SnT2013, Vienna, Austria, 17-21 June 2013 Air-blast data for Sayarim calibration explosions facilitate new method

of source identification and TNT yield estimation

Yefim Gitterman

CONCLUSIONS

1) a new simple and cost-effective method was developed for yield estimation of chemical (calibration) explosions based on a novel scaling relationship for the air-blast Secondary Shock delay;

2) Spectral analysis of teleseismic P-waves from N. Korea nuclear tests at ISN stations revealed spectral minima, interpreted as P+pP interference, that correspond to both cases: vertical borehole of the depth ~2 km, or

horizontal shaft in the mountain of the height ~2 km; the result should be verified by other data and methods.

Acknowledgements The research was supported by the Israel Ministry of Immigrant Absorption.

Evaluation of source depth for N. Korea nuclear tests from ISN teleseismic data

Two underground nuclear explosions conducted by North Korea in 2009 and 2013 were recorded by several stations of Israel Seismic Network. Pronounced minima (spectral nulls) at 1.25 Hz were

revealed in the spectra of teleseismic P-waves. For a ground-truth explosion with a shallow source depth (relatively to an earthquake), this phenomenon can be interpreted in terms of the

interference between the down-going P-wave energy and the pP phase reflected from the Earth's surface. Based on the null frequency dependency on the near-surface acoustic velocity and the

source depth, the depth of the both N. Korea tests was estimated as ~2 km, different from the value ~1 km informed by USGS. Abstract Large on-surface explosions were conducted by the Geophysical Institute of Israel at Sayarim: 82 tons of strong HE explosives in August 2009, and 10&100 tons of ANFO explosives in January 2011 (initiated and supported by the

CTBTO). The main goal was to provide strong controlled sources in different wind conditions, for calibration of IMS infrasound stations. High-pressure gauges were deployed at 100-600 m to record air-blast properties and provide

reliable yield estimation. The rarely reported Secondary Shock (SS) phenomenon was clearly observed at the gauges, and numerous seismic and acoustic sensors. Empirical relationships for peak pressure, impulse, and SS time

delay were developed and analyzed. The parameters, scaled by the cubic root of estimated TNT equivalent charges, were found uniform for all explosions, except of SS delays, clearly separated for 2009 and 2011 shots, thus

demonstrating clearly dependence on the type of explosives with different detonation velocity. Additionally air-blast records from non-Sayarim shots, were used to extend the charge and distance range for the SS delay relationship, and

showed consistency with Sayarim data. Obtained results evidence that measured SS delays can provide important information about an explosion source character, and can be used as a new simple cost-effective yield estimator.

Classic method of yield estimation from air-blast basic parameters

Accurate and reliable TNT yield is an important Ground Truth (GT) parameter of a calibration explosion.

Free-field high-pressure records and basic air-blast parameters (Pm, I+, +) were used for accurate yield estimation.

Preparation of the infrastructure at the SMR: leveling and cleaning

the site, trench for pressure gauge cables, recorder concrete bunker

Sample air-blast

pressure record

and calculated

impulse

The closest gauge G1 before and after

the 2009 explosion

This high-cost and time-consuming method needs 2-3 weeks of infrastructure preparation, gauge

calibration, and protection of expensive equipment from air-blast impact at close distances

New method of yield estimation from air-blast Secondary Shock delay

Based on these SMR observations of high-pressure gauges, acoustic sensors,

accelerometers and seismometers, and using the charge cubic root scaling law a novel

empirical relationship was developed for the scaled delay Dt versus the scaled

distance R (for the TNT equivalent charge W):

Δt = tSS – tMS , Dt = Δt/W1/3 (s/kg1/3), R = r/W1/3 (m/kg1/3) ANFO

Linear RMS fit regression curve was obtained for two Sayarim surface

hemispherical ANFO explosions, for the range 0.1-37 km (56 points):

Dt = 0.0057565log(R) + 0.0032

ANFO

IMI-TNT

A good agreement is found of non-Sayarim ANFO shots with Sayarim ANFO fit curve.

SS delay dependence on explosives (energy, VOD) is revealed

The data for Sayarim 82-ton shot (2009) with IMI explosives, are clearly separated

from the ANFO data, showing also a linear relationship, but significantly lower (with

much smaller SS delays).

The possible reason: IMI explosives are stronger than ANFO, and have

a higher VOD (7.5 km/s vs 4 km/s).

Secondary Shock delay as a new yield estimator

Observations of on-surface Sayarim explosions and previous experiments demonstrate a special character

of the SS delay, as a reliable, stable air-blast feature, different from other basic parameters (Pm, +, I+):

• easily measured, by any simple low-cost acoustic or seismic sensor, not calibrated; even video;

• a sensor can be deployed at remote locations, in the low over-pressure range, and should not be

protected from the blast impact (like the expensive high-pressure gauge system);

• it is a differential parameter, and not necessary to know detonation time, it is critical in a blast accident;

• on the same reason – doesn’t depend on atmospheric conditions - speed and direction of the wind;

• depends on the explosives VOD, thus serving in some cases as a blast source indicator.

The obtained results evidence that this air-blast parameter - SS delay – can be used

as a new yield estimator, based on the developed scaled relationship – for surface

chemical explosions (ANFO, TNT, gas, etc.).

Secondary Shock delay and nuclear explosions

Nuclear explosions do not appear

to produce secondary shocks.

It could be explained by a different

source phenomenology, because a

nuclear test provides an

instantaneous and point-like source

of energy release.

Therefore theis technique could not

be used for the yield estimation.

Nevertheless, some hypothetic

cases can be imagined, when the

second shock can be used for

identification of the source:

nuclear or chemical.

