ABSTRACTS
 


Session 1 : IMS AND IDC

Infrasound processing at the IDC and its its near-term development
Petr FIRBAS et al

Technical aspects of automatic and interactive processing of infrasound data at the IDC
Nicolas BRACHET et al

Operation and maintenance of IMS infrasound stations
Timothy John DALY

Establishment of CTBTO IMS infrasound network - a status report
Thomas HOFFMANN

Progress of IS 15 and IS 16 project and infrasound research at NCI
Songbei LI

Status of International Monitoring System and the perspective for future installations
Gerardo SUAREZ


Session 2 : INSTRUMENTATION

Study of high-wind noise reducing system using multiple sensors
Matthieu LE PRADO et al

Infrasound measuring chain - Calibration device
Damien PONCEAU et al


Session 3 : NETWORK DETECTION CAPABILITY - CALIBRATION ISSUES

Ambient infrasound noise estimates and meteorological conditions
Roger BOWMAN et al

Metadata for global geophysical monitoring networks
Roger BOWMAN

Detection of distant atmospheric explosions : implications for the design of IMS infrasound array stations
Douglas R. CHRISTIE et al

A summary of signal detections at the I 56 US and I 57 US arrays
Michael HEDLIN et al


Session 4 : DATA PROCESSING

A quantitative evaluation of PMCC's detection capability
Yves CANSI et al

Infrasound source location using the neighbourhood algorithm
David BROWN

An analysis of seismo-acoustic data of seismic and infrasound array stations of Songino - Ulaanbaatar
Bayarsaikhan CHIMEDTSEREN

Towards automatic infrasonic source locations with site-specific, temporally varying travel time curves
Milton GARCES et al

Classification of infrasound events using optimized feature vectors and RBF neural networks
Fredric M. HAM

One year of data processing of IS 21 station (Marquesas Islands, French Polynesia)
Olivier HYVERNAUD et al

Automatic blast detection using wavelet neural network
Lucas MOREIRA et al

Grid methods applied to location of infrasound sources
Dominique REYMOND

Near-field infrasound source localization
Curt SZUBERLA et al


Session 5 : MODELLING

Infrasound associated with the large 2004 Sumatra earthquake and tsunami
Lars CERANNA et al

On the effect of atmospheric fine structure on the azimuth and grazing angles of infrasonic signals at long distances from explosions
Sergey KULICHKOV et al

On physical simulation of long-range infrasonic propagation in the atmosphere
Sergey KULICHKOV et al

Broadband propagation modelling and scattering
David NORRIS


Session 6 : SOURCES

Acoustic estimation of breaking ocean wave height and period
David FEE et al

The inaudible noise of wind turbine
Lars CERANNA et al

Atmospheric processes controlling the number of infrasound detections
Laslo EVERS

Surf infrasound : the movie
Milton GARCES et al

Infrasonic remote sensing of breaking ocean waves in the surf zone
Robert GIBSON et al

An analysis of infrasound signals from 11th December 2004 meteorite at Lanzhou in China
Hee-il LEE et al

Infrasound generated by the M7.8 Northern Chile earthquake of 13 June 2005
Alexis LE PICHON et al

Infrasound produced by tsunami-generating earthquakes
Sara Mc NAMARA et al

Seismo-acoustic, wave acoustic observations and synergy
Dominique REYMOND et al

Seismo-acoustic analysis of quarry blasts in the I 31 KZ region and the influence of atmospheric conditions on infrasonic wave propagation
Alexandr SMIRNOV et al

Infrasound observations in Alaska and Antarctica
Curt SZUBERLA et al

Thunder, waterfalls and volcanic eruptions : a menagerie of infrasound in the Ecuadorian jungle
Hugo A. YEPES et al


Session 7 : VOLCANO SESSION

Multi-technology observations of the Manam volcano eruption in January 2005 - Potential for an enhanced volcanic ash warning system (Poster)
David BROWN et al

Detection of infrasound from the eruption of Manam volcano on January 27, 2005
Paola CAMPUS et al

The CTBT IMS infrasound network and monitoring of volcanoes
Paola CAMPUS

Diurnal variations in infrasonic Tremor signals from Kilauea volcano, Hawaii
David FEE et al

The physics of volcano-acoustic sources
Miton GARCES

Seismo-acoustic study of Mount Saint Helen
Michael HEDLIN

Madagascar infrasound station I 33 S and the observed signals
Gérard RAMBOLAMANANA et al

Monitoring of Indonesian volcanoes with infrasound : preliminary results
Bruno FEIGNIER et al


Session 1 : IMS AND IDC

Infrasound processing at the IDC and its its near-term development
Petr FIRBAS and Nicolas BRACHET
CTBTO/PTS, IDC Division      Vienna - AUSTRIA

The presentation will provide information about the current status of development of the infrasound data processing at the Vienna International Data Centre (IDC) of the Comprehensive Nuclear-Test-Ban Treaty (CTBT) Preparatory Commission (PrepCom).
The IDC is currently running the automatic processing of data from 31 infrasound stations on the testbed. This set of stations represents more than one half of the total network of the International Monitoring System (IMS). In 2005 substantial efforts were focused on the enhancements and testing of the automatic processing, in particular its parts for the signal detecting and event building.
The first priority in 2005, however, has been given to the new tools for interactive processing of the infrasound data, with the objective of reinstating the infrasound data processing in the IDC operations in 2006 and to be fully integrated together with seismic and hydroacoustic technologies. The first version of the interactive software (Geotool PMCC) has been developed and tested. A dedicated group of analysts is being trained for the interactive review of infrasound signals and also provides feedback on the newly developed versions of the interactive tools. The re-introduction of the full infrasound processing into the Operations pipeline (including building of events using all waveform technologies and interactive processing) will be done as smoothly as possible in order not to interfere with the ongoing production of the daily reviewed event bulletin.

Technical aspects of automatic and interactive processing of infrasound data at the IDC
Nicolas BRACHET and Petr FIRBAS
CTBTO/PTS, IDC Division      Vienna - AUSTRIA

The presentation will provide some technical information regarding the latest developments of infrasound data processing at the International Data Centre (IDC) in Vienna. This covers various aspects of the automatic and interactive system : signal detection, detection categorization, event building and interactive review of the results. Practical examples will be shown to illustrate the work accomplished to prepare the reinstatement of infrasound data processing in IDC operations in 2006.

Operation and maintenance of IMS infrasound stations
Timothy John DALY
CTBTO/PTS, International Monitoring System Division     Vienna - AUSTRIA

This presentation will discuss some of the operational aspects of the IMS Infrasound Monitoring Unit and the PTS Operations Centre. With more than half of the IMS Infrasound stations certified, focus is being placed on operation and maintenance of the stations. A brief description of the structure of the IMS Infrasound Monitoring Unit will be provided. The PTS Operations Centre will be presented including its structure, operating procedures, and some of its tools. The maintenance concept of the IMS Infrasound Monitoring Network will be discussed. Finally, some of the failures and successes achieved in the operation and maintenance of the IMS Infrasound Stations will be presented.

