In situ and remote sensing signature of meddies east of the mid-Atlantic ridge

In situ and remote sensing signature of meddies east of the mid-Atlantic ridge I. Bashmachnikov, 1 F. Machı´n, 2 A. Mendonc¸a, 1 and A. Martins 1 Received 18 July 2008; revised 21 February 2009; accepted 9 March 2009; published 19 May 2009. [1]Mediterranean Water eddies (meddies) are thought to play an important climatic role. Nevertheless, their dynamics are not sufficiently known because of difficulties encountered in their observation. Though propagating below the main thermocline, a number of pieces of evidence of sea surface manifestation of meddies are collected. The present work is based on joint in situ and altimetry data analyses to prove that the meddies can be followed with remote sensing data for long periods of time. The in situ observations are based on data from an oceanographic cruise, which crossed three meddies, and reanalysis of historical data sets, including RAFOS floats paths. Suggested methodology permitted us to obtain uninterrupted tracks for several meddies for a period from several months to more than 2 years. It was found that the dynamically calm region to the north of the Azores current presents favorable conditions for meddy tracking. The meddy surface signal may become shattered and difficult to follow during interaction with a strong dynamic structures (the Azores current/surface vortexes) or peaking topography. Theoretical considerations support the observations and lead to the conclusion that the dynamic signature of meddies at the sea surface is an intrinsic property of meddy dynamics. Citation: Bashmachnikov, I., F. Machı´n, A. Mendonc¸a, and A. Martins (2009), In situ and remote sensing signature of meddies east of the mid-Atlantic ridge, J. Geophys. Res.,114, C05018, doi:10.1029/2008JC005032. 1. Introduction [2] Mediterranean Water eddies (meddies) represent warm salty anticyclonically rotating lenses of the modified Mediterranean Water (MW). In the subtropical NE Atlantic meddies typically have horizontal dimensions between 40 and 150 km, and, in vertical, may have one or two cores situated at 700–900 m and at 1000–1200 m. Close to the Iberian coast, their salt and temperature anomalies often exceed 1 practical salinity unit (psu) and 4°C, respectively, but tend to decrease as the meddy progresses away from the generation region. The core of a meddy is typically highly mixed and characterized by low potential vorticity [Richardson et al., 2000]. [3] Meddies are generated through instability of the Mediterranean underwater current at the Iberian continental slope and the Corringe bank, and then propagated mainly westward or southwestward at middepth gradually loosing their heat and salt contents to the surrounding water. Because of their high stability meddies are met thousands of kilometers away from the generation region (Figure 1) and are quite frequent features in the NE Atlantic [Richardson et al., 1989; Richardson and Tychensky, 1998; Iorga and Lozier, 1999; Siedler et al., 2005]. They are thought to play an important climatic role, through maintenance of a substantial portion of the observed MW salt flux into the ocean. Richardson et al. [1989], Arhan et al. [1994], and Bower et al. [1997], using independent data and analyses techniques, estimated that about 15 to 20 meddies are generated on a single year. The associated salt transport may support up to 50% of the observed MW salt flux. At the same time, their observations were concentrated at the southern and central parts of the Iberian Peninsula and did not cover some important regions of meddy formation, as Portima˜o canyon or Gorringe bank [Serra and Ambar, 2002]. Having analyzed a number of hydrographical sections from the Iberian basin, Maze et al. [1997] did not find any stable westward advection patterns west of 12°W, thus, concluding that 100% of the MW salt flux is related to meddy transport. This agrees with the results by Shapiro and Meschanov [1996], who also noted that some of the MW flux around the northern flank of the Josephine seamount can equally be related to advection. [4] The wide range of the estimates above is partly a result of uncertainty in number of meddies generated per year, as well as, insufficient knowledge of their life histories and propagation patterns. The establishment of an in situ observational network is impeded by the comparatively small spatial scales of the eddies, consequently their remote detection might be helpful. In fact, although propagating below the main thermocline, meddies often have a clear surface signature (Table 1). The last column of Table 1 shows that meddies surface dynamic signal in most cases exceeds 50% of that in the core region at middepth. If this signature is sufficiently stable, remote sensing techniques can become a useful tool to track the meddies. JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 114, C05018, doi:10.1029/2008JC005032, 2009 1 Institute of Marine Research, Department of Oceanography and Fisheries, University of the Azores, Horta, Faial, Azores, Portugal. 