BRAHSS Publications

Journal papers and book chapters

Dunlop, R.A., Noad, M.J., Cato, D.H., Kniest, H., Miller, P.J.O, Smith J.N., Stokes, M.D. (2013)
Multivariate analysis of behavioural response experiments in humpback whales (Megaptera novaeangliae)
Journal of Experimental Biology, 216, 759-770

Available from (open access): jeb.biologists.org/content/216/5/759.full.pdf+html.
Cover: jeb.biologists.org/content/216/5.toc

The behavioural response study (BRS) is an experimental design used by field biologists to determine the function and/or behavioural effects of conspecific, heterospecific or anthropogenic stimuli. When carrying out these studies in marine mammals it is difficult to make basic observations and achieve sufficient samples sizes because of the high cost and logistical difficulties. Rarely are other factors such as social context or the physical environment considered in the analysis because of these difficulties. This paper presents results of a BRS carried out in humpback whales to test the response of groups to one recording of conspecific social sounds and an artificially generated tone stimulus. Experiments were carried out in September/October 2004 and 2008 during the humpback whale southward migration along the east coast of Australia. In total, 13 "tone" experiments, 15 "social sound" experiments (using one recording of social sounds) and three silent controls were carried out over two field seasons. The results (using a mixed model statistical analysis) suggested that humpback whales responded differently to the two stimuli, measured by changes in course travelled and dive behaviour. Although the response to tones was consistent, in that groups moved offshore and surfaced more often (suggesting an aversion to the stimulus), the response to social sounds was highly variable and dependent upon the composition of the social group. The change in course and dive behaviour in response to tones was found to be related to proximity to the source, the received signal level and signal-to-noise ratio (SNR). This study demonstrates that the behavioural responses of marine mammals to acoustic stimuli are complex. In order to tease out such multifaceted interactions, the number of replicates and factors measured must be sufficient for multivariate analysis.

Cato, D.H., Noad, M.J., Dunlop, R.A., McCauley, R.D., Gales, N.J., Salgado Kent, C.P., Kniest, H., Paton, D., Jenner, K.C.S., Noad, J., Maggi, A.L., Parnum I.M., Duncan, A.J. (2013)
A study of the behavioural response of whales to the noise of seismic air guns: design, methods and progress
Acoustics Australia, 41, 91-100

Available from: acoustics.asn.au/joomla/australian-acoustics-journal-april-2013.html.

The concern about the effects of the noise of human activities on marine mammals, particularly whales, has led to a substantial amount of research but there is still much that is not understood, particularly in terms of the behavioural responses to noise and the longer term biological consequences of these responses. There are many challenges in conducting experiments that adequately assess behavioural reactions of whales to noise. These include the need to obtain an adequate sample size with the necessary controls and to measure the range of variables likely to affect the observed response. Analysis is also complex. Well designed experiments are complex and logistically difficult, and thus expensive. This paper discusses the challenges involved and how these are being met in a major series of experiments in Australian waters on the response of humpback whales to the noise of seismic airgun arrays. The project is known as BRAHSS (Behavioural Response of Australian Humpback whales to Seismic Surveys) and aims to provide the information that will allow seismic surveys to be conducted efficiently with minimal impact on whales. It also includes a study of the response to ramp-up in sound level which is widely used at the start of operations, but for which there is little information to show that it is effective. BRAHSS also aims to infer the longer term biological significance of the responses from the results and the knowledge of normal behaviour. The results are expected to have relevance to other sources and species.

Dunlop, R.A., Noad, M.J., Cato, D.H. (2012)
Behavioural-response studies: problems with statistical power
In Effects of Noise on Aquatic Life, ed. Popper A.N. and Hawkins, A. Springer, New York, 293-297

