What is AquaCloudSustainable growth
Fish Welfare
Data plattforms
Innovation in the maritim sector
Del denne Casen
Hello. My name is Nina Stangeland and I'm the managing director of NCE Seafood Innovation. We are a business cluster with the different stakeholders in the seafood industry, where we work together to contribute to the sustainable growth and development of the seafood industry, by focusing on innovation. In this podcast series of five together with Lørn.Tech, NCE Seafood Innovation, and AquaCloud, we try to share our knowledge and experience on the digitalization of the seafood industry. AquaCloud is a seafood industry project where we gather different farmers in the Norwegian aquaculture industry to collect data from different locations, with the aim to contribute to more insights about the Norwegian aquaculture industry, and we hope to contribute to innovation, butterfish health, and fish welfare in the Norwegian aquaculture, and also contribute to sustainable growth and development of the industry. Enjoy your learning and enjoy our experience in elaborating this project in this podcast series.
Silvija Seres: Hello, and welcome to a Lørn podcast in collaboration with NCE Seafood Innovation. My name is Silvija Seres and my guests today are Trond Kathenes and Jørn Torsvik.
Trond Kathenes: Thank you.
Jørn Torsvik: Hello.
Silvija: Both your names sound very easy in Norwegian, but then I got a complete tongue twister when I was trying to say them in English. I'm going to say just a few words about the series and then I'm going to ask you to introduce yourself and your goals with your project. So, Lørn is doing a series with the Norwegian cluster called NCE Seafood Innovation, which is a national cluster for companies that collaborate in aquaculture, which is salmon farming, and it's Norway's second-largest export product or industry with 1.1 million tons of salmon, being exported every year. You guys are handling values and growth just as important, and perhaps, more environmentally friendly than our largest industry, which is energy, and oil. So, we are going to talk about digitalization in your industry with a specific focus on a product called AquaCloud which is a data platform related to the production data in your industry. This is the first out of six conversations related to this platform. With that, I'm hoping that you two could tell us just a few words about who you are and what makes you unique. Maybe we start with Trond.
Trond: Yes, I'm Trond Kathenes and I'm the Chief Digital Officer in Grieg Seafood and then obviously responsible for the digital initiatives and strategies that we are running the company. One of the main activities we started with around the digitalization was the AquaCloud initiatives, ensuring that several companies gathered or collected the data, sharing the data into AquaCloud for, first of all, the goal of actually being able to predict the lice exposure towards our operational units as such. So, that's a quick runner, but Jørn you can now go ahead with who you are.
Jørn: Yes, I'm Jørn Torsvik, I'm the general manager of AquaCloud, which is this initiative we're going to be talking about. I ended up in aquaculture a couple of years ago after working many years in financial services, and also in digitalization. It was really interesting to join this industry which is so completely different from banking where it's just zeros and ones and everything is virtual, and where there's a lot of physical equipment, there are living creatures in these cases, in the sea, that you need to take care of. The risks and considerations are completely different.
Silvija: I think that you said a little bit more about who you are than Trond said. You were a Navy officer before.
Trond: Yes, that's correct. So kind of a strange background, being educated under the Naval Academy, I had a military career as commanding officer on Coastal Corvettes and so on. And then, at a certain point in my career, I started purchasing projects, and then, it's involved into the IT industry and how to utilize data to get more out of the tools that we gain. That's where my interest in technology really took off. From there on, I jumped out of the Navy and IBM captured me and different numbers of big companies within the IT management area. Then, with some crossovers and going back and forth, I ended up in this industry 10 years ago, and it has been a tremendous journey, actually, both in the company I work in, but also in the industry as such, being a part of the development and the early stages of digitization. Now there is a growing maturity to interact well into and we contingently open up for colleagues out to the sites and so on. So, it's really exciting and enjoyable.
Silvija: It's a very important industry for Norway, and perhaps, also one of the industries where Norway has a unique advantage in the world, both geographically, climate-wise, environmentally, and in other ways historically and culturally. So, it's nice to be working with something that we know exceptionally well. I'm just curious because you both come from different Industries. You've ended up in this cross-section of fisheries, or in this case, aquaculture, which I've just learned from you before we started the podcast, is a subset of the fisheries industry which allows you to explain, but you also focus very much on digitalization, and I would like you to tell us a little bit about the history of the project AquaCloud. If we start a little bit with defining the industry, and then tell us about what's this project is about. Maybe we start with Trond again.
