Guest essay by Sheldon Walker
At the moment, I am working on a statistical proof which will clearly show that the recent slowdown was real, and that it is statistically significant. Dark forces have been trying to prevent me from finishing this proof. “He who must not be named” (othewise known as Tamino, aka Grant Foster, oh no, I just named him) has been casting Harry Potter type spells at me, like “autocorrelation”, “heteroscedasticity”, and “multicollinearity”. Luckily I have a good supply of magic jellybeans, and I have managed to stay safe so far.
Sometimes I like to take a break from working on the serious stuff, and one of the things that I enjoy doing is experimenting with data visualisation. Sometimes it is possible to create striking images, which convey an idea or message in a way that a statistical test just can’t do.
I need to make everybody aware that the following images are not statistical tests. When you look at them, you may say “that really looks like there was a slowdown”, but the image does not provide an objective result which can be measured. You are getting a subjective impression.
You might now understand why I chose the title for this article. If it looks like a slowdown, and walks like a slowdown, and quacks like a slowdown, is it a slowdown? Subjective impressions are not always wrong. In fact, they are often right. We could not function efficiently as humans, if we could not rely on our subjective impressions. That pile of papers “looks” like it might fall over. That car “looks” to be out of control. It “looks” like it might rain.
Enjoy looking at the images, and see if they give you a subjective impression. Whatever you do, don’t tell a warmist that the images prove that there was a slowdown. You are likely to get a 30 minute lecture on how there is no statistical proof that a slowdown ever existed. You may even have to endure Tamino’s sermon on “how things are not always what they look like”. In other words, who are you going to believe, Tamino, or your own lying eyes?
Sorry, I need to give you some “objective” data about the images. For the first image I used 741 trends. These trends came from the date range 1970 to 2017. I only used trends which were 10 years long or greater. The maximum trend length was 47 years Every possible starting year, was paired with every possible ending year. So some example trends were 1970 to 1980, 1970 to 1981, 1970 to 1982, …, 1971 to 1981, 1971 to 1982, …, 2001 to 2011, 2001 to 2012, …, 2006 to 2016, 2006 to 2017, 2007 to 2017.
Hopefully I have explained that clearly. Every possible trend which is 10 years or greater in length, from the data range 1970 to 2017.
Then I picked the trends that I thought were “in” the slowdown. I made this decision based on other testing that I have done. I used the date range 2001 to 2015 for the slowdown. The trends in this group started in 2001 or later, AND ended in 2015 or before. There were not many of these trends, only 15 of them. Since there are not many of them, I will list them all.
| Start | End |
| 2001 | 2011 |
| 2001 | 2012 |
| 2001 | 2013 |
| 2001 | 2014 |
| 2001 | 2015 |
| 2002 | 2012 |
| 2002 | 2013 |
| 2002 | 2014 |
| 2002 | 2015 |
| 2003 | 2013 |
| 2003 | 2014 |
| 2003 | 2015 |
| 2004 | 2014 |
| 2004 | 2015 |
| 2005 | 2015 |
All that you need to know now, is that the slowdown trends were plotted in red, and the non-slowdown trends were plotted in blue.
There is a bonus point for anybody who can name the famous French monument that looks like Graph 1.
Graph 1
This graph surprised me when I plotted it. So I thought that other people might enjoy it. I didn’t expect the slowdown trends to be clustered so much to the left. I would call that a striking image.
A problem that people like me have, is that we are always anticipating the negative comments that other people might make. I thought that people might complain that Graph 1 was “unfair”, because the slowdown trends only go up to 14 years in length, but the non-slowdown trends go up to 47 years. The huge mass of non-slowdown points above 14 years in the middle, makes the red points look like they are squashed into the left hand side, making it look more like a slowdown.
I think that this would be a fair comment, and the last thing that I want to be is unfair. So I removed all of the non-slowdown points above trend length 14 years, and plotted a second graph. See graph 2.
Graph 2
This graph is probably not as striking as Graph 1, but it still manages to create the impression that the slowdown trends are “unusual” when compared to the non-slowdown trends.
Please leave a comment after the article describing what your subjective impression is.
If we were to measure things properly, and retrofit the best sited stations with the same type of LIG thermometers as used at those stations in the late 1930s/early 1940s and take measurements at each individual best sited station in the same manner and practice as used at the station in question, so that modern day RAW observational data at each best sited station could be compared directly with historic observed data at each best sited station individually, without the need for any corrections/adjustments to RAW data, we would probably find that there was no or all but no rise in temperature at each of the best sited stations since the late 1930s/early 1940s.
How much of a slow down does one need?
