RESEARCH | REPORTS PALEOCEANOGRAPHY vergent results and our newly generated data by considering geographic location, choice of fora- minifera species, and changes in thermocline – depth (see supplementary materials). Reduced El Niño Southern Oscillation ENSO variability is asymmetric (the El Niño warm phase is more extreme than the La Niña during the Last Glacial Maximum cold phase) (14), so temperature variations in the equatorial Pacific are not normally distrib- Heather L. Ford,1,2* A. Christina Ravelo,1 Pratigya J. Polissar2 uted (7, 15), and statistical tests that assume normality (e.g., standard deviation) can lead to El Niño–Southern Oscillation (ENSO) is a major source of global interannual variability, but erroneous conclusions with respect to changes its response to climate change is uncertain. Paleoclimate records from the Last Glacial in variance. Therefore, we use quantile-quantile Maximum (LGM) provide insight into ENSO behavior when global boundary conditions (Q-Q) plots—a simple, yet powerful way to vi- — (ice sheet extent, atmospheric partial pressure of CO2) were different from those today. sualize distribution data to compare the tem- In this work, we reconstruct LGM temperature variability at equatorial Pacific sites perature range and distribution recorded by two using measurements of individual planktonic foraminifera shells. A deep equatorial populations of individual foraminifera shells to thermocline altered the dynamics in the eastern equatorial cold tongue, resulting in interpret possible climate forcing mechanisms. reduced ENSO variability during the LGM compared to the Late Holocene. These results Sensitivity studies using modern hydrographic suggest that ENSO was not tied directly to the east-west temperature gradient, as data show how changes in ENSO and seasonality previously suggested. Rather, the thermocline of the eastern equatorial Pacific played a modify temperature distributions that can be di- decisive role in the ENSO response to LGM climate. agnosed as changed slopes on Q-Q plots (Fig. 2) (also see supplementary materials). For our loca- he equatorial Pacific mean climate state nifera in a sediment sample to quantify tropical tions, the sensitivity studies indicate that season- (average oceanic and atmospheric proper- variability during the late Holocene (<6000 years ality weakly affects the temperature distributions, ties across the basin) is characterized by a ago) and LGM. This distribution has been used whereas ENSO has a large influence (Fig. 2); con- T strong east-west sea surface temperature to accurately reconstruct the mean and seasonal sequently, changes in temperature distribution (SST) gradient that is tightly coupled to variability at several locations, and our sample between the late Holocene and LGM can be at- the thermocline and the winds that drive warm size (40 to 70 individuals) is sufficient to cap- tributed to changes in ENSO. on January 20, 2015 water to the west and cause cold water to up- ture climate variability at our study sites (10). In the WEP, our data show that LGM surface well in the east (1). Because wind strength and Three prior individual foraminifera studies of and subsurface temperatures were cooler by ~2.3° the zonal SST gradient are mutually dependent, the LGM suggest either increased [site V21-30 (11); to 2.4°C compared with the late Holocene (Fig. 3). perturbationsinthisocean-atmospherelinkini- site CD38-17P (12)] or decreased [site MD02-2529 Cooler surface waters are consistent with ad- tiate and propagate El Niño–Southern Oscilla- (13)] ENSO variability compared with the late justment to reduced PCO2 forcing (16). Cooler tion (ENSO) events (2). Theoretical and modeling Holocene. We synthesized these apparently di- subsurface temperatures could be interpreted studies suggest that the mean state should strongly affect ENSO (3–5) by altering the balance of sev- eral positive and negative ocean-atmosphere feed- www.sciencemag.org C 27 20°N 7 C 29 backs that determine ENSO behavior (2, 3, 6, 7). 228C 28 10°N 27 However, climate models disagree on how these Site 806 Site 849 3 26 feedbacks interact when the mean state changes 0° 25 21 24 (8). Here we examine climate variability during 23 the Last Glacial Maximum (LGM) (~20,000 years 10°S 1 22 21 ago) as an opportunity to investigate ENSO behav- 20°S 20 ior during an altered mean state when ice sheets 120°E 180° 120°W 60°W Mean Annual SST (°C) covered North America and partial pressure of Downloaded from 10°N 28C CO2 (PCO2) levels were ~100 parts per million lower 28C −2.7 EPWP 28C than during preindustrial times (9). 