Beschreibung: The space-time structure and predictability of the El Niño/Southern Oscillation (ENSO) phenomenon was investigated. Two comprehensive datasets were analyzed by means of an advanced statistical method, one based on observational data and the other on data derived from an extended-range integration performed with a coupled ocean-atmosphere general circulation model. It is shown that a considerable portion of the ENSO-related low-frequency climate variability in both datasets is associated with a cycle involving slow propagation in the equatorial oceanic beat content and the surface wind field. The existence of this cycle implies the ability of climate predictions in the tropics up to lead times of about one year. This is shown by conducting an ensemble of predictions with our coupled general circulation model. For the first time a coupled model of this type was successfully applied to ENSO predictions.

Beschreibung: A hybrid coupled model (HCM) of the tropical ocean–atmosphere system is described. The ocean component is a fully nonlinear ocean general circulation model (OGCM). The atmospheric element is a statistical model that specifies wind stress from ocean-model sea surface temperatures (SST). The coupled model demonstrates a chaotic behavior during extended integration that is related to slow changes in the background mean state of the ocean. The HCM also reproduces many of the observed variations in the tropical Pacific ocean-atmosphere system. The physical processes operative in the model together describe a natural mode of climate variability in the tropical Pacific ocean–atmosphere system. The mode is composed of (i) westward-propagating Rossby waves and (ii) an equatorially confined air–sea element that propagates eastward. Additional results showed that the seasonal dependence of the anomalous ocean–atmosphere coupling was vital to the model's ability to both replicate and forecast key features of the tropical Pacific climate system. A series of hindcast and forecast experiments was conducted with the model. It showed real skill in forecasting fall/winter tropical Pacific SST at a lead time of up to 18 months. This skill was largely confined to the central equatorial Pacific, just the region that is most prominent in teleconnections with the Northern Hemisphere during winter. This result suggests the model forecasts of winter SST at leads times of at least 6 months are good enough to be used with atmospheric models (statistical or OGCM) to attempt long-range winter forecasts for the North American continent. This suggestion is confirmed in Part II of this paper.

Beschreibung: We have investigated the seasonal cycle and the interannual variability of the tropical Indian Ocean circulation and the Indian summer monsoon simulated by a coupled ocean-atmosphere general circulation model in a 26- year integration. Although the model exhibits significant climate drift, overall, the coupled GCM simulates realistically the seasonal changes in the tropical Indian Ocean and the onset and evolution of the Indian summer monsoon. The amplitudes of the seasonal changes, however, are underestimated. The coupled GCM also simulates considerable interannual variability in the tropical Indian Ocean circulation, which is partly related to the El Niño/Southern Oscillation phenomenon and the associated changes in the Walker circulation. Changes in the surface wind stress appear to be crucial in forcing interannual variations in the Indian Ocean SST. As in the Pacific Ocean, the net surface beat flux acts as a negative feedback on the SST anomalies. The interannual variability in monsoon rainfall, simulated by the coupled GCM, is only about half as strong as observed. The reason for this is that the simulated interannual variability in the Indian monsoon appears to be related to internal processes within the atmosphere only. In contrast, an investigation based on observations shows a clear lead-lag relationship between interannual variations in the monsoon rainfall and tropical Pacific SST anomalies. Furthermore, the atmospheric GCM also fails to reproduce this lead-lag relationship between monsoon rainfall and tropical Pacific SST when run in a stand-alone integration with observed SSTs prescribed during the period 1970–1988. These results indicate that important physical processes relating tropical Pacific SST to Indian monsoon rainfall are not adequately modeled in our atmospheric GCM. Monsoon rainfall predictions appear therefore premature.

Beschreibung: A 26-year integration has been performed with a coupled ocean-atmosphere general circulation model (CGCM). The oceanic part resolves all three oceans in the latitude band 70°N–70°S but is dynamically active only between 30°N and 30°S. The atmosphere is represented by a global low-order spectral model. The coupled model was forced by seasonally varying insolation. Although the simulated time-averaged mean conditions in both atmosphere and ocean show significant deviations from the observed climatology, the CGCM realistically simulates the interannual variability in the tropical Pacific. In particular, the CGCM simulates an irregular ENSO with a preferred time scale of about 3 years. The mechanism for the simulated interannual variability in the tropical Pacific is related to both the “delayed action oscillator” and the “slow SST mode.” It therefore appears likely that either both modes can coexist or they degenerate to one mode within certain locations of the parameter space. This hypothesis is also supported by calculations performed with simplified coupled models, in which the atmospheric GCM was replaced by linear steady-state atmosphere models. Further, evidence is found for an eastward migration of zonal wind anomalies over the western Pacific prior to the extremes of the simulated ENSO, indicating a link to circulation systems over Asia. Because an earlier version of the CGCM did not simulate interannual variability in the tropical Pacific, additional experiments with a simplified coupled model have been conducted to study the sensitivity of coupled systems to varying mean oceanic background conditions. It is shown that even modest changes in the background conditions can push the coupled system from one flow regime into another.

