Our current understanding of the Earth’s interior structure and processes is limited to observations made at the surface that are mapped to the subsurface using inverse methods. The complexity of geophysical inverse problems mainly arises from the existence of many free parameters that sometimes have traded off with each other. This can cause inaccuracies, low resolution and non-uniqueness problems in geophysical models. The main focus of my dissertation is on how we can use two independent geophysical data types – geodesy and seismology – to increase knowledge, resolution and accuracy of Earth’s structure, and of interseismic and coseismic processes in the earthquake cycle.
For example, in my first project (Chapter 2) I search for repeating earthquakes (REs) using similarity search on recorded seismic waveforms from the northern San Francisco Bay Area. Evidence from the San Andreas fault and elsewhere indicates that REs are correlated with, and likely driven by, aseismic slip (creep) at depth. This is complementary knowledge to the geodetic observation of creep at the surface. The source information of REs can also be used to constrain the interseismic slip models inverted from geodetic data such as GPS and InSAR.
By using a new fast similarity search algorithm, that I developed specifically for probing big seismic data sets (described in Chapter 3), we found 198 RE groups, including periodic and nonperiodic repeating earthquake 'families', and repeating event pairs. Our results can not only help us to map the depth and extent of creep on several major faults but also reveal previously unknown structural complexity – e.g. that subparallel strands of the Maacama fault are active and creep simultaneously. Source parameters and locations of these REs can be used to update seismic hazard models, by better constraining the creeping areas of faults in the region, and to improve community models of fault geometry.
In a second major project (Chapter 4), I aim to reconcile earthquake source parameters and locations determined by long-period teleseismic source inversions with those obtained from InSAR data. The latter observes earthquakes in situ and thus, we presume, accurately locates them. Previous studies suggest that the discrepancies between these two catalogs arise from the existing inaccuracy in Earth models and are caused by the historic (and circular) problem that earthquake locations estimated using inaccurate velocity models are themselves inaccurate, and vice-versa.
In several case studies of various locations (e.g. California, Iceland, central Italy) we observe and quantify the biases of the S40RTS Earth velocity model that cause a delay or early arrival of the predicted seismic waves to the seismic stations at certain azimuths. We gather these misestimations of predicted seismic wave arrivals as corrections that can be applied to teleseismic source inversions in order to improve location accuracy. The similarity of corrections that we observe for events in the same region suggests they could be used as regional corrections. We also show that these corrections not only can be used to accurately locate global events but also can help us to accurately obtain the source mechanisms of these events. In future, by gathering these corrections for all the events with existing InSAR source models (i.e. more than 120 global events so far) we might be able to increase accuracy of velocity models of the upper mantle, e.g. by using finite-frequency tomography.