International Journal of Greenhouse Gas Control 109 (2021) 1033392(Kim et al., 2019; Rhino et al., 2016) and modeled by means of computational fluid dynamics (CFD) (Dewar et al., 2013; Drange et al., 1993) (Umer, add this paper to here, Sato, T.; Sato, K. Numerical pre-diction of the dilution process and its biological impactsonCO2 ocean sequestration. J. Marine Sci. Technol. 2002, 6, 169-180.) and a kind of integral plume models (Dissanayake et al., 2018). The data of changes in pH from these small-scale plume models has been successfully imple-mented, as the input parametric data for plume further movements, in the large-scale ocean models (Blackford et al, 2020) once CO2 leaks from seabed. On the other hand, the CO2 mitigation from storage reservoirs to sediments were widely investigated from the field observations (Furre et al., 2017), the Lab experimental studies (Rillard et al., 2015; Tongwa et al., 2013) and numerical simulations (Discacciati et al., 2002; Du et al., 2016; Cai et al., 2009; Chidyagwai and Rivi`ere, 2010). However, leakage from sediments into the ocean turbulent bottom boundary layers (TBBL) has been less focused, which, however, is one of the key processes for assessments of the biological impacts of leaked CO2 on the ocean, as a rich population of marine organisms resides within both the ocean TBBL and shallow sediments. It is the dynamics of CO2 migration and dissolution crossing the interface between shallow sediments and oceanic TBBL that dominate the leakage sources to the plumes evalua-tions in the ocean (Caudron et al., 2012). In order to ensure the effective environmental monitoring of offshore CCS storage sites, the Strategies for Environmental Monitoring of Marine Carbon Capture and Storage (STEMM-CCS) project was launched in 2016 (Blackford et al., 2018), following previous field experiments QICS (Taylor et al., 2015) and ECO2 (Furre et al., 2017). STEMM-CCS is a scientific research project to simulate a sub-seafloor CO2 leak under real-life conditions in the North Sea (Blackford et al., 2018). One of the primary objectives of the project is to produce experimental data for the development and calibration of numerical models to simulate the leakage dynamics of CO2 out of the geoformations with the knowledge of dispersion time, pathways through faults and high permeable zones. In terms of marine physiochemical and biological impacts, for instance, the changes in pH due to leaked CO2 is one of the most significant data for the assessment of leakage and model calibrations. Adequate knowledge of CO2 dispersion through complex structure geoformation and dissolution characteristics of developed plume are mandatory and vital towards the development of the leakage prediction models. Sup-ported by the project, a so-called Arbitrary Navier-Stokes-Darcy mul-ti-fluid flow model (AnsdMF) has been developed for simulations of CO2 transportations through the sediments with complex structures into the turbulent ocean. In this study, the AnsdMF model is applied to simulate and predict the dynamic processes of initiations and developments of CO2 and CO2 solution plumes from injection ports through sediments to the ocean current in STEMM-CCS experiment. The numerical model settings, including the data collections, analysis, and reconstruction of sediments, are discussed in Section 2. The overall methodology of the numerical model including sub models of mass, momentum, and interfacial in-teractions among fluid-fluid-solid are explained in Section 3. The anal-ysis and discussions on the model diagnostic simulation results are made in Section 4 with the CO2 injection through pipe and leakage scenario results. Finally, the conclusions are drawn in Section 5. 2. Model sediment setup and data collection The CO2 release experiment of STEMM-CCS project was carried out in the vicinity of the Goldeneye platform located in a sandstone for-mation of Early Cretaceous near Scotland beneath Moray Firth (56-60 ̊N) (Dean and Tucker, 2017). In this project, gas CO2 supplied from gas tanks was controllably released from 3m underneath the seabed (Flohr et al., 2020). Ships, remotely operated vehicles (ROVs), autonomous underwater vehicles (AUVs), the gas sampling system, and related equipment and sensors were employed to measure the changes in physicochemical properties in seawater and brine and monitor the leakages of CO2. The gas release started on 11/05/2019 at 15:19hrs, which is set as the start of day 0 of the experiment (Flohr et al., 2021) in this study. The data of CO2 injection rate and the total CO2 injected, as shown in Fig. 1, are collected from field experiment as the input data of the modeling. 2.1.Porosity and grain size distribution To set up the model sediments, AnsdMF requires data of particle size distributions and the porosity distributions of sediments. The data ob-tained from a project pre-survey were collected, which are those sampled in 18 boxes at different locations around CO2 injection site down to 5.0 meters. The porosity data from core samples used in this study are plotted in Fig. 2, of which are the data from landers (POS527/ si83 and gravity corer GC-06-Station-102), pockmarks (GC-01-Station- 97 st.90/91 Obj 103), and well (GC-07-Station-103). The data shows that the average porosity is around 53% with +10% and –20% for the sediments deeper than 16 cm (Fig. 2 left), while, the decent distributions from surface to the depth of 16cm for shallow sediments (Fig. 2 right). Another set of data requested from setting the model is the particle size distributions, which are taken from Particle size analysis (PSA) for various samples collected at different depths of sediments (Lichtschlag et al., 2021). A brief discussion on the characteristics of the particle distribution will help for model setting and for modelling the dynamic process of CO2 gas penetration through sediments. From the data, shown in Fig. 3, it can be found that the deeper sediments, 3–4 m below the seabed (data of GC3 398-412cm), are mostly fine sand or laminated mud. The surface sediments down to about 0.5 m, however, are mostly particles of 60-100 μm with some smaller particles of less than 10 μm. In between those layers is a mixture of substances with varied grain sizes dominated within 8 to 100 μm. It must be noted that the very small particles within the sediments are more sensitive to additional distur-bances, such as the penetration of CO2 into the sediments, while the larger particles have a relatively larger inertia to withstand the distur-bances and keep their original positions. This behavior of particles can be utilized for diagnostic setting of the model sediments for prediction of CO2 dispersion with comparisons of observation data, such as the eruption time. 2.2.Reconstruction of model sediments by data The heterogenous sediments are sediments with complex structure and various pore throat size distributions. The model requires the permeability distribution, intrinsic permeability at the first stage, to predict the resistant forces of fluids and porous solids. For the hetero-geneous sediment or rock, the throat size varies depending on the inter- Fig. 1.Gas injection rate and total gas injected over the days from start in-jection collected from field experiment (Flohr et al., 2021). U. Saleem et al.