Emmy-Noether-Project "Bridging Geodesy and Seismology"

Project Objectives

Deep-rooted joint use of space-geodetic and teleseismic data to characterize crustal earthquakes as finite rupture processes globally

 

Seismic waveforms and surface displacement data have slightly different constraints on earthquake source characteristics
H. SudhausSeismic waveforms and surface displacement data have somewhat different constraints on earthquake source characteristics.

 

Today, we regularly use joined data sets of near-field geodetic displacement data and far-field seismic waveform data for earthquake source models, but this combination is never truly 'deep-rooted'.

Usually a source inference starts with a first-order source model using one single data set, e.g, an inferred moment tensor, or with a simple dislocation model from GNSS or InSAR data or using aftershock locations for rupture plane inferences.

The reason is that the inverse problem at this stage is non-linear and common non-linear modelling algorithms for the different data are not existing. From these first-order models source properties are taken to further analyse data for higher-order source models including more data.
The final models are usually very dependent from the first inferences and model biases and parameter trade-offs are inherited. For this reason we often find different solution for the same earthquake when using different data sets and or different models, which is in quantitative science very unsatisfying.

In the SRCMOD data base different earthquake source models can be compared.

In our project we develop a common modelling framework for near-field and far-field data, which will enable a data combination from the beginning. This development is done within the framework of the open-source pyrocko software project in partnership with the German Research Centre of Geosciences GFZ in Potsdam

rigorously propagate data errors in our analyses and development of modelling standards for first-order earthquake source parameters

Model ensembles of Bayesian models
Bayesian ensembles of model parameters (black dots) and optimum model solutions of different source studies.H. Sudhaus

Results of our data analyses are never perfect. When we measure far traveling seismic waves at globally distributed stations, we often have noise at the locality of the seismic station which is introducing data errors.

The solid earth through which the seismic waves travel, is known to us only to some extent and we make errors in modelling wave travel times and the surface displacement. Further more are the assumptions we make on the shape of the earthquake rupture (e.g. planar rupture) only valid to some extent.

In other words, we have imperfect data, imperfect knowledge of the medium, which is rupturing and how it ruptures and so we can have only imperfect earthquake source models.

A very important part of our earthquake research is therefore to estimate how imperfect the models, how uncertain they can be as a quality measure of the results. Admitting uncertainty and efforts to estimate them as realistically as possible will render our results also more useful to their further use in science: geologic interpretation, hazard assessments and earthquake simulations.

This topic is an overarching one and of course important for all the project topics and cooperations, particularly for the cooperation the SRCMOD project and our KAUST partners and the IASPEI seismic moment tensor group.

integration of the geodetic and seismological data to better resolved hazardous seismic faulting and harmless aseismic faulting processes at crustal faults

Dummy image
Slow and steady motion of curbstones caused by tectonic creep at the Hayward fault (San Francisco). missyleone

 

Slow movements at faults were hard to detect in the times before space observations with GNSS and/or InSAR have been possible.

These slow movements, e.g. postseismic afterslip or creep, are important in the budget of the tectonic fautling, however, as they release tectonic stress in a harmless manner and reduce the earthquake potential.

We want to integrate these faulting processes better in the earthquake research by improving an integrated modelling of seismic waveform data and geodetic surface displacements.

A target area is the very active North Anatolian fault zone, Turkey, and we particularly cooperate here with the British COMET group centered at the University of Leeds, who have both a profound knowledge in the area and a large regional surface displacement data base of seismic and interseismic surface motion.




 

Building of finite-source earthquake catalog databases with uncertainty communication

Dummy image
Through the data combination we will be able to provide more robust and added-value earthquake source models

 

 

Dummy image
Baysian ensembles of model parameters (black dots) and optimum model solutions of different studies.                                         Sudhaus

 

The most used and very valuable catalogs on earthquake source models are provided by seismological institutes that producing first-order seismic moment tensor models of earthquake ruptures worldwide for earthquake at least larger than magnitude 5, e.g. USGS (United States Geological Survey), Geofon (German seismological database) and others.

These earthquake source models give the time, location, magnitude and the mechanism of the earthquake (thrust, strike-slip or normal faulting earthquakes) in form of a point source.

Using the combination of seismic waveform data and surface displacement data

With the combination of the global data sets of seismic waveform data and surface displacement data enables us to robustly represent shallow crustal earthquakes as finite fault ruptures with a nucleation point. Potentially, we better locate the earthquakes and therefore have better information on the causative fault and resolve the extension of the fault.

Be able to model the rupture directivity

With this we will be able to model also the rupture directivity, which is for larger earthquake very relevant for the seismic hazard and to expected damage in the epicentral region. As an outcome of the project we want to have started building a catalog with such finite rupture models and have progressed towards a routinely analysis for these events.

developing of the corresponding analyses modules for open-source community-software

Dummy image
Example of a pyrocko scriptUrheberinfo

 

 

All technical developments are meant to be freely and timely available also for the science community and easy to implement. To achieve this we are in this project building up on the seismological open-source python-based toolbox pyrocko developed and maintained at GFZ Potsdam.

Geofon

SRCMOD Database