10°N −2.0 −2.3 We use deep-sea sediment samples from Ocean WEP -2.0 27C Drilling Program (ODP) site 806 in the western −2.8 −3.5 26C 29C−2.8 equatorial Pacific (WEP) warm pool and ODP site −2.4 −1.6 25C −1.6 −1.0 0° 0° 24C 849 in the eastern equatorial Pacific (EEP) cold −2.4 ~0 −0.9 −2.0 28 COLD TONGUE tongue to examine tropical variability during −3.0 C 28C −1.6 23C discrete time intervals (Fig. 1). Site 806 is located 4C 2 2 2 10°S C 21C in the heart of the warm pool on the equator, 28C −2.8 27C where SST variability is small and primarily at 27C 10°S interannual-to-decadal frequencies. In contrast, 120°E 130°E 140°E 150°E 160°E 100°W 90°W 80°W 70°W site 849 is located in the EEP cold tongue exten- Fig. 1. Reconstruction of the zonal Holocene-LGM temperature gradient. (A) Ocean Drilling Program sion, where variability is large and dominated sites 806 and 849 indicated on a mean annual SSTmap, using the Met Office Hadley Centre’sHadISST1.1 equally by seasonal and canonical ENSO frequen- data set (32). Numbers indicate individual foraminifera study sites: V21-30 (1), CD38-17P (2), and MD02- cies (2). Here we use the distribution of Mg/Ca- 2529 (3). Inset maps of (B)WEPand(C) EEP show anomaly of published LGM [18 to 20 thousand years based temperatures measured on individual ago (ka)] minus Holocene (4 to 6 ka) temperatures reconstructing the zonal SSTgradient. Locations using shells of surface- (Globigerinoides sacculifer)and Mg/Ca proxy are denoted by circles, whereas locations using the alkenone proxy are denoted with stars. subsurface-dwelling (Globorotalia tumida)forami- Generally, there was a reduced zonal temperature across the Pacific, though there is some spatial hetero- geneity in the EEP.The WEP and EPWP have similar cooling magnitudes (~2.7° and ~2.3°C, respectively), 1 Ocean Sciences Department, University of California, Santa which suggests that LGM cooling was largely dominated by radiative cooling. In contrast, the EEP cold Cruz, CA 95064, USA. 2Biology and Paleo Environment, Lamont-Doherty Earth Observatory, Palisades, NY 10964, USA. tongue region cooled less (~1.6°C) during the LGM, indicating that radiative cooling was partially com- *Corresponding author. E-mail: [email protected] pensated by dynamic components (a deep thermocline and reduced upwelling). SCIENCE sciencemag.org 16 JANUARY 2015 • VOL 347 ISSUE 6219 255 RESEARCH | REPORTS as a colder or shallower thermocline (17). How- and our Q-Q reanalysis of these data (fig. S13) dients (26). Although subsurface variability is a ever, because modeling (18), faunal (19), and geo- suggests that the high variance reported at V21-30 strong indicator of ENSO in today’socean(6, 7), chemical (20) studies suggest that the thermocline (11) reflects enhanced seasonality during the comparing the magnitude of subsurface varia- was actually deeper during the LGM in compar- LGMincomparisontothelateHolocene.Together, bility from foraminifera proxies may not be the ison with today, our cool subsurface tempera- these data indicate that ENSO was reduced and best indicator of ENSO strength when there tures indicate a colder thermocline. Equatorial seasonality was enhanced during the LGM in are substantial changes in thermocline depth. thermocline waters originate in mid-latitude re- comparison with today, which is corroborated At site CD38-17P (east of site 849 and south gions where they subduct, move equatorward, by a recent modeling study (23). Some climate of V21-30) (Fig. 1), Neogloboquadrina dutertrei and upwell along the equator (21). Our G. tumida models indicate a strong inverse relationship show no change in mean temperature but greater subsurface temperatures suggest that these mid- between the amplitude of the seasonal cycle and subsurface temperature variability during the latitude thermocline source water regions were ENSO (3); that is, when the seasonal cycle is strong, LGMincomparisonwiththeLateHolocene(12). cooler during the LGM, also probably as a response ENSO is weak. Our interpretation suggests that The lack of a change in mean temperature agrees to PCO2 forcing. Surface and subsurface cooling in this inverse relationship may be a robust behavior with our G. tumida observation, and we similarly the WEP occurred without changes in variability, over long time scales. suggest that N. dutertrei must have also occupied consistent with PCO2 radiative forcing as the most The EEP G. tumida subsurface temperatures a shallower portion of the thermocline during likely agent of WEP temperature change (16). exhibit no change in mean temperature but do the LGM in comparison to today. This change in In the EEP, our data from site 849 show that show an increase in variability (Fig. 4). Although habitat explains the apparent increase in varia- average surface temperatures were only ~1.2° an increase in thermocline variability is consist- bility, and consequently, N. dutertrei from site to 1.3°C cooler during the LGM compared with ent with enhanced ENSO (Fig. 2E), this signal is CD38-17P do not support increased ENSO varia- the Holocene and that SST variability was re- conflated with a deepening of the EEP cold tongue bility during the LGM (figs.