Document [ to be completed by the Secretariat ] WG-EMM-08/27 Date submitted [ to be completed by the Secretariat ] 4 July 2008 Language [ to be completed by the Secretariat ] Original: English Agenda Agenda Item No(s): 7

Title Trophic study of Ross Sea Antarctic toothfish (Dissostichus mawsoni) using carbon and nitrogen stable isotopes Author(s) S.J. Bury1, M.H. Pinkerton1, D.R. Thompson1, S. Hanchet2, J. Brown1, and I. Vorster1 Affiliation(s) 1 National Institute of Water and Atmospheric Research Ltd Email: [email protected] 2 NIWA, PO Box 893, 217 Akersten Street, Nelson, New Zealand Published or accepted for Yes No ü publication elsewhere? ABSTRACT This report amalgamates stable isotope analyses of fish (n=476), squid (n=50) and octopod (n=17) samples obtained from long-line fishing vessels from four CCAMLR SSRUs (88.1C, 88.1H, 88.1I and 88.1J) during two fishing seasons 2005/6 and 2006/7. The species sampled were: 6 fish: Antarctic toothfish (Dissostichus mawsoni, n= 100), Patagonian toothfish (Dissostichus eleginoides, n=8), deep sea cod/blue antimora (Antimora rostrata, n=103), icefish (Chionobathyscus dewitti, n=83), moray (or eel) cod (Muraenolepis microps, n=75), and Whitson’s grenadier (Macrourus whitsoni, n=107); 4 squid: Galiteuthis glacialis (Gg, n=3), Kondakovia longimana (Kl, n=20), Psychroteuthis glacialis (Pg, n=20) and the Colossal squid, Mesonychoteuthis hamiltoni (Mh, n=7); and 3 benthic octopods: Octopodid sp. 1 (Oct-1, n=3), Octopodid sp. 2 (Oct-2, n=5) and Cirroctopus glacialis (Cg, n=9).

Length and SSRU were the most significant variables in explaining the variation of δ15N and δ13C. Positive relationships between length and δ15N indicate that, very generally, larger fish consume prey of a higher trophic level than smaller fish. There were substantial residual within-species variations in δ15N and δ13C. Dissostichus mawsoni exhibited a range of 7 ‰ (9–16 ‰) in δ15N, which is equivalent to two trophic steps. All fish, except Antimora rostrata (2.7 ‰ range) showed a d15N range greater than 3.4 ‰ spanning more than one trophic step. This implies that the diet of all species sampled was variable, or that individual species were eating a similar diet which itself varied in size and trophic status. Overall, Dissostichus mawsoni and Dissostichus eleginoides occupied a trophic level equivalent to orca (Orcinus orca) and Weddell seals (Leptonychotes weddellii). Antimora rostrata, Muraenolepis microps and Macrourus whitsoni all occupied a trophic level below them. Chionobathyscus dewitti occupied the lowest trophic level of all fish analysed. There was considerable isotopic overlap in both δ15N and δ13C for all four fish prey species. Squids, excluding Mesonychoteuthis hamiltoni were found to be at a lower trophic level than fish species sampled, whereas on average octopods occupied a similar trophic level to the four fish prey species. The squid δ13C signature was more depleted (indicating a pelagic signature) than the octopods, which were all benthic feeders. Large variations in d13C for each species (around 3 ‰ for each species) indicated a variation in source of carbon within individual species. Species with enriched d13C may be feeding further north in warmer waters or may have a stronger benthic compared to pelagic source of carbon.

There was no significant difference in Dissostichus mawsoni δ15N and δ13C values between the Northern Area, Ross Sea Slope and Terra Nova Bay Trench. In contrast, all of the four potential prey species caught in the Northern Area had enriched 13C values compared to the Ross Sea Slope, most likely due to warmer temperatures to the north. Since this increased δ13C signature is not picked up by Dissostichus mawsoni, then this suggests that Dissostichus mawsoni either move between and feed equally within the Northern Area and the Ross Sea Slope, or that they predominantly feed on the Ross Sea Slope. SUMMARY OF FINDINGS AS RELATED TO NOMINATED AGENDA ITEMS Fishing for Antarctic toothfish (Dissostichus mawsoni) in the Ross Sea is likely to affect 7 prey species through trophic linkages. Stable isotope data can help to elucidate these relationships. 7 Dissostichus mawsoni and Dissostichus eleginoides occupy a trophic level equivalent to orca (Orcinus orca) and Weddell seals (Leptonychotes weddellii). WG-FSA Spatial homogeneity in Dissostichus mawsoni δ15N and δ13C values between the Northern Area, Ross Sea Slope and Terra Nova Bay Trench contrasted with enriched 13C values for all four potential prey species only in the Northern Area suggests that Dissostichus mawsoni either move between and feed equally within the Northern Area and the Ross Sea Slope, or that they predominantly feed on the Ross Sea Slope. This paper is presented for consideration by CCAMLR and may contain unpublished data, analyses, and/or conclusions subject to change. Data in this paper shall not be cited or used for purposes other than the work of the CCAMLR Commission, Scientific Committee or their subsidiary bodies without the permission of the originators and/or owners of the data. 1. INTRODUCTION

Antarctic toothfish (Dissostichus mawsoni) are the major finfish resource currently exploited in the Ross Sea, with an exploratory fishery operating since 1996/7 in the Convention for the Conservation of Antarctic Marine Living Resources (CCAMLR) sub-area 88.1 (the western Ross Sea) conducted by independently observed commercial longliners (Hanchet et al. 2006, 2007; Ministry of Fisheries, 2006). Information on the trophic links between Dissostichus mawsoni and both their prey and predators is currently fragmented and a more comprehensive picture is needed to assess the potential impact of the fishery on associated species (e.g. Pinkerton et al. 2006, 2007c, d). The overall aim of this study was to investigate whether stable isotope analysis of muscle tissue samples supports the hypothesis that demersal fish dominate the diet of Dissostichus mawsoni caught in the commercial fishery of the Ross Sea.

We present stable isotope data from 476 fish, 50 squid and 17 octopod samples obtained from long-line fishing vessels from four CCAMLR Small Scale Research Units (SSRUs) (88.1C, 88.1H, 88.1I and 88.1J) during two fishing seasons, 2005/6 and 2006/7. The species sampled were:- 6 fish: Antarctic toothfish (Dissostichus mawsoni, n=100), Patagonian toothfish (Dissostichus eleginoides, n=8), deep sea cod/blue antimora (Antimora rostrata, n=103), icefish (Chionobathyscus dewitti, n=83), moray (or eel) cod (Muraenolepis microps, n=75), Whitson’s grenadier (Macrourus whitsoni, n=75); 4 squid species: Galiteuthis glacialis (Gg, n=3), Kondakovia longimana (Kl, n=20), Psychroteuthis glacialis (Pg, n=20) and colossal squid, Mesonychoteuthis hamiltoni (Mh, n=7); and 3 benthic octopod species: Octopodid sp. 1 (Oct-1, n=3), Octopodid sp. 2 (Oct-2, n=5) and Cirroctopus glacialis (Cg, n=9). In total the report integrates 543 carbon and nitrogen stable isotope analyses which collectively represent 476 fish, 50 squid, and 17 octopod analyses.

1.1 Distributions and biology of Dissostichus mawsoni

Dissostichus mawsoni are endemic to the Antarctic seas, and like their common prey species Chionobathyscus dewitti, Macrourus whitsoni and Muraenolepis microps, they are generally found in higher latitudes south of the Antarctic Convergence (Gon & Heemstra, 1990). In contrast, Dissostichus eleginoides are widespread throughout the southern ocean, typically extending north from the Antarctic Convergence into subantarctic waters of the Atlantic, Pacific, and Indian Oceans, whilst Antimora rostrata are even more widespread extending north to the equator in abyssal depths (Gon & Heemstra, 1990). All fish species sampled in this report are circumpolar in distribution and are known to overlap in the areas immediately to the south and north of the Antarctic Convergence, especially in the area to the north of the Ross Sea (Hanchet et al. 2003). A wide variation in fishing effort and catch of Dissostichus mawsoni has been noted between years due to annual variations in sea-ice distribution, improved knowledge of fishing grounds, and SSRU catch limits (Hanchet et al., 2007). The variability in area and depth of fishing between seasons has a large influence on length and age composition of Dissostichus mawsoni caught each year. Details of the age and length of maturity and the timing of the spawning cycle are poorly known because the areas is covered by ice for at least 6 months of the year (Sullivan et al., 2005).

Dissostichus mawsoni are primarily a demersal species, but adults are neutrally buoyant and are known to migrate in the pelagic zone over deep water. They grow rapidly, reaching about 60 cm total length after five years and about 100 cm after ten years. After about 25 years, growth slows down at a length of about 150 cm. The maximum recorded age is 48 years and maximum length is 250 cm (Sullivan et al., 2005). Analyses of fish caught over the last ten years in the Ross Sea showed that generally smaller fish were caught from the southern more shallow regions, compared to deeper or more northerly locations (Dunn & Hanchet, 2007). From 1997 to 2007 the majority of Dissostichus mawsoni have been caught on the Slope (72%) compared to the northern

2 Ross Sea (22%) and on the Shelf (7%: Dunn & Hanchet, 2007). Smaller fish tend to be found in the shallower inshore regions, compared to the deeper offshore regions where only larger fish are caught (Philipps et al., 2005). Dunn et al. (2006) reported that Dissostichus mawsoni longer than 120 cm caught in the northern areas of the Ross Sea had a higher mean age-at-length than those caught on the Shelf or the Slope, i.e. large fish in the north had on average experienced a slower growth rate than similar sized fish on the slope. They suggested that the apparent differences between areas in mean age-at-length may be related to a trade-off between somatic growth and reproductive productivity. Patchell (2002) and Fenaughty (2006) noted that fish caught in the north were generally more sexually mature than those caught on the shelf and slope.

A toothfish species profile for the Ross Sea, covering aspects of the biology, fisheries and stock assessment of both Dissostichus mawsoni and Dissostichus eleginoides has been completed by Everson (2002), Fenaughty (2006) and Hanchet (2006). Details of the exploratory fishery and stock assessment are provided by Dunn & Hanchet (2007) and Hanchet et al. (2003, 2006, & 2007). Dissostichus mawsoni make up the greatest proportion of the annual catch in Subarea 88.1 (usually > 85%), whilst Dissostichus eleginoides usually only make up about 1-6 % of the catch with catch rates decreasing further south (Hanchet et al., 2007; Ministry of Fisheries, 2006). Most Dissostichus eleginoides have been sampled in the north of Subarea 88.1. A discussion on the hypothetical lifestyle of Dissostichus mawsoni has been summarised by Hanchet et al. (submitted) and the Ministry of Fisheries (2006) consolidated results from New Zealand’s Ross Sea research programme and discuss the effects of the fishery on the Ross Sea ecosystem.

1.2 Dissostichus mawsoni diet as determined by stomach content analyses

Early studies on the diet of Antarctic toothfish in the Ross Sea were largely opportunistic and based on small sample sizes of fish taken from coastal waters, usually in McMurdo Sound (Calhaem & Christoffel 1969; Eastman 1985; Yukhov 1971). Antarctic silverfish (Pleuragramma antarcticum) were found to dominate the diet, with mysids, cephalopods, amphipods and other (mainly Notothenioid) fish remains also frequently present. More recently, the diet of Dissostichus mawsoni in the Ross Sea fishery, which is most intense in the continental slope region and near oceanic seamounts (Hanchet et al. 2006), has been reasonably well studied by analysing stomach contents (Fenaughty et al. 2003; Stevens 2004, 2006; summarised by Hanchet 2006). Fenaughty et al. (2003) found that icefish (Channichthyidae) and Macrourus whitsoni were the most common prey items identified, with some squid, prawns, rocks and bait also found in the stomachs. Stevens (2004) carried out the first study of Dissostichus mawsoni diet using the Index of Relative Importance (IRI) method (Cortés 1997). The most important prey species were Macrourus whitsoni, which dominated the overall diet and had an IRI of 47%. This was followed by icefish (mainly Chionobathyscus dewitti: IRI of 21%), cephalopods (IRI of 11%), and eel cods (Muraenolepis spp.) and morid (deep-sea) cods (Antimora rostrata) each with an IRI of about 5%. In continental slope waters, Macrourus whitsoni, Chionobathyscus dewitti, and cephalopods predominated in the diet, while on oceanic seamounts Chionobathyscus dewitti were replaced by Antimora rostrata. Stevens (2004) also noted more benthic prey items (including stones) in fish caught on the slope and more benthopelagic prey items from the seamounts. Stevens (2006) analysed the stomach contents of 190 sub-adult (50–100 cm total length) fish. Fish were again the main prey item with an IRI of 92%, with the remainder being mainly cephalopods (mostly glacial squid, Psychroteuthis glacialis). Icefish were the most important prey with an IRI of 28%, while small notothens (Nototheniidae) were more numerous with an IRI of 16%. Macrourus whitsoni, dragonfish (Bathydraconidae), and eel cods were all important, and there were a significant number of stomachs where the fish could not be identified.

It should be noted, however, that stomach contents analysis only gives an instantaneous indication of diet and may not indicate long-term diet since this can vary between years, seasonally, or from day to day depending on several factors including prey availability. Stomach

3 analysis can also give results that are biased towards less digestible and/or more easily identified prey items. Even where organic material in stomachs can be identified, several limitations of conventional dietary approaches can confound such analyses (Hyslop, 1980). We therefore sought to complement the current stomach content dietary information by carrying out carbon and nitrogen stable isotope analysis of predator and potential prey muscle tissue to obtain an integrated, long-term indication of Dissostichus mawsoni diet (sensu Fry & Sherr, 1984; Peterson & Fry, 1987), which is discussed in more detail in section 2.2.1.

