Understanding economic success of unconventional production from shales requires an explanation of the relationship between induced seismicity and hydraulic fracturing. This thesis deals with observing and analyzing synthetic and real microseismic monitoring data acquired during hydraulic fracturing. The thesis is based on observation and analyses of source mechanisms of induced microseismic events that have recently become regularly inverted and interpreted in the oil and gas industry. The results of analyses are interpreted with the geomechanical model of the relationship between hydraulic fracturing and induced seismicity. The study of source mechanisms starts with detailed analyses of spatial distribution of full moment tensor inversion stability. It was mapped based on synthetically computed condition numbers in the vicinity of different monitoring arrays including dense arrays at the surface and sparse arrays with sensors in the boreholes. Stability of inversion was tested under several conditions, mainly dependency on size and geometry of monitoring array and level of noise in the data. In this part of the thesis it is shown that dense surface arrays may provide very stable inversion of source mechanisms which may be interpreted. The study shows that an increasing percentage of non-shear...

Η μεταπτυχιακή διπλωματική εργασία έχει ως στόχο να δημιουργήσει ένα ανοικτού κώδικα λογισμικό (scisola) -σε python-, το οποίο να παρέχει τις ακόλουθες λειτουργίες: Αυτοματοποίηση του υπολογισμού ΤΣΡ με σύνδεση στο σύστημα του SeisComP3. Αποθήκευση των αποτελεσμάτων σε βάση δεδομένων. Δυνατότητα παραμετροποίησης του λογισμικού από το χρήστη με χρήση γραφικού περιβάλλοντος (configuration). Δυνατότητα προβολής των αποτελεσμάτων με χρήση γραφικού περιβάλλοντος (review). Δυνατότητα τροποποίησης των αποτελεσμάτων βάση επιλογής του χρήστη με χρήση γραφικού περιβάλλοντος (revision). / This master thesis contributes by providing an open-source software application (scisola) -written in python-. It supports: Automatic calculation of moment tensors; the seismic event notification, station information and the corresponding waveforms are provided by the SeisComP3 system through different utilities and services like slinktool and seedlink server respectively. The moment tensor is calculated through the ISOLA software in parallel mode using multiple threads through multiprocessing python libraries for much faster calculations. Result storing in database for better data management. Extensive configuration changes based on the needs of each researcher, through a user friendly graphical interface. A graphical overview of the moment tensor calculation and the corresponding data fit. Moment tensor revision in case the user wishes to alter the automatically suggested result.

Earthquakes form by sudden brittle failure of rock mostly as shear ruptures along a rupture plane. Beside this, mechanisms other than pure shearing have been observed for some earthquakes mainly in volcanic areas. Possible explanations include complex rupture geometries and tensile earthquakes. Tensile earthquakes occur by opening or closure of cracks during rupturing. They are likely to be often connected with fluids that cause pressure changes in the pore space of rocks leading to earthquake triggering. Tensile components have been reported for swarm earthquakes in West Bohemia in 2000. The aim and subject of this work is an assessment and the accurate determination of such tensile components for earthquakes in anisotropic media. Currently used standard techniques for the retrieval of earthquake source mechanisms assume isotropic rock properties. By means of moment tensors, equivalent forces acting at the source are used to explain the radiated wavefield. Conversely, seismic anisotropy, i.e. directional dependence of elastic properties, has been observed in the earth's crust and mantle such as in West Bohemia. In comparison to isotropy, anisotropy causes modifications