Fossil coral snapshots of ENSO and tropical Pacific climate over the late Holocene

Fossil coral snapshots of ENSO and tropical Pacific climate over the late Holocene Kim Cobb Georgia Inst. of Technology Chris Charles Scripps Inst of Oceanography Larry Edwards Hai Cheng University of Minnesota with thanks to: Norwegian Cruise Lines Palmyra Research Consortium NOAA

Line Islands ENSO El Niño-Southern Oscillation (December 1997 SST Anomalies) Locations of annually-resolved climate proxy records that extend to 1500A.D. ENSO-like Decadal Variability? (SST anomalies for proposed ~12-13yr pan-tropical climate variability, from Cobb et al, 2001) Zhang et al, 1997 Mantua et al, 1997 ENSO-like Global Warming?? (SST trend from 1949-1991, in degrees/decade, from Latif et al, 1997)

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 δ 18 O ƒ(sst, δ 18 O sw ) Site

ENSO and coral δ 18 O in the Central Tropical Pacific (CTP) SST and rainfall anomalies during the 1982 El Nino During El Niño events, positive SST and precipitation anomalies both contribute to negative coral δ 18 O anomalies in the CTP Interpretation of coral δ 18 O on lower frequencies relies on assumption that warm SST drives higher precipitation in the CTP, and vice versa

Climate Proxy: Coral δ 18 O 18 δ O = 18 16 18 ( O/ O) spl ( 18 16 ( O/ O) O/ std 16 O) std x1000 Palmyra Coral δ 18 O ( ) O ( ) -4.7-5.1-5.5-5.9 Drilled in May 1998 1995 Coral δ 18 O=ƒ(SST,δ 18 O SW ) 1990 Palmyra Coral Calibration Coral δ 18 O = -0.23(SST) R = -0.81 1985 1980 26 27 28 29 30 SST ( C)

What about regional-scale SST? Palmyra coral vs. NIÑO3.4 SST 3 2 R = -0.66-0.6-0.4 SST Anomoly ( C) 1 0-1 -0.2 0.0 0.2 δ 18 O ( ) SST Anomoly ( C) -2 1 0-1 -2 R = -0.84 NIÑO3.4 SST Palmyra coral 1900 1920 1940 1960 1980 2000 Year (A.D.) -0.3-0.2-0.1 0.0 0.1 0.2 0.3 δ 18 O ( )

Modern Line Islands coral δ 18 O records Christmas Island (Evans et al., 1999) Palmyra Island (Cobb et al., 2001) Fanning Island (Dunbar et al., in prep; Nurhati et al., in prep) Southern Line Islands (upcoming cruise, fall 2006, needed for resolving 20 th century tropical Pacific climate trends) R(Christmas, NIÑO3.4) = 0.90 R(Palmyra, NIÑO3.4) = 0.84

20 th century trends in the tropical Pacific How has anthropogenic CO2 forcing affected tropical Pacific climate? La Niña-like: Clement et al., 1996; Cane et al., 1997 El Niño-like: Timmermann et al., 1999 collection of evidence from tropical Pacific supports shift towards El Niño-like mean state KEY POINT: understanding how - temperature - precipitation - circulation contributed to late 20 th century coral δ 18 O trend will help interpret low-frequency fossil coral δ 18 O variability

Fossil corals collected during fieldtrips in 1998, 2000, and 2005 Beached fossil corals range from gravel-sized to 2m-diameter (~100-150yrs). rare longer cores: ENSO & decadal variability common short cores: mean climate

Number of cores 20 15 10 5 Palmyra Fanning Christmas The Line Islands Fossil Coral Collection Strategy - rare long cores (40-100y) for variability - common short cores (10-20y) for mean climate 0 20 40 60 80 100 120 Length of core (yr) Palmyra 40 cores U/Th dated 28 cores undated Christmas 18 cores U/Th dated 63 cores undated Fanning 33 cores undated 6000 5500 5000 4500 4000 3500 3000 2500 2000 1500 1000 500 0 Years B.P.

orbital forcing Medieval Warm Period African aridification Little Ice Age 6000 5500 5000 4500 4000 3500 3000 2500 2000 1500 1000 500 0 Years B.P.

How reliable are young fossil corals? young fossil U/Th dated fidelity unknown Palmyra Island Coral Collection modern absolutely dated fidelity proven Medieval Warm Period (MWP) Little Ice Age (LIA)

Testing the U/Th dating and fidelity of a 20 th century fossil coral U/Th Dating - based on decay of 238 U to 230 Th - accuracy depends on closed system behavior - analytical errors typically < 0.5% δ 18 O ( ) -5.4-5.2-5.0-4.8-4.6 U/Th date 1= 1895+20-2 U/Th date 2= 1903+14-2 Absolute date = 1915 U/Th date 3= 1917+37-2 U/Th date 4= 1920+30-2 U/Th date 5= 1925+20-2 Absolute date = 1936 1910 1920 1930 1940 Fossil NB9 Modern -4.4 Cobb et al, EPSL, 2003 Large errors associated with the presence of non-radiogenic 230 Th can be overcome with multiple U/Th dates and by dating samples of known age. Coral δ 18 O records are spliced together within dating uncertainty.

