Rebecca Dunlop

Quantifying the risk of collision between humpback whales and vessels

As many large whale species and populations recover from exploitation, there is a substantial increase in the numbers of whales inhabiting populated coastlines. During the time these coastlines have developed and become more populated, there has also been a large increase in the number, size, and speed of vessels. This has resulted in an increased probability that large whales will collide with vessels. When large ships collide with whales, they can injure or kill the whales but are unlikely to damage the ship. In collisions with smaller vessels, there is a higher risk of damage to the vessel, injury to the whale and, most importantly, injury to passengers and crew. Therefore, both the International Whaling Commission (IWC), and Conservation and Scientific Committees, are examining ship strike as an emerging and important issue. The IWC, for example, has focused on developing a strategic plan to mitigate ship strike impacts, and aims, by 2020, to achieve a permanent reduction in ship strikes.

Strategies to mitigate for collisions between whales and vessels are not used globally, as there must be some identifiable collision risk. The easiest way to identify and quantify a collision risk for a species within a particular area is to use simple temporal estimates of species density overlayed on shipping routes and lanes; known as a “static model”. An increase in species density close to heavily used shipping channels would be given a high collision risk. However, these models do not account for the movement of the whales relative to the ships in that whales may avoid the ship to prevent collision. Further, given there is no inclusion of behavioural response data, it is difficult to say how mitigation measures such as a reduction in vessel speed would reduce the risk collision without making generalised assumptions. “Dynamic” models include information on how whales behave around different types of vessels in terms of their avoidance strategies, which factors dictate the use of these strategies (e.g., a female with a calf may use a different strategy to a group of adults), and which cues they use (e.g., received level of noise, vessel proximity, vessel speed and trajectory). From these dynamic interaction models, the risk of collision can be quantified much more accurately as well as changes in the risk with changes in vessel speed. However, dynamic models require much more information that the basic static model meaning there are few available.

The PhD project will collect behavioural response data from a field site based at Caloundra, on the Sunshine coast. Here, the shipping channel is relatively close to shore, and is located within humpback whale migratory corridor. Ships are moving in and out of Moreton Bay daily. During the humpback migration, ships are moving at speed, and close to, migrating humpback whale groups. This offers an opportunity to collect behavioural response data on the response of groups to fast-moving ships, as well as the factors that contribute to this response such as the vessel’s speed, size, proximity to the group, and received level of noise. These data will be used to generate both static and dynamic models of the risk of collision risk between humpbacks and vessels and compare these models. Once models are created, various mitigation measures will be introduced to the models, such as reduction in vessel speed, and the risk of collision compared. Outcomes will inform the assessment of risk for industry (reputational risk for port authorities), the environment (risk of whales injured or killed, and safety (human injuries and possible fatalities) and develop globally applicable mitigation measures to reduce these risks.

Humpback whale breeding behaviour and sexual signalling

Humpback whales are renowned for their complex acoustic communication repertoire. For example, male humpback whales utilise a wide and varied acoustic communication repertoire whilst undertaking breeding interactions. They use song, which likely functions as a sexual selection signal directed at females, and/or use social sounds, which likely function as female sexual selection signals as well as male-male interaction signals. Song also may be a male-male interaction signal in that eavesdropping males can gain information from the singing male, even if the song is not directed at them.

To find females, males switch tactics between singing and ‘seeking’ (i.e., actively seeking out a female and joining with her, which can lead to fighting with other males (Dunlop and Frere 2023). Their choice of tactic is significantly related to the density of other males within their ‘social circle’. In low male densities, where competition for females is low, males tend to sing. In higher male densities, males will cease to sing and switch to the ‘seeker’ tactic. This is likely because of the balance of costs and benefits of each tactic. If choosing to sing, the male may attract a female, however, the risk is this male may attract other eavesdropping males that can interrupt his song and displace him from the area if alone, or from the female, if with a female (Dunlop and Noad 2016, 2021). In higher male densities, the seeker tactic may be more successful given the increased competition. If quietly seeking out a female rather than advertising using song, there is less risk of attracting an eavesdropping male. However, despite the fact much is known about these breeding behaviours, the information contained within the song, in terms of singer’s fitness, is currently unknown.

Following these studies, the PhD project will determine if there are parameters in the song that are likely to encode the singer’s fitness. It will utilise behavioural datasets of singers and their breeding interactions that have been collected during various field seasons from the late 90’s to mid-2000’s. Song parameters that may signal fitness, such as unit peak frequencies, unit duration, phrase repetition rate, source level, will be compared across different male breeding to test the hypothesis that fitter males are those ones that successfully join a female whilst not attracting male competition, whereas less fit males are those that attracted male competitors. Ultimately, this will improve our understanding of acoustically-mediated breeding behaviour in humpback whales.

There is also the potential to collect more focussed data during this PhD. For example, collecting fitness information on individual singing males, such as body condition using drone photogrammetry and testosterone levels using biopsy samples. This may provide an opportunity to further test specific findings from the song analysis. This will depend on the student’s ability to seek project funding noting that many past students in the lab have had successful grant applications.