The graphs are interesting but not very useful, because you have made the common error of plotting the same data multiple times! This makes it look as though you have many warming trends and few cooling ones, whereas this is not true. There have been warming periods and cooling periods, and to give a true plot you need to plot single data items once, say one point for each year. The length of each period is then quantised at one year, and the number of each is a true representation of what has happened. The last 20 years of data would be particularly interesting plotted this way.
Regards
davezawadi February 9, 2018 at 1:26 am
Thanks, Dave. Whether this is an “error” depends on the nature of the investigation.
For example, suppose we’re looking at 10-year trends. We want to determine whether a particular 10-year trend, say zero ± 0.05°C per decade warming, is common or uncommon in the historical record.
The only way to do this is the way used in the head post. You need to count (and plot) ALL of the possible 10-year trends in the historical record, not merely the non-overlapping trends.
You could think of it another way, as the odds of finding the given trend (e.g. 0 ± 0.05°C/decade) starting on a random date in the historical record … the only way to do that is to investigate and/or plot ALL the random dates, not just non-overlapping dates.
Regards,
w.
Ah! Eet eez ze ‘ockey – ‘ow eez eet you say? – steeck!!
The thing that jumps out at me like a huge phallus is the overwhelming number of warming trends when normally you might expect 50/50. Is that significant?
Hi zazove,
warmists think that I am denying global warming, just because I claim that there was a slowdown.But if you look at my Eiffel Tower graph, it clearly shows that the average warming rate from 1970 to 2017 was 1.78 degrees Celsius per century. That is clearly warming. We do live in warming times, with the occasional slowdown.
Remember that during a slowdown, it is normally still warming. It is warming, but at a slower rate than normal.
Now you know why I get called a denier. In my opinion, warmists are the biggest deniers on the planet.
Of course. We are in an interglacial period during which it warms, and warms, and warms…until it doesn”t. Then back into another ice age.
Hi Sheldon
“It is warming, but at a slower rate than normal.”
It is entirely “normal” for the rate to waver up and down a bit; the result of short-term fluctuations like the ENSO cycle for example – what else would one expect. Sometimes slower, sometimes faster. There was a faster rate up to the peak of the recent El Nino, now let us all hope the warming rate slows for at least a few years.
No-one disagrees with the statement above: Nick agrees, Toneb agrees, Tamino agrees. But you know, so what, the rate fluctuates. A few atmospheric wiggles don’t indicate anything about the underlying trend – zero significance. They signify there was an El Nino – but we already knew that.
Get fixated if there is ever a multi-year reversal in the trend of Ocean Heat Content, that *would* be significant, until then…
Rudi
Probably significant you saw it as a (huge) phallus
Was thinking the same thing 😀
Middle finger.
Statisticians use a more technical term: witch’s hat.
Yeah, me too. Never had one “jump out at me”. Not yet, anyway.
Log or anti-log scales can help.
You didn’t say what temperature data you have used for these charts…
Also I kind of miss your definition of slowdown. How do you decide that a part of the series is still in a given trend and on what basis do you consider it a slowdown?
Apart from that, two simple observations:
1) I would think it logical to expect to find the slowdown trends to the left of the chart, simply by the way your chart is constructed (0-1 degree is on the left) and how you (probably) define a slowdown trend.
2) Based on the temperature series that you used the most common trend, i.e. dominant or typical trend, seems to be about 1.8 degrees / century. Which is interesting but will only tell us something when you do the same for two periods: before (say) 1950 (say 1880-1950) and after 1950 (1950-2020). And then compare the dominant trends.
Statistical evidence proof.
You are using a run-length trend comparison, which is sensitive to how you decide to determine that a trend has ended. It may be better & easier to do something similar: for each year since the start of you data set you simply determine the trend over x-years using that year as starting point.
Say you have data starting 1880, and say you like to do this for 30 year periods, then you would simply determine the temperature delta for 1880-1910, 1881-1911, 1882-1912 etc. and then you chart them. Also you can then simply calculate the average 30-year delta for that data set. That should give you a fairly representative ‘ trend’ for the entire data set.
And again this will only tell you something when you compare the trend for two (or more periods). After all we know that it has been warming since 1800, the question really is whether there has been any significant increase in warming rate after circa 1950 (what caused that is still yet another issue). So you really have to compare warming rates between two periods. Something like 1880-1950 versus 1950-2020 might be useful.
Also consider: how will you know that any difference between the two periods is significant?
Earlier statistical tests, both parametric (when using the proper statistics for the underlying data-set, as argued by Keenan) as well as non-parametric (usually random walk variants) showed that the difference is not significant (or barely significant), so you should verify that whatever approach you take can be compared in a meaningful way.
