Recreational vessels without Automatic Identification System (AIS) dominate anthropogenic noise contributions to a shallow water soundscape

Recreational boating is an increasing activity in coastal areas and its spatiotemporal overlap with key habitats of marine species pose a risk for negative noise impacts. Yet, recreational vessels are currently unaccounted for in vessel noise models using Automatic Identification System (AIS) data. Here we conduct a case study investigating noise contributions from vessels with and without AIS (non-AIS) in a shallow coastal area within the Inner Danish waters. By tracking vessels with theodolite and AIS, while recording ambient noise levels, we find that non-AIS vessels have a higher occurrence (83%) than AIS vessels, and that motorised recreational vessels can elevate third-octave band noise centred at 0.125, 2 and 16 kHz by 47–51 dB. Accordingly, these vessels dominated the soundscape in the study site due to their high numbers, high speeds and proximity to the coast. Furthermore, recreational vessels caused 49–85% of noise events potentially eliciting behavioural responses in harbour porpoises (AIS vessels caused 5–24%). We therefore conclude that AIS data would poorly predict vessel noise pollution and its impacts in this and other similar marine environments. We suggest to improve vessel noise models and impact assessments by requiring that faster and more powerful recreational vessels carry AIS-transmitters.


Study area
The presence of porpoises was based on knowledge from Mikkelsen et al. (2017), but also confirmed by visual observations of animals during fieldwork and detection of porpoise clicks in sound recordings.
The study area was approx. 20 km from the city Aarhus with the largest commercial harbour in Denmark, and 10 km from Ebeltoft ferry harbour. Nine marinas, incl. two in Aarhus, are located between 13 and 20 km from the study area. The main shipping route connecting Aarhus to Kattegat runs approx. 1.5 km south of the recording station.

Recorder deployment
The recorder was suspended 1.5 m from the seabed between an anchor and a mid-water buoy with the hydrophone facing upwards. To maintain visibility of the recorder position from the tracking point, but minimize artificial noise and drag of the recording gear, a surface buoy with a slack 10 m rope was attached to the mid-water buoy.

Theodolite tracking
The theodolite was placed at the same location every day at the high point Burs klint (56°6.153N, 10°32.23E, 47 m above sea level), where there was a good overview of the marine study area. To ensure that the exact position of the theodolite was known, the position of a lighthouse (Hjelm Lighthouse, 56°8.00N 10°48.29E) and a reference pole (measured by differential GPS by a certified surveyor in 2015, as used in Mikkelsen et al. 2017) were used daily as fix points. A laptop running the software Cyclops Tracker (version 2004, Eric Kniest, University of Newcastle, Australia) delivered real-time display of the tracked vessels via a communication port connection to the theodolite, and stored the data for later processing. Variations in sea level according to tide were corrected for in Cyclops Tracker by tracking sea level on a tide pole approx. every hour. In case of multiple vessels present at the same time, tracking of the motorised vessel closest to the recorder was prioritised. The 2 km criteria led to discard of 26% of the theodolite tracked vessels.

Vessel types
Registered vessels within the study area (max 2 km from the recorder) and the study period were grouped into nine types (Table S1): Cargo vessels, Tankers, Dredgers, Other vessels (e.g. fishing vessels and navy vessels), Motorboats, Speedboats, Rigid-hulled inflatable boats (RHIBs), Dinghies, Sailboats.
Vessels tracked by AIS were automatically labelled with vessel type in the AIS data, whereas theodolite tracked vessels without AIS were instead labelled manually during observation and confirmed later from photos. Here, motorboats were characterised by having a cabin, whereas speedboats were not fitted with a cabin, but had a windshield in the front. Dinghies were characterised as open aluminium boats with small outboard engines and without cabin or windshield. All inflatable boats were grouped in the category 'RHIBs'. Sailboats were all vessels that had sails, including those that were travelling by motor power (including a catamaran and a wooden ship with masts).

Vessel tracks
Multiple passes of the same vessel were treated as separate events, as there were sufficient distinct vessels to dilute any resulting pseudo replication. AIS vessel tracks always extended beyond the 2 km criterion, as we had collected available AIS data for a 20 km area around our recorder. We did not extrapolate tracks of non-AIS vessels out to 2 km, as recreational vessels often travel unpredictably (Nowacek et al., 2001) and due to acoustic shading by the sandbank of many of them. Vessel tracks were also used to overlay AIS data and theodolite data in time and space to determine if a small vessel had voluntarily installed an AIS ( Supplementary Fig. S4). If both were available for a given vessel, the AIS data was used for further analysis. The comparison between the theodolite and AIS track for unique vessels allowed for estimating the difference between the two measures, and was used to obtain an error estimate for theodolite positions (assuming that AIS positions were correct).

Harbour porpoise reaction thresholds
We used response thresholds of porpoises derived from three previous studies reporting avoidance, who investigated how tagged wild porpoises reacted to vessel noise. These authors found a significant reduction in foraging when noise levels exceeded 96 dB re 1µPa (RMS) in a third-octave band at 16 kHz, quantified as the 90 th percentile of noise in 0.5 s bins over 1 min windows. To implement this threshold, we calculated the 90 th percentile of 1 s bins over a 1 min sliding window. For all thresholds, a blanking time was applied such that, if a high noise level was detected, a subsequent detection could not be made until the noise level had been below the threshold for at least 1 minute. This blanking time was used to avoid that a vessel pass counted multiple times, as it was expected that an animal would react during vessel approach phases, when noise levels initially increased and exceeded porpoise reaction thresholds.

An index of vessel noise (Jakob Tougaard 2016)
The received noise level (RL) from a single vessel at range r1 can be modelled as: 1 = 1 − log 10 1 (1) Where SL is the source level and absorption is ignored.
If there are N vessels, at respective ranges r1, r2,… rN, the combined received level, expressed as intensity, is given as: = 10 log 10 (10 1 10 + 10 2 10 + ⋯ 10 10 ) Combining with (1) yields: The figure below shows a simple example, where one vessel is stationary at 1000 m and another varies from 100 m to 10 km. A simple assumption that the RL is determined by the closest vessel will at most be 3 dB off, when the two vessels are at the same range from the receiver.
Vessel 1 at 1000 m Vessel 2 at variable range Figure S1. Water depth (a) and vessel passes (b) within the study area, i.e. red circle = 2 km around sound recorder (red filled circle). The yellow triangle on (b) marks the position of the theodolite, while green lines mark interpolated tracks of non-AIS vessels and blue lines mark AIS vessel tracks.   Figure S4. Examples of recreational vessels that were tracked with theodolite, but were also found to have an AIS by overlaying theodolite tracks with AIS tracks. Figure S5. Examples of how recorded TOLs (bottom, centred at 2 kHz) were divided into the five vessel categories (b), depending on vessel presence in the study area (a). Figure S6. Examples of how exceedance of the three response thresholds for porpoises (b-d) in Fig. 6 were found (red dots), using a blanking time of 1 minute (i.e. max one exceedance pr. minute). Figure S7. TOLs at 0.125 (grey), 2 (orange) and 16 kHz (green) during the closest point of approach (CPA) for each vessel. Vessel types are AIS vessels (circle) and non-AIS motorised vessels (triangle). Data is sorted by the levels in the 2 kHz band on the x-axis (from high to low).