2. METHODS

2.1 Sampling locations and methods

During the 2005/6 and 2006/7 fishing seasons, fish, squid and octopod muscle samples were collected for stable isotope analyses by New Zealand Ministry of Fisheries scientific observers from trip numbers 2185 (Vessel Janas), 2186 and 2330 (Vessel Avro Chieftain), and 2183 and 2333 (Vessel San Aspiring). In total, over the 2 sampling seasons, five hundred and nineteen muscle tissue samples were obtained from a combination of: Dissostichus mawsoni (n=142), Dissostichus eleginoides (n=8), Antimora rostrata (n=103), Chionobathyscus dewitti (n=83), Muraenolepis microps (n=76) and Macrourus whitsoni (n=107). Table 1 summarises the sampling times and collections of each fish species. In addition, 4 species of squid (n=50) and three species of octopod (n=17) samples were also obtained from stomachs of Dissostichus mawsoni mostly during the same sampling seasons. The species names, numbers and sizes are given in Table 2. This report integrates isotope data from fish, squids and octopods collected during both the 2005/6 and 2006/7 fishing seasons.

4 Table 1: Fish species from which stable isotope samples data were obtained from the 2005-6 and 2006-7 sampling seasons. New Zealand Ministry of Fisheries common names and species codes are given. Species code Common name Scientific name 2005-6 2006_7 Total sampling sampling sampled TOA (ATO) Antarctic Dissostichus 142 0 142 toothfish mawsoni TOP (PTO) Patagonian Dissostichus 2 6 8 toothfish eleginoides ANT (VCO) Blue antimora Antimora rostrata 103 0 103 (deep sea cod) CHW (ICX) Icefish Chionobathyscus 48 35 83 dewitti MRL (MOY) Moray cod Muraenolepis 1 75 76 microps WGR Whitson’s Macrourus 107 0 107 grenadier whitsoni (Rat-tail) Total 403 116 519

Table 2: Cephalopod species obtained from Dissostichus mawsoni stomachs sampled from longline fishing vessels mainly during the 2005-6 and 2006-7 sampling seasons. “N =”, refers to the number of stable isotope analyses carried out on muscle samples from the cephalopods. Measurements of cephalopods beaks, which are an indicator of overall specimen size, are given as LRL = lower rostral length, CL = crest length and HL = hood length. Cephalopods Species (common name) N = LRL CL HL (mm) (mm) (mm) Octopods (n=17) Octopodid sp. 1 3 9-15 4-7 (benthic) Octopodid sp. 2 5 9-18 4-8 Cirroctopus glacialis 9 11-27 5-14 Squids (n=50) Galiteuthis glacialis 3 4-5 Kondakovia longimana 20 11-19 Psychroteuthis glacialis 20 7-8 Mesonychoteuthis 7 15-41 hamiltoni (Colossal squid)

Figure 1 shows the locations of all fish species sampled during the 2005/6 and 2006/7 fishing seasons. Fish were caught from SSRUs: 88.1A, 88.1C and 88.1E (which we have designated “Northern Area” - primarily sampled from a region of seamounts, ridges and banks 900-1700m deep); 88.1H and 88.1I (here referred to as “Ross Sea Slope” - primarily sampled along the shelf/slope break 900-1600m deep); and 88.1J (which were all located within the “Terra Nova Bay Trench” a depression which is 1200 m maximum depth, extending about 800 km offshore from Terra Nova Bay, where the depth range of samples was 700-1200m). Most Dissostichus mawsoni were sampled in the 2005/6 fishing season during which 75% of the fish were caught on the Ross Sea Slope and 25% in the Northern Area. From 1997 to 2007 the majority of Dissostichus mawsoni have been caught on the Slope (72%) compared to the northern Ross Sea (22%) and the Shelf (7%: Dunn and Hanchet, 2007). During the 2 years of sampling reported here, in 2005/06 ice conditions were good, allowing vessels access to most of the main fishing grounds in Subarea 88.1, whilst in 2006/7, ice conditions were bad and restricted the fishing of

5 some of the southern SSRU’s in January and early February (Hanchet et al., 2007). Table 3 shows the range of water depths for each fish species caught during sampling from the long-line fishing boats. As observed by Hanchet et al. (2007) fishing depth varied between different SSRU’s: fishing was deepest in the northern SSRU’s and shallowest in the southern SSRU’s bordering the coast and ice-shelf. Figures 2-5 plot the locations of all fish, squid and octopod samples caught during the 2005/6 and 2006/07 fishing seasons.

The sampling protocols for the collection of samples were as listed in Pinkerton et al. (2007a). A “mini-fillet” of muscle tissue of approximately 50g was sampled fresh, placed in a sealed bag labelled with fish and set number, and then stored frozen at -20ºC.

Table 3: Sample depths of fish, squids and octopods sampled during the 2005/6 and 2006/7 fishing seasons in the Ross Sea. SSRU = Small Scale Research Unit. Species Scientific Name (n= Sample area number of species sampled for Average Minimum name (and which depth information was depth maximum depth SSRU location) available) (m) depth (m) (m) All species (n=194) 1410 1670 940 Dissostichus mawsoni (n=37) 1275 1460 1205 Dissostichus eleginoides (n=7) 1155 1300 940 Northern Area Antimora rostrata (n=62) 1460 1665 1240 (SSRU 88.1A, Chionobathyscus dewitti (n=33) 1415 1520 1125 88.1C & 88.1E) Muraenolepis microps (n=15) 1315 1660 1115 Macrourus whitsoni (n=33) 1555 1665 1240 Squids, 3 sp. (n=7) 1355 1520 1100 All species (n=294) 1290 1570 910 Dissostichus mawsoni (n=18) 1335 1570 920 Antimora rostrata (n=41) 1290 1480 1160 Ross Sea Slope Chionobathyscus dewitti (n=49) 1270 1475 1115 (SSRU 88.1H, Muraenolepis microps (n=59) 1330 1550 1070 88.1I) Macrourus whitsoni (n=74) 1260 1480 1110 Octopods, 3 sp. (n=15) 1190 1440 910 Squids, 4 sp. (n=38) 1310 1520 910 Terra Nova Bay All species (n=47) 840 1190 725 Trench Dissostichus mawsoni (n=45) 830 1140 725 (SSRU 88.1J) Squids, 2sp. (n=2) 1150 1190 1115

Northern Area Ross Sea Slope Terra Nova Bay Trench

6 Figure 1: Sampling locations of all fish, squid and octopod species sampled for stable isotope analysis in the Ross Sea during the 2005/6 and 2006/7 fishing seasons.

ANT CHW

MRL WGR

Figure 2: Location of four fish species: Antimora rostrata (ANT), Chionobathyscus dewitti (CHW), Muraenolepis microps (MRL) and Macrourus whitsoni (WGR) sampled during 2005/6 and 2006/07 fishing seasons. The plotted symbols represent locations where fish sets were laid out and represent the catch of several individual samples.

TOA TOP

Figure 3: Location of Dissostichus mawsoni (TOA) and Dissostichus eleginoides (TOP) sampled during the 2005/6 and 2006/7 fishing seasons. The plotted symbols represent locations where fish sets were laid out and represent the catch of several individual samples.

7

Gg Kl

Pg Mh Figure 4: Location of the squids Galiteuthis glacialis (Gg), Kondakovia longimana (Kl), Psychroteuthis glacialis (Pg) and Mesonychoteuthis hamiltoni (Mh) predominantly sampled during the 2005/6 and 2006/7 fishing seasons from Dissostichus mawsoni stomach content analysis. The plotted symbols represent locations where fish sets were laid out and represent the catch of several individual samples of Dissostichus mawsoni from which the squid samples were obtained.

8 Oct-1 Oct-2

Cg

Figure 5: Location of the benthic octopods Octopodid sp. 1 (Oct-1), Octopodid sp. 2 (Oct-2) and Cirroctopus glacialis (Cg) predominantly sampled during the 2005/6 and 2006/7 fishing seasons from Dissostichus mawsoni stomach content analysis. The plotted symbols represent locations where fish sets were laid out and represent the catch of several individual samples of Dissostichus mawsoni from which the octopod samples were obtained.

2.2 Analytical techniques

2.2.1 Stable isotope theory The complex interplay of physical, biological and chemical processes in the environment produces distinct isotopic signatures in naturally occurring materials. These natural abundance signatures are increasingly used as tracers in environmental studies. Carbon and nitrogen isotope ratios can track nutrient fluxes between ecosystems and provide information on the trophic structure of foodwebs. Carbon isotopes are a powerful tool for identifying carbon sources and fluxes within ecosystems (Fry & Sherr, 1984; Peterson & Fry, 1987), whilst nitrogen isotope ratios often show distinct enrichments (i.e. increases in δ15N) of on average 3.4 ‰ per successive trophic level and have strong applications in food web and dietary studies (DeNiro & Epstein, 1981; Minagawa & Wada, 1984; Vander Zanden & Rasmussen, 2001). Table 4 explains the concept of stable isotope delta notation and how it is calculated and Table 5 illustrates some of the known factors that can affect isotope signatures in the marine environment.

At high latitudes, most of the carbon isotope variation in phytoplankton, which is subsequently reflected higher up the food chain, is related directly to variations in dissolved CO2 [CO2,aq] (Deuser et al, 1968; Rau et al., 1989). At high latitudes where water temperatures are lower, [CO2,aq] concentrations are higher, and phytoplankton have a lower CO2 demand, phytoplankton exhibit lower d13C values than at temperate and tropical latitudes (Goericke & Fry, 1994; Rau et al., 1982, 1989; Sackett et al., 1965). This is due to greater isotopic fractionation associated with 12 13 CO2 fixation, where the lighter isotope C, is preferentially fixed relative to C (Raven et al,

9 1993). The corollary of this, is that in Antarctic waters phytoplankton become more enriched in 13C as water temperatures increase to the north. Between 50˚ N and 50˚ S, the strong relationship between phytoplankton d13C and latitude breaks down, largely due to greater variability in: phytoplankton growth rates (affected by temperature, light intensity and nutrient supply) (Bigidaire et al, 1997; Burkhardt et al., 1999; Hinga et al., 1994; Kukert & Riebesell, 1998; Laws et al., 1995; Nakatsuka et al., 1992; Popp et al., 1998); species compositions (Falkowski, 1991; François et al., 1993; Goericke et al., 1994; Laws et al., 1995, 1997; Rau et al., 1996, 1997, Wong & Sackett, 1978); and other biological factors, such as cell geometry (Popp et al., 1998; Rau et al., 1997) and cell size (Rau et al., 1996); all of which have a more dominant effect in warmer waters. However, within the Southern Ocean sector, d13C signatures are extremely useful indicators of provenance, where typically a muscle d13C value of -25‰ would signify residence in predominantly Antarctic waters, whereas a -17 to -20‰ signature would indicate a predominantly “sub-tropical waters” habitat over the previous 6-9 months to a year (Cherel & Hobson, 2007; Cherel et al., 2006; Goericke & Fry 1994; Rau et al., 1982,1989; Sackett et al., 1965).

In marine systems benthic algae are on average 13C-enriched by about 5 ‰ compared to average planktonic algae (-17 ±4 ‰ compared to -22 ‰ ±3 ‰ respectively) (France, 1995). A stagnant water boundary layer exists around benthic algae (Jørgensen & Revsbech, 1985; Riber & Wetzel, - 1987) restricting the rate of CO2 or HCO3 diffusion from surrounding waters (Andrews & Abel, 1970; Smith & Walker, 1980). This means that 13C, which is normally discriminated against during photosynthetic carbon uptake, accumulates in the stagnant boundary layer producing a 13 12 13 more positive C/ C ratio in both the CO2 and HCO3- pool. The net effect is a higher δ C signature of benthic algae (Keeley & Sundquist, 1992), which can then be reflected in the isotopic values of benthic consumers (France, 1995; Hobson et al., 1995). A more positive δ13C value in a consumer may indicate strong benthic, as opposed to pelagic trophic links, or as Fry & Sher (1984) conclude, it could alternatively indicate a higher number of trophic links. If this was the case however, one would also expect to see more trophic links reflected in the d15N signal, which is a stronger indicator of trophic position (DeNiro & Epstein, 1981; Minagawa & Wada, 1984; Vander Zanden & Rasmussen, 2001). In most instances it is possible to deduce the likely mechanisms for enriched 13C values by interpreting them in the context of δ15N signals and other supportive biological and oceanographic information.

Table 4: Stable isotope notation: delta (d) values

Stable Isotope Notation: Delta (d) values

In isotope geochemistry relative abundances of isotopes measured in the environment are represented in relation to an international standard. The relative abundances of isotopes are expressed using “delta notation”, written as δ, in units of parts per thousand, more commonly referred to as “per mil” written as ‰.

