Waveform Modeling and Comparisons with Ground Truth Events David Norris BBN Technologies 1300 N. 17 th Street Arlington, VA 22209 [email protected] 703-284-1348 14 Nov 2007
Dec 16, 2015
Waveform Modeling and Comparisons with Ground Truth Events
David Norris
BBN Technologies
1300 N. 17th Street
Arlington, VA 22209
703-284-1348
14 Nov 2007
Motivation
• Quantify terrain effects on predicted travel times and waveform parameters
• Provide seamless integration between terrain and atmospheric specifications
• Improve overall propagation modeling capabilities and suite of tools that can be utilized by analysts and researchers
InfraMAP – Infrasonic Modeling of Atmospheric Propagation
Terrain Modeling Approaches
• Goal is to quantify the lower boundary condition for propagation modeling over complex terrain
• General approaches:
• Stair-stepping or Terrain Masking– Terrain realized as a series of stair steps– Results in the terrain being modeled as a series of knife-edge diffractors– All surface reflections approximated by tip diffraction– Assumes perfectly reflecting surface
Terrain Modeling Approaches (Cont.)
• Piecewise Conformal Mapping– Propagation coordinates are transformed into terrain-following arcs– Each arc is applied over a defined piece of the total terrain profile– Helmholtz equation and fundamental form of PE solution do not change– The mapping in implemented by applying a modified index of refraction
over a uniform grid
Terrain Modeling Approaches (Cont.)
• Piecewise Linear Shift Mapping– New coordinate system based on shifting height to follow terrain– Recomputed at each range step – Results in rederivation of Helmholtz equation– Additional terms must be numerically addressed in PE solution
r
z
T(r)
r
U(r) = z – T(r)
Terrain Masking Approach
Range (km)
He
igh
t (km
)
PE with absorption: Amplitude Field (dB re 1 km)
0 50 1000
5
10
15
20
-100
-80
-60
-40
-20
0
Range (km)
He
igh
t (km
)
PE with absorption: Amplitude Field (dB re 1 km)
0 50 1000
5
10
15
20
-100
-80
-60
-40
-20
0
Range (km)
He
igh
t (km
)
PE with absorption: Amplitude Field (dB re 1 km)
0 50 1000
5
10
15
20
-100
-80
-60
-40
-20
0
Range (km)
He
igh
t (km
)
PE with absorption: Amplitude Field (dB re 1 km)
0 50 1000
5
10
15
20
-100
-80
-60
-40
-20
0
Range (km)
He
igh
t (km
)
PE with absorption: Amplitude Field (dB re 1 km)
0 50 1000
5
10
15
20
-100
-80
-60
-40
-20
0
Range (km)
He
igh
t (km
)
PE with absorption: Amplitude Field (dB re 1 km)
0 50 1000
5
10
15
20
-100
-80
-60
-40
-20
0
Mt. Everest
Mt. Fuji
Mt. Everest
Mt. Fuji
• Terrain masking integrated into Parabolic Equation (PE) and Time-Domain Parabolic Equation (TDPE) models
• Test case of PE predictions with 10 km wedge
2.0 Hz 0.2 Hz
Hypothetical Scenario
• Source just East of Mt. Fuji
• 400 km propagation to West
• PE predictions at 2 Hz– Significant interference pattern
near source due to nearby terrain features
– Effects at range not obvious 0 50 100 150 200 250 300 3500
500
1000
1500
2000
2500
Range (km)
Hei
ght (
m)
ETOPO2 Topography
Distance = 400.0 kmBearing = 272.8 deg
Range (km)
He
igh
t (km
)
PE with absorption: Amplitude Field (dB re 1 km)
0 100 200 300 4000
20
40
60
80
100
120
140
-100
-80
-60
-40
-20
0
Range (km)
He
igh
t (km
)
PE with absorption: Amplitude Field (dB re 1 km)
0 100 200 300 4000
20
40
60
80
100
120
140
-100
-80
-60
-40
-20
0No Terrain Terrain
Hypothetical Scenario (cont.)
