Record-breaking hot years almost impossible without humans - and Anthony Watts' conspiracy theory

Recent record hot years are almost impossible in the absence of human-induced forcings, a new study shows once again.  A multi-national team of scientists led by Andrew King from the University of Melbourne did a study of hot weather years, and found human influences on heat extremes going back to the 1930s globally, and earlier in some places. The new paper has just been published in Geophysical Research Letters.

Anthony Watts, a climate conspiracy blogger at WUWT, didn't read the paper, but he decided it made unfounded claims (archived here). Anthony decided, in his foolishness, that the team, led by a University of Melbourne scientist, and including other researchers from Universities in Melbourne, South Korea, Switzerland, Oxford UK, New Zealand and the New South Wales, must be wrong because Professor Chris Turney (not an author) led an expedition on a ship that got caught in the ice in Antarctica a couple of years ago. Need I say more about the weirdness of science deniers?

Record hot years influenced by human activity - going back to the 1930s globally

The authors weren't looking at heatwaves (a duration of days), they restricted their study to timescales of at least a month. They performed an analysis of hot years looking at the globe as a whole and selected regions. They found that by using the technique knows as Fractional Attributable Risk (FAR), they could determine years in which there was a detectable anthropogenic influence on especially hot years globally and in different regions of the world.

Below are charts showing the hot years (marked with an X) that would be unlikely to have been as hot except for human-induced warming. As always, click the charts to see them more clearly.

Figure 1 | Modelled and observed annual temperature timeseries. Timeseries of observed (grey line), historicalNat ensemble (blue), historical and RCP8.5 ensemble (orange) temperature anomalies for 1901-2015 (or shorter periods for some observational series) from a 1961-1990 climatology. Record-breaking years (excluding the first year in each timeseries) that may be attributed to anthropogenic climate change are marked with black crosses. Source: King16

You'll see that in Australia, where the new head of the CSIRO has proposed we be blindfolded so we won't know what is happening, human influence can be detected more than a century ago. Same for central England. From the press release at ScienceDaily.com, the researchers list the years shown in the very top panel of Figure 1 above, where human activity had a detectable influence globally:

According to the new study, record-breaking hot years attributable to climate change globally are 1937, 1940, 1941, 1943-44, 1980-1981, 1987-1988, 1990, 1995, 1997-98, 2010 and 2014.

With regard to the earlier years in the list in the 1930s and early 1940s, the authors wrote that as well as a detectable human influence, these may have also been partly influenced by warm phases of the Atlantic Multidecadal Oscillation (AMO) and the Interdecadal Pacific Oscillation (IPO).

Moving windows of 21 years

The authors used moving windows of 21 years in their FAR analysis to help detect anthropogenic influence. They explain in the paper:

Windows of 21 years were chosen as the use of shorter periods would reduce sample sizes and the use of longer periods would lead to larger changes in the anthropogenic forcing between the start and end of the windowed period.

What I think they are saying is that the world is warming so quickly, that using windows larger than 21 years means the disparity between the start and end years is big.

Fraction attributable risk or FAR

One of the ways climate scientists work out the likelihood of an extreme event is fraction attributable risk, or FAR. This can be used to indicate the fraction of the risk that can be attributed to global warming. It is an attempt to show how likely it is that an extreme event would occur without the extra CO2 that we've added, compared to how likely it is that it occurred because there is now extra CO2. In a recent paper on the subject, Bellprat and Doblas-Reyes wrote FAR as the equation:

FAR = 1-P_nat/P_all

Where:

• FAR = fraction attributable risk
• P_nat= the probability with only natural radiative forcing (ie not including any forcing resulting from human activity)
• P_all= the probability with all radiative forcing, including from human activity

So if the probability of an event occurring naturally is one in a thousand or 0.001 and the probability that it occurs with human activity =0.99, then FAR = 1-0.001/0.99 or 1-0.001, which is 1.0. If a similar event took place recently, then it would be attributed almost solely to human activity.

If FAR is 1, then it is getting close to impossible for the event to have occurred without the extra greenhouse gases. If FAR is zero, then the event could happen with or without any human-induced forcing. If FAR = 0.5 then the chance of a weather event occurring is doubled with AGW.

