Yes, The Tornado Warning Was Warranted - BUT

I've was contacted by two readers in the Kansas City area this morning asking if their phones triggering in the middle of the night for a tornado warning at Lake Lotawana, Missouri, was warranted. They were awakened from sound sleeps and went to the basement.

While that was certainly inconvenient, yes, the warning was warranted. [But, there is more to the story. A critical piece of information came in while I was writing this piece and I thought it might be helpful to readers to see how that data tipped the scales.] 

I'm writing this because storm warnings are certainly in the news yesterday and today. Let's begin with the fierce derecho that affected Pennsylvania and New Jersey at this time yesterday. More than a million people lost power and there were a number of deaths. Watch this heroic policeman assisting people when the derecho reached Atlantic City's beach. A tree falls and nearly hits them!
The obvious question: Given three hours of notice, why were they on the beach in the first place?!

A sliver of the answer may be found in a comment to an article posted this morning at another weather site:

In spite of incredible progress in the field of warning of these storms the last twenty years, meteorologists still have to deal with,

• I can get a better forecast from the "Farmer's Almanac", or,
• I wish I could keep my job while being wrong half of the time, or, 
• There are too many false alarms.

The last complaint has real validity and, you may be surprised to learn, most meteorologists would agree. An unnecessary "false alarm" at night would be a real negative as I doubt people would go to the basement a second time for a future nocturnal tornado warning, once they felt we had let them down at 3am. 

So, I looked up the radar and other information and came to exactly the same conclusion as the Kansas City office of the National Weather Service: the tornado warning was exactly what they should have done. [see more below]

Here's the pertinent radar information. The first indication of a tornado appeared in the Doppler wind data (right) at 2:21am. Please click to enlarge. 

An area of weak rotation appeared (circled). However, there is no indication of a tornado in the reflectivity data (left, the type of radar that is usually shown during television weathercasts). 

At 2:23am (kudos to the NWS in KC for having their radar set up to update frequently), strong rotation appears in a tiny area (right). 

At this point, the meteorologist is in a terrible situation. There is still no indication of a tornado in the reflectivity data (left) and the rotation has strengthened so quickly it may be a radar artifact -- not representative of the atmosphere's true state. This happens from time to time. The meteorologist also knows that a false alarm at 2:23am is going cause all kinds of trouble*. So, he/she decides not to issue a tornado warning. Since there is no indication of a tornado in the reflectivity data, I would agree. 

2:24am: As suspected, the strong rotation may have been incorrect sensing by the radar. No significant rotation is seen at right. But, now, the reflectivity data (left) is showing a suspicious pattern that may indicate a tornado. 

Oy! The meteorologist holds off. That is certainly a defensible position. 

2:26am: The rotation is back (right) but the reflectivity (left) is less impressive. 

One could go either way at this point and have it be a defensible decision. During the day, I probably would have warned on this. At night, especially since it was not a typical tornado situation, I probably would not have. 

2:28am: The rotation is what I would call moderate, but there is still no tornado signature in the 

reflectivity data. Regardless, given the persistence of the rotation, I probably would have warned at this point. 

2:29am. The reflectivity data again shows the potential for a tornado. 

2:30am: While broader, the rotation is pretty clear at this point. This is the data upon which the NWS issued its tornado warning. 

I had gotten about this far in writing this piece when the NWS issued its damage survey. They go out into the field and survey the potential path of a tornado (if any) based on any path inferred from radar. They learned there was a very weak tornado. 

So, there was a tornado. And, there was a tornado for a minute or two (perhaps) after the tornado warning was issued. This will go down in the books as a successful warning, as it should. 

[I now have the info that there was a tornado] But there is one more thing to consider: The tornado was on the ground for several minutes before the warning was issued. I didn't know that until I was nearly finished writing this piece. So, I went back to the radar data to see if there was a "debris signature." I was almost certain there would not be because of the weak intensity of the storm. 

Yet, there was. At 2:24am, it is quite evident. It is the blue-gray area surrounded by magenta. Since the tornado was initially in a heavily wooded area (see green tint to map above), it is likely tree limbs, leaves, etc., that are producing the damage signature. With this additional piece of the puzzle, it is clear the tornado warning should have been issued at 2:24am. 

So, no, I now cannot say this tornado warning was a success.

*Still, I have considerable empathy for the meteorologists who were put in this fix by Mother Nature. 

If they had issued a tornado warning at 2:30 in the morning and no tornado had occurred, politicians would have been upset. The newspaper may editorialize. People will swear off going to the basement ever again. Yes, all of this really occurs in these situations. Believe me, I have counseled meteorologists upset with themselves in these situations on more times than I can count over my 47-year career.

Before Doppler radar, there was zero chance of warning of these storms. We could only warn of this particular storm because it was close to the radar's location in Pleasant Hill, Missouri. There are many meteorologists who believe we shouldn't try to warn of this type of tornado (known as a "QLCS Tornado") at all. I disagree, but understand their point of view. 

Still, weather science must go where the data takes us. Since there were no injuries and no significant damage, I am hopeful we will learn and do better the next time.