Journal of Marine Systems 17 Ž1998. 275–288 Estimates of Southern Ocean primary production—constraints from predator carbon demand and nutrient drawdown J. Priddle ) , I.L. Boyd, M.J. Whitehouse, E.J. Murphy, J.P. Croxall British Antarctic SurÕey, Natural EnÕironment Research Council, High Cross, Madingley Road, Cambridge CB3 0ET, UK Received 15 October 1995; accepted 15 July 1996 Abstract In view of the wide range of estimates for the total primary production for the Southern Ocean south of the Subantarctic Front—current estimates range from 1.2 to 3.5 Gtonne C yeary1 —we have examined two indirect methods for assessing primary production. First, we have estimated the primary production needed to sustain the carbon requirements of the endotherm top predators in the ecosystem. Estimation of the carbon requirements for crabeater seals of about 7 Mtonne C yeary1 is extrapolated to a value for all endotherm predators of 15–30 Mtonne C yeary1. Current data indicate that 70–80% of the diet of this suite of predators is zooplankton Žpredominantly the euphausiid krill., making for highly efficient transfer from primary production to top predators. Our best estimate of Southern Ocean primary production by this method is of the order of 1.7 Gtonne C yeary1, or an averaged areal primary production of about 30–40 g C my2 yeary1. Our second approach is to estimate primary production from the drawdown of inorganic nutrients, based on the limited suite of studies from which an annual nutrient deficit can be calculated. Again, this indicates annual primary production of the order of 1.5 Gtonne. Although both methods have inherent uncertainties, taken together they provide a relatively robust constraint on annual primary production. For both methods to underestimate primary production by the 1–1.5 Gtonne C implied by the higher current estimates, carbon export from the Southern Ocean pelagic ecosystem would need to be much higher than is normally found in other oceans. Resume ´ ´ Au regard a` la vaste gamme des estimations de la production primaire totale pour l’Ocean ´ ´ Austral au sud du Front Subantarctique—les estimations actuelles varient de 1.2 a` 3.5 Gt C yy1 —nous avons examine´ deux methodes indirectes ´ pour evaluer la production primaire. Nous avons en premier lieu estime´ la production primaire necessaire pour subvenir aux ´ ´ besoins en carbone des predateurs superieurs endothermes dans l’ecosysteme. Une estimation d’environ 7 Mt C yy1 pour les ´ ´ ´ ´ besoins de la population de phoques crabiers est extrapolee ´ a` une valuer de 15–30 Mt C yy1 pour l’ensemble des predateurs ´ endothermes. Les donnes sont constitues ´ actuelles indiquent que 70–80% de la ressource alimentaire de ces predeteurs ´ ´ de zooplancton Žprincipalement du krill., ce qui permet un transfert efficace de la production primaire vers les predateurs ´ superieurs. Notre meilleure estimation de la production primaire de l’Ocean est de’ordre de 1.7 Gt ´ ´ Austral par cette methode ´ C yy1, soit une moyenne pour la zone d’environ 30–40 g C my2 yy1. Notre second approche est d’estimer la production à partir de la consommation d’elements nutritifs mineraux, en nous basant sur l’ensemble restreint d’etudes pour lesquels un ´ ´ ´ deficit annuel de sels nutritifs peut etre donne egalement une production primaire annuelle de l’ordre ´ ˆ calcule. ´ Cette methode ´ ´ ) Corresponding author. Tel.: q44-1223-221-597; Fax: q44-1223-221-254; E-mail: [email protected] 0924-7963r98r$ - see front matter q 1998 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 4 - 7 9 6 3 Ž 9 8 . 0 0 0 4 3 - 8 276 J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 de 1.5 Gt. Bien que les deux methodes presentent des incertirtudes propres, elle encadrent, prises ensemble, la production ´ ´ primaire annuelle de maniere exactes, les deux methodes ´ assez robuste. Si les plus fortes estimations actuelles etaient ´ ´ devraient sous-estimer de 1 a` 1.5 Gt C la production primaire annuelle, ce qui impliquerrait une exportation de carbone par l’ecosysteme tres a` ce qu’elle est normalement dans les autre oceans. q 1998 Elsevier Science B.V. All ´ ´ pelagique ´ ´ superieure ´ ´ rights reserved. Keywords: Predator; Carbon requirement; Nutrient 1. Introduction Current estimates of the primary production of the Southern Ocean south of the Subantarctic Front Žreferred to here as the Antarctic Southern Ocean, ASO., range from 1.2 to 3.5 Gtonne C yeary1 ŽHuntley et al., 1991; Smith, 1991.. A recent estimate based on ocean colour measurements puts total ocean primary production of 45–50 Gtonne C yeary1 ŽLonghurst et al., 1995.. Their biogeochemical provinces do not match the delineation of the ASO used here, but their most southerly zones APLR and ANTR had an estimated combined primary production of 2.24 Gtonne C yeary1 , but this excludes the Polar Frontal Zone which forms the southern part of their larger SANT province Žtotal production 3.63 Gtonne C yeary1 .. It seems likely that Longhurst et al. Ž1995. would predict a primary production of 3.5–4 Gtonne C yeary1 for the ASO as defined here and by other studies. In most estimates, the ASO has a primary production lower than that implied by its area Žabout 35 = 10 6 km2 or around 10% of the world’s ocean area., which suggests that it should have an annual primary production of around 4 Gtonne C yeary1 . The ASO is also the largest area of the world’s ocean where nutrients are not utilized fully by phytoplankton production during the growth season—the so-called ‘High Nutrient–Low Chlorophyll’ ŽHNLC. phenomenon Žsee Chisholm and Morel, 1991.. Consequently there is a need to understand the dynamics of carbon cycling within the Southern Ocean in the context of the global carbon cycle, because such areas can be identified as the only locations where nutrient concentrations do not limit annual primary production and thus increases in biological carbon drawdown are at least possible ŽPriddle et al., 1992.. The uncertainty in our current understanding of the total primary production in the Southern Ocean is a drawback when considering global oceanic carbon fluxes. The database for ‘conventional’ measurements of primary production is patchy in space and time ŽPriddle et al., 1992; Longhurst et al., 1995., and such ‘point estimates’ may, in any case, provide poor integrations of annual productivity ŽWiggert et al., 1994.. The use of ocean colour remote sensing estimates provides a coverage which is closer to synoptic, but for the Southern Ocean poor coverage with the Coastal Zone Colour Scanner ŽCZCS—the only satellite dataset. is compounded by bio-optical problems which may compromise phytoplankton biomass retrieval algorithms ŽMitchell and HolmHansen, 1991; Fenton et al., 1994; Sagan et al., 1995.. The calculation of primary production data from chlorophyll biomass is again dependent on scarce measurements of photosynthetic parameters Žcf. Longhurst et al., 1995.. Here we examine Southern Ocean primary production from two contrasting standpoints. First, we have combined estimates of the carbon flux to top predators with a simple model of trophic efficiency to predict a likely range for the primary production needed to sustain the pelagic foodweb. Second, we compile the relatively restricted dataset on annual drawdown of inorganic nutrients and interpret these in terms of primary production. This again suggests a relatively narrow range for Southern Ocean annual primary production, and we compare the two approaches and their likely sources of error. We emphasise that we are putting neither approach forward as proxy measurements per se, but at the same time they provide a valuable constraint on regional estimates which are otherwise poorly controlled. 