Lateral transports of mass, inorganic nutrients and dissolved oxygen in the Cape Verde Frontal Zone in summer 2017

. The circulation patterns in the conﬂuence of the North Atlantic Subtropical and Tropical gyres delimited by the Cape Verde Frontal Zone (CVFZ) in summer 2017 were examined. Hydrology, dissolved oxygen ( O 2 ) and inorganic nutrients data collected in a closed box embracing the CVFZ allowed estimating transports of water masses, O 2 and inorganic nutrients for the ﬁrst time. Higher transports occurred mainly at the surface and central waters, and were moderately affected by the Cape Verde Front located in the southeastern part of the domain. Thus, the front conditioned the meridional transports, acting 5 as a barrier between North and South Atlantic Central waters. Speciﬁcally, − 3 . 2 ± 1 . 7 Sv entered through north and east and 6 . 7 ± 1 . 7 Sv left through west and south transects. At intermediate levels, the most important source came from the south with − 2 . 2 ± 1 . 5 Sv of modiﬁed Antarctic Intermediate water, moderately affected by the circulation pattern above. The transports of O 2 and inorganic nutrients conditioned by their distributions behaved quite similar to mass transports. The most intense and important transports of O 2 and inorganic nutrients occurred in the deepest layer of central waters and in the shallowest two 10 layers of intermediate waters where inorganic nutrients accumulated and large differences in concentrations of O 2 were found, especially in the deepest layer of central waters between the northeast and southeast zones. In these three layers, transports of O 2 and inorganic nutrients came from the east and south and they left northward and westward. This circulation pattern


INTRODUCTION
The Cape Verde Basin (CVB) is located in the eastern boundary of the North Atlantic Ocean where the subtropical region meets the tropical one. This area is influenced by the south-eastern extension of the North Atlantic subtropical gyre, NASG (Stramma and Siedler, 1988), the north-eastern extension of the North Atlantic tropical gyre, NATG (Siedler et al., 1992), and the upwelling region off NW Africa (Ekman, 1923;Tomczak, 1979;Hughes and Barton, 1974;Hempel, 1982). Inside by the CUF, establish distinct biogeochemical domains with substantial differences in O 2 and inorganic nutrient concentrations at the CVB.
In recent years, several authors have addressed the distribution of O 2 and inorganic nutrients and also their link to the physical processes taking into account the water mass distributions and transports in this margin of the North Atlantic Ocean (Pelegrí et al., 2006;Machín et al., 2006;Pastor et al., 2008;Álvarez and Álvarez-Salgado, 2009;Peña-Izquierdo et al., 2012;15 Pastor et al., 2013;Peña-Izquierdo et al., 2015;Hosegood et al., 2017;Burgoa et al., 2020). These physical processes range from the regional scale, usually associated to advective processes, to mesoscale and smaller scales, usually associated to mixing processes (Pastor et al., 2008). These works consider remineralization as the main biochemical process inside NASG and NATG.
This manuscript aims to address the circulation patterns and the physical processes behind the distribution of O 2 and inor- 20 ganic nutrients at the CVFZ, a domain where in situ data availability has historically been very limited. Secondly, the location and intensity of the CVF in summer is assessed, a front acting both a barrier and a source of meso-and submesoscale variability, having influence on the transports of mass, O 2 and inorganic nutrients. The rest of the manuscript is organized in three main sections. Section 2 presents the methodology, where the sampling performed during the FLUXES-I cruise and the modelling description are given. Then, section 3 presents results in terms of the hydrography and water masses, absolute velocity field, 25 and transports of mass, dissolved oxygen and inorganic nutrients. The manuscript is closed with a discussion of the results and the main conclusions.