It is best to use longer periods (30 years or even 40) because studies into the random walk properties of the climate measurement time-series showed that the typical random-walk length is fairly long (see for example https://wattsupwiththat.com/2011/02/14/pielk-sr-on-the-30-year-random-walk-in-surface-temperature-record/)
It appears to me that neither Graphs 1 or 2 show normally distributed data, like a bell curve. What keeps this from happening? Solar cycles? UHI? Recovery from Little Ice Age? The data appears to show variable external input to the system in question.
You can’t get a normal distribution from this kind of display. The total record length is 47 years. You get one and only one data point. For 46 years, you get two, 45 years you get three, etc. It’s always going to tend to look like a triangle.
Try a test:
take a ten year old flat data set. Add a massive trend to it (like, +0.02 per month, that is, 2.4K in a decade).
Past it at the end of current data.
Run the data through your tool, just to check if the massive trend you know is here, shows.
My bet is: you will have a faint hint, but not enough “prove” it.
If however you do, repeat with a smaller trend, and again a smaller, until it stops showing. This way you’ll know the minimal strength of trend your tool can detect and “prove”.
I mentioned in an earlier post that there is another kind of autocorrelation that is operative here, which is jut from arithmetic. The trend from 2001-2011 is not independent to that from 2002-2012; eight years are common to both, and obviously influence the trend. You have said that 2001-2015 is a “slowdown” (which seems to mean trend<1 °C/cen). So it isn't really saying much more that all those sub-periods are also lower trend.
It's an odd fixation that you have there, and I wonder how long WUWT will want to indulge it. It really isn't surprising that the trend might dip below 1°C/Cen. That is still warming. You have watered the "slowdown" so far that it really isn't notable.
As commented above, you didn't say what dataset you are using, but it looks like the GISS that you have used before. I think a better visualisation is the triangle plot that you have experimented with. Here (from here) is one for GISS since 1960. It shows a color square for each of the combinations of months that you describe. The SW-NE lines represent constant duration periods, and I have marked in black the period of ten years. The color scheme shows a band around 1.7 °C/Cen in gray, and you can see why it shows up as the peak of your Eiffel Tower. Anyway, what you are calling the showdown is where it goes through a greenish region in periods ending after 2010, and even almost touches some dark brown spots (zero trend).
But it isn’t unusual. There are three other such regions in those 58 years – stretch a bit, and you could almost say they are regularly spaced. It really isn’t a big deal.
Hi Nick,
I do have a fixation. It is with the truth.
I have a fixation because warmists refuse to admit the truth, that there was a slowdown.
I will keep pointing out their lies, until they admit that they are wrong.
Once warmists tell the truth, then my job is mostly done. I will have to find something else to occupy my time.
You say that I have watered down the “slowdown” so far that it really isn’t notable. I can notice it, and so can a lot of other people.
I am making no claim that the slowdown lessens global warming in a permanent way.
It was a temporary slowdown. I think that it was caused by ocean cycles. We should acknowledge ocean cycles as a source of temperature and warming rate variation. We do with El Nino.
Why is it so hard for warmists to admit that we had a fairly small temporary slowdown.Answer that question Nick, and you will discover the “secret” of the slowdown.
I am using GISTEMP.
“Why is it so hard for warmists to admit that we had a fairly small temporary slowdown.”
It isn’t. We had a fairly small temporary slowdown. The puzzle is why anyone thinks that matters.
The triangle plot puts it in context. Slowdowns happen.
don’t worry sheldon, the slowdown coming with the drop in the amo will be far more noticeable . north sea received a big push of warm water that had been sitting north of shetland and off the west coast of norway at the start of uk winter due to the early winter north winds.
north sea temps are still high (ignore satellite temps, they measure skin temps,just a few feet down it is much warmer) but about to drop significantly as following the warm surface waters came a huge pulse of cold water from further north. in the space of two months bottom water temps went from 9 c to 6.5 c and stopped the nephrops fishery in its tracks.
the cold water is coming 🙂
Hi Nick,
you are the only warmist who has ever admitted that the recent slowdown exits. What does that say about warmists? And what does it say about you?
It matters, Nick, because the truth matters. If you don’t believe that the truth matters, then how can we trust anything that you say?
I see a poodle-like dog in the lower left part of the above triangle. The interpretation of that Rorschach Ink-blot test is that the dog days of summer will be upon us.
Sheldon,
“you are the only warmist who has ever admitted that the recent slowdown exits”
Nonsense. For example, A Scott below links to a paper in Nature by Mann, Santer and a galaxy of “warmists”.The title:
“Making sense of the early-2000s warming slowdown”
Hi again Nick,
You said “But it isn’t unusual. There are three other such regions in those 58 years – stretch a bit, and you could almost say they are regularly spaced. It really isn’t a big deal.”
