Saving Hope… – Cetacean Strandings and Satellites

A composite drawing image of Hope the Blue Whale's skeleton swimming between two waters.  Artwork: NaturPhilosophieIt’s unclear why so many great whales beach en masse around the World.  And it’s really difficult to estimate their number, especially when the cetacean strandings occur along remote shores.  But there is hope.  Scientists are now developing techniques to monitor marine populations from space.

Keeping Hope

A black and white photograph showing the stranding of Hope the blue whale in 1891, off the coast of Wexford, United Kingdom.  Image: Natural History Museum, London
The beaching of Hope, off Wexford, Ireland in 1891.  Source: NHM, London

On the morning of 25 March 1891, a blue whale (Balaenoptera musculus) migrating up the east coast of Ireland was caught by low tide and remained stranded on a sandbar just outside the harbour town of Wexford.

Whales can live up to 100 years.

Hope was a young 15-year-old whale, not fully grown.  However, that did not stop locals from taking a great interest in the 25.2-metre leviathan from the deep that had been grounded so close to a British shore.

A photograph of Hope's skeleton above the Hintze Hall, at the Natural History Museum in London,  Image: NaturPhilosophieAfter two pitiful days struggling in the shallows, a lifeboat pilot named Ned Wickham rammed an improvised harpoon under the whale’s flipper, finally putting the poor animal out of its misery.

Her 4.5-tonne skeleton was carefully conserved. 

Hope the blue whale now graces the Hintze Hall of the National History Museum (NHM) in London, as its star exhibit.

A symbol of humanity’s power to shape a sustainable future.

 

On Fatal Shores

The reasons for whales strandings are unclear, but a combination of factors may contribute.  Navigational errors among pods are not uncommon, especially when the whales are chasing food or coming close to shore to escape larger predators, such as orcas.

Around the globe, cetacean strandings happen all year roundThey are a regular occurrence in places like the Falkland Islands, Tasmania, Patagonia and New Zealand.

The old, the sick and injured are particularly vulnerable.  But most events usually involve only one or two individuals.

 

Largest Stranding Events

Two maps pinpointing the location of  the Farewell Spit in New Zealand.  Source: AFPThe coast of New Zealand is well accustomed to the phenomenon.  The Department of Conservation (DOC) responds to about 85 stranding events a year.

According to DOC records, since 1840, over 5,000 whales and dolphins have beached themselves on the shores.

Even then, the largest ever recorded mass stranding events were probably underestimated.

In 1918, 1,000 whales beached themselves on the Chatham Islands.

In 1985, 450 stranded at Great Barrier Island, off the coast of Auckland.

And in 2017…

The Stranding at Farewell Spit, 2017

Satellite image showing the geography of the Farewell Spit, New Zealand. Image: NASA World Wind

The stranding at Farewell Spit is the second largest whale beaching event in recorded history.

On the morning of February 10, 2017, more than 416 pilot whales (Globicephala melas and Globicephala macrorhynchus) were discovered on a crescent of land in Golden Bay on New Zealand’s South Island.

By dawn, over 70% of them had perished.  A further 240 individuals stranded themselves later the next day, bringing the total number of stranded pilot whales to 656.

The reason why such a large number of whales beached en masse is unclear.  Shallow waters can pose a risk to pilot whales in particular because of the way in which they navigate and communicate.

So, how can we begin to save whales like Hope from stranding?

 

Counting Great Whales from Space

The sei whale (Balaenoptera borealis) is a baleen whale, the third-largest rorqual after the blue whale and the fin whale. They migrate annually from cool, subpolar waters in summer to temperate, subtropical waters in winter. Image: World Wildlife Fund (WWF)

Efficient detection of cetacean strandings is a massive challenge.

Most resources tend to be concentrated on ground-based activities associated with carcass sampling and analysis, rather than on the initial detection of whales in peril.  The scale of those events can also be hard to identify, particularly in remote and inaccessible regions.

The new study “Using remote sensing to detect whale strandings in remote areas: The case of sei whales mass mortality in Chilean Patagonia”, published in Plos One by a team of researchers from four Chilean research institutes and the British Antarctic Survey (BAS) illustrates a new technique for analysing satellite images designed to help scientists detect and count whales stranded on shores around the World from space.

