Earth and Planetary Science Letters, 114 (1993) 549-554Elsevier Science Publishers B.V., Amsterdam

549

[CL]

Express Letter

Pre- and post-bomb radiocarbon in fish otoliths

John M. KalishMAE Fisheries Greta Point, P.O. Box 297, Wellington, New Zealand

Received December 9, 1992; revision accepted January 4, 1993

ABSTRACT

Measurements of radiocarbon in seawater dissolved inorganic carbon (D1C), or suitable proxies such as hermatypiccorals, are a valuable source of information on carbon flux and ocean circulation. However, knowledge of the globaldistribution of both pre- and post-bomb radiocarbon is limited due to the sources of these data. Suitable hermatypic coralsare restricted to shallow tropical and subtropical waters and oceanographic collections of seawater are prohibitivelyexpensive. What is needed is a proxy for ocean radiocarbon that can be collected at most latitudes and depths, and whichcan he reliably aged. Here 1 report accelerator mass spectrometry analyses of radiocarbon from selected regions of fishotoliths and show that such measurements are suitable for determining both pre- and post-bomb radiocarbon in all oceansand at most depths. Radiocarbon data obtained from otoliths can extend our knowledge of carbon flux in the oceans andatmosphere and help to develop further understanding of the fate of atmospheric CO, and ocean circulation. The datapresented here represent the first pre- and post-bomb time series of radiocarbon levels from temperate waters. Further-more, 1 demonstrate that the dramatic increase in radiocarbon in the atmosphere and oceans, attributable to theatmospheric testing of thermonuclear bombs during the 1950's and 1960's, provides a chemical mark on fish otoliths that issuitable for the validation of age in fishes.

1. Introduction

Measurements of cosmogenically produced andanthropogenic radiocarbon in the atmosphere andocean are invaluable in determining the flux ofcarbon through the biosphere and ocean radio-carbon can also serve as a tracer to investigateocean circulation. Retrospective determinationsof °C can help to understand climate change,and analyses of "C in dated tree rings and handsof hermatvpic corals have proven to he goodproxies for past "C levels [I-6]. Further impetusto investigating historical radiocarbon levels de-rives from two major anthropogenic alterations toatmosphere and ocean "C. The rate of fossil fuel

Present address: Division of Botany and Zoology. Aus-tralian National University. G.P.O. Box 4. Canberra, A.C.T.2601, Australia.

CO„ input to the ocean, the major reservoir ofinorganic carbon, has been estimated by deter-mining changes in ocean radiocarbon inventories[71. The ability to quantify these rates is a by-product of the dilution of cosmogenically pro-duced radiocarbon by ' 4 C-free CO 2 released intothe atmosphere from the burning of fossil fuelsand the incorporation of this CO, into the oceanas dissolved inorganic carbon (D1C). Althoughthe gradual dilution of cosmogenically producedradiocarbon has been obscured by atmospherictests of thermonuclear bombs in the 1950's and1960's. great benefits can he obtained from theseevents through the monitoring, of the flux ofbomb-derived "C in the atmosphere and ocean.

Historical records of ocean 14 C are typicallyobtained from hermatypic corals: however, themajor limitation of corals as recorders of thesedata is that suitable corals are restricted to theupper 50 in of tropical and subtropical oceans.

0012-821X/93/$06.00 c 1993 - Elsevier Science Publishers B.V. All rights re'i'ned

550 J.M. KALISH

Improved estimates of the input of bomb I4 C tothe ocean would result from more widespreadestimates of both pre- and post-bomb "C. Mol-luscan shells can provide an alternative to corals[5,8], but suitable material is often limited indistribution and difficult to date. What is neededis a proxy for ocean 14 C that can be collected atmost latitudes and depths, and which can bereliably aged.

