AACAI NEWSLETTER NO. 89 APRIL 2002 Calibrating Marine Radiocarbon Dates: A Guide to Australian ∆R Values Introduction radiocarbon dates obtained on marine samples, including the ability to account for Every year hundreds of radiocarbon dates regional differences from the global model are obtained on Australian marine shells with the input of a local/regional ∆R value and corals by archaeologists and (Stuiver et al. 1986). Reimer and Reimer geomorphologists. As a first approximation (2000, 2001) recently summarised all of the it is common practice to correct these dates available global ∆R values in a world wide for marine reservoir effect by simply web database. subtracting a generalised factor of 450 ± 35 years to make them comparable to coeval In this paper, I briefly discuss the principles terrestrial (e.g. charcoal) samples. of marine reservoir correction before presenting a guide to regional and Gillespie calculated this correction value in subregional Australian ∆R values extracted the 1970s (see Gillespie 1975; Gillespie from the Reimer and Reimer (2000, 2001) and Polach 1979; Gillespie and Temple database and recent work presented in Ulm 1977). Since that time several studies have (2002). suggested the possibility of significant deviations in regional marine reservoir Background signature from this generalised value (e.g. Hughes & Djohadze 1980; Murray-Wallace A basic assumption of the radiocarbon 1996; Spennemann & Head 1996; Ulm et dating method is that the concentration of al. 1999; Woodroffe & Mulrennan 1993; radioactive carbon (14C) in the biosphere is Woodroffe et al. 1986:75, 77). uniform through space and time. Early in the development of the radiocarbon dating In the time that has elapsed since method, however, it was recognised that Gillespie’s pioneering study, researchers marine shells exhibited a systematic age have gained a much more sophisticated difference to contemporary terrestrial appreciation of the complexity of global samples on a regional basis that allowed marine carbon reservoirs. One of the most calculation of a regionally specific age significant developments was the offset. calculation of a global model of marine 14C activity that enabled the calibration of Global variation in marine reservoir effects 10 AACAI NEWSLETTER NO. 89 APRIL 2002 evident in marine shell carbonates are certain coral species with well-defined principally caused by incomplete mixing of annual growth structures can provide the upwelling water of ‘old’ inorganic most accurate determination of local marine carbonates from the deep ocean where reservoir effects (Reimer and Reimer long residence times (>1000 years) cause 2000). Unfortunately, few such coral studies depletion of 14C activity through radioactive are available and they are limited largely to decay, resulting in very old apparent 14C tropical regions with long-term coral ages (Mangerud 1972). Estuarine records. reservoirs are even more complex with the interaction and incomplete mixing of 14C In recent years, regional marine reservoir from both terrestrial reservoirs and marine effect has commonly been expressed as a reservoirs from tidal action (e.g. Little ∆R value (e.g. Higham & Hogg 1995; 1993). Phelan 1999). Stuiver et al. (1986; Stuiver & Braziunas 1993; Stuiver et al. 1998) Regional differences in marine reservoir modelled global marine 14C activity using a effect are generally determined through one simple box diffusion global carbon cycle or a combination of three methods: model of marine reservoir responses to variation in atmospheric 14C activity. 1. direct radiocarbon dating pre-AD Regional deviations from the modelled 1955 live-collected shell specimens marine calibration curve (∆R) were of known historical age; calculated using radiocarbon ages on live- collected marine shell samples of known 2. radiocarbon dating shell/charcoal historical age (Stuiver et al. 1986: Table 1). paired samples from high integrity ∆R is the difference between the archaeological contexts that are conventional radiocarbon age of a sample assumed to be contemporaneous; of known age from a specific locality (P) and the equivalent age predicted by the 3. radiocarbon dating and/or paired global modelled marine calibration curve radiocarbon and uranium-thorium (Q); therefore ∆R= P-Q (Stuiver et al. 1986: (230Th/234U) dating of live corals or 982). long-lived live shells with clear annual growth bands. Once calculated, the ∆R value can be applied to marine calibration curves to In the first method, marine shell specimens calibrate dates obtained on marine shell of known historical age must be live- (and other marine-derived sample materials collected prior to AD 1955 and the date and such as fish bone, marine mammal bone location of live-collection known with