Trond: If we're going from the top as you indicate, it's the fishery, this is the part of, obviously, what we get out of the sea, but officially is divided into the traditional fishery, and then there is the aquaculture, which is breeding Salmon control environments at sea and even also partly on the shore. The industry as such has come very far away when it comes to being efficient in production and so on, but still, to be able and to be allowed to grow we need to get control over some of our biggest challenges when it comes to the amount of environmental impact, the sea, lice exposure, and so on. Then a number of the players in the industry gathered to see how can we now utilize technology and data acquisition to get a better understanding of what's happening in the water column, and how can we collect data and share data to see how we can utilize and find out based on the big data sets that there is something in the water column that can help us to predict how exposed the fish is at any given time to lice or other species or organisms that is impacting the fish health and welfare indication. That was the very start of the project, that they were going to focus on the lice issue as such. The gathering of all of these data also opens up for more opportunities to find even more, and that's where we have gained in different companies today also, gaining experience from the AquaCloud, and then take some of the knowledge back to our own companies and working even more into deep analysis, on our data set as such. What we have seen now is just the beginning of our industry to work closer together with Offspring for AquaCloud, and how we sharing data in a totally new way to learn as much as possible of what's going on down in the water column.
Silvija: Water column analysis I think is a really interesting image, and I'm going to let you Jørn follow up on his comment. I just want to give you a kind of personal side note, we did a series here in Lørn with innovators in Bodø, and one of the parties that we spoke with is a family-owned fish farmer, and I remember their description, they export most of their products to a particular customer in America who buys everything, and their conversation with the board of this customer, and their confusion and shock when they saw the level of infrastructure that we have in Norway: 5D and internet of things and full broadband far out in the sea, and it's really interesting for us all to learn what can this be used for to make production efficiencies, but also, as you said sustainable growth. I'm looking forward to hearing more examples when we get there. Where do you come into this story, Jørn?
Jørn: That's a good question. All the years I've been in aquaculture I have been working with digitalization. First, I worked with one of the large equipment vendors in the industry, and now I work for AquaCloud as well. At this point is essentially really exciting to be in the industry because since the last two or three years the digitalization rate and the digital maturity of the industry have gained speed and traction. When these visionary companies and others came together five years ago to Asia to initiate the AquaCloud project, that was quite a big leap to take because Big Data was the really hot bus that year, and coming together to collaborate and share the data was something, at this level, completely new. Of course, there have been initiatives before, there is governmental reporting and those kinds of things, but coming to bed together voluntarily as companies to share the data openly to solve common challenges in this sustainability area, it's an exciting initiative. We will follow up on that. We moved on from the lice prediction algorithms, but still, the core data platform is hard, and we also have different other initiatives to support that data platform in terms of data quality.
Silvija: I have to admit I used to joke a little bit about some of the projects related to Artificial Intelligence applied to fisheries or aquaculture. I remember a couple of years ago there was an award to the most Innovative organization or startup in Trondheim, and I think the prize was given to a company that did some sort of very smart image recognition and could recognize individual salmons, measure their size, make conclusions on whether the size is legal either for fishing or whatever it was. I know of many other projects where people optimize the amount of food being given to the fish based on the dynamics of the fish, and they can also find out whether the fish is stressed or not and relate the psychological health of the fish to the environmental conditions that they thrive in. So, in some ways, I'm wondering if we've been better at digitalizing fish health than digitalizing human health in this country.
Trond: Well, we have come a long way with the digitization of the fish. I'm not that sure that we are ahead of human health, somebody else needs to answer that one, but I can tell you that all of the tools or the mechanisms that you are listing here: the image recognition, the analysis of the fish behavior, pallet recognition when you're feeding the fish, and all these aspects, are things that are already put into use in our production at Grieg Seafood for instance, and other older players in the industry as well. So yes, we are taking advantage of all these kinds of startups and inviting them in to help us to kind of gain results, and also, gain more maturity in the organization, to get the organization to understand the benefit and the possibilities of taking these kinds of tools and technology into use.
Silvija: I just want to understand a little bit better the basic problem. So, you talked about lice infections and I think most of us just can't imagine really what's going in a fish farm. I see a circle on the surface and I don't know how far down I should think, how big, I don't know what's the process of dealing with that kind of an animal, and I understand also that given the pressure of having many fishes together, there is a bigger exposure to disease, and then we need to make sure that we treat fishes just as kindly as we treat animals on land, so we're trying to solve this problem, but please, teach me a little bit about that. How does a fish farm work? If we start there.
