Using physics-informed machine learning methods to understand Earth surface processes and predict high-resolution landscape changes in the Arctic under climate change.
Helmholtz AI project call showcase: Physics-Constrained Deep Learning for Quantifying Surface Processes across the Arctic
Researchers from the German Research Centre for Geosciences (GFZ) and the Alfred Wegener Institute for Polar and Marine Research (AWI) are developing a framework to predict the impact of climate change on Arctic deltas. Read more in this week’s Helmholtz AI project showcase.
Could you introduce yourself, giving your affiliation, area of work, and of course, the project title?
My name is Hui Tang. I’m a senior scientist in the Earth Surface Process Modelling section at GFZ. Our group's work focuses on the development and application of data science methods to quantify and understand different Earth surface processes. Our project title is 'Physics-Constrained Deep Learning Framework for Quantifying Surface Processes across the Arctic Region (PCDL-QuaSPA)'.
In simple words, what specifically is your project about? And, how and why do you think it is a high risk, high gain endeavour?
The goal of the PCDL-QuaSPA project is to develop a framework understanding Earth surface processes and predicting high-resolution landscape changes in the Arctic under climate change with a coarse-resolution dataset. In general, there is only a very limited set of labelled data in the Arctic. Therefore, we need to find an intelligent way to combine domain-specific scientific knowledge to reduce data requirements for training the data science model. We use physics-informed machine learning methods to achieve that. Then, it can help us solve some long-standing research questions, such as how global warming will change surface processes in the Arctic system like the Lena Delta (Figure 1).
Figure 1: One of our tasks in the PCDL-QuaSPA project is to reconstruct and predict how the behaviour of Arctic deltas changes under the influence of climate, like the Lena Delta in this figure.
How important has the Helmholtz AI funding and platform been to carry out this project?
The Helmholtz AI funding gives us a chance to support and train our young scientists (Erik Ngai Ham Chan) to work on edge-cutting research as well as use the updated GPU facility.