F
Infrasonics

This appendix provides some additional technical details about infrasonics theory and practice, along with identifying some multiuse applications of the International Monitoring System (IMS) infrasonic data.

INTRODUCTION

Low-frequency acoustic signals in the Earth's atmosphere are traditionally termed infrasound when they have frequencies below 10 Hz. Although subaudible, they do propagate as regular acoustic signals. However, because of the low frequency, there is little physical absorption of signal energy. There is little excess loss beyond geometric spreading, which is always present.

Infrasonic systems must typically detect pressure changes levels one millionth of the total atmospheric pressure in the midst of pressure changes from a variety of sources. These ''contaminating" sources include high winds, turbulence, eddies, thermal plumes, atmospheric gravity waves, and Earth vibrations. For example, normal wind speeds from 5 to 15 m/s can generate quasi-static dynamic pressures from two to three orders of magnitude greater than most infrasonic signals.

Atmospheric acoustic or gravity waves are lower-frequency infrasonic waves indicating that dual processes can influence atmospheric propagation. For normal sound waves the restoring force controlling propagation is the compressibility of the medium. At longer periods, gravitational restoring forces start to become important. Hence, the term acoustic or gravity wave is sometimes used for lower-frequency infrasound. Figure F.1 indicates the approximate frequency range at which this transition takes place.

Lamb waves are atmospheric surface waves that travel along the Earth's surface at the speed of sound in air and can be compared to Stoneley waves in seismology. They represent solutions to the atmospheric equations that have a totally horizontal wave vector, attenuate exponentially with altitude and with depth into the Earth, and attenuate slowly with range (Lamb, 1911; Pierce and Posey, 1971). In the earlier monitoring period, with rather large source yields, Lamb waves were quite distinct with a long-period cycle arriving before the main acoustic signal (see Figure F.2). For CTBT monitoring the source size of interest is smaller, around 1 kiloton (kt). Based on the analysis of Pierce and Posey (1971), Lamb waves for smaller sources may not be as robust as they are



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 125
Research Required to Support Comprehensive Nuclear Test Ban Treaty Monitoring F Infrasonics This appendix provides some additional technical details about infrasonics theory and practice, along with identifying some multiuse applications of the International Monitoring System (IMS) infrasonic data. INTRODUCTION Low-frequency acoustic signals in the Earth's atmosphere are traditionally termed infrasound when they have frequencies below 10 Hz. Although subaudible, they do propagate as regular acoustic signals. However, because of the low frequency, there is little physical absorption of signal energy. There is little excess loss beyond geometric spreading, which is always present. Infrasonic systems must typically detect pressure changes levels one millionth of the total atmospheric pressure in the midst of pressure changes from a variety of sources. These ''contaminating" sources include high winds, turbulence, eddies, thermal plumes, atmospheric gravity waves, and Earth vibrations. For example, normal wind speeds from 5 to 15 m/s can generate quasi-static dynamic pressures from two to three orders of magnitude greater than most infrasonic signals. Atmospheric acoustic or gravity waves are lower-frequency infrasonic waves indicating that dual processes can influence atmospheric propagation. For normal sound waves the restoring force controlling propagation is the compressibility of the medium. At longer periods, gravitational restoring forces start to become important. Hence, the term acoustic or gravity wave is sometimes used for lower-frequency infrasound. Figure F.1 indicates the approximate frequency range at which this transition takes place. Lamb waves are atmospheric surface waves that travel along the Earth's surface at the speed of sound in air and can be compared to Stoneley waves in seismology. They represent solutions to the atmospheric equations that have a totally horizontal wave vector, attenuate exponentially with altitude and with depth into the Earth, and attenuate slowly with range (Lamb, 1911; Pierce and Posey, 1971). In the earlier monitoring period, with rather large source yields, Lamb waves were quite distinct with a long-period cycle arriving before the main acoustic signal (see Figure F.2). For CTBT monitoring the source size of interest is smaller, around 1 kiloton (kt). Based on the analysis of Pierce and Posey (1971), Lamb waves for smaller sources may not be as robust as they are

OCR for page 125
Research Required to Support Comprehensive Nuclear Test Ban Treaty Monitoring FIGURE F.1 Frequency ranges and signal pressures of infrasonic and acoustic or gravity waves. FIGURE F.2 Sample infrasound record of an atmospheric explosion showing a low-frequency Lamb wave arriving before the main acoustic signals. The timescale is indicated in minutes.

