Introduction
The pelagic ecosystem of the Southern Ocean is one of the least known environments on Earth (White & Piatkowski Reference White and Piatkowski1993). In terms of fish assemblages, there is a pronounced separation between oceanic and continental shelf waters. The oceanic ecosystem is dominated by the families Myctophidae, Paralepididae and Bathylagidae, composed mainly of meso- and bathypelagic species (DeWitt Reference DeWitt1970, Loeb et al. Reference Loeb, Kellermann, Koubbi, North and White1993). They are generally distributed in the oceanic waters north of the shelf break, but sometimes they are also found close to the Antarctic Continent in trenches or inner shelf depressions (Hubold Reference Hubold1990, Loeb et al. Reference Loeb, Kellermann, Koubbi, North and White1993). According to Kock (Reference Kock1992), both meso- and bathypelagic fish are most abundant and diverse between the Antarctic Convergence and the northern limit of the pack ice in winter, i.e. within the Ice Free Zone. Conversely, the pelagic fish communities inhabiting the continental shelf waters are dominated by the early life stages of the suborder Notothenioidei, which is endemic to Antarctic waters and composed almost exclusively of demersal species (Loeb et al. Reference Loeb, Kellermann, Koubbi, North and White1993). Indeed, true epipelagic families are not present in Antarctic surface waters, except for larvae and juveniles of notothenioids dwelling in the water column before settling as adults, probably as a result of niche separation or to improve dispersal. Other than early life stages of notothenioids, there are only a few other pelagic forms which have been secondarily derived from the demersal notothenioid families, such as the Antarctic silverfish, Pleuragramma antarcticum Boulenger, and the Antarctic toothfish, Dissostichus mawsoni Norman (Eastman Reference Eastman1993, Loeb et al. Reference Loeb, Kellermann, Koubbi, North and White1993).
Generally, fish are not evenly distributed over their geographical or vertical range, but form temporary or permanent assemblages or communities influenced by both biotic and abiotic factors (Kock Reference Kock1992). In the pelagic environment, the biogeographical boundaries are often associated with particular hydrological properties that have a strong influence on population dynamics and distribution of lower and middle trophic level communities (Koubbi Reference Koubbi1993). Rather, physical and biological processes have in some cases very similar space-time scales, indicating a close link between environmental parameters and pelagic communities (Verity & Smetacek Reference Verity and Smetacek1996).
In the pelagic ecosystem of the Southern Ocean, postlarval and juvenile notothenioids and several mesopelagic species have often been encountered in association with krill aggregations (Kock Reference Kock1992). A specific epipelagic krill-fish community developing temporarily in proximity to krill aggregations has been described from the Antarctic Peninsula and South Georgia (Slosarczyk & Rembiszewski Reference Slosarczyk and Rembiszewski1982, Slosarczyk Reference Slosarczyk1983). On the other hand, several studies provide data on distribution patterns of larval fish assemblages in relation to hydrographic conditions, mainly from the Antarctic Peninsula, Weddell Sea and South Georgia (reviewed in Loeb et al. Reference Loeb, Kellermann, Koubbi, North and White1993) or from sub-Antarctic islands such as Kerguelen and Crozet (Koubbi et al. Reference Koubbi, Ibanez and Duhamel1991, Koubbi Reference Koubbi1993). In comparison to the areas mentioned above, the knowledge of the pelagic fish community in the Ross Sea is far from complete. Indeed, available data on larval fish assemblages in this Antarctic sector are scarce (DeWitt Reference DeWitt1970, Efremenko Reference Efremenko1987, Vacchi et al. Reference Vacchi, La Mesa and Greco1999), and rarely focused on the nature and influence of the neighbouring physical or biological environment.
In the permanent pack ice zones of the Weddell Sea and Ross Sea, P. antarcticum is overwhelmingly the most abundant fish species, playing a key role in the pelagic system in the absence or relatively low abundance of krill, Euphausia superba (DeWitt Reference DeWitt1970). As a consequence of its wide distribution and high abundance on the continental shelf of the high Antarctic zone, P. antarcticum represents one of the most common prey of typical Antarctic top predators, such as whales, seals (Burns et al. Reference Burns, Trumble, Castellini and Testa1998, Casaux et al. Reference Casaux, Baroni and Ramon2003), penguins and flying birds (Mund & Miller Reference Mund and Miller1995, Ainley et al. Reference Ainley, Wilson, Barton, Ballard, Nur and Karl1998) and fishes (Eastman Reference Eastman1985, La Mesa et al. 2004a). Pleuragramma antarcticum, in turn, spending its entire life cycle in the water column, feeds almost exclusively on zooplankton, such as copepods, euphausiids, amphipods amd mysids (DeWitt & Hopkins Reference DeWitt and Hopkins1977, Eastman Reference Eastman1985, Hubold & Ekau Reference Hubold and Ekau1990, Hubold & Hagen Reference Hubold and Hagen1997, Granata et al. Reference Granata, Guglielmo, Greco, Vacchi, Sidoti, Zagami and La Mesa1999).
