Variability in the El Niño-Southern Oscillation Through a Glacial-Interglacial Cycle


Uplifted coral reef terraces of the Huon Peninsula, Papua New Guinea. In this area, the land is moving upwards at a rate of ~2m/1000 years. Consequently, fringing coral reefs along the coast get uplifted, and become sub-aerially exposed. These ancient reefs now form a succession of 'steps', or terraces, in the coastal landscape, with the youngest reefs closest to the coast, and the oldest reefs at higher elevation further back from the coast. The oldest reefs in this image are about 250,000 years old and are seen as terraces towards the top-left of the photo. For our study, we collected fossil corals from reefs up to 130,000 years old, seen here as the terraces from the present-day coast up to the top of the first very distinctive set of cliffs near the middle of the photo. [Photo: Sandy Tudhope]

Variability in the El Niño-Southern Oscillation Through a Glacial-Interglacial Cycle

Science, v.291(5508), pp 1511-1517, February 23, 2001




Alexander W. Tudhope,1,2* Colin P. Chilcott,1 Malcolm T. McCulloch,3 Edward R. Cook,4 John Chappell,3 Robert M. Ellam,5 David W. Lea,6 Janice M. Lough,2 Graham B. Shimmield7





1Department of Geology & Geophysics, Edinburgh University, Edinburgh, EH9 3JW, UK.

2Australian Institute of Marine Science, Townsville, Queensland 4810, Australia.

3Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australia.

4Tree-Ring Laboratory, Lamont-Doherty Earth Observatory, New York 10964, USA.


5Scottish Universities Environmental Research Centre, East Kilbride, Glasgow G75 0QF, UK.

6Department of Geological Sciences and Marine Science Institute, University of California, Santa Barbara, CA 93106, USA.

7Dunstaffnage Marine Laboratory, Oban, Argyll, PA34 4AD, UK.

*To whom correspondence should be addressed. E-mail:


The El Niño-Southern Oscillation (ENSO) is the most potent source of interannual climate variability. Uncertainty surrounding the impact of greenhouse warming on ENSO strength and frequency has stimulated efforts to develop a better understanding of the sensitivity of ENSO to climate change. Here we use annually banded corals from Papua New Guinea to show that ENSO has existed for the past 130,000 years, operating even during "glacial" times of substantially reduced regional and global temperature and changed solar forcing. However, we also find that during the 20th century ENSO has been strong compared with ENSO of previous cool (glacial) and warm (interglacial) times. The observed pattern of change in amplitude may be due to the combined effects of ENSO dampening during cool glacial conditions and ENSO forcing by precessional orbital variations.


Walking up through the raised reef terraces on the Huon Peninsula. The scientists are walking on top of Holocene reefs (about 6,000- 10,000 years old), and are going towards older reefs. The ridge at the top of the photograph is the reef crest of the reef that grew ~125,000 years ago during the Last Interglacial. The cliffs and terraces between the scientists and the uppermost ridge are former reefs which grew during the last glacial period [Photo: Sandy Tudhope]

    Fig. 1. Location of the study sites

Fig. 2. (A) Comparison of coral skeletal d18O, rainfall and SST time series. The rainfall record is from Madang, the SST record is the IGOSS nmc blended ship and satellite data for the 1° square centered on 147.5°E, 6.5°S, and coral H95-64 is from the Huon Peninsula (Fig. 1). All records are interpolated to a resolution of eight samples per year (thin lines). In addition, the coral record has been smoothed with a binomial filter (thick line) to remove seasonality. Historical El Niño events are shown by shading. The double peak in wet/warm season (i.e., more negative d18O) values seen in some years in the coral record reflects the distinct double peak in wet season rainfall caused by the passage of the Inter-Tropical Convergence Zone southward over the area at the start of the wet season then back northward at the end of the wet season. (B) Fossil coral skeletal d18O time series for comparison with the modern records. Inferred paleo-El Niño events are shown by shading. Coral H95-16 has the weakest ENSO signal of all the fossil corals sampled. Although close to the sampling and analytical resolution of these records, double peaks in wet/warm season values in some years are suggestive of similar seasonal development to that of the modern era and imply excellent preservation of the proxy climate signal.

