With a clearer picture of where the fish are, fishermen can plan more efficient routes that save both time and fuel, in addition to promoting more sustainable fishing practices. We have created the AI solution that can transform the Norwegian fishing industry.
Co-author: Aleksander Stangeland
In an era where digital innovation is transforming most sectors, Knowit has directed its focus towards Norway's second-largest export industry – the fishing industry. Fuel and crew account for over half of the operating costs of fishing vessels (1), and a significant amount of time at sea is spent searching for fish. At the same time, it is estimated that 5,000 tons of fish are discarded annually in Norway due to bycatch (2).
With this in mind, Knowit has combined its expertise in AI, data-driven value creation, and maritime domain knowledge to create a solution that can contribute to a more efficient and sustainable fishing industry.
Since 2018, Knowit has been working on digital decision support using ocean and maritime positional data. The developed solution utilizes advanced AI models and big data to provide forecasts on the whereabouts of different fish species for the next seven days. The goal is to create a "weather forecast for fish" that will assist fishermen in planning their fishing routes. This is done through an app that fishermen can use while fishing, showing them where and when it is best to fish based on the targeted species and the gear they use.
Image: Better catch predictions with AI
The digitization of the fishing industry has led to increasing reporting requirements for fishermen. As the first in the world, the Norwegian Directorate of Fisheries has made catch data from the Norwegian fishing fleet open to everyone. The motivation behind this is to drive innovation and promote open management of Norway's marine resources, but the growing reporting requirements can make life more challenging for many fishermen without providing them with tangible value in return. We believe there is untapped potential in the data reported from fishing vessels. With our solution, it may be possible to utilize this data to simplify the daily lives of fishermen.
By gaining better insights into the movements of different fish species, it also becomes possible to reduce the impact of fishing on marine ecosystems. Bycatch, the unintentional capture of non-target species, is not only a waste of resources but can also lead to a loss of biodiversity. Although it is illegal to discard bycatch at sea, it still occurs to a significant extent.
The predictions of fish locations make it possible to fish more targeted towards specific species, thereby minimizing bycatch.
The predictions are generated by AI models trained on over 10 years of historical data from various sources. This includes catch reports, vessel position data, ocean currents, temperature, salinity, chlorophyll levels, and other oceanic data that fishermen use when searching for fish.
Video: Some of the ocean data used to predict fish migration.
The oceanic data is also made available as supporting information layers in the solution, allowing fishermen to see the underlying data in addition to the predictions. To strengthen trust in the forecasts, we believe it is important for our models to be interpretable, meaning it should be easy to see which data is used in the model and how different variables affect the results.
The user interface of the solution has been developed to be easy and intuitive to use. Throughout the process, user experience has been a priority, and there has been a focused effort to create a tool that is easy to use, regardless of the equipment used on different boats.
Knowit's solution provides daily insights and is easily accessible on mobile devices and other screens that fishermen have with them while fishing. We have collaborated closely with the industry to ensure that the solution meets the requirements and needs of fishermen. Through dialogues with fishermen, we have understood the need to prioritize features such as a night mode, as much fishing occurs during the night, and a responsive interface that adapts to mobile phones, tablets, and monitors.
Video: The solution is easy to use in the form of an app.
Image: Unsplash
Referanser:
(1) Profitability survey for the fishing fleet - time series (NO)
(2) Robert has to throw good fish in the bin: - We do not prioritize this fish (NO)