Based on several studies of ichthyoplanktonic communities living on the continental shelves around Antarctica (Keller Reference Keller1983, Hubold Reference Hubold1984, Reference Hubold1985, Kellermann Reference Kellermann1986b, Morales-Nin et al. Reference Morales-Nin, Garcia and Lopez1998), larvae and juveniles of P. antarcticum represent by far the bulk of catches, accounting for more than 90% of specimens. On the wide continental shelf of the western Ross Sea, for example, the abundance of early life stages of P. antarcticum may constitute 95–99% of the ichthyplankton (Guglielmo et al. Reference Guglielmo, Granata and Greco1998, Vacchi et al. Reference Vacchi, La Mesa and Greco1999, Granata et al. Reference Granata, Cubeta, Guglielmo, Sidoti, Greco, Vacchi and La Mesa2002). Larvae and juveniles of this species are generally found in the upper 200 m depth and their spatial distribution is largely affected by water masses and general circulation.
To understand better the mechanisms involved in the geographical distribution of P. antarcticum within the Ross Sea, we present here a thorough analysis of its abundance and distribution in relation to oceanographic parameters and water masses.
Material and methods
Study area
The study area covers about 164 000 nm2 in the western Ross Sea, between 69°–78°S and 164°–180°E (Fig. 1). According to Lewis & Perkin (Reference Lewis and Perkin1985), this area can be subdivided into four main sectors: a) open ocean, comprising ocean water deeper than 2000 m and extending up to about 71°S, b) continental slope, where oceanic waters mix with shelf waters coming from the Ross Sea, extending between approximately 71° and 73°S, c) continental shelf, a wide area extending between 73° and 76°S and comprising Pennell, Mawson and Crary banks and an inner-shelf depression such as Drygalski Basin, d) continental shelf adjacent to the Ross Ice Shelf (CSARIS), extending from 76°S to the margin of the Ross Ice Shelf and including Ross and southernmost Crary Banks as well as the JOIDES Basin.
Fig. 1 Study area in the south-western Ross Sea, showing bathymetry and topography of the main basins and banks.
The bottom topography of the continental shelf of the western Ross Sea is irregular, consisting of wide areas of relatively shallow banks with tops about 300 m deep (Iselin, Mawson and Pennell banks in the northernmost sector, Crary and Ross banks in the southernmost sector), separated by deep and narrow trenches extending in a south-west/north-east direction (i.e. Drygalski and JOIDES basins). The Ross Sea is almost entirely covered by sea ice for at least nine months of the year (Smith & Schnack-Schiel Reference Smith and Schnack-Schiel1990). However, persistent katabatic winds off the Ross Ice Shelf (150°W–160°E) cause the formation of a large coastal polynya during spring and reduce ice concentrations in winter (Bromwich et al. Reference Bromwich, Liu, Rogers and van Woert1998). The front of the Ross Ice Shelf is free of sea ice during summer, and sea ice in the western part of Ross Sea is generally advected northward into warmer water. A smaller winter polynya is kept open by katabatic winds in Terra Nova Bay. The winter polynyas are sites where High Salinity Shelf Water (HSSW) forms; this water has a temperature equal to the sea surface freezing point (about -1.90°C) and the highest salinity (close to 35) of the Southern Ocean because brine is rejected during sea ice formation fed by katabatic winds. HSSW occupies the deep layer of western Ross Sea, in particular of Drygalski and JOIDES basins, and exiting from the continental shelf it becomes, by turbulent mixing with Circumpolar Deep Water (CDW) on the bottom of the continental slope, a high salinity form of Antarctic Bottom Water (AABW). CDW is characterized by relatively high temperature (about +1.5°C) and salinity (about 34.70) off the Ross Sea continental shelf, and, when intruding onto it, it forms the freshened and cooled modified CDW (MCDW) by mixing with surrounding waters. MCDW is typically identified on the continental shelf, if present, by the subsurface temperature maximum.
A strong cyclonic gyre characterizes summer surface circulation on the continental shelf, with surface current flowing westward along the Ross Ice Shelf and then northward along the Victoria Land coast (Smith & Schnack-Schiel Reference Smith and Schnack-Schiel1990). In the deeper layer, below 200–300 m, the continental shelf is intersected by a double anticlockwise circulation with the western gyre ending near 176°W and the eastern one beginning at about 172°W, both extending below the Ross Ice Shelf (Locarnini Reference Locarnini1994). The numerical model simulation of Dinniman et al. (Reference Dinniman, Klinck and Smith2003), limited by the absence of a Ross Ice Shelf cavity, is in general agreement with this circulation pattern. In addition, Dinniman et al. (2003) point out that much of the flow is constrained to follow topography, and anticyclonic flow is produced around depressions and cyclonic flow around banks. On the continental slope, just off the shelf break, an oceanographic front (i.e. the slope front) separates shelf and open ocean waters (Jacobs Reference Jacobs1991); stronger currents than in most of the adjacent areas are found associated with this feature and are directed westward along the slope.
