Acoustic anomaly in Oklahoma
© OKGEOSURVEY
Here the anomaly can be seen at station SC14, boxed in red. It repeated once in this instance, as it usually does (June 26, 2019). (Each horizontal line represents an hour.)
"What is that?" asked Dr. Walter, pointing to the seismograms displayed on a flat screen TV hanging on the wall of the OGS seismic lab. There were a series of red marks that indicate automatic picks by the computer as potential earthquakes.

However, these marks were spaced very regularly, so regularly that at first glance they looked like some sort of mechanical noise.

The problem with that assumption was that they were showing up on stations all across the state, all at the same time. Anything that widespread is usually associated with a correspondingly large scale event, like an earthquake.

This pattern we were seeing looked nothing like an earthquake, or even a series of earthquakes. Other potential causes we guessed at were military aircraft, meteor shower, or something related to gas pipelines. We dubbed this acoustic pattern 'The Anomaly.'

Curious about what we were seeing, and what our sensors were hearing, we attempted to locate a potential source for individual pulses from the pattern.

This produced nothing but garbage location potentials with errors so high it was useless. We also attempted locating a source using the first arrival time of the pattern at various stations where it was clearly discernible.

This yielded a more interesting result. It still failed to yield a location, however, by plotting out the first arrival times on a map, the anomaly arrived in a 'ping pong' like pattern back and forth across the state, mainly in a swath from Tulsa, across OKC and toward Lawton.

Oklahoma acoustic anomaly -  pattern, twice. Two pulses followed by a gap, followed by 24 pulses.
© OKGEOSURVEY
This image shows the anomaly pattern, twice. Two pulses followed by a gap, followed by 24 pulses.
In following days, we noticed the anomaly showing up again, and at around the same time each day, as well (11:11am local time). This led to checking the data back in time to see what days we might have not noticed it.

Indeed, the anomaly had been occurring regularly. Oddly, though, sorting through the data backwards in time, the anomaly showed up less regularly and on fewer stations. After checking a period of three weeks with no activity, it seemed that March 15th was the first day that the anomaly occurred.

In performing this investigation, more curious details emerged. The anomaly has been detected by sensors of various types (i.e.: Streckeisen STS-2, Guralp CMG-6T, Guralp CMG-C3ESPC, Guralp CMG-3T, RaspberryShake, etc.), on various networks, including networks the OGS doesn't manage (i.e.: Kansas, Texas, etc.)

Also, the pattern of the anomaly is quite specific, though it does often repeat once (never more than that, though): pulse, 20s pause, pulse, 80s pause, then 24 repeats of pulse, 20s pause. There have been very few variations to this pattern.

Each individual pulse ranges in intensity from what would be the equivalent of a magnitude 1 earthquake, up to magnitude 2.2 (this hasn't been investigated exhaustively, these values are from random sampling). Furthermore, we noted that the anomaly has never occurred on a Sunday, and it also didn't occur over the 4th of July weekend.

We contacted the Oklahoma Corporation Commission about the anomaly, and inquired if they had received any reports of strange noises. Yes, in fact, they had received reports, and had even sent personnel to investigate some of the reports. However, they are as much in the dark as to a potential cause as we are. Contacting Tinker AFB, Vance AFB, and Altus AFB also didn't lead to any answers.

Oklahoma acoustic anomaly graph
© OKGEOSURVEY
Move out plot for the anomaly on Saturday, July 13, 2019. Each data point represents a station, showing when it detected the first ‘pulse’ of the anomaly, and how far away that station is from the first station to detect the anomaly that day. Note that the move out plot doesn’t make a nice, straight line, as would be typical for point sources like earthquakes or blasts.
The anomaly became even more curious after pursuing another avenue of inquiry. Picking a day with a significant number of data points (41 stations had clearly detected the anomaly on July 13, 2019), we generated a move out plot. Move out plots are simple distance versus time plots where the slope shows velocity.

We still couldn't determine the origin of the anomaly (even making first phase picks on the same 'pulse' of the anomaly on each of the stations that clearly detected it failed to yield an origin location without an error area larger than the state of Oklahoma). So, for a move out plot we instead used the 'origin' as the first station that detected the anomaly that day.

The plot might still be useful to determine velocity. And indeed, it did yield a velocity for the anomaly on that day: approximately 370m/s. This is near the speed of sound in air. That suggested that the anomaly wasn't primarily moving as energy through the ground (vibrations move much faster through the earth than the air), but rather was moving through the air. Also, the overall trend of detection suggested that the anomaly moved from the Southeast to the Northwest.

The precision of timing in the pattern and regularity of its timing most days suggest that it is associated with some sort of widespread activity in the state that has recently been implemented (since March 15th). Also, over time it is occurring more regularly and more strongly (becoming more widespread). As for an exact cause, however, we're still just guessing.

Oklahoma acoustic anomaly screenshot
© OKGEOSURVEY
This shows the anomaly as it appeared on multiple stations, June 24, 2019.