Laura Mansfield giving her talk at EGU
Research group
Laura, Lilli, Peter, Simon, Edward and Milan attended the EGU General Assembly in Vienna. The week was very exciting and we enjoyed attending many talks and posters. We were also lucky to have wonderful sunny weather.
Talks and posters by group members:
- Laura gave an invited talk on Uncertainty Quantification of Machine Learning Parameterisations, showing results on quantifying parametric uncertainties in a machine learning emulator for atmospheric gravity waves and the resulting uncertainty on stratospheric circulation.
- Laura also co-convened a session on Advancing Earth System Models using Machine Learning, where we gained insights into how others are using machine learning to improve climate models and their workflows. The room was jam packed!
- Simon gave a talk titled “Unravelling the role of increased model resolution on surface temperature fields using explainable AI” in the “Machine Learning for climate” session.
- Peter presented preliminary results from his work on a machine-learned parameterisation of sub-grid physics with built-in stochasticity. The response to his (tiny) poster was positive
- Edward gave a talk titled “Precipitation rate, convective diagnostics and spin-up compared across physics suites in the model uncertainty model intercomparison project (MUMIP)”, presenting data from the MUMIP (researchers from all modeling centers co-authoring the dataset).
- Lilli shared preliminary results from her current work on Discovering convection biases in global km-scale climate models using computer vision. In her poster, she compared the two nextGEMS models ICON and IFS-FESOM to satellite observations to better understand how well the two models simulate convective clouds.
- Lilli also gave a talk on Reconstructing 3D cloud fields from multispectral satellite images using deep learning where she presented results from her work as part of the Frontier Development Lab summer research sprint last summer.
- Milan convened a short course on using Julia for simulating the oceans, atmospheres and ice, presenting a live demo of SpeedyWeather, and also gave an invited talk on “Challenges and perspectives of climate data compression in times of kilometre-scale models and generative machine learning”
We all greatly enjoyed networking and exchanging ideas with groups from other universities and institutions! Thank you to everyone who participated in that!
Laura presenting her research
Lilli presenting her research