Satellites show no global warming for 17 years 5 months

richardscourtney says:
February 7, 2014 at 4:05 am
Well, quite!
And the Idiot, in the same post you referenced, quoted wikipedia!!! Apart from the probability that the article he quotes has been Connolley-ised to death, my son and daughter, who both graduated in the last 10 years, were told in no uncertain terms that any referencing of wikipedia in any shape or form would lead to a loss of marks (quite apart from a loss of credibility).

Mr. Dohbro says linear regression is not an appropriate method of determining the trend on stochastic data. However, it was developed precisely for that purpose, and it is one of many methods that might be used. Mr. Dohbro does not suggest what method he would prefer, besides saying that it should be non-linear. One might, for instance, use a higher-order fit, or one might use an AR(n) auto-regressive model, though the former would reveal little over less than half a century and the latter is unnecessary in global temperature records because for obvious reasons they exhibit no seasonality.
However, I have used linear regression because that is what the IPCC uses and that is what Professor Jones uses. As anyone versed in Socratic elenchus will appreciate, using the methods of one’s opponents in debate leaves them less wriggle-room when – as here – results based on their own methods are at odds with their conclusions.
Andy L says he suspects that the Great el Niño of 1998, more strongly represented in the RSS dataset than in the other four, may be partly responsible for levelling what might otherwise be a slight uptrend. He asks what the RSS trend is from 1999, after the Great el Niño event of 1998 had ended. The answer is that there is an insignificant uptrend of 1/25 K since 1999.
The self-proclaimed “Village Idiot”, in a characteristically spiteful, unconstructive and ill-informed comment, raises – as is its wont – a scatter-gun series of half-baked objections. It says I call myself a scientific non-scientist. Well, at least I attempt to examine the actual data rather than adopting some arbitrary standpoint that is as aprioristic as it is indefensible.
“Village Idiot” says I “prefer” the RSS dataset. No, as it well knows, I regularly post here using all five datasets, three terrestrial and two satellite. However, it is possible that over the past couple of decades RSS may be the most accurate of the datasets because it alone correctly represents the relative magnitudes of the Great el Niño, which was severe enough to cause widespread coral bleaching, and of the subsequent el Niños, which were not. I suspect, but have not yet verified, that the other datasets apply self-correcting dampers to their data to reduce the magnitude of sudden anomalies such as that of 1998.
It (the “Idiot”, that is) goes on to say that I have “constructed a graph that points down”. I have not “constructed” anything. I have merely taken the data off the satellites and plotted them to high precision on a graph. I have then determined the least-squares linear-regression trend, using two distinct algorithms to ensure that identical answers are given by both, and have plotted the trend-line in bright blue. As should be obvious, the trend-line does not “point down” much: it is as near horizontal as makes no difference. If the “Idiot” were capable of determining a linear-regression trend, it would no doubt be able to process the data for itself to verify that the trend-line on the graph was accurately determined. Before finalizing and using the 350 lines of code that generate these very clear temperature graphs, I had the linear-regression routine verified by a professor of epidemiological statistics to ensure that it was sound.
Next, the “Idiot” says he has looked at the 37-month running average on the five global-temperature datasets and has observed that they “peaked” – whatever that may mean – in 2006. Well, my algorithm is capable of plotting the average of any combination of the five datasets, and it shows that taking the average of all five datasets there has been no global warming for 13 years. The “Idiot”, when examining a trend to the present, ought not to have relied upon a 37-month mean, for by definition that will show a plot of what has occurred in the past three years.
Next, the “Idiot” cites a speech made by me in October 2009, in which I said that for the nine years since 1 January 2001 there has been “statistically significant and rapid cooling”. Sure enough, in the 8 years 9 months from 1 January 2001 to 30 September 2009 (my speech was on 14 October) the RSS dataset shows a statistically-significant cooling of 0.16 K, equivalent to 1.87 K/century of cooling.
The “Idiot” also says I said in that same speech that all global warming had stopped since 1995. And so it had – all but a statistically-insignificant 0.4 K warming over the period from January 1995 to September 2009.
The only statement made by the “Idiot” is that it does not believe me. But, as it will learn when it grows up, science is not a belief system. The “Idiot” need do no more than attend Statistics 101, as I did, and then obtain the RSS data from the link plainly shown on the graph, and then plot the data and determine and position the least-squares linear-regression trend line. When it has taken these necessary steps, it will be able to determine for itself whether or not the graph as displayed in the head posting is correct and fair, and it will find that it is.
The “Idiot” is by no means the first troll to make the imprudent and impudent assumption that because I have not received a certificate to the effect that I have received socialist “training” in science I do not know any. I am one of the few mathematicians in the world who has made substantial sums by using math in innovative ways. The “Idiot” – unless it is paid very well by those who spend so much time and money trying to discredit anyone who does not unquestioningly accept every daft tenet of the New Religion – would be better off going to play in the Little Leagues in future. It is not man enough, and not knowledgeable enough, to take an effective or constructive part this debate. Indeed, it does not even have the courage to reveal its own identity, preferring to snivel furtively behind an admittedly appropriate pseudonym.

If Gary Mount would like to let me have his email address, I shall be very happy to send him a high-resolution version of the graph showing no global warming for 17 years 5 months. The narrow width of the standard WordPress columns at WUWT does not allow graphs such as this to be displayed as well as one might like.

The PDO flattened it and the PDO+AMO combo will drive it down. Add in solar cycle and depressed cycle 25 for good measure.

Kathy says:
February 7, 2014 at 6:05 am
17 years and 5 months is an odd interval. What does the data show about warming over 20 years, 15 or 30?
Hi Kathy,
You can calculate the linear trend between any 2 points, and more specifically from any given starting point up until the present. If you do this exercise from January 1979 until the present, you get a positive trend of 0.126 degC/decade. If you start your analysis in later and later months, the calculated trend fluctuates up and down a bit, reaching a high 0.149 if you start in December 1983. Thereafter it meanders generally downwards, and the first starting month for which it goes negative is September 1996, which is 17 years and 4 months before the last month of January 2014.
Whether you should count both the start and end months in your time interval is the subject of another discussion. Should the trend be regarded as being from 1st September 1996 until 31st January 2014, then that is (almost) 209 months. If you only regard the trend as being between the midpoints of the months then it is 208 months.
Regards
Richard