2 Institut de Cie`ncies del Mar, CSIC, Barcelona, Spain. Copyright 2009 by the American Geophysical Union. 0148-0227/09/2008JC005032 C05018 1of16 [5] The idea was first expressed by Ka¨se and Zenk [1987], who detected over a meddy anticyclonic looping trajectories of surface satellite tracked drifters. Afterward, the results were supported by numerical modeling [Ka¨se et al., 1989]. Later, Stammer et al. [1991], using Geosat altimetry data, was able to follow several meddies in the Iberian basin. The strongest one could be tracked for more than 1 year. Oliveira et al. [2000] showed that all four Iberian basin meddies, detected with in situ data during 1994, were also visible in the Topex/Poseidon (T/P) ground tracks. On along-track sea level profiles the meddies could be seen as positive bell-like sea level anomalies with about 10 cm of amplitude. For some of the tracks the anomalies resided at the same place for up to 20–30 days, and could be repeatedly resampled. The signal was exceptionally clear even when a track crossed a meddy 50–60 km away from the center. At the very side of a meddy (90–100 km away), the meddy surface signature, still visible, lowered to the level of the background noise and was difficult to detect. Tournadre [1990] estimated that the probability to see an eddy from altimeter data depends on satellite cross-track distance, repetitivity of the track, as well as, eddy diameter and propagation speed. The author predicted that the joint missions of Geosat and T/P are able to detect 95% of 130 km rings and 60% of 100 km rings, propagating at the speed of 2–10 cm s 1 , with 80% probability. The similar computations, accepted as Archiving, Validation, and Interpretation of Satellite Oceanographic data (AVISO) recommendations, state that robust detection of mesoscale structures can be made with at least three altimetry satellites on orbit (with both, T/P and ERS orbit parameters), while one satellite is generally insufficient for mesoscale eddy detection (http://www.aviso.oceanobs.com/en/altimetry/ multi-satellites/index.html). This means, that a robust detection of meddies is possible at least since 2000, when data from T/P, ERS and GFO satellites could be merged. [6] In the present work we establish observational evidence that most of the historical meddies did have a surface signal sufficiently stable and pronounced to be followed Figure 1. Mediterranean Water (MW) salt tongue at 1000 m [Antonov et al., 2006]. Color scale represents water salinity. Bathymetry is represented with 2000 m depth contour (white line). Large white squares mark the center positions of some meddies from RAFOS float trajectories for the dates specified (see also Tables 1 and 2). Black/white arrows are the altimetry-derived currents for the same period (the reference vector is also shown). Black/white dots represent remotely tracked center positions of the corresponding vortices. Black squares are the result of the exact correspondence of the remote/in situ derived centers. Dashed straight lines represent the approximate position of the AzC jet. Black empty circles mark the positions of three meddies observed during OPALINA cruise. C05018 BASHMACHNIKOV ET AL.: REMOTE SENSING SIGNATURE OF MEDDIES 2of16 C05018 with altimetry for long periods of time, as well, as discuss the background conditions which should be met for successful continuous tracking. 2. Materials and Methods [7] Within the framework of Ocean Dynamics and related Productivity of the Northeast Subtropical Atlantic Near the Azores region using ENVISAT, ERS, SeaWiFS, NOAA, and in situ data (OPALINA) (PDCTE/CTA/49965/2003) project, an oceanographic cruise aboard the R/V Arquipelago was accomplished in August 2005 (Figure 2). The region of observations is in the vicinity of the mid-Atlantic ridge (MAR), between the Azores archipelago and the Azores current (AzC). Thirty six conductivity-temperature-depth profilers (CTD) casts were made down to 2000 m depth, the data further processed with the standard SBE processing procedure. During the cruise three meddies were identified as positive temperature-salinity anomalies and a negative potential vorticity (PV) anomaly. Here, Ertel PV is computed as [Pedlosky, 1987; Pingree and Le Cann, 1993]: PV ¼N2g*zþfðÞ;ð1Þ where gis gravity acceleration, Nis buoyancy frequency and fis Coriolis parameter. For typical azimuthal current speeds in meddies of around 