The cruise
FLUXES-I cruise was carried out on summer 2017, from July 14th to August 8th aboard the R/V Sarmiento de Gamboa as part 30 of the research project FLUXES (Carbon Fluxes in a Coastal Upwelling System -Cape Blanc, NW Africa). The 35 sampling stations were selected to form a closed box (pink dots in Fig. A1). The rosette sampler was a SBE 38 equipped with 24 Niskin bottles of 12 L. Conductivity-temperature-depth data (CTD) were collected with a SBE 911+ from the sea surface down to more than 2000 m depth with a vertical resolution of 1 dbar. Stations 1, in the northeast corner, and 29, in the southeast corner, were very shallow, so both were discarded from the analyzes. The average distance between neighboring CTD stations was some 84 km. In addition, 39 expendable bathythermograph probes (XBT, T5 by Sippican) were deployed between most CTD stations (blue dots in Fig. A1). In all cases, the XBTs reached down to 2000 m thanks to a specific setup in the WinMK21 acquisition program and to a reduced boat speed during deployment (5 kn). Some XBTs were also removed (12,19,30,31,38,5 39 and 40) due to recording errors. The northern transect (N) spanned from station 2 to 12 at 23 • N; the western one (W) was located at 26 • W from station 12 to 19; the southern zonal transect (S) at 17.5 • N ran from station 19 to 28, while the eastern transect (E) closed the box approximately at 18.57 • W from station 28 to 3.
Practical salinity, S P (UNESCO, 1985) was calibrated by analyzing 51 water samples with a Portasal model 8410A salinometer with an accuracy and precision within the values recommended by WOCE. An oxygen sensor SBE43 was interfaced 10 with the CTD system during the cruise. In total, 417 samples were collected to calibrate the dissolved oxygen sensor on the rossette. The final precision obtained was ±0.53 µmol kg −1 . The dissolved inorganic nutrients analyzed in this work were nitrates (NO 3 ), phosphates (PO 4 ) and silicates (SiO 4 H 4 ). 419 water samples were collected by the Niskin bottles to 25 mL polyethylene bottles in all the stations and depths. The samples were frozen at -20 • C until their analyzes in the base laboratory. NO 3 , PO 4 and SiO 4 H 4 concentrations were determined using a segmented flow Alliance Futura analyzer following the 15 colorimetric methods proposed by Grasshoff et al. (1999).

Remote sensing and climatological datasets
ASCAT on Metop L4 daily global wind field (Bentamy and Fillon, 2012) with a spatial resolution of 0.25 • made available by CERSAT (Centre ERS d' Archivage et de Traitement; http://cersat.ifremer.fr) was employed as the source of wind data to estimate the Ekman transport. Freshwater flux was estimated from the average rates of evaporation and precipitation extracted 20 from the Weather Research and Forecasting model (WRF (Powers et al., 2017)), provided with a spatial resolution of 0.125 • and a temporal resolution of 12 h.
The climatological mean depths of the neutral density field during the summer season were calculated from the climatological temperature and salinity extracted from the World Ocean Atlas 2018 (Locarnini et al., 2018;Zweng et al., 2018, WOA18).
Moreover, WOA18 was also employed to generate a climatological neutral density field during the summer season to estimate 25 a climatological geostrophic velocity field. On the other hand, 4 new stations in the N transect and 12 new stations in the S transect were produced with the WOA18 summer field with the aim to extend the N and S transects of the inverse model to the African coast. SEALEVEL_GLO_PHY_L4_REP_OBSERVATIONS_008_047 product issued by Copernicus Marine Environment Monitoring Service, CMEMS (http://marine.copernicus.eu) provided Level-4 Sea Surface Height (SSH) and derived variables as 30 surface geostrophic currents, measured by multi-satellite altimetry observations over the global ocean with a spacial resolution of 0.25 • . These data capture the mesoscale structures and are helpful to validate the near-surface geostrophic field. GLORYS 12V1 (GLOBAL_REANALYSIS_PHY_001_030 product also issued by CMEMS was used to diagnose the average dynamics during 2017 with a horizontal resolution of 1/12 • at 50 standard depths. 4 https://doi.org/10.5194/os-2020-98 Preprint. Discussion started: 5 November 2020 c Author(s) 2020. CC BY 4.0 License.
