Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.

org/page/policies/terms
DOI:10.1190/leedff.2024.43.issue-4

Special Section:
Gravity, electrical, and magnetic methods

ISSN 1070-485X
April 2024 · Volume 43, No. 4
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms
DOI:10.1190/leedff.2024.43.issue-4

 
     
       
     
        
     
   

  
 A SUPERIOR IMAGE
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

NEXT MONTH
NOT NEXT YEAR
Conventional Workflow DUG MP-FWI Imaging

LAND
DOI:10.1190/leedff.2024.43.issue-4

MARINE

Imagine producing a better subsurface image, ready for interpretation, in weeks,

not months. Using superior physics, DUG Multi-parameter FWI Imaging replaces
the traditional processing and imaging workflow, making all that possible now.
Don't wait another day to ask us how.

SURANCE SURANCE
AS AS
COMPASS
COMPASS

SE
SE

Let’s talk—contact info@dug.com dug.com/fwi

RVICES
RVICES

• • IS
IS O
•

O 27001
•

9001
 The Leading Edge
Table of Contents
Special Section: Gravity, electrical, and magnetic methods Departments
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

205................ Editorial Calendar

208 �������������Introduction to this special section: Gravity, electrical, and magnetic methods
I. Filina, M. Fedi, J. Sun, and A. Morgan 206................ President’s Page
207................ Foundation News
210 �������������Magnetic data — What am I looking at?
M. P. Bates, E. K. Biegert, and A. B. Reid 260................ Reviews
261................ Membership
218 ��������������Potential fields as a tool to characterize the inaccessible areas of the earth: The case of Pine Island–Ellsworth
Mountains area, West Antarctica 262................ Meetings Calendar
G. Ferrara, F. Ferraccioli, and M. Fedi 264................ Seismic Soundoff

228 �������������Analytic continuation: A tool for aeromagnetic data interpretation

J. B. Thurston and B. Fornberg On the cover: Glaciers and rock outcrops
in Marie Byrd Land, West Antarctica.
Credit: NASA/Michael Studinger.

235 �������������An artificial intelligence workflow for horizon volume generation from 3D seismic data
A. Abubakar, H. Di, Z. Li, H. Maniar, and T. Zhao
DOI:10.1190/leedff.2024.43.issue-4

246 �������������Spectrometric borehole logging in mineral exploration and mining

M. Queißer, M. Tudor, J. Schubert, A. R. Domula, H. Märten, T. Heinig, T. Rothe, and A. Stadtschnitzer

Episode 217:
Advancing subsurface knowledge through
Listen free On demand ❘
microseismic insights
with Joël Le Calvez, Principal Geologist at SLB

Episode 216:
Rethinking data – Geophysics in the era of change
with Lindsey Heagy, Assistant Professor, Department of Earth,
Ocean, and Atmospheric Sciences, University of British Columbia

Episode 215:
Strengthening diversity in the geosciences
with Isaac Crumbly, Associate Vice President for Careers and
Collaborative Programs, Fort Valley State University

Episode 214:
The untapped potential of the earth's hidden
commons with Iain Stewart, Professor of Geoscience
Communication, University of Plymouth, UK seg.org/podcast

202 The Leading Edge April 2024

 Explore more at
shearwatergeo.com
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms
DOI:10.1190/leedff.2024.43.issue-4

MORE TO
EXPLORE
Just 29% of the Earth’s surface is covered by land, we cover the rest.
We explore and analyse what’s beneath the seabed. This provides
the knowledge needed to make informed decisions for responsible
use of the Earth’s resources.

We’re explorers at heart.

 The Leading Edge
SEG Board of Directors
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

PRESIDENT TREASURER DIRECTOR AT LARGE

Arthur C. H. Cheng Mike Mellen Olga I. Nedorub
The Chinese University of Hong Kong Houston, TX, USA ConocoPhillips
Houston, TX, USA Houston, TX, USA

PRESIDENT-ELECT PAST PRESIDENT DIRECTOR AT LARGE

John Eastwood Kenneth M. Tubman Catherine Truffert
Calgary, AB, Canada SAExploration IRIS Instruments
Houston, TX, USA Orléans, France

FIRST VICE PRESIDENT DIRECTOR AT LARGE DIRECTOR AT LARGE

Mauricio Sacchi Sergio Chávez-Pérez Constantine Tsingas
University of Alberta Mexican Petroleum Institute Saudi Aramco
Edmonton, AB, Canada Ciudad de México, Mexico Dhahran, Kingdom of Saudi Arabia
DOI:10.1190/leedff.2024.43.issue-4

SECOND VICE PRESIDENT DIRECTOR AT LARGE CHAIR OF THE COUNCIL

Marianne Rauch Ana Curcio Allen J. Bertagne
Houston, TX, USA Proingeo SA BRT Energy Advisors LLC
Buenos Aires, Argentina Houston, TX, USA

VICE PRESIDENT, PUBLICATIONS DIRECTOR AT LARGE

Kyle Spikes Lillian G. Flakes
The University of Texas at Austin GeoSoftware
Austin, TX, USA Richardson, TX, USA

The Leading Edge® (Print ISSN 1070-485X; Online ISSN 1938-3789) is published monthly by the Society of Exploration Geophysicists, 125 W. 15th St., JIM WHITE, Executive Director
Suite 100, Tulsa, Oklahoma 74119 USA; phone 1-918-497-5500. Periodicals postage paid at Tulsa, OK and at additional mailing offices. SCOTT SUTHERLAND, Managing Director, Business and
Geoscience Technology
Print subscriptions for professional members of the Society in good standing are included in membership dues paid at World Bank IV rates. JENNIFER COBB, Managing Director, Publications and Membership
Dues for Active and Associate Members for 2024 vary depending on the three-tiered dues structure based on World Bank classification of the JENO MAVZER, Director, Journals
member’s country of citizenship or primary work residence. Students may purchase print subscriptions by paying a discounted fee during dues STEVE BROWN, Managing Editor
renewal or by contacting members@seg.org. Nonmembers may obtain online subscriptions by contacting books@seg.org. Copies of single issues KELSY TAYLOR, Associate Editor
of The Leading Edge® may be purchased by inquiry to books@seg.org. MARIA GEE, Graphic Designer
Corporations, universities, and other institutions can purchase access to The Leading Edge® via standalone subscriptions or one of several SARAH WEATHERS, Publishing Platform Manager
available subscription package options. SEG’s subscription pricing features three academic and three corporate tiers, providing affordable STACY BAKER, Publishing Platform Analyst
access to applied-geophysics content for institutions of any size. SEG subscription packages offer significant savings on the price of single-
publication subscriptions. Special pricing is available for organizations with multiple sites and for institutional consortia. For more information on Editorial information: 1-918-497-5503, sbrown@seg.org
package pricing, including single-site and multisite institutional subscriptions, please contact Patrick Riley at sales@seg.org or 1-918-497-5531. Advertising information and rates:
Institutional subscriptions for print delivery are available only with the purchase of a digital subscription. Rates are subject to change without 1-918-497-5524
notice. For more information, please visit https://seg.org/subscriptions. Kathy Gamble, kgamble@seg.org
Advertising rates for The Leading Edge® will be furnished upon request. No advertisement will be accepted for products or services that cannot Subscription information:
be demonstrated to be based on accepted principles of the physical sciences. For more information, contact kgamble@seg.org. Members, members@seg.org
Statements of fact and opinion are made on the responsibility of the authors and advertisers alone and do not imply an opinion on the part of Nonmembers, books@seg.org
the officers or members of SEG. Copyright 2024 by the Society of Exploration Geophysicists. The Leading Edge®, SEG®, and the SEG logo are Institutions, Patrick Riley, 1-918-497-5531,
registered marks of the Society of Exploration Geophysicists. All rights reserved. Material may not be reproduced without written permission. priley@seg.org
Printed in the USA. POSTMASTER: Send changes of address to
The Leading Edge
125 W. 15th St., Suite 100
Tulsa, OK 74119 USA

204 The Leading Edge April 2024

 The Leading Edge THE LEADING EDGE
Editorial Calendar EDITORIAL BOARD
CHAIR
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Chester J. Weiss
Sandia National Laboratories
Albuquerque, NM, USA
cjweiss@sandia.gov
To learn more about submission opportunities
scan the QR code shown at right
or visit library.seg.org/TLE-sections. Heather Bedle
University of Oklahoma
Norman, OK, USA
hbedle@ou.edu
Issue Special section theme Due date Guest editors
All Technical standalone articles ongoing TLE Editorial Board

Niels Grobbe
May 2024 Rock physics past due Gregor Baechle University of Hawai‘i at Mānoa
Honolulu, HI, USA
Jeremie Dautriat ngrobbe@hawaii.edu
Laurent Louis1
June 2024 Subsurface uncertainty past due David Lubo-Robles
Matt Walker
Madhav Vyas1 Joël Le Calvez
July 2024 General submissions past due TLE Editorial Board SLB
Sugar Land, TX, USA
Chester J. Weiss1
DOI:10.1190/leedff.2024.43.issue-4

jcalvez2@slb.com

August 2024 Geophysics and sustainability past due Julia Correa

Aleksei Titov
Vladimir Kazei1
Madhav Vyas
September 2024 Focus on the Mediterranean region 15 Apr 2024 Walter Rietveld BP America
Ivica Mihaljevic Houston, TX, USA
Maha Khattab madhav.vyas@bp.com
Ramesh Neelamani1
October 2024 Geophysical methods 15 May 2024 Rich Krahenbuhl
in archaeology Michael Wilt1
Vladimir Kazei
November 2024 Optical fiber 15 Jun 2024 Erkan Ay Aramco Americas
Houston, TX, USA
Joël Le Calvez1 vladimir.kazei@aramcoamericas.com
December 2024 Reservoir characterization 15 Jul 2024 Satinder Chopra
Tom Davis
Heather Bedle1
Laurent Louis
1
TLE Editorial Board coordinator Aramco Americas
Houston, TX, USA
laurent.louis@aramcoamericas.com

Ramesh (Neelsh) Neelamani

ExxonMobil
Houston, TX, USA
ramesh.neelamani@exxonmobil.com

TLE publishes special sections and standalone articles covering all aspects of applied
geophysics and related disciplines. Submission of articles is open to all. Please submit articles via
Michael Wilt
the online manuscript submission system at https://mc.manuscriptcentral.com/tle. Submission Lawrence Berkeley National Laboratory
instructions are available at https://library.seg.org/TLE-authors. For full descriptions of special Berkeley, CA, USA
mwilt@lbl.gov
section themes, see https://library.seg.org/TLE-sections. TLE Editorial Board coordinators work
with guest editors to coordinate and support the review process and also may serve as guest editors.
For additional assistance, contact tle@seg.org.

April 2024 The Leading Edge 205

 President’s Page
CCUS: Challenges
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

and opportunities
This month’s author:
Shan Zhou, Purdue University,
Department of Political Science

C arbon capture, usage, and storage

(CCUS) technologies can move car-
bon dioxide produced from fossil-fueled
a collaborative effort among the private
sector, local, state, and federal government
to develop a coherent and systematic regu-
mechanisms to ensure long-term safety and
environmental protection.
An emerging challenge for policy-
power plants and other industrial facilities latory approach for CCUS. makers and stakeholders is how to pri-
to geologic storage sites at a distance. Due From a social justice perspective, the oritize social equity considerations in
to the pressing societal need to reduce deployment of CCUS technologies may CCUS planning and implementation to
greenhouse gas emissions, CCUS has exacerbate existing disparities if not care- ensure that benefits and burdens of these
gained wide attention as a flexible, cost- fully managed. On one hand, CCUS initia- projects are equitably distributed. In this
effective, and rapid approach to climate tives can create job opportunities, stimulate process, governments need to engage with
DOI:10.1190/leedff.2024.43.issue-4

change mitigation and energy security. economic development, and generate tax diverse stakeholders, including commu-
CCUS is particularly appealing to indus- revenues in some areas. Private landowners nity groups, environmental justice advo-
tries for which decarbonizing has been who lease their land for CCUS develop- cates, and industry representatives, to
historically challenging. ment typically receive compensation from build consensus, address competing
Regulatory authority of CCUS in the developers. However, the process of obtain- interests and priorities, and make effective
United States is shared largely between ing consent from the broader community, and equitable siting decisions of CCUS
state and federal governments. The federal including adjacent landowners and uncom- infrastructure. Marginalized communi-
government provides financial incentives pensated residents, can lead to conflicts ties often lack access to information,
and technical assistance to catalyze CCUS and tensions within the community. resources, and political power, which can
deployment. An example is the recent On the other hand, CCUS projects are lead to distrust and misunderstanding.
federal income tax credits authorized under often considered locally unwanted land An inclusive and participatory community
the Inflation Reduction Act of 2022, which uses and thus face strong public opposition engagement approach needs to be
significantly boost the economic viability in the planning and siting process. The “not employed to ensure that all stakeholders
of CCUS projects. The federal government in my backyard” phenomenon is prevalent have a voice in shaping CCUS policies
can also become involved when developers in CCUS deployment as many are con- and projects. This may involve conducting
seek to store CO2 beneath federal lands. cerned with risks associated with CCUS, outreach and education efforts, providing
U.S. state governments are responsible such as carbon leaks and induced seismic opportunities for public comments and
for establishing rules and standards for events. Historically marginalized com- feedback, and establishing community
the siting, permitting, construction, and munities, including low-income neighbor- advisory boards or task forces to facilitate
operations of CCUS infrastructure. hoods and communities of color, often bear dialogue and communication.
CCUS injection-well permitting is regu- a disproportionate burden of environmental Despite these challenges, many
lated by state governments and the U.S. pollution and health risks associated with already see CCUS as a viable option to
Environmental Protection Agency’s industrial activities (e.g., fossil-fueled help build more resilient and inclusive
Underground Injection Control program. power plants and ethanol plants), which communities that are better equipped to
In addition, local land use planning and are large point sources of CO2 . These com- address the challenges of climate change,
permitting policies (e.g., zoning ordi- munities may face disproportionate harms economic development, and social
nances) can play an important role in from CCUS because CCUS infrastructure, inequality. Efficient and equitable deploy-
shaping CCUS projects. Due to the frag- such as pipelines and injection wells, tends ment of CCUS technologies requires a
mented authority among multilevel govern- be located in or close to their neighborhood. coordinated and dedicated effort from all
ments, developers often encounter difficulty Communities hosting CCUS infrastruc- stakeholders, including geoscientists,
navigating the complex CCUS regulatory ture also face the formidable challenge of policymakers, community leaders, and
regime. There remains an opportunity for devising effective monitoring and liability industry representatives.

206 The Leading Edge April 2024

 F o u n d at i o n N e w s
SEG Trustee Associates honored during Foundation event
Sarah Hewitt1
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

O n 31 January 2024, the SEG Foundation recognized its

Trustee Associates for their significant charitable contribu-
tions to the Foundation and their continued support of SEG
the SEG Foundation, with support from Trustee Associates,
corporate partners, and members, are working to recruit and
support the best and brightest in the field of applied geophysics
programs. The celebration was held at the Houston Country Club to continue to make a positive impact in our world.
in Houston, Texas. It was generously sponsored by SEG supporters, The SEG Foundation continues to invite charitable participa-
Geophysical Pursuit, SLB, and TGS. tion and new members in support of advancing the science of
An SEG Foundation Trustee Associate is a Society member applied geophysics, supporting members, and growing our
who makes a commitment to cumulatively contribute US$10,000 positive global impact. In addition to the Foundation’s Trustee
or more to the annual fund or $25,000 or more to SEG programs Associates, cumulative and annual contributions from members
within a five-year period. Trustee Associates are contributors who of any amount are welcome and will contribute to furthering
have demonstrated a strong shared commitment to the current the Foundation’s cause. Learn more and make a donation by
and future vitality of the Foundation and Society. They support visiting https://seg.org/foundation.
our growing positive impact on the science, members, and the
world. When one gives back to SEG as a Trustee Associate, the
gift is used to support the important work and crucial programs
of the Society. This includes scholarships, field camp grants,
student expositions, life-changing global humanitarian efforts,
DOI:10.1190/leedff.2024.43.issue-4

as well as mentoring, professional networking, and growth oppor-

tunities. New Trustee Associates will also receive a significant
level of matching funds for joining the program in 2024 due to
the generosity of some existing donors.
The evening’s Mistress of Ceremonies Maria Angela Capello
was introduced by the Foundation’s Managing Director Sarah
Hewitt. Capello took the opportunity to highlight many of the
SEG programs that are supported by the Foundation. Particular
emphasis was on the impact of the expanding SEG EVOLVE
program, which currently includes curricula in the areas of world-
wide exploration, carbon capture, and geothermal project execution.
The evening also included a social, dinner, and recognition of Trustee Associates enjoyed a recent celebration and recognition event.
service to former Foundation
Chairman of the Board David
Bartel by incoming Chair
Rocky Detomo.
SEG’s science and contri-
butions to our communities
have been dynamically evolv-
ing for the past few decades
but never quite as fast or as
significantly as is the case
today. Fundamental global
changes in energy transforma-
tion are underway, and SEG’s
programs provide support to
help our future and current
fellow professionals under-
stand and adapt to this shift-
ing landscape and shape their
career trajectories. SEG and

1
SEG Foundation, managing director. E-mail: shewitt@seg.org.

April 2024 The Leading Edge 207

 Introduction to this special section:
Gravity, electrical, and magnetic methods
Irina Filina1, Maurizio Fedi2, Jiajia Sun3, and Alan Morgan4
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

https://doi.org/10.1190/tle43040208.1

G eophysics plays an important role in all aspects of geologic

analysis, spanning from regional exploration and tectonic
mapping to local prospect-level studies and energy transition
behavior. Studying subglacial geology beneath the WAIS requires
geophysical methods. Magnetic surveying is perfectly suited for
surveying inaccessible areas because of the efficient way in which
projects. Out of a variety of geophysical techniques, seismic data can be captured, e.g., by a plane or helicopter. The presented
methods often play the most central role, while less expensive multiscale analysis of magnetic data not only resulted in outlining
nonseismic methods are less frequently applied. One of the objec- the large tectonic zones, such as the Pine Island Rift, Byrd
tives of SEG’s Gravity and Magnetics Committee is to promote Subglacial Basin, and Bentley Subglacial Trench, but also made
nonseismic geophysical methods and showcase their value in it possible to identify structural features beneath the ice, such as
various geologic applications. As members of that committee, we, contacts-like or fault sources, dykes, sills, volcanic necks/conduits,
the editors of this special section, assert that gravity, magnetic, and spherical source distributions, as well as to estimate their
and electrical methodologies are powerful yet often undervalued depth and delineate several interesting tectonic trends. The study
tools. We present this special section focused on nonseismic maps various geologic features that can be further linked to
geophysical methods and the impact they can make on various potential magmatic sources that influence the formation and
geoscience projects. Included here are a regional tectonic study evolution of the West Antarctic Rift System.
in Antarctica, a local mining exploration mapping project in The last paper in our special section, by Thurston and
Canada, and an analytical methodology capable of providing Fornberg, describes the analytic continuation of potential fields
critical information about subsurface geology and helping to to a vertical complex plane using rational complex-series expan-
DOI:10.1190/leedff.2024.43.issue-4

determine the depths of important geologic structures. Although sion (Padé approximation). The paper presents synthetic examples
our scope is to provide a broader discussion of different types of of potential geologic magnetic sources, such as point and 2D
nonseismic techniques, all of the papers presented in this special sources, sheets with finite depth extent, thin slabs, and contacts.
section focus on magnetic surveying. The use of Padé approximation provides stable downward con-
We open our special section with “Magnetic data — What tinuation for distances that are significantly greater than those
am I looking at” by Bates et al. This paper was initiated by the achievable by Taylor series or by spectral methods. The analyti-
SEG Gravity and Magnetics Committee and intends to address cally continued fields for these objects exhibit polarity flips near
the confusion that may exist in the exploration community related the source location, which can be used to interpret the source
to the variety of magnetic deliverables and their sometimes confus- geometry based on the shape of the approximating function.
ing names. People often refer to the result of magnetic surveying In addition to synthetic examples, the authors also ground truth
as “the magnetic map,” but when asked about the type of map or their methodology by application to the helicopter-based mag-
what corrections (if any) were applied to it, a puzzled look is the netic survey acquired over the sulphide ore deposits in a
frequent response. This paper aims to address the general confusion Canadian greenstone belt. It is noteworthy that in the study
about different magnetic parameters and naming conventions area, magnetic rocks associated with ore deposits are overlain
— i.e., addressing what exactly “the magnetic map” shows — using by nonmagnetic sedimentary rocks, which prevents surface-
a set of aeromagnetic data over the Oka Complex of Quebec, based geologic mapping and requires geophysical exploration
Canada. The Gravity and Magnetics Committee intends to follow to map subsurface features. The result of the authors’ analytic
up with a similar paper for the gravity methodology, and we are continuation methodology agrees well with the recent wells in
happy to expand to other methods if we receive requests from the the study area, thus revealing the value of the Padé approximant
broader community. for subsurface analysis.
The second paper presents a magnetic study in Antarctica. In closing, we thank the authors who submitted papers to this
Ferrara et al. analyzed airborne magnetic data sets acquired by special section. We hope that the collected papers will help readers
the British Antarctic Survey over the West Antarctic Rift System. develop a greater appreciation of magnetic surveying in particular
Despite being one of the largest rifts on earth, this geologic region and nonseismic geophysical methodologies in general. We remain
remains poorly understood because it is buried under at least a committed to our goal of promoting the value of nonseismic
kilometer-thick ice cover. It is important to understand the complex geophysical methods to the broader geoscience community, as we
geologic setting beneath the West Antarctic Ice Sheet (WAIS) are highly confident that gravity, magnetic, and electromagnetic
in order to evaluate its impacts on ice sheet dynamics, to assess methods will play a critical role in addressing some of the chal-
the ice sheet response to climate change, and to predict its future lenges associated with the energy transition.

1
University of Nebraska-Lincoln, Lincoln, Nebraska, USA. E-mail: ifilina2@unl.edu.
2
Università di Napoli Federico II, Naples, Italy. E-mail: fedi@unina.it.
3
University of Houston, Houston, Texas, USA. E-mail: jsun29@central.uh.edu.
4
Bell Geospace, Katy, Texas, USA. E-mail: alan_morgan75@hotmail.com.

208 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms
DOI:10.1190/leedff.2024.43.issue-4
 Magnetic data — What am I looking at?
Martin P. Bates1, Edward K. Biegert 2, and Alan B. Reid3
https://doi.org/10.1190/tle43040210.1
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Abstract magnetic field of the rocks (remanent magnetization) and the

For a number of years in geophysical surveying, the use of magnetization induced by an externally applied magnetic field
certain technical terms that describe data has not been consistent. (such as that generated by the earth’s geodynamo, electric currents
This is particularly apparent in the field of magnetic surveying, in the ionosphere, and the field from nearby rocks).
which is the most commonly practiced technique. Across the The total vector magnetic field that is created by combining
world, there are multiple companies and clients that collect and the external field, remanent magnetic field, and induced mag-
employ magnetic data acquired via ground, airborne, or marine netic field is called the magnetic induction or magnetic flux
platforms. However, what they mean by certain terms is either density, usually denoted by vector B. It is the intensity of B
imprecisely defined, ambiguous, or significantly different. Terms only, expressed in units of weber •meter−2 (Wb/m 2) or tesla (T),
that are accurate and consistent will help the end user understand that is measured by the magnetometers employed in most
what they are actually looking at. We describe the most commonly surveys. By convention, the value of magnetic intensity is usually
used terms for magnetic data and discuss how these terms should expressed in nanoteslas (nT).
and should not be used.
Total magnetic intensity
Introduction The total magnetic intensity (TMI) of a magnetic field is the
For most aeromagnetic and marine surveys, only the intensity, magnitude of the vector sum of all the constituent magnetic fields
or magnitude, of the magnetic field is measured. Neither the at the measurement point. So, TMI is equal to √(B BT ), where
DOI:10.1190/leedff.2024.43.issue-4

direction of the magnetic vector, nor the magnetic gradient tensor B is the vector sum of the earth’s magnetic field and all other local
components, are normally recorded. Over time, it has become and regional sources of remanent magnetic and induced magnetic
common to say that magnetic surveys measure the magnetic field field with significant magnitude. TMI is a scalar quantity without
when in fact they mostly only measure the intensity of the magnetic directional sense (i.e., not a vector), but it varies in 3D space.
field. More recently, some surveys are being conducted that The TMI, expressed as quantity, is a standard final product
capture the vector of the magnetic field. However, for the purposes of magnetic surveys. Figure 1 illustrates the TMI of the Oka
of this paper, we only consider surveys of magnetic intensity. Complex, dominated by a bean-shaped magnetic high and its
Examples of each of the different representations of magnetic associated smaller magnetic low to the north. It is set within an
intensity are illustrated by using proprietary aeromagnetic data area of more complex magnetic character that is associated with
acquired by Sander Geophysics Ltd. (SGL) over the Oka Complex the surrounding mostly gneissic terrain. The apparently simple
in Quebec, Canada. This complex is a composite igneous pluton dipolar high-low anomaly associated with the Oka Complex
that has yielded dates of 107 and 119 Ma (Shafiquall et al., 1970). provides a good example of a well-defined magnetic anomaly to
It comprises two carbonatite cores surrounded by bands of alkali illustrate the various terms described later. Note the high value
igneous and carbonatite rocks in a ring-like arrangement that of intensity in the tens of thousands of nanoteslas that results
intrude country rocks composed of anorthosite and gneiss (Gold, from inclusion of the International Geomagnetic Reference Field
1967). An interpretation of these data is available from the SGL (IGRF) in the measured intensity.
website (Sander Geophysics, 2019). The carbonatite cores coincide At a minimum, the TMI data would have been corrected
with low magnetic intensity surrounded by bands of higher for aircraft maneuver effects and short-term temporal variations
intensity modeled with high magnetic susceptibility, averaging in the earth’s magnetic field. Temporal corrections are usually
0.3 in the System Internationale (SI). Modeling of magnetic and determined from a combination of a diurnal correction and/or
gravity data indicates that the intrusion is steep sided and fans leveling. A diurnal correction is achieved by monitoring the
out slightly on the southeastern side. temporal changes of the magnetic field while surveying using
one or more static magnetic base stations. Leveling is carried
Magnetic field terminology out by analysis of data acquired at intersections between the
If we consider a group of magnetic dipoles in a volume, the principal “survey” or “traverse” lines and intersecting (often
magnetization M is the vector sum of all the dipoles divided by perpendicular) more widely spaced “control” or “tie” lines. Various
the volume. Magnetization has the unit ampere•meter−1 (A/m) different leveling algorithms are employed. However, the short
in SI. That magnetization is the vector sum of the intrinsic time to travel from one traverse line to the next along any given

Manuscript received 7 December 2023; revision received 8 February 2024; accepted 12 February 2024.
1
Sander Geophysics Ltd., Ottawa, Ontario, Canada. E-mail: mbates@sgl.com.
2
Gentleman Geoscientist, Houston, Texas, USA. E-mail: eb723475@gmail.com.
3
Reid Geophysics Ltd., Eastbourne, Sussex, UK. E-mail: alan@reid-geophys.co.uk.