2. Carbon demand by top predators A special feature of interest in the context of oceanic carbon cycling is the unusual combination of features of the ASO food-web. Classically, this contains relatively short pathways between relatively J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 large organisms—diatoms, large euphausiid crustaceans, and mammalian and avian predators. Although such a picture understates the complexity of the ASO pelagic food web, it also highlights features important in the context of carbon cycling. The high proportion of endotherm top predators in the ASO has been implicated as a means whereby autotrophically-fixed carbon may be returned rapidly to the atmosphere ŽHuntley et al., 1991. although the magnitude of this ‘leak’ has been questioned ŽMoloney, 1992; van Franneker et al., 1994, van Franneker et al., 1997; Banse, 1995.. This paper is not concerned directly with that particular argument, but we return to the magnitude of CO 2 flux from air-breathing predators at the end. However, the apparent dominance by birds and mammals of the top predator level in the Southern Ocean pelagic system provides a means of describing the underlying primary production. Because some of these predators are accessible on land for part of the year, they have been the subject of physiological studies which are not carried out so easily with aquatic predators Žfish and pelagic cephalopods.. On the basis of these measurements, and using biomass and demographic data, we argue that overall carbon demand by Southern Ocean endotherm predators can be estimated with a moderate level of accuracy. We then combine this information with a range of scenarios of food-web efficiency to ‘back predict’ a most plausible range of primary production. 2.1. The estimation of carbon demand by all warmblooded top predators We started our calculation of carbon flux to endotherm predators by considering the single species for which we have the best data, and fortunately one which is of considerable importance in the Southern Ocean ecosystem. The crabeater seal Ž Lobodon carcinophagus. is very abundant ŽErickson and Hanson, 1990., is predominantly a krill predator ŽLaws, 1984., and has a body size which falls around the midpoint of the suite to avian and mammalian predators for which we eventually wish to estimate carbon demand. Our calculation of the flux of carbon to crabeater seals in the Southern Ocean is set out in detail in Appendix A. Net of losses during assimilation, demand was in the range of 3.9–9.3 Mtonne C yeary1 with a most probable value of 6.0 Mtonne C yeary1 ŽTable 1.. The assimilation efficiency of crabeater seals feeding on krill is 83.8 " 2.2% ŽNordøy et al., 1994.. This implies that the total flux of carbon to crabeater seals is 7.2 Mtonne yeary1 Žrange 2.8–11.1 Mtonne yeary1 .. The mass of carbon sequestered within body tissues was 0.27 Mtonne Žrange 0.01–0.03 Mtonne. and the ratio of carbon flux to sequestered carbon was 8.4:1 Žrange 7.3:1–9.0:1.. Thus growth efficiency, expressed in terms of carbon sequestered as a percentage of carbon assimilated, was 3.7%. This is within the range estimated for mammals ŽHumphries, 1979.. This analysis only considered crabeater seals. We do not have similar data to calculate carbon budgets for other major endotherm predator species, and therefore need to apply our estimates for crabeater seals to other predators. Based on the rough calculations presented in Table 2, crabeater seals are likely to constitute about one-third of the total metabolic biomass of endotherm predators in the Southern Ocean, the remainder being composed of whales Žmainly 0.75 million minke whales, Balaenoptera Table 1 Estimate of carbon flux to crabeater seals including maximum and minimum values Source Carbon flux ŽMt yry1 . Minimum carbon flux ŽMt yry1 . Maximum carbon flux ŽMt yry1 . Y Metabolism Y Growth Y Reproduction ? Metabolism ? Growth Total 2.973 0.013 0.136 2.879 0.014 6.028 1.126 0.005 0.057 1.091 0.004 2.283 4.570 0.016 0.225 4.429 0.017 9.257 See Appendix A for details of calculations. 277 278 Table 2 Calculation of the approximate metabolic biomass for the main species and groups of top predators in the Southern Ocean Mean body mass Žkg. Population size Žmillions. Metabolic biomassa Žmillion tonnes. Source Major proportion of diet b Crabeater seal 180 12 2.65 zooplankton Weddell seal Southern elephant seal Antarctic fur seal Minke whale Penguins Odontocete whales Total 200 250 35 5000 4.5 ; 3500 2 1 3 0.75 50 c 1 0.49 0.30 0.16 3.09 0.87 0.29 d 7.85 Erickson and Hanson Ž1990.; Laws Ž1984.; BAS unpublished data Erickson and Hanson Ž1990. SCAR Ž1991.; Boyd et al. Ž1994. Boyd Ž1993.; Payne Ž1979. International Whaling Commission; Folklow and Blix Ž1992. Croxall Ž1984. Kasamatsu and Joyce Ž1995. fish fish and squid zooplankton zooplankton zooplankton fish and squid The numbers given are approximate and are for illustrative purposes only. According to these figures, crabeater seals metabolic biomass accounts for 34% of the total endotherm predator metabolic biomass. Species not included in the calculations are assumed not to have a significant overall effect on the proportion of carbon flux due to crabeater seals Žespecially when rounded to be one-third of the total flux due to endotherm predators. due either to their small size andror their low numbers. a Metabolic biomass was calculated as the product of the approximate population size and the mean mass 0.88 in the case of the mammals and the mean mass 0.75 in the case of the penguins ŽNagy, 1987.. b The major proportion of the diet is indicated for each species or group, but this is not exclusive. In particular, some penguin species are fish or squid feeders, and part of the diet of Antarctic fur seals comprises fish. c The figure for the number of penguins is an estimate with a large potential error. However, errors in this value will have a relatively small impact on the estimate of total metabolic biomass. d The biomass for odontocete whales is calculated based on data presented by Kasamatsu and Joyce Ž1995. ŽTable IX. further adjusted for metabolic body mass and for the proportion of time these animals spend in the Southern Ocean Žestimated by Kasamatsu and Joyce to be 90 days.. J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 Species J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 acutorostrata., other seals Ž; 6 million individuals. and seabirds, the greatest biomass of which will be penguins. Crabeater seal body mass falls at around the midpoint of the range for this suite of predators, and as a crude ‘rule of thumb’ we suggest that allometric differences across the size range will balance out, allowing us to extrapolate across the range of predators. On this basis, total carbon flux to endotherm predators is likely to be in the region of 15 to 30 Mtonne C yeary1 . Based on efficiency of 3.7% Žcalculated for crabeater seals feeding on krill., this is equivalent to an overall top predator production of 0.55–1.11 Mtonne C yeary1 . In the following calculations, we will use the upper bound value only. 2.2. Food-web efficiency In order to use the carbon flux to top predators to estimate the primary production needed to sustain the entire pelagic ecosystem, we must estimate the efficiencies of the various intervening transfers. In order to do this, we have used a very simple food web model, coupled with plausible values of transfer efficiencies. 