Geosciences Friday Seminars

Fall 2021

Time and Location

Day: Fridays

11:45 am

Location: Contact instructor for details.

Contact Information

Ronni Grapenthin


Date Presenter/Topic  

August 27

Ronni Grapenthin, UAF,  Department of Geosciences and The Geophysical institute
Seminar Overview


September 3 UAF Faculty and Students
Lightning Talks                                                                            

September 10

Ezgi Karasozen, UAF, Geophysical institute
"How can we Better Assess Earthquakes"     

As earthquake seismologists, our goal is to constantly improve our understanding of earthquakes, and related seismic hazards. Although it sounds like a trivial task, we often face challenges that persistently require developing novel strategies. It is sometimes a noisy station, or it can be a limiting assumption in a routine technique that we frequently adopt. Sometimes it is a puzzling earthquake sequence that requires a more in-depth analysis. In this talk, l present examples from case studies that span a variety of these challenges from all over the world: a seismic array in Turkey that suffers from anthropogenic noise, a mountain-wide orogeny in Iran that suffers from earthquake location errors, and an interesting earthquake sequence from Alaska that suffers from unfavorable data coverage. These studies highlight the importance of data availability and variability, the need to constantly developing and improving techniques, and the value of effectively combining different datasets and techniques. Overall, our results show how we can employ different strategies to transform our understanding of the kinematics of faulting, to better assess earthquakes, and also to improve interpretations of the associated seismic hazard.

September 17

Diego Melgar, University of Oregon

"Early Warning for Great Earthquakes from Characterization of Crustal Deformation Patterns with Deep Learning"


Although infrequent, large (Mw7.5+) earthquakes can be extremely damaging and occur on subduction and intraplate faults worldwide. Earthquake early warning (EEW) systems aim to provide advanced warning before strong shaking and tsunami onsets. These systems estimate earthquake magnitude using the early metrics of waveforms, relying on empirical scaling relationships of abundant past events. However, both the rarity and complexity of great events make it challenging to characterize them, and EEW algorithms often under predict magnitude and the resulting hazards. Here we propose a model, M-LARGE, that leverages the power of deep learning to characterize crustal deformation patterns of large earthquakes in real time. We demonstrate the algorithm in the Chilean Subduction Zone by training it with more than six million different simulated rupture scenarios recorded on the Chilean GNSS network. M-LARGE successfully performs reliable magnitude estimation on the testing dataset with an accuracy of 99%. Furthermore, the model successfully predicts the magnitude of five real Chilean earthquakes that occurred in the last 11 years. These events were damaging, large enough to be recorded by the modern HR-GNSS instrument in the last decade, and provide valuable ground truth. M-LARGE tracks the evolution of the source process and can make faster and more accurate magnitude estimation, significantly outperforming other similar EEW algorithms.


Plain Language Summary:

Great earthquakes are infrequent but devastating natural disasters. To mitigate their effects, earthquake early warning (EEW) systems aim to provide advance warning of strong shaking and tsunami. However, many of the most sophisticated EEW algorithms operating globally have a difficult time characterizing large earthquakes quickly and accurately enough to issue a meaningful warning -- this is most evident from the failure of EEW during the 2011 M9 Tohoku Oki, Japan earthquake. Here we propose a model, M-LARGE, that learns earthquake's surface deformation patterns from millions of simulations, and then apply it to unseen events. Our model shows a high accuracy of 99% performing on the testing dataset, and accurately estimates the magnitude of five real large historical events in Chile. The M-LARGE outperforms currently operating similar EEW algorithms

September 24 Sarah Stamps, Virginia Polytechnic Institute and State University

"Clues About the Break-up of the African Continent"


The African continent is slowly fragmenting along the East African Rift System (EARS) across east Africa. What drives this fragmentation process and how exactly is the Earth's lithosphere moving along the EARS? In this seminar, I present results that shows how the tectonic plates of the EARS are slowly moving and deforming, which causes hazardous seismic and volcanic activity.  Using modern-day, high precision positioning measurements from Global Navigation Satellite Systems (GNSS) and advanced numerical methods, we investigate the physics driving surface motions to understand the driving forces causing the break-up of Africa. We find that deformation is characterized by east-west extension and the major forces driving extension are derived from topography gradients.