210 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
 tie line means that minimal changes in the magnetic field may human infrastructure alone when the intensity of those combined
occur during that period. This enables adjustments to be made to fields is small compared to the intensity of IGRF.
the traverse lines that bring them to a consistent level. Note that Other terms in common usage for this same quantity include
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

any heading effects (variations associated with the direction of residual magnetic intensity, crustal magnetic intensity, total field
travel of the measurement platform) are also resolved via the anomaly, anomalous magnetic field, and even, but incorrectly,
leveling process. Other corrections such as adjustment for survey TMI. It is the difference from what would otherwise be predicted
altitude may or may not be applied. However, to be described as by the model of the earth’s magnetic field (IGRF). Therefore, this
total intensity, the data must include the expected intensity of data must not include the IGRF intensity. At what stage the IGRF
earth’s magnetic field, which is normally represented by the current intensity was removed in data processing is not implied. However,
IGRF (Alken et al., 2021), along with the anomalous field that we will expect to see intensities that are commonly in the approxi-
arises due to effects of the local geology and/or impact of human mately hundreds of nanoteslas and occasionally more where
infrastructure. For example, a mid-magnetic latitude survey may particularly strong magnetic bodies occur. If the nature of the
be expected to have values in the 50,000 to 60,000 nT range. In magnetic anomaly is not stated, then AMI is assumed to be the
general, the data will have intensities that fall between about sum of induced and remanent magnetism.
20,000 nT close to the magnetic equator and values in the region Figure 2 illustrates the AMI of the Oka Complex. In this
of 80,000 nT near the magnetic poles. Note that the TMI still example, AMI is superficially similar to TMI. The subtraction
incorporates a degree of temporal variation because IGRF itself of IGRF has removed a gentle south–southwest-dipping slope
varies slowly over time. TMI may have been calculated without that is too subtle to notice relative to the anomaly amplitude. On
removal of IGRF, or IGRF may have been removed at some stage the other hand, the overall intensity has dropped significantly
in data processing but added back at a later stage. The term itself and now includes areas of both positive and negative intensity.
does not differentiate between these two possibilities. Negative intensity occurs where the TMI is less than predicted
by the IGRF.
Anomalous magnetic intensity
DOI:10.1190/leedff.2024.43.issue-4

TMI is a scalar quantity. TMI should not be confused with Derivatives of the magnetic intensity
anomalous magnetic intensity (AMI), a scalar quantity formed In general usage, the word derivative simply means some-
by subtracting the scalar intensity of the earth’s magnetic field, thing that is derived from something else. However, in the
usually represented by the IGRF, from the scalar intensity of geophysical world, we recommend that this term is reserved
the total magnetic field (Blakely, 1995). Note that AMI is not for values that represent gradients or rates of change in a specific
equivalent to the magnitude of the anomalous magnetic field direction and are therefore a vector quantity. The terms gradients
because it was created by subtracting two scalars, not two vector and derivatives can be used interchangeably to describe vector
fields. However, AMI can be a reasonable approximation for quantities described by magnitude and direction. Vertical and
the magnetic intensity that arises due to the local geology and/or horizontal derivatives of first, second, or higher order are

Figure 1. The TMI of the magnetic field associated with the Oka Complex in Quebec, Canada. Figure 2. The AMI of the Oka Complex.
All other examples are from these same data.

Special Section: Gravity, electrical, and magnetic methods April 2024 The Leading Edge 211
 commonly calculated from either TMI or AMI. The derivative Therefore, we recommend 1NHD and 1EHD for geographic
terminology does not imply one or the other. In some instances, first horizontal derivatives (1NHD of TMI, 1NHD of AMI,
derivatives may be directly measured by taking the difference etc.). Sometimes the horizontal derivatives are expressed with
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

between two physically separated magnetometers. In other respect to the principal survey line orientation, especially if they
cases, in particular for vertical derivatives, they may be calculated are directly measured by an array of sensors. One direction along
through analysis in the wavenumber domain after transformation the line is arbitrarily assigned to be forward, and the derivative
by Fourier transforms. The term derivative in itself does not parallel to the line is termed longitudinal, positive forward. The
differentiate between one that is mathematically calculated or orthogonal derivative perpendicular to the line is termed lateral,
one that is directly measured. In all instances, first derivatives positive 90° clockwise from forward.
represent rates of change in space of the magnetic intensity in Any orthogonal pair of first horizontal derivatives may be
a given direction. If no order of a derivative is provided, it is added vectorially together to provide a total horizontal gradient
generally assumed to be the first derivative of the intensity. If (THG). The THG is often used as a magnitude, but it is a vector
no direction is provided, it is generally assumed to be a vertical in the horizontal plane oriented by some angle in the plane.
derivative. However, it is recommended that both are always Figure 3 illustrates the 1VD of the Oka Complex AMI. Note
specified. The unambiguous term first vertical derivative (1VD) how the detailed form of the anomaly as captured by the shortest
is recommended so these assumptions need not be made. The wavelengths is now apparent. The magnetic lows associated with
direction of the gradient is positive when intensity is increasing the two carbonatite cores can be seen. Figures 4–6 illustrate the
in the downward direction (assuming a north-east-down coor- geographic horizontal derivatives and THG after reduction to
dinate system). Higher derivatives are rates of change in space the pole (RTP) (1EHD of RTP, 1NHD of RTP, and THG of
of the lower derivatives in some given direction. For example, RTP). The horizontal derivatives highlight the boundaries of the
the second vertical derivative is the vertical gradient of 1VD. Oka Complex with well-defined positive and negative features,
Using numbers in the naming convention allows easy extension while THG combines them into one continual high feature around
to higher-order derivatives. the edge of the complex.
DOI:10.1190/leedff.2024.43.issue-4

Derivatives may be calculated for TMI, AMI, and other

forms of intensity described later. To differentiate them, we Analytic signal and total gradient of magnetic intensity
recommend combining terms such as 1VD of TMI and 1VD of A useful parameter is the scalar magnitude of the gradient
AMI for vertical derivatives. of magnetic intensity (calculated as the square root of the sum
Horizontal derivatives, usually expressed as orthogonal pairs, of the squares of the vertical and the two horizontal derivatives
are often shown in a geographic reference frame (normally positive of TMI or AMI). This is sometimes erroneously referred to as
when intensity is increasing to the north and to the east). the analytic signal. Superficially, these are similar by construc-
Although higher orders are generally not employed for horizontal tion. The horizontal and vertical gradients of a magnetic field
derivatives, the naming convention should still allow for it. caused by 2D sources form an analytic signal and are related to

Figure 3. The 1VD of the Oka Complex AMI. Figure 4. The west to east horizontal derivative of AMI RTP.

212 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
 each other by Hilbert transforms (Nabighian, 1984). However, of field direction and is therefore not affected by assumptions
the sum-squared magnitude of the gradients from 3D sources of the type of magnetization as described later.
does not form an analytic signal. Therefore, the term total gradi-
Reduction to the pole and to the equator
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

ent amplitude (TGA) of the magnetic intensity is recommended

for general use because sources that vary in 3D will be the norm RTP is a calculation of the magnetic intensity that would
in most geologic settings. in theory be observed if the survey were conducted in the presence
Figure 7 illustrates the TGA of AMI of the Oka Complex. of a vertically inclined magnetic field (such as what occurs at
TGA forms a well-defined high directly at the location of the the north magnetic pole) rather than in the presence of the local
causative body, much like RTP. However, it is largely independent magnetic field declination and inclination. As such, RTP is a
projection and is equivalent to a phase shift of the signal. The
calculation is normally based on the assumption that AMI arises
from pure induction of magnetically susceptible material in the
presence of the earth’s magnetic field (so the magnetization
within the body is parallel to the local inducing field) and that
remanent magnetization is negligible. RTP is useful for simplify-
ing the geometry of magnetic anomalies, centering a more or
less symmetric magnetic high above the causative body if the
assumption holds true. Because remanent magnetization has a
vector field direction that is independent of the local earth’s
magnetic field direction, rotating the inducing field from the
local direction to the vertical does not usually result in vertical
remanent magnetization within the body. The resulting RTP
anomaly may remain asymmetric in appearance. Other factors
DOI:10.1190/leedff.2024.43.issue-4

such as anisotropy of magnetic susceptibility and the shape of

the causative body may also prevent the generation of simple
symmetric anomalies when reduced to the pole. Note that
asymmetric RTP anomalies due to significant remanence can
be leveraged to determine the influence of the remanence (e.g.,
Roest and Pilkington, 1993).
Close to the equator where magnetic inclination is less than
15° to 20°, the standard RTP algorithms are only approximate,
and other specially designed algorithms are employed. The usual
Figure 5. The south to north horizontal derivative of AMI RTP. term for this is low-latitude RTP (e.g., Keating and Zerbo, 1996).

Figure 6. The THG of AMI RTP. Figure 7. The TGA of AMI.

Special Section: Gravity, electrical, and magnetic methods April 2024 The Leading Edge 213
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Figure 8. The AMI RTP. Figure 9. The AMI RTE.

DOI:10.1190/leedff.2024.43.issue-4

A similar process known as reduction to the equator (RTE)

is a calculation of the magnetic intensity that would in theory be
observed if the survey was conducted where the field is inclined
horizontally with northerly declination, such as at the magnetic
equator. RTP and/or RTE may be performed on TMI or AMI.
It is perfectly valid to calculate the vertical and horizontal deriva-
tives and total gradients of RTP and RTE.
Figures 8 and 9 illustrate AMI of the Oka Complex RTP
and RTE. The RTP forms a well-defined high directly at the
location of the causative body, much like TGA. This suggests that
the assumption of dominant induced magnetization inherent in
the standard algorithm is valid. By contrast, the RTE forms a
well-defined low directly at the location of the causative body,
with intensity highs to both the north and south. RTP and RTE
allow comparisons to be made between anomalies generated in
the presence of different orientations of the earth’s inducing
magnetic field.

Regional versus residual magnetic intensity

It is common to separate longer-wavelength signal from
shorter-wavelength signal as a means toward geologic interpreta-
tion. However, what constitutes a regional intensity is scale Figure 10. Regional AMI RTP after 100 passes of a 3 × 3 Hanning convolution filter.
dependent. For example, on a large scale, the removal of IGRF
from TMI to create AMI is a regional and residual split. Thus, can be calculated. Therefore, there is no such thing as the regional
AMI is a type of residual magnetic intensity, but it is not the intensity and the residual intensity. Each pair must be defined
only possibility. The magnetic intensity data may also be divided and described to be understood, and only the general notion of
based on a specific wavelength by using various low- and high- what these names mean can be implied from the terminology
pass filters rather than a model, as is the case with IGRF. So, itself. Residuals of varying kinds can be very useful for spatial
TMI may be divided into a regional and residual, but so may analysis, such as gradient detection, and understanding the
AMI. The longest wavelength removal amounts to the subtraction distribution of magnetic sources with depth.
of the average. At the other end of the spectrum, any subdivision A regional intensity can be defined and removed by modeling
of a data set may have a local regional trend removed that is with sources or by fitting a polynomial surface. The resulting
specific to that subset, and local regional versus residual pairs residual anomalies can then be modeled and/or inverted.

214 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms
DOI:10.1190/leedff.2024.43.issue-4

D ATE
T H E
AV E
S
Mark your calendars and start making your plans to
attend IMAGE ‘24, the premier meeting for geoscientists
and energy professionals to connect and advance
understandings, expand global networks, and create
solutions for increased energy demand and challenges.

Registration opens in May.

onjunction wi
nc t
I

h:

IMAGEevent.org
 Alternatively, the division may be achieved by using low- and passes of a Hanning filter. In this example, the resulting smoothed
high-pass filtering. However, it should be remembered that regional intensity is the regional. This is subtracted from the unfiltered
and residual intensities calculated this way are not magnetic data to define the residual. Separation of anomalies associated
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

anomalies that arise from the causative bodies. To match a model with shallow geology (residual) and deeper-seated geology
response to a residual made by a low-pass filter, the model response (regional) is achieved.
must be filtered in exactly the same way. The short forms REG
and RES can be used to refer to regional and residual data however Tilt angle
they are generated. In general, tilt angle refers to the inclination or dip angle of
Figures 10 and 11 illustrate a regional versus residual split of a surface. Thus, tilt angle could refer to the dip angle of a potential
AMI of the Oka Complex after RTP. This is achieved by multiple field surface. What is often called the tilt derivative is a scalar
quantity with no directional sense, with a tangent equal to the
ratio of 1VD divided by the total first horizontal derivative
(Miller and Singh, 1994). This parameter may be better thought
of as the tilt of the derivative, and we recommend the short form
TILT.
TILT normalizes the derivatives so it is equally sensitive to
both shallow and deep sources. Using simple models and the
assumption that the magnetic field is vertical (usually achieved
by RTP), the location and depth to the source body may be
estimated. For example, Salem et al. (2007) show how the TILT
will be zero above a vertical contact at a depth equal to half the
distance between +45° and –45° contours.
Figure 12 illustrates the TILT of AMI of the Oka Complex
DOI:10.1190/leedff.2024.43.issue-4

RTP. Note the good separation of anomalies and definition of

their edges.

Figure 11. Residual AMI RTP after 100 passes of a 3 × 3 Hanning convolution filter.
2024
Events in the Americas
30 Apr–2 May – Virtual
Summit on Geophysical Detection of Explosive
Remnants of War: Solving Current Challenges of
Unexploded Ordnance (UXO) and Demining
Registration open

6–9 May – Houston, TX, USA

Offshore Technology Conference (OTC)
Registration open

22–23 May – Rio de Janeiro, Brazil

Carbon Capture, Utilization, and Storage
(CCUS) Brazil
Register by 22 Apr and save

4–5 Jun – Fort Worth, TX, USA

SPE/SEG Workshop: Injection Induced Seismicity
Register by 7 May and save
Current information as of 14 March. Please check
seg.org/events for more information and more events.
Figure 12. TILT of AMI RTP.

216 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
 Conclusion
The methods in which magnetic survey data are processed to correct for
undesirable nongeologic signal or are manipulated for the purpose of interpreta-
ModelVision
tion are always evolving. Therefore, attempting to capture all past, current, and Magnetic & Gravity
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

future possibilities as a rigid set of predefined terms is very challenging. However, Interpretation System
when certain terminology is employed in fundamentally different ways by those
engaged with magnetic data, significant misunderstandings can occur, especially All sensors Minerals
by nonexpert end users. By addressing the definition of the most commonly used Processing Petroleum
and abused terms, the authors hope to provide better clarity of data products 3D modelling Near Surface
going forward.
3D inversion Government
Acknowledgment Visualisation Contracting
The authors would like to thank Martin Mushayandebvu of SGL for helpful Analysis Consulting
insight and providing the examples of data. Utilities Education

Data and materials availability

Data associated with this research are confidential and cannot be released.

Corresponding author: mbates@sgl.com

References
Alken, P., E. Thébault, C. D. Beggan, H. Amit, J. Aubert, J. Baerenzung, T. N. Bondar,
et al., 2021, International Geomagnetic Reference Field: The thirteenth generation:
Earth, Planets and Space, 73, no. 49, https://doi.org/10.1186/s40623-020-01288-x.
DOI:10.1190/leedff.2024.43.issue-4

Blakely, R. J., 1995, Potential theory in gravity and magnetic applications: Cambridge
University Press.
Gold, D. P., 1967, Alkaline ultrabasic rocks in the Montreal area, Quebec, in P. J. Wyllie,
ed., Ultramafic and related rocks: John Wiley and Sons, 288–302.
Keating, P., and L. Zerbo, 1996, An improved technique for reduction to the pole at low
latitudes: Geophysics, 61, no. 1, 131–137, https://doi.org/10.1190/1.1443933.
Miller, H. G., and V. Singh, 1994, Potential field tilt — A new concept for location of potential
field sources: Journal of Applied Geophysics, 32, nos. 2–3, 213–217, https://doi.
org/10.1016/0926-9851(94)90022-1.
Nabighian, M. N., 1984, Toward a three-dimensional automatic interpretation of potential
field data via generalized Hilbert transforms: Fundamental relations: Geophysics, 49,
no. 6, 780–786, https://doi.org/10.1190/1.1441706.
Roest, W. R., and M. Pilkington, 1993, Identifying remanent magnetization effects in
magnetic data: Geophysics, 58, no. 5, 653–659, https://doi.org/10.1190/1.1443449.
Salem, A., S. Williams, J. D. Fairhead, D. Ravat, and R. Smith, 2007, Tilt-depth method:
A simple depth estimation method using first-order derivatives: The Leading Edge,
26, no. 12, 1502–1505, https://doi.org/10.1190/1.2821934.
Sander Geophysics, 2019, https://www.sgl.com/resources-TechnicalPapers.Interpretation.
html, accessed 28 February 2024.
Shafiqullah, M., W. M. Tupper, and T. J. S. Cole, 1970, K-Ar age of the carbonatite
complex, Oka, Quebec: The Canadian Mineralogist, 10, no. 3, 541–552.

Tensor Research

support@tensor-research.com.au
www.tensor-research.com.au
Tel:

Special Section: Gravity, electrical, and magnetic methods April 2024 The Leading Edge 217
 Potential fields as a tool to characterize the inaccessible areas
of the earth: The case of Pine Island–Ellsworth Mountains area,
West Antarctica
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Giuseppe Ferrara1, Fausto Ferraccioli 2, and Maurizio Fedi1

https://doi.org/10.1190/tle43040218.1

Abstract
We seek to characterize the geo-
physical properties of the West
Antarctic Rift System. While it is
recognized as one of the largest rift
systems on the earth, the West
Antarctic Rift System is poorly under-
stood as it is covered by the thick West
Antarctic Ice Sheet, making any geo-
logic evaluation difficult. We used the
potential-field data sets acquired by the
British Antarctic Survey, which
included aeromagnetic surveys, to con-
duct a multiscale analysis to identify
DOI:10.1190/leedff.2024.43.issue-4

the main structural lineaments, i.e.,

contacts, dikes, sills, volcanic necks Figure 1. Regional map of West Antarctica showing the Pine Island-Ellsworth-Whitmore Mountains area, overlapping the sub-ice
and conduits, and spherical source topography (modified after Morlighem et al., 2020). Continental blocks as proposed by Dalziel (1992). The red dashed circle
distributions. Several interesting areas represents the study area of the present work.
were defined in which contact-type
sources were associated with faults bordering major rift systems The West Antarctic Rift System is a region of thinned,
(Pine Island, Byrd Subglacial Basin, and Bentley Subglacial subsided continental crust (Lucas et al., 2020) characterized
Trench) or related to the tributaries in the Pine Island Rift. mainly by large sedimentary basins (Jordan et al., 2010), which
Moreover, we detected magmatic sources close to rifting zones have an influence on glacier flows. Although the geologic setting
based on their shape and orientation (Bentley Subglacial Basin of the rift system is characterized by few outcrops, geophysical
edges), helping to define the depth of the sub-ice topography. surveys, through aerial and satellite measurements, may provide
Our results helped improve our knowledge of the structural important information at the local and regional level.
regional geology of the West Antarctic Rift System. This work West Antarctica consists of crustal blocks of different tectonic
opens interesting scenarios about the extent and position of origins, all of which have been affected by complex convergent
magmatic sources and how they contribute to the topography of margin processes (Jordan et al., 2020), including the Antarctic
this sector of the West Antarctic Rift System. Peninsula, Marie Byrd Land (MBL), the Thurston Island-Eights
Coast (TI) block, and the Haag-Ellsworth-Whitmore block
Introduction (Figure 1). The Antarctic Peninsula, MBL, and TI block are all
The West Antarctic Rift System is the largest but also the forearc and magmatic-arc terranes that developed along the
least understood rift system in the world (Storti et al., 2008) due paleo-Pacific subduction margin of Gondwana (Grunow et al.,
to its inaccessibility and coverage of the West Antarctic Ice Sheet 1991; Dalziel, 1992; Dalziel and Grunow, 1992; Mukasa and
(Jordan et al., 2010), which varies from a few meters to about Dalziel, 2000). The fourth West Antarctic crustal block, the
3 km thick. The rifted bedrock, significantly below sea level (Ross Ellsworth-Whitmore Mountains, is considered an allochthonous
et al., 2012), makes the ice sheet particularly sensitive to subglacial continental fragment (Jordan et al., 2017; Lloyd et al., 2020),
intrusion of warm ocean water along tectonic rift structures, which was originally located between southernmost Africa and
thereby inducing bottom melting, bringing an important check East Antarctica within Gondwanaland (Schopf, 1969; Grunow
for ice stability (LeMasurier, 2006, 2008; Bingham et al., 2012; et al., 1987; Randall and Mac Niocaill, 2004; Lloyd et al., 2020).
Fretwell et al., 2013; Lindow et al., 2016). The area of the Pine Island Glacier and Thwaites Glacier is mostly

Manuscript received 16 December 2023; revision received 27 February 2024; accepted 5 March 2024.
1
Università degli Studi di Napoli Federico II, Department of Earth, Environment and Resources Sciences, Naples, Italy. E-mail: giuseppe.ferrara@
unina.it; fedi@unina.it.
2
National Institute of Oceanography and Applied Geophysics, Geophysics Section, Trieste, Italy. E-mail: fferraccioli@ogs.it.

218 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
 influenced by Mesozoic wide-mode rifting phase and a Cenozoic had been weakened previously. The lithospheric thinning could
narrow-mode oblique rifting causing localized extension in the derive from the initial position of the West Antarctica crustal
Amundsen Sea Embayment (LeMasurier et al., 1990, 2008; blocks before the rupture of this area of Gondwana occurred about
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Winberry and Anandakrishnan, 2004, Siddoway, 2008; Storti 180 Ma (Behrendt et al., 1996).
et al., 2008; Jordan et al., 2010; O’Donnell et al., 2019). The Pine The second phase of rifting during the Cenozoic was character-
Island Glacier lies entirely within the TI crustal block, while ized by widespread volcanism between 28 and 30 Ma (LeMasurier,
Thwaites Glacier sits at the junction of the West Antarctic Rift 1990; LeMasurier, 2008; Jordan et al., 2010), resulting in slow
System, MBL, and the TI (O’Donnell et al., 2019; Lucas et al., mantle velocities and anomalously high elevations of thinned
2020; shown in Figure 1). crust (Siddoway, 2008). Lithospheric extension associated with
The timing of rifting in this region of the West Antarctic Rift this narrow-mode rifting, localized around the Pine Island Rift,
System has been classically interpreted in response to the initial has resulted in some of the thinnest crust (approximately 20 km)
breakup of Gondwana during the Jurassic (Schmidt and Rowley, in the West Antarctic region, already weakened previously (Storti
1986; Dalziel, 1992; Elliot, 1992; Storey, 1996; Storti et al., 2008; et al., 2008; Jordan et al., 2010; O’Donnell et al., 2019).
O’Donnell et al., 2019). The rifting phase developed in two distinct Granot et al. (2013) suggested that the Byrd Subglacial Basin
periods with different characteristics: the first one, during and Bentley Subglacial Trench relate to Eocene-Oligocene trans-
Cretaceous, is wide-mode rifting, causing major crustal thinning current faults that have been reactivated by mid-Miocene to recent
in the entire West Antarctic Rift System, while the second event, extension (Lucas et al., 2020). Similar to the Bentley Subglacial
during Cenozoic, was narrow-mode rifting (Cooper et al., 1987, Trench and Byrd Subglacial Basin, the Pine Island Rift may have
1991; Behrendt et al., 1991; Tessensohn, 1994; Behrendt, 1999; originated as an Eocene-Oligocene transcurrent fault that under-
Karner et al., 2005; Storti et al., 2008; Jordan et al., 2010; went mid-Miocene up to recent extension (Granot et al., 2013;
O’Donnell et al., 2019). During the early Cretaceous, a wide-mode Lucas et al., 2020). Jordan et al. (2010) suggested that the mor-
rifting phase dominated in which extensional deformation was phology may be graben- or half-graben-like structures. The depth
widespread. The rift phase results from the trench collision between of these structures is approximately 1600 m for Pine Island Rift,
DOI:10.1190/leedff.2024.43.issue-4

the Pacific-Phoenix ridge and simultaneous subduction zone along 2600 m for the Byrd Subglacial Basin, and 2200 m for the Bentley
the South Pacific margin (Dalziel et al., 2001; Lindow et al., Subglacial Trench. (Figure 2b).
2016). Subduction stopped between 110 and 94 Ma (Mukasa and Potential-field analysis is particularly useful to infer the main
Dalziel, 2000; Larter et al., 2002). This phase of passive rifting structural and geologic features at small and large scales and
was followed by the movement of the New Zealand microcontinent represents a low-cost approach to reveal useful information even
away from Antarctica around 85 Ma (Jordan et al., 2010; Lindow in areas of difficult access (e.g., Olesen et al., 2010; Milano et al.,
et al., 2016). This first phase of rifting consisted of considerable 2020; Kelemework et al., 2021). Therefore, in this study we show
intracontinental extension and thinning (Siddoway, 2008). that the analysis of an aeromagnetic data set with a multiscale
Behrendt (1999) and others suggested that, during this phase, approach can be a successful tool to interpret the main structural
the rifting extension was localized to regions where the lithosphere features of the West Antarctic Rift System.

Figure 2. Geophysical surveys at West Antarctic sector (from red dashed circle in Figure 1). The black rectangle indicates the study area. The black dashed line surrounds the Pine Island
Rift (PIR), the Byrd Subglacial Basin (BSB), the Bentley Subglacial Trench (BST), and the Ellsworth-Whitmore Mountains (EWM) structures. Also present in the sector are the Hudson
Mountains (HM) and Sinuous Area (SA). Red triangles mark the locations of known subaerial Cenozoic volcanoes (LeMasurier, 1990; Lucas et al., 2020). (a) Aeromagnetic map (Frémand
et al., 2022). (b) Sub-ice topography map from radar survey.

Special Section: Gravity, electrical, and magnetic methods April 2024 The Leading Edge 219
 Data and method R2, defined by the zeros of the first-order vertical derivative of
As regards the geophysical surveys, airborne magnetic, radar the field at different altitudes; and a subset of multiridges, R3,
(Figure 2), and gravity data were collected simultaneously using identified by the modulus maxima of the field at different altitudes.
the British Antarctic Survey Twin Otter aircraft and were draped By a simple geometric approach (e.g., Fedi et al., 2009; Milano
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

at a constant elevation of 2.4 km above the bedrock, as defined et al., 2016), we can estimate the depth to the source and its hori-
by coincident AGASEA/BBAS airborne data (Vaughan et al., zontal position whenever some ridges intersect each other, as
2006; Ferraccioli et al., 2007a, 2007b; Nitsche et al., 2007; Corr shown in Figure 3c. The multiscale analysis is very stable with
et al., 2021; Jordan et al., 2023). The flight line direction was respect to the high‐frequency noise, as it is performed at different
mostly north–south oriented with east–west-oriented tie lines. altitudes by upward continuation of the field at the measurement
The nominal flight line spacing was 30 km. A detailed flight level. It is well known that upward continuation acts as a smoothing
grid across the trunk of Pine Island Glacier was flown with 3 km filter (e.g., Blakely, 1996). In our case, we will consider upward
line spacing. Radial lines were also flown from a field camp in continuations up to 30 km, with a 1 km step.
the middle of the Pine Island Glacier catchment (Frémand et al., Besides these geometrical features, the data on the ridges
2022). These data were used for regional interpretation. themselves may be interpreted by a method that transforms the
In this inaccessible area, with a complex tectonic setting, we potential field (T) in the Scaling function τ (Florio et al., 2009).
use multiscale methods to interpret the magnetic data. In par- This quantity is a function of the altitude z:
ticular, we estimated the source type and depth information using
the multiridge analysis of a multiscale data set (e.g., Milano et al.,
∂ log​(T​(z)​)​ ____n​(z)​z
2019; Paoletti et al., 2020). ​τ​(z)​ = _
​ ​= ​  ​​ , (1)
The multiridge analysis (Fedi et al., 2009) is a multiscale
∂ logz z− ​z0​ ​(​z​)​
method based on the analysis of the lines formed by joining
selected points of the field, and/or of its n-order derivatives, at where z 0 is the depth to the source and n is the degree of homo-
different altitudes. These lines are called ridges, and the term geneity of T. Estimating n is important because we may recover
DOI:10.1190/leedff.2024.43.issue-4

“multiridge analysis” means that they refer to different subsets: a from it the structural index N:
subset of multiridges, R1, identified by the zeros of the horizontal
derivatives of the field at different altitudes; a subset of multiridges, N = –n – q, (2)

Figure 3. (a) Aeromagnetic map: the black rectangle indicates the study area, and the black dashed line surrounds the Pine Island Rift (PIR), the Byrd Subglacial Basin (BSB), the Bentley
Subglacial Trench (BST), and the Ellsworth-Whitmore Mountains (EWM) structures. The red line indicates the location of the analyzed profile in Figure 3b. Red triangles mark the locations
of known subaerial Cenozoic volcanoes. (b) First-order vertical derivative of the magnetic field data. (c) Multiridge analysis with upward continuation at 20 km with a 1 km step. Ridges
corresponding to the zeros of the horizontal field derivative and to the zeros of the field vertical derivative are indicated by cyan and yellow lines, respectively. The geometric approach
consists of fitting the ridges by straight lines, prolonging them into the source region (red lines), and identifying the sources at their intersection. The estimated depth refers to the
topography (i.e., estimated depth is 5 km). (d) ScalFun method chart. The degree of homogeneity n is given at the intersection of the scaling function with the vertical axis. The structural
index N is computed through equation 2.