The behavioural-response study (BRS) is an experimental design used by field biologists to determine the function and/or behavioural effects of conspecific, heterospecific, or anthropogenic stimuli. Although this has been used for many years in studies of acoustic stimuli and whales, many of these studies have been limited by lack of true replication, otherwise known as "pseudoreplication" (Hurlbert 1984 ; Kroodsma 1989 ; Searcy 1989) . The term "pseudoreplication" covers a specific range of experimental errors in which the sample size (n) used in a particular statistical test is not sufficient to test the hypothesis (McGregor 2000) ; either the stimulus is not representative (i.e., a true replicate) of the class of stimuli or the subject animals are not representative of the class or animals to which they belong. In a review of playback studies with marine mammals, Deeke (2006) found that only 17 of the 46 studies reviewed avoided pseudoreplication. A frequently used experimental design is a "before, during, and after" (BDA) design in which the subject is monitored before, during, and after presentation of a stimulus. One of the most common errors in the analysis of BDA data is to treat subjects within each BDA period as independent samples rather than as a repeated-measures design. Also, subjects may interact biologically within an experimental BDA (which relates to internal validity errors) or each subject may be repeatedly presented with the stimulus; therefore, samples are not independent, although they are treated as independent samples in the analysis. A number of variations of a stimulus may be used within the experiment, but the analysis wrongly pools the samples and treats these multiple recordings as one treatment. Finally, conclusions may be wrongly drawn about the function of, or response elicited by, a specific stimulus without true replication of the experiment. Invalid conclusions are then made about how applicable the result is to the population or species. Part of the problem is balancing the need to obtain adequate sample size to ensure statistically significant results against the high cost and logistical difficulty of such experiments. Many previous studies have been inconclusive through inadequate sample size, but this was not evident until after the data were analysed. Hence it is important to make a reliable estimate of the sample size required as part of the experimental design, such as by using a statistical power analysis. A power analysis calculates the number of samples needed to detect a change based on the variance in the response, but this requires some initial information about the statistical variability expected. Ideally, the power analysis should be carried out as part of a pilot study and, based on the results of this analysis, the sample size needed for the full experiment should be determined. In reality, field experiments are limited by time, money, or logistics, and such an ideal situation rarely exists.

Conference proceedings

Cato, D.H., Noad, M.J., Dunlop, R.A., McCauley, R.D., Kniest, H., Paton, D., Salgado Kent, C.P., Jenner, K.C.S. (2013)
Behavioral response of humpback whales to seismic air guns
International Congress on Acoustics and meeting of the Acoustical Society of America, 2-7 June 2013, Montreal

A study of the responses of humpback whales to seismic air guns is being conducted in Australian waters and two of four major experiments have been completed. It aims to assess the impact of seismic surveys on the whales and the effectiveness of ramp-up in mitigation. In separate trials, whales were exposed to a 20 cu in air gun, ramp-up in level from 20 to 440 cu in with an air gun array, and a "hard start" of 140 cu in. Trials exposing whales to air gun treatments were balanced by controls without air guns firing. Whales were tracked from land using theodolites. Behavioural observations were made from these land stations, from three small vessels and from the source vessel. Vocalising whales were tracked with an array of hydrophones. Dtags were attached to some of the whales. Observations were made before, during and after exposure. Characterization of the sound field throughout the area and the exposure at each whale were determined from propagation measurements and recordings on the hydrophone array and several moored acoustic recording systems. Some preliminary results will be discussed. [Work supported by E&P Sound & Marine Life Joint Industry Program and the U.S. Bureau of Ocean Energy Management].

Cato, D.H., Noad, M.J., Dunlop, R.A., McCauley, R.D., Gales, N.J., Salgado Kent, C.P., Kniest, H., Paton, D., Jenner, K.C.S., Noad, J., Maggi, A.L., Parnum I.M., Duncan, A.J. (2012)
Project BRAHSS: Behavioural Response of Australian Humpback whales to Seismic Surveys
Proceedings of Acoustics 2012, 21-23 November 2012, Fremantle, Australia

Available from (search Cato paper 137): acoustics.asn.au/conference_proceedings/AAS2012/index.htm.

BRAHSS is a major project aimed at understanding how humpback whales respond to noise, particularly from seismic air gun arrays. It also aims to infer the longer term biological significance of the responses from the results and knowledge of normal behaviour. The aim is to provide the information that will allow seismic surveys to be conducted efficiently with minimal impact on whales. It also includes a study of the response to ramp-up in sound level. Ramp-up is widely used at the start of operations as a mitigation measure intended to cause whales to move away, but there is little information to show that it is effective. BRAHSS involves four experiments with migrating humpback whales off the east and west coasts of Australia with noise exposures ranging from a single air gun to a full seismic array. Two major experiments have been completed off the east coast, the second involving 70 scientists. Whale movements were tracked using theodolites on two high points ashore and behavioural observations were made from these points and from three small vessels and the source vessel. Vocalising whales were tracked underwater with an array of hydrophones. These and other moored acoustic receivers recorded the sound field at several points throughout the area. Tags (DTAGs) were attached to whales with suction caps for periods of several hours. Observations and measurements during the experiments include the wide range of variables likely to affect whale response and sufficient acoustic measurements to characterise the sound field throughout the area. The remaining two experiments will be conducted further off shore off the west coast in 2013 and 2014.

