Detailseite
Projekt Druckansicht

Brückenschlag zwischen Geodäsie und Seismologie zur verbesserten und automatisierten Untersuchung von Bruchvorgängen

Fachliche Zuordnung Physik des Erdkörpers
Geodäsie, Photogrammetrie, Fernerkundung, Geoinformatik, Kartographie
Förderung Förderung von 2015 bis 2022
Projektkennung Deutsche Forschungsgemeinschaft (DFG) - Projektnummer 276464525
 
Erstellungsjahr 2022

Zusammenfassung der Projektergebnisse

The Bridges project aimed at a better combined use of new space-geodetic measurements of surface displacements caused by shallow crustal earthquakes and globally recorded seismic waveforms for analyzing the first-order characteristics of these earthquakes. A motivation for this project was that often a large variety of different models for individual signicant earthquakes exists, due to the use of various methods and different data. We suggested that with better standards, and easy to have and to handle software tools this variability of results would likely be reduced. Therefore all methods here are documented and open-source and published at pyrocko.org. Also, we wanted to support a better error handling and reporting of uncertainties associated to results of earthquake source modeling. Some earthquake characteristics are hard to determine but important in order to understand the process with respect to the host rock and fracture physics, for instance size, slip and speed of the rupture. We build the needed common model framework used to synthesize static surface displacements and dynamic ground motion with the same rupture model, such that these independent observation could be used simultaneously in the inferences and like this ensure the earthquake models to be more accurate and more informative. Furthermore, this framework allows study also slow faulting at depth. Now, aseismic motion can be analyzed on the same basis of assumptions as seismic ones. For instance, our model enables 3D block modeling of fault zones with a realistic elastic-viscoelastic medium. With our improved methods we robustly estimate first-order models of earthquake rupture. Such a low-parametric rupture model is formed by a rectangular rupture plane with a defined geographic location and depth, with a length and a width, an orientation, a slip direction and amount, as well as the rupture nucleation and the rupture velocity. The medium around the rupture is a definable horizontally layered elastic medium. We enable a rigorous data error propagation for a Bayesian model uncertainty estimation. Like this we prove in our publications that we retrieve more accurate information on the subsurface rupture processes and therefore learn more about the subsurface geology, compared to analyses based on either one of these data sets alone. An automated routine modeling of nite sources with combined data is planned. Using surface displacement data in the earthquake'sneareld, we gain sensitivity for rupture complexity, e.g. segmentation of rupture, which requires to model multiple low-parametric models. This, however, comes at the price of a higher dimensionality of the problem and usually much more uncertain solutions. To inform the kinematic modeling independently, we accompany it with results of a newly developed multiarray backprojection of seismic waves, a method that tracks high(-frequency) seismic energy emission during an earthquake in space and time. These emissions are high if slip and/or rupture speed changes, so at the start and stop, as well as at any strong irregularities during the rupture. Our multi-array method uses the entire global seismological network from which it forms a set of virtual local station arrays. Because stations globally are used, this backprojection achieves a high spatial resolution and can be successfully applied to earthquakes with magnitude less than 7. The Bridges project did major ground work for more accurate earthquake source inferences, for everybody to use. Rupture at depth can be characterized with increased condfience. We hope with the available software tools, a generally higher standard in earthquake source modeling will evolve. Time will tell. As usual in science, some results point to yet more problems. Here we nd that with an improved rupture model, estimated with more and better data, we became signicantly more sensitive for shortcomings of the 1D medium model places around the rupture, causing inconsistencies in the modeling results. There is a silver lining though. We got sensitive to inaccuracies of our earth models, which opens a door to learn more about the properties of the earth's crust and reduce the discovered inconsistencies in the same time.

Projektbezogene Publikationen (Auswahl)

 
 

Zusatzinformationen

Textvergrößerung und Kontrastanpassung