220 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
 where q is the field differentiation order with respect to the derivatives. We selected 287 solutions, of which we estimated the
magnetic field or the gravity first-order derivative (e.g., Milano magnetic source depths and types (through the structural index,
et al., 2016; Paoletti et al., 2022). If N is an integer, it is also N). All the found depths to the source do not exceed the Curie
constant versus the altitude, and the field obeys a global homogene- depth, as modeled in Dziadek et al. (2021).
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

ity law. In this case, N is a useful quantity because it defines The maps in Figure 4 indicate the solutions found with the
different source types, such as the contact (N = 0), the sill or dike multiridge analysis for which the scaling function determined a
(N = 1), the horizontally infinite cylinder and the vertically infinite structural index between 0 and 1. The identified solutions were
cylinder (N = 2), and the sphere (N = 3). Note that the depth to 152 out of a total of 287, at varying depths down to 15 km. When
the source z 0 will correspond to the center (sphere, horizontal 0 < N < 1, we are analyzing the field from contact-like sources
cylinder, sill) or to the top (vertical cylinder, dike, contact) of the and/or faults with a finite throw. These solutions are located in
source, depending on the value of N. some key areas of the study, which we indicate in Figures 4–6 by
However, n, and so N, may be either integer or fractional. ellipses numbered 1 to 7. The first one (ellipse 1) regards Pine
When n is fractional, which applies to complex sources, the field Island Rift (Figures 4 and 7), a sector rich in volcanoes and
is inhomogeneous and n changes with the altitude. Inhomogeneity characterized by a high Neogenic magmatism. This area is char-
means that the homogeneity holds only locally — that is, at any acterized by an articulated sub-ice morphology (as shown in
altitude there is a homogenous field with fractional n approximat- Figures 4b and 7) corresponding to the tributaries (i.e., valleys
ing the field (Fedi et al., 2015). Thus, in case of local homogeneity, carved out by the movement of ice) of the rift, with north–south
n(z) and z 0 (z) will relate at each altitude to a different point of and northwest–southeast directions.
the source, and we will need all the n and z 0 values at different The second area (ellipse 2) is between Pine Island and the
altitudes to fully characterize the source. For more complete Byrd Basin northern edge (Figure 4), where the topography has
information, refer to Fedi et al. (2015) in which the authors a horst and graben configuration (Figure 4b) and the sector is
describe the cases of the inhomogeneous fields generated by sources affected by the last tectonic phase, the Cenozoic extension. Indeed,
such as disk, fault, and finite cylinder. a northeast–southwest trend of contact-type solutions can be
DOI:10.1190/leedff.2024.43.issue-4

noted (see also Figures 7 and 8).

Results and discussion The third area (ellipse 3) is localized on the northwest part
The multiridge analysis was performed on 89 profiles extracted of the Byrd Basin edge (Figure 4) in which there are geologic
from the aeromagnetic data, within the area shown in Figure 2 features developed during the formation of this passive rift during
(black rectangle). The profiles used to analyze the source depth the Mesozoic (Figure 4b). We note an east–west alignment of
and estimate N are not relative to the orientation of flight lines contact-type solutions, with depths varying from 5 to 10 km,
but are chosen on the map of the magnetic field, according to the parallel to the basin edge.
anomaly trends. We considered only the source depths matching The fourth area (ellipse 4) is between the Byrd Basin and Bentley
each other on orthogonal profiles. The analysis was carried out Trench, also containing the Sinuous Area (Figure 4), a sector
on the magnetic field data and on their first- and second-order defined by an elevated topography compared to the lateral rifts

Figure 4. (a) Aeromagnetic map: the black rectangle indicates the study area, and the black dashed line surrounds the Pine Island Rift (PIR), the Byrd Subglacial Basin (BSB), the Bentley
Subglacial Trench (BST), and the Ellsworth-Whitmore Mountains (EWM) structures. In this map, we represent the solutions of the multiridge analysis (black circles) with 0 < N < 1. The
circle size is proportional to the calculated source depth. (b) Sub-ice topography with the solutions of the multiridge analysis (black circles) with 0 < N < 1. Numbers from 1 to 5 and related
red ellipses indicate the areas commented in the study. Red triangles mark the locations of known subaerial Cenozoic volcanoes.

Special Section: Gravity, electrical, and magnetic methods April 2024 The Leading Edge 221
 (Byrd Basin and Bentley Trench), strongly dependent to the orienta- of 287 identified. Even for this type of source, there is a high
tion of the last Cenozoic rifting phase, with a northeast–southwest- concentration of these solutions in the first area (ellipse 1; Figures 4,
articulated morphology. The solutions identified for 0 < N < 1 follow 5, and 8). Indeed, in this area, high magmatism and shallow
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

this trend, with mainly shallow magnetic sources (Figure 7). intrusive bodies occur, characterizing dike-type sources (Figure 5a),
Finally, the fifth area (ellipse 5) is located in the continental which may have a control over the tributary systems.
slope of the Ellsworth-Whitmore Mountains block toward the The third area (ellipse 3; Figures 4 and 5), as described before,
Bentley Trench (Figure 4) with deep continental structures. and the sixth (ellipse 6), corresponding to the northern and
The maps in Figure 5 indicate the found solutions by the scaling southern Byrd Basin edges, are interesting because of the crustal
function method, for 1 < N < 2. The identified solutions are 36 out thinning, which in this sector reaches minimum levels (Jordan
DOI:10.1190/leedff.2024.43.issue-4

Figure 5. (a) Aeromagnetic map: the black rectangle indicates the study area, and the black dashed line surrounds the Pine Island Rift (PIR), the Byrd Subglacial Basin (BSB), the Bentley
Subglacial Trench (BST), and the Ellsworth-Whitmore Mountains (EWM) structures. In this map, we represent the solutions of the multiridge analysis (red circles) with 1 < N < 2. The circle
size is proportional to the source depth calculated. (b) Sub-ice topography with the solutions of the multiridge analysis (red circles) with 1 < N < 2. Numbers from 1 to 6 and related red
ellipses indicate the areas commented in the study. Red triangles mark the locations of known subaerial Cenozoic volcanoes.

Figure 6. (a) Aeromagnetic map: the black rectangle indicates the study area, and the black dashed line surrounds the Pine Island Rift (PIR), the Byrd Subglacial Basin (BSB), the Bentley
Subglacial Trench (BST), and the Ellsworth-Whitmore Mountains (EWM) structures. In this map, we represent the solutions of the multiridge analysis with 2 < N < 3, magenta circles, and
with N = 3, white circles. The circle size is proportional to the source depth calculated. (b) Sub-ice topography with the solutions of the multiridge analysis with 2 < N ≤ 3. Numbers from 1
to 7 and related red ellipses indicate the areas commented in the study. Red triangles mark the locations of known subaerial Cenozoic volcanoes.

222 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
 et al., 2010), and because of the weakening of the crust character- ridges. The solutions with N = 3 (white circles), 11 out of 287,
ized by two long overlapping rift phases (Mesozoic-Cenozoic). indicate the center of spherically distributed sources. The most
Furthermore, we may identify both superficial and deep magmatic interesting area is again the first area (ellipse 1; Figures 4–8),
bodies, which could intrude directly from the upper part of the which regards Pine Island Rift. Indeed, within this sector, the
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

mantle (as shown also in Figure 8). seventh (ellipse 7) area indicates a deep, extended, high-amplitude
Maps in Figure 6 show the solutions estimated for 2 < N < 3 magnetic source, which is located near the southern edge of the
(magenta circles), with 88 out of 287 solutions indicating the rift, corresponding to West Antarctic Ice Sheet thinning and
top or center of cylindrical linear sources, such as sill, pipes, or warm ocean inclusion.
DOI:10.1190/leedff.2024.43.issue-4

Figure 7. (a) Sub-ice topography with all the solutions of the multiridge analysis (black circles). Red triangles mark the locations of known subaerial Cenozoic volcanoes. The circle size is
proportional to the calculated source depth. We analyze the profile from A to E (orange line): (b) magnetic field and (c) sub-ice topography and solutions of the multiscale analysis (star
markers). The star colors refer to the values of the structural index (black 0 < N < 1, red 1 < N < 2, magenta 2 < N < 3), respectively.

Figure 8. (a) Sub-ice topography with all the solutions of the multiridge analysis (black circles). The red triangles mark the locations of known subaerial Cenozoic volcanoes. The circle size
is proportional to the calculated source depth. We analyze the profile from F to G (orange line): (b) magnetic field and (c) sub-ice topography and solutions of the multiscale analysis (star
markers). The star colors refer to the values of the structural index (black 0 < N < 1, red 1 < N < 2, magenta 2 < N < 3, gray N = 3), respectively.

Special Section: Gravity, electrical, and magnetic methods April 2024 The Leading Edge 223
 Even the third area (ellipse 3; Figures 4–6), close to the Corresponding author: giuseppe.ferrara@unina.it
northern Byrd Basin edge, shows a remarkable interest: it is
generated by magmatic structures related to the Neogene rifting References
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

phase (LeMasurier, 2008; Jordan et al., 2010) that led to the Behrendt, J. C., 1999, Crustal and lithospheric structure of the West
formation of the Byrd Basin. Furthermore, we identified high- Antarctic Rift System from geophysical investigations — A review:
Global and Planetary Change, 23, no. 1–4, 25–44, https://doi.
amplitude magnetic anomalies with an east–west orientation
org/10.1016/S0921-8181(99)00049-1.
(Figure 6a) with relatively deeper solutions, for 2 < N < 3 (see also Behrendt, J. C., W. E. LeMasurier, A. K. Cooper, F. Tessensohn, A.
Figures 7 and 8). Trehu, and D. Damaske, 1991, Geophysical studies of the West
For the sake of clarity, we show in Figures 7 and 8 two Antarctic Rift System: Tectonics, 10, no. 6, 1257–1273, https://doi.
magnetic field profiles, and the estimated depth solutions, inter- org/10.1029/91TC00868.
cepting the most important structures in the area, such as Pine Behrendt, J. C., R. Saltus, D. Damaske, A. McCafferty, C. A. Finn, D.
Island, Byrd Basin, Sinous and Bentley Trench areas (Figure 7), Blankenship, and R. E. Bell, 1996, Patterns of late Cenozoic volcanic
and tectonic activity in the West Antarctic Rift System revealed by
and the northern and western sectors of the Byrd Basin (Figure 8).
aeromagnetic surveys: Tectonics, 15, no. 3, 660–676, https://doi.
In the Pine Island area (northern part, profile A-B in Figure 7), org/10.1029/95TC03500.
contact-type sources could define a faulted basement deepening Bingham, R., F. Ferraccioli, E. C. King, R. D. Larter, H. D. Pritchard,
southward, while dike-type intrusions seem to be shallower A. M. Smith, and D. G. Vaughan, 2012, Inland thinning of West
southward (profiles C-D, D-E in Figure 7). In Figure 8, isolated Antarctic Ice Sheet steered along subglacial rifts: Nature, 487, no.
sources prevail, with different Structural Index, thus not defining 7408, 468–471, https://doi.org/10.1038/nature11292.
specific structural trends. Blakely, R. J., 1996, Potential theory in gravity and magnetic applica-
tions: Cambridge University Press, https://doi.org/10.1017/
CBO9780511549816.
Conclusion Cooper, A. K., P. J. Barrett, K. Hinz, V. Traube, G. Letichenkov,
In this study we provide a new geologic interpretation of the and H. M. J. Stagg, 1991, Cenozic prograding sequences of the
Antarctic region extending from Pine Island to the Ellsworth Antarctic continental margin: A record of glacio-eustatic and
DOI:10.1190/leedff.2024.43.issue-4

Mountains area, obtained with a multiscale analysis of potential- tectonic events: Marine Geology, 102, no. 1–4, 175–213, https://
field data. Out of a total of 89 profiles defined in the study area, doi.org/10.1016/0025-3227(91)90008-R.
we identified 287 solutions for the sources within a 15 km depth. Cooper, A. K., F. J. Davey, and J. C. Behrendt, 1987, Seismic stratigraphy
and structure of the Victoria Land Basin, Western Ross Sea,
More than half of the identified anomalies (152) refer to contact-
Antarctica: Circum Pacific Council Publications, 58.
like and fault sources, 80 of them are related to dike-like sources, Corr, H., F. Ferraccioli, and D. Vaughan, 2021, Processed airborne
44 solutions to cylinder-like sources (i.e., conduits or volcanic radio-echo sounding data from the BBAS survey covering the Pine
necks), and 11 to spherical ones. Island Glacier Basin, West Antarctica (2004/2005) (Version 1.0);
These results allow a new regional interpretation. In particular, [Data set]: NERC EDS UK Polar Data Centre, https://doi.
the contact-type sources (0 < N < 1) may be associated mainly org/10.5285/db8bdbad-6893-4a77-9d12-a9bcb7325b70.
with faults corresponding to the tributaries related to Pine Island Dalziel, I. W. D., 1992, Antarctica; a tale of two supercontinents?:
Annual Review of Earth and Planetary Sciences, 20, 501–526, https://
Rift and bordering the major rift structures in the area, such as
doi.org/10.1146/annurev.ea.20.050192.002441.
Pine Island, Byrd Basin, and Bentley Trench. As regards planar Dalziel, I. W. D., and A. M. Grunow, 1992, Late Gondwanide tectonic
and linear surfaces (1 < N < 2 and 2 < N < 3), they can represent rotations within Gondwanaland: Tectonics, 11, no. 3, 603–606,
magmatic sources typical of rifting zones, based on their shape https://doi.org/10.1029/91TC02365.
and orientation, but also characterizing the sub-ice topography. Dalziel, I. W. D., and L. A. Lawver, 2001, The lithospheric setting of
In addition, the source-types maps allowed us to locate magmatism the West Antarctic Ice Sheet, in R. B. Alley and R. A. Bindschadle,
trends defining their depths. eds., The West Antarctic Ice Sheet: Behavior and Environment:
AGU, Antarctic Research series: 77, 29–44, https://doi.org/10.1029/
Lastly, by integrating these types of results, we had the oppor-
AR077p0029.
tunity to interpret the extent of the areas affected by magmatism Dziadek, R., F. Ferraccioli, and K. Gohl, 2021, High geothermal heat
in the West Antarctic Rift System. This work opens up many flow beneath Thwaites Glacier in West Antarctica inferred from
interesting scenarios because it keeps open the debate on the depth aeromagnetic data: Communications Earth & Environment, 2, 162,
extension of magmatic sources in this sector of Antarctica, and https://doi.org/10.1038/s43247-021-00242-3.
their contribution to the sub-ice articulated topography of this Elliot, D. H., 1992, Jurassic magmatism and tectonism associated with
sector of the West Antarctic Rift System. Gondwanaland break-up: An Antarctic perspective: Geological
Society of London, Special Publications, 68, no. 1, 165–184, https://
In conclusion, inaccessible areas can be explored through
doi.org/10.1144/GSL.SP.1992.068.01.11.
potential-fields analysis, particularly with tools that characterize Fedi, M., G. Florio, and V. Paoletti, 2015, MHODE: A local-homogeneity
complex geologic structures well in a multiscale framework, such theory for improved source-parameter estimation of potential field:
as the multiridge method. Joint analysis with other kinds of data Geophysical Journal International, 202, no. 2, 887–900, https://doi.
can allow a more complete geologic description of the area. org/10.1093/gji/ggv185.
Fedi, M., G. Florio, and T. A. M. Quarta, 2009, Multiridge analysis of
Data and materials availability potential fields: Geometric method and reduced Euler deconvolution:
Geophysics, 74, no. 4, L53–L65, https://doi.org/10.1190/1.3142722.
Data associated with this research are available and can be
Ferraccioli, F., T. A. Jordan, E. Armadillo, E. Bozzo, H. F. J. Corr, G.
obtained by contacting the corresponding author. Caneva, C. N. Robinson, N. P. Frearson, and I. Tabacco, 2007a,

224 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
 Collaborative aerogeophysical campaign targets the Wilkes Subglacial Larter, D. R., A. P. Cunningham, P. F. Barker, K. Gohl, and F. O.
Basin, the Transantarctic Mountains and the Dome C region: Terra Nitsche, 2002, Tectonic evolution of the Pacific margin of Antarctica:
Antarctica Reports, 13, 1–36. Late Cretaceous tectonic reconstructions: Journal of Geophysical
Ferraccioli, F., T. A. Jordan, D. G. Vaughan, J. Holt, M. James, H. Corr, Research, Solid Earth, AGU, 107, no. B12, EPM 5-1–EPM 5-19,
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

D. D. Blankenship, J. D. Fairhead, and T. M. Diehl, 2007b, New https://doi.org/10.1029/2000JB000052.

aerogeophysical survey targets the extent of the West Antarctic Rift LeMasurier, W. E., 1990, Late Cenozoic volcanism on the Antarctic
System over Ellsworth Land: Proceedings of the 10th International Plate: An overview, in W. E. LeMasurier, J. W. Thomson, P. E.
Symposium of Antarctic Earth Sciences. Baker, P. R. Kyle, P. D. Rowley, J. L. Smellie, and W. J. Verwoerd,
Florio, G., M. Fedi, and A. Rapolla, 2009, Interpretation of regional eds., Volcanoes of the Antarctic Plate and Southern Oceans: AGU,
aeromagnetic data by the scaling function method: The case of Antarctic Research series, 48, 1–17, https://doi.org/10.1029/
Southern Apennines (Italy): Geophysical Prospecting, 57, no. 4, AR048p0001.
479–489, https://doi.org/10.1111/j.1365-2478.2009.00807.x. LeMasurier, W. E., 2006, What supports the Marie Byrd Land Dome?
Frémand, A. C., J. A. Bodart, T. A. Jordan, F. Ferraccioli, C. Robinson, An evaluation of potential uplift mechanisms in a continental rift
H. F. J. Corr, H. J. Peat, R. G. Bingham, and D. G. Vaughan, 2022, system, in D. K. Fütterer, D. Damaske, G. Kleinschmidt, H. Miller,
British Antarctic Survey’s aerogeophysical data: Releasing 25 years and F. Tessensohn, eds., Antarctica: Springer, 299–302, https://doi.
of airborne gravity, magnetic, and radar datasets over Antarctica: org/10.1007/3-540-32934-X_37.
Earth System Science Data, 14, no. 7, 3379–3410, https://doi. LeMasurier, W. E., 2008, Neogene extension and basin deepening in
org/10.5194/essd-14-3379-2022. the West Antarctic Rift inferred from comparisons with the East
Fretwell, P., H. D. Pritchard, D. G. Vaughan, J. L. Bamber, N. E. African rift and other analogs: Geology, 36, no. 3, 247–250, https://
Barrand, R. Bell, C. Bianchi, et al., 2013, Bedmap2: Improved ice doi.org/10.1130/G24363A.1.
bed, surface and thickness datasets for Antarctica: The Cryosphere, Lindow, J., P. J. J. Kamp, S. B. Mukasa, M. Kleber, F. Lisker, K. Gohl,
7, no. 1, 375–393, https://doi.org/10.5194/tc-7-375-2013. G. Kuhn, and C. Spiegel, 2016, Exhumation history along the eastern
Granot, R., S. C. Cande, J. M. Stock, and D. Damaske, 2013, Revised Amundsen Sea coast, West Antarctica, revealed by low-temperature
Eocene-Oligocene kinematics for the West Antarctic Rift System: thermochronology: Tectonics, 35, no. 10, 2239–2257, https://doi.
Geophysical Research Letters, 40, no. 2, 279–284, https://doi. org/10.1002/2016TC004236.
org/10.1029/2012GL054181. Lloyd, A. J., D. A. Wiens, H. Zhu, J. Tromp, A. A. Nyblade, R. C.
DOI:10.1190/leedff.2024.43.issue-4

Grunow, A. M., D. V. Kent, and I. W. D. Dalziel, 1987, Mesozoic evolu- Aster, S. E. Hansen, et al., 2020, Seismic structure of the Antarctic
tion of West Antarctica and the Weddell Sea Basin: New paleomag- upper mantle imaged with adjoint tomography: JGR: Solid Earth,
netic constraints: Earth and Planetary Science Letters, 86, no. 1, 125, no. 3, https://doi.org/10.1029/2019JB017823.
16–26, https://doi.org/10.1016/0012-821X(87)90184-1. Lucas, E. M., D. Soto, A. A. Nyblade, A. J. Lloyd, R. C. Aster, D. A.
Grunow, A. M., D. V. Kent, and I. W. D. Dalziel, 1991, New paleo- Wiens, J. P. O’Donnell, et al., 2020, P- and S-wave velocity structure
magnetic data from Thurston Island: Implications for the tectonics of central West Antarctica: Implications for the tectonic evolution
of West Antarctica and Weddell Sea opening: Journal of Geophysical of the West Antarctic Rift System: Earth and Planetary Science
Research: Solid Earth, 96, no. B11, 17935–17954, https://doi. Letters, 546, 116437, https://doi.org/10.1016/j.epsl.2020.116437.
org/10.1029/91JB01507. Milano, M., M. Fedi, and J. D. Fairhead, 2016, The deep crust beneath
Jordan, T. A., F. Ferraccioli, and P. T. Leat, 2017, New geophysical the Trans‐European Suture Zone from a multiscale magnetic model:
compilations link crustal block motion to Jurassic extension and Journal of Geophysical Research. Solid Earth, 121, no. 9, 6276–6292,
strike-slip faulting in the Weddell Sea Rift System of West Antarctica: https://doi.org/10.1002/2016JB012955.
Gondwana Research, 42, 29–48, https://doi.org/10.1016/j. Milano, M., M. Fedi, and J. D. Fairhead, 2019, Joint analysis of the
gr.2016.09.009. magnetic field and total gradient intensity in central Europe: Solid
Jordan, T. A., F. Ferraccioli, D. G. Vaughan, J. W. Holt, H. Corr, D. D. Earth, 10, no. 3, 697–712, https://doi.org/10.5194/se-10-697-2019.
Blankenship, and T. M. Diehl, 2010, Aerogravity evidence for major Milano, M., Y. Kelemework, M. La Manna, M. Fedi, D. Montanari,
crustal thinning under the Pine Island Glacier region (West and M. Iorio, 2020, Crustal structure of Sicily from modelling of
Antarctica): Geological Society of America Bulletin, 122, no. 5-6, gravity and magnetic anomalies: Scientific Reports, 10, 16019, https://
714–726, https://doi.org/10.1130/B26417.1. doi.org/10.1038/s41598-020-72849-z.
Jordan, T. A., T. R. Riley, and C. S. Siddoway, 2020, The geological history Morlighem, M., E. Rignot, T. Binder, D. Blankenship, R. Drews, G.
and evolution of West Antarctica: Nature Reviews Earth & Environment, Eagles, O. Eisen, et al., 2020, Deep glacial troughs and stabilizing
1, 117–133, https://doi.org/10.1038/s43017-019-0013-6. ridges unveiled beneath the margins of the Antarctic Ice Sheet:
Jordan, T. A., S. Thompson, B. Kulessa, and F. Ferraccioli, 2023, Nature Geoscience, 13, no. 2, 132–137, https://doi.org/10.1038/
Geological sketch map and implications for ice flow of Thwaites s41561-019-0510-8.
Glacier, West Antarctica, from integrated aerogeophysical observa- Mukasa, S. B., and I. W. D. Dalziel, 2000, Marie Byrd Land, West
tions: Science Advances, 9, no. 22, eadf2639, https://doi.org/10.1126/ Antarctic: Evolution of Gondwana’s Pacific margin constrained by
sciadv.adf2639. zircon U-Pb geochronology and feldspar common-Pb isotopic com-
Karner, G. D., M. Studinger, and R. E. Bell, 2005, Gravity anomalies positions: Geological Society of America Bulletin, 112, no. 4, 611–627,
of sedimentary basins and their mechanical implications: https://doi.org/10.1130/0016-7606(2000)112<611:MBLWAE>2.0
Applications to the Ross Sea basins, West Antarctica: Earth and .CO;2.
Planetary Science Letters, 235, no. 3-4, 577–596, https://doi. Nitsche, F. O., S. S. Jacobs, R. D. Larter, and K. Gohl, 2007, Bathymetry
org/10.1016/j.epsl.2005.04.016. of the Amundsen Sea continental shelf: Implications for geology,
Kelemework, Y., M. Milano, M. La Manna, G. de Alteriis, M. Iorio, oceanography, and glaciology: Geochemistry, Geophysics, Geosystems,
and M. Fedi, 2021, Crustal structure in the Campanian region 8, no. 10, https://doi.org/10.1029/2007GC001694.
(Southern Apennines, Italy) from potential field modelling: Scientific O’Donnell, J. P., G. W. Stuart, A. M. Brisbourne, K. Selway, Y. Yang,
Reports, 11, 14510, https://doi.org/10.1038/s41598-021-93945-8. G. A. Nield, P. L. Whitehouse, et al., 2019, The uppermost mantle

Special Section: Gravity, electrical, and magnetic methods April 2024 The Leading Edge 225
 seismic velocity structure of West Antarctica from Rayleigh wave Schmidt, D. L., and P. D. Rowley, 1986, Continental rifting and transform
tomography: Insights into tectonic structure and geothermal heat faulting along the Jurassic Transantarctic Rift, Antarctica: Tectonics,
flow: Earth and Planetary Science Letters, 522, 219–233, https:// 5, no. 2, 279–291, https://doi.org/10.1029/TC005i002p00279.
doi.org/10.1016/j.epsl.2019.06.024. Schopf, J. M., 1969, Ellsworth Mountains: Position in West Antarctica
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Olesen, O., M. Brönner, and J. Ebbing, 2010, New aeromagnetic and due to seafloor spreading: Science, 164, no. 3875, 63–66, https://doi.
gravity compilations from Norway and adjacent areas: Methods and org/10.1126/science.164.3875.63.
applications: Geological Society, London, Petroleum Geology Siddoway, C. S., 2008, Antarctica: A keystone in a changing world:
Conference Series, 7, 559–586, https://doi.org/10.1144/0070559. USGS National Research Council Polar Research Board.
Paoletti, V., E. Hintersberger, I. Schattauer, M. Milano, G. P. Deidda, Storey, B. C., 1996, Microplates and mantle plumes in Antarctica: Terra
and R. Supper, 2022, Geophysical study of the Diendorf-Boskovice Antarctica, 3, no. 2, 91–102.
Fault System (Austria): Remote Sensing, 14, no. 8, 1807, https://doi. Storti, F., M. L. Balestrieri, F. Balsamo, and F. Rossetti, 2008, Structural
org/10.3390/rs14081807. and thermochronological constraints to the evolution of the West
Paoletti, V., M. Milano, J. Baniamerian, and M. Fedi, 2020, Magnetic Antarctic Rift System in central Victoria Land: Tectonics, 27, no. 4,
field imaging of salt structures at Nordkapp Basin, Barents Sea: 2006TC002066, https://doi.org/10.1029/2006TC002066.
Geophysical Research Letters, 47, no. 18, e2020GL089026, https:// Tessensohn, F., 1994, Geological contributions from Antarctica, in G.
doi.org/10.1029/2020GL089026. Hempel, ed., Antarctic Science: Springer, 215–228, https://doi.
Randall, D. E., and C. Mac Niocaill, 2004, Cambrian palaeomagnetic org/10.1007/978-3-642-78711-9_14.
data confirm a Natal Embayment location for the Ellsworth– Vaughan, D. G., H. F. J. Corr, F. Ferraccioli, N. Frearson, A. O’Hare,
Whitmore Mountains, Antarctica, in Gondwana reconstructions: D. Mach, J. W. Holt, D. D. Blankenship, D. L. Morse, and D. A.
Geophysical Journal International, 157, no. 1, 105–116, https://doi. Young, 2006, New boundary conditions for the West Antarctic Ice
org/10.1111/j.1365-246X.2004.02192.x. Sheet: Subglacial topography beneath Pine Island Glacier: Geophysical
Ross, N., R. Bingham, H. F. J. Corr, F. Ferraccioli, T. A. Jordan, A. Research Letters, 33, no. 9, 2005GL025588, https://doi.
Le Brocq, D. M. Rippin, D. Young, D. D. Blankenship, and M. J. org/10.1029/2005GL025588.
Siegert, 2012, Steep reverse bed slope at the grounding line of the Winberry, J. P., and S. Anandakrishnan, 2004, Crustal structure of the
Weddell Sea sector in West Antarctica: Nature Geoscience, 5, no. West Antarctic Rift System and Marie Byrd Land hotspot: Geology,
6, 393–396, https://doi.org/10.1038/ngeo1468. 32, no. 11, 977–980, https://doi.org/10.1130/G20768.1.
DOI:10.1190/leedff.2024.43.issue-4

2 02 4
Middle East & Africa Events
23–25 Apr – Muscat, Oman 21–24 Oct – Al Khobar, Saudi Arabia 10–12 Dec – Abu Dhabi, UAE
Next Generation Geophysical Land AAPG/EAGE/SEG: GEO 4.0 – Addressing the Challenges in
and Shallow Water Acquisition Digitalization in Geosciences Processing Data for Land and
Registration open Symposium Shallow Water in the Middle East
Call for abstracts closes 1 May Call for abstracts opens soon
21–23 May – Abu Dhabi, UAE
Recent Advances in Geophysical 5–7 Nov – Muscat, Oman
Reservoir Characterization and Challenges and Advances in
Monitoring of CO2 Sequestration in Velocity Model Building
Carbonate Reservoirs Call for abstracts closes 24 Jun
Early bird savings end 20 Apr
19–21 Nov –
8–10 Oct – Al Khobar, Saudi Arabia Al Khobar, Saudi Arabia
SEG | SPLWA Symposium: Deep Reservoir Imaging Current information as of 14 March.
Seismic Petrophysics and Characterization Please check seg.org/events
Call for abstracts closes 1 May Call for abstracts closes 24 Jun for up-to-date information.

