Do three massive coral species from the same reef record the same SST signal? A test from the Dry Tortugas, Florida Keys

Paleoclimatologists have reconstructed century-long records of sea surface temperature (SST) in the Pacific using the Sr/Ca of massive corals, whereas similar reconstructions in the Atlantic have not proceeded at the same pace. Past research in the Florida Keys has focused on Montastrea spp., an abundant and fast-growing massive coral, thus a good candidate for climate reconstructions. However, coral records from the Florida Keys are complicated by freshwater flux, which varies the Sr/Ca in seawater, thus confounding the Sr/Ca to SST signal. In this research, we compared the monthly Sr/Ca variations in three massive corals species (Montastraea faveolata, Diploria strigosa, and Siderastrea siderea) from the same reef in the nearly pristine Dry Tortugas National Park (24.70N, 82.80W) at the southwestern extent of the Florida Keys. This location is ideal for a calibration study as hourly water temperature records are available and the remote reef is far from mainland freshwater influence. These corals experienced the same environmental conditions (water depth, clarity, Sr/Ca of seawater, etc.) but differ in the mean annual growth rates (0.86 ±0.10 (1σ) cm/year M. faveolata; 0.67 ±0.04 (1σ) cm/year D. strigosa; 0.44 ±0.04 (1σ) cm/year S. siderea). The mean Sr/Ca values are not the same but decrease with mean annual growth rates (9.201 ±0.091 (1σ) mmol/mol M. faveolata; 9.177 ±0.081 (1σ) mmol/mol D. strigosa; 8.964 ±0.12 (1σ) mmol/mol S. siderea), thus supporting the “vital effect” or biological differences during calcification between coral species. The amplitude of the seasonal cycle in Sr/Ca varies with the slower growing S. sidereahaving the largest mean amplitude and D. strigosa the smallest (0.340 mmol/mol S. siderea; 0.284 mmol/mol M. faveolata; 0.238 mmol/mol D. strigosa). We confirmed our sampling methods by conducting several intracolony and intercolony coral Sr/Ca replication tests and found a high correlation in all tests (>0.95 S. siderea; >0.90 D. strigosa; >0.83 M. faveolata; p < 0.05). The weighted linear regression of monthly coral Sr/Ca to mean monthly SST revealed that S. sidereacaptured the seasonal and interannual variability in SST (r = -0.97, -0.61 monthly and monthly anomalies, respectively, p < 0.05). The other corals have reduced correlation with monthly anomalies and do not capture the seasonal variability with the same fidelity as S. siderea. All three corals were sampled along the thecal wall following the same procedures; however, each coral species has a different skeletal structure, density, and micro-scale growth patterns. We hypothesize the thecal wall of S. siderea calcifies at a continuous rate along the time-growth axis whereas the wall of D. strigosa and M. faveolata reflects a more complex signal. Of the three species, the slow growing S. siderea provides a robust reconstruction of mean monthly SST for the Dry Tortugas thus suitable for longer centennial-scale reconstructions.