in wave amplitudes and shear-wave splitting. In this work, effects of seismic anisotropy on true or apparent tensile source components of earthquakes are investigated. In addition, earthquake source parameters are determined considering anisotropy. It is shown that moment tensors and radiation patterns due to shear sources in anisotropic media may be similar to those of tensile sources in isotropic media. In contrast, similarities between tensile earthquakes in anisotropic rocks and shear sources in isotropic media may exist. As a consequence, the interpretation of tensile source components is ambiguous. The effects that are due to anisotropy depend on the orientation of the earthquake source and the degree of anisotropy. The moment of an earthquake is also influenced by anisotropy. The orientation of fault planes can be reliably determined even if isotropy instead of anisotropy is assumed and if the spectra of the compressional waves are used. Greater difficulties may arise when the spectra of split shear waves are additionally included. Retrieved moment tensors show systematic artefacts. Observed tensile source components determined for events in West Bohemia in 1997 can only partly be attributed to the effects of moderate anisotropy. Furthermore, moment tensors determined earlier for earthquakes induced at the German Continental Deep Drilling Program (KTB), Bavaria, were reinterpreted under assumptions of anisotropic rock properties near the borehole. The events can be consistently identified as shear sources, although their moment tensors comprise tensile components that are considered to be apparent. These results emphasise the necessity to consider anisotropy to uniquely determine tensile source parameters. Therefore, a new inversion algorithm has been developed, tested, and successfully applied to 112 earthquakes that occurred during the most recent intense swarm episode in West Bohemia in 2000 at the German-Czech border. Their source mechanisms have been retrieved using isotropic and anisotropic velocity models. Determined local magnitudes are in the range between 1.6 and 3.2. Fault-plane solutions are similar to each other and characterised by left-lateral faulting on steeply dipping, roughly North-South oriented rupture planes. Their dip angles decrease above a depth of about 8.4km. Tensile source components indicating positive volume changes are found for more than 60% of the considered earthquakes. Their size depends on source time and location. They are significant at the beginning of the swarm and at depths below 8.4km but they decrease in importance later in the course of the swarm. Determined principle stress axes include P axes striking Northeast and Taxes striking Southeast. They resemble those found earlier in Central Europe. However, depth-dependence in plunge is observed. Plunge angles of the P axes decrease gradually from 50° towards shallow angles with increasing depth. In contrast, the plunge angles of the T axes change rapidly from about 8° above a depth of 8.4km to 21° below this depth. By this thesis, spatial and temporal variations in tensile source components and stress conditions have been reported for the first time for swarm earthquakes in West Bohemia in 2000. They also persist, when anisotropy is assumed and can be explained by intrusion of fluids into the opened cracks during tensile faulting. / Erdbeben entstehen durch plötzlichen Sprödbruch des Gesteins, meist als Scherbruch entlang einer Bruchfläche. Daneben werden für einige Beben v.a. in vulkanischen Gebieten auch Mechanismen beobachtet, die scheinbar vom Modell des Scherbruches abweichen. Ursachen dafür beinhalten komplexe Bruchgeometrien und tensile Erdbeben. Bei tensilen Erdbeben kommt es während des Bruchs zum Öffnen oder Schließen der Bruchfläche und damit zu Volumenänderungen. Erdbeben mit tensilen Anteilen stehen wahrscheinlich oft im Zusammenhang mit Fluiden, welche zur Durckänderung im Porenraum von Gesteinen und