Feeding habits of harp and hooded seals in Greenland waters

Results of stomach contents analyses of harp and hooded seals collected in West Greenland waters in the period 1986-1993 are reviewed, and compared with published data and circumstantial information from local hunters. The diet of harp seals in this region is variable but consists mainly of pelagic crustaceans (Thysanoessa spp. and Parathemisto libellula) and small fish species like capelin (Mallotus villosus) , sandeel (Ammodytes spp.), polar cod (Boreogadus saida) and Arctic cod (Arctogadus glacialis) . Species of importance for commercial fisheries in Greenland, such as Northern prawn (Pandalus borealis), Atlantic cod (Gadus morhua), and Greenland halibut (Reinhardtius hippoglossoides) playa minor role in the diet of harp seals in this area. Variation in the diet of hooded seals is less well documented, but in addition to the species also taken by harp seals, larger demersal fishes like Greenland halibut, redfish (Sebastes spp.), cod, and wolffish (Anarhichas minor) are apparently important prey items. Kapel, Fo. 2000. Feeding habits of harp and hooded seals in Greenland waters. NAMMCO Sci. Publ. 2:50-64.


Introduction
During the 1970's and 1980's, harp seals (Phoca groenlandica) and hooded seals (Cystophora cristata) were often on the front pages because of an intense debate concerning the commercial harvesting of seal pups at the whelping patches.Partly because of this debate and the following decreasing demand for seal skins, and partly because of management decisions leading to catch regulations, the commercial catches of these seals decreased dramatically (particularly off Newfoundland), and the Northwest Atlantic stock of harp seals has apparently increased (Stenson et al. 1993, Stenson et al. 1996).Also the stock(s) of hooded seals in the Northwest Atlantic may have increased, but the evidence for that is less convincing (Bowen et al. 1987, Hammill et al. 1992, Stenson et al. 1997, Myers and Stenson 1996).
The rise in population size of harp seals (and other seals) may lead to another debate.The fishing industry and local fishermen are very concerned about the potential competition from consumption by these increasing seal stocks.This is not a new concern, but it is particularly relevant when seal stocks are expanding.
However, there were few quantitative analyses of seal diet in Greenland, and with limited temporal or regional coverage.In order to improve the knowledge, material from harp seals has been collected since 1985 as part of three parallel projects: The ' Heavy Metal Project' (HMP) by Greenland Environment Research Institute, a  Similar material for quantitative analysis of hooded seal diet is limited (Kapel 1982(Kapel , 1995)).

--
The purpose of this presentation is to summarize the information available in these earlier reports and meeting documents on the feeding habits of harp and hooded seals in Greenland waters during their migrations and stay in these regions, i. e. outside of the whelping and moulting seasons.

MATERIAL AND METHODS
In connection with collection of j aws of harp and hooded seal for analyses of age composition of catches in Greenland, hunters provided written info rmation on stomach contents of thousands of harp and hooded seals in Greenl and in the period 1970-1983(Kapel MS 1973, Kapel and Ge isler MS 1979, Kapel 1982).The hunters' information was not particularly specifi ed, and it is not known how they evaluated the stomach contents, but their reports were used for comparison with results of analyses of stomach samples, or as indication of food preferences when samples were lacking.
During the period 1985-1993 stomachs were sampl ed from 1,175 harp sea ls caught in Greenland (Table 1, Fig. I).In connection with thi s sampling program, stomachs from 44 hooded seals were also obtained.
All samples were frozen as soon as possible for later shipment to the laboratory in Copenhagen.
In the laboratory the material was treated as described in previous papers (Kapel andAngantyr MS 1989, Angantyr andKapel 199 1, Kapel MS 1994): For all species or taxonomi c groups (Table 2) the number of specimens were either counted or estimated.The total number of each fish species was calculated by adding the number of fresh fishes to the number of intact skull s and half the number of free otoliths in the stomach.For cephalopods the total number was calculated by adding the number of fresh specimens to the highest number of either lower or upper beaks.
For larger species of crustaceans (crabs, prawns), the number of specimens was counted.When small species (euphausids, amphipods, mysids) were present in large numbers the amount was estimated by counting a subsample.
The fresh weight of the stomach contents was sometimes calculated using otolith-fresh weight equations (e.g.Hark6nen 1986), but most often standard va lues fo r each species were used.Details on the values used were given in previous reports (e.g.Angantyr and Kapel 199 1).
The composition of the diet was expressed in terms of mean estimated weight percentage ("calculated biomass") calculated on the basis of the number and average weight of each species or taxonomic group identified in the stomach.Weight percentages are presented in this paper for harp seal only.Samples from several sampling localities are pooled to give an overview of variations between regions and seasons.For hooded seal, hunters' information, expressed as per cent occurrence (i.e. the proportion of stomachs in which a species is recorded) is compared with the limited stomach analysis material, expressed as calculated biomass or mean per cent volume (the visually estimated volume of a prey species in relation to the volume of all food items found in the stomach).