226 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
 2024
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Asia-Pacific Events
23–24 Apr – Kuala Lumpur, Malaysia Offshore Technology Conference
Borehole Geophysics:
New Developments for Global Energy
JOIN THE CONVERSATION ON THE PROGRESS
19–22 May – Shenzhen, China OF FUTURE OFFSHORE PROJECTS
GEM 2024 Shenzhen: International Workshop on
Gravity, Electrical, and Magnetic Methods and
Their Applications Be at the Forefront of
Energy Technologies and Innovations
3–5 Jun – Korla, China
1st Tarim Ultra-Deep Oil and Gas Exploration
Technology Workshop REGISTER TO AT TEND
23–27 Jun – Changchun, China
20th International Conference on Ground 6–9 May 2024 NRG Park Houston, Texas, USA
Penetrating Radar (GPR 2024)
3–5 Jul – Kuala Lumpur, Malaysia #OTC2024
DOI:10.1190/leedff.2024.43.issue-4

Advances in Geophysical Solutions for

Complex Geological Settings
Current information as of 21 March. Please check
seg.org/events for up-to-date information and more events.

DIGITALIZATION IN
GEOSCIENCES SYMPOSIUM
Reservoir characterization Shaping Innovation for
a Sustainable Future
With rapid advances in computing power, data integration, machine
learning, and 3D seismic analysis, reservoir modeling is transforming
SHARE 21-24 OCTOBER 2024

YOUR
AL KHOBAR, SAUDI ARABIA
subsurface analysis across energy sources and solutions, from
hydrocarbon exploration to geothermal energy and carbon sequestration.
The editors of this special section welcome submissions highlighting
DISCOVERIES
innovative workflows and real-world applications of reservoir charac- SUBMIT ABSTRACTS BY
terization for all aspects of energy, including carbon capture, hydrogen 01 MAY 2024
storage, and other sectors critical to realizing net-zero carbon goals.
Share your research on using reservoir models for renewable energy,
next-generation characterization techniques, and the expanding role of
this multidisciplinary field in building a sustainable future.

Submission deadline: 15 July 2024

Publication of issue: December 2024
Submission portal: https://mc.manuscriptcentral.com/tle
Information for authors: https://library.seg.org/page/leedff/ifa

Special section editors: SCAN THE QR CODE TO SUBMIT

Heather Bedle (coordinator): hbedle@ou.edu
Satinder Chopra: satinder.chopra@samigeo.com
HOSTED BY P R I N C I PA L S P O N S O R ORGANISING SOCIETIES

Tom Davis: tdavis@mines.edu

Special Section: Gravity, electrical, and magnetic methods April 2024 The Leading Edge 227
 Analytic continuation: A tool for aeromagnetic data interpretation
Jeffrey B. Thurston1 and Bengt Fornberg 2
https://doi.org/10.1190/tle43040228.1
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Abstract Cartesian coordinates (Figure 1). He went on to show that airborne

Extending potential-field data to vertical complex-plane slices data are well described by the Cauchy-Riemann differential equa-
provides opportunities, unique to analytic functions, for subsurface tions. This led to introduction of the modulus of complex-valued
imaging. We begin with an existing numerical method for dif- data (now widely known as the analytic signal) as a means for
ferentiation by integration using Cauchy’s integral formula. An reducing ambiguity owing to nonvertical sources and inducing fields.
example from data over northern Ontario, Canada, illustrates Proceeding in the complex plane, we find that high-order and rapidly
better conditioning compared to derivatives computed using finite convergent analytic power-series expansions improve the accuracy
differences. Next, these quantities are used for analytic continuation and stability of downward continued data. These provide crisp,
using rational complex-series expansion (Padé approximation). unambiguous images of complex-plane singularities.
This method provides stable downward continuation to distances
that are significantly greater than are achievable by Taylor series
and spectral methods. A synthetic test confirms that the rational
approximant faithfully reproduces the magnetic response of a thin
sheet. A sign reversal of the total field, coincident with the origin
of the sheet, features prominently on this example. We also provide
synthetic results for sheets with finite depth extent, point sources,
thin slabs, and contacts. These also exhibit polarity flips near the
DOI:10.1190/leedff.2024.43.issue-4

source location. Further, these provide a template for interpreting

the source geometry using the shape of the approximating function.
Additional synthetic tests illustrate an automatic method for
locating closely spaced random assemblages of isolated poles. We
apply our methods to a deposit-scale heliborne survey over a
recently discovered volcanogenic massive sulphide ore deposit in
a Canadian greenstone belt. The analytically continued profile
over the deposit suggests the presence of a subvertical thin sheet
of unknown depth extent. This interpretation coincides with a
sulphide ore body that has been previously delimited by several
drill holes. Some known prospects near this deposit are also
imaged on maps and profiles of the Padé approximant.

Introduction
Wang et al. (2023) provide a comprehensive review of the vast
inventory of methods for downward continuation. A recurring
theme among most of the surveyed techniques is the limited
extent to which experimental data can robustly identify the sources.
A notable departure from this conventionally accepted limitation
is described by Fedi and Florio (2011), who used normalized full
gradients for continuing into a quasi-harmonic region. Accelerated
Figure 1. Illustration of the theory and method for differentiation using integration. Panel
series-convergence methods have also received attention recently (a) shows how an airborne profile may be cast in the complex plane. The imaginary (vertical)
as a means for extended downward continuation. For example, axis is the sensor’s depth below the survey datum, and the real axis is its horizontal
Zhou et al. (2021) and Chong et al. (2022) use real-valued, position. Cauchy’s integral formula specifies that derivatives are proportional to the
comparatively low-order, power series. Analytic-function theory coefficients of an integral transform around the periphery of C (the method is described in
offers opportunities to enhance this approach. detail in Fornberg, 1981). Panel (b) provides the framework for using the integral transform.
In the interest of efficiency, we use the fast Fourier transform for this task. Alternatives,
Nabighian (1972) was among the first to cast potential-field data such as trapezoidal quadrature methods, are discussed by Fornberg (1981). Note, the
as an analytic function of the complex independent variable, z = x + iy, path C is restricted to the upper half plane to avoid poor conditioning resulting from
where x and y are, respectively, (real-valued) horizontal and vertical interpolation points continued below the plane by spectral downward continuation.

Manuscript received 29 November 2023; revision received 14 February 2024; accepted 19 February 2024.
1
Calgary, Alberta, Canada. E-mail: jbthurston@gmail.com.
2
University of Colorado Boulder, Boulder, Colorado, USA. E-mail: bengt.fornberg@colorado.edu.

228 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
 Cauchy’s integral formula: High-order differentiation personal communication, 2024). Fornberg (1981) adapted this
Derivatives of potential-field data are typically computed theory to a framework suited to typical airborne geophysical
using either finite-difference (FD) or spectral methods. The FD experiments. Details, including source code, have been left to
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

method solves a linear system for stencil coefficients that are then this reference. Aided by the schematic in Figure 1, we summarize
convolved with the signal to compute the derivatives. FD approxi- our implementation.
mations are surveyed in Fornberg and Flyer (2015, Chapter 1). For an analytic function f (z), Cauchy’s integral formula pro-
Spectral methods differentiate an approximate Fourier series. vides its N th order derivative at z 0:
These are described in Blakely (1996) and, from an FD perspective,
N! f ​(ξ)​
in Fornberg and Flyer (2015, Chapter 2). Both techniques are ​f​N​(​z0​ ​)​ = ​_ ​​∮ ​ ​_​dξ​​.
2π C ​(ξ − ​z​ ​)​ ​
N+1 (1)
0

applicable to many forms of digital input and are well known to

signal processors in a variety of disciplines. It is also well known Thus, if a signal, f (z), is known on the periphery of a circle C,
that these classical methods, when used for evaluating high-order numerical integration around it (Figure 1b) gives a result pro-
derivatives, are ill-conditioned. However, because gravity and portional to a specified derivative order N at the circle’s center,
magnetic signals are harmonic functions, they can be well approxi- z 0 . Because the integrand is smooth and periodic, the trapezoidal
mated by analytic functions in vertical complex planes. Then, rule is spectrally accurate. We align the real axis, along which
Cauchy’s integral formula can be used to improve conditioning y = 0, with the survey datum. Apart from the signal recorded
of high-order derivatives. It seems the earliest published method along the real axis, we also require signal along several profiles
for numerical differentiation of analytic functions is described by for y ≠ 0. These data, off the real axis, are obtained with spectral
Abate and Dubner (1968). At about the same time, Voskoboynikov continuation. In principle, we could continue equal distances
and Nachapkin (1969) published, in Russian, a method also based both upward and downward, which would allow z 0 to lie on the
on analytic functions of complex variables (Roman Pasteka, real axis. However, downward continuation in the Fourier domain
is quite ill-conditioned and often requires some measure of
regularization. We have achieved the best results by preserving
DOI:10.1190/leedff.2024.43.issue-4

bandwidth prior to contour integration. So, we first upward

continue 2r – 1 times. The results of this, together with the
measured data, provide 2r profiles along levels in the interval
0 ≤ y < 2r. This is followed by interpolation around a circle of
radius r (Figure 1b). Note that, with the irregular spacing between
the continued profiles, interpolation is only required along the
horizontal axis. That is, around the periphery of C, the red
(interpolated) and black (input) open circles reside at the same
elevation. In this configuration, derivatives are at a distance r
above the survey’s acquisition level, that is z 0 = x + ir. Optimum
values for r are almost always in the range of two to four of the
intervals between the samples along the profile in the x direction.
Refinement within this range requires some experimentation
to settle on the value that produces the highest-fidelity results
for a specific survey.
We show by example the importance of the superior condi-
tioning provided by Cauchy’s integral formula. Figure 2 shows
fourth derivatives of the vertical component of gravity from an
airborne gravity and magnetic survey acquired on behalf of the
Canadian and Ontario geological surveys (Ontario Geological
Survey and Geological Survey of Canada, 2011). The Black Thor
deposit, outlined by the green rectangle, is discussed by Rainsford
et al. (2017). The signal on the FD derivatives (Figure 2a) is
predominantly narrowband, comprising quasi-sinusoidal fluctua-
tions with only small variations in peak-to-peak amplitude. This
is consistent with its narrowband amplitude spectrum (Figure 2c),
Figure 2. (a) Derivatives calculated using FD and Cauchy’s integral formula. Note examples
characterized by a bandwidth to central frequency ratio less than
of narrow and broadband signal. (b) The difference between the derivatives shown in (a).
(c) Amplitude spectra of the FD and Cauchy derivatives. The bandwidths (i.e., frequency ​ ​)​/ ​(​‾
unity (i.e., ​(​Δ ​fFD ​ ​)​ < 1​​). On the other
fFD _ hand, the Cauchy
bands with amplitudes above –3 dB) of the FD derivatives (ΔfFD ≈ 0.014 m–1) and the Cauchy derivatives are broader band (i.e., Δ ​ f​C​ )​ ​/ (​ f​C​ ​ ≈ 1​)​ . The wider band
derivatives (ΔfC ≈ 0.021 m–1) are annotated. Their mean frequencies, defined is evident on the Cauchy derivatives (Figure 2a) as inflections
as the midpoint of the bandwidths, are ​‾f​FD​ ≈ 0.017 ​m​−1​and ​‾
f​FD​ ≈ 0.021 ​m​−1​, so that and ripples superimposed on the narrowband carrier. Subtracting
‾Δ ​f​ ​
​_ ‾
​
Δ f
​ ​ ​ the FD from the Cauchy derivatives isolates the signal in the
​ _ C​ ≈ 1​.
​ FD ​ ≈ 0.75​and _
​‾
f​FD​ ​f​C​ upper part of the band of the Cauchy derivatives (Figure 2b).

Special Section: Gravity, electrical, and magnetic methods April 2024 The Leading Edge 229
 Next, we use these higher-fidelity, higher-order derivatives for
​c​2​ c​ ​1​ c​ ​0​ ⎡​b1​⎤​
⎢⎥
downward continuation. ​− c​3​

[​c​4​ ​c​3​ ​c​2] [−​ c​5]

​ c​​ ​3​​ ​c​2​​ ​c​1​​ ​ ​b2​​ ​ = ​ ​​− c​4​​ ​.​ (5)
The Padé approximant: Analytic continuation in the complex plane ​ ⎣b​ ​⎦​ ​
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

3
Fornberg and Piret (2020) is an illustrated introduction
to complex analysis. In the following we make frequent refer- Then, the numerator’s coefficients, a 0 , a1, and a2 , are given by
ences to Section 3.2.9 of this text. We begin with Taylor-series
expansions, which prove useful when all that is known are a0 = c0
approximations to the first few coefficients (i.e., the lower- a1 = c 1 + c 0 b 1 (6)
order derivatives). However, differentiation with Cauchy’s a 2 = c 2 + c 1b 1 + c 0 b 2
integral formula affords reliable higher-order derivatives, and
so more terms can be included in the series expansion. Using these coefficients in equation 3, and setting z = 0 – iΔy,
Furthermore, accelerated convergence is achieved by converting continues the data a distance Δy straight downward.
the expansion to a Padé approximation, which is a rational
function comprising the ratio of two finite Taylor series. The Test scenarios
starting point is a truncated Taylor series (Fornberg and Piret, The approximant f (z) is a rational complex function. This
2020, equation 3.14): experiment examines the behavior of this functional form when
approximating potential-field data.
N+M
​f​(z)​ = ​∑ ​ ​cn​​z​​ ​,​ Isolated singularities
n
(2)
n=0
We first study source geometries comprising combinations of
F(z):
where the coefficients ​cn​​ = ​​_
1 d ​f ​n(​ z)​
​ ​​_ |
​ ​​  ​​​ have been obtained from
DOI:10.1190/leedff.2024.43.issue-4

n ! ​d z​​n​ 1
z=0 ​F​(z)​ = A ​e​−iθ​​_ ​.
​ z − ​z​​ ​ ​ q
(7)
( )
Cauchy’s integral formula. The algorithmic steps to follow convert
p

the coefficients cn in equation 2 to an and bn in (Fornberg and Piret, With q = 1, F(z) has an isolated pole at zp, where limz→zp F(z) = ∞.
2020, equation 3.13) The real part of F(z) is the total-magnetic intensity of a thin
sheet with its top at zp = (xp,iyp) and its bottom at infinity (Gay,
​∑ N ​a​​z​​n​ 1963). By superimposing the response of various configurations
​f​(z)​ = ​_ (3) of this function, we model four source shapes and study their
n=0 n
​
M
​∑ n=0​​bn​​z​​ ​
n
Padé approximants.
Thin sheet with infinite depth extent: One singularity. We
(with normalization b 0 = 1) in such a way that the Taylor expansion, consider this to be the impulse response of the approximant. It is
of equation 3, around z = 0 coincides with equation 2 if truncated used to indicate how well a complex rational function, with six
to the same number of terms. This makes little difference over degrees of freedom, approximates the actual field of an isolated
small continuation distances. Thus, for example, there is not much pole. The results are comparable (cf. Figures 3a and 3b), where
reason to use a Padé series, in favor of a purely Taylor series, for Figure 3a is the field of a pole calculated using equation 5 on the
correcting the influences of deviations from a preplanned drape profile y = 0. This is then continued upward, differentiated using
surface (see, e.g., Pilkington and Thurston, 2001); such deviations Cauchy’s integral formula, and subsequently downward continued
are usually only on the order of a few meters. However, for the using the Padé approximants. Figure 3b is the field of a pole at
large continuation distances we contemplate, a rational approxima- all points in the section calculated using equation 7 directly. This
tor is indispensable. offers a fundamental guide for interpreting analytically continued
The best results are obtained by specifying the free parameters data. That is, the thin-sheet approximant (Figure 3a) is a realistic
M and N such that M = N or M = N + 1. Our experience has been portrayal of the actual field of an isolated pole (Figure 3b). Notably,
that good quality, finely sampled airborne data can be stably the top of the sheet coincides with an abrupt change from positive
differentiated five, and often six, times. For this example, we use to negative amplitudes. It is characterized by symmetrical, rounded
the first through fifth derivatives and set N = 2 and M = 3, in lobes of opposite polarity.
which case, equation 3 becomes Thin sheet with finite depth extent: Proximal singularities. This
test provides constraints for what we classify as an isolated body.
We do this by adding a proximal pole to the model. These poles
​a​0​+ ​a​1​z + ​a​2​​z​ ​ 2

​f​(z)​ = _______________
​  
   .​​ (4) are vertically separated by a distance Δh (= 100 m in this instance)
1 + ​b1​​z + ​b2​​​z​ ​+ ​b3​​​z​ ​
2 3
and have opposite susceptibilities. Together these mimic a thin
vertical sheet with finite depth extent, equal to Δh. When analyti-
The coefficients b1, b 2 , and b3 in the denominator of equation 4 cally continued, this source produces the image shown in Figure 3c.
are found by solving The top and bottom each correspond to a change in polarity that
resembles the approximant’s impulse response (Figure 3a).

230 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
 Empirically, we have determined these poles are resolvable wherever location. This source’s focused response (i.e., narrow side lobes)
the depth extent (Δh) is more than twice the depth to the top of can be distinguished from the isolated pole (Figure 3a) and so
the sheet. Put another way, we consider a pole to be isolated if the provides a means for detecting compact sources. Laterally align-
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

distance to its nearest neighbor is greater than its depth. ing several such poles lengthens the slab (Figure 3e). This is
Infinitesimally thick slabs: Coincident singularities. We next imaged as an elongated version of a single, coincident, pole
set q = 2 in the denominator of equation 7. F(z) then becomes (Figure 3e).
the response of a source with infinitesimal thickness. This is
often used to model the effect of a horizontal cylinder (Gay, Nonisolated singularities
1965). This function has two poles at zp, i.e., one pole with a Here, we model sources using the complex logarithm (Fornberg
multiplicity of 2. The approximant (Figure 3d) images two poles and Piret, 2020, Section 2.2.3)
(i.e., two polarity flips) immediately surrounding the actual pole
F(z) = Ae–iθ log(z – zp). (8)

The real part of equation 8 is commonly

used for forward and inverse modeling
bodies resembling deeply seated geo-
logic contacts (e.g., Reid et al., 1990).
At z = zp, the singularity is classified as
a branch point, at which the imaginary
part becomes multivalued. However,
away from branch points, complex
logarithms admit power-series expan-
sions, so rational functions should still
DOI:10.1190/leedff.2024.43.issue-4

serve as reliable approximants. Figure 3f

features a polarity flip coincident with
the edge. Lateral stretching differenti-
ates nonisolated singularities from the
more bulbous symmetrical lobes that
characterize isolated singularities (cf.
Figures 3a, 3d, and 3f).

Scattered singularities
Here we illustrate the capability of
the Padé continuation approach using
responses that originate from randomly
distributed sources. Our synthetic source
uses di / r-type singularities. With r
denoting distance, these functions sat-
isfy the 3D Laplace equation. In an x,y
plane, we place these types of singulari-
ties at the locations shown by the red
circles in each of the subplots of Figure 4.
The constants di, i = 1,2,…,6 were chosen
arbitrarily as –1,1 –2,2, 1,1. This poten-
tial field was then sampled, in nine
separate tests, at 25 equispaced points
around the circle indicated in each
subplot. Applying to these data M = N = 7
Padé approximations give the singularity
locations (zeros of the Padé denomina-
tors) shown by crosses, here displaying
only those falling below the surface. We
see in each case that singularities in the
Figure 3. (a) Response of the Padé approximant for an isolated pole modeling a thin vertical sheet of infinite depth extent at
the magnetic pole (i.e., θ = π / 2 in equation 7). (b) Forward model of an isolated pole. The remaining panels show the Padé vicinity of where the field was sampled
approximants for (c) a thin sheet with finite depth extent, (d) a horizontal cylinder, (e) a thin slab, and (f) a deep-seated contact. become quite accurately located both in
The intensities of the images are linearly scaled between plus (red) and minus (blue) unity. horizontal position and depth.

Special Section: Gravity, electrical, and magnetic methods April 2024 The Leading Edge 231
 Examples from the Flin Flon–Snow Lake the polarity reversal at about 250 m above sea level (asl). The ore
volcanogenic massive sulphide belt body comprises a thin massive sulphide horizon extruded near
The Paleoproterozoic-aged Flin Flon–Snow Lake Greenstone the unconformity, underlain by disseminated to semimassive
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

belt straddles the border between eastern Saskatchewan and sulphide mineralization (Masun and Rennie, 2021). These
western Manitoba. This district hosts dozens of mines and more sulphide-rich rocks correspond to elevated susceptibilities spanning
than 400 million tonnes of Cu-Zn-rich massive sulphide ore. 100 m (Mahmoodi et al., 2018). The vertical extent of this volcanic
Historically, discoveries were made primarily in the belt’s exposed sequence is small relative to its depth. In light of our test cases,
northern portion. To the south, the greenstone facies rocks are it seems likely that the top and bottom are indistinguishable on
unconformably overlain by a southwestward thickening, non- the analytically continued signal. However, the source’s top is
magnetic and barren, Phanerozoic sedimentary sequence. In the constrained to the vicinity immediately above and below the
past few decades, prospectors have enjoyed increasingly successful change in polarity. This interpretation could be further refined
results targeting buried ore using a variety of electromagnetic and with a (forward/inverse) model incorporating the geometrical
potential-field methods. Lo (2019) provides a comprehensive constraints indicated by the approximant and the downhole
summary of examples of subsedimentary discoveries. rock-property measurements.
We use data from the recently discovered sub-Phanerozoic
Bigstone deposit, located 85 km west of Flin Flon, Manitoba, Near-deposit exploration
Canada, near the western end of the belt, where the sedimentary Ongoing exploration has led to the discovery of additional
sequence is tens of meters thick. A reserves estimate has been prospective targets (Foran Mining Corp., 2023). The Babbage
published (Masun and Rennie, 2021), and the data used to compile prospect sits roughly 500 m southward along the Bigstone trend
the report are off-confidential. A well-understood subsurface, (labeled on Figure 4a). Images of this prospect are shown in planar
overflown with modern airborne data, provides a good opportunity (Figure 6a) and cross-sectional (Figure 6b) views. The image in
to ground truth our methods. We do this by comparing a 2011 Figure 6a is of a grid of isolines extracted from the continued profiles
heliborne magnetic survey (Saskatchewan Ministry of Energy at the 280 m asl horizon. Figure 6b shows the continued magnetics
DOI:10.1190/leedff.2024.43.issue-4

and Resources, Assessment File Number 63L11-0093), with along the portion of the flight line over it. In cross section, Babbage
results from a 2015 drilling campaign (Saskatchewan Ministry and Bigstone produce comparable responses (cf. Figures 5b and
of Energy and Resources, Assessment File Number MAW00758). 6b). Note, the polarity flip of the latter is about 20 m deeper, and
The total-magnetic field is shown in Figure 5a. Also shown are thus is not imaged on the 280 m asl slice (Figure 6a). If this is also
the location of the ore body and the trace of an airborne profile a sheet-like host rock, its top is likely about 20 m shallower than
over it. Figure 5b shows its Padé approximant overlain with the the top of Bigstone; again, its depth extent is unknown.
2015 borehole paths. These intersect a feature that most closely The Marconi prospect is about 500 m east of Bigstone (their
resembles the signature of a thin sheet (isolated singularity), with locations are shown on Figure 5a). Drill holes from the late 1990s

Figure 4. Subplots showing the performance of the Padé continuation approach with multiple bodies. The source type and the parameters are described in the main body.

232 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
 revealed mineralization coinciding with the boundary between to a Padé approximant for analytically continuing downward from
a magnetic granodioritic intrusion and relatively nonmagnetic the sensor elevation to considerable distances below the earth’s
felsic volcanics. The white circles (Figure 6a) and the arrow surface. This has the effect of abruptly reversing the sign of the
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

(Figure 6c) show the surface projection of this edge, which at computed field near the source locations, making them readily
depth corresponds to a polarity reversal from positive to negative. identifiable. Synthetic examples provide a template for quantifying
Following the areal extent of the polarity reversal may help to source shape, location, and resolving power. We recommend cali-
locate the contact between the granodioritic and felsic rocks. bration using an atlas of templates mimicking the relevant geologic
setting. We further recommend using the constraints imposed by
Summary and discussion the continued data for refining the interpretation by forward/
Compared to classical methods, Cauchy’s integral formula and inverse modeling of the measured signal (i.e., total-field strength).
Padé approximation offer alternatives for differentiating and con- This provides the opportunity to impose lithologic and structural
tinuing potential-field data These methods distinguish themselves constraints from in-situ measurements.
with superior conditioning, improved stability, and higher accuracy. The results shown in Figures 4 and 5 were computed using a
The derivatives computed using Cauchy’s integral formula are used six-term expansion, necessitating fifth-order derivatives. In our
to generate a rapidly converging power series, which is converted
DOI:10.1190/leedff.2024.43.issue-4

Figure 6. (a) Depth slice of the computed total field at 280 m asl. At this level, the Babbage
and Marconi deposits correspond to sign reversals of the total field (these were overflown
Figure 5. (a) Total-magnetic intensity over the southern portion of the airborne survey overlain by profiles B to B' and C to C', respectively). Cross-sectional views of the continued data
with locations of the deposit and prospects (described in the main body). (b) The continued over (b) the Babbage and (c) the Marconi prospects. Babbage bears resemblance to the
profile, A-A' (looking north), along the part of Line 1300 that is over the known deposit. The response of a thin sheet (cf. Figures 2b and 3c), suggesting a setting similar to Bigstone.
dipping solid black lines are the paths of the drill holes that are in the immediate vicinity of the The Marconi deposit is spatially associated with the edge of a granodiorite intrusion
profile. Some of these intersect the plane and are grayed wherever they are masked by (i.e., previously discovered by drilling. The surface projections of two historical boreholes are
north of) the trace of the profile. The white rectangle near the eastern end is discussed in the close to the polarity flip in both planar and section views. The depth slice (a) may provide
Summary and discussion section. guidance for mapping the areal extent of this intrusion.

