UPDATE: Thanks to the comments below and on Facebook, I now have a link to the VORTEX2 "tornado eye" radar image I was looking for, and a note that Brad Panovich once blogged about the vortex hole -- only his may have really shown the eye of the tornado, because it was from higher-resolution TDWR data.

Researchers from the VORTEX2 project (which used real-time radars that were not "operational" - meaning that it doesn't transmit 24/7) have noted a new component to a supercell thunderstorm called a "low reflectivity ribbon (LRR)." The LRR, although recently noted by the VORTEX2 project in a 2009 radar image, was associated with radar data [PDF] as early as 2008 (examples shown below).

It is what it looks like -- a weaker (lesser/smaller raindrops) area of the storm, in the middle of the heavy rain. Researchers are unsure what relationship, if any, it has to tornado formation, but it's neat that we are discovering more about thunderstorms with this newer technology. I believe that I may have spied an LRR on the Vance Airforce Base NEXRAD radar in Oklahoma on May 1st (image below):

But here's what's even more interesting: Brett Kreais, a meteorology student at Central Michigan University, posted a radar image from the same radar, and same storm, which appears to show the "eye" of a tornado -- something that I believe I have seen on high-resolution non-operational radars (although I can't find a link at the moment). More typically you would look to confirm a tornado by observing a debris ball on NEXRAD.

I pulled some more data from the storm and took a couple of 3D screen shots. Not surprisingly, it is co-located with the Doppler velocity "couplet" (winds moving fast towards, and fast away from, the radar). Is this really the "eye" of a tornado? It looks like it at first glance. Perhaps readers (who are more experts at interpreting radar than I) will comment on my story, but I believe it could simply be drier air entrained as the mesocyclone began to rotate itself out of existence (see radar loops - reflectivity & velocity).

Below are 2-D and 3-D shots of the same data that Brett showed:

The 2-D cross section does look like a cross-section of a tornado...

And when you look down the reflectivity "hole" from above, it seems plausible you're looking down the eye of the tornado:

But perhaps the most important thing to remember is that, on the velocity data, there is no "calm" area. I'd like to show the image below with a lot of caution, because maps like these often get misinterpreted. What you're looking at here is rotation towards and away from the radar (red vs. green), but you're only looking at speeds lower than a certain amount, so you can't look at the space between green and red and say "there's the tornado" because we arbitrarily "blanked out" the higher wind speeds with the GRLevel Analyst software. You also can't point to the red or green tubes and call them tornadoes, because the wind is only going one direction in those areas.

What you are looking at is a massive, strong circulation that reaches the ground (or confirmed by spectrum width and rotation), but the radar doesn't have the resolution to show the tornado; you're probably seeing the mesocyclone. So why does the middle of the mesocyclone appear empty? It's not completely empty, or the velocity data wouldn't show up there. As I said above, it may just be drier air that was wrapped into the storm.

Report a Typo