Wizards of the Norwegian Coast

SeaBee research infrastructure conducts annual field campaign on the island of Runde, strengthening collaboration and furthering development

By GRID-Arendal

December 16^th, 2022

As August morphed into September, a team of 17 descended onto the island of Runde along the rocky Norwegian west coast. An incredible amount of expertise and knowledge was concentrated in one place, with scientists, drone pilots, technical and communications specialists working together to accomplish a common goal - further developing the SeaBee project's research infrastructure and its novel ways of mapping and monitoring the coastline. It was no wonder they were able to complete all their planned tasks and more in the early autumn sun. Meet the "wizards" of the Norwegian coast!

"One of my personal motivations for SeaBee and this work is that I strongly believe that, with a better understanding and management of the oceans, we’ll have a better life. In fact, I think it’s the only way humanity and our society will make it into the future. We need to understand the ocean better. We need to play 'on team' with nature, and in order to do that, we need to understand how it works and where the nature types are."

"It’s been inspiring to meet up with all the other teams, seeing what they’re doing, collaborating, troubleshooting, talking about how to fine-tune our methodologies, and how to collect the best types of data that in science can provide the best foundation for intelligent and meaningful analysis in the future."

A magical week

Hege Gundersen and Kristina Kvile work at NIVA as researchers. Hege co-leads the SeaBee project with Kasper, and Kristina is the Work Package lead for WP5 - focusing on groundtruthing and field validation for the infrastructure's applications. Both of them spent a significant time in the water during the week in Runde, with the important role of groundtruthing as drones flew and floated above the water.

Kristina played a key role in organising the fieldwork in Runde as WP lead and a newcomer when it comes to the project. She works closely with marine ecosystems, ensuring the best possible choice of location for a successful and enjoyable week!

Here we have access to nice kelp forests and eelgrass beds and a lot of different seaweeds. So we can test the SeaBee approach and we can collaborate together with all the different institutions within SeaBee – from the drone pilots to the ground people collecting data. It’s a good opportunity for collaboration and for making a good data set together!

Hege also works with kelp as much as she can through ecology and ecosystem services, and was happy to be out in the field with the team. She also has experience with modelling of kelp issues, so she understands the importance of collecting good data.

My favourite thing about being in SeaBee is to be able to run this quite big project. At least, assist those that do it. And I work with so many skilled people. Really really skilled people, on different aspects, from the technology to biology, and like now being out in the field, being together, putting our brains together. Having some hypotheses, maybe solve some of them. Creating some new ideas for future work.

Skimming the surface

Both of SeaBee's two Otter Unmanned Surface Vehicles were present for the fieldwork at Runde. One of the Otters is hosted at NTNU in Trondheim, whereas the other sits with NIVA at their Grimstad office.

The Otter that resides with NIVA, nicknamed "RobOtto", is equipped with three different sensors.

On the front, it has a downward-looking single beam echosounder, also known as a habitat mapping echosounder. It's made to measure the depth and also to classify the seafloor type. This sensor can distinguish between harder and softer seabed types, as well as the thickness and density of aquatic plants like seagrass and kelp.

The second sensor measures the water's salinity and temperature along the surface.

Finally, the third sensor is a C3 Submersible Fluorometer with three optical sensors. The first of the optical sensors measures the presence of chlorophyll A, which tells us about the presence of algae in the water. Then there's a turbidity sensor, which measures the amount of particles floating around. The final optical sensor measures what is known as "CDOM" - Color Dissolved Organic Matter. This usually comes in from land sources, such as rivers and streams into the sea.

The Otters are configured so that they follow a route quite accurately, usually in a "lawn-mower" pattern and along the shoreline. The Otters can adapt to the local geography and topography, avoiding obstacles and reaching speeds of up to 6 knots.

"The Otter is kind of an intermediate platform, between the flying drones and the manual on-the-ground verification of what’s there."

Malin and Martin are masters students studying ocean resources at NTNU and are helping to deploy one of the Otters with a hyperspectral imaging camera that faces downwards. They're hoping to get some transects of kelp and seaweed.

