F. Schütte et al.: Hidden vortices: near-equatorial low-oxygen extremes driven by high-baroclinic-mode vortices 121 
17.5°N 
190 
15°N 
u 
12.5°N 
1 
Fr 
10°N 
km 
1.5°N 
5°N 
2.5°N | 
AHr 
= min. 1 CTD available 
© 3x Mooring 
“) High-baroclinic-mode vorticies (HBVs) 
el 
Sn 30°W 
25°W 
helle 
20°W 
15°W 
170 
150 
= 
130 8 
Q 
-110 5 
S 
90 
70 
5 
Figure 1. Map of the eastern tropical North Atlantic. Shaded are minimum dissolved oxygen (DO) values in the upper 200 m of the climato- 
logical DO distribution from the World Ocean Atlas 2023. The small squared boxes indicate regions of 0.5° boxes for which at least one CTD 
station is available. These boxes are colored with their minimum DO concentration in the upper 200 m (from multiple CTDs, if available) 
only if the minimum DO concentration is less than 60 umol kg7!. Red circles suggest the occurrence of high-baroclinic mode vorticies as 
analyzed in detail in the manuscript. The yellow points mark the positions of the moorings analyzed in the manuscript. The black contour 
indicates the first baroclinic Rossby radius of the deformation (in km), calculated from the World Ocean Atlas data, following Chelton et 
al. (1998). 
water of the HBVs is analysed in Sect. 4.5 and the origin and 
temporal evolution of the HBVs based in model simulations 
is shown in Sect. 4.6. We give a detailed discussion in Sect. 5 
and provide a summary in Sect. 6. 
2 Data 
Data from moored, shipboard and satellite observations, cli- 
matological data as well as the output of an actively eddy- 
ıng ocean-biogeochemistry model from the tropical North 
Atlantic were used within this study as described in the fol- 
lowing. 
2.1 Moored observations 
Multi-year moored observations from three different loca- 
vions 11°N/21° W; 11° N/23° W and 4° N/23° W were used 
in this manuscript (Fig. 1). The mooring at 11° N/21° W 
was equipped with DO (AADI Aanderaa optodes of model 
types 3830 and 4330) and CTD (Conductivity, temperature, 
depth) sensors (Sea-Bird SBE37 microcats) which were at- 
tached next to each other on the mooring cable between 2012 
co 2018. Eight of these optode/microcat combinations were 
installed evenly distributed in the depth range between 100 to 
800m, delivering multi-year time series of temperature, 
salinity and DO with a temporal resolution of up to 5 min. 
At 800m depth, an upward looking Acoustic Doppler Cur- 
rent Profiler (ADCP) was installed to record velocity in the 
depth range between about 60 and 800 m. During the 2nd de- 
https:/doi.org/10.5194/os-22-119-20 716 
ployment period (May 2014 to September 2015), no velocity 
observations were available due to a failure of the ADCP. 
Before and after a deployment period, optodes and microcats 
were calibrated against CTD-O measurements during CTD 
casts and onboard lab measurements as described in Hahn 
et al. (2014, 2017). The correction against reference mea- 
surements, thereby considering potential sensor drifts (Bittig 
et al., 2018), allowed best data quality and yielded average 
root mean square calibration errors of 0.003 °C, 0.006 and 
3 umol kg"! for temperature, salinity and DO, respectively. 
Only quality controlled data that was flagged good was used 
for further analysis. ADCP measurements were quality con- 
trolled against a percent good criterion (20% threshold) 
and were checked for plausibility and evident outliers due 
to surface reflection. ADCP bin depths were corrected us- 
ing a mean sound speed profile following the approach by 
Shcherbina et al. (2005). This mooring is used to study hy- 
drographic, DO and velocity temporal variability (on daily 
to intraseasonal time scales) related to low-oxygen extreme 
events. The other moorings at 11° N/23° W and 4° N/23° W 
are part of the prediction and research moored array in the 
tropical Atlantic (PIRATA), which were equipped with DO 
(AADI Aanderaa optodes of model types 3830 and 4330) 
sensors at 300 and 500m depth from 2009 to 2024. At 
11° N/23° W additionally a DO sensor at 80 m depth was in- 
stalled between 2017 to 2024. The DO sensors deliver hourly 
data and are calibrated and processed in the same way as de- 
seribed above. 
Ocean Sci... 22. 119-143, 2026