Coupling of Extratropical Mesoscale Eddies in the Ocean to Westerly Winds in the Atmospheric Boundary Layer

1095 Coupling of Extratropical Mesoscale Eddies in the Ocean to Westerly Winds in the Atmospheric Boundary Layer WARREN B. WHITE AND JEFFREY L. ANNIS Scripps Institution of Oceanography, University of California, San Diego, La Jolla, California (Manuscript received 19 December 2001, in final form 8 November 2002) ABSTRACT The sea surface temperature (SST) signature in mesoscale eddies in the western boundary current extensions around the globe and in the Antarctic Circumpolar Current are found to alter the surface stress associated with background westerly winds, producing wind stress curl (WSC) residuals of eddy scale that are capable of modifying the eddy dynamics. This is revealed by examining satellite-derived mesoscale sea level height (SLH), SST, and neutrally stable zonal surface wind (ZSW) residuals together for 18 months. In the presence of background westerly winds on basin scales, warm mesoscale eddies reduce the stability of the marine atmospheric boundary layer, increasing the zonal air sea momentum flux measured by satellite scatterometry. Warm SST residuals of 0.8 C are capable of producing westerly ZSW residuals of 1.2 m s 1 under background westerly winds of 6 ms 1. Alternatively, this means increasing the otherwise neutrally stable drag coefficient by 40%, consistent with in situ measurements. The resulting feedback from atmosphere to ocean through the resulting mesoscale WSC residuals ( 5.0 10 7 Nm 3 ) produces residual Ekman pumping that can be on the same order as the residual SLH tendency in the eddy field. Moreover, the spatial phasing of the mesoscale WSC residuals acts, on average, to displace the mesoscale eddies equatorward with meridional coupling phase speeds of 0.01 m s 1 while suppressing their amplitude. 1. Introduction Mesoscale eddies achieve their largest magnitude in the western boundary current extensions in each ocean basin and in the Antarctic Circumpolar Current (ACC) in the Southern Ocean (Fig. 1). Initially, the mesoscale eddy field in the midlatitude North Pacific Ocean was studied extensively using in situ upper-ocean temperature measurements collected from volunteer observing ships (White and Bernstein 1979). The mesoscale eddy field in the Kuroshio Oyashio current extension was found to be dominated by zonal wavelengths ranging from 400 to 1200 km and by periods ranging from 6 to 18 months (Bernstein and White 1977; Talley and White 1987). The source of mesoscale eddy activity in the Kuroshio Oyashio extension was observed to be mixed baroclinic barotropic shear instability (Bernstein and White 1982; Bennett and White 1986; Tai and White 1990), with mesoscale eddy activity governed largely by Rossby wave dynamics, propagating eastward or westward depending on the background absolute vorticity gradient (Mizuno and White 1983; Qiu et al. 1991). Similar results have been obtained for the mesoscale eddies in the ACC using satellite altimetry and Corresponding author address: Dr. Warren B. White, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093-0230. E-mail: wbwhite@ucsd.edu radiometry (Hughes 1996; Hughes et al. 1998; Hill et al. 2000). These mesoscale eddies are associated with sea surface temperature (SST) signatures that alter the stability of the marine atmospheric boundary layer above them and have the potential for affecting the surface wind stresses associated with the background winds. Work on this subject began in the eastern Pacific Ocean, when Greenhut (1982) found SST discontinuities altering the surface fluxes and associated drag coefficients accompanying the background wind. From aircraft observations taken during the Joint Air Sea Interaction (JASIN) experiment, Guymer et al. (1983) found mesoscale variability in SST producing similar results. From aircraft measurements taken during the Frontal Air Sea Interaction Experiment (FASINEX), Friehe et al. (1991) focused on background winds directed across an SST front, finding larger surface wind stresses and lesser buoyant stability in the marine atmospheric boundary layer over the warm side of the front than on the cold side, with surface wind stress magnitudes increasing by a factor of 2. From these measurements, they computed 10-m drag coefficients on both sides of the front, finding nearly a 100% increase when passing from the cold side to the warm side. Friehe et al. (1991) proposed that this rather extraordinary increase in surface wind stress across the front could produce a feedback to the ocean, thereby altering the dynamics of the front. 2003 American Meteorological Society

1096 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 33 FIG. 1. (a) The Northern Hemisphere geographical distribution of the rms of mesoscale SLH, SST, and ZSW residuals in extratropical mesoscale eddy field, filtered for zonal wavelengths of 400 1200 km and for periods 1 month, observed over the 18-month record for Jul 1999 Dec 2000. (b) As in (a) but for the Southern Hemisphere. Contour levels are 0.02 m, 0.08 C, and 0.05 m s 1. Shading is for effect. Black boxes indicate regions where the mesoscale eddy fields are examined in detail.

