Too Hot in Washington: A Climate Mystery?

After the purportedly all-time record high May temperature was reported at D.C.’s Reagan National Airport (the “official” Washington weather station), I noted something very funny with its temperatures. When 2015’s monthly temperatures are stated as departures from the 1981-2010 average, which is the standard reporting procedure, the departures from average at Reagan (DCA) were always warmer than those at Dulles Airport (IAD), about 25 miles to the northwest.

Note that the departures are from the average at each location. So when DCA reports a temperature of three degrees below average, that is three degrees below the average at Reagan. Similarly for the departure at Dulles. It’s compared to the average at Dulles.

So, we’re not talking about raw temperature here. Of course DCA is going to be intrinsically warmer than IAD. It’s several hundred feet lower in elevation and its being additionally heated by the bricks, buildings, and the pavement of urban Washington, as well as the waste heat from all your money changing hands. What is interesting is that the departures from normal, or “anomalies,” at DCA were all running hot compared to the departures from normal at IAD.   

For example, last month was 1.7°F above the average at DCA, while Dulles was 0.7°F below its average, for a difference of +2.4°, where the expected difference should be near zero. The record hot May departure from normal was 1.5° greater at Reagan than Dulles. Adjusting the Reagan temperature down to where it should be would take it out of the record books.

I noted this to Jason Samenow, a former student of mine and a great forecaster who runs the Washington Post’s excellent Capital Weather Gang (  He agreed something was weird, and I then set out to look to see if there was some point in time where this obvious error suddenly started. I soon found it, in January, 2014.  Since then, DCA’s monthly departures from normal have averaged 2.2°F warmer than Dulles.

As I was about to write this for Cato@Liberty, Jason and his crew published their take on it–a worthwhile and informative read. In addition to what I had found, they discovered that there was a similar anomaly difference from when IAD opened up in 1962 through the early 1980s. That makes the current era look not-so-unique.

Let us all praise the Interwebs, for someone soon alerted Jason to this document: “Temperature Irregularities at Washington National Airport during the Heating Season: 1978-85.” It was published in the journal of the National Weather Association, a smaller competitor to the more widely known American Meteorological Society. One of the authors, the late William (Bill) Klein, was one of the world’s most accomplished scholars in the field of statistical meteorology.

The paper found that the instrumentation at DCA back then was drifting warmer and warmer, in addition to the expected urban warming. When it was replaced in 1985, the anomaly differences largely stopped—until 2014.

So there’s some reasonable evidence that DCA is reading way too hot, at least for monthly departures from normal. The question is, why has it taken so long to be noticed? There are 6 million people in the metro area, and, judging from the traffic on Jason’s site, an awful lot of weather nerds. Maybe people weren’t looking because they didn’t want to know. As it stood, the DCA record was providing wonderful “confirmation bias” to a city obsessed with global warming—so if the DCA temperatures are warming like crazy, that’s consistent with the belief system of the inhabitants. 

The sudden warming of the temperature anomalies at DCA compared to the Dulles anomalies could have many causes. There might have been a coding (programming) error inserted in January 2014, related to the fact that “normal” temperatures are adjusted every ten years to reflect the average for the last thirty. The new normals, for 1981-2010, were published in November, 2012, when they replaced the 1971-2000 values.

There’s also evidence, shown to me by Ian Livingston, one of Jason’s associates, of a completely unrealistic change in the raw temperature (as opposed to departures from normal) difference between DCA and IAD. That figure changes a remarkable 4°F in the last seven years. I know of no way a local record could warm that much in such a short time, compared to a nearby neighbor. 

Then there’s the number of days with a temperature of 90°F or above. Normally there are about seven more at DCA per year than at Dulles.  So far this year, the difference is 26 days. Livingston has also looked at this and it’s apparent that when this difference is unrealistically large, like it is this year, the departure from the local normal at DCA is also much larger than it is at IAD, when it should be close to zero.

It’s a mystery, OK, but apparently the can of worms that I opened, and that Jason’s crew has expanded on, has touched off a firestorm of interest amongst weather and climate nerds, both professional and amateur. Some of them are some of the smartest people I know, so I hope to have a comprehensive handle on  the cause of the onset of DC’s local (but probably not real) fever in January, 2014.