Open Access
Foroozan, Roozbeh
Graduate Program:
Energy and Geo-Environmental Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
February 04, 2010
Committee Members:
  • Derek Elsworth, Dissertation Advisor
  • Derek Elsworth, Committee Chair
  • Barry Voight, Committee Member
  • Jamal Rostami, Committee Member
  • Chris Marone, Committee Member
  • Fluid dynamics
  • Volcano
  • Basin
  • Fault
  • Geodesy
  • Back Analysis
  • GPS
This study aims to investigate the interaction between hydraulic transport and mechanical deformation characteristics of geological settings. Numerical models developed here can quantitatively provide constrained information to further explore the phenomena and processes which cannot be easily captured experimentally. They deliver tools to probe into the processes from the observations and provide knowledge from data. The developed models link rheology, permeability, compressibility, deformation, stress, and pore pressure. The magnitude and interaction of these controlling parameters govern the fluid dynamics in volcanoes, geothermal or petroleum reservoirs, carbon sequestration settings, faulted zones in basins and other geological natural systems. Part I of this thesis (Chapters I-V) explores the magma transfer dynamics of Soufriere Hills Volcano (SHV) in Montserrat, West Indies. To this aim, ground deformation data are used to test the hypotheses concerning the magmatic processes in a geodynamic context. The purpose of this study is to determine the location, geometry, and extent of sub-surface deformation sources including magma reservoirs, dykes and conduit using field measurements. It is also to explore the relationships of magmatic storage-transport processes such as the basalt injection/storage, andesite reservoir, and volatiles degassing. A Numerical HMC model for cyclic behavior of SHV is calibrated and validated using geodetic data. In Part II (Chapter VI) scale- and process-appropriate numerical models are developed to follow the evolution of mechanical and transport properties in a generic basin during its faulting. The study is to show that the evolution of transport properties of rock due to shear is an important, likely crucial, feedback in promoting localization of fault structures. In Chapter I, defining the concept of “apparent depth”, a transfer function is developed so that the results of inversions assuming a homogenous medium can be transferred into a heterogeneous case representing Montserrat Iceland. Ground deformation data from high quality continuous GPS (cGPS) records from 1999 to 2008 are used to show that the SHV magmatic system has multiple crustal pressure sources. As For a dual chamber volcanic model, a deep chamber located within the mid crust which undergoes volume changes approximately an order of magnitude larger than those of a shallow one can explain the apparent depth signature of the deformation field. Building upon the model presented in Chapter I, in Chapter II measured cGPS velocity histories and mean magma efflux are fitted to model predictions by searching the full parameter space of potential chamber depth combinations by a least-squares optimization. Two geometry sets are considered in dual magma chamber models. The first comprises two vertically-stacked spherical pressure point (Mogi-type) chambers and the second a horizontal circular sill co-centered with a Mogi-type chamber. We confirm that SHV chamber depths and shapes are only weakly constrained by measured surface displacements. Applying the additional constraint of presumed constant influx to the lowermost chamber fixes an optimal geometry of dual interconnected Mogi-type sources at depths of about 5 and 19 km. This corresponds to a constant input of magma to the lowermost chamber of ~1.2 m3/sec which remains steady throughout the three successive episodes of eruption and pause. The magma compressibility varies with depth of chamber. The eruption-pause episodes are characterized by the synchronous deflation-inflation of both chambers with volume changes of the deep chamber larger than the shallow chamber. Using the inversion results of Chapter II, Chapter III suggests that the shallow chamber controls the periodic system behavior: surface magma efflux resumes when the shallow chamber reinflates to its initial threshold pre-eruptive volume (thus triggering re-opening of an eruptive feeder dike), and ceases when it has lost 18±4 Mm3 of its volume (sealing the conduit and staunching the magma flow). These observations are consistent with eruption re-initiation controlled by magma overpressures exceeding threshold strength of the shallow host-rock and of re-cessation triggered by a threshold underpressure. In this way Chapter III provides evidence that the periodic behavior of SHV from 1995 to 2008 has been controlled by shallow processes. To accommodate the mentioned volume change, the initial volume of the shallow chamber should be larger than 1.3 km3. The inferred geometry and boundary conditions from the previous chapters are used in a flow dynamics and transport model in chapter IV. A transient 1-D FEM model is devolved here in an attempt to replicate the three fairly regular cycles of dome growth and repose of Soufrière Hills Volcano. The model captures the key processes controlling the eruption and repose of the volcano including magma crystallization, volatile exsolution, and subsequent rheological stiffening of the compressible magma ascending through an elastically inhomogeneous wallrock. The extreme sensitivity of the highly nonlinear feedbacks of conduit and magma chamber processes result in sudden transitions between active and repose phases after slight variations of the governing parameters. To match the observed eruptivity, the free parameters of the model, including the magma chamber sizes, conduit diameter, and crystallization rate can be estimated. The volumes of the upper and lower chambers are confined to 2 km3 and 10 – 20 km3, respectively. The conduit diameter and crystallization rate control the efflux rates and also the over-pressure in the chambers before the eruptions. To the contrary, the period of eruptions depend mainly on the chamber sizes. The average pressures in the chambers are slightly higher than the lithostat pressure and the upper chamber cycles between conditions of over- and under-saturation during periods of eruption and repose, respectively. As a result, abrupt changes in compressibility of magma exert repetitive cycles of expansion and contraction to the chamber wallrock which can be a deciding factor in the formation of the upper chamber. Part II, Chapter V, examines the role of basin shortening on the development of hydro-mechanical compartments in sedimentary basins on passive margins. A coupled flow deformation model is used to follow the evolution of an idealized prismatic basin during lateral shortening. This includes the deformation-induced generation (compaction) and dissipation (hydraulic fracturing) of pore fluid pressures and the resulting natural evolution of underlying décollement and fault structures. This model is used to examine the influence of strata stiffnesses, strain softening, permeability-strain dependence and contrast, and deformation rate on the resulting basin structure and fluid charge. Thrust faulting develops as overpressures evolve to trigger failure. A décollement forms within the system at the boundary with the substrate where overpressures drive failure in extension, by hydrofracturing. Failure in the basin overlying the décollement initiates from these overpressured sediments at the décollement. Where the evolution of permeability with shear strain is artificially suppressed, pervasive shear develops throughout the basin depth as fluid pressures are stabilized everywhere to the lithostat. Conversely, where permeability is allowed to increase with shear strain/rupture, faulting first nucleates at the décollement and localizes upwards through the section. Correspondingly, permeability evolution with shear is an important, likely crucial, feedback in promoting localization, as failure is concentrated at the limits of the upward-migrating fault-tip. Elevated pore pressures approaching the lithostat are localized at the hanging wall boundary of the faults. Faults extend to bound blocks that vertically offset in to yield graben-like structural highs and lows and evolve with distinctive surface topography and separate pore pressure signatures. Up-thrust blocks have elevated fluid pressures and reduced effective stresses at their core, and down-thrust blocks the converse. These models indicate that fluid migration within compressional systems is intimately connected to both the mechanical and fluid transport behavior of the component sediments via strong feedbacks.