Biological sampling
Ichthyoplankton samples were collected during three cruises of the RV Italica carried out in the western sector of the Ross Sea in the summers of 1997–98, 1999–00 and 2003–04. In 1997–98, sampling was from 7 December 1997–4 January 1998 in the northern area of the Ross Sea, extending between 71°–75°S and between 164°E–178°W. Overall, 35 hauls were carried out between 130 m and the sea surface over bottom 262–2112 m deep at a vessel speed of 2.2–4.5 knots and for a towing time of 30–60 min. The sampling gear was a Hamburg Plankton Net (HPN, which is a modification of the Isaacs-Kidd-Midwater-trawl, (Hydro-Bios, Kiel, West Germany) with 5 m2 mouth area and mesh size of 500–1000 μm, equipped with a time-depth recorder (Benthos Inc, USA). The net deployment was normally a typical double oblique haul. The volume of filtered water from a flowmeter at the mouth of the net was 806 897 m3.
In 1999–2000, sampling was from 16 January–4 February 2000 between 70°–77°S and 164°E–176°W. Overall, 59 HPN hauls were undertaken between 300 m and the sea surface (bottom depth from 258–2441 m). The vessel speed and towing time ranged between 2.2–3 knots and 45–90 min, respectively. Total volume of filtered water was 601 951 m3.
In 2003–04, sampling was from 30 December 2003–27 January 2004 between 68°–77°S and 167°E–176°W. Overall, 33 hauls were made between 230 m and the sea surface (bottom depth 204–3438 m). The sampling gear was a Krill Midwater Sampling Trawl (KMST) with an approximately 12 m2 mouth opening and a mesh size of 1000 μm, equipped with a flowmeter (General Oceanics Inc). The gear was towed at speed of 2.5–3.3 knots and for a towing time of 30–70 min. Total volume of filtered water was 1 809 122 m3.
Oceanographic measurements
During the 1999–2000 and 2003–04 cruises a Sea Bird Electronics SBE 911plus CTD (Conductivity-Temperature-Depth) probe was lowered from the sea surface to about 300 m depth at the biological stations. The CTD probe was calibrated before and after the cruise, and it measured pressure (accuracy about 1 dbar or 104 Pa), temperature (0.002°C) and conductivity (0.0003 S m-1). Between CTD stations, Sippican T-7 XBT (eXpendable BathyThermograph) probes were launched, measuring temperature (with a nominal accuracy of 0.15°C) from the sea surface to 760 m depth or the sea bottom, whichever was shallower. Data were collected according to standard procedures and quality checked, derived variables were computed using standard algorithms and finally averaged in 5 decibar (c. 5 m) bins.
During the 1999–2000 survey the temperature was continuously measured by an Aanderaa modified 3444 sensor mounted in a water intake located in the ship hull at c. 4 m below the sea surface. Resulting data were calibrated against CTD data collected at same time and position.
Sample analysis
For each haul, fish samples were separated from other taxa, sorted by species and stored in buffered 4% formaldehyde. The taxonomic diagnosis were carried out using the BIOMASS identification keys of early life stages of Antarctic fish (Kellermann Reference Kellermann1990). On the basis of body morphology, the ontogenetic development of larval stages of P. antarcticum was divided in four groups (Russell Reference Russell1976, Koubbi et al. Reference Koubbi, Duhamel and Camus1990): a) larvae with yolk sac (stage 1), b) larvae with pre-flexion notochord, total loss of yolk reserve and exogenous nutrition (stage 2), c) postlarvae with post-flexion notochord, anal and dorsal fins partially developed (stage 3), d) postlarvae in metamorphosis and with dorsal fin rays fully formed (stage 4). Finally, specimens with pigmentation becoming more diffuse or silvering were considered juveniles (Russell Reference Russell1976).
Each specimen caught was measured to the nearest 0.1 mm both as total length (TL) and standard length (SL). The body shrinkage of specimens due to formaldehyde preservation was not considered. In case of large catches of P. antarcticum and/or narrow length range in a single haul, subsamples of 50 specimens were randomly selected and measured. All the measurements were made by means of a stereomicroscope linked to a monitor by a CCD camera using image analysis software.
