Slashdot Mirror
2016 Was Second Hottest Year For US In More Than 120 Years of Record Keeping (climatecentral.org)
Last year was the second hottest year for the United States in more than 120 years of record keeping, according to the National Climatic Data Center, marking 20 above-average years in a row. While Georgia and Alaska recorded their hottest year, every state had a temperature ranking at least in the top seven. Climate Central reports: The announcement comes a week before the National Oceanic and Atmospheric Administration, which released the U.S. data, and NASA are expected to announce that 2016 set the record for the hottest year globally. Both the global record and the U.S. near-record are largely attributable to greenhouse gas-driven warming of the planet. In addition to the pervasive warmth over the last year, the U.S. also had to deal with 15 weather and climate disasters that each caused more than $1 billion in damage. Together, they totaled more than $46 billion in losses and included several disastrous rain-driven flooding events. These events, along with continued drought, lay bare the challenge for the country to learn how to cope with and prepare for a changing climate, said Deke Arndt, the climate monitoring chief of NOAA's National Centers for Environmental Information. The temperature for the contiguous U.S. was 2.9 degrees Fahrenheit above the 20th century average for 2016, displacing 2015 and ranking only behind 2012, when searing heat waves hit the middle of the country. More notable than the back-to-back second place years, Arndt said, was that 2016 was the 20th consecutive warmer-than-normal year for the U.S. and that the five hottest years for the country have all happened since 1998. Those streaks mirror global trends, with 15 of the 16 hottest years on record occurring in the 21st century and no record cold year globally since 1911.
Yes, better calibration helps getting more accurate results with less uncertainties, but it's highly unlikely that all calibration errors were in the same direction and skewed over time into the opposite direction.
Good lord folks -- I'm tired of this discussion coming up periodically and people talking out of their asses.
Look: thermometers dating back the mid-1800s were highly precise and did not need frequent calibration. They had perfected glass tubes of mercury and could make very even marks on them by that point. A late 19th-century thermometer had precision that was easily within +/-0.1 degree. (For specialized applications and laboratory thermometers, there were plenty that were manufactured by the late 1800s to be read down to 0.01 degree.)
And a sealed glass tube doesn't need repeated calibration if it's not disturbed or damaged. The issue here is simply how good the (somewhat permanent) calibration was. By the first couple decades of the 1900s, there were standards organizations which existed that would do standardized calibrations (i.e., where you could get a standard calibrated thermometer or send one away to be checked for calibration). We have actual logbooks when many thermometers were checked for accuracy. We have actual logbooks where thermometers were replaced and the old thermometers were compared with the new ones in terms of their scales and calibration. Etc., etc.
Just because you cannot fathom that people 100 years ago could read a thermometer or manufacture an even glass tube doesn't mean they didn't. They did. We still have many of these thermometers today to prove it. A 1900-era thermometers is about as accurate as a 1900-era RULER.
In fact, in terms of precision AND accuracy, what you should be questioning instead is MODERN electronic thermometers, which DO need frequent calibration and are frequently only accurate to maybe +/- 1 degree even when calibrated properly. But they're used for convenience because they no longer need a human to go look at it and write it down. Ask any meteorologist who knows anything about temperature measurement, and he'll likely tell you that stuff they were using decades and even over a century ago (often accurate to +/-0.1 degree) is more accurate than the stuff weather records are generated with now. (And regardless, that +/-1 degree or whatever is plenty to generate an average over several years to compare temperature records.)
No -- the real issue in dealing with old records is questions of siting and distribution. Historical thermometers weren't always located in the best of places, but again, most were, and we generally have records of those that were. The biggest statistical issue is that we didn't have such even distribution for samples all over the globe, so there's some sampling bias. Again, there's a lot of work statisticians do to take this into account when looking at long-term global averages.
Anyhow, I personally have complete confidence that those statistical analyses are good and reflect the overall trend. But people who are arguing that old thermometers were bad and needed frequent calibration simply have ABSOLUTELY no clue what they're talking about.