10 International standards d (‰) = {R (samp) – R (std)/R (std)} * 1000 d13C Pee Dee Belemnite (PDB) (0 ‰), the where, original d13C standard is now exhausted and secondary standards, 13 12 R (samp) is the isotopic ratio (e.g. C/ C or such as NBS-21 graphite, are currently 15N/14N) of the sample used in its place 13 12 R (std) is the isotopic ratio (e.g. C/ C or 15 14 15 N/ N) of an international standard d N N2 in air (0 ‰)

Table 5: Factors affecting isotopic signatures in the marine environment. STW = sub tropical waters, AAW = Antarctic waters Element Good Environmental Effect on isotopic ratio Key references indicator factor of Latitude/ High latitude samples Cherel & Hobson, 2007; Water tend to be d13C depleted Cherel et al., 2006; temperature (e.g. STW -17‰ to - Goericke & Fry 1994; Rau 20‰, AAW -25‰) et al., 1982,1989; Sackett Carbon Energy et al., 1965 Source Terrestrial vs. Terrestrial samples are Hobson, 1999; Hobson et Marine d13C depleted compared al., 1994; Romanuk & to marine samples Levings, 2005 Benthic vs. Benthic samples are d13C France, 1995; Hobson et Pelagic enriched by about 5 ‰ al., 1994,1995 compared to pelagic samples (e.g. benthic algae -17 ±4 ‰; planktonic algae -22 ‰ ±3 ‰) Trophic Small (average +0.4 ‰) DeNiro & Epstein, 1978 Enrichment enrichment per trophic level Nitrogen Trophic Trophic Larger (average +3.4 ‰) DeNiro & Epstein, 1981; Position Enrichment enrichment per trophic Minagawa & Wada, 1984; level Vander Zanden & Rasmussen, 2001

A key feature of the application of isotopes to ecology is their utility in determining the relative contributions of several sources to a mixture, e.g. the proportions of different food sources ingested by consumers (Marguillier et al.., 1997; Melville & Connolly, 2003; Michener & Schell, 1994; Nagelkerken & van der Velde, 2004a, b; Vizzini et al., 2002). The proportional contributions of n+1 different sources to an end-product mixed signature can be uniquely determined using n different isotope tracers (e.g. δ13C, δ15N, δ34S) and linear mixing models based on mass balance equations (Phillips & Gregg, 2001; 2003). In reality multiple sources exist in most environmental situations, where n+1 sources usually exceed the possible number of isotope tracers. Phillips & Gregg (2003) have devised and written IsoSource (available in the public domain at http://www.epa.gov/wed/pages/models.htm), a multi-source mixing model that

11 enables a range of source contributions to be determined when the number of sources is too large to permit unique solutions from stable isotope mixing models.

Muscle samples of fish contain a mixture of protein and lipid. Relative abundances of stable isotopes of carbon and nitrogen in proteins reflect the long-term diet of the animal (on the order of months to a year), and can hence be used to investigate its trophic position in the food web. Lipid synthesis strongly discriminates against the 13C isotope (De Niro & Epstein, 1977, 1978), leading to more negative d13C in lipid-rich tissues that is independent of the organism’s diet. Lipids are therefore known to be 13C depleted relative to proteins and carbohydrates (Rounick & Winterbourn, 1986). Lipids and proteins, or tissues such as fat and muscle, have been shown to have d13C values that differ by as much as 2 ‰ or more (Parker, 1964; Tieszen et al, 1983; van der Merwe 1982; van der Merwe & Vogel, 1978) and more recent studies show that lipid d13C values can differ by as much as 3-8 ‰ (Jardine & Cunjak, 2006; Logan & Lutcavage, 2006; Schell, 2002). It is well known that the proportion of lipid in muscle samples varies with a number of physiological factors of the individual animal, including, for example, age, sex, reproductive stage, and condition. These factors introduce biases to the stable isotope results due to the “lipid contamination issue”, which are not straightforward to correct for retrospectively (Hebert & Keenleyside, 1995; Logan et al., 2008; Mintenbeck et al, 2008, Post et al., 2007). The preferred and recommended method is therefore to lipid extract the sample prior to stable isotope analysis (Ricca et al, 2007).

2.2.2 Lipid extraction, sample preparation and analysis

Approximately 100 mg was sub-sampled from the 50g frozen fillets, being careful to avoid cross contamination between successive samples. The sample was then freeze-dried to remove water, wrapped in a GF/C filter and secured using a stapler. All samples were uniquely labelled using pencil on a small piece of paper which was stapled to the filter. Lipid extraction was performed on a DIONEX 200 accelerated solvent extraction system (ASE) at NIWA Hamilton’s organic chemistry laboratory. Samples were transferred to 22 mL s/s ASE cells and extracted three times with dichloromethane at 70ºC and 1500psi for a static hold time of 5 minutes. Following extraction, samples were then heated to 40ºC in an oven overnight to evaporate any traces of solvent. Samples were then shipped back to NIWA Wellington where they were dried in an oven at 50ºC in their GF/C envelopes for 24 hrs. After drying, samples were put in a dessicator until weighing for isotope analysis.

2.2.3 Verification of Lipid Extraction Process

Lipid-extracted biota samples are expected to have a C:N ratio less than 3.5 (McConnaughty & McRoy 1979; Hebert & Keenleyside 1995; Post et al., 2007), reflecting the elemental composition of protein. A number of the original lipid-extracted Dissostichus mawsoni samples (41 out of 141), analysed in our first batch of analyses had C:N ratios in excess of 3.5, with d13C signatures ranging from –31.5 ‰ to –26.5 ‰ indicating that lipid extraction was incomplete (Bury et al., 2007, 2008). The lipid-contaminated samples were exclusively limited to Dissostichus mawsoni, which are known to be relatively lipid-rich compared to other species, using excess lipids as their buoyancy control in the absence of swim bladders (Eastmann, 1993; La Mesa et al., 2004). Total lipid extraction was achieved for the remainder of the samples in all subsequent batches, as confirmed by C:N ratios consistently lower than 3.5. All samples with C:N ratios >3.5 were removed from the data set, resulting in a range of d13C values for Dissostichus mawsoni of -26.5 ‰ to -23 ‰, consistent with other published data (Burns et al., 1998).

As an internal check on the efficiency and reproducibility of the lipid-extraction process we routinely ran a selection of four internal fish standards (2x Dissostichus mawsoni, 1 x 12 Chionobathyscus dewitti and 1x Macrourus whitsoni) within each lipid-extraction batch. Following lipid-extraction these four standards were analysed for %C and %N content, C:N ratios and stable isotope values. Precision for fish standards are reported below and were consistent with analysis of international standards that had not been through lipid-extraction procedures. We are therefore confident that all subsequent lipid extraction batches produced complete lipid removal and consistent isotopic values.

All stable isotope analyses were carried out on a DeltaPlus (Thermo-Finnigan, Bremen, Germany) continuous flow, isotope ratio mass spectrometer at the NIWA stable isotope laboratory in Wellington (Bury, 1999). Solid samples were weighed out into tin boats and combusted in an NA 1500N (Fisons Instruments, Rodano, Italy) elemental analyser combustion furnace at 1020°C in a flow of oxygen and He carrier gas. Oxides of nitrogen were converted to N2 gas in a reduction furnace at 640°C. N2 and CO2 gases were separated on a Porapak Q gas chromatograph column before being introduced to the mass spectrometer detector via an open split Conflo II interface (Thermo-Finnigan, Bremen, Germany). CO2 and N2 reference gas standards were introduced to the mass spectrometer with every sample analysis. ISODAT (Thermo-Finnigan) software was 15 13 used to calculate δ N values against atmospheric air, and δ C values against the CO2 reference gas relative to PDB, correcting for 17O. Percent C and % N values were calculated relative to a solid laboratory reference standard of DL-Leucine (DL-2-Amino-4-methylpentanoic acid, C6H13NO2, Lot 127H1084, Sigma, Australia) at the beginning of each run. Internal standards were routinely checked against National Institute of Standards and Technology (NIST) standards. Accuracy and precision data for NIST standard analyses are given in Pinkerton et al. (2007a).

Repeat analysis of NIST standards produces data accurate to within 0.1-0.5 ‰ for δ15N and 0.3- 0.4 ‰ for δ13C and a precision (1 s.d.) of better than 0.5 ‰ for N and 0.25 ‰ for C. For % N and C content, data are accurate to within 0.4%, with a precision (1 s.d.) usually better than 0.3% for N and 0.2% for C. Analysis of fish standards which had undergone triple-phase lipid-extraction (n=24 for each of the four standards) consistently produced C:N ratios <3.5 (average 3.2 ± 0.1) with a δ15N and δ13C precision (1 s.d.) of 0.3 ‰ and 0.5 ‰ respectively. In addition to the four internal fish standards we also ran some replicate analysis of samples, the results of which are presented in Table 6. Delta 15N replicate analysis produced a maximum 1 standard deviation variation of ±0.37 ‰, with values usually below ± 0.2 ‰, whilst d13C replicate analysis gave a maximum variation of ±0.19 ‰, with values mostly below ±0.15 ‰.

Table 6: Replicate stable isotope analysis of a selection of fish muscle samples. Fish Fish Replication %N average d15N average %C average d13C average species sample of analysis ±1 standard ±1 standard ±1 standard ± 1 standard code i.d. deviation deviation deviation deviation TOA 14 4 13.39 +/- 0.07 10.21 +/- 0.19 44.24 +/- 0.13 -24.68 +/- 0.13 TOA 17 2 11.80 +/- 0.19 13.79 +/- 0.37 46.46 +/- 0.90 -26.37 +/- 0.15 TOA 210 2 13.83 +/- 0.09 13.38 +/- 0.16 43.58 +/- 0.07 -23.47 +/- 0.07 TOA 220 2 14.12 +/- 0.16 12.84 +/- 0.06 43.96 +/- 0.03 -24.60 +/- 0.06 ANT 201 2 14.38 +/- 0.03 11.11 +/- 0.11 44.46 +/- 0.13 -22.99 +/- 0.00 MRL 31 4 14.27 +/- 0.19 11.62 +/- 0.25 44.89 +/- 0.93 -24.55 +/- 0.19 WGR 230 2 14.11 +/- 0.08 10.78 +/- 0.17 42.77 +/- 0.12 -23.14 +/- 0.14 WGR 248 2 14.16 +/- 0.04 11.80 +/- 0.22 43.42 +/- 0.06 -23.28 +/- 0.05

2.3 Statistical Analyses

A stepwise generalised linear model (GLM) was carried out to determine the best predictors of δ15N and δ13C. The response variables δ15N and δ13C were treated independently for the current analysis. The variables offered to the model included sex (not Muraenolepis microps), length,

13 length as a function of sex (not Muraenolepis microps), area, and depth. Total length was used for all species except Macrourus whitsoni, where snout-anus length (cm) was used. Unsexed fish were excluded from analysis. Sex was not used for Muraenolepis microps as too few samples were sexed (F=33, M=2, unsexed=41). Area here was SSRU, and was always one of 88.1C, 88.1E, 88.1H, 881.I and 88.1J.

Linear, 2-degree and 3-degree polynomial transformations were also considered for the continuous data (i.e., length and depth) in evaluating each of the models. Stable isotope composition in each species was considered separately. We had sufficient samples for analysis of the predator species (Dissostichus mawsoni), and the prey (Antimora rostrata, Macrourus whitsoni, and Chionobathyscus dewitti). We excluded fish where no length information was available (15 samples).

The Akaike Information Criteria (AIC) was used to determine the significance of each variable, where residual sum of squares was penalised by twice the number of parameters times the residual mean square of the initial model (Akaike,1974). Terms were added in order of greatest reduction in AIC in a stepwise manner, until a final model was chosen from a sequence of steps that minimised the AIC statistic. Diagnostic plots were examined for each model run to evaluate departures from normality.

3. RESULTS AND DISCUSSION

3.1 Biological Measurements

For each of the fish samples labelled with set number and station number, the following data were obtained from the CCAMLR observer database: vessel name, sampling area, position (including latitude, longitude, and depth) and date. For most of these samples we also obtained information on fish length (down to the nearest cm), fish weight (kg) and sex (Tables 7–8). In addition, a total of 519 tissue samples were processed for stable isotope analysis, producing a C and N stable isotope dataset of 476 samples that were reliably lipid-extracted (Table 7). In all the subsequent analysis, fish that were originally coded as CHW (identified as Chionobathyscus dewitti) and ICX (identified to the family level only as Channichthyidae) have been amalgamated and plotted under the code “icefish” (CHW) which universally refers to Chionobathyscus dewitti. Fish identification at NIWA has subsequently confirmed that all the icefish sampled were Chionobathyscus dewitti (Peter Marriott, Peter McMillan pers. com). Similarly, all fish identified to the family level as MOL (“Muraenolepididae”) were reassigned as MRL (Muraenolepis microps).

Table 7: Number of fish species sampled and number of individuals for which biological measurements were carried out (refer to Table 1 for explanation of fish codes). *141 Dissostichus mawsoni samples were analysed, but 41 samples were excluded from the dataset as they had C:N ratios > 3.5, signifying residual lipid-contamination. Fish species Total no. Sex Snout-anus Total fish Fish C, N code of fish determination length length weights stable individual measurements measurements taken isotope s sampled made made analysed TOA 142 142 0 142 142 100 (141)* TOP 8 8 8 8 8 8 ANT 103 102 0 102 102 103 CHW/ICX 83 66 0 67 67 83 MRL/MOY 76 39 0 72 72 75 WGR 107 107 107 35 103 107

14 Total 519 464 115 426 494 476 (517)*

Table 8: Number of fish species sampled in 2005/6 and 2006/7 seasons according to vessel trip number (refer to Table 1 for explanation of fish codes). Fish Total no. 2183 2185 2186 2330 2333 species of fish San Janas Avro Avro San code individuals Aspiring Chieftain Chieftain Aspiring sampled TOA 142 112 0 30 0 0 TOP 8 2 0 0 4 2 ANT 103 26 54 23 0 0 CHW/ICX 83 28 20 0 28 7 MRL/MOY 76 0 0 1 28 47 WGR 107 28 55 24 0 0 Total 519 196 129 78 60 56

3.2 Stable Isotope Fish Species Results

Carbon and nitrogen stable isotope data for all fish species and for all SSRU areas sampled, incorporating fish caught during 2005/6 and 2006/7 fishing seasons, are plotted in Figure 6.

