Investigation of Small Tornado Event on December 31st, 2021 (author: Jacob Hinson)

To ring in the final day of the year, some interesting elements came together to make an interesting day for some people in the Southeast. A few sporadic cells came together and managed to find enough shear (change in wind speed/direction as you go up in the atmosphere) to initiate tornadic conditions.

Fig. 1 – Radar loop from 1755Z to 2325Z (12:55pm to 6:25pm EST) on December 31st, 2021


In Figure 1, we see the radar loop of the Southeast during our region of interest. The storms that produced the most damage were within the “marginal” risk category as provided by the Storm Prediction Center (Fig. 2) at 20Z (3pm EST).

While this risk generally means there is low change of anything happening, it does not exclude it entirely. This day definitely exemplified that case, as there are limitations to forecast ability as well as representation to the public about severe weather threat.

Fig. 2 – SPC Day 1 Categorical Outlook for 12-31-21 at 20Z


Another thing to consider when looking at the forecast for these storms is that severe weather is not typical in the winter months for the southeast. Some tornadoes may occur during early winter in the rust belt, but the southeast can generally expect to stay free of them, making an event like this even more dangerous.

Fig. 3 – SPC Storm Reports for 12-31-21


There were a total of 4 tornado damage reports on the 31st, one of which garnered a damage survey by the NWS. The damage was categorized as EF-0 to EF-1 on a scale from EF-0 to EF-5, with winds of around 70-100mph.

Fig. 4 – Damage from early in the Carroll county tornado (EF-0) (Credit: NWS)


Despite the low rating, these events can still manage to destroy people’s lives. Homes can be destroyed by trees or water, and someone’s car could be their only form of transportation to their job.

Fig. 5 – Tornadic damage from the end of the Carroll county tornado (EF-1) (Credit: NWS)


The tornado lifted a carport off of its foundations and collapsed it on top of itself.

Fig. 6 – Damage path from the Carroll county tornado
The tornado was on the ground for about 2.5 miles, doing EF-0 and EF-1 damage the whole time.


Fig. 7 – Screengrab from GRLevel2 of Carroll county cell further along in its life (Credit: Sammy Maggio)


The WSR-88D in Peachtree City had a good look at the storm as it approached Marietta, GA. In the top right the dual-polarization product Correlation Coefficient is picking up that are not uniform in nature, alluding to debris being lifted and thrown around. While this data is low resolution, you can still see the velocity couplet in the bottom left image. The bottom right image shows Spectrum Width, a more advanced dual polarization product that compares the range of wind motions that are observed. It increases, becoming gray, then orange, then yellow to white as it sees higher turbulence. This is telling us that the environment around the storm is not just wind flowing in one direction, but wind twisting around in the atmosphere.

Fig. 8 – RadarScope Super-Res Reflectivity and Velocity scan (Credit: Jacob Hinson)


The typical supercell structure with a well defined hook echo cannot be identified, but thanks to doppler radar, we know that the storm is rotating. On the bottom of Fig. 8, we can see a velocity couplet, identified by wind coming toward the station (green) and going away from the station (red) right next to each other (just above Powder Springs). This wind can be measured as “gate-to-gate velocity” to get an idea of how fast the storm is rotating. These measurements are not the best analog for windspeed at the surface however. Most radar beams are curving away from the surface of the Earth. At the current 0.5 degree tilt, the radar is imaging the storm at 2700ft above the surface. Inferring tornadic strength from radar imagery can be a slippery slope, especially for a fringe case such as this where the parent storm hasn’t even reached a 50dBZ reflectivity return.

Combined with the information from Fig. 7, we can tell that the tornado that this had produced was lifting small debris at least that high. I had started to record the storm and alert friends at Kennesaw State University of its arrival. You can see KSU as a blue dot on both of the radar images.

Fig. 9 – Sounding from KFFC (Peachtree City) on 12-31-21 at 12Z (7am EST)


The atmosphere did not have much instability to create such an event, so much that the NWS even discredited the chance for severe potential in their mesoscale discussion #2089. However, even with surface-based CAPE (Convective Available Potential Energy) at a meager 5 J/kg¬3 (meteorologists are only ever paying attention to values above 1000 J/kg¬3), the storm found enough shear to become unstable. Shear can help a storm begin to rotate and form a supercell structure, and this can come from change in wind direction with height, or simply change in wind speed with height. Evidently, there was enough shear this day to cause the initiation of tornadoes, which is why everyone needs to stay vigilant and have a way to get alerts. The storm that produced these tornadoes was captured by Stu Ostro in the Tweet below.