Stable isotopic signature of El Ni�o-Southern Oscillation events in eastern tropical Pacific reef corals

Abstract

In this study we measured stable oxygen and carbon isotope ratios in the skeletons of massive reef-building corals (Porites lobata and Pavona gigantea) at four widespread locations in the tropical eastern Pacific, each with a unique marine climate. Annual variation in sea surface temperature (SST) varied from 5–7°C at upwelling sites (Gulf of Panama and Galapagos Islands) to 2–3°C where upwelling was absent [Cano Isand (Costa Rica) and Gulf of Chiriqui (Panama)]. Annual range in salinity was high in the gulfs of Panama and Chiriqui (up to 15‰) and low at Cano Island and Galapagos (2–3‰). We characterize the isotopic signatures of recent (15–40 y long records) El Nino/Southern Oscillation (ENSO) events, particularly the very severe 1982–1983 event. Subannual sampling at 1–2 month resolution reveals that the δ18O signal at Cano Island records strong to very strong ENSOs. In the Gulf of Chiriqui, this signal is governed mainly (80%) by salinity and thus is a weak recorder of ENSO events: only the 1951–1952, 1957–1958 and 1972–1973 events appear as significant δ18O anomalies over the period 1940–1984. In the Gulf of Panama, high variation in both SST and salinity contribute to the δ18O signal. ENSO events at this site are poorly recorded, probably due to ameliorating effects of cool upwelled water during the early stage of the ENSO event. The δ18O record in Galapagos, however, shows a strong correlation with SST and accurately records all but the most severe 1982–1983 ENSO event. Thus, ENSOs are most clearly recorded at sites where salinity variation is minimal. At Cano only strong ENSOs are recorded while mild to strong events appear in the Galapagos record. Nowhere did the δ18O signal accurately record the full range of temperatures that occurred during the 1982–1983 ENSO; however, a stress band on the coral skeleton was evident at all sites. By comparing the δ18O records and skeletal features across sites it may be possible to identify the occurrence of strong to very severe ENSOs prior to instrumental records. The relationship between δ18O and δ13C was examined and found to be significantly positive and in phase (i.e., either depleted or enriched at the same time) at three of the four sites studied. Weak correlations at some sites may be explained by high variability in water column clarity leading to depleted δ13C when waters are cool and surface isolation high.