Changes in U.S. climate data do nothing to debunk global heating

Posted on August 13, 2007

6


A while back, NASA’s Goddard Institute for Space Studies (GISS) revised their temperature data for the United States based on the work of Steve McIntyre of Climateaudit.org. In the process, 1998 (which had been trumpeted as the hottest year on record) was downgraded from the hottest year on record in the United States to the second hottest year on record. The new hottest year became 1934. Similarly hot years moved around a little, with 1921 moving from #4 to #3 (and 2006 falling from #3 to #4), and the rest of the top 10 staying pretty much stable. The predictable result? Deniers like this National Post article using the corrected data to justify screaming “The global data is wrong! The global data is wrong!” (Note that the Post article says, in the second to last paragraph on page 2 of the story that the corrections only apply to the U.S. and claims that this matters the most because the U.S. has the longest and best-kept meteorological records. Too bad that India has several that go back to 1785, scientific observations started in Australia in 1788, 1872 marks the accepted start of in Japanese scientific measurements, and in France in 1765, pointing out that there was global, scientific meteorological observatories internationally before 1880 and that the Europeans created a widespread meteorological observation system well before the Smithsonian Institution set up observation posts across the U.S. starting in 1849.)

Unfortunately, just because the data for the U.S. was wrong and needed correcting doesn’t mean that all the global data is wrong.

Look at this image. It’s a graph of the corrected U.S. data. There’s no denying that it was hot between 1920 and 1960 – a simple counting up of the number of data points hotter than +0.5 degrees C variance gives you 9 years over that 40 year period. But if you count up the number of years since 1980 that have been hotter than +0.5 degrees C variance, you find that there are 14 years. Using a rough extrapolation tells us that there will be 22 years hotter than +0.5 degrees C variance between 1980 and 2020. Better yet, if you grab the actual GISS temperature data from this link and run a linear best fit of the data in your favorite spreadsheet program, you’ll find that the linear trend is positive in temperature. In fact, when I ran the Microsoft Excel “linest” function on the yearly data, I got a positive slope of .0048 degrees C per year (and a -9.3162 Y-axis intercept, although this number is provided just for verification of my math – feel free to check it if you don’t believe me). In other words the long-term U.S. trend in temperature is toward increasing temperatures.

But even if the long-term U.S. temperature trend had actually flattened, GISS revised their global graphs to correct for the updated U.S. data less than a week after the U.S. data was revised. In fact, this image shows the corrected global graph, using nothing but meteorological station data. Notice that the global trend hasn’t changed.

I’m glad that there are people like Mr. McIntyre out there to keep the scientists honest. After all, sometimes even the best of us make mistakes, and uncompromising reviews of data and models and reconstructions are necessary to make sure that mistakes are caught and corrected. And in this case, science worked. Unfortunately for the global heating deniers (and fortunately for everyone else who knows that global heating must be addressed as soon as possible), those who used this minor correction to turn into Chicken Littles did their arguments no good whatsoever.

(Thanks to fellow S&R blogger Sam for putting me onto this one)

Advertisements
Posted in: Uncategorized