128 F. Schütte et al.: Hidden vortices: near-equatorial low-oxygen extremes driven by high-baroclinic-mode vortices
coastal upwelling region north of 12° N, where very low-DO
extremes can also be observed near the coast (see Fig. 1 or
Schütte et al., 2016b). In order to scale for the different num-
ber of CTD-O profiles in the four regions shown in Fig. 4a
(5%, 9%, 61% and 25 % of the profiles for the boxes 30-
27° W, 27-24° W, 24-21° W, 21-18° W), we estimated the
relative distribution and calculated the 10-percentile thresh-
old in every box (Fig. 4c). This threshold is lowest in the
open ocean (24-21° W), whereas the mean DO distribution
is increasing from the eastern boundary towards west. This
counterintuitive distribution of low-oxygen extremes, which
is against the mean DO gradient, suggests that DO depleted
water generally cannot be purely advected from a remote re-
gion at the eastern boundary, that is poor in DO. Locally en-
hanced biological activity associated with enhanced DO con-
sumption must play a role as well.
The two events with the lowest dissolved oxygen (DO)
concentrations were measured as 17 umolkg”! by a CTD
at 60m depth at 8 umolkg”! were recorded by a mooring
at 80m depth at 11° N/23° W. These two low-oxygen ex-
t(remes were well below the climatological average mini-
mum DO concentration for the whole ETNA (40 umolkg”!
in the deep OMZ, Brandt et al., 2015). We shall note, that no
CTD-O profiles were available in this data set for the eastern
boundary region within about 2° longitude off the African
coast.
4.2 Association of low-oxygen events with subsurface
high-baroclinic mode vorticies
For the majority of the ship based data and for the mooring at
L1°N/21° W additional observations of hydrography, zonal
and meridional velocity are available indicating the passage
of anticyclonically and cyclonically rotating vortices associ-
ated to the low-oxygen events. At the mooring position the
low-oxygen events #01, #02, #03, #04 and #07 were most
likely related to the passage of subsurface intensified vor-
ices, whereof events #02, #04 and #07 were associated with
anticyclonic vortices and events #01 and #03 with cyclonic
vortices (Fig. 5). Note, that we explicitly refer here to the no-
ation vortex, since we could not derive the vortices’ radii in
order to differentiate between mesoscale and submesoscale.
For the anticyclonic vortices, meridional velocity was ob-
served with maximum northward and southward flow taking
place at the beginning and the end of each low-DO period.
Zero crossing was observed in between at around the time,
when DO was at its minimum (Fig. 5e-h). Corresponding
üme series of potential density derived from hydrographic
observations, conducted next to the DO sensors, indicated a
depression of isopycnal surfaces in the depth range below
L00 m. Time series of velocity and potential density agree
well with the dynamical understanding and passage of west-
ward propagating eddies (van Leeuwen, 2007) through the
mooring site. Zonal velocity was either small or showed max-
imum flow during time periods of minimum DO, depending
Ocean Sei... 22. 119-143. 2026
whether the eddy has crossed the mooring site either with
its core or with one of its meridional flanks. Zonal velocity
vanished at the beginning and the end of each of the three
events.
During events #01 and #03, that are associated with the
passage of subsurface intensified cyclonic vortices, we found
a depression of isopycnal surfaces above 150m and a heave
of isopycnal surfaces below (cf. McGillicuddy, 2015, de-
noted as eddies of type Thinny). This is associated with
a maximum in stratification at about 150m depth. The
time series of zonal and meridional velocity, respectively,
showed maximum values at a similar depth with a transi-
tion from westward to eastward (event #01) and southward
to northward (event #03) velocities during the time of max-
imum stratification. In contrast to the anticyclonic vortex
events (#02, #04 and #07), the DO minima during the pas-
sage of the two cyclonic vortex events (#01 and #03) were of
similar intensity at 100 and 200 m depth, with no separation
rom the deep OMZ at 300 m by an intermediate DO maxi-
mum. Though, during both events the minimum DO at 100m
was well below the average DO concentration that was ob-
served for time periods without any vortex event. We shall
explicitly note, that the characteristics for zonal and merid-
jonal velocity during event #01 were swapped compared to
the other eddy events (#02, #03, #04 and #07). We can only
speculate whether this cyclonic vortex has crossed the moor-
ing site in a more meridionally directed pathway.
The vertical structure of these vortices could not be iden-
tified for the near surface layer and the deep ocean, since
moored hydrographic and velocity observations were only
available between 100m (60m for velocity) and 800m
depth. This made it challenging to distinguish among surface
intensified and subsurface intensified (but at shallow depth)
vortices. The most likely subsurface intensified vortex was
associated with event #02, showing extreme velocity (both
zonal and meridional) slightly below the shallowest depth of
available observation accompanied by an oxygen minimum
of 39 umolkg!. Notably, none of these vortices exhibited a
clear surface signature in satellite data that could be unam-
biguously associated with the subsurface features
4.3 Horizontal extent of the low-oxygen high-baroclinic
mode vorticies
The ship-based data, which cover the region spatially, are sig-
nificantly better suited than the stationary moored data for
assessing the spatial extent of the HBVs. Repeated merid-
ional ship sections between 6-12°N along 23° W, available
over a distance of at least 300 km, captured 15 events with
DO concentrations well below 60 umolkg”! in the upper
200m (Table 1, Fig. 6). All DO minima were found di-
rectly below the shallow oxycline at depths between 45 and
90 m (corresponding to surfaces of potential density between
00 = 26.2 and 26.4kg m7*). The meridional resolution of
CTD-O measurements did not allow for a proper identifica-
https://doi.org/10.5194/os-22-119-2026
