**Miles Parsons \* and Mark Meekan**

Australian Institute of Marine Science, Perth, WA 6009, Australia; m.meekan@aims.gov.au **\*** Correspondence: m.parsons@aims.gov.au; Tel.: +61-(8)-6369-4053

Received: 20 October 2020; Accepted: 19 November 2020; Published: 27 November 2020

**Abstract:** Vessel noise is an acute and chronic stressor of a wide variety of marine fauna. Understanding, modelling and mitigating the impacts of this pollutant requires quantification of acoustic signatures for various vessel classes for input into propagation models and at present there is a paucity of such data for small vessels (<25 m). Our study provides this information for three small vessels (<6 m length and 30, 90 and 180 hp engines). The closest point of approach was recorded at various ranges across a flat, ≈10 m deep sandy lagoon, for multiple passes at multiple speeds (≈5, 10, 20, 30 km h<sup>−</sup>1) by each vessel at Lizard Island, Great Barrier Reef, Australia. Radiated noise levels (RNLs) and environment-affected source levels (ASLs) determined by linear regression were estimated for each vessel and speed. From the slowest to fastest speeds, median RNLs ranged between 153.4 and 166.1 dB re 1 μPa m, whereas ASLs ranged from 146.7 to 160.0 dB re 1 μPa m. One-third octave band-level RNLs are provided for each vessel–speed scenario, together with their interpolated received levels with range. Our study provides data on source spectra of small vessels to assist in understanding and modelling of acoustic exposure experienced by marine fauna.

**Keywords:** vessel noise; radiated noise levels; monopole source levels; propagation loss

## **1. Introduction**

Through evolutionary time, sound has become an important sensory cue for many marine taxa. The efficient transmission of sound underwater has meant that a wide variety of species have developed frequency-specific hearing sensitivity and rely on the detection of acoustic cues and subtle changes in the biophony of their local soundscape during vital life functions [1–5]. These important signals, such as the spawning calls of fishes or the sound of healthy habitat in which larvae will settle, can be masked by anthropogenic noise, disrupting natural behaviors [6–8]. Sound produced by vessels is a major element of marine anthrophony and has been recognized as a chronic stressor [9], negatively impacting communication, health and behavior of many species [4,10–15]. As human populations have increased, so too has anthrophony in oceans and inland waterways [16–21], creating what has now been termed the 'Ocean soundscape of the Anthropocene' [22].

Management strategies that aim to mitigate the impact of vessel noise on marine fauna [23–28] require information about source levels and vessel movements. Although the Automatic Information System (AIS) can be used to track passages of the majority of commercial vessels [29,30], noise is also dependent on vessel size, speed, load and power, as well as other design characteristics [24,31,32]. This requires characterization of source signatures from different types and sizes of vessels.

At present, there is little data on the noise characteristics of small (<25 m length) vessels [26,33–37]. This is important because in coastal waters, these vessels often vastly outnumber larger ships ferrying commercial cargos. The data required to accurately model the propagation of signals from small vessels are rarely reported, or typically provided as one or two measures at limited numbers of speeds [36]. For this reason, we lack data on the variability in noise among vessels of different classes (e.g., monohull, catamarans, tugs, landing craft) within this size range or even different passes of the same vessel. This is problematic for management strategies that aim to set useful guidelines to mitigate

noise for boating activities [38], particularly in shallow coastal waters, inland waterways, and coral reefs, where small vessels have the potential to significantly change the local soundscape and, due to proximity, are more likely to affect fishes, invertebrates and small marine mammals [38–42].

To address these issues, our study aimed to characterize the source spectra of three small vessels under 10 m length that are commonly used in shallow coastal marine environments. We took multiple measurements at the closest point of approach (CPA) at multiple ranges and speeds, over multiple passes, in shallow water. Source characteristics of noise can be specific to a vessel and have multiple engine-, propeller- and hull-related origins [43] and their impacts on fauna are frequency-dependent. Therefore, one-third octave levels were also calculated, and their propagation across the measured ranges investigated.

