In late June of 2013 I ran a 3-post series on the frequency of tornadoes in the United States and in Missouri (starts here). Statistics show a marked increase in the total number of tornadoes, but I concluded that the increase derived from improved detection, which was catching many small ones that had previously escaped notice.

Source: Tornado Climatology page of the National Climate Data Center

Regarding the most powerful tornadoes (EF 3-5), I said “…it seems pretty clear that there is no overall trend toward increasing frequency. In fact, if anything, the trend has been towards decreased frequency – there are fewer highly destructive tornadoes now than there were during the mid-1960s.”

Click on chart for larger view.

Now a post in Andy Revkin’s Dot Earth blog takes up the issue. Richard Muller, the physicist and climate change denier who gained national attention by conducting an independent analysis confirming that humans were warming the planet, wrote an op-ed piece saying what I said: there are fewer of the most destructive tornadoes. Paul Markowski and five others wrote a reply claiming that the apparent decline comes from a tightening of reporting standards over time. Tornado frequencies prior to the 1970s were created retrospectively by meteorology graduate students, and suffered from “grade inflation.” In 2003, new policies required additional (and expensive) analysis of every severe tornado. According to Markowski et al, this led to the systematic and intentional under-counting of severe tornadoes after 2003 in order to avoid increased costs. Markowski and his colleagues are meteorologists at several different institutions.

Errors plague all attempts to develop historical climate data. The basic thrust of the scientific process is that a hypothesis is produced, then somebody criticizes it. If the criticism is valid, a new and better hypothesis wins adoption. Then a new criticism arises, followed by more improvement.

It seems to me, however, that systematic and intentional under-counting is a fairly serious accusation, roughly equivalent to falsifying data. Although Markowski et al seem positioned to be “in the know,” they make the accusation without citing any supporting evidence. If the NOAA data is intentionally under-reporting tornado severity, it needs to be proved and corrected.

For now, I stick with my original statement: for whatever reason, the data show no overall trend toward increasing frequency of severe tornadoes. If anything, the trend shown in the data is towards decreased frequency. The data may need to be corrected, but for now, this is what they show.

Sources:

The MoGreenStats Blog. “Tornadoes In and Around Missouri,” blog entry by John May, 6/24/2013.

The MoGreenStats Blog. “Tornadoes Seem More Frequent, But They Aren’t,” blog entry by John May, 6/29/2013.

The MoGreenStats Blog. “The Most Powerful Tornadoes Are Less Frequent,” blog entry by John May, 7/1/2013.

Revkin, Andrew. 2013. “A Closer Look at Tornadoes in a Human-Heated Climate.” Dot Earth (blog), New York Times, 12/9/2013, http://dotearth.glogs.nytimes.com/2013/12/09/a-closer-look-at-tornadoes-in-a-heated-climate.

Muller, Richard. 2013. “The Truth About Tornadoes.” New York Times, November 20, 2013. http://www.nytimes.com/2013/11/21/opinion/the-truth-about-tornadoes.html.

Markowski, Paul, Harold Brooks, Yvette Richardson, Robert Trapp, John Allen & Noah Diffenbaugh. 2013. “A ‘Truthier’ Interpretation of ‘The Truth on Tornadoes.’” A letter published in “A Closer Look at Tornadoes in a Human-Heated Climate.” Dot Earth (blog), New York Times, 12/9/2013, http://dotearth.glogs.nytimes.com/2013/12/09/a-closer-look-at-tornadoes-in-a-heated-climate.

U.S. Tornado Climatology, National Climate Data Center, http://www.ncdc.noaa.gov/oa/climate/severeweather/tornadoes.html#history.