“Global temperatures have not increased since 1998.”
That point has been a common argument among climate skeptic communities in the blogosphere for the past few years. It gained prominence recently in an article in the New Statesman by David Whitehouse, a journalist and former BBC science correspondent.
“Global warming has, temporarily or permanently, ceased,” Whitehouse wrote. “Temperatures across the world are not increasing as they should according to the fundamental theory behind global warming – the greenhouse effect. Something else is happening and it is vital that we find out what or else we may spend hundreds of billions of pounds needlessly.”
So have temperatures really stopped increasing? And would stagnating temperatures despite increasing CO2 emissions cast doubt on the theory of anthropogenic warming?
There are two ground-based global temperature datasets: NASA’s GISSTemp and the Met Office Hadley Center’s HadCRU. Both for the most part use the same network of temperature stations. They differ subtly in their methodologies for correcting heat island effects and other station-specific anomalies and in how each interpolates temperatures in regions with few or no stations, such as polar regions.
Two global temperature records from satellites are plagued by much greater uncertainty because of orbital decay and other factors, and those are not discussed here.
The figure below shows temperature records since 1975, when the “modern period” of global warming began as greenhouse gas forcings started to exceed other climate forcings. Both GISSTemp and HadCRU show a strong positive warming trend of roughly 0.18 degrees C per decade over the past 32 years. GISSTemp shows that seven of the 10 warmest years on record have occurred in the past decade, and HadCRU shows nine of the ten warmest years in the same period. GISSTemp identifies 2005 as the warmest year on record, with 1998 and 2007 tied for second. HadCRU has 1998 as the warmest year, followed by 2005 and 2003, with 2007 placing sixth. However, the variance in temperature between these years is all within the margin of error of the respective dataset, so year-by-year rankings are not that meaningful. The overall trends, however, are robust and unambiguous.
|Temperature anomalies are measured in degrees celsius relative to the 1951-1980 average temperature. 2007 HadCRU data is estimated using November and December 2007 values from GISS, as these are not yet available from HadCRU.|
To determine if warming has recently stopped, consider the data from the past eight years, from 2000 to 2007. This is a more meaningful comparison than 1998 to 2007, as 1998 temperatures were anomalously high as a result of the “El Niño of the century” (pdf), a natural cyclical event that produced an enormous temperature spike relative to surrounding years. Choosing an El Niño year as that start of the dataset would amount to rather egregious cherry picking (though both GISS temp and HadCRU would still show a warming trend over the decade).
|Temperature anomalies are measured in degrees Celsius relative to the 1951-1980 average temperature.|
Over the past eight years, Earth has warmed 0.025 degrees C per year according to GISS, and 0.014 degrees C per year according to HadCRU, so GISS shows slightly faster warming than over the long-term trend of 0.018 degrees C per year, and HadCRU shows warming slightly slower.
However, these trends are somewhat meaningless: eight years is too short a period to determine anything about the trajectory of climate change. As the 1975 to 2007 temperature record shows, annual temperatures are subject to numerous spikes and dips. One cannot reasonably argue that global warming stopped in 1983 or 1992.
Likewise, the argument that global warming stopped in 1998 is mistakenly conflating the large degree of annual variability (the temperature “noise”) with the long-term trend. The difference between temperatures reported by GISS and HadCRU is no larger in recent years than it has been in the past, so there is no reason to believe that one is systematically over- or under-estimating recent temperatures.
Climate scientists writing for RealClimate clearly demonstrate the danger of relying on such short time series in the figure below, plotting a running eight year linear trend from 1975 to 2007.
|Annual GISSTemp record with an eight-year trend for each specific year. Major volcanic eruptions and the 1997/1998 El Niño event are identified. Figure taken from RealClimate.|
What about Whitehouse’s second argument? Is it unusual to have a period of stagnating temperatures while CO2 levels are increasing? Here again there is a long term positive correlation, but considerable short-term noise.
Atmospheric CO2 concentrations have increased steadily (and roughly linearly) over the past 32 years, with no appreciable dips or spikes in the annual average value. It is therefore realistic to conclude that dips and spikes in temperature from year to year are driven more by natural factors (e.g., El Niño, volcanic eruptions like Pinatubo, etc.) than by anthropogenic greenhouse gases, and that the long-term warming trend is primarily driven by human emissions.
|Annual average atmospheric CO2 concentrations taken from the Mauna Loa observatory. Temperature anomalies are measured in degrees Celsius relative to the 1951-1980 average temperature.|
What is unclear at this point is whether Earth has warmed faster or more slowly than expected over the past eight years. Whether the GISS or HadCRU dataset proves more accurate is somewhat inconsequential, as both show the same long-term trends.
There is a considerable danger in inferring too much from these short-term temperature fluctuations, and the period in question is too short to draw either dire or comforting conclusions.
In the end, reporters covering climate change should keep in mind that the relationship between temperature and human greenhouse gas and aerosol emissions is considerably more complicated than we often assume, and natural forcings tend to dominate year-to-year variability.