Telcost fish possess three pairs of otolithswithin the paired membranous labyrinths and theyact, in concert with sensory maculae, as bothgravity and auditory receptors. The sagitta, usu-ally the largest otolith, is composed of aragoniteand typically less than 1% organic material. Inmany fish species, the dominant features in sagit-tae are alternating opaque and translucent zones,which together comprise annual increments. Thequantification and measurement of annual incre-ments in otoliths is a major source of informationon the age and growth of fishes [9]. Furtherresolution on the age and growth of fish is possi-ble in many species clue to the ability to resolvedaily increments-bipartite structures composedof an organic matrix-dominant and an aragonite-dominant zone [9]. The suitability of annual anddaily increments in otoliths for the age determi-

nation of fishes is based on the finding that,under normal circumstances, otolith aragonite isnot resorbed after deposition, thus providing arecord of the fish's age and growth [10]. Becauseotolith aragonite is not resorbed during the life ofthe fish, it also retains a chemical record that canbe related to both the biology of the fish and thephysical and chemical properties of its environ-ment. Several studies have investigated trace ele-ments [11-14], stable isotopes [15-17], and ra-dioisotopes [18-20] in fish otoliths in relation to arange of biological and environmental hypothe-ses. Results from research on the growth andcomposition of otoliths suggest that it would bepossible to extract data on both pre- and post-bomb "C from these structures. In this study, Ireport accelerator mass spectrometry analyses of"C from selected regions of otoliths from fish ofknown age to determine if such measurementsare suitable for estimating levels of pre- andpost-bomb '4C.

2. Experimental method

Sagittal otolith pairs from specimens of a ma-rine telcost, Pagrus auratus, were selected fromarchives at the Ministry of Agriculture and Fish-

TABLE 1

Otolith data from Pagrus auratus collected off the east coast of North Island, New Zealand. Values of & 4 C are reported with +1standard deviation

Date spawned Collect. date Fish age (yrs) Otolith wt. (mg) 613C (%c,PDB) ANC (aloe)

1918 1972 54 948 - 2.5 -48.4+ 6.41928 1972 44 7()4 - 2.0 - 39.6 + 6.91932 1972 40 538 - 2.6 - 58.5 + 6.41943 1992 49 810 - 3.0 55.0 + 7.21943 1992 49 810 3.1 -51.1+ 6.61950 1992 42 675 2.6 65.9 + 6.21955 1992 37 689 - 4.1 - 50.6 + 6.71960 1972 12 246 -2.5 - 33.9 + 6.81962 1972 I() 213 -3.0 - 18.6 + 7.01963 1972 9 255 -2.0 - 11.3 + 7.01964 1972 8 210 -2.0 27.8 + 7.61966 1972 6 142 -?.2 55.3 + 8.21969 1973 4 88 -2.3 64.9 F 11.61969 1973 4 88 -2.1 75.6 + 11.91972 1992 353 -2.7 94.0+ 8.51980 1992 12 398 98.4 + 8.81985 1992 7 189 - 2.3 96.5 + 9.81990 1992 2 94 -3.1 81.7+10.51990 1992 2 94 2.9 76.5 + 10.4

0ocb

0

00 0

n if

0

°if:br

PRE- AND POST-BOMB RADIOCARBON IN FISH OTOLITHS 551

eries, Wellington, New Zealand. Throughout itslife P. auratus lives in shallow coastal waters and,on the basis of capture-mark-recapture studies,the majority of fish make net movements of lessthan 50 km. An additional reason for selecting P.auratus was because of the extensive research onthe age and growth of the species [21]. Individualsfor this study were collected from the east coastof the North Island, New Zealand, between EastCape and Hawke Bay (39°S, 179°E), where coastalwaters are greatly influenced by adjacent oceanicwaters and there appears to be no significantupwelling.

Otolith-based ages were determined from rightsagittae using procedures that have been vali-dated for this species by a capture-mark-recap-ture experiment [211. Right sagittae were sec-tioned through the core along the transverse planeand otoliths were baked at 280°C for about 5 min.This increased contrast between the opaque andtranslucent zones of annual increments. Thenumber of annual increments was counted byviewing baked sections with a binocular micro-scope at 50 x magnification. Otoliths were con-sidered acceptable for "C analysis only if therewas agreement on the fish's age among three

otolith readers. It was assumed that there wereno errors in the estimation of age. The date whena fish was spawned was determined on the basisof the fish age at capture. The range of datesspawned encompassed the early industrial periodand the post-bomb decrease in ocean 14 C (Table1).

