Sapulpa, OK Mini-Supercell Tornado
Between 4pm and 6pm CST on Sunday, October 31, 1999 a mini-supercell thunderstorm spawned three F0 (preliminary ratings - lowest on the Fujita Scale) tornadoes in Creek County Oklahoma. While very difficult to diagnose in real-time, examination of thunderstorm's structure and environment reveals the supercell nature of the event. This event description is somewhat technical and very preliminary. It's intended to quickly share ideas with other area meteorologists, emergency managers and spotters. Hyperlinks will take you to definitions of some of the more technical terms.
The radar image below is from the KINX WSR-88D (Tulsa - Inola NEXRAD) at 4:07 pm CST . The 4-panel image contains radar reflectivity (upper images) and storm-relative velocity data (lower images). The upper images are from 0.5 and 3.4 degree beam elevations. The lowest elevation data show a small hook echo, while the data from the 3.4 degree slice show an elevated reflectivity core. This structure is typical for supercell storms and suggests strong storm-relative inflow and a significant updraft. The unusual aspect is that these features are tiny as compared with classic supercells. The situation is further complicated in that the cell orientation and motion are opposite of a more typical event...this cell is moving west-northwest. Only very weak rotation can be seen in the low-level storm-relative velocity data (lower left pannel: red is air motion away the radar in Inola / green represents air motion toward the radar).
The radar data above are approximately 10 minutes prior to the first tornado touchdown...3 miles northeast of Sapulpa on Hilton Road. The developing mesocyclone passed very near the Riverside Airport in Jenks at 4:10 pm. Data from the sensors there show the wind shifted to the northwest and blew at 13 knots...providing significant storm-relative inflow. The barometric pressure at Jenks fell and rose rapidly for short time as the developing circulation passed. Also of great importance to this event was the arrival of a wind speed max aloft. Data from the KINX WSR-88D showed that the winds at 7 to 14 thousand feet above the ground approximately doubled in speed from 15 knots to 30 knots at some levels and from 30 knots to 55 knots at others. This speed change occurred over about a 30 minute period and briefly resulted in an environment capable of supporting supercell storm structure. Below is an estimated hodograph for the event. The pattern of the hodograph, while again opposite in look as compared to a classic event, shows the supercell environment.
The first tornado touched down at approximately 4:17 pm CST. An eyewitness who saw the developement described low clouds converging on a single spot in the cloud base, followed by steadily increasing rotation just prior to the tornado. The radar data below is from the time of the tornado and is again in a 4-panel format. The images on the left are from the 0.5 degree slices and show a continued tiny hook echo and a small area of relatively strong rotation. The rotational signature in this area was approximately 1.6 miles wide, suggesting a small mesocyclone, rather than a non-supercell tornado (rotational signatures from non-supercell events are very small and often not seen at all on radar). The mesocylone was shallow, only about 5 thousand feet deep, which is about half of what is typically observed. The strength of the signature could only be considered moderate at best for a classic storm, but given the small nature of the cell and taking into account the circulation diameter, the values were of tornadic magnitude (rotational velocity of 35 knots - shear 0.033 1/s).
This event was unusual from many aspects and will be investigated further. Often times its the large tornadoes that get studied, but its the small events like this one that are the most difficult to predict and warn for.