Harp seal feeding in Greenland
Considerable geographical and seasonal variation was demonstrated, but some general patterns in feeding habits emerged.These are illustrated in Table 3 and Fig. 2, and can be summarized as follows: In the coastal waters of Southwest Greenland two prey items dominate the diet of harp seals: capelin (Mallotus villosus) and "krill" (euphausids).In the early summer (May) krill appears to be the predominant food; later, in June and in the autumn months, capelin constitutes the major part of the food .Codfishes, mostly cod (Gadus morhua) and Greenland cod (G.ogac), other fish species, and prawns (particularly Northern prawn Pandalus borealis) are also taken but altogether they account for less than 20% of the diet in terms of biomass.In the winter months, however, these 'secondary prey species' appear to contribute with more than one third of the food biomass, although capelin is still the most important prey item.
In the western part of Central West Greenland, i.e. the region around the southern entrance to the Disko Bay, the pattern resembles that found in Southwest Greenland: In June-July pelagic crustaceans (euphausids and amphipods, Parathemisto sp.) dominate the diet, followed by cape lin, and prawns.In the autumn, however, capelin is the predominant prey item, whereas the winter diet is composed of a number of other fish species, pelagic crustaceans, and squid (mainly Gonatus fabricii).
Few samples were obtained from offshore waters in Southwest and Central West Greenland (Table 3).They show considerable variation between areas and months, with sandeel (Ammody tes sp.) as a very important food item, supplemented by Parathem is to, Pandalus, redfish (Sebastes sp.), and squid (Kapel MS 1974).
In addition to the above-mentioned species, the diet of harp seals feeding in inshore waters of West Greenland includes a number of other fish species, prawns, and squid (see Table 2).Some samples indicate that the relative importance of these 'secondary food items' is greater in the autumn and winter, but other samples demonstrate that capelin is the predominant prey also at this time of the year.

Hooded seal feeding in Greenland
Much less information is available on hooded seal diets in Greenland than for harp seals.
Accounts in the literature, and information gathered from hunters were reviewed by Kapel (1982), and are summarised in Table 4a and Fig. 3. Results of exam ination of the contents of the few hooded seal stomachs obtained in connection with sampling of harp seal stomachs are presented in Table 4b and Fig. 3).
South Greenland hunters provided information on the stomach contents of 1,236 hooded seals taken during the spring hunt, 386 of which (3l.2%) were empty.Of the 850 stomachs with contents, 828 (97.4%) contained fish, whereas crustaceans and squid were reported in only 16 and 6 stomachs -less than 2% and 1 %, respectively.For many stomachs, the contents were only given as "fish", but the fish species reported most frequently were redfish, cod or Greenland cod, and capelin.
Six stomachs of hooded seals caught in Southwest Greenland have been examined in the laboratory.One was empty, and in the remaining five cape lin was the predominant food.Three of the seals were taken in Apri l, two in June, and one in November; they were all very young animals (0-1 year old) -in contrast to the above mentioned hunters' samples that were dominated by seals of age 2 or older (Kapel 1982).
Hunters ' information on stomach contents is available for 236 hooded seals caught in Southeast Greenland (Ammassalik district) 1970-74.Of 215 caught in late July-early August; 201 (93.5%) were empty.This is not surprising because the seals were in the late stage of moulting, during which hooded seals are known to have reduced feeding activity (Rasmussen 1960, Potelov et al. this volume).Of the 21 reports from this district later in the year (autumn and winter), only 5 (24%) stomachs were empty.The prey species reported most frequently by the Ammasalik hunters was redfish (in 83% of the stomachs with contents), followed by Green land halibut (Reinhardtius hippoglossoides, 7%).Neither crustaceans nor squid were reported by the hunters.
In Southeast Greenland, 29 hooded seal stomachs were collected in early September 1991.Laboratory examination revealed that 9 (31 %) were empty, and that the rest contained only small amounts of food remains.The predominant prey item was squid, particularly identified by the eyes found in 80% of these stomachs, and constituting about 50% (calculated biomass) of the stomach contents.Prawns (Pandalus sp.) dominated in five stomachs (25% occurrence, 9% of biomass).The stomach contents ofa three year old seal were dominated by redfish, and fish otoliths (most frequently polar cod) were found in five other stomachs.Most of the seals in this samp le were young of the year (70%), whereas these constituted only a minor part (3%) of the 1970-74 samples.
From Northwest Greenland, hunters' information is available for 6 14 hooded sea ls, of which 153 (25%) had empty stomachs.The highest proportion of empty stomachs (42%) was found in July, which probably relates to recent moulting condition, the lowest in late autumn (18% ) (Kapel 1982).In this region, Greenland halibut was the prey reported most frequently (in 61 % of the stomachs with contents), followed by wolffish (Anarhichas sp.) (11 %), and pelagic crustaceans (12%).
Laboratory analyses of stomach contents of hooded seals caught in the northern part of West Greenland are few: One stomach from Vaigat (Central West Greenland) contained 64% capelin, 32% polar cod, and 4% other fish remains (calculated biomass); the contents of seven stomachs from Uummannaq and Upernavik were dominated by Greenland halibut (87% and 65%, respectively), followed by redfish (Uummannaq, 4%) or polar cod (Upernavik, 25% biomass).