Otolith aragonite deposited during the firstyear of life was isolated from the left sagitta ofeach fish. This material would have been de-posited during the year when a fish was spawnedbecause the majority of P. auratus spawn be-tween November and February. Sample weightsranged from 25 to 30 mg. Otolith carbonate wasconverted to CO, by reaction with 100% phos-phoric acid and a sample of the CO, was ana-lyzed by mass spectrometry to determine 6"C.The remainder was graphitized and 14 (2 was de-termined by accelerator mass spectrometry (AMS)at the Institute of Geological and Nuclear Sci-ences, Lower Hutt, New Zealand [22]. Radiocar-bon values are reported as which is theage- and fractionation-corrected permit deviationfrom the activity of nineteenth century wood [23].Determinations were made via the NBS oxalicacid standard in conjunction with the ANU su-

H

1,

is

ae)1.Il

e

1-

a

d

t-I

)f)fisd

160-

ci Fiji coral120 -

o Heron Island coral

80- • New Zealand fish otoliths

40-4

0-

-40--00 o

0

°l°0

-801910 1920 1930 1940 1950 1960

Date of calcification (years A.D.)

Fig. 1. A 14 C I- 1 sd) of Pagrus auratus otolith aragonite from New Zealand over the time period 1918 to 1990. PC measurementsfrom hermatypic corals sampled in Fiji and heron Island, Australia are presented for comparison (5,8).

19170 1980 1990

552 J.M. KALISH

crose standard. Reported errors are 1 standarddeviation, and include both counting errors andlaboratory random errors.

3. Results and discussion

Otolith AH C values plotted against date ofcalcification (spawning) result in a curve similarto those derived from corals (Fig. 1). The nearpre-anthropogenic (1918) A 14 C value of —48.4 ±6.8%c was not significantly different from themean of pre-1920 values (-55.5%0) measured inlower latitudes of the southern hemisphere asreported in Toggweiler et al. [5]. (The recordfrom the Galapagos was not included due to theeffect of upwelling.) There was a net decrease inotolith 014 C between 1918 and 1950 to a mini-mum value of — 65.9 ± 6.6%0; however, thesedata provide only weak evidence for a Suesseffect (Se , the decline in atmospheric A 14 C levelsbetween 1900 and 1950) due to their variability.Based on linear regression, the estimated Se be-tween 1918 and 1950 in the ocean off NewZealand was approximately — l6%o. This value ishigher than the Se measured at tropical southernhemisphere locations [5], but it is within the rangeof values determined from corals at sites nearBermuda ( — 25%c) [4] tild Florida ( — I 1.5%0[3]. The effect of the atmospheric detonation ofthermonuclear bombs on ocean I4 C is shown bythe steep rise in z..1 14 C that begins in about 1955.The increase in bomb 14 C measured in otolithsmimics trends determined in hermatypic coralsfrom Fiji [5] and Heron Island, Australia [8] untilabout 1966. After 1966 the rate of increase inbomb "C is greater for the southern hemispherecorals than for otoliths and the peak coral 0 14 C isgreater by at least 40%r. After 1980 there is agradual decrease in otolith ..1 14 C, indicating oceanmixing and the loss of I4 C from seawater DIC toother pools; however. this decrease is more grad-ual than that determined in the corals from HeronIsland and Fiji. This may result from severalfactors, including the ocean transport of bomb"C from lower to higher latitudes, more rapidloss of DIC to other carbon pools in the tropics,or a greater rate of dilution of bomb "C bydeeper water in the tropics. Finally, a directindication that otolith A H C values are a goodproxy for ocean ./.1 14 C is provided by comparing

TABLE 2

Measurements of A N C made on South Pacific surface watersamples from latitudes comparable to collection locations forPagrus auratus otoliths. * The AN C value from Jan. 1954 toMar. 1958 is the mean of eight measurements

Lat. Long. Collect.date

Al 4c

(%o)Rcf.

