SATELLITE OBSERVATIONS OF COOL OCEAN-ATMOSPHERE INTERACTION* BY SHANG-PING XLE New observations from space reveal surprisingly robust patterns of air-sea coupling over cool oceans where such coupling has been thought to be weak. igh sea surface temperature (SST) is generally tropical ocean-atmosphere interaction that gives rise required for deep convection that reaches the to phenomena ranging from the El Nino-Southern Htropopause. In the current climate, the SST Oscillation (ENSO; Neelin et al. 1998) to the northward threshold for deep convection is somewhere around displacement of the intertropical convergence zone 26°-27°C, depending upon region and season (Gra- (ITCZ; Xie and Philander 1994; Philander et al. 1996). ham and Barnett 1987; Waliser et al. 1993). Over such Atmospheric general circulation models (GCMs) show warm oceans, SST changes cause deep convective ad- considerable skills in simulating this deep adjustment justment, which in turn excites the dynamic response to SST anomalies associated with ENSO (e.g., Alexander of predominantly first baroclinic mode structure with et al. 2002). strong surface wind signals. This robust relation be- Over cool oceans where deep convection does not tween SST, deep convection, and wind has led to the occur,1 the atmospheric adjustment to changing SST rapid advance in understanding and modeling the differs markedly from that over warm oceans. Cool ocean-atmosphere interaction is poorly understood, and this lack of understanding is a stumbling block in the current effort to study non-ENSO climate vari- international Pacific Research Center Contribution Number 239 and School of Ocean and Earth Science and Technology Contribu- ability. Unlike their success in simulating the South- tion Number 6261. ern Oscillation, atmospheric GCMs disagree among AFFILIATION: XIE—International Pacific Research Center and themselves in their atmospheric response to SST Department of Meteorology, University of Hawaii at Manoa, anomalies in the extratropics (see Kushnir et al. 2002 Honolulu, Hawaii for a review). CORRESPONDING AUTHOR: Shang-Ping Xie, IPRC/SOEST, University of Hawaii at Manoa; 1680 East West Road, Honolulu, HI 1 96822 While the SST threshold is a convenient way to divide the warm E-mail: [email protected] and cool regimes for air-sea interaction, local SST is not the DOI: 10.1 175/BAMS-85-2-195 only factor for deep convection, which is also influenced by In final form 21 September 2003 other factors such as large-scale subsidence. It may be more ©2004 American Meteorological Society physically appropriate to divide the warm and cold regimes ac- cording to whether there is significant deep convection or not. AMERICAN METEOROLOGICAL SOCIETY FEBRUARY 2004 BAI1S" | 195 Unauthenticated | Downloaded 10/08/21 07:00 PM UTC Observational efforts so far have failed to yield con- ments are unaffected by clouds except by those with clusive evidence of SST effects on the planetary-scale sizable precipitation. Based on microwave remote- atmospheric circulation over cool oceans because of a sensing data, several recent studies have identified SST high level of weather noise in the extratropics on one variations induced by ocean dynamics, and they have hand, and short and sparse in situ observations on the mapped the effects of these variations on the atmo- other. Most often negative correlations between SST sphere (Xie et al. 1998; Wentz et al. 2000; Chelton and surface wind speed variability are observed in the et al. 2001; Xie et al. 2001; Liu 2002; White and Annis extratropics for seasonal means and on the basin scale 2003; O'Neill et al. 2003; Vecchi et al. 2004). These (Fig. 1). Researchers have viewed this as being indica- new satellite observations reveal a rich variety of pat- tive of one-way forcing from atmosphere to ocean terns of ocean-atmosphere coupling in the Indian, through wind-induced changes in surface turbulence Pacific, and Atlantic Oceans, from the equator to the heat flux (sidebar 1). midlatitudes. Because individual papers tend to focus To achieve an adequate sample size in a grid box, on one particular phenomenon in one particular re- climate datasets typically have resolutions of several gion, the full spectrum of cool ocean-atmosphere hundred to 1000 km and of a month to a season. interaction has not been identified. Major ocean currents like the Kuroshio and Gulf The present paper synthesizes these recent stud- Stream are only around 100-200 km wide and form ies and compares atmospheric response over differ- sharp SST fronts that are poorly represented in these ent ocean conditions in different parts of the World datasets. It is on these narrow oceanic fronts, however, Ocean. By assembling these observational facts, we that ocean dynamics become important and cause wish to see whether there is a common atmospheric large SST variations. Thus, conventional climate response pattern, and if so, by what mechanisms and datasets may severely underrepresent such dynami- under what conditions this response takes place. A cally induced SST anomalies and their atmospheric