Establishment of CTBTO IMS infrasound network - a status report
Thomas HOFFMANN
CTBTO/PTS, International Monitoring System Division      Vienna - AUSTRIA

Steady progress has been made in the establishment of the IMS infrasound monitoring network. To date, 90 % of the site surveys for the 60 infrasound stations of the network have been completed, 53 % of the stations are transmitting continuous data to Vienna, and 51 % of the stations were certified. With more than one half of the network now certified, the global distribution pattern of infrasound stations begins to form an extended continuous sub-network in the southern hemisphere, where most of Australia, the Pacific Region, South America, the Atlantic Region and Southern Africa are connected through a band of certified infrasound stations. Another continuous sub-network has formed in the North American Region. This presentation will provide an overview of recent progress made in the establishment of the global infrasound network, and also present some of the challenges and difficulties that still lie ahead for the program.

Progress of IS 15 and IS 16 project and infrasound research at NCI
Songbei LI
Signal Processing Department - North China Institute of Computing Technology     Beijing - CHINA

Three parts are included in the presentation.
First part : a briefing about the progress of IS15 & IS16 construction.
Second part : a review of the history of infrasound research at the Signal Processing Department.
Third part : to introduce infrasound research in the latest years. The topic focuses on the development and testing of infrasound monitor equipments and signal processing toolkit. The analysis of some observation data is given.

Status of International Monitoring System and the perspective for future installations
Gerardo SUAREZ
CTBTO/PTS, International Monitoring System Division      Vienna - AUSTRIA

The International Monitoring System (IMS) of the Comprehensive Test-Ban Treaty (CTBT) is composed of 321 stations with four different types of monitoring instruments : infrasound, hydroacoustic, seismic and radionuclide. Additionally, the system is supported by 16 radionuclide laboratories. The implementation of the IMS began in late 1997. To date, about 65 % of the IMS stations in all technologies are installed and an additional 14 % is under construction. The final stage in the construction of a monitoring station is the certification, when it is verified that station meets all the technical specifications specified by the signatories of the Treaty. Approximately 45 % of the stations are certified to date : 60 % of the primary seismic network ; 32 % of the auxiliary seismic network ; 55 % of the hydroacoustic and 45 % of the radionuclide. On the basis of current projections and assuming a budget equivalent to those of the last few years, it is predicted that at the end of 2007 approximately 90 % of the IMS network will be completed and transmitting data to the International Data Centre. The distribution of stations that exists today allows excellent regional coverage in many areas. It is expected that at the end of 2007, all technologies will have a global coverage.


Session 2 : INSTRUMENTATION

Study of high-wind noise reducing system using multiple sensors
Matthieu LE PRADO1, Thomas HOFFMANN2 and Paola CAMPUS2
1 CEA/ DIF / DASE / TMG     Bruyères-le-Châtel - FRANCE
2 CTBTO/PTS, IMS, Infrasound Monitoring Section      Vienna - AUSTRIA


An infrasound network is being installed for the Comprehensive Nuclear Test-Ban Treaty Organisation (CTBTO). In order to achieve a homogenous distribution of IMS infrasound stations over the globe, it is necessary to implement some of them at locations with high winds. These sites are exposed to considerable infrasound noise, which degenerates the detection capabilities for micropressure signals of interest. In order to reduce the wind noise influence to acceptable levels, each microbarometer of an infrasound element is connected to a wind noise reducing pipe array. Nevertheless, the standard pipe arrays are most of the time not enough efficient because of the huge infrasound noise level and the large coherence length of the eddies. Therefore, outstanding wind noise reducing system has to be used to optimise the detection capability.
We present here a theoretical and experimental study of a noise reducer constituted of several sensors that are, according to the need, either summed or individually processed. These noise reducers beneficiate of a large inlet distribution diameter, adapted to low frequency measurements, while improving the high-frequency measurements possibilities realized with each of the sensors. Moreover, the frequency response of the multi sensor noise reducer is not dependent of the inlet distribution span, contrary to the standard pipe arrays. At a first step, we compare the multi-sensor-noise-reducers to the large pipe arrays. As a second step, we show a parametric study of the noise reducer constituted with several sensors.

Infrasound measuring chain - Calibration device
Damien PONCEAU and Matthieu LE PRADO
CEA/ DIF / DASE / TMG    Bruyères-le-Châtel - FRANCE

Infrasound measuring chains installed for the Comprehensive Nuclear Test Ban Treaty Organisation (CTBTO) need to be calibrated periodically. The DASE designed a portable instrument, called "IS Calibrator", in order to perform infrasound sensor transfer function measurement all over the CTBT band [0.02-4Hz].
This "IS Calibrator" is able to generate pressure signals over the appropriated band with a flat and known response in amplitude and phase. This device consists in three components. The first one is a mechanical assembly which converts a voltage into a pressure signal applied to the infrasound sensor with a flat response within the CTBT required 5 % uncertainties in amplitude and phase over its functionning band. Then, a reference microphone is used to measure pressure amplitude at a particular frequence. As a consequence, the "IS Calibrator" response is known and flat all over the CTBT band. The last element is a board processing device used to adjust automatically the pressure to the desired value and to select the signal generated: Available shapes are sine waves, pulses and Maximum Length Sequences (MLS).
The DASE also developed procedures to dynamically calibrate infrasound measuring chains thanks to this device. These procedures were routinely used to calibrate some sensors under controlled pressure and temperature conditions in order to prove the accuracy of the IS Calibrator. This year, the IS Calibrator performed about 300 transfer function measurements.
Today, the "IS Calibrator" is a portable and accurate device able to calibrate infrasound sensors and measurement chains in the CTBT band.


Session 3 : NETWORK DETECTION CAPABILITY - CALIBRATION ISSUES

Ambient infrasound noise and meteorological conditions
Roger BOWMAN and Gordon SHIELDS
S A I C (Science Applications International Corporation)     San Diego - USA

We characterize the ambient infrasound noise environment in the frequency band of 0.03 to 7 Hz for all operational infrasound stations of the International Monitoring System as well as research stations in the United States. Our first study included 21 stations for which data were available in January 2004 and spanned about one year. Our current study includes seven additional stations that were available by December 2004 and spans two years. The ambient noise at infrasound stations is highly variable by season, time of day, station, and site type. Noise spectra for an individual station may vary by more than four orders of magnitude at any given frequency. We define preliminary infrasound noise models that can be used as baselines for evaluating ambient noise at current and new infrasound stations. We also define a simple metric for ranking the performance of infrasound stations in terms of the percentage of time that the station's noise is below the 25th percentile of network noise.

Metadata for global geophysical monitoring networks
Roger BOWMAN
S A I C (Science Applications International Corporation)     San Diego - USA

Quantitative analysis of data from global geophysical monitoring networks is dependent on accurate and complete metadata. Basic, parametric metadata, such as sensor locations, sampling rate, and instrument response, are required to make any use whatsoever of the time-series data. Other types of metadata, such as vegetation and topography, are required for more sophisticated interpretation of the signals and ambient noise recorded by sensors. Management of metadata requires a robust process for acquiring, compiling, and validating metadata ; for disseminating metadata to users ; for reporting and resolving errors ; and for disseminating updates to users.

