Faqi Liu

From SEG Wiki
Jump to: navigation, search
Faqi Liu
Faqi Liu 2020 headshot.png
BSc Mathematics
MSc Mathematics
PhD Geophysics
BSc university Qufu Normal University
MSc university Chinese Academy of Sciences
PhD university University of Texas at Austin

The 2020 SEG Reginald Fessenden Award [1]

by Mrinal Sen and Bin Wang

Faqi Liu has identified and analyzed a broad spectrum of reverse time migration (RTM) artifacts, including ellipses, rabbit ears, and cigar wave paths. Subsequently, he developed algorithms to mitigate such effects from seismic images. These fundamental ideas were applied in the 1980s in full-waveform inversion but first revealed by Faqi for RTM. Faqi has developed methods that are now part of the standard seismic imaging workflow for many companies and researchers in our industry.

Biography Citation for the 2020 SEG Reginald Fessenden Award

A mathematician by training, Faqi Liu never thought he would choose geophysics as his career. That lasted until he graduated with a master’s degree in applied mathematics from the Chinese Academy of Sciences following a bachelor’s degree in mathematics from Qufu Normal University in China. Even though his master’s thesis was on dip moveout by wavefield transformation, he had focused mostly on the mathematics and its numerical implementation without much interest in the geophysical background. The real turning point came in fall 1994, when he got an opportunity to work as a visiting scientist at DELPHI at the Delft University of Technology, The Netherlands. In 1996, Faqi enrolled at the University of Texas at Austin to pursue his PhD in geophysics, where he strengthened his background in computational seismology and published a seminal paper on multiple attenuation in coupled plane wave domain.

Upon graduation from the University of Texas in fall 1999, Faqi started his industry career as a research geophysicist with Conoco’s Seismic Imaging Center in Ponca City, Oklahoma, and later moved to Conoco-Phillips in Houston, Texas, after the merger of Conoco and Phillips. Faqi started with developing ray-tracing software in both isotropic and transversely isotropic media for Kirchhoff depth migration. He later switched to one-way wave equation migration. In the early 2000s, limited computing power challenged the large-scale application of prestack wave equation migration. Faqi proposed a composed shot record migration using plane wave construction, called source plane wave migration, producing comparable results with much less computation.

Faqi joined Hess Corporation in 2005 where he continued his endeavor in depth imaging. In early 2006, he began to work on reverse time migration (RTM). Starting from scratch, he developed the initial version of the RTM software and obtained the first RTM image in about six months. Soon after, he noticed the strong low-frequency artifact with large amplitude in an image that badly contaminates the real signal, and a simple filter is incapable of removing it properly. This triggered his scientific curiosity. He tried to understand the underlying physics and look for a mathematical remedy for it. After some experiments and discussions with his master’s degree advisor Guanquan Zhang, who came to the United States from China for the 2006 SEG Annual Meeting, Faqi invented an imaging condition for RTM by wavefield separation, which was able to successfully eliminate the back-scattered artifact from the image. The implicit separation of the wavefields at the imaging step seems physically straight forward, but is not mathematically intuitive. Faqi provided detailed demonstration of both its theory and the numerical implementation in a Geophysics paper published in 2011 after he first presented the research at an SEG Annual Meeting a few years earlier. This work is important not only because it has been adapted in production by many companies, including both exploration and production companies and geophysical service firms, it also has inspired many geophysicists to extend its application beyond depth imaging — for example, to full-waveform inversion (FWI) of reflection data. It is worthwhile to mention that the wavefield separation-based imaging condition is very efficient interms of both CPU time and memory consumption. In his last few years with Hess, Faqi worked on FWI and, at the SEG summer research workshop on FWI in 2012, was among the first groups to realize and present on the importance of integrating tomography into an FWI workflow.

Faqi went on to work with PGS in early 2017 in its Houston office, where he led a group and developed its least-squares full-wavefield migration for data containing both primaries and multiples. Most recently, Faqi has been employed at TGS NOPEC as the research manager responsible for developing innovative algorithms for velocity model building, FWI, and depth imaging.

SEG recognizes Faqi’s innovations by presenting him the 2020 SEG Reginald Fessenden Award. It has been our privilege and pleasure to write this citation for Faqi.


  1. Honors and Awards Ceremony Program, SEG Annual Meeting, 13 Oct 2020, Houston