20 cm s 1 and core radius of 20 km, relative vorticity z1*10 5 s 1 f8*10 5 s 1 , and to the first order of accuracy the PV can be estimated as the buoyancy frequency normalized by g*f. [8] From the CTD casts geostrophic currents were computed, using an inverse box model technique. The method permits to compute absolute geostrophic velocity fields consistent with both the thermal wind equation and conservation of water properties, as mass, salt, etc. [Wunsch, 1996]. Water proprieties are assumed to be preserved inside each of the selected isoneutral density layers [Jackett and McDougall, 1997], presented in Figure 3, as well as in the domain as a whole. The basic formulation of the mass conservation equation is: ZZ A rvrdxdy þZZ A rvodxdy ¼0;ð2Þ where ris the density field, Ais the area between selected density layers and hydrographic stations. In the equation, the geostrophic velocity field is divided into a reference level velocity (v o ) and a velocity relative to this level (v r ). The reference level is chosen as the layer with expected minimum (but nonzero) current velocity. In absence of other information, v o is computed to compensate relative transport imbalances within the water column. Large v o are considered to result from ageostrophic component of the flow, for which the thermal wind equation fails. To avoid the related errors, v o values should be limited to a certain range. Evaluation of possible errors in v o , e.g., possible accuracy to which the conservation equation could be satisfied, is performed through a priory estimation of the covariance of the noise (R nn ) and the covariance of the unknowns (R xx ). R xx is estimated through the variance in the direct velocity measurements in the eastern North Atlantic [Mu¨ller and Siedler, 1992; Machı´n et al., 2006], which for the reference level g ref = 27.922 (the lower interface for Mediterranean Water) gives a quite small value of (0.02) 2 m 2 s 2 . In this case our data shows an AzC transport of 10 ± 1 Sv, Table 1. Observations of Surface Signatures of Meddies a Reference Meddies’ Names Instrumentation Showing Deep and/or Surface Signatures R(km) Vc (cm s 1 )Vs (cm s 1 )Vs/Vc (%) At the Formation Region (Iberian Basin) Ka¨se et al. [1989] model Numerical model <50 10 7 70 Stammer et al. [1991] A CTD, SLA, mooring N.D. 5 3 60 B CTD, SLA, mooring N.D. 7 2 29 D CTD, SLA, mooring N.D. 12 4 33 Pingree and Le Cann [1993] Smeddy CTD, XBT, PF, SST 13 20 8 40 Schultz Tokos et al. [1994] Aska (A) CTD, RAFOS, SF 17 23 12 56 B1 CTD, RAFOS, SF 25 25 10 40 B2 CTD, RAFOS, SF 30 31 18 58 Oliveira et al. [2000] A1 RAFOS, SF, SLA, SST 18 23 23 100 A2 RAFOS, SF, SLA, SST 37 20 17 83 A3(Pinball) RAFOS, SF, SLA, SST 17 23 23 100 Paillet et al. [2002] Ulla CTD, XBT, LADCP, RAFOS, DDB, SF 15 17.5 7.5 43 Away From the Formation Region Tychensky and Carton [1998] (south of the Azores) Hyperion CTD, XBT, SF, SLA <60 22.5 18 80 Ceres CTD, XBT, SF, SLA <50 23 30 130 b Encelade CTD, XBT, SF, SLA <70 16 6 38 Le Cann et al. [2005] (Azores-Biscay Rise) A2 CTD, RAFOS, PF, SF <40 15 13 87 Mean values 21 18 13 72 a Ris the radius of maximum azimuthal velocity at the depth of the meddy core (tentative values), Vc is the maximum azimuthal velocity value at the meddy core, and Vs is the maximum azimuthal velocity at the sea surface (meddy signal). CTD, conductivity-temperature-depth profilers; SLA, sea level anomalies; N.D., no data available; XBT, expandable bathythermograph profilers; PF, profiling floats; SST, sea surface temperature; SF, surface floats; LADCP, Lowered Acoustic Doppler Profiler; DDB, deep-drogued floats. b Maximum at the surface is due to surface vortex, aligned with the meddy. C05018 BASHMACHNIKOV ET AL.: REMOTE SENSING SIGNATURE OF MEDDIES 3of16 C05018 corresponding well to the estimates in [Siedler and Onken, 1996]. As a sensitivity test, we have tried another reference level located at g ref = 27.200 (the lower interface of North Atlantic Central water). This results in the residuals of one order higher: (0.23) 2 m 2 s 2 , and small AzC transport of 5 ± 5 Sv. Thus, from the two reference levels, only g ref = 27.922 is acceptable. Similarly, the noise uncertainty in the salt anomaly equation is computed as a function of the noise uncertainty in the mass and salt variability equations [Ganachaud, 2003]. In mass equations R nn can be estimated as (2.2*10 9 ) 2 kg 2 s 2 , which indicates that the noise could be the same order as the mass transports in the region, and the results of mass conservation will have comparatively low reliability. [9] The in situ data are supplemented with AVISO altimetry data set (AVISO data available at http://las.aviso. oceanobs.com/las/servlets/dataset). In the data set, the absolute dynamic topography