2.4 Understanding and predicting SST variations outside the tropical Pacific

2.4.1 The atmospheric bridge

One example of global climate interaction is the "atmospheric bridge", where atmospheric teleconnections associated with ENSO drive anomalous ocean conditions outside of the equatorial Pacific through changes in the heat, momentum, and fresh water fluxes across the air-sea interface. The resulting SST anomalies can also feed back on the initial atmospheric response to ENSO. As part of the GFDL-Universities Consortium project, we developed a coupled AGCM-mixed layer ocean model and used it to conduct experiments to study the atmospheric bridge and other air-sea interaction processes. In the "MLM" experiment, observed SSTs were prescribed as boundary conditions in the tropical Pacific (15N-15S, 172E-South American Coast), and the remainder of the global oceans were simulated using the variable-depth mixed-layer model. As Fig. 2.12 shows, the simulated SLP and SST anomalies associated with ENSO are fairly realistic, with stronger cyclonic circulation and cold water over the central North and South Pacific and warm water along the west coast of the Americas. ENSO-induced changes in the Walker circulation also lead to warm SSTs in the north tropical Atlantic and Indian Oceans.

Composites of SLP and SST from observations and model simulations

Fig. 2.12 (a) Observed and (b) simulated El Niño minus La Niña composite of SLP (contour interval of 1 mb) and SST (shading interval of 0.2 C) for DJF(0/1), where 0 indicates the ENSO year and 1 the next year. The composite is based on 9 El Niño and 9 La Niña events during 1950-1999. The observed values are from the NCEP reanalyses and the model results from an average of 16 MLM integrations.

How useful is this effect in actually predicting the interannual variations of SSTs outside the tropical Pacific basin? Figure 2.13 provides one measure of the forecast skill. It shows the correlation of the observed seasonal-mean SST anomalies, over the 50-year 1950-1999 period, with those predicted by the MLM model. The predicted field for each season represents a 16-member ensemble-mean. Results are shown separately for the 50 winter (JFM) and 50 summer (JAS) forecast cases over which the correlations were calculated. The correlations are generally higher than 0.4 in the central north and south Pacific oceans, and also in the tropical Indian and tropical Atlantic oceans. This is encouraging, and also sheds some light on why the LIM SST forecast models described in section 2.1, that are based solely on SST correlations between different tropical locations, perform as well as they do: the "bridge" effect is implicitly included in them. It is also interesting to compare Fig. 2.13 with Fig. 2.12 in areas such as the north Atlantic, where Fig. 2.12 suggests a bridge effect but Fig. 2.13 shows it to be unimportant.

Predictability of seasonal SST anomalies via the atmospheric bridge

Fig. 2.13 Predictability of seasonal SST anomalies worldwide via the "atmospheric bridge" from the tropical Pacific ocean, in winter (top panel) and summer (bottom panel). Values plotted are the correlations of observed SST anomalies with those predicted by an atmosphere-mixed layer ocean coupled model with specified observed SSTs in the central and eastern tropical Pacific ocean. See text for further details.

2.4.2 The re-emergence of long-lived subsurface temperature anomalies

The atmospheric changes associated with ENSO influence upper-ocean processes that affect the subsurface temperature structure and mixed-layer depth (MLD) long after the ENSO signal decays. Thermal anomalies that form in the surface waters of the extratropics during winter partially reemerge in the following winter, after being sequestered beneath the mixed layer in the intervening summer. SST anomalies generated via the atmospheric bridge recur in the following winter in central North Pacific via this reemergence mechanism (Fig. 2.14). The MLD is substantially deeper in the central North Pacific during El Niño than La Niña winters, but the reverse is true in the subsequent winter. During El Niño winters, enhanced buoyancy forcing (surface cooling) and mechanical mixing creates a colder and deeper mixed layer. After the MLD shoals in late spring, the cold water stored beneath the surface layer as part of the reemergence process increases the vertical stability of the water column, reducing the penetration of the mixed layer in the following fall and winter.

Composite El Niño minus La Niña ocean temperature

Fig. 2.14 The composite El Niño minus La Niña ocean temperature from Nov(0)-Jun(2) and the composite mixed layer depth (m) during El Niño (red line) and La Niña (blue line) from the MLM experiment in the central North Pacific (180-160W, 28N-42N).

The reemergence process is not confined to the central North Pacific, nor does it occur solely in conjunction with the atmospheric bridge. Figure 2.15 shows the evolution of the leading pattern of North Pacific SST variability in observations and in two distinct AGCM - ocean model experiments. In the first experiment, the mixed layer ocean model is active over the entire globe, including the tropical Pacific, and thus does not include ENSO. In the second experiment, observed SSTs are specified in the tropical Pacific, but the remainder of the world oceans are simulated by a slab model without mixed-layer physics. The observational column in Fig. 2.15 shows that the dominant large-scale SST anomaly pattern that forms in the eastern two-thirds of the North Pacific during winter recurs in the following winter without persisting through the intervening summer. Experiment 1 (middle column) reproduces this behavior, but Experiment 2 (right column) does not. These results suggest 1) that the winter-to-winter SST correlations are due to the reemergence mechanism and not due to similar atmospheric forcing of the ocean in consecutive winters, and 2) that the SST anomalies in the tropical Pacific associated with El Niño are not essential for reemergence to occur.

Evolution of the leading pattern of North Pacific SST variability

Fig. 2.15 The evolution of the leading pattern of SST variability over 20N-60N in the Pacific as indicated by extended EOF analyses of monthly SST anomalies from January through April of the following year. The results are presented as the correlation between the leading principal component (time series of EOF1) with SST anomalies at the individual grid points for March, September and March of the following year. (The other months, which are not shown, indicate a similar evolution). Results are presented for (a) observations for 1950-1995, (b) an AGCM-global MLM simulation, and (c) 4 TOGA-50m slab simulations. The contour interval is 0.2 with values > 0.4 shaded red and those < -0.4 shaded blue.

Back | Forward