All species showed a large variation in d15N. Dissostichus mawsoni exhibited a range of 7 ‰ (9– 16 ‰) in δ15N, which is equivalent to two trophic steps. All fish, except Antimora rostrata (2.7 ‰) showed a d15N range greater than 3.4 ‰ spanning more than one trophic step: Chionobathyscus dewitti had a range of 4.7 ‰, Muraenolepis microps 3.5 ‰ and Macrourus whitsoni 5.8 ‰. This implies that the diet of all species sampled was variable, or that individual species were eating a similar diet, which itself varied in size and trophic status.

A plot of the average species values (Figure 6: lower plot) showed that overall, Dissostichus mawsoni (δ15N: 13.55 ±1.10, 1 s.d., n=100) and Dissostichus eleginoides (δ15N: 13.49 ±1.47, 1 s.d., n=8) occupied the same top trophic level, with Antimora rostrata (δ15N: 10.19 ± 0.59, 1 s.d., n=103), Muraenolepis microps (δ15N: 11.00 ± 0.7, 1 s.d., n=75) and Macrourus whitsoni (δ15N: 10.61 ± 0.81, 1 s.d., n=107) all occupying a trophic level below them. Chionobathyscus dewitti occupied the lowest trophic level of all fish analysed, giving an average δ15N value of 9.33 ± 1.24, 1 s.d. (n=83). It should be noted that there was considerable isotopic overlap for both δ15N and δ13C for all sampled fish other than Dissostichus mawsoni and Dissostichus eleginoides.

Large variations in d13C for each species (around 3 ‰ for each species) indicated a variation in source of carbon within individual species. Fish with enriched d13C may have been feeding further north in warmer waters or may have had a stronger benthic compared to pelagic source of carbon.

Dissostichus eleginoides, which were only caught in Subarea 88.1C, had a lower range of δ15N values (10-15 ‰) than Dissostichus mawsoni, possibly indicating a less variable diet. They also had more enriched d13C values (-22.72 ± 0.97, 1.s.d., n=8) than Dissostichus mawsoni (-24.63 ± 0.58, 1.s.d., n=100), indicating they were most likely predominantly feeding to the north in warmer waters. Their carbon signatures were similar to those of king penguins (-22.3 ±0.2) feeding in oceanic waters sampled close to Kerguelen Island (Cherel and Hobson, 2007) and it is therefore likely that they migrated south to the Ross Sea from the Polar Frontal Zone.

15 18

16 Dissostichus mawsoni (n=100) Dissostichus eleginoides (n=8) 14 Antimora rostrata (n=103) N 5

1 12

d Chionobathyscus dewitti (n=83)

10 Muraenolepis microps (n=75)

8 Macrourus whitsoni (n=107)

6 -28 -26 -24 -22 -20 d 13C

18

16 Dissostichus mawsoni (n=100) Dissostichus eleginoides (n=8) 14 Antimora rostrata (n=103) N 5 1

12

d Chionobathyscus dewitti (n=83)

10 Muraenolepis microps (n=75)

8 Macrourus whitsoni (n=107)

6 -28 -26 -24 -22 -20 d 13C

Figure 6: Delta 13C vs. δ15N plot for all SSRU areas sampled, incorporating fish caught during 2005/6 and 2006/7 fishing seasons. All Dissostichus mawsoni samples with C:N ratios in excess of 3.5 have been removed from the data set. The upper plot shows all data points and the lower plot shows averages for each species (±1 s.d.).

16

Dissostichus mawsoni (n=100) Dissostichus eleginoides (n=8) 18 18

16 16

14 14 N N 5 5 1 1

12 12 d d

10 10

8 y = 0.0066x + 12.702 8 y = 0.0557x + 7.516 R2 = 0.0277 R2 = 0.6227 6 6 0 50 100 150 200 0 50 100 150 200 Total fish length (cm) Total fish length (cm) Figure 7: Delta 15N versus fish length (total fish length or snout-anus length) for Dissostichus mawsoni and Dissostichus eleginoides. The plots show data compiled from the two sampling periods 2005/6 and 2006/7 incorporating all SSRU areas, although note that Dissostichus eleginoides was only caught in the Northern Area.

In order to establish whether there was a relationship between fish size and trophic level as has been reported in previous studies we first plotted δ15N against fish length (Figures 7 & 8). For Dissostichus eleginoides (r2 = 0.62) and Chionobathyscus dewitti (r2 = 0.53) there appeared to be a reasonably strong relationship between fish size and δ15N values, implying that as these fish increase in size they increase their trophic status. Extremely weak relationships, however, were observed for Dissostichus mawsoni, Antimora rostrata, Macrourus whitsoni and Muraenolepis microps. A more complex statistical analysis using a generalised linear model (GLM) was then employed to further investigate factors that were significant in explaining both δ15N and δ13C variations for all fish species (with the exception of Dissostichus eleginoides, for which we only had 8 samples). Specific details and parameters of the model are outlined in the methods, numbers of samples used in the analysis are given in Table 9, and Table 10 lists the results of the model runs.

17 Antimora rostrata (n=102) Chionobathyscus dewitti (n=69) (no length data for 1 sample) (no length data for 14 samples)

14 14

12 12 N N 5 5 1 1

10 10 d d

8 8 y = 0.0192x + 9.1082 y = 0.1693x + 2.8611 R2 = 0.0742 R2 = 0.5295 6 6 0 20 40 60 80 0 20 40 60 80 Total fish length (cm) Total fish length (cm)

Muraenolepis microps (n=68) Macrourus whitsoni (n=107) (no length data for 8 samples) 14 14

12 12 N 5 1 N

10 5 1 d 10

d

8 8 y = 0.0154x + 10.284 y = 0.0214x + 10.104 R2 = 0.0072 R2 = 0.0179 6 6 0 10 20 30 40 0 20 40 60 80 Snout-anus length (cm) Total fish length (cm)

Figure 8: Delta 15N versus fish length (total fish length or snout-anus length) for Antimora rostrata, Chionobathyscus dewitti, Macrourus whitsoni and Muraenolepis microps. The plots show data compiled from the two sampling periods 2005/6 and 2006/7 incorporating all SSRU areas.

Table 9: Number of samples used for statistical analyses. F = female, M= male, U=unclassified Sex (N=) Length Depth SSRU (N=) (N=) (N=) Species N = F M U TOA 102 54 46 2 100 100 100 ANT 103 75 24 4 102 103 103 CHW 83 50 15 18 69 82 82 MRL 76 33 2 41 68 76 76 WGR 107 80 27 0 107 107 107

Table 10: Summary of regression results for within-species variation in δ15N and δ13C for all species. All regressions are significant at the p<0.001 level. In the model columns, L=length, Z=depth, “^2” indicates quadratic model, “^3” indicates a cubic model. δ15N δ13C Correlation Reduction Correlation Reduction in Species N = model coefficient in variance model coefficient variance TOA 100 SSRU, L^2 0.47 0.22 SSRU, L^2 0.52 0.27 ANT 99 SSRU, L 0.72 0.51 SSRU, Z 0.75 0.57

18 CHW 65 SSRU, L 0.81 0.65 SSRU, L^3 0.88 0.78 MRL 61 SSRU, L 0.62 0.38 SSRU 0.83 0.69 WGR 107 SSRU, L 0.51 0.26 SSRU 0.82 0.67

3.2.1 Variation of δ15N and δ13C for Dissostichus mawsoni (TOA)

The generalised linear model suggested two factors were significant in explaining the variation of δ15N in Dissostichus mawsoni: SSRU and length (as a quadratic expression). The best model had reduction in residual deviance of 23%, 5 d.f, F=5.3, p<0.001 (highly significant). The diagnostics were reasonable, showing no major departures from normality. The predicted effects of length and SSRU are given in Figure 9 (upper). While treating length as a function of sex, and including depth, resulted in small improvements in the model fit, they were not significant given the increased degrees of freedom. The 3-degree polynomial functions of length were not significant. The generalised linear model suggested two factors were significant in explaining the variation of δ13C in Dissostichus mawsoni: SSRU, and length (quadratic function). The best model had reduction in residual deviance = 35%, 5 d.f, F=9.9, p<0.001 (highly significant). The diagnostics were reasonable, showing no major departures from normality. The predicted effects of length and SSRU are given in Figure 9 (lower).

5 5 1 1 N d

d 0 0 e 1 1 t c i d e r P 5 5 0 0

0 50 100 150 200 881C 881H 881I 881J

Length (cm) SSRU

0 0 2 2 - - 4 4 C d 2 2 - -

d e t c i d e 8 8 r 2 2 - - P 2 2 3 3 - -

0 50 100 150 200 881C 881H 881I 881J Length (cm) SSRU Figure 9: Top row: relative effects on δ15N for Dissostichus mawsoni (TOA) model for length (left) and SSRU (right). Bottom row: relative effects on δ13C for Dissostichus mawsoni, length (left) and SSRU (right). Dashed lines indicate twice the standard error in the fitted results. Solid lines to the bottom of the left panels show the distribution of data in the sample.

3.2.2 Variation of δ15N and δ13C for Macrourus whitsoni (WGR)

The generalised linear model suggested two factors were significant in explaining the variation of δ15N in Macrourus whitsoni: SSRU and length (as a linear expression). The best model had reduction in residual deviance = 26%, 3 d.f, F=12.3, p<0.001 (highly significant). The diagnostics were reasonable, showing no major departures from normality. The predicted effects of length and SSRU are given in Figure 10 (upper). The generalised linear model suggested only one factor was significant in explaining the variation of δ13C in Macrourus whitsoni: SSRU. The best model had reduction in residual deviance = 67%, 2 d.f, F=108, p<0.001 (highly significant). The

19 diagnostics were reasonable, showing no major departures from normality. The predicted effects of SSRU are given in Figure 10 (lower).

5 1 5 1 N 0 d 1

d 0 e 1 t c i d e r 5 P 5 0 0

0 5 10 15 20 25 30 35 881C 881H 881I

Length (cm) SSRU

4 2 - 6 C 2 - d

d e 8 t 2 c - i d e r 0 P 3 - 2 3 -

881C 881H 881I

SSRU

Figure 10: Top row: Relative effects on d15N for Macrourus whitsoni (WGR) model for length (left) and SSRU (right). The solid line to the bottom of the left panel shows the distribution of data in the sample. Note that there are no data to which to fit the model at lengths below 12 cm and greater than 31 cm (snout-anus length). Bottom row: Relative effects on d13C for Macrourus whitsoni model for SSRU. Dashed lines indicate twice the standard error in the fitted results.

3.2.3 Variation of δ15N and δ13C for Chionobathyscus dewitti (CHW)

The generalised linear model suggested two factors were significant in explaining the variation of δ15N in Chionobathyscus dewitti: length (as a linear expression) and SSRU. The best model had reduction in residual deviance = 65%, 1 d.f, F=28.4, p<0.001 (highly significant). The diagnostics were reasonable, showing no major departures from normality. The predicted effects of length are given in Figure 11 (upper). The generalised linear model suggested two factors were significant in explaining the variation of δ13C in Chionobathyscus dewitti: length (cubic function) and SSRU. The best model had reduction in residual deviance = 77%, 3 d.f, F=36.1, p<0.001 (highly significant). The diagnostics were reasonable, showing no major departures from normality. The predicted effects of length are given in Figure 11 (lower).

20 5 5 1 1 N C 0 0 d d 1 1

d d e e t t c c i i d d e e r r 5 5 P P 0 0

15 20 25 30 35 40 45 50 881C 881E 881H 881I Length (cm) SSRU 4 4 2 2 - - 6 6 C C 2 2 - - d d

d d e e 8 8 t t 2 2 c c - - i i d d e e r r 0 0 P P 3 3 - - 2 2 3 3 - -

15 20 25 30 35 40 45 50 881C 881E 881H 881I

Length (cm) SSRU

Figure 11: Top: Relative effects on δ15N for the Chionobathyscus dewitti (CHW and ICX combined) model for length (in 88.1H), and by SSRU (for snout-anus length of 40 cm). Bottom: Relative effects on δ13C for the Chionobathyscus dewitti (CHW and ICX combined) model for length (in 88.1H) and by SSRU (for snout-anus length 40 cm). Dashed lines indicate twice the standard error in the fitted results. The solid lines to the bottom of the panels show the distribution of data. Note that there are no data to which to fit the model at lengths below 28 cm and greater than 46 cm.

3.2.4 Variation of δ15N and δ13C for Antimora rostrata (ANT)

The generalised linear model suggested two factors were significant in explaining the variation of δ15N in ANT: SSRU and length (as a linear expression). The best model had reduction in residual deviance = 51%, 3 d.f., F=33.4, p<0.001 (highly significant). The diagnostics were reasonable, showing no major departures from normality. The predicted effects of length and SSRU are given in Figure 12 (upper). The generalised linear model suggested two factors were significant in explaining the variation of δ13C in Antimora rostrata: SSRU, and depth (linear function). The best model had reduction in residual deviance = 60%, 5 d.f., F=28.4, p<0.001 (highly significant). The diagnostics were reasonable, showing no major departures from normality. The predicted effects of depth and SSRU are given in Figure 12 (lower).

