Assessing the reproducibility and reliability of estuarine bivalve shells (Saxidomus giganteus) for sea surface temperature reconstruction:: Implications for paleoclimate studies

Studies using oxygen isotopes (delta O-18) of mollusk shells to determine paleotemperature need to assume water delta O-18 values, which could severely influence calculated temperatures. We analyzed aragonitic shells of the Butter Clam, Saxidomus giganteus (DeShayes, 1839), to determine the reproducibility of the isotopic signal between individuals and to assess how precisely temperature could be calculated given known salinity and temperature. Furthermore, carbon isotopes are also investigated as an environmental proxy. The abundance of well-preserved S. giganteus shells in archeological and geological deposits in northwestern North America makes them a particularly suitable species for paleoclimate studies. Seasonally resolved stable oxygen isotope profiles in three S. giganteus shells collected from the same site in Puget Sound (Washington, USA) were well correlated (0.77 < R-2 < 0.87). Although there were differences up to 0.5 parts per thousand in high resolution delta O-18 profiles of the three shells, the difference between the average delta O-18 of each shell was less than half of this (0.19 parts per thousand) and half of what has been reported for between-colony coral variability. delta C-13 profiles on the other hand were more complex, with shell delta C-13 being about 2.5 parts per thousand lower than expected equilibrium values. However, this roughly conforms to the idea that about 10% of the shell carbon originates from metabolic CO2. Both delta O-18 and delta C-13 indicate that S. giganteus do not grow during periods of reduced salinity. Despite the excellent reproducibility of delta O-18 between shells, and the fact that salinity effects were duly considered, calculated temperature still differed from instrumental temperature. Applying different salinity-delta O-18 water relationships to average shell delta O-18, and considering salinity from the shell collection site and a nearby offshore station resulted in calculated average water temperatures ranging from 1.7 to 6.4 degrees C warmer than measured. Although we could not determine if S. giganteus precipitate their shells in isotopic equilibrium, we believe that the difficulty in predicting temperature arose from not being able to accurately determine delta O-18 of the water at the time of shell precipitation. These data highlight the difficulties inherent to using stable isotope profiles of estuarine biogenic carbonates as environmental proxies. (c) 2005 Elsevier B.V. All rights reserved.