S1: From pore space to whole Earth: understanding complex geoprocesses through Model-Data Integration in the 21st century


 Ensemble data assimilation for estimating earthquake and geodynamic states and parameters

Y. van Dinther1, H. D. Montero, C. Marsman, A. Banerjee, P. J. van Leeuwen, R. Govers, and F. Vossepoel

  • 1. Department of Earth Sciences, Utrecht University, Utrecht, Netherlands

 The large growth in observational data and compute power for physics-based models transforms our ability to understand, quantify, and forecast Earth processes. Geophysicist aim to efficiently exploit and combine both advances to unravel complex geo-processes across spatial and temporal scales. However, both streams of information face challenges when dealing with uncertainties. To be able to effectively extract information we need to deal with uncertainties in a mathematically rigorous manner. This is done in ensemble data assimilation methods, which have revolutionised e.g., weather forecasting and are now also explored in solid Earth geophysics. Sequential methods incorporating observations step by step have not yet been explored much in solid Earth sciences. Generally, we anticipate too few observations are available or that their errors are too large to be useful. We often find both issues are efficiently tackled through ensemble data assimilation. I will demonstrate this using different applications estimating states and parameters governing Earth processes on human and historical time scales. We use Ensemble Kalman Filters (EnKF) to estimate the fault stress state well enough to have very low miss rates when forecasting synthetic earthquakes and slow slip events. Exploring Adaptive Gaussian Mixture Filters and Particle (Flow) Filters addressing non-Gaussian prior distributions just prior to earthquakes suggests including model errors addresses parameter bias. Yet large parameter biases in rate-and-state friction parameters needs to be addressed by estimating both states and parameters. Long enough observational time series would then allow to separate error contributions from friction and shear stresses to estimate current and future shear stresses and slip rates. Latest results on large-scale laboratory experiments will address the importance of an inaccurate representation of the governing physics. Finally, zooming in time we focus on incorporating geodetic observations prior to and after the 2011 M9.0 Tohoku earthquake to quantify geodynamic parameters and understand processes. We find with an Ensemble Smoother with Multiple Data Assimilation that a power law rheology with a single set of parameters can explain surface velocities before and after the earthquake. These methods show great promise for probabilistically quantifying a range of earthquake, geodynamic and solid Earth processes.

S2: Vom Erdkern bis zur hohen Atmosphäre und noch weiter: Erkenntnisse aus langen geophysikalischen Zeitreihen in der Wissenschaft und Industrie


Massentransporte im System Erde und ihre Signale in gravimetrischen Zeitreihen

B. Meurers

  • Universität Wien, Institut für Meteorologie und Geophysik, Wien, Österreich

Das Schwerefeld der Erde erfährt zeitliche Änderungen durch die Wechselwirkung mit extraterrestrischen Massen sowie durch Massentransporte im System Erde. Letztere beeinflussen das Schwerefeld der Erde direkt, wenn sie mit Dichteänderungen verbunden sind und indirekt als Folge von Deformation sowie von Änderungen der Erdrotation. Die Erfassung und Interpretation von zeitlichen Schwereänderungen ist somit von großer Bedeutung für die Erforschung geodynamischer Prozesse, die in einem weiten zeitlichen und räumlichen Skalenbereich und in allen Regionen des Erdinneren, vom Erdkern bis zur Erdoberfläche, sowie in der Atmosphäre ablaufen. Dazu gehören Erdgezeiten, atmosphärische und hydrologische Prozesse, tektonische Deformation oder Vulkanismus. Moderne Instrumentation erlaubt die Detektion von Signalen mit Amplituden in der Größenordnung von nur 1 nm/s² im Zeitbereich. Allerdings überlagern sich die Signale verschiedener Quellen und ihre Trennung stellt eine besondere Herausforderung dar. Sie gelingt in vielen Fällen durch Anwendung geeigneter, oft interdisziplinärer Methoden, die häufig die Erfassung sehr langer Zeitreihen erfordern. Der Vortrag diskutiert die grundsätzliche Problematik und zeigt an Hand ausgewählter Beispiele Lösungen und Resultate aktueller Forschung.

S3: Crustal Fluids and Seismicity: Observations, Modelling and Geophysical Imaging


N. Shapiro[1], C. Journeau[2], J. Soubestre[1], G. Farge[3], W. Frank[4], C. Jaupart[5], O. Melnik[1], V. Lyakhovsky[6]

  • ISTerre, CNRS, Grenoble, France
  • University of Oregon, Department of Earth Sciences, Eugene, United States
  • Seismology Laboratory of UC Santa Cruz, Santa Cruz, United States
  • Massachusetts Institute of Technology, Cambridge, United States
  • Institut de Physique du Globe de Paris, Paris, France
  • Geological survey of Israel, Jerusalem, Israel

Low frequency (LF) seismicity in form of swarms of earthquakes or nearly continuous tremors is regularly observed in association with volcano-magmatic systems and during recent decades has been discovered in subduction zones. In this presentation we will focus on LF signals originating in the fluid-rich parts of the lower crust and upper mantle. This deep LF seismicity can be used to detect the presence of such fluids as well as to understand their migration patterns and interaction with surrounding rocks. In the first part, we will consider how complex DLF signals can be studied with methods based on the data of seismic networks and adapted to detect spatially coherent signals without clear impulsive offsets, to measure their time-frequency properties, and to locate their sources. Examples of the network-based analysis of volcanic and tectonic DLF seismicity will be shown with comparing results obtained in Guerrero, Mexico and in Kamchatka, Russia. We will present physical models for the source processes of DLF events and discuss how their time and space variability can be explained by the influence of deep fluids and their migrations.

S4: Seismic Noise and Coda Waves


Advancing Gravitational Wave Astronomy through seismic noise reduction

K.-S. Isleif, the Einstein Telescope Collaboration

  • Helmut-Schmidt-Universität, Messtechnik, Hamburg, Germany

 Gravitational wave astronomy, akin to seismology, interprets waves in spacetime to uncover the universe's most intensive events. Gravitational waves offer insights into cosmic phenomena such as merging black holes, neutron stars, or the collapse of whole galaxies. Since 2015, we've been „hearing“ these spacetime ripples with ground-based instruments and we've been studying over 100 cosmic sources until today. Nonetheless, ground-based detectors encounter challenges, particularly a significant noise level at lower frequencies, mainly due to Earth's seismic noise.


This presentation will address the seismic challenges in current ground-based detectors, which achieve a displacement sensitivity of 10-18 m/√Hz at audio frequencies. Future detectors, like the Einstein Telescope, aim to enhance this sensitivity, significantly below 10 Hz, necessitating additional seismic noise reduction and management of new noise types, like Newtonian Noise. Newtonian noise directly results from seismic noise: Mass density changes in the Earth induce gravitational fluctuations, limiting the envisioned performance of future gravitational wave detectors. Innovative strategies are required to address this, such as deploying extensive underground seismic sensor arrays to measure seismic noise and predict Newtonian Noise. In this context, we will discuss the development of new seismic sensors, seismic network designs, and noise cancellation systems. These advancements in sensor technology and research methodologies are not just pivotal for the future of gravitational wave astronomy, they might also hold significant potential for advancing the field of seismology.