41.1°S 178.0°E Jan. 1954— —47 * [251Mar. 1958

42.7°S 96.0°W Nov. 1957 —40 [24144.9°S 166.0°W Mar. 1974 80 [7]38.6°S 170.1°W Mar. 1974 110 [7]

surface water measurements of A 14 C near 39°S inthe Pacific [7,24,25] (Table 2) with otolith-baseddata (Table 1).

The similarity of 014 C measured in fish otolithsto the 6,14 C of seawater DIC would be expectedbased on our understanding of the sources ofotolith carbon [10J. Also, evidence from studies ofstable carbon in fish otoliths suggests that about70% of otolith carbon is derived directly fromseawater DIC and that it is incorporated viabranchial pathways [17,26]. The remaining 30%of otolith carbon would be derived from metabolicsources, specifically ingested food. Pagrus auratusis omnivorous, consuming a range of benthic andpelagic invertebrates as well as fishes. Becausethe majority of these dietary items would havez.‘"C levels similar to those measured in seawater[27,28], these alternative sources of ' 4 C would notbe expected to alter A"C values measured inotoliths. However, significant effects due to diet,on otolith 014 C, might be expected in fish thatconsumed detritivores; in deep-sea species thatconsumed prey that ingested material derivedfrom surface waters; in fishes, such as tuna, thatmight forage at fronts associated with upwelledwaters; and in other similar circumstances.

The presence of teleost fishes in all marineenvironments provides a rich source of data onocean 14 C that will make it feasible to modelmore accurately the global flux of carbon andunderstand the fate of atmospheric CO,, theprincipal greenhouse gas. Furthermore, thelongevity of many fish species [18-20], combinedwith our ability to accurately determine the ageof some species, makes it possible to extend theocean record to pre-bomb periods in a wide

PRE- AND POST-BOMB RADIOCARBON IN FISH OTOLITHS

range of environments. The existence of otolitharchives at fisheries laboratories throughout theworld ensures the accessibility of this data source.

Additional benefits to be derived from themeasurement of 14 C in fish otoliths relate to thedetermination and validation of fish age. Age isamong the most important parameters for deter-mining the production of fish populations. Fishages arc generally determined through the inter-pretation of periodic features on calcified struc-tures such as otoliths, scales, spines or vertebrae[9]; however, despite the great emphasis on ageestimation, it is infrequent that a method of agedetermination is proven to be accurate [29]. Thelack of age validation studies can be attributed tothe high cost and logistical difficulty of mark-re-capture studies. The application of the bomb "Cchronometer and AMS to the validation of fishages provides an alternative to traditional valida-tion studies and, in many cases, provides a methodfor age validation where standard techniques arenot suitable. The method is based on the occur-rence of an anthropogenic chemical marker inotoliths, specifically the sharp increase in A"C inthe 1950's and 1960's that should he evident inthe otoliths of fish species that inhabit at least theupper 200 m of the water column. The rationaleis similar to that employed in age determinationstudies of molluscs using the bomb "Cchronometer [8,30]; however, as in the presentstudy, only the earliest formed portions of theotoliths need he analyzed. An age 'calibrationcurve' based on otolith A H C in fish of known age,such as P. auratus, could be produced for anocean region and L.1 14 C determinations from otherspecies could be applied to the relationship. Al-ternatively, more reliable results would be ob-tained by producing a sigmoid A 14 C versus pre-sumed annual increment curve and. ultimately.,assigning ages to the fish. In many cases agevalidation via the measurement of ' 4 C would bemore effective than mark-recapture methods.Mark-recapture studies only provide direct evi-dence of the frequency of annual increment for-mation between the time of marking and recap-ture, whereas determination of "C levels in fishotoliths can he used to determine the absoluteage of a fish (if samples are available with pre-sumed birth dates in the range of about 1958 to1970). In those instances where the fish are not

553

long-lived ( < 30 yrs) it should be feasible to uti-lize the historical otolith collections held at manyfisheries laboratories. Furthermore, estimation ofage using the bomb chronometer eliminates thestress related to capture and tagging, which cangreatly affect fish growth and bias results.