Detection of distant atmospheric explosions : implications for the design of IMS infrasound array stations
Douglas R. CHRISTIE, Brian L. N. KENNETT and Chris TARLOWSKI
Research School of Earth Sciences - The Australian National University     Canberra, ACT 0200 - AUSTRALIA

The infrasound component of the International Monitoring System (IMS) is a network of 60 stations distributed as uniformly as possible over the face of the globe. On average, the distance between stations located in continental land mass areas in the Northern Hemisphere is about 2 500 km. In contrast, the distances between stations on opposite sides of the vast open ocean areas in the Southern Hemisphere may exceed 6 000 km. These remote open ocean areas in the Southern Hemisphere are by far the most difficult areas to monitor. It is clearly important to have a good understanding of the properties of infrasonic waves from explosive sources located at distances in the range from about 1 000 to 4 500 km. This investigation is motivated by the observation that there have been very few observations in recent years of infrasound from surface mining explosions located at distances of more than 1 500 km. This surprising result suggests that there may be some fundamental problems with the basic design of IMS infrasound monitoring stations or with currently used infrasonic data processing procedures. Stations in the IMS infrasound monitoring network have been established in a wide variety of configurations. There has been a tendency in recent years to increase the number of array elements and to decrease the overall aperture of the arrays. In this study, we report the results of a detailed investigation of the properties of infrasound from distant explosions recorded at two typical neighbouring IMS infrasound stations in Australia, IS 05 Hobart and IS 07 Warramunga. The results of this study show that a number of factors have a direct influence on detection and location capability for infrasound from distant explosions. These include the decay of the higher frequency components in the infrasonic signal, problems associated with the spatial coherence of signals at higher frequencies, the type of signal detected (i.e., stratospheric or thermospheric), ambient noise characteristics and spatial aliasing of higher frequency signals.
The main conclusion of this investigation is that the optimal detection passband for the monitoring of explosion-generated infrasound at a given IMS station depends primarily on the distance to the source and the direction and intensity of the upper atmospheric wind component along the path between the source and the monitoring station. The optimum passband for the monitoring of surface explosions at regional and near-regional distances will usually be in the range from about 0.5 to 2.0 Hz. The optimum passband for the monitoring of far-regional and distant explosions will often be at lower frequencies in the range from about 0.4 to 1.0 Hz. In some circumstances, detection of distant explosions will be restricted to longer period thermospheric waves and the optimum monitoring passband will be in the range from about 0.03 to 0.1 Hz. The results also suggest that arrays should have at least 8 elements and the overall aperture of IMS infrasound array stations should not be less than 2.0 km.

A summary of signal detections at the I 56 US and I 57 US arrays
Stephen ARROWSMITH and Michael HEDLIN
Laboratory for Atmospheric Acoustics, Institute of Geophysics and Planetary Physics, University of California     San Diego - USA

The Laboratory for Atmospheric Acoustics (L2A) at the University of California, San Diego, operate the I56 US and I57 US IMS infrasound arrays. The data received at L2A undergo routine analysis for signals in two separate frequency bands. Here, we present a summary of the typical detections observed at the two arrays. In addition, we focus on some more significant signals including a bolide that was recorded by both arrays, signals from wind farms and the many thousands of detections we observe each year from surf and microbaroms.


Session 4 : DATA PROCESSING

A quantitative evaluation of PMCC's detection capability
Yves CANSI, Julien VERGOZ and Estelle SCHISSELE-REBEL
CEA/ DIF / DASE / LDG     Bruyères-le-Châtel - FRANCE

The PMCC detector is now routinely used at the French NDC since 1998 and at the IDC since 2004. These operational implementations have allowed the report of various examples of infrasound events detected with a low signal to noise ratio. Nevertheless, up to now, no quantitative evaluation of PMCC detection capability for CTBT infrasound stations has been performed. We present a study of the performances of PMCC in terms of detection capability, propagation parameter measurements and source identification. This study is based on an original approach using "synthetic" signals constructed from real signals embedded in real noise sequences. This approach allows us to precisely control the actual signal to noise ratio, and consequently to precisely determine the detection threshold using a large set of "synthetics". Furthermore, other algorithms dedicated to array data processing, such as the MUSIC or CAPON beamformers, are known to exhibit high resolution performances. The performance evaluation has also been carried out using these algorithms, and a comparison with the PMCC results will be presented.

Infrasound source location using the neighbourhood algorithm
David BROWN
Geospatial and Earth Monitoring Division, Geoscience Australia     Canberra - AUSTRALIA

The Neighbourhood Algorithm of Sambridge et al (M. Sambridge, B. Kennett, pageoph, 2001) for parameter space exploration is used to tackle the problem of estimating spatial and temporal source location of distant infrasonic events. Using an acceptable misfit function, 4-dimensional hypocentral parameter space is searched to determine a solution with acceptable fits to array observations of measured back-azimuth and signal time across each station. Contrary to many other source location procedures, derivatives of travel-time information are not used, which may make for a more stable algorithm. This source location algorithm is the next stage of the INFER processing software for infrasound signal processing.
A two stage procedure is employed. First, the entire globe is discretized using a geodesic tessellation, and a matrix of travel-times from each point to each station is generated. The user is required to provide a travel-time and azimuthal deviation model from each point of the Tessellation to every station for each phase. During algorithm operation the latitude/longitude of each point is used as an index or key into the matrices via a minimal perfect hash algorithm. A suite of keys that could be a potential source location is found using a fast table-look up procedure.
The Second stage operation employs the Neighbourhood Algorithm to hypothesise a more precise spatial and temporal source location using the initial coarse tessellation as a starting point.
A 'conflict resolution' algorithm is used to prevent bogus associations between detections on different stations prior to the NA stage.
The source location procedure is applied to several test cases. Although large memory requirements are a drawback, preliminary results indicate the new algorithm to be fast and reliable, with the conflict resolution module reducing false associations.

An analysis of seismo-acoustic data of seismic and infrasound array stations of Songino - Ulaanbaatar
Bayarsaikhan CHIMEDTSEREN
Research Center of Astronomy and Geophysic of the Mongolian Academy of Sciences     Ulaanbaatar - MONGOLIA

We have compiled and analyzed seismo-acoustic signals from mine blast for 2000 and 2005 in order to determine detection seismo-acoustic signals of explosion by seismic and infrasound stations.
Several large mines in the region routinely generate explosions that are detected seismically and with infrasound. The mines range in distance from 40-500 km from the seismic, infrasound array. The corresponding number of infrasound detections is found to be depending upon the season and local winds. During the winter months, when the direction of the zonal component of the stratospheric wind is from north-west to south-east, a strong stratospheric dust develops IS 34 MN the number of infrasound detections is high.
Also we observed shooting noise in both seismic and infrasound stations from military training area which is our infrasound stations is located.
The focus of this paper will be presentation of IMS seismic and infrasound stations, analysis result of signals of quarry blasts.

Towards automatic infrasonic source locations with site-specific, temporally varying travel time curves
Milton GARCES 1, C. HETZER 1, Sara McNaMARA 1, D. DROB 2 and Nicolas BRACHET 3
1 Infrasound Laboratory, University of Hawaii     Manoa - USA
2 Naval Research Laboratory     Washington DC - USA
3 CTBTO/PTS, IDC, Waveform Development and Software Integration Section, Waveform Development Unit     Vienna - AUSTRIA

Automatic seismic source location algorithms have been available for many years. The development of similar infrasonic source location capabilities has been hampered by the fact that the propagation media - the atmosphere - is heterogeneous, anisotropic, and dynamic at infrasonic temporal (minutes to hours) and spatial (10-1-105 km) scales. The automatic and timely generation of atmospheric travel-time tables is a prerequisite for the development of these capabilities. Software based on the tau-p method of Garces et al. (1998) has been developed that generates sets of travel-time curves at specified locations on demand, based on MSISE and/or G2S climatological models. These curves divide the atmosphere into upper-atmosphere and lower-atmosphere propagation regions, are geographically and temporally specific, and account for variations in propagation characteristics with azimuth. Using these curves, we present a comparative study of multiple-station infrasonic locations for events in the infrasonic reference event database.

