HR: 17:00h
AN: NG12D-04 [PDF]
TI: Data-mining Analysis and Multi-dimensional Visualization
of Earthquake Clusters in a GRID-Like Interactive Environment
AU: * Yuen, D A
EM: davey@krissy.geo.umn.edu
AF: Minnesota Supercomputing Instiute and Dept. of Geology and Geophysics, University of Minnesota,
Minneapolis, MN 55455 United States
AU: Dzwinel, W
EM: dzwinel@uci.agh.edu.pl
AF: A.G.H., Computer Science Institute, Krakow, 30-159
Poland
AU: Bollig, E F
EM: bollig@msi.umn.edu
AF: Minnesota Supercomputing Instiute and Dept. of Geology and Geophysics, University of Minnesota,
Minneapolis, MN 55455 United States
AU: Kadlec, B F
EM: kadlec@msi.umn.edu
AF: Minnesota Supercomputing Instiute and Dept. of Geology and Geophysics, University of Minnesota,
Minneapolis, MN 55455 United States
AU: Ben-Zion, Y
EM: benzion@usc.edu
AF: Dept. of Earth Sciences, University of Southern California, Los Angeles, CA 98809 United States
AU: Yoshioka, S
EM: yoshioka@geo.kyushu-u.ac.jp
AF: Dept. of Earth and Planetary Sciences, Kyushu University, Fukuoka, 813
Japan
AB:
We have developed an interactive web-based scheme for data-mining
the spatio-temporal patterns of many earthquakes. This novel technique is
based on cluster analysis of the multi-resolutional structures
of earthquakes. The interactive scheme is based on a client-server
paradigm in which we have used the off-screen rendering technique
to facilitate the visual interrogation. A powerful 3-D visualization package
Amira ( www.amiravis.com ) is also used to visualize the complex clusteral patte
nrs in a reduced dimensional
space. We have applied our method to observed and synthetic tic seismic
catalogs. The observed data represent seismic activities situated
around the Japanese islands in the 1997-2003 time interval. The synthetic
data were generated by numerical simulations for various cases of a heterogeneous
fault governed by quasi-analytical 3-D elastic dislocation models
.At the highest resolution, we analyze the local cluster structure in the data
space of seismic events for the two types of catalogs by using an agglomerative clustering algorithm. We demonstrate that
small magnitude events produce
local spatio-temporal patches corresponding to neighboring large events.
Seismic events, quantized in space and time, generate the multi-dimensional
feature space of the earthquake parameters. Using a non-hierarchical clustering
algorithm and multi-dimensional scaling, we explore the multitudinous
earthquakes by real-time 3-D visualization and inspection of multivariate
clusters. At the resolutions characteristic of the earthquake parameters,
all of the ongoing seismicity before and after largest events accumulate to
a global structure consisting of a few separate clusters in the feature space
. We show that by combining the clustering results from low and high resolution
spaces, we can recognize precursory events more precisely. We will discuss
how this WEB-IS ( Web-Interrrogative system ) would work.
One can also access this by going to the URL http://boy.msi.umn.edu/web-is/. Its implementation and deployment in light of
future GRID-computing will be discussed in terms of the recently developed Narada-Brokering (distributed messaging )
system of publishing and subscribing . This will provide
a scalable infrastructure for several applications involving
a set of nodes communicating with each other.
.
UR: http://boy.msi.umn.edu/web-is
DE: 7215 Earthquake parameters
DE: 7223 Seismic hazard assessment and prediction
SC: Nonlinear Geophysics [NG]
MN: 2003 Fall Meeting