An intractable climate archive — Sclerochronological and shell oxygen isotope analyses of the Pacific geoduck, Panopea abrupta (bivalve mollusk) from Protection Island (Washington State, USA)

Annual growth increment patterns of cardinal teeth (CT) of Panopea abrupta (Conrad) can reportedly provide information about past climate variations. However, little is known about the intra-annual timing and rate of shell growth necessary to interpret such records. In addition, it remains unclear whether actual temperatures can be reliably inferred from δ18O values of geoduck {gooʹe-duk} shells. This study compared high-resolution environmental records (hourly to monthly resolved temperature, bi-weekly to monthly δ18Owater and salinity data) with temperatures reconstructed from oxygen isotope values of the outer shell layer (Tδ18OOSL) and cardinal tooth portions (Tδ18OCT) of different contemporaneous specimens alive at the same locality. Results indicate that shell growth mainly occurred between March/April and November/December with a maximum during May–August. This finding must be considered when comparing the “annual” growth increment width chronologies to environmental parameters. In addition, intra-annual δ18Oshell values require the calculation of weighted averages instead of arithmetic means. During ontogeny, the duration of the growing season remained nearly unchanged; an important finding for paleoclimate studies based on inter-annual growth patterns. Seasonal shell growth was strongly correlated with temperature (R = 0.93, R2 = 0.86, p < 0.0001). Presumably due to individual differences in the exchange rate between the extrapallial fluid (EPF) and the ambient water, the outer shell layer of some specimens formed out of oxygen isotopic equilibrium, particularly during summer (high growth rates, increased 18O depletion of the EPF). This resulted in a Tδ18OOSL difference of up to 2 °C among different specimens. In addition, a bias was observed in different specimens toward daytime or nighttime temperatures, particularly during summer. Such a bias may be related to individual differences in the physiological activity at ultradian time-scales or to elevated predation pressure. More importantly, CT portions (= inner shell layer) formed in isotopic disequilibrium with the ambient water. Typically, reconstructed temperatures differed by more than 3–4 °C from actual water temperatures. Within specimens, Tδ18OOSL and Tδ18OCT were offset by ca. 2 °C. Some Tδ18OCT also exhibited unexplained inter-annual trends, so that Tδ18OCT among specimens varied by up to 4 °C. Given the δ18Oshell inconsistency between and among shells, a small seasonal temperature amplitude barely exceeding 4 °C and the error bars of Tδ18O of geoducks at this setting on the order of ± 2 °C (error bars of the paleothermometry equation + variable δ18Owater values + precision error of the mass spectrometer), the geochemical record of a single P. abrupta may not serve as a suitable paleoclimate archive. A reliable approximation to paleotemperatures may only be achieved by exclusively sampling the outer shell layer of multiple contemporaneous specimens, so that the Tδ18OOSL variance among shells can be quantified.