The stations spread across the continental United States to collect temperature data have become the flashpoint in the latest climate-change skirmish.
A new analysis, released online by former TV meteorologist, climate-change skeptic and blogger, Anthony Watts, suggests many of these stations are collecting inaccurate records. Add statistical fiddling by climatologists, and the result is an artificial doubling of the rate of warming for the lower 48 U.S. states over the past 30 year years, he and colleagues write.
Watts goes even further, saying that when other factors, such as urbanization, are taken into account, the rate of warming declines even further.
"The trend value is not much greater than zero. Some regions of the USA are actually cooling," Watts told LiveScience in an email, referring to the change in U.S. climate.
However, climate research has shown that the planet, including the United States, is heating up significantly. Data from the U.S. National Oceanographic and Atmospheric Administration (NOAA) shows the lower 48 have warmed by 0.58 degrees Fahrenheit (0.32 degrees Celsius) per decade between 1979 and 2008. A representative of NOAA acknowledged that challenges exist to collecting historical weather data, but said these don't invalidate the changes in temperature seen over time. Other climate scientists pointed out that the study had not been vetted by other scientists, and said it was receiving undue attention. [Extreme Weather Facts: Quiz Yourself]
An inconvenient truth
Watts' research was released earlier this week, as was another analysis, headed by University of California at Berkeley physicist Richard Muller, which came to the opposite conclusion. Muller's group traced warming back to 1753 and found it closely correlated with carbon dioxide, the principal greenhouse gas, indicating humans are the primary cause.
It's important to note neither has been peer reviewed, the process used to vet scientific papers for publication. "You should not be writing about them at all," Kevin Trenberth, a senior scientist the National Center for Atmospheric Research in Boulder, Colo., told LiveScience in an email.
Even so, by challenging NOAA's warming rate, Watts touches on some important, and well-known, issues for temperature data, said Thomas Peterson, principal scientist at NOAA's National Climatic Data Center (NCDC). That said, U.S. weather records capture an image of temperature trends that is "quite accurate," Peterson said.
At this point, temperature readings alone aren't necessary to validate climate change, since seasonal events like flower blooms and bird migrations are shifting, Peterson said. "If we did not have any thermometers here on the planet, we would know darn well that it is warming because of all these other indicators." [10 Surprising Results of Global Warming]
The history of weather
The data collected by weather stations is just that "weather," which describes things like a storm or temperatures this summer; years of weather make climate, so in order to know anything about climate, researchers need records that go back many years. A network of long-standing weather stations, called the U.S. Historical Climatology Network, allows climate scientists working for the NOAA and elsewhere to do that. This network contains about 1,200 stations with thermometers located around the country with records that can date back into the late 19th century.
While researchers want a large-scale picture of climate, temperature varies according to the immediate environment — a nearby paved surface or shade from a tree can affect readings — and the timing of a measurement — 5 p.m. is typically warmer than 9 a.m. As a result, keeping measurements consistent over decades is a challenge.
In a previous survey, Watts found numerous problems with the placement of the monitoring stations, and a U.S. Government Accountability Report, published a year ago, found 42 percent of stations did not meet at least one standard regarding their location, such as being too close to extensive paved surfaces or obstructions such as buildings or trees.
However, a study published in 2010 by NCDC researchers in response to these concerns, found no evidence that the temperature trend was inflated as a result, and other work has come to similar to conclusions, Gavin Schmidt, a climate scientists at NASA's Goddard Institute for Space Studies, told LiveScience in an email.
"This is of course not the answer that Watts et al want to hear, and so they keep talking about it as if this work doesn't exist," Schmidt wrote.
The controversy extends to a statistical process, called homogenization, which climate scientists use to correct for bias in the data, which Watts' analysis says further inflates the warming trend. However, the homogenization methods used by NCDC have been heavily reviewed and ranked among the best internationally, according to Peterson.
"There is no network in the world that does not have this problem, so scientists all over the world are working on this," Peterson said.
Watts' work contradicts the unpublished research carried out by Muller, a former climate skeptic who recently made a "total turnaround" on man-made climate change. He and and colleagues used historical temperature records, going back all the way to 1753, as well as modern records to reconstruct global and regional warming over land.
Globally, average temperature has risen 2.7 degrees Fahrenheit (1.5 degrees Celsius) in that time, and he and colleagues found the upward path of temperature closely matches the increase in carbon dioxide in the atmosphere. If bad station data, not carbon dioxide, were responsible for the warming trend that would be "a truly remarkable coincidence," Muller said.