21 5 5 1 1 N 0 0 d 1 1

d e t c i d e r 5 5 P 0 0

0 20 40 60 80 881C 881H 881I

Length (cm) SSRU

0 0 2 2 - - 2 2 2 2 - - C d 4 4

2 2 d - - e t c i 6 6 d 2 2 e - - r P 8 8 2 2 - - 0 0 3 3 - -

1000 1200 1400 1600 1800 881C 881H 881I Depth (m) SSRU Figure 12: Top row: Relative effects on δ15N for Antimora rostrata (ANT) model for length (left) and SSRU (right). Bottom row: Relative effects on δ13C for Antimora rostrata model for depth (left) and SSRU (right). The dashed lines indicate twice the standard error in the fitted results. The solid lines to the bottom of the left panels show the distribution of data in the sample.

3.2.5 Variation of δ15N and δ13C for Muraenolepis microps (MRL)

The generalised linear model suggested two factors were significant in explaining the variation of δ15N in Muraenolepis microps: SSRU and length (as a linear expression). The best model had reduction in residual deviance = 38%, 3 d.f, F=6.8, p<0.001 (highly significant). The diagnostics were reasonable, showing no major departures from normality. The predicted effects of length and SSRU are given in Figure 13 (upper). The generalised linear model suggested one factor was significant in explaining the variation of δ13C in Muraenolepis microps: SSRU. The best model had reduction in residual deviance = 69%, 4 d.f, F=31.0, p<0.001 (highly significant). The diagnostics were reasonable, showing no major departures from normality. The predicted effects of depth and SSRU are given in Figure 13 (lower).

5 5 1 1 N 0 0 d 1 1

d e t c i d e r 5 5 P 0 0

20 30 40 50 60 881C 881H 881I

Length (cm) SSRU

22 0 2 - 2 2 - 4 2 - 6 2 - 8 2 - 0 3 -

881C 881H 881I

SSRU

Figure 13: Top row: Relative effects on δ15N for Muraenolepis microps (MRL) model for length (left) and SSRU (right). Bottom row: Relative effects on δ13C for Muraenolepis microps model for SSRU. The dashed lines indicate twice the standard error in the fitted results. The solid lines to the bottom of the left top panel show the distribution of data in the sample.

3.3 Stable Isotope Fish and Cephalopod Results

Stable isotope averages of individual squid and octopod species were compiled with the averaged fish species isotope data (Figure 14, Table 11). With the exception of the colossal squid (Mesonychoteuthis hamiltoni, average δ15N: 11.52‰ ± 1.1, 1 s.d., n=7), which plots one trophic level below Dissostichus mawsoni and Dissostichus eleginoides, (both with average δ15N values of 13.5‰), the squids are at the lowest trophic level of all species sampled. Figure 15 shows the average δ15N squid value (excluding Mesonychoteuthis hamiltoni), which was 7.8‰ ± 0.59, 1 s.d. (n=43), occupying a lower trophic level than Chionobathyscus dewitti (δ15N: 9.33‰ ± 1.24, 1 s.d., n=83). On average octopods occupy a similar trophic level (δ15N: 9.89‰ ± 1.29, 1 s.d, n=17) to the four main species of fish sampled (excluding Dissostichus spp.) and are trophically higher than the sampled squids. The squid species sampled are known to be pelagic, which explains their more depleted δ13C signature (-25.18‰ ± 0.63, 1 s.d., n=43) than the octopods (δ13C: -23.45 ± 1.26, 1 s.d., n=17), which are benthic feeders. The enriched δ15N value of Mesonychoteuthis hamiltoni compared to the other squid species reflects its larger size and higher trophic feeding level. In addition, the more positive δ13C value (-22.98 ± 2.10, 1 s.d., n=7) is indicative of a feeding habitat in sub-tropical waters and subsequent migration to the Ross Sea where the specimens were caught.

23 16 Dissostichus mawsoni (n=100) Dissostichus eleginoides (n=8) Antimora rostrata (n=103) 14 Chionobathyscus dewitti (n=83) Muraenolepis microps (n=75) 12 Mh Macrourus whitsoni (n=107) N 5

1 Octopodid sp. 1 (n=3)

d 10 Octopodid sp. 2 (n=5) Cirroctopus glacialis (n=9) 8 Octopod Galiteuthis glacialis (n=3) Kondakovia longimana (n=20) Squid s Psychroteuthis glacialis (n=20) 6 Mesonychoteuthis hamiltoni (n=7) -28 -26 -24 -22 -20 d 13C

Dissostichus mawsoni (n=100) 16 Dissostichus eleginoides (n=8) Antimora rostrata (n=103) 14 Chionobathyscus dewitti (n=83) Muraenolepis microps (n=75) 12 Macrourus whitsoni (n=107) N

5 Octopodid sp. 1 (n=3) 1

d Octopodid sp. 2 (n=5) 10 Cirroctopus glacialis (n=9) Galiteuthis glacialis (n=3) 8 Kondakovia longimana (n=20) Psychroteuthis glacialis (n=20) 6 Mesonychoteuthis hamiltoni (n=7) -28 -26 -24 -22 -20 d 13C

Figure 14: Average d13C and d15N values of fish, octopods and squids from samples collected during the 2005/6 and 2006/7 fishing seasons in the Ross Sea area, combining data from all SSRU’s. Top figure shows 1 s.d. error bars, whilst these have been removed in the bottom figure for clarity of plots.

24 Dissostichus mawsoni (n=100)

Antimora rostrata (n=103) 16 Chionobathyscus dewitti (n=83)

14 Muraenolepis microps (n=75)

Macrourus whitsoni (n=107) 12

N Octopods (3 sp., n=17) 5 1

d 10 Squids (3 sp., n=43) Mesonychoteuthis hamiltoni (n=7)

8 Pleuragramma antarcticum, lipid corrd. (n=4) Burns et al., 1998 Dissostichus mawsoni (n=5), lipid corrd. 6 Burns et al., 1998 -28 -26 -24 -22 -20 d 13C

Figure 15: d13C and d15N average values of Dissostichus mawsoni and their potential prey over the whole area sampled during the 2005/6 and 2006/7 fishing seasons. Lipid-corrected isotopic data for Dissostichus mawsoni and Pleuragramma antarcticum from Burns et al. (1998) are also plotted for comparison.

Table 11 collates data from this study with published stable isotope analyses of fish and mammals in the Ross Sea. There are no published stable isotope cephalopod data from the Ross Sea for comparison with our data. Within each group, data have been listed in order of decreasing trophic status. Trematomus loennbergii (Burns et al., 1998), Trematomus newnesi (Krahn et al., 2006) and Pagothenia borchgrevinki (Burns et al., 1998; Krahn et al., 2006) were all at similar trophic levels to Antimora rostrata, Macrourus whitsoni, and Muraenolepis microps, whilst Rhigophilia dearborni (Burns et al., 1998), a sedentary demersal fish, had similar δ15N values to both Dissostichus species. Dissostichus mawsoni and Dissostichus eleginoides both had higher δ15N values than any of the reported mammal species (Burns et al., 1998; Krahn et al., 2006; Zhao et al., 2004), indicating that they have similar trophic importance in the food web to these apex predators.

25 Table 11: Comparison of δ15N, δ13C and C:N ratio values in fish and cephalopods from this study with fish and mammals sampled in the Ross Sea from published literature. (1) This study; (2) Burns et al. (1998); (3) Zhao et al. (2004); (4) Krahn et al. (2006). Mammal data presented is for adult animals. Data published in Burns et al. (1998) and Zhao et al. (2004) have both been lipid-corrected using the equation given in Post et al. (2007). Krahn et al. (2006) samples were lipid extracted prior to analysis.

15 13 δ N (‰) δ C (‰) C:N ratio Mean ± 1 Mean ± 1 Mean ± 1 Scientific name Common Name N = s.d. Min Max s.d. Min Max s.d. FISH Dissostichus mawsoni (1) Antarctic toothfish 100 13.6 ± 1.1 9.2 16.0 -24.6 ± 0.6 -26.1 -22.9 3.2 ± 0.1 Dissostichus mawsoni (2) Antarctic toothfish 5 13.5 ± 0.2 n.d. n.d. -25.9 ± 0.6 -30.3 -27.1 7.2 ± 1.2 Dissostichus eleginoides (1) Patagonian toothfish 8 13.5 ± 1.5 10.1 14.8 -22.7 ± 1.0 -24.2 -21.1 3.1 ± 0.1 Rhigophilia dearborni (2) Antarctic eel pout 1 13.2 -23.6 Trematomus loennbergii (2) Deepwater notothen, Scaly rockcod 2 11.9 ± 2.2 10.3 13.4 -22.7 ± 1.6 -23.8 -21.5 Muraenolepis microps (1) Moray Cod 75 11.0 ± 0.7 9.4 12.9 -24.9 ± 0.5 -25.7 -23.3 3.2 ± 0.2 Pagothenia borchgrevinki (2) Bald notothen 3 11.0 ± 0.2 -22.5 ± 0.5 -23.5 -21.8 Pleuragramma antarcticum (2) Antarctic Silverfish 4 10.9 ± 0.2 n.d. n.d. -25.3 ± 0.4 -26.0 -21.5 6.4 ± 0.6 Trematomus newnesi (4) Dusky notothen, dusky rockcod 2 10.9 ± 0.7 10.4 11.4 -25.2 ± 0.7 -25.7 -24.7 Pagothenia borchgrevinki (4) Bald notothen 2 10.7 ± 0.8 10.1 11.2 -25.2 ± 1.2 -26.0 -24.3 Macrourus whitsoni (1) Whitson's Grenadier, Rat-tail 107 10.6 ± 0.8 7.7 13.5 -24.7 ± 0.8 -26.1 -22.9 3.1 ± 0.1 Antimora rostrata (1) Blue Antimora, Deep Sea Cod 103 10.2 ± 0.6 8.6 11.3 -24.3 ± 1.0 -26.3 -22.6 3.1 ± 0.1 Chionobathyscus dewitti (1) Icefish 83 9.3 ± 1.2 7.2 11.9 -25.2 ± 0.8 -26.5 -23.6 3.1 ± 0.1 CEPHALOPODS Mesonychoteuthis hamiltoni (1) Colossal squid 7 11.5 ± 0.1 10.0 13.3 -23.0 ± 2.1 -25.1 -19.2 Octopodid sp. 1 (1) 3 10.5 ± 0.9 9.6 11.3 -23.9 ± 0.1 -24.0 -23.8 3.1 ± 0.0 Cirroctopus glacialis (1) 9 10.4 ± 1.2 8.4 12.2 -23.0 ± 1.6 -24.2 -19.5 3.2 ± 0.1 Octopodid sp. 2 (1) 5 8.6 ± 0.6 8.2 9.7 -24.0 ± 0.4 -24.4 -23.4 3.2 ± 0.1 Galiteuthis glacialis (1) 3 8.1 ± 0.4 7.7 8.6 -24.7 ± 0.9 -25.7 -24.1 3.1 ± 0.0 Psychroteuthis glacialis (1) 20 7.9 ± 0.6 6.9 8.7 -25.3 ± 0.6 -26.4 -23.2 3.1 ± 0.1 Kondakovia longimana (1) 20 7.6 ± 0.6 6.8 8.9 -25.1 ± 0.6 -25.7 -23.9 3.2 ± 0.1 MAMMALS Orcinus orca (4) Killer whales, Orca 15 13.3 ± 0.4 12.6 14.1 -23.7 ± 0.2 -24.1 -23.4 Leptonychotes weddellii (2) Weddell seal 12 13.1 ± 0.2 12.5 14.3 -22.5 ± 0.1 -23.0 -21.7 3.8 ± 0.1 Leptonychotes weddellii (3) Weddell seal 22 13.0 ± 0.9 10.5 14.7 -21.9 ± 0.8 -23.7 -20.6 Hydrurga leptonyx (3) Leopard Seal 2 12.3 ± 0.5 11.9 12.8 -21.8 ± 0.3 -21.9 -21.6 Omatophoca rossii (3) Ross Seal 33 10.3 ± 0.5 9.2 11.4 -21.2 ± 0.4 -22.0 -20.3 Lobodon carcinophagus (3) Crabeater seal 30 8.4 ± 0.6 7.1 9.5 -23.5 ± 1.0 -25.7 -21.6

26

Dissostichus mawsoni (n=100)

Antimora rostrata (n=103) 18 Chionobathyscus dewitti (n=83)

16 Muraenolepis microps (n=75)

Macrourus whitsoni (n=107) 14

N Octopods (3 sp., n=17) 5 1

d 12 Squids (3 sp., n=43) Mesonychoteuthis hamiltoni (n=7)

10 Pleuragramma antarcticum, lipid corrd. (n=4) Burns et al., 1998 Dissostichus mawsoni (n=5), lipid corrd. 8 Burns et al., 1998 -28 -26 -24 -22 -20 d 13C

Figure 16: Isotope prey polygon for Dissostichus mawsoni over the whole area sampled. Fractionation factors of +3.4 ‰ (d15N) and +0.4 ‰ (d13C) have been applied to all prey items.