279 The model ŽFig. 1a. contains five compartments, and is structured in two layers. Carbon flux in the bottom layer may either be directly from phytoplankton to zooplankton, or through an intermediate trophic level which is dominated by the microbial web but may also contain smaller mesozooplankton—the taxonomic composition of all compartments in the lower layer is defined only loosely. In a similar way, carbon flux in the upper layer may be directly from zooplankton to top predators Ži.e., endotherms., or via an intermediate trophic level containing aquatic predators such as fish and cephalopods. We have used different transfer efficiencies ŽTE. Žproportion of production in a lower trophic level which appears as production in the next level up. for the different links in the food-web. The TE for the grazers in the microbial web is taken as 40%, the midpoint of the range 30–50% quoted by Fenchel Ž1987., although we are aware that very much higher efficiencies have been quoted for some microheterotrophs. The TE of 12% used for transfers to zooplankton is a slightly pessimistic version of the range of 15–20% quoted by Banse Ž1995. for euphausiids. A TE of 10% for fish and squid is based on an analysis of marine Fig. 1. Ža. Diagram of the foodweb model used to predict primary production from top predator production. Carbon flux between phytoplankton and zooplankton, and between zooplankton and top predators can take either a ‘direct’ route, or an ‘indirect’ route via an intermediate trophic level. Partition of carbon flux between these routes for the top and bottom of the foodweb is independent. The ‘transfer efficiencies’ indicated on the diagram represent the production at the higher trophic level as a proportion of production in the lower component—values used are explained in the text. Žb. Contour plot of the values of primary production ŽGtonne C yeary1 . for the ASO predicted by different partitioning of carbon flux through the food web model. The graticule indicates our most probable estimates of the proportion of carbon flow through the higher and lower trophic levels, and the corresponding primary production. 280 J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 fisheries ŽPauly and Christensen, 1995.. For the consumption of zooplankton by top predators, we have used a TE of 3.5% based on the calculation of 3.7% growth efficiency for crabeater seals in this paper. The value of TE s 2.5% for fish and squid consumption by top predators reflects the fact that part of this diet will involve an additional step Žseals feeding on squid feeding on fish feeding on zooplankton, for example.. The value for TE is arrived at by assuming that the transfer to the top predators is also 3.5% efficient, and that the more complex pathway Žwith overall TE s 0.35%. accounts for about one-third of the overall flux along this route. Taking the two extreme behaviours of the model, a three-level transfer—the classical ‘diatom–krill–whale’ pathway— has the most efficient transfer, with 0.42% of primary production converted into top predator production, whereas carbon flow passing through all five ‘levels’ in the model reduces the transfer efficiency to 0.012%. This very simple model is designed specifically for the purposes of this paper, and is structured to highlight the effects of different trophic pathways. In order to separate the effects of trophic complexity in the top and bottom layers, we have made these independent. As a consequence, all carbon flux from phytoplankton to zooplankton enters the same zooplankton compartment, irrespective of the pathway along which it travels ŽFig. 1a.. The top level has a similar structure, with a direct and an indirect route but all flows directed towards a single top predator compartment. Apart from this property, the model is very similar to other models describing carbon flow through the Southern Ocean pelagic ecosystem ŽHuntley et al., 1991; Banse, 1995.. The model is also closed in the sense that all carbon sequestered by primary production enters the food web and is either respired or ends up in top predator production. We have not included other loss processes, although some of the transfer efficiencies in the bottom layer are deliberately pessimistic Žcompare our 12% transfer efficiency to zooplankton with the 15–20% for euphausiids used by Banse, 1995., and note also that we have used the higher end of the range of predator carbon demand to prime our model. If loss processes are underestimated, then our estimate of primary production could be lower than the real value—a factor discussed later in the paper. In describing the model, we noted the overall transfer efficiencies for the two extreme modes of the model. In order to use the top predator data to predict primary production, we vary the share of carbon flux between direct and indirect pathways independently for the top and bottom layers of the foodweb. The share is described by the proportion flowing along the indirect route—this statistic has a value 100% is when all carbon passes through an intermediate trophic level between either the phytoplankton and zooplankton, or zooplankton and top predators. Taking the upper value of 30 Mtonne C yeary1 from our range of top predator carbon demand Žequivalent to production of 1.11 Mtonne C yeary1 ., 100% indirect flow in both layers implies the implausibly high primary production of ) 8 Gtonne C yeary1 ŽFig. 1b.. At the opposite extreme, with zero indirect flow the foodweb can be sustained by primary production of 0.5 Gtonne C yeary1 . The efficiency of transfer in the top layer has much greater impact than at the bottom, because of the low efficiency used for top predator production and the high efficiency ascribed to the microbial web ŽFig. 1b.. In order to use this approach, we need to narrow our choice of values of the percent indirect flow, to reflect plausible trophic behaviour. Croxall et al. Ž1985. provide the only large-scale attempt to assess the relative contributions of different types of prey to the food supply of Southern Ocean endotherm predators. They estimate that overall 70% of the diet of birds and seals breeding around the Scotia Sea comprises krill, with the remainder dominated by fish and squid. This figure may underrepresent the importance of krill, which is the major component of the diet of baleen whales not covered by Croxall et al. Ž1985., but