Assessment of Simulated and Aquarius-Observed Sea Surface Salinity Using Salinity Measurements in Dry Tortugas National Park, Florida, U.S.A.: Application for Pseudocoral δ18O

Nazanin Chaichitehrani, Louisiana State University, Oceanography&Coastal Sciences, Baton Rouge, LA, United States and Kristine L DeLong, Louisiana State University, Geography & Anthropology, Baton Rouge, LA, United States
Abstract:
Salinity plays a critical role in ocean physics thus is a target for paleoclimatologic and paleoceanographic reconstruction. Here we assess the quality of space-borne sea surface salinity (SSS) determinations and simulated SSS versus SSS measurements from an open ocean coral reef site, Dry Tortugas National Park (DTNP). The oxygen isotopic composition of seawater (δ18Osw) is related to SSS, thus SSS can be used to understand δ18Osw variability when measurements of δ18Osw are sparse. In marine carbonates such as corals, δ18Ocoral varies with temperature and δ18Osw creating a bivariate system, which is difficult to calibrate with two variables. Accurate determinations of SSS from satellites and simulations can be substituted for local SSS, converted to δ18Osw, in bivariate forward models to estimate δ18Ocoral or pseudocoral thus improving calibrations of δ18Ocoral for locations and time intervals without in situ observations. Monthly and daily Aquarius-retrieved SSS data Level-3 (Official Version 3.0) with spatial resolution are compared with local SSS in DTNP obtained from Water Quality Monitoring Project for the Florida Keys National Marine Sanctuary and southwest Florida shelf, which includes DTNP, for the concurrent interval from 2011–2014. Our statistical analysis shows a satisfactory agreement between daily Aquarius SSS and local SSS (r2=0.68; RMSE=0.24 psu). Additional SSS data are obtained from the National Data Buoy Center DRYF1 station in DTNP. Monthly-simulated SSS (Global Ocean Physics Reanalysis GLORYS2V3) obtained from the MyOcean WebPortal compares relatively well with DRYF1 monthly SSS data (r2=0.68; RMSE=0.35 psu) for the earlier interval from 1998–2002. Our analysis indicates that Aquarius-retrieved and simulated SSS can be utilized as a substitute for local SSS in bivariate forward models to calculate pseudocoral δ18Ocoral as well as forward models other marine carbonates for locations without SSS observations.