ICTP-East African Institute for Fundamental Research
KIST2 Building CST
Nyarugenge Campus
University of Rwanda
Kigali, Rwanda
GEO@EAIFR Webinar Series 2026
Subsurface to inner core structure using Bayesian inversion of Geophysical observations
Starts: 24 March 2026 16: 00
Ends: 24 March 2026 17:00
Central Africa Time (CAT)
Speaker:
Surya Pachhai
Abstract:
Earth’s internal structure and dynamic processes cannot be observed directly but must be inferred by analyzing measurements taken at the surface. These measurements include, but are not limited to, gravity anomaly, electrical properties, seismic waves generated by earthquakes, the whole Earth’s vibrational pattern in response to large earthquakes, and even ambient noise.
Advanced computer models utilize these observations to image the electrical, density, and seismic properties of the material that lies deep beneath our feet. Over the years, the availability of high-quality data and the development of more powerful computers have significantly enhanced our understanding of the Earth’s interior. However, robust estimation of Earth’s internal structure and its proper physical interpretation are challenging. This is particularly because the measurements contain noise, model complexity is unknown, different models can explain the same data, and it is often challenging to know how certain we are about the results.
In this talk, I will describe a Bayesian method that helps address these challenges. This method estimates not only the structure of the Earth but also how much uncertainty is present in the data and the model itself. By combining prior scientific knowledge with observations, this approach produces a range of possible solutions rather than a single fixed answer. In this approach, model complexity is also treated as unknown, avoiding overfitting (making the model too complex) or underfitting (making it too simple) the data. I will share examples of how I applied this method to study different parts of the Earth, including the thickness of subsurface ice using gravity, subsurface seismic velocity structure using microtremors, groundwater variations using electrical resistivity, the deep mantle using body waves, and the inner core using normal modes.
Register in advance for this event by clicking here.