17 th Century Splice splice - absolute δ 18 O values agree (give consistent picture of mean climate) - large, well-reproduced El Niño events in mid-17th century - low-amplitude, less well-reproduced decadal variability

14 th -15 th Century Splice SB7 vs. CH9 R = 0.68 SB5 vs. CH5 R = 0.71 SB6 vs. CH5 R = 0.69-5.4 splice -5.2-5.0 δ 18 O( ) -4.8-4.6-4.4-4.2 SB7-0.06 CH9-0.06 SB5 + 0.19 SB6-0.11 CH5 + 0.04 1320 1340 1360 1380 1400 1420 1440 1460 Year (A.D.) Appreciable error in mean climate estimates from single corals

12th-13th centuries splice - 1180-1245AD: most regular ENSO period of reconstruction (5y period)

2σ error mean δ 18 O δ 18 O Anomaly ( ) -0.2 0.0 0.2 Cobb et al., 2003 Palmyra coral δ 18 O reconstruction to date 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000-5.75 "Medieval Warm Period" "Little Ice Age" 2 corals -5.50 5 corals 1 coral 3 corals 1 coral 2 corals -5.25 coming coming soon soon -5.00-4.75-4.50-4.25 6 4 2 0 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 Year A.D. δ 18 O ( ) # of El Niño events/30y What controls ENSO & tropical Pacific climate variability over the last millennium?

The Ocean Thermostat : heating La Niña-like cooling El Niño-like Can solar (heating) and volcanic (cooling) forcing shape tropical Pacific climate over the last millennium? Cane-Zebiak model w/ uniform heating winds annual mean SST, begin uniform heating SST increases in West, but SST in East is buffered by upwelling of cool waters, zonal SST gradient increases Bjerknes feedback: trades strengthen, zonal SST gradient increases Clement et al., 1996

Does ENSO respond to solar and volcanic forcing? model mean (with ensemble spread) 18: 447-456

Mann et al., 2005 updated with new data 1.2 1.0 Z-C ensemble response Palmyra coral 40y lowpass model Coral mean δ 18 O -5.5 NIÑO3 Anomaly (ºC) 0.8 0.6 0.4-5.0 δ 18 O ( ) 0.2-4.5 0.0 1000 1200 1400 1600 1800 Year (A.D.) -4.0 Large model ensemble spread still encompasses coral data, but large volcanic eruption in 1258AD a more direct test of thermostat response.

Tropical Pacific link to US West drought? Cook et al., 2004 Evidence of La Niña-like conditions from ~900-1250AD: Mono Lake lowstands (Stine et al., 1994) high Warm Pool temperatures (Sr/Ca MD81, Stott et al., 2004) decreased Peru runoff (lithic counts, Rein et al., 2004) cool Santa Barbara basin temperatures (G. bull. δ 18 O, Field et al., in prep) Question: So how important were centennial-scale changes in tropical Pacific climate over the last millennium?

New sequences from Christmas and Fanning Islands Age (kybp) 6.24 6.22 6.20 δ 18 O ( ) -5.2-5.0-4.8-4.6-4.4-4.2 6.2 kypb Fanning (4ºN) fossil coral 460-yr-old Christmas fossil coral Palmyra (6ºN) modern coral 2σ error mean δ 18 O Fanning (4ºN)? -5.2-5.0-4.8-4.6-4.4-4.2 δ 18 O ( ) -4.0-3.8 1520 1540 1560 Date (A.D.) -4.0 Christmas (2ºN) modern coral -3.8 1900 1920 1940 1960 1980 2000 Date (A.D.) 6ky: cooler, drier conditions?; reduced ENSO activity 0.5ky: statistically indistinguishable mean, ENSO variability wrt modern GOAL: A unified tropical Pacific climate reconstruction from Line Islands fossil corals?

Late Holocene trends in ENSO variance ENSO variance (% change from present) # events/100y 90 60 30 0-30 -60 25 20 15 10 5 0 7 7 6 6 5 5 4 ENSO variance changes relative to present Modelled ENSO variance 4 3 Lake Pallcacocha Moy et al., 2002 3 2 2 1 Clement et al., 1999 1 0 0 140 130 120 110 # events/500y Equally important are reconstructions of tropical Pacific mean climate through the late Holocene. Are corals up to the challenge? Reduce error bar by using many small corals? Are we up to the challenge? Tudhope et al., 2001 Woodruffe et al., 2003 McGregor et al., 2004 Correge et al., 2000 Cobb et al., 2003, in progress Age (kybp)

Conclusions 20 th century trends consistent with trend towards El Niño-like mean conditions large range of ENSO variability observed over last millennium governing mechanisms unclear cooler, drier conditions 900-1200A.D. causes unknown (solar?), regional climate responses consistent (hydrological changes) tentative evidence of cooler, drier conditions and reduced ENSO at 6kybp, but significant scatter remains in mid- to late-holocene fossil coral data new collections highlight potential to reconstruct ENSO and mean climate in central tropical Pacific quasi-continuously for last 6ky