Changes in the Performance of the IMS Infrasound Network due to Seasonal Propagation Effects David Norris and Robert Gibson BBN Technologies 1300 N. 17.

Changes in the Performance of the IMS Infrasound Network due to Seasonal

Propagation Effects 

David Norris and Robert Gibson

BBN Technologies 1300 N. 17th Street

Arlington, VA 22209

Infrasound Technology WorkshopLa Jolla, California

27-30 Oct 2003

Outline

• Network Performance Issues

• Simulation Approach

• Results (multimedia!)

• Conclusions

Network Performance Factors

• Network coverage

• Array performance and signal-to-noise ratio

• Propagation effects and uncertainties

Network Coverage

• Lack of azimuthal coverage can lead to elongated error ellipses

• Example: Pacific bolide 23 Apr 01:

IS57

NVIAR

DLIAR

IS59

IS10

IS26

Source

Signal-to-Noise Ratio

• Some stations are inherently noisier than others– Full exposure vs. cover (e.g. tree canopy)– Island vs. mainland– Regional wind conditions (e.g. Windless Bight vs. Palmer)

• Signal gain– Nominal array beamforming gain: 10log(N)– Nominal Bandwidth gain: 5log(W)

• To improve SNR– Increase number of sensors– Improve wind filter– Advanced signal processing

Propagation factors

• Stratospheric arrival– Shorter path– Less absorption– Duct presence depends

on stratospheric winds

With wind Counter wind

• Thermospheric arrival– Longer path– More absorption– Duct always present

Stratosphericduct

thermosphericduct

Propagation Parameterizations

• Signal Strength– Empirical equations that

account for effect of stratospheric winds

– Received pressure (P) function of range (R), yield (W) and winds at 50 km (Vs)

– 38 dB difference between 50 m/s upwind and downwind propagation

)(019.0)log(36.1)log(68.037.3)log( sVRWP +−+=

•Mutschlecner, J. et al., “An Empirical Study of Infrasonic Propagation,” Los Alamos National Laboratory report LA-13620-MS, 1999.

•Stevens, J. et al., “Infrasound Scaling and Attenuation Relations from Soviet Explosion Data and Instrument Design Criteria from Experiments and Simulations,” Proceedings of the 21st Seismic Research Symposium, Las Vegas, NV, 1999.

Propagation Parameterizations

• Azimuthal uncertainty– Theoretical formulations,

based on• Signal-to-noise ratio• Signal and noise

coherence• Array geometry

– Empirical formulations

•R. Shumway and S. Kim, “Signal Detection and Estimation of Directional Parameters for Multiple Arrays,” Defense Threat Reduction Agency Technical Report DSWA-TR-99-50, 2001.

•Blandford, R., “Detection and Azimuth Estimation by Infrasonic Arrays as a Function of Array Aperture and Signal Coherence,” AFTAC report, 1998.

•C. Szuberla, “Array Geometry and the Determination of Uncertainty,” Infrasound Technology Workshop, Kailua-Kona, HI, 2001.

•Clauter, D. and R. Blandford, “Capability Modeling of the Proposed International Monitoring System 60-Station Infrasonic Network,” Proceedings of the Infrasound Workshop for CTBT Monitoring, Santa Fe, NM. Los Alamos National Laboratory report LA-UR-98-56, 1997.

InfraMAP

• InfraMAP is a software tool kit– Infrasonic Modeling of Atmospheric Propagation

• Designed for infrasound researchers and analysts

• Supports infrasonic-relevant R&D– Sensitivity studies– Network performance– Modeling specific sources of interest

IMS Coverage Simulation

• Goal: – Characterize seasonally-dependent effects of stratospheric

ducting on localization accuracy (AOU).

• Previous studies– Clauter, D. and R. Blandford, “Capability Modeling of the

Proposed International Monitoring System 60-Station Infrasonic Network,” Proceedings of the Infrasound Workshop for CTBT Monitoring, Santa Fe, NM. Los Alamos National Laboratory report LA-UR-98-56, 1997.

– E. Blanc and J. L. Plantet, “Detection Capability of the IMS Infrasound Network: A More Realistic Approach,” Proceedings of the Informal Workshop on Infrasounds, Bruyeres-Le-Chatel, France, 1998.

• Simulation parameters:

IMS Coverage Simulation

Variable Value/Description Comments

Background Station Noise 0.5 PaLow wind noise condition. Assume uniform across network

Array configuration 4 element, 1 km baseline Standard IMS array configuration

Array Gain, 10log(N) 6.0 dB Assume correlated signal across array

Bandwidth Gain, 5log(W)Strato (W=4 Hz): 3 dB

Thermo (W= 2 Hz): 1.5 dBProcessing over 1 sec window

Signal velocity UncertaintyStrato: 0.01 km/s

Thermo: 0.02 km/sOn order of that assumed in Blandford, 1998

Azimuthal Uncertainty Fit to data in Clauter and Blandford, 1997

SNR Detection Threshold 2

Source Yield 10kT

Received Signal Strength LANL wind-corrected eqn.Stratospheric winds at 50 km found from HWM averaged along propagation path

IMS Performance

Number of station detections

15

0

5

10

IMS Performance

Number of station detections

15

0

5

10

IMS Performance

Area of Uncertainty Radius (km)

0

400

300

200

100

IMS Performance

Area of Uncertainty Radius (km)

0

400

300

200

100

Conclusions• Performance of IMS network strongly dependent on seasonally

varying flow of stratospheric winds– Winter

• Northern Hemisphere: East flow• Southern Hemisphere: West flow

– Summer• Northern Hemisphere: West flow• Southern Hemisphere: East flow

• Localization capabilities of a given station improve in direction of stratospheric headwinds

• Recognized area of poor coverage: Southern Ocean• Shift in “Hole” in coverage:

– January: Off West coast of South America – July: Off of East coast of New Zealand

• AOU Radius southeast of Easter Island– January: > 400 km– July: < 100 km

Future Research

• Include station configuration– Number and location of elements– Wind filter properties

• Characterize local station background noise

• Improved characterization/modeling of propagation effects

– Signal strength– Azimuthal bias and uncertainty– Signal velocity and associated uncertainty