April Come She Sure Did
May 15, 2016
Another month, another new record — for rising temperatures.
The latest NASA data reveals that April was the warmest month ever recorded on Earth. The new record marks the 12th consecutive month of record-high global temperatures, as the scientific consensus remains that human activity is contributing to detrimental climate change across the globe.
NASA data uses the average global temperatures between 1951 and 1980 as a control. April 2016 was 1.11 degrees Celsius above that 1951-1980 average, the sixth straight month that the average global temperature has exceeded that average.
According to Slate, Gavin Schmidt, director of NASA’s Goddard Institute for Space Studies, concluded that scientists can already predict with near certainty that 2016 will be the hottest year on record — a claim that’s hardly astonishing because 15 of the 16 hottest years on record have occurred since 2001, according to AccuWeather.
Climate Denial Crock of the Week 
May 15, 2016 at 7:58 am
The 12th consecutive month in 137 years of records topping the global temperature.
The odds for that to happen in a balanced climate would be
1 in 43,716,643,078,717,303,412,870,881.
May 15, 2016 at 12:30 pm
Tamino would like to talk to you about autocorrelation. 🙂
But yeah, it’s ludicrously warm. After a winter where we surprisingly got a decent snowpack in the mountains here, it’s melting out much faster than normal.
May 17, 2016 at 1:22 pm
Intended as a reply to the “how did he get that number” discussion below: 137^12 is the number Sir Charles posted. It’s true that if you had 137 years of monthly numbers and you randomly picked one January, one February, etc., then the odds that you pick the highest month for each month would be 1 in 137^12.
But there’s this thing called autocorrelation in time series that’s a rather technical issue to deal with mathematically. An expert in statistics who blogs under the name Tamino (tamino.wordpress.com) is actually about to start offering an online course in statistics where he will focus on time series. His blog is heavy on climate science, so anyone who reads this blog should probably check out his.
May 17, 2016 at 1:39 pm
In general, autocorrelation describes data that has a seemingly random distribution when points are chosen at random, but when plotted against their domain (typically time, sometimes spatial location or something else) they aren’t really random because they can only change so quickly over time/distance/whatever. Here’s a really silly example, but it illustrates the point. Suppose I have a distribution function that is the height relative to sea level of any random point on the land surface of the earth. But if I take a bunch of measurements all within a few miles of my house, they aren’t going to look anything like the global distribution, because obviously the height of terrain is spatially autocorrelated.
What autocorrelation means in a time-series such as this context, is that “hot months are more likely near other hot months”.
May 15, 2016 at 6:22 pm
That is one batshit number ( and I mean that in the nicest way ).
How exactly did you arrive at that?
All I could find out about that number was that it’s the 22nd term of a sequence defined by (6n + 5)^12 where n is a +ve integer starting from zero.
May 15, 2016 at 9:08 pm
Who’s giving Morin Moss and I “thumbs down” for commenting on Sir Charles’ math? That “batshit number was PFTA—-Plucked from Thin Air—-as a joke, and I suspect Morin Moss’s came from the same place. Why are we trying to confuse people? Or does everyone get the joke but me and that’s why Sir Charles got 8 “thumbs up”?
The question of a particular month being a record high is a simple yes or no—-like flipping a coin—-so the chances are 1 in 2 or .5 or 1/2. Two consecutive months is 1/2 times 1/2 or 1 in 4, three is 1/2 times 1/2 times 1/2 or 1 in 8, and so on.
I made a a typo with the 1 in 8096—-it’s actually 1 in 4096 that record highs would occur for 12 months in a row. The series continues:
13 = 1 in 8192
14 = 1 in 16384
15 = 1 in 32768
16 = 1 in 65536 etc.
The longer it goes, the more astronomical the odds become. If it continues for 24 consecutive months, the odds of that happening would be around 1 in nearly 17,000.000, which is actually more scary than inventing impossible numbers
May 15, 2016 at 9:47 pm
I don’t know why Sir Charles picked that particular, very large number but I wasn’t kidding.
If you start with zero, the 1st number that results from (6n+5)^12 would be 5^12 or 244140625 and the 23rd) (not 22nd as I posted above would be ((6×22)+5)^12 which is 137^12.
And that is exactly 43,716,643,078,717,303,412,870,881
May 16, 2016 at 9:08 am
just because your formulation is right doesn’t mean it applies to this situation.
May 16, 2016 at 9:35 am
Maybe it does (sort of) if you take into account the “autocorrelation” that Greg W brought up and try to factor in the probabilities of “twelve in a row” at any point over the 137 years of records—-that’s what Morin Moss is doing and Sir Charles must have done, and that crazy number apparently IS what you get.
The math needed to do that makes my Dumb Old Head hurt too much, so I’ll take their word for it and stick with the simple “extended coin flip series” I outlined earlier, and repeat that the odds against so many record hottest months in a row of 12, 15, 24, or anything in between are still pretty darn high. Looking at all that’s going on with the climate, betting on the string to continue is a much safer bet today than it would have been a few decades ago (ignoring the odds).
May 15, 2016 at 8:10 am
“May, she will stay”
May 15, 2016 at 11:08 am
Reblogged this on A Green Road Daily News.
May 15, 2016 at 12:45 pm
Thanks for rounding that number off for us. LOL. It DOES sound like it was PFTA, though, which can now also be interpreted as Plucked From Trump’s (Anal orifice). the difference being that if The Donal d says it, it must be true.
The actual odds of this happening in a “balanced” climate are 1 in 8096, which are way longer odds than the longest ever posted at a a horse race.
May 15, 2016 at 12:46 pm
(speaking in reply to Sir Charles)
May 16, 2016 at 1:20 pm
NASA data uses the average global temperatures between 1951 and 1980 as a control. April 2016 was 1.11 degrees Celsius above that 1951-1980 average, the sixth straight month that the average global temperature has exceeded that average.
I don’t get what this quotation is saying. The last month to be below the 1951 – 1980 baseline, i.e. to have a 0 or negative anomaly in the NASA GISS data, was September of 1992. That month shows a -1 anomaly, or 1/100th of a degree C. The link actually says the six months were “one per cent above the 1951-1980 average”, but that doesn’t even make sense. 1% of what? The best I can figure is they meant to say 1 degree C not 1%, and they also meant to say seven straight months, not six.