FE. Basan et al. 
Marine Pollution Bulletin 198 (2024) 115891 
D1-SE-VIN 
02-DK-ANH 
03-DK-HRF | 
D4-DE-FN3 
05-DE-ES1 
06-DE-FN1 
07-NL-TEX 
D8-BE-WST 
)9-UK-DOW 
LO-SC-ARB 
LL-SCHEL, 
12-SC-MOR | 
13-NO-LOV 
L4-NO-NTR + 
15-SC-CNS | 
16-DK-TN1 
17-DK-TN4 
18-DK-EDA 
AIS VS wind 
Wind VS Current 
| 
L_—— 
BLEI A a 
0.5 -04 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 
Correlation Coefficient 
Fig. 11. Spearman correlation coefficients of AIS counts (for a radius of 35 km) and modelled wind (blue) and modelled wind and modelled currents (yellow) for 
each JOMOPANS station. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) 
complexity of flow noise. Vortex shedding and cable vibration may have 
ınfluenced the scaling factors as well as other contributions to the SPL in 
addition to flow noise. 
Table 3 summarises the one-third octave bands of highest correlation 
and the corresponding scaling factors for wind and flow. 
5. Discussion & conclusion 
the shallower southern North Sea compared to deeper northern stations, 
„onsistent with previous model results. A limitation of this study was 
that fewer stations were operational in the northern North Sea and 
Norwegian Trench. Therefore, it would be advantageous to increase 
monitoring efforts in remote areas of the North Sea where validation of 
soundscape maps has not been possible (Putland et al., 2022). 
The highest SPLs at all stations were measured between 100 Hz and 
500 Hz. These frequencies are consistent with radiated noise from ves- 
sels, which were expected to dominate ambient noise across the North 
Sea region due to high levels of shipping traffic. Indeed, strong corre- 
'ations with local shipping density were often observed in the mid- 
frequency band (50 Hz - 1250 Hz; Spearman coefficients of 0.3-0.7:; 
;g, 9). 
The strongest correlations were with wind at frequencies above 500 
Hz (Spearman coefficients >0.8). Flow noise also correlated to mea- 
surements in low frequencies up to 500 Hz (Spearman coefficients from 
I) to 0.9). Flow noise — which is caused by turbulence around the hy- 
Adrophone and is not present in the environment — was found to occur at 
almost all stations, with highest correlations between TOL and current 
7elocity at stations that are greatly impacted by tidal currents (especially 
I8-BE-WST where highest current velocities occur). It was possible to 
derive linear scaling laws between the TOLs and the logarithm of wind 
speed and of current velocities. As these scaling factors were largely 
zonsistent between the stations, they provide a simple method of 
describing the relationship between TOLs and wind speed and 
:ontamination from flow noise. 
Several statistical tests were carried out to evaluate geographical 
patterns. The pairwise Kolmogorov-Smirnov test identified similarities 
between monitoring locations, with geographic clusters per three 
broadbands revealed. For example, large parts of the North Sea were 
similarly affected by wind noise, which corresponded to low variability 
This international study offered unprecedented insight into the 
North Sea soundscape over a wide spatial and temporal scale. The 
:‚emporal variability of SPLs at individual stations was found to be higher 
at shallow stations in the southern North Sea whereas more persistent 
SPLs were measured at the deeper northern stations. Propagation loss in 
shallower waters is greater than in deeper waters (for frequencies where 
ship noise dominates, <1-2 kHz). The higher number of vessels oper- 
ating in the southern North Sea combined with increased transmission 
losses thus results in higher variability of received sound levels. In 
deeper water however, sound travels further and ships from a greater 
distance contribute to received sound levels. Individual ship passages 
:hen tend to average out. Depth also determines which frequencies 
»ropagate best under water (Jensen and Kuper, 1983). The dominant 
requencies are lower at the deep stations in Norway (13-NO-LOV and 
14-NO-NTR) than at the shallower stations in the southern North Sea. At 
nost stations the depths are <50 m and therefore the optimal fre- 
quencies that propagate the furthest are expected to be between 500 and 
1000 Hz, which coincides with the frequency range of maximum ship 
noise energy. At very shallow stations we can also observe cut-off effects, 
where low frequencies below 100 Hz do not seem to propagate at all (e. 
g. at 02-DK-ANH). 
The temporal variability was relatively low compared to the spatial 
variability of SPI,s across the different stations. SPI.s were also higher in