doi:10.1016/j.ocemod.2006.02.004
Copyright © 2006 Elsevier Ltd All rights reserved.
Spatial and temporal structure of Tropical Pacific interannual variability in 20th century coupled simulations
References and further reading may be available for this article. To view references and further reading you must
purchase this article.
Antonietta Capotondia, 
, 
, Andrew Wittenbergb and Simona Masinac
aNOAA/Earth System Laboratory, CIRES/Climate Diagnostics Center, R/CDC1, 325 Broadway, Boulder, CO 80305, United States
bGeophysical Fluid Dynamics Laboratory, Princeton, NJ, United States
cIstituto Nazionale di Geofisica e Vulcanologia, Bologna, Italy
Received 27 May 2005;
revised 6 February 2006;
accepted 7 February 2006.
Available online 9 March 2006.
Abstract
Tropical Pacific interannual variability is examined in nine state-of-the-art coupled climate models, and compared with observations and ocean analyses data sets, the primary focus being on the spatial structure and spectral characteristics of El Niño-Southern Oscillation (ENSO). The spatial patterns of interannual sea surface temperature (SST) anomalies from the coupled models are characterized by maximum variations displaced from the coast of South America, and generally extending too far west with respect to observations. Thermocline variability is characterized by dominant modes that are qualitatively similar in all the models, and consistent with the “recharge oscillator” paradigm for ENSO. The meridional scale of the thermocline depth anomalies is generally narrower than observed, a result that can be related to the pattern of zonal wind stress perturbations in the central-western equatorial Pacific. The wind stress response to eastern equatorial Pacific SST anomalies in the models is narrower and displaced further west than observed. The meridional scale of the wind stress can affect the amount of warm water involved in the recharge/discharge of the equatorial thermocline, while the longitudinal location of the wind stress anomalies can influence the advection of the mean zonal temperature gradient by the anomalous zonal currents, a process that may favor the growth and longer duration of ENSO events when the wind stress perturbations are displaced eastwards. Thus, both discrepancies of the wind stress anomaly patterns in the coupled models with respect to observations (narrow meridional extent, and westward displacement along the equator) may be responsible for the ENSO timescale being shorter in the models than in observations. The examination of the leading advective processes in the SST tendency equation indicates that vertical advection of temperature anomalies tends to favor ENSO growth in all the CGCMs, but at a smaller rate than in observations. In some models it can also promote a phase transition. Longer periods tend to be associated with thermocline and advective feedbacks that are in phase with the SST anomalies, while advective tendencies that lead the SST anomalies by a quarter cycle favor ENSO transitions, thus leading to a shorter period.
Keywords: El Niño phenomena; Climatic changes; Permanent thermocline; Winds; Surface temperature
Regional terms: Tropical Pacific Ocean
Fig. 1. Standard deviation of interannual SST from observations (top-left) and different CGCMs. In all cases the interannual SSTs have been computed by subtracting the monthly climatology from the monthly averages of the model outputs. Contour interval is 0.25 °C. Values larger than 0.75 °C are shaded. The CGCMs are sorted by decreasing period (see Table 2).
Fig. 2. Evolution of the Niño3.4 index (area-average SST over the region 5°S–5°N, 170°–120°W) for observations (a), CCSM3 (b), GFDL-CM2.0 (c), and GISS-EH (d). The climatological seasonal cycle has been removed, and a 5-month boxcar (equal weighted) smoother has been applied, according to Trenberth (1997). The horizontal dashed lines indicate ±0.5 °C.
Fig. 3.
(a) Power spectra of the Niño3.4 index derived from different observational datasets: HadISST, over the period 1950–2000 (black line), NOAA.ERSST over the period 1880–2000 (red line), and NOAA.ERSST over the period 1950–2000 (blue line). (b) Power spectra of the Niño3.4 index from CCSM3 (red), GFDL-CM2.0 (blue), and IPSL-CM4 (green). The spectrum from the HadISST dataset is shown for comparison. (c) Same as in (b), but for GISS-EH (red), PCM (blue), and MRI-CGCM2.3.2 (green). (d) Same as in (b), but for UKMO-HadCM3 (red), CSIRO-Mk3.0 (blue), and CNRM-CM3 (green). The numbers on the top axis indicate the period in years.