subsequent evidence that fish and squid may be the dominant component of diet in winter for predators which are krill-feeders in summer may counterbalance this. We also note that the overall contribution of crabeater seals to the prey uptake by endotherm predators will not be well-represented by the Scotia Sea data, and again these are almost exclusively krill predators. We therefore examine a scenario for 70–80% krill Ž‘zooplankton’ in our model. in predator diet, or an ‘indirect’ flow of 20–30%. For the lower layer of the foodweb, we follow the suggestion of Banse Ž1995. of an indirect J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 flow of about 50% for the Southern Ocean as a whole, and note our earlier observation that the output of this simulation is, in any case, less sensitive to the behaviour of this part of the food web. Taking our most plausible picture of transfer through the foodweb, and using top predator production of 1.11 Mtonne C yeary1 , we suggest a value for Southern Ocean annual primary production of the order of 1.7 Gtonne C yeary1 ŽFig. 1b.. Even with 100% flow through the microbial loop, this would only just exceed 2 Gtonne C yeary1 for the 20–30% indirect flow for the top layer, and conversely direct transfer from phytoplankton to zooplankton lowers the value to about 0.9 Gtonne C yeary1 ŽFig. 1b.. If we keep the indirect flow in the lower layer at 50%, then it is easy to see that increasing importance of fish and squid in the diet of top predators implies rapidly increasing primary production—about 3 Gtonne C yeary1 at 50% indirect flow, and 4 Gtonne C yeary1 at 70%. Increasing the complexity of the upper part of the food web, and thereby decreasing its efficiency, has a strong amplifying effect because it demands greater zooplankton production and thus greater primary production. To take a simple example, we can run the model for crabeater seal produc- 281 tion alone Žabout 0.26 Mtonne C yeary1 .. If its diet is 100% krill, then it only requires about 0.1 Gtonne C yeary1 primary production to maintain crabeater seal production. However, for an equivalent piscivorous seal the ‘krill–fish–seal’ food chain increases to this value 1.1 Gtonne C yeary1 , and it would be 11 Gtonne yeary1 for the food chain ‘krill–fish–fish– seal’. Although the model is clearly sensitive to both the values of TE and the way carbon is routed through the food web, we believe that there are sufficient data to remove much of the uncertainty from our estimates. In particular, the predominance of top predators feeding on krill and other zooplankton precludes the higher values of primary production estimated for the ASO. Conversely if all endotherm predators were feeding at the top of the most complex chain covered here they would require primary production close to the value estimated by Longhurst et al. Ž1995. for the entire world ocean. 3. Drawdown of inorganic nutrients The uptake of dissolved nitrogen, silicon and phosphorus have been correlated with primary pro- Table 3 Averaged annual mixed layer nutrient concentration deficitsa for Southern Ocean study areas SiŽOH.4 PO4 NO 3 Source Winter-to-summer South Georgia Bransfield Strait ŽBIOMASS. Weddell Sea ice edge 25–30 b – 17 0.75 ; 0.8 0.5 5 ; 10 6 Whitehouse et al. Ž1996. Priddle et al. Ž1994. Jennings et al. Ž1984. Residual winter conc.a Elephant Island ŽAMLR. Bransfield Strait Bellingshausen Sea MIZ Weddell Sea MIZ Weddell-Scotia MIZ Ross Sea ice edge 5–21b 13 ) 10 - 10 7 20 0.25–0.432 b 0.35 ) 0.4 – 0.3 1 1.5–7 b 4.6 )5 ;5 2.5 15 Silva et al. Ž1995. BAS unpublished data Whitehouse et al. Ž1995. Bianchi et al. Ž1992. Nelson et al. Ž1987. Nelson and Smith Ž1986. Sediment trap a Bransfield Strait ŽRACER. – ; 0.25 ; 10 Karl et al. Ž1991. a a Integrated values divided by mixed layer depth, concentration units millimoles per cubic meter. ‘Winter-to-summer’ data are for studies where the seasonal changes in mixed layer concentrations can be determined directly, whilst ‘Residual winter conc.’ indicates where calculations are based on differences between mixed layer values and winter concentrations predicted from values at the temperature-minimum layer, or immediately beneath the summer mixed layer. The single sediment trap result is included for comparison with the other Bransfield Strait data b Range of values derived for different water masses. 282 J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 duction in a variety of pelagic ecosystems, and may provide the means to integrate primary production over the annual cycle Že.g., Jennings et al., 1984.. Although the number of sites in the Southern Ocean for which suitable data are available is limited, we consider that it is worthwhile to examine the implications of these measurements for primary production. In the context of this paper, it is also important to note that any estimate of primary production based on nutrient drawdown will be independent of predator carbon demand. For Southern Ocean studies, there are two approaches to estimating annual nutrient drawdown. For a very small number of studies, it has been possible either to construct the entire annual cycle ŽJennings et al., 1984; Whitehouse et al., 1996. or to use data for the main growing season Žquasi-annual data. ŽPriddle et al., 1994.. The remaining estimates used here depend on the ability to predict winter mixed layer ŽWML. concentrations on the basis of the residue of the WML, characterised in Antarctic Surface Water by a minimum in the vertical temperature profile, below the summer mixed layer. We have assembled data from nine studies in the Southern Ocean, of which six are based on the residual WML concentration technique and only three are annual or quasi-annual studies. The overall consensus from published data suggests that the average annual changes in concentration of SiŽOH.4 , PO4 and NO 3 are 10–20, 0.5–0.75 and 5–10 mmol my3 yeary1 respectively ŽTable 3.. To convert these data to primary production, we apply conversion factors based on the stoichiometry of nutrient uptake. For silicate, we use a C:Si molar ratio of 4, which appears to be a reasonable compromise value for a range of Southern Ocean phytoplankton communities ŽPriddle et al., 1995, and references therein.. For phosphate, we use the classical Redfield ratio of C:P s 106. If the averaged deficits for these two nutrients are scaled up to the entire Southern Ocean and a mixed layer depth of 50 m, the resulting estimates of primary production are 1–1.5 Gtonne C yeary1 . For nitrate, the picture is confused slightly because this only represents a part, itself poorly constrained, of the nitrogen pool utilised in primary production ŽSambrotto et al., 1993; Banse, 1994; Priddle et al., 1995.. Conformity with the Redfield ratio ŽC:N s 6.625. suggests relatively low primary production, less than 1 Gtonne C yeary1 . Decreasing the uptake f-ratio Žnitrate as a proportion of total nitrogen uptake. produces a corresponding increase in the primary production implied by the nitrate drawdown. Southern Ocean primary production prediction which corresponds to the values obtained for silicate and phosphate drawdown corresponds to an f-ratio of between 0.5 ŽC:NO 3-N f 13. and 0.75 ŽC:NO 3-N f 9.. 