For each positive haul, the relative abundance of each developmental stage of P. antarcticum was standardized to number of individuals per 1000 m3 of filtered seawater. To provide a measure of spatial distribution, the frequency of occurrence was calculated as well, taking into account only the positive hauls. According to Hubold (Reference Hubold1984), the age groups of P. antarcticum are defined according to length frequency data analysis: specimens between 7 and 20 mm SL were referred to age 0, specimens between 20 and 50 mm SL were referred to age 1+, specimens between 50 and 85 mm SL were referred to age 2+.
Finally, for each summer season investigated, all stations of occurrence for each larval stage and juveniles were plotted (as number of individuals per 1000 m3) and superimposed onto a bathymetric and sea surface temperature maps, to investigate the influences of water depth, benthic topography and water masses on the spatial distribution of the species. In particular, the relationship between standardized larval abundance and temperature gradient calculated at different depth layers (0, 30, 50, 100, 200 m) was evaluated by applying the Pearson product-moment correlations to each set of variable combinations.
Results
Abundance
A total of 3829 specimens of P. antarcticum were collected during the summer season 1997–98, constituting more than 75% of total catch with a frequency of occurrence of 86.2%. The bulk of catches were made on or in the vicinity of Mawson Bank, as well as in the coastal waters of Terra Nova Bay. Overall, standardized abundance ranged between 0.03 and 69.9 ind. 10-3 m3 (mean 7.7 ± 2.9). Early and late postlarvae (stages 3 and 4) were the most abundant specimens, accounting for 0.03–18.3 ind. 10-3 m3 (mean 4.3 ± 1.3) and 0.03–69.9 ind. 10-3 m3 (mean 5.8 ± 4.3), respectively.
In 1999–2000, a total of 73 889 specimens of P. antarcticum were caught, accounting for 99.8% of total catch with a frequency of occurrence of 78.9%. As in the summer season 1997–98, most catches were recorded on the Mawson Bank, other than in the south-eastern area of the Ross Sea, namely offshore the Ross Ice Shelf. On thirty positive hauls standardized abundance varied substantially, ranging between 0.05 and 2659.9 ind. 10-3 m3 (mean 124.2 ± 93.3). Larvae with pre-flexion notochord (stage 2) and late postlarvae (stage 4) were the most abundant stages of development, both showing a wide range of abundance. Standardized abundance of stage 2 was between 0.05 and 2659.1 ind. 10-3 m3 (mean 262.3 ± 197.1), whereas that of stage 4 ranged between 0.05 and 25.1 ind. 10-3 m3 (mean 0.4 ± 0.1).
47 574 specimens of P. antarcticum were collected in the season 2003–04, representing 95.7% of fish caught. The frequency of occurrence was significantly lower than those of the previous years, accounting for 63.3% of hauls. Most of catch was again located in the relatively shallow area of Mawson Bank seaward of Terra Nova Bay. The overall abundance of P. antarcticum ranged between 0.01 and 800.5 ind. 10-3 m3 (mean 42.7 ± 42.1). Pre-flexion larvae (stage 2) represented the overwhelmingly bulk of catches, attaining a standardized abundance of 0.02–800.2 ind. 10-3 m3 (mean 67.2 ± 66.6). On the other hand, the abundance of juveniles ranged between 0.01 and 2.0 ind. 10-3 m3 (mean 0.2 ± 0.1).
Length frequency distributions
In the summer season 1997–98 (7 December–4 January), the length frequency distribution of specimens caught was made up of three size groups extending from 7–64 mm SL (Fig. 2a), each of them roughly corresponding to different age classes of 0, 1 and 2+ years old fish (see below). The first size group consisted of small pre-flexion larvae (7 specimens, age 0) of 7–8 mm SL caught only in the coastal waters of Terra Nova Bay (Station PHN3). The second group represented the overwhelmingly bulk of the catch and was made up of 3784 postlarvae (stages 3 and 4, age 1) ranging from 19 mm to 50 mm SL. The size frequency distribution of postlarvae was bimodal, with modes at 32 mm and 43 mm. The third size group was a few juveniles (6 specimens, age 2+) ranging from 53 mm to 64 mm SL.
Fig. 2 Length frequency distributions of early life stages of P. antarcticum collected in the summer seasons a. 1997–98, b. 1999–2000, and c. 2003–04.
In the summer season 1999–2000 (16 January–4 February) the length frequency distribution was again composed of three discrete size groups ranging from 9 mm to 82 mm SL (Fig. 2b). Pre-flexion larvae (stage 2, age 0) largely dominated the catches, with 72 913 specimens caught (98.7% of total catches) ranging from 9 mm to 18 mm SL, with a single mode at 15 mm. The second group was made up of 965 postlarvae (stages 3 and 4, age 1) ranging from 31 mm to 57 mm SL. The length frequency distribution in this case was characterized by two modes at 37 and 50 mm SL. The last size group was a few juveniles (6 specimens, age 2+) ranging between 68 mm and 82 mm SL.
