Lab 12: Revisiting the 2013 El Reno Tornado with Satellite Observations - Cassidy Johnson

Introduction

On May 31, 2013, the widest tornado ever recorded wreaked havoc in El Reno, OK. It touched down at 23Z and was on the ground for 40 minutes. 

Figure 1: El Reno supercell on 31 May 2013. Photographed by Brennan Joseph/From NOAA's El Reno report.

Using UCAR's satellite imagery archive, we can see the visible satellite imagery from 17:45Z to 23:45Z in figure 2. At 21:55Z, there are clearly some overshooting convective tops in central Oklahoma, which are a pre-cursor of severe weather. When the tornado touched down at 23Z, the convective top is so large that it is covering nearly the entire state of Oklahoma. We are also able to see some of the classic lumpy, cauliflower-like texture in the center of these convective clouds. As we learned in class, this is a structure that can be seen in satellite imagery when trying to identify severe weather.

Figure 2: Visible satellite imagery from 17:45Z to 23:45Z. From UCAR-MMM satellite imagery archive.

The area that we can see the largest convective plumes are coincident with the hook echo that was prominent on the radar reflectivity. The hook echo is located in Canadian County, OK which is in central Oklahoma, which is also where we see large overshooting cloud tops.

Figure 3: Radar reflectivity of the El Reno supercell from 23:00Z to 23:46Z on 31 May 2013.

For this lab, I will be using Worldview to analyze the El Reno tornado, and I will also discuss some limitations of using Worldview for mesoscale storm forecasting.

Worldview Satellite Analysis

NASA's Worldview provides daily satellite images using Aqua, Suomi NPP, and Terra. Aqua provides an image over Oklahoma at 1:30pm CDT (18:30 UTC) while Suomi NPP and Terra provides images over Oklahoma at 10:30am CDT (15 UTC). Figure 4 below shows Aqua's daily image over Oklahoma on May 13th, 2013 at 18:30 UTC. One obvious limitation that we can see right away is the lack of any overshooting tops or deep convection over Oklahoma. This image was taken 4.5 hours before the infamous El Reno tornado touched down, so it makes sense that no deep convection has occurred yet. Aqua's satellite imagery is not very useful when it comes to forecasting tornadic events, unless the tornadic events occurred soon after 1:30am/pm local time (LT).

Figure 4: NASA Worldview Aqua visible satellite imagery at 18:30Z on 31 May 2013.

Although the time is not perfectly lined up with the highest tornadic activity time, I decided to take a look at the cloud top temperatures through the IR channel. I was expecting that there would be at least some cold cloud top temperatures over Oklahoma at 18:30Z (4.5 hours before the main tornadic event), and this is exactly what we can see in figure 5. As we learned in class, cold cloud top temperatures correspond to deep convection. So, in the image below, we can see that there are some points that have colder cloud top temperatures, and these areas eventually did turn into much deeper convection. Again, not being able to have satellite imagery at the exact time of the deepest convection is a big limitation in forecasting using satellites.

Figure 5a: NASA Worldview satellite cloud top temperatures imagery at 18:30Z on 31 May 2013.

Figure 5b: NASA Worldview satellite cloud top temperature scale.

To further illustrate the limitation of Worldview, I decided to look at the IR satellite imagery from UCAR's satellite imagery archive for a better idea of the cloud top temperatures. Obviously, as we can see in figure 6, the coldest cloud tops start forming at 21:32Z and are very visible by the next timestamp at 22:02Z. At 22:32Z and 23:02Z, the coldest cloud tops are visible as white dots in the middle of the dark blue and purple colors. It makes sense that the coldest cloud tops occurred during the peak of the tornadic activity because, like I said before, we learned in class that deep convection is associated with colder cloud top temperatures.

Figure 6: Infrared satellite imagery from 17:45Z to 23:45Z. From UCAR-MMM satellite imagery archive.

Finally, I decided to take a look at the precipitation estimate, which is shown in figure 7. Throughout this supercell's lifetime, it was clear that the most precipitation was recorded during the June 1, 2013 evening observation. The highest precipitation estimates occurred in Southeastern Kansas. In Oklahoma, the highest precipitation estimates were recorded at around 20mm. It is surprising to me that there are not higher precipitation measurements in the vicinity of El Reno tornado, and that the highest precipitation measurements didn't even occur in Oklahoma. Perhaps this is showing us that tornadic activity is not necessarily synonymous with high rainfall totals. The highest rainfall appears to have occurred after the peak tornadic activity, when the storm moved eastward.

Figure 7a: NASA Worldview satellite precipitation estimate imagery at 18:30Z on 31 May 2013.

Figure 7b: NASA Worldview satellite precipitation estimate scale.