A new player has joined the ocean science community in Sweden: VOTO – Voice Of The Ocean. SCOOT is of course delighted that VOTO is a strong believer in automated methods of observations. VOTO has already a fleet of six ocean robots: 4 diving SeaExplorers by Alsemar and 2 surface going Sailbuoys by Offshore Sensing. SCOOT’s director Seb Swart and his team joined VOTO’s SeaExplorer course, read their report!
In March the Killer Shrimp Invasion Challenge was opened on Kaggle by Ocean Data Factory Sweden (coordinated by SCOOT): Use innovative machine learning (ML) solutions to predict the spread of the so-called “Killer Shrimp” (Dikerogammarus Villosus) into the Baltic Sea. The challenge is now closed , the referees have completed the evaluation and announced that the winner is…
Dimitriy, research engineer at IBRAE RAS in Moscow, Russia! Dimitriy nicked the winner’s place with a score of 0.99954 (out of a maximum possible 1) and by complying to all rules. Dimitry will be presenting his model in an open webinar on June 16th, see below:
The Benefits and Challenges of Open Innovation and Citizen Science for Solving Ocean Challenges through Applying ML to Ocean Data
Join us on June 16th for a Zoom discussion (Please register here):
- 15:30 to 15:40 Introduction to Ocean Data Factory Sweden and afternoon discussion
- 15:40 to 16:10 The Killer Shrimp Invasion Challenge on Kaggle: An online competition tackling the spread of invasive marine species through machine learning
- 16:10 to 16.20 Break
- 16.20 to 16.50 The Koster Seafloor Observatory on Zooniverse (still open to join!): Using citizen science and machine learning to annotate subsea imagery
- 16.50 to 17.00 Concluding remarks
The image above shows our autonomous surface vessel SB Kringla being retrieved after the first of two magnificent field seasons. Now she will be upgraded with both an acoustic sensor for measuring currents and an alternative “high wind” sail.
The new sensor will be a Doppler Current Profiler by Aanderaa AS, measuring from the surface down to 30 -50 meters (depending on conditions). SCOOT is proud to be an early adopter of this sensor on an autonomous platform. Measuring the currents in the ocean surface layer can for example tell interesting things about air-sea interaction and the transfer of wind energy to the ocean, especially together with surface wind data, which SB Kringla is already measuring. The new sail is slightly smaller and more optimised for the high wind conditions in the Southern Ocean. Sails are easily switched between deployments. The upgrade (sensor and sail) will be done by Sailbuoy’s manufacturer, Offshore Sensing AS in Norway.
The Doppler Current Profiler will be used, among other applications, to further enhance our understanding of the role of vertical shear mixing in the S Ocean, which was hypothesised in a recent publication in Geophysical Research Letters, using data from SB Kringla: “Submesoscale Fronts in the Antarctic Marginal Ice Zone and Their Response to Wind Forcing”.
SB Kringla is the first Sailbuoy to be deployed on the Southern hemisphere. SCOOT’s director Sebastiaan Swart often refers to her as “the cute one”, which does not imply that she is spared in any way! Kringla has repeatedly earned battle scars and proven herself as an ultra robust platform in the harshest conditions of the Southern Ocean (see posts from first season and second season). Data from SB Kringla (and the diving Seaglider robots) can be viewed at the website of the project Robotic Observations and Modelling of the Marginal Ice Zone.
The upper video shows a snapshot of SB Kringla during first operator training in Gothenburg. The lower video shows a Sailbuoy battling a storm on the North Sea.
Louise Biddle summarizes the Austral winter-spring season for this robotics heavy project:
The ROAM-MIZ (Robotic Observations And Modelling in the Marginal Ice Zone) project aims to capture high resolution observations of how the upper ocean near and under sea ice responds to sea ice growth or melt. The team, led by SCOOT’s director Sebastiaan Swart at the University of Gothenburg, deployed multiple platforms in, and at the edge of, the marginal ice zone in austral winter and spring 2019. Using the South African icebreaker, the SA Agulhas II, as a deployment platform, three profiling ocean gliders (Seagliders), a surface glider (Offshore Sensing Sailbuoy), multiple wave-recording surface floats (SWIFT buoys) and two ice-moored buoys were deployed between 56-60°S. Between the 10 platforms, over 370 days of data were collected from this remote and under-sampled region of the Southern Ocean, with the longest continuous time-series of over 130 days.
The challenges that the deployment team and platforms faced included rapid freezing up of sensors in temperatures that dropped to minus 20°C (before wind chill), rough seas and rapid movement of sea ice floes. With support from collaborators around the world including CSIR (South Africa), CalTech (USA) and University of East Anglia (UK), this project has demonstrated that autonomous instrumentation holds the capability to explore and monitor some of the most inaccessible oceans on Earth. Data is now being processed by the team based at the University of Gothenburg, with the “live data” viewable on roammiz.com and the first paper recently published in Geophysical Research Letters. Projects update can be followed on @PolarGliders .