All data treatments (in situ, operational and modelling), interpolations with Data-Interpolating Variational Analysis (DIVA, (Troupin et al., 2012)), the analyzes, the graphical representations, and the inverse model were coded in MATLAB (MATLAB, 2019). Additionally, the Smith-Sandwell bathymetry V19.1 (Smith and Sandwell, 1997) was used in all maps and full-depth vertical sections.

5
The XBTs deployed between CTD stations doubled the in situ temperature (t) horizontal resolution. A unified in situ temperature dataset from surface to 2000 m depth was constructed with the CTD and XBT profiles. The interpolation of t was optimally generated in each vertical section with DIVA once the horizontal and vertical correlation lengths, L x and L y , and also the signal to noise ratio, λ, were defined for all transects. The vertical correlation length was L y = 50 m, the horizontal correlation length was within the range L x = 110 − 120 km and λ = 4. Once the interpolated t was validated by a comparison 10 with the original XBT profiles, the same L x , L y , and λ values were employed for the remaining hydrological and biochemical variables, i.e., S P , O 2 , NO 3 , PO 4 and SiO 4 H 4 .
After these interpolations, S P was converted into Absolute Salinity (S A , McDougall et al. (2012)), while t was converted into Conservative Temperature (CT , McDougall and Barker (2011)) as TEOS10 variables (IOC et al., 2010). In addition, neutral density (γ n , Jackett and McDougall (1997)) was used as the density variable.

Velocity field and transports
The geostrophic velocity was calculated along the perimeter of the box referenced to γ n = 27.962 kg m −3 , where the geostrophic velocity was initially considered to be null. This isoneutral surface was the deepest common isoneutral for all the stations and it is located at around 2000 m depth. An inverse box model (Wunsch, 1978) was then applied to estimate the reference level velocities at the box boundaries. This inversion method provides the reference level geostrophic velocity field to estimate the 20 absolute water mass transport through each transect of the closed box, a method that has been widely applied in the Atlantic Ocean (Martel and Wunsch, 1993;Paillet and Mercier, 1997;Ganachaud, 2003a;Machín et al., 2006;Pérez-Hernández et al., 2013;Hernández-Guerra et al., 2017;Fu et al., 2018).
Instead of working with the whole box, it was decided to close the box at the coast to avoid any issues in the temporal evolution of structures when closing the eastern transect with the northern one. To do so, WOA18 climatological nodes were 25 used extending eastward the transects N and S. Therefore, the geostrophic velocities at the reference level were obtained from the inversion for transects N, W and S, while for the E transect those velocities were actually the annual mean extracted from GLORYS and interpolated to the reference level γ n = 27.962 kg m −3 (obtained from WOA18).
The absolute velocity field allows transport estimates of O 2 and inorganic nutrients for the whole box. On the other hand, the uncertainties of transect E for the mass transports were calculated with the velocity variance estimated 5 https://doi.org/10.5194/os-2020-98 Preprint. Discussion started: 5 November 2020 c Author(s) 2020. CC BY 4.0 License. from the annual mean velocity of GLORYS. The uncertainties of transport estimates of O 2 and inorganic nutrients are relative to those of mass transport.

