LAST summer, headlines blared: "Hottest July in the history of the US!" The National Climatic Data Centre (NCDC) of the US National Oceanic and Atmospheric Administration said so, so it must be true. Last week, the NCDC was reporting the same, with the added alarm that last year was the warmest year on record and one of the top two years for extreme weather in the US.
Climate activists are linking this to man-made global warming, ignoring the fact that the area reported on in the NCDC reports, the US contiguous states (that is, continental America, not including Alaska), is only 2% of the earth’s surface. So trends that may, or may not, be real in the US in no way indicate global phenomena. In fact, the UK Met Office has admitted that there has been no global warming for 16 years and, last week, announced that temperatures are expected to stay relatively stable for another five years.
Regardless, all NCDC temperature proclamations must be taken with a large pinch of salt. Here’s why.
Until the use of thermocouples became common in the US climate network, temperatures were determined with mercury thermometers that have, at best, accuracy to within about 0.5°C. Even today, many US stations record temperatures only to the closest whole degree Fahrenheit (0.56°C). Thus, breaking the 1936 high temperature record by about 0.1°C, as the NCDC claimed occurred last July, is not meaningful. This change falls within the uncertainty of the measurement. It is akin to being alarmed that the moon has moved a millimetre closer when we can only measure the earth-moon distance to within a few centimetres.
All that was recorded for most of the US record was minimum and maximum temperatures for each day. The NCDC’s so-called average daily temperatures were derived by simply computing the average of the minimum and maximum temperatures. But this is not a true average since it does not take into account how temperatures varied throughout the day.
Trusting the NCDC averaging method to come to "hottest ever" conclusions is a mistake because higher minimums at night will yield a higher daily average, even if daytime highs do not rise. This is what happened in July last year, when, because NCDC records indicate that the US was less cool at night than in July 1936, the average they computed for July last year was higher than in 1936. Yet, as demonstrated by Roy Spencer of the University of Alabama at Huntsville, the NCDC records show that daytime high temperatures in July 1936 far surpassed last year’s. So, July last year was not the warmest month in the US’s 118-year instrumental record.
Last week, the NCDC’s credibility was further damaged when Chico, California-based meteorologist Anthony Watts said he had discovered huge differences between its "State of the Climate" (SOTC) reports released each month and the actual database of NCDC temperatures. For example, the July 2012 SOTC report, issued in early August, announced that a new record had been set, with the average July temperature for the contiguous US being 25.33°C, or 0.11°C higher than in July 1936. However, the NCDC says the July 2012 average was actually 24.96°C, or 0.37°C less. This is then 0.26°C cooler than the 25.22°C claimed as the previous monthly record in 1936.
What is going on?
It turns out that NCDC does not wait for all the data to be received before computing and announcing the US average temperature and its rank compared with other months and years. When the data from lower technology sources finally arrives, the NCDC updates its temperature database, typically "cooling" the country when all the data is used.
But the NCDC does not tell the public and the press if, when the complete data set is analysed, the temperature announcements in previous SOTCs are no longer correct.
Strangely, the NCDC changes temperature data even from the distant past without notification. For example, the NCDC now asserts that the average temperature in July 1936 was 24.68°C, more than half a degree cooler than the 25.22°C it claimed for the month in the July 2012 SOTC report.
This allows it to continue to say that July 2012 set a record.
Watts found that, in the 23 monthly SOTC reports between October 2010 and November last year (three SOTC reports did not list average temperatures), 22 of them do not match the NCDC database, presumably due to later updating, when all the data is received and analysed. And, in all cases except one, the country cooled when all the data was incorporated.
Watts says: "It is mind-boggling that this national average temperature and ranking is presented to the public and to the press as factual information and claims each month in the SOTC, when in fact the numbers change later."
So, we don’t really know how much, if any, warming has occurred in the US over the past century. Since the US records are considered to be the most accurate part of the Global Historical Climatology Network, we really do not even know that global warming has occurred at all in the past century.
• Harris is executive director of the International Climate Science Coalition. Ball is a Victoria, British Columbia-based environmental consultant and former climatology professor at the University of Winnipeg. Both are advisers to the Frontier Centre for Public Policy.