Multicentury long Records of SST Variability
in the Tropical and sub-Tropical Atlantic (extracted from massive brain corals at
seasonal resolution) Anne L. Cohen and Michael S. McCartney (Woods Hole Oceanographic Institution) in collaboration with Struan R. Smith (Bermuda Biological Station for Research) Our proposed research (described below) is aligned with the long term goals of Atlantic Clivar in that we seek to (a)
Assemble quantitative proxy data sets that may be used to test, improve and
initialize models of coupled Atlantic climate variability, (b) Utilize these data
to describe and model coupled atmosphere-ocean-land interactions in the
Atlantic sector, quantify their influences on and interactions with the
regional and global climate system, and determine their predictability (e.g.,
NAO, TAV and also, Pacific ENSO) We propose to do this by way of an
Atlantic-wide "paleo"-observing system of the ocean Our plan is to extend the existing record of observed Atlantic SSTs back at least 500 years to enable exploration of variability in the ocean/atmosphere system on decadal, multidecadal and centennial timescales. The record of observed SST variability in the North Atlantic extends back through about 1860. However, observational and sampling errors and time-dependent biases so degrade the record that despite substantial effort spent minimizing these biases, considerable uncertainty remains in the dataset prior to 1942. Our proposed project responds to the goals of NOAA/Atlantic CLIVAR to observe, model and understand patterns of climate variability with a focus on large-scale recurrent patterns of variability (NAO, TAV). It also responds to recent calls for improvement in the records of past climate upon which are based the modeled predictions of future climate (for example, Smith et al. Science, 2002; Stott and Kettleborough Nature, 2002). Seasonally-resolved Atlantic SSTs over the past five centuries will be reconstructed at multiple sites selected for their importance as centers of SST variability modes and air-sea interaction. Our source of paleodata will be the aragonitic skeletons of massive, long-lived corals, specifically brain corals, which are found throughout the tropical and subtropical North Atlantic and Caribbean. The plan outlined here constitutes the third phase of a long term project to establish an Atlantic-wide Paleo-Observing System. In the first phase, a 2-year NOAA funded pilot study we showed that the skeletal chemistry and density of brain corals is sensitive to SST variability of the surrounding ocean, with especially high resolution during the cool season (September through March) (Cohen et al., 2002; Figure 1 below).
Figure 1 (left) shows our Bermuda coral density profiles 1955-1998
compared against the Station S record of wintertime SSTAs in the western
subtropical Atlantic. We are currently extending the coral record back through
1700 A.D. These corals live and accrete skeleton continuously for several hundred years. We are now extending our Bermuda proxy SST record back 320 years with a larger brain coral sampled off Bermuda. The combination of this record with independently generated proxy NAO records (Appenzeller et al. 1998, Cook et al. 1998, Glueck and Stockton 2001) will enable us to make a start at several key issues in the North Atlantic, including an assessment of low frequency variability in the associations amongst NAO, midlatitude SSTs and terrestrial temperatures over a longer time period than is possible with the available instrumental data. We are using the temperature-dependent changes in skeletal density to reconstruct the paleorecord. Density analysis is a tool with which we can reconstruct records from multiple corals at multiple sites at a third of the labor and analysis cost than is required of the chemical analyses. With Nick Bates, Alexander Amat and Struan Smith (BBSR) we continue to study the relationship between coral density and ocean temperature with a view to improving our reading of the coral proxy record. The third phase of our project planned for the next 3-5 years, is to move on to a full science program at multiple Atlantic sites. Likely sites identified using EOTs (Empirical Orthogonal Teleconnection - van den Dool et al. 2000) are Barbados, in the tropical and oppositely-signed lobe of the Atlantic SST tripole, Trinidade/Tobago in the southern subtropics and Cape Verde in the eastern North Atlantic. The western tropical North Atlantic (WTNA) is of climatic interest both in its own right and in its relation to the Bermuda site (in the oppositely signed lobe of the tripole). It is hypothesized to be an area of positive feedback to the atmosphere and is likely central to ocean-atmosphere coupled dynamics in the Atlantic sector. Within our PaleoClimate Observing system it will be a site parallel to the modern instrumental time series site given first priority in COSTA for expansion of the Pirata array into the WTNA. The WTNA site is also important for its potential relation to the NAO, for example by an atmospheric bridge from tropics to subtropics, or by some feedback from subtropical fluctuations back to the tropics, or perhaps interhemispheric linkages (for example modulated by anomalous Amazonian subsidence, Robertson, pers. comm. 1999). The WTNA also feels remote forcing tied to ENSO. Hurricane development correlates with WTNA SST. Furthermore, the WTNA is one end member defining the decadal fluctuations of cross-equatorial SST gradient (sometimes called the tropical SST dipole), whose fluctuations are linked to climate anomalies in tropical Africa and America. Teleconnected phenomenon like these of course suggest addition sites that may be developed into future phases of this long term project.
We will use the Barbados corals to produce a multicentury long record of WTNA SST which is directly comparable with that we're producing at Bermuda, and with coral-based ENSO reconstructions from the equatorial Pacific. The application of this record will be two-fold: first, comparison of the Bermuda and Barbados proxy climate records will determine the amplitude and phasing of tropical and sub-tropical SST over centuries. For example, the extraordinary large high to low NAO transition in winters 95 and 96 was coincident with extraordinarily warm tropical SSTs (and with increased hurricane activity, Landsea et al. 1999). Are these short term anomalies or a "flip" of the system from one long-lived state to another? How often do such extreme WTNA warm pool events occur? As another example, is the tropical leg of the SST tripole observed in the modern instrumental record a robust association, evidenced by oppositely signed SST anomalies in proxy data from Bermuda and Barbados? The second application will explore the connection between the western TNA and Pacific ENSO variability on interannual/interdecadal time scales. What part of the long term WTNA SST variability correlates with the Pacific ENSO region? Are there "decadal ENSO" footprints in WTNA SST variability? Can we deduce the relative contributions of ENSO correlated and NAO correlated signals in the WTNA in the paleo record and how do they compare with those of the modern instrumental record. We will then explore the suitability of other priority sites in the Atlantic - Cape Verde representing the eastern tropical North Atlantic; the Canary Islands for the eastern subtropical North Atlantic; the eastern tropical south Atlantic SST action center (the other end member for the fluctuation of cross equatorial SST gradient; Ascension Island and St Helena are possible target sites), sites which potentially index the subtropical Atlantic SST variability (there being recent literature beginning to focus on inter-hemisphere correlations of subtropical and tropical SST; the Brazilian Islands Trinidade and Martin Vaz are well-placed for this purpose, and possible the western equatorial Atlantic (various Brazilian Islands near the equator that might be useful in the search for warm/cold anomalies crossing from one hemisphere to the other). Our intention is that by the end of three years we will have multi-century long records of SST variability at key Atlantic sites including Bermuda and Barbados, a start at reconstructions in the eastern tropical North Atlantic and the southern tropical regions, and targeted additional sites to complete the Atlantic Paleo-Observing System. |