Paper: Swart, S., du Plessis, M. D., Thompson, A. F., Biddle, L. C., Giddy, I., Linders, T., et al. ( 2020). Submesoscale fronts in the
Antarctic marginal ice zone and their response to wind forcing. Geophysical Research Letters, 47, e2019GL086649.
Ocean Data Factory Sweden (coordinated by SCOOT) releases an open challenge: Use innovative machine learning (ML) solutions to predict the spread of the so-called “Killer Shrimp” (Dikerogammarus Villosus) into the Baltic Sea. Join the Killer Shrimp Invasion Challenge online on Kaggle!
- a data scientist interested in applying your knowledge to environmental challenges?
- a marine scientist interested in using Machine Learning in your work?
- generally passionate about the ocean and keen to learn more about Machine Learning and Marine Science?
Then this is the right challenge for you! The winner of this competition will be awarded a €150 prize, along with the opportunity to present their winning model at the next ODF Sweden Grand Meeting. The event will be in the beautiful city of Gothenburg in June 2020 (date tba).
About The Koster seafloor observatory
At the Koster seafloor observatory we want to know how climate change and human activities influence Sweden’s marine ecosystems.
For the last 20 years, we have used Remotely Operated Vehicles and cameras to monitor the Kosterhavets National Park (click images above for example videos!). A highly-diverse and unique marine reserve in the south of Sweden. Now, we need your help identifying the habitat, species, and presence of litter recorded in over 3,000 hours of underwater videos.
Your answers will allow us to filter out crucial information from the recordings and study how the fauna on the sea floor has changed in response to warmer waters, fishing activities, and changes in environmental protection.
Submerge yourself among large sponges, starfish, and cold-water corals to help us understand the health of the Koster’s sea floor and how we can better protect it.
Words from project owner Victor Valdes
“We are excited to bring you a new way to dive into Sweden’s first marine national park. By classifying the underwater videos you will explore our unique marine biodiversity and enable us to better understand the health of Sweden’s marine ecosystems.”
Ocean Data Factory Sweden (coordinated by SCOOT) will support the AI/UX Hackathon about the Baltic Sea Guillemots (“Sillgrisslor”), taking place on November 21-22 in Gothenburg. ODF Sweden will be in the Hackathon jury and our partner SMHI will deliver local oceanographic data.
The Guillemots are fascinating seabirds high up in the food chain. They can tell us a lot about the state of the fish and the rest of the Baltic Sea ecosystem. Gathering and annotating data is a heavy task for the scientists of the Baltic Seabird Project. In last year’s trial, huge leaps were made with the help of citizen science. The live stream hosted by WWF produced some 1300+ comments and insights. With the use of AI, recognition of individuals and frequent events could be automated, leaving the extra-ordinary and rarer observations to researchers and citizen science.
Register and more info here. Deadline for registration is November 7.
The Saildrone is one of the most advanced autonomous ocean surface platforms. SCOOT is very pleased to announce that they are visiting us in Gothenburg. They will give an open presentation on October 1st. Welcome!
DATE: Tuesday 01 October 2019, 12:15 pm
LOCATION: Carl Skottsbergs Gata 22B
SPEAKERS: Sebastien de Halleux, COO, Liz Douglas, International Partnerships
Working with a variety of institutions and governments globally, Saildrone designs and manufactures wind and solar powered unmanned surface vehicles (USVs) which make cost-effective ocean data collection possible at scale and in near real-time. Saildrone USVs carry a full suite of ocean and meteorological sensors, with the option to include other instruments for research such as air-sea fluxes, biomass, bathymetry, and carbon and . Saildrones have been successfully deployed for months in challenging regions of the global oceans.
In this workshop, Sebastien and Liz will detail Saildrone’s capabilities and report on recent missions in the Southern Ocean, the North Sea, the Atlantic, and upcoming missions in the Mediterranean. Typical investigations in these regions include measuring air-sea fluxes and how they may be impacting global energy, hydrological, and carbon cycles. Sebastien and Liz will give a live demo of the Mission Portal, followed by an open discussion regarding in-situ observation needs and challenges from participants.
ODF Sweden (coordinated by SCOOT) is now gaining pace! First innovation cycle, lasting to end of 2019, will focus environmental monitoring and management. We will deliver learning and support for digitalization in marine environmental management to the Swedish Agency for Marine and Water Management (SwAM). The collaboration between ODF Sweden and SwAM is a part of the ”Three Archipelagos” project. We are excited about this opportunity.
Do you want to learn and test what/where/how autonomous platforms can measure in the ocean. Do you already have experience and want to meet other experts? This is the workshop for you! More info about available resources and schedule here.
SCOOT’s fleet of autonomous vessels and robots will be available, together with support vessels, technicians and experts. Specifically this includes SCOOT’s flagrobot, the AUV Ran. SCOOT is proud of the coverage Ran’s achievements has got in media, e.g. in Roling Stone, Public Radio International and Swedish public service television.
Our host for this workshop is Institute of Marine Research, belonging to Swedish University of Agricultural Sciences (SLU).
Please contact us if you interested in participating!