Thursday, May 23, 2013

More Than 200 Lives Likely Saved By Moore Warnings

Update: Sunday, 5:55pm. According to the New York Times,  there were 1,000 homes destroyed and 1,200 damaged (h/t Dr. Kevin Kloesl) and the initial report (see below) of the number of homes damaged was an order of magnitude high.

Using the methodology below, a comparison would Joplin would yield 45 deaths in Moore. Because there were 24 deaths, weather science approximately cut the death toll in half over what would have been expected. That is a substantial achievement!

IMPORTANT: The numbers are undergoing revision and it appears the numbers I was originally given for Moore are too high. See comments. When the numbers are final, I will revise this posting.


For the purpose of investigating warning system effectiveness, the Moore and Joplin tornadoes are a rare “apples to apples” situation. Because they were so similar, we can draw some conclusions about the warning system's effectiveness in Moore.

Let’s begin with some basic statistics:

Population of Moore:  55,000
Joplin (before the tornado): 50,000

Homes damaged in Moore:  12,000 to 13,000*
Homes damaged in Joplin: 7,850

Homes with Basements in Moore: Nearly Zero
Homes with Basements in Area of Joplin Tornado: 30%

Was a Business District Affected in Moore?  Yes, I-35 Corridor
Was a Business District Affected in Joplin?:  Yes, Range Line Rd.

Time of Day for Moore:  Daylight in May
Time of Day for Joplin: Daylight in May

Intensity of Tornado in Moore: EF-5
Intensity of Tornado in Joplin: EF-5

Based on the above statistics, one would surmise there should have been more fatalities in Moore than in Joplin, especially since about 30% of the homes damaged by the Joplin tornado had basements, the most effective type of shelter. Yet it turned out the reverse was true and by a large margin. 

Number of Deaths in Moore:  24
Number of Deaths in Joplin: 161, worst in the tornado warning era

Why? The relative effectiveness of the warning system for those two tornadoes.

As a rough basis of comparison, using a number of 12,500 homes damaged in Moore versus 7,850 in Joplin, yields 59% more homes destroyed in Moore. Since Joplin had 161 fatalities X 1.59 (larger number of homes affected) would predict 256 deaths in Moore due to the greater exposure.

Those who have read When the Sirens Were Silent know all of the things that went wrong with the warning system as the Joplin tornado formed, approached, and moved across the city. 

Moore was a textbook example of how the warning system should perform. Given a predicted death toll of 256 and subtracting the tragic number of 24 deaths, approximately 232 lives were saved by the warning system.

We mourn for the loss of life in Moore and pray for a quick and complete recovery for those who were injured. The only solace we might take is that the result of the Moore storm could have been so much worse. 

* Thanks, Keli Tarp. 

Addition: The Norman office of the National Weather Service has a rundown of all of its tornado forecasts and warnings here


  1. Interesting, but wouldn't there be a big difference in the number of people in their homes given Joplin was on a Sunday and Moore was a workday?

  2. No. Note that I highlight that a business area in both cities was destroyed.

    The people killed in businesses in Joplin was the highest percentage of any for which we have data. So, no, more people were not killed in homes in Joplin.

    Even if there were a slight difference, it does not account for an 85% difference in deaths.

  3. Has there been an analysis of deaths by address/locale/institution/structure type/etc. in Joplin? Will there be a comparable analysis for Moore? (Perhaps done for Moore in the earlier F5?)

    Pure speculation on my part, but I would not be surprised to learn that the "typical citizen" in Moore was/is much more attuned to the reality of tornado risk than the "typical citizen" was in Joplin. Another research topic.

  4. Joplin is very much in tornado alley and had previously experienced fatal tornadoes and wind storms.

    There is a point I wish to emphasize: 30% of the damaged homes in Joplin had basements (the best kind of shelter) and none of the Moore homes had basements. That would lead one to conclude the higher death toll would have been in Moore. But, it was the reverse.

    Please read: When the Sirens Were Silent. The softcover is sold out. The ebook is only $2.99.

  5. It appears that the Moore and Joplin tornadoes parallel one another in much the same way that the Greensburg and Udall (Kansas) tornadoes cited in "Warnings" do -- even down to the same numerical estimate of lives saved by timely warnings (232). It would be interesting to see radar images of the two supercells compared as you did in "Warnings" for Greensburg and Udall.

    I suspect also that the Moore tornado might suggest an answer to the "reverse analysis" question I have posed before -- how many people would NOT have died in Joplin had the warning system worked exactly as it should have. The answer: 137 (161 - 24).


  6. I just saw that they revised the damaged and destroyed home total in Moore from 12,000 to 1,200. That will make a big difference in the stats of this blog post.

  7. Yes, Shawn. However, that is Moore only. I'm waiting for the numbers from Draper Lake and Newcastle to revise the post.

  8. The person who told me my number of homes damages was too high promised to get me the final number by the end of the day (Friday). Nothing. So, I'm leaving the post alone at this point.


Note: Only a member of this blog may post a comment.