Model selection and comparison with observations

The FAR technique is one in which observations are compared with climate models, using natural only as well as human forcings. That way, scientists can compare what would have happened with only natural forcings, with what would happen with anthropogenic forcings, and with what happened in reality. The scientists chose models that "have at least three historical simulations, one historicalNat simulation and one RCP8.5 simulation available for analysis". They then regridded them onto a common 2x2 grid (except for the CET and global series) and masked them before area averaging, as shown in Figure 1 above.

The paper by Bellprat and Doblas-Reyes (which was touted on denier blogs with no understanding) discussed how model reliability needs to be taken into account, and that unreliable models add to uncertainty. To avoid this problem, and after performing the above, Andrew King and his colleagues selected models that captured the observed temperature variability. They defined this as "no more than one-third of the historical simulations being significantly (p < 0.05) different than the observational series...for a given region". It's those models that were selected for further analysis. In addition, to account for sampling uncertainty,

the 21-year moving natural- and all-forcings ensembles associated with each year were bootstrapped (with replacement) 10000 times on 50% sub-samples of complete model simulations. These 10000 bootstrapped estimates were used to calculate 10th and 90th percentile FAR estimates.

Human-induced cooling detected

It wasn't just warming that was detected. The scientists were able to detect the influence of smog, or when the smog lifted. From the press release:

Aerosols in high concentrations reflect more heat into space, thereby cooling temperatures. However, when those aerosols are removed from the atmosphere, warming returns rapidly. The researchers observed this impact when they looked at five different regions: Central England, Central Europe, the central United States, East Asia and Australia.

There were cooling periods, likely caused by aerosols, in Central England, the central United States, Central Europe and East Asia during the 1970s before accelerated warming returned, and aerosol concentrations also delayed the emergence of a clear human-caused climate change signal in all regions studied except Australia, according to the study.

Describing likelihood

The authors described the levels of likelihood:

• A significant anthropogenic influence is detected if the 10th percentile FAR value is greater than zero (i.e. human-induced climate change has very likely increased the likelihood of record-breaking hot events). 
• A significant and substantial anthropogenic influence is detected if the 10th percentile FAR is greater than 0.5 (i.e. human-induced climate change has very likely at least doubled the likelihood of record-breaking hot events). 
• Similarly, anthropogenic influences can be said to have significantly reduced the likelihood of hot events if the 90th percentile FAR is below zero.This happens in the mid-20th century in Central Europe and East Asia due to the effect of anthropogenic aerosols.

Anthony Watts didn't read the paper and wrote irrelevancies

Anthony Watts didn't read the paper. It looks as if he didn't even read the press release he copied and pasted. He wrote:

They suggest that from 2000 onward, heat waves should be anywhere from 2 to 10 times more common. By that reasoning, this graph of record high temperature should also be increasing during that period, but it isn’t:

I couldn't find anything in the paper or the press release about heat waves being from 2 to 10 times more common after 2000. I've no idea where Anthony got that from. The paper wasn't even about heat waves. It was about record hot years. If anyone can find where the paper or press release discusses this, do let us know.

Anyway, what does Anthony Watts do? Did he put up a chart of record hot years or months around the world? Nope. He figured that the USA is the entire world, and put up some charts about heat waves in the USA. Then he put up a chart from John Christy, showing the "average number of days with a temperature 100F or more divided by the number of stations".  I think Anthony was trying to argue that this hasn't happened, with the last three decades the hottest in the record:

Figure 2 | Dedadal land surface temperature anomaly for the contiguous USA. Data source: NOAA ClimDiv

As he often does, Anthony effectively accused the scientists of misconduct, writing:

In short, the entire study is an attempt to pull a desired result out of a set of data. Reading the paper, it seems clear to me that the conclusion existed before the paper was written.  Here is the press release for what it’s worth.

No, it doesn't. It's Anthony Watts who cherry-picked data to try to tell his readers that the USA isn't getting mighty warm.

From the WUWT comments

The conspiracy theorists were out in force, as you'd expect, with comments about climate science being a hoax and scientists committing fraud. WUWT is a climate conspiracy blog after all. Rather than inform people about climate science, Anthony Watts promotes Velikovskyism "in the context of learning". I didn't see a single comment from anyone who gave the slightest sign of accepting science. His blog is virtually 100% for climate conspiracy wackos.

Mycroft  thinks it's been too hot to get any sleep. He's not far wrong.

March 8, 2016 at 12:39 pm
Jesus! How do these clowns sleep at night?