• PE predictions 1 km above sea level at 2 Hz
– First arrival with terrain 30 km shorter in range
– First arrival structure more complex with terrain
• TDPE predictions over 1-3 Hz at 207 km range
– Waveform with terrain more disperse than without terrain
– Decrease in peak amplitude in terrain case
Terrain
0 50 100 150 200 250 300 350 400-100
-80
-60
-40
-20
0
Range (km)
Am
plit
ud
e (
dB
re
1 k
m)
PE with absorption: Amplitude vs. Range, Height: 1 km
terrainno terrain
800 850 900 950 1000 1050 1100-4
-3
-2
-1
0
1
2
3
4
Absolute Travel Time (sec)
Am
plit
ud
e (
Pa
)
Time-domain PE with absorption: Amplitude vs. Time, Height: 1 km
terrainno terrain
Henderson Event
• Two major chemical explosions occurred in Henderson, Nevada on May 4, 1988
• Explosions resulted from plant fire which ignited stores of ammonium perchlorate
• Estimated surface burst yields of 0.7 and 1.8 kT
“Infrasonic Signals from the Henderson, Nevada, Chemical Explosion,” Mutschlecner, J. and R. Whitaker, LA-UR-06-6458, 2006
Detection at SGAR
• Infrasound detections– St. George, Utah
(SGAR), range 159 km– Los Alamos, New
Mexico (LANL), range 774 km
• SGAR– Both major explosions
observed with similar waveforms
– Two arrivals• First arrival consistent
with surface wave (group velocity 332 m/s)
• Second arrival consistent with stratospheric or thermospheric path (group velocity 275 m/s)
(meters)
-6000
0
6000
longitude (deg)
latit
ud
e (
de
g)
ETOPO2: Topography (meters)
UNITED STATES
S R S
R
-115 -110 -105
32
34
36
38
40
42
0 20 40 60 80 100 120 140 1600
200
400
600
800
1000
1200
1400
1600
1800
Range (km)
Hei
ght (
m)
ETOPO2 Topography
Distance = 168.6 kmBearing = 49.3 deg
Range (km)
He
igh
t (km
)
0 50 100 1500
20
40
60
80
100
120
140
-100
-80
-60
-40
-20
0
Range (km)
He
igh
t (km
)
0 50 100 1500
20
40
60
80
100
120
140
-100
-80
-60
-40
-20
0No Terrain Terrain
PE Prediction: 0.2 Hz
• PE prediction at 0.2 Hz through climatology (HWM/MSIS)• No energy predicted to reach SGAR due to lack of stratospheric duct• Terrain results in deeper shadow zone
Small-Scale Atmospheric Variability• Mean Zonal and Meridional Winds specified using NRL-G2S
• Horizontal Wind perturbations specified using Spectral gravity wave model– Range-dependent Horizontal Correlation Length: 50 km
• Total wind field along propagation path realized from sum of mean and perturbed components
NRL-G2S Mean Total RealizationGravity Wave Perturbation
range (km)
he
igh
t (km
)
0 50 100 150 200 2500
20
40
60
80
100
120
-40
-30
-20
-10
0
10
20
30
40
range (km)
he
igh
t (km
)
0 50 100 150 200 2500
20
40
60
80
100
120
-40
-30
-20
-10
0
10
20
30
40
range (km)
He
igh
t (k
m)
0 50 100 150 200 2500
20
40
60
80
100
120
-40
-30
-20
-10
0
10
20
30
40
Wind along propagation Path
Range (km)
He
igh
t (km
)
PE with absorption: Amplitude Field (dB re 1 km)
0 50 100 1500
20
40
60
80
100
120
140
-100
-80
-60
-40
-20
0
Range (km)
He
igh
t (km
)
PE with absorption: Amplitude Field (dB re 1 km)
0 50 100 1500
20
40
60
80
100
120
140
-100
-80
-60
-40
-20
0
No Terrain Terrain
PE Prediction: 0.2 Hz
• PE prediction at 0.2 Hz through climatology (HWM/MSIS)• Horizontal wind perturbations introduced by gravity waves model• Terrain results in more distinct scattering paths
500 550 600 650 700-1
-0.5
0
0.5
1
Absolute Travel Time (sec)
Am
plit
ud
e (
Pa
)
500 550 600 650 700-1
-0.5
0
0.5
1
Absolute Travel Time (sec)
Am
plit
ud
e (
Pa
)
Waveform Comparison
• Bandpass filter 1-3 Hz
• Terrain introduces more spreading in waveform arrival -4
-2
0
2
4
Am
plit
ud
e (
Pa
)
No Terrain
Terrain
Data
Interface with Atmospheric Specifications
• How terrain/atmospheric boundary condition is handled may have significant effect on prediction performance
– Masking of atmospheric profiles near terrain
– Mesoscale atmospheric effects over regions such as mountains
Terrain
Flat Earth Assumption
Terrain Blocking Assumption
EffectiveSoundspeed
Terrain matching Assumption
Conclusions and Future Research
• Preliminary Conclusions– Terrain appears to disperse down-range waveforms in cases where
significant terrain features are in proximity of source
• Issues for further research– Comparison of terrain masking, conformal mapping, and linear shift
mapping approaches – Sensitivity to ground impedance specification and its variability over
typical propagation paths– Required Topographic database resolution