Special Section: Gravity, electrical, and magnetic methods April 2024 The Leading Edge 233
 experience, a notable broadening of the bandwidth of Cauchy Fedi, M., and G. Florio, 2011, Normalized downward continuation of
potential fields within the quasi-harmonic region: Geophysical
derivatives, vis-à-vis their classical counterparts, becomes evident
Prospecting, 59, no. 6, 1087–1100, https://doi.org/10.1111/​
for derivative orders greater than 2. Thus, Cauchy’s integral formula j.1365-2478.2011.01002.x.
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

facilitates inclusion of at least three additional stable and accurate Foran Mining Corp., 2023, https://foranmining.com/projects/near-
expansion terms that are not conventionally feasible to calculate. mine-exploration, accessed 10 November 2023.
These higher-order terms extend the utility of the method. Recall, Fornberg, B., 1981, Numerical differentiation of analytic functions:
in the complex plane, a polynomial always has the same number ACM Transactions on Mathematical Software, 7, no. 4, 512–526,
of zeros as its degree. Thus, the rational function specified by https://doi.org/10.1145/355972.355979.
Fornberg, B., and N. Flyer, 2015, A primer on radial basis functions
equation 4 has up to two zeros (equal to the order of the numerator’s
with applications to the geosciences: Society for Industrial and
polynomial) and three poles (equal to the order of the denominator’s Applied Mathematics, CBMS-NSF Regional Conference Series
polynomial). Poles are produced by sources, and zeros result from in Applied Mathematics CBMS-NSF, no. 87, https://doi.
interference of adjacent sources. Allowing for a trio and pair, org/10.1137/1.9781611974041.
respectively, in the region below each sample point provides a Fornberg, B., and C. Piret, 2020, Complex variables and analytic func-
robust means for imaging assemblages comprising multiple sources, tions: An illustrated introduction: Society for Industrial and Applied
without the need for any a priori knowledge of the geometry of Mathematics, https://doi.org/10.1137/1.9781611975987.
Gay, P., Jr., 1963, Standard curves for interpretation of magnetic anomalies
the source configuration.
over long tabular bodies: Geophysics, 28, no. 2, 161–200, https://doi.
We conclude with a discussion of unwanted signal. Downward org/10.1190/1.1439164.
continuation enhances short wavelengths. This can undesirably Gay, S. P., Jr., 1965, Standard curves for interpretation of magnetic
overemphasize random (e.g., instrument) noise as well as systematic anomalies over long horizontal cylinders: Geophysics, 30, no. 5,
(e.g., leveling) errors. The examples shown in Figures 5 and 6 do 818–828, https://doi.org/10.1190/1.1439656.
not appear to be significantly affected by imperfections of this Lo, B., 2019, Geophysical exploration beneath the Phanerozoic cover of
nature. On the profiles shown in Figures 5b, 6b, and 6c, there the Flin Flon–Snow Lake VMS belt: CSEG Recorder, 44, no. 4, 1–23.
Mahmoodi, O., R. O. Maxeiner, R. Morelli, and O. Boulanger, 2018,
are no overwhelming broadband spurious fluctuations that char-
DOI:10.1190/leedff.2024.43.issue-4

New airborne geophysical surveys in the Creighton–Flin Flon area,

acterize random noise. Moreover, the depth slice shown in in Summary of investigations 2018, Volume 2, Saskatchewan
Figure 6a does not bear any obvious herringbone patterns caused Geological Survey, Saskatchewan Ministry of Energy and Resources,
by imperfect line-to-line leveling. However, there are artifacts Miscellaneous Report 2018-4.2, Paper A-7.
that we presume arise from sources that do not conform to our Masun, K. M., and D. W. Rennie, 2021, Technical report on the Bigstone
assumed model. Because there are no counterparts to Cauchy’s Project, East Central Saskatchewan, Canada, NI 43-101 Technical
integral formula or Padé approximation in three independent Report: Foran Mining Corporation.
Nabighian, M. N., 1972, The analytic signal of two-dimensional magnetic
variables, the method is restricted to the vertical complex plane.
bodies with polygonal cross section: Its properties and use for auto-
We have thus implicitly assumed that the responses all correspond mated anomaly interpretation: Geophysics, 37, no. 3, 507–517, https://
to singularities in the plane of the profile. Where this is untrue, doi.org/10.1190/1.1440276.
the rational approximator is incapable of providing a realistic Ontario Geological Survey and Geological Survey of Canada, 2011,
continued image. An example of nonconforming behavior lies Ontario airborne geophysical surveys, gravity gradiometer and mag-
within the white rectangle on the western end of Figure 5b. Note, netic data, grid and profile data (ASCII and Geosoft formats) and
the striped maximum just above 250 m asl, which lacks an atten- vector data, McFaulds Lake area: Ontario Geological Survey,
Geophysical Data Set 1068.
dant minimum. This physically improbable response (it decays
Pilkington, M., and J. B. Thurston, 2001, Draping corrections for
both upward and downward from this peak) likely arises from aeromagnetic data: Line versus grid-based approaches: Exploration
3D sources adjacent to the plane of the profile. For the time being, Geophysics, 32, no. 2, 95–101, https://doi.org/10.1071/EG01095.
these remain beyond our capability to resolve. Rainsford, D. R. B., P. A. Diorio, R. L. S. Hogg, and R. T. Metsaranta,
2017, The use of geophysics in the Ring of Fire, James Bay Lowlands
Data and materials availability — The chromite story: Proceedings of Exploration 17: Sixth Decennial
Data associated with this research can be accessed at http:// International Conference on Mineral Exploration, 649–662.
Reid, A. B., J. M. Allsop, A. J. Granser, A. J. Millett, and I. W. Somerton,
mineral-assessment.saskatchewan.ca/pages/basepages/main.aspx.
1990, Magnetic interpretation in three dimensions using Euler
deconvolution: Geophysics, 55, no. 1, 80–91, https://doi.
Corresponding author: jbthurston@gmail.com org/10.1190/1.1442774.
Voskoboynikov, G. M., and N. I. Nachapkin, 1969, Method of special
References points for the interpretation of potential fields. Izvestiya Akademii
Abate, J., and H. Dubner, 1968, A new method for generating power Nauk SSSR: Fizika Zemli, 5, 28–39 (in Russian).
series expansion of functions: SIAM Journal on Numerical Analysis, Wang, T., S. Liu, H. Cai, and X. Hu, 2023, Comparison and application
5, no. 1, 102–112, https://doi.org/10.1137/0705008. of downward continuation methods for potential field data:
Blakely, R. J., 1995, Potential theory in gravity and magnetic applica- Geophysics, 88, no. 6, G141–G159, https://doi.org/10.1190/
tions: Cambridge University Press, https://doi.org/10.1017/ geo2022-0520.1.
CBO9780511549816. Zhou, W., C. Zhang, and D. Zhang, 2021, Depth estimation of potential
Chong, Z., L. Qingtian, Y. Jiayong, W. Siqi, S. Danian, Z. Wenna, and field by using a new downward continuation based on the continued
L. Lichun, 2022, Method for downward continuation of gravity and fraction in space domain: Earth and Space Science, 8, no. 6,
magnetic data based on Padé approximation: U.K. Patent GB2595741B. e2021EA001789, https://doi.org/10.1029/2021EA001789.

234 The Leading Edge April 2024 Special Section: Gravity, electrical, and magnetic methods
 An artificial intelligence workflow
for horizon volume generation from 3D seismic data
Aria Abubakar1, Haibin Di1, Zhun Li1, Hiren Maniar1, and Tao Zhao1
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

https://doi.org/10.1190/tle43040235.1

Abstract presence of potential reservoirs of natural resources, especially

Horizon-based subsurface stratigraphic model building is a in oil and gas exploration.
tedious process, especially in geologically complex areas where For both sequence stratigraphy and facies analysis, geoscien-
seismic data are contaminated with noise and thus are of weak tists rely heavily on horizon-based seismic interpretation to
and discontinuous reflectors. Seismic interpreters usually use construct the subsurface structural and/or stratigraphic models
stratal (proportional) slices to approximately inspect 3D seismic from seismic images. However, as it is probably the most time-
data along seismic reflectors yet to be interpreted. We introduce consuming task in seismic interpretation, horizon interpretation
an artificial intelligence workflow consisting of three deep learning is always restricted to a few key seismic events that delineate the
steps to provide a conditioned seismic image that is easier to overall structure of the region, as well as within a small interval
interpret, a stratigraphic model that outlines major formations, of interest, usually around the reservoir layers. As it becomes
and moreover a relative geologic time volume that allows us to impractical to manually interpret every peak/trough in a 3D
automatically extract an infinite number of horizons along any seismic volume into an individual horizon, over the past decades,
seismic reflectors within a seismic cube. Depending on the avail- researchers have been developing automatic horizon extraction
ability of interpreters, the proposed workflow can either run fully methods, aiming to extract as many horizons as possible from
unsupervised without human inputs or using sparse horizon the seismic volume, while reducing human efforts. These automatic
DOI:10.1190/leedff.2024.43.issue-4

interpretation as constraints to further improve the quality of horizon extraction methods fall mainly into three categories. The
extracted horizons, providing flexibility in both efficiency and first and simplest category is to use stratal slices (also called
quality. Starting from only seismic images and a few key horizons phantom horizons) to interpolate between two existing horizons
interpreted on very sparse seismic lines, we demonstrate the (Zeng et al., 1998a, 1998b). This method relies on existing inter-
workflow to generate a stack of complete horizons covering the preted horizons as the upper and lower bonds for interpolation
entire seismic volume from offshore Australia. and is extremely easy to implement. However, the extracted
horizons do not always follow the seismic events precisely, and
Introduction the error will accumulate with increasing distance between the
Originated in the 1970s, seismic stratigraphy interpretation upper and lower existing interpretation. The second category of
is the method of choice for understanding the depositional solutions is based on merging local horizon patches automatically
system and mapping the strata, facies, and reservoir distribution extracted along peaks and troughs (Borgos et al., 2003; Monsen
beyond well control. It forms the basis of modern analysis of et al., 2007). The last category is based on seismic structural
sedimentary rocks, especially for hydrocarbon exploration attributes, such as tracking along structure tensor or dip fields
(Mitchum et al., 1977). There are two key branches under (Lomask et al., 2006; de Groot et al., 2010; Wu and Hale, 2015),
seismic stratigraphy analysis: sequence stratigraphy and seismic and extracting as iso-surfaces from relative geologic time (RGT)
facies. A seismic sequence represents a package of sediments volumes (Stark, 2004; Wu and Zhong, 2012). In this study, we
deposited under similar geologic conditions during a specific propose a deep learning workflow that consists of three steps
time interval (Posamentier and Vail, 1988). These sequences — seismic conditioning, key horizon extraction, and RGT estima-
are bounded by unconformities, which are surfaces representing tion — to extract an infinite number of horizons from a 3D seismic
gaps in the geologic record due to sea level change, erosion, or volume. Depending on the availability of interpreters, the proposed
nondeposition. By identifying seismic sequences and uncon- workflow can either run fully unsupervised without any human
formities, geoscientists can reconstruct the geologic history of inputs or use sparse horizon interpretation as constraints to further
an area and understand the factors that influenced sediment improve the quality of the extracted horizons, providing flexibility
deposition. Seismic facies analysis is another essential aspect in both efficiency and quality.
of seismic stratigraphy. It involves the classification of seismic In a traditional seismic interpretation workflow, seismic image
reflections based on their amplitude, frequency, continuity, conditioning is routinely used to improve the accuracy and effi-
and geometry (Mitchum et al., 1977). Different seismic facies ciency of horizon interpretation when using traditional horizon
correspond to distinct lithologies or depositional environments, picking tools. This is enabled by removing the noise and artifacts
allowing geoscientists to infer the nature of the subsurface in the data, as well as improving the continuity of seismic events
sediments. Such information is critical for predicting the and clarity of faults. Similarly, conditioning the seismic image

Manuscript received 1 December 2023; accepted 30 January 2024.

1
SLB, Houston, Texas, USA. E-mail: aria.abubakar@gmail.com; haibin.di@outlook.com; zli68@slb.com; hmaniar@slb.com; tao.zhao@alumni.ou.edu.

April 2024 The Leading Edge 235

 also provides better performance for downstream machine learning RGT volume, one can conveniently identify the order of the
(ML) tasks, such as ML-based horizon picking and fault detection. sequence and extract horizons as iso-surfaces. We use a deep
Although there are abundant classical methods for seismic con- learning model to generate an RGT volume under the constraints
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

ditioning (Fehmers and Höcker, 2003; Hale, 2011; Zhang et al., of the key horizons interpreted from the previous step to improve
2016), it usually relies on great expertise to parameterize these the quality of the horizons in regions with complex structures.
methods and requires balancing among different conditioning In the rest of the paper, we use an example from the Stybarrow
objectives. To make seismic image conditioning a fully automated oil field in offshore Australia to demonstrate the proposed work-
process, we use a deep learning model trained in a self-supervised flow, while also providing more technical details on the deep
fashion to prepare the seismic data for the next steps. learning methods for each of the three steps.
Deep learning methods have greatly advanced seismic stra-
tigraphy interpretation in the past few years (Zhao, 2018; Peters Data set
et al., 2019; Wu et al., 2019; Di et al., 2020; Abubakar et al., The Stybarrow oil field is located in a northeast–southwest-
2022). These methods enable geoscientists to map the major tilted fault block, bound to the east and west by faults and with
stratigraphic layers or facies by training deep learning models structural closure to the southwest in Western Australia
with a few labeled examples. In this study, we use deep-learning- (Figure 1a). It was discovered in 2003 by BHP Billiton (now
based seismic stratigraphy interpretation to generate complete BHP), which drilled Stybarrow-1 to target the turbiditic Late
coverage of key horizons in a 3D seismic volume, which serve Jurassic Macedon Sandstone in a fault-bound structural closure.
as constraints when performing the fully automatic horizon After an 18.6 m net thickness oil column was discovered, the
extraction over the entire seismic volume. field was appraised with Stybarrow-2 to Stybarrow-4, and the
The concept of RGT was introduced in Stark (2004), where Stybarrow 3D seismic survey was collected between 2003 and
the author used phase unwrapping to generate a 3D volume 2005. The subset used in this paper covers an area of 4.875 by
representing the estimated relative age of each seismic sample in 11.875 km and the depth from 800 to 2572 ms two-way trav-
the volume. With a sequence of sedimentary rocks, RGT indicates eltime. The seismic survey consists of 391 inlines, 951 crosslines,
DOI:10.1190/leedff.2024.43.issue-4

the relative age of a sedimentary rock layer compared to the other and 887 samples per trace. As shown in Figure 1b, the seismic
layers (Catuneanu, 2017). As younger layers sequentially deposit data are imaged of good quality. One of the dominant features
on older layers, RGT can be determined by tracking the vertical in the 3D seismic data is the pinch-out, above which the struc-
order of rock layers (Dawes and Dawes, 2013). With a determined tures are relatively continuous and clearly imaged. However, the

Figure 1. The Stybarrow survey in Western Australia used for testing the proposed workflow.

236 The Leading Edge April 2024

 zones below the pinch-out contain numerous faults and relatively using multiple self-supervised tasks such as denoising, occlusion,
poor signal-to-noise ratio (S/N). distortion, etc. During training, the parameters related to the
self-supervision tasks are continuously and randomly modified in
Workflow description
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

a controlled manner to ensure that the model excludes inconsistent

Figure 2 illustrates the proposed workflow for generating a features in the survey and only learns meaningful consistent
horizon volume corresponding to any 3D seismic data. The attributes of the data. Such an approach endows the model with
workflow consists of three components,
detailed in the following.
Seismic image conditioning. We
define conditioning as transformations
to simultaneously accomplish the tasks
of random noise attenuation, coherent
migration artifact attenuation, smooth-
ing and/or filling in broken horizons,
and accentuating faults. The condition-
ing approach also must maintain resolu-
tion between finely spaced consecutive
horizons, maintain fault integrity, and
within reason, locally maintain relative
amplitudes. Here, seismic image con-
ditioning is conducted employing a
single model built using a deep convo-
lutional neural network. The network
DOI:10.1190/leedff.2024.43.issue-4

is trained in an unsupervised manner Figure 2. The proposed workflow for horizon volume generation from 3D seismic data.

Figure 3. A stand-alone deep convolutional model can be used for “conditioning” surveys to allow for better structural interpretation, as well as further use in downstream ML-based
interpretation tasks. See the text for definition of conditioning. (a) Original section, (b) ML-conditioned section, and (c) the residual.

April 2024 The Leading Edge 237

 capabilities to detect and extract widely varying noise/artifact cube (Figure 5). Finally, we extract the stratigraphic boundaries
characteristics and strengths. Furthermore, the nature of the throughout the survey to extract the horizons while correlating
training tasks also provides the model with the ability to improve them with the corresponding seismic peak/trough events.
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

horizon continuity in poor S/N zones while maintaining fault For the Stybarrow data set, six key horizons are identified.
integrity. It is noteworthy that the model will also attenuate Manual picking of the six horizons is on eight inline and 10
coherent noise and migration artifacts. crossline sections (Figure 6), and the corresponding stratigraphic
Figure 3 shows an example of how the imaging quality of the image contains seven sequences. Figure 7 displays the predicted
Stybarrow survey is improved by employing the conditioning ML stratigraphy cube as well as the extracted six horizons. All of them
model. Note that the method sharply attenuates migration artifacts, match the seismic events closely and clearly capture the deposi-
significantly boosts the horizon continuity in deeper zones, and tional features from the seafloor to the unconformity. Additionally,
effectively sharpens the fault planes. While a certain amount of Figure 8 compares the stratigraphy prediction from original and
signal change is observed locally at a short scale, such changes conditioned seismic images, which verifies the added values of
have minimal effect on structural interpretation tasks including preconditioning seismic images in producing artifacts (denoted
the two described in the following. by circle and arrow). Although a few small pieces are missing as
Key horizon extraction. With the conditioned seismic volume, the machine prediction is rejected while snapping to the seismic
the next step is to extract a set of key horizons that outline the events, it has minimal effect for ensuing interpretation workflows
major formations and are to be used for constraining the following including the horizon-constrained RGT estimation described in
step of RGT estimation. In general, we formulate the task as a the following.
multiclass segmentation problem. To address the challenge of RGT estimation. The workflow to generate the RGT volume
sparse horizons in a seismic image, the segmentation is targeted is illustrated in Figure 9. In addition to the step of key horizon
at the stratigraphic formations bounded by these target horizons. extraction noted earlier, a 3D seismic survey is fed into another
Figure 4 illustrates the corresponding workflow. More specifically, processing branch — FlowNet inference, from which a flow field
it starts with picking the key horizons in a few sections. Next, for each pair of seismic traces will be obtained. The following is
DOI:10.1190/leedff.2024.43.issue-4

the horizons are converted into stratigraphic images with the loop-tie optimization, in which both the key horizons and flow
horizons as the stratigraphic boundaries. With the converted fields will be fed into an optimization engine to iteratively improve
stratigraphic images at the selected sections as training labels, the RGT volume until it satisfies both the flow field correlation
the semisupervised learning workflow for stratigraphy interpreta- and the key horizons constraints. In this workflow, the flow field
tion by Di et al. (2020) is applied to generate the stratigraphy between each pair of seismic traces is obtained by applying a
pretrained FlowNet on the 3D seismic survey. The training and
implementation details of FlowNet can be found in Li and
Abubakar (2020a, 2020b).
Figure 10 illustrates the optimization process in detail. The
RGT cube is assumed to be in a Cartesian grid, which consists
of inline, crossline, and depth dimensions. The inline dimension
is indexed by i = 1, 2, …, n, with n representing the number of
grids along the inline dimension. Crossline and depth dimensions
are indexed by j = 1, 2, …, m and k = 1, 2, …, l, respectively.
RGTi,j,k is used to indicate the RGT value at inline i, crossline j,
and depth k. Flow fields involved in this study are oriented along
both inline and crossline directions, annotated as F in and F cr,
Figure 4. The subworkflow for deep-learning-based key horizon extraction. respectively. ​F ​i,jin​ indicates the flow from trace RGTi,j to RGTi+1,j
along the inline direction, and F ​ i,j​cr​​indi-
cates the flow from trace RGTi,j to
RGTi,j+1 along the crossline direction.
Starting from a flow field trace ​Fi,j​in​​
along an inline direction between two
RGT traces RGTi,j and RGTi+1,j , an
accurate flow field will provide the pixel-
wise vertical shift distance of seismic
horizons from RGTi,j and RGTi+1,j .
Specifically, the horizon located at
depth k on RGTi,j will move to depth​
k − ​Fi,j,k
​in ​​ on RGTi+1,j , where ​Fi,j,k
​in ​​ is the
shift displacement. Let ​RGT​i+1,j,k−​F​ ​​​ be
in
i,j,k

the value of RGTi+1,j sampled at depth​

Figure 5. The semisupervised learning for stratigraphy interpretation (Di et al., 2020).

238 The Leading Edge April 2024

Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Figure 6. (a) Map view of training lines and (b) key horizon labels used for the step of deep-learning-based key horizon extraction.
DOI:10.1190/leedff.2024.43.issue-4

Figure 7. The 3D view of extracted six key horizons in the Stybarrow survey.

Figure 8. Comparison of stratigraphy interpretation from an (a) original and (b) conditioned seismic image. Note the improved continuity and reduced artifacts when the seismic image is
preconditioned (denoted by the circle and arrow).

April 2024 The Leading Edge 239

 k − ​Fi,j,k
​in ​​. A good estimate of an RGT cube should satisfy the
following equation: ​Loss = ​Loss​Fin​​+ ​Loss​Fcr​​+ b​(​​Loss​Ain​​+ ​Loss​Acr​​)​. (7)
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

​RGT​i,j,k​− ​RGT​i+1,j,k−​F​ ​​ = 0, k = 1, … , l​.

in (1) The principle of superposition in stratigraphy states that for
i,j,k

a sequence of sedimentary layers, the older layer is at the base and

As the depth ​k − ​Fi,j,k
​in ​​ might not be at the exact grid location, layers above are progressively younger with ascending order (Dawes
linear sampling is applied to obtain an estimate of ​RGT​i+1,j,k−​F​ ​:​ in and Dawes, 2013). This requires that the RGT on a single trace
i,j,k

is monotonically increasing as the depth increases. Furthermore,

​RGT​i+1,j,k−​F​ ​​ = ​∑ l​k′​=1​RGT​i+1,j,​k′​​max​(0,1 − ​|k − ​Fi,j,k
in ​in ​− ​k′​|​)​. (2) ensuring that each RGT trace is a monotonic function can sig-
i,j,k

nificantly reduce the complexity of the algorithm to extract

In this study, flow field is obtained by training and inferencing horizons from the RGT cube.
seismic FlowNet. Based on equation 1, the following mean square Given a seed point, horizon extraction from an RGT cube
error loss functions can be defined by including all the trace pairs, is a process of finding the iso-surface so that all the points on
along both inline and crossline dimensions: the surface have the same RGT value as the seed point. An RGT
value search on a monotonic curve can be accelerated in different
​​∑ mj=1​∑ lk=1​(​​RGT​i,j,k​− ​RGT​i+1,j,k−​F​ ​)
2
​Loss​F​in​​ = ​∑ n−1
i=1 ​
in​ ​, (3) ways by taking advantage of the assumption of monotonicity.
i,j,k

This will make horizon extraction much faster than on a non-

​​∑ lk=1​(​​RGT​i,j,k​− ​RGT​i,j+1,k−​F​ ​)
2
​Loss​Fcr​​ = ​∑ ni=1​∑ m−1
j=1 ​
cr​ ​. (4) monotonic curve.
i,j,k

To ensure that the monotonic constraint is satisfied for each

In addition to the flow field-based loss functions in equations 3 trace, we parameterize the RGT value as a running total of a
and 4, loss functions based on seismic amplitude can be included in sequence of positive numbers. Let a real number wi,j,k ∈ ℝ be the
the optimization. Let Ai,j,k represent a seismic amplitude that shares parameter to be optimized at inline i, crossline j, and depth k.
the same grid location with RGTi,j,k. For every trace, an RGT value The RGT trace value at inline i, crossline j will be formulated as
DOI:10.1190/leedff.2024.43.issue-4

should correspond to a seismic amplitude value. Let A ​ ′​ini,j​​be the seismic

amplitude values of trace Ai+1,j to which RGTi,j corresponds. ​A′​ini,j​​ can RGTi,j,1 = wi,j,1, (8)
be calculated be linear sampling Ai+1,j at all RGTi,j. For a good RGT
cube, ​A′​i,jin​​ should have maximum cross correlation with Ai,j. The ​RGT​i,j,k​ = ​RGT​i,j,k−1​+ ​e​​w​ ​​, k = 2, … , l​.
i,j,k (9)
following amplitude-based loss function can be defined:
in
For any value of wi,j,k ∈ ℝ, e​ ​w​ ​ ​​​ is positive, and this guarantees that
i,j,k

​in ​ = ​∑ k+s
​Ci,j,k ​A​ ​*​A′​ ​k = 1, … , n − 1​,
​k′​=k−s i,j,k
(5) RGTi,j is monotonically increasing along depth.
i,j,​k′​
The final loss is differentiable with respect to each parameter
​Loss​Ain​​ = ​∑ n−1​​∑ mj=1​∑ lk=1​e​−​C​ ​​, (6) wi,j,k. Automatic differentiation of equation 7 with respect to wi,j,k
in
i,j,k
i=1
can be implemented in many deep learning frameworks as an
where C​ i,j,k
​in ​​is cross correlation, and s is the half size of the window industry standard. Once the loss function reaches a preset thresh-
for calculating cross correlation. The final loss function can be old, the optimization is completed and the RGT cube can be
defined in the following, where b is a scaling factor to balance calculated using equations 6 and 7.
between flow-based loss and amplitude-based loss: The flow field generated by seismic FlowNet (Li and Abubakar,
2020a) can vary in scale. The distance
that a flow field can extend ranges from
one grid interval to n – 1 grid intervals,
where n is the number of grids along
one of the two lateral dimensions.
Selecting the flow field with shorter
working distance requires more com-
puter memory and runtime to complete
the optimization, as more variables and
computing steps are involved. On the
other hand, using a flow field that is too
short may cause the optimization to be
trapped in local minima. Identifying
the proper flow field range is important
to balance convergence stability, com-
puting time, and output resolution.
To achieve stable convergence and
save computing resources, optimization
Figure 9. The subworkflow of RGT volume estimation. in this study is performed on a decimated

240 The Leading Edge April 2024

 grid. Every 10th trace is kept for RGT optimization. This requires Results and applications
the input of a flow field with a working distance of 10 grid intervals. Figure 11 demonstrates a 3D view of the generated horizon
The flow field is derived by simply aggregating all the flow fields cube. It not only successfully tracks the continuous horizons in
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

in between with distance of one trace interval. Another method the top zone and clearly reflects the pinch-out, it also captures
to obtain such a flow field is to feed the seismic FlowNet with well the geologic complexities in the presence of faults in the
image pairs that are 10 grids apart and output the required flow deep zone.
field. Once the optimization is completed on the sparse grids, RGT The generated horizon cube can be used for other tasks of
with full resolution can be obtained using interpolation. seismic interpretation, such as
DOI:10.1190/leedff.2024.43.issue-4

Figure 10. The loop-tie optimization.

Figure 11. The generated horizon volume corresponding to the Stybarrow seismic survey by the proposed workflow.

April 2024 The Leading Edge 241

 • Seeded horizon tracking. Figure 12 demonstrates an example Conclusions
of tracking four horizons in the Stybarrow data set. All of In this paper, we presented an integrated workflow for convert-
them match the original seismic events with high accuracy, ing 3D seismic data into the corresponding horizon volume via
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

especially the two horizons in the middle where the seismic deep-learning-accelerated seismic data conditioning, key horizon
quality is relatively low. extraction, and RGT estimation. As tested over the public
• Strata extraction. A stratum allows detailed analysis of rock Stybarrow survey in Western Australia, it captures the structural
relationships within a chronostratigraphic framework and complexity well in the presence of faults and generates a horizon
moreover the distribution of facies and lithology to identify volume that is of high quality and can further assist other seismic
potential stratigraphic traps. Figure 13 demonstrates the interpretation tasks such as seeded horizon tracking, strata extrac-
extracted stratum that clearly reflects the pinch-out and tion, and structure-constrained static property modeling.
indicates the lateral variations in deposition.
• Structure-constrained ML-based static property modeling. Acknowledgments
By treating the horizon cube as a low-frequency model, We would like to thank Chester J. Weiss, Mrinal K. Sen, and
incorporating it into ML-based static property modeling one anonymous reviewer for their insights and suggestions on
has led to significantly improved consistency of machine improving the quality of this work. Thanks also go to Geoscience
prediction, particularly in the absence of sufficient training Australia for providing the Stybarrow seismic survey and SLB
wells (Di et al., 2022). for granting permission to publish.
DOI:10.1190/leedff.2024.43.issue-4

Figure 12. An example of seeded tracking of four horizons based on the generated horizon volume.

