The radar rotates around 360 degrees and scans a volume of air 1 degree wide, 1 deg tall, and 1 km long. So the scan volume varies with distance from the radar, but let’s call it roughly a one kilometer cube. The closest radar for this event was KSLX in St Louis, but other radar in Missouri and Kansas could also see the same area.
The radar scans the lower 1 deg of space, then rises up and scans the next 1 deg of space, and continues for 7 degrees for this event. The scan volume changes for the radar mode, but in this case, the radar was in clear air mode. For the lower 3 degrees, the radar scans once around for reflectivity and once around for velocity. The reflectivity scan takes 55 seconds, and the velocity scan takes 30 secs. The radar does a short calibration run ,that takes a few degrees of angle, so it starts the scan at an arbitrary azimuth.
We have the event end time from a video as 2:51:53 UTC, and it takes a minimum of 60 seconds or so for rocks to fall from the top of the stratosphere, at 18 km, to the altitude the radar scans. So we see a radar scan that started at 2:52, which is just about perfect timing for this event.
So, it starts with a scan at 0.5 degrees reflectivity mode, and 0.5 degrees velocity mode. These modes are very cluttered due to ground returns and are not very useful. The area of interest is about 277° azimuth from the radar, so the 1 degree reflectivity scan. 0.88° actually started at an Azimuth of 265°, which is perfect. The scan is still early in timing, since it scans the 277° azimuth at about 69 seconds after dark flight starts. So, hits in this scan should be the early larger rocks.
The radar is calibrated to 1-millimeter-sized droplets of water, for every 1 cubic meter. A small amount, such that a light cloud, which is many small droplets every cubic meter, shows up as about 5 dB. Every 3 dB doubles the amount of droplet volume. So, for a single rock to register on the radar it has to be bigger than the equivalent of 1 mm drops every square meter, over a volume about 1 x 1 x 1 km. Or, you can have two rocks, half the size… or a few hundred rocks of pebble size. The scale is not linear… but you get the idea.
The rocks in clear air radar cross section (RSC) must be larger than the 1 mm drop standard. We do get some returns in the 0.88 deg scan on the west side of the fall zone, which would correspond to large (> 1kg) rocks, which should be on the far west side of the fall zone… so everything is consistent. It does not mean the hits are meteorites, but if it were a cloud it would likely show up in the radar scans 9 mins before or 9 mins after this scan. Clouds tend to last thru multiple scans. These hits are only in this 0.88° reflectivity scan. A good sign.
The next scan is the 0.88° velocity scan. Unfortunately, it starts at about 285° azimuth, so it does not look at 277° until the end of the scan, which is too late for the larger rocks, and too early for the smaller rocks. So, the timing of this scan is not favorable.
The 1.5° scan, actually 1.36°, scans the 277° AZ 359 seconds after dark flight, and has a nice block of hits in the fall zone, that should correspond to rocks in the 30-50 gram size. We continue looking at the scans and find a few more hits moving eastward, which is consistent with the winds, and a good sign. This gives us some confidence the hits could be meteorites. One bad sign is that other radars that can scan the area have no returns that match the St Louis radar. There could be explanations, such as the the rocks just do not have large enough RSC at the longer ranges, to register. One sign of this is the hits on KSLX are low dB. So, we get multiple scans consistent with timing and winds in the fall zone, which gives us some confidence, but no confirmation from other radars.
For a Google Earth importable KMZ file, including all the Doppler data for the St. Louis event, visit the main event page here: Meteor Event – St. Louis