Classification of infrasound events using optimized feature vectors and RBF neural networks
Fredric M. HAM
Florida Insitute of Technology     Melbourne - USA

Results are presented for a bank of Radial Basis Function (RBF) neural networks, to discriminate between different infrasonic events. Each module in the bank of RBF networks is responsible for classifying one of several events, and thus, is trained to identify only that particular event. However, each module is also trained to not classify all other events (i.e., negative reinforcement). Output thresholds of each module are set according to specific Receiver Operating Characteristic (ROC) curves. Preprocessing of the infrasound signals is carried out by extracting the optimal set of cepstral coefficients and their associated derivatives that form the feature vectors used to train and test the RBF networks. When the optimal feature vectors are used for training and testing, a significant increase in the classification accuracy can be achieved over the ad hoc method of selecting a sequential set of cepstral coefficients and a subset of their associated derivatives.

One year of data processing of IS 21 station (Marquesas Islands, French Polynesia)
Olivier HYVERNAUD, Johann BAILLEUL and Dominique REYMOND
Laboratoire de Géophysique Papeete     French Polynesia - FRANCE

The new infrasonic mini-array IS 21, (Marquesas Islands) has been working routinely since December 2004. One year of automatic processing and studies of particular events are presented.

Automatic blast detection using wavelet neural network
Lucas P. MOREIRA 1, Francisco A. O. NASCIMENTO 2 and Adson F. da ROCHA 2
1 Seismological Observatory of University of Brasília     Brazília - BRAZIL
2 Electrical Engineering Department of University of Brasília     Brazília - BRAZIL

An automatic blast detection system based on wavelet neural network is presented. The wavelet transform approach to signal feature extraction drastically reduces the neural network complexity, decreasing the processing time and keeping its high detection performance, making possible a real-time implementation. The proposed system was tested with quarry blast signals recorded by the IMS station I 09 BR, with distance around 20 km from the station and weight around 4000 kg, and a good performance was observed.

Grid methods applied to location of infrasound sources
Dominique REYMOND
Laboratoire de Géophysique Papeete     French Polynesia - FRANCE

A method of source location, using an algorithm of grid has been applied to infrasonic events. The evolution and very rapid increase of speed of computational power during these last years, allows data processing to use grid methods with reasonable time calculation. This straightforward method, based on the search of the minimum minimorum residual, computed with L1 or L2 norm; it is shown to be extremely robust, and can work with various type phases (acoustic, seismic and hydro-acoustic). It can be applied to mini arrays networks, even in degraded situations of running with stations failures inside a mini-array. This very basic method could be improved with more realistic velocity models.

Near-field infrasound source localization and discrimination
Curt A. L. SZUBERLA, Kenneth M. ARNOULT and John V. OLSON
Geophysical Institute, University of Alaska     Fairbanks - USA

We present fast methods of discriminating and localizing near-field sources based on time difference of arrival (TDOA) information for an arbitrary array of sensors. Discrimination stems from a statistical measure of the planarity of a signal propagating across an array. Once a signal is deemed to be in the near field, an ensemble of sub-array localizations is calculated analytically from conic sections. The statistical properties of this ensemble lead us to a seed simplex for a numerical optimization procedure. The optimization yields a simplex that bounds a region in coordinate and velocity space corresponding to the TDOA information of the signal of interest. Since TDOA information is generally estimated via cross correlation techniques, the method described constrains neither the functional form, nor the amplitude characteristics of the signal present at the array. The same is true of the noise characteristics. This is a distinct advantage for infrasound array data, as terrain and vegetation often introduce amplitude characteristics that deviate significantly from r-2 behavior and infrasound signals are not often well represented by simple functional forms. The method is shown in application to array data derived from University of Alaska Fairbanks infrasound arrays.


Session 5 : MODELLING

Infrasound associated with the large 2004 Sumatra earthquake and tsunami
Lars CERANNA 1 and Alexis LE PICHON 2
1 Federal Institute for Geosciences and Natural Resources (BGR)     HANNOVER - GERMANY
2 CEA/ DIF / DASE / LDG     Bruyères-le-Châtel - FRANCE

The large M9 Sumatra earthquake from December 26th 2004 has generated strong seismic and acoustic signals which are recorded at the IMS infrasound station I 52 GB in the Diego Garcia atoll. The Progressice MultiChannel Correlation (PMCC) algorithm is performed to analyze both the seismic and the low frequent infrasound signals associated with this event. Focusing on the acoustic signals and applying an inverse location procedure to reconstruct the source location of the individual infrasound wave trains two different types of sources are obvious. One is related to the vibration of land masses on the island of Sumatra near epicenter region and the other one is emitted by the co-seismic generated tsunami reaching the Bay of Bengal. Measurements of infrasound related to large continental earthquakes with respect to the source mechanism are well described in the literature. In contrast infrasound from tsunamis is not clear since a tsunami has only moderate amplitudes in the range a mete, and surf noise generated at coastal reefs seems unlikely due to the very low observed frequencies and its high attenuation.
The aim of this presentation is to demonstrate that the low frequent infrasound reaching I 52 GB ~ 3 hours after origin time and lasting almost ~ 4 hours is caused by the tsunami. This is performed by using different approaches for forward modeling based on the Chebyshev pseudo-spectral method to explain the onsets kinematically related to the tsunami, and to illustrate how a low amplitude wave phenomena can produce a strong pressure impulse that can be recorded over 3 000 km.

On the effect of atmospheric fine structure on the azimuth and grazing angles of infrasonic signals at long distances from explosions
Sergey KULICHKOV 1, Igor CHUNCHUZOV 1, Oleg POPOV 1, Vitali PEREPELKIN 1 and Anatoly BARISHNOKOV 2
1 Oboukhov Institute of Atmospharic Physics Russian academy of sciences     Moscow - RUSSIA
2 Research Institute of Pulse Techniques     Moscow - RUSSIA

The effect of the atmospheric fine structure on the azimuth and grazing angle of infrasonic signals recorded at long distances from surface explosions is studied both theoretically and experimentally. The data on infrasonic signals recorded at a distance of 297 km from surface explosions equivalent to 20-70 t of TNT are analyzed. The experiments were carried out during different seasons. In these experiments, explosions were set off by series (up to 7 explosions in a row). The time intervals between explosions within each series varied from 10 to 20 min. The measurements were taken at the same observation site. Variations in the azimuths and grazing angles of infrasonic signals are revealed for both the experiments carried out within one series and carried out during different seasons. Both the methods of parabolic equation code and normal waves code were used in a theoretical modelling. The data on temperature and wind stratifications were obtained from the rocket sounding of the atmosphere during the time periods closest to those during which the experiments were carried out. The vertical spectrum of atmospheric inhomogeneities was taken in the form of the conventionally used dependence Kz-3 (Kz is the vertical wave number). It is shown that, due to the presence of atmospheric fine-layered inhomogeneities, fluctuations in the phase of infrasonic waves mainly affect the errors in determining the azimuths and grazing angles of infrasonic signals.

