Arguments
Study finds slim odds of record heat, but not as slim as reported
Posted on 27 January 2016 by John Abraham
No, this isn’t another article about how damn hot 2015 was. Although just between us, I may have lost a bet to climate seer Joe Romm because he correctly predicted 2015 would blow 2014 out of the water. Instead, this is a post about the probability that temperature records keep getting broken if climate change is natural.
A paper just published in Nature Scientific Reports by Michael Mann and Stefan Rahmstorf and their colleagues considered this question. In particular, they wanted to know how likely recent temperature records and the string of records would be if the climate was completely driven by natural variations. Not even including the crazy-hot 2015, what did the authors find?
Well it depended on the “record” they looked at. For instance, the likelihood that 13 of the hottest 15 years would be in the past 15 years is 1 in 10,000. The likelihood that 9 of the 10 hottest years occur in the past decade is 1 in 770. The results are similar regardless of whose temperature dataset is used.
Some media stories reported that the temperature records were even more unlikely. The reason this study arrived at different results is that they took into consideration the fact that temperature records are not like coin-flips. Each year is not independent of another year or the prevailing situation.
For instance, there are natural events that alter the temperature such as volcanoes, variations in the solar output, or even internal variability such as the La Niña/El Niño cycle. If climate change were all natural, you could get these natural events to align just by chance, and this could give a naturally-occurring record. So, while it’s very unlikely that this could occur, it is much more likely than if we just treated each year as a coin flip.
The authors then asked how likely it would be to have a string of records given the reality of human-caused warming. They found that it’s 83% likely that 9 of the last 10 years would be the hottest on record and 76% likely that 13 of the past 15 years were hottest on record. They also found that the odds of 2014 being the hottest year on record was 40%. Without human influences, that chances 2014 would have been the hottest year is approximately 1 in a million. Mann said:
The press reports last year about the unlikely nature of recent global temperature records raised some very interesting questions, but the scientists quoted hadn’t done a rigorous calculation. As a result, the probabilities reported for observing the recent runs of record temperature by chance alone were far lower than what we suspected the true probabilities are.
Dr. Stefan Rahmstorf (left) and Dr. Michael Mann (right). Photograph: Phil Coates
Although the new odds of producing recent runs of record temperatures by chance are greater than the odds previously reported in the news (between 1 in 27 million and 1 in 650 million), they’re still incredibly slim at between 1 in 5,000 and 1 in 170,000. Including the data for 2015, which came in after the study was completed, makes the odds even slimmer.
In calculating the odds, the previous reports did not take into account that the data did not end simply because December 31 occurred, but that trends overlap into previous and subsequent years. This needs to be taken into account to determine the real probabilities of chance causing the warming events. As Mann noted,
I take "13 of the past 15 years were hottest on record" to mean that 13 of the past 15 years are in the top 15 hottest years of the temperature record. Including 2015, it's 14 of the past 15. Another way to put it is that every single year of the 21st C. (which started in 2001) was hotter than every single year of the 20th C., with the single exception of 1998.