It’s not easy to see an object, even one as large as a great whale, from several hundred kilometres up in space, but the international team believes the capability of modern satellites now makes this a practical task.

Being able to detect the strandings more effectively will inform the ongoing conservation of whales. It will also flag out potentially deteriorating ocean conditions, something the fishing industry will also be keen to know about.

 

The Chilean Patagonia Case, 2015

On this map, triangles indicate the location of whale strandings in the 2015 mass stranding event recorded by boat, aerial or ground survey.  The underlying map is a composite of Landsat8 satellite images and shows the complexity of the coastline in the area. The white boxes refer to the extent of the VHR satellite imagery used in this study.  The blue triangles show the locations of the three whales identified using WorldView 3 imagery in 2017. Source: Plos One  Image: USGS

In 2015, the carcasses of 343 sei whales were spotted on remote beaches in Patagonia, Chile.

However, the survey was conducted from planes and boats.  This work was also carried out many weeks after the cetacean deaths actually occurred.

An analysis of high-resolution satellite images of the area taken much closer in time to the stranding has now identified many more bodies.

One reason for the uncertainty is that researchers arrived very late on the scene to run the tests that could establish the cause.  Part of it was because the stranding occurred in such a thinly populated area of central Patagonia called Golfo de Penas (or “Gulf of Sorrows”).

The Gulf of Penas is difficult to access.  It has multiple fjords, channels and islands, and the deaths only came to light by accident when an unrelated expedition chanced on the carcasses many weeks after the event happened.

Scientists in protective suits are attending to the decomposing body of a stranded sei whale. The carcass has turned a orange/pink colour.
The carcass of a decomposing sei whale. Image: British Antarctic Survey

By then, the sei whales had already started to decompose.

Nonetheless, a team on the ground were able to conclude that the cetaceans had probably been poisoned by toxic algae.

 

Achieving Better Estimates

The Fretwell et al (2019) investigation was undertaken as a proof of principle exercise by the British Antarctic Survey (BAS) and various Chilean organisations.

WorldView-2 is a commercial Earth observation satellite, launched on 8 October 2009.  Image: DigitalGlobe

The researchers used pictures from the WorldView-2 spacecraft which can discern features larger than 50 cm across from an altitude of 700 km.

Studying satellite images covering thousands of kilometres of coastline, the researchers were able to identify the shape and size of the whales.

For example, a whale that may be 10-15 metre in length would produce a good outline of the animal’s overall shape, including its distinctive fluke.

Photographic evidence from the boat surveys showed many dead whales had partially decomposed and had changed colour. 

WorldView-2 provides panchromatic imagery of 0.46 metre (18 in) resolution, and eight-band multi-spectral imagery with 1.84 m (72 in) resolution.  Diagram: SatImagingCorp.com

As decomposition set in, the carcasses of the animals turned a pink and orange hue (in wavelengths visible to the human eye).

So it was also possible to detect their colour from space.

 

In the Gulf of Sorrows

Considering the case of sei whales mass mortality in Chile, it was very difficult to give a precise total for the number of whales involved.

Although planes and boats had surveyed the Gulf of Penas extensively and counted over 340 dead whales, the complex local geography meant that some bodies had almost certainly been missed.

High-resolution satellite imagery allowed scientists to do a count much closer in time to the event itself.

 

Seeing Back in Time…

At the top are whales from Image 1 and the bottom are from Image 2. The upper row displays detailed pan-sharpened images (spatial resolution 0.5 m) shown using visible bands of the satellite imagery. The second row is the multispectral band (2 m resolution pixels) used in the spectral analysis with the area of pixels chosen for each profile highlighted in red. The third row are the spectral profiles for each image showing Top-Of-Atmosphere radiance in the Y axis and wavelength in nanometers in the X axis. Image: DigitalGlobe, MAXAR Technologies (2019)

The team examined two archival WorldView-2 images of the gulf of Penas, dating back to March and April 2015.

“The aerial survey was done on a huge scale and was very impressive, but it’s possible some of the carcasses got washed back out to sea in storms and simply weren’t counted.

The 343 number was only ever a best estimate.”