The length frequency distribution of specimens caught during the summer season 2003–04 (30 December–27 January) closely resembled the previous years (Fig. 2c). Three discrete size groups were caught overall ranging from 7 mm to 74 mm SL. As found in the 1999–2000 survey, the bulk of the catch was of pre-flexion larvae (47 358 specimens composing 99.5% of total catch) measuring from 7 mm to 19 mm SL (age 0). The length frequency distribution was unimodal at 12 mm. Forty postlarvae (stages 3 and 4, age 1) formed the second size group ranging from 28 mm to 50 mm SL, with a single mode at 33 mm. The third size group was made up of 158 juveniles (age 2+) from 53 mm to 74 mm SL. The size frequency appeared multi-modal, with the largest mode at 65 mm.
Age and growth
Based on the length frequency distributions recorded in each of the seasons investigated, as well as on literature data (Hubold Reference Hubold1984), the three different size groups were considered to belong to age groups 0, 1 and 2+. Using the differences between mean lengths of co-occurring age groups we have estimated the average growth rates of each age group over a period of one year for the 1997–98, 1999–2000 and 2003–04 summer seasons. In 1997–98, differences of 28.8 mm and 20.7 mm between age groups 0 and 1 and 1 and 2+ yielded an average growth rate of 0.08 mm and 0.06 mm day-1 for the first and second year of life, respectively. In 1999–2000, age groups 0 and 1 and 1 and 2+ differed by 29.8 mm and 31.9 mm, providing a similar average growth rate of c. 0.08 mm day-1 in both groups 1 and 2+. Similarly, in 2003–04, a difference of 24.1 and 26 mm between age groups 0 and 1 and 1 and 2+ yielded an average growth rate of 0.07 mm day-1 in both age groups.
To estimate growth rate of pre-flexion larvae in their first period of life, we followed the increase in mean length of 0 age group in consecutive days within the same season (i.e. 2003–04), assuming that the length frequency distribution was not biased by immigration or size selective mortality. Hence, fitting a linear relationship to length data (Fig. 3), we derived a mean daily increment of 0.16 mm. Coupling such a growth rate to the mean length at hatching of about 9 mm reported in literature for P. antarcticum in the Ross Sea (Vacchi et al. Reference Vacchi, La Mesa, Dalu and MacDonald2004), we were able to back-calculate the hatching period of pre-flexion larvae in the respective years of sampling. Taking into account the mean length of age 0 group over the whole sampling period, the hatching was estimated to occur in early December, mid December and late December in the summer season 1997–98, 1999–2000 and 2003–04, respectively.
Fig. 3 Linear relationship fitted to mean length increase of the 0 age group recorded in consecutive days throughout the summer season 2003–04.
Spatial distribution and oceanographic features
To investigate the geographical distribution of early life stages of P. antarcticum in the study area, a map of standardized abundance was made for each age group (0, 1 and 2+ old fish) linking together data from all surveys. In general terms, the overall pattern of distribution seems to be characteristic of each age group, as evidenced by the low spatial overlap between them.
The pre-flexion larvae (i.e. specimens of age 0) were caught mainly in the western and southern sectors of the Ross Sea, close to or in the vicinity of the continent and of the Ross Ice Shelf (Fig. 4a). Close to the Victoria Land Coast, they were collected in the northern area of the Drygalski Basin, on the shelf of Crary Bank and in Terra Nova Bay. Off the Ross Ice Shelf, they were caught in the JOIDES Basin, around the Ross Bank and in the Challenger Basin.
Fig. 4 Map of standardized abundance (number of ind. 10-3 m3) of P. antarcticum a. pre-flexion larvae (age 0), b. postlarvae (age 1), and c. juveniles (age 2+) collected in the western Ross Sea.
Pleuragramma antarcticum postlarvae (i.e. specimens of age 1) were distributed on a wide area of the continental shelf, but mainly in the northern and south-eastern area of the western Ross Sea (Fig. 4b). Below 75°S, postlarvae were found mainly in Terra Nova Bay, as well as in a few stations located on the Ross Bank and Challenger Basin. Although in low abundance, postlarval specimens were also collected in the southern areas of the Crary and Pennell banks and in the JOIDES Basin. However, the majority of catches of postlarvae were recorded in a relatively small area around the Mawson Bank, as well as the northern part of Drygalski and JOIDES basins. A few specimens were also taken at stations located in the oceanic waters close to or just offshore to shelf break.
Juvenile fish (age 2+) were caught occasionally and in low abundance. They were generally distributed on the banks or in close proximity of them (Fig. 4c). Most specimens were collected mainly on the easternmost banks of the western Ross Sea, namely the northern tip of Mawson Bank and the easternmost sectors of Pennell and Ross banks.