Earlier work in the Ross Sea, where we aimed to “quantify relative trophic dependence of Dissostichus mawsoni on 3 key prey species (Macrourus whitsoni, Chionobathyscus dewitti, Antimora rostrata)”, led us to conclude that we had sampled insufficient fish species to be able to quantify the proportions of prey items ingested by Dissostichus mawsoni (Bury et al., 2007; Pinkerton et al., 2007a,b). Applying a fractionation factor of +3.4‰ to δ15N values and +0.4‰ to δ13C values of potential prey items of Dissostichus mawsoni, we plotted an isotope predator-prey polygon which showed that Dissostichus mawsoni plotted outside of the polygon. This indicated that we had failed to sample some important prey items (see Phillips and Gregg, 2003, Figure 6A). During this current study, we attempted to solve the problem by sampling and analysing additional potential prey species, namely, Muraenolepis microps (for which we only had one sample in the previous season), squids and octopods. Plotting these additional data on a Dissostichus mawsoni prey polygon isotope plot, however, still showed Dissostichus mawsoni plotting just outside of the prey polygon (Figure 16).

A logistical short-coming of the sampling in the Ross Sea observer programmes during both the 2005/6 and 2006/7 sampling seasons was that we were unable to obtain any Pleuragramma antarcticum samples, which are known to be a component of the Dissostichus mawsoni diet from published stomach content analyses (Calhaem and Christoffel, 1969; Fuiman et al., 2002; Pakhomov and Tseytlin, 1992; Stevens 2006). Searching the literature, there are only four published papers which report Pleuragramma antarcticum stable isotope data, which we have summarised in Table 11. The d13C data reported in Burns et al., (1998, Table 6 therein) were unfortunately from non-lipid-extracted samples and the raw data reported in the paper cannot be directly compared to our data. We have therefore applied a lipid-correction factor using Burn’s (1998) un-corrected δ13C and C:N data and the equation given in Post et al. (2007), which is reported as one of the better normalisation models (Mintenbeck et al., 2008). The Pleuragramma

27 antarcticum average, plus a lipid-corrected Dissostichus mawsoni average value (Burns et al., 1998), have both been plotted on Figures 15 and 16.

The Burns et al. (1998) average Dissostichus mawsoni δ15N value (13.5‰ ± 0.2, n=5) is very similar to our measured value (13.6 ± 1.1, n=100), however their δ13C values were 1.3 ‰ more depleted than our values (-25.9 ± 0.6 compared to -24.6 ± 0.6 respectively). This may well be due to the fact that the Post et al. (2007) equation used to lipid-correct Burn’s data is known to under- correct δ13C values, which leaves them more depleted than they should be if lipids had been removed from the sample (Mintenbeck et al., 2008; Figure 3 therein). Indeed, Mintenbeck et al. (2008) quote that lipid-normalisation with the Post et al. (2007) model can produce deviations of up to 2.2‰. Based on this comparison, whilst the δ15N Burns et al. (1998) values seem robust for Pleuragramma antarcticum, (see comparable δ15N data in Table 12), we do not have total confidence in the complementary δ13C values. We concluded that these lipid-corrected Burns et al. (1998) Pleuragramma antarcticum data were not robust enough to include in the ISOSOURCE model. We are therefore still unable to model the isotope data to apportion % prey items to Dissostichus mawsoni since the predator plots outside our current prey polygon.

We are currently awaiting Pleuragramma antarcticum stable isotope analyses from samples collected during the International Polar Year (IPY) NIWA Ross Sea Voyage (Feb-Mar 2008), and hopefully these data will enable us to use the ISOSOURCE model and apportion % prey items to Dissostichus mawsoni.

Table 12: Isotopic values of Pleuragramma antarcticum from published data. N.D. signifies “no data” available in the referenced paper. * lipid-corrected Burns et al. (1998) value using equation given in Post et al. (2007). Reference N = Region (life C:N d15N (1 s.d.) d13C (1 s.d.) stage) ratio Burns et al. (1998) 4 Mc.Murdo 6.4 10.9 (+/- 0.2) -25.3 (+/- 0.4)* Sound Hodum & Hobson (2000) 10 Prydz Bay N.D. 9.6 (+/- 0.4) -24.3 (+/- 0.3) (juvenile) Hodum & Hobson (2000) 13 Prydz Bay N.D. 10.7 (+/- 1.2) -23.9 (+/- 0.7) (adult) Cherel (2008) 5 Weddell Sea 3.8 10.6 (+/- 0.3) -24.7 (+/- 0.4) (adult)

3.4 Isotopic variation of fish samples between the 3 main sample areas

A comparison of the δ13C and δ15N values of Dissostichus mawsoni segregated out into the three main sample areas (Northern Area, Ross Sea Slope and Terra Nova Bay Trench) revealed very little variation between the areas in either d15N or δ13C values (Figure 17). This could indicate that either Dissostichus mawsoni had a similar diet in each of the three areas, or that they moved between, and fed within, all 3 areas i.e. they were not solely resident within any of the 3 areas of sampling. The fact that there was little variation in δ13C is a strong indication that the fish were not exclusively resident in each of the three areas, otherwise we would expect to see enriched 13C values in samples caught in the Northern Area compared to those caught further south. The colder and warmer water δ13C signatures may have become integrated through latitudinal movement of fish between the three areas.

28

18 18

16 16

14 14 N N 5 5 1

12 1

12 d d

10 10

8 Northern Area 8 Northern Area (n=37) Ross Sea Slope Ross Sea Slope (n=18) Terra Nova Bay Trench Terra Nova Bay Trench (n=45) 6 6 -28 -26 -24 -22 -20 -28 -26 -24 -22 -20 d13C d 13C

Figure 17: Comparison of the isotopic values of Dissostichus mawsoni in the three main sampling areas: Northern Area, Ross Sea Slope and Terra Nova Bay Trench. The figure on the left shows the isotopic values of all Dissostichus mawsoni sampled, whilst the plot on the right shows average values (± 1 s.d.) for each of the three designated areas.

In contrast, assigning isotopic values for Antimora rostrata, Chionobathyscus dewitti, Macrourus whitsoni and Muraenolepis microps to the Northern Area and the Ross Sea Slope (none of these species were sampled in the Terra Nova Bay Trench area) showed strong isotopic segregation, i.e. there were two distinct isotopic populations of each species (Figures 18 and 19). All four of the fish species sampled in the Northern Area had markedly enriched 13C signatures compared to the Ross Sea Slope. Figure 19 illustrates that there was a consistent 13C offset for each species between the Northern Area and the Ross Sea Slope. The offset was so regular it is likely to be chemically-driven, with the most plausible explanation being water temperature/latitudinal effects producing a more positive δ13C value to the north (refer to Table 5 and preceding discussion).

29 Antimora rostrata Chionobathyscus dewitti

18 18 Northern Area Northern Area Ross Sea Slope 16 Ross Sea Slope 16

14 14 N N 5 5 1 1

12 12 d d

10 10

8 8

6 6 -28 -26 -24 -22 -20 -28 -26 -24 -22 -20 d13C d13C

Macrourus whitsoni Muraenolepis microps

18 18 Northern Area Northern Area 16 Ross Sea Slope 16 Ross Sea Slope

14 14 N N 5 5 1 1

12

12 d d

10 10

8 8

6 6 -28 -26 -24 -22 -20 -28 -26 -24 -22 -20 13 d C d13C Figure 18: Comparison of isotopic data for four fish species (Antimora rostrata, Chionobathyscus dewitti, Macrourus whitsoni and Muraenolepis microps) caught in the Ross Sea Slope and Northern Areas.

18

16

A. rostrata - North 14 A. rostrata - Slope

N C. dewitti - North 5 1

12

d C. dewitti - Slope

M. microps - North 10 M. microps - Slope

M. whitsoni - North 8 M. whitsoni - Slope

6 -28 -26 -24 -22 -20 d13C

30

Figure 19: d13C and d15N averages (± 1 s.d.) of Antimora rostrata, Chionobathyscus dewitti, Macrourus whitsoni and Muraenolepis microps caught in the Ross Sea Slope and Northern Areas. Note that “Macrourus whitsoni – North” plots over “Muraenolepis microps – North”, and “Macrourus whitsoni – Slope” plots over “Muraenolepis microps – Slope”.

The more positive δ13C values of species sampled in the Northern Area is also evident from a total species plot comparing stable isotope data between this area and the Ross Sea Slope (Figure 20). The upper plots in Figures 21 and 22 represent averages for each species plotted in Figure 20. Fractionation factors have been applied to the average values of potential prey items of Dissostichus mawsoni to produce predator-prey polygon plots for each area (Figures 21 & 22: lower plots).

Ross Sea Slope Northern Area 18 Dissostichus 18 mawsoni (n=18) 16 Antimora rostrata (n=41) 16 Chionobathyscus 14 dewitti (n=49) 14 Muraenolepis N N

5 microps (n=59) 5 1 12 1

12 d Macrourus whitsoni d (n=74) 10 10 Octopods, 3 sp. (n=15) 8 Squids, 3 sp. (n=34) 8

6 Mesonychoteuthis 6 hamiltoni (n=4) -28 -26 -24 -22 -20 -28 -26 -24 -22 -20 13 d C d 13C

Figure 20: Comparison of d13C and d15N for all fish, octopods and squids caught over the Ross Sea Slope and in the Northern Area.

31 18

16

14 N 5 1

12 d 10

8

6 -28 -26 -24 -22 -20 d13C

Dissostichus mawsoni (n=37) Antimora rostrata (n=62) Chionobathyscus dewitti (n=33) Muraenolepis microps (n=15) Macrourus whitsoni (n=33) Octopods, 3 sp. (Ross Sea Slope) (n=15) Squids, 2 sp. (n=4) Mesonychoteuthis hamiltoni (n=3) Pleuragramma antarcticum, lipid corrd. (n=4) Burns et al., 1998 Dissostichus mawsoni (n=5), lipid corrd. Burns et al., 1998

18

16

14 N 5

1 12

d 10

8

6 -28 -26 -24 -22 -20 d13C

Figure 21: d13C and d15N isotope plots for all species sampled in the Northern Area. The upper graph plots average stable isotope values (± 1 s.d.) for all species sampled (excluding Dissostichus eleginoides), and the lower graph presents a Dissostichus mawsoni prey polygon, assuming a 3.4 ‰ fractionation factor for nitrogen and a 0.4 ‰ fractionation factor for carbon in potential prey items.

32 18

16

14 N 5 1

12 d 10

8

6 -28 -26 -24 -22 -20 d13C

Dissostichus mawsoni (n=18) Antimora rostrata (n=41) Chionobathyscus dewitti (n=49) Muraenolepis microps (n=59) Macrourus whitsoni (n=74) Octopods, 3 sp. (n=15) Squids, 3 sp. (n=34) Mesonychoteuthis hamiltoni (n=4) Pleuragramma antarcticum, lipid corrd. (n=4) Burns et al., 1998 Dissostichus mawsoni (n=5), lipid corrd. Burns et al., 1998

18

16

14 N 5

1 12

d 10

8

6 -28 -26 -24 -22 -20 d13C

Figure 22: d13C and d15N isotope plots for all species sampled on the Ross Sea Slope. The upper graph plots average stable isotope values (± 1 s.d.) for all species sampled (excluding Dissostichus eleginoides), and the lower graph presents a Dissostichus mawsoni prey polygon, assuming a 3.4 ‰ fractionation factor for nitrogen and a 0.4 ‰ fractionation factor for carbon in potential prey items.

33 3.5 Implications of prey polygon data for the Northern Area and Ross Sea Slope

If we exclude the questionable Pleuragramma antarcticum data from the literature, the predator- prey polygon in the lower plot of Figure 21 shows Dissostichus mawsoni clearly plotting outside of the potential prey polygon. It is not possible through the isotopic prey polygon plot to constrain the diet of Dissostichus mawsoni in the Northern area based on the demersal fish and cephalopods measurements obtained in this study. That is, we cannot explain the diet of Dissostichus mawsoni from the analysed prey caught in the Northern area. In contrast, it appears that it may be possible to model the Dissostichus mawsoni diet (and derive % prey contributions) from demersal fish and cephalopod samples caught by the fishery on the Ross Sea Slope (Figure 22). For both areas, Pleuragramma antarcticum seem likely to be a key prey item and it is therefore important to obtain some reliable stable isotope data for this species.

Considering the isotopic data alone, there are two feasible explanations for these results: · Dissostichus mawsoni are acquiring their prey from all three areas, migrating between the Northern Area, the Ross Sea Slope and the Terra Nova Bay Trench, over the stable isotope muscle-integrated time-frame of 6-9 months. · Dissostichus mawsoni are predominantly feeding on the Ross Sea Slope with only minor feeding in other areas, such that their diet, averaged over a period of months (i.e. their muscle isotopic signature), appears to be similar between all three areas sampled. Some Dissostichus mawsoni may move to the north for spawning aggregations, but may not feed there either much, or for long. Some Dissostichus mawsoni may move to the south (to the Shelf area) to take advantage of seasonally high aggregations of Pleuragramma antarcticum, squids and other potential prey, but they feed in this area for only a short while before returning to the slope area to obtain the majority of their diet.

These possible explanations could produce similar Dissostichus mawsoni isotopic signatures between the three sampled areas. In order to investigate the validity of these theories, we would need to analyse a range of tissues with different turnover rates (e.g. blood, liver, muscle, cartilage) to establish whether Dissostichus mawsoni switch their diet (see MacNeil et al., 2005) and if so, to what extent, and for how long. It would also be desirable to try to establish the distances and timescales that Dissostichus mawsoni migrate over, to establish if it is likely that they could move between the suggested areas over the timescales indicated.