Beschreibung: Two extended integrations of general circulation models (GCMs) are examined to determine the physical processes operating during an ENSO cycle. The first integration is from the Hamburg version of the ECMWF T21 atmospheric model forced with observed global sea surface temperatures (SST) over the period 1970–85. The second integration is from a Max Planck Institut model of the tropical Pacific forced by observed wind stress for the same period. Both integrations produce key observed features of the tropical ocean-atmosphere system during the 1970–85 period. The atmospheric model results show an eastward propagation of information from the western to eastern Pacific along the equator, although this signal is somewhat weaker than observed. The Laplacian of SST largely drives the surface wind field convergence and hence determines the position of large scale precipitation-condensation heating. This statement is valid only in the near-equatorial zone. Air-sea heat exchange is important in the planetary boundary layer in forcing the wind field convergence but not so important to the main troposphere, which is heated largely by condensation heating. The monopole response seen in the atmosphere above about 500 mb is due to a combination of factors, the most important being adiabatic heating associated with subsidence and tropic-wide variations in precipitation. The models show the role of air-sea heat exchange in the ocean heat balance in the wave guide is one of dissipation/damping. Total air-sea heat exchange is well represented by a simple Newtonian cooling parameterization in the near-equatorial region. In the wave guide, advection dominates the oceanic heat balance with meridional advection being numerically the most important in all regions except right on the equator. The meridional term is largely explained by local Ekman dynamics that generally overwhelm other processes in the regions of significant wind stress. The principal element in this advection term is the anomalous meridional current acting on the climatological mean meridional SST gradient. The eastward motion of the anomalies seen in both models is driven primarily by the ocean. The wind stress associated with the SST anomalies forces an equatorial convergence of heat and mass in the ocean. Outside the region of significant wind forcing, the mass source leads to a convergent geostrophic flow, which drives the meridional heat flux and causes warming to the east of the main wind anomaly. West of the main anomaly the wind and geostrophic divergence cause advective cooling. The result is that the main SST anomaly appears to move eastward. Since the direct SST forcing drives the anomalous wind, surface wind convergence, and associated precipitation, these fields are seen also to move eastward.

Beschreibung: The ECMWF-T21 atmospheric GCM is forced by observed near-global SST from January 1970 to December 1985. Its response in low level winds and surface wind stress over the Pacific Ocean is compared with various observations. The time dependent SST clearly induces a Southern Oscillation (SO) in the model run which is apparent in the time series of all variables considered. The phase of the GCM SO is as observed, but its low frequency variance is too weak and is mainly confined to the western Pacific. Because of the GCM's use as the atmospheric component in a coupled ocean-atmosphere model, the response of an equatorial oceanic primitive equation model to both the modeled and observed wind stress is examined. The ocean model responds to the full observed wind stress forcing in a manner almost identical to that when it is forced by the first two low frequency EOFs of the observations only. These first two EOFs describe a regular eastward propagation of the SO signal from the western Pacific to the central Pacific within about a year. The ocean model's response to the modeled wind stress is too weak and similar to the response when the observed forcing is truncated to the first EOF only. In other words, the observed SO appears as a sequence of propagating patterns but the simulated SO as a standing oscillation. The nature of the deviation of the simulated wind stress from observations is analyzed by means of Model Output Statistics (MOS). It is shown that a MOS-corrected simulated wind stress, if used to force an ocean GCM, leads to a significant enhancement of low frequency SST variance, which is most pronounced in the western Pacific.

Beschreibung: The time history of upper-ocean temperatures in the tropical Pacific has been used as a predictor in a statistical prediction scheme to forecast SST anomalies in this region. The temperature variations were taken from the output of an oceanic general circulation model that was forced by observed winds for the period 1961 to 1985. Since such model data are presently used as initial conditions in prediction experiments with coupled ocean–atmosphere models, it is of particular interest to investigate up to what lead time tropical Pacific SST is predictable without the coupling of an atmosphere model to the ocean model. We compared our results with those obtained by the persistence forecast and with those obtained by using the wind stresses themselves as predictors in a statistical forecast model. It is shown that using the upper ocean temperatures from the ocean model forced by observed winds gives significantly better skills at lead times of 6 to 12 months compared to persistence and to the pure wind-stress model. Off-equatorial heat content anomalies at 5°N are shown to contribute significantly to the predictability at these lead times, while those at 12°N do not.