Manolin COO Natalie Brennan joins Oliver Fredriksson, Writer and Editor at FoodUnfolded, which is powered by the European Institute of Innovation and Technology (EIT) Food. This interview originally appeared on www.FoodUnfolded.com.
Oliver: Many have heard of forecasting in terms of weather, but probably not in the context of fish farming. How do predictive analytics work in aquaculture?
Natalie: The principles are actually the same: You look at factors like the occurrence of a disease or a toxic algae bloom in the past to predict the likelihood of an occurrence in the future.
In practice, we create forecasts for fish farms. For example, we recently developed a forecast for Pancreas Disease (PD), which is a disease found in salmon that can be transmitted between farms during production, and limits salmon growth. To predict the development of this disease, we collect data to identify which factors (like sea temperature, the number of boats to a farm, etc.) impacted the past development of PD.
We then run algorithms over these factors to assess the relationships between them. Based on these relationships, the computer can forecast when there is a low, medium, or high chance of developing PD in the near future – allowing fish farm operators to manage the farm activities like harvesting, feeding, or cleaning accordingly to prevent PD. Essentially, this technology gives farmers the tools to use past events as a means to make the most educated decisions in the future on how to maximize and maintain the welfare of their fish.
So, how exactly could digitalization help with fish welfare then?
To understand how it can impact fish health, it’s important to understand how fish health is broadly maintained in the first place. Essentially, the health of a fish is dependent on 3 main areas:
A fish can get sick or die when the conditions in these areas go too far outside their ideal level. To give you an example, if the water temperature is slightly too warm, and there is an increase in harmful bacteria in the water, then an action like moving the fish from one pen to another (usually somewhat harmless) could be enough to push the fish over a critical threshold of what the fish can withstand before falling ill or dying.
By understanding the relationships between these factors and how it impacts the fish, our forecast models help fish farmers see challenges or threats to fish health sooner. It gives them the opportunity to take preventative measures to keep the fish healthy, rather than dealing with the consequences once fish are already in a vulnerable position.
Fish farmers already do some of this analysis in their heads, and with years of experience they get very good. But there are limits, as the human mind can only manage so many factors and the relationships between them. This is where computers and data science techniques come in. Combined with a farmer's knowledge, it provides a much more powerful analytic tool to give the fish the best possible living conditions.
Aside from maintaining fish welfare, how could going digital benefit fish farmers?
Data provides more certainty when making high-stakes decisions, like whether to do another parasite treatment or to harvest. Digitalization speeds up tasks like filling out regular reports to the authorities, assessing how the overall fish health is going, or changing production plans. This helps farms run more efficiently with fewer costs and higher yields. It also helps communication within the team, allowing leaders to see the data behind certain decisions and to go back and assess how effective past choices have been.
On top of all this, it helps to fast-track the 30+ years of experience that older farmers have. New employees can now look back and see the decisions made or the dynamics at a site and resulting outcomes to better inform daily decisions. Not to mention it lessens the amount of time spent in front of an Excel sheet, allowing for more time out on the farm.
If going digital means more recording of information, could consumers expect to see greater product transparency as a result?
In short, yes, definitely. With more fish and environmental data recording from farming operations and other sources, consumers will be able to know more about both where their product comes from and what has happened to it. Combined with technology like blockchain, this data can then be safeguarded against change or corruption, as well. It provides added transparency to consumers by raising the level of accountability farms have to their product. Increasing transparency opens the door for a more open and productive dialogue between consumers and industry.
You work primarily with the Norwegian salmon farming industry, has it been difficult gaining support for a software product in a traditionally hands-on industry?
It has been an interesting journey for sure, and because we are tying data science to traditional fish farming, there has certainly been a lot of education on both sides. Visiting farms in small remote Norwegian towns helps us to understand the issues and limitations they face on a case-by-case basis. We are personally involved and try to understand as much as we can about how each company has grown and overcome challenges, and this has helped bridge that critical gap between data scientists and farmers. It’s incredibly important to build those relationships so we can work alongside farmers to provide the digital data tools that will help with some of those challenges.
Many are concerned about AI or digitalization taking jobs from people. Have you felt this sentiment reflected in the aquaculture industry?
Hmm, no I wouldn’t say so. I think more of the issue is how AI will benefit them, and how do they get there from where they are today. Farming, in my opinion, will always need farmers. There are certain things that will never be able to replace the farmer's touch and intuition.
Sustainability is driving a number of the large protein-producing industries today. Can increased digitalization make any positive steps towards sustainability in fish farming?
Absolutely, it does in a few ways:
Lastly, where do you personally hope to see the future of fish farming heading?
Being a relatively young industry (in the west, at least), I hope to see it learn and build upon both the mistakes and successes of agriculture. The industry is growing rapidly with huge potential; especially if we continue to develop offshore technology.
As consumers, we are much more demanding and aware of the importance of sustainable food production than ever. My personal hope is that this focus and awareness on sustainable food production and environmental management becomes part of the core DNA that makes up aquaculture food production. I hope technology, transparency and trust continue to also be a critical part of where aquaculture is heading: A food production method that represents the values of many today.
And if I am honest, the time I have spent in the Norwegian industry, visiting farms and talking with all sorts of people in the industry, has shown me that it is heading in that direction. It’s an exciting space to work in and I look forward to more passionate people bringing new ideas to the industry in the time to come
For more data on the state of Norwegian salmon farming, subscribe to Manolin’s newsletter.