1097 In the present study, we find mesoscale eddies in the western boundary current extensions and in the ACC associated with SST residuals that alter the surface wind stresses accompanying the background westerly wind field. We find warm (cool) SST residuals in these mesoscale eddies associated with positive (negative) neutrally stable zonal surface wind (ZSW) residuals at 10- m height derived from satellite scatterometry by the SeaWinds QuikScat project (Weiss 2000). These ZSW residuals are an artifact of the method of producing neutrally stable winds for a marine atmospheric boundary layer that is not neutrally stable. Their presence in these satellite winds indicates that the background westerly wind imparts greater stress to the ocean over warm eddies than over cold eddies. So, we estimate the corresponding change in the drag coefficient for the background westerly winds of basin scale that are directed over warm and cold mesoscale eddies. This yields results consistent with those of Friehe et al.; that is, a background westerly wind of basin scale is associated with a surface wind stress field of mesoscale. Here, we find the corresponding wind stress curl (WSC) residuals exerting a significant feedback on the dynamics of these mesoscale eddies, producing an Ekman pumping that significantly alters the observed sea level height (SLH) tendency in the eddy field, capable of modifying not only their propagation characteristics but their stability as well. 2. Data and methods We examine SLH from TOPEX/Poseidon and ERS- 1/2 altimetry (Ducet et al. 2000; Le Traon et al. 2001), SST from multichannel advanced very-high resolution radiometry (Smith et al. 1996), and neutrally stable zonal and meridional winds at 10-m height zonal surface winds from the SeaWinds QuikScat project (Weiss 2000) for 18 months from July 1999 through December 2000. The ZSW and meridional surface winds (MSW) are available on a daily basis, while SLH and SST are available every 10 days. We compute the surface wind stresses and wind stress curl estimates from daily ZSW and MSW estimates using a bulk formula with a winddependent drag coefficient under neutrally stable conditions, increasing with wind speed but independent of air sea temperature differences (Large and Pond 1981). We average the daily ZSW and MSW estimates, and the corresponding WSC estimates, onto the same 10-day grid as SLH and SST estimates, interpolating all five variables onto a common 0.25 latitude longitude grid. Then we form residuals by subtracting the 10-day estimates from an annual long-term mean from January to December 2000. We temporally low-pass filter and spatially bandpass filter the residuals for periods 1 month and for zonal wavelengths of 400 1200 km, respectively (Kaylor 1977). The temporal filtering operation suppresses atmospheric synoptic-scale variability in all the residuals, most particularly that in the ZSW, MSW, and WSC residuals, while the spatial filter allows us to isolate the dominant signal in the mesoscale eddy spectrum (e.g., Bernstein and White 1977; Hughes 1995). Since we observe only two to three cycles of mesoscale variability at each grid point over the 18-month record, we obtain statistical confidence in the results by conducting comparisons over latitude longitude domains (15 lat by 60 lon) that encompass 12 16 mesoscale eddies. We do this for four independent domains, focusing on mesoscale eddies in the Kuroshio Oyashio current extension, the Gulf Stream current extension, the Brazil current extension, and the ACC south of Africa. This allows us to examine the association between mesoscale SLH, SST, ZSW, and WSC residuals over 48 64 independent eddies, yielding equivalent numbers of effective degrees of freedom (Emery and Thomson 2001). 3. Geographical distribution of the rms of mesoscale SLH, SST, and ZSW residuals The geographical distributions of the root-meansquare (rms) of mesoscale SLH, SST, and ZSW residuals in the extratropical Northern Hemisphere (Fig. 1a) and Southern Hemisphere (Fig. 1b) display largest estimates in the western boundary current extensions in each ocean basin and along the ACC in the Southern Ocean, achieving maximum rms estimates of 0.10 m, 0.6 C, and 0.3 m s 1. Mesoscale eddy activity in SLH, SST, and ZSW residuals can also be seen to be relatively intense off the east and west coasts of southern Australia and along the South Pacific convergence zone (SPCZ). In the ACC, the mesoscale eddy activity is most intense south and east of Africa, and south of Tasmania and New Zealand, with weaker eddy activity over the rest of the ACC. The nominal correspondence in global distributions of the rms of mesoscale SLH, SST, and ZSW residuals in the strong current regions (delineated by rectangular boxes, Fig. 1), where mesoscale eddies arise principally from shear instability in the mean flow, indicates that covarying SLH and SST residuals influence ZSW variability in the overlying marine atmospheric boundary layer. 