The CTBT/IMS infrasonics array I53US is located in the forested area north of the University of Alaska campus in Fairbanks, Alaska. The geographic coordinates of the center of the microphone array are: 64.8671 deg. North latitude, 147.8559 deg. West longitude. Previous infrasonic arrays have been operated in this same area for 20 years with great success at observing signals of very low amplitude down to 0.01 Pa. In the winter at Fairbanks there is a strong surface temperature inversion that produces very stable windless conditions at the locations of the I53US microphones. In May of 2002 work was begun on the installation of the I53US infrasonic array.

Figure 1 below is an aerial view looking toward the east of the both the University of Alaska campus, with the eight story Geophysical Institute building shown to the  left of center, and the boreal forest on the far left  of the G.I building where the I53US infrasonic array is installed, The locations of two of the microphone sites are identified in the picture as I53H2 and I53H3. From the picture it is clear that there is line-of-sight radio telemetry from the microphone sites to the roof of the G.I. building. All of the microphone sites are in bosky areas.

An eight microphone array with 5 microphones in a pentagonal pattern and three microphones in an inner triangular pattern was installed at I53US. The aperture of pentagonal array pattern is 1.7 km, while that of the inner triangle array is about one tenth as large. The I53US array is similar to but not identical that at I55US in Antarctica. With the use of 8 sensors in an array with a large pentagon pattern of 5 sensors surrounding an inner smaller triangular pattern of 3 sensors results gives an excellent impulse response diagram with virtually no spatial aliasing. This produces very much better infrasonic signal detection capabilities over the entire frequency band from 0.02 to 10 Hz as specified by the CTBTO as compared with that of the original 4 sensor CTBT prototype array design.

The infrasonic pressure time-series data are telemetered from each site to the Geophysical Institute building. The monitoring of the data from both I55US and I53US is done at the CTBT infrasonic hub in an air-conditioned room in the Geophysical Institute building using both visual observation of the real-time data and an automatic computer system. Over the several years of operation of I55US and I53US the data reliability has been near the 99% level. The infrasonic data from both I55US and I53US are stored in the CTBT IDC in Vienna and locally on hard drives. Eventually the infrasonic data is all transferred to CD-ROMs for archiving at the GI of UAF.   

Eight Model 5 Chaparral microphones with Geotech digitizers comprise the I53US array.  Commercial power is available at each microphone site. . The microphone, the digitizer, an ac to dc power panel and a Free-Wave radio at each site are housed in 40 inch insulated square boxes made of ¼ inch thick steel to protect the equipment pictured in Figure 2 below. There is a 15 foot high steel mast attached to each instrument box with a GPS unit and the telemetry antenna mounted on its top. The four pipes radiating from the box at ground level are used to connect the microphone to its noise-reducing pipe array.

Wind noise-reduction at each site is accomplished using the standard CTBT design of a set of four rosettes of 24 radial pipes vented at the end of each pipe as shown in Figure 3. Each site’s wind-reducing system of pipes is 18 meters in width with a total of 96 low impedance inlets, 24 inlets installed on each rosette. The design of the total noise reducing pipe array used at each microphone is shown in Figure 4. The low impedance vent at the end of each radial pipe is shown in Figure 5.

The Model 5 Chaparral microphones are built with a special intake manifold on top that will also mate with a portable field calibration device. The manifold also provides four import fittings for connection to the noise-reduction pipe array. In Figure 6 below the Chaparral microphone’s   white manifold is shown in the foreground of the picture with a port hole on the top for use in the insertion of the portable calibrator. The calibrator itself is shown in Figure 6 in the background of the picture. It is mounted on another Chaparral microphone.  At both I55US and I53US all the Chaparral microphones were calibrated in-situs after installation at the microphone sites. The calibrations were done at the three frequencies of 1.0 Hz, 0.10 Hz, and 0.02 Hz using a sinusoidal pressure input of 0.1 Pa peak-to-peak.

