Abstract
Simulations of El Niño–Southern Oscillation (ENSO) variability with the Commonwealth Scientific and Industrial Research Organisation (CSIRO) Mark 3 coupled climate model, which is not flux adjusted and has an ocean north–south resolution of approximately 0.9°, are described. Major indices, periodicity, and spatial patterns of the modeled ENSO compare well to those observed over the last 100 yr. This good simulation is achieved despite some deficiencies in the model climatology, in particular the climatological tropical Pacific sea surface temperature (SST). The model SST climatology has a “cold tongue” that is too strong and extends too far into the western equatorial Pacific, a common problem experienced by many climate models. Although this cold tongue problem also affects the model rainfall climatology over the tropical ocean, the ENSO–rainfall teleconnection pattern is realistically simulated, particularly over the Indonesian and northeast Australian regions, where in reality rainfall is significantly affected by ENSO cycles.
Comparisons between modeled and observed equatorial thermocline structure reveal that the model thermocline depth (depth of the 20°C isotherm) is shallower, whereas the spread or thickness (depth difference between 16° and 22°C isotherms) of the modeled thermocline is greater, than the observed. The former is favorable, whereas the latter is unfavorable, for generating strong ENSO variability, because a shallower thermocline with smaller spread of isotherms and steeper slope makes it easier for the equatorial upwelling to draw the colder subthermocline water to the surface. On balance, the model is capable of producing ENSO cycles with realistic amplitude. This model capability is further highlighted by what is called here a “super-ENSO” pair: a super-El Niño event followed by a super-La Niña event, both with a Niño-3.4 index (SST average over 5°S–5°N, 120°–170°W) exceeding 3°C in amplitude.
The pairing of the two superevents is unique, and the dynamics are explored. It is found that during the super-El Niño event, the surface zonal wind stress, SST, and the equatorial upwelling anomalies are proportionately large. In contrast, during the super-La Niña event, the response of SST anomalies to easterly and upwelling anomalies is disproportionately large. It is demonstrated that this exceptionally large cooling of SST is linked to an exceptionally strong shallowing of the equatorial thermocline depth, and that the shallowing is induced by the exceptionally strong westerly wind anomalies associated with the super-ENSO. In the context of the recently proposed recharge–oscillator paradigm, which is shown to operate in the present model, the strong shallowing can be seen as a result of the discharge of the equatorial Pacific warm water volume in response to the exceptionally strong westerly anomalies associated with the super-El Niño event.
Corresponding author address: Dr. Wenju Cai, Division of Atmospheric Research, CSIRO, PMB 1, Aspendale, VIC 3195, Australia. Email: wenju.cai@csiro.au