Fig. 3. ENSO variability since 1880. Modern coral skeletal d18O and instrumental climate records have been filtered with a Gaussian bandpass filter to reveal the 2.5- to 7-year (ENSO) components of variability. For all parameters, the polarity of the y axis has been set so that the El Niño phase of the Southern Oscillation would be expected to result in a downward anomaly in the curve. The "local" SST record is the reconstruction for the 1° square centered on 146.5°E, 5.5°S from the GISST2.3b data set of the UK Meteorological Office; the NINO3.4 OS SST is the optimally smoothed SSTreconstruction for the NINO3.4 region in the equatorial central-western Pacific. Darwin sea level pressure and NINO3.4 SST are widely recognized indices of ENSO activity. The pattern of relatively weak and irregular ENSO activity in the middle of the 20th century (but with a major event in the early 1940s) is a well-known feature of historical ENSO variability.

Download modern coral data as an Excel file

Download modern coral data as individual text files:

Madang dt91-7.txt Laing dur-2.txt
Huon h95-64.txt  

Fig. 4. Paleo-ENSO variability from fossil corals. (A) (Left) Seasonal resolution (thin lines) and 2.25-year binomial-filtered (thick lines) skeletal d18O records from all fossil corals used in this study, with the record from modern coral DT91-7 shown for comparison. (Right) 2.5- to 7-year (ENSO) bandpass-filtered coral d18O time series. (B) Standard deviation of the 2.5- to 7-year (ENSO) bandpass-filtered time series of all modern and fossil corals discussed in this study. Asterisk indicates that the time series is <30 years long. The horizontal dashed lines indicate maximum and minimum values of standard deviation for sliding 30-year increments in the modern coral records. Black bars, modern corals; gray bars, fossil corals.

Download fossil coral data as an Excel file

Download fossil coral data as individual text files:

m93tbfc.txt(2-3ka) laingfc2.txt(2-3ka)
h95-16.txt(~6.5ka) h96-27.txt(~38-42ka)
h96-18.txt(~38-42ka) h96-36.txt(~38-42ka)
h96-6.txt(~38-42ka) h97-10.txt(~85ka)
h95-14.txt(~112ka) h94-2.txt(~118-128ka)
h95-58.txt(~118-128ka) h95-6.txt(~130± 2 ka)
h95-18.txt(~130± 2 ka) h95-25.txt(~130± 2 ka)

Data Description


Fig. 5. (A) "Strength" of ENSO variability in coral d18O records for eight time periods over the past 150 ka. Solid black bars show the standard deviation ( units, on the 0 to 0.1 y axis scale) of the ENSO (2.5 to 7 year) bandpass-filtered coral d18O records from each time period. Shaded bars provide a measure of the occurrence of high-amplitude events for each period. The darker bars indicate the percentage of the data in the ENSO bandpass-filtered data that exceeds 0.15 absolute amplitude; lighter bars indicate the percentage of data that exceeds 0.10 amplitude (both plotted relative to the 0 to 25 scale on the y axis). m, number of corals combined for each group; n, total number of years represented by corals in each group. The horizontal dashed lines indicate the maximum and minimum standard deviation for sliding 30-year increments of modern coral d18O 2.5- to 7-year bandpass-filtered time series. (B) Estimate of global sea level (plotted as meters below present sea level) derived from benthic foraminiferal d18O. Bars indicate paleo-sea level estimated from the elevation, age, and uplift rate of corals analyzed in this study. These bars include uncertainty in the water depth in which the corals grew. Estimates of uplift rate are based on an assumed sea level of +5 m at 123 ka (circle and bar). (C) Sea surface temperature record for the western equatorial Pacific (ODP Hole 806B, 159°22'E, 0°19'N, 2520-m water depth) based on Mg/Ca composition of planktonic foraminiferans. The horizontal line indicates modern SST. (D) ENSO variability estimated from application of the Zebiak-Cane-coupled ocean-atmosphere model forced only by changing orbital parameters. Shown here is power in the 2- to 7-year (ENSO) band from multitaper spectral analysis of nonoverlapping 512-year segments of model NINO3 SST index. Power is approximately equal to 100× variance. Although there is considerable variation at sub-orbital wavelengths (2 of power estimates ~±71 based on a control run with no change in orbital parameters), the main precessionally related features, including the trend of increasing ENSO amplitude and frequency through the Holocene, were found to be statistically significant. (E) The precessional component of orbital forcing. For one cycle, the timing of perihelion is indicated as follows: a, boreal autumn; b, boreal winter; c, boreal spring; d, boreal summer.

To read or view the full study, please visit the Science website. It was published in Science Vol. 291, Issue 5508, pp.1511-1517, February 23, 2001.

Contact Us
National Oceanic and Atmospheric Administration
22 October