res. Part II Top. Stud. Oceanogr. 51(17–19):2041–2072. doi:10.1016/j.dsr2.2004.07.022.

Marr, J.W.S. 1962. The natural history and geography of the Ant-arctic krill (Euphausia superba Dana). Discov. Rep. 32: 33–464.

Miller, A.K., and Trivelpiece, W.Z. 2008. Chinstrap penguins alterforaging and diving behavior in response to the size of theirprinciple prey, Antarctic krill. Mar. Biol. (Berl.), 154(2): 201–208. doi:10.1007/s00227-008-0909-z.

Miller, D.G.M., Barange, M., Klindt, H., Murray, A.W.A., Hamp-ton, I., and Siegel, V. 1993. Antarctic krill aggregation charac-teristics from acoustic observations in the southwest AtlanticOcean. Mar. Biol. (Berl.), 117(1): 171–183. doi:10.1007/BF00346440.

Reiss, C.S., Cossio, A.M., Loeb, V., and Demer, D.A. 2008. Varia-tions in the biomass of Antarctic krill (Euphausia superba)around the South Shetland Islands, 1996–2006. ICES J. Mar.Sci. 65(4): 497–508. doi:10.1093/icesjms/fsn033.

Siegel, V. 2000. Krill (Euphausiacea) demography and variabilityin abundance and distribution. Can. J. Fish. Aquat. Sci. 57(S3):151–167. doi:10.1139/cjfas-57-S3-151.

Siegel, V., and Kalinowski, J. 1994. Krill demography and small-scale processes: a review. In Southern Ocean ecology. The BIO-MASS perspective. Edited by S.Z. El-Sayed. Cambridge Univer-sity Press, Cambridge, UK. pp. 145–163.

Stanton, T.K., and Chu, D. 2000. Review and recommendations forthe modelling of acoustic scattering by fluid-like elongated zoo-plankton: euphausiids and copepods. ICES J. Mar. Sci. 57(4):793–807. doi:10.1006/jmsc.1999.0517.

Stanton, T.K., Chu, D., and Wiebe, P.H. 1996. Acoustic scatteringcharacteristics of several zooplankton groups. ICES J. Mar. Sci.53(2): 289–295. doi:10.1006/jmsc.1996.0037.

Warren, J.D., Santora, J.A., and Demer, D.A. 2009. Submesoscaledistribution of Antarctic krill and its avian and pinniped preda-tors before and after a near gale. Mar. Biol. (Berl.), 156(3):479–491. doi:10.1007/s00227-008-1102-0.

Watkins, J.L., and Brierley, A.S. 2002. Verification of acoustictechniques used to identify and size Antarctic krill. ICES J.Mar. Sci. 59(6): 1326–1336. doi:10.1006/jmsc.2002.1309.



Date post:	02-Oct-2020
Category:	Documents
Upload:	others
View:	1 times
Download:	0 times

Abundance and distribution of Antarctic krill Euphausia...

Documents