Characteristics and constrains of the inversion model
The inverse model is consistent with both geostrophy and conservation of mass, salt and heat, allowing an adjustment of Ekman transports and freshwater flux. The model is made up of 8 layers divided by the free surface and 8 isoneutrals (26.46, 26.85, 5 27.162, 27.40, 27.62, 27.82, 27.922 and 27.962 kg m −3 ), taken essentially from those defined by Ganachaud (2003a) for the North Atlantic Ocean (Fig. A2). The inverse model takes into account mass conservation and salinity anomaly conservation per layer and also over the whole water column (Ganachaud, 2003b). Heat anomaly is introduced only in the deepest layer where it is also considered conservative. Salinity and heat are added as anomalies to improve the conditioning of the model and reduce the linear dependency between equations (Ganachaud, 2003b). Therefore, the inverse model is composed of 19 10 equations (9 for mass conservation, 9 for salt anomaly conservation and 1 for heat anomaly conservation). Those equations are solved using a Gauss-Markov estimator for 69 unknowns, comprised of 65 reference level velocities, 3 unknowns for the Ekman transport adjustments (one per transect), and 1 unknown for the freshwater flux. In addition, it is necessary to provide a priori the uncertainties related to the noise of the equations (R nn ) and the unknowns (R xx ) in order to solve this undetermined system composed of 19 equations and 69 unknowns. 15 The noise of each equation is dependent on the layer thickness, on the density field, and on the uncertainties of the unknowns (Ganachaud, 1999(Ganachaud, , 2003bMachín et al., 2006). Thus, an analysis of the annual variability of the velocity is performed in the mean depths of the 8 layers. The velocity variance of each layer is estimated from the annual mean velocity provided by GLORYS. These variances are transformed into transport values multiplying by density and the vertical area of the section involved. The uncertainty assigned to the total mass conservation equation is the sum of the uncertainties from the rest of the 20 8 mass conservation equations. The equations for salt and heat anomaly conservation depend on both the uncertainty of the mass transport, on the variance of these properties and, specifically, on the layer considered (Ganachaud, 1999;Machín, 2003).
The uncertainties for salt and heat anomaly equations follow this equation (Ganachaud, 1999;Machín, 2003): R nn (Cq) = a * var(C q ) * R nn (mass(q)) where R nn (Cq) is the uncertainty in the anomaly equation of the property (salt or heat anomaly); var(C q ) is the variance of this property; a is a weighting factor (4 in the heat anomaly, 1000 in the salt anomaly and 10 6 in 25 the total salt anomaly) and q is a given equation corresponding to a given layer. Then, these variability estimates are included in the inverse model as the a priori uncertainty on the noise of equations in terms of variances of mass, salt anomaly and heat anomaly transports.
On the other hand, the variance of the velocities in the reference level is used as a measure of the a priori uncertainty for the unknowns. These variances are calculated from the GLORYS annual mean velocity. However, Machín et al. (2006) concluded 30 that the final mass imbalance is quite independent on both the reference level considered and also on the a priori uncertainties in the reference level velocities.
The initial Ekman transports were estimated from the average wind stress during the days of the cruise. A 50% uncertainty was assigned to the initial estimate of Ekman transports, related to the errors in their measurements and to the variability of the wind stress. An uncertainty of 50 % of the initial value of the freshwater flux, which is 0.0935 Sv, was also prescribed (Ganachaud, 1999;Hernández-Guerra et al., 2005;Machín et al., 2006). Both the Ekman transports, the freshwater flux and their uncertainties were added to the inverse model in the shallowest layer for mass and salt anomaly, and also in the total mass transport and total salt anomaly transport equations.
Dianeutral transfers between the layers were considered to be negligible as compared to other sources of lateral transports, so 5 they were not included in the inverse model. Furthermore, the inverse model provides information only from the box boundaries and cannot be used to provide any details of dianeutral fluxes spatial distribution within the box (Burgoa et al., 2020). Straight lines represent the Θ − S A relationship between the salty and warm NACW on the northern side of the front and the fresh and cold SACW at its southern side, a distribution similar to that proposed by Tomczak (1981). The main water mass in transects N and W was NACW, while SACW was dominant in transect S. Both NACW and SACW appeared along transect E. On the other hand, water masses were also well distinguished at intermediate levels, with colder and fresher AAIW over 20 warmer and saltier MW. MW was sampled mainly in transect N and slightly in transects E and W, while AAIW was the main intermediate water sampled in the remaining transect S (Fig. A3).