Fig. 4. Mean temperature sections along the equator for the INGV ocean analysis (a), the TAO observations (b), and the nine CGCMs (c–k). Contour interval is 2.5 °C. The thick solid line in each panel highlights the 15 °C isotherm. The 15 °C isotherm from the INGV analysis is shown in panels (b–k) as a dotted line for comparison. The section from the TAO buoy data has been obtained using linear interpolation.
Fig. 5.
Leading eofs of thermocline depth (diagnosed as the depth of the 15 °C isotherm) for the INGV ocean analysis, covering the period 1958–2000 (top), and for the GFDL/ARCs analysis over the period 1980–2000 (bottom). Negative values (blue shading) indicate shallower than average thermocline, while positive values (orange shading) indicate deeper than average thermocline. Contour interval is 10 m, but the −5 m and +5 m contours are also indicated. Dot-dash lines in all panels indicate the 10°S and 10°N latitudes for reference. (c) Same as in (a) but for the UKMO-HadCM3 model. (c′) Same as in (a′), but for the UKMO-HadCM3 model. (d) Same as in (a) but for the PCM. (d′) Same as in (a′) but for the PCM. (e) Same as in (a), but for GISS-EH. (e′) Same as in (a′) but for GISS-EH. (f) Same as in (a) but for the CNRM-CM3 model. (f′) Same as in (a′), but for the CNRM-CM3 model. (g) Same as in (a) but for the CSIRO-Mk3.0. (g′) Same as in (a′) but for the CSIRO-Mk3.0. (h) Same as in (a), but for MRI-CGCM2.3.2. (h′) Same as in (a′) but for MRI-CGCM2.3.2. (i) Same as in (a) but for the GFDL-CM2.0 model. (i′) Same as in (a′), but for the GFDL-CM2.0 model. (j) Same as in (a) but for IPSL-CM4. (j′) Same as in (a′) but for IPSL-CM4. (k) Same as in (a), but for CCSM3. (k′) Same as in (a′), but for CCSM3.
Fig. 6.
(a) Leading principal components associated with EOF1 (solid line) and EOF2 (dot-dash line) for the INGV ocean analysis over the period 1958–2000. The Niño3.4 index is also shown for comparison (red dotted line). The correlation coefficient between the Niño3.4 index and PC1 is 0.96, with the Niño3.4 index leading PC1 by 1 month. (a′) Lag-correlation between PC1 and PC2. Negative (positive) lags are for PC1 leading (lagging) PC2. (b) Same as (a), but for CCSM3. The correlation coefficient between Niño3.4 index and PC1 is 0.97, with the Niño3.4 index leading PC1 by 2 months. (b′) Same as (a′), but for CCSM3. (c) Same as (a), but for the UKMO-HadCM3. The correlation coefficient between Niño3.4 index and PC1 is 0.84, with the Niño3.4 index leading PC1 by 3 months. (c′) Same as (a′), but for the UKMO-HadCM3 model.
 |
 |
 |
Fig. 7. (a) Regression coefficient of monthly zonal wind stress anomalies upon the Niño3.4 index, based upon monthly values of band-pass filtered data, as described in the text, for the INGV ocean analysis. Contour interval is 2.5 × 10−3 Pa °C−1 The dot-dash line indicates the longitude of the “center of mass” of the regressed zonal wind stress within 2°S–2°N, as described in the text. Dashed lines indicate the 10°S, 10°N latitudes, for reference. (a′) Zonal average of (a) between 154°E and 120°W, displayed as a function of latitude. (b) Same as in (a), but for the GFDL/ARCs analysis. (b′) Same as in (a′), but for the GFDL/ARCs analysis. (c) Same as in (a), but for the UKMO-HadCM3. (b′) same as in (a′), but for the UKMO-HadCM3. (d) Same as in (a), but for the PCM. (d′) Same as (a′), but for the PCM. (e) Same as in (a), but for GISS-EH. (e′) Same as (a′), but for the GISS-EH. (f) Same as in (a), but for the CNRM-CM4 model. (f′) Same as in (a′), but for the CNRM-CM4 model. (g) Same as in (a), but for CSIRO-Mk3.0. (g′) Same as in (a′), but for CSIRO-Mk3.0. (h) Same as in (a), but for the MRI-CGCM2.3.2. (h′) Same as in (a′), but for the MRI-CGCM2.3.2. (i) Same as in (a), but for the GFDL-CM2.0 model. (i′) Same as in (a′), but for the GFDL-CM2.0 model. (j) Same as in (a), but for IPSL-CM4. (j′). Same as in (a′), but for IPSL-CM4. (k) Same as in (a), but for CCSM3. (k′) Same as in (a′), but for CCSM3.