Quantifying the risk of collision between humpback whales and vessels

Funded project

As many large whale species and populations recover from exploitation, there is a substantial increase in the numbers of whales inhabiting populated coastlines. During the time these coastlines have developed and become more populated, there has also been a large increase in the number, size, and speed of vessels. This has resulted in an increased probability that large whales will collide with vessels. When large ships collide with whales, they can injure or kill the whales but are unlikely to damage the ship. In collisions with smaller vessels, there is a higher risk of damage to the vessel, injury to the whale and, most importantly, injury to passengers and crew. Therefore, both the International Whaling Commission (IWC), and Conservation and Scientific Committees, are examining ship strike as an emerging and important issue. The IWC, for example, has focused on developing a strategic plan to mitigate ship strike impacts, and aims, by 2020, to achieve a permanent reduction in ship strikes.

Strategies to mitigate for collisions between whales and vessels are not used globally, as there must be some identifiable collision risk. The easiest way to identify and quantify a collision risk for a species within a particular area is to use simple temporal estimates of species density overlayed on shipping routes and lanes; known as a “static model”. An increase in species density close to heavily used shipping channels would be given a high collision risk. However, these models do not account for the movement of the whales relative to the ships in that whales may avoid the ship to prevent collision. Further, given there is no inclusion of behavioural response data, it is difficult to say how mitigation measures such as a reduction in vessel speed would reduce the risk collision without making generalised assumptions. “Dynamic” models include information on how whales behave around different types of vessels in terms of their avoidance strategies, which factors dictate the use of these strategies (e.g., a female with a calf may use a different strategy to a group of adults), and which cues they use (e.g., received level of noise, vessel proximity, vessel speed and trajectory). From these dynamic interaction models, the risk of collision can be quantified much more accurately as well as changes in the risk with changes in vessel speed. However, dynamic models require much more information that the basic static model meaning there are few available.

The PhD project will collect behavioural response data from a field site based at Caloundra, on the Sunshine coast. Here, the shipping channel is relatively close to shore, and is located within humpback whale migratory corridor. Ships are moving in and out of Moreton Bay daily. During the humpback migration, ships are moving at speed, and close to, migrating humpback whale groups. This offers an opportunity to collect behavioural response data on the response of groups to fast-moving ships, as well as the factors that contribute to this response such as the vessel’s speed, size, proximity to the group, and received level of noise. These data will be used to generate both static and dynamic models of the risk of collision risk between humpbacks and vessels and compare these models. Once models are created, various mitigation measures will be introduced to the models, such as reduction in vessel speed, and the risk of collision compared. Outcomes will inform the assessment of risk for industry (reputational risk for port authorities), the environment (risk of whales injured or killed, and safety (human injuries and possible fatalities) and develop globally applicable mitigation measures to reduce these risks.

Humpback whale breeding behaviour and sexual signalling

Humpback whales are renowned for their complex acoustic communication repertoire. For example, male humpback whales utilise a wide and varied acoustic communication repertoire whilst undertaking breeding interactions. They use song, which likely functions as a sexual selection signal directed at females, and/or use social sounds, which likely function as female sexual selection signals as well as male-male interaction signals. Song also may be a male-male interaction signal in that eavesdropping males can gain information from the singing male, even if the song is not directed at them.

To find females, males switch tactics between singing and ‘seeking’ (i.e., actively seeking out a female and joining with her, which can lead to fighting with other males (Dunlop and Frere 2023). Their choice of tactic is significantly related to the density of other males within their ‘social circle’. In low male densities, where competition for females is low, males tend to sing. In higher male densities, males will cease to sing and switch to the ‘seeker’ tactic. This is likely because of the balance of costs and benefits of each tactic. If choosing to sing, the male may attract a female, however, the risk is this male may attract other eavesdropping males that can interrupt his song and displace him from the area if alone, or from the female, if with a female (Dunlop and Noad 2016, 2021). In higher male densities, the seeker tactic may be more successful given the increased competition. If quietly seeking out a female rather than advertising using song, there is less risk of attracting an eavesdropping male. However, despite the fact much is known about these breeding behaviours, the information contained within the song, in terms of singer’s fitness, is currently unknown.

Following these studies, the PhD project will determine if there are parameters in the song that are likely to encode the singer’s fitness. It will utilise behavioural datasets of singers and their breeding interactions that have been collected during various field seasons from the late 90’s to mid-2000’s. Song parameters that may signal fitness, such as unit peak frequencies, unit duration, phrase repetition rate, source level, will be compared across different male breeding to test the hypothesis that fitter males are those ones that successfully join a female whilst not attracting male competition, whereas less fit males are those that attracted male competitors. Ultimately, this will improve our understanding of acoustically-mediated breeding behaviour in humpback whales.

There is also the potential to collect more focussed data during this PhD. For example, collecting fitness information on individual singing males, such as body condition using drone photogrammetry and testosterone levels using biopsy samples. This may provide an opportunity to further test specific findings from the song analysis. This will depend on the student’s ability to seek project funding noting that many past students in the lab have had successful grant applications.