Hydrography and water masses
In vertical sections, the x-axis direction is selected according to the path followed by the vessel during the cruise, starting in the northeast, running along the north, west, south and east sections to end up in the northeast corner again. The N/W, W/S and S/E corners are indicated with three vertical pink lines at stations 12, 19 and 28, respectively (Figs. A2, A4, A5, A6 and 25 A7). The upper panel in Figure A2 shows the γ n section estimated only with the CTD dataset while the lower panel shows the γ n field until 2000 m once CTD and XBT temperature fields are merged and salinity field is included by interpolating the salinity profiles at XBT positions. This last high resolution γ n field is the one used to calculate the mass transport. The γ n section shows the isoneutrals which define the 8 layers in the model (

Absolute velocity
The absolute velocity field perpendicular to each transect and with sign depending on geographic criteria (positive sign is northward and eastward) was estimated with the reference level velocity at 27.962 kg m −3 provided by the inversion in transects N, W, and S and, by the annual mean extracted from GLORYS in transect E (Fig. A7). This velocity distribution was validated comparing the average of the SSH with the mass transport in the shallowest layer of each transect (red bars in Fig. A8). Before 5 validation, the high variability in the area and the time spent in its sampling was taken into account: transect N was carried out from July 14 to 21, transect W from July 21 to 26, transect S from July 26 to August 3 and transect E from August 3 to 8.
Hence, the average of SSH during the sampling time period of each transect was selected to analyze the mesoscalar structures  20 The mass transports integrated per layer and transect are represented in units of 10 9 kg s −1 (equivalent to Sv) in Figure A9. These transports integrated for the different water levels of each transect are also compiled in Table A1. Positive/negative transport values indicate outward/inward transports from/to the closed box. In the shallowest two layers, −2.5±0.6 and −0.2± 0.7 Sv entered through transects N and E and, 3.8±0.8 and 1.2±0.5 Sv left through transects W and S, respectively. However, in the next two layers the mass transports reversed their direction in transects N and S. In these two deepest CW layers, the 25 largest mass transport input, −1.0 ± 1.3 Sv, was from transect E, whereas the largest output continued to be through transect W. In the three layers of IW and in the layer of DW, the direction of mass transport in each transect was rather similar:  Figure A10 and collected in Table A2. The transport 30 of O 2 entered through transects N and E and it left mainly through transect W and to a lesser extent through transect S at SW levels. However, the transports of the three inorganic nutrients at SW levels were small through transects N and E in spite of outward transports through transects W and S had significant values. In the shallowest CW layer, −200 ± 74, −15 ± 6, −0.95 ± 0.40 and −6 ± 2 kmol s −1 of O 2 , NO 3 , PO 4 and SiO 4 H 4 entered from N and left mainly westward with 125 ± 51, 34 ± 14, 2.00 ± 0.80 and 12 ± 5 kmol s −1 of O 2 , NO 3 , PO 4 and SiO 4 H 4 , respectively. In the second CW layer, transports of O 2 entered from transect N and E and left mainly through transect W. The incoming transports of inorganic nutrients in the second CW layer was negligible despite they showed outward transports through transects W, N and S. In the deepest CW layer where inorganic nutrients accumulated, especially NO 3 and PO 4 , the highest inward inorganic nutrients transport was 5 obtained from S and E with values up to −38±33, −2.2±1.9 and −33±29 kmol s −1 of NO 3 , PO 4 and SiO 4 H 4 , respectively.