In conclusion, measurements of "C in fishotoliths can provide both pre- and post-bombestimates of ocean "C in a broad spectrum ofenvironments and can serve as an adjunct to datafrom hermatypic corals in shallow tropical andsubtropical seas. These data can be applied tostudies of carbon flux in the atmosphere andoceans, and make it feasible to investigate oceancirculation in most regions on time scales rangingfrom years to decades. Fish biologists can alsobenefit from AMS analyses of otolith "C due tothe utility of these data for the validation of fishage.

Acknowledgements

I thank L. Paul for access to the otolith archivesand R. Sparks and J. McKee for sample process-ing. AMS analyses were carried out by the Insti-tute of Geological and Nuclear Sciences, LowerHutt, New Zealand under contract to MAF Fish-eries.

References

1 H.E. Suess, The radiocarbon record in tree rings of thelast 8000 years, Radiocarbon 22, 200-209, 1980.

2 M. Stuiver and P.M. Quay, Atmospheric 14 C changesresulting from fossil fuel CO 2 release and cosmic ray fluxvariability, Earth Planet. Sci. Lett. 53. 349-362. 1981.

3 E.M. Druffcl and T.W. Linick. Radiocarbon in annualcl i.9) 7Ni1. rings of Florida, Geophys. Res. Lett. 5, 913-916,1978.

1 Y. Nozaki, D.M. Rye, K.K. Turekian and R.E. Dodge, A200-year record of carbon-13 and carbon-14 variations inBermuda coral, Geophys. Res. Lett. 5, 825-828. 1978.

5 J.R. Toggweiler, K. Dixon and W.S. Broeckcr. • Fhe Peruupwelling and the ventilation of the South Pacific thermo-cline. J. Geophys. Res. 96. 20467-20496, 1991.

6 E.R.M. Druffel. Bomb radiocarbon in the Pacific: Annualand seasonal timescale variations, J. Mar. Res. 45, 667-698.1987.

7 W.S. Broecker, T.-II. Peng. G. Ostlund and M. Stuiver.The distribution of bomb radiocarbon in the ocean. J.Geophys. Res. 90. 6953 -6970, 1985.

8 N.H. Landman, E.R.M. Druffel, J.K. Cochran. D..1. Don-ahue and A.J.T. Jull. Bunch-produced radiocarbon in the

is!!

tierfor

1 to

M;ed

thetedofof

nutDM

via)(70

tits.tndJSC

iveternot

iniet,hathatiedhatled

incondelindthetheiedtgetheide

554 J.M. KAI.ISH

shell of the chambered nautilus: rate of growth and age atmaturity, Earth Planet. Sci. Lett. 89, 28-34, 1988.

9 R.C. Summerfelt and G.E. Hall, eds., Age and Growth ofFish, 544 pp., Iowa State University Press, Ames, Iowa,1987.

10 K. Simkiss, Calcium metabolism of fish in relation toageing, in: Ageing of Fish, T.B. Bagenal, ed.. pp. 1-12,Unwin, London. 1974.

11 J.S. Edmonds, M.J. Moran, N. Caputi and M. Morita.Trace element analysis of fish sagittae as an aid to stockidentification: pink snapper (Chrysophrys auratus) inWestern Australian waters, Can. J. Fish. Aquat. Sci. 46,50-54, 1989.

12 J.M. Kalish, Otolith microchemistry: Validation of theeffects of physiology, age and environment on otolithcomposition. J. Exp. Mar. Biol. Ecol. 132, 151-178. 1989.

13 D.W. Townsend, R.I.. Radtke, M.A. Morrison and S.D.Folsom, Recruitment implications of larval herring over-wintering distributions in the Gulf of Maine inferred usinga new otolith technique, Mar. Ecol. Prog. Ser. 55, 1-13,1989.