OCR for page 125
Research Required to Support Comprehensive Nuclear Test Ban Treaty Monitoring for larger sources. The period will approach that of the main acoustic signal and the amplitudes will be less than that of the main acoustic signal. In addition, turbulence and viscosity in the boundary layer can limit their propagation. SENSOR DESIGN The unwanted fluctuations can largely be eliminated by sampling over an area and not merely at a point, thus averaging out disturbances with small length scales. Since these "filters" use the spatial distribution to help remove unwanted quasi-static pressure noise (also called pseudo-sound, because at a point it mimics infrasound), they are known as spatial filters, infrasonic filters, or infrasonic noise reducers. Figure F.1 summarizes a number of the concepts introduced above and also indicates the approximate frequency ranges for some phenomena producing infrasonic and local pressure signals. It is clear that geophysical signals can appear in the same passband as explosive signals. Figure F.3 is an overview of the wavelengths of various signal types as a function of period; it emphasizes the differences in wavelengths of different sources of pressure fluctuations, with Rayleigh wave-coupled pressure signals having large wavelengths and pressure signals associated with winds and turbulence having small scale sizes. The dimensions of some spatial filters are indicated on the ordinate. (Figures F.1 and F.3 are intended for summary overviews and treat infrasonic monitoring factors in a simplified way.) Past efforts at creating a single sensor capable of covering a large area (e.g., in the form of large pancake-like devices) have not proven practical nor have large numbers of individual sensors. Pneumatic sampling has been effective. In this approach, pressure signals enter ports or porous openings, pass down pipes or tubes, and are then summed together at the sensor analogous to the summing junction of an operational amplifier. A great range of geometries has been tested, the most widely used types being long pipes with periodic sensing ports at intervals, covering a 300 m distance. More recently, using pipes or porous garden hose, 8 or 16 m radial tubes from a central sensor have allowed cost-effective areal sampling. Such filters greatly reduce unwanted noise, but there is still a need to distinguish true infrasonic signals from the remaining local noise, which does not propagate at acoustic frequencies. Beamsteering processing helps to accomplish this in an effective way. COMPREHENSIVE NUCLEAR TEST BAN TREATY INFRASOUND MONITORING If a source produces infrasonic energy, there is a good chance that energy can be detected by remote sensors as is the case for hydrophones in the marine environment. Of particular interest are atmospheric explosions from a fraction of a kiloton to a few kilotons in size. The low-frequency components of the near-field blast wave, as modified by propagation, become the long-range propagating infrasound signal. Experience shows that for 1 kt explosions, detection ranges can be between 2000 and 5000 km depending on noise level and time of year. Measurements are made with arrays of sensors, usually four or more, so that traditional beam-forming techniques can be applied to determine the bearing to observed signals. Thus, an important aspect of detection is the use of correlation among sensor elements in the array. Single-sensor measurements are generally not useful because one cannot tell if increased power is from local noise or from genuine distant sources. Also, the use of power detections alone is not as beneficial as correlation due to the nature of low-level winds, which provide much of the background noise. Bearing estimates are normal outputs of infrasonic array processing and are used to aid in source location. Of course, bearing accuracy is dependent on array size, signal-to-noise ratio, and frequency. In CTBT applications, sources of interest must be detected by at least two arrays. Single-station events generally will not be passed on for further examination, and this serves as a filter for uninteresting events. With two stations and the same event, the two bearings should intersect and provide a preliminary location. Noise at an array can come from ground-level winds for which noise-reducing hoses are effective, although for the CTBT bandpass, work needs to be done to define the optimum noise-reducing configuration. There can also be "noise" from

OCR for page 125
Research Required to Support Comprehensive Nuclear Test Ban Treaty Monitoring FIGURE F.3 Wavelengths of various types of infrasonic signals. In this and Figure F.1, there is considerable overlap between the properties of explosions and geophysical phenomena. background sources not of interest in particular types of measurements. One particular background source, especially for southern ocean island stations, is microbaroms with a frequency peak around 0.2 Hz. This part of the spectrum is of interest in explosion monitoring because it is close to the frequency of the main acoustic arrival for 1 kt explosions. The adopted pressure range curve that the United States presented in Geneva was based on nuclear detections from 0.2 to 112 kt and conventional chemical explosions of less than 1 to 6 kt free air equivalent. These data were given as zero to peak amplitude (scaled by W0.5) as a function of distance (where W is yield in kilotons). Although the influence of middle-atmosphere winds on observed amplitude is well documented out to ˜2000 km, the data above, in raw form, have a standard deviation of about a factor of 2. This has led some to question whether the wind effects survive to 5000 to 10,000 km. This question requires some additional consideration. SOUND SOURCES Natural Sources There are several natural sources of infrasound that can complicate CTBT monitoring in two ways. First, a background of infrasonic signals from nonexplosive sources can mask the existence of an explosive signal. Second, natural signals similar to explosion signals could cause false alarms or mask the nuclear explosion signal. Thus, there will be great value in documenting the