is produced by adding timeindependent mean dynamic topography to the sea level anomalies from the corrected altimeter measurements. Geostrophic currents are computed from the sea level regular grid, obtained by objective space-time interpolation of various satellite tracks. The resultant regular space-time grid has 1/3 of degree mesh interval and weekly time step. The spatial-temporal resolution is good enough to trace surface signals of meddies. The meddy tracing is done with the method of ‘‘crawling squares,’’ which proved to be rather stable in tracing corresponding altimetry anomalies [Bashmachnikov et al., 2009]. The method derives the next (or previous) meddy position as a local minimum of the second derivative of sea level height (SLH), not more than one grid point away from the one obtained at a given time step. Using the second derivative, instead of SLH, allows incorporation in the computations of five SLH grid points (in zonal and meridional directions), instead of one. This ensures that we follow some integral part of the structure and adds to the stability of the track. Dynamically, second derivatives from SLH can be regarded as a proxy for relative vorticity. We also consider the situation, when at a certain moment the meddy surface signal fails to appear, Figure 2. CTD stations of OPALINA cruise (diamonds) overlaid on AVISO altimetry-derived geostrophic currents centered at 17 August 2005. The scale is relative vorticity in 10 5 s 1 . The suggested meddies’ positions are marked with letter ‘‘M.’’ White dotted lines represent the 2000 m bathymetry contours. The Azores islands are also shown. C05018 BASHMACHNIKOV ET AL.: REMOTE SENSING SIGNATURE OF MEDDIES 4of16 C05018 for example, seeded in between altimeter tracks. To avoid unwanted jumps to another closest minimum two precautions are undertaken. First, the SLH derivative filed is composed from three corresponding consecutive fields, summed with decreasing weights as we step back in time. Second, a limitation is imposed on the signal propagation speed, e.g., the next tracking point cannot move more than one grid point away from the previous meddy position. This limitation serves to avoid sudden jumps to another anticyclonic structure when, at a particular moment, the surface signature of a meddy becomes hardly visible. With 7-day AVISO time step, this means that we consider meddies to have propagation velocities not exceeding 4 cm s 1 . Most frequently observed velocities of meddy propagation usually do not reach this value. Still, more exclusive cases of short-period fast translations may occur, when a meddy can move with the speed of up to 10 cm s 1 [Richardson and Tychensky, 1998]. To catch up with the meddy in this situation, a possibility is foreseen for a track to jump over up to three grid points (to the closest vorticity minimum), when the SLH second derivative at the tentative newly found tracking point abruptly falls to less than 50% of its previous value. 3. In Situ Signatures of the Observed Meddies [10] The general water structure and its spatial variability at the time of OPALINA cruise are investigated. TS diagrams show that the CTD stations can be grouped into those with lower salinities below the main thermocline and those with higher ones (Figure 4, gray and black dots, respectively). The first region (stations 101–109) encompasses the southwestern part of the Azores plateau and the area to the west of MAR, and the second one (stations 110–133) covers the region south of the Azores and east of MAR (Figure 2). This reflects the known MW distribution pattern [Richardson et al., 2000]. In the inverse model each of the subregions is treated as an ‘‘independent’’ basin with its own conservation balance. Vertically, the water column is separated in two layers with independent conservation balances: the main thermocline layer (26.800 < g n < 27.200) of the North Atlantic Central Water (NACW), and Figure 3. Isopycnals of neutral density (g n ) along the cruise box. The 27.922 isopycnal is taken as the reference level for inverse model computation. Upper ticks indicate the stations numbering, while black vertical lines indicate the limits of CTD sections. N, northern; W, western; S, southern; E, eastern; M, meddies. C05018 BASHMACHNIKOV ET AL.: REMOTE SENSING SIGNATURE OF MEDDIES 5of16 C05018 the intermediate level (27.200 < g n < 27.922) of the Subarctic Intermediate Water (SAIW), Mediterranean Water (MW) and Labrador Sea Water. As it is already discussed above, the base of the intermediate water level (g n = 27.922, at about 1600 m deep) is assumed to be the best choice for the reference level. At this level isopycnals have minimum tilt along the box boundaries (Figure 3), and computed v o ’s are small