4. Phase relationships among mesoscale SLH, SST, ZSW, and WSC residuals To demonstrate that covarying SLH and SST residuals in the extratropical mesoscale eddy field influence the ZSW residuals along the western boundary current extensions and in the ACC south of Africa, we overlay SST residuals onto SLH residuals, ZSW residuals onto SST residuals, and WSC residuals onto SLH residuals over the four domains delineated by the rectangular boxes in Fig. 1. These four domains are in the Kuroshio Oyashio current extension (Fig. 2a), the Gulf Stream current extension (Fig. 2b), the Brazil current extension

1098 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 33 FIG. 2a. The distribution of mesoscale eddies in the Kuroshio Oyashio current extension, revealed in mesoscale SLH, SST, ZSW, and WSC residuals mapped from 30 to 45 N and 140 E to160 W for the 10-day period bracketing 14 Mar 2000. This occurs during late winter when the depth of the near-surface mixed layer extends to the top of the main pycnocline and when the background westerly winds overlie the current (see Fig. 3a). The overlay of contours of mesoscale (a) SST residuals on colors of mesoscale SLH residuals, (b) ZSW residuals on colors of mesoscale SST residuals, and (c) WSC residuals on colors of mesoscale SLH residuals. Color levels for SLH residuals are 0.04 m; color levels and contour intervals for SST residuals are 0.2 C; contour intervals for ZSW residuals are 0.2 m s 1 ; and contour intervals for WSC residuals are 1.5 10 7 Nm 3. Positive contours are solid and negative contours are dashed. Zonal- and meridional-lag cross correlation between (d) mesoscale SST and SLH residuals, (e) mesoscale SST and ZSW residuals, and (f) mesoscale SLH and WSC residuals, with contours of 0.2, computed over the entire domain. Correlations 0.32 are significant at the 95% confidence level for 40 effective spatial degrees of freedom over the domain (Snedecor and Cochran 1980).

1099 FIG. 2b. As in Fig. 2a but for the Gulf Stream current extension, mapped from 35 to 50 N and 80 to 20 W for the 10-day period bracketing 29 Nov 2000, during the middle of winter when the westerly winds are strong near the current (see Fig. 3b). (Fig. 2c), and the ACC south of Africa (Fig. 2d). These regions are chosen for two reasons: first, they display robust mesoscale eddy activity and, second, they are subject to significant background westerly wind during most seasons of the year (Peixoto and Oort 1992), which is a necessary requirement for SST residuals to force overlying ZSW residuals through the boundary layer stability mechanism, as we shall see below. In all four regions, the spatial phase relationships between these four variables are quantified by examining zonal- and meridional-lag cross-correlation matrices computed over each domain. These cross correlations are com-

1100 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 33 FIG. 2c. As in Fig. 2a but for the Brazil current extension, mapped from 30 to 45 S and 70 to 10 W for the 10- day period bracketing 12 Jun 2000, during the middle of winter when the westerly winds are strong near the current (see Fig. 3d). puted for residuals whose magnitudes exceed half the standard deviation. This allows the cross-correlation matrices to focus on eddies that are robust. In each region, contours of warm SST residuals encircle yellow-to-red colors of high SLH residuals [(a) in Figs. 2a d] as confirmed by zonal- and meridionallag cross-correlation matrices [(d) in Figs. 2a d]. Contours of westward ZSW residuals encircle yellow-to-red colors of warm SST residuals [(b) in Figs. 2a d] as confirmed by zonal- and meridional-lag cross-correlation matrices [(e) in Figs. 2a d]. Contours of cyclonic WSC residuals (positive in the Northern Hemisphere and negative in the Southern Hemisphere) both overlay and are displaced poleward of yellow-to-red colors of high SLH residuals [(c) in Figs. 2a d] as confirmed by zonal- and meridional-lag cross-correlation matrices [(f)

