The tornado outbreak in Oklahoma on May 20, 2013 occurred on the last day of a series of three consecutive days with significant severe weather. On this day, the most deadly tornado of the year for the United States occurred in Moore, Oklahoma. The tornado that tracked through Moore had developed near Chickasha, Oklahoma and then tracked east-northeastward across Moore and south Oklahoma City. The tornado was on the ground for nearly 40 minutes, resulting in numerous injuries, billions of dollars in damage, and 24 deaths.
The model data was also very impressive the morning leading up to the event. By 6AM, instability was beginning to increase and a dryline was beginning to take shape along the Oklahoma/Texas border.
By 6am, a developing dryline was apparent across Western Oklahoma. Warm temperatures and high dewpoints suggested that instability was already very high.
Throughout the morning, southerly winds began increasing. These southerly winds allowed additional moisture to be advected northward. In addition, temperatures warmed rapidly due to daytime heating and at the mid levels, some cold air advection was occurring. All of these effects contributed to the destabilization of the atmosphere, which was evident in the 12pm NAM analyses.
Temperatures rose quickly through the morning across the state. In Central Oklahoma, temperatures rose from the lower 60s at 6am to the lower 90s by noon; dewpoints rose from the upper 60s in the morning to the lower 70s at noon. Both of these increases, as well as cold air advection at the mid levels, helped to rapidly increase the CAPE to 3800 J/kg by noon.
The upper air sounding from Norman, OK on the morning of the 20th already indicated a high degree of instability and shear in place. As the morning progressed, daytime heating and moisture advection contributed to further destabilization, as can be shown by the increase in CAPE and the steeper lapse rates in the sounding taken at 1pm.
By noon on the 20th, a dryline was beginning to move across Western Oklahoma. Just ahead of the dryline, low level moisture and CAPE was increasing. By early afternoon, the dryline was approaching Central Oklahoma and convergence ahead of the dryline was increasing. This instability, coupled with the convergence ahead of the dryline, allowed thunderstorms to begin developing. Once thunderstorms developed, they developed very rapidly and began to move northeast.
The vertical shear causes vorticity (which is essentially the amount of spin in a fluid) to be created. Once created, low level convergence acts on the vorticity to “stretch” it, producing a tornado. The convergence itself doesn’t create vorticity, but amplifies existing vorticity. Other effects can (de)amplify vorticity, however, in the context of this discussion, stretching is the most relevant effect. As the tornadoes developed, they began to move northeast into Moore and Oklahoma City. The tornado touched down in Newcastle, Oklahoma at 2:56pm CDT and tracked northeast into Moore and Oklahoma City. To show another view, here’s a visible satellite image of the storms.
The area impacted most on its path was the Moore area. We took some radar imagery from when the storm impacted the Moore area to show some of the features of the storm as observed by radar.
This is the radar scan at the time that the tornado began to move into Moore. The storm has the classic supercell structure with the largest portion of the storm to the north of Moore and the hook echo approaching Moore. Below is a schematic of the structure of a Classic Supercell.
At the time of the Base Reflectivity image above, the strongest shear was located to the southwest of Moore and it continued to track northeastward into the city of Moore.
One of the advantages of the dual polarization upgrade that was completed a few years ago is that meteorologists are able to discriminate between meteorological and non meteorological echos in radar data. Unlike traditional weather radar, dual polarization radar transmits not only a horizontally polarized pulse, but also a vertically polarized pulse as well. The correlation coefficient product compares the similarity in the power returned in each of the pulses. The closer the correlation coefficient is to 1, the more similarity there is between the two signals. Meteorological echos usually have correlation coefficient values greater than .8. Values less than .8 indicate echos that are possibly non meteorological. In the context of supercells, low correlation coefficient values in the vicinity of the hook echo of a supercell indicate possible debris being detected by the radar. In fact, correlation coefficients in the hook echo were around .54, indicating the likelihood of debris. This type of signature is known as a Tornado Debris Signature. Some studies have suggested a connection between the height of a tornado debris signature and the strength of a tornado.