In the tropical eastern-central Pacific Ocean, the annual cycle in sea surface temperature (SST), surface winds and pressure, and clouds are alternatively dominated by an antisymmetric (with respect to the equator) monsoonal mode in February and August and a quasi-symmetric equatorial-coastal mode in May and November, both having a period of one year. The monsoonal mode is forced by the differential insolation between the northern and southern hemisphere. The surface wind variation of the monsoonal mode tends to lead SST variation in late spring/fall. The equatorial-coastal mode originates from atmosphere-ocean interaction. Its development is characterized by contemporaneous intensification and spatial expansion (westward and poleward phase propagation).
The interaction between the forced monsoonal mode and the coupled equatorial-coastal mode plays a critical role in the annual cycle. From October to February, the decline of the northern winter regime of the monsoonal mode initiates and sustains the amplification of the equatorial-coastal mode, causing annual weakening of the cold tongue. From April to June, the enhancement of the poleward SST gradient associated with the decay of the equatorial-coastal mode initiates the eastern North Pacific summer monsoon. Atmosphere-ocean interaction is directly responsible for the annual weakening and re-establishment of the cold tongue, whereas the annual cycle in insolation regulates the interaction indirectly through the forced monsoonal mode.
The solar radiation flux at the surface under clear skies can be decomposed into a symmetric and an antisymmetric component with respect to the equator as shown in Fig. 1. The antisymmetric component represents the contrast in insolation between the northern and southern hemisphere, and is forced by the annual cycle in solar declination angle. The symmetric component, on the other hand, is caused by the semiannual cycle in solar declination (maximized on the equator at equinoxes) and the annual cycle in the sun-earth distance (maximum in northern winter solstice when the earth is closest to the sun and minimum in northern summer solstice). In the vicinity of the equator the amplitude of the semiannual harmonic is about 50% larger than that of the annual harmonic. Away from the equatorial regions, however, the antisymmetric annual harmonic dominates (Fig. 1).
The atmospheric response to the insolational forcing is strikingly different over the eastern and western hemisphere oceans. Regardless of the dominant semiannual variation of solar radiation in the equatorial region, SST and clouds in the equatorial eastern Pacific, as well as the Atlantic, have a dominant annual cycle with a warm and wet season occurring in boreal spring and a cold and dry season in boreal fall (Wyrtki 1965, Hastenrath and Lamb 1978, Horel 1982, Mitchell and Wallace 1992). Why does the periodicity of annual variation there differ from that of insolational forcing? How does the solar radiational forcing affect the annual variation in that region? What roles do the oceanic processes interacting with atmosphere play?
This paper presents an observational analysis of the peculiar annual cycle in the tropical eastern-central Pacific and explores its physical causes; in particular, the causes responsible for annual weakening and reestablishment of the equatorial cold tongue and the onset of the eastern North Pacific monsoon. There are a number of papers (for a brief review refer to Horel 1982 and Mitchell and Wallace 1992) and several comprehensive atlases (e.g., Wyrtki and Meyers 1975, Hastenrath and Lamb 1977, Weare et al. 1980, Janowiak et al. 1985, Sadler et al. 1987) that documented the annual variation of SST and atmospheric variables in the tropical Pacific by describing features of the climatological monthly means. The present analysis focus on the departure of climatological monthly mean from the long-term mean condition (the annual mean), which will be termed as annual perturbation in this paper. The analysis of the annual perturbations will provide additional understanding of the nature of the interaction between solar radiational forcing and atmosphere-ocean coupling in the annual cycle of the equatorial Pacific cold tongue.
Section 2 describes data and Section 3 briefs characteristics of the background mean state. Section 4 describes analysis procedure, beginning by showing that the annual perturbations of SST and surface winds in the eastern-central Pacific are alternatively dictated by a symmetric and an antisymmetric component during the course of the year. To explore the origin and evolution of these components, the annual perturbation was further partitioned into two orthogonal modes: a quasi-symmetric, equatorial-coastal mode and an antisymmetric, monsoonal mode. The spatiotemporal structures and the origin of the two modes are investigated in sections 5 and 6, respectively. The results indicate that the monsoonal mode originates from the insolational contrast between the northern and southern hemisphere, whereas the equatorial-coastal mode is a result of ocean-atmosphere-land interaction. The causes of the weakening and re-establishment of the cold tongue are examined in section 7. It is demonstrated that the interaction between the forced monsoonal mode and the coupled equatorial-coastal mode plays a critical role in the annual cycle of the cold tongue and eastern North Pacific monsoon. The last section summarizes major findings and discusses the limitations of the analysis and issues that require further investigations.
The annual perturbations (departures of climatological monthly means from the corresponding annual means) of SST, SLP, surface winds, and clouds as measured by OLR and HRC in the tropical eastern-central Pacific are partitioned into two modes: one asymmetric and one quasi-symmetric about the equator.