58-
In recent years, a considerable amount of additional data on harp seal feeding patterns in Canadian waters have been published or presented at meetings in international fora : Murie and Lavigne (1991) and Beck et al. (1993) studied the diet of harp seals collected in the Gulf ofSt.Lawrence, Lawson et al. (1995Lawson et al. ( , 1998) ) and Lawson and Stenson (1997) reviewed a large material of stomach contents of harp seals in nearshore and offshore waters around Newfoundland and Labrador, and Lawson and Stenson (1995) discussed historic variation in the diet of harp seals in this region.The present knowledge on the diet of harp seals during their stay in Atlantic Canada is thus based on extensive sampling, and is well documented.
Less is known on harp seal feeding in Arctic Canada.Sergeant (1973) presented stomach contents data on 16 harp seals caught in the Canadian Arctic (at four different localities).Polar cod occurred in 6 of these, mysids in 5, Parathemisto in 4, and euphausids in 3 stomachs.Foy et al. (1981, cited in Sergeant 1991) reported on harp seal feeding along the Labrador coast.They found that immature and adult harp seal feeding in bays at the Labrador north coast in late May-June were taking mainly capelin, whereas the diet of juveniles feeding near the offshore archipelago consisted mainly of euphausids.Immatures and adults feeding further offshore had a more varied diet including fi sh, euphausids, and bottom living decapods .In November-January harp seals feeding in the bays were feeding on a variety of small Gadidae, polar cod, and capelin, whereas invertebrates were of secondary importance.Smith et al. (1979) working in southeastern Baffin Island (63-64 0 N) in July-September found that the diet of four young-of-the-year harp seals consisted of 64% pelagic crustaceans (mysids and Parathemisto), and 36% fish (mainly polar cod).One adult female harp seal was feeding on the pelagic shrimp Sergestes arctic us.
In the western Hudson Strait, Beck et al. (1993) found that the stomachs of 14 harp seal caught in September-October near the south coast were dominated by capelin, with polar cod, Greenland cod, sculpins, and flatfishes as secondary elements.One seal feeding offshore, near Salisbury Island, had taken Parathemisto and polar cod in almost equal amounts.
In the High Arctic, Finley et al. (1990) analysed 63 stomachs of harp seals at Pond Inlet (73 0 N) and Grise Fiord (76 0 N) between mid-August and early October.Polar cod occurred in all of them, and accounted for 84% of all food items found in the stomachs (% frequency).Arctic cod was found in 63% of the stomachs, but in low numbers (5% frequency).In terms of biomass, however, polar cod and Arctic cod contributed with 66% and 33%, respectively.Other fish species and invertebrates were also found, but accounted for only a minor part of the diet (altogether 16% frequency, and about 1% by weight).
The analyses of the material collected in Greenland is in accordance with a general conclusion that can be drawn from the studies in coastal areas of the Canadian Arctic, namely that capelin is a predominant prey in the southern part of the range, polar cod in the northern part, and that krill and Parathemisto make significant contributions to the diet in some seasons or areas.Sergeant (1973Sergeant ( , 1991) ) stated that small crustaceans were taken particularly by young harp seals.In some of the summer samples from Southwest Greenland crustaceans seem to constitute a greater part of the diet of young animals, but in other samples no indication of an age dependent difference in diet was found (Angantyr and Kapel 1991).In general, the analyses of the Greenland samples indicate that the relative importance of the 'primary' food items (pelagic crustaceans vs. pelagic fish species) in the diet of harp seals in West Greenland depends on the season and locality, rather than on the age of the seals (Kapel MS 1994).Some of the samples indicate, however, --------------------------------------------------------59 NAMMCO Scientific Publications, Volum e 2 that the food items referred to above as 'secondary' may constitute a larger fraction of the diet of adult seals than of immature sea ls.
All these studies indicate great seasonal and regional, and probably also year-to-year or longterm variation in the feeding of harp seals, and thus stress the need for continued samp ling and analyses in order to document the feeding patterns better and elucidate any short-or long-term changes in food preference or avail ability.For some seasons or areas, e.g.offshore waters, the present knowledge is based on a very limited data set.
In order to translate the knowledge on feeding patterns to an evaluation of the consumpti on at the popul ation level, information on population size and dynamics , on the spatial distribution of the seal during the year and on the energy requirements of harp seals, is also needed.
The present population size of the Northwest Atlantic stock is fairly well known (Stenson et al. 1993, Stenson et al. 1996), and its composition and dynamics can be modelled.A crucia l point may be the question on the spatia l distribution of the stock within its general distribution area.The general distribution and migration pattern of harp seals is fairly well known (see e.g.Sergeant 1991), but when it comes to estimating the number occurring in a given area at a given time our knowledge is far from satisfactory, and ways to improve that knowledge are not easy to find -or extremely resource demanding.Satellite telemetry of harp seals has already revealed new details about the movements of a number of individuals (Stenson and Sjare MS 1997), but a major effort in satellite tagging is required to illustrate the dispersal of the entire population all the year round.An alternative approach may be to make reasonable guesses and check the ranges of the results by simulations.
A fair amount of data on the energy requirements of harp seals, and calculations of the energy budget of harp seal populations has been publi shed (e.g.Lavigne et al. 1985, Keiver et al. 1984), which may be app licable in the present context.But again, the allocation to certain regions or seasons is highly speculative.
Using avai lable knowledge on diet, abundance, and energy requirements, Stenson et al. 1997 estimated the consumption by harp seals during their stay in the southern range of the distribution, and Hammi ll and Stenson (forthcoming) made similar ca lcul ations for all seal species occurring in the waters of Atlanti c Canada.They made assumptions of the part of the seal populations present in these waters during winter and summer, and using similar assumptions one could attempt to calculate the consumption by harp and hooded sea ls during their stay in the Arctic.A next step would be to make assumptions on distribution in Arctic Canada, offshore waters and coastal Greenland.
Questions relating to possible interactions between an increasing population of harp seals and commercial fisheries in Greenland is part of the background for conducting the studies on harp sea l diet reported in this and previous papers.The most important species for the commercial fishing fleet in Green land are Northern prawn, cod, and Greenland halibut.According to the present ana lysis, none of these species constitute a signi ficant fractio n of the stomach contents of harp seals in Greenland waters.
It is known that other seals sometimes eat parts of larger fish without taking the head with the oto li ths (Bou lva andMcLaren 1979, Harkonen 1986).If harp sea ls often feed that way, this analysis is biased.However, it is generally believed that they take their prey by suction, i.e. swa llow the items whole (Ronald and Healey 1981).It appears, therefore, that harp seals in Greenland waters were of little importance as direct competitors to the commercial fishery in the late 1980 's. Whether the diet of harp seals in Greenland waters has changed in recent decades is not known ; but it is likely that the fraction of Atlantic cod in the diet was higher previously, when that species was extremely abundant off West Greenland.