Figure 13. An example of extracting the stratum of the pinch-out based on the generated horizon volume.

242 The Leading Edge April 2024

 Data and materials availability Monsen, E. M., H. G. Borgos, P. L. Guern, and L. Sonneland, 2007,
Data associated with this research are available and can be Geologic-process-controlled interpretation based on 3D Wheeler
diagram generation: 77th Annual International Meeting, SEG,
obtained by contacting the corresponding author.
Expanded Abstracts, 885–889, https://doi.org/10.1190/1.2792549.
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Peters, B., J. Granek, and E. Haber, 2019, Multiresolution neural networks

Corresponding author: haibin.di@outlook.com for tracking seismic horizons from few training images: Interpretation,
7, no. 3, SE201–SE213, https://doi.org/10.1190/INT-2018-0225.1.
References Posamentier, H. W., and P. R. Vail, 1988, Eustatic controls on clastic
Abubakar, A., H. Di, A. Kaul, C. Li, Z. Li, V. Simoes, L. Truelove, and deposition II — Sequence and systems tract models, in C. K. Wilgus,
T. Zhao, 2022, Deep learning for end-to-end subsurface modeling B. S. Hastings, C. A. Ross, H. Posamentier, J. Van Wagoner, and
and interpretation: An example from the Groningen gas field: The C. G. St. C. Kendall, eds., Sea-level changes: An integrated approach:
Leading Edge, 41, no. 4, 259–267, https://doi.org/10.1190/ Society of Economic Paleontologists and Mineralogists, 125–154.
tle41040259.1. Stark, T. J., 2004, Relative geologic time (age) volumes — Relating every
Borgos, H. G., T. Skov, T. Randen, and L. Sonneland, 2003, Automated seismic sample to a geologically reasonable horizon: The Leading
geometry extraction from 3D seismic data: 73rd Annual International Edge, 23, no. 9, 928–932, https://doi.org/10.1190/1.1803505.
Meeting, SEG, Expanded Abstracts, 1541–1544, https://doi. Wu, X., and D. Hale, 2015, Horizon volumes with interpreted constraints:
org/10.1190/1.1817590. Geophysics, 80, no. 2, IM21–IM33, https://doi.org/10.1190/
Catuneanu, O., 2017, Sequence stratigraphy: Guidelines for a standard geo2014-0212.1.
methodology: Stratigraphy & Timescales, 2, 1–57, https://doi. Wu, X., and G. Zhong, 2012, Generating a relative geologic time volume
org/10.1016/bs.sats.2017.07.003. by 3D graph-cut phase unwrapping method with horizon and uncon-
Dawes, R. L., and C. D. Dawes, 2013, Geology of the Pacific Northwest: formity constraints: Geophysics, 77, no. 4, O21–O34, https://doi.
Basics — Statigraphy and relative ages, https://commons.wvc.edu/ org/10.1190/geo2011-0351.1.
rdawes/Basics/stratigraphy.html, accessed 12 March 2024. Wu, H., B. Zhang, T. Lin, D. Cao, and Y. Lou, 2019, Semiautomated
de Groot, P., A. Huck, G. de Bruin, N. Hemstra, and J. Bedford, 2010, seismic horizon interpretation using the encoder-decoder convolutional
The horizon cube: A step change in seismic interpretation: The Leading neural network: Geophysics, 84, no. 6, B403–B417, https://doi.
Edge, 29, no. 9, 1048–1055, https://doi.org/10.1190/1.3485765. org/10.1190/geo2018-0672.1.
DOI:10.1190/leedff.2024.43.issue-4

Di, H., Z. Li, and A. Abubakar, 2022, Using relative geologic time to Zeng, H., M. M. Backus, K. T. Barrow, and N. Tyler, 1998a, Stratal
constrain convolutional neural network-based seismic interpretation slicing, part I: Realistic 3-D seismic model: Geophysics, 63, no. 2,
and property estimation: Geophysics, 87, no. 2, IM25–IM35, https:// 502–513, https://doi.org/10.1190/1.1444351.
doi.org/10.1190/geo2021-0257.1. Zeng, H., S. C. Henry, and J. P. Riola, 1998b, Strata slicing; Part II,
Di, H., Z. Li, H. Maniar, and A. Abubakar, 2020, Seismic stratigraphy Real 3-D seismic data: Geophysics, 63, no. 2, 514–522, https://doi.
interpretation by deep convolutional neural networks: A semisuper- org/10.1190/1.1444352.
vised workflow: Geophysics, 85, no. 4, WA77–WA86, https://doi. Zhang, B., T. Lin, S. Guo, O. E. Davogustto, and K. J. Marfurt, 2016,
org/10.1190/geo2019-0433.1. Noise suppression of time-migrated gathers using prestack structure-
Fehmers, G. C., and C. F. W. Höcker, 2003, Fast structural interpretation oriented filtering: Interpretation, 4, no. 2, SG19–SG29, https://doi.
with structure-oriented filtering: Geophysics, 68, no. 4, 1286–1293, org/10.1190/INT-2015-0146.1.
https://doi.org/10.1190/1.1598121. Zhao, T., 2018, Seismic facies classification using different deep convo-
Hale, D., 2011, Structure-oriented bilateral filtering of seismic images: lutional neural networks: 88th Annual International Meeting, SEG,
81st Annual International Meeting, SEG, Expanded Abstracts, Expanded Abstracts, 2046–2050, https://doi.org/10.1190/
3596–3600, https://doi.org/10.1190/1.3627947. segam2018-2997085.1.
Li, Z., and A. Abubakar, 2020a, Seismic Flownet: Using optical flow
field for dense horizon interpretation: 82nd Annual Conference and
Exhibition, EAGE, Extended Abstracts, https://doi.
org/10.3997/2214-4609.202010777.
Li, Z., and A. Abubakar, 2020b, Complete sequence stratigraphy from
seismic optical flow without human labeling: 90th Annual International
Meeting, SEG, Expanded Abstracts, 1248–1252, https://doi.
org/10.1190/segam2020-3427292.1.
Lomask, J., A. Guitton, S. Fomel, J. Claerbout, and A. A. Valenciano,
2006, Flattening without picking: Geophysics, 71, no. 4, P13–P20, © 2024 The Authors. Published by the Society of Exploration Geophysicists. All article
https://doi.org/10.1190/1.2210848. content, except where otherwise noted (including republished material), is licensed
Mitchum, R. M. Jr., P. R. Vail, and S. Thompson III, 1977, Seismic under a Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA) license.
stratigraphy and global changes of sea level, Part 2: The depositional See https://creativecommons.org/licenses/by-sa/4.0/. Distribution or reproduction of
sequence as a basic unit for stratigraphic analysis, in C. E. Payton, ed., this work in whole or in part commercially or noncommercially requires full attribution
AAPG Memoir: Seismic stratigraphy — Applications to hydrocarbon of the original publication, including its digital object identifier (DOI). Derivatives of this
exploration, https://doi.org/10.1306/M26490C4. work must carry the same license.

April 2024 The Leading Edge 243

Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Experience the Best Networking,

Technology, and Solutions for Geophysicists

Opening Plenary Session

Back to Basics: Finding and Developing Oil and Gas
DOI:10.1190/leedff.2024.43.issue-4

Mark Finley, Deborah Byers, Trey Lowe, Kim McHugh,

Fellow in Energy Independent Board Chief Technology Vice President of
and Global Oil, Rice Member, Energy Officer and Vice Rockies Business Unit,
University’s Baker Transition and President, Chevron
Institute Women in Energy Devon Energy

Attend These Special Events for New Insights and Opportunities

Special Session: Unconventionals in Focus: Enverus Special Session: Best of ARMA
Presents the Latest Trends and Future Outlook ARMA specialists will provide insights in different aspects of
Uncover which basins and operators have the most low-cost running room subsurface science and engineering for developing energy resources and
as well as the implications for commodity and E&P asset markets. This talk solving its techno-economic challenges. Application of rock mechanics
will identify which operational strategies are driving the most value extraction principles, measurements, and simulations to characterize, comprehend,
across key shale basins like the Permian, Williston, and Eagle Ford. quantify, and forecast the reservoir response will be highlighted.

Special Session: Unconventional Technology Special Session: The New Permian - Operations of the
Paving the Way for Geothermal Development Future: Changes in How We Work in the Permian
Unconventional O&G technologies, including hydraulic fracturing, Leading operators plan to highlight key aspects related to the future
steam injection, and directional drilling, have paved the way for of operations in the Permian Basin, including induced seismicity, water
hydrocarbon production. These technologies are now being utilized for management, and AI powering workflows with new technologies.
enhanced geothermal systems (EGS), which are expected to be a major
part of the transition. Learn how these technologies are leading the way Panel Session: New Technologies and AI in
to convert heat into sustainable, reliable energy. Reservoir Modeling, Drilling, and Completions
Operators share their experiences and insights in using new
Special Session: Hydraulic Fracturing - 2024 Technical technologies in reservoir modeling, drilling, and completions of
Work Highlights unconventional resources. Specifically, operators will discuss new methods
Explore the latest technical work related to hydraulic fracturing. Keynote for reservoir modeling, AI-assisted drilling, and refracs in the Barnett.
speakers from HFTC 2024 will provide a summary of their papers, followed
by an enriched Q&A session.
 17–19 June 2024 • George R. Brown Convention Center • Houston, Texas

Registration Now Open

Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Scan to see the full Save $125 By 17 May

program details
SPONSORING ORGANIZATIONS:

URTeC.org

Access Global Expertise with URTeC Short Courses

URTeC Short Courses offer a wealth of information in a short period of time.
Leverage the expertise from technical instructors sharing the knowledge and skills to help you succeed.
DOI:10.1190/leedff.2024.43.issue-4

The Techniques and Business Aspects Introduction to Unconventional Reservoir

of Carbon Capture, Utilization and Geohazards Assessment Engineering Fundamentals
Storage, and the CO2 Storage Resources Instructor: Agatha Onuwaje, Shell International for Geoscientists
Management Systems (SRMS) Exploration and Production Inc. Instructor: Yogashri Pradhan, Chevron
Instructors: Rawdon Seager and
Nicholas Fulford, GaffneyCline

Explore Diverse and Compelling Technical Sessions

Reflecting the Industry’s Evolving Landscape
Monday Tuesday Wednesday
• Operators Forum: A Focus on Key • Operators Forum: Well Spacing to • Operators Forum: Well Spacing to
Unconventional Basins and Their Field Development and Stacked Pay Field Development and Stacked
Development I Development Strategies I Pay Development Strategies II
• Geophysical Analysis: Operational • Operators Forum: Quantifying, • Geophysical Analysis: Seismic
Hazard and Fracture Identification Mitigating and/or Overcoming Applications to Aid Drilling
• Geophysical Analysis: Complex Parent-Child Effects I and II and Production
Fracture Characterization • Geomechanics: Rock and Hydraulic
• Geomechanics: Geomechanics in Well Fracture Modeling I and II
Design and Impact of Hydraulic Fracturing

Hungry for More?

Join your colleagues for Topical Luncheons and informative discussions from leading experts
covering the latest challenges and opportunities in unconventional resource development.

A Bigger, Faster Disaster: Making U.S. Energy Outlook Back to the Future: Parts I, II, & III
AI Behave with Massive Subsurface Speaker: Manuj Nikhanj, Enverus — The Evolution of Unconventional
Data Sets Reservoir Technologies Over the Last
Speaker: Robert Gibson, S&P Global Three Decades
Speaker: Thomas Blasingame,
Texas A&M University
 Spectrometric borehole logging in mineral exploration and mining
Manuel Queißer1, Matthew Tudor1, 2, Jens Schubert1, Alexander R. Domula1, Horst Märten1, Thomas Heinig1, Tobias Rothe1, and Alfred Stadtschnitzer3
https://doi.org/10.1190/tle43040246.1
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Abstract only a point-like volume, which is embedded in a larger body of

Borehole logging tools based on pulsed neutron technology a complex heterogenous mineralogy with often only a few samples
are used to infer geophysical parameters such as bulk density. If per hole. This increases uncertainty of the mineralogy between
fitted with a gamma-ray spectrometer, they may also measure sampling points, in particular, the type of deposit and its spatial
elemental mass fractions of a variety of formation elements. variation.
Spectrometric borehole logging tools of this kind are established In mining, obtaining a more detailed picture of the elemental
in the oil and gas industry. With the demand for mineral resources composition hidden behind the visible quarry face or borehole
surging, cost- and time-efficient measurement techniques are wall at high depth resolution is advantageous. In exploration,
increasingly needed in mineral exploration. Their ability to detect obtaining a near real-time quantitative picture of host rock versus
multiple elements that are building blocks of minerals makes mineralized zone, and even a quantification of the elemental
spectrometric borehole logging tools popular in mineral exploration composition of the mineralized zone, is no less desirable.
and mining grade control. In this article, the potential of the Geophysical borehole logging is a measurement technique that
OreLog pulsed neutron gamma spectrometric downhole logging can meet these needs and may complement assay analysis.
tool is demonstrated at Erzberg Mountain (Styria, Austria). Geophysical borehole logging is well established (Hearst et al.,
Erzberg hosts the largest known siderite deposit in the world, 2000; Walls et al., 2004; Ellis and Singer, 2008; Williams and
and open-pit mining is used to extract iron ore. The aim was to Paillet, 2023). Various sensor types are used to probe different
obtain a more detailed and faster geochemical characterization characteristics of the subsurface through sonic (Box and Lowrey,
of iron ore deposits prior to blasting. Elemental logs were acquired 2003; Wong et al., 2009), seismic techniques (Goetz et al., 1979;
DOI:10.1190/leedff.2024.43.issue-4

with highest accuracies for Fe, Ca, and Mn, followed by Si, Mg, Stewart, 2001), resistivity and self potential (Archie, 1942; Salazar
K, and S. The tool mapped a 2D cross section of formation element et al., 2008; Jackson et al., 2012), gamma ray (Howell and Frosch,
concentration within a few hours, which agrees with orogenetic 1939, 1940; Hesselbo, 1996), and nuclear magnetic resonance
fault and geologic contact lines. Mineralogical information (Brown and Gamson, 1960; Kleinberg et al., 1994). Further, there
retrieved from elemental logs confirmed the interlaced mineralogy is the gamma-gamma technique combined with the measurement
of the site. It is dominated by a carbonate body, hosting an inter- of the photoelectric cross section of the formation (Minette et al.,
growth of Mg-rich siderite (sideroplesite) and a solid solution of 1986). It uses an active gamma source and is not only able to
dolomite/ankerite/kutnohorite. In addition to the tool’s potential measure density but also able to provide geochemical information
for enhanced real-time mining control, the benefit of continuous (i.e., elemental composition of the formation). Another active-
depth-resolved elemental logs is seen in the possibility to deduce source technique can be summarized as neutron-induced gamma
a more refined and reliable geochemical and mineralogical model spectroscopy. It is capable of simultaneously yielding elemental
of the deposit. The information-rich data stream opens applications mass fractions of an array of elements including Al, Mg, Si, K,
beyond grade control and exploration, such as data assimilation S, Ca, and Ti (Hertzog et al., 1989) by using spectrometry of
with gravity models or geologic model building. gamma radiation induced by the bombardment of the formation
with neutrons. Initially developed for hydrocarbon exploration,
Introduction the principles of neutron-induced gamma spectroscopy have been
Conventional assessment of ore grade in mining and explora- successfully engineered into logging tools geared toward mineral
tion relies on laboratory analysis of drill-hole samples (assay exploration and mining. This includes applications in coal and for
analysis). To this end, X-ray fluorescence (XRF) spectroscopy is elements such as Fe (Borsaru, 1993), Ni (McDowell et al., 1998),
widely used. The advantage of assay analysis is the high accuracy and Cu (Charbucinski et al., 2003). A significant advancement
that is achievable for many key elements. A main drawback is emerged when radionuclide neutron sources such as 252Cf were
that sampling and analysis are relatively lengthy and costly, which replaced with pulsed neutron generators (Smith et al., 2013).
restricts the number of samples taken and hence the depth resolu- These generators not only enhance safety and mitigate environ-
tion and depth coverage for deep-lying ore bodies (greater than mental risk but also enable precise temporal control of the neutron
100 m). Therefore, for economic reasons, a selection of depth source. This time-lapse mode enables the analysis of gamma
intervals of core material to be sent to the laboratory for analysis photons and scattered neutrons from different neutron-matter
must be made, which bears the risk of missing fertile ore bodies interactions, facilitating more accurate quantification of a broader
(i.e., lithium-fertile pegmatites). Moreover, assay analysis samples range of elements (Zauner, 2021).

Manuscript received 5 October 2023; revision received 28 February 2024; accepted 5 March 2024.
1
Umwelt- und Ingenieurtechnik GmbH Dresden, Dresden, Germany. E-mail: manuelqueisser@web.de; m.tudor@uit-gmbh.de; j.schubert@uit-gmbh.de;
a.domula@uit-gmbh.de; h.maerten@uit-gmbh.de; y.wendt@uit-gmbh.de; t.rothe@uit-gmbh.de.
2
TU Bergakademie Freiberg, Geotechnical Institute, Freiberg, Germany.
3
VA Erzberg GmbH, Eisenerz, Austria. E-mail: alfred.stadtschnitzer@vaerzberg.at.

246 The Leading Edge April 2024

 A geochemical downhole tool of
this latter kind was used to measure the
complex deposit at the Eisenerz site, an
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

open-pit iron mine near Eisenerz,

Austria. The chief motivation of the
measurement was to assess the tool’s
potential to complement or extend assay
analysis by testing its spectrometric
capability to detect mass fractions of
key elements at the Erzberg deposit.
This is to enhance efficiency in pre-
screening the ore body, offer grade
control before blasting takes place, and
obtain a more detailed picture of the
mineralogy of the site, which is helpful
in planning mining activities.
In the rest of the article, the logging
tool is briefly introduced, the physics of
logging and data analysis are briefly
sketched out, and the results from the
Erzberg campaign are presented.

Description of the borehole logging tool

DOI:10.1190/leedff.2024.43.issue-4

OreLog, a borehole logging tool

that employs pulsed fast thermal neu-
tron analysis (PFTNA), was used. It is
based on the activated prompt fission
neutron tool, which was developed for
uranium exploration more than 10 years
ago (Märten et al., 2014). The PFTNA
tool houses a miniaturized pulsed
deuterium-tritium (DT) neutron gen-
erator (Figure 1). PFTNA, as imple-
mented in the tool, combines time-
resolved neutron detection at various
detector locations (source, central, near,
and far) and in several energy groups
(fast and thermal) with time-resolved
gamma-ray spectrometry (Figure 1a)
(Würz and Buth, 1973). This distin-
guishes between gamma rays from
inelastic scattering of fast neutrons and
thermal neutron capture. This allows
the simultaneous in-situ measurement
of mass fractions of formation elements
(in weight percent [wt%]) including Al,
Ba, Ca, Cl, Cu, Fe, Gd, H, K, Mg, S,
and Si, along with geophysical param-
eters such as bulk density, porosity, and
hydrogen index. The tool can be oper-
ated in passive mode (neutron generator
Figure 1. Principle of the PFTNA tool. (a) The design overview shows the major components. The inset shows the fast neutron off) to acquire spectra of natural gamma
density (red is high and purple is low) from MCNP simulations. (b) The logging setup including the logging van with the operator,
radiation from U, Th, and K, which, for
the winch, and the logging PC. The PFTNA tool is inside the borehole. Here, a tripod was used. For high throughput applications, a
crane or similar means would be used to manipulate the tool. Also shown is a Bonner sphere neutron detector to monitor neutron example, can be used to infer lithology.
dose equivalent radiation for safety. The lower illustration depicts a burst of fast neutrons penetrating the formation. The zoom It can also be operated in active mode
illustrates the subsequent emission of gamma photons from elements in the formation after capturing thermalized neutrons. (neutron generator on) for time-resolved

April 2024 The Leading Edge 247

 prompt gamma-ray spectrometry, which Table 1. Key parameters of the OreLog PFTNA tool.
is used for the measurements presented
in this article. Table 1 summarizes some Measurement principle Time-resolved prompt-gamma count
from neutron scattering and activation
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

key parameters of the PFTNA tool.

The main components of the tool Tool length 3.1 m
are shown in Figure 1a and include the Tool diameter 76 mm
neutron generator, an inertial measure- Tool mass 36 kg
ment unit (to measure azimuth and
Supply voltage 110 or 220 VAC
inclination of the borehole), five 3He
neutron detectors, a detector for fast Power consumption (active mode) 20 W
neutrons, and a gamma-ray spectrom- Neutron source flux 10 s
8 –1

eter with a large CeBr scintillation

3
Neutron cloud range in formation 20 to 50 cm from tool center
crystal coupled to a photo multiplier.
Logging speed 1 to 4 m min–1
All components are housed in a pres-
sure-proven cylindrical housing Maximum logging depth Tested to 70 bars
(Table 1) made of Zr to minimize
interaction with fast neutrons. In active mode, the tool is operated
at a logging speed of several meters per minute. The design is
compact and highly integrative, so the tool fits a variety of borehole
diameters including 96 mm diameter (HQ ). Given that Cl is a
very effective neutron absorber, the borehole must be free of casing
containing Cl such as PVC.
In the following, a brief description of the working principle
DOI:10.1190/leedff.2024.43.issue-4

of the PFTNA tool used here is given. More details can be found
in Zauner (2021). As the tool is pulled upward in active mode,
the DT neutron generator produces a cloud of fast neutrons with
a nominal energy of 14 MeV and a flux of 108 s-1 in pulses with
1 kHz repetition rate and 10% duty cycle. The cloud of fast
neutrons penetrates the formation (Figure 1b). The penetration
depth (inset in Figure 1a) depends on the water saturation in the
formation and borehole, in particular, the H and Cl content. It Figure 2. Capture gamma spectra of various elements from MCNP simulation.
is between 20 and 60 cm (Zauner, 2021). The fast neutrons lose
kinetic energy (are thermalized) within a few microseconds due mechanically manipulates the borehole logging tool (Figure 1b).
to inelastic collisions with formation nuclei such as H (formation A terminal box houses a winch controller to read winch speed,
water), Ca, or Fe (Ellis and Singer, 2008). The elastic collisions direction, and cable tension during logging. It also houses the
in the initial microseconds after the burst give rise to gamma power supply unit for both the winch and the borehole tool and
photons with element-specific energies. In the few hundreds of provides communication between tool, winch, and PC. The winch
microseconds following this initial burst phase, gamma photons and the terminal box are mounted in the logging vehicle (Figure 1b).
with characteristic energies are still being produced, but they The principle of elemental mass fraction retrieval is as follows.
originate from inelastic neutron capture and activation processes. In very simplified terms, the neutrons emitted by the tool are
Gamma photons from all three kinds of processes (burst, capture, allocated among different elements in the formation (e.g., via
and activation) are detected by the scintillator crystal photomul- neutron capture). For a given chemical element involved in the
tiplier arrangement. The signal processing electronics convert the neutron interaction, this gives rise to an emission of a number of
resulting voltage pulses into eight time-lapse gamma count spectra, gamma photons, which is proportional to the corresponding
covering the 1 ms cycle of the neutron generator. Pulse heights elemental volume number density in the probed volume (Hertzog
are converted to gamma photon energies, ranging from 0 to 10 et al., 1989). To determine the mass fraction of each element
MeV, with energy dependent on the element and pulse count rate expected in the formation at a given depth, employing Monte
dependent on the elemental mass fraction in the formation. To Carlo N-Particle Transport (MCNP) simulations, a combination
give a picture of the element-specific spectra, Figure 2 shows of modeled spectra of all expected elements (Figure 2) is fitted to
modeled capture gamma spectra for various elements. In parallel the measured gamma spectra. The MCNP simulations consider
to the gamma acquisition, the neutron detectors (near, central, the geometric aspects and neutron physics, accounting for factors
and far) probe the neutron cloud, yielding time-resolved neutron such as borehole diameter and water saturation as comprehensively
count spectra (32 time bins). as possible. The resulting analysis links the height of peaks in the
Gamma and neutron count spectra are accumulated over 10 s gamma spectrum to the relative yield of elements within the
and 300 ms, respectively, before being transmitted to the logging volume probed by the neutron cloud, thereby estimating the mass
PC via the geophysical wireline inside the steel cable that fraction of each element. Because the tool itself contributes to the

248 The Leading Edge April 2024

 spectrum, a tool-specific spectrum is acquired in a water tank. The Erzberg site
This tool background spectrum is considered in the analysis by Figure 3a shows the location of Erzberg Mountain in Styria,
including it in the fit. The neutron spectra provide information Austria. Figure 3b shows an aerial view of the site, a producing
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

about the neutron dynamic in the formation such as decline time, open-pit mine for iron ore. It represents the largest open-pit ore
which is used to derive the geophysical formation properties. mine in Central Europe and hosts the largest siderite deposits in
The physics of neutrons and gamma photons is complex and the world. The mine is operated by VA Erzberg GmbH, but
contingent upon the geologic composition, in particular the mining activity dates to at least the 17th century. More than
assortment of elements in the formation. For instance, not all 12 million tons of ore and overburden are blasted each year to
capture gamma photons produced will arrive at the detector due produce 3.2 million tons of sinter feed. In the following, the
to scattering or absorption processes. Consequently, establishing geology of the site is briefly summarized.
the relationship between the number of emitted neutrons and the The Paleozoic basement of the Mesozoic platform carbonates
observed gamma counts for a particular element remains uncertain. of the Upper Austroalpine nappe hosts the Greywacke Zone, in
This necessitates calibrating the tool, or more accurately, the data, which the siderite mineralization occurs (Schönlaub, 1982;
for each site. Presently, the PFTNA tool undertakes this calibration Neubauer et al., 1994). The rock series of the Greywacke Zone
by using a limited set of assay data from a representative formation. ranges from carbonates, to metapelites, to metamorphosed acid
Efforts are underway to make tool calibration independent of volcanics and are of Ordovician to Carboniferous age. Coarse-
assay data entirely. grained siderite ore exhibits discordant contacts with the unmin-
The result of the data processing is output in LAS format, eralized limestones, and metasomatic-epigenentic structures are
containing profiles (logs) of weight percentages of detectable dominant. Weak metamorphic overprints and Eoalpine tectonic
elements, geophysical data, tool inertial data, and diagnostic data structures are ubiquitous in the Erzberg deposit. The siderite body
(e.g., neutron generator high voltage). at the Erzberg mine is generally hosted by fine-grained limestones
DOI:10.1190/leedff.2024.43.issue-4

Figure 3. The geography and geology of Erzberg Mountain. (a) Location of Erzberg near the town of Eisenerz, Styria, Austria. (b) Aerial photo of the Erzberg mine from 2022. (c) Vertical view of
the bench face (quarry wall) with tentative interpretation. The red arrows mark the top of the drill holes with their abbreviated ID above, from EBA-001 to EBA-011. Each is covered with a cone.

April 2024 The Leading Edge 249

 of Devonian age. The agreed main mechanism of ore formation
is syngenetic hydrothermal metasomatism (Laube et al., 1995;
Prochaska, 2012). Following dissolution of Mg-rich host rocks,
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

fluids rich in Fe, Mn, and Mg were transported over great distances
via faults and cracks into the Silurian/Devonian limestone base-
ment. There, over time, carbonates underwent ion exchange
reactions, leading to the formation of the current mineralogy of
the iron deposits such as siderite, ankerite, and dolomite with
pockets of quartz and various sulfate minerals. As a result, the
Erzberg deposits are characterized by folded and fractured carbon-
ates, including calcite CaCO3, dolomite CaMg(CO3)2 , ankerite
Ca(Fe, Mn, Mg)[CO3]2 and siderite, specifically Mg-rich siderite
(Fe, Mn, Mg)CO3, and metamorphic siliceous schists. Subordinate
amounts of quartz, hematite, pyrite, chalcopyrite, fahlore, cin-
nabar, and barite also occur.