and stable along all the sections (0.5 ± 2 cm s 1 ). [11] The computed geostrophic currents are binned into five density layers (corresponding approximate depth layers are given in brackets): surface–26.85 (0–300 m), 26.85– 27.20 (300–600 m), 27.20–27.62 (600–900 m), 27.62– 27.82 (900–1200 m), 27.820–27.922 (1200–1600 m). The results showed that mass conservation is achieved in all the layers. The computed box-mean transport suggested a net eastward mass flux through the area at all the water levels, which primarily is a result of the eastward water transport by the AzC in the southern part of the region (Figure 5). Salt anomaly is not fully conserved in the surface layer (<300 m) and the upper MW layer (600–900 m). Aside of high possible error associated with salt flux computations; this may equally be a sign of unbalanced salt fluxes in and out of the region. Particularly, the inward salt flux across the eastern CTD section may be associated with the meddy entering the ‘‘box,’’ as discussed below. Water transport for the two meridional sections, obtained using the inverse model technique, is presented at Figure 6. In the southern parts of the western and eastern sections an intensive eastward transport in the upper layer is associated with the AzC. [12] The anticyclonic vortex, well pronounced in altimetry, has been crossed by the cruise track only in its eastern part. Even so, the CTD cast inside the vortex (station 131) demonstrates a pronounced positive salinity and temperature anomaly (0.4 psu and 2.0–2.5°C, respectively). The anomaly occupies the layer from 700 to 1000 m, with the maximum values at 900 m (Figures 4, 7a, 7b, and 7c). The salinity-temperature anomalies are coupled with low PV, and indicates that a meddy was crossed (M131). Figures 7a, 7b, and 7c also show, just below the meddy core, two high-salinity side-blobs. Those maxima stretch away from the core, perpendicular to the bottom slope (north of the meddy) and may be a result of meddy interaction with bottom topography. The anticyclonic rotation of the current vectors is centered at station 131, and extends up to the surface and down to the bottommost level (Figure 6). [13] Besides M131, two other meddy-like structures have been observed. The one, centered at station 111 (M111), is characterized by moderate positive salinity and temperature anomalies (0.2 psu and 1.5–2.0°C, respectively), and a twocore structure (Figures 7d, 7e, and 7f). The main core, centered at 1000 m, corresponds to the region of comparatively low PV, though its minimum is shifted to the north. CTD-based geostrophic flows show the dynamic signature Figure 4. TS diagram for OPALINA cruise. The CTD stations from 101 to 108 are presented in gray, and the ones from 109 to 133 are presented in black. Thick lines are the casts inside the meddies M106 (light gray line), M111 (dark gray line), and M131 (black line). The s o density isolines are overlaid. C05018 BASHMACHNIKOV ET AL.: REMOTE SENSING SIGNATURE OF MEDDIES 6of16 C05018 of the meddy to be disguised by the Azores front–current system (stations 111–115) and a southwestward mean flow between stations 110–111 (Figures 2 and 7). At the same time, in the segments of typical AzC influence (stations 112–114, as well as, 116–117 and 120–121 (not shown)), the ratios of the mean current speed in second to fourth vertical layers (300–600 m, 600–900 m and 900–1200 m) to the one in the upper layer (0–300 m) are about 60%, 45% and 20%, correspondingly. Between stations 111–112 the referred ratios are only 35%, 13% and 3%, respectively. This difference, most pronounced in the MW layer, we link to interaction with deep anticyclonic rotation around station 111, provoked by M111. [14] Another meddy-resembling structure is associated with station 106 (M106). It has rather small positive salinity and temperature anomalies (0.10–0.15 psu and 1.0°C, respectively), and a pronounced negative PV anomaly (Figures 7g, 7h, and 7i). The core is situated between 700 and 1000 m depth, but the salinity anomaly extends throughout the water column (from the surface down to 1200– 1400 m). Contrary to stations 105 and 107, where relative vorticity smoothly decreases away from the surface, at station 106 the temperature-salinity anomalies are associated with negative relative vorticity maximum in 300–900 m layer, decreasing in absolute measure toward the surface and toward the bottom. Thus, both thermohaline and current patterns support the suggestion of meddy origin of the structure. 