The source depth is an important parameter of a nuclear explosion, conducted

discreetly, exhibiting a violation of the CTBT. Its accurate estimation contributes

to understanding of test design and technological features.

Usual location procedures based on regional and teleseismic records show

large errors for shallow tests.

Some nuclear explosions demonstrate at short-period teleseismic P-wave

records pronounced spectral minima (nulls) near 1-2 Hz resulting from strong

destructive interference between down-going P-wave energy and the pP-wave

reflected from the Earth’s surface.

P+pP interference

The interference produces a spectral modulation

minimum at the frequency

f1 = Vp/(2h)

where Vp(m/s) is the compressional (P-wave)

velocity of the medium above the source and

h (m) is the depth

Spectral null data for previous nuclear explosions

Pronounced spectral minima near 1 Hz

were found at teleseismic records of

Nevada tests (Kulhanek, 1971), 1.5-1.8

Hz for Semipalatinsk tests (Kulhanek,

1973), and ~1.7 Hz for a Pakistan test

(Gitterman et al., 2002).

From paper:

Gitterman, Pinsky, Hofstetter, 2002.

Signal Processing for Indian and

Pakistan Nuclear Tests Recorded at

IMS Stations Located in Israel.

Pure Appl. Geophys., Vol. 159,

no. 4, pp. 779-801.

Seismograms (vert.) and P-wave spectra at ISN stations

from the first Pakistan nuclear test on 28 May 1998

North Korea nuclear tests: observations at Israel Seismic Network (ISN)

The 2006 test was too

small and P-waves were

not observed at ISN

stations, but two larger

explosions in 2009 and

2013 were recorded

Seismograms of the 2nd

North Korea nuclear

explosion in 2009

recorded at ISN stations

(vertical), band-pass

filtered 1-3 Hz

P-waves were observed at some stations after a narrow BP filtering; at some stations signals were not revealed

2009 test: f1~1.2-1.3 Hz 2013 test: f1~1.2-1.3 Hz vertical seismograms are BP filtered 0.7–2 Hz, teleseismic P-waves are aligned

Clear semblance in spectral minima at ~1.2-1.3 Hz

was found for 2009 and 2013 tests, supposedly due to

P+pP interference

Acoustic (P-wave) velocity for granites of the test site is

Vp~5.1 km/s (Bonner et al., 2008, BSSA, V.98, No.5)

Then the depth of both tests is estimated roughly as

h = Vp/(2f1) ~ 2 km

Radionuclides (krypton-85 and xenon-133) were found in atmosphere after the 2006 test, but not

revealed after much stronger shots in 2009 and 2013, thus indicating possible much deeper sources.

For nuclear explosions of this size (2-10 kT) a source depth of less 1 km (~0.5-0.8 km) is sufficient to

provide full containment. A larger depth can be suspected in order to prevent exit of radioactive gases

to the atmosphere that can be detected by IMS stations and provide sensitive information about design

of a clandestine explosion.

Most of underground nuclear explosions were conducted at depths less than 1 km. However, there were

a number of tests in the USSR and USA (23) at the depths 1.4-2 km, and even ~2.5-2.8 km.

The waveform similarity for 2009 and 2013 tests

supposedly indicates the same source depth

The GII depth estimation is based only

on one method and several closely

placed teleseismic stations and should

be verified by other data and methods

North Korea nuclear tests – supposed horizontal shaft case

Satellite imagery of the test site showing horizontal

tunnels in the 2,200 m tall Mt. Mantap

(YONHAP news agency, 2013/02/04 17:46 KST)

Hypothetical reflection of pP wave from the mountain surface on

the height h relatively to the source can interfere with the straight

P-wave and produce the same spectral nulls effect.

Some optimal directions for the reflected pP-wave can be sugested.

Seismology Division, Geophysical Institute of Israel

yefimg@gii.co.il

Calibration surface explosions at Sayarim Military Range

conducted by GII in collaboration with IDF, supported by US Army SMDC and PTS CTBTO (2011)

Different explosives were used

explosives

strong cast IMI bulk ANFO

VOD, m/s

7500 ~4000

density, g/cm3

1.6 0.8

TNT equivalent, tons

96.0 76.8

IMI ANFO

26 Aug. 2009, 82 tons 26 Jan. 2011, 102 tons

Observations of air-blast Secondary Shocks

In all Sayarim surface explosions, distinct air-blast secondary shocks (SS) were observed at high-pressure gauges

(the known effect), and also at acoustic and seismic sensors and even at video-records (new interpretation):

Secondary

shock

delay

Main

shock

Δt = 0.35 s

at near-source distances (0.1-0.6 km) at local distances (up to 37 km)

82-ton (2009) by a high-pressure gauge (0.4 km) 102-ton ANFO (2011) by accelerometer (0.3 km) 102-ton ANFO (2011) by 3C seismo-acoustic station (WGC, USA) (2 km)

102-ton ANFO (2011) Visualization of audio-channel from a home video-camera at 9 km

A new special air-blast

parameter is proposed –

the secondary shock delay

The delay was found increased for

larger charges & distances

Secondary Shocks for non-Sayarim shots We extended charge and distance ranges, including WSMR extra-large shots Distant Image (1991)&Minor Uncle (1993) > 2000 tons ANFO, recorded at 28-60 km

Distant Image

Minor Uncle Distant Image, 20 June 1991, 2,210 tons ANFO, station Rimfire

(pressure sensor and 2 seismic channels RADIAL and VERTICAL)

TNT equivalent:

WTNT = 0.82W=1,812,200kg

Distance r = 28,000m,

SS delay: dt = 2.024 sec

Sayarim experiment layout

SS delay data for Sayarim ANFO shots

SS delay data for all shots and different explosives