The asymmetric mode is an invisible monsoonal mode that has relatively large variance (due mainly to the contrast between the northern and southern hemisphere) occurring in January-March and July-October. The amplitudes of this mode in SST and SLP increase with latitude. It is forced by the differential insolation between the northern and southern hemisphere. It represents the tropical part of a global response of the atmosphere-ocean-land system to external solar forcing. The surface wind variation of the monsoonal mode leads that of SST in early summer and early winter, suggesting that the transition of SST from cold to warm or vice versa is partially caused by the changes in the trade winds.
The quasi-symmetric mode is an equatorial-coastal mode that has relatively large variance (due mainly to the contrast between the equator and midlatitudes) occurring in April-June and November-December. This mode has maximum amplitudes of SST and SLP occurring on the equator and near the South American coast. It exhibits a distinctive annual cycle with a wet and warm season in March-April and a dry and cold season in September-October. This mode has little to do with the annual march in insolation. The latter on the equator is forced by the semiannual cycle in solar declination and a minor annual cycle in earth-sun distance. The coherent structure and evolution between SST and atmospheric boundary layer circulation indicate that this mode results from large-scale air-sea interaction. The development of this mode is characterized by simultaneous spatial expansion and intensification with the maximum warming staying around (4oS, 95oW) where the oceanic thermocline depth is a minimum. The westward propagation is a manifestation of the interaction between the Walker circulation and SST (Horel 1982), whereas the poleward (mainly northward due to the proximity of the Central-North America landmass to the ocean) propagation is speculated to be a result of suppression of poleward Ekman transport of upwelled cold water and the dispersion of the warming area by reflected Rossby waves. This process is possibly enhanced by spring warming of the land mass over Mexico and by reduced evaporational latent heat loss.
It is the evolution of the equatorial-coastal mode that forms a distinct equatorial regime of annual cycle in the tropical oceans of the western hemisphere (the eastern Pacific and Atlantic Oceans). In sharp contrast, the forced antisymmetric monsoonal mode dominates the annual variation of the tropical oceans in the eastern hemisphere (the Indian and western Pacific Oceans). The peculiar annual cycle in the eastern Pacific originates from the coupling of the atmosphere and ocean. This is reflected in the fact that the changes in the northeast and southeast trades affect equatorial and coastal upwelling and surface latent heat flux; on the other hand, the decay of the equatorial mode (equatorial cooling and off-equatorial warming) influences the onset of the eastern North Pacific summer monsoon. The interplay between the two modes and the two media are also exemplified by the development of cold tongue-monsoon (or ITCZ) complex during northern summer as described by Mitchell and Wallace (1992) and Murakami et al. (1992).
The existence of cold tongue is a result of complex atmosphere-ocean-continent interaction. On the one hand, the SST distribution, in particular the thermal contrast between the cold tongues and warm pools, is a principal geophysical factor that determines the climatic locations of the ITCZ and South Pacific convergence zone (SPCZ) (Wang and Li 1993). The convective heating released in these convergence zones along with the SST gradient-induced pressure gradient force in the boundary layer drives the Pacific trades. On the other hand, the oceanic Ekman-layer divergence, induced by the equatorial easterly stress and along-shore wind stress of the trades, produces equatorial and coastal upwelling, which maintains the cold tongue and associated oceanic thermal contrasts. The influence of the air-sea interaction on the free tropospheric circulation, however, is restrained by the stable atmospheric stratification. The coherency in the annual variations between the 850 hPa equatorial zonal wind and SST is considerably weaker than that between the surface zonal wind and SST (figure not shown). The variation of 200 hPa equatorial zonal wind is independent of SST and is primarily controlled by the midlatitude conditions (Murakami and Wang 1993).
The present analysis is focused on the annual perturbation which is the departure of the climatological monthly mean from annual means. The analysis procedure is valid up to the limit where the nonlinear transient effects are small. The analysis of annual perturbation alone is based on the assumption that the physical processes responsible for the maintenance of the annual mean state are different from those governing the annual perturbation. The annual mean fields are highly asymmetric with regard to the equator. The important question of how the annual mean states form remains to be addressed. In addition, there are several issues that require further investigations, including the nature of unstable development of the equatorial-coastal mode, the relative importance of surface net heat flux versus surface wind forcing in the SST variation in the eastern Pacific, the relative importance of the continental heating over Mexico and the SST gradients north of the cold tongue in the onset of the eastern North Pacific summer monsoon, and the relative importance of the equatorial zonal versus meridional wind forcing in the annual cycle of the Pacific cold tongue. The role of the antisymmetric mode in triggering and promoting annual warming of the cold tongue may have an implication for the ENSO cycle. This, however, needs to be studied by careful analysis of long-term records.
Back to top | Back to menu