4. Assessment and comparison of the two approaches At the present level of knowledge, neither method employed here can be sustained as a proxy for primary production. However, both are useful as yardsticks against which to measure the fairly wide range of estimates of Southern Ocean primary production in the current literature. We need to look at the likely errors embodied in each technique, before comparing the two approaches. We do this by an evaluation of the weaknesses of each, and by considering the implications of primary production values at the extremes of the current range of estimates Žabout 1.2 to 3.5 Gtonne C yeary1 : Huntley et al., 1991; Smith, 1991.. 4.1. Predator carbon demand and trophic efficiency Our approach embodies two components—an extrapolation from one Ždominant. top predator species to the carbon demand for the entire suite of endotherm predators, and the prediction of primary production from this carbon demand using a range of food-web scenarios. We feel that the range of estimates of carbon demand for crabeater seals is the best possible for current data, and that this species is a suitable case study for such an approach to trophic dynamics. It is the most abundant Antarctic Žand global. pinniped species, is predominantly a krill predator, and occupies a midposition in the size spectrum of endotherm predators Žsmall birds to large whales.. Our extrapolation to give a carbon demand for all endotherm predators is based on the likely share of total biomass ascribed to crabeater seals. We have implicitly taken allometric considerations into account in dividing the J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 biomass, and hence carbon flux between the crabeater seal, a midsized predator, and the remainder of the size range from penguins to whales. The translation of predator carbon demand into primary production is dependent on the structure and parameters of the food-web model. We have already noted that the model does not explicitly include loss processes other than those arising from the processing of carbon within and between trophic levels. This includes the assumption that migration out of the Southern Ocean does not represent a significant route for carbon loss, in addition to the lack of explicit treatment of vertical particulate flux. However, we have already noted that we have erred towards caution in the transfer efficiencies specified in the bottom layer of the foodweb, and that top predator production used for all calculations is our upper limit, so that the absence of nontrophic carbon flux should not be viewed as the major omission it might at first seem. This leaves the sensitivity of the model to different interpretations of the trophic structure. We have commented on this already, and noted that the model is sensitive to properties of the top of the model food web, whilst changes in the complexity of the bottom half of the have little effect. Even 90% routing through the microbial web does not increase the value of Southern Ocean primary production to ) 2 Gtonne C yeary1 if predators take F 20% ‘non-krill’ prey. For a 50:50 partition of carbon flow at the Phytoplankton-to-Krill level, a 50:50 partition or worse at the top level is required to imply a primary production value significantly greater than 2 Gtonne yeary1 . If all of the zooplankton carbon reaches the top predators via fish and squid, and the microbial loop accounts for 50% of the carbon flux in the bottom layer, this implies a primary production of around 5.5 Gtonne C yeary1 . The upper bound of current primary production estimates implies that 50–60% of the diet of top predators consists of fish and squid—a scenario which is clearly at variance with observations of Croxall et al. Ž1985. and which conflicts with the fact that crabeater seals and minke whales, both almost exclusively krill predators, together constitute two thirds of the endotherm predator biomass. Therefore, if the food web structure, partitioning and efficiencies used here are accepted as realistic, then either the predicted primary produc- 283 tion is also close to the real value, or higher estimates imply considerable loss of photosyntheticallyfixed carbon from the food web. 4.2. Nutrient uptake The utility of the nutrient drawdown data depends crucially on two properties. First, the stoichiometry of nutrient uptake and carbon fixation need to be defined. As a part of this, we need to satisfy ourselves that primary production is not being underestimated because of nutrient recycling—we have already explored this problem in the case of nitrate. Second, we need to be in a position to judge how far our rather limited dataset can be extrapolated to the Southern Ocean as a whole. The three nutrient datasets each have unique properties. Silicate is taken up by a large, and very important, class of phytoplankton Ždiatoms., so that primary production by phytoplankton which do not sequester silicon is not registered. On the other hand, it is likely that most of the biogenically-fixed silicon is not recycled within the euphotic zone ŽTreguer et ´ al., 1995.. Phosphate is taken up by all phytoplankton, but recycling may result in nutrient drawdown underestimating primary production. We have used these two properties to evaluate the utilisation of nitrate, where it is clear that nitrate is not the only source of nitrogen for primary production. As is shown by the variation in primary production estimates for nitrate uptake, the degree of recycling has major impact—greater levels of recycling imply increased primary production for a given nutrient drawdown. From this point of view, our estimates of primary production may be low, and perhaps closer to new than to recycled production. The question of the representativeness of the data is easier to address. Most of the sites are neritic or close to islands, or are ice-edge studies. These are likely to be in the higher band of primary productivity, albeit in areas where export production is more important than in other areas of the Southern Ocean, so that recycling is likely to be less significant. Overall, we would suggest that our nutrient deficit data will overestimate the overall production in the Southern Ocean. Examining the top and bottom of the range of estimates of Southern Ocean primary production 284 J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 suggests extreme values for nutrient drawdown. Thus for a total annual primary production of 3.5 Gtonne C yeary1 , drawdown for silicate, phosphate and nitrate Ž f-ratio s 0.75. would be expected to be 40, 1.5 and 18 mmol my3 respectively—all values which are significantly higher than observations for sites where primary production and hence nutrient drawdown would be high. At the bottom end of the range of primary production estimates Žabout 1.2 Gtonne C yeary1 ., the corresponding nutrient deficits would be around 10, 0.4 and 5 mmol my3 —at the lower limit of the observed data. 4.3. A consensus between the two estimates We regard the fact that two contrasting and independent methods present similar results supports our estimated value for ASO primary