Eunseo Choi

GeoFramework: Coupling multiple models of mantle convection within a computational framework

Solver coupling can extend the capability of existing modeling software and provide a new venue to address previously intractable problems. A software package has been developed to couple geophysical solvers, demonstrating a method to accurately and efficiently solve multiscale geophysical problems with reengineered software using a computational framework (Pyre). Pyre is a modeling framework capable of handling all aspects of the specification and launching of numerical investigations. We restructured and ported CitcomS, a finite element code for mantle convection, into the Pyre framework. Two CitcomS solvers are coupled to investigate the interaction of a plume at high resolution with global mantle flow at low resolution. A comparison of the coupled models with parameterized models demonstrates the accuracy and efficiency of the coupled models and illustrates the limitations and utility of parameterized models.

Deformation in transcurrent and extensional environments with widely spaced weak zones

Previous mechanical models of the western U.S. have concluded that plate boundary forces cannot generate far-field deformation. Such models have ignored preexisting large-scale lithospheric strength variations, an assumption that appears to be inconsistent with seismically determined variations in lithospheric structure. We have formulated a three-dimensional viscous flow model with imposed plate motions, but include lateral zones of low viscosity. These models show that strain rates are concentrated in weak zones with adjacent blocks experiencing little deformation. Deformation can extend far inboard of plate boundaries, contrary to the result of previous studies with rheologically homogeneous plates, and apparently compatible with the variation is seismic velocity and GPS determined deformations in western U.S. These results suggest that plate boundary forces cannot be neglected in the deformation of the western U.S., including the Cenozoic extension of the Basin and Range Province.

Prediction of ground motion and dynamic stress change in Baekdusan (Changbaishan) volcano caused by a North Korean nuclear explosion

Strong ground motions induce large dynamic stress changes that may disturb the magma chamber of a volcano, thus accelerating the volcanic activity. An underground nuclear explosion test near an active volcano constitutes a direct treat to the volcano. This study examined the dynamic stress changes of the magma chamber of Baekdusan (Changbaishan) that can be induced by hypothetical North Korean nuclear explosions. Seismic waveforms for hypothetical underground nuclear explosions at North Korean test site were calculated by using an empirical Green’s function approach based on a source-spectral model of a nuclear explosion; such a technique is efficient for regions containing poorly constrained velocity structures. The peak ground motions around the volcano were estimated from empirical strong-motion attenuation curves. A hypothetical M7.0 North Korean underground nuclear explosion may produce peak ground accelerations of 0.1684 m/s2 in the horizontal direction and 0.0917 m/s2 in the vertical direction around the volcano, inducing peak dynamic stress change of 67 kPa on the volcano surface and ~120 kPa in the spherical magma chamber. North Korean underground nuclear explosions with magnitudes of 5.0–7.6 may induce overpressure in the magma chamber of several tens to hundreds of kilopascals.

Constraints on shallow mantle viscosity from morphology and deformation of fast‐spreading ridges

[1] We show that the morphology and the deformation of a fastspreadingridgeconstrainshallowuppermantleviscosity. A fast spreading center is simulated in a numerical model that couples tectonic deformation due to plate spreading and periodic dike emplacements. The amount of magma intruded into dikes or extruded is enough to make a 7 km thick crust and the axial lithosphere is fixed at ∼2 km thickness. For low viscosities (≤10 19 Pa·s),the model spreadingcenter is marked by an axial high with relief governed by the axial density structure. For high viscosities (≥10 20 Pa·s) the axial morphology is a valley and extensional brittle deformation is distributed away from the near‐axis region. The observed morphology and deformation of fast spreading ridges are not consistent with the high mantle viscosity estimated for very dry olivine, as some suggest may result from partial melting related dehydration. Citation: Choi, E., and W. R. Buck (2010), Constraints on shallow mantle viscosity from morphology and deformation of fast‐spreading ridges, Geophys. Res. Lett., 37, L16302, doi:10.1029/2010GL043681.

One-sided transform basins and "inverted curtains": Implications for releasing bends along strike-slip faults

[1] Some basins associated with bends along strike‐slip faults grow only on one side of the fault. We ascribe this asymmetry to a characteristic 3D geometry. The strike‐slip fault is planar and vertical, except at the “inverted curtain” where the fault has a sinusoidal bend that decreases linearly downward and vanishes below a prescribed depth. We model numerically the deformation around inverted curtains with different dips. We consider a crustal block traversed by a fault with lower cohesion than the surrounding rock and subject it to strike‐slip motion, first with purely elastic models and then with Mohr‐Coulomb elasto‐plastic ones. We found that when a curtain is vertical, both sides of the bend subside forming a symmetrical basin. In contrast, basins are formed only on the hanging wall side of a releasing bend of a non‐vertical curtain, in agreement with field examples. For the same curtain geometries, the basins are broader and shallower in an elastic crust than in an elastic‐plastic one. While releasing bends are stable over a range of dips, restraining bends are “shunted” by new straight faults except in shallow‐dipping curtains. These model results show that bends on transcurrent faults with inverted curtain geometries lead to asymmetric ridges and basins. Conversely, such a characteristically asymmetric basin (ridge) is symptomatic of a transform bend with oblique slip that attenuates with depth and is rooted below the subsided (uplifted) area. One implication is that master transcurrent faults may remain primarily or solely responsible for earthquake hazard along segments with anomalous vertical deformation.

Accelerating DynEarthSol3D on tightly coupled CPU-GPU heterogeneous processors

DynEarthSol3D (Dynamic Earth Solver in Three Dimensions) is a flexible, open-source finite element solver that models the momentum balance and the heat transfer of elasto-visco-plastic material in the Lagrangian form using unstructured meshes. It provides a platform for the study of the long-term deformation of earths lithosphere and various problems in civil and geotechnical engineering. However, the continuous computation and update of a very large mesh poses an intolerably high computational burden to developers and users in practice. For example, simulating a small input mesh containing around 3000 elements in 20 million time steps would take more than 10 days on a high-end desktop CPU. In this paper, we explore tightly coupled CPU-GPU heterogeneous processors to address the computing concern by leveraging their new features and developing hardware-architecture-aware optimizations. Our proposed key optimization techniques are three-fold: memory access pattern improvement, data transfer elimination and kernel launch overhead minimization. Experimental results show that our proposed implementation on a tightly coupled heterogeneous processor outperforms all other alternatives including traditional discrete GPU, quad-core CPU using OpenMP, and serial implementations by 67%, 50%, and 154% respectively even though the embedded GPU in the heterogeneous processor has significantly less number of cores than high-end discrete GPU. HighlightsWe accelerate Dynamic Earth Solution 3D program on CPU-GPU heterogeneous processors.We propose data transformation to improve GPU memory performance.We propose to merge kernels to minimize kernel launch overhead.We show performance gain over implementations on discrete GPU and multi-core CPU.