Bill Partin decides that the climate science of the past 200 years has been a leftist plot from communist countries like New Zealand, Australia, South Korea, Switzerland, and the UK:

March 8, 2016 at 3:32 pm
They truly believe they are saving humanity from itself. Since most of them are far left cretins, they naturally use Alinsky’s framework as the means to attain their goals, subjugation of the human race to an elitist ruling class (them).

Tom Halla thinks the paper was only about the USA. He's wrong:

March 8, 2016 at 12:56 pm
Are they using raw or corrected records from the 1930’s? As various reports of what the raw records were from the 1930’s, at least in the US, it was warmer then than in the 2000’s. 

Brendan is a climate fraud conspiracy nutter:

March 8, 2016 at 12:59 pm
We tortured the data until it confessed!
But of course they got their main aim – unquestioned parroting of their press release in the Fairfax papers across Australia and no doubt similar newspapers across the world. 

Paul Westhaver is incoherent:

March 8, 2016 at 1:21 pm
Conveniently, the data set reaches ALLLL the way back to 1895 and the 2000’s is just plain unusual.
I hardly think this pseudo science ought to be published here and could lead to the discredit of WUWT if the likes of Michael Mann choose to use it as evidence of Watt’s scientific acumen. /sarc 

Andrew says something about aeroplanes, aerosols and satellites for some reason known only to himself:

March 8, 2016 at 1:22 pm
Hmmm, yes – those aerosols. I’m reminded of the Swedish shopkeeper sketch – asked for deodorant he asked “ball or aerosol?” “Neither – I want it for my armpits.”
Has anyone asked the aerosols whether they also work at 33,000 ft? I’ve seen dark smoke hang over cities – I’ve never seen it at cruising altitude. Can’t see how it would affect the satellite readings. 

Owen in GA needs to read some paleo-climate papers, and learn something about past climate forcings:

March 8, 2016 at 1:26 pm
They don’t have a clue about “natural” climate drivers, but claim to have considered them all? When did I miss the trove of papers describing the climate drivers that created the Minoan Warm Period and the subsequent cold period, the Roman Warm Period and the Dark Ages Cold Period, the Medieval Warm Period and the Little Ice Age? I think any paper claiming to have “…examined weather events that exceeded the range of natural variability…” needs to first be able to fully explain the previous natural variability.

Bob Tisdale, who thinks global warming is caused by El Ninos and blobs, doesn't understand the first thing about climate. He wouldn't know a FAR from a near:

March 8, 2016 at 1:57 pm
Anthony writes: “In short, the entire study is an attempt to pull a desired result out of a set of data. Reading the paper, it seems clear to me that the conclusion existed before the paper was written.”
These are the fundamental problems with the hypothesis of human-induced global warming. As a result, the climate science community still can’t realistically differentiate between natural and manmade global warming. 

Trebla opts for sarcasm:

March 8, 2016 at 3:16 pm
You mean humans had an impact on the climate back in the 1930? Phew! I thought it was all my fault. 

This must be the first time taxed has heard about the cooling effect of aerosols:

March 8, 2016 at 3:22 pm
So let me get this right.
They are claiming that we would have had a lot more man made warming, have it not been for the man made cooling. Had they just come back from the bar when they wrote this.? 

Pamela Gray is another one who made a meaningless incomprehensible comment. (She's been away from WUWT for a spell, and returned recently to more vociferously pledge her science denial):

March 8, 2016 at 6:54 pm
I would think that records set is more a function of station growth and dropout, plus station changes from one kind of sensor to another kind of sensor. It seems these researchers ran over this first encountered pathology without even saying, “pardon me but may I pass.” 

References and further reading

Andrew D. King, Mitchell T. Black, Seung-Ki Min, Erich M. Fischer, Daniel M. Mitchell, Luke J. Harrington, Sarah E. Perkins-Kirkpatrick. Emergence of heat extremes attributable to anthropogenic influences. Geophysical Research Letters, 2016; DOI: 10.1002/2015GL067448 (subs req'd)

• Human influence on climate dates back to 1930s, new research finds - press release at ScienceDaily.com
• We traced the human fingerprint on record-breaking temperatures back to the 1930s - article by Andrew King and Mitchell Black at The Conversation
• Scientists trace human influence on extreme weather events back to 1937 - article by Tom Arup at Sydney Morning Herald

Bellprat, O., and F. Doblas-Reyes (2016), Unreliable climate simulations overestimate attributable risk of extreme weather and climate events, Geophys. Res. Lett., 43, doi:10.1002/2015GL067189.