Publications

Highlights

(For a full list see below or go to Google scholar)

Nonlinear inversion in the scattering-source space

Yonghyun Chung

Manuscript in Preparation (Hopefully 2027)

1. Conventional Born-approximation-based linearized inversion neglects higher-order sensitivities, often producing nonlinearity artifacts and limiting convergence and resolution.

2. We introduce a scattering-source space and a nonlinear sensitivity operator to quantify and directly incorporate nonlinear effects into the inversion formulation.

3. We apply the method to geophysical field data to resolve seismogenic fault structures and demonstrate its superiority over conventional methods.



Efficient 3D MT inversion at seismogenic-fault scales using data-space regularization

Yonghyun Chung and Tae Jong Lee

Manuscript in Preparation (Hopefully 2027)

1. Three-dimensional MT inversion can mitigate errors from 2D assumptions, but its computational cost limits applications to localized areas.

2. We adapt a previously developed data-space regularization framework to enable efficient 3D MT inversion at the scale of seismogenic faults.

3. We validate the method using 3D MT data acquired near the Gyeongju and Pohang earthquake source regions in Republic of Korea.



GPU-powered 3D earthquake simulation

Yonghyun Chung and Jaejoon Lee

Manuscript in Preparation (Hopefully 2027)

1. We present a GPU-powered simulator for high-performance 3D earthquake simulation.

2. It combines OpenMP-based domain decomposition, GPU-thread parallelism, and direct inter-GPU communication to overcome memory and communication bottlenecks.

3. On four Tesla P100 GPUs, it achieves approximately 130× the simulation speed of the CPU-based MPI–OpenMP implementation under equivalent conditions.



Data-space first-arrival traveltime tomography using damped wavefields

Yonghyun Chung and Sukjoon Pyun

Manuscript in Preparation (Hopefully 2027)

1. Damped wavefields suppress later arrivals and provide finite-frequency sensitivity to first-arrival traveltimes, reducing reliance on high-frequency ray approximations.

2. Formulating the inversion in data space enables a computationally efficient Gauss-Newton solution.

3. By combining these advantages, the proposed method improves computational efficiency and the spatial resolution of the reconstructed images, as demonstrated using field data.



A versatile framework for data space regularization

Yonghyun Chung

Manuscript in Preparation (Hopefully 2027)

1. Regularization can reduce the computational efficiency of data-space inversion because constructing model covariance is computationally demanding.

2. To address this challenge, I propose a computationally efficient regularization framework for data-space inversion.

3. By achieving structural constraint parity with model-space regularization, this framework removes a key barrier to the widespread adoption of data-space inversion.



Three-dimensional inversion of controlled-source electromagnetic data using general measures to evaluate data misfits and model structures

Yonghyun Chung, Soon Jee Seol and Joongmoo Byun

Geophysical Prospecting (2024)

doi:10.1111/1365-2478.13576

1. We developed a new 3D controlled-source electromagnetic inversion algorithm that allows general measures to be made selectively available.

2. The behaviours of general measures are examined focusing on weighting mechanism and user-defined parameters.

3. Our findings highlight the seamless integration of general measures, contributing to improved robustness against data outliers and enhanced spatial properties provided in output models.



Three-dimensional modelling of controlled-source electromagnetic surveys using an edge finite-element method with a direct solver

Yonghyun Chung, Jeong-Sul Son, Tae Jong Lee, Hee Joon Kim and Changsoo Shin

Geophysical Prospecting (2014)

doi:10.1111/1365-2478.12132

1. We propose a fully three-dimensional edge-finite-element algorithm for simulating CSEM surveys.

2. We compare the performance between direct and iterative solvers.

3. The results show the superior efficiency of direct solvers, particularly for multisource problems.



 

Full List

Peer-review Journal publication


Yonghyun Chung (Hopefully 2027), Nonlinear inversion in the scattering-source space, Manuscript in Preparation

Yonghyun Chung and Tae Jong Lee (Hopefully 2027), Efficient 3D MT inversion at seismogenic-fault scales using data-space regularization, Manuscript in Preparation

Yonghyun Chung and Jaejoon Lee (Hopefully 2027), GPU-powered 3D earthquake simulation, Manuscript in Preparation

Yonghyun Chung and Sukjoon Pyun (Hopefully 2027), Data-space first-arrival traveltime tomography using damped wavefields, Manuscript in Preparation

Yonghyun Chung (Hopefully 2027), A versatile framework for data space regularization, Manuscript in Preparation

Yonghyun Chung, Soon Jee Seol and Joongmoo Byun (2024), Three-dimensional inversion of controlled-source electromagnetic data using general measures to evaluate data misfits and model structures, Geophysical Prospecting, 72(8), doi:10.1111/1365-2478.13576

Yonghyun Chung, Jeong-Sul Son, Tae Jong Lee, Hee Joon Kim and Changsoo Shin (2014), Three-dimensional modelling of controlled-source electromagnetic surveys using an edge finite-element method with a direct solver, Geophysical Prospecting, 62(6), doi:10.1111/1365-2478.12132