The Dead Do Tell Tales

Figure 1: Discussing strategy for optimising data collection during the 2023 pilot whale mass stranding event on the Isle of Lewis, Scotland. Photo credits to SMASS.

Student Spotlight - Rachel Lennon

How can we use post mortem reports of stranded cetaceans to better understand their health?

Whales, dolphins, and porpoises, collectively referred to as cetaceans, are charismatic ocean megafauna that often evoke powerful feelings of connection in most people. These beautiful species have complex social structures that are reflective of our own, making them easy to relate to and impossible to ignore. We have an innate desire to know how cetaceans are faring in our waters and a strong drive to protect them. This is perhaps most evident in the “Save the Whale” movement of the 80’s that was key to instating the IWC moratorium on whaling in 1982.

Being top predators, cetaceans are also environmental sentinels with changes in them reflecting changes lower down the trophic level. Therefore, it is often through public interest that these animals are monitored, informing on the health of entire ocean ecosystems. It is also something that we must do, with key legislation like the Wildlife and Countryside Act of 1981 and the Conservation of Habitats and Species Regulations of 2017, requiring that we monitor these species, understand the threats against them, and try to minimise them as best we can.

Our seas are becoming ever more threatening places to these species, with increasing underwater noise, bycatch, plastic and chemical pollution, and reduction of resources from interactions with fishing fleets. To mitigate the effects of these anthropogenic stressors, we first need to understand what their impact is and how they affect cetacean health.

A key component to consider when thinking about the impact of anthropogenic stressors is that they never work alone. Often these stressors are acting in tandem with synergistic or antagonistic relationships, resulting in cumulative effects. For wild animals like cetaceans, they are already experiencing environmental stressors from reproductive costs, competition within and between species, battling diseases and parasites, and from simply keeping themselves warm and fed. All these life tasks have energetic costs on an individual, costs that they are well equipped to deal with.

Figure 2: A bottlenose dolphin attacking a harbour porpoise, illustrating antagonistic interactions between species. Photo credits to SMASS

However, when we start to add noise pollution to the environment that disrupts a whale’s ability to catch fish, when entanglement in ghost gear can cause increased drag to the point that swimming becomes impossible, or when high chemical contaminants in the water reduce the immune response to an infection, individuals can start to lose the ongoing energetic battle. If this is what cetaceans are experiencing in the ocean then it is important to monitor stressors concurrently to ensure accurate reflections of the environment. A key way of doing this is to determine the effects that stressors are having on an individuals health, and thus determine what energetic toll they are having.

Establishing health presents a challenge for cetaceans, since you often only catch a glimpse of a passing fin when trying to monitor them. Instead, opportunistically obtained stranding data offers a unique opportunity to assess mortality trends in cetaceans and to determine health. Schemes like the Scottish Marine Animal Stranding Scheme (SMASS) are long-term networks collecting data on stranded cetaceans and carrying out post-mortems and diagnostic tests. The plethora of data from such schemes offers a valuable source for assessing the health of these animals. Current assessments of health in cetaceans use morphometrically assessed body condition (think BMI for whales) providing a snapshot of body size. This is known to naturally fluctuate throughout an individual’s life history and vary depending on species.

Figure 3: SMASS team carrying out a post-mortem on two common dolphins, collecting important data used to assess health. Photo credits to SMASS.

Development of a pathology driven health indices can provide an alternative, multidimensional approach that would accurately capture changes to the physiology of an individual. However, due to a lack of homogenisation of post mortem data, developing a health index that is robust and universally applicable presents a challenge. My research aims to develop a multidimensional health index derived from pathology and additional ancillary datasets of stranded cetaceans. This would provide a value of health that better reflects all stressors that an animal is dealing with, including its parasite burden, if it had any diseases, and how its organs were functioning. Having a full view of an individual’s health allows for us to quantify the interacting effects of all those ever-increasing stressors.

Figure 4: Author collecting samples from a pilot whale to be used in developing a health index. Photo credits to George Morris.

The first stage of this research is to develop a framework that extracts prose from post mortem reports and turns them into codable features. The framework is being developed in collaboration with experienced marine mammal pathologists to ensure that all relevant elements of pathology are captured in the framework and that the margins of each feature are appropriate. For instance, presence of a lesion on the skin could be coded as a binary presence/absence feature whereas parasite burden would need more detail and may be ordered into low, medium, and high. To ensure applicability across different stranding schemes, the developed framework is being validated with a systematic questionnaire that will be sent to pathologists across Europe. Thereby ensuring that the developed code is objective and most accurately captures relevant features. The validated framework will then be used on a test batch of 1000 post mortem reports to generate a pathology dataset.

Figure 5: Stepwise approach to developing codable framework of pathology features.

Using unsupervised machine learning, the coded features from the new dataset will be condensed into their principal components. This will highlight key features of significance, allowing pathologists to prioritise target regions during resource limited post mortems. The principal component analysis (PCA) will also generate eigenvectors, a set of values that can be extracted and used as our indicators of health.

Once our new health index has been established, we can compare this to commonly used proxies (like BMI) to ensure that the new index captures health more cohesively. Once validated, the health index can be a response against which the impacts of multiple stressors can be measured, allowing for quantification of how stressors are impacting cetacean health.

Figure 6: Stepwise approach to generating health index from coded pathology.

With a robust, multidimensional health assessment we can begin to understand how stressors effect cetaceans in real time. Quantifying the impact of stressors is instrumental in evidencing policy that will enforce reductions of such stressors. Subsequently, it will also allow for more accurate monitoring of the success or any implemented mitigation. In an increasingly busy ocean, it is vitally important that we can appropriately measure the effects that human activities are having on its wildlife, so that we can ultimately ensure their protection.

Rachel is a student of University of Glasgow.

Email: r.lennon.3@research.gla.ac.uk

Twitter handle: @rach_lenn.