Assessing coral Sr/Ca–SST calibration techniques using the species Diploria strigosa

Abstract

The coral species Diploria strigosa is a promising archive for paleoclimate reconstruction in the tropical Atlantic, but limited work has been accomplished thus far to quantify the relationship between strontium to calcium ratios and sea surface temperatures in this species. In this study, we collected three modern D. strigosa coral colonies from Anegada, British Virgin Islands to refine the use of this species as a paleoclimate archive. Three regression techniques including ordinary least squares, reduced major axis, and weighted least squares were used to assess how assumptions implicit in these regression methods influence the calibration. Our analysis demonstrated that regression slopes were strongly impacted by the assumptions about data uncertainty. We recommend using the unbiased weighted least squares regression technique for calibrating the Sr/Ca paleothermometer, especially when precise local sea surface temperature data are unavailable. Combining our Sr/Ca data with previously published data from this species results in a multi-site, multi-colony temperature sensitivity of −0.046(±<0.001) mmol/mol/°C for mean-centered monthly data, −0.059(±0.001) mmol/mol/°C for monthly anomalies, and −0.063(±0.004) for annual anomalies. We suggest these relationships are the best available characterization of the temperature dependence of Sr/Ca in D. strigosa, and can be utilized in future paleoclimate reconstructions.