Trond: Yes, I'll take this one then Jørn.
Jørn: Okay!
Trond: Well, first of all, the fish farm is obviously what you see on the surface, is the rings, and then it's all about the underwater condition, the typography beneath the sea level, that actually gives the size and depth of the fishes to the farmers and such, so it can be between 50 or 75 meters deep, and it could be up to 160-200 meters of diameter, but the most vital thing is that you're only allowed to have 2.5% of the biomass, or fish, in the total volume inside the cage. So, it's 97,5% of water. That's the maximum of fish that it can have, and this is something that we are working very hard to keep track of. Over the years there has been a limited possibility to understand the whole complexity of the water column in the pen, and that's where the new technology comes in when a new data collection comes in, and also together with the scientists and the different organizations in research and so on, we are looking into what is the current condition in that specific pen, and what does the different micro-mechanisms that is in the tongue, how lives that influence, the fish, and the fish welfare? A happy fish is a good fish, right? Can we foresee the potential negative exposure by seen changes in the conditions in the water column, then we might be able to put in some precautions and foresee that the fish still has high welfare as it is. So, it's all about understanding the details in the water column.
Silvija: I'm just going to try to regurgitate this a little bit just to make sure that I understood it. So, these water columns are going 50 to 75 meters below, they can be 160 to 200 meters in diameter, and the rule is that you can't have more than a 2,5% of fish within one column like that. There are some other variables I'm not quite sure about, there is a mesh of some sort that allows more or less water to pass. How much is the dynamics of the external conditions and the internal conditions in a column like this?
Trond: It depends on the surroundings, the sea current, and the specific position that you are allowed to put your forearm into, and then, obviously, there is a mosque that, as the fish grow, the pen size goes up, so that you had enough throughput of water through the pen, so the fish can get all the nutrition it gets from the surroundings.
Silvija: Can I ask you one more dummy question? Do you move the fish from one pen to the other when they reach a certain size, or do they live their whole life within one pen?
Trond: Normally they live their whole life in one pen. We change the net as it grows.
Silvija: So, then there is the variable of food, and the weather conditions are reasonably stable, maybe that's one of the advantages of Norway of having the gulf stream and relatively cold water. So, by optimizing what do you make sure that the fish is as happy as possible? How do you keep them happy?
Trond: Yes, that's the big thing, right? By identifying the different marker mechanisms, the amount of them, in the water column, then you can see in the data set different markers and how the markers influence each other to then create good conditions or negative conditions. So, if you getting a negative condition, which they're obviously in some areas, you can do something. In other words, you have to kind of protect the fish in another measure. For instance, if there is an alga blooming in the top layer of the sea you stop feeding the fish and the fish goes down and stay down since there are no feed coming, and then, when the alga is done and It's drifting course away from the location, you start feeding again. So, that's one example of what you can do. You can also put in some lice nuts, as we call it, or carpets in a way around the pen, which is about beneath the lice layer because the lice are done from, let's say, 5 to 1 or 0 meters. That's where the main concentration of the lice is, and then, you run and put some carpet around the pen, or "presenning" as we call it in Norwegian, and then the lice will not come into the pen, and then, obviously, the fish is not exposed. So, there are several measures that we can add to the scenario.
Silvija: How do you convert this good farmer sense into data?
Jørn: That's really interesting, let me get back to that. I just wanted to comment that one of the three key things is to balance these measures because you want to keep the lice out, so you put in these carpets, but then again, you also want the fish to have sufficient oxygen levels for example, which these corporates limit because it limits the fall through of fresh, oxygenated water. So, you need to balance and do the right message. You need data and input from your sensors, and you need to know what's happening in the water.
Silvija: These sensors help me imagine them, so, they measure what?
Jørn: With the typical distances we had through a long time are about some of the basic environmental conditions, so, when we mention oxygen, it's important, then, salinity and temperature as well. Temperature is probably one of the most important factors acting in the growth of the fish. What's happening now is that we gain more and more sensors measuring different things and we were talking about the computer vision some a minutes ago, and that's also just a sensor giving us more information about what's going on in the pens, so, fish behavior, detecting pellets if they're falling through the fish layer, those of the fishes that have stopped eating, etc. Of course, that does impact on operational efficiency because feed is expensive, but also the environment because the pellets are falling through into the bottom, and they're being eaten by different species that are not supposed to have the feed of the salmon. So, being able to measure the fish appetite and feed accurately is good for business and it's also really good for the environment. So it's a win-win case.