October 1

Revathy Parameswaran, UAF, Geophysical Institute
"Seismic Failures and Evolution of Stresses on Conjugate Faults Systems: A Case Study from South Iceland"

Seismic ruptures on conjugate faults are prominent along active and relic rift-transform systems. One such setting presents itself along the section of the Mid-Atlantic ridge (MAR) that extends over Iceland. The MAR mutates into the obliquely rifting Reykjanes Peninsula (RP) in southwest Iceland, and splits into two parallel rifting arms (Western and Eastern Volcanic Zones; WVZ, EVZ) connected by the sinistral South Iceland Seismic Zone (SISZ), that exhibits conjugate NS-EW bookshelf faulting. All recorded major seismic events in South Iceland seem to have ruptured NS-trending faults stitching across the EW-oriented SISZ. However, the western end of the SISZ, referred to as the Ölfus region, that joins the RP and the WVZ at the Hengill triple junction, shows deviations from the established seismic pattern. We begin by analyzing the seismic activity in Ölfus centered around the 1998 M5.1 Hjalli-Ölfus earthquake. Seismic data analysis, relative relocations, faulting mechanisms, and stress inversions point to a significant EW-directed seismic activity. This is contrary to the documented trend of earthquakes in the SISZ, where most moderate to large earthquakes (2000, 2008 earthquake sequences) broke NS faults that lie to the east of Ölfus. Therefore, it is reasonable to evaluate the causality of variations in the stress field along the length of the western segment of SISZ. For instance, a possible cause for the November 1998 EW seismicity in Ölfus could be reduced normal stress on the fault zone, following the June 1998 NS seismic activity in the Hengill area that extended south to Ölfus. The foreshocks of the November 1998 earthquake occurred on a ~N-S fault until a day prior to the mainshock when they shifted to the ~ENE direction. The subsequent aftershocks are also mainly restricted to the ~ENE direction. Multiple relocations of the mainshock using various constraints indicate that the event likely occurred close to the junction of the conjugate ~ENE-WSW and ~N-S faults. Associated spatiotemporal variations in stresses in Hjalli-Ölfus indicate an eastward rotating SHmax prior to seismic swarms over defined time-windows during the study period, followed by a coseismic northward rotation.

October 8

No Seminar


October 15

No Seminar  

October 22

Zach Ross, California Institute of Technology
"Evidence for Latent Crustal Fluid Injection Transients in Southern California From Long-Duration Earthquake Swarms"

Earthquake swarms are manifestations of aseismic driving processes deep in the crust. We examine the spatiotemporal distribution of aseismic processes in Southern California using a 12-years catalog of swarms derived with deep learning algorithms. In a core portion of the plate boundary region, which is not associated with elevated heat flow, we identify 92 long-duration swarms ranging from 6 months to 7 years that constitute 26.4% of the total seismicity. We find that 53% of the swarms exhibit ultra-slow diffusive patterns with propagating backfronts, consistent with expectations for natural fluid injection processes. The chronology of the swarms indicates that the aseismic driving processes were active at all times during 2008–2020. The observations challenge common views about the nature of swarms, which would characterize any one of these sequences as anomalous. The regional prevalence of these sequences suggests that transient fluid injection processes play a key role in crustal fluid transport.
October 29

 No Seminar


November 5

Emily Montgomery Brown, USGS Cascades Volcano Observatory (CVO)
"Multifaceted Deformation at Long Valley Caldera"