1101 FIG. 2d. As in Fig. 2a but for the ACC as it traverses south of Africa from 30 to 45 S and 10 to 70 E for the 10-day period bracketing 10 Sep 2000, during late winter when the westerly winds are strong near the current (see Fig. 3c). in Figs. 2a d]. The mesoscale SLH, SST, ZSW, and WSC residuals in these four regions range over 0.2 m, 0.8 C, 1.2 m s 1, and 1.5 10 7 Nm 3, associated with residual meridional geostrophic flows ranging over 0.2 m s 1 [i.e., k(g/ f ) SLH, where k is the zonal wavenumber corresponding to a wavelength of 600 km, g is gravity (9.8 m s 2 ), and f is the Coriolis parameter ( 1.0 10 4 s 1 )]. The ZSW response to covarying mesoscale SLH and SST residuals in these eddy fields [(b) in Figs. 2a d] is fundamentally different from the deep diabatic heating scenario (White and Chen 2002) observed on basin

1102 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 33

1103 scales (e.g., White et al. 1998; White 2000a,b, 2001) that yields an MSW response to SST residuals over the extratropical ocean. Here, the ZSW and MSW residuals are artifacts of the scatterometer source. Yet, they are indicative of the influence that mesoscale SST residuals exert on the stability of the planetary boundary layer in the atmosphere and on the resulting air sea turbulent exchange of momentum. The satellite scatterometer winds are derived from radar backscattering cross sections (Li et al. 1989), directly measuring the friction velocity, which is proportional to the square root of wind stress. These friction velocities have been transformed into 10-m winds using the neutral-stability criterion (Weiss 2000). Moreover, since the scatterometer cross section is measured relative to the surface current, the ZSW is measured relative to the surface current (Liu et al. 1979). Thus, the SST/ZSW phase relationship in these mesoscale eddy fields [(b) in Figs. 2a d] can be explained by warm (cool) SST residuals destabilizing (stabilizing) the marine atmospheric boundary layer in the presence of background westerly winds, instigating more (less) stress at the air sea interface compared to that expected under the neutral stability criterion (Friehe et al. 1991). Already, a similar mechanism has been observed operating in instability waves along the equatorial front in the eastern Pacific Ocean (Xie et al. 1998; Chelton et al. 2001). The magnitude of the corresponding increase in wind stress is significant, with warm SST residuals of 0.8 C producing westerly ZSW residuals of 1.2 ms 1 under westerly winds of 6ms 1, or alternatively increasing the otherwise neutrally stable drag coefficient by 40%, that is, nearly doubling the drag coefficient when transiting from a cold eddy to a warm eddy, consistent with the direct measurements of drag coefficient across a front by Friehe et al. This latter estimate of drag coefficient is derived by equating the bulk formula for stress under neutrally stable conditions, where the background wind is augmented by the neutrally stable wind, to the bulk formula for stress under stable or unstable conditions, where the background wind remains unchanged but the drag coefficient is altered. Thus, a neutrally stable drag coefficient of 1.4 10 3 increases (decreases) to 2.0 10 3 ( 1.0 10 3 ) when a background wind of 6 ms 1 is directed over a warm (cool) mesoscale eddy that is 0.8 C warmer (cooler) than the surrounding ocean. 5. Response of ZSW residuals to SST residuals over the extratropical global ocean To determine where over the extratropical global ocean the SST residuals in the mesoscale eddy field drive ZSW residuals directly overhead, we display the distribution of pattern correlation between the two variables over both Northern and Southern Hemispheres (Fig. 3). The pattern correlations are computed over a 10 latitude longitude box centered at each grid point for the 10-day period of 14 March 2000 in the Northern Hemisphere (Figs. 3a,b) and for the 10-day period of 10 September 2000 in the eastern Atlantic and Indian ocean sectors of the Southern Ocean and of 12 June 2000 for the South Pacific Ocean (Figs. 3c,d). We also display corresponding distributions of pattern correlations between SLH and SST