14 J.M. Kalish, Determinants of otolith chemistry: seasonalvariation in the composition of blood plasma, endolymphand otoliths of bearded rock cod Pseudophycis harhatus,Mar. Ecol. Prog. Ser. 74, 137-159, 1991.

15 I. Devereux, Temperature measurements from oxygen iso-tope ratios of fish otoliths, Science 155. 1684-1685, 1967.

16 S.A. Mulcahy, J.S. Killingley, C.F. Phleger and W.H.Berger. Isotopic composition of otoliths from a ben-thopelagic fish, Coryphaenoides acrolepis, Macrouridae:Gadiformes, Oceanol. Acta 2, 423-427, 1979.

17 J.M. Kalish, I3 C and 18 0 isotopic disequilibria in fishotoliths: metabolic and kinetic effects, Mar. Ecol. Prog.Ser. 75, 191-203, 1991.

18 J.T. Bennett. G.W. Boehlert and K.K. Turekian. Confir-mation of longevity in Sehastes diploproa (Pisces: Scor-paenidae) using 210 Pb/ 226 Ra measurements in otoliths,Mar. Biol. 71, 209-215, 1982.

19 S.E. Campana, K.C.T. Zwancnburg and J.N. Smith.21 °Pb/ 226 Ra determination of longevity in redfish, Can. .1.Fish. Aquat. Sci. 47, 163-165, 1990.

20 G.E. Fenton, S.A. Short and D.A. Ritz, Age determina-tion of orange roughy, Hoplostethus atlanticus (Pisces:Trachichthyidae) using 210 Pb: 226 Ra disequilibria, Mar.Biol. 109, 197-202, 1991.

21 R.I.C.C. Francis, L.J. Paul and K.P. Mulligan, Ageing ofadult snapper (Pagrus auratus) from otolith annual ringcounts: validation using tagging and oxytetracycline injec-tion, Aust. J. Mar. Freshwater Res. 43, in press, 1992.

22 G. Wallace, R.J. Sparks, D.C. Lowe and K.P. Pohl, TheNew Zealand accelerator mass spectrometry facility, Nucl.Instrum. Meth. B29, 124-128, 1987.

23 M. Stuiver and H.A. Polach. Reporting of 14 C data. Ra-diocarbon 19, 355-363, 1977.

24 G.S. Bien, N.W. Rakestraw and H.E. Suess, Radiocarbonconcentration in Pacific Ocean water, Tellus 12, 436-443,1960.

25 T.A. Rafter and B.J. O'Brien, Exchange rates between theatmosphere and the ocean as shown by recent C-14 mea-surements in the South Pacific, in: Nobel Symposium 12,Radiocarbon Variations and Absolute Chronology, I.U.Olsson, ed., pp. 355-377, Wiley, New York, 1970.

26 J.M. Kalish, Oxygen and carbon stable isotopes in theotoliths of wild and laboratory-reared Australian salmon(Arripis maw), Mar. Biol. 110, 37-47. 1991.

27 W.G. Pearcy and M. Stuiver, Vertical transport of carbon-14 into deep-sea food webs, Deep-Sea Res. 30, 427-440,1983.

28 P.M. Williams, E.R.M. Druffel and K.L. Smith, Dietarycarbon sources for deep-sea organisms as inferred fromtheir organic radiocarbon activities. Deep-Sea Res. 34,253-266, 1987.

29 R.J. 13eamish and G.A. McFarlane, The forgotten require-ment for age validation in fisheries biology, Trans. Am.Fish. Soc. 112, 735-743, 1983.

30 K.K. Turekian, J.K. Cochran, Y. Nozaki, I. Thompson andD.S. Jones, Determination of shell deposition rates ofArctic(' islandica from the New York Bight using natural228 Ra and 228 Th and bomb-produced I4 C, Limnol.Oceanogr. 27, 737-741, 1982.