damit zum Auslösen des Bebens führen. Sie wurden auch im Vogtland während eines Erdbebenschwarms im Jahr 1997 beobachtet. Die Beurteilung und sichere Bestimmung tensiler Anteile von Erdbeben sind Ziel und Gegenstand dieser Arbeit. Bei Standardverfahren zur Bestimmung von Erdbebenmechanismen werden isotrope Gesteinseigenschaften angenommen. Momententensoren beschreiben dabei Kräfte, die das abgestrahlte Wellenfeld erklären. Allerdings wird seismische Anisotropie, d.h. Richtungsabhängigkeit elastischer Eigenschaften, in der Erdkruste und im Mantel wie z.B. im Vogtland beobachtet. Anisotropie bewirkt im Vergleich zu isotropen Medien Veränderungen der Wellenamplituden und -polariserungen sowie das Aufspalten von Scherwellen. In der vorliegenden Arbeit werden daher der Einfluss seismischer Anisotropie auf wahre oder scheinbar auftretende tensile Quellanteile untersucht und Erdbebenmechanismen unter Berücksichtigung seismischer Anisotropie bestimmt. Es wird gezeigt, dass Momententensoren und Abstrahlmuster von Scherbrüchen in anisotropen Medien denen von tensilen Brüchen in isotropen Medien ähneln können. Umgekehrt treten Ähnlichkeiten tensiler Beben in anisotropen Gesteinen mit Scherbrüchen in isotropen Medien auf. Damit existieren Mehrdeutigkeiten beobachteter tensiler Quellanteile. Die Effekte von Anisotropie hängen von der Orientierung des Bruches und vom Grad der Anisotropie ab. Außerdem beeinflusst Anisotropie das Moment eines Bebens. Herdflächenorientierungen können auch dann verlässlich bestimmt werden, wenn man Isotropie statt Anisotropie annimmt und die Spektren von Kompressionswellen verwendet. Bei Hinzunahme der Spektren von Scherwellen können Uneindeutigkeiten auftreten. Abgeleitete Momententensoren zeigen systematische Artefakte. Beobachtungen tensiler Quellanteile von Beben im Vogtland im Jahr 1997 können nicht allein durch moderate Anisotropie erklärt werden. Weiterhin wurden früher bestimmte Momententensoren induzierter Beben nahe der Kontinentalen Tiefbohrung, Bayern, unter Annahme anisotroper Parameter reinterpretiert. Die Beben werden einheitlich als Scherbrüche charakterisiert, obwohl deren Momententensoren tensile Bestandteile enthalten, die als scheinbar angesehen werden. Die Resultate unterstreichen die Notwendigkeit, seismische Anisotropie zu berücksichtigen, um tensile Komponenten von Erdbeben eindeutig zu bestimmen. Ein daher neu entwickelter Inversionsalgorithmus wurde getestet und erfolgreich auf 112 Erdbeben der letzten intensiven Schwarmepisode im Jahr 2000 im Vogtland an der deutsch-tschechischen Grenze angewandt. Die Herdparameter wurden unter Verwendung isotroper und anisotroper Geschwindigkeitsmodelle ermittelt. Die Beben zeigen Lokalmagnituden zwischen 1,6 und 3,2. Sie weisen zueinander ähnliche Herdflächenlösungen mit linkslateralem Versatz auf steil einfallenden, etwa Nord-Süd orientierten Bruchflächen auf. Die Fallwinkel nehmen oberhalb 8,4km Tiefe ab. Für über 60% der betrachteten Erdbeben werden tensile Quellanteile mit Volumenvergrößerung beobachtet. Die tensilen Komponenten zeigen Abhängigkeiten von Herdzeit und -ort. Sie sind zu Beginn des Schwarms sowie in Tiefen unterhalb 8,4km besonders signifikant und nehmen später an Bedeutung ab. Abgeleitete Hauptspannungsachsen enthalten P Achsen mit nordwestlicher und T Achsen mit südwestlicher Streichrichtung. Sie ähneln denen in Mitteleuropa. Es werden tiefenabhängige Fallwinkel beobachtet. Die Änderungen erfolgen für die P Achsen graduell von 50° hin zu flacheren Fallwinkeln bei tieferen Beben. Sie erfolgen jedoch abrupt für die T Achsen von etwa 8° oberhalb einer Tiefe von etwa 8,4km zu 21° einfallend unterhalb dessen. Mit dieser Arbeit werden erstmals zeitliche und räumliche Veränderungen tensiler Quellanteile und Spannungszustände im Vogtland für Erdbeben im Jahr 2000 beobachtet. Diese haben auch dann Bestand, wenn seismische Anisotropie berücksichtigt wird. Sie können durch Fluide erklärt werden, die in die Bruchflächen eindringen.

