F. Basanı et al. 
Table 2 
Summarised information on location, equipment and conditions at the measurement sites, “Shipping volume” describes the sum of unique vessels being within a 35 km 
radius throughout the year 2019 around each station for time steps of 15 min referenced to the 08-BE-WST station; 13-NO-LOV was located outside the modelled area 
or currents, wind and AIS; Measurement setups are given to be autonomous (a) or cabled (c). 
Station Position Measurement Shipping volume referenced to Max. bottom current Mean 10 m Water depth No. of measured days 
LAT | LON setup 08-BE-WST velocity wind speed [m] (2019 & 2020) 
[ms”!] Im 1 
Marine Pollution Bulletin 198 (2024) 115891 
01-SE-VIN 57.623 11.572 
01b-SE- 57.676 11.671 
HON 
)2-DK- 56.927 11.200 
ANH 
\3-DK- 55.575 7.438 
HRF 
14-DE- 55.195 7.158 
FN3 
)5-DE- 55.626: 14.099 
ES1 
J6-DE- 54.015 6.588 
FN1 
)7-NL- 53.316 4043 
TEX 
)8-BE- 51.353 2.445 
WST 
I9-UK- 53.529 1.053 
DOW 
„0-SC- 56.500 —2.380 
ARB 
11-SC- 57.976 —3.536 
HEL 
12-SC- 58.575 —2.120 a 
MOR 
;3-NO- 68.910 14.380 
LOV 
4-NO- 58.2537 5.839 
NTR 
L5-SC- 56.647 —0.094 
CNS 
16-DK- 56.919 11.758 
TN1 
L7-DK- 56.902 11.6482 
TN4 
18-DK- 55474 5.110 
EDA 
c/a 
% 
70 % 
77 % 
21% 
J).16 
6.67 45.0 316 
5.22 30.0 182 
7.50 12.0 200 
8.13 14.9 211 
8.17 21.8 
30.0 328 
21.0 127 
21.0 633 
18.0 347 
50.0 286 
30.0 104 
258.0 
340.0 
47.2 329 
45.0 
I.21 
2.37 
16% 
;/a 
13 % 
0.23 
3% 
7.42 
‘/a 
14 9% 
3.56 
33 % 
100 % 
2.63 
‘/a 
7.99 
27 % 
).44 
Ra 
323 
I.10 
2.31 
5% 
9% 
23 % 
).12 
3% 
3.32 
533% 
).21 
1.17 
48 Yı 
20% 
0.25 
he North Sea of 2019 is provided in Fig. C in the supplementary 
material. 
ınderlying distribution of sound levels was revealed, and outliers and 
multimodalities were detected. 
Broadband sound pressure levels at the stations were compared 
]uantitatively. The three chosen bands (D1, D2 and D3) resemble the 
decadal bands covering the frequencies in the 20 Hz - 160 Hz, 0.2 kHz - 
_.6 kHz and 2 kHz to 16 kHz one-third octave bands. These frequency 
bands were also the focus of executed modelling during JOMOPANS, 
and broadly represent the hearing range of different animal groups (de 
Jong et al., 2022). To account for missing data at some stations, the low- 
frequency band (D1) only includes frequencies down to 25 Hz and the 
highest band (D3) only includes frequencies up to 10 kHz. Temporal 
variability in these frequency bands was examined using the 5th, 50th, 
and 95th percentile. 
To quantitatively identify clusters, the broadband SPLs were 
:ompared pairwise using a two-sample Kolmogorov-Smirnov (KS) test 
(Smirnov, 1939). The Test Variable, as a convenient measure of simi- 
larity between distributions, gives a numerical indication on how similar 
the cumulative distribution functions of the respective broadband levels 
(D1-D3) were at all stations (ranging from 0 to 1, where 0 represents 
identity). The Kolmogorov-Smirnov statistic is given by: 
Dim = SUp|S. (x) Sm(x)] 
where S, and S„ are the cumulative distribution functions of 
broadband sound pressure level time series at two stations and sup is the 
zsupremum function. 
The KS Test statistic provides an estimate on how similar the SPL 
distributions are in terms of location and shape (width and peak values). 
2.4. Ocean current data 
Ocean current velocity data were derived from the operational hy- 
drodynamic model of BSH (Brüning et al., 2021). The layer closest to the 
seafloor was selected as all hydrophones were placed 2-3 m above the 
seafloor (cf. Fischer et al., 2021). The modelled currents were forced by 
wind, pressure gradients and tidal forces. The temporal resolution of the 
nodelled data was 15 min and included data on current velocity and 
direction. The spatial resolution of the model was three nautical miles, 
and for each individual sound measurement station the output at the 
nearest grid point was selected. The maximum bottom current velocities 
per station are given in Table 2 and a map of bottom currents is shown 
Fig. D in the supplementary material. 
3. Methods 
3.1. Comparison of acoustic measurements 
The available data and variability in the measured TOLs were 
visualised as spectrograms and Spectral Probability Density (SPD) plots. 
SPD plots comprehensively depict characteristics of ambient sound 
levels by presenting the empirical probability density of the measured 
TOLs. As proposed by Merchant et al. (2013) this presentation was 
combined with conventional percentiles and spectral averages. The