Fig. 8. (Top) ENSO period (defined as the timescales shorter than 10 yrs corresponding to the maximum power in the spectrum of the Niño3.4 index) versus the meridional scale of the wind stress Ly, for all the CGCMs (solid dots) and the INGV ocean analysis (open circle). (Middle) ENSO period versus longitudinal location of the zonal wind stress (computed as the “center of mass” C, defined by Eq. (1) in the text) for all the experiments and the INGV analysis. (Bottom) ENSO period versus Tp, the value of the period predicted by using the multiple regression in Eq. (2).
Fig. 9.
(a) Regression of the mean vertical advection of temperature anomalies term, 
(wmTp), upon the Niño3.4 index for the INGV ocean analysis. Positive lags indicate that the Niño3.4 index leads the vertical advective tendency. Contour interval is 0.1 °C/month. The ±0.05 contours are also included. Negative values are indicated by dashed contours, while positive values are shown by solid lines. Absolute values larger than 0.05 °C/month are shaded. (c) Regression of the anomalous zonal advection term 
(upTm) upon the Niño3.4 index, displayed as a function of longitude and phase lag for the INGV ocean analysis. Contour interval and shading as in (a). (b) As in (a), but for the GFDL/ARCs analysis. (d) As in (c), but for the GFDL/ARCs analysis. The dot-dash lines in each panel indicate phase lags of −90°, 0°, and +90°. Models are sorted by period (from longest to shortest). The dotted line indicates the lag of maximum correlation between the local SST and the Niño3.4 index. (e) Same as in (a), but for the CNRM-CM3. (h) Same as in (c), but for the CNRM-CM3. (f) Same as in (a), but for CSIRO-Mk3.0. (i) Same as in (c), but for the CSIRO-Mk3.0. (g) Same as in (a), but for MRI-CGCM2.3.2. (j) Same as in (c), but for MRI-CGCM2.3.2. (k) Same as in (a), but for GFDL-CM2.0. (n) Same as in (c), but for GFDL-CM2.0. (l) Same as in (a), but for IPSL-CM4. (o) Same as in (c), but for IPSL-CM4. (m) Same as in (a), but for CCSM3. (p) Same as in (c), but for CCSM3.
Fig. 10. Time mean vertical velocity, averaged between 2°S and 2°N, and displayed as a function of longitude and depth for the GFDL/ARCs (a), and INGV (b) ocean analyses, and the CGCMs whose output included the vertical velocity field. The dot-dash line indicates 50 m depth, for reference.
Fig. 11. Solid lines show the longitudinal structure of the mean SST, averaged between 2°S and 2°N for CCSM3 (a), GFDL-CM2.0 (b), CSIRO-Mk3.0 (c), CNRM-CM3 (d), IPSL-CM4 (e), MRI-CGCM2.3.2 (f). The dotted line in each panel shows the longitudinal SST profile for GFDL/ARCs.
Table 1.
List of the models used in this study, including information about atmospheric and tropical oceanic resolutions (longitude × latitude × # of vertical layers) and period covered by the 20th century simulation
Table 2.
Interannual timescale (T), meridional width of the zonally-averaged τx regressed upon the Niño3.4 index (Ly), amplitude of the interannual zonal wind stress (τ) normalized by the standard deviation of the Niño3.4 index, “center of mass” of τx regressed upon the Niño3.4 index (C), and period predicted using a multiple regression algorithm that relates T to C and Ly (Eq. (2)), for the INGV ocean analysis and the CGCMs
All the quantities are defined in the text. The ocean analyses and the CGCMs are sorted by period.
Table 3.
Fitted values, jackknifed ranges, and bootstrapped 90% confidence intervals for the parameters in Eq. (2)
Corresponding author. Tel.: +1 303 497 6105; fax: +1 303 497 6449.
Abstract
Purchase PDF (4971 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (14)
Purchase PDF (1540 K)