On physical simulation of long-range infrasonic propagation in the atmosphere
Sergey KULICHKOV 1, Igor CHUNCHUZOV 1, Oleg POPOV 1, Vitali PEREPELKIN 1 and Anatoly BARISHNOKOV 2
1 Oboukhov Institute of Atmospharic Physics Russian academy of sciences     Moscow - RUSSIA
2 Research Institute of Pulse Techniques     Moscow - RUSSIA

The results of a physical modelling of long-range sound propagation in the atmosphere are given. The property of similarity between the profiles of effective sound speed stratification for the atmospheric boundary layer and for the land-the lower thermosphere thickness was used in the modelling. A detonation generator was used as a source of acoustic pulses. The spectral maximum for the initial acoustic pulse was between 40 and 60 Hz. The maximum amplitude of acoustic pressure reached 120 Pa at a distance of 20 m from the source. The initial pulses were generated by series: during each series, 100 pulses were generated one after another with a one-minute interval. The measurements were taken in different azimuthal directions at distances up to 7 km from the source. Atmospheric stratification was determined from the measurements from a mast and also with the use of an acoustic locator (sodar) and a temperature profile meter up to a height of 600 m. The experimental results show a structure similarity between the signals recorded in the atmospheric boundary layer and in the land-the lower thermosphere thickness.

Broadband propagation modelling and scattering
David NORRIS BBN Technologies     Arlington - USA

Advanced broadband propagation prediction of infrasound includes the application of Parabolic Equation (PE) and Time-domain Parabolic Equation (TDPE) models. In this study, the effects of scattering on PE and TDPE models are evaluated. Specifically, the focus is on : acoustic energy levels within shadows zones, waveform pulse duration, and waveform peak amplitude and rise time. Scattering effects will be accounted for by incorporating a model that realizes horizontal velocity perturbations due to turbulent gravity wave structure. The study will include comparison of TDPE waveform predictions with measurements from infrasonic events.


Session 6 : SOURCES

Acoustic estimation of breaking ocean wave height and period
Jerome AUCUN 1, David FEE 1, Milton GARCES 1, M. MERRIFIELD 1 and Robert GIBSON 2
1 Infrasound Laboratory, University of Hawaii     Manoa - USA
2 BBN Technologies     Arlington - USA

Simultaneous infrasonic, visual, and ocean-bottom pressure sensor observations of large swells on the island of Kauai and small to medium-sized surf on the island of Hawaii yielded a clear relationship between breaking wave height and low-frequency atmospheric sound amplitudes in the 0.5-20 Hz frequency range. These experiments confirmed that infrasound can be generated by barreling waves as well as by waves crashing against rocky shorelines and exposed ledges. The envelope of surf infrasound signals yields breaking wave periods comparable to those derived from nearshore sea surface elevation measurements. The results of these experiments demonstrate that low-frequency sound may be used for real-time estimates of the amplitude, period, and spatial distribution of surf in the littoral zone, with a potential application to the identification of breaking wave types.

The inaudible noise of wind turbine
Lars CERANNA, Gernot HARTMANN and Manfred HENGER
Federal Institute for Geosciences and Natural Resources     Hannover - GERMANY

A temporary field experiment was carried out north of Hannover in summer 2004. It was aimed at the measurement of infrasonic signals from a single wind turbine in the near field up to a distance of two kilometers. Four mobile stations with MB2000 microbarometers were used for this purpose. Results from similar measurements have already been published, however, they are based on microphone data which do not properly reflect the noise conditions in the frequency range below a few Hz. Consequently, the microbarometer measurements can be considered as an extension of the microphone based results to low frequencies. The site of the wind generator was advisedly chosen since there is no other one within a radius of ~ 10 km. Such conditions enabling un-biased acoustic measurements are rather rare in Germany. The power spectral density curves clearly show an increase of the noise at the blade passing harmonics corresponding to both the rotational speed of the blades and the wind speed. These results are compared with theoretical estimations of the RMS pressure amplitudes with respect to the considered type of wind turbine. The findings confirm the theoretical model in the CTBT relevant frequency range below four Hz.
The presentation will also inform on the impact of a single wind turbine and of wind parks to the background noise of an infrasound array using the analytical formula for a set of wind generator configurations. Taking the increasing number and size of these facilities into account this aspect is fundamental for the site selection of an infrasound array or vice versa for the planed installation of a wind park in the vicinity of a pre-existing infrasound array.

Atmospheric processes controlling the number of infrasound detections
Laslo EVERS
Royal Netherlands Meteorological Institute (KNMI)     AE De Bilt - THE NETHERLANDS

Infrasound in the Netherlands is continuously monitored with a wide variety of array configurations. The number of microbarometers per array range from 6 to 16 while the aperture of the arrays varies between 30 and 1 500 meter. In this presentation we'll concentrate on a year of data of the De Bilt Infrasound Array (DBN). Within DBN the microbarometers are placed two by two on the sides of an equilateral triangle with an aperture of 70 meters. Events are triggered by applying a detector based on Fisher statistics in both the frequency and time domain. Taking the size of the array into account, standard processing is done in the frequency range of 1 to 15 hertz. The number of detections per day highly vary throughout the year and range between 0 to 10's of coherent arrivals. The events originated from close by the array, near field infrasound, to distances of 100's of kilometers being in the far field. Parameter estimation, based on beamforming, derives the event characteristics like: arrival time, apparent sound speed, back azimuth and frequency. These parameters are compared with the state of the atmosphere obtained from wind and temperature models up to 65 km height from the European Center of Medium Range Weather Forecasting (ECMWF). Clear infrasonic signatures of the state of the atmosphere are observed in terms of the number of infrasound detections and its characteristics. In this presentation special attention will be paid to the infrasonic signature of a sudden stratospheric warming based on the lack and increase in the number of detections.

Surf infrasound : the movie
Milton GARCES 1, David FEE 1, Jerome AUCAN 1 and Robert GIBSON 2
1 Infrasound Laboratory, University of Hawaii     Manoa - USA
2 BBN Technologies     Arlington - USA

A public outreach DVD presents the deployment and preliminary conclusions derived from two surf infrasound experiments performed in the Hawaiian islands on January and March of 2005. The duration of the movie is ~ 15-20 minutes and provides an introduction to the subject of surf infrasound. A DVD player and speakers are needed for the presentation.

Infrasonic remote sensing of breaking ocean waves in the surf zone
Robert GIBSON 1, Milton GARCES 2, Jerome Aucan 2, David Fee 2, Joydeep Bhattacharyya 1and Sheila Shah 1
1 BBN Technologies     Arlington - USA
2 University of Hawaii     Manoa - USA

Infrasonic measurements, ground truth data collection, and ocean modelling have been conducted in near-shore environments in Hawaii in order to quantify relationships between infrasound and surf activity. An overview of the objectives and scope of the investigation is presented. Empirical analyses have been conducted in order to determine correlations between key observables, and implications for signal processing parameters are discussed. Additional results of the experiments are presented in a companion presentation by Aucan et al (these proceedings).
Sponsored by the Defense Advanced Research Projects Agency. Support of the State of Hawaii is acknowledged.

An analysis of infrasound signals from 11th December 2004 meteorite at Lanzhou in China
Hee-il LEE 1, Il-young CHE 1 and Alexis LE PICHON 2
1 Korean Institute of Geoscience and Mineral Resources (KIGAM)     Daejeon - REPUBLIC OF KOREA
2 CEA/ DIF / DASE / LDG     Bruyères-le-Châtel - FRANCE

A large meteorite was observed at the Lanzhou Province of China on December 11, 2004. The infrasound signals produced by this meteorite were well recorded by the Korean Infrasound Network which consisted of 3 seismo-acoustic array stations : BRDAR, CHNAR and KSGAR. However, these stations are aligned along the direction to the source. In order to improve the accuracy of the source location, we also used the data recorded at the IMS infrasound station I34MN in Mongolia. A preliminary PMCC analysis of the infrasound signals shows a moving signature of the source. Using the azimuths at each station and a rough propagation model, the location of the final explosion is estimated near coordinates 37.45° N-105.29° E. A second location (37.54° N-104.25° E) is obtained using appropriate celerity models and wind-corrected azimuths derived from 3D ray-tracing modeling in a more realistic atmosphere (MSIS90/HWM-93). Both locations are discussed. A possible trajectory of the meteorite is close to SE-NW direction. These locations are not confirmed due to the lack of information. Based on the observed signal amplitudes and dominant frequency, the estimated energy release from this meteorite ranges from 100 to 400 t of TNT.

