Numerical Simulation of Geological Reactive Porous Flow

Yizhuo Sun

Research output: ThesisDoctoral Thesis

3 Downloads (Pure)


Reactive porous flow is a fundamental process related to various geological problems from the deep Earth (Part I) to the surface (Part II). In this thesis, I investigate the influence of reactive porous flow on redistributing crucial chemical components within porous media. To achieve this goal, I performed numerical simulations using the source code, MuPoPP (Multiphase Porous flow and Physical Properties) which is based on the FEniCS platform, a popular open-source finite element computing platform for solving partial differential equations (PDEs). In Part I (Chapter 3 & 4), the distribution and storage of carbonate-rich melts at the base of the upper mantle are tested. The seismic low-velocity layers (LVLs), located atop the mantle transition zone (MTZ), contain small amounts of partial melt, possibly derived from melting of subducted carbonate-bearing oceanic crusts. Petrological and geochemical evidence from inclusions in superdeep diamonds supports the existence of slab-derived carbonate-rich melts, which may potentially explain the origin of the observed melt in the LVL. However, the presumptive reducing nature of the ambient mantle can be an impediment to the stability of carbonate-rich melts. The theory of redox freezing claims that the carbonate-rich melt in the LVL will be reacted with the metallic iron in the reduced mantle and reduced to diamond. To reconcile this apparent contradiction, in Chapter 3, the stability and migration of carbonate-rich melts atop a stalled slab are tested as a function of melt percolation, redox freezing, amount of carbon supplied by subduction, and the metallic Fe concentration in the mantle. The simulation results demonstrate that carbonate-rich melts in the LVL can potentially survive redox freezing over long geological time scales, by propagating through channels. The results also show that the amount of subducted carbon exerts a stronger influence on the stability of carbonate melts than does the mantle redox condition. This work was published in 2020 (Sun et al., 2020a). In chapter 4, the seismic observation of an LVL beneath northeast Asia is combined with the numerical model of reactive porous flow from Chapter 3 and the geochemical model of the mantle-melt interaction. The northeast Asian LVL are characterized by small average degrees of partial melting (0.8±0.5 vol%), assisted by the volatiles. Based on this inferred melt fraction, numerical simulations of melt channeling suggest that despite the effect of buoyancy and chemical reaction, the melt can remain stable in the LVL over geologically significant periods of time, thus providing some constraints on the time of origin of these LVLs. Using the inferred melt fraction to model trace element abundances, we show that long term interaction between this melt and the surrounding mantle can produce a HIMU like geochemical signature. In Part II (Chapter 5), the same equations for reactive porous flow are applied to study geological carbon sequestration (GCS), the process by which pumped, supercritical CO2 is slowly assimilated into a subsurface reservoir. The influence of three parameters is tested: reservoir porosity, composition of the reservoir rock, and the chemical reaction rate, on the mode of fluid flow and efficiency of CaCO3 precipitation during GCS. The simulation results demonstrate that the mode of porous flow switches from propagation of a planar front at low porosities to propagation of channels at porosities exceeding 10%. Based on the results, deep saline aquifers containing 10 vol% or more porosity and 10 wt% or more anorthite in the bulk are recommended as suitable candidates for future sites of carbon capture and sequestration projects. This work was published in 2020 (Sun et al., 2020c).
Original languageEnglish
Awarding Institution
  • Royal Holloway, University of London
  • Chiarella, Domenico, Supervisor
  • Hier-Majumder, Saswata, Supervisor
Award date1 Sept 2021
Publication statusUnpublished - 2021


  • Porous Flow
  • Numerical Simulation
  • Low-velocity Layer
  • Deep volatile cycling
  • Geological Carbon Sequestration

Cite this