Digital Arctic Shipping - New data products and visualisation services

The Arctic is undergoing the most rapid changes in the climate system worldwide, demonstrated by the thinning and reduction of sea ice, the melting of ice sheets and glaciers, thawing of permafrost, and an increased possibility for extreme weather events and other hazards. While the role of these Arctic changes in increasing risks of extreme weather events remains strongly debated, these interlinked processes are expected to increase the chance of natural hazards connected to unstable sea ice conditions and unexpected occurrences of icebergs. Such rapid and hard to predict changes pose a significant risk for shipping activities in this region, whether it be cargo, fisheries or tourism. Therefore, improved access to new and customised data products from innovative combinations of available in-situ measurements, remote sensing imagery and forecasting models is critical to further development of safe and sustainable Arctic shipping. The overall objective of the Digital Arctic Shipping project is to develop a suite of data products in collaboration with Arctic shipping companies and local authorities and to demonstrate the use of these products on vessels traveling in ice infested areas. This includes using drone mounted sensors custom-built for Arctic operations, in combination with near real-time data from satellites, and automatic identification of ice dynamics patterns (divergence, convergence). Examples of products include maps of ice types, ridges, leads and motion, which are all essential parameters for navigation in sea ice. Products will be presented to scientists and personnel onboard vessels operating in ice covered areas, together with data from other sources such as automatic weather stations (AWS) and data supplied by other actors (e.g. Automatic Identification System (AIS)). The project has participated in two field research cruises north of Svalbard in 2020 and 2021. During these two cruises project personnel collected data using drones, ship-borne radar and by physical measurements on the ice. Drones were used to collect high-resolution images over different types of ice, in the same area as the physical ice measurements and in the area surrounding the vessel. A series of ice stations with ice cores and measurements of temperature, salinity and density of the ice were collected. In addition, a Ku-band radar was used to collect radar images of the ice conditions around the vessel. A set of photographs were taken for both cruises to further document the ice conditions at the ice station locations and along the route. Data from the ice stations will be used for validation of the drone data and in the construction of a 3D model of the ice using drone photogrammetry. In 2022 the project participated in a research cruise in the Greenland Sea. This cruise focused on use of different types of drones to collect data from various imaging sensors around the research vessel, testing machine learning algorithms and communication of drone data in real time from the drone to the vessel. The algorithms for classifying different types of sea ice in drone image have developed by means of machine learning (ML). These classification algorithms will ultimately run in the on-board computing device of the drones, which was tested during the last cruise in 2022. Furthermore, the project has enhanced visualisation services for the types of data collected or generated, and for combining project data with data from other sources. Data, algorithms and services from the project have been demonstrated to different actors operating in ice infested areas of the Arctic. Our results can form the basis for future products and services in support of safer navigation under difficult ice conditions. Further developments will be carried out through new research and innovation projects.

The main outcome of the project is improved competence and capability to observe sea ice conditions around vessels operating in ice infested areas under Arctic conditions in support of safe navigation. This includes knowledge and technology to apply machine-learning techniques to high-resolution and multi-sensor payloads to classify imagery onboard the drone, and to transfer drone data back to the vessel in near real-time using radio communication. Other major results include integration, visualisation and analysis of satellite, drone, AIS and other data to support scientific studies and planning of field operations. Furthermore, the project has contributed to improving sea ice forecasts using advanced data assimilation and ML techniques, as well as extending GIS systems that can be used by scientists, vessel operators and other stakeholders both during, planning and execution of field operations and in the subsequent post-cruise data analysis. Results have been presented in scientific presentations and papers and will be utilised in ongoing and future projects for scientific studies and services development in support of safe navigation in the Arctic and to support safe scientific operations on the sea ice.

The Digital Arctic Shipping project will develop new sea ice data products and visualisation services for Arctic shipping. Automated and intelligent sensors will be used to collect data in difficult and dangerous navigation areas. The consortium has the following observation technologies available for the project: drones carrying multispectral or radar sensors, single or dual-band ship radars, in situ measurements of snow and ice, and a broad range of satellite sensors (e.g. synthetic aperture radar, passive microwave and optical) from Chinese, European and US space programmes. The heterogeneity and vast amount of data generated by these sensors requires scalable ICT frameworks to efficiently manage and extract key sea ice parameters that can be translated into useful information for Arctic shipping. The chosen technologies and tools will be co-developed together with ship operators with extensive knowledge of navigation in Arctic regions. Specifically, we will apply novel machine learning techniques such as deep learning for automatic detection of riskful conditions. The derived data products will be organised in a data management system and distributed to targeted user groups using standard protocols and communication networks. We will demonstrate the developed data collection, management, analysis and distribution services through case studies targeted at the Arctic shipping industry. Each case study will be co-developed by scientists, local authorities and shipping companies and result in new data products and visualisation services for supporting safer Arctic shipping. Demonstrations of products and services will focus on cargo traffic in the Northern Sea Route and other ship traffic areas in the Arctic. The project partners have collaborated on Arctic and Climate science for the past 15 years and will strengthen and extend this collaboration to provide digital services for customised data delivery for Arctic shipping.