I went back to 1970, you have gone back to 1960. That means that you have gone back to the Great Hiatus, which even warmists admit was a time of slowdown/cooling. You have put my slowdown in the same class as the Great Hiatus. Thank you Nick.
You don’t seem to be getting the main point that I am trying to make, Nick. It IS a BIG DEAL, mainly because warmists WON’T ADMIT THE TRUTH. The slowdown SHOULDN’T be a big deal.
It is like alcoholism. You have a huge problem when you won’t admit that you are an alcoholic. You still have a problem after you admit that you are an alcoholic, but at least the healing can begin.
Be the first one, Nick. Say “Hello everybody. My name is Nick, and I am a slowdown-denier. I would like to apologise to all skeptics for belittling their slowdown. It is humbling to realise how many years I wasted following the wrong path. I am now taking life one slowdown at a time.
RE: “Be the first one, Nick. Say “Hello everybody. My name is Nick, and I am a slowdown-denier.”
Beautiful! Made me laugh out loud! Thanks!!
Sheldon, Nick just said there was a slowdown. In fact, he agreed with the very words you used.
Hi Nick,
I have been thinking about your comment, that I have an odd fixation.
My original comment, that I have a fixation with the truth, is definitely correct. But I have thought of another aspect. Imagine that you talked to somebody, and they made a comment which you knew to be false (I am not talking about a mistake, they said that they did something, but you know that they didn’t). Would you trust other comments made by that person? Would you lend a (large) sum of money to that person? I would not feel comfortable lending money, and I would be suspicious of anything that they said.
I am willing to admit that the slowdown is relatively subtle. But there are simple ways that you can detect it. My graph in this article is one way. If you plot a LOESS line with GISTEMP data then you can see the line flatten around 2002 to 2012. You only see this if you tell the LOESS method to do the local regressions over about 10 years.
If you tell the LOESS method to do the local regressions over about 20 years, then you don’t see the line flatten around 2002 to 2012.
There are other ways to detect the slowdown, but we don’t need to talk about those here.
Let me be open about my views on global warming.
– I believe that CO2 is a greenhouse gas.
– I believe that humans are responsible for most of the increase in CO2 levels above 280ppm.
– I believe that the earth has warmed by about 1 degree Celsius in the last 100 years.
– The average warming rate from 1970 to 2017 was 1.78 degrees Celsius per century.
You can see Nick, what a huge denier I am (people have been calling me a denier since before 2009 – even though I held all of those beliefs the whole time). So now, when somebody calls me a denier, I write them off as a stupid moron.
Where I differ from a classical warmist, is in
– the issue of catastrophic global warming, and
– what we should do about global warming.
With my beliefs, I am agreeable to doing something about global warming.
I like the idea of renewable energy. If it can be made reliable, and not too expensive, then you won’t be able to stop me using it. I am prepared to pay a small premium for renewable energy, because it should benefit the planet, but I don’t want to pay 10x the current cost, and even 2x the current cost worries me slightly.
So I believe in the recent slowdown. It may not be big, and it may not be long, but I can detect it in many ways.
So why do warmists refuse to believe in a small temporary slowdown (which really poses little threat to global warming)?
I see it as a refusal to believe in anything which might go against the global warming agenda (so it is more of a political thing).
Because warmists deny a thing that I can plainly see evidence for, I think that they are lying, either consciously or unconsciously.
If they are lying about that, then what else will they lie about?
Remember that we are talking about major expenses dealing with global warming (billions of dollars), and major changes to peoples lives (often negative changes).
If warmists are not honest about everything (including a “trivial” slowdown), then how can I trust them about all of the big decisions?
Do you consider the possibility that you are plainly seeing something that isn’t there?
I suspect though the real problem here is you are talking about something being “really there” in a different sense to that of a skeptical statistician. Many things exist in the data, but a skeptic wants to test if that thing really is there and not just a chance event. Hence the need for statistically significant evidence.
+1
Excellent!
I’m impressed by the moose/poodle you have drawn in the lower part there.
But the rest tells us nothing other than the ciimate/weather changes.
…I have a fixation because warmists refuse to admit the truth, that there was a slowdown….
Hmm. I happen to agree with Mr Walker that there has been a slow-down, but I find the presentation of graphics to show this is rather diminished in authority if you declare your interest beforehand. Everyone knows that it is easy to construct a graphic that looks striking but is not necessarily the whole truth, and many warmists have done the same thing to push their view of the climate…
A much more convincing approach would be to discuss what would count as a slowdown, and try to get agreement over a wide group of interested parties, both skeptics and warmists. Does Nick Stokes agree that this is a good presentation, for instance? Only once you have got agreement is it worth doing the calculation and presenting the picture.
