Scientific American Needs to Do a Better Job at Communicating Science

A report in Scientific American discusses summer heat in Madison, Wisconsin:

To [Jason] Schatz [from the University of Wisconsin in Madison], the summer of 2012 exemplified temperatures Madison and the rest of southern Wisconsin could see more frequently in July and August by the middle of the century.

That summer, Madison had 39 days above 90 degrees Fahrenheit. The annual average is usually 12 days over 90, according to National Oceanic and Atmospheric Administration data.

Schatz cited a 2011 report by the Wisconsin Initiative on Climate Change Impacts that predicted the southern part of the state will likely experience about 25 days above 90 F per year by 2050. Northern Wisconsin, which usually experiences about five days over 90 degrees, would likely exceed 90 F 12 times per year by midcentury.

Using the NOAA database, here are the number of days at or above 90 F for Madison during the past century.

Yes, Madison did have “39 days above 90 degrees Fahrenheit” in 2012.  Actually, if climate scientists want to be precise, “above 90 degrees Fahrenheit” means “>90º F” (i.e., greater than 90º F), not “>=90º F” (i.e., greater than or equal to 90º F), which does make a difference in using the NOAA database.  Regardless, I found the correct dataset that yields “39 days” in 2012, and it is the greater than or equal to 90º F data.  It doesn't matter, because both the “>90º F” and “>=90º F” datasets tell the same story.

Namely, there has been absolutely no hint of a significant increasing trend in the number of “days above 90º F per year” for Madison during the last 100 years.  In fact, the correlation is slightly negative toward fewer of these hot days – not more.  Since 1970, the decline in the annual numbers of these hot days has been almost statistically significant.  There is also a negative correlation toward fewer hot days over the last three decades as well.  This seems to be an inconvenient truth.

The only rational conclusion is that the number of “days above 90º F per year” is not increasing.  Actually, on the balance of probabilities, the number is decreasing.  I missed that basic fact in this Scientific American article.

The article ends with this delight:

Although the summer of 2012 was the second hottest since 1939, according to NOAA, it was an 'isolated warming incident.' The summer was the only time in the 21st century that Madison experienced temperatures exceeding 100 F. The three hottest days happened in a single heat wave between July 4 and 6, with highs reaching 102 and 104.

Only time in the 21st century?  You mean all 14 years of it so far?

Yes, there were three days in July 2012 above 100º F in Madison.  Yes, these were the only three days in the 21st century above 100º F in Madison.  But here is a factoid to consider: the record number of days above 100º F in Madison during July didn't take place in 2012.  Rather, these days occurred in July 1936 – when there were four days above 100º F.  And there have been no days above 100º F since July 2012.

Since Scientific American is a prominent publication, another chart is in order.  This one shows the number of days exceeding 100º F each year in Madison over the past century – including, of course, the aforementioned hot year of 2012.

All those overlapping circles at the bottom, which extend right up to the present, are years with no days exceeding 100º F.  The record year was back in 1936, when there were five days.  Only two years earlier, in 1934, there were an additional three days (the same as in 2012).  Thus, in only three years during the 1930s, Madison saw eight days exceeding 100º F.  For perspective, the region has had only eight days >100º F in total since 1978.  Of course, there is certainly no trend in days exceeding 100º F in Madison over the past century.  Strange how the article didn't mention that fact, either.

This Scientific American article is another textbook failure in properly conveying and contextualizing science information to the public.  Stuff like this shouldn't be published.