The Proceedings of the Eighth International Conference on Creationism (2018)

experimental data from the literature for olivine, we were able to calibrate our ISV model including pressure sensitivity within the reported experimental uncertainty. Considering uncertainty is important, because of the lack of experimental recrystallization data at higher pressures. However, Hansen et al. (2012) demonstrated that the recrystallization process weakens more than 90% of total hardening in olivine aggregates under deformations at 1473 K in temperature and 300 MPa in pressure. Based on this observation, we chose pressure dependent parameters that cause the volume fraction of recrystallization gradually to decrease as the pressure increases. By this means the weakening from recrystallization in this model is made to agree with experimental measurements. The comparison between the experimental stress-strain behavior and our calibrated model is plotted in Fig. 4, and the constants and parameters are summarized in Table 1. Once the recrystallized volume fraction increases, the stress-strain behavior shows a concave downward shape (i.e., the slope decreases). In the present study, olivine’s properties alone were applied to the whole mantle, even though the mantle has multiple mineral constituents as well as important phase transitions. Hence, it is necessary for the present model to make large extrapolations of olivine’s mineralogical and mechanical properties because the large ranges of pressure and temperature across the mantle. In order to alleviate the unrealistic predictions, some parameters for the hardening variables were slightly reduced to the strength level that the few experimental studies (e.g., Girard et al. 2016) provide for mantle-like temperature and pressure conditions. G. Simulation setup in TERRA Once the ISV model has been calibrated, the initial and boundary conditions for the finite element analysis using TERRA need to be identified. In particular, for the purposes of this work we sought to capture the mechanics instability and associated runaway behavior exhibited in the Flood cataclysm. In the TERRA2D simulations, we used a rectangular grid with approximately 131,000 elements (512 on the surface and 256 through the depth) corresponding to a domain 5,780 km in width and 2,890 km in depth as displayed in Fig. 6 (the element size was approximately 11 km). The TERRA3D calculations used a grid with approximately 1,400,000 cells with the shape of hexagonal prisms (33 spherical layers of cells in radial direction with 42,250 cells in each shell). Regarding the initial microstructural states when the Flood started, more studies on these microstructural effects are frankly needed, because it is likely that microstructural rearrangements from flow and other processes in the mantle already had been taking place after the creation week until the Flood started. For example, the grains of the mantle rocks could have grown during that time from Day 3 of Creation until the Flood (~1600 years). Although these initial microstructural effects are not included in this paper, such effects might have played a crucial role in triggering the Flood itself (Horstemeyer et al. 2002). Regarding the initial grain size when the simulation started, it was simply set to 100 µm throughout the entire mantle. To alleviate numerical singularities, the statically and dynamically recrystallized volume fractions at the initial stage were set to a very low number, 5 x 10 -7 , instead of zero. Also, the initial hardening variables were set to zero. From a Biblical standpoint we have reason to suspect some uncertainty with the initial conditions in our simulations. The Bible seems to imply high rate, large, and complex deformation on Day 3 of Creation Week when God gathered the waters below the firmament together into one place and caused the dry land to appear. Such an event almost certainly would have generated major heterogeneities throughout the mantle. Because we have no way even of guessing what those variations in microstructural properties were like, in this paper, we therefore assume very simple and largely uniform initial conditions and focus on the mantle dynamics and related weakening effects that develop from that admittedly simplified initial state. In future simulations we will explore the different possible deformational histories prior to the Flood. The viscosity in our numerical models depends on the local temperature, pressure, and strain rate. Assuming that the strain rate was very low before the Flood began, we prescribe a depth- dependent reference viscosity η using following equation: (19) where η 0 is the reference viscosity, is the activation energy for the mantle’s viscous response, P is the pressure, is the activation volume for the pressure dependence, R is the gas constant, and T is the absolute temperature. In order to match closely the earth’s current viscosity structure, separate values for E * and V * are applied to the upper mantle, mantle transition zone, and lower mantle (Baumgardner 1994). And the reference viscosity η 0 was set to 10 23 Pa·s. The initial viscosity structure used in these simulations is shown in Fig. 5. Also because we do not know what the temperature distribution within the earth might have been at the onset of the Flood, we assume an extremely simple initial temperature distribution as depicted in Fig. 6. This distribution corresponds to a horizontal variation in temperature across the computational domain of a single period of the cosine function as follows: (20) where T is the absolute temperature, x and y are the horizontal and vertical points of the mantle domain in the TERRA2D, respectively, T ref is the depth dependent reference temperature profile to match the temperature gradient implied by the physical parameters from the Preliminary Reference earth Model (PREM) (Dziewonski and Anderson 1981), w and h are the width and height of the TERRA2D domain, respectively. We set the initial temperature to 300 K on the top surface and 3700 K at the core-mantle boundary with this complex thermal gradient in between. It is this non-uniform temperature distribution in the horizontal direction that drives the flow in the mantle which eventually becomes unstable because of the rock weakening mechanisms. The horizontal temperature gradient in the upper boundary layer produces flow toward the center of the domain which causes the middle part of the boundary layer to thicken. Eventually this middle part of the boundary layer becomes gravitationally unstable, and runaway of a cold blob of rock that had been in that boundary layer ensues. Cho et al. ◀ Strength-reducing mechanisms in mantle rock during the Flood ▶ 2018 ICC 714