synthesis of similarities and differences in the atmo- effect. spheric adjustment over different regions of the The vast oceans can only be sparsely observed by World Ocean can help shed light on the dynamics of traditional ship-based measurements, but the rapid cool ocean-atmosphere interaction and stimulate fu- advance in space-based microwave remote sensing is ture studies. revolutionizing ocean observations, providing global Over cool oceans, the direct effect of SST variations fields of key ocean-atmospheric variables at unprec- is likely to be trapped within the planetary boundary edented resolutions in space and time. Unlike visible layer (PBL), the lowest 1-2 km of the atmosphere, and infrared remote sensing, microwave measure- because this layer is capped by a stable layer, often in FIG. I. SST-wind relation in the North Pacific and Atlantic Oceans, (left) COADS SST (color shade), surface wind vectors, and SLP regressed upon the Pacific decadal oscillation index (Mantua et al. 1997). (right) COADS SST (color in °C) and NCEP surface wind (m s_l) composites in Jan-Mar based on a cross-equatorial SST gradient index (Okumura et al. 2001). 196 | BAI1S* FEBRUARY 2004 Unauthenticated | Downloaded 10/08/21 07:00 PM UTC the form of a temperature inversion (e.g., Norris 1997 at 0.25° resolution. QuickSCAT provides a daily 1998). In the present review, we focus on the satellite coverage over 90% of the World Ocean while the TMI observations of SST, surface wind velocity, and takes about 2 days to cover the global Tropics. boundary layer clouds—variables most relevant to air-sea interaction. We begin with a description of the WIND RESPONSE. This section examines the SST wind responses to SST variations, and then continue effect on winds. We organize our discussion accord- with a survey of cloud responses to these variations. ing to the sign and size of correlation between SST Finally, we discuss how SST-induced anomalies in the and wind speed [denoted as r(T, W) hereafter], which, boundary layer might lead to a deep response that as will become clear, is a useful indicator of whether reaches the upper troposphere. SST variations are merely forced by the atmosphere Unless we mention otherwise, the following mi- or whether the SST changes exert an influence on sur- crowave measurements were used in the studies we face wind. are reviewing: sea surface vector wind by the SeaWinds scatterometer on the QuickSCAT satellite Positive correlation between temperature and wind (Liu 2002); and SST, cloud water, and scalar wind speed. In the eastern equatorial Pacific and Atlantic, speed by the Tropical Rain Measuring Mission there is a SST front centered along 1°-2°N and sepa- (TRMM) satellite's microwave imager (TMI; Wentz rating the equatorial cold tongue from the warmer et al. 2000). QuickSCAT data are available since July water to the north. Tropical instability waves (TIWs), 1999 on a 0.25° grid, and TMI data since December due to hydrodynamic instabilities of equatorial ocean With its huge heat content and modes of SST variability with the negative correlation in Fig. I slow dynamic adjustment, the preferred spatial patterns. The in the extratropics can be inter- ocean is generally considered as coupled SST-wind pattern in preted as ocean passively respond- important for climate variability Fig. I b, for example, can be ing to wind-induced latent and with time scales longer than a maintained by a so-called wind- sensible heat flux, with SST season. On interannual to evaporation-SST feedback with decreasing as the prevailing interdecadal time scales, SST the help of weather noise (Chang westerlies intensify (Frankignoul variability is generally caused by et al. 1997; Xie and Tanimoto 1985; Barsugli and Battisti 1998). the following mechanisms: surface 1998). Surface heat flux is the Thus, understanding two-way effects due to surface heat flux and central agent for such a thermo- ocean-atmosphere interaction in Ekman advection that involves dynamical feedback, in contrast to the extratropics requires answers only the ocean mixed layer, and the Bjerknes feedback that gives to the following questions. First, subsurface effects by horizontal rise to ENSO with ocean dynamics do the ocean dynamics matter and and vertical advection due to as a key element. (The SST-wind where? Recent ocean GCM studies thermocline variability. For relation in the equatorial Pacific suggest that much of SST variabil- example, the subsurface effects near the date line in Fig. la is ity over the Kuroshio-Oyashio dominate the equatorial upwelling broadly consistent with the Extenstion (KOE) east of Japan is zone as seen in ENSO where Bjerknes feedback.) due to ocean dynamic effects. The ocean wave adjustment sets the Matters become very different deep winter mixed layer there time scale (Neelin et al. 1998). in the extratropics, where atmo- allows the ocean memory of past Surface effects become domi- spheric internal variability is often wind changes to have a marked nant over open off-equatorial organized in space forming effect on SST (Xie et al.