-
The role of hooded sea ls is perhaps even more difficult to assess, because the knowledge on population size and distribution is far from complete, and quantitative data on feeding very limited.Besides the info rmati on from Greenland presented here, Ross (M S 1993) analysed the contents of 67 hooded seal stomachs collected around Newfoundland.Seven seals taken at the whelping patch had empty stomachs.The four most important f ish species were Greenland halibut, redfish, Atlantic herring (Clupea harengus) and polar cod, accounting fo r 96% of the stomach contents (expressed as both per cent occurrence and as calculated biomass); squid occurred in 70% of the stomachs and constituted 52% of the number of all items fo und, but only 1 % of the REFERENCES: biomass.Hammill et al. (1997) and Hammill and Stenson (forthcoming) used thi s info rmation to estimate the consumption by hooded seals in the Gulf of St. Lawrence and the entire Atlantic Canada region.
It is worth noting that although the number of hooded seals in the Northwest Atlantic is only about one tenth of that of harp seals, the average body mass is 2-3 times as large and their diet appears to include a much higher fracti on offi sh species exploited commercially in Greenland.It is quite possible that the consumption by hooded seals in some areas of Greenl and constitutes a more direct conflict with f isheries than in the case of harp seals.

Fig
Fig. 1.Map of Greenland showing the geographical areas referred to and th e locations jimll which stomach samples were obtained.

Table 3
. Harp seal food composition in Greenland in terms of calculated biomass (Weight %), by area and month; subareas (s) southern, (n) northern, (w) western, and (e) eastern part.

Table 4a .
Hooded seal food composition in Greenl and, information fro m hunters expressed as % occurrence in all reports, and in reports of stomachs with contents *The percentages of the various fish species are adjusted proportionally to account for "unspecified fish".Table4b.Hooded seal stomach contents in Greenl and samples in terms of estimated volume per cent (Vol.%) and calculated biomass (Weight %).In most of the samples from NorthwestGreenland, polar cod and to some extent Arctic cod (Arctogadus glacialis) dominated the harp seal diet.In the Uummannaq district, however,