GNGTS 2018 - 37° Convegno Nazionale

522 GNGTS 2018 S essione 2.3 FEASIBILITY STUDIES TOWARDS A PROBABILISTIC SEISMIC HAZARD ANALYSIS FOR HYDROCARBON EXTRACTION AND WASTEWATER INJECTION ACTIVITIES A. Garcia-Aristizabal 1 , M. Anselmi 2 , L. Faenza 2 , B. Lolli 1 , A. Morelli 1 , I. Munafò 3 , L. Sandri 1 , L. Zaccarelli 1 1 Istituto Nazionale di Geofisica e Vulcanologia, sezione di Bologna, Bologna, Italy 2 Istituto Nazionale di Geofisica e Vulcanologia, sezione ONT, Roma, Italy 3 Istituto Nazionale di Geofisica e Vulcanologia, sezione di Roma1, Roma, Italy In recent years, great efforts have been devoted to the study of the human-induced earthquakes, owing to the effect that these earthquakes can have in terms of seismic hazard. In the past, different authors proposed classification schemes for distinguishing different mechanisms for fluid-induced seismicity generation (see, e.g. McGarr 2000). For example, it has been suggested that when the anthropic activities are responsible for a very small part of the stress field perturbations, the seismic events can be classified as “ triggered ”, while when the anthropic activities are responsible for the most of stress perturbations driving to the event occurrence, it can be classified as “ induced ”. In practice, we can rather consider that it may exist a continuum of cases depending on both the characteristics of technological operations and the local stress state. Outstanding features of induced seismic sequences is that events can be very shallow and seismicity rates may have a non-stationary character as a consequence of trends in industrial activity. The purpose of our analyses is to implement probabilistic seismic hazard analyses when considering possible effects of georesource development activities. It is recognized that the hazard implications of shallow, small-to-moderate events at short hypocentral distances merit careful evaluation. In implementing a probabilistic seismic hazard framework that takes into account possible seismicity induced by the exploitation of georesources, it is necessary to (1) assess the rate at which seismic events occur (considering both the natural and the induced components) and identify a possible relationship between induced seismicity rate and parameters characterizing the anthropic activity; (2) quantify the scaling laws dominating the size distribution of the seismic events (including the proper evaluation of event magnitudes and the determination of the maximum event size), and (3) evaluate the ground motion prediction equations governing the seismic energy attenuation processes. The induced seismic sequences associated with the injection of fluids can be observed both during and after fluid injections. Seismicity rates observed during the injection phases show a different temporal behavior if compared with the one usually observed after the end of injection phases. During injection phases, seismicity is usually “clustered” in time and space and shows non-stationary behavior. Non-stationarity is generally associated with the variability of the forcing associated with fluid injection. However, it is also possible that the induced events produce local stress changes that, in turn, can influence the evolution of the seismic sequence. Therefore, the complexity of the spatial and temporal distribution of the induced seismic sequences is probably linked to the combined contribution of different correlated processes such as the perturbations caused by fluid injection and the interactions between events. During the injection phase, the seismic sequence can be modelled assuming that the dominant factor for the generation of induced seismicity is the disturbance of stress caused by the injection of fluids (e.g., Langenbruch et al. , 2011; Garcia-Aristizabal, 2018). In particular, in this application, we are considering the method proposed by Garcia-Aristizabal (2018), which models the inter- event distribution between earthquakes, allowing the parameters of the probabilistic model to vary according to covariates chosen among the factors that characterize the injection of fluids (such as the injection rate). Besides the spatio-temporal behaviour, another key feature of the earthquake occurrence is the frequency-magnitude distribution, and, specifically for this case, the analysis is done at the local scale and in relation to the industrial activity. The starting point of this analysis is the use of a homogeneous data set, but the calculation of the magnitude of small events is a complex and challenging task. We are therefore interested in implementing reliable methods for magnitude