top of page
Writer's pictureMalcolm Kennedy

From sea to space: ways to monitor whales


As human impacts on whales become an increasingly critical issue, monitoring whales’ distributions and numbers are bound to be a key part of an effective response. The data gained through monitoring can help direct policy, as well as corporate decision-making, toward less harmful practices.


As monitoring efforts scale up, it is important to not exacerbate disturbances on whale populations, by using non-invasive methods as much as possible, while staying cost-effective. Some techniques provide useful data but can add stress to whales, especially at scale. For instance, satellite tags affixed using airguns to blue whales in the Gulf of St Lawrence, Quebec and off the coast of California have yielded accurate real-time location data for tagged individuals. However, the tagging process is costly, limited in scale, and stressful for the whales being tagged since it involves tailgating the whales in a motorboat. Similarly, sonar has proven itself a useful tool for gathering marine data — but for whales, it is disruptive at best, lethal at worst. To avoid introducing unnecessary risks or stressors to whales, Whale Seeker is committed to focusing on remote sensing techniques, that is, techniques that detect whales at a distance, rather than by physically interacting with them.


Passive acoustic data, which is recorded using hydrophones (underwater microphones) has proven one important channel for remote monitoring of whales. Whales rely heavily on sound for communication, and the vocalizations they produce can be picked up by hydrophones, transmitted onshore, and processed to determine not only the presence of whales, but their species, and approximate location. As is often the case, our ability to gather acoustic data has superseded our ability to put it to use. In the Canadian Atlantic, for instance, most hydrophones are privately owned, and little of what is stored is ever retrieved, let alone processed. However, there have been some promising initiatives to apply machine learning to make sense of underwater soundscapes. A study published earlier this year showed that the endangered North Atlantic Right Whale (NARW) could be reliably identified from acoustic data alone. This came in the wake of a similar application for orca calls last year.


Acoustic data can be gathered all day and night, regardless of weather conditions, and has advanced our understanding of whale distributions and behaviours tremendously. However, the acoustic medium does not lend itself well to two key goals of many conservation and impact mitigation efforts. First is estimating the number of individuals in a given area. Second is deployments for real-time ship-borne whale monitoring, since most ships produce enough background noise to render any signal recorded nearby unintelligible. For these two applications, visual data has proven more useful.


Visual data gathered from aboard vessels or aircraft allow for detailed observation not only of species identity, but of group size and behaviour. Aircraft surveys typically fly along lines or “transects” that are arranged so as to avoid re-counting the same whales multiple times. However, traditional aircraft surveys are expensive, weather-sensitive, and expose crews to potential deadly dangers. Unmanned aerial vehicles (UAVs) provide a promising and less dangerous source for image acquisition but are not yet straightforwardly more cost-effective than traditional aircraft surveys.

In recent years, the decreasing cost of satellite imagery has provided a new and promising data source. Satellites represent a truly non-invasive monitoring technique and can yield data of increasingly high resolution. One drawback of satellite imagery is that clear images depend on minimal cloud cover, and whereas aircraft can fly below the cloud cover if it is high enough, this option is not available for satellites.


As with acoustic data, as image data becomes more accessible, our ability to process these images has struggled to keep up. Human annotation of aerial imagery is costly and requires significant expertise to yield consistent results. Whale Seeker is helping to tackle this discrepancy by automating the detection process so that our data processing power can keep up with the new and exciting developments in data acquisition.


In addition to acoustic and visual data, a number of modalities beyond the human senses have been deployed for whale detection. These include thermal cameras, which have been used in California to survey a portion of the grey whale migration route. The cameras are coupled with an algorithm that automatically detects potential whale spouts, which are then verified by humans. This has allowed for more accurate population estimates than ever before. Thermal cameras have also been deployed to monitor the endangered orca population in BC, Canada.


Whale Seeker is excited about the range of new monitoring solutions, and committed to using and combining them in a way that is cost-effective, results-driven and non-invasive to whales.

bottom of page