Publication Date:
2012-07-06
Description:
EGU2008-A-04391
Paired measurements of δ18O and Sr/Ca in coral aragonite are routinely used for deriving estimates of δ18Osw and, by extension, sea surface salinity variations over the past centuries. However, in practice, the accuracy (or the error) of these estimates is often difficult to assess.
Here, we use simulated proxy data and Monte-Carlo simulations to investigate the accuracy of δ18Osw estimates from paired coral δ18O and Sr/Ca measurements. First, we estimate expected values of coral Sr/Ca and δ18O from instrumental or reanalysis data of sea surface temperature (SST) and sea surface salinity (SSS). We then add the typical analytical errors onto the expected Sr/Ca (δ18O) data as random numbers and compute δ18Osw+error from the noisy proxy data for a 1000 sample Monte Carlo. From this simple Monte Carlo simulation, the range of correlation coefficients between δ18Osw+error and expected δ18Osw is estimated. As expected, we find that this range mainly depends on the magnitude of the actual SSS variations at a given site, as well as on the slope of the δ18Osw-SSS relationship.
A comparison with real coral-based δ18Osw reconstructions from multiple sites indicates that correlations between reconstructed δ18Osw and instrumental SSS fall within the range of correlation coefficients predicted based on our Monte-Carlo simulation. Thus, our simple simulation exercise may help to assess the feasibility of δ18Osw and salinity reconstructions from corals in different climatic settings, provided that (i) some instrumental data of δ18Osw and/or SSS is available, and (ii) the slope of the
δ18Osw-SSS relationship is known.
Type:
Conference or Workshop Item
,
NonPeerReviewed
Format:
text
Permalink