Conference presentations

Dunlop, R.A., Noad, M.J., McCauley, R.D., Kniest, H., Cato, D.H. (2013)
The behavioural response of humpback whales (Megaptera novaeangliae) to a 20 cu in air gun stimulus
The 20th Biennial Conference Society of Marine Mammalogy, 9-13 December 2013, Dunedin, New Zealand

To determine the effects of anthropogenic noise on cetacean behaviour, behavioural response experiments are often carried out. One source of anthropogenic noise is from oil and gas industry seismic exploration. The BRAHSS (Behavioural Responses of Australian Humpback whales to Seismic Surveys) study aims to quantify the behavioural responses of humpback whales to various sizes of seismic air gun array. In the first experiment, a 20 cu in air gun was used as the stimulus. The objectives were to determine if there was any measureable behavioural response to noise from a 20 cu in air gun related to the proximity of the source vessel or received level and determine if social, environmental and temporal effects influenced this response. Each targeted humpback whale group was subjected to one of five different treatments; "active east" or "active north", where the source vessel travelled east or north whilst firing the air gun every 11 seconds, "control east" or "control north", where the source vessel travelled east or north with the air gun deployed but not firing, or "unexposed", where the group was followed through the study site but there was no source vessel. Response variables such as group dive time, group course and speed and group surface behaviour were analysed to determine if any of these were related to "treatment", "source vessel proximity" and "received level". Measures of social environment (for example nearest neighbour, nearest singing whale and group density), physical environment (wind speed, background noise and water depth) and temporal factors (time of day) were also included in the analysis. Much of the behavioural change observed was dependent on their social context and physical environment. The response to the 20 cu in air gun was difficult to separate from the response to these factors.

Kavanagh, A.S., Dunlop, R.A., Blomberg, S.P., Goldizen, A.W., Noad, M.J., Cato, D.H. (2013)
What drives the variability in humpback whale (Megaptera novaeangliae) movement and diving behaviour on migration?
The 20th Biennial Conference Society of Marine Mammalogy, 9-13 December 2013, Dunedin, New Zealand

Humpback whales undertake one of the longest migrations of any animal. While information is available on their travel speeds and surfacing rates, little is known about what factors drive these behaviours during migration. This study examines the effects of multiple social and environmental factors on the movements (speed and course) and diving behaviour (dive duration and surfacing interval) of humpback whales off the east coast of Australia. This research was carried out as a baseline study for a large behavioural response study examining the response of whales to seismic airgun noise (BRAHSS). Our methodology included both land-based observations and acoustic monitoring. Two hundred and eight hours of focal survey data were collected on 96 different groups of whales when no exposure experiments were occurring. Factors including group composition, involvement in social interactions, wind conditions, water depth and the proximity of singing whales to the focal group were found to influence the movements and diving behaviour of whales. For example, groups tended to spend less time in deep dives when being joined by other animals or when there was more than one singing whale in the area, suggesting that social environment significantly influences the whales' diving behaviour. Environmental conditions also affected dive behaviour, with the duration of dives increasing with water depth and wind speed. Results provide important baseline information on the drivers of natural behaviour in migrating humpback whales and demonstrate that multiple factors are simultaneously influential. Such information is not only important to our general understanding of their biology but also to behavioural response studies examining the effects of anthropogenic disturbances on humpback whales. To detect the effects of any disturbance, to put these effects into context and infer biological significance a clear understanding of natural behaviour is vital.