In Figure 7 there is a diagram of the geometry of the eight-sensor I53US infrasonic array. The array pattern is an outer pentagon of 5 microphones and an   inner centered-triangle of three microphones. The 5 microphones at the apexes of the pentagon are each approximately 1000 meters from the center of the array while the inner 3 microphones are each approximately 100 meters from the center. The center of the array is about 1500 meters to the NW of the G.I. building. Local topography to some extent prevented the establishment of a perfect pentagon array pattern with an equilateral triangle pattern at the center. The northernmost microphone I53H1 is taken as the reference site with coordinates ( 0 east  , 0 north.). The East and North coordinates of the 8 microphones are given in kilometers in the table below.

I53US Microphone Coordinates

                        Sensor     H1          H2          H3          H4           H5           H6           H7           H8

                        East         0.0        1.0803    0.8663   - 0.2355   - 0.8243   - 0.0027   0.2092     0.0610

                        North      0.0      - 0.2904  - 1.4345  - 1.7622   - 0.7709   - 0.8632   - 0.9801  - 1.0652

In  wave number space  the output Y(k)  of a microphone  array  to an infrasonic signal X(k)  impinging upon the array is given by the convolution of X(k) with the array impulse response H(k). The array impulse response H(k) is the array’s response to a unit impulse input at t = 0. In order to make an unambiguous detection of an incoming infrasonic signal impinging upon a microphone array it is necessary to design the array pattern so that it’s impulse response has most of the power in it’s central maximum at K = 0. If there is a lot of power distributed among many other maxima in the H(k) impulse response diagram in K space then  spatial aliasing is possible and may affect the detection capability of the array. The pentagon-triangle array design for I53US and I55US, as shown in Figure 7, has an excellent impulse function as can be seen in Figure 8 below. This array design has nearly circular symmetry so that the uncertainties in the determination of trace-velocity and azimuth of arrival of an incoming infrasonic signal is nearly independent of azimuth. In Figure 8 the principal maxima in the middle of the diagram, that can be seen at coordinates X scale units = 50 and Y scale units = 50, is at the origin at Kx = 0 and Ky =0 in K space.

An example, using a Microbarom event at I53US on October 21, 2005, of the convolution of an input signal Y(k) with the microphone array impulse response function H(k) is used to illustrate the role that the impulse response H(k) plays in signal detection. The input signal is shown in Figure 9 is a time series from a microbarom wavetrain of three minutes of pressure data. The trace velocity of this signal was 0.353 km/sec and its azimuth was 188 degrees with a coherence value of 0.86. The detector is a Bartlett estimator based on the spectral matrix of the signal data. The detector algorithm output is a three dimensional matrix of detector values [D], (ranging from 0 to 1 ), as a function of  three parameters: frequency, slowness and azimuth of the signal. A Matlab script performs a sweep in slowness and azimuth within a designated band of frequencies.

In Figure 10 a Cartesian plot is shown of the contours of the value of the detector D, as derived from the microbarom data. There is a strong maximum in the contours of D in Figure 10 at an ordinate value at 28.3 units for slowness with an abscissa value of 94 units for azimuth. This maximum represents the clear detection of the microbarom signal at a slowness of 2.83 sec/km and an azimuth of 188 deg.

In order to show the relationship between the impulse response diagram of the I53US array, that is shown in Figure 8, and the detection of a infrasonic microbarom signal by the stdetect analysis that is portrayed in Figure 10, a polar coordinate plot of the contours of the value of the detector D is given in Figure 11. By the comparison of the impulse response contour pattern of Figure 8 with the detector D contour pattern in Figure 11 one can clearly see the results of the convolution of the input signal Y(k) with the impulse response H(k). In polar coordinate plots in Figures 8 and 11 one can see an almost identical pattern of contours with the exception that the contours of D in Figure 11 are shifted off the origin at ( 0, 0) in the direction of propagation of the signal of 188 degrees by an amount equal to the magnitude of the slowness vector of 2.83 sec/km. The shift of the center of the D contour maximum off the origin, in Figure 11, is indicated by the red arrow that points in the direction of propagation of the microbarom signal that is depicted in Figure 9. The magnitudes of principal maximum in the impulse response diagram and in the detector diagram are much larger than all the other contour maxima in both diagrams. This leads to an unambiguous detection of signals with no spatial aliasing in the CTBTO frequency passband from 0.02 to 10 Hz.