Mass, inorganic nutrients and O 2 transports
Despite the third CW layer is a layer with relative low O 2 concentrations, there was also an important inward transport of −140 ± 124 kmol s −1 of O 2 in transects S and E. On the other hand, it is noted a significant outward transport of 31 ± 42, 2.20 ± 2.95 and 23 ± 31 kmol s −1 of NO 3 , PO 4 and SiO 4 H 4 northward and westward in the deepest CW layer. In addition, in the shallowest IW layer where inorganic nutrient concentrations are high and O 2 concentration is relatively low, the transports 10 were more intense than in the deepest CW layer: −220 ± 175, −51 ± 41, −3.2 ± 2.6 and −31 ± 25 kmol s −1 entered from S and E and they left northward and westward with 175 ± 165, 40 ± 38, 2.75 ± 2.60 and 30 ± 28 kmol s −1 of O 2 , NO 3 , PO 4 and SiO 4 H 4 , respectively. Finally, from the second IW layer to the deepest layer, almost all transport of O 2 and inorganic nutrients entered from S and left mainly through the W transect (Fig. A10). 15 In this work, the dynamics related to the water masses and their O 2 and inorganic nutrient concentrations in the limit between the eastern NASG and the NATG during the summer are analyzed. The sampled water masses in the northern area of the CVB latitudinal change between SACW and NACW, from south to north, below the mixing layer and above 700 m (Pelegrí et al., 20 2017). At IW levels a second latitudinal change is observed between AAIW and MW from south to north (Zenk et al., 1991), with better defined cores: AAIW cores occupy the upper layers whereas MW cores does the lower layers ).

While transects N and S present well defined water masses, transects W and E reflect a transition between transects N and S.
In fact, transect E presents the highest variability in the domain. The variability at SW and CW levels could be related with the 25 position of the CVF, with the proximity of the CUF in the north of the domain and with the meso-and sub-mesoscale structures associated to these two frontal systems, as the upwelling filaments off Cape Blanc (Meunier et al., 2012;Lovecchio et al., 2018). At IW levels, some variability was observed in transects N, W and E due to the MW. However, AAIW is well defined with properties rather constant throughout transect S.
The characteristic of the water masses are conditioned by their origin and by the path followed in their way to the CVB. Their 30 properties are summarized in Figures A11 and A12 where the relationships between in situ measurements of S A , O 2 , NO 3 and PO 4 are displayed. SACW is sampled in transect S and is characterized by relatively low O 2 and high inorganic nutrients concentrations due to its old age (third column of Fig. A11). On the other hand, NACW is a saltier water mass with low concentrations of NO 3 and PO 4 and high of dissolved O 2 . At least in this summer 2017, SACW was not confined above γ n = 26.85−26.86 kg m −3 (equivalent to σ θ = 27.8 kg m −3 ), in the upper CW levels, and NACW was located below it in the lower CW levels, as previously reported (Voituriez and Chuchla, 1978;Peña-Izquierdo et al., 2015). In transects W and E, especially in E, properties of both NACW and SACW were clearly distinct and coexist below γ n = 26.85 kg m −3 (350 m depth). This fact is observed in Figure A11, especially in charts of dissolved oxygen versus S A , dissolved oxygen versus NO 3 and dissolved 5 oxygen versus PO 4 at transect E, where two "patches" are shown, one on top of the other, corresponding to NACW (up) and SACW (down). Additionally, the NACW and SACW property distributions shown in Figure A11 compare well with those shown by Pastor et al. (2008) and Pelegrí and Benazzouz (2015). The shallowest SACW (between 26.46 and 26.85 kg m −3 ) in transect S had concentrations of NO 3 and PO 4 higher than 12.5 and 0.9 µmol kg −1 and lower than 36.5 g kg −1 and 115 µmol kg −1 of S A and O 2 . In contrast, the shallowest NACW (in the same range of γ n of shallowest SACW) in transect 10 N had concentrations of NO 3 and PO 4 lower than 12.5 and 0.9 µmol kg −1 and higher than 36.5 g kg −1 and 115 µmol kg −1 of S A and O 2 . In the same way, the deepest SACW/NACW in transect S/N had concentrations of NO 3 and PO 4 higher/lower than 30 and 2 µmol kg −1 and lower/higher than 35.25 g kg −1 and 95 µmol kg −1 of S A and O 2 (Fig. A11).