OCR for page 125
Research Required to Support Comprehensive Nuclear Test Ban Treaty Monitoring characteristics of various signal types and using this information in the development and application of discrimination algorithms. Table 3.1 indicates a number of natural source types observed by researchers. In the future, additional human-related infrasonic sources may evolve from industrial or transportation methods using high energies. Further research on the range of background infrasonic sources can have significant payoffs in terms of expanded discrimination methodologies. Explosive Sources Explosions are impulsive releases of energy within a relatively small source volume. Although the close-in signals from explosions are shock waves, they do contain low-frequency components that can travel to large distances with measurable signal levels. Larger atmospheric nuclear explosions were easily heard at ranges in excess of 10,000 km. Thus, large explosions provide signals well suited to study long-range, low-frequency acoustic propagation in the Earth's atmosphere. Both nonnuclear events and nuclear explosions are of interest to CTBT monitoring. Nonnuclear sources include explosive volcanic eruptions and bright meteors known as bolides. Infrasonic data from the Mount St. Helens eruption in 1981 can be found in Donn and Balachandran (1981). The volcanic sources may have associated seismicity, which aids in discrimination. Bolides can deposit all of their energy in the atmosphere or survive entry and impact to the ground. In the former case there would not necessarily be another IMS technology detection (e.g., seismic or hydroacoustic), making this an important potential source of false alarms. (There could be ground observer reports or information from satellite sensors.) Current estimates of meteor influx rates give one 15 kt event per year (ReVelle, 1996). Human-made sources include mining activity, industrial accidents, and sonic booms. Mining operations can use multimillion-pound charges to move large amounts of rock. Many of these events are fired over a finite interval with multiple firing points. These factors reduce the coupling to the atmosphere compared to the same amount of explosive fired as a single source, giving lower amplitudes at a given range. These events do represent human-made sources of great concern because of their size. The CTBT includes guidelines for notification of mining events above a certain size so that they can be identified easily. Detection of an infrasound signal can confirm the fact that a surface explosion took place, even if no preannouncement occurs. This may be a confidence-building measure. However, seismic and infrasound signals from the event must be compared to address the possibility that signals from a conventional surface explosion have been used to mask a buried nuclear explosion. In cases where seismic and infrasound stations are colocated, the infrasound stations may be useful in identifying signals recorded by seismic instruments that arise because the passage of the infrasound signal causes the Earth's surface to move. Identification of such arrivals can remove signal detections from the data base that must be addressed by the seismic association process and thus increase the effectiveness of that process. SIGNAL PROPAGATION The Earth's atmosphere is dynamic, with regions where the wind speed is a significant fraction of the sound speed (the sound speed at the surface is about 340 m/s). This means that the effects of winds must be considered for atmospheric propagation. This is not the case for seismic propagation where the medium is static. The winds of interest for propagation are in the middle and upper atmosphere at about 50 km and 100–120 km altitude. In the middle atmosphere the wind speed can be up to 70 m/s, and in the upper region it can be greater than 100 m/s. In the middle region, the winds are seasonal, blowing east in the winter and west in the summer (in the Northern Hemisphere). During midsummer and midwinter, the magnitude of the zonal component is much larger than the meridional component. The winds at 50 km couple to the ambient sound speed to form a wind duct at that altitude. For sources on or near the ground, energy moving east in the winter (with wind) can be totally refracted around 50 km altitude and be directed back to the Earth's surface, where, for infrasonic frequencies, it is reflected back into the atmosphere. Total refractions can begin where the

OCR for page 125
Research Required to Support Comprehensive Nuclear Test Ban Treaty Monitoring sound speed plus the wind speed in the direction of propagation exceeds the sound speed on the ground. Thus, greater wind speeds can form ducts that turn more energy. The process of multiple refraction and subsequent reflection of energy from the Earth's surface is referred to as multi-hop propagation, and the reflections are called bounces. The regions between the first few bounce points are called zones of silence because in a ray picture, no rays—and therefore no sound—falls in them. Ray acoustics is a high-frequency form of the acoustic equations and ignores diffraction effects that can direct sound into the zones of silence. In the opposite direction, against the wind, the only refractions occur at 100 km or higher where the thermal structure alone or with wind refracts the energy. Because of wind dynamics in atmospheric propagation (compared to seismic propagation) the concept of specific phase arrivals is not as useful. Four types of arrivals have been observed and provide a useful distinction. Surface or Lamb (L) waves travel along the Earth's surface and have average group velocities of 330 to 340 m/s. Tropospheric arrivals (T) travel between the surface and troposphere and have travel speeds of 320 to 330 m/s. Stratospheric arrivals (S) travel between the surface and 40 to 60 km and have travel speeds of 280 to 310 m/s. Ionospheric arrivals (I) travel between the surface and 80 to 110 km and have speeds of 220 to 270 m/s. These values are approximate heights of total refraction as given in a ray picture and are reflected in the range of average travel speeds. With wind propagation, as described above, larger measured signals from a source can be observed downwind than would be measured in the opposite direction, against the wind. This is demonstrated well by Reed (1969) using atmospheric nuclear explosions as sources for infrasonic measurements at several stations around the Nevada Test Site. In addition, average travel speeds are affected by the region through which energy travels. Dual application of infrasound data for other research goals should be encouraged without distracting from the primary mission of CTBT monitoring. If done wisely, there is potential to enhance the role of the IMS to the benefit of the scientific and technical community. The worldwide, extensive, integrated data sets provided by global infrasound monitoring will constitute a unique resource for monitoring the global environment.