Noad, M.J., Kniest, H., Dunlop, R.A., McCauley, R.D., Cato, D.H. (2013)
Real-time, dynamic mitigation against temporary threshold shifts in humpback whale hearing during a large behavioural response study using air guns
The 20th Biennial Conference Society of Marine Mammalogy, 9-13 December 2013, Dunedin, New Zealand

Behavioural response studies (BRS) are used to determine the behavioural reactions of marine mammals to various acoustic stimuli. Some acoustic sources, however, are capable of producing high level sounds that have the potential to cause hearing damage in very close animals. In Australia, we conducted a large BRS using a small array of seismic air guns to examine the reactions of humpback whales to oil and gas exploration activities (BRAHSS). An array of six air guns was used to produce four steps in a ramp-up sequence with source sound exposure levels (SEL) of approximately 200, 204, 213 and 216 dB re1µPa2.s @ 1m. In a second treatment, the third step only was used. We designed a mitigation system to avoid exposures greater than 183 dB re1µPa2.s cumulative SEL, the current best estimate for the onset of temporary threshold shift from impulsive sounds in baleen whales. This contrasts with the more usual technique of setting a shut down range associated with a particular single impulse received level. Using land based theodolites and source vessel observers, we tracked every whale group in the area while the air guns were operating using VADAR software. VADAR used an empirical propagation model to calculate the single shot received level for each group every 11 secs taking into account the current source level. It then calculated a real-time cumulative SEL for each group and we shut down the air guns if this exceeded 180 dB for any group allowing a 3 dB margin of error. A detailed post-field recalculation of received levels shows that the maximum cumulative SEL received by any group was 183 dB re1µPa2.s. Avoiding harm to animals is a cornerstone of conducting ethical experiments and this study shows that, in the right circumstances, a real-time, dynamic mitigation system can be used effectively and safely.

Williamson, M.J., Dunlop, R.A., Kavanagh, A.S., Noad, M.J., Cato, D.H. (2013)
The short term effect of small research boats on the behaviour of humpback whales (Megaptera novaeangliae)
The 20th Biennial Conference Society of Marine Mammalogy, 9-13 December 2013, Dunedin, New Zealand

Corresponding author: m.williamson3@uq.edu.au. When conducting cetacean research, techniques such as tagging and biopsy are often used to collect behavioural and other data. These techniques usually require the research vessel to closely approach the animals to deploy a tag or biopsy dart onto the body of the whale. Even if tagging and biopsy are not part of the research methodology, accurate measurement of the behaviour of the animals requires the research vessel to remain in close proximity. This may change the targeted individual's behaviour. In studies of whale behaviour, it is important to take into account the effect of the research methodology. Any significant effect must be accounted for in the analysis. In this study we used behavioural observations from land-based platforms to quantify the effects of research boats and Dtag application on the behaviour of humpback whales. This was part of a larger project to assess the effect of seismic survey sounds on the behaviour of humpback whales as they migrated along the east Australian coast (BRAHSS). Land-based surveys of humpback whale focal groups were carried out during the southward migration in 2010 and 2011. Data from land platforms were collected from 42 different humpback whale groups that were approached by research vessels, of which 14 were Dtagged. Observations were taken before the research vessel approached a whale group, during the approach, and after the approach while the vessel followed the group and stayed approximately 100 m away. An additional 40 groups that were not followed by research boats were used as controls. Statistical mixed models with multiple predictor variables such as approach type, boat type and tag success, were used for analysis. The change in response variables such as swim speed, course, surface time and dive time with and without research vessels present were evaluated.

Cato, D.H., Dunlop, R.A., Noad, M.J., McCauley R.D., Kniest, H., Paton, D., Salgado Kent, C.P. (2013)
Addressing Challenges in Studies of Behavioural Responses of Whales to Noise
Third International Conference on the Effects of Noise on Aquatic Life, August 11-16, 2013, Budapest, Hungary

Studying the effects of noise on whale behaviour presents many challenges. Well-designed experiments, with adequate sample size and the necessary controls to obtain reliable results, are complex, logistically difficult and expensive. The experimental design should be able to detect and measure any change in whale behaviour, as well as determine the factors that drive this change in behaviour. Some factors will be related to the noise exposure, such as received level and proximity of the source, but the behaviour may also depend on other factors such as the social context and the physical environment. Attempting to account for all experimental, social, environmental, temporal and measurement predictor variables can lead to a complex analysis model. The results also need to be placed into the context of the normal behaviour of the whales to aid in evaluating the biological significance of the response. To be effective, the team of investigators need to have adequate expertise to cover the range of disciplines, from the acoustics to the biology. This paper discusses how these challenges are being met in project BRAHSS (Behavioural Response of Australian Humpback whales to Seismic Surveys) and presents insights gained from completing two of the four planned experiments. Some preliminary result of the first experiment (using a 20 cu in stimulus) will be discussed in the context of highlighting the issues and challenges in analysing this kind of data. As much of the behaviour observed was dependent on the social context and physical environment, it was difficult to separate direct response to the air gun from these factors.

