Experimental, Computational and Machine Learning Methods in Geophysics

A thermal plume in syrup imaged by thermochromic liquid crystal. Credit: Carolina Lithgow-bertelloni

I am also very interested in developing and applying novel methodologies, including both end-to-end and generalizable approaches to address various challenging problems in geophysics.

Specificlly, I focus on:

• State-of-the-art laboratory fluid dynamic experiments and analysis workflow tailored for viscous convection applicable to planetary interiors (Bao and Lithgow-Bertelloni 2024; Bao and Lithgow-Bertelloni Submitted; Bao and Lithgow-Bertelloni Submitted). They are based on real-world physics and especially suitable to investigate multi-scale flow.

Cartoon of the 3D-scanning Stereoscopic Particle Image Velocimetry (SSPIV) system I built at UCLA

• Digital twins of lab experiments based on a hybrid-adjoint method (Bao 2024). These fully quantified numerical models are tightly constrained by 4D lab measurements, but they further extend the results in space and time and even to unmeasured quantities.

SSPIV measurement that covers 68% volume (left), full velocity reconstructed with hybrid-adjoint (middle), and reconstruction of unmeasured temperature field (right) Ra=1e6~1e8

• Machine learning and AI methods to analyze, simulate and invert complex geophysical systems. These include data mining from geophysical and geochemical data (Bao et al. 2022; Bao et al. 2024; Bao et al. In revision for Science), as well as forward and inverse modeling with physics-infomred neural networks (PINNs) (Bao 2024) and neural operators (Bao et al. In preparation).

3D view of a numerical simulation of Rayleigh Benard convection at Ra=1e5, viscosity contrast up to 10 (hotter less viscous)

2D central-plane view of the simulation above. Left panel is prediction from PINNs, right is the misfit from the reference simulation.

Related Work

1.

Bao, Xiyuan and Carolina R. Lithgow-Bertelloni. 2024. “Self-correction of the optical distortion effect of thermal plumes in particle image velocimetry.” Physics of Fluids

2.

Bao, Xiyuan and Carolina R. Lithgow-Bertelloni. Submitted. “A method for measuring and analyzing four-dimensional flow fields in viscous rayleigh-benard convection.”

3.

Bao, Xiyuan and Carolina R. Lithgow-Bertelloni. Submitted. “Morphology, interactions, and evolution of laminar thermal plumes revealed by four-dimensional velocity measurements and analysis: Application to earth’s mantle.”

4.

Bao, Xiyuan. 2024. “PhD thesis: Hotspots from top to bottom.” https://escholarship.org/uc/item/4dx424br.

5.

Bao, Xiyuan, Carolina R. Lithgow-Bertelloni, Matthew G Jackson, and Barbara Romanowicz. 2022. “On the relative temperatures of earth’s volcanic hotspots and mid-ocean ridges.” Science 375(6576): 57–61.

6.

Bao, Xiyuan, Tushar Mittal, and Carolina R. Lithgow-Bertelloni. 2024. “Determining mid-ocean ridge geography from upper mantle temperature.” Earth and Planetary Science Letters 641: 118823.

7.

Bao, Xiyuan, Mathurin D. Wamba, and Andreas Stracke. In revision for Science. “A deep-mantle geochemical zoning map beneath east africa and indian ocean.”

8.

Bao, Xiyuan, Sophie Coulson, Natasha Valencic, Jerry Mitrovica, Sönke Dangendorf, Mostafa Mousavi, and Andrew Lloyd. In preparation. “Inverting sea level fingerprints for ice mass changes using neural operators: How much can we recover and how unique is it?”