Gradcracker - Where the STEM jobs are
My Gradcracker Dashboard Login
My Account Activity
Upcoming webinars I have registered for
0 of 13
Alerts from my Followed employers
Employers I am Following
Opportunities in my shortlist

New to Gradcracker?

Register here

Degree Apprenticeships Placements / Internships Graduate Jobs
Company Hubs A-Z Sector A-Z
STEM Societies
Webinars Toolkit


Imaging Geophysicist

"In my first six months at CGG I’ve worked on aspects of both seismic signal processing and earth velocity model building for imaging."

MSci in Mathematics, Imperial College London

Current project: North Rona Ridge 3D, a rich azimuth seismic survey. The 3600 km² project focuses on delivering high-resolution seismic data in a prospective but under explored area north-west of the Shetland Isles.

Current focus: Using multi-layer reflection tomography to build a velocity model for imaging the seismic data.

I am a mathematician with a love of reading, baking and puzzles.

I’ve always been interested in problem solving and the connections between different branches of mathematics, but also between different branches of science.

I was drawn to CGG because of its scientific work, allowing me to continue to immerse myself in science without remaining in academia. I’d never thought about working in geophysics before, and knew very little about the subject, but the training I received was thorough and allowed me to immediately begin to contribute to the team. CGG has a friendly atmosphere, and everyone is very welcoming and helpful as you learn on the job.

Although as a geophysicist I do not really use my maths in the way that I did at university, the skills I developed to assess and solve novel and interesting problems are used daily. The ability to adapt to new problems and draw on knowledge from across different topics; and being able to think analytically about a problem which at first seems impenetrable and persevere. I think my knowledge of mathematical concepts and understanding of how to approach new topics and learn about them has also helped me a lot in understanding the physics of seismic imaging and signal processing.

In my first six months at CGG I’ve worked on aspects of both seismic signal processing and earth velocity model building for imaging. This has included denoise (removing the many different types of noise), demultiple (removing multiple reflections of the geology and sea surface) and deblending (separating out different source-receiver combinations as the data was acquired with triple sources using simultaneous source technology – see image!); as well as data preparation and imaging for tomographic updates and FWI. I have just started working on velocity model updating using reflection tomography, which is all new to me, so exciting to learn some new things!

Multi-layer tomographic inversion for velocity modelling

Multi?layer tomography is an extension of non-linear slope tomography, and uses a hybrid model format that uniquely defines the velocity and anisotropy parameters for each model layer as a b-spline mesh while also carrying the precise information for the layer boundaries as horizons. It is an incredibly versatile tool for building accurate and stable velocity models of the earth, which we can use for imaging seismic data. It principally uses reflection data to drive the update as this is what we create our image from but a whole range of different data types can be input as additional terms in the cost function or as constraints.

The input data typically consists of combined measurements of the steepness (dip), direction of steepest slope (azimuth) and the residual move out (velocity error) of imaged geological reflections. These are kinematically demigrated (taken back to the pre-imaged domain using the model the data was imaged with) to become kinematic invariants i.e. they become independent of the imaging model so can be re-imaged with a new starting model before performing the update. This is an incredibly powerful step because it allows measurements made on different images to be combined; the starting model for the update can be refined and improved as the model building progresses without needing to re-image the data each time; and the starting model can be very different from the initial imaging model.

The image below shows the uplift that multi-layer tomography can bring over conventional methods.

Follow your favourite employers
Save jobs to your shortlist
Receive personalised alerts
Access our live webinars
Register now