production. Despite the fact that we have used the higher estimate of top predator carbon demand to run our trophic model, the model itself is likely to underestimate primary production because it contains no explicit losses other than those in the transfer between trophic compartments. By contrast, the estimation of primary production based on nutrient drawdown is likely to oÕerestimate primary production, because the studies used are biased towards more productive neritic and ice-edge systems. Therefore, the two methods are likely to be biased in opposite directions, but both indicate similar values which suggests to us that our best estimate of ASO primary production is correct. We can ask whether there is a combination of circumstances which might lead to both methods underestimating ASO primary production. The higher values Ž) 2.5 Gtonne C yeary1 . in the current range of estimates imply food web dynamics and nutrient drawdown which appear to differ strongly from observations. For our value of primary production based on top predator carbon demand to underestimate primary production, either the food web has to be very much less efficient than we suggest, or the percentage of photosynthetically-fixed carbon which ‘escapes’ from the food web must be high. The proportion of annual primary production which is exported Žcrudely but not technically equivalent to that which does not pass through the food web. is normally considered to be no more that 10%, so that we appear to be correct in assuming that such losses can be accommodated within errors in our values for transfer efficiencies. For carbon export to represent an addition to estimated primary production to effectively add 1 Gtonne C yeary1 to ASO primary production, export needs to rise to about 40% in our model. At the same time, the documented nutrient drawdown must ‘hide’ a large part of the primary production. This can only come about if this is being fuelled by nutrients which are not included in our calculations. It would certainly imply a high, but perhaps not improbable proportion of primary production carried out by nonsiliceous phytoplankton. It would suggest that 40% of the phosphorus is recycled, and that uptake f-ratio for nitrogen should average 0.45 for the Southern Ocean as a whole. We do not have firm data to refute any part of this scenario, but the possibility of 40% carbon export appears to merit urgent attention! On this basis, we would suggest that the primary production in the Southern Ocean south of the Subantarctic Front approximates to our estimate of 1.7 Gtonne C yeary1 , or an averaged areal primary production rate of around 30–40 g C my2 yeary1 . Finally, we indicated earlier that we were not approaching this problem from the specific point of view of how much primary production is respired by air-breathing, endotherm predators Žcf. Huntley et al., 1991; Banse, 1995.. However, our estimate of carbon demand for the entire suite of such predators in the Southern Ocean does allow us to provide a limit for the amount of carbon respired. The maximum value would be 96.5% Žbased on transfer efficiency and diet. of the upper bound of carbon demand, or 28.95 Mtonne C yeary1 . This is slightly under 2% of our best estimate for total annual primary production in the Southern Ocean, or 0.8% of the primary production value used by Huntley et al. Ž1991., and agrees with the broad conclusion reached by Banse Ž1995., although van Franneker et al. Ž1997. suggest that the value might be even lower at 0.3–0.6%. Our value is very much less than the CO 2 efflux of up to 860 Mtonne C yeary1 , which is implied by the most extreme case presented by Huntley et al. Ž1991., where 20–25% of primary production of 3.5 Gtonne C yeary1 was respired by endotherm predators. We consider that examination of the partition of carbon in the Southern Ocean food web is a key activity for understanding the biogeo- J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 chemical carbon cycle, but that the specific problem of the role of endotherm predators as a conduit from aquatic primary production to atmospheric CO 2 is best addressed by quantifying carbon demand at the top of the food web. Acknowledgements We are grateful to several of our colleagues for discussion of aspects of this paper, especially Ray Leakey, Paul Rodhouse and Martin White. The comments by the two referees were most helpful, and suggested several improvements to the original which we hope we have implemented to their satisfaction. Appendix A. A carbon budget for crabeater seals The crabeater seal Ž Lobodon carcinophagus. is the most abundant pinniped species in the world and is a dominant member of the top predator community of the Southern Ocean. It is endemic to the Southern Ocean, and its diet consists mainly of krill. The carbon budget for crabeater seals has been assembled as follows. A.1. Population size The most recent estimate of crabeater seal population size was given by Erickson and Hanson Ž1990.. Using information from surveys carried out between 1968 and 1983, they concluded that the size of the crabeater seal population was most likely to be 11–12 million with a minimum of 7 million, which is lower than previous estimates of somewhere between 15 and 30 million ŽLaws, 1984.. Since the size of the crabeater seal population is critical to estimates of carbon flux, estimates are made for minimum Ž7 million., median Ž12 million. and maximum Ž15 million. population sizes. A.2. Age-specific biomass The age-specific biomass Žtotal biomass for each year class. has been estimated from the product of age-specific individual mass and the number of individuals in each one-year age class. Age-specific mass was estimated from growth curves fitted to 285 mass-at-age data from crabeater seals shot for dog food between 1967 and 1977 in the area of Marguerite Bay, Antarctica Peninsula ŽBritish Antarctic Survey unpublished data.. The von Bertalanffy growth curve Žvon Bertalanffy, 1938. was used to smooth and summarise these data. Information about population age structure was obtained from Siler survivorship functions fitted to crabeater age structures by Boveng Ž1993.. This gave the proportion of individuals in each age class and the product of these proportions with the total population size gave the age-specific biomass. A.3. Carbon sequestered in seal biomass The proportion of fat and protein in crabeater seals of each age-specific mass was calculated from the relationships in the carcasses of grey seals Ž Halichoerus grypus: Reilly and Fedak, 1990. and ringed seals Ž Phoca hispida: Stirling and McEwan, 1975.. Water content was assumed to be 0.52 body fresh mass, based on data from female southern elephant seals under a wide range of feeding conditions and body fat content ŽBoyd et al., 1993.. Green et al. Ž1993. found slightly higher water contents in preweaning crabeater seal pups but their measurement projected to the time of weaning supports the view that adult water contents are likely to lie between 0.50 and 0.55, and these values were also used in the calculations. Estimation of the mass of carbon sequestered in seal biomass was made from the product of the carcass composition by mass and the proportion by mass of carbon in protein Ž0.529. and fat Ž0.776. ŽGnaiger and Bitterlich, 1984.. The annual requirement for carbon in the growth of seal tissue was estimated from the sum of the annual growth increments between years. Since pups are dependent upon maternal resources up to weaning, growth in advance of weaning was included as part of the net flux based upon a mass of pups at weaning of 77 to 130 kg with a mean of 97.6 kg ŽGreen et al., 1993.. Apart from the biomass increase due to pregnancy, seasonal changes in biomass were not included in this analysis because these were assumed to be time lags between the assimilation and oxidation of food which were smoothed out within years. The costs of storage were also assumed to be insignificant. 