At IW, between 27.40 and 27.922 kg m −3 , transect S was also the one with the least variability due to the predominance of AAIW. In contrast, transects E and especially N and W had a larger variability due to the coexistence of AAIW and MW, 15 represented by sudden salinity increases especially in charts of dissolved oxygen versus S A up to more than 27.7 kg m −3 However, MW presents some variability in its properties, in accordance with the ranges documented by Burgoa et al. (2020). 20 The shadow zone proposed by Luyten et al. (1983) between well-ventilated NASG and less-ventilated NATG next to a highly productive coastal upwelling system leads to the development of an oxygen minimum zone (OMZ) within CW and IW levels (Karstensen et al., 2008). In our domain, the OMZ is centered in transect S and it is located between 100 and 800 m depth with its core around 400 m between isoneutrals 27.1 and 27.3 kg m −3 . That distribution matches well with the one provided by Peña-Izquierdo et al. (2015) and it has high concentrations of NO 3 and PO 4 . In addition, below the location of the CVF in As soon as the CVF is crossed, especially in transect E, a sudden and abrupt change due to the coexistence of NACW and SACW was observed. The predominance/lack of SACW/NACW south of the front in transect S suggests that CVF functions as a barrier against lateral transports across the front. On the other hand, the analyzes performed here cannot assess a lateral 30 transport along the CVF.
In summer 2017, in the second CW layer between 26.85 and 27.162 kg m −3 , an inversion of transport was observed in transects N, S and E. This inversion at about 300 m depth in transects N and E and approximately 450 m depth in transect S had been observed in temperature and salinity before (Voituriez and Chuchla, 1978;Pastor et al., 2012;Peña-Izquierdo et al., 2012;Peña-Izquierdo et al., 2015). Some of these authors relate it to different circulation patterns above and below the inversion that 35 11 https://doi.org/10.5194/os-2020-98 Preprint. Discussion started: 5 November 2020 c Author(s) 2020. CC BY 4.0 License.
we have also tried to analyze. In the first two layers of the water column, which group SW and the shallowest CW, incoming transports of −2.5 ± 0.6 and −0.2 ± 0.7 Sv were estimated in transects N and E while outward transports of 3.8 ± 0.8 and 1.2 ± 0.5 Sv were obtained in transects W and S. That is, in the first two layers, approximately 90% of the transport that entered the domain and came from the north, left the domain in transect W with 75% directed to the open ocean. In addition, north of Cape Blanc, this transport can be driven, at least in the first layer, by the wind field (Benazzouz et al., 2014). On 5 the other hand, in the two deepest CW layers the circulation patterns change with a significant input of −1.0 ± 1.3 Sv from the east. In these two layers, where the inversion is given between 26.85 and 27.162 kg m −3 , the main circulation was also westward with a slight northward transport increasing in transect N. These estimated circulation patterns in CW coincide with the climatological streamlines of the geostrophic field estimated from WOA18 for the summer season (Fig. A13). In transects N and W, a west-southwest circulation of NACW was observed throughout the entire water column. However, the cyclonic 10 northward circulation that intensifies in the deepest CW layer produced that only SACW was found in transect S and it also induced the coexistence between NACW and SACW in transect E (Fig. A13). On the other hand, at IW and DW layers, the main transport comes from the south with −2.4 ± 1.6 Sv. This northward transport starts to be noticed in the last CW layer, specially in transects E and S.
In spite of the 21.31% of the relative error associated to the imbalance of 2.6 ± 4.5 Sv (Tab. A1) mainly related with the 15 lack of sypnopticity in the sampling, more than 60% of the mass transport occurred in SW and CW layers, with the exception of transect S where transport barely exceeds 45%. Transports perpendicular to the African coast were 10% more intense than north-south/south-north transports at SW and CW levels. In general, around 70-75% of the mass transport were westward perpendicular to the African coast at SW and CW levels (Fig. A14). More than 75% of the southward mass transport in transect N reduced to a 50% when it crosses transect S at SW and CW levels. In the opposite way, south-north, more than 50% 20 of the mass transport at IW and DW levels moves north in transect S, and it flows northward through transect N with half of its intensity. On the other hand, 20-30% of the mass transport perpendicular to the African coast at IW and DW levels moved away from the coast (Fig. A14).
