The Science and Public Policy Institute has been asked to comment on the apparent inconsistency between the news that July 2012 was the warmest July since 1895 in the contiguous United States and the news that the Meteorological Office in the UK has cut its global warming forecast for the coming years. The present paper is a response to that interesting question.
Early in August 2012, the NOAA issued a statement to the effect that July 2012 had been the hottest month in the contiguous U.S. since records began in 1895. NOAA said the July 2012 temperature had been 77.6 degrees Fahrenheit, 0.2 F° warmer than the previous July record, set in 1936.
However, NOAA's statement was based on incomplete information that has since been revised. Updated data available at the NCDC website (NCDC is the division of NOAA that maintains national climatic data for the United States) show that July 2012's temperature was not 77.6 °F, as NOAA had previously claimed, but 76.9 °F, half a degree Fahrenheit below the record 77.4 °F set in July 1936
Even this revised value may be a considerable exaggeration. In response to criticisms of the siting of U.S. temperature monitoring stations, in 2008 NOAA introduced a new network of carefully sited stations with up-to-date, standardized, properly monitored equipment. The Climate Reference Network, as it is called, shows that the July temperature for the continental U.S. was 75.6 °F, lower by 1.3 °F than stated by the NOAA in August 2012 based on incomplete data from its older, poorly-sited stations influenced by urban heat-island effects, and lower by 2 full Fahrenheit degrees than the 77.6 °F that NOAA had published in August 2012.
For some unaccountable reason, NOAA has not issued any statement correcting its original false claim that July 2012 was the warmest July since 1895. Indeed, the accurately-sited and properly-maintained Climate Reference Network is consistently delivering significantly lower contiguous-US temperatures than the older network: yet, again unaccountably, NOAA finds it expedient not to draw attention to the inconvenient results from its new network.
To summarize: NOAA, on incomplete data from an outdated, ill-sited, poorly-maintained network, stated in August 2012 that the July 2012 temperature in the contiguous U.S. had been 77.6 °F. The temperature in July 1936, the previous record, had been 77.4 °F. By the end of 2012, corrections to the NOAA dataset, but still based on the outdated network, showed that the July 2012 temperature in the contiguous U.S. had been 76.9 °F, half a degree below the July 1936 record. Also, the new Climate Reference Network shows that the July 2012 temperature in the contiguous U.S. had been 75.6 °F, two full degrees Fahrenheit below the erroneous value the NOAA had erroneously trumpeted as a new July record.
So to the UK Met Office. It has issued a revised projection of temperature change to 2017. Its previous projection, made last year, had predicted that global temperature would rise by around 1 Fahrenheit degree more than its revised projection, based on a new computer model, "HadGEM3". If the new projection proves correct, by 2017 the global warming in the 20 years since 1997 will have been statistically indistinguishable from zero for two full decades.
For comparison, the least-squares linear-regression trend on the HadCRUt4 monthly global mean surface temperature anomaly data for the past 18 years is statistically indistinguishable from zero; similarly for the HadCRUt3 for the past 19 years; and similarly for the RSS satellite lower-troposphere dataset for the past 23 years.