Posted on 24 August 2007 by John Cook

There's been much talk across the blogosphere of late regarding the top ten hottest years since NASA discovered an error in their temperature data. The skeptic argument 1934 is the hottest year on record has been bandied around and even cracked the Skeptical Science Top Ten - it's currently the #9 most used argument and likely to climb further over the next few weeks. Now granted, I started this database in May so it's only an indication of the arguments used over the last 4 months. Nevertheless, the 1934 argument's sudden ubiquitiousness is a surprise.

Two misconceptions commonly appear. The first is NASA's data error was caused by a Y2K bug. That the error began in January 2000 has nothing to do with Y2K - it's a mixup over data corrections with the NOAA. NASA obtain much of their temperature data from the NOAA who adjust the data to filter out inhomogeneities, urban warming and time-of-observation bias (more on NOAA adjustments). From January 2000, NASA were mistakenly using unadjusted data. A few websites even go so far as to claim "global warming is just a Y2K bug".

More serious is the misconception that the data corrections are significant at a global level. NASA's temperature adjustments only apply to 48 U.S. states. As the USA comprises only 2% of the globe, this has had infinitesimal effect globally. The change to the global temperature is less than a thousandth of a degree and has made no difference to the global trend.

While some of the original blogs mention the corrections as USA only, the emphasis is on NASA's error and changes to the top ten list. This has led some to erroneously get the impression that 1934 is the hottest year globally. From this, the Chinese whispers phenomenon took hold and many blogs now report 1934 is globally the hottest year on record. It's a shame such a misconceived argument should make our top ten list - I imagine this is how Google engineers feel when spam websites make their top ten search engine results.

More on 1934...