On-the-fly tsunami simulations using machine-learning and AI
Project ID: 2228cd1387 (You will need this ID for your application)
Research Theme: Physical Sciences
UCL Lead department: Institute for Risk and Disaster Reduction (IRDR)
Lead Supervisor: Fatemeh Jalayer
Project Summary:
Why this research is important? Rapid and accurate tsunami forecast is vital for an effective early warning system. This leads to saving lives through preparedness and prompt evacuations. A tsunami forecast provides information such as inundation distance, flow depth, and arrival time. Therefore, it requires rapid collection and process of observational data from several sources, eg, GNSS, seismic stations, ocean bottom sensors, and tidal gauges.
The tsunami early warning systems are mainly based on decision matrices, envelopes and/or pre-calculated scenarios. With increased high-performance and exascale computational capabilities, there is considerable momentum towards real-time data assimilation and on-the-fly tsunami simulations. However, these simulations are computationally expensive, they generally do not cover non-seismically induced tsunamis, and might not be sufficiently accurate for near source tsunamis.
Recent advances in artificial intelligence (AI) and machine learning provide opportunities for trustworthy, responsible, more accurate and rapid data processing and method developments. The main objective of this project is to study and develop fast yet accurate machine learning-based on-the-fly tsunami simulations accompanied by transparent and comprehensive uncertainty characterization/quantification.
The student will be trained in tsunami modelling, probabilistic risk analysis and forecasting, Artificial Intelligence, advanced machine learning techniques, and uncertainty characterization and quantification.
You will work with different organizations including several Tsunami Early Warning Service Providers, local authorities, and the Global Tsunami Model research community to collect data and test/validate your developed models. You will have access to IRDR’s new HPC GPU server and will have the possibility of applying for access to supercomputers in BSC (Spain) and Cineca (Italy) which host HPC PTHA software and workflows.
Who we are looking for • You have a Master of Science degree, preferably in Earth Sciences, Computer Science or related fields. • You are fluent in python, and advanced machine-learning tools. • You have knowledge of geophysical hazards.