Abstract
Coral geochemical archives provide high-resolution records of past environmental variability, but accurate salinity reconstructions require localized calibrations that are rarely paired with overlapping in-situ observations. The Flower Garden Banks National Marine Sanctuary (FGBNMS) offers a unique opportunity to test these methods because long-term coral growth coincides with continuous, site-specific environmental monitoring.
This study uses paired δ18O and Sr/Ca measurements from a Siderastrea siderea coral to reconstruct sea surface salinity (SSS) from 2002–2022 and evaluate proxy performance using daily in-situ sea surface temperature (SST) and SSS records. Local calibrations were developed from these overlapping datasets, enabling direct comparison between modeled and observed environmental signals.
Despite precise sampling and direct in-situ calibration, the proxy reconstruction did not reproduce observed SSS behavior. Modeled SSS exhibited large mean offsets, exaggerated amplitudes, and loss of the characteristic seasonal frequency pattern observed in the in-situ record. Rigorous uncertainty propagation could not account for these significant differences. Evaluation of methods steps found that strong seasonal SST variability dominated the coral δ18O signal, producing a residual thermal overprint that obscured the comparatively small salinity-driven component.
These results demonstrate that coral-based SSS reconstruction at the FGBNMS is limited by the region’s environmental regime: large seasonal temperature swings and minimal salinity variability. Such conditions allow SST to overwhelm salinity signals in the coral geochemistry, reducing the efficacy of standard δ18O-Sr/Ca methods. Regions with more favorable conditions, such as strong salinity gradients paired with muted temperature variability, may produce more reliable calibrations for future coral-based paleosalinity studies.