286 J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 A.4. Carbon flux due to metabolism Metabolic rate has not been determined for crabeater seals but studies of metabolic rate in other pinnipeds have shown sufficient consistency across species to suggest that they may also be applicable to crabeater seals after controlling for differences in body mass. BMR was assumed to vary with mass Ž M . according to the equation where metabolic rate s 293Ž M . 0.75 ŽKleiber, 1961.. Although this equation strictly only applies to interspecific comparisons, it may be accepted as a conservative view of the degree to which body mass affects metabolic rate within a single species ŽLindstedt and Boyce, 1985.. Most studies have found metabolic rates for pinnipeds when ashore of 2 to 3.3 times the BMR predicted from mass ŽKleiber, 1961.. Based on these measurements, an average metabolic rate of crabeater seals while hauled out was set at 2.8 times BMR with a minimum of 2.0 and a maximum of 3.3. When at sea, measurements of metabolic rate vary from 4.4 to 6.9 times BMR, although it is believed that lower metabolic rates may be possible for phocid pinnipeds Že.g., Ponganis et al., 1993.. An average at-sea metabolic rate of 4 times BMR was used in our calculations, with 2.5 and 5 times BMR as the upper and lower extremes. The average metabolic rate chosen here also corresponds well with that predicted for animals the size of crabeater seals for the allometric relationship between mass and field metabolic rate given by Nagy Ž1987.. The proportion of time spent in and out of the water was calculated from the frequency of haulout in each hour of the day given by Bengtson and Stewart Ž1992. and supported by Nordøy et al. Ž1995.. Since crabeater seals are known to feed almost exclusively on krill Ž Euphausia superba. ŽLaws, 1984. the ratio of lipid and protein in krill was used to estimate the composition of the metabolic substrate. Clarke Ž1980. showed that, on average, krill contain 4% lipid and 10.5% protein, and using energy densities of lipid of 39.34 kJ gy1 and protein of 18.0 kJ gy1 , the percentage of energy from lipid was: El s Ž 39.34 = 4 . r Ž 39.34 = 4 . q Ž 18.0 = 10.5 . s 45.4% and it follows that metabolic energy derived from protein was 53.6% of the total budget. The carbon contents of lipid and protein were used to estimate the carbon flux due to metabolism Žsee above; Gnaiger and Bitterlich, 1984.. A.5. Carbon flux due to reproduction The calculation of the carbon flux due to reproduction assumed that only animals greater than age-3 years breed and that there was an annual pregnancy rate of 0.87 ŽBengtson and Siniff, 1981; Laws, 1984.. It also assumed that materials transferred from mothers to pups and which were assimilated into pup growth were accounted for in the growth of the 0 q age class. This analysis concentrated upon the metabolic costs of lactation, which has been estimated as 14.6 and 18.1 MJ dayy1 for grey and northern elephant seals Ž Mirounga angustirostris. respectively ŽAnderson and Fedak, 1987; Costa et al., 1986.. Given that this is likely to vary with body size and that crabeater seals have a lactation duration midway between that of grey and elephant seals ŽGreen et al., 1993., then a lactation energy expenditure of 17 MJ dayy1 was used for crabeater seals, with lower and upper extremes of 15 and 19 MJ dayy1 . Carbon flux due to reproduction in males was assumed not to be significantly greater than normal as suggested by studies of metabolic rate in male pinnipeds during the breeding season ŽBoyd and Duck, 1991.. References Anderson, S.S., Fedak, M.A., 1987. Grey seal, Halichoerus grypus, energetics: females invest more in male offspring. J. Zool., London 211, 667–679. Banse, K., 1994. Uptake of inorganic carbon and nitrate by marine plankton and the Redfield ratio. Global Biogeochem. Cycles 8, 81–84. Banse, K., 1995. Antarctic marine top predators revisited: homeotherms do not leak much CO 2 to the air. Polar Biol. 15, 93–104. Bengtson, J.L., Siniff, D.B., 1981. Reproductive aspects of female crabeater seals Ž Lobodon carcinophagus. along the Antarctic Peninsula. Can. J. Zool. 59, 92–102. Bengtson, J.L., Stewart, B.S., 1992. Diving and haulout behaviour of crabeater seals in the Weddell Sea, Antarctica, during March 1986. Polar Biol. 12, 635–644. J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 Bianchi, F., Boldrin, A., Cioce, F., Dieckmann, G., Kuosa, H., Larsson, A.-M., Nothig, E.-M., Sehlstedt, P.-I., Socal, G., ¨ Syvertsen, E.E., 1992. Phytoplankton distribution in relation to sea ice, hydrography and nutrients in the northwestern Weddell Sea in early spring 1988 during EPOS. Polar Biol. 12, 225–235. Boveng, P.L., 1993. Variability in a crabeater seal population and the marine ecosystem near the Antarctic Peninsula. Unpublished PhD Thesis, Montana State University, Bozeman, MT, 110 pp. Boyd, I.L., 1993. Pup production and distribution of breeding Antarctic fur seals Ž Arctocephalus gazella. at South Georgia. Antarct. Sci. 5, 17–24. Boyd, I.L., Duck, C.D., 1991. Mass changes and metabolism in territorial male Antarctic fur seals. Physiol. Zool. 64, 375–392. Boyd, I.L., Arnbom, T., Fedak, M.A., 1993. Water flux, body composition, and metabolic rate during molt in female southern elephant seals Ž Mirounga leonina.. Physiol. Zool. 66, 43–60. Boyd, I.L., Arnbom, T., Fedak, M.A., 1994. Biomass, diet and energy consumption of the South Georgia stock of southern elephant seals. In: Le Beouf, B.J., Laws, R.M. ŽEds.., Elephant Seals: Population Ecology, Behavior and Physiology. University of California Press, Berkeley, pp. 98–117. Chisholm, S.W., Morel F.M.M. ŽEds.., 1991. What controls phytoplankton production in nutrient-rich areas of the open sea? Žpapers from the symposium of the American Society of Limnology and Oceanography.. Limnol. Oceanogr. 