Carl Heiland Lecture Series
In Spring 2025, the Carl Heiland Lecture Series will be on Wednesdays from 12:00-12:50 PM in Hill Hall 202. Each week, we will be joined by a distinguished speaker from academia, industry, or government on a topic pertinent to the geosciences. The lecture series is a public event open to all members of the Mines community and beyond.
Some Heiland lectures will be offered in person on campus and others will only be offered virtually via Zoom. A Zoom link will be available for all presentations so that you can attend from wherever you are in the world.
January 15 - Nicholas Holschuh
A Complex Ice Flow History in Northern Greenland Inferred from Novel Ice Penetrating Radar Data
Nicholas Holschuh
Amherst College
In person: HH 202; Zoom webinar link: https://mines.zoom.us/j/95432461641
Abstract
Radio-echo sounding data from some of the earliest geophysical surveys in Greenland captured anomalous structures deep within the Greenland ice sheet. There has been active debate about how these structures form, with most studies assuming they comprise deformed, meteoric (i.e., glacial) ice. But these studies were only informed by observations of structure geometry from radar. In this talk, I will demonstrate some of the advanced imaging capabilities of modern ice-penetrating radar, now capable of capturing the 3D-scattering distribution from dielectric contrasts within the Greenland Ice Sheet. These new data show that many (but not all) of the previously identified structures appear to be capped by debris, scattering radio waves in a way that is qualitatively different from typical glacial layering. I will use these data to challenge existing models of structure formation, and identify new questions about how ice, rock, and water interact at the base of glaciers.
Speaker Bio
Dr. Nick Holschuh is an assistant professor of geology at Amherst College. His primary research interest is in improving our understanding of the physics of glacier sliding using observational geophysics. That work has taken him to Antarctica four times, most recently to Thwaites Glacier, where he has used ice-penetrating radar, active and passive seismic techniques, and gravimetry to measure the glacier subsurface.
January 22 - Joanna Millstein
How Observations Improve Models of Iceberg Calving
Dr. Joanna Millstein
Colorado School of Mines
In-person: BE 243; Zoom link: https://mines.zoom.us/j/95432461641
Abstract
Fractures in glacier ice, from crevasses at the glacier surface to icebergs calving at the continent’s edge, compromise the structural stability of glaciers and ice sheets, contributing to their mass loss. The current mechanical understanding of fracture initiation and propagation in glaciers is notably lacking in observationally validated and empirically derived models, largely due to a scarcity of data over relevant spatial and temporal scales. Here, I will describe two interconnected research projects that strive to reconcile the limited understanding of fracture propagation with the need for realistic numerical models of ice sheet change. The first project analyzes major Antarctic calving events through four decades of observational iceberg data, employing extreme value theory to develop a statistical model. Our statistical model captures the stochastic nature of iceberg calving and establishes recurrence intervals over which we expect to see icebergs of varying sizes. The second project implements this framework within the Community Ice Sheet Model (CISM), a numerical framework designed for simulating large-scale ice sheet dynamics. By integrating physics-based calving criteria with a stochastic parameterization, the model successfully reproduces observed patterns of sudden change in Antarctica’s ice sheets. These complementary approaches establish an exciting path for determining and assessing future changes to the Antarctic ice sheet.
Dr. Joanna Millstein is a Postdoctoral Fellow in Geophysics at Mines. Her research investigates the rheology, fracture, and stability of glacier ice and ice sheets. She uses observational data to derive statistical and mechanical models that describe the physical processes of ice. When she isn’t working with the Mines Glaciology Laboratory in the Green Center, she is an instructor for an Arctic Geophysics field course at The University Centre in Svalbard.
March 5 - Dave Monk
The Evolution of Seismic Acquisition and Artificial Intelligence
Dave Monk
ACTeQ
In-person: HH 202 ; Zoom link: https://mines.zoom.us/j/95432461641
Abstract
Artificial Intelligence (AI) is viewed by many as having a huge impact on geophysics in general. It is the “lens” of future geophysics. In this presentation I look backwards at the history of seismic acquisition and show how the technology has changed, and I illustrate some of the short comings of commonly used AI systems in use today. While A.I. may be useful in some applications where solutions or answers are not necessarily based on scientific data, some solutions often fall short of being accurate or correct.
Speaker Bio
Dr. Dave Monk holds a PhD in Physics from Nottingham University in the UK, and served as director of geophysics at Apache Corporation, until his retirement in October 2019. Author of more than 200 technical papers, he is a past president of the SEG. He currently serves as a technical advisor for ACTeQ (a seismic survey design software company which he co-founded), GTI (a seismic node manufacturer). He recently took a position on the Board of Directors for DUG, a technology company based in Australia.
