F. Schütte et al.: Hidden vortices: near-equatorial low-oxygen extremes driven by high-baroclinic-mode vortices 137

T- 35 Longitude and Latitude 13
7-30 12
251 Oper, A un
an \ 10 ®
1507 Eddy Radius 600
de
x 50 AAN ;
ı
Swirl Velocity
=
nn 0.2
E91
3 a a
7 Le er
Propagation Speed
= ;
© 0.2
SE 0.1
oA ZEE M lm ma.
. Rossby Number
———
0.2}
T9.1!
PP
1/lsolation
u — ZPO

Potential Vorticity
[el] — Core — -Outside —— 20y mean EN
& —
% Salinity 600
) a nn IN . En
35.3 + m
CN 1
35.21 nm ES
35.1- EP A A A A
= 506 Dissolved Oxygen
> zz
= 150 — T *
3,100 ZZ De
D 0.2 Phosphate, 0-100m
5 v ol ä —
a ) m SIE
9 ie —
Ss
u |

_

Biomass, 0-100m
iA UT
A
An
_ I
5 Particulate Oraanic Phosphorus Flux, at 100m
% 10
12 6 Me
8 5
31 ©
9 nm
0 100 200 300 400 500 600 &
Time [davs]

Figure 10. Time series of different variables related to the core of the modelled subsurface intensified eddy shown in Figs. 6g-1 and 9. Time is
given as elapsed days since eddy detachment from the African coast. Vertical gray lines in each panel denote time points for horizontal maps
shown in Fig. 9 (0, 211, 346 and 581 d). (a) Longitude (black line) and latitude (blue line), (b) Eddy radius, (c) Eddy swirl velocity, (d) Eddy
propagation speed, (e) Rossby number, (f) Inverse of isolation parameter (black line). Black-white dashed line denotes threshold, below
which the water is trapped in the eddy core (swirl velocity > propagation speed). (g) Potential vorticity, (h) salinity, (i) DO, (j) phosphate,
(k) biomass, (1) flux of particulate organic phosphorus. Variables are given for the following layers. In panels (a)-(f) and (h)-(i): isopycnal
surface 26.5kg m7*, In panel (g): isopycnal surface 26.6 kg m”. In panels (j) and (k): integral over 0-100 m. (1) at 100 m. For the right
column (panels g-1), black lines show value in eddy core, red lines show mean values outside the eddy (average between 1 and 3° of
longitude/latitude around the eddy core position) and green lines show 20-year model mean that is given at the respective position of the
eddy core. In panels (bH(f). gray dots at zero line denote time points, where no estimate was possible.

shown in Fig. 6; others not shown) and around 0.1-0.4 in
the GFDL CM2.6 model simulation (exemplarily shown in
Figs. 6 or 10e).
The observed cyclonic HB Vs appeared with a stratification
maximum at about 150m and a cyclonic velocity structure
with maximum EKE at a similar depth. Shallow DO minima
were found at 100 and 200 m throughout the transition, but
without any clear separation from the deep OMZ at 300m.
This less intensified DO minimum at 100m and the missıng
 intermediate DO maximum at 200 m is a substantial difference
 to the DO distribution observed within anticyclonic
HB Vs. However, enhanced DO consumption has been shown
to be a reasonable driver for DO depletion in a cyclonic HBV,
chat was observed in the western subtropical North Atlantic
(Li et al., 2008). Pure upwelling or upward mixing of low-DO
 water from the deep OMZ cannot explain such vertihttps://doi.org/10.5194/os-22-119-20

 26

cally homogeneous distribution of low-DO between 100 and
300 m in cyclonic HBVs. These processes would imply either
 a shallowing of isopycnal surfaces or a weakened stratification
 within this depth range, which is contradictory to
the observed deepening of isopycnals above 150m, shallowing
 of isopycnals below 150m and consequently the intensified
 stratification at 150m (Fig. 5, Event #03). Due to
the increased stratification, the thickness of the intermediate
 DO maximum layer (that is associated with isopycnal
26.6 kg m?) is reduced and very likely not resolved by the
sparsely distributed number of DO sensors.
We could not collocate any clear signals in SLA or SST
from satellite observations with the in situ observed HBVs.
Shipboard observations showed a strongly weakening velocity
 signature toward the surface. Also the simulated HBVs
from GFDL CM2.6 model showed similar characteristics as

Ocean Sci... 22. 119-143. 2026