residuals, together with those for average zonal wind for each 10-day period. This reveals significant positive pattern correlation between SLH and SST residuals, and between SST and ZSW residuals, where the background westerly winds exceed 4 ms 1. These significant correlations transcend the western boundary current extensions and the ACC examined in Fig. 2, extending into the interior ocean and the eastern boundary region. Thus, SST residuals in the mesoscale eddies drive ZSW residuals nearly everywhere over the extratropical oceans where robust westerly winds and mesoscale eddies can be found. Thus, even though the 10-day background westerly wind field is basin scale, the mesoscale SLH/SST residuals produces a wind stress field that has a significant mesoscale component. We also can see easterly trade winds encroaching into the extratropics in Fig. 3. Where this occurs, significant negative pattern correlations can be seen between SST and ZSW residuals (Fig. 3), indicating that negative (positive) ZSW residuals overlie warm (cool) SST residuals for mesoscale eddies in the trade wind belt. 6. Magnitude of the residual WSC feedback on extratropical mesoscale eddies The baroclinic Rossby wave equation, which nominally governs the propagation of extratropical mesoscale eddies (e.g., Mizuno and White 1983), can be written as FIG. 3. (a) The distributions of pattern correlations over the Kuroshio Oyashio current extension between mesoscale SLH and SST residuals, and SST and ZSW residuals, together with the average zonal wind for 10 days bracketing 14 Mar 2000. (b) As in (a) but for the Gulf Stream current extension, (c) as in (a) but for the Antarctic Circumpolar Current for 10 days bracketing 10 Sep 2000, and (d) as in (a) but for the Brazil current extension for 10 days bracketing 12 Jun 2000. Pattern correlations are computed over a 10 lat lon box at each grid point. Correlations of 0.6 are significant at 90% confidence level for 8 effective degrees of freedom in the box. Additional confidence is gained by adjacent correlations of similar sign and magnitude. Contour levels for the pattern correlations are 0.2, with positive correlations shaded and negative correlations unshaded. Correlations greater than 0.6 are shaded darker. Contour levels for average zonal wind are 2 m s 1, with mean westerly wind unshaded and mean easterly wind shaded.

1104 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 33 / t K C / x R ( / 0)[WSC /( 0 f )], (6.1) where represents the SLH residual; C R is the baroclinic Rossby wave phase speed, which includes the meridional gradient of the mean potential vorticity (e.g., Yang 2000); is the density difference between the upper ocean above the main pycnocline and the deep ocean below, here computed over the upper 3000 m of ocean based on the Levitus climatology (Levitus et al. 1998); 0 is the mean density in the ocean; f is the Coriolis parameter, where f k residual relative vorticity; and K 1 is the e-folding dissipation timescale. The divergence of the residual Ekman flow [i.e., ( / 0 )WSC /( 0 f ) in Eq. (6.1)] produces Ekman pumping, which contributes to the residual SLH tendency (i.e., / t). It competes with both the Rossby wave propagation (i.e., C R / x) and the dissipation (i.e., K ). Normally, the background westerly winds do not produce WSC residuals large enough to influence the mesoscale vorticity balance in Eq. (6.1) due to the mismatch in spatial scale between the two fields. The question now is whether the magnitude of mesoscale WSC residuals in Figs. 2a d are large enough to affect the residual SLH tendency in any significant way. To answer this question directly, we plot the zonal profile of the observed residual SLH tendency along a constant latitude near the core of the western boundary current extensions and the ACC south of Africa, together with that expected from residual Ekman pumping [i.e., ( / 0 )WSC /( 0 f )], both displayed in units of m s 1 (Fig. 4). This finds the rms of residual Ekman pumping to be 66%, 20%, 29%, and 19% of that in the observed residual SLH tendency in the Kuroshio Oyashio current extension, the Gulf Stream current extension, the Brazil current extension, and the ACC south of Africa, respectively. The effect can be much larger, or smaller, for individual mesoscale eddies in the different current regimes. Note as well, this favorable comparison extends into