The ubiquitous five second period infrasonic signals called “microbaroms”, which are generated by standing sea waves in marine storms, are the cause of the low-level natural-infrasound background in the passband from 0.02 to 10 Hz. At infrasonic arrays all over the Earth during all seasons, in periods of no local wind-noise, microbaroms can be detected in the infrasound signal observations even though the microphone arrays may be thousands of kilometers from the source storms. Microbaroms are narrow band signals with a center frequency of 0.2 Hz where the band width of the signals is such that  df/f ~ 1. The pressure versus time traces of microbaroms appear as an amplitude modulated string of “pearls” or wave packets of 0.2 Hz oscillations with a packet length of from 2 to 5 cycles. An example of the pressure trace of a microbarom signal observed at one sensor H6 from the I53US array can be seen in Figure 1 below. There is often a visible change of phase in the wave train from one wave packet to the next packet. Hilbert transform techniques indicate that the mean packet length for microbaroms is  about 2 wave lengths. The coherence of microbaroms is limited to spatial scales of 2 to 5 km. The inter-microphone coherence has been found to drop significantly across the microphone array between pairs of sensors that are separated along the direction that is perpendicular to the wave propagation.

Microbaroms are radiated in all directions from a marine storm center. However, they may or may not be observed at an infrasonic array depending on the strength and direction of the upper level winds along the propagation path between a marine storm and an infrasonic array. The microbarom signal strength observed at an infrasonic array is greatest in the direction that is down-wind from the source. An example of a microbarom-producing marine storm on October 21, 2005 in the Gulf of Alaska is shown in Figure 2 as a map of the surface- pressure contours. The long black arrow in Figure 2 from the center of the storm shows the direction of propagation of the microbarom signals to I53US at Fairbanks. During the day the location of the storm changed, drifting toward the south from an azimuth of 220 deg. to 180 deg over a 24 hour period. This motion of the source storm is consistent with in the change in the direction of the microbarom signals received at I53US as plotted in Figure 3 below. The azimuth of the received microbarom signals drifted at about 1.7 degrees per hour following the storm’s motion. In Figure 4 an expanded view of the azimuth versus time diagram in Figure 3 is given for the short time interval from 16:49 to 17:10 UT using an analysis window of 60 seconds. This expanded view is presented to illustrate that successive wave packets can come from quite different directions that are within the band of azimuths subtended by the source storm as seen from I53US. For the data in Figure 4 the azimuthal width of the storm was from about 20 degrees from 185 to 205 degrees.

The detector plot for the storm of Oct. 21, 2005 is shown in Figure 5. The data were filtered in the microbarom passband from 0.1 to 0.5 Hz. An analysis window of 60 seconds length was used in the analysis with a 30 second over-lap for a total of 2880 sliding analysis windows. Only the inner three microphones in the array were used for the microbarom analysis of Oct.21, 2005. In the next to the bottom panel of Figure 5 for azimuth it is clear that the microbarom signals come from a narrow band of directions around 200 degrees. In the upper left panel the probability-density-function for the correlation value MCCM shows the very high inter-microphone microbarom signal coherence over the small inner 3 sensors of the I53US array. The bottom panel shows that the microbarom wave packets are plainer and thus from a distant source.

The statistical characteristics of the microbarom parameters of trace-velocity, azimuth and MCCM value are shown respectively by the histograms in Figure 6, 7 and 8 for the Oct.21, 2005 event. The mean microbarom trace-velocity was 0.462 km/sec with an SDT of 0.066 km/sec. The mean azimuth of arrival over a 24 hour period of microbaroms was 200 deg. with a STD of 13.6 deg. The mean MCCM value was 0.947 for the microbarom event. There is an intrinsic uncertainty that can be calculated in the estimates of the trace-velocity and the azimuth for the I53US microphone array. The calculated uncertainties depend on the array geometry, the measured trace-velocity, the azimuth and the sigma-tau value as derived by the Datascan least-squares signal processing algorithm.

In Figure 9, uncertainty diagrams are given in the trace-velocity plane and the slowness plane for input values of the means of trace-velocity, azimuth and sigma-tau for the Oct. 21 microbarom event. The intrinsic uncertainty in the estimates of trace-velocity and azimuth are respectively +/- 0.066 km/sec and +/ - 6.0 degrees. Returning now to Figure 6 we see that the standard deviation on the histogram of trace-velocity was 0.066 km/sec. which is in agreement with the intrinsic uncertainty associated with the array geometry. In Figure 7 the histogram of azimuth has a standard deviation of 13.6 degrees for the actual microbarom event. This value of 13.6 is twice as large as the calculated intrinsic uncertainty in the estimate of azimuth of +/ - 6.0 deg. Thus the spread in the azimuth for an actual microbarom event can be seen as a true measure of the angular width of the marine storm source region as seen from the infrasonic array.

If the propagation conditions are suitable then it is sometimes possible to triangulate on the source region of a microbarom storm using several highly separated infrasonic arrays. This was done for the tropical storm “Barbara” in the Pacific Ocean of June 20, 2001. by using microbarom azimuth data from I59US in Hawaii, I53US data from Fairbanks and data from an array near Los Angeles. The intersection of the three azimuth directions from the three infrasonic arrays is shown in Figure 10 to be in the Pacific where tropical storm “Barbara” was located.

At I53US and I55US diagrams are made, in the pass band from 0.1 to 10 Hz, for microbarom signals that have an MCCM value greater than 0.7 as summary diagrams for the entire month as polar plots of Trace-velocity versus Azimuth. Examples of two of these plots, for winter months, are given respectively for October, 2004 at I53US and for September, 2004 at I55US in Figures 11 and 12. In each of these figures the plotted points are in red if the value of sigma-tau for the wave packet was less than 0.15 sec. and in blue if sigma-tau was greater than 0.15 sec. In Figure 11 for I53US at Fairbanks the microbarom storms can be seen to lie in Gulf of Alaska or the Bering Sea. In Figure 12 for I55US in Windless Bight, Antarctica the microbarom storms lie in The Ross Sea to the northwest of the array.