In practice, if a particular image or calculation wil not support one side of teh other, I suspect that opponents will not agree. At that point you can zero in on the actual mathematical manipulation which is the root of the disagreement, which is a much more productive exercise than creating a graph and then taking snipes at it…
Hi dodgy geezer,
so you don’t want skeptics to be able to present evidence for a slowdown. Will you also stop warmists from presenting evidence for global warming?
Your idea about “diminished authority if you declare your interest beforehand”, will encourage people to lie about their interests. Suddenly, I will become “Sheldon Walker the extreme warmist”. Then I can present evidence for slowdowns.
I think that you should analyse each piece of evidence as it comes. For example, look at my graph, and think about why the slowdown points are clustered on the left. Don’t just assume that the graph is showing the truth. Question how I made the graph. Did I do something to bias the graph in some way? Was it fair of me to pick the date interval 2001 to 2015 for the slowdown?
If people have the skills, I strongly recommend that they download the data and plot it themselves. If you can use Excel then that is all that you need. Excel is the only program that I use. If you plot your own graphs then you will have more trust in what they are showing. I also want to say, analysing the data and making graphs is a lot of fun. It is a bit like being an artist, and I consider many of my graphs to be a “work of art”. If you are using your “works of art” as evidence about global warming, then it is important that it is a true and real reflection of reality. I always try to follow that rule.
It is hard enough to just calculate a significant trend in 2001-2015, and you propose to calculate something about its derivative (specifically, its sign). Sorry, I don’t buy it.
The outcome will be just another hockey stick: a pure artifact of the calculation procedure.
“A much more convincing approach would be to discuss what would count as a slowdown, and try to get agreement over a wide group of interested parties, both skeptics and warmists. ”
You are a dreamer.
During the “pause”, it was just pointed out the physical fact that a zero speed was a stop. That is just basic definition, not presuming anything about what would happen next (as “pause” or “inversion” imply). Well, warmists denied that. I repeat: warmist cannot even agree to a very basic physical definition of the word “stop”. Says it all about them.
I was under the impression that theories were there to be challenged. In which case, there is nothing wrong with stating that intent.
Why would that be a surprise? By definition the years with the lowest trend will be on the left.
What became of the hockey stick and its time frame; surely it cannot still be seriously bandied about.
You know when you really get under Tamino’s skin when he deletes comments and observations that clearly contradict him.
The pause is his major obsession and he gets very upset when you point it out.
Which is rather funny because he gets extremely upset about something that he denies exists.
–
A pause is simply a flat line between two points in time.
A slowdown is a different beast.
Pauses always go from the present backwards and new pauses are always occurring in most time series.
The beauty of a pause in an uplifting temperature series is that it gets longer both ways once a dip in temperature occurs.
The more it dips the further back in time the pause can go.
–
Tamino simply refuses to believe his eyes and the observations because he is obsessed with the idea of AGW. Any admission of any pause raises the possibility that the theory may be wrong.
Hence he, and many other believers, are put in the position of denying actual factual observations and statements.
Despite numerous journal articles commenting on the pause and speculating on its causes.
In his case deleting posts that point out this simple fact.
–
The best attempt he jokingly makes is to denigrate it by calling it a fake or apparent pause.
He resorts to the only tactic available, taking a longer time period than the pause and saying, see, there is an upward trend hence no pause.
The buffoonery of this argument besmirched his otherwise fine mathematical brain.
In any time period that has an overall uplifting trend from the start to the finish choosing a period longer than a pause will obliterate it.
It does not mean that a pause does not exist, or that multiple pauses do not exist.
It does not mean AGW may not be true.
It does mean that someone is trying to be very precious in arguing for an unsustainable argument.
–
I object to the word slowdown in your article.
Pauses are pauses . The fact that they disappear with a new temperature rise does not mean that they have disappeared from the temperature record.
Depending on your data set you can go back and work out exactly how long they lasted.
A slowdown is simply a lower than the average trend.
–
A last word.
A pause can reappear and incorporate an old pause.
Hence if we did have a further 2 La Nina’s of reasonable size ( big actually) a new pause would occur that could stretch back to the old one and then beyond.
Fairly unlikely at the moment.
Still it is fun watching warmists deny an obvious and apparent and real pause by definition and therefore tying themselves in very untidy, awkward and unedifying knots.
–
Perhaps John Cook could do a new survey on how many climate journals mention the pause, whether they accept it or not. Probably 97%, but Tamino still refuses to get it.
Monty Python redux.