Noad, M.J., Cato, D.H., Dunlop, R.A., McCauley, R.D. (2013)
An interdisciplinary approach to measuring behavioural impacts of seismic surveys on humpback whales
27th European Cetacean Society Conference, 8-10 April 2013, Setúbal, Portugal

Seismic air guns are devices used to find oil and gas deposits under the seafloor. They produce regular loud percussive sounds for extended periods, and there is concern that these sounds may have a negative impact on nearby marine mammals. Determining the impacts of various industrial activities in the marine environment can be very difficult, particularly behavioural effects on marine mammals. We are currently two years into a four to five year behavioural response study exposing humpback whales to the sounds of air guns as they move through our study site on the east coast of Australia. The aims of the study are to record the behavioural responses of the whales to the air guns and determine which factors involved in the interaction are most strongly correlated with the degree of behavioural change (e.g. received level, proximity of the airgun, social behaviour of the whale group), whether or not these behavioural changes are likely to have longer term biological effects, and to test the efficacy of "ramp up" as a mitigation tool. The whales were followed from three different types of platform: elevated land stations, small boats and high resolution behavioural tags attached to the whales (Dtags). The study followed a "before, during, after" design with focal groups of whales observed and tracked as they migrated past a source vessel. Treatments included firing a 20 cubic inch (in3) gun towed on two different paths, firing a 140 in3 array of air guns, and "ramping up" from 20 to 440 in3. All active treatments were balanced by controls with the source vessel moving but the air guns not firing. Acoustic loggers, drifting vertical arrays and an array of moored hydrophone buoys were also deployed to measure the acoustic environment. Future experiments will include the use of a full seismic array.

Godwin, E.L., Dunlop, R.A., Noad, M.J. (2013)
Comparing multiple methods for measuring the behaviour of our humpback whales (Megaptera novaeangliae)
27th European Cetacean Society Conference, 8-10 April 2013, Setúbal, Portugal

Measuring the behaviour of marine mammals is challenging due to the fact they are underwater and not observable most of the time. Given the limitations of many sampling methods and the difficulties associated with measuring animal behaviour, it may be more practical to use a combination of methods rather than just one. However, the inherent strengths and limitations of these different methods, as well as the variability in the scale of data captured needs to be correctly accounted for. Little is known on whether data captured from these different methods are similar and therefore comparable, or different and in need of some sort of correction. The experimental design of the Behavioural Response of Australian Humpback whales to Seismic Surveys (BRAHSS) project incorporated a number of different sampling methods; land-based and boat-based visual observations of surface behaviours and movements of humpback whale groups as well as Digit al Tags (DTAGs), which recorded ultra-fine scale underwater movements and acoustic behaviour. This study used data from the BRAHSS project to compare measures of behaviour collected using all three sampling methods. Land-based observations tended to underestimate rates of surface behaviour (such as blow rate) compared to boat-based observations (blow rate was not measured from the tags). Broader scale movement variables (course and speed) were measured similarly from land-based and boat-based platforms. Diving behaviour fell into two distinct types ("short" dives <60 seconds and "long" dives >60 seconds). When compared with dive data obtained from tags, both land-based and boat-based observations were found to reliably estimate the number of long dives however land-based observations underestimated the number of short dives. This study demonstrates that measuring behaviour using different methods produces some discrepancies in data collected; however these discrepancies are more prominent in some measures of behaviour than others.