the interior ocean near the eastern edge of the profile (Fig. 4) in the Kuroshio Oyashio current extension, the Gulf Stream current extension, and the Brazil current extension. Looking back at the meridional- and zonal-lag crosscorrelation matrices of SLH and WSC residuals in Figs. 2a d, we can see mesoscale cyclonic (anticyclonic) WSC residuals consistently displaced poleward (equatorward) of positive SLH residuals in both Northern and Southern Hemispheres, as indicated in the accompanying cartoon (Fig. 5). Since the mesoscale WSC residuals can be expected to produce residual Ekman pumping via Eq. (6.1), this north south alignment of the WSC residuals with respect to the warm mesoscale eddies (Fig. 5) will tend to displace the mesoscale eddies equatorward, that is, pump them up (down) on the poleward (equatorward) side of a typical anticyclonic eddy. The same goes for the cyclonic eddies. Thus, the feedback by the residual mesoscale WSC residuals to the mesoscale eddy field is to produce equatorward coupling phase speeds throughout. Conducting a scaling analysis by relating the residual zonal surface wind stress to the residual ZSW through the bulk formula, and relating the residual ZSW to residual SST and then to residual SLH through their observed statistical associations, finds the meridional coupling phase speed to be on the same order (i.e., 0.01 m s 1 ) as the zonal Rossby wave phase speed [i.e., C R is Eq. (6.1)]. Moreover, since this displacement of the WSC residuals from the SLH residuals is generally less than 90 of phase [(f) in Figs. 2a d], they can also be expected to suppress the intensity of the eddies, enhancing their intrinsic dissipation. 7. Discussion and conclusions We examine mesoscale eddies over the extratropical global ocean on period scales of 6 18 months and wavelength scales 400 1200 km in satellite SLH, SST, and ZSW residuals for 18 months from July 1999 to December 2000. We find covarying SLH and SST residuals in these eddies associated with ZSW residuals when and where the eddies underlie background westerly winds of basin scale. We focus on the eddies located in the Kuroshio Oyashio current extension, the Gulf Stream current extension, the Brazil current extension, and the ACC south of Africa. Since mesoscale eddies there are driven primarily by shear instability in these currents, the link between SST and ZSW residuals there is interpreted to be a response of the latter to the influence by the former on the stability of the marine atmospheric boundary layer, as observed by Friehe et al. (1991) and others. This alters the surface wind stress associated with background westerly winds of basin scale, giving it a mesoscale component. Since the ZSW residuals derive from scatterometer frictional velocities, eastward (westward) ZSW residuals arise over warm (cool) SST residuals in the presence of background westerly winds, with the opposite occurring over background easterly trade winds. A similar result has been observed with scatterometer-derived winds in instability waves observed in the eastern equatorial Pacific Ocean (Xie et al. 1998; Chelton et al. 2001). In the western boundary current extensions and in the ACC south of Africa, we find warm SST residuals of 0.8 C producing westerly ZSW residuals of 1.2 ms 1 under background westerly winds of 6 ms 1, or alternatively increasing the otherwise neutrally stable drag coefficient by 40%. This amounts to increasing the neutrally stable drag coefficient of 1.4 10 3 to 2.0 10 3 over warm eddies and decreasing it to 1.0 10 3 over cold eddies. Since the evolution of the mesoscale eddies are governed largely by baroclinic Rossby wave dynamics, stable or not, the obvious question is whether these residual ZSW responses produce mesoscale WSC residuals large enough to effect a sig-