I think you better keep your ‘trends’ of exactly the same length. In effect you then determine the slope from that point onwards, for a fixed period of time. And then you can show how those mini-trends for the last 20 years tend to cluster to the left side of the chart, whereas most of the rest clusters elsewhere. Because you intend to show that the last 20 years was different, 10 years may work. But be aware that 10 years is quite short, longer would be much better.
More problematic: even if you can show that all of the last ones’ cluster on the (extreme) left of the chart, how do you prove that that is significant?
What are the odds that this is just caused pure by chance? And how do you calculate that,… correctly?
Remember that the underlying data is not normally distributed. Similar phenomena (like, say, weather, turbulence, even incidence & strength of volcanos) tend to have distributions with interesting properties such as imaginary ‘trends’, ‘too many’ extreme events, self-similarity (repeating similar behavior at multiple zoom-factors) and clustering in time and space.
Statistical “evidence” is based upon the question: is the observed results unusual under the null hypothesis?
The author’s null hypothesis a PRIOR trend exists. He then hopes to demonstrate that the observed result is unusual under the hypothesis that a prior trend exists. (i.e. the data is inconsistent with a trend and the prior trend has probably stopped.)
If the null hypothesis is poor, so is the resulting test conclusion. (Read about the statstical testing behind “New Coke” in the ’80s.)
A trend in historical data is NOT evidence of a trend unless you also assume a deterministically mean-reverting climate process (ARIMA models, Orenstein-Uhlenbeck processes, etc. fit in this category). But climate processes are demonstrably NOT deterministically mean-reverting and the null hypothesis presented here is invalid. As a result, the findings have no meaning.
Put another way, if I ask whether rabbit horns are long or short, either answer is nonsense.
The real problem is the historical temperature data has been so corrupted, all this is nonsensical.
As a believer in medium-run irregular cycles (AMO and solar) stacked with shorter-run irregular cycles (ENSO), I notice that there are many creative and alternative ways to approximate the admittedly complicated slope changes on these stacked cycles. This is yet another one and all of them will be updated eventually with actual declines in the slope term as indicated most clearly by the AMO turning point in progress.
Perhaps my brain hasnt awakened yet and joined me for coffee, but isnt it to be expected that the “slowdowns” would cluster at the low end of the ‘degrees warming per century’? I was completely unsurprised.
Graph 1 is the view from the east-facing window of an asylum room at Saint-Rémy-de-Provence 1889: Van Gogh’s “Starry Night”.
My subjective impression is that it looks more like a Hershey’s Kiss than the Eiffel Tower…
Tamino, I thought he was a Nuccitelli clone. Both a extremely dishonest people
For my part I’d like to see evidence of a “slow down” in global warming that actually demonstrated a slow down rather than focusing on piecemeal trends. You can find periods when the trend was slower of flat, yet as you say it makes no difference to the long term trend.
Take your 2001 – 2015 slow down. 15 years when the trend was less than half the long term trend. Yet long term trend was completely unaffected by it. The trend from 1975 to 2001 is almost identical to the trend from 1975 to 2015.
The point is that a lessening of the trend, slowdown, hiatus, or whatever you want to call it was not supposed to happen. The model projections were wrong. This is the ugly fact that Huxley identified as the bane of a lovely hypothesis. What the warmists don’t want to acknowledge is that the assumptions they made about CO2 to build the model were wrong.
A slowdown was not supposed to happen?. What do you mean? The modellers expected the temps to go up at a totally uniform rate – year after year? Simple experience and common sense would dictate that it wouldn’t happen that way.
I’m mystified by this subject matter.
Sheldon: If it looks like noise, and walks like a noise, and quacks like noise, why isn’t it noise?
Well, it isn’t exactly a bell-shaped curve. I suspect if you constructed artificial linear data 45 years long with a trend of 1.7 K/century and then added random noise (normally distributed noise) to each data point, and then repeat your analysis, you will get something that resembles your Figure. (If I remember correctly there is little autocorrelation in annual temperature to worry about. This wouldn’t work for monthly temperatures, which are highly autocorrelated. To change the magnitude of the noise include variable multiplication factor before adding noise so that you have different data sets with different amounts of noise.) Repeat your analysis for each dataset. You can do this with EXCEL if you add the free statistics package.
I’ll bet you observe something like your Efffiel Tower figure. The question is whether or not the extremes on the wings for short periods are much wider than expected from the tighter trends from longer data.
——————–
The other thing you could do is look at the confidence intervals for all of the 10-year, 11-year … n-year trends. If the average 95% confidence interval were +/- 1 K/century for 10-year trends, then all of the points within 0.7 to 2.7 K/decade would lie within this interval. Do more than 5% lie outside?