The spatial distribution of P. antarcticum age groups in relation to bottom depth at the sampling site is summarized in Fig. 5a–c). Although the spatial distribution of each age group varied substantially in different years, a general trend is evident. In particular, the smallest larvae (age 0) showed a wider range of distribution than other age groups, being caught across the entire bottom depth range sampled, although they appear to be more abundant over deep bottom waters (500–1000 m). Postlarvae (age 1) were again collected on a wide bottom depth range, but they were mainly concentrated on relatively shallower bottom waters with respect to the previous age group (namely between 250 m and 750 m). Finally, juveniles of P. antarcticum were restricted to a very narrow bottom depth range (i.e. almost exclusively between 250 m and 500 m bottom depth), showing a strong preference for relatively shallow bottom waters over bank tops.
Fig. 5 Spatial distribution of early life stages of P. antarcticum in relation to bottom depth recorded in the summer seasons a. 1997–98, b. 1999–2000, and c. 2003–04. Abundance is standardized as ind. per 10-3 m3 of filtered seawater.
Taking into account the oceanographic features such as the water temperature of the upper layer recorded throughout different years of sampling, we were able to investigate the influence of these characteristics on the spatial distribution of early life stages of P. antarcticum. As an example, the standardized abundance of pre-flexion larvae (age 0) and post-larvae (age 1) estimated during the 1999–2000 cruise were plotted along with the surface sea temperature (Fig. 6). Almost all specimens were caught on the continental shelf, particularly in the southern and western areas. Similarly, the standardized abundance of post-larvae (age 0) and juveniles (age 2+) estimated during the 2003–04 cruise were plotted along with the sea temperature recorded at 200 m depth (Fig. 7). The warmest area is intersected by the nutrient-rich Circumpolar Deep Water (CDW), which intrudes on the shelf area along the western side of the Mawson Bank and on the Pennell Bank. Most specimens of P. antarcticum were captured in areas reached by the modified CDW.
Fig. 6 Standardized abundance of pre-flexion larvae (age 0) and postlarvae (age 1) during 1999–2000 in relation to sea surface temperature (°C).
Fig. 7 Standardized abundance of postlarvae (age 1) and juveniles (age 2+) during 2003–04 in relation to sea temperature (°C) recorded at 200 m depth.
We found that generally the bulk of catches were recorded in close proximity to oceanographic fronts or, in other words, larval distribution was positively influenced by transition zones characterized by abrupt changes of water temperature, driven in turn by circulation. The Pearson correlation coefficient between larval abundance and temperature gradient was statistically significant for postlarvae collected during the 1999–2000 cruise (0 m, r = 0.34, P < 0.05; 30 m, r = 0.48, P < 0.01; 50 m, r = 0.37, P < 0.05) and during the 2003–04 cruise (30 m, r = 0.44, P < 0.05).
Discussion
The continental shelf of the Ross Sea, one of the broadest shelf areas in the Southern Ocean, averages 500–600 m deep, with inner shelf depressions or trenches reaching depths of more than 1200 m. Its southernmost area is permanently covered by the Ross Ice Shelf, whereas northward it is characterized by a strong seasonality of pack ice coverage and light regime. The hydrographical and biological features of the south-western Ross Sea largely depend on atmospheric conditions that drive sea ice dynamics and upper ocean stratification (Arrigo & Van Dijken Reference Arrigo and van Dijken2003). Strong katabatic winds blow from the continent over the front of the Ross Ice Shelf, giving rise to an offshore movement of sea ice that leads to the formation of the Ross Sea polynya. In spring, indeed, melting and advection of sea ice northward of the Ross Ice Shelf form an area of open water, progressively expanding in size throughout summer.
Summarising the life history pattern of P. antarcticum in terms of biological and physiological adaptations of each stage of development to the pelagic environment, it is possible to provide a reasonable explanation for the predominance of this species throughout the high Antarctic shelf waters.
The life cycle of P. antarcticum begins in winter (August), when adults migrate inshore to spawn off the great ice shelves of Antarctica (Kellermann Reference Kellermann1986b). Compared to other nototheniids of the high Antarctic zone, both absolute and relative fecundities of P. antarcticum are unusually high, attaining about 18 000 eggs/female and 160 eggs/g, respectively (Hubold Reference Hubold1991, Kock & Kellermann Reference Kock and Kellermann1991). Unlike other species, most of which spawn large eggs on the sea bottom, P. antarcticum spawn pelagic eggs of small size (about 2 mm) floating more or less freely in the platelet ice under the sea-ice cover (Vacchi et al. Reference Vacchi, La Mesa, Dalu and MacDonald2004). As a result, egg predation by other fish species, commonly reported in benthic feeders from the Ross Sea (La Mesa et al. Reference La Mesa, Vacchi, Castelli and Diviacco1997, 2004b), is probably prevented or largely reduced by the inaccessibility of this unusual brooding site.