Infrasound generated by the M7.8 Northern Chile earthquake of 13 June 2005
Alexis LE PICHON 1, P. MIALLE 1, Julien VERGOZ 1, Jocelyn GUILBERT 1 and D. DROB 2
1 CEA/ DIF / DASE / LDG     Bruyères-le-Châtel - FRANCE
2 E.O. HULBURT CENTER FOR SPACE RESEARCH, NAVAL RESEARCH LABORATORY     Washington DC - USA

On June 13, 2005, a major earthquake (M7.8) occurred in Tarapaca, Chile, within region of high mountains. At large distances from the epicenter, this event produced coherent infrasonic waves detected by the I08O-Bolivia (at 410 km), I09BR-Brazilia (2 300 km) and I41PA-Paraguay (1 420 km) IMS stations. The observed azimuth variations and the long signal durations suggest that wide regions in the Andes Mountains radiated infrasonic waves. From these observations, the main source regions of infrasound are reconstructed. The input parameters of the location procedure include the measured azimuths and arrival times, the origin time and coordinates of the epicenter. The velocity models used describe the propagation of the seismic surface waves and the propagation of infrasound in the direction of each station. The long range propagation is simulated using the WASP-3D ray theory-based method which account for the spatio-temporal variations of Ground to Space (G2S) horizontal wind terms along the ray paths. The reconstructed source regions as seen from each station are compared. Such an event, recorded by several infrasound stations, offers an opportunity to validate velocity models in the atmosphere as well as to improve the understanding of the amplification of ground displacement caused by the topography.

Infrasound produced by tsunami-generating earthquakes
Sara McNAMARA and Milton GARCES
Infrasound Laboratory, University of Hawaii     Manoa - USA

Several tsunami-generating earthquakes have produced infrasound which suggests that infrasound analysis could be a possible discriminant for tsunami genesis. We will present the infrasonic signals generated from the Great Sumatran (M9) event of December 26th, 2004, the March 28th, 2005 Nias (M8.7) event, the April 10th, 2005 Mentawai events (M6.7 and 6.7) and the August 16th, 2005 (M7.2) Miyagi-Ogi event. The Sumatra and Nias events were large and were recorded at multiple stations. Three separate arrivals were produced: the seismic arrival, the T-phase arrival, and the infrasound arrival. Both low (< 0.1 Hz) and high (> 1 Hz) frequency infrasound was generated. The Mentawai events, which had no associated tsunamis, produced only infrasound arrivals. For the events in the Sumatra region, which have been described in more detail elsewhere (1), (2), all of the stations were over 3 000 km away. In contrast, the Miyagi-Ogi event, while much smaller, was recorded at station IS30 which is only ~360 km away. The proximity of the station to the Miyagi-Ogi event has allowed for detailed analysis of detections from a range of azimuths. This allows the discrimination of infrasound produced from the earthquake vibrations of land massees and from the tsunami and its interaction with the coastline and bathymetry.
(1) M. GARCES, P. CARON, C. HETZER, A. LE PICHON, H. BASS, D. DROB, and J. BHATTACHARYYA, Deep Infrasound Radiated by the Sumatra Earthquake and Tsunami, EOS, 86, No 35, 30 August 2005.
(2) A. LE PICHON, P. HERRY, P. MIALLE, J. VERGOZ, N. BRACHET, M. GARCES, D. DROB and L. CERANNA, Infrasound associated with the 2004-2005 large Sumatra earthquakes and tsunami, Geophysical Research Letters, 2005 (in press).

Seismo-acoustic, wave acoustic observations and synergy
Dominique REYMOND, Jacques TALANDIER and Olivier HYVERNAUD
Laboratoire de Géophysique Papeete     French Polynesia - FRANCE

The excitation of atmospheric layer by seismic waves is shown through examples of moderate size earthquakes to giant earthquakes, like Alaska march 1964, Chile 1960 and Sumatra dec. 2004 ; on the other hand, in the range of low energy observations, the interactions between oceanic swell, seismic waves and long period infrasonic excitation are demonstrated.

Seismo-acoustic analysis of quarry blasts in the I 31 KZ region and the influence of atmospheric conditions on infrasonic wave propagation
Läslo EVERS 1, Irina ARISTOVA 1 and Alexandr SMIRNOV 2
1 Netherlands Meteorological Institute - KNMI     AE De Bilt - THE NETHERLANDS
2 Institute of Geophysical Research     Almaty - KAZAKHSTAN

IS 31 "Aktyubinsk" infrasound station has been operating in Kazakhstan since 2001. A new "Akbulak" seismic array situated 188 km south-east IS 31 began operating in 2003. It consists of 9 highly sensitive borehole seismic stations. Large quantity of industrial explosions required for mining operations is carried out in this region. Signals of those events are recorded both by the infrasound station and seismic array. Seismo-acoustic analysis of the blasts records in one of the chromite winning quarry is presented in the report. Data for the first six months of 2004 has been examined. Seasonal change of the atmospheric conditions influence onto the infrasonic wave propagation has been studied using records of infrasound and seismic components of the blasts signals.

Infrasound observations in Alaska and Antarctica
Curt A. L. SZUBERLA, John V. OLSON, Charles R. WILSON and Kenneth M. ARNOULT
Geophysical Institute, University of Alaska     Fairbanks - USA

The infrasound group of the University of Alaska Fairbanks operates CTBTO/IMS infrasound stations I 53 US (Fairbanks, Alaska) and I 55 US (Windless Bight, Antarctica). These arrays have been consistently providing data for the past three and four years, respectively. Data from the two locations are collected and catalogued at the Geophysical Institute of the University of Alaska Fairbanks in parallel with their transmission to the IDC. Once received at the Geophysical Institute, the data undergo routine analysis for signals that show high correlations and high signal to noise levels using a generalized cross-correlation estimator and an F-statistic estimator. A further discrimination is made between near- and far-field signals. The data are divided into three frequency bands to mitigate microbarom contamination. Daily and monthly summaries of the detection levels are produced for all data recorded at the arrays. We will present a précis of the typical signal content at each station, along with examples of representative signals arising over the years from various natural and manmade sources.