Despite the CVFZ is in the middle of the study area preventing transports at CW in the north-south direction, the estimated total transport of 7.4 ± 2.5 Sv towards the open ocean is comparable to the 7.0 ± 2.6 Sv estimated by Burgoa et al. (2020) for 25 fall 2002 at 26 • W just north of this domain. Thus, lateral transport in the domain seems to be only moderately affected by the location of the CVF, in the southeastern corner of the domain with an angle of 45 • with respect to the African coast (Fig. A14).
On the other hand, O 2 and inorganic nutrients transports are conditioned by their concentrations and the dynamics in each layer. The inorganic nutrient transports had an inversion between 26.85 and 27.162 kg m −3 on the second CW layer. This inversion was less marked and at a greater depth in O 2 transport. In the first CW layer, O 2 and inorganic nutrients came from 30 the north, but below the inversion they came mainly from the east and south. However, the transport was mainly westward, to the open oligotrophic ocean, in all CW layers. In the third CW layer, there was also important northward transport of inorganic nutrients through transect N. In IW and DW, these transports came from the south and east and they left also mainly westward as with mass transports. The most intense and important transports of O 2 and inorganic nutrients occur in the deepest CW layer and the shallowest two IW layers (Fernández-Castro et al., 2018;Burgoa et al., 2020). In these three layers, O 2 enters 35 with −280±262 and −340±206 kmol s −1 from east and south and it leaves with 95±124 and 255±238 kmol s −1 northward and westward. In the same way, inorganic nutrients entered with −58 ± 54 /−3.5 ± 3.3 /−34 ± 32 and −65 ± 39 /−4.2 ± 2.5 /−42 ± 26 kmol s −1 of NO 3 / PO 4 / SiO 4 H 4 from east and south and left with 35 ± 46 / 2.40 ± 3.13 / 28 ± 37 and 57 ± 53 / 3.9 ± 3.6 / 40 ± 37 kmol s −1 of NO 3 / PO 4 / SiO 4 H 4 northward and westward. These transports do not match with the lateral advection ranges of Burgoa et al. (2020) north of our domain where inorganic nutrients transports are smaller than in the CVB 5 but they are in agreement with the climatological ranges reported by Fernández-Castro et al. (2018).
In summary, the differences between NASG and NATG regions which are well defined by O 2 and inorganic nutrients concentrations in CW layers and in the shallowest two IW layers, condition the transport of these properties. In the shallowest CW layer, transports of O 2 and inorganic nutrients come from the north and they leave mainly westward by the presence of the CVFZ. The estimated net outflows through the east and south in this layer can be justified by the variability sampled in 10 transects S and E related with the southeastern cyclonic gyre and with meander-like structures along the African slope. In the deepest CW layer and the rest of IW layers, the main inputs of inorganic nutrients from the east and south leave latter the box through the north and west transects with smaller transport values but which can change the properties of the more oligotrophic waters north of CVFZ. These inorganic nutrients transport values in the deepest CW layer and in the rest of IW layers might be explained, on the one hand, by the local remineralization in the southeastern of the domain, where the OMZ is located; 15 moreover, the transport intensities are higher in transects S and E than in transects N and W. On the other hand, despite the fact that the southeastern waters have lower concentrations of O 2 than those of the northwest in the deepest CW layer and the shallowest IW layer, the incoming transports from the east and south are higher than the transports northward and westward.
This fact may also be due to a greater intensity of incoming transports from south and east in these two levels and below, hindering their ventilation during this season.     Surface−1897 −1.0 ± 1.6 7.4 ± 2.5 −1.3 ± 2.0 −2.5 ± 2.7 2.6 ± 4.5 Table A2. Transports of O2, NO3, PO4 and SiO4H4 in kmol s −1 with their errors relative to mass transport for SW, CW, IW, and DW across northern, western, southern and eastern transects for FLUXES-I cruise. Positive/negative values indicate outward/inward transports.
Total row of each inorganic nutrient or O2 is its integrated transport in all the water column. The last column is the disajust or imbalance of each water level.