Silvija: Then you connect this to the production numbers, which are what?
Jørn: Well, production numbers are basically what we were talking about, but it's also a lot about the inventory. So, you have the number of fish you put in a pen and you have the estimated growth of that fish, so you have an estimate of the complete biomass in your cage at any point in time. With the computer vision that might change because we will get a more accurate reading of the actual size and weight of the fish, so, instead of using a model over almost two years of salmon growth, we can do accurate readings of their current size, which is good. That's important production data. Of course, we record along with that the conditions in which they are living at every point in time, the environmental conditions. There is also information regarding fish health, so, diseases, vaccines, we collect mortality, the reasons of mortality, why do the fish die, or when the fish gets removed for specific courses when they are sent to harvest because when they are at the end of their life they're supposed to be food, so we record that and what feed they get through their lifetime. So, when the fish is going to harvest you can take out more or less the fish's CV, and you can read all that happened through the fish since it basically hatched and then went into the sea.
Silvija: Can I just go back to one question? I'm just fascinated by the relevance of fish lice. Can you just help me understand? It's a fish that is lice-infected to death, or is it just not going to try very much? Can we eat it? In what sense do lice affect the fish?
Trond: It's unpleasant for the salmon to have it on their skin. You can see that it doesn't actually feel well up to a certain amount, and then if you got too many lice on it it's also eating the skin of the fish, so you got wounds on the fish and that's not the scenario you want to get at all. That's why it's very important actually to try to avoid that with a device that is driving down or setting down into the pen.
Silvija: Do you medicate to get rid of it, or do you blow some air through the thing? What do you do?
Jørn: There are many different measures. Previously, medications were used a lot, but the lice have gone resistant to the most commonly used medication, so, new measures have been invented, typically mechanical, so we take the fish out of the sea and we use warm water to paralyze the lice, and then, they fall off. It can be flushed off with semi high-pressure water, we have so nice little brushes that brush the lice off the fish, so that's used very, very much. Of course, there is more fancy stuff recently developed like the lice laser, that shoots laser beams and kills the lice falling off fish. So, there are different measures.
Silvija: From fish brushes to lice lasers, there is an incredible world out there. So, now we go back to AquaCloud. We have this data about the size of the fish, the number of the fish, oxygenation of the water, and the temperature, and as we correlate this data we can figure out if there is an even better way to bring up the fish. Is that the goal? You told me that your goal is sustainable growth, what does that mean?
Jørn: Do you want to take this one, Trond?
Trond: Yes, that's easy. The aim is actually to build up the total knowledge of what's going on, what influences the fish as such. One thing is in the freshwater stage of the fish, which is the first 18 months of the growth of the life cycle of the fish, and then it comes to the sea. In the freshwater, you have good control over the facilities that we are using, and then, there is a more open question as to what is coming through the water to influence. So, by adding some sources and building up the data sets to see the markers and identify the markers as early as possible, then you can understand the fish in a much better way and how the fish behaves based on the current condition in the pen. That's why we are trying for, why we are using Machine Learning, why we are adding all the algorithms so that we can understand the fish the outermost, so we are as early as possible to try to give the fish the best possible way of living. It's as easy as that, but as complex as that.
Jørn: Thinking about sustainable growth it's a big term because growth is about really a significant growth of the industry because there is a goal of producing five million tons of salmon in 2050, which is crippling of the current production rate. So, to get there, we need to do a lot of things: we need to keep salmon competitive in terms of producing it efficiently, the price per kilo is competitive against other sources of animal protein, it needs to be attractive to customers, customers are more and more environmentally conscious, so it needs to be tracked and proved that it's been brought up in an environment in a good way and it's produced in a licensed environment. You need licenses to produce salmon and there are conditions to that, limitations of handing out licenses. It depends on solving some of the issues like for example the sea lice problem, which is a huge limitation today because the total impact in the area on a wild salmon population is impacted by the density of new forms. So, we need to fix that if we're going to get more licenses and all this brings together, it's being attractive to customers, being within government regulations, and produced in a sensible business. All of this works hand-in-hand.
Trond: Everything is also interconnected, right? One thing is that we need to be sustainable in there, but at the same time, we also need to ensure that our feed providers, for instance, actually are producing sustainably themself as well. So, there are no child workers involved in the production and all these aspects to it, and that's the total picture of sustainability. It's a vital process in our industry, same as when we take the fish from the farm to the customer, how do we do that in the most environmentally positive way? There are many elements to it and everything is then interconnected.
