CRP El Niño member
Mr. Robert Dunbar
Stanford University
Stanford, USA
Scientific Background
Our group has been studying interannual to centennial climate variability using high resolution marine archives for ~20 years. We have focused on reconstructing the history of tropical sea surface temperatures using the skeletons of reef-building corals, and to a lesser extent, on atmospheric and oceanic variability recorded in lacustrine or continental margin sediments.
Dunbar first began analyzing corals in the late 1970's. At this time he conducted the first stable isotopic studies of a branching coral, using specimens collected from the Gulf of Panama in the Pacific, an ENSO sensitive site, as well as fossil specimens from the west coast of Costa Rica. This work led to several new projects focusing on ENSO variability. We switched our archive from branching corals to massive corals, in particular to specimens in the genera Porites and Pavona, in order to take advantage of multi-century growth records. We developed chronologies extending back 300 years at Secas Island, Panama, and nearly 400 years at Urvina Bay, Galapagos Islands. The Secas record includes near-monthly stable isotopic analyses that show pronounced annual and interannual variability associated with the inter-tropical convergence zone (ITCZ). The ITCZ migrates northward and southward in a highly regular fashion, producing Panama's dry and wet seasons. The pattern is altered during strong ENSO warm mode events, when the ITCZ remains locked at a lower latitude than normal, producing drought in part of Panama and Costa Rica. The Secas record revealed significant ENSO variability as well as interdecadal variability, the origin of which remains unknown.
Our work in the Galapagos revealed that 90% of the coral oxygen isotopic signal was generated by sea surface temperature anomalies (SSTa). At this site SSTa are primarily controlled by the state of the ENSO cycle. A 380 year annually-resoled isotopic record shows significant decade-to-decade and century-to-century variability in the character of ENSO. In particular, we observed an increase in the mean return time for El Niņo events going back through time, from about 3.5 years during the most recent decades, to about 6 years during the 1600's. The Galapagos record remains one of the longest continuous coral paleoclimate records from the tropical Pacific Ocean.
After developing some long records of Pacific climate variability, we turned our attention to signal-to-noise and reproducibility issues. We conducted detailed in-situ calibration experiment in the Galapagos Islands, using a series of marked corals, in-situ water temperature and salinity recorders, and hand-collected seawater samples. This work serves to demonstrate the sensitivity of coral isotopic composition to actual measured environmental changes. We also conducted a replication study at Clipperton Island in the eastern Pacific. To assess the use of corals in the development of paleoclimatic reconstructions in this region, we developed oxygen isotope time series from multiple specimens of the massive coral Porites lobata. Six near-monthly isotopic records from different sized (age) colonies where produced for the interval 1986-1994. Our results show that the average oxygen isotopic disequilibrium offset (vital effect) from equilibrium seawater composition for individual corals can vary by up to 0.4 ppmil. However our results also suggest that the vital effect offset is constant over time. Similar offsets are observed in the tops of old (age >100 y) and young (age <10 y) colonies, further suggesting that the biologically mediated vital effect offset does not change as a Porites colony ages. All El Niņo events of the past 100 years appear to be recorded by coral skeletal isotopic composition at this site.
In the mid-1990's we expanded our region of interest to include the Indian Ocean, in part to examine ENSO teleconnections insofar as they influence the Indian Ocean monsoon systems. In particular, we found that A 194-year annual record of skeletal oxygen isotopic composition from a coral growing at Malindi, Kenya, preserves a history of sea surface temperature (SST) change that is coherent with instrumental and proxy records of tropical Pacific climate variability over interannual to decadal periods. This variability is superimposed on a warming of as much as 1.3 degrees C since the early 1800s. These results suggest that the tropical Pacific imparts substantial decadal climate variability to the western Indian Ocean and, by implication, may force decadal variability in other regions with strong El Niņo/Southern Oscillation teleconnections.
We have also begun examining the origin of climate signals in corals
at the crystal-genesis spatial scale. To do this, we are employing micro-scale analytical
techniques that have only recently become available. For example, in thermodynamic
equilibrium with sea water the Sr/Ca ratio of aragonite varies predictably with temperature
and the Sr/Ca ratio in coral have thus become a frequently used proxy for past Sea Surface
Temperature (SST). However, biological effects can offset the Sr/Ca ratio from its
equilibrium value.
We report high spatial resolution ion microprobe analyses of well defined skeletal elements
in the reef-building coral Porites lutea that reveal distinct monthly oscillations in
the Sr/Ca ratio, with an amplitude in excess of ten percent. The extreme Sr/Ca variations,
which we propose result from metabolic changes synchronous with the lunar cycle, introduce
variability in Sr/Ca measurements based on conventional sampling techniques well beyond
the analytical precision. These variations can limit the accuracy of Sr/Ca paleothermometry
by conventional sampling techniques to about 2 degrees C. Our results may help explain
the notorious difficulties involved in obtaining an accurate and consistent calibration of
the Sr/Ca vs. SST relationship. Our work in this area is continuing as we start to make use
of nanometer-scale secondary ion emission mass spectrometers.
