Emma Boland

Polar Oceanography with an interest in machine learning and advanced modelling techniques.

Research Area

I study the Polar Oceans using a combination of computer models and observations. The polar oceans are hugely important to our climate, being the only place on the planet where old deep waters come to the surface, exchange heat and carbon with the atmosphere, and then return to the global ocean interior, locking away that heat and carbon for hundreds to thousands of years. Most of the heat and carbon we have added to the climate system has ended up in the ocean, so it’s crucial to understand what controls these processes now and how it might change in the future.

I use a variety of techniques to carry out my research, including adjoint modelling techniques, analysing climate models and observations, and machine learning techniques. Adjoint modelling is an exciting tool for examining how climate models behave that can find connections that traditional approaches miss. I’m enthusiastic about using machine learning to help do more with the sparse observations we have in the polar regions and to inform where we should prioritise more observations.

Project Interests

I'm looking to develop projects that use techniques such as machine learning or GPU-based numerical models to generate new insights into observational priorities for the Southern Ocean, based on the fusing of observational and simulated data. Polar observations are hard won, so we need to know that any new observations are well-targeted to provide valuable scientific insights. Possible research questions include: How can we optimise observations of Southern Ocean warming? What can we learn about ice shelf melt from remote observations? How can we efficiently track the Southern Ocean carbon sink?

I'm also interested in projects that look at the future state of the Arctic. We expect the Arctic to be seasonally-ice free by around 2050. Some studies suggest this could cause freshwater leaving the Arctic to dramatically increase, others that it will remain fairly constant. Using cutting edge modelling and machine learning techniques, we can simulate many possible future conditions and start to answer questions like: How stable is the structure of the future Arctic? Will freshwater flood the sub-polar North Atlantic or will it just shift location? How big an impact will future conditions have on biological productivity?"