Mapping seafloors for a better understanding of glaciers

Martin Jakobsson will use AI to map seafloors for a better understanding of marine glaciers in a warmer climate.

Knowledge of the depth, bathymetry, and shape, morphology, of the seafloor is crucial for several marine research areas, such as marine biology, oceanography, marine geology, and geophysics. Bathymetry is also fundamental for many underwater constructions, such as laying all critical communication cables.

An example of the importance of bathymetry is glaciers' exposure to warmer ocean currents where they flow into the ocean. Deep channels can allow warmer water currents to reach and melt them from below, while shallow areas may protect against the same water. Glaciers resting on the seabed leave traces, "submarine glacial landforms," providing information, for example, on how quickly they retreated as the climate warmed after the last glacial period.

Automate identification and classification of seabeds 

As a Wallenberg Scholar, Martin Jakobsson aims to develop methods using artificial intelligence and machine learning to automate the identification and classification of seabed forms, focusing on glacial landforms. The main goal is to increase knowledge of marine glacier dynamics to better predict their development in a warmer climate. The greatest uncertainty in estimates of global sea-level rise lies in how much glaciers and ice sheets in contact with the ocean will lose in mass. The methods will also have applications far beyond studies of glacial landforms. Another goal of the project is to establish machine learning in marine geoscience research at Stockholm University.

Jakobsson will leverage his over two decades of experience in marine geophysical seafloor mapping and research on the marine cryosphere to step into working with artificial intelligence and machine learning through collaboration with computer scientists. Machine learning algorithms will be trained with echo sounder data collected during expeditions starting from 2007 when a multibeam echo sounder was installed on the icebreaker Oden. Jakobsson's research group also has access to data through the international network established within the "International Bathymetric Chart of the Arctic Ocean, IBCAO." Today, they lead the IBCAO work through a global mapping project called Seabed 2030, which aims to have the world's ocean floors mapped by 2030.