Characterizing Western Atlantic SST and Ocean Variability: Replicating and Extending a ~440-yr-long Coral-based SST Reconstruction at Bahamas.


OCCI Funded Project: 2009


Northern hemisphere (NH) reconstructions suggest that the late 20th century was the warmest time of the last 500 years, and, with decreasing confidence, of the last 1,300 years1,2. Global temperature reconstructions have greater uncertainties because there are fewer data sets from the southern hemisphere and low latitudes. Moreover, these reconstructions are based largely on terrestrial records from extra-tropical or high elevation sites. Temperature changes of the earth’s surface, however, most closely follow those of the global tropics, which are ~75% ocean. Improving global temperature reconstructions and reducing uncertainties, therefore, requires the development of well-replicated, well-dated, multi-century long records from the low-latitude oceans. With such records, we can provide a better context for understanding recent warming trends, provide better estimates of the temperature response to natural (volcanic/solar) and anthropogenic (greenhouse gas and aerosol) external forcing, and provide a better assessment of the persistence of internal modes of climate variability with changing background climate.

One such mode is the Atlantic Multidecadal Oscillation (AMO), which appears to be a basin-wide, quasi-periodic (~65-80 years) oscillation of North Atlantic SSTs. Model simulations suggest the AMO may be a persistent mode of internal ocean variability associated with the Atlantic Meridional Overturning Circulation3. However, the brief ~150 year instrumental record makes assessing SST variability on these multidecadal timescales highly uncertain. Although a coral-based 18O reconstruction extending to ~ AD 1750 suggests that AMO-associated atmospheric variability may have been a persistent feature of the last 250 years4, that record diverges significantly from tree-ring based estimates of the AMO prior to AD 18005. We have recently published 439-year long coral based SST record6. It is derived from annual coral skeletal extension rates estimated from computer assisted tomography (CAT) scanning. As discussed in the proposal text, our results suggest that unforced multidecadal variability may have been absent from ~ AD 1550 - 1750, a finding that has implications for the persistence of the AMO, and one that we can evaluate further with new coral records.

The main goal of this proposal is to core and analyze a 6-ft tall coral colony we have identified at the Bahamas, which we conservatively estimate be ~500 years old or longer, as well as multiple new shorter cores (70-200 yrs) from massive, slow-growing, long-lived corals (all S. siderea) from the same area. We will use SST estimates from the new coral colony to replicate our published record.  While not as exciting to us as beginning work on coral colonies from other regions, we believe that replicating the long coral record will be a key outcome that will help us attain NSF funding to generate multiple, multi-century length, annually resolved records at several sites across the tropical/subtropical Atlantic. Furthermore, new Bahamas coral colony is approximately twice the size of our analyzed coral, and may actually be close to 900 years old, doubling the length of our published reconstruction.

Although our method of estimating SST does not resolve the seasonal cycle, one advantage it has over geochemical methods is that CAT scanning is relatively rapid and inexpensive compared to geochemical techniques. As a result, we can generate many records of varying lengths from multiple colonies at each site, and, using techniques applied in dendrochronology, provide realistic error estimates on our reconstructed SSTs.

Our research is motivated by a series of questions such as: How much have low-latitude SSTs changed since the Little Ice Age? Have low-latitude SSTs ever been as warm as today’s during the last five centuries? What is the response of low-latitude SST to external forcing? Does background climate influence decadal-multidecadal internal variability? Are interactions between climate modes stationary? This proposal represents the first step in securing funds to address these and related questions.