Dr Jennifer Jackson

BAS Whale Expert

In one picture, the analysis recorded slightly fewer whales than in the aerial survey work.

But in the second picture examined by the researchers, the result was substantially more.

Actually, the count was nearly double!

 

Number of Whales in Satellite Survey compared to Ground Survey (from boat and air) in the same areas

Ground SurveySatellite SurveyClass 1Class 2Class 3
Image #130221128
Image #214231436

Source: Fretwell et al., 2019, Using remote sensing to detect whale strandings in remote areas: The case of sei whales mass mortality in Chilean Patagonia, Plos One.

 

The number of whales counted in the second image, of Byron and Wager Island, was considerably more than in the aerial survey of the same area.  Twenty-three potential whale carcasses were manually counted in the satellite image, 14 with high confidence (class 1), three with moderate confidence (class 2) and six with low confidence (class 3).

Overall, the objects on Image 2 were more clearly whale-like than on Image 1, probably due to the lack of crowding evident in Escondido Sound.

Of the 23, seven were within 100 m of the position of a whale recorded in the aerial survey, whilst a further six were within 300 m, and two others were within a kilometer.  If it is assumed that the remaining 14 whales recorded in the aerial survey correspond to the nearest whale in the satellite image, this leaves nine extra whale-like objects on the satellite image that were not recorded in the aerial survey.

Of these, three had a low confidence ranking (class 3), whilst five had a high (class 1) ranking and one a moderate (class 2).

In areas of thin cloud, whale-like features could generally still be detected with some confidence.

Fretwell et al. 2019

 

How Useful is Satellite Observation?

Not only is it possible to see the big baleen whales, like the sei whale, from orbit, but with the WorldView satellites constellation now offering 30 cm resolution, the task should become progressively easier, even when it comes to smaller whale species.

All this detail could facilitate a more rapid response.

Involving Automation

Of course, it would be even easier if an automated detection system could be developed.

The team tried this by training a computer to look for the spectral (light) signature of a dead whale.

In the Gulf of Penas, the sei whales turned pink and orange as they decomposed…

As it turned out, however, this approach proved less successful than manual inspection of the pictures.

“There are many more satellites planned to be launched with 50cm and 30cm resolution, so if we could automate the system it might be able to find these stranding events almost as they happen.”

Peter Fretwell,

BAS Remote Sensing Specialist

Watching over Hope…

Sei whales have continued to wash up in the Gulf of Penas every year since 2015.

Examples of stranded whales in WorldView3 imagery from the Gulf of Penas in March 2017, showing the improved shape discrimination using this higher resolution data. Image: Fretwell et al. 2019/MAXAR Technologies

Getting to the scene of a stranding early will provide even greater certainty regarding the cause of an event. 

Cetacean strandings can be useful markers of the status of a population.  Albeit gruesome, the dissection of washed-up carcasses provides marine biologists with an opportunity to investigate the general health of the animals, and study aspects of their behaviour and dietary habits.

Even the pattern of strandings provides information on which whales are present in an area and their likely numbers.

With technology, we could prevent cetacean strandings… Image: British Antarctic Survey (BAS)

Therefore, it is vital for marine biologists to understand fully what is happening off-shore.

Algorithms are only bound to improve.

Scientists could easily monitor any beach in the World, especially the most remote coastlines.

“The technology is getting better all the time.  In this study, we were using 50 cm resolution images, but the satellites now can see 30 cm.  In the future, we’d like also to be able to analyse the pictures automatically, rather than manually; and I’m sure as more minds are applied to the problem, this will become possible.”

Dr Carlos Olavarría

Centre for Advanced Studies in Arid Zones (CEAZA), La Serena, Chile.

 

For scientists, the monitoring of whales from orbit is set to become a powerful tool with which to assess the state of our oceans.

“Whales are often top predators and they are very involved in the marine ecosystem.  If they are suffering because of any actions of humans then we need to monitor that and mitigate it if possible.  It’s important ecologically.”

Andrew Baillie,

Cetacean Strandings Officer at London’s NHM

 

In time, we must hope that with increasingly sophisticated use of technology the “Gulf of Sorrows” can become the “Gulf of Hope”…