That’s because there’s more than one history of global temperature. Three receive the most citations. NASA’s record begins in 1880, as does another history from the U.S. Department of Commerce, developed at the Department’s National Climatic Data Center (NCDC). But the most widely referenced history (and the one primarily used by the U.N.’s Intergovernmental Panel on Climate Change (IPCC)) is compiled by the Climate Research Unit (CRU) at England’s University of East Anglia. It goes back to 1856.
The vast majority of the underlying temperature observations that go into each of these compilations is the same, but each organization has developed its own techniques for how the raw observations are geographically combined and adjusted for confounding factors such as urbanization, missing values, etc. As a result, annual values in each temperature history differ slightly.
So let’s take a look at where the average temperature in each stands through September 2005, and what the prospects are for setting a record for the year as a whole, given that there are still three months of data to be added.
In the table below are all of the relevant numbers.
|Source||Record||2005 (Jan-Sep)||Additional Monthly Anomaly||Number of Observed Exceedences Needed to Set Record||Length of Record||Percent Chance that 2005 will set record|
The GISS anomalies are calculated from the average for the period 1951 through 1980, the NCDC anomalies are relative to the average from 1880 through 2004, while the CRU temperatures are the departures from the 1961-1990 mean. But this matters little in our analysis. In each history, the record-warm year is 1998. The September anomaly in the CRU data is not yet available but we have estimated it based upon the values reported by GISS and NCDC.
The “Additional Monthly Anomaly” is the increment relative to the January-September average that each month in the period October through December must average in order to have 2005 become the warmest year. It is negative for the GISS temperatures because they are currently above the record value.
The “Number of Observed Exceedences” is the number of times during the period of record that the average anomaly (relative to the first 9 months of the year) during last three months of the year reached or exceeded the value required to have 2005 set the record.
The “Percent Chance” is “Number of Observed Exceedences” divided by the “Total Number of Observations”.
This last column is where the rubber meets the road. Based upon the previous behavior of the climate system (as captured by the global average temperature in each compilation), there is a nearly three-in-four chance that 2005 will finish as the warmest year in the NASA GISS global temperature history, but less than a one-in-five chance that the NCDC record will be set, and virtually no chance that the CRU will report 2005 as the hottest year measured. Both Rind and Taylor are going to turn out to be correct.
The significance of all of this depends on whom you talk to. The press, as already foreshadowed by Eilperin’s Washington Post article, will surely trumpet the record-setting values from the GISS data, while noting that the other datasets (probably) will have placed 2005 as the second warmest year on record. The various scientists interviewed will point out that this occurred even in the absence of a strong El Niño (the primary reason 1998 was so hot) and that this is further evidence that the earth is warming from an enhancing greenhouse effect.
So, what else is new? We already know that the world is warming and that it will continue to do so for the foreseeable future (with or without any greenhouse gas emission controls). Record temperatures will continue to be set every couple of years or so. In fact, if it weren’t for the 1998 El Niño, a new record high global average temperature would have been established in 4 of the last 5 years (including 2005). The big news is that 2005 will further establish that the rate at which temperatures have been rising during the past 30 years or so has been remarkably constant with a value of about 0.17ºC per decade, and it shows no sign of speeding up. Climate models share this constancy of warming; they just predict different rates. Unless that behavior is wrong, the additional warming until 2100 will be about 1.6°C, near the low end of projections made by our friends at the United Nations, and, frankly, too small to worry about, given that the energy structure of our society is likely to change dramatically in 100 years’ time. We’ll bet that no one points that out in December, when the warmth-of-2005 stories will proliferate like Santas.
Goddard Institute for Space Studies Temperature Data
National Climatic Data Center Temperature Data
Climate Research Unit Temperature Data