Sheldon: The question I’m trying to get you to ask yourself is not whether there was a slowdown. Any point even 0.1 K/century less than the 45-year average is technically a slowdown (and above is a speed up). Is there a meaningful slowdown – a statistically significant slowdown that requires some type of explanation besides noise?
Now, if you add enough noise to the suggested artificial data mentioned above, you can produce a spread in 10-year artificial trends as wide as you want. However, the spread in the 25+ year artificial trends will also widen and perhaps be much wider than you observe in your real data. If you can’t get the spread in 10-year and 25+ year artificial trends with a given noise multiplier to agree with observations, then something is causing a slowdown on the 10-year time scale that isn’t explained by random noise. Remember than 2.5% of data points in your random noise lie outside each end of a 95% confidence interval, so the one or two widest points will vary in your artificial data with the random noise.
The statistical test you need to do is Kendall’s Tau. Also known as Kendall-Mann (not that Mann, this guy was in the 1940’s.) In the test every point in the series is compared with every other point. Auto-correlation is not a problem, and no assumption is made about the shape of the trend. This test is common in the literature evaluating precipitation trends.
Test the series from after the El nino, la nina of the late 90’s to before the similar event of 2017. If there is no trend, there is no trend.
I thought “he who must not be named” was Steve McIntyre
I don’t see what you are trying to prove with the graphs.
The big spike shows that the longer the period the closer its rate will be to the rate for the entire period. That is what I would expect for almost any dataset.
The “slowdown” as defined by you is a period of small to negative warming, so I would expect those points to be on the left. This does nothing to prove that what you are defining as a “slowdown” is one.
IMHO the thing to do is start by defining what one means by a slowdown. Since the dividing line between weather and climate is generally regarded as 30 years I suggest one starts by determining the rate of change of temperature for each 30 year period, i.e. 1970-2000, 1971-2001 etc. A slowdown is then any period where the rate of change is less than that for the previous period.
A “who’s who” of some of the biggest names in catastrophic anthropogenic global warming scare mongering have already acknowledged and agreed the Pause is Real …
“It has been claimed that the early-2000s global warming slowdown or hiatus, characterized by a reduced rate of global surface warming, has been overstated, lacks sound scientific basis, or is unsupported by observations. The evidence presented here contradicts these claims.”
“John C. Fyfe, Gerald A. Meehl, Matthew H. England, Michael E. Mann, Benjamin D. Santer, Gregory M. Flato, Ed Hawkins, Nathan P. Gillett, Shang-Ping Xie, Yu Kosaka and Neil C. Swart”
https://www.nature.com/articles/nclimate2938.epdf?referrer_access_token=rO_LAj7Squh3f_qt6natqdRgN0jAjWel9jnR3ZoTv0OqExA1EwYluYLwiaayT9ble9FcNagQ1ss5L1V0KiWd-xzbFQjp8p3e-nUsgU7jNuUykRRWZpgMltUfROWf3xSKeGSSY7TvMiWdaeBCmNzlbQKCodQ3ivWje8eZYAs8Dr1uu8L-i3CHt8f_jYiil5eUpRpdxdWDCSCvqts_NYB_l8yUG-b6Qu0dtrZLMnaUyec%3D&tracking_referrer=www.nature.com
Sheldon,
Another plot to put it in perspective is just the 10-year trend over time. Here is is, shown relative to end year. The graph goes up and down, which should surprise no one. But the average is positive
Nick is one of the few warmists who acknowledges the pause, but then of course tries to prove it does not exist.
He has a number of blogs up on it on his site (type in pause on his site).
One of his tricks of diversion is to talk of 10 year time periods but either select only ones that do not show a pause eg starting one in 1996 or talking of overall trends for a longer period than the 10 year trend is in and pretending that the trend is therefore positive.
As above where he shows the 10 year trend graph over a range of 37 years.
Tamino’s trick.
Hiding the real 10 year and longer flat pauses in the verbiage of a mendacious longer time period.
Again, like Tamino, smart enough to know the pauses really exist for 10 year and longer periods.
Hypocritical enough to always avoid direct mention and acknowledgement of them.
It is this sort of chicanery thar desperate people resort to that stops us having sensible arguments about the valid points that they do make.
“As above where he shows the 10 year trend graph over a range of 37 years.”
These are exactly the years used by Sheldon. All of them. I was reproved in the triangle plot above for including an extra decade at the start.
So the question is: “… Which statement has merit? 1. “…how things are not always what they look like.” or 2. “If it looks like a slowdown,…” it is a lowdown.
Walker’s essay to objectively answer this question is interesting but requires a lot of mental contortions to understand. I prefer a simple analysis of the HadCrut4 time-series below. The green curve is the derivative (degrees C per decade) of the red HadCrut4 time-series trend line. The green curve demonstrates 2 slowdowns and 2 speedups since 1900. What more is needed?
