Download the publication
MIT2024-07 Hector's Dolphin Acoustic Deterrence Devices in Trawl and Set Net Fisheries (PDF, 4,025K)
Summary
Two types of acoustic devices are regularly utilised in New Zealand waters to alert marine mammals to the presence of fishing gear, and/or to deter them from areas of risk or potential feeding opportunities. The two devices are the Netguard Dolphin Pinger (60 – 120 kHz model) and the STM Dolphin Deterrent Device (DDD) (DDD03H model). In 2025, JASCO Applied Sciences (Australia) (JASCO) undertook a measurement and modelling study (Warren et al. 2025) for Seafood New Zealand and the New Zealand Department of Conservation to verify the source characteristics and calculate underwater sound propagation associated with the two acoustic devices. This study reports on at-sea measurement trials conducted in April 2026 off the coast of Timaru, New Zealand that aimed to test the modelled assumptions (Warren et al. 2025) in a real world scenario by measuring the propagation of sound emitted by the devices and comparing it to the modelled outputs.
Four Netguard Pingers (with 60 m spacing) were mounted to the float-line of a bottom-mounted set net, and in a separate trial, two STM DDDs were suspended at 3 m above the seabed, 30 m apart to approximate their positions when mounted on the wing-ends of a trawl net. Acoustic measurements were undertaken using both static and mobile receivers to consider whether the propagation environment varied over time (temporal variability) and to examine how sound propagated over distance and depth (spatial variability). The static receiver was a baseplate JASCO Autonomous Multichannel Acoustic Recorder (AMAR) that was deployed onto the seabed. The mobile receiver was a JASCO Ocean Sound Meter-2 (OSM2) with a cabled hydrophone that was dipped from a vessel to near-surface, mid-water, and near-bottom depths. Recording bandwidths were 10 – 256,000 Hz for both receivers. Measurements were undertaken during calm weather, with wind speeds below 10 knots and wave heights of less than 0.1 m (Sea State 1). Nonetheless, significant drift in vessel location occurred during the mobile measurements. Careful notes and GPS positions were taken for each mobile measurement. For the static receiver, and for each dip obtained with the mobile receiver, average and maximum 1-s SPL values across the recording period were obtained, as well as an ‘average maximum’ value, which considered peaks in sound level in the acoustic data series, with each peak in SPL assumed to be driven by emissions of the Pingers or DDDs.
Netguard Pinger results
Overall, the ambient noise environment across all frequency bands remained approximately consistent across the measurement period, as recorded by the static receiver. There were no notable trends in level across the recording period, such as might be caused by tidal movement or the arrival of another sound source (biological, geological or anthropogenic). Analysis focused on the 63 kHz-centred decidecade frequency band for the Netguard Pingers as this is the band that ‘pings’ are emitted in. The maximum received levels in the 63 kHz-centred decidecade frequency band, and to a lesser extent, the average maximum received level in the band, were higher at closer distances to the nearest Netguard Pinger which aids in confirming that the intended sound sources (the Netguard Pingers) were driving the value of the 63 kHz-centred decidecade frequency band, and external sound sources did not unduly influence the results.
Overall, the Netguard Pingers functioned as intended, and the entire net region was ensonified by the ‘pings’ emitted by the four devices. This achieves the primary aim of attaching Pingers to set nets, which is to provide an alert system to marine mammals to raise awareness of an object of potential risk. The Pinger spacing applied (60 m) was more than adequate in this environment to avoid ‘gaps’ in the sound field along the length of the net.
The real-world underwater environment was more conducive to sound propagation than the environment applied within the modelling, with higher received levels recorded at all horizontal distances than were predicted by the modelling. The better real-world propagation could have been a result of different sediment properties (the model considered pure sand), bathymetric features (the model considered a perfectly flat seafloor), or sound speed profile parameters in the water column (the model considered a profile for the Timaru area for the month of July as a conservative choice).
Within the modelling study, the environment with the longest propagation had a range to detectability of a Netguard Pinger of 60 m for Hector’s and Māui dolphins. Within this measurement study, the real-world underwater environment was more conducive to sound propagation than the environment applied within the modelling and the detectability threshold relevant to Netguard Pinger emissions was found to be exceeded at approximately 60 to 120 m. Whilst this was the case for this specific location, the conducive propagation environment is unlikely to be present at all locations in the New Zealand marine environment. Therefore, it is not recommended that the results from this trial should be extrapolated to other locations around New Zealand, or considered to be the norm for this location given the calm sea conditions. Nonetheless, the results from this measurement trial and the original modelling study could be considered in combination to produce future recommendations for spacing of Netguard Pingers on set nets.
While underwater noise measurements and modelling can speak to the temporal and spatial propagation of the sound fields emitted by the acoustic devices, to be confident that the devices function as effective bycatch mitigation tools, the responses of marine mammals to the emitted sounds must be confirmed. This is identified as an area in need of future research as there is limited understanding of the behaviour of free-ranging marine mammals in New Zealand, let alone in relation to active fisheries operations. Animal-mounted telemetry tags that record fine-scale movement, in combination with passive acoustic monitoring equipment at fishery sites, could form a basis for initial investigations on the performance of Pingers as a bycatch mitigation tool.
STM DDD results
During the measurements of the STM DDDs, the ambient noise environment across all frequency bands remained approximately consistent across the measurement period, as recorded by the static receiver. Analysis for the STM DDDs focused on the 25 kHz-centred decidecade frequency band, which was the band with the longest modelled propagation and is a local maxima within the source spectra. Towards the beginning of the measurements, a passing vessel contributed sound energy below 3000 Nonetheless, the maximum received levels in the 25 kHz-centred decidecade frequency band, and to a lesser extent, the average maximum received level in the band, were higher at closer distances to the STM DDDs which aids in confirming that the DDDs were the primary driver of the level of the 25 kHz-centred decidecade frequency band in the underwater environment.
In order to compare to modelled results, the 25 kHz-centred decidecade frequency band were corrected to an estimate of the broadband received level. As with the Netguard Pinger measurements, the real-world underwater environment was more conducive to sound propagation than the environment applied within the modelling, with higher received levels recorded than were predicted by the modelling.
The original work conducted by Warren et al. (2025) verified the sound source characteristics of the STM DDDs, and this measurement study has validated the overall outputs of the modelling study in terms of the propagation of sound, albeit with less transmission loss affecting propagation than was expected from the modelled scenarios. The sound source characteristics and verified modelling approach could now be used in future modelling work to consider the ensonified areas arising from multiple DDD sources, such as might occur across a fleet of fishing vessels operating in a specific area. Such modelling could contribute to management-level decision-making by quantifying acute and cumulative noise exposure that would be experienced by local marine mammal species, potentially allowing for a balance to be struck between adequate deterrence from risk areas and excessive noise pollution within the habitat. Further research to understand how free-ranging marine mammals respond to the DDD sound source would also be valuable, to ensure the devices function as effective bycatch mitigation tools.
Publication information
Warren, V.E., S.C. Connell, and C.R. McPherson. 2026. MIT2024-07 Hector's Dolphin Acoustic Deterrence Devices in Trawl and Set Net Fisheries: Measurements of Sound Propagation At-Sea in New Zealand’s Coastal Waters. Document 04323, Version 2.0 FINAL. Technical report by JASCO Applied Sciences for Department of Conservation, Te Papa Atawhai. 67 pp.