University of Mississippi
National Center for Physical Acoustics
University, MS 38677


Roger Waxler
University of Mississippi
National Center for Physical Acoustics
University, MS 38677


This document describes the supplementary information for the Acoustics Today paper “Infrasound from Tornados: Theory, Measurement, and Prospects for their use in Early Warning Systems: Supplemental Information” by Talmadge and Waxler.


This is a Google Earth “KML” file.  It contains the tornado track data from the National Weather Service Weather Forecast Office in Norman Oklahoma.  The original webpage page for this outbreak is located at


This directory contains a number of recordings converted to WAV format from sensors SN 56 (left channel) and SN 68 (right channel),which were located near the CPEG and STW tornado tracks. The format is 24-bit IEEE floating point, and the units are volts.  An approximate conversion to Pascal can be made by multiplying these values by 40.  Care must be taken when reading these files into programs such as MATLAB, as the default behavior for these programs can be to rescale the input values so there is no clipping.



This is a 1000-sps stereo WAV file.  Since the sampling rate is very low, it may not play correctly using default audio players.  Audacity is one free-to-download audio editor known to this sound file correctly.  Some programs, such as MATLAB, will rescale the files if you use the default wav reader

The remaining files have been resampled from 1000-sps to 22050-sps, and should play on any standard audio player.  The following table gives the range and direction of the tornado track associated with these recordings.

Sound File Range [km] Bearing [deg]

33 219

29 216


11 145

* No tornado on the ground.

The approximate direction of the tornado from the array is noted on Figure 1 below.

Note that the tornado meta data provided by the National Weather Service is approximate.  Thus the time and positions of the tornado vortices must be viewed as approximate.  Similarly, the stop time of the CPEG tornado and start time for the STW tornado are only approximately denoted on this map, and denote periods for which there was significant tornadic damage. In all likelihood, tornado vortices were present during part of the interval 22:35-22:50 for which the metadata suggests no tornado is present.



Figure 1:  Google Earth image showing the direction of the tornado vortex (when present) during the recording period for each sound file.  The dashed lines correspond to the approximate direction to a tornado vortex, were it present.

If, as this study suggests, the low frequency and infrasound signals represent tornadic activity, this data could be a more reliable predictor of the presence of a tornado vortex than the National Weather Service “ground truth”.