1105 FIG. 4. Zonal profiles of SLH tendency ( SLH/ t) and Ekman pumping [( / )WSC/( 0 f )] at representative latitudes from each of the four regions depicted in Figs. 2a d, each profile spanning the domain. (a) Zonal section from Fig. 2a at 40.1 N. (b) Zonal section from Fig. 2b at 42.6 N. (c) Zonal section from Fig. 2c at 34.9 S. (d) Zonal section from Fig. 2d at 37.4 N. The rms of the residuals for each zonal profile is given in the inset. nificant feedback to the mesoscale eddies themselves, as observed by Chelton et al. (2001) in tropical instability waves in the eastern equatorial Pacific Ocean. Coupled Rossby waves have been observed operating on annual, biennial, and interannual period scales of basin-scale dimensions (White et al. 1998; White 2000a,b, 2001). Therein, surface wind anomalies have been observed responding to SST anomalies according to the deep diabatic heating scenario (e.g., White and Chen 2002), wherein anomalous SST-induced convection extends upward into the mid- to upper-level troposphere, yielding a deep troposphere response. However, on mesoscales, the residual SST-induced convection appears confined to the marine atmosphere bound-

1106 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 33 FIG. 5. Schematic diagram of an anticyclonic (warm) mesoscale eddy in a background westerly wind field, displaying the increase in scatterometer-derived (neutrally stable) ZSW over the warm SST residual in the eddy. This yields corresponding WSC residuals, cyclonic north of the eddy and anticyclonic south of the eddy, which drive residual SLH tendencies that tend to displace the eddy toward the equator. In reality, the warm eddy decreases the stability of the planetary boundary layer in the atmosphere, thereby increasing the surface frictional velocity associated with the background westerly winds. This enhanced frictional velocity measured by the satellite scatterometer yields greater neutrally stable ZSW at 10-m height. ary layer (e.g., Friehe et al. 1991), where it alters the efficiency of the air sea momentum transfer exerted by the background wind. Thus, any feedback from atmosphere to ocean, driven by the associated mesoscale WSC residuals, would be expected to generate coupled Rossby wave activity among the mesoscale eddies very different from that observed in the basin-scale Rossby waves (e.g., White et al. 1998). The resulting feedback from atmosphere to ocean acts through the mesoscale WSC residuals to produce an Ekman pumping, which can alter the baroclinic Rossby wave equation that governs the eddy activity. We find the rms of residual Ekman pumping to range from 19% to 66% of that in the residual SLH tendency for mesoscale eddies in the western boundary current extensions and in the ACC south of Africa. The effect can be much larger, or smaller, for individual mesoscale eddies in these current regimes. The feedback is greatest in the Kuroshio Oyashio current extension and smallest in the ACC south of Africa. Moreover, we find the spatial phasing of this Ekman pumping feedback acting, on average, to displace the eddies equatorward, and to suppress their amplitude. A simple scaling argument finds the meridional coupling phase speed (i.e., 0.01 m s 1 ) to be on the same order as the Rossby wave phase speed. It remains to construct a coupled Rossby wave model for this brand of mesoscale coupling, operating it under realistic conditions over an entire year to determine the cumulative effect that it has on individual eddies. Since we find this mesoscale boundary layer coupling extending into the interior ocean as well, it remains to determine the influence that it has on the evolution of mesoscale eddies over the entire globe. Acknowledgments. This research was supported by the National Aeronautics and Space Administration (NASA) under Contract JPL 1205106 and by the National Science Foundation (OCE-9920730). Warren White is also supported by the Scripps Institution of Oceanography. Our thanks extend to Andrea Fincham who developed the final figures. REFERENCES Bennett, A. F., and W. B. White, 1986: Eddy heat flux in the subtropical North Pacific. J. Phys. Oceanogr., 16, 728 740. Bernstein, R. L., and W. B. White, 1977: Zonal variability in the distribution of eddy energy in the mid-latitude North Pacific Ocean. J. Phys. Oceanogr., 7, 123 126., and, 1982: Meridional eddy heat flux in the Kuroshio extension current. J. Phys. Oceanogr., 12, 154 159. Chelton, D. B., and Coauthors, 2001: Observations of coupling between surface wind stress and sea surface temperature in the eastern tropical Pacific. J. Climate, 14, 1479 1498. Ducet, N., P. Y. Le Traon, and G. Reverdin, 2000: Global high resolution mapping of ocean circulation from the combination of TOPEX/ POSEIDON and ERS-1/2. J. Geophys. Res., 105, 19 477 19 498. Emery, W. J., and R. E. Thomson, 2001: Data Analysis Methods in Physical Oceanography. Elsevier Science, 638 pp. Friehe, C. A., and Coauthors, 1991: Air sea fluxes and surface layer

