Personalized ocean forecasts in a two-way data flow system

We all know the significance of the weather forecast in our day-to-day decision making, but for those working in close interaction with the ocean, a similar ocean forecast can be just as important. For example, search and rescue operations rely on precise predictions of drift trajectories based on the ocean currents, winds and waves; aquaculture industry can plan feeding routines based on the forecast of the local flow conditions; and who would not like a forecast of the swimming temperatures during the summer vacation? The research project HAVVARSEL aims to develop localized and personalized ocean forecasts based on the examples just mentioned. The ocean dynamics can be described by similar physical laws as the atmosphere, which means that it is possible to forecast the ocean in the same way as the weather. The main difficulty, however, is that there is a lack of observations from the ocean that can tell us from which ocean state to start the forecast. This is why HAVVARSEL will develop web services facilitating a two-way dataflow for ocean information. The users will be able to upload ocean observations of local currents, temperature, salinity and so on. This information can then be used and assimilated into the existing ocean forecast, enabling an improved localized and personalized forecast that can be sent in return. The project partners SINTEF Digital, Institute of Marine Research and the Norwegian Meteorological Institute join forces to develop infrastructure and improved simulation tools to make localized ocean forecasts for three selected demonstrators. The demonstrators will be developed in dialog with a stakeholder group from industry, public service, and media, to ensure that the forecast information will fit the stakeholders' information needs. Web services for receiving and providing observations from pilot users are operational and curation of stakeholder data sets are well underway. Development of methods for producing localized and personalized forecasts have also been started. The web portal havvarsel.no, which so far can display sea surface temperature, salinity, surface wind and current, has been launched and well received by the general public. An API has also been developed. Havvarsel.no will be expanded with new results and data sets from the project.

HAVVARSEL intend to develop an improved ocean forecast web service. It will improve operational models of ocean state (such as temperature, salinity, currents, sea level and waves) by assimilation of user provided data. Further, the project will pave the way for local personalised forecasts. Ocean 'weather service' demands very large computer resources. Further, methods and computer capacity for including more irregular observations in the forecasts by assimilation have improved. Combined with the development of cheaper instrumentation and better IoT and cloud technologies, which now is accessible for local industry and public, gives a window of opportunity to develop locally more precise forecasts directed towards a wide set of user groups. The partners SINTEF DIGITAL, Institute of Marine Research and the Norwegian Meteorological Institute join forces to develop the methods needed, and to build competence towards, utilising and handling this increasing amount of observational data from new providers, while ensuring GDPR compliance. Three pilot demonstrators will be developed; on search-and-rescue, on aquaculture management information and on swimming temperature in the sea. The latter directed towards media, tourism and the public in general. The demonstrators will be developed in dialog with a stakeholder group to ensure that the forecast information will fit the stakeholders information needs. We will have an emphasis on FAIR (findable, accessible, interoperable, reuseable) data management and easily accessible interfaces for other entrepreneurs to access data and develop their own downstream products.