Alexey Romanenko

Using systematically characterized low-frequency earthquakes as a fault probe in Guerrero, Mexico

Studies of low-frequency earthquakes (LFEs) have focused on detecting events within previously identified tectonic tremor. However, the principal LFE detection tools of matched-filter searches are intrinsically incapable of detecting events that have not already been characterized previously as a template event. In this study, we therefore focus on generating the largest number possible of LFE templates by uniformly applying a recently developed LFE template detection method to a 2.5 yearlong data set in Guerrero, Mexico. Using each of the detected templates in a matched-filter search, we then form event families that each represents a single source. We finally develop simple, empirical statistics to select the event families that represent LFEs. Our resulting catalog contains 1120 unique LFE sources and a total of 1,849,486 detected LFEs over the 2.5 yearlong data set. The locations of the LFE sources are then divided into subcatalogs based on their distance from the subduction trench. Considering each LFE as a small unit of slip along the subduction interface, we observe discrete episodes of LFE activity in the region associated with large slow-slip events; this is in direct contrast to the near-continuous activity observed 35 km farther downdip within the previously identified LFE/tremor sweet spot.

High-Performance Tsunami Wave Propagation Modeling

Strongest earthquake of December 26, 2004 generated catastrophic tsunami in Indian Ocean. This shows that, in spite of recent technology progress, population at coastal zone is not protected against tsunami hazard. Here, we address the problem of tsunami risks mitigation. Note that prediction of tsunami wave parameters at certain locations should be made as early as possible to provide enough time for evacuation. Modern computational technologies can accurately calculate tsunami wave propagation over the deep ocean provided that initial displacement (perturbation of the sea bed at tsunami source) is known. Modern deep ocean tsunameters provide direct measurement of the passing tsunami wave in real time, which help to estimate initial displacement parameters right after the tsunami wave is recorded at one of the deep ocean buoys. Therefore, fast tsunami propagation code that can calculate tsunami evolution from estimated model source becomes critical for timely evacuation decision for many coastal communities in case of a strong tsunami. Numerical simulation of tsunami wave is very important task for risk evaluation, assessment and mitigation. Here we discuss a part of MOST (Method of Splitting Tsunami) software package, which has been accepted by the USA National Ocean and Atmosphere Administration as the basic tool to calculate tsunami wave propagation and evaluation of inundation parameters. Our main objectives are speed up the sequential program, and adaptation of this program for shared memory systems (OpenMP) and CELL architecture. Optimization of the existing parallel and sequential code for the task of tsunami wave propagation modeling as well as an adaptation of this code for systems based on CELL BE processors (e.g. SONY PlayStation3) is discussed. The paper also covers approaches and techniques for programs optimization and adaptations, and obtained results.

Service-oriented tsunami wave propagation modeling tools

A big variety of methods and services for Tsunami Modeling require developing special approaches supporting integration of these heterogeneous components with respect to the various development platforms and architectures. In this paper, we are considering a way of transforming the Tsunami Wave Propagation Modeling Software to a Service-Oriented Architecture (SOA). The transformation was implemented based on the original Virtual MVC-design pattern (VMVC-pattern) that is demarcating a Functional (View) and an Implementation (Model) task by inducing an Integrator (Controller). The paper describes basic features of the MOST (Method of Splitting Tsunami) software package that was used as an initial Propagation Software Engine. It also includes an analysis the initial and output tsunami data and the description of a VMVC-based prototype in which control and resource management operations can be implemented via WEB-interface.

Modern Hardware to Simulate Tsunami Wave Propagation

Problem of tsunami risks evaluation, assessment and mitigation is here addressed. Modern computational technologies are able to calculate accurately tsunami wave propagation over the deep ocean provided that initial displacement (perturbation of the sea bed at tsunami source) is known. Modern deep ocean tsunameters provide direct measurement of the passing tsunami wave in the real time mode, which help to estimate initial displacement parameters right after the tsunami wave is recorded at one of the deep ocean buoys. Therefore, fast tsunami propagation code that can calculate tsunami evolution from estimated model source becomes critical for timely evacuation decision for many coastal communities in case of a strong tsunami. In this paper we discuss a part of MOST (Method of Splitting Tsunami) software package, which has been accepted by the USA National Ocean and Atmosphere Administration as the basic tool to calculate tsunami wave propagation and to evaluate of inundation parameters. Our main objectives are to speed up the existing sequential program, and to adapt this program for shared memory systems (OpenMP), CELL BE architecture and GPU. For caring out this research we use SMP server and a system build on IBM CELL BE CPU. We also use NVIDIA Tesla C1060 board which acts as co-processor for CPU. Optimization of the existing parallel and sequential code for the task of tsunami wave propagation modeling as well as an adaptation of this code for GPU and CELL BE processor is discussed. The obtained results are up to 170 times performance gain for one time step. These are very promising.