36, 1507– 1965. Clarke, A., 1980. The biochemical composition of krill, Euphausia superba Dana, from South Georgia. J. Exp. Mar. Biol. Ecol. 43, 221–236. Costa, D.P., Le Boeuf, B.J., Huntley, A.C., Ortiz, C.L., 1986. The energetics of lactation in the northern elephant seal Mirounga angustirostris . J. Zool., London 209, 21–33. Croxall, J.P., 1984. Seabirds. In: Laws, R.M. ŽEd.., Antarctic Ecology. Academic Press, London, pp. 533–621. Croxall, J.P., Prince, P.A., Ricketts, C., 1985. Relationships between prey life-cycles and the extent, nature and timing of seal and seabird predation in the Scotia Sea. In: Siegfried, W.R., Condy, P.R., Laws, R.M. ŽEds.., Antarctic Nutrient Cycles and Food Webs. Springer, Berlin, pp. 516–533. Erickson, A.W., Hanson, M.B., 1990. Continental estimates and population trends of Antarctic ice seals. In: Kerry, K.R., Hempel, G. ŽEds.., Antarctic Ecosystems. Ecological Change and Conservation. Springer, Berlin, pp. 253–264. Fenchel, T., 1987. Eology of Protozoa—the Biology of Free-living Phagotrophic Protists. Springer, Berlin. Fenton, N., Priddle, J., Tett, P., 1994. Regional variations in bio-optical properties of the surface waters in the Southern Ocean. Antarct. Sci. 6, 443–448. Folklow, L.P., Blix, A.S., 1992. Metabolic rates of minke whales Ž Balaenoptera acutorostrata. in cold water. Acta Physiol. Scand. 146, 141–150. Gnaiger, E., Bitterlich, G., 1984. Proximate biochemical composition and caloric content calculated from elemental CHN analysis: a stoichiometric concept. Oecologia 62, 289–298. 287 Green, B., Fogerty, A., Libke, J., Newgrain, K., Shaughnessy, P., 1993. Aspects of lactation in the crabeater seal Ž Lobodon carcinophagus.. Aust. J. Zool. 41, 203–213. Kleiber, M., 1961. The Fire of Life: an Introduction to Animal Energetics. Krieger, Huntington, New York. Humphries, W.F., 1979. Production and respiration in animal populations. J. Anim. Ecol. 48, 427–453. Huntley, M.E., Lopez, M.D.G., Karl, D.M., 1991. Top predators in the Southern Ocean: a major leak in the biological carbon pump. Science 253, 64–66. Jennings, J.C., Gordon, L.I., Nelson, D.M., 1984. Nutrient depletion indicates high primary productivity in the Weddell Sea. Nature 309, 51–54. Karl, D.M., Tilbrook, B.D., Tien, G., 1991. Seasonal coupling of organic matter production and particle flux in the western Bransfield Strait, Antarctica. Deep-Sea Res. 38, 1097–1126. Kasamatsu, F., Joyce, G.G., 1995. Current status of odontocetes in the Antarctic. Antarct. Sci. 7, 365–379. Laws, R.M., 1984. Seals. In: Laws, R.M. ŽEd.., Antarctic Ecology. Academic Press, London, pp. 194–212. Lindstedt, C.C., Boyce, M.S., 1985. Seasonality, fasting endurance, and body size in mammals. Am. Nat. 125, 873–878. Longhurst, A., Sathyendranath, S., Platt, T., Caverhill, C., 1995. Estimate of global primary production in the ocean from satellite radiometer data. J. Plankt. Res. 17, 1245–1271. Mitchell, B.G., Holm-Hansen, O., 1991. Bio-optical properties of Antarctic Peninsula waters: diferentiation from temperate ocean models. Deep-Sea Res. 38, 1009–1028. Moloney, C.L., 1992. Carbon and the Antarctic marine food web. Science 257, 259. Nagy, K.A., 1987. Field metabolic rate and food requirement scaling in mammals and birds. Ecol. Monogr. 57, 111–128. Nelson, D.M., Smith, W.O., 1986. Phytoplankton bloom dynamics of the western Ross Sea ice edge: II. Mesoscale cycling of nitrogen and silicon. Deep-Sea Res. 33, 1389–1412. Nelson, D.M., Smith, W.O., Gordon, L.I., Huber, B.A., 1987. Spring distributions of density, nutrients, and phytoplankton biomass in the ice edge zone of the Weddell-Scotia Sea. J. Geophys. Res. 92, 7181–7190. Nordøy, E.S., Martensson, P.-E., Folklow, L.P., Blix, A.S., 1994. ˚ Distribution, diving behaviour and food utilization in crabeater seals off Queen Maud Land in Antarctica. In: Antarctic Communities: Species, Structure and Survival. Sixth SCAR Biology Symposium, Venice, 30 May–3 June 1994. Abstracts volume, p. 201. Nordøy, E.S., Folklow, L., Blix, A.S., 1995. Distribution and diving behaviour of crabeater seals Ž Lobodon carcinophagus. off Queen Maud Land. Polar Biol. 15, 261–268. Pauly, D., Christensen, V., 1995. Primary production required to sustain global fisheries. Nature 374, 255–257. Payne, M.R., 1979. Growth of the Antarctic fur seal Arctocephalus gazella. J. Zool., London 187, 1–20. Ponganis, P.J., Kooyman, G.L., Castellini, M.A., 1993. Determinants of the aerobic dive limit of Weddell seals: analysis of diving metabolic rates, postdive end tidal pO 2 ’s, and blood and muscle oxygen stores. Physiol. Zool. 66, 732–749. Priddle, J., Smetacek, V., Bathmann, U., 1992. Antarctic marine 288 J. Priddle et al.r Journal of Marine Systems 17 (1998) 275–288 primary production, biogeochemical carbon cycles and climate. Philos. Trans. R. Soc. London B 338, 289–297. Priddle, J., Brandini, F., Lipski, M., Thorley, M.R., 1994. Pattern and variability of phytoplankton biomass in the Antarctic Peninsula region—an assessment of the BIOMASS cruises. In: El-Sayed, S.Z. ŽEd.., Southern Ocean Ecology: the BIOMASS Perspective. Cambridge Univ. Press, Cambridge, pp. 49–61. Priddle, J., Leakey, R., Symon, C., Whitehouse, M., Robins, D., Cripps, G., Murphy, E., Owens, N., 1995. Nutrient cycling by Antarctic marine microbial plankton. Mar. Ecol. Progr. Ser. 116, 181–198. Reilly, J., Fedak, M.A., 1990. Measurement of body composition in living grey seals by hydrogen isotope dilution. J. Appl. Physiol. 69, 885–891. Sagan, S., Weeks, A.R., Robinson, I.S., Moore, G.F., Aiken, J., 1995. The relationship between beam attenuation and chlorophyll concentration and reflectance in Antarctic waters. DeepSea Res. II 42, 983–996. Sambrotto, R.N., Savidge, G., Robinson, C., Boyd, P., Takahashi, T., Karl, D.M., Langdon, C., Chipman, D., Marra, J., Codispoti, L., 1993. Elevated consumption of carbon relative to nitrogen in the surface ocean. Nature 363, 248–250. SCAR, 1991. Report of the workshop on southern elephant seals. Scientific Committee for Antarctic Research, Monterey, CA, 22–23 May 1991, 29 pp. Silva, N.S., Helbling, E.W., Villafane, ˜ V., Amos, A.F., HolmHansen, O., 1995. Variability in nutrient concentrations around Elephant Island, Antarctica, during 1991–1993. Polar Res. 14, 69–82. Smith, W.O., 1991. Nutrient distributions and new production in polar regions, and contrast between the Arctic and the Antarctic. Mar. Chem. 35, 245–258. Stirling, I., McEwan, E.H., 1975. The caloric value of whole ringed seals Ž Phoca hispida. in relation to polar bear ŽUrsus maritimus. ecology and hunting behaviour. Can. J. Zool. 53, 1021–1027. Treguer, P., Nelson, D.M., Van Bennekom, A.J., DeMaster, D.J., ´ Leynaert, A., Queguiner, B., 1995. The silica balance in the ´ world ocean: a re-estimate. Science 268, 375–379. van Franneker, J.A., Bathmann, U., Mathot, S., 1994. Southern Ocean carbon flux: the role of top predators. Trans. Am. Geophys. Union ŽEOS. 75 Ž3., 177. van Franneker, J.A., Bathmann, U.V., Mathot, S., 1997. Carbon fluxes to Antarctic top predators. Deep-Sea Res. II 44, 435– 455. von Bertalanffy, L., 1938. A quantitative theory of organic growth ŽInquiries on growth laws II.. Human Biol. 10, 181–213. Whitehouse, M.J., Priddle, J., Woodward, M., 1995. Spatial variability of inorganic nutrients in the marginal ice zone of the Bellingshausen Sea during the austral spring. Deep-Sea Res. II 42, 1047–1058. Whitehouse, M.J., Priddle, J., Symon, C.J., 1996. Seasonal and annual change in seawater temperature, salinity and nutrient distributions around South Georgia, South Atlantic. Deep-Sea Res. 43, 425–443. Wiggert, J., Dickey, T., Granata, T., 1994. The effect of temporal undersampling on primary production estimates. J. Geophys. Res. 99, 3361–3371.
© Copyright 2026 Paperzz