1107 turbulence around a sea surface temperature front. J. Geophys. Res., 96, 8593 8609. Greenhut, G. K., 1982: Stability dependence of fluxes and bulk transfer coefficients in a tropical boundary layer. Bound.-Layer Meteor., 24, 253 264. Guymer, T. H., J. A. Businger, K. B. Katsaros, W. J. Shaw, P. K. Taylor, W. G. Large, and R. E. Payne, 1983: Transfer process at the air-sea interface. Philos. Trans. Royal Soc. London, 308A, 253 273. Hill, K. L., I. S. Robinson, and P. Cipollini, 2000: Propagation characteristics of extratropical planetary waves observed in the ATSR global sea surface temperature record. J. Geophys. Res., 105, 21 927 21 945. Hughes, C. W., 1995: Rossby waves in the Southern Ocean: A comparison of TOPEX-POSEIDON altimetry with model predictions. J. Geophys. Res., 100, 15 933 15 950., 1996: The Antarctic circumpolar current as a waveguide for Rossby waves. J. Phys. Oceanogr., 26, 1375 1387., M. S. Jones, and S. Carnochan, 1998: Use of transient features to identify eastward currents in the Southern Ocean. J. Geophys. Res., 103, 2929 2943. Kaylor, R. E., 1977: Filtering and decimation of digital time series. Institute of Physical Science and Technology, University of Maryland Tech. Rep. BN 850, 14 pp. Large, W. G., and S. Pond, 1981: Open ocean momentum flux measurements in moderate to strong winds. J. Phys. Oceanogr., 11, 324 336. Le Traon, P. Y., G. Dibarboure, and N. Ducet, 2001: Use of a highresolution model to analyze the mapping capabilities of multiplealtimeter missions. J. Atmos. Oceanic Technol., 18, 1277 1288. Levitus, S., and Coauthors, 1998: Introduction. World Ocean Database 1998, Vol. 1, NOAA Atlas NESDIS 18, 346 pp. Li, F., W. Large, W. Shaw, E. J. Walsh, and K. Davidson, 1989: Ocean radar backscatter relationship with near-surface winds: A case study during FASINEX. J. Phys. Oceanogr., 19, 342 353. Liu, W. T., K. B. Katsaros, and J. A. Businger, 1979: Bulk parameterization of air sea exchanges of heat and water vapor, including the molecular constraints at the interface. J. Atmos. Sci., 36, 1722 1735. Mizuno, K., and W. B. White, 1983: Annual and interannual variability in the Kuroshio Current system. J. Phys. Oceanogr., 13, 1847 1867. Peixoto, J. P., and A. H. Oort, 1992: Physics of Climate. American Institute of Physics, 520 pp. Qiu, B., K. A. Kelly, and T. M. Joyce, 1991: Mean flow and variability in the Kuroshio extension from Geostat altimetry data. J. Geophys. Res., 96, 18 491 18 507. Smith, E. A., J. Vazquez, A. Tran, and R. Sumagaysay, 1996: Satellitederived sea surface temperature data. American Geophysical Union, Electronic publ. 95274e. Snedecor, G. W., and W. G. Cochran, 1980: Statistical Methods. Iowa State University Press, 507 pp. Tai, C.-K., and W. B. White, 1990: Eddy variability in the Kuroshio extension as revealed by GEOSAT altimetry: Energy propagation away from the jet, Reynolds stress, and seasonal cycle. J. Phys. Oceanogr., 20, 1761 1777. Talley, L. D., and W. B. White, 1987: Estimates of time and space scales of 300 m temperature in the mid-latitude North Pacific from the TRANSPAC XBT program. J. Phys. Oceanogr., 17, 2168 2188. Weiss, B., 2000: Level 2B Data Software Interface Specification (SIS- 2). Internal JPL SeaWinds Project Doc. 686-644-3 Rev B, JPL Doc. D-16079. White, W. B., 2000a: Coupled Rossby waves in the Indian Ocean on interannual timescales. J. Phys. Oceanogr., 30, 2972 2988., 2000b: Tropical coupled Rossby waves in the Pacific ocean atmosphere system. J. Phys. Oceanogr., 30, 1245 1264., 2001: Evidence for coupled Rossby waves in the annual cycle of the Indo Pacific Ocean. J. Phys. Oceanogr., 31, 2944 2957., and R. L. Bernstein, 1979: Design of an oceanographic network in the Midlatitude North Pacific. J. Phys. Oceanogr., 9, 592 606., and S.-C. Chen, 2002: Thermodynamic mechanisms responsible for the troposphere response to SST anomalies in the Antarctic circumpolar wave. J. Climate, 15, 2577 2596., Y. Chao, and C.-K. Tai, 1998: Coupling of biennial oceanic Rossby waves with the overlying atmosphere in the Pacific basin. J. Phys. Oceanogr., 28, 1236 1251. Xie, S.-P., M. Ishiwatari, H. Hashizume, and K. Takeuchi, 1998: Coupled ocean-atmosphere waves on the equatorial front. Geophys. Res. Lett., 25, 3863 3866. Yang, H., 2000: Evolution of long planetary wave packets in a continuously stratified ocean. J. Phys. Oceanogr., 30, 2111 2123.