3D Wave-packet decomposition implemented on GPUs

Decomposition of seismic data into wave-packet representations has been successfully used for 2D data compression, interpolation and de-noising. In this paper we present a fast implementation of a 3D wave-packet decomposition using graphical processing units (GPUs). This allows for the similar processing of 3D seismic gathers. We discuss parallel implementation of the wave-packet transform on GPUs as opposed to existing algorithms (sequential and MPI-parallel). A few computational steps had to be modified adapting them for a GPU platform. The code has been tested on a 3D data set of size 2563, where we obtain speedup of about 40 times compared to the sequential code performance.

Implementation of Mac-Cormack scheme for the fast calculation of tsunami wave propagation

In this paper, we continue the discussion about of fast numerical simulation of tsunami wave propagation. In case of seismic event offshore Japan, tsunami approaches the nearest shore in approximately 20 minutes. As the calculation of tsunami wave propagation normally serves as a part of any warning system, speeding up this software results in overall better performance for tsunami warning. Implementation of the Mac-Cormack scheme to solve the shallow water system is here described. Field programmable gates array (FPGA) microchip is used to achieve better performance. The numerical solution is compared with the available exact solution to the problem considered.

Optimal structure of the cost functional for coastal profile evolution problems

Behaviour-oriented diffusion models, available in literature, reasonably describe the time evolution of the cross-shore position of coastal profiles (depth as a function of the distance to shore line). In previous works, we were able to identify the time-independent, but space-dependent coefficients, in 1D (and later in 2D) behaviour-oriented diffusion models for the time dynamics of long-term coastal profile evolution. Models with such coefficients even predict the depth profile evolution for some years ahead. We now focus on the structure of the so-called Cost Functional (CF), which is used to identify the coefficients above. An optimized version of CF provides smaller relative error predicting the depth profile evolution.

Determination of initial tsunami wave shape at sea surface

Any Tsunami Early Warning System (TEWS) should provide reliable prediction of the wave parameters as early as possible. Numerical calculation of the tsunami wave propagation over the particular part of the ocean can be implemented rather efficiently using modern high-performance system. It is possible when the initial wave shape at tsunami source is given. In this paper we discuss and show the fast algorithm for reconstructing the initial sea surface displacement at tsunami source. Based on the Fourier approximation theory, the presented algorithm treats the measured wave profile as a linear combination of synthetic (calculated) waves from the so-called unit sources. Using an artificial bathymetry, we evaluate the algorithm parameters focusing on its precision and performance. Moreover, a large tsunami sources (several hundred kilometers long) are also simulated.

Brain's tumor image processing using shearlet transform

Brain tumor detection is well known research area for medical and computer scientists. In last decades there has been much research done on tumor detection, segmentation, and classification. Medical imaging plays a central role in the diagnosis of brain tumors and nowadays uses methods non-invasive, high-resolution techniques, especially magnetic resonance imaging and computed tomography scans. Edge detection is a fundamental tool in image processing, particularly in the areas of feature detection and feature extraction, which aim at identifying points in a digital image at which the image has discontinuities. Shearlets is the most successful frameworks for the efficient representation of multidimensional data, capturing edges and other anisotropic features which frequently dominate multidimensional phenomena. The paper proposes an improved brain tumor detection method by automatically detecting tumor location in MR images, its features are extracted by new shearlet transform.

Modern simulation tools for real time numerical simulation of ocean-related processes

In this paper we focus on two particular ocean-related problems, valuable progress in which is expected due to extended facilities of the modern computer architectures. The first one concerns real time tsunami risk mitigation by using the advantages of such computer architectures as are Graphic Processing Units (GPU) and Field Programmable Gates Arrays (FPGA). Among many important aspects of tsunami simulation we study the time consumable calculation of the wave propagation over a given water area. A comparison of performances achieved at a range of architectures is given. Secondly, we discuss the coastal profile evolution. As it has been found relatively recently, behavior-oriented diffusion models reasonably describe the time evolution of the cross-shore position of coastal profiles. Two time-independent coefficients in the governing equation, which embody the relevant physical properties, are identified simultaneously. Earlier, the authors have validated and calibrated numerically the proposed model, processing two sets of real data, the first one being measured over 10 years at Duck, in North Carolina (USA), the second one obtained over 39 years measurements at Delfland (Holland). Here, the model dependence on the alongshore position of the observation point is studied. The coefficients of the model equation are determined by means of a certain iteration process. As it was observed, the achieved convergence is now better than when several separate observations along the coast are involved.