Logging at Erzberg
Elemental logging with the PFTNA tool at Erzberg took
place from 10 October 2022 to 13 October 2022 in the open-pit
mine at one of the benches named Antoni, shown in Figure 3c.
Figure 3c shows a vertical view of the bench face and a visual
interpretation of the lithology of the section. A fault, possibly a
postgenesis fault, and a geologic contact, which could be syngenesis
DOI:10.1190/leedff.2024.43.issue-4

folding, separate what appear to be three major mineral types:

(1) a calcite-dominated section, (2) a siderite-dominated section,
and (3) an intermediate section with a more complex heterogeneous
mineralogy (approximately between EBA-004 and EBA-007).
Eleven blast holes with a nominal depth of 23 m and diameter
of 152 mm were logged, starting from hole EBA-001 and finishing
with hole EBA-011. Their positions are marked in Figure 3c. All Figure 4. Elemental logs. (a) Logging results (black), repeat run (blue), and assay data (lab,
holes were uncased and drilled every 6 m with an inclination of red) for borehole EBA-005. (b) Scatter plots of Fe and Ca mass fractions from elemental
70°. Following the normal practice at VA Erzberg GmbH, samples logs and assay data.
of drill cuttings were taken every 3 m for each hole, except for
EBA-006. The drill cuttings were dried and processed into pow- assay data (Figure 4b). The mass fractions for Fe vary between
dered pellets that were analyzed with a Malvern Panalytical 0% and 40%, with sharp concentration changes in the upper 10 m.
AxiosmAX XRF spectrometer to obtain elemental mass fractions The tool measured Si mass fractions up to 12 wt%. Although in
for each sampling depth, resulting in assay data. These assay data line with assay data, the agreement for Si is slightly lower than
are used to compare depth profiles of elemental mass fractions for Fe. For Mg, assay data range between 1 and 5.5 wt% and are
obtained with the PFTNA tool. in fair agreement with the logs. Patterns in the depth distributions
The tool was guided into the drill hole by one person, while a are partially reproduced. Reasonable agreement between assay
second person operated the winch and the logging PC from inside data and Mn and K mass fractions is found for most holes including
the logging vehicle (Figure 1b). Once the tool was inserted into EBA-005. The repeatability for K is not as refined as with other
the hole, depth calibration was performed (depth reading was elements. Due to very low concentrations, detecting S was chal-
zeroed). After that, the tool was lowered to the hole bottom, where lenging. Nonetheless, logged elemental mass fractions of S are
the neutron generator was activated. The tool was pulled up and a compatible with assay data.
vertical scan of the formation (logging) commenced. The nominal Figure 5 shows the elemental logs for all logged boreholes in
logging speed used was 1 m/minute. When the tool reached the a cross-sectional mode, revealing the spatial context of correlation
hole top, the neutron generator shut down automatically. and anticorrelation of elemental mass fractions in the form of
spatial coherence. For example, Fe and Ca contents are comple-
Elemental logs mentary. At places where Fe mass ratios decrease, Ca mass fractions
Figure 4a shows a log with mass fractions for all elements increase (e.g., the section from EBA-001 to EBA-004). As with
detected with the PFTNA tool, along with the bulk assay data Fe, the measured Ca mass fractions also sharply vary vertically.
for borehole EBA-005, which is in the middle of the logged bench Both Ca and Fe logs show sharp edges and plateaus. For example,
face section. Also shown is the repeat run, which is coherent with around the middle of hole EBA-002 and near the top of EBA-003
the first run. A variety of structures are reproduced in all holes Fe sharply transitions to Ca. Overall, Ca concentrations decrease
including EBA-005. Fe and Ca mass fractions largely agree with toward the last logged holes (EBA-007 to EBA-011), with about

250 The Leading Edge April 2024

 half of the maximum Ca concentrations measured in the first five Low end-member mass fractions in line with assay data of
or so holes. A relationship between Ca and K can be observed approximately 1 wt% Fe, 2 wt% Ca, and 2 wt% Mn were measured
with neighboring holes. Because Fe and Mn correlate and Fe and (Figure 4a, uppermost 10 m and bottommost 5 m, respectively).
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Ca anticorrelate, Ca and Mn generally anticorrelate. Note that these should not be seen as lower detection limits but
The sensitivity to a particular element in the formation strongly as representative low mass fractions, assuming highly accurate
depends on the type and concentration of that element but also assay data and that the volumes probed by the PFTNA tool and
on the type, concentration, and spatial distribution of other nuclei assay analysis have equal matrix properties.
present in the formation, as well as water saturation and borehole Comparing logs from the PFTNA tool, an in-situ logging
properties (casing, diameter, water saturation, salinity, etc.). Due tool, with assay data is not a direct match. Because the measure-
to this matrix effect, the detection limit of any neutron-induced ment is integrated over a volume of the formation immediately
gamma spectroscopy tool strongly varies with element and mea- surrounding the tool, each point in the log represents a volume,
surement site. At comparable concentrations, elements with a which is a function of logging speed and the number of spectra
high absorption cross section for thermal neutrons will capture averaged. Typically, with a standard configuration (six spectra
more neutrons and thus reduce sensitivity to collocated elements averaged, i.e., 6 × 10 s worth of gamma spectra, 2 m/minute
with a significantly lower absorption cross section (Würz and logging speed), this volume roughly corresponds to a cylinder
Buth, 1973; Hertzog et al., 1989). Elements with a lower absorption of 2 m × 0.4 m². In contrast, the assay data probe a significantly
cross section but higher concentrations may be equally detectable smaller volume of core material of a localized sample from the
as elements with a higher absorption cross section but lower drill-hole cuttings. Given the heterogeneity in mineralogy,
concentrations. For example, the thermal neutron absorption cross this mismatch may in part help explain the disagreements
section of Ca is about six times smaller than for Fe. However, Ca between assay and logging data, particularly evident in the
volume concentrations are high, as indicated by the assay data in middle section of EBA-005 (Figure 4a). To directly compare
Figure 4a, leading to excellent sensitivity to Ca. Mn is present in assay and logging results, in a scatter plot (crossplot), the data
modest amounts only but has a large neutron capture cross section. should correspond to the same location and resolution. This is
DOI:10.1190/leedff.2024.43.issue-4

For this reason, the best agreement between the PFTNA tool not the case. However, to make them more comparable, before
and assay data is expected and observed for Fe, Ca, and Mn. Mg depicting the data in a crossplot, the logs have been downscaled
and Si have lower neutron absorption cross sections and thus are by averaging them into seven bins, which are collocated with
associated with poorer agreement with assay data. the assay bins. The result is shown in Figure 4b. Given the low
number of samples (seven), the linear
least-squares fit, and the R-squared
value, the goodness of fit should be
regarded with care.
The measurement uncertainty of
the XRF analysis (used to produce the
assay data utilized here as a reference)
is a combination of systematic error
(including sample preparation, calibra-
tions standard bias, and matrix effects
of assayed sample volume) and statisti-
cal error (including instrument back-
ground and count uncertainty)
(Rousseau, 2001). For Erzberg, analyz-
ing the same sample multiple times
yielded repeatability, with standard
deviation of repetition below 0.2 wt%
for Fe (less than 0.15% for Ca, less than
0.17% for Si, less than 0.03% for Mg,
and less than 0.02% for Mn), indicating
no significant matrix effects. Adopting
the other sources of uncertainty from
Rousseau (2001), the total uncertainty
of the assay data would be about 1%
(relative). However, it is worth noting
that the standard deviation of repeat-
Figure 5. Qualitative depiction of element mass fractions from the logging tool (30 bin color coded), presented as a tomographic ability was determined from a relatively
cross section through Antoni. Gray tracks depict mass fractions from assay analysis. The farther away they are from the center, small number of samples (three sets of
the higher the mass fraction. core samples analyzed five times each)

April 2024 The Leading Edge 251

 with closely aligned mass fractions Table 2. MAE between the two subsequent runs in wt% for holes with a repeat run.
(e.g., approximately 31 wt% for Fe and
2 wt% for Mn). Upon conducting our Hole ID Fe Ca Si Mg K Mn S
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

own XRF analysis of rock samples EBA-001 N/A N/A N/A N/A N/A N/A N/A
encompassing a diverse spectrum of EBA-002 N/A N/A N/A N/A N/A N/A N/A
elemental mass fractions, we observed EBA-003 1.70 1.35 0.71 0.21 0.25 0.13 0.007
a scattering effect corresponding to
EBA-004 1.26 3.14 0.76 0.18 0.25 0.15 0.009
several wt%. Although the calibration
standards covering the same concentra- EBA-005 1.26 2.32 0.70 0.23 0.30 0.09 0.007
tion range tightly cluster along a line EBA-006 1.10 2.04 0.63 0.14 0.29 0.11 0.009
in a cross-validation exercise (Figure 2 EBA-007 0.90 1.39 0.63 0.13 0.24 0.08 0.08
in Rousseau, 2001). In addition, the
EBA-008 1.40 0.52 0.73 0.14 0.33 0.12 0.01
calibration standards themselves may
have a systematic uncertainty, contrib- EBA-009 0.91 1.08 0.48 0.16 0.25 0.09 0.01
uting to the total uncertainty. EBA-010 1.25 0.8 0.63 0.16 0.25 0.10 0.007
Consequently, this leads us to consider EBA-011 1.53 0.74 0.72 0.18 0.21 0.11 0.007
that the total uncertainty may surpass
the initially estimated 1%.

Logging repeatability and speed performance

The repeatability of two subsequent measurements was deter-
mined by comparing the corresponding logs. A major source of
uncertainty is the manual operation and handling of the tool.
DOI:10.1190/leedff.2024.43.issue-4

Differences on the order of centimeters between runs propagate

into the depth calibration and thus into repeatability. Good
repeatability of measured mass fraction between two repeat runs
is found and is independent of the hole. For example, for Fe,
depending on the hole, the mean absolute error (MAE) is between
0.9 and 1.7 wt% (Table 2).
The time it takes to perform a measurement of a log is an
important commercial parameter of a logging tool and is mainly
dictated by the winch speed (logging speed). To assess its impact
on the measurement accuracy, logging runs at three different
logging speeds (1, 2, and 4 m/minute) were carried out for borehole
EBA-011 (Figure 6a). Logging speed is related to depth resolution.
For the PFTNA tool used here, the nominal accumulation time
per gamma spectrum is fixed at 10 s. To improve count statistics
for the elemental gamma peaks, a standard of six subsequent
spectra (60 s worth of spectral data) are accumulated per point in
the log. This effectively corresponds to sliding window averaging,
reducing the depth resolution of the log. At a logging speed of
1 m/minute, this entails approximately 1 m effective depth resolu-
tion (Figure 6a). With an MAE of 1.26 wt%, the agreement Figure 6. Speed dependance of the Fe log for EBA-011. (a) Accumulation time is six spectra
between the logs for 1 and 2 m/minute (Figure 6a) is comparable (60 s) in a sliding-window manner. (b) Accumulation time is a single spectrum (10 s). Values
to a repeat run (Table 2). At 4 m/minute, the repeatability is still in brackets depict the effective depth resolutions (Res.) resulting from the accumulation of
adjacent spectra and the MAE in wt% relative to a logging speed of 1 m/minute. Lab: assay data.
satisfactory (MAE = 1.96 wt%; Figure 6a), although the maximum
absolute error is 7.4 wt% (at 7.5 m depth). This likely results from
the correlation of the spectra from other depth sections in the used per point, corresponding to a depth resolution of 0.17 m at
sliding window. For a logging speed of 4 m/minute, 4 m of forma- 1 m/minute logging speed. Initially, the higher spatial frequency
tion per point contributes to the log, which effectively leads to a structures (e.g., spikes) indicate a potential reduction in the smear-
smearing of the profile of logged elemental concentration. ing effect and an enhancement in depth resolution. This necessitates
The PFTNA tool used here recently underwent a major a good spectral fit, which in turn needs minimal statistical uncer-
upgrade, resulting in enhanced sensitivity. This means the accu- tainty (good count statistics of the gamma spectra). However,
mulation time per point may be reduced to gain vertical resolution, utilizing a single spectrum, the accumulation time is six times
significantly reducing the smearing effect. The result is shown in shorter, and the count statistics are poorer. When increasing the
Figure 6b, where a single spectrum (10 s accumulation time) is logging speed from 1 to 4 m/minute, one would expect the

252 The Leading Edge April 2024

 averaging effect to reduce effective depth resolution and potentially Therefore, the strong drop of Fe mass fraction around 7.5 m depth
diminish high-frequency structures. Because this is not the case is likely an artifact.
for 2 m/minute at about 7.5 m depth (Figure 6b), this suggests Both the MAE for 2 m/minute (0.47 wt%) and 4 m/minute
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

an issue with the fit itself. In fact, the model spectrum poorly (0.61 wt%) relative to 1 m/minute are smaller than when merging
matched the full energy peaks of Fe near 7.6 MeV (Figure 2). six spectra (Figure 6a) and about half the typical MAE for a
That is, the modeled Fe spectrum underestimates the measured repeat run at the same speed (Fe in Table 2). Given the potential
peaks, leading to underestimated Fe mass fractions. The Fe peaks fitting issues due to low statistics in this case, a smaller MAE
in the capture gamma spectrum for this depth appear to have alone does not necessarily indicate better repeatability.
particularly bad statistics, which could explain the poor fit. In summary, a higher logging speed reduces the effective
depth resolution, while an increase in the number of points in the
log increases vertical depth resolution at the cost of a reduced
number of gamma spectra per point, which decreases count
statistics, possibly leading to artifacts in the log. Based on the
Erzberg case, a single spectra (10 s accumulation time) per depth
point is deemed too short.

Geochemistry of the logged section

Besides Fe (the target element in the Erzberg mine), other
elements crucial for identifying the geochemistry and mineralogy
of the deposits were detected with the logging tool. Overall,
element concentrations from the tool confirm that the deposits
are dominated by carbonates, in particular, ankerite and calcite.
The ternary plot in Figure 7, derived from all 11 logs, shows that
DOI:10.1190/leedff.2024.43.issue-4

samples agglomerate near Fe+Mn (siderite), Ca (calcite), and near

the 50% molar fraction mark between Ca and Fe+Mn, indicating
a significant presence of ankerite. Only two samples were classified
as magnesium carbonates. The occurrence of the solid solution at
Figure 7. Ternary plot derived from elemental mass fractions of the PFTNA logs converted the logged section, notably Ca(Fe,Mg,Mn)(CO3)2 associated with
to molar fractions (see Gittins and Harmer, 1997). CC: calcium carbonate, FC: ferrous CaMg(CO 3) 2 (dolomite), CaMn(CO 3) 2 (kutnohorite), and
carbonate, MC: magnesium carbonate, FCC: ferrous-rich calcium carbonate. Ca(Fe2+,Mg)(CO3)2 (ankerite) mixed with FeCO3 (siderite) means
that specific elements can be swapped to a degree within the
crystal structure and still form a stable
crystal (e.g., Ca with Fe). This leads to
elemental impurities of end-member
minerals (e.g., calcite always has a small
amount of Fe and/or Mg), as evident in
Figure 7.
Figure 8 shows stratigraphic sec-
tions constructed from superposing Fe
and Ca logs on the image of the bench
face. The boundaries of the Fe and Ca
mass fractions are coherent among
neighboring boreholes and with the
stratigraphy. The outline of the geologic
contact in Figure 3c has been adapted
to reflect the step changes in Fe and Ca
mass fractions across the section. The
anticorrelation of Fe and Ca mass frac-
tions as seen in the logging data reflects
the hydrothermal metasomatic miner-
alization, where the host carbonate body
is locally transformed into siderite (and
magnesite) due to ion exchange between
Fe and Ca (and Mg). The decline of Ca
Figure 8. The 30 bin color-coded Ca and Fe logs from Figure 4 overlaying the photo of the bench face (quarry wall). Also shown mass fractions toward the last holes
are the tentative interpretations of the geologic contacts. logged is evident from Figure 8 and in

April 2024 The Leading Edge 253

 line with dominance of iron carbonates (siderite and ankerite) in
that section, as opposed to the calcite-dominated first section (see
also Figure 3c).
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Automatic mapping of elemental logs to mineralogy

The goal is to assign the elemental logs (i.e., depth-resolved
mass fractions) to the corresponding mineral at a given depth,
essentially obtaining mineralogy logs. To that end, a major assump-
tion is made. The cloud of fast neutrons emitted into the formation
penetrates a volume, which strongly depends on the properties
of the drill hole and the formation, but roughly extends a few tens
of centimeters radially and about 50 cm vertically (Figure 1a).
For a given volume, from the borehole logging data alone it is not
possible to assign for example Ca to either the carbonate body or
an ankerite inclusion within the carbonate body. The gamma
photons received by the tool may originate from either one or
both mineral types. Therefore, in the following, it is assumed that
practically the whole of the sampled volume is occupied by a single
mineral species. This, except for mineral transition zones, is reason-
able, given a priori knowledge of the lithology from the drone
imagery of the bench face (Figure 8) and extensive geologic
mapping (Schulz et al., 1997; Prochaska, 2012; Boch et al., 2018).
For an automatic mineral classification of the elemental logs,
DOI:10.1190/leedff.2024.43.issue-4

partial least-squares discriminant analysis (PLS-DA) (Barker

and Rayens, 2003) has been applied. PLS-DA is a version of
partial least-squares regression (PLS). PLS and PLS-DA are
based on the well-known principal component analysis (PCA).
The basic idea of PLS is sketched out in the following in a rough
and simplified way with loss of generality, rather than using rigor-
ous mathematical jargon. The basic idea of both PLS for regression
and PLS-DA used here for classification is to relate a set of
dependent variables (in this case, mineral classes) to a multidi-
mensional set of independent variables (mass fractions of seven
elements, hence seven dimensions) by using a linear regression
model. In very simplified terms, this can be written as

Y = BX + error, (1)

where Y depicts a matrix (in this case, a vector) containing the

dependent variables (N mineral classes) and X depicts a matrix Figure 9. Simplified principle of PLS-DA for a bivariate example. (a) First latent variable,
(dimension: N samples × number of independent variables, i.e., which corresponds to the PC, reducing dimensionality from 2 to 1. The chemical symbols
7). PLS finds a linear combination (PCs or latent variables) of the depict the mass fraction in wt%. (b) Mass fractions labeled with mineral and corresponding
numerical class. In this example, there are only two classes. Also shown is the computed
mass fractions in hyperspace with components contained in B latent variable. (c) Fit of the scores resulting in a trained model. New mass fraction data are
called loadings (Figure 9a), for which the covariance between Y classified by computing the scores and using the model to compute the mineral class. For
and BX (the latent variables) is maximum. The latent variables example, a result of 0.83 would be rounded to 1 and correspond to siderite.
represent the new coordinate axes, and the data points projected
onto them are called scores. Training data are generated using a low confidence. Figure 10 shows the result of a PCA with the
subset of the mass fractions labeled with their known mineral biplot of the first two principal components (PCs) of all samples
class. The labels are replaced with numerical values (0 and 1 in labeled with their corresponding mineralogy, indicating a strong
the bivariate example) (Figure 9b). A linear least-squares fit of a clustering of the main minerals calcite, ankerite, and siderite along
plane in hyperspace to the scores yields the model, which can the first PC. More than 90% of the variance is explained by the
then be used to predict mineral classes for new data (Figure 9c). first two PCs. The consistency with the logs suggests that labeling
For each depth in the Erzberg elemental logs, the ratios of of mineralogy, while being approximative, is valid. This encourages
the mass fractions (e.g., Ca/Fe) and their combination were linked the use of PLS-DA. Trained with the labeled data depicted in
to the most likely mineralogy by using literature data (Lafuente Figure 10, PLS-DA yields mineralogy logs as shown in Figure 11
et al., 2015). K and S logs were excluded due to their relatively for borehole EBA-005, which penetrates the section with the

254 The Leading Edge April 2024

 most diverse mineralogy. The boundar-
ies between mineralogy are consistent
with the interpreted fault and the geo-
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

logic contact (based on visual and ele-

mental information, respectively). At
shallow depths (less than 10 m), the low
Mn/Ca (Figure 4a) and high Ca/Fe
suggest calcite (CaCO3) to dominate at
least in the topmost 7 m. At places,
Mg/Fe of about 3, Si/Fe of about 10,
and high K content (Figure 4a) indicate
clay minerals, which are typically
found in fault structures. Examples
of this are Mg₃[(OH)₂|Si₄O₁₀] (talc),
Al 2 O3 (SiO2) 2 (H 2 O) 2 (kaolinite), and
(K,H 3 O)Al 2 (Si 3 Al)O 10 (H 2 O,OH) 2
(illite). In the intermediate part of EBA-
005 (10 to 13 m), Fe/Mn, Fe/Mg, and
Figure 10. Biplot of all 11 boreholes for Fe, Ca, Mg, Mn, and Si depicting the first two PCs (PC1 and PC2) and the percentage of variance Si/Fe suddenly drop, suggesting a step
they explain. Each sample is labeled with the mineralogy that best corresponds to the relative weight fractions of these elements from change in lithology. The elemental ratios
the literature. Also shown are the loadings (depicted as arrows and in the table in the inset) of the elemental mass fractions. between 10 and 13 m of Fe/Ca from
about 1 to 3, Ca/Mg from about 10 to 5, and the large Ca/Mn greater than 10
suggest the presence of ankerite (Lafuente et al., 2015). Overall, the section
DOI:10.1190/leedff.2024.43.issue-4

below 13 m in EBA-005, up to a depth of 17 m, marks a transition zone between

the original calcite and the siderite mineralization. Fe/Ca of about 3 makes
ankerite more likely than siderite (Figure 11). Below 17 m, Fe/Ca increases to
greater than 30, indicating a dominance of siderite (Lafuente et al., 2015).

Summary and conclusion

Borehole logging tools utilizing gamma spectrometry offer continuous
qualitative and quantitative measurements of key chemical elements. Their
measurement accuracy depends on the unique mineralogic and petrologic
properties of the logged formation. A PFTNA tool using gamma spectrometry
was used to log blast holes in the open-pit iron mine of Erzberg, Austria, and
achieved best accuracies for Fe, Ca, Mn, and Si. Accounting for tool setup, at
a logging speed of 1 m/minute, the effective measurement time was two holes
per hour or about 40 m/hour. The results indicate that a reduction by up to a
factor of 4 is possible without reducing the sensitivity to important elements at
the cost of reduced depth resolution. The swift operation of the logging tool
enables a cross-sectional geochemical image on the order of 100 m to be acquired
within a few hours. This demonstrates high potential for improved ore/waste
discrimination at the Erzberg site and other production mines. The continuous
(spatially integrative) probing of the subsurface offered by the tool enabled
identifying stratigraphic and mineralogic boundaries.
PFTNA logging is furthermore deemed a valuable complement to borehole
assaying. For example, in explorational drilling, the tool can provide quantita-
tive insights into the mineralized zone in near real time, aiding in decisions
regarding whether further analyses, such as XRF assaying for highest accuracies,
are warranted. Such swift provision of geochemical data to geologists on site
can significantly streamline decision making in drilling strategies, saving
valuable time.
The wealth of data enables quantitative statistical analysis and thus a more
detailed picture of the mineralogy of the volume probed. It is valuable in
Figure 11. Classified mineralogy around borehole EBA-005. Also shown
are the outlines of the fault (red dotted), the geologic contact interpreted constraining ore deposit models, including multiparameter models such as
from the photo (white), and the geologic contact as interpreted from the gravity inverse models (Chen and Zhang, 2022). The geochemical constraints
logs in Figure 7 (blue) and the outline of the borehole. provided by the tool may also be useful in carbon sequestration, where

April 2024 The Leading Edge 255

 knowledge of the mass fractions of mineralizing cations (includ- offshore New Jersey, in G. S. Mountain, K. G. Miller, P. Blum, C.
ing Na+, Ca 2+, and Mg2+) in the rock matrix could help constrain W. Poag, and D. C. Twichell, eds., Proceedings of the Ocean Drilling
Program, scientific results, 150, https://doi.org/10.2973/odp.proc.
reactive transport simulations of mineral precipitation from
sr.150.032.1996.
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

captured CO2 and its impact on porosity (Johnson and Nitao, Howell, L. G., and A. Frosch, 1939, Gamma-ray well-logging:
2003; Xu et al., 2011). Geophysics, 4, no. 2, 106–114, https://doi.org/10.1190/1.1440486.
Howell, L. G., and A. Frosch, 1940, Detection of radioactive cement in
Acknowledgments cased wells: Transactions of the AIME, 136, no. 1, 71–78, https://
The authors would like to thank the Austrian Environmental doi.org/10.2118/940071-G.
Agency (Umweltbundesamt) for their kind assistance. Jackson, M. D., A. P. Butler, and J. Vinogradov, 2012, Measurements
of spontaneous potential in chalk with application to aquifer char-
acterization in the southern UK: Quarterly Journal of Engineering
Data and materials availability Geology and Hydrogeology, 45, no. 4, 457–471, https://doi.
Data associated with this research are confidential and cannot org/10.1144/qjegh2011-021.
be released. Johnson, J. W., and J. J. Nitao, 2003, Reactive transport modeling of
geologic CO2 sequestration at Sleipner, in J. Gale, and Y. Kaya, eds.,
Corresponding author: manuelqueisser@web.de Greenhouse Gas Control Technologies: Elsevier, 327–332.
Kleinberg, R. L., W. E. Kenyon, and P. P. Mitra, 1994, Mechanism of
NMR relaxation of fluids in rock: Journal of Magnetic Resonance,
References Series A, 108, no. 2, 206–214, https://doi.org/10.1006/jmra.1994.1112.
Archie, G. E., 1942, The electrical resistivity log as an aid in determining
Lafuente, B., R. T. Downs, H. Yang, and N. Stone, 2015, The power of
some reservoir characteristics: Transactions of the AIME, 146, no.
databases: The RRUFF project, in T. Armbruster, and R. M. Danisi,
1, 54–62, https://doi.org/10.2118/942054-g.
eds., Highlights in mineralogical crystallography: De Gruyter, 1–30,
Barker, M., and W. Rayens 2003, Partial least squares for discrimination:
https://doi.org/10.1515/9783110417104-003.
Journal of Chemometrics, 17, no. 3, 166–173, https://doi.org/10.1002/
Laube, N., H. E. Frimmel, and S. Hoernes, 1995, Oxygen and carbon
cem.785.
isotopic study on the genesis of the Steirischer Erzberg siderite deposit
Boch, R., X. Wang, T. Kluge, A. Leis, K. Lin, H. Pluch, F. Mittermayr,
DOI:10.1190/leedff.2024.43.issue-4

