Marine Seismic Sources Part X: Seismic Surveys and Fish

Author Martin Landrø and Lasse Amundsen

The possible effects of seismic surveys on fish and fisheries have been given a great deal of media attention in Norway. Here, we summarise results from the Fish Rock experiment offshore Scotland.

Firemore Bay on the coast of North-West Scotland was the site of a 2001 study on the impact of seismic shooting on fish. Source: Ron Washbrook/Dreamstime.com

The possible effects of seismic surveys on fish and fisheries have been given a great deal of media attention in Norway. In the last edition of GEO ExPro we reported on results from two investigations offshore Norway in 2009 and 1992, carried out to study if airgun activity affected fish distribution and commercial fisheries.

For the 2009 seismic survey, the main conclusions according to the Norwegian Petroleum Directorate were that seismic shooting resulted in increased catches for some species and smaller catches for others. It appeared that pollack, to some extent, may have withdrawn from the area, while other species seemed to remain. Here, we summarise results from the Fish Rock experiment offshore Scotland.

Time amplitude signature of triple G. guns recorded at 16m distance. Spectrum of the sound signature. Source: Adapted from Wardle et al (2001)

The Fish Rock Experiment

Although behavioural studies of fish suggest that there might be some changes in behaviour associated with seismic surveying, a study by Wardle et al (2001) found results to the contrary. Bangs did not chase fish away, but they did cause an involuntary sudden bending of the body, or C-starts (see box).  

The investigators used a video system to examine the behaviour of fish on an inshore rocky reef, ‘Fish Rock’ in Firemore Bay, Loch Ewe, on the west coast of Scotland, in response to shooting a stationary triple G. airgun (three synchronised airguns, each gun 150 cu in (2.5 l) and 2000 psi). Fish inhabiting this reef include juvenile saithe that leave for the open sea when they are about three years old, adult pollack, juvenile cod, with some flatfish, wrasse and gobies. The water depth is 10–20m.  

The G. guns represent a type of gun now commonly used by survey companies in arrays and clusters for seismic survey work. The guns were fired once per minute for eight periods on four days at different positions. The peak–peak sound level was 210 dB re 1 μPa @ 16m from the source and 195 dB @ 109m from the source. We note that a 210 dB equivalent pressure would be received at about 100m below a full-scale seismic airgun array generating about 250 dB re to 1 µPa @ 1m.  

Firing the G. guns had little effect on the day-to-day behaviour of the resident fish, which were not sufficiently irritated by the shooting to move away from the reef.  

However, reef fish watched by the TV camera showed involuntary reactions in the form of C-starts at each explosion of the guns at all ranges tested. When the explosion was not visible to the fish, the C-start reaction was cut short and the fish continued with what they were doing before the stimulus. In one experiment, when the guns were suspended mid-water (5m depth) and just outside visible range at 16m, the fish receiving a 6 ms peak to peak, 206 dB pressure swing exhibited a C-start and then continued to swim towards the gun position, their intended swimming track apparently unaltered.  

The long term day-to-night movements of two tagged pollack were observed. One of them showed little variation at the onset of and during gun firings, but was never closer than 35m from the guns. The other pollack showed detailed reactions to the gun only when it is brought close to the fish. At one onset of firing, the pollack was about 10m from the gun location. After the first firing the fish moved rapidly away from the gun by approximately 30m.  

It is noted that the fish involved In the Fish Rock experiment are mainly inshore and reef species, closely associated with a home territory and not easily moved. In contrast, other open-sea experiments have found indications of large-scale influences resulting in apparent movements of commercial fish species, for example, making them more or less accessible to fisheries.

The C-start response When a fish receives a strong sound stimulus, or is attacked, an alarm reaction or an escape reaction is triggered. The reaction is known among scientists as the ‘C-start’ response, and it is perhaps one of the best-studied fish behaviour patterns. The ‘C-start’ takes its name from the shape the fish makes as it rapidly starts its escape from danger. The drawing shows that a threatened fish will perform two distinct movements to change direction and speed away from the threat. First, it will curve its body into the shape of a ‘C’ away from the source of sound. Second, it will then straighten its body from the curved position. The straightening motion allows the fish to push water off the full broadside of its body to quickly swim away from its attacker.   Field experiments have demonstrated that sound energy transmitted from air guns initiates this type of response on various fish species. In particular, intense infrasound results in escape reactions.

Startle Threshold

Fish react to manmade sound in various ways. The weakest form of response is minor changes in swimming activity where the fish change their direction or increase their speed. The most significant response on sound is an escape reaction where the fish initially show the C-start response.  

Few controlled studies of startle response thresholds in different fish species and auditory groups due to sound pulses of varying frequencies have been made. It has been reported, however, that for hearing generalists the C-start response is triggered at a far-field sound pressure of 174 dB re 1 μPa @ 10 Hz, and 154 dB @ 100 Hz. These numbers indicate that the threshold for triggering rapid escape responses is significantly above the absolute hearing threshold. Further, Karlsen (2010) gives startle behaviour responses for codfish starting at around 160–175 dB for a frequency of about 100 Hz. His observation indicates that their startle threshold values are around 80 dB above known auditory thresholds.