Jørn: We really at least don't have all the answers on how to handle this, but what we do know, is that access to high-quality comparable data is vital to understand this and going forward. We took the learnings from the lead Silas prediction projects we did a couple of years ago and we use that knowledge to see where can we contribute as a platform, and that is to provide data to help the standardization, so that data is more comparable and standard across different providers and we make that data available to research institutes, entrepreneurs and startups, or farmers themselves. Anyone who has great ideas, that has great knowledge and insight, can use data to give these answers back to the industry. So, we don't have the answers, we probably don't have all the questions even, but we try to provide that infrastructure to let others stay cold and focused on their ideal value proposition.
Silvija: I was just going to add that the mission here is partly to build a data culture in the space of aquaculture, where the more people in your cluster, in your NCC food greater environment, that agreed to be a part of this data-producing society, the better we will be able to understand our happy fish.
Trond: Absolutely, and there is an add-on to that as well. One thing is to collect the data and another thing is to share them and work together while sharing and analyzing the data, test each other's algorithms and models for analysis, and how you approach a challenge and work across the companies in the industry. That's the next level, but we are already started, and that's the benefit of AquaCloud, is the cultural sharing, the culture of sitting a dagger and gain insight into the data that will benefit the industry, not only the individual companies to the industry.
Silvija: I have to comment on our international audience here and it has to do with the Norwegian public sector platforms for data. We have one of the best platforms in the world in something called Altinn, which is an all public tax-related, financial data flows, related to individuals and it's a super interesting collaboration between our banks, tax authorities, regulatory authorities, and administrative authorities, and I'm wondering if you guys are making an Altinn in for fish.
Jørn: Well, to some extent, yes, and on the other hand, absolutely not. This is a voluntary effort by companies who want to come together and want to share, which typically means that we have more velocity because there are no regulations involved. So, we can act sooner, we can adapt data structures and contents, and what we do with it more rapidly. I think we have, of course, some challenges because we can not mandate to get all the data, but we have some advantages in terms of what we can do and how fast we can do that.
Silvija: You can avoid some of the politics of our electronic patient journals. I think it's a really interesting point that you're talking about velocity because here we have Jørn from the cluster and the kind of the platform owner, and we have Trond, who's one of the most ambitious users of this platform and perhaps also one of the earliest of both developers and users. So, Norwegian fish farming is not necessarily like all other kinds of fish farming. I know other countries are over-farmed, and therefore, you know they have an irreversible lice problem perhaps, so, what you're trying to do here is grow from 1.1 million tons to 5 million tons, but in a way that works with nature. Is that correctly summarized?
Trond: That's correctly summarized and that is also the clear statement for a government and regulators as well, so this is something that we have a common goal on definitely.
Silvija: So, what's your request to the people that are listening to this? I think, in some ways, you've already said it, but, you have some data and you'd like to have even more data, you have good users and you'd like to have even more users.
Jørn: Yes. We want as many as possible Norwegian farmers to join the platform to share the data and to also get access to share data sets between each other. We are talking with lots of innovators and data consumers out there, but we're really happy to have more in that conversation.
Silvija: What do people need to do to join your data revolution?
Jørn: Well, as a farmer you can contact us through our website AquaCloud.ai, or if you are a consumer we will soon be releasing our developer portal, you can subscribe to data sets that we make available.
Silvija: And you help people analyze as well? You don't need to have a master's degree in Computer Science to be able to do this?
Jørn: Well, it will help on the analytic side. We try not to go into the analysis, we work somewhat with our platform just to ensure the quality and try to see possible uses of the data, but how to connect, correlate, and find those deep insights in the data, we leave that to all our partners.
Silvija: Trond and Jørn, as an outsider to your industry, I have to say, it's been incredibly inspiring to talk to you. I think for all of us Norwegians listening to this, learning more about the real advantages and strengths of our second largest industry is a real necessity. Thank you so much for contributing here.
Jørn: Thank you.
Trond: Thank you very much.
Du har nå lyttet til en podkast fra Lørn.tech- en læringsdugnad om teknologi og samfunn. Nå kan du også få et læringssertifikat for å lytte til denne podkasten på vårt online universitet Lørn.university.
LØRN AS, c/o MESH,
Tordenskioldsgate 2
0160 Oslo, Norway
Bibliotek
Om LØRN
© 2024 LØRN AS