Thunder, waterfalls and volcanic eruptions : a menagerie of infrasound in the Ecuadorian jungle
Jeffrey B. JOHNSON 1, Hugo A. YEPES 2, Jonathan M. LEES 3, Alexander GARCIA 2, Mario RUIZ 3, Jaime LOZADA 2, Ethelwoldo JUA 2 and Andres CADENA 2
1 Department of Earth Sciences, University of New Hampshire     USA
2 Instituto Geofisico, Escuela Politecnica Nacional     Quito - ECUADOR
3 Department of Earth Sciences, University of North Carolina     USA

Since 1999, infrasonic studies have been an important component of volcanic monitoring at three active Ecuadorian volcanoes (Guagua Pichincha, Tungurahua, and Reventador). During the current, ongoing 2005 eruption of Reventador Volcano in Ecuador's eastern lowlands, three low-frequency sensitive microphones have been deployed at distances of ~ 2.0, 4.0, and 7.5 km from the inferred vent. These three dispersed sensors routinely record correlated infrasound in the 0.5 to 20 Hz bandwidth consistent with a vent source. Signals which include discrete short-duration explosive pulses (with recorded amplitudes in excess of 120 Pa at ~ 2 km) and extended-duration tremor (broadband and/or harmonic lasting for periods up to several hours) are attributed to vigorous degassing in the crater. These infrasonic transients enable scientists at the IG-EPN improved understanding of the enigmatic seismic signals produced by the volcano and recorded coincidentally.
Additionally the dispersed infrasonic network has been able to identify coherent, consistent infrasound from non-volcanic sources. A continuous (24 hours/day) infrasound radiator has been identified from the vicinity of San Rafael Falls where the Rio Quijos plunges over 150 m. This infrasound source, which peaks at ~ 1 Hz, is identifiable at all three stations, situated 2 km, 5 km and 8 km from the waterfall. Attempts are now being made to correlate acoustic energy from this source with river flow flux data from a nearby gauging station. Finally, the infrasound network has been successful at identifying non-stationary sources, which are attributed to thunder from electrical storms. Discrete thunder waveforms are poorly correlated across the network owing to their diffuse source volumes, but the signal onsets are impulsive enabling accurate time picks for the initiation of lightning strikes. Our data set, which includes hundreds of thunder signals, enables us to effectively locate the lightning onset.


Session 7 : VOLCANO SESSION

Multi-technology observations of the Manam volcano eruption in January 2005 - Potential for an enhanced volcanic ash warning system (Poster)
David BROWN 1, Andrew TUPPER 2, Douglas CHRISTIE 3 and Ima ITIKARAI 4
1 Geospatial and Earth Monitoring Division, Geoscience Australia     Canberra - AUSTRALIA
2 Darwin Volcanic Ash Advisory Centre     Darwin - Northern Territory - AUSTRALIA
3 Research School of Earth Sciences, Australian National University     AUSTRALIA
4 Rabaul Volcano Observatory     Rabaul - PAPUA NEW GUINEA

Ash Events from Volcanic activity pose a significant threat to aviation. The Darwin Volcanic Ash Advisory Centre (VAAC) and the Rabaul Volcano Observatory (RVO) have the task of monitoring the environment for potentially hazardous activity throughout all of Indonesia, Papua New Guinea, the Philippines and islands in the southwest Pacific and eastern Indian Ocean.
The violent eruption of Manam Volcano in Papua New Guinea on January 27, 2005 is one of the most significant eruptions in the past 15 years. This eruption provides an opportunity to compare observations from various technologies and identify areas where observations from one technology can fill the gap in others. For example, satellite data can be compromised by significant cloud cover, and infrasonic data may be compromised by adverse seasonal winds.
The acoustic observations, primarily from IS07 Warramunga in Central Australia, are compared with pilot observations, RVO records, and post-analysis satellite observations of the Manam ash cloud. Preliminary results indicate a generally good correlation between recorded infrasound from Manam at IS07 in the frequency band from 0.03 to 0.1 Hz and known ash events from satellite post-analysis and pilot observations. There are several noteworthy infrasonic episodes that cannot be correlated with corresponding volcanic ash clouds in the post analysis satellite observations. These additional acoustic observations may correspond to significant ash events that cannot be identified in the satellite observations due to significant cloud cover that was known to exist over the area at the time.
Infrasound can clearly play an important role in the volcano ash monitoring problem. Work is proceeding rapidly on the establishment of the 60-station infrasound component of the International Monitoring System. Many of the stations in this global network are located in areas that are well suited to the monitoring of volcanic activity. It must be anticipated that data from the global infrasound network would enhance exi?? sting volcanic ash warning systems.

Detection of infrasound from the eruption of Manam volcano on January 27, 2005
Paola CAMPUS 1, Douglas R. CHRISTIE 2 and David J. BROWN 3
1 CTBTO/PTS, IMS, Acoustic Monitoring Section, Infrasound Monitoring Unit     Vienna - AUSTRIA
2 Research School of Earth Sciences, The Australian National University     Canberra - ACT - AUSTRALIA
3 Geoscience Australia     Canberra - ACT - AUSTRALIA

One of the largest explosive volcanic eruptions in recent years occurred on January 27, 2005 on the small island of Manam near the north coast of Papua New Guinea. This violent eruption started at about 14:00 UT (midnight local time) and continued as a series of large explosions for a period of about two hours. The volcano monitoring station on Manam was destroyed and estimates of the height of the volcanic ash plume generated by this eruption range from about 21 to 24 km. Manam is a very active volcano and there were numerous reports of minor eruptions in the period before and after the main explosive eruption. Infrasonic signals from the eruption on 27 January were detected at a large number of IMS infrasound monitoring stations. Most of these signals were detected only at long periods. In some cases, however, the signals were observed at both low and high frequencies. This investigation is concerned with a detailed study of the properties of infrasonic waves generated by the Manam eruption at distances ranging from a few thousand km to more than ten thousand kilometres. The results of this investigation highlight the detection capability of the IMS infrasound system. The results also suggest that data from the IMS infrasound monitoring system could play a valuable role in a global volcanic ash warning system.

The CTBT IMS infrasound network and monitoring of volcanoes
Paola CAMPUS
CTBTO/PTS, IMS, Acoustic Monitoring Section, Infrasound Monitoring Unit     Vienna - AUSTRIA

The International Monitoring System of the Comprehensive Nuclear-Test-Ban Treaty includes an Infrasonic Network consisting of 60 stations worldwide distributed.
Although the primary purpose of this Network is to monitor the occurrence of nuclear tests around the world, it is becoming more and more evident that it can give as well a potentially significant contribution to the monitoring of volcanic eruptions around the world. Some examples are presented and discussed.

Diurnal variations in infrasonic Tremor signals from Kilauea volcano, Hawaii
David FEE 1, Sara McNAMARA 1, Milton GARCES 1, W. BORTZ 1, R. HOBLITT 2 and F. TRUSDELL 2
1 Infrasound Laboratory, University of Hawaii     Manoa - USA
2 US Geological Survey - Hawaiian Volcano Observatory - Hawaii National Park     Hawaii - USA