Les séismes de type longue période (LP) sont aujourd’hui enregistrés sur la plupart des volcans dans le monde entier. Malgré cela, le mécanisme à leur source n’est encore que très peu compris. A l’heure actuelle les modèles proposés pour expliquer leur origine sont : 1) la résonance d’une fracture remplie de gaz ou de fluides excités par des instabilités dans l’écoulement des fluides ou par la rupture fragile du magma ; 2) la fracturation lente des sédiments peu consolidés à la surface des volcans, dans des conditions de transition entre le ductile et le fragile. L’outil le plus utilisé pour comprendre leur nature est aujourd’hui l’inversion du tenseur des moments. Au cours des dernières années, les inversions du tenseur des moments se concentraient principalement sur la compréhension du mécanisme physique à l’origine des séismes LP qui souvent supposaient des milieux géologiques très simples, voire homogènes. Des études récentes ont montré l’influence des sédiments peu consolidés à la surface des volcans sur la propagation des ondes à basse fréquence et en conséquence, sur l’inversion du tenseur des moments quand ils ne sont pas pris en compte dans le processus d’inversion. Le but de cette thèse est de mieux comprendre les processus physiques qui génèrent les séismes LP et de quantifier les incertitudes liées à leur interprétation. / Long-period (LP) seismic events are abundantly recorded during rest and unrest periods at many volcanoes worldwide. However, their source mechanism is still poorly understood. Models which have been proposed so far to explain their origin are: 1) the resonance of a fluid-filled cavity triggered by fluid instabilities or the brittle failure of magma; 2) slow-rupture earthquakes occurring in the low consolidated materials composing the shallow portion of the volcanic edifice. Nowadays the main tool used to get insights into their nature is moment tensor (MT) inversion. MT inversions carried out in the past years focused mainly on the understanding of the physical origin of LP events and often supposed a relative simple geological structure of the medium. Recent studies highlighted the strong influence of shallow unconsolidated materials on the retrieved MT solutions and the importance of considering geological inhomogeneity in the inversion process. The principal aim of this thesis is to gain a better understanding of the source processes that generate LP events and to quantify the uncertainties related to the MT inversion process.

Title: Automated determination of earthquake source parameters Author: Jiří Vackář Department: Department of Geophysics Supervisor: prof. RNDr. Jiří Zahradník, DrSc., Department of Geophysics Abstract: The thesis deals with methods for automated inversion of seismic source parameters. We studied the influence of structure model used and show an ex- ample how the existing model can be improved. We have developed a new, fully automated tool for the centroid moment tensor (CMT) inversion in a Bayesian framework. It includes automated data retrieval from ArcLink server or local data storage. Step-like disturbances are detected using modeling of the distur- bance according to instrument parameters and such components are automati- cally excluded from further processing. Frequency ranges for the filtration and time windows for the inversion are determined automatically due to epicentral distance. Full-waveform inversion is performed in a space-time grid around a provided hypocenter. A data covariance matrix calculated from pre-event noise yields an automated weighting of the station recordings according to their noise levels and also serves as an automated frequency filter suppressing noisy frequency ranges. The method is tested on synthetic and observed data. It is applied on a dataset from the Swiss seismic...