4. Remote Sensing Signatures of the Meddies 4.1. Altimetry Geostrophic Flows via in Situ [15] Geostrophic currents derived from the altimetry observations (Figure 2) are compared with those obtained from the CTD sections. To be compatible, the altimetry currents were averaged and projected in the directions perpendicular to the CTD sections. Furthermore, we divided the eastern section into two segments (from stations 126 to 129 and from stations 129 to 133). The southern segment is the area of influence of the AzC and the northern one is that of the meddy M131. For both segments correlation between the altimetry and in situ currents is generally high throughout the water column. At the same time, for the southern segment the correlations are significant (at 5% significance level) only for the upper 600 m layer and decrease with depth, whereas for the northern segment the correlations increase with depth until reaching the significant maximum in the 600–1200 m layer. Similarly, for the depth-cumulative transports (e.g., the total transport from the sea surface to a Figure 5. (a) Mass fluxes and (b) salt anomaly fluxes obtained with the inverse model for each of the CTD sections. The horizontal lines mark approximate borders of North Atlantic Central Water (NACW), Mediterranean Water (MW), and Labrador Sea Water (LSW), as used for computations. Mass fluxes obtained with the inverse model for (c) NACW (g n < 27.200) and (d) MW level (g n < 27.820). The computations are done for the reference level g n = 27.922. Numbers indicate mass transports in Sv through each of the section; for positive values the fluxes are outward from the box. C05018 BASHMACHNIKOV ET AL.: REMOTE SENSING SIGNATURE OF MEDDIES 7of16 C05018 certain depth), the southern segment correlations reach maximum (100%) in the upper 600-m layer and gradually decrease with increasing layer thickness, while the northern segment correlations monotonically increase with increasing layer thickness. Though these correlations are derived from a small number of data-points, they are significant, stable and consistently changing, proving their validity. [16] The results above suggest that the general current structure is similar throughout the whole 1600-m water column. However, over the meddy, the sea level variability reflects better the baroclinic flow structure at the intermediate water levels than those at the upper ones. This may result from the sea level topography being considerably affected by the barotropic component of the flow, correlated with the deep meddy structure. Equally, as an integral measure of the density structure throughout the water column, the sea level spatial variability may reflect the density anomalies in the intermediate or deep water layers, if those are dominating throughout the water column. 4.2. Surface Signature of Meddy M131, Traced Over 2 Years [17] Figure 2 shows a clear altimetry signal of M131. The data suggest that the meddy center was located 30 to 60 km west of station 131, and the radius of its dynamic influence was around 160–180 km. With the method referred in the previous section, the anticyclonic structure could be traced forward to December 2005, and backward to October– December 2003 (Figure 8). After four months of stagnation to the northeast of the Josephine seamount, during February 2004, the meddy broke down with steady westward propagation. At the stagnation stage the meddy was difficult to detect with altimetry, and its signature became more pronounced as it started its progress to the west. In December 2004 the meddy reached the Azores plateau, east of Santa Maria island. During the period of active movement, the M131 average propagation speed was about 2 km d 1 , e.g., typical for a meddy [Richardson et al., 2000]. In January 2005 the meddy squeezed itself turning around the southern tip of the plateau, and regained a circular form west of Santa Maria in February. Then it started moving southwest along the southern flank of the plateau, generally following the 1000–2000 m isobaths. During June 2005 it trapped a substantial part of the AzC meander which made it much more visible on the surface. This is confirmed by our observations in July 2005: the ship thermometer registered a considerable increase in sea surface temperature of about 0.3–0.6°C at both sides of the surface manifestation of M131 (Figure 9). Since then, the meddy slowly propagated southwest. In October 2005, after merging with another AzC meander, it was quickly translated south, crossing the AzC. At this stage it is very difficult to distinguish the meddy from the meander structures, solely on the basis of altimetry observations. In December 2005 the meddy had already crossed the AzC and was moving westward/ southwestward, leaving the Atlantis seamounts to the east. The M131 trajectory is coherent with one of the previously reported major meddy paths [Shapiro and Meschanov, 1996; Richardson et al., 2000]. [18] The suggested M131 path was checked against