(Austria): Mineralium Deposita, 30, 285–293, https://doi.org/10.1007/

et al., 2018, Aragonite-calcite veins of the ‘Erzberg’ iron ore deposit
BF00196364.
(Austria): Environmental implications from young fractures:
Märten, H., A. Marsland-Smith, J. Ross, M. Haschke, H. Kalka, and
Sedimentology, 66, no. 2, 604–635, https://doi.org/10.1111/sed.12500.
J. Schubert, 2014, Advancements in exploration and in-situ recovery
Borsaru, M., 1993, Nuclear techniques for in situ evaluation of coal and
of sedimentary-hosted uranium deposits: International Symposium
mineral deposits: Nuclear Geophysics, 7, 555–574.
on Uranium Raw Material for the IAEA, IAEA, Extended Abstracts,
Box, R., and P. Lowrey, 2003, Reconciling sonic logs with check-shot
https://doi.org/10.13140/2.1.4322.2400.
surveys: Stretching synthetic seismograms: The Leading Edge, 22,
McDowell, G. M., A. King, R. E. Lewis, E. A. Clayton, and J. A. Grau,
no. 6, 510–517, https://doi.org/10.1190/1.1587672.
1998, In‐site nickel assay by prompt gamma neutron activation
Brown, R. J. S., and B. W. Gamson, 1960, Nuclear magnetism logging:
wireline logging: 68th Annual International Meeting, SEG, Expanded
Transactions of the AIME, 219, no. 1, 201–209, https://doi.
Abstracts, 772–775, https://doi.org/10.1190/1.1820589.
org/10.2118/1305-G.
Minette, D. C., B. G. Hubner, J. C. Koudelka, and M. Schmidt, 1986,
Charbucinski, J., J. Malos, A. Rojc, and C. Smith, 2003, Prompt gamma
The application of full spectrum gamma-gamma techniques to density/
neutron activation analysis method and instrumentation for copper
photoelectric cross section logging: Presented at the 27th Annual
grade estimation in large diameter blast holes: Applied Radiation
Logging Symposium, SPWLA.
and Isotopes, 59, no. 2–3, 197–203, https://doi.org/10.1016/
Neubauer, F., R. Handler, S. Hermann, and G. Paulus, 1994, Revised
S0969-8043(03)00163-5.
lithostratigraphy and structure of the Eastern Graywacke Zone
Chen, T., and G. Zhang, 2022, Mineral exploration potential estimation
(Eastern Alps): Mitt. Österr. Geol. Ges., 86, 61–72.
using 3D inversion: A comparison of three different norms: Remote
Prochaska, W., 2012, Siderite and magnesite mineralizations in Palaeozoic
Sensing, 14, no. 11, 2537, https://doi.org/10.3390/rs14112537.
strata of the Eastern Alps (Austria) field trip guide: Presented at the
Ellis, D. V., and J. M. Singer, 2008, Well logging for earth scientists, second
29th Meeting of Sedimentology, IAS.
edition: Springer, https://doi.org/10.1007/978-1-4020-4602-5.
Rousseau, R., 2001, Detection limit and estimate of uncertainty of
Gittins, J., and R. E. Harmer, 1997, What is ferrocarbonatite? A revised
analytical XRF results: The Rigaku Journal, 18, no. 2, 33–47.
classification: Journal of African Earth Sciences, 25, no. 1, 159–168,
Salazar, J. M., G. L. Wang, C. Torres-Verdín, and H. J. Lee, 2008,
https://doi.org/10.1016/s0899-5362(97)00068-7.
Combined simulation and inversion of SP and resistivity logs for the
Goetz, J. F., L. Dupal, and J. Bowler, 1979, An investigation into dis-
estimation of connate-water resistivity and Archie’s cementation
crepancies between sonic log and seismic check shot velocities: Journal
exponent: Geophysics, 73, no. 3, E107–E114, https://doi.
of the Australian Petroleum Production and Exploration Association,
org/10.1190/1.2890408.
19, no. 1, 131–141, https://doi.org/10.1071/AJ78014.
Schönlaub, H. P., 1982, Die Grauwackenzone in den Eisenerzer Alpen
Hearst, J. R., P. H. Nelson, and F. L. Paillet, 2000, Well logging for
(Österreich): Jahrbuch der Geologischen Bundesanstalt, 124, no. 2,
physical properties: A handbook for geophysicists, geologists, and
361–423.
engineers, second edition: Wiley.
Schulz, O., F. Vavtar, and K. Dieber, 1997, Die Siderit-Erzlagerstätte
Hertzog, R., L. Colson, O. Seeman, M. O’Brien, H. Scott, D. McKeon,
Steirischer Erzberg: Eine geowissenschaftliche studie, mit
P. Wraight, et al., 1989, Geochemical logging with spectrometry
wirtschaftlicher und geschichtlicher Betrachtung: Arch. f. Lagerst.
tools: SPE Formation Evaluation, 4, no. 2, 153–162, https://doi.
Forsch. Geol. B.-A., 20, 65–178.
org/10.2118/16792-pa.
Smith, C. P., P. Jeaneau, R. A. M. Maddever, S. J. Fraser, A. Rojc, M.
Hesselbo, S. P., 1996, Spectral gamma-ray logs in relation to clay mineral-
K. Lofgren, and V. Flahaut, 2013, PFTNA logging tools and their
ogy and sequence stratigraphy, Cenozoic of the Atlantic margin,

256 The Leading Edge April 2024

 contributions to in-situ elemental analysis of mineral boreholes:
Presented at the TOS Forum.
Stewart, R., 2001, VSP: An in-depth seismic understanding: CSEG The journey
begins here
Recorder, 26, no. 7.
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Walls, J., J. Dvorkin, and M. Carr, 2004, Well logs and rock physics in
seismic reservoir characterization: Offshore Technology Conference,
Extended Abstracts, https://doi.org/10.4043/16921-MS.
Williams, J. H., and F. L. Paillet, 2023, Geophysical logging for Explore the Future of Geoscience in China
hydrogeology: The Groundwater Project, https://doi.org/10.21083/
UQGA6966.
Wong, J., L. Han, R. R. Stewart, L. R. Bentley, and J. C. Bancroft,
International Workshop on Gravity,
2009, Geophysical well logs from a shallow test well and automatic Electrical and Magnetic Methods
determination of formation velocities from full-waveform sonic and Their Applications (GEM)
logs: CSEG Recorder, 34, no. 4, 20–29. 19–22 May • Shenzhen, China
Würz, H., and L. Buth, 1973, Theoretische und experimentelle
Untersuchungen zur Lagerstättenprospektion mit Hilfe von neu- 1 st Tarim Ultra-Deep Oil and Gas
troneninduzierter Gammastrahlung: Institut für Neutronenphysik Exploration Technology Workshop
und Reaktortechnik, Projekt Actiniden, Kernforschungszentrum 3–5 June • Korla, China
Karlsruhe, Gesellschaft für Kernforschung MBH, Report KFK
1771. Workshop on Fiber Optics Sensing
Xu, T., L. Zheng, and H. Tian, 2011, Reactive transport modeling for Energy Applications
for CO2 geological sequestration: Journal of Petroleum Science 21–23 July • Xi’an, China
and Engineering, 78, no. 3–4, 765–777, https://doi.org/10.1016/j.
petrol.2011.09.005. Advancing Carbon Capture, Utilization, and
Zauner, M. J., 2021, Deposit characterization based on pulsed neutron Storage (CCUS): East Asia Perspective
induced borehole n-/γ-spectroscopy: Ph.D. thesis, TU Clausthal. 11–13 November • Qingdao, China
DOI:10.1190/leedff.2024.43.issue-4

Current information as of 14 March. Please check

seg.org/events for up-to-date information and more events.

Upcoming Magnetotellurics for Natural

Resources: From Acquisition

Virtual Courses Through Interpretation

23–24 Apr • Karen Christopherson

Distinguished Instructor Short Course (DISC) Near-surface Interferometry

Storage of Carbon Dioxide in Saline 30 Apr–1 May • Sherif M. Hanafy
Aquifers – Building Confidence by
Passive Surface Wave Methods
Forecasting and Monitoring
Using Ambient Noise: From Basic
9–10 Apr • Phil Ringrose
1D Soundings to High-Resolution 3D
Basic Seismic Interpretation Imaging
9–12 Apr • Donald A. Herron and Robert C. Wegner 7–8 May • Koichi Hayashi

The Strategic Opportunity of Introduction to Field Safety Leadership

Sustainability in the Energy Sector 13–14 May • Kevin M. Bohacs
16–17 Apr • Maria Angela Capello
A Practical Understanding of
Inversion for Exploration Geophysics
seg.org/events/events-calendar 28–29 May • John Bancroft
Select ‘Course’ from event format

April 2024 The Leading Edge 257

 Hey members!
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

did you know ?

DOI:10.1190/leedff.2024.43.issue-4

Your SEG membership entitles you to purchase member subscriptions

to SEG Library journals at significant discounts.
Bundled options allow you to select only the publications you need and to save
on Geophysics, Interpretation, and Journal of Environmental
and Engineering Geophysics single-article prices.

Visit seg.org/benefits for options and pricing.

 Visit seg.org and be sure to log in with your e-mail address as your username.
After renewing or joining, visit seg.org/benefits and scroll to the bottom of the page.
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

1 View the available subscriptions, and click to select

the bundle you want to add.

Then (1) click the plus sign to increase the quantity

2
DOI:10.1190/leedff.2024.43.issue-4

to 1, (2) add it to your cart, and (3) view your cart.

3 Click to checkout.

Enter your payment information.

Click to continue.
4

5 Click Pay Now.

 Reviews
C o o r d i n at e d by Julie Aitken
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

Earth’s Core: Geophysics of a Planet’s Deepest Interior, by Vernon F. as machine learning (ML) algorithms that are designed to process
Cormier, Michael I. Bergman, and Peter L. Olson, ISBN 978- today’s enormous data sets.
0-128-11387-5, Elsevier, 2021, 324 p., US$98 (print), $98 (e-book). The beginning chapters of the book review the main steps
in remote sensing, starting with an analysis of active-source

I f I were naming this volume, I would call it “From core to

crust” because that is what it covers. I was very impressed by
the breadth and depth of the different earth-science topics. I
(e.g., reflected sunlight) and passive-source (e.g., thermal emis-
sion) electromagnetic wavebands. Richards describes how each
band penetrates the atmosphere before being recorded and
have rarely seen a volume that covers so many different elements analyzed in terms of specific surface characteristics such as
of the earth sciences including gravity and magnetics, crystology, vegetation type. Each of these steps is hampered by noise and
and seismology. errors that require corrections, some of which have analogies in
I would recommend this volume to the broader earth-science exploration geophysics. Corrections include spectral absorption
community. The author’s approach is to cast a data net as wide as compensation, geometrical adjustments for earth rotation, and
possible over the earth sciences and draw it into a planetary model reconciliation for instrumentation defects or radiation patterns.
for the earth’s core. Unlike the predominantly scalar seismic images in exploration
There are nine high-quality chapters spanning 324 pages geophysics, images produced in remote sensing display several
in this 11 × 8 inch volume. The color figures are well rendered, different wavebands at each pixel simultaneously. Besides discuss-
but some of the black and white figures are less so. However, ing more conventional image processing methods, Richards
this is rather minor. Mathematics are presented without deriva- describes how to extract information from multivalued pixel
tion. Each chapter ends with a references section. I would have data using spectral domain image transforms. For example,
DOI:10.1190/leedff.2024.43.issue-4

preferred that all of the references were put at the end of the spectral transforms allow information to be condensed into a
volume. Four chapters have an appendix, and some others have color representation for photointerpretation, that is, visual
a further readings section. analysis by skilled analysts.
I would like to call out chapter 3, titled “Geodynamo geo- Richards next addresses classic and advanced artificial
magnetic basics.” It summarizes both the present and paleomag- intelligence-based methods. Machine-assisted classifiers have
netic fields, as well as “geomagnetic jerks” and field reversals on perhaps been the area of greatest advances in remote sensing image
several time scales. analysis over the past several decades. They perform pixel-reso-
lution, multiband, quantitative analysis by using supervised,
— Patrick Taylor unsupervised, and hybrid methods. The meticulous discussions
Davidsonville, Maryland of the underlying mathematics and statistics of the considered
ML and deep learning (DL) algorithms is impressive. However,
the highlight of the textbook is the chapter on best practices that
Remote Sensing Digital Image Analysis, by John A. Richards, ISBN make ML/DL methodologies more robust, efficient, and accurate
978-3-030-82329-0, Springer, 2022, 567 p., $69.99 (print), $54.99 for field data. Many of these useful hints could potentially be
(e-book). applied to automated seismic data processing of vectorized data
sets, such as those generated by seismic experiments of distinct

R emote sensing refers to geospatial technologies that aid in

extracting information from objects at or near the earth’s
surface. Remote sensing and geophysics are both branches of
geometries and frequency content generated in 4D projects, or by
filtering of distinct frequency bands.
This is a complete textbook on remote sensing, enriched and
earth science, with the main difference being that in the former, battle hardened by generations of students with engineering,
sensors are located in space and the energy recorded is delivered computer science, and mathematics backgrounds. Its hands-on
in sets of electromagnetic wavebands. Thanks to an ever-expanding style focuses on modern methodologies, error analysis, and the
number of satellites roaming above us, remote sensing is an practical skills required for efficient and accurate image analysis
increasingly important technology for applications in agriculture, and interpretation. The mathematical material in the later chap-
resource exploration, weather forecasting, mapping of environ- ters may represent a challenge to some readers and could poten-
ments, and more. tially be moved into the appendix section. Otherwise, the book
When learning about a related discipline, nothing beats is very approachable. Its illustrations are well done, and engaging
starting with a classic textbook written by a distinguished authority examples and exercises reinforce the key concepts. The magnitude
in the field. Professor John Richards of both the Australian of enticing analogies to technologies in our familiar seismic
National University and University of New South Wales published imaging worlds makes this book fascinating for TLE readers.
the first of six editions of this book in 1986. As with the 2022
edition reviewed here, Richards rigorously keeps the material — Andreas Rueger
current and continually incorporates new scientific advances such Golden, Colorado

260 The Leading Edge April 2024

 The Leading Edge
Membership
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

A pplications for Active membership have been received from

the candidates listed below. This publication does not consti-
tute election but places the names before the membership at large
Requirements for membership
Active: Eight years of professional experience practicing
or teaching geophysics or a related scientific field. Mem-
in accordance with SEG’s Bylaws, Article III, Section 5. If any bership applications and details of other types of member-
member has information bearing on the qualifications of these ship, including Associate, Student, and Corporate, may
candidates, it should be sent to the SEG president within 30 days. be obtained at https://seg.org/membership.

For Active membership Kumar, Pankaj (India)

Albertin, Uwe (USA) Kumar, Sandeep (India)
Alomar, Reem (Saudi Arabia) Lall, David (India)
Arboleda Zapata, Mauricio (Germany) Lee, Dawoon (Republic of Korea)
Biharie, Dheeraj Ritesh (Suriname) Macelloni, Leonardo (USA)
Boden, John (USA) Maiti, Saumen (India)
Buchanan, Bruce (USA) Mishra, Satyabrata (Australia)
Buchanan, Matthew (USA) Montealegre-Cazares, Conrado (Mexico)
Candy, Cliff (Canada) Natarajan, Dhanvanthri (USA)
Carpentier, Stefan (Netherlands) Nelson, Nwoyi (Nigeria)
Chandoe, Shahied (Suriname) O’Connell, Kevin (USA)
DOI:10.1190/leedff.2024.43.issue-4

Dadwal, Eena (India) Oni, Ademakinwa (Nigeria)

David, Sandrine (Norway) Osumeje, Joseph (Nigeria)
de Seil, Carol (Suriname) Oyediran, Ibrahim (Nigeria)
Diaz, Laura (USA) Pan, Jianwei (China)
Dunlop, Rick (Australia) Pelemo-Daniels, Oyedoyin (USA)
Eckroth, Joshua (USA) Raghuvanshi, Harsha (India)
Elliott, Julia (USA) Ramdhiansing, Winitadevi (Suriname)
Endres, David Mack (USA) Ray, Jordan (USA)
Falebita, Dele (Nigeria) Roy, Somali (India)
Gadela, Dharma (India) Sadeghian, Rocco (UK)
Gangaram-Panday, Payal Vaishnavi (Suriname) Sahoed, Kimberly (Suriname)
Ghasemichimeh, Mohammadreza (Turkey) Schoen, Eric (USA)
Gilbert, John Rusty (USA) Schultz, Whitney (USA)
Goebel, Thomas (USA) Singh, Jagmeet (India)
Hincapie, Jaime (USA) Singh, Ram (India)
Holmes, Dylan (USA) Strong, Shaun (Australia)
Hough, Lannie (USA) Tata, Santaram (UK)
Huang, Guangtan (China) Tiwari, Dhananjay (USA)
Huang, Mengqi (USA) Vadapalli, Uma (India)
Ismail, Afsal (India) Wang, Rosalynn (USA)
Jayasangar, Thivyaadarshini (Malaysia) Wawiye, Francis (Kenya)
Johnson, Elizabeth (USA) West, Richard (Canada)
Khangembam, Nabakumar (India) Zhang, Mengli (USA)
Kiraz, Mert (USA) Zhou, Jessie (Australia)
Kisoensingh, Maneesha (Suriname)

April 2024 The Leading Edge 261

 The Leading Edge
Meetings Calendar
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

See https://seg.org/events/events-calendar for the latest updates.

APRIL 2024 21–23 MAY

Recent Advances in Geophysical Reservoir
JULY 2024
9–10 APR Characterization and Monitoring of CO2 3–4 JUL
8th SEG Virtual Student Conference (Europe) Sequestration in Carbonate Reservoirs Advanced Geophysical Solutions for Complex
https://seg.org/calendar_events/8th-seg-virtual- https://seg.org/calendar_events/co2- Geological Settings
student-conference-europe sequestration-carbonate-reservoirs-2024 https://seg.org/calendar_events/advanced-
Virtual Abu Dhabi, UAE geophysical-solutions-for-complex-geological-
settings
23–24 APR 22–23 MAY Kuala Lumpur, Malaysia
Borehole Geophysics: New Developments for Carbon Capture, Utilization, and Storage Brazil
Global Energy https://ccusevent.org/latinamerica/2024 14–19 JUL
https://seg.org/calendar_events/borehole- Rio de Janeiro, Brazil Geophysical Research for Gigatonnes
geophysics-for-exploration-development-ccs CO2 Storage Workshop
Kuala Lumpur, Malaysia
JUNE 2024 https://seg.org/calendar_events/geophysical-
research-toward-gigatonnes-co2-storage
23–25 APR 3–5 JUN Golden, Colorado, USA
Next Generation Land and Shallow Water SEG 1st Tarim Ultra-Deep Oil & Gas Exploration
Acquisition Workshop Technology Workshop 21–23 JUL
DOI:10.1190/leedff.2024.43.issue-4

https://seg.org/calendar_events/new-generation- https://seg.org/calendar_events/2024-seg-1st- 2024 SEG Workshop on Fiber Optics Sensing for

aquisition-2024 tarim-ultra-deep-oil-gas-exploration-technology- Energy Applications
Muscat, Oman workshop https://seg.org/calendar_events/2024-seg-
Korla, China workshop-on-fiber-optics-sensing-for-energy-
30 APR–2 MAY applications
Summit on Geophysical Detection of Explosive 4–5 JUN Xi‘an, China
Remnants of War: Solving Current Challenges of SPE/SEG Workshop: Injection Induced
Unexploded Ordnance (UXO) and Demining
https://seg.org/calendar_events/summit-on-
Seismicity
https://www.spe-events.org/workshop/injection-
AUGUST 2024
geophysical-detection-of-explosive-remnants-of- induced-seismicity 26–29 AUG
war-solving-current-challenges-of-unexploded- Fort Worth, Texas, USA International Meeting for Applied Geoscience &
ordnance-and-demining Energy (IMAGE)
Virtual 17–19 JUN https://www.imageevent.org/image-2024
Unconventional Resources Technology Houston, Texas, USA

MAY 2024 Conference (URTeC)

6–9 MAY
https://urtec.org/2024
Houston, Texas, USA
SEPTEMBER 2024
Offshore Technology Conference 11–13 SEP
https://2024.otcnet.org 23–27 JUN SEG Workshop on Reservoir Geophysics
Houston, Texas, USA 20th International Conference on Ground Website coming soon
Penetrating Radar (GPR 2024) Qingdao, Shandong, China
19–22 MAY https://gpr2024.jlu.edu.cn/webinfo/viewenglish
GEM 2024 Shenzhen: International Workshop on
Gravity, Electrical, and Magnetic Methods and
Changchun, China
OCTOBER 2024
Their Applications 24–27 JUN 8–10 OCT
https://seg.org/calendar_events/gem-2024 Net-Zero Emissions Workshop SEG | SPWLA Symposium: Seismic Petrophysics
Shenzhen, China https://seg.org/calendar_events/net-zero- https://seg.org/calendar_events/seismic-
emissions-workshop petrophysics
Virtual Al Khobar, Saudi Arabia

262 The Leading Edge April 2024

 21–24 OCT
AAPG/EAGE/SEG: GEO 4.0 – Digitalization in
NOVEMBER 2024 18–20 NOV
International Geomechanics Conference
Geosciences Symposium 4–6 NOV https://www.igsevent.org
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

https://geo4event.com 4D Forum – Insight to Actions: Creating Value, Kuala Lumpur, Malaysia

Al Khobar, Saudi Arabia Reducing Cycle Time, and Optimizing Production
and Injection in a Digital World 19–21 NOV
28–31 OCT https://seg.org/calendar_events/4d-forum- Deep Reservoir Imaging and Characterization
Summit on Drone Geophysics insight-to-actions-creating-value-reducing-cycle- https://seg.org/calendar_events/deep-
https://seg.org/calendar_events/2024-summit- time-and-optimizing-production-and-injection-in- reservoir-2024
on-drone-geophysics a-digital-world Al Khobar, Saudi Arabia
Virtual Galveston, Texas, USA

22–24 OCT 5–7 NOV

DECEMBER 2024
Advances in Quantitative Interpretation Challenges and Advances in Velocity Model 3–5 DEC
https://seg.org/calendar_events/advances-in- Building 3rd SEG/EAGE Workshop on Geophysical Aspects
quantitative-interpretation-3 https://seg.org/calendar_events/velocity-model- of Smart Cities
Kuala Lumpur, Malaysia building-2024/ Website coming soon
Muscat, Oman Seoul, South Korea

11–13 NOV 10–12 DEC

Advancing Carbon Capture, Utilization, and Addressing the Challenges in Processing Data
Storage: East Asia Perspective for Land and Shallow Water in the Middle East
Website coming soon Website coming soon
DOI:10.1190/leedff.2024.43.issue-4

Qingdao, Shandong, China Abu Dhabi, UAE

Optical fiber
Distributed temperature, strain, acoustic,
and pressure sensing (DxS) technologies
revolutionize the way we monitor and
analyze temperature, acoustic, and pres-
sure signals over large distances. Unlike
traditional sensors, DxS utilizes optical
fibers as sensitive elements, transform-
ing them into continuous high-resolution
distributed sensors. The editors of an
upcoming TLE special section on optical
fiber welcome submissions pertaining to the latest developments
in DxS and optical fiber related to the geosciences, the oil and gas
industry (including underground pipeline monitoring for leaks and
intrusions), and railway, highway, and building safety.

Submission deadline: 15 June 2024

Publication of issue: November 2024
Submission portal: https://mc.manuscriptcentral.com/tle
Information for authors: https://library.seg.org/page/leedff/ifa

Special section editors:

Joël Le Calvez (assistant editor): jcalvez2@slb.com
Erkan Ay (associate editor) : erkan.ay@shell.com

April 2024 The Leading Edge 263

 S e i s m i c S o u n d o ff — C o o r d i n a t e d by Andrew Geary
The untapped potential of the earth’s hidden commons
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

I n this thought-provoking episode, expert geoscience com-

municator Iain Stewart explores the “hidden commons” of the
subsurface. Stewart shares his vision of the subsurface as a new
Stewart: Yes. What really perplexes
me is we have this fascination for
space. We stare up at the heavens and
frontier, not just for resource exploitation but as a space for sustain- look at that starry sky. Here’s the
able development and urban innovation. This episode gives geo- thing: it is broadly empty up there.
scientists a new language to describe the importance of their work The planets that we look at are mil-
to the world. Listen to the full episode at https://doi.org/10.1190/ lions of miles apart. We will never
seismic-soundoff-episode214. reach most of them. Yet, the irony is
that underneath our feet is a bunch
Andrew Geary: What do you mean by the term hidden of stuff that we can access that is
commons? useful. So, why does space have that
emotional appeal? How can we trans-
Iain Stewart: The idea of the tragedy of the commons is a very late that into an emotional appeal for the subsurface? If we have
obvious and well-rehearsed story about resource availability and a bit of imagination, we can start to sell the idea of what would
exploitation. For me, the hidden commons is the subsurface. One happen if we took a fraction of the budget to go down into the
thing that I think geoscience has been poor at exploiting is that earth. I think selling the idea that if you become a geoscientist
we have a unique domain that no one else touches — the subsur- in the 21st century, the frontier is not up there in the sky but
face. We often fight over the surface domains. A lot of the issues underneath your feet. If we can find narratives and stories for
that geoscience talks about deal with the surface and very near- that, then we make geoscience a much more attractive option.
DOI:10.1190/leedff.2024.43.issue-4

surface environment. However, the subsurface world is ours. No

one else wants it. That’s the problem. Geary: You mention that geoscientists deal with the
Ownership of the subsurface is an interesting debate. Who dilemma of being the earth’s stewards. They are trying
owns the subsurface as you go down a few kilometers? There’s the to find oil and gas and use them well. They are also exploit-
old idea that if you own a parcel of land, you own from the heavens ing those by mining and getting the oil and gas out of
above to the center of the earth. Whenever that’s tested in court, the subsurface. What do you think about this?
it’s thrown out. You cannot own something that’s 1000 km down.
You can own something a few meters down. Stewart: They are end members of a continuum. I think that one
The hidden commons is a huge resource. As we move into extreme thought is we are nasty people who extract stuff from
the future and think about sustainable development and all its the earth. At the other end, there is the notion that we are guard-
different dimensions, the parts of the subsurface that are going ians of the planet because we have spent all of our time studying
to be critical is the realm of geoscientists. Therefore, we should its 4.5 billion year history. We know how it operates as a system.
be exploiting our knowledge of the hidden commons. The chal- Your average geoscientist probably sits somewhere in the middle.
lenge is to justify and sell that to the public and politicians. Over the course of their career, they may move from one area into
In many parts of the world, the future of cities is probably another. I do not know a geoscientist who doesn’t care about the
going to lie in the deep subsurface. As climate changes and earth. Even people who work in the front line of oil exploration
places get warmer, more of life’s activities are going to start and mineral exploration do not want to see themselves as polluters
happening in the subsurface of cities. If you are building in or despoilers of the earth because that’s the reason they got in-
the subsurface of cities, you need geoscientists. The hidden terested in the first place. However, they carry a responsibility to
commons is that notion of a new frontier that is completely deliver stuff for society.
unique to ourselves. My take is that if we can push and emphasize the fact that
we are earth’s stewards, even if we are resource exploiters, we
Geary: You talk about building empathy for the subsur- want to do so in a sustainable way. If we can convey this notion
face. Do you think that is the proper direction for geo- that we are stewards or guardians, that is a more positive and
scientists to get people to care about what is happening progressive view than we are just there to take stuff from the
in these hidden commons? earth. I don’t think that is a tenable viewpoint in this part of the
21st century.

264 The Leading Edge April 2024

Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

www.z-terra.com

Beam Tomography
A fast and stable alternative to Full Waveform Inversion (FWI)
DOI:10.1190/leedff.2024.43.issue-4

Z-Terra North Inc. Z-Terra Inc. Z-Terra BV

2500, 715– 5 Avenue SW 17171 Park Row, Suite 247 Stationsplein 6, 5th Floor
Calgary, Alberta T2P 2X6 Houston TX 77084 The Hague
Canada USA Netherlands

Argentina Brazil SE Asia Middle East Colombia Eastern Europe

Av. Lima 575, 8th floor Av. Nilo Pecanha 50 Phileo Damansara 1, NW Sulaibikhat Area Carrera 14B Dr. Anibal Teoharie Street nr. 9
Buenos Aires, Argentina Edificio De Paoli, Sala 1616 Petaling Jaya 9, Jalan 16/11 Block 3, Building 25 Bogota, Colombia Bucharest, Romania
Rio de Janeiro, RJ Brazil Kuala Lumpur, Malaysia Shuwaikh KW Kuwait
Downloaded 07/23/25 to 122.161.72.155. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/page/policies/terms

528 & VE564 TM TM

EXPLORE OUR ADVANCED SOLUTIONS

DOI:10.1190/leedff.2024.43.issue-4

FOR EFFICIENT LAND SEISMIC ACQUISITION

SOLVING YOUR TOUGHEST

OPERATIONAL AND IMAGING
CHALLENGES
Our most recent land solutions have been engineered to increase
productivity, reduce costs, and ensure the success of seismic projects.

• Unparalleled reliability
• The lowest power consumption and weight
• Innovative productivity dashboards
• Optimized vibrator fleet management
• Unmatched data integrity and data quality
• Supported by our unique portfolio of land services

Nantes, France
sales.nantes@sercel.com
Houston, USA
#TrueBroadband sales.houston@sercel.com
For more information #TrueFidelity www.sercel.com