### **2. Materials and Methods**

The International Standards Organization (ISO) protocols for the measurement of vessel radiated noise levels (RNLs) and monopole source level (MSL) focus on large vessels in deep water (see ISO 17208-1 [44] and 17208-2 [45]). The ISO criteria require a minimum water depth equal to the greater of 150 m or 1.5 times the overall ship length. For the highest standard of estimates, this comprises the deployment of three hydrophones positioned vertically at depths that result in 15◦, 30◦, and 45◦ angles from the sea surface at a CPA distance of either 100 m or one overall ship length, whichever is greater. Neither are these requirements achievable, nor is the procedure applicable, in shallow water. Indeed, meeting these requirements in Australia would require vessels to travel a significant distance offshore, which may not be appropriate for all classes of small vessels. Standards for measurements of RNL in shallow water are under development. However, we had sufficient replication of measurements to accurately estimate both RNLs and affected source levels (ASLs), in lieu of any current shallow-water ISO protocols.

### *2.1. Study Site*

Lizard Island is a granitic island located approximately 30 km off the north Queensland coastline (14◦40.88 S, 145◦27.82 E, Figure 1). The Lizard Island group comprises four late-Permian granite islands—Lizard, Palfrey, South and Bird Islands—which, together with the surrounding fringing reef, encircle an up to 10 m-deep flat, sandy-bottomed lagoon [46]. Tidal range at Lizard Island reaches a maximum of 3 m and current speeds into the lagoon can be >30 cm s−<sup>1</sup> [47,48]. Measurements were collected in the 10 m deep area of the lagoon, to the south of Lizard Island (Figure 1).

**Figure 1.** (**a**) Map of Australia with expansion of Cape York Peninsula, Queensland; (**b**) expansion of Lizard Island on the Great Barrier Reef; (**c**–**e**) expansions of the island lagoon with the vessel tracks from three consecutive survey days (**c**, **d**, and **e**, respectively) displayed in white and the positions of pairs of seafloor-mounted SoundTraps shown by the red dots.

#### *2.2. Vessel Recordings*

Vessel recordings were acquired using Ocean Instruments ST300 SoundTraps. These are piston-phone calibrated passive acoustic pressure sensors with a flat response of ±3 dB over the 20 Hz to 60 kHz system bandwidth, calibrated by the manufacturer using a 121 dB re 1 μPa source at 250 Hz. Divers deployed 10 ST300s on the seafloor of the lagoon, each orientated vertically, attached to the top of a star picket, and positioned approximately 50 cm above the sand. Two ST300s were deployed, 1 m apart, at each of five sites, forming a 100 m-long transect with relative spacing of 0, 10, 33.5, 52.1 and 100 m from the first site, running approximately north-south, at the northern end of the lagoon (Figure 1, red dots). All Sound Traps recorded 290 of every 300 s, at a sampling frequency of 48 kHz. Gaps between the recordings were kept to separate files into manageable sizes and minimize the potential for losing recordings due to buffering issues, and the 97% duty cycle minimized the likelihood of missing the CPA of a vessel pass. At each SoundTrap, a tight vertical line was run from the ST300 to a surface buoy where the exact GPS location was recorded with a Garmin 64SX.

Between the 1st and 3rd December 2019, three vessels (Figure 2, characteristics shown in Table 1) each conducted ten transects at speeds as close as possible to 5, 10, 20, and 30 km h−<sup>1</sup> (1.4, 2.8, 5.6, and 8.3 m s<sup>−</sup>1, respectively). For each speed, five transects were conducted across the southern end of the line of SoundTraps and five across the midway point (Figure 1, white lines), totaling 100 potential recordings of each vessel at each speed. Vessel transects were planned to be orthogonal to the line of SoundTraps, though in reality were not completely perpendicular to the SoundTrap transect (Figure 1). Each vessel conducted all transects over a three-hour period, with one vessel completed each day. Vessel positions were recorded using a handheld Garmin 64SX. Wind over the three days remained at Beaufort scale two or below.

**Figure 2.** Photos of research vessels (**a**) *Primrose*, (**b**) *Macquarie 2* and (**c**) *Kirsty K*.


**Table 1.** Specifications of vessels recorded in Lizard Island lagoon.