Based on direct observation of eggs with embryos and yolk-sac larvae in the field (Vacchi et al. Reference Vacchi, La Mesa, Dalu and MacDonald2004), as well as on estimated growth rates of early larvae (Guglielmo et al. Reference Guglielmo, Granata and Greco1998, present data), the hatching period of P. antarcticum in the Ross Sea lasts from late November throughout December, in agreement with data from the Weddell Sea and Antarctic Peninsula (Keller Reference Keller1983, Kellermann Reference Kellermann1986a, Hubold Reference Hubold1990). The yolk-reserve in spring hatching notothenioids such as Pleuragramma is probably sufficient to sustain the larvae through the temporal variability of the pack ice retreat and the concomitant onset of the production cycle, because the hatching period observed in different years and/or different sites varies little. Nevertheless, in comparison to other species, the larvae of nototheniids have generally small yolk supplies (Kock Reference Kock1992). Thus, the reproductive cycle of P. antarcticum is closely coupled to the seasonal zooplankton production during the short spring–summer season and it is strongly affected by the interannual variability of physical and oceanographic characteristics.
From a metabolic perspective, the early larvae of Pleuragramma exhibit low lipid levels, consisting mainly of structural lipids such as phospholipids. Thus, the early larvae cannot rely on their energy stores and their survival is totally linked to external energy sources like copepod offspring (Wöhrmann et al. Reference Wöhrmann, Hagen and Kunzmann1997). Interestingly, antifreeze proteins (AFPs) levels are also low in the early larvae, which probably evolved other mechanisms to prevent freezing in the ice-laden waters, such as an undamaged integument and intestinal epithelium and a delayed development of the gills (Cziko et al. Reference Cziko, Evans, Cheng and DeVries2006). However, during their second summer, postlarvae begin to store large lipid deposits (mainly triacylglycerols) in specific ventral sacs, as well as antifreeze glycoproteins (Wöhrmann et al. Reference Wöhrmann, Hagen and Kunzmann1997). The stored lipids provide a twofold advantage, including near neutral buoyancy to a swim bladderless fish and energy reserves for a period of food deprivation, like winter (Maes et al. Reference Maes, Van de Putte, Hecq and Volckaert2006).
The importance of the presence and extent of coastal polynyas in the early life cycle of this species has been indicated by several authors (Hubold Reference Hubold1984, Reference Hubold1985, Kellermann Reference Kellermann1986a, Guglielmo et al. Reference Guglielmo, Granata and Greco1998, Maes et al. Reference Maes, Van de Putte, Hecq and Volckaert2006) and it is confirmed by present data. In conjunction with expansion of polynyas in the south-western Ross Sea, that starts in early November in front of the Ross Ice Shelf and off Terra Nova Bay, phytoplankton blooms exhibit a rapid increase until December (Arrigo & Van Dijken Reference Arrigo and van Dijken2004), providing favourable food conditions for the newly hatched larvae of P. antarcticum. Coupled with phytoplankton blooms there is a gradual increase of zooplankton and particularly of cyclopoid copepods (Oithona similis and Oncaea curvata) and pteropods (Limacina helicina) that constitute the preferential food of early larvae of P. antarcticum (DeWitt & Hopkins Reference DeWitt and Hopkins1977, Hubold Reference Hubold1985, Kellermann Reference Kellermann1987, Hubold & Hagen Reference Hubold and Hagen1997). However, the interannual variability of ice conditions may delay or prevent local onset of plankton production, causing a negative cascading effect on the higher trophic levels. Indeed, we found significant differences in the three years investigated both in terms of total catch and catch per unit of effort, possibly linked to physical (i.e. ice cover and/or extension of polynyas in the Ross Sea) and biological (i.e. primary production) cues. The abundance of early life stages was very low in the summer season 1997–98, when the early larval stage forming the young of the year group (age 0) was practically absent. Accordingly, mean annual primary production recorded in the study area during 1997–98 was very low and phytoplankton blooms dramatically decreased (Arrigo & Van Dijken Reference Arrigo and van Dijken2004), owing to an unusual heavy ice cover throughout the spring and summer season (Fig. 8a). On the other hand, the peak of abundance of P. antarcticum was recorded in the summer season 1999–2000, in which both annual primary production and extent of mean open water in the western Ross Sea were normal (Arrigo & Van Dijken Reference Arrigo and van Dijken2004) (Fig. 8b). Finally, the annual formation of the Ross Sea polynya was significantly delayed in the summer season 2003–04 (Fig. 8c), affecting negatively the abundance of the early life stages, especially as far as the frequency of occurrence was concerned.