At the northern extent of the Cascadia subduction zone, the subducting Explorer and Juan de Fuca plates interact across a translational deformation zone, known as the Nootka fault zone. The Seafloor Earthquake Array Japan-Canada Cascadia Experiment (SeaJade) was designed to study this region. In two parts (SeaJade I and II, deployed from July – September 2010 and January – September 2014), seismic data from the SeaJade project has led to several important discoveries. Hypocenter distributions from SeaJade I and II indicate primary and secondary conjugate faults within the Nootka fault zone. Converted phase analysis and jointly determined seismic tomography with double-difference relocated hypocenters provide evidence to several velocity-contrasting interfaces seaward of the Cascadia subduction front at depths of ~4-6 km, ~6-9 km, ~11-14 km, and ~14-18 km, which have been interpreted as the top of the oceanic crust, upper/lower crust boundary, oceanic Moho, and the base of the highly fractured and seawater/mineral enriched veins within oceanic mantle. During SeaJade II, a MW 6.4 mainshock and subsequent aftershocks, known as the Nootka Sequence, highlighted a previously unidentified fault within the subducted Explorer plate. This fault reflects the geometry of the subducting plate, showing downward bending of the plate toward the northwest. This plate bend can be attributed to negative buoyancy from margin parallel mantle flow induced by intraslab tearing further northwest. Seismic tomography reinforces the conclusions drawn from the Nootka Sequence hypocenter distribution. Earthquakes from the entire SeaJade II catalogue reveal possible rotated paleo-faults, identifying the former extent of the Nootka fault zone from ~3.5 Ma. / Graduate

In this project machine learning (ML) interatomic potentials are trained and used in molecular dynamics (MD) simulations to predict the physical properties of total energy, mean squared displacement (MSD) and specific heat capacity for systems of bulk Aluminium and Silicon. The interatomic potentials investigated are potentials trained using the ML models kernel ridge regression (KRR) and moment tensor potentials (MTPs). The simulations using these ML potentials are then compared with results obtained from ab-initio simulations using the gold standard method of density functional theory (DFT), as implemented in the Vienna ab-intio simulation package (VASP). The results show that the MTP simulations reach comparable accuracy compared to the DFT simulations for the properties total energy and MSD for Aluminium, with errors in the orders of magnitudes of meV and 10-5 Å2. Specific heat capacity is not reasonably replicated for Aluminium. The MTP simulations do not reasonably replicate the studied properties for the system of Silicon. The KRR models are implemented in the most direct way, and do not yield reasonably low errors even when trained on all available 10000 time steps of DFT training data. On the other hand, the MTPs require only to be trained on approximately 100 time steps to replicate the physical properties of Aluminium with accuracy comparable to DFT. After being trained on 100 time steps, the trained MTPs achieve mean absolute errors in the orders of magnitudes for the energy per atom and force magnitude predictions of 10-3 and 10-1 respectively for Aluminium, and 10-3 and 10-2 respectively for Silicon. At the same time, the MTP simulations require less core hours to simulate the same amount of time steps as the DFT simulations. In conclusion, MTPs could very likely play a role in accelerating both materials simulations themselves and subsequently the emergence of the data-driven materials design and informatics paradigm.

Une approche donnant de meilleurs résultats pour les potentiels interatomiques en apprentissage-machine à la volée est proposée en comparant trois approches pour la recherche de processus activés par la technique d'activation-relaxation. Tout d'abord, nous discutons de l'intérêt et des enjeux de l'utilisation des potentiels en apprentissage-machine et justifions l'utilisation de l'apprentissage à la volée pour la recherche de processus activés. Cela nous mène à présenter la forme générale des potentiels en apprentissage-machine, quelques modèles via leurs descripteurs de configuration atomique, paramètres et hyperparamètres ainsi que la méthode de l'apprentissage à la volée. Ensuite, nous présentons les méthodes d'exploration utilisées et les détails d'intégration du potentiel à la volée. Enfin, nous menons une étude comparative des trois approches pour un système de Si et de SiGe avec diffusion de lacune. La méthodologie proposée de potentiel de haute précision permet d'étendre la gamme de problèmes pouvant être étudiés par la technique d'activation-relaxation. / An approach giving better results for on-the-fly machine learning interatomic potential proposed by comparing three approaches for exploration of activated processes by the activationrelaxation technique. We first discuss the interest and challenges of on-the-fly machine learning potential and justify the use of on-the-fly learning for the search for activated processes. This leads us to present the general form of machine learning potentials and some models via their atomic configuration descriptors, parameters and hyperparameters as well as the on-the-fly learning method. Then, the exploration methods used are defined and the details of the integration of the potential are presented. Finally, a study is conducted comparing the three approaches for a Si and SiGe system with vacancy diffusion. The proposed methodology of high-precision potential allows to extend the range of possible problems to be studied by the activation-relaxation technique.