Salone, Rosanna (2024) Integration of Geophysical and Thermo-Fluid Dynamic Numerical Modelling to Study Natural Gas Emissions in Complex Geological Contexts. [Tesi di dottorato]

	Text Salone_Rosanna_36_COMPLETA.pdf Visibile a [TBR] Repository staff only Download (8MB)
Preview	Text Salone_Rosanna_36_PARZIALE.pdf Download (6MB) | Preview

Collection description

A natural degassing system is a heterogeneous geological environment in which the release of gases, such as carbon dioxide, methane, sulfur dioxide and other volatile compounds, can have far-reaching environmental, geological, and human implications, making the study and understanding of such degassing systems crucial for hazard assessment, sustainable energy resource exploration and management, climate change mitigation and, more recently, geological carbon sequestration and storage. Over the last two decades, there has been a growing interest in the study of the Earth’s degassing processes. Many laboratory and field studies (e.g., gas sampling, remote sensing, geological fieldwork and geophysical surveys) have been performed to gather comprehensive data and insights about the origin, composition, behavior and impact of gas emissions in both volcanic and non-volcanic areas. In particular, thanks to current numerical modelling approaches for geophysical data analysis, many geophysical methods have been successfully and increasingly applied to detect gas migration in the subsurface, providing insight into the pathways that gases follow as they move from their source to the surface, as well as to characterize subsurface reservoirs and assess their potential for resource extraction. On the other hand, thermo-fluid dynamic numerical modelling has proven to be an indispensable tool for studying the dynamic processes occurring in such geological contexts and their possible changes over time. Therefore, in this PhD thesis an innovative approach that combines geophysical surveys and numerical simulations has been proposed to study complex geological systems characterized by natural gas rise phenomena. The aim was to establish an effective method for analyzing the upwelling fluid dynamics and understanding the behavior of natural gas in complex geological environments. Specifically, a multi-methodological geophysical prospecting, consisting of high-resolution 2D/3D electrical resistivity and induced polarization tomographies and self-potential surveys, is proposed to identify shallow permeable structures (i.e., faults, fractures) that could provide preferential pathways for fluid/gas migration towards the surface. Then, the use of constraints derived from different types of data (e.g., geological, geochemical and/or hydrogeological data), is suggested to build an accurate 3D petrophysical model of the system under study, which is successively used to simulate, through numerical modelling, the physical processes likely to occur in the system and their temporal evolution. Furthermore, as an additional challenge of the present research, a Random Forest Machine Learning algorithm is proposed to predict resistivity data from numerical simulation results, with the aim of establishing a numerical relationship between resistivity and petrophysical/thermodynamic properties of the studied system, which could provide a better understanding of the complex interaction between fluid flows and geological structures. The performance of the whole proposed approach has been tested in some orogenic and volcanic areas of the Southern Apennines (Italy), where diffuse CO2 and/or CH4 outgassing occurs near fault and fracture segments. The geophysical models of the studied systems, obtained from the joint interpretation of the acquired geoelectrical data, have identified anomalous sectors that correlate well with the possible presence of subsurface fluids (CO2 and/or CH4) in shallow buried fault and fracture systems, highlighting their role as preferential pathways for fluid and gas migration to the surface. The thermo-fluid dynamic numerical modelling, on the other hand, has proved highly effective in describing complex fluid flow processes, in particular in reproducing the behavior of CO2 fluxes in porous and fractured media. Finally, the trained Random Forest algorithm proved successful in predicting resistivity values, providing a powerful tool for tuning the petrophysical and thermodynamic parameters of the model and improve the understanding of ongoing system dynamics through geophysical monitoring. Based on the results, the developed research provides new insights into the activity of permeable structures and preferential fluid flow pathways, making it promising for the study of many geological processes, including fluid circulation in geothermal systems, natural gas risk assessment, crustal earthquakes and anthropogenic CO2 storage.

Downloads

Actions (login required)

View Item