All this proves to me (this article and it’s responses) as a common rank amateur with a solid technical college education is that … any set of cherry-picked data can be assembled into a graph which purports to depict … the “truth”. Lying eyes are nothing compared to the mathematicians at NOAA or NASA or a slide show by Al Gore.
I have lived 62 years in Northern CA … and as far as I can tell … the climate is same as it ever was, same as it … ever was. Some summers are blazing hot, others are fog-swept freeze fests. Some winters are as dry as Death Valley, others are monsoons of nonstop precipitation. Meh. The vast unwashed masses (like me) look at the Warmists -vs- rationalists as great popcorn entertainment … nothing more. Hint: check the opinion polls
What I am missing in the article is a clear statement about what the data actually are. Have I overlooked something obvious? The first descriptive sentence should in my opinion be a reference to the source, so that everyone who is sufficiently interested can collect the actual numerical values and carry out their own “peer review”. If I can’t find the numerical data I normally am not prepared to offer a comment.
So, what are we watching? Global averages, NH averages, SH averages Southern and northern tropics, NW Europe, continental USA, etc etc. I’d like to be sure before I invest time in replicating even a small proportion of this work.
Sheldon,
I would subjectively interpret this graph as showing the moderate warming of long length in time is the most common (mode) trend. Very high and very low warming are less common and tend to have short durations. However, the significant question that this doesn’t answer is whether the most recent period of slowed warming is the typical short duration type, or is a harbinger of a change in the pattern. We need to be able to predict the future from past events, not just summarize the past.
Vaguely resembles the Statue of Liberty.
The starting data is worth examining in detail. How was it derived for each year? Is there any real world significance in taking the max and the min on a day and dividing by 2? Is there any real world significance in taking multiple temperature readings over minutes, averaging them and then computing averages of those averages over a day? If average temperatures over time and/or geography do have a real significance are we taking that significance into account when we use the data for other purposes?
WARM WORDS
Even more chilling 10 years later – and we can now see evidence of its implementation all around us.
Pretty evil stuff imho
I am amazed that the discussion has continued when the OP analysis of data is flawed, and of no statistical value! The whole of “climate science” seems to me to be unaware of what can and cannot be done with “data”, and are prepared to use any method to produce false conclusions which match their political belief. I will give another example which is easy to understand, but is being used to make false claims:
We are told that air pollution is killing “tens of thousands” of people. This fantastic conclusion is produced by the following method:
Estimate how many days of life shortening due to “air pollution” is possible for an average citizen, which has no measurement procedure and is an estimate, lets estimate 20 days. Now calculate how many average lifetimes are lost by multiplying this number by the total population and dividing by the average lifetime. Answer so many thousands of people! Is this meaningful or a deliberate attempt to deceive? No death certificate has ever read “cause of death, Air pollution”.
This figure is then used by “pressure groups” to attempt to change our society in ways which are not necessarily good for the citizens, for example banning cars!
This is and always has been the kind of procedure that climate science is prepared to use on the rest of us too. It propagates falsities (and I will single out the term “greenhouse gas”) which have no scientific basis which can be demonstrated, and shows that those using the ideas do not even understand how a greenhouse (or whatever process they are comparing) works in the first place!
I will not even waste my time talking about computer modelling, but the BBC has just announced that its long term forecast has been extended from 7 to 10 days as the modelling has improved that much with acceptable accuracy.
My conclusion – If you hear any of the usual words like “model, statistics, greenhouse, CO2 level” etc. you should stop and do something else, it is more likely to be useful. And why do the “climate scientists” not go away and learn some mathematics and physics before launching even more nonsense at us, oh I know, its the money they get paid for the valueless research which they claim to carry out!
nice cherry pick … start at the end of a 30 year cooling 🙂 useless
Your math is a ‘running’ or ‘moving’ average. It is a filtering technique. It is designed to remove variation. Your graph demonstrates how the variation is reduced as you increase the number of years in the average.
You are so fixated on the ‘slowdown’ in the lower left that you don’t discuss the ‘speed up’ in the lower right. You know. The region where the warming reaches 3.5 deg C / 100 yrs. Funny how that is about as far from the average as 0.0 deg C / 100 yrs. Almost as if the triangle is telling us something about the variation that is being removed by the averaging.
Moreover, the triangle is just the base of the graph. It stops about 25 to 30 year averaging. I am not sure the dataset is large enough to say anything. But you are illustrating that 30 year averaging takes out the variations in the last 47 years of data.
Given that the global climate models predict these “slowdowns,” What do you think they mean scientifically?