The latter two hypotheses are difficult to reconcile in the context of current understanding of the movement and feeding of Dissostichus mawsoni in the Ross Sea. Tag-recapture data from Dissostichus mawsoni released in 2006 and captured in 2007 (Dunn et al., 2007) revealed that recaptured fish were caught close to their release locations a year earlier. Tag-recapture data however, does not provide any information on the movement of the fish between tagging deployment and recapture. It is possible that fish caught in the same location may have moved some distance between tagging and recapture, returning to a similar location on a yearly basis. Indeed data does exist to show that Dissostichus mawsoni do travel large distances. Sullivan et al. (2005) report that one adult tagged fish at McMurdo Sound was recaptured 500nm to the north off Cape Adare (SSRU 88.1H), whilst a second was recaptured after 18 years 1300 nm to the northeast in SSUR 88.2E. There are still large uncertainties in the impact of movements and spatial structure of Dissostichus mawsoni populations (Dunn & Hanchet, 2007). Further information on migration distances and timescales would greatly assist these stable isotope interpretations.

34

4. SUMMARY

· This study integrates 543 carbon and nitrogen stable isotope analyses which are broken down into 476 fish, 50 squid, and 17 octopod analyses. · All species showed large variations in δ15N and δ13C. Dissostichus mawsoni δ15N values spanned two trophic steps and all other fish spanned more than one trophic step. This implied that the diet for each species either varied in composition, or size, or both. · Dissostichus mawsoni and Dissostichus eleginoides occupied the same top trophic level, as orca (Orcinus orca) and Weddell seals (Leptonychotes weddellii). Antimora rostrata, Muraenolepis microps, and Macrourus whitsoni all occupied a trophic level below them. Chionobathyscus dewitti occupied the lowest trophic level of all fish analysed. Squids (excluding Mesonychoteuthis hamiltoni) were found to be the lowest trophic level of all species sampled. · Positive relationships between length and δ15N indicated that larger fish were consuming prey of a higher trophic level than smaller fish. · Higher δ13C values were found in fish prey samples caught in the Northern Area compared to the Ross Sea Slope. This was not reflected in Dissostichus mawsoni δ13C and δ15N values, which were very similar between areas. This suggests that Dissostichus mawsoni either moved between and fed equally between sample areas, or they predominantly fed on the Ross Sea Slope with short excursions to other areas. · The diet of Dissostichus mawsoni could not be constrained using an isotopic prey polygon plot. This means that we were most likely still missing some important prey items in their diet. Stomach content analyses indicate that Pleuragramma antarcticum are the most likely missing component.

5. CONCLUDING COMMENTS

The large sample number for each species makes this study unique in terms of previously reported stable isotope studies in oceanographic systems. In most published studies it is far more usual for the number of specimens of each species to number far less than those reported here. Our relatively large sample sizes have enabled us to capture the true variability of isotope values for each species, where the number of specimens analysed for each fish species exceeded 70 (with the exception of D. eleginoides, where n=8). In addition, more than 60 specimens of cephalopods were analysed.

For future fish sampling and isotope work on the Ross Sea ecosystem and diet of Dissostichus mawsoni we strongly recommend: (1) simultaneous stomach content analysis and tissue sampling of predator and prey items and, (2) multiple tissue sampling (e.g. muscle, liver and cartilage) of Dissostichus mawsoni, and all its potential prey items, as muscle tissue sampling alone does not reveal seasonal differences or possible diet switching of individual organisms, (3) simultaneous muscle sampling of pre-tagged and post-tagged fish to investigate possible changes in δ15N values during movement of fish. It would be highly desirable to try to establish the distances and timescales that Dissostichus mawsoni migrate over, to establish if it is likely that they could move between the suggested areas over the timescales indicated.

By improving our knowledge of the relationships between Dissostichus mawsoni and their prey over different geographic regions and investigating seasonal variations via seasonal and multiple tissue sampling, we are more likely to be able to improve our predictions of the potential impacts of the toothfish fishery on the Ross Sea ecosystem.

35

6. ACKNOWLEDGMENTS

This work was funded by the Ministry of Fisheries project code ANT200504 Objectives 1 & 2 and ANT200603 Objective 1, and by and the New Zealand government Foundation for Research Science and Technology project C01X0505. We are grateful to the New Zealand Ministry of Fisheries and CCAMLR observers Sarel Du Plessis, Marli Dee, Donavan Cole, David Bilton, and Rodney Hansen for collection of tissue samples in the Ross Sea region, and temporary worker Chris Gibbons for his help in collating the samples. We acknowledge the work of Peter Marriott, Peter McMillan, Matt Evans, Greg Olsen, and Scott Nodder (all NIWA) on this project.

7. REFERENCES

Akaike, H. (1974) A new look at statistical model identification. IEEE Transactions on Automatic Control AU-19: 716–722. Andrews, J.J. and Abel, K.M. (1979) Photosynthetic carbon metabolism in seagrasses. Plant Physiology 63: 650-656. Bigidaire, R.R., Fluegge, A., Freeman, K.H., Hanson, K.L., Hayes, J.M., Hollander, D., Jasper, J.P., King, L.L., Laws, E.A., Milder, J., Millero, F.J., Pancost, R., Popp, B.N., Steinberg, P.A., and Wakeham, S.G. (1977) Consistent fractionation of 13C in nature and in the laboratory: growth rate effects in some haptophyte algae. Global Biogeochemical Cycles 11: 279-292. Burkhardt, S., Rievesell, U., and Zondervan, I. (1999) Stable carbon isotope fractionation by marine phytoplankton in response to daylength, growth rate, and CO2 availability. Marine Ecology Progress Series 184, 31-41. Burns, J.M., Trumble, S.J. and Castellini, M.A. (1998) The diet of Weddell seals in McMurdo Sound, Antarctic as determined from scat collections and stable isotope analysis. Polar Biology 19: 272-282. Bury, S., Pinkerton, M., Hanchet, S., and Thompson, D. (2007) Trophic study of Ross Sea Antarctic toothfish (Dissostichus mawsoni) using carbon and nitrogen stable isotopes. Oral presentation at Annual New Zealand Marine Science Conference, 29-31 August 2007, Hamilton, New Zealand. Bury, S., Pinkerton, M.H., Brown. J., Vorster, I., (2008) Progress Report. Stable isotope analysis of Southern Ocean 2006/7 field season fish tissue samples: preliminary results including repeat analysis of high lipid-content 2005/6 Toothfish samples. Advanced Progress Report for Ministry of Fisheries Research Project ANT2006-03: Objective 1, March, 2008, 17pp. Bury, S.J. (1999) The DeltaPlus stable isotope mass spectrometer: advancing research in environmental processes. Water and Atmosphere 7(2): 22–25. Calhaem, I., Christoffel, D.A. (1969) Some observations of the feeding habit of a Weddell seal, and measurements of its prey, Dissostichus mawsoni, at McMurdo Sound, Antarctica. New Zealand Journal of Marine and Freshwater Research 3: 181–190. Cherel, Y. and Hobson, K.A. (2007) Geographical variation in carbon stable isotope signatures of marine predators: a tool to investigate their foraging areas in the Southern Ocean. Marine Ecology Progress Series 329: 281-287. Cherel, Y., Phillips, R.A., Hobson, K.A. and McGill, R. (2006) Stable isotope evidence of diverse species-specific and individual wintering strategies in seabirds. Biology Letters 2: 301-303. Cortés, E. (1997) A critical review of methods of studying fish feeding based on analysis of stomach contents: application to elasmobranch fishes. Canadian Journal of Fisheries and Aquatic Science 54: 726–738.

36 DeNiro, M.J. and Epstein, S. (1977) Mechanism of carbon isotope fractionation associated with lipid synthesis. Science 197: 261-263. DeNiro, M.J. and Epstein, S. (1978) Influence of diet on the distribution of carbon isotopes in animals. Geochimica et Cosmochimica Acta Vol 42: 495-506. DeNiro, M.J. and Epstein, S. (1981) Influence of diet on the distribution of nitrogen isotopes in animals. Geochimica et Cosmochimica Acta Vol 45: 341-351. Deuser, W.G., Degens, E.T. and Guillard, R.R.L. (1968) Carbon isotope relationships between plankton and seawater. Geochimica et Cosmochimica Acta 32: 657-660. Dunn, A. and Hanchet, S.M. (2007) Revised input parameters and implications for the Antarctic toothfish (Dissostichus mawsoni) stock assessment in subareas 88.1 and 88.2. Paper presented to CCAMLR Fish Stock Assessment working group, 22 June 2007. WG-SAM- 07/6, 32pp. Dunn, A., Hanchet, S.M. and Ballara, S.L. (2007) An updated descriptive analysis of the toothfish (Dissostichus spp.) tagging programme in Subareas 88.1 & 88.2 for 2006_7. Paper presented to CCAMLR Fish Stock Assessment working group, 21 September, 2007. WG- FSA-07/40, 25pp. Dunn, A., Horn, P.L. and Hanchet, S.M. (2006) Revised estimates of the biological parameters for Antarctic toothfish (Dissostichus mawsoni) in the Ross Sea. WG-FSA-SAM-06/08, 14 pp. Eastman, J.T. (1985) The evolution of neutrally buoyant notothenioid fishes: their specializations and potential interactions in the Antarctic marine food web. In: Siegfried, W.R., Condy, P.R., Laws, R.M. (eds) Proceeding of the 4th SCAR Symposium on Antarctic Biology. Springer, Berlin, pp 430-436. Eastmann, J.T. (1993) Antarctic Fish Biology: evolution in a unique environment. London, UK, Academic Press. Everson, I. (2002) Fish species profiles – toothfish. SC-CAMLR-XXI/BG/30. Falkowski, P.G. (1991) Species variability in the fractionation of 13C and 12C by marine phytoplankton. Journal of Plankton Research 13: 21-28. Fenaughty, J.M. (2006) Geographic differences in the condition, reproductive development, sex ratio and length distribution of Antarctic toothfish (Dissostichus mawsoni) from the Ross Sea, Antarctica (CCAMLR subarea 88.1). CCAMLR Science 13: 27-45. Fenaughty, J.M., Stevens, D.W. and Hanchet, S.M. (2003) Diet of the Antarctic toothfish (Dissostichus mawsoni) from the Ross Sea, Antarctica (Subarea 88.1). CCAMLR Science 10: 113–123. France R.L. (1995) Carbon-13 enrichment in benthic compared to planktonic algae: foodweb implications. Marine Ecology Progress Series 124: 307-312. François. R., Altabet, M.A. and Goericke, R. (1993) Changes in the δ13C of surface water particulate organic matter across the subtropical convergence in the SW Indian ocean. Global Biogeochemical Cycles 7: 627-644. Fry, B., and Sherr, E.B. (1984) δ13C measurements as indicators of carbon flow in marine and freshwater systems. Contributions in Marine Science 27: 13–47. Fuiman, L.A., Davis, R.W. and Williams, T.M. (2002) Behaviour of midwater fishes under the Antarctic ice: observationis by a predator. Marine Biology Online First: 1-14. Goericke, R., and Fry, B. (1994) Variations of marine plankton δ13C with latitude, temperature, and dissolved CO2 in the world ocean. Global Biogeochemical Cycles 8: 85-90. Gon, O. and Heemstra, P.C. (1990) Fishes of the Southern Ocean. J.L.B. Smith Institute of Ichthyology, Grahamstown, South Africa: 462 pp. Hanchet, S.M. (2006) Species profile for Antarctic toothfish (Dissostichus mawsoni). Paper presented to CCAMLR Fish Stock Assessment working group, 20 September 2006. WG- FSA-06/26, 22pp.

