Predicting airgun impact on marine life

Overview

A conventional airgun array as used in the seismic data acquisition industry is a powerful source of acoustic waves, so powerful that it is important to assess any potential adverse impact on marine life.

Animals react in complex ways to sound. Here we will simply quote relevant sources of information to help quantify the overall impact as implemented in Gundalf and we will refer the reader to

  • "Marine Mammal Noise Exposure Criteria: Updated Scientific Recommendations for Residual Hearing Effects" by Southall et. al. (2019), Aquatic Mammals 2019, 45(2), 125-232, DOI 10.1578/AM.45.2.2019.125.

  • "Technical Guidance for Assessing the Effects of Anthropogenic Sound on Marine Mammal Hearing" National Oceanic and Atmospheric Administration (NOAA), (2016) NOAA Technical Memorandum NMFS-OPR-55 July 2016.

  • "Marine mammals and noise" Richardson, Greene, Malme and Thomson, (1995), Academic Press ISBN 0-12-588441-9.

  • "Marine Mammal Noise Exposure Criteria: Initial Scientific Recommendations" by Southall et. al. (2007), Aquatic Mammals (33) 4, p. 411-509 ISSN 0167-5427.

Until relatively recently, there was no really high quality broadband data (up to at least 25 kHz.) available to calibrate sophisticated airgun modelling systems like Gundalf. In 2003, the IFRC funded the acquisition of such data and permitted this to be made available to the Gundalf designers by Phil Fontana of Veritas. These data were used to demonstrate that Gundalf is within a few db. of the measured value at 18kHz. For full details see:- PETEX 2004 presentation by Les Hatton

These data were supplemented by the JIP sponsored Svein Vaage experiment in 2009-13.

Given then that Gundalf can predict an airgun array spectrum out to at least 25 kHz. with good accuracy, we can begin to include other factors which impact on the environment to produce a reasonable prediction of the overall contribution of an airgun array to marine environmental noise. The principle additional factors we must include are:-

  • Absorption: Absorption in the sea has a negligible effect on the principle bandwidth of interest in exploration seismology which is typically 0-200 Hz. however it plays an increasingly important part at the higher frequencies so this effect is included in swept-area plots in Gundalf under option.

  • Spreading: The way the wave-front spreads as it propagates has a direct effect on the amplitude at all frequencies, In practice, it is an extraordinary difficult phenomenon to model as it requires detailed knowledge of the sub-surface and also the salinity variation in the sea itself however it is parameterised in Gundalf so that the user can input anything from super-cylindrical (perhaps intense focussing effects) to super-spherical. In practice, for many environments, the spreading factor is likely to be rather closer to spherical (point source spreading) than cylindrical (line source spreading).

    This has been supplemented with more sophisticated empirical spreading models, particularly building on the Marsh-Schulkin models, Marsh H.W. and Schulkin M. (1962) Shallow-water transmission, J. Acoust. Soc. Am. 34(6), p. 863-864).

  • Animal audiogram data: As has been pointed out by numerous authors including Jeremy Nedwell and Gary Hampson (who kindly made their original data and research available to the Gundalf designers), animals respond to sound in a highly frequency dependent way and this should be allowed for when calculating environmental noise impact of airgun source arrays. These data are often shown in audiogram form as can be seen in the diagram below.

Graphic1

In the last few years, there has been considerable research into how such sensitivity information should be used. Originally, audiogram data was used to weight the predicted spectra appropriately. This is no longer considered to be the most appropriate way of allowing for this phenomenon, so from Gundalf 8.1d, the data was applied in the form of "M-Weighting" (see Southall et. al. p.433-) and even more recently in the form of "Cetacean Auditory Weighting Functions", (see NOAA (2016), Southall et al (2019)).