1 Location and Characterization of Infrasonic Events Roger Bowman 1 , Greg Beall 1 , Doug Drob 2 , Milton Garces 3 , Claus Hetzer 3 , Michael O’Brien 1 , Gordon Shields 1 1. Science Applications International Corporation 2. Naval Research Laboratory 3. University of Hawaii Infrasound Technology Workshop University of California, San Diego October 27-30, 2003
27
Embed
1 Location and Characterization of Infrasonic Events Roger Bowman 1, Greg Beall 1, Doug Drob 2, Milton Garces 3, Claus Hetzer 3, Michael O’Brien 1, Gordon.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Location and Characterization of Infrasonic Events
Roger Bowman1, Greg Beall1, Doug Drob2, Milton Garces3, Claus Hetzer3, Michael O’Brien1, Gordon Shields1
1. Science Applications International Corporation2. Naval Research Laboratory
3. University of Hawaii
Infrasound Technology Workshop
University of California, San Diego
October 27-30, 2003
2
Outline
• Challenges• Approach• Data sets• Atmospheric models• Travel-time tables• Characterization and visualization• Ongoing work• Summary
3
Challenges in Infrasound Monitoring
Source
Propagation
Receiver
Challenge Approach Lack of signals of interest Scale atmospheric nuclear explosion and
embed in ambient noise Abundance of clutter Characterize and reject clutter Time-variant propagation media
Use NRL’s G2S atmospheric models
Lack of monitoring stations Use all current stations Characterize performance of all stations Simulate detection performance of
complete network Lack of ground truth events Assemble ground truth data sets
4
Location Approach
HWM/MSISEmodels
NRL G2Smodels
Ray tracing(tau-p)
Travel-time tables
Uncertaintyestimation
Stations
Canonicallocation data set
Signal observations
Location algorithm
Travel-time tables
Ray tracing(tau-p)
Event locations
Location evaluation
Event times Arr
ival
tim
esA
zim
uths
5
Project Network
Acrobat Document
• All stations available in June 2003
6
Canonical Location Data Set
• Focuses on signals with ground truth locations• Waveforms and arrivals• Multiple station detections
– For assessing location, azimuth, and travel-time estimates– Chemical explosions: GT1-101 (3 events)– Moving sources: GT100 (3 events)
• Single station detection– For assessing azimuth and travel-time estimates– Mining explosions GT10-15 (5 events)– Chemical explosions: GT1-20 (5 events)– Gas pipe explosion: GS1 (1 events)– Earthquakes: GT5-10 (2 events)
• Meridional winds for a location in the southwest United States
• 0000 UT for January 1-25, 2003
9
Travel-Time Tables: PIDC
• Prototype International Data Center (PIDC) ca. 2001• Use HWM and MSISE climatological models
– Horizontal Wind Model (HWM)– Mass Spectrometer, Incoherent Scatter – Extended (MSISE)
• Use David Brown’s ray tracing program• Include travel times for “I” phase only• Depend on azimuth and season
– 1o azimuthal resolution; 1.8o radial resolution
• Use uncertainties based on possible phase misidentification
10
Travel-Time Tables: Automatic Processing
• Use HWM/MSISE climatological models• Use Milton Garces’ tau-p ray tracing program• Include travel times for stratospheric (Is),
thermospheric (It) and undetermined (I) phases• Depend on azimuth, month and time of day
– 19 stations x 4 times of day x 12 months x 3 phases =2,736 tables!
– 1o azimuthal resolution; 1.5o radial resolution
– 0o-120o range
• Use uncertainties based on variability of G2S models for each month
11
HWM/MSISE Travel-Time Table: DLIAR
• January
• 0000 UT
• Back-azimuth: 200o
• 2 out of 33,840 curves
12
HWM/MSISE Travel-Time Table: DLIAR
• 0000 UT
• Is phases do not exist for some azimuths
• Longer travel times westbound from source to receiver
13
HWM Travel-Time Uncertainties
• Non-Gaussian distribution of predicted travel times
• Scatter in modeled travel times increases monotonically with range
• Characterize uncertainty by standard deviation at two ranges
• Interpolate for other ranges
14
Accounting for Range Dependence
Distance
Tim
e
0
Met 1at receiver
Met 2mid- dist
Met 3near max dist
2) Times shiftedbased on offset of
fits to adjacentsegments
3) Final fit toshifted points of
all segments
Illustration of rudimentary range-dependence in Tau-P ray tracing
1) Rays traced formulitple met profiles
• Accounts for variation of atmospheric model along range
• Use 1-D ray tracing for four models along profile
• Final curve is 4th degree polynomial
15
Travel-Time Tables: Interactive Analysis
• Use Naval Research Laboratory’s Ground-to-Space (G2S) models
• Dependent on azimuth, date and time of day– Tables calculated for stations as needed
• Include travel times for stratospheric (Is), thermospheric (It) and undetermined (I) phases
• Use uncertainties based on variability of travel-time with take off angle for G2S models for each month
16
G2S Travel-Time Table: DLIAR
• 1000 km range
• January 23, 2003
• 2000 UT
• Similar to HWM travel times
17
HWM and G2S Travel Time Tables
• 2000 km range
• January 23, 2003 2000 UT.
• January, 1800 UT
• Azimuth range for existence of Is phases differs
• All G2S travel times are larger than HWM in this example
18
Source-Size Estimation
• Implemented Brown (1999) formula in libmagnitude– M = log10P + 1.36log10R – 0.019v
– Where • P is pressure
• R is range
• v is wind velocity
• Preliminary version uses wind at infrasound stations from G2S model
19
Visualization Tools for Characterization
Infra EventMapping
Array Tool
Feature Plotting
Feature Animation
Analyst Review Station
• Seismic• Hydroacoustic• Infrasound
• libinfra• libPMCC• Spectrograms
• Frequency• Apparent velocity• Azimuth
20
Infra Mapping Tool
• Supports “tip-and-queue” processing
• Integrated with Analyst Review Station (ARS)– Arrival information
sent back and forth
• Zoom capability
• Topography resolution varies with map scale
21
Array Tool - Features
• Watusi explosion at NTS
• “libPMC” features
• “libinfra” features
• Waveforms
22
Array Tool - Spectrograms
Array Tool
• Watusi explosion at NTS
• Standard spectrogram
• Coherence spectrogram separates coherent signal from incoherent noise
• Waveforms
23
Feature Animation Tool
• Maps features to: – x-axes– y-axes– Color– Saturation– Animation sequence
• Supports 3-D animations
24
Feature Animation Tool (2)
0.8 Hz 4.8 Hz
…can animate over any variable
25
Ongoing Work
• Location– Test location algorithm using new travel time curves– Complete travel-time tables for location event data set– Quantify changes in capability to estimate location and