37 Hanchet, S.M., Rickard, G.J., Fenaughty, J.M., Dunn, A. and Williams, M.J. (submitted) A hypothetical life cycle for Antarctic toothfish Dissostichus mawsoni in the Ross Sea region. Submitted to CCAMLR Science (scheduled publication September 08). Hanchet, S.M., Stevenson, M.L., and Dunn, A. (2006). A characterisation of the toothfish fishery in Subareas 88.1 & 88.2 from 1997/98 to 2005/06. Paper presented to CCAMLR Fish Stock Assessment working group, 21 September 2006, Hobart, Australia. WG-FSA-06/29, 25 pp. Hanchet, S.M., Stevenson, M.L., and Dunn, A. (2007). A characterisation of the toothfish fishery in Subareas 88.1 & 88.2 from 1997/98 to 2006/07. Paper presented to CCAMLR Fish Stock Assessment working group, 20 Sept 2007. WG-FSA-07/28, 19 pp. Hanchet, S.M., Stevenson, M.L., Horn, P.L., Blackwell, R.G. (2003) Assessment of toothfish (Dissostichus mawsoni and D. eleginoides) stocks in the Ross Sea (CCAMLR Subareas 88.1 and 88.2). New Zealand Fisheries Assessment Report 2003/43, 27 pp. Hebert, C.E. and Keenleyside, K.A. (1995) To normalise or not to normalise? Fat is the question. Environmental Toxicology and Chemistry 14: 801-807. Hinga, K.R., Arthur, M.A., Pilson, M.E.Q. and Whitaker, D. (1994) Carbon isotope fractionation by marine phytoplankton in culture: the effects of CO2 concentration, pH, temperature, and species. Global Biogeochemical Cycles 8: 91-102. Hobson, K.A. (1999) Tracing origins and migration of wildlife using stable isotopes: a review. Oecologia 120: 314-326. Hobson, K.A., Ambrose, Jr., W.G., and Renaud, P.E. (1995) Sources of primary production, benthic-pelagic coupling, and trophic relationships within the Northeast Water Polyna: insights from δ13C and δ15N analysis. Marine Ecology Progress Series 128: 1-10. Hobson, K.A., Piatt, J.F. and Pitocchelli, J. (1994) Using stable isotopes to determine seabird trophic relationships. Journal of Animal Ecology 63:786-798. Hyslop, E.J. (1980) Stomach contents analysis – a review of methods and their application. Journal of Fisheries Biology 17: 411-429. Jardine, T.D., and Cunjak, R.A. (2006) Lipid extraction in stable isotope ecology: a call for consensus. Abstracts of 5th International Conference on “Applications of stable isotope techniques to ecological studies”, Queens University Belfast, U.K., August 2006, p166. Jørgensen, B.B. and Revsbech, N.P. (1985) Diffusive boundary layers and the oxygen uptake of sediments and detritus. Limnology and Oceanography 30: 111-122. Keeley, J.E. and Sundquist, D.R. (1992) Carbon: freshwater plants. Plant Cell Environment 15: 1021-1035. Krahn, M.M., Pitman, R.L., Burrows, D.G., Herman, D.P. and Ronald, W.P. (2006) Assessing the feeding ecology of Antarctic Type C killer whales using chemical tracers. International Whaling Commission Scientific Committee Paper SC/58/13, 16pp. Kukert, H. and Riebesell, U. (1998) Phytoplankton carbon isotope fractionation during a diatom spring bloom in a Norwegian fjord. Marine Ecology Progress Series 173: 127-137. La Mesa, M., Eastman, J.T. and Vacchi, M. (2004) The role of notothenioid fish in the food web of the Ross Sea Shelf waters: a review. Polar Biology 27: 321-338 Laws, E.A., Popp, B.N., and Bidigaire, R.R. (1997) Effect of growth rate and CO2 concentration on carbon isotopic fractionation by the marine diatom Phaeodactylum tricornutum. Limnology and Oceanography 42: 1552-1560. Laws, E.A., Popp, B.N., Bidigaire, R.R., Kennicutt, M.C. II and Macko, S.A. (1995) Dependence of phytoplankton carbon isotopic composition on growth rate and [CO2]aq: theoretical considerations and experimental results. Geochimica et Cosmochimica Acta 59: 1131-1138. Logan, J.M, and Lutcavage, M.E. (2006) Diet and movement of the Atlantic bluefin tuna (Thunnus thynnus) through carbon and nitrogen stable isotope analysis. Abstracts of 5th International Conference on “Applications of stable isotope techniques to ecological studies”, Queens University Belfast, U.K. August 2006, p153.

38 Logan, J.M., Jardine, T.D., Miller, T.J., Bunn, S.E., Cunjak, R.A. and Lutcavage, M.E. (2008) Lipid corrections in carbon and nitrogen stable isotope analyses: comparisons of chemical extraction and modelling methods. Journal of Animal Ecology 77(4): 838-846. MacNeil, M.A., Skomal, G.B. and Fisk A.T. (2005) Stable isotopes from multiple tissues reveal diet switching in sharks. Marine Ecology Progress Series 302: 199-206. Marguillier, S., van der Velde, G., Dehairs, F., Hemminga, M.A. and Rajagopal, S. (1997) Trophic relationships in an interlinked mangrove-seagrass ecosystem as traced by δ13C and δ15N. Marine Ecology Progress Series 151: 115–121. McConnaughty, T. and McRoy, C.P. (1979) Foodweb structure and the fractionation of carbon isotopes in the Bering Sea. Marine Biology 53: 257-262. Melville, A.J. and Connolly, R.M. (2003) Spatial analysis of stable isotope data to determine primary sources of nutrition for fish. Oecologia 136: 499-507. Michener, R.H. and Schell, D.M. (1994) Stable isotope ratios as tracers in marine aquatic foodwebs. In: Lajtha, K. and Michener, R.H. (eds) Stable isotopes in ecology and environmental science. Blackwell, London, pp 138-158. Minagawa, M. and Wada, E. (1984) Stepwise enrichment of 15N along food chains: Further evidence and the relation between δ15N and animal age. Geochimica et Cosmochimica Acta 48: 1135–1140. Ministry of Fisheries (2006) The Ross Sea toothfish fishery: a recent characterisation and the fishery status. Ministry of Fisheries publication, Wellington, New Zealand, 48pp. Mintenbeck, K., Brey, T., Jacob, U., Knust, R. and Struck, U. (2008) How to account for the lipid effect on carbon stable-isotope ratio (δ13C): sample treatment effects and model bias. Journal of Fish Biology 72, 815-830. Nagelkerken, I., and van der Velde, G. (2004a) Are Caribbean mangroves important feeding grounds for juvenile reef fish from adjacent seagrass beds? Marine Ecology Progress Series 274: 143–151. Nagelkerken, I., and van der Velde, G. (2004b) Relative importance of interlinked mangroves and seagrass beds as feeding habitats for juvenile reef fish on a Caribbean island. Marine Ecology Progress Series 274: 153–159. Nakatsuka, T., Handa, N., Wada, E., and Wong, C.S. (1992) The dynamic changes of stable isotopic ratios of carbon and nitrogen in suspended and sedimented particulate organic matter during a phytoplankton bloom. Journal Marine Research 50: 267-296. Pakhomov E.A. and Tseytlin, V.B. (1992) Diet of seven species of Antarctic fishes and estimation of their daily rations. Journal of Ichthyology 32: 31-41. Parker, P.L. (1964) A rapid method of total lipid extraction and purification. Plant Physiology 36: 133-138. Patchell, G.J. (2002) Information on the spawning of Dissostichus mawsoni from Subareas 88.1 and 88.2 in the 2001/02 season. Sealord Group Ltd. unpublished manuscript presented to the Fish Stock Assessment Working Group of CCAMLR, WG-FSA-02/31, 17pp. Peterson, B.J. and Fry, B. (1987) Stable isotopes in ecosystem studies. Annual Review of Ecological Systems 18: 293–320. Phillips, D.L., and Gregg, J.W. (2001) Uncertainty in source partitioning using stable isotopes. Oecologia 127: 171–179. Phillips, D.L., and Gregg, J.W. (2003) Source partitioning using stable isotopes: coping with too many sources. Oecologia 136: 261–269. Phillips, N.L., Dunn, A. and Hanchet, S.M. (2005) Stratification of catch-at-length data using tree based regression. An example using Antarctic toothfish (Dissostichus mawsoni) in the Ross Sea. WG-FSA-SAM-05/8, 16 pp.

39 Pinkerton, M.H., Bradford-Grieve, J., Hanchet, S.M., Williams, M. and Wilson P. (2006) Modelling the effects of fishing in the Ross Sea. Final Research Report to Ministry of Fisheries, New Zealand, for project ANT2004-05. NIWA, Wellington, New Zealand, 169pp. Pinkerton, M.H., Bury, S.J., Hanchet, S.M. and Thompson D. (2007b) Stable isotope analysis of Southern Ocean fish tissue samples to investigate trophic linkages of Antarctic toothfish (Dissostichus mawsoni). Paper presented to the Ecosystem Monitoring and Modelling Working Group, CCAMLR, 16–27 July 2007, Christchurch. EMM-07-19, 28pp. Pinkerton, M.H., Bury, S.J., Vorster, I., and Hanchet, S.M. (2007a) Stable isotope analysis of Southern Ocean fish tissue samples: preliminary results. Final research report for Ministry of Fisheries Research Project ANT2005-04 Objective 1, March 2007, 32pp. Pinkerton, M.H., Dunn, A. and Hanchet, S.M. (2007c) Ecological risk management and the fishery for Antarctic toothfish (Dissostichus mawsoni) in the Ross Sea, Antarctica. Paper presented to the Ecosystem Monitoring and Modelling Working Group, CCAMLR, 16–27 July 2007, Christchurch. EMM-07-24, 22pp. Pinkerton, M.H., Hanchet, S.M. and Bradford-Grieve, J.M. (2007d) A balanced trophic model of the ecosystem of the Ross Sea, Antarctica for investigating effects of the Antarctic toothfish fishery. Paper presented to the Ecosystem Monitoring and Modelling Working Group, CCAMLR, 16–27 July 2007, Christchurch. EMM-07-18, 24pp. Popp, B.N., Laws, E.A., Bigidaire, R.R., Dore, J.E., Hanson, K.L. and Wakeham, S.G. (1998) Effect of phytoplankton cell geometry on carbon isotope fractionation. Geochimica et Cosmochimica Acta 62: 69-77. Post, D.M., Layman, C.A., Arrington, D.A, Takimoto, G., Quattrochi, J. and Montana, C.G. (2007) Getting to the fat of the matter: models, methods and assumptions for dealing with lipids in stable isotope analysis. Oecologia 152: 179-189. Rau, G.H., Riebesell, U., and Wolf-Gladrow, D. (1996) A model of photosynthetic 13C fractionation by marine phytoplankton based on diffusive molecular CO2 uptake. Marine Ecology Progress Series 133: 275-285. 13 Rau, G.H., Riebesell, U., and Wolf-Gladrow, D. (1997) CO2aq- dependent photosynthetic C fractionation in the ocean: a model versus measurements. Global Biogeochemical Cycles 11: 267-278. Rau, G.H., Sweeney, R.E. and Kaplan, I.R. (1982) Plankton 13C/12C ratio changes with latitude: differences between northern and southern oceans. Deep-Sea Research 29: 1035-1039. Rau, G.H., Takahashi, T. and Des Marais, D.J. (1989) Latitudinal variations in plankton δ13C: implications for CO2 and productivity in past oceans. Nature 341: 516-518. Raven, J.A., Johnston, A.M., and Turpin, D.H. (1993) Influence of changes in CO2 concentration and temperature on marine phytoplankton 13C/12C ratios: an analysis of possible mechanisms. Global and Planetary Change 8: 1-12. Riber, H.H. and Wetzel, R.G. (1987) Boundary layer and internal diffusion effects on phosphorus fluxes in lake periphyton. Limnology and Oceanography 32: 1181-1194. Ricca, M.A., Miles, A.K. Anthony, R.G., Deng, X., and Hung, S.S.O. (2007) Effect of lipid extraction on analyses of stable carbon and stable nitrogen isotopes in coastal organisms of the Aleutian archipelago. Canadian Journal of Zoology 85: 40-48. Romanuk, T.N. and Levings, C.D. (2005) Stable isotope analysis of trophic position and terrestrial vs. marine carbon sources for juvenile Pacific salmonids in nearshore marine habitats. Fisheries Management and Ecology 12:113-121. Rounick, J.S. and Winterbourn, M.J. (1986) Stable carbon isotopes and carbon flow in ecosystems. BioScience 36(3): 171-177. Sackett, W.M., Eckelmann, W.R., Bender, M.L. and Be, A.W.H. (1965) Temperature dependence of carbon isotope composition in marine plankton and sediments. Science 148: 235-237.

40 Schell, D. (2002) Temporal isotopic records in marine mammals – old stories and new. Plenary presentation at 3rd international conference on “Applications of stable isotopes to ecological studies”, Flagstaff, Arizona, April-May 2002. Smith, F.R. and Walker, N.A. (1980) Photosynthesis by aquatic plants: effects of unstirred layers in - relation to assimilation of CO2 and HCO3 and to carbon isotopic discrimination. New Phytology 86: 245-259. Stevens, D.W. (2004) Stomach contents of the Antarctic toothfish (Dissostichus mawsoni) from the western Ross Sea, Antarctica. CCAMLR WG-FSA-04/31. 15pp. Stevens, D.W. (2006) Stomach contents of sub-adult Antarctic toothfish (Dissostichus mawsoni) from the western Ross Sea, Antarctica. WG-FSA-06/27. 15 pp. Sullivan, K.J., Smith, NW.McL. and Waugh, S. (Comps.) (2005) Report from the Mid-Year Fishery Assessment Plenary, November 2005: stock assessments and yield estimates. Unpublished report held at NIWA, Greta Point library, Wellington, New Zealand, 62 pp. Thompson, D.R., Stevens, D.W., Pinkerton, M.H. and Bury, S.J. (2007) Stable isotope analysis of Ross Sea cephalopod tissue samples: preliminary results. Final research report for Ministry of Fisheries research project ANT200504-Objective 2, December 2007, 18pp. Tieszen, L.L., Boutton, T.W., Tesdahl, K.G. and Slade, N.A. (1983) Fractionation and turnover of stable carbon isotopes in animal tissues: Implications for δ13C analysis of diet. Oecologia 57: 32-37. Van der Merwe, N.J. (1982) Carbon isotopes, photosynthesis and archaeology. American Scientist 70: 596-606. Van der Merwe, N.J. and Vogel, J.C (1978) 13C content of human collagen as a measure of prehistoric diet in woodland North America. Nature 276: 815-816. Vander Zanden, M.J. and Rasmussen, J.B. (2001) Variation in δ15N and δ13C trophic fractionation: implications for aquatic food web studies. Limnology and Oceanography 46: 2061–2066. Vizzini S., Sara G., Michener, R.H. and Mazzola, A. (2002) The role and contribution of the seagrass Posidonia oceania (L.) Delile organic matter for secondary consumers as revealed by carbon and nitrogen isotope analysis. Acta Oecoligica 23: 277–285. Wong, W.W. and Sackett, W.M. (1978) Fractionation of stable carbon isotopes by marine phytoplankton. Geochimica et Cosmochimica Acta 42: 1809-1815. Yukhov, V.L. (1971) The range of Dissostichus mawsoni Norman and some features of its biology. Journal of Ichthyology 11:8-18. Zhao, L., Castellini, M.A., Mau, T.L. and Trumble, S.J. (2004) Trophic interactions of Antarctic seals as determined by stable isotope signatures. Polar Biology 27: 368-373.

41