A portable infrasound array deployed ~13 km from the active Puu Oo crater complex of Kilauea Volcano, Hawaii recorded near-continuous infrasonic tremor from August 18th - 26th, 2005. Results from the PMCC detection algorithm show the tremor appears to originate from an extremely well defined direction, pointing to the active Puu Oo vent. Spectral analysis of the data shows the infrasonic tremor appears to be concentrated below 3 Hz and exhibits clear diurnal amplitude variations. Vent diameters along the crater and active lava tube system were on the order of tens of meters or less, whereas acoustic wavelengths are on the order of hundreds of meters, so individual open vents may be treated as compact acoustic sources.
The recording system consisted of four Chaparral 2.2 microphones on high gain (2V/Pa), an ultrasonic wind sensor, and a 24-bit digitizer with GPS time synchronization. The station was deployed as a 60m-aperture triangular array with a central element, and within a dense tropical forest that sheltered the station from surface winds. Because the wind sensor was deployed in a heavily forested area, both the wind direction and wind speed measurements are not representative of the region's wind patterns. However, due to the penetration depth of atmospheric boundary layer turbulence at long periods, the effects of strong winds can still be observed in the infrasound data below ~0.5 Hz.
Comparison of spectral amplitudes between the 0.01-0.5 Hz band (a proxy for wind speed) and the 1.5-3 Hz band (where infrasonic tremor is concentrated) shows that during periods of high wind the infrasonic tremor amplitudes and PMCC detections decrease substantially. For this part of Hawaii, these periods of strong winds correspond to the increase in NE trade wind speed during the middle of the day. The Puu Oo crater complex is located SE of the array location, thus the trade winds blow perpendicular to the path between the source and instruments. When the trade winds decreased during the night both the infrasonic tremor amplitude and PMCC detections increase. This diurnal variability may be due to a combination of wind effects and different boundary layer conditions during the day and night.
Because of its persistence, infrasonic signals from Kilauea volcano may be used to assess the range-dependent capabilities of infrasound monitoring. This pilot study demonstrates that adequate site and instrument selection and recognition of the regional wind and boundary layer patterns is necessary to permit robust infrasonic remote sensing of volcanic processes.

The physics of volcano-acoustic sources
Miton GARCES
Infrasound Laboratory, University of Hawaii     Manoa - USA

The intensity of infrasonic waves produced by volcanic activity ranges from very low amplitude pressure signals (mPa) to violent shock waves produced during explosive eruptions (MPa). Recorded waveforms vary from simple single pulses to complicated, long lasting signals where echoes and/or multiple pulses may be present. Whether echoes occur, are sustained, and are recorded depends on the elasticity of the surrounding walls, the attenuation of the fluid, the depth of the source, and the relative position of the sensor. A shallow explosion would release most of its energy to the atmosphere. In this case, echoes would be primarily associated with reflections from crater walls or nearby mountains. A deep explosion in a vesiculated magma column may not be multiply reflected (and thus maintain resonance) in a conduit if it has to propagate through a heavily attenuating magma-gas mixture. Yet highly vesiculated foams, with their low sound speeds and their sensitive dependence of gas exsolution and viscosity on ambient pressure, are extremely unstable under any fluid flow conditions. Due to the decrease in density and sound speed with increased vesiculation, an acoustic pulse arriving from some depth in a moving magma column would encounter an increase in Mach number as it approaches a highly vesiculated region. When this pulse reaches the foam, the pressure perturbation and its associated streaming may induce rapid exsolution and trigger a fragmentation enhanced explosive eruption that could lower the fragmentation void fraction threshold and enhance jet flow. Lowering of the fragmentation threshold may permit conduit reverberation.
Cavitation may occur when a fluid is excessively tensed. Flow acceleration through a constriction (choked flow), or the passage of an intense sound pulse can induce cavitation and produce a bubble oscillation. The precondition of existing bubbles for cavitation lend vesiculated foams particularly vulnerable to collapse. Sound from periodic turbulent vortices induced by surface discontinu?? ities or shear (Aeolian tones, edge tones, vortex sheets) may occur at depth in the melt or at the ground-air interface. Avalanches, landslides, and pyroclastic flows would also generate acoustically active turbulent structures, as well as a sound from impact and explosive gas release. Jet noise can be produced by fumaroles, lava tubes, and eruptions. Jet flow resonance, known as screech, may occur within a supersonic jet and be observable during vigorous eruptions. Vigorous lava fountaining events radiate discrete infrasonic pulses which may be indicative of oscillations in the pressure driving the fluid flow. Infrasound from the oscillation of a lava tube or lava lake may be produced by the movement of the magma. Sound from lava falls, as seen through skylights in Pu'u O'o, may be enhanced by ringing of the air in a lava tube.
As in the ocean, standing waves in a molten lava lake may generate sound efficiently if they slam into walls or if they entrain periodic flow into confined regions. As in a furnace, pressure and thermal oscillations may be induced in a lava tube when the gas in the tube is overburned, leading to a low pressure with gas overdrawing, followed by a fiery pressure increase during subsequent overburning.

Seismo-acoustic study of Mount Saint Helens
Robin MATOZA and Michael HEDLIN
Laboratory for Atmospheric Acoustics, Institute of Geophysics and Planetary Physics, University of California     San Diego - USA

Following the eruptions in October, 2004 we deployed infrasound arrays 13 km to the north of Mount Saint Helens and 240 km to the east. Each array comprises four MB2000 microbarometers in a centered triangle spanning 100 meters. A Guralp broadband seismometer was co-located with each array. Our intent was to collect close-in recordings of acoustic and seismic emissions from the volcano to accurately document the source. The more distant array was deployed to determine how readily activity at the volcano could be monitored remotely.
We observe clear acoustic signals from the direction of MSH at both of our arrays. These signals coincide with activity recorded by other monitoring systems (e.g. visual and seismic). We also observe periodic sequences of acoustic signals from the direction of the crater that do not correlate other known activity at the volcano. We report our preliminary findings from our analysis of these signals.

Madagascar infrasound station I 33 S and the observed signals
Gérard RAMBOLAMANANA, Andry RAMANANTSOA and Raveloson ANDRIAMIRANTO
Institut et Observatoire de Géophysique d'Antananarive     Antananarive - MADAGASCAR

The I 33 S CTBT/IMS infrasonic station, located at Imerintsiatosika, 35 km from Antananarivo, MADAGASCAR, commonly observes infrasound associated with ocean tides, cyclones, volcano eruption of Karthala Comoros and quarry blasts. These signals are divided into two main range of frequency. In the range of Low Frequency (0-0.5Hz) we have observed signals from ocean tides which occur every day, signals from cyclones during rain season and signals from volcano eruption from Karthala, Comoros situated at 900 km from I 33 S. In the range of High Frequency (0.5-4 Hz), signals are generated by quarry blasts and unknown event occurred at 39° azimuth. The PMCC method is used to process data.

Monitoring of Indonesian volcanoes with infrasound : preliminary results
Jocelyn GUILBERT 1, Prih HARJADI 2, Masatje PURBAWINATA 3, Suzon JAMMES 1, Alexis LE PICHON 1 and Bruno FEIGNIER 1
1 CEA/ DIF / DASE / LDG     Bruyères-le-Châtel - FRANCE
2 Badan Meteorologi dan Geophisica     Jakarta - INDONESIA
3 Directorate of Volcanology and Geological Hazard Mitigation     Bandung - INDONESIA

In conjunction with other systems of observation, the infrasound technology has already shown its potentiality for the detection of volcanic activity. Monitoring volcanic activity is essential in Indonesia where volcanoes are aligned along the arc of subduction over more than 5 000 km. Due to the mutual experience and knowledge on meteorological observations, infrasound observation and modeling, the BMG (Indonesia), VSI (Indonesia) and the CEA (France) have decided to cooperate to validate this technology in the Indonesian context. The Kalimantan Island (Borneo) was identified as the best location for the installation of an infrasound array. Considering its central position, all active volcanoes in Indonesian would be detected with a good azimuthal resolution. Several months of PMCC detection bulletins are presented. These detections are analyzed along with all available reports of volcanic activity of Indonesia and Philippines. The detection capability of the infrasound array is demonstrated and examples will be presented. Detection results from other IMS stations around Indonesia are used for identifying potential sources of infrasound. This comparison proves that the Kalimantan array is ideally situated to uniformly monitor the volcanic activity along the Indonesia-Philippines arc.