Modelling parametric uncertainty in large-scale stratigraphic simulations.

We combine forward stratigraphic models with a suite of uncertainty quantification and stochastic model calibration algorithms for the characterization of sedimentary successions in large scale systems. The analysis focuses on the information value provided by a probabilistic approach in the modelling of large-scale sedimentary basins. Stratigraphic forward models (SFMs) require a large number of input parameters usually affected by uncertainty. Thus, model calibration requires considerable time both in terms of human and computational resources, an issue currently limiting the applications of SFMs. Our work tackles this issue through the combination of sensitivity analysis, model reduction techniques and machine learning-based optimization algorithms. We first employ a two-step parameter screening procedure to identify relevant parameters and their assumed probability distributions. After selecting a restricted set of important parameters these are calibrated against available information, i.e., the depth of interpreted stratigraphic surfaces. Because of the large costs associated with SFM simulations, probability distributions of model parameters and outputs are obtained through a data driven reduced complexity model. Our study demonstrates the numerical approaches by considering a portion of the Porcupine Basin, Ireland. Results of the analysis are postprocessed to assess (i) the uncertainty and practical identifiability of model parameters given a set of observations, (ii) spatial distribution of lithologies. We analyse here the occurrences of sand bodies pinching against the continental slope, these systems likely resulting from gravity driven processes in deep sea environment.

Comparative assessment of continuum-scale models of bimolecular reactive transport in porous media under pre-asymptotic conditions.

We compare the ability of various continuum-scale models to reproduce the key features of a transport setting associated with a bimolecular reaction taking place in the fluid phase and numerically simulated at the pore-scale level in a disordered porous medium. We start by considering a continuum-scale formulation which results from formal upscaling of this reactive transport process by means of volume averaging. The resulting (upscaled) continuum-scale system of equations includes nonlocal integro-differential terms and the effective parameters embedded in the model are quantified directly through computed pore-scale fluid velocity and pore space geometry attributes. The results obtained through this predictive model formulation are then compared against those provided by available effective continuum models which require calibration through parameter estimation. Our analysis considers two models recently proposed in the literature which are designed to embed incomplete mixing arising from the presence of fast reactions under advection-dominated transport conditions. We show that best estimates of the parameters of these two models heavily depend on the type of data employed for model calibration. Our upscaled nonlocal formulation enables us to reproduce most of the critical features observed through pore-scale simulation without any model calibration. As such, our results clearly show that embedding into a continuum-scale model the information content associated with pore-scale geometrical features and fluid velocity yields improved interpretation of typically available continuum-scale transport observations.

Microscale simulation and numerical upscaling of a reactive flow in a plane channel.

A bimolecular homogeneous irreversible reaction of the kind A+B→C is simulated in a plane channel as a base example of reactive transport processes taking place at the microscale within porous and/or fractured media. The numerical study explores the way microscale processes embedded in dimensionless quantities such as Péclet (Pe) and Damköhler (Da) numbers propagate to upscaled coefficients describing effective system dynamics. The microscale evolution of the reactant concentrations is obtained through a particle-based numerical method which has been specifically tailored to the considered problem. Key results include a complete documentation of the process evolution for a wide range of Pe and Da, in terms of the global reaction rate, space-time distribution of reactants, and local mixing features leading to characterization of effective reaction and dispersion coefficients governing a section-averaged upscaled model of the system. The robustness of previously presented theoretical analyses concerning closures of volume-averaged (upscaled) formulations is assessed. The work elucidates the dependence of the effective dispersion and reactive parameters on the microscale mixing and reactive species evolution. Our results identify the role played by Da and Pe on the occurrence of incomplete mixing of reactants, which affects the features of the reactive transport scenario.