Images from video tape of three 30–40 cm saithe swimming towards the airgun at 16m range and about 4m from the sea bed and 10m from the surface. When the gun fires they show the typical C-start, veer off course and then continue swimming in the direction of the gun. All three saithe show the reaction in the same TV frame (Frame 2). Note the sound pulse, lasting 6ms, travels 30m during one TV frame of 20ms and the visual range is about 6m. The first three images are 20ms apart, the fourth frame is 5s later.   Source: Wardle et al (2001)

To precisely relate any startle response thresholds of fish to marine seismic activity is a difficult task since seismic sound propagation in the sea is a rather complex subject. Often, one oversimplifies the problem and assumes that sound attenuates with spherical spreading, 20 log(R), or cylindrical spreading, 10 log(R), where R is the distance from the seismic source. Spherical spreading applies to loss in deep oceans and cylindrical spreading to an ideal waveguide with perfectly reflecting boundaries at the sea surface and the water bottom. Such conditions are often encountered in the oceans for sound that strikes the bottom at angles greater than a critical angle. It follows that spreading loss in a waveguide such as the sea, with constant speed of sound, follows a spherical spreading law at short distance and cylindrical spreading at longer distance. A combination of the two spreading laws gives for distances R greater than the ocean depth Z in metres (m), the asymptotic loss behaviour:  

20log(Z) + 10log(R/Z)  

However, it is well know that the bathymetry and the composition of the ocean bottom, whether soft or hard, is important for long-range propagation. In addition, sound propagation changes with the oceanographic conditions and thereby the season. Therefore, if one wants to scientifically consider sound propagation under specified ocean conditions, one has two options: to measure or model seismic sound propagation. Obviously, it would be costly to measure sound propagation from seismic activity in the water column everywhere where seismic is acquired. The realistic alternative is to develop mathematical-acoustic simulation models which describe how sound propagates in the sea at long distances from the seismic source. Inputs to such acoustic models are source information and available geological and oceanographic information.

Auditory and startle thresholds Auditory and startle thresholds for codfish, which are hearing generalists with medium hearing ability. The audiogram (black curve) gives the faintest sounds that can be heard at each frequency. The startle response level (red curve) is assumed to be around 80 dB above the known hearing threshold. The red curve is displayed as a smoothed version of the black curve, added around 80 dB. Fish species react very differently to sound. Therefore, any generalisation about the effects of sound on fish should be made with care. The reactions of fish to anthropogenic sound are expected to depend on the sound spectrum and level, as well as the context (e.g. location, temperature, physiological state, age, body size, etc.)

Simulation Models

Combined with knowledge of fish hearing and their startle thresholds, simulation models can be used to estimate the distance various fish species are affected by seismic activity. Such simulation models are under development. Among others, the Norwegian Petroleum Directorate (NPD) has commissioned SINTEF and the University of Oslo to develop an acoustic-biological model to predict the impact of seismic sound on the fish population. The ultimate goal is to develop an acoustic-biological model to use in the design and planning of seismic surveys, such that possible disturbance to fishing interest is minimised. Maybe, in the future, in a similar way as acousticians compute noise maps around airports due to airplane takeoff, underwater acousticians will be able to compute sound propagation maps in the sea due to seismic shooting?  

In summary, seismic surveys may introduce a behavioural change of fish in the vicinity of the seismic source. The radius of the affected zone will depend on many variables, like the local physical conditions of the sea, the food supply for the fish and the behavioural patterns of the fish. Fish with natural habituation will be more steadfast than shoals of fish migrating through an area. Therefore it may be difficult to accurately determine the exact impact of seismic on the behaviour of fish. However, as long as their prey does not vanish, the steadfast fish will return.

References:  

C. S. Wardle, T. J. Carter, G. G. Urquhart, A. D. F .Johnston, A. M. Ziolkowski, G. Hampson, D. Mackie 2001 Effects of seismic air guns on marine fish: Continental Shelf Research 21, 1005–1027.  

J. Dalen et al 2008 Kunnskapsstatus og forskningsbehov med hensyn til skremmeeffekter og skadevirkninger av seismiske lydbølger på fisk og sjøpattedyr: Rapport til Oljedirektoratet, Fiskeridirektoratet og Statens Forurensningstilsyn fra spesielt nedsatt forskergruppe.  

H. E. Karlsen 2010 Hørsel og reaksjoner på lyd hos fisk http://www.olf.no.
 

Martin Landrø is a professor in Applied Geophysics at the Norwegian University of Science and Technology (NTNU), Department of Petroleum Engineering and Applied Geophysics, Trondheim, Norway.   Lasse Amundsen is Chief Scientist Exploration Technology at Statoil. He is adjunct professor at the Norwegian University of Science and Technology (NTNU) and at the University of Houston, Texas.