Fig. 8 Monthly average sea ice cover around the Antarctic Continent recorded in a. January 1998, b. 2000, and c. 2004 (Cavalieri et al. Reference Cavalieri, Parkinson, Gloerson and Zwally2004). Squares indicate the area investigated in the present study.
During summer, the prevalent local north-westward currents induced by local winds shift early larvae of P. antarcticum on the continental shelf of the Ross Sea to the seaward limit of polynya. There, the early larvae are distributed mainly in the well stratified upper layer 0–100 m depth (i.e. the Antarctic Surface Water, AASW), probably in proximity to the summer thermocline (Guglielmo et al. Reference Guglielmo, Granata and Greco1998). Such a distribution has been described in the Weddell Sea and off the Antarctic Peninsula as well (Hubold Reference Hubold1984, Reference Hubold1990, White & Piatkowski Reference White and Piatkowski1993, Morales-Nin et al. Reference Morales-Nin, Garcia and Lopez1998). Residence in the relatively warm and productive summer surface layer of the shelf water allows a rather rapid development of early larvae, which attain an average growth rate of 0.15–0.25 mm day-1 (Keller Reference Keller1983, Hubold Reference Hubold1985, Guglielmo et al. Reference Guglielmo, Granata and Greco1998, present data). Indeed, during their first year of life, the early larvae seem to invest dietary energy mainly in somatic growth (Wöhrmann et al. Reference Wöhrmann, Hagen and Kunzmann1997). Interestingly, in the Ross Sea early larvae were distributed at a wide range of bottom depths, being caught both in the relatively shallow waters of banks and in the deep waters of basins, as reported in the Weddell Sea (Hubold Reference Hubold1985, White & Piatkowski Reference White and Piatkowski1993).
Driven by the local currents and the cyclonic surface summer circulation of the Ross Sea, larvae of P. antarcticum were mostly found in the southernmost area, whereas postlarvae and juveniles were progressively caught north-westward. Later stages tended to progressively occupy a wide area of the continental shelf, often in close proximity to the slopes surrounding the banks and near the shelf break. A similar distribution of postlarvae and juveniles of P. antarcticum has been observed in the Weddell Sea and off the Antarctic Peninsula (Hubold Reference Hubold1984, Reference Hubold1985, Kellermann Reference Kellermann1986a, White & Piatkowski Reference White and Piatkowski1993). Indeed, highest plankton concentrations and consequent zooplankton and fish aggregations are generally located in areas where abrupt bottom topography creates oceanographic fronts, like above seamounts, canyons and shelf breaks (Genin Reference Genin2004). In the Ross Sea, in particular, juveniles of P. antarcticum, which are able to swim actively against current, were mostly found in association with MCDW intrusions onto the continental shelf (see Fig. 7). This is probably due to the favourable food conditions in such areas, as recent studies (e.g. Dinniman et al. Reference Dinniman, Klinck and Smith2003) indicate a close relationship between the location and strength of MCDW intrusions and phytoplankton blooms. Furthermore, the strong frontal zone (the so-called slope front) delimiting the continental shelf of the Ross Sea probably acts as a boundary, limiting offshore transport of postlarvae and juveniles to low productivity oceanic waters.
Finally, on the continental shelf of the Ross Sea, postlarvae and juveniles of P. antarcticum were segregated both vertically and horizontally, reducing the competition for food and the possibility of cannibalism (Hubold Reference Hubold1984). Juveniles were distributed almost exclusively on relatively shallow areas (i.e. over bank tops), whereas postlarvae were distributed also over deep basins. In addition, juveniles were caught at greater depths than postlarvae, which generally were collected close to the thermocline (Granata et al. Reference Granata, Cubeta, Guglielmo, Sidoti, Greco, Vacchi and La Mesa2002, present data). The competition for food is also mitigated by an ontogenetic shift of feeding habits. Indeed, the postlarvae of P. antarcticum feed mainly on copepods, whereas juveniles and sub-adults are able to switch their diet on early life stages of krill (DeWitt & Hopkins Reference DeWitt and Hopkins1977, Hubold Reference Hubold1985, Kellermann Reference Kellermann1987, Hubold & Hagen Reference Hubold and Hagen1997).
In conclusion, the high fecundity of adult fish and the rapid growth recorded in the first year of life, as well as the subsequent spatial and trophic segregation between postlarvae and juveniles, which are able to consume a wide range of different food organisms, may be the key adaptations of P. antarcticum to the pelagic environment and provide a possible explanation for the overwhelming dominance of this species in the high Antarctic shelf waters.
Acknowledgements
This work has been partially supported by the Italian National Programme for Research in Antarctica (PNRA). We wish to thank all the people involved in data gathering, and in particular Prof G. Spezie and Dr E. Paschini for the support and instrumentation provided in the hydrological data collection, and Dr A. Sala and Mr V. Palumbo for the net sampling.
