Modeling fluid flow in fault zones: two different-scale cases from Majella Mountain and East Pacific Rise

A quantitative assessment of how rock discontinuities (i.e., fractures and faults) control the migration of geofluids is critical in several areas of geological and environmental sciences. In this project I concentrated my attention to two problems at the extreme range of the spectrum of fluids properties, in two different geological settings: the flow of CO2 from a natural reservoir and the melt migration in mid-ocean transform fracture zones. The specific targets of my studies were: -a fault zone exposed in the Roman Valley Quarry (Lettomanoppello, Italy); -the fracture zone of the Siqueiros Transform Fault (East Pacific Rise); The thesis is structured starting from a general overview of the two study cases and then describing in detail the methods used and the results achieved for each scenario. To model the migration of CO2 in a fault zone, I created a new pipeline that starts from the field data and then uses open source software and new developed codes to model the fluid flow in a fault zone considering all its components: core and damage zones. I selected the Roman Valley Quarry as study site because of the great exposure of the inner structure of two oblique slip normal faults. Besides, the massive presence of fluid migration in the form of tar in the fracture systems makes this site a good natural analogue for studies of fluid flow in fractured media. The work on the fault zone of Roman Valley Quarry can be divided in three main parts: 1. Collection of quantitative information on the fractures/fault distribution; 2. Application of state-of-art modelling techniques to infer the hydraulic properties of the fractured reservoir from field data; 3. Numerical modeling of flow of CO2 in the fractured reservoir; I modeled the data collected in the field to infer the hydraulic properties of the fractured reservoir (i.e. the Bolognano Formation). I employed a hybrid numerical technique, modeling the damage zones as a fractured medium and the core as a continuum medium. This allowed me to: 1. characterize the hydraulic differences observed in the field between the southern part and the northern part of the fault; 2. characterize the hydraulic parameters of the footwall and hangingwall damage zones; 3. use these values to model the fluid flow in the whole fault zone, coupling damage zones and core. I built Discrete Fracture Networks (DFN) models using both commercial (Move®) and open source software (dfnWorks, developed by Los Alamos National Laboratory). Move® has been used to model the hydraulic differences found in the field between the northern and southern sector of the fault. dfnWorks has been used to infer the hydraulic parameters of the fracture systems of the damage zones of the fault. These parameters have then been used to upscale the properties of the fractures to an equivalent continuum medium, in order to simulate the fluid flow in the entire fault zone, coupling the core and the damage zones. The numerical model of CO2 flow in the fault zone was developed using the open source software PFLOTRAN. To better reproduce a real-life case study, I simulated the injection of CO2 into the footwall of the Roman Valley Quarry fault. Hydrostatic initial conditions have been imposed, according to the pressure distribution in the domain. An injection mass rate has been imposed at the location of the well to simulate the injection of CO2. First, I run a number of simulations to test the workflow and verify the consistency of the numerical results. Once I obtained a stable numerical framework, I run several numerical experiments. Results from numerical experiments show that the distribution of the CO2 in the domain appears mainly controlled by the permeability distribution in the damage zone of the fault. The CO2 in fact accumulates in the high permeability fault footwall, where the injection happens and reaches the maximum values of the pressure and saturation close to the core of the fault, that is characterized by a low permeability. Although developed and calibrated for the specific site of the Roman Valley Quarry fault, the methodology developed in this study can be extended to different geological contests. Although not originally part of my PhD proposal, the involvement on the Off-Axis Seamounts Investigations at Siqueiros (OASIS) project was a unique opportunity to learn how to collect, process and employ geophysical data to characterize and model fluid flow (in this case, magmatic melts) near a fault zone. The OASIS (Off-Axis Seamount Investigations at Siqueiros) expedition is a multidisciplinary effort to systematically investigate the 8˚20’N Seamount Chain to better understand the melting and transport processes in the southern portion of the 9˚-10˚N segment of the East Pacific Rise (EPR). The 8˚20’N Seamount Chain extends ~160 km west from its initiation, ~15km northwest of the EPR-Siqueiros ridge transform intersection (RTI). To investigate the emplacement of the 8˚20’N Seamounts, shipboard EM-122 multibeam, BGM-3 gravity, and towed magnetometer data were collected using the R/V Atlantis in November 2016. Multibeam data show that the seamount chain is characterized by the emplacement of discrete seamounts in the distal portion of the chain, while east of 105˚20’ W, the chain is a nearly-continuous volcanic ridge comprised of small cones and coalesced edifices with some evidence for rift zones, craters and calderas on the larger constructs. Isostatic anomalies, calculated along several profiles crossing the main edifices of the seamount chain, indicate that the seamounts formed within 100 km of the EPR ridge axis. Excess crustal thickness variations of 0.5 to 1 km, derived from the Residual Mantle Bouguer Anomaly, suggest an increase in melt flux eastward along the chain. Consistently high emplacement volumes are observed east of -105 ˚20’ W, ~130 km from the ridge axis corresponding with lithosphere younger than 2 Myr. Inverted three-dimensional magnetization data indicate that the seamounts have recorded a series of magnetic reversals along the chain, which correlate to reversals recorded in the surrounding seafloor upon which the seamounts were built. However, reversals along the eastern portion of the chain appear skewed to the west indicating that seamount formation is likely long-lived. The geophysical observations indicate that the overall seamount chain is age progressive, and suggest a coeval volcanism in a region 15-100km from the EPR. The seamounts do not follow absolute plate motions, but are located consistently 15-20 km north of the Siqueiros fracture zone, which further suggests that their formation is linked to the location and tectonic evolution of the Siqueiros-EPR ridge-transform intersection. These findings have implications for the location/origin of the melt region feeding the EPR as well as how melt is transported near a fracture zone. In fact, the seamounts chain does not follow an hotspot reference frame, but instead runs parallel to the fracture zone, at a constant distance. This observation is unusual, compared to the other seamounts in the region. Evidences of plate direction rotation in variation of the main trend of the chain can be observed. They can be attributed to events of plate rotations that characterized the evolution of the Siqueiros Transform Fault as according to Pockanly et al., (1997): the seamount chain may have formed with the first event of plate reorganization from 3.5 Ma, with the inset of the volcanism close to the RTI. We could think about a melt migration model that takes into account the tectonic evolution of the area, pointing out the role of a stress concentration in the vicinity of the RTI as triggering mechanism for the volcanism.