Paton, D., Cato, D.H., Noad, M.J., Dunlop, R.A., McCauley, R.D., Kniest, H., Salgado Kent, C.P., Jenner, K.C.S. (2013)
Behavioral response of humpback whales to seismic air guns
Advantage New Zealand: 2013 Petroleum Conference 28 April - 1 May 2013, Auckland, New Zealand

The behavioural impacts of seismic air guns on whales are not well understood. Current impact mitigation protocols for seismic surveys are based largely on avoiding damage to whales' hearing, and do not take into account behavioural effects, which may occur at greater ranges. BRAHSS is a major scientific research project aimed at understanding how humpback whales respond to seismic air gun arrays and to infer longer term biological significance of the responses. The project is designed to record the behavioural responses of the whales to the air guns and determine which factors involved in the interaction are most strongly correlated with the degree of behavioural change (e.g. received level, proximity of the airgun, social behaviour of the whale group). In addition the study will test the efficacy of "ramp up" as a mitigation tool. BRAHSS involves four experiments with migrating humpback whales off the east and west coasts of Australia with noise exposures ranging from a single air gun to a full seismic array. Two major experiments have been completed off the east coast, the second involving 70 scientists. The experiments followed a "before, during, after" design with focal groups of whales observed and tracked as they migrated past a source vessel. Whale movements were tracked using theodolites from three land stations and behavioural observations were made from these points and from three small vessels and the source vessel. Vocalising whales were tracked underwater with an array of hydrophones, and other moored acoustic receivers recorded the sound field at several points throughout the area. Tags (DTAGs) were attached to whales with suction cups for periods of several hours. Observations and measurements during the experiments include the wide range of variables likely to affect whale response and sufficient acoustic measurements to characterise the sound field throughout the area. The remaining two experiments will be conducted off the west coast and further off shore in 2013 and 2014.

Dunlop, R.A. (2012)
High resolution data from DTAGS on the response of humpback whales to noise from seismic air guns
Meeting of the Acoustical Society of America, 14-18 May 2012. Journal of the Acoustic Society of America, 131, 3522 (abstract only) (Invited paper)

The BRAHSS (Behavioural Response of Australian Humpback Whales to Seismic Surveys) series of experiments uses a multi-platform approach to determine the behavioural and acoustic response of humpbacks to seismic air gun array sounds. One of the data collection platforms utilises the DTAG, or acoustic digital recording tag. DTAGs are small suction cup tags which contain a hydrophone as well as x, y and z plane accelerometers, magnetometers and depth sensors. Data from the DTAGs allows fine scale movement data of the tagged whale (dive profile, pitch, roll and heading movements, fluking rates) to be viewed as a pseudotrack whilst simultaneously listening to sound field of the whale (air gun shots and vocal sounds from nearby whales). This presentation summarises some results of responses of the whales to air gun sounds and compares these with visual observations made at the time. Since DTAGs provide a continuous record of whale vocalizations and 3D movements, vocal and physical reactions are detectable in response to the air gun sounds, whereas visual observation are limited to the appearance of the whales at the surface.

Cato, D.H., McCauley, R.D., Noad, M.J., Dunlop, R.A., Kniest, H., Gales, N.J, Salgado Kent, C.P, Paton, D., Noad, J., Duncan, A.J., Maggi, A. (2011)
Acoustic issues in studies of behavioural response of humpback whales to seismic ramp-up and hard start
Meeting of the Acoustical Society of America, 14-18 May 2012. Journal of the Acoustic Society of America, 131, 3457 (abstract only)

Two large behavioural response studies (BRS) have been conducted with humpback whales migrating along the east Australian coastline (in project BRAHSS: Behavioural Response of Australian Humpback Whales to Seismic Surveys). Whales were exposed to four stages of ramp-up with nominally 6 dB increase in level at each step, and a hard start nominally 12 dB above the first stage. Observations of behaviour were made by theodolite teams ashore and small boats following specific whale groups, DTAGs, and binoculars from the source vessel. The sound field throughout the area was recorded using five buoys that radioed data back to the shore station, four autonomous receivers and two drifting systems with a vertical array of four hydrophones. Measurements show that the propagation loss at the site is variable and includes patches of anomalously high loss. This complicates estimation of the sound levels received by whales, but may not be unusual in near shore environments. This paper presents preliminary results of the project to illustrate acoustic issues involved in designing and executing comprehensive BRS, including characterization of sources and the acoustic environment experienced by the whales, and monitoring cumulative exposure at individuals for mitigation.