March 12 - Andrei Swidinsky
Stochastic Control of Natural Resource Operations using Geophysical Measurements and Deep Reinforcement Learning: Fundamental Concepts and a Few Illustrative Examples
Andrei Swidinsky
University of Toronto
In-person: HH 202; Zoom link: https://mines.zoom.us/j/95432461641
Abstract
Stochastic control is concerned with optimal decision-making in the presence of uncertainty. One approach to solving such control problems is deep reinforcement learning (DRL), which, along with supervised and unsupervised learning, is one of the three main categories of machine learning. DRL agents learn optimal behavior through repeated interaction with their surrounding environments, and have achieved human-level performance in challenging domains such as algorithmic trading, navigating self-driving cars and playing classic board & video games. To date, DRL has only been sparingly used for natural resource projects but because applied geophysicists – and more generally, applied geoscientists – acquire data to make various engineering and business decisions, I posit that the machinery of DRL is an ideal match for the exploration, development, production and/or management of subsurface resources.
One key feature that differentiates DRL from other machine learning paradigms is that “time really matters”, meaning that decisions are made on-line according to the current situation (such as each move on a chess board). Various geoscientific applications like geothermal energy production, groundwater management or geological carbon storage can benefit from such on-line sequential decision-making for optimal well placement or flow control, and time-lapse geophysical measurements – like gravity, electromagnetics or seismic – can be key inputs for this process. Likewise, optimal exploration strategies for net-zero technology-enabling commodities such as critical minerals also require sequential decisions concerning what type of data to collect as exploration proceeds. In this talk, I will outline the efforts of GeoDecisions, my research group at the University of Toronto, in applying DRL to subsurface resource problems: first through fundamental concepts and subsequently by a few illustrative optimal control examples.
Speaker Bio
Dr. Andrei Swidinsky completed his undergraduate education in theoretical physics at the University of Guelph, and his graduate studies in geophysics at the University of Toronto. Upon finishing his doctorate in 2011, he spent the following two years as a postdoctoral research fellow at the Helmholtz Centre for Ocean Research Kiel (Geomar). From 2013–2021, Swidinsky was a faculty member at the Colorado School of Mines, Department of Geophysics (as an assistant professor from 2013–2019 and as a tenured associate professor from 2019–2021). Since July 2021, he has been an associate professor and the Teck Chair in Exploration Geophysics at the University of Toronto, Department of Earth Sciences.
March 26 - Sean Bader
April 2 - Erin Wirth
April 16 - Luca Duranti
April 23 - Susanne Ouellet
Meaningful monitoring of Tailings Dams with Distributed Acoustic Sensing
Susanne Ouellet
Lumidas
In-person: HH 202; Zoom link: https://mines.zoom.us/j/95432461641
Abstract
Tailings dams are designed and built to retain waste generated from mining. The safety of tailings dams requires comprehensive monitoring solutions that can detect subtle precursors to potential failure modes. Distributed acoustic sensing (DAS) is an emerging fiber-optic sensing technology providing real-time monitoring capabilities across extensive areas with nanostrain sensitivity. By implementing complementary signal processing algorithms, DAS technology can detect strain, temperature, and seismic perturbations through Rayleigh backscattering of light along tens of kilometers of cable, making it particularly suitable for monitoring dams and auxiliary infrastructure. In 2019, nearly six kilometers of fiber-optic cable were installed at ~1 m depth along an active upstream tailings dam in northern Canada. Energy from the ambient seismic wave field was used with DAS to infer changes in shear wave velocities of up to ~2%, corresponding to springtime thaw and rainfall. In a separate study, we demonstrate how the same technology, with different signal processing techniques, can be used to detect submillimeter near-surface slope displacements, correlating with nearby geotechnical instrumentation. This dual monitoring capability, combined with high spatiotemporal resolution, enables earlier detection of potential instabilities, supporting proactive dam safety management and risk mitigation.
Speaker Bio
Dr. Susanne Ouellet is the founder of Lumidas, a cleantech startup specializing in advanced geotechnical monitoring of critical infrastructure using distributed-fiber-optic sensing technologies. Dr. Ouellet’s research at the University of Calgary focused on advancing tailings dam performance monitoring using DAS. She is a registered professional geotechnical engineer in Alberta with over ten years’ experience in research and consulting. Her background gives her a deep insight into the practical challenges and technological needs of the industry, insights she is now channeling into Lumidas.
April 30 iury Simões-Sousa
iury Simões-Sousa
Woods Hole Oceanographic Institute
In-person: HH 202; Zoom link: https://mines.zoom.us/j/95432461641
