1. Fossil coral records of ENSO and “ENSO-like” variability over the last millennium Kim M. Cobb, Georgia Institute of Technology Christopher D. Charles, Scripps Institution of Oceanography R. Lawrence Edwards, Hai Cheng, UMN

2. “ENSO-like” ___________ variability Palmyra Island ENSO “El Niño-Southern Oscillation” December 1997 SST Anomalies Locations of annually-resolved climate proxy records in the tropics that extend to 1500A.D. 1) change in ENSO characteristics vs. change in tropical Pacific baseline 2) forced (sun, volcanoes) vs. intrinsic variability

3. Research Objective: To generate >100-yr-long, high-resolution, high-fidelity climate proxy records from the tropical Pacific Ocean Materials: Modern and Fossil Corals Methods: Dating: U-Th Climate proxy: Coral skeletal  18 O ƒ(SST,  18 O sw ) Site

4. ENSO and coral  18 O at Palmyra +4° +2° 0° -2° 16°N 8°N 0° 8°S 16°S Palmyra 180°150°E150°W120°W90°W SST and rainfall anomalies for the 82-83 El Niño event Color = SST anomaly (°C), contours = rain anomaly (CI 10 cm/month) El Niño (La Niña) events are recorded as negative (positive) anomalies in Palmyra coral  18 O. Central tropical Pacific warmer SST = -  18 O more precip = -  18 O

5. Most Porites corals at Palmyra grow in ~10m water depth. Large fossil Porites corals are strewn on ocean-facing beaches.

6. Modern The Palmyra Island Coral Collection Little Ice Age (LIA)Medieval Warm Period (MWP) Already run Future work

7. Calibrating the Palmyra modern coral  18 O record

8. (Evans et al., 1999) (McPhaden and Zhang., 2002) (Deser, 2000) (Mantua et al., 1997) “ENSO-like” decadal variability Palmyra Coral  18 O (Cobb et al., 2001) CTP corals instrumental records

9. 5-coral splice 3-coral spliceModern Single records Mean climate error ENSO error Decadal error 1°C Palmyra Coral Sequences Cobb et al., 2003b 2-coral splice

10. ENSO 2-7yr # El Niño events/30yr # La Niña events/30yr Low-frequency 8yr+ MWP: little change in ENSO, cooler, drier mean climate LIA: vigorous ENSO little change in mean

11. Proxy-proxy comparison for last millennium 1°C - rate, magnitude of 20th century climate change unprecedented - no simple relationship b/t tropical Pacific climate and NH temp, solar, or volcanic forcing - changes in the meridional temperature gradients implied (inc. during LIA, dec. during MWP)

12. Dai and Wigley, 2000 Building an interpretive framework for the Palmyra fossil coral  18 O Starting assumptions: The teleconnections associated with “ENSO-like” variability resemble those of ENSO. The relationship between CTP temperature anomalies and CTP  18 O sw have remained constant. Hypothesis: (MWP - “La Niña”-like, LIA - “El Niño”-like) Research Question: Is “ENSO-like” centennial-scale variability during the last millennium consistent with evidence from ENSO-sensitive proxy records?

13. “ENSO-like” MWP, LIA scorecard “ENSO-like” centennial-scale variability may have contributed to the MWP and LIA, but firmer conclusions await more rigorous testing. Research Question: Is “ENSO-like” centennial-scale variability during the last millennium consistent with evidence from ENSO-sensitive proxy records?

14. Conclusions Most intense ENSO activity occurred during 17 th century, no change in the CTP mean state. Cooler, drier conditions in the CTP characterized the early centuries of the last millennium. ENSO characteristics can change dramatically from decade to decade, absent any known forcing. No simple relationship emerges between ENSO character and NH temperature or mean coral  18 O. Changes in the tropical Pacific zonal SST gradient during the LIA (El Niño-like) and MWP (La Niña-like) are consistent with some key paleo-data, but can be more rigorously tested. Data available at: http://www.ngdc.noaa.gov/paleo/paleo.html

16. Gunung Buda Speleothem Project Goal: to reconstruct annual to centennial-scale rainfall variability from an ENSO- sensitive region over the last glacial cycle Materials: ~40 speleothems (0.1 – 1.8m long) Dating: U-series geochemistry Climate proxies: growth rate,  18 O,  13 C, Sr/Ca, Mg/Ca, optical properties Approach: 1. On-site geochemical calibration 2. Reproducibility tests 3. billions and billions of datapoints

18. 2.3yr

20. U/Th Dating Constraints for 17 th Century splice 10 years younger than best fitbest fit Assigned Date (A.D.)

21. U/Th Dating Constraints for 5-coral splice best fit 10 years younger than best fit Assigned Date (A.D.)

23. SST (°C) Palmyra Coral  18 O (‰) 1995 1990 1985 26 27282930 1980 Coral  18 O = -0.23(SST) R = -0.81 Coral  18 O=ƒ(SST,  18 O SW ) Climate Proxy: Coral  18 O Palmyra Coral Calibration Drilled in May 1998

24. (Evans et al., 1999) (McPhaden and Zhang., 2002) (Deser, 2000) (Mantua et al., 1997) Decadal Coherence Palmyra Coral  18 O (Cobb et al., 2001) CTP warmer CTP colder

28. Reproducibility of 14th-15th century  18 O records Interannual reproducibility good, decadal reproducibility lower

31. Palmyra modern coral

32. Palmyra 17 th century spliced coral

33. Palmyra 14 th century spliced coral

34. Error of fossil coral  18 O records 7 pairs of overlapping corals Interannual±0.04‰ (±0.2°C) Decadal±0.08‰ (±0.4°C) Mean climate±0.12‰ (±0.5°C) Strength of splicing approach: When overlapping corals agree, we have more confidence in the climate reconstruction.