All this has to happen without damaging the Otter's sensors, since they hang lower than the Otter's main body frame. That's where Martin comes in - he's all kitted up in diving gear and ready to spend the afternoon swimming alongside the Otter, both to make sure it doesn't get stuck, as well as calibrating the sensors with a reference plate.

Martin is going to be following the Otter with this reference plate and very sneakily place it in front of the Otter as I’m driving it.

Keep it spinning

Medyan is an instrument engineer at NIVA, and his job with SeaBee is to fly drones, keep them airworthy, keep updated with all the training, and make sure that all the sensors are ready to go.

Traditionally, researchers would have to go out and look at what's there, and then make an assessment of habitats based on working out in the field.

"It's not just hard work, we enjoy being out in the field, but often you're only able to physically sample a small area."

The rotor drones Medyan flies can carry various sensors, such as multispectral sensor and very high detail "normal" camera sensors, called RGB sensors. An RGB sensor collects lighting in red, green, and blue, but they're all "blended up". If you take that image and then try to separate those colours, you're never sure which one belongs to which band.

With the multispectral sensors, you actually have 5 or more cameras in one camera, with each camera is taking in a very specific wavelength or band of wavelengths - the light that's reflecting back. Scientists then look for a correlation between what's on the ground and the specific "signatures" of kelp, seagrass, and other aspects they wish to monitor.

"It's nice to have areas of rock because then we can get more accurate positioning. Very often, over more than a few meters of water, it's very difficult to get accurate positioning of the pictures."

Seeing in colour

During the week in Runde, Pål Kvale and Joe Garret from NTNU flew an DJI N600 drone carrying a hyperspectral camera to capture images of vegetation along the coast. Totaling around 15kg including payload and batteries, it's quite a lump!

Joe runs the hyperspectral cameras, in this case a SPECUM AFX10. It has 224 bands or different colors of light and they range from 400 nanometers to about 1000 nanometres. So it can see differences in colour much more finely than our human eyes.

Pål builds, flies, and maintains drones at NTNU. In SeaBee, he has a similar role to Joe, in addition to advising the purchase of drones for the project. For him, one of the big challenges that SeaBee is working on solving is bringing together lots of different drones and technology to complete the task of creating an unmanned infrastructure all along the coast of Norway.

Once we can be sure that what we really see is eelgrass or kelp with a drone, then you don’t need people to spend so much time groundtruthing. You can cover much larger areas!

"I think we’re really gonna get something out of the data this time around. And some nice maps!"

Birds of a feather

While most of the other SeaBee members focus on the shoreline, Sindre Molværsmyr from NINA focuses on novel ways to map and monitor some of its inhabitants, namely seabirds.

"There are a lot of seabirds in the same habitats, and we use a lot of the same methods and drones in the same areas. That's how we fit in."

Today, scientists mainly use small drones flown from small boats out over colonies to help map the entire colony. This enables them to create a map of all next locations. Traditionally, if you wanted to count a bird colony, you'd have to go on land and walk in "transects" much like a drone does and count every single next manually. In large colonies, this can take 4-5 hours to do properly.

“You end up disturbing the entire colony for 5 hours, while with the drone we can often fly over the colony in 30 minutes and not disturb the birds at all.”

During the field campaign in Runde, Sindre went out in the afternoons with a drone to create a 3D model of the cliff face that hosted the gannet colony on the island.

"I have a 3D model now, it's ok, a bit low resolution, and not good enough. We've made a new script and we'll try again."

Like the rest of the fieldwork, it looks like it was a success!

A new addition

On the final day, the team gathered together for one last test: taking off and landing with SeaBee's new DeltaQuad VTOL drone.

The DeltaQuad VTOL drone (Vertical Take Off and Landing drone) is an unmanned fixed-wing vehicle with four propellers that allow it to take off and land in a confined space, such as a small island, a rocky shore or in steep terrain. What sets the DeltaQuad apart from other fixed-wing drones is its ability to combine the pros of a fixed-wing drone (eg. covering larger distances, resilient in suboptimal weather conditions) and those of a rotor drone (smaller space needed to deploy and land safely, greater control).