Long Valley caldera is a well-known restless caldera that has experienced multiple episodes of tumescence recorded in several decades of deformation monitoring.  The superimposition of multiple deformation signals from volcanic, tectonic, and hydrologic processes have complicated our ability to provide satisfactory interpretations.  Incorporating multi-disciplinary data, time series decomposition methods, and finite element modeling, different components of the deformation can be extracted and highlighted to help in understanding how they influence the interpretation of the volcanic system and the implications for hazards.
November 12  Andy Aschwanden, UAF, Geophysical Institute
"Towards credible sea-level predictions"
Compared to 10-20 years ago, our understanding of how fast ice sheets can respond to a warming climate has grown substantially, and so have the capabilities of ice sheet models. Sea-level contribution projections with numerical models are now an integral part of IPCC’s sixth assessment report.
Despite all efforts, however, the models are still far from capable of making accurate predictions. In this talk, we will dive into the sixth assessment report to uncover the main model deficiencies and outline a path towards credible sea-level predictions. We take a Bayesian approach to reduce uncertainties in mass loss projections from the Greenland ice sheet. Our results show that a two-step calibration can significantly reduce uncertainties in mass loss projections and improve the skill of ice sheet models.
November 19

Eric Petersen, UAF, Geophysical Institute
"Melt and Morphology on Debris-Covered Glaciers: Field Studies on the Kennicott Glacier"


Melt from debris-covered glaciers represents a significant source of freshwater for millions of people worldwide, in regions like Alaska, the Andes, and High Mountain Asia. It is thus of great importance to understand the processes which affect surface melt on debris-covered glaciers, and be able to represent them in models. Two important competing factors are the ability of sufficiently thick debris to suppress melt rates and the ability of melt hot spots such as exposed ice cliffs to remove ice. We interrogate these processes through a series of studies constrained by field data on the Kennicott Glacier, Alaska. Our work includes (1) the development of a degree day melt model to predict distributed sub-debris melt, (2) a novel technique for measuring conductive and convective heat fluxes in supraglacial debris, and (3) monitoring of ice cliffs to determine controls on ice loss, morphologic evolution, and cliff survival. From this work we show that energy balance models for glaciers may be greatly improved by treatment of convection in debris. We also show how the survival and evolution of ice cliffs is controlled by their size, geometry, and the conditions at their base—whether they are undercut by a supraglacial stream, rest at the surface of a melt pond, or experience reburial by debris deposition. Ice cliffs on debris-covered glaciers may also be a viable analog for “icy scarps” observed exposing buried ice sheets in the mid-latitudes of Mars.
November 26

No Seminar Thanksgiving


December 3

Jenny Nakai, USGS Alaska Volcano Observatory (AVO)
"Near trench 3D seismic attenuation offshore Northern Hikurangi subduction margin, North Island, New Zealand"


We image seismic attenuation near the Hikurangi trench offshore New Zealand, using ocean bottom and land-based seismometers, revealing high attenuation above a recurring shallow slow-slip event and within the subducting Hikurangi Plateau. The Hikurangi subduction margin east of the North Island, New Zealand is the site of frequent shallow slow slip events. Overpressured fluids are hypothesized to lead to slow slip at shallow depths close to the oceanic trench. Seismic attenuation, energy loss of seismic waves, can be used to detect high temperatures, melt, the presence of fluids, and fractures. We use local earthquake P- and S-waves from 180 earthquakes to invert for t*, and subsequently invert for Qp and Qs, offshore the North Island directly above the area of slow slip. We image Qp and Qs to ~25 km depth, increasing resolution of previously identified coastal low Q, and finding a new region of even higher attenuation directly above the shallow slow slip event of 2014-2015, beneath the offshore seismic array. This highest attenuation is downdip of a subducting seamount, and is spatially correlated with a high seismic reflectivity zone and Vp/Vs>1.85, all of which provide evidence for the presence of fluids. The Qp and Qs is low at the trench and in the subducting plate, suggesting that seismic wave scattering due to faults, fractures, and the inherent heterogeneous composition of the Hikurangi Plateau, a large igneous province, plays a role in seismic attenuation.