CTD-XBT data obtained from the NODC database [Boyer et al., 2006]. Most of these profiles contain only temperature data and do not reach 900 m depth, limiting our analysis to temperature anomalies in the 600–800 m layer, where the Figure 6. Geostrophic water transport (kg s 1 ) across the cruise sections for different depth layers obtained from inverse modeling and altimetry data. (left) Western section, and (right) eastern section. Circles are the CTD the stations. Stations 111 and 131 are the closest to the centers of the corresponding meddies. C05018 BASHMACHNIKOV ET AL.: REMOTE SENSING SIGNATURE OF MEDDIES 8of16 C05018 Figure 7. (a, d, g) Vertical cross section of salinity, (b, e, h) respective salinity anomalies, and (c, f, i) buoyancy frequency (s 1 10 3 ). The CTD sections (eastern (Figures 7a, 7b, and 7c), western (Figures 7d, 7e, and 7f), and northern (Figures 7g, 7h, and 7i)) are presented. Vertical lines and numbers above mark the CTD stations. The cores of meddies M131, M111, and M106, as crossed by the cruise, are marked with dashed ellipses. Lines below schematically represent the bottom topography. Figure 8. Tentative backward (black) and forward (gray) tracing of the M131 center estimated from altimetry data. Diamonds mark the beginning of a month and are labeled with corresponding month and year. Large squares represent the monthly mean positions where significant negative correlation between the in situ water temperature and the distance from the tentative meddy position is obtained. Altimetry currents over the meddy are plotted for the months marked in bold. Reference vector is also shown. C05018 BASHMACHNIKOV ET AL.: REMOTE SENSING SIGNATURE OF MEDDIES 9of16 C05018 of Marine Research (IMAR/DRCT/REF. U&D/MED.M1.1.2/008/2005// BPD/002/007) and a doctoral fellowship sponsored by DRCT together with University of the Azores (UAz/DRCT/REF. U&D/MED.M3.1.1/I/003/ 2005/A) for support of this work. We are especially grateful to the scientific and technical staff of R/V Arquipelago and IMAR-DOP/UAz, for help in preparation and accomplishment of the scientific mission. Special thanks are also due to Ana Filipa (IMAR-DOP/UAz) for help in altimetry data preparation. F. Machı´n is supported by the Juan de la Cierva Programme and CANOA Project (CTM2005–00444/MAR), both funded by the Spanish Ministry of Education and Science. We are also grateful for the anonymous reviewers, whose comments significantly added to the quality of the paper. References Antonov, J. I., R. A. Locarnini, T. P. Boyer, A. V. Mishonov, and H. E. Garcia (2006), World Ocean Atlas 2005, vol. 2, Salinity, NOAA Atlas NESDIS, vol. 62, edited by S. Levitus, 182 pp., NOAA, Silver Spring, Md. Arhan, M., A. C. de Verdie`re, and L. Memry (1994), The eastern boundary of the subtropical North Atlantic, J. Phys. Oceanogr.,24, 1295–1316, doi:10.1175/1520-0485(1994)024<1295:TEBOTS>2.0.CO;2. Bashmachnikov, I., C. Mohn, J. L. Pelegrı´, A. Martins, F. Machı´n, F. Jose, and M. White (2009), Interaction of Mediterranean Water eddies with Sedlo and Seine seamounts, subtropical Northeast Atlantic, Deep Sea Res., Part II, doi:10.1016/j.dsr2.2008.12.036, in press. Bower, A. S., L. Armi, and I. Ambar (1997), Lagrangian observations of meddy formation during a Mediterranean Undercurrent Seeding Experiment, J. Phys. Oceanogr.,27, 2545–2575, doi:10.1175/15200485(1997)027<2545:LOOMFD>2.0.CO;2. Boyer, T. P., J. I. Antonov, H. E. Garcia, D. R. Johnson, R. A. Locarnini, A. V. Mishonov, M. T. Pitcher, O. K. Baranova, and I. V. Smolyar (2006), World Ocean Database 2005, NOAA Atlas NESDIS, vol. 60, edited by S. Levitus, 190 pp., NOAA, Silver Spring, Md. Carton, X. (2001), Hydrodynamical modelling of oceanic vortices, Surv. Geophys.,22, 179–263, doi:10.1023/A:1013779219578. Cenedese, C. (2002), Laboratory experiments on mesoscale vortices colliding with a seamount, J. Geophys. Res.,107(C6), 3053, doi:10.1029/ 2000JC000599. Ganachaud, A. (2003), Large-scale mass transports, water mass formation, and diffusivities estimated from World Ocean Circulation Experiment (WOCE) hydrographic data, J. Geophys. Res.,108(C7), 3213, doi:10.1029/2002JC001565. Iorga, M. C., and M. S. Lozier (1999), Signature of the Mediterranean outflow from a North Atlantic climatology: 2. Diagnostic velocity fields, J. Geophys. Res.,104, 26,011–26,029, doi:10.1029/1999JC900204. Jackett, D. R., and T. J. McDougall (1997), A neutral density variable for the world’s oceans, J. Phys. Oceanogr.,27, 237–263, doi:10.1175/15200485(1997)027<0237:ANDVFT>2.0.CO;2. Ka¨se, R. H., and W. Zenk (1987), Reconstructed Mediterranean salt lens trajectories, J. Phys. Oceanogr.,17, 158–163, doi:10.1175/15200485(1987)017<0158:RMSLT>2.0.CO;2. Ka¨se, R. H., A. Beckmann, and H. H. Hinrichsen (1989), Observational evidence of salt lens formation in the Iberian basin, J. Geophys. Res.,94, 4905–4912, doi:10.1029/JC094iC04p04905. Le Cann, B., M. Assenbaum, J. C. Gascard, and G. Reverdin (2005), Observed mean and mesoscale upper ocean circulation in the midlatitude northeast Atlantic, J. Geophys. Res.,110, C07S05, doi:10.1029/ 2004JC002768. Locarnini, R. A., A. V. Mishonov, J. I. Antonov, T. P. Boyer, and H. E. Garcia (2006), World Ocean Atlas 2005,vol.1,Temperature, NOAA Atlas NESDIS, vol. 61, edited by S. Levitus, 182 pp., NOAA, Silver Spring, Md. Machı´n, F., A. Herna´ndez-Guerra, and J. L. Pelegrı´ (2006), Mass fluxes in the Canary Basin, Prog. Oceanogr.,70(2 –4), 416 –447, doi:10.1016/ j.pocean.2006.03.019. Maze, J. P., M. Arhan, and H. Mercier (1997), Volume budget of the eastern boundary layer off the Iberian Peninsula, Deep Sea Res., Part I,44, 1543–1574, doi:10.1016/S0967-0637(97)00038-1. Mu¨ller, T., and G. Siedler (1992), Multi-year current time series in the eastern North Atlantic Ocean, J. Mar. Res.,50, 63–98. Oliveira, P. B., N. Serra, A. F. G. Fiu´za, and I. Ambar (2000), A study of meddies using simultaneous in-situ and satellite observations, in Satellites, Oceanography and Society,Elsevier Oceanogr. Ser., vol. 63, pp. 125– 148, Elsevier, Amsterdam. Paillet, J., B. Le Can, X. Carton, Y. Morel, and A. Serpette (2002), Dynamics and evolution of a northern meddy, J. Phys. Oceanogr.,32, 55– 79, doi:10.1175/1520-0485(2002)032<0055:DAEOAN>2.0.CO;2. Pedlosky, J. (1987), Geophysical Fluid Dynamics, 2nd ed., 710 pp., Springer, New York. Pingree, R. D. (1995), The droguing of Meddy Pinball and seeding with ALACE floats, J. Mar. Biol. Assoc. U. K.,75, 235–252. Pingree, R. D. (2002), Ocean structure and climate (eastern North Atlantic): In situ measurement and remote sensing (altimeter), J. Mar. Biol. Assoc. U. K.,82, 681–707, doi:10.1017/S0025315402006082. Pingree, R. D., and B. Le Cann (1993), A shallow meddy (a smeddy) from the secondary Mediterranean salinity maximum, J. Geophys. Res.,98, 20,169–20,185, doi:10.1029/93JC02211. Richardson, P. L., and A. Tychensky (1998), Meddy trajectories in the Canary Basin measured during the Semaphore Experiment, 1993– 1995, J. Geophys. Res.,103, 25,029– 25,045, doi:10.1029/97JC02579. Richardson, P. L., D. Walsh, L. Armi, M. Schro¨der, and J. F. Price (1989), Tracking three meddies with SOFAR floats, J. Phys. Oceanogr.,19, 371– 383, doi:10.1175/1520-0485(1989)019<0371:TTMWSF>2.0.CO;2. Richardson, P. L., M. S. McCartney, and C. Maillard (1991), A search for meddies in historical data, Dyn. Atmos. Oceans,15, 241–265, doi:10.1016/0377-0265(91)90022-8. Richardson, P. L., A. S. Bower, and W. Zenk (2000), A census of meddies tracked by floats, Prog. Oceanogr.,45(2), 209– 250, doi:10.1016/S00796611(99)00053-1. Schultz Tokos, K., H. H. Hinrichsen, and W. Zenk (1994), Merging and migration of two meddies, J. Phys. Oceanogr.,24, 2129–2141, doi:10.1175/1520-0485(1994)024<2129:MAMOTM>2.0.CO;2. Serra, N., and I. Ambar (2002), Eddy generation in the Mediterranean undercurrent, Deep Sea Res., Part I,49, 4225 –4243, doi:10.1016/ S0967-0645(02)00152-2. Shapiro, G. I., and S. L. Meschanov (1996), Spreading pattern and mesoscale structure of Mediterranean outflow in the Iberian basin estimated from historical data, J. Mar. Syst.,7, 337 –348, doi:10.1016/09247963(95)00011-9. Siedler, G., and R. Onken (1996), Eastern recirculation, in The Warm Water Sphere of the North Atlantic Ocean, edited by W. Krauss, pp. 339 –364, Gebru¨der Borntraeger, Berlin. Siedler, G., L. Armi, and T. J. Mu¨ller (2005), Meddies and decadal changes at the Azores front from 1980 to 2000, Deep Sea Res., Part II,52, 583– 604, doi:10.1016/j.dsr2.2004.12.010. Stammer, D., H. H. Hinrichsen, and R. H. Ka¨se (1991), Can meddies be detected by satellite altimetry?, J. Geophys. Res.,96, 7005 –7014, doi:10.1029/90JC02740. Tournadre, J. (1990), Sampling oceanic rings by satellite radar altimeter, J. Geophys. Res.,95, 693– 697, doi:10.1029/JC095iC01p00693. Tychensky, A., and X. Carton (1998), Hydrological and dynamical characteristics of Meddies in the Azores region: A paradigm for baroclinic vortex dynamics, J. Geophys. Res.,103, 25,061–25,079, doi:10.1029/ 97JC03418. Wunsch, C. (1996), The Ocean Circulation Inverse Problem, 442 pp., Cambridge Univ. Press, New York.  I. Bashmachnikov (corresponding author), A. Martins, and A. Mendonc¸a, Institute of Marine Research, Department of Oceanography and Fisheries, University of the Azores, P-9901-862, Horta, Faial, Azores, Portugal. ([email protected]) F. Machı´n, Institut de Cie`ncies del Mar, CSIC, Passeig Marı´tim de la Barceloneta 37-49, E-08003 Barcelona, Spain. C05018 BASHMACHNIKOV ET AL.: REMOTE SENSING SIGNATURE OF MEDDIES 16 of 16 C05018