etc.) for specific regions. The ∆R value can confidence. After AD 1955 natural levels of also be used in widely available computer 14C activity in marine environments were calibration programs such as CALIB enriched as a result of detonation of nuclear (Stuiver and Reimer 1993) and OxCAL and thermonuclear weapons in the (Bronk Ramsey 1995). atmosphere. Dating shell/charcoal paired samples is potentially problematic because Australian ∆R Values: A Guide it must be assumed that the samples selected are contemporaneous and that Reimer and Reimer (2000, 2001) recently association is not simply the result of post- launched a world wide web database of ∆R depositional processes, excavation values (available at http://calib.org/marine). procedures or erroneous interpretation. In the Marine Reservoir Correction Therefore such data need to be considered Database Reimer and Reimer (2000, 2001) carefully on an individual basis. Owing to recalculated world-wide ∆R values using these problems ∆R values calculated from the latest calibration dataset of Stuiver et al. shell/charcoal pairs are not included in this (1998). For Australia, Reimer and Reimer guide (or in the Reimer and Reimer (2000) present pooled regional ∆R values database). Recent studies have for Northeast, North and Northwestern, demonstrated that radiometric dating of Southern and Southeastern Australia 11 AACAI NEWSLETTER NO. 89 APRIL 2002 (Figure 1, Table 1). These regional values the regional values, the subregional ∆R combine between two and 11 individual ∆R value is the error-weighted mean of the values and cover very broad geographical values available. Error-weighted means regions composed of potentially different were calculated using the procedures and marine reservoir conditions. Therefore, in methods outlined by Ward and Wilson addition to the regional ∆R values I have (1978), with all sample groups presented calculated subregional ∆R values where forming statistically indistinguishable two or more individual ∆R values are groups. available for a specific area (Table 2). Like Figure 1. Map of Australia, showing regional and subregional ∆R values. ∆R values in bold denote regional values. Those without bold are subregional values. Table 1. Regional average ∆R values (after Reimer and Reimer 2000; Ulm 2002). # ∆R Regional Region Values Average Northeast Australia 12 12±7 North & Northwestern Australia 9 64±24 Southern Australia 11 61±29 Southeastern Australia 2 -1±70 Table 2. Subregional average ∆R values (after Reimer and Reimer 2000; Ulm 2002). 12 AACAI NEWSLETTER NO. 89 APRIL 2002 # ∆R Subregional Subregion Values Average Torres Strait 3 49±45 Gulf of Carpentaria 2 52±42 Kimberley Region 7 74±34 Southwest Western Australia 4 66±46 Spencer Gulf 5 58±47 Gulf of St Vincent 2 55±60 Central Queensland 7 10±7 References Discussion Bronk Ramsey, C. 1995. Radiocarbon calibration and The choice of a particular ∆R value to analysis of stratigraphy: The OxCal program. Radiocarbon 37(2): 425–430. calibrate a particular radiocarbon date must Gillespie, R. 1975. The suitability of marine shells for be based on a consideration of the radiocarbon dating. Unpublished PhD thesis, environment in which the sample material University of Sydney, Sydney. was formed. For example, in a study of ∆R Gillespie, R. and H.A. Polach, 1979. The suitability of marine shells for radiocarbon dating of Australian values for a number of estuaries in central prehistory. In R. Berger and H. Suess (eds), Queensland I recently calculated estuary- Proceedings of the Ninth International Conference specific values of up to ∆R= -155 ± 55 (see on Radiocarbon Dating, pp. 404–421. Los Ulm 2002 for detailed discussion). In this Angeles: University of California Press. case, the blanket application of the regional Gillespie, R. and R.B. Temple, 1977. Radiocarbon dating of shell middens. Archaeology and or subregional ∆R value would have Physical Anthropology in Oceania 12: 26–37. produced calibrated ages approximately Higham, T.F.G. and A.G. Hogg, 1995. Radiocarbon 200 years too young. dating of prehistoric shell from New Zealand and calculation of the ∆R value using fish otoliths. In the absence of additional information, it Radiocarbon 37: 409–416. Hughes, P.J. and V. Djohadze, 1980. Radiocarbon is assumed that temporal changes in ∆R for Dates from Archaeological Sites on the South a specific region coincide with changes in Coast of New South Wales and the Use of the global model ocean (Stuiver et al. 1998: Depth/Age Curves. Occasional Papers in 1135). Time-factored ∆R(t) (t=time) values Prehistory 1. Canberra: Department of Prehistory, Australian National University. can be calculated through large-scale Ingram, B.L. 1998. Differences in radiocarbon age studies of annual coral records and/or between