Final Summary Report CICOR Cooperative Agreement 1998-2002
The aim of this project is to develop a proxy for the North Atlantic Oscillation (or NAO) in the skeletons of massive, long-lived reef corals. In the past two years we have sought to calibrate skeletal chemistry and structural variations in 3 small braincorals from Bermuda against the instrumental records of NAO and SST. We chose this site because it is well-situated in an area identified by Sutton and Allen (1997) for the growth of SST anomalies which subsequently propogate from the western sub-tropics to the eastern sub-polar region. The case for our initial Bermuda focus has since been strengthened by the recent successes in ALCM simulations which hindcast the observed NAO history When the observed SST is applied as a boundary condition (three independent model studies: Rodwell et a1. 1999, Hurrell, pers. comm. And Mehta, pers. comm. We chose to work on the braincoral Diploria labyrinthiformis despite the difficulties presented by its complex skeletal architecture and slow growth rate for reasons of abundance, longevity and strong annual growth bandingl.
Three small D.labyrinthiformis colonies were collected live in April 1999 at 50ft at the edge of the south-eastern Bermuda platform. We chose this specific part of the reef because of its exposure to the open ocean and least influence by local lagoonal processes. Well-defined high (HDB) and low density bands (LDB) are revealed in x-ray, the number and spacing of which indicates an average age of 40 years and an average growth rate of 3 mm/year. LDB s are approximately 3 times wider than HDBs. Two δ18O profiles were constructed by analysing samples drilled at 0.2 mm intervals along a single septatheca in the coral calyx, a sampling strategy developed in the first year of the project. Both profiles show strong annual δ18O cycles with average amplitude of 1.5 per mil, equivalent to 7.5 °C which agrees well with temperatures recorded in situ at our Bermuda sampling site. To determine the seasonality of density band formation, we used NIH image (a software program developed for use on CAT scans) to map hi h-freuency (seasonal) density changes against the δ18O profiles. HDBs coincide with high temperatures ( depleted δ18O) lndlcating that the densest skeleton (usually signifyng least calcification) is accreted during the summer months and the least dense skeleton (singnifying most calcification) is accreted during cooler months. Our observation agrees with the findings of Logan and Tomascik (1991) for this species on Bermuda even though it is unusual for corals, in general, to increase calcification in cooler water.
The consistent correlation between skeletal density and δ18O indicated that skeletal density itself may be a proxy for SST. We smoothed the density profile using a 5-point running average which essentialy decreases the "sampling" resolution from tri-monthly to annual and compared the result against actual climate data. Skeletal density, sampled at annual resolution, is significantly, positively correlated with the wintertime (J,F,M) NAO Index over the past 40 years. It is better correlated with the NAO Index than it is with wintertime SSTs for Sutton and Allen s (1997) region 3. The coral accretes a more dense skeleton when the NAO Index is positive although the reason for this is as yet unclear and requires further investigation. In the meantime results of this study indicate that low-frequency skeletal density changes in Bermuda D.labyrinthiformis are a good 1st-order proxy for NAO.
Wintertime δ18O anomalies between 1955 and 1998 were calculated using the heaviest δ18O value in each annual δ18O cycle. δ18O anomalies are also well-correlated with NAO but consistently precede NAO anomalies by approximately 3 years. Examination of the skeletal structure of the septothecae in petrographic thin-section shows that pore spaces evident when the skeleton is first accreted are subsequently filled-in over a period of several years. The effect of infilling on the climate signal would be to dampen the annual cycles and create a consistent phase-offset between the actual climate forcing and the measured geochemical response in the direction that we observe.
In summary, our investigation of the North Atlantic climate signal in Bermuda braincorals indicates that Best regards,the coral skeleton records the history of NAO in two independent ways, by changes in skeletal density and in skeletal chemistry. Accurate interpretation of both datasets requires that we understand how and when the corals accrete skeleton. We propose to utilize both tools to extract a multi-century long history of NAO activity from a 500-year old D. labyrinthiformis that we recently sampled at Bermuda.
1Abundance. Our long-range (10-year) plan is to explore past Atlantic climate variability at multiple sites chosen for their importance as SST and ocean-interaction centers. D. Iabyrinthiformis occurs throughout the tropical and sub-tropical western Atlantic. The same techniques we develop for extracting accurate climate data from Bermuda colonies can be applied to samples from other sites.
Longevity. We recently sampled a D. labyrinthiformis colony at Bermuda that stands more than one meter high. With a growth rate of 2 mm/year, we estimate this colony to be about 500 years old. Similar sized colonies occur throughout the Caribbean which increases our confidence that similarly long climate records can be constructed both in the sub-tropical and tropical Atlantic.
Growth banding. Of the three species of Atlantic braincoral , x-rays of D. labyrinthiformis reveal the strongest annual growth banding. Since annual bands are the primary method by which corals are aged and specific years identified, this feature is important when deciding on a species.