This blog is mostly aimed at a source of criticism and fact checking for the blog 'real science' run by someone who goes by the name Steven Goddard. It is intended that material presented here is informative, neutral, impersonal and well sourced such that any of my claims can be checked and criticized in their own right if necessary.

Friday, 28 March 2014

Poor arguments used in climate science debates.

Merely my observations. A few common fallacies used when debating.

1) Quote mining:

Quoting famous scientists, authors, celebrities, the man on the street e.c.t in such a way to advance a point. Generally this is ill-advised as essentially anyone can be made to support any proposition by skilfully quote mining them, i.e taking the person out of context. Typically quotations are judged on how profound they sound, rather than actual utility. Instead of quotations which might be appropriate in an English literature exam, I would recommend a reference instead. Here is a webpage that gives an exemplar method of citation <> Obviously it is not expected in a forum setting, but a URL to a journal article is not too much to ask.

2) Appeals to conflict of interest or conspiracy

In the first case, such an objection could be justified. However it is not the case simply to assert a conflict of interest without justification. All scientific papers and journal articles are plagued by this phenomena, although the peer review process tends to minimize the effect for high impact journals, so be weary of open access, or articles not submitted for peer review. Certainly using conflict of interest as a way of simply discarding a journal article without even reading it, is completely inappropriate. This however, generally only applies to situations where confirmation bias is reasonably low which boils down to the peer review system once again. In the case of conspiracy of suppression, there are statistical means to implying publication bias for instance, however it is all too easy to assert a conspiracy without evidence, because by definition a conspiracy theory is un-falsifiable. Anything that can be applied without distinction to any publication is not a useful determinant of validity.

3) Giving undue weight to individual scientists.

The same applies to giving undue weight to a single scientific study. An argument from an individual scientist given to the public will seem convincing regardless of the position, usually because the scientist is arguing from a position of knowledge, whereas the public come from a position of relative ignorance. Thus appeals to the general public, and media are no substitute for the peer review process and the opinions of other scientists, and colleges in the community. Now, by the same token one should not simply dismiss the opinion of a scientist because he or she disagrees with most of his or her colleagues. However new and innovative ideas would still be examined by said colleagues. This is the concept of reliability, and repeat-ability.

4) Poor referencing, often to blogs or media articles.

Now of course, the reference should suit the purpose. It is fine to link to a newspaper that objectively reports on a phenomena in science; i.e scientific journalism. However, blogs and newspapers are not the place for original research and should never be referenced to in the context of a review of evidence (or indeed a 'systematic review' in the case of a paper published examining all the relevant evidence for a particular concept).  This holds true particularly if the concept or idea conflicts with current understanding. For instance, I could get away with talking about Newtonian mechanics by referencing to a informative website, or the general media - I wouldn't be expected to go back to Newton's principia. However in the case of paradigm shifting ideas, I would be expected to provide a more appropriate reference.

5) Ad homenin attacks, tone critisms.

Often these just annoyingly get in the way. In the first case, an ad homenin attack is critisising a persons character with the aim of refuting their idea. As always, an idea or concept is true regardless of the author. In the second case, tone criticisms are a specific example of an ad homenin attack. I.e attacking the presentation of an argument, rather than the argument itself. That is not to say tone isn't important, or cannot be critsised, but shouldn't be done so under a veil of refuting a concept or idea.

6) Arguments from ignorance - the big one.

Or arguments for intuition. Allow me to construct an analogy. If you push an object, it moves faster. If you push an object going pretty fast, it goes even faster. So if you push an object moving at the speed of light it will go faster surely?! This is a simple one that most people should be aware of, but counter-intuitive concepts are extremely pervasive in science. For instance, try and explain to me concisely why moving faster on a bicycle increases stability, and why a person might fall of if they were to move too slowly? Or perhaps, why drinking alcohol actually lowers your body temperature not raises it. Or why the sea ice in antartica is growing despite a generally warming southern ocean? I think I have made my point. Just because something is counter intuitive does not mean it is 'obviously' wrong. I would be weary of even using the word 'obvious' at all. But before making an argument from intuition, please consider how realistic the view is, that no climate scientist would have thought about such an 'obvious' objection.

7) Fitting evidence to the conclusions.

It is possible to tell sometimes when people are doing this. Please, before making an argument consider whether you have checked a reasonable body of evidence. And I don't mean 'from both sides of the argument'. The idea that simple dichotomies exist in climate science is incredibly naive, a more accurate reality is different models/methods producing different results that have different interpretations.

Tuesday, 25 February 2014

Robert Wagner article . Opening paragraph

So, I read this article that seems to have caught people's attention recently:

I'm going to have a look at some of the problems with this article, and offer a few refutations. Note I will be quoting from the article, but note this is allowed under copyright law as fair use.

"Science is the belief in the ignorance of the experts.
It is my understanding that this quote originates from Richard Feynman. I think it is important to give the full context of this quote. As far as I can tell it originated from a lecture to a teachers association in 1966. Now I have to admit, I have not yet been able to fully ascertain the reliability of the source I am about to give for the Richard Feynman lecture. However there are sources which do seem to corroborate the existence of this lecture as the origin of the quote. So with that caveat in mind I will give a bit of context to feynman's quote (courtesy of

"We have many studies in teaching, for example, in which people make observations, make lists, do statistics, and so on, but these do not thereby become established science, established knowledge. They are merely an imitative form of science analogous to the South Sea Islanders' airfields--radio towers, etc., made out of wood. The islanders expect a great airplane to arrive. They even build wooden airplanes of the same shape as they see in the foreigners' airfields around them, but strangely enough, their wood planes do not fly. The result of this pseudoscientific imitation is to produce experts, which many of you are. [But] you teachers, who are really teaching children at the bottom of the heap, can maybe doubt the experts. As a matter of fact, I can also define science another way: Science is the belief in the ignorance of experts.
The overriding context being a presentation to science teachers about the teaching of science and the danger of 'going through the motions' in a way that isn't critical, reflective or willing to question. I haven't managed to fully wrap my head around the actual quote. But I think the following interpretation provided by Wagner is dubious:

"Science isn't about joining the herd. Science isn't about confirming someone else's work. Science is about looking at the world, looking at the current explanation, deciding that the world is wrong and you are right, and then going out and proving it.
It seems to me like feynman would agree with some of this (incidentally its worth reading the proceeding paragraph to the one quoted by Feynman and preferably the whole thing), science isn't about joining the herd, that is definitely true. I would take issue with the 2nd bit though, because science is about confirming (or falsifying) other people's work. In fact the crux of the scientific method is the idea of empirical reliability. I.e if something is reliable, an independent scientist would expect to be able to achieve a similar result. If a result is not reliable, then it will not be accepted; and rightly so. A well known example of this was cold fusion, the results that were claimed could not be verified by any other scientists so the idea was subsequently dropped. As for the 3rd part, I would also take great exception to this. You do not 'decide' the world is wrong, you do an experiment, or observe a phenomena that casts doubt on conventional theories. And this has been done countless times, for instance Young cast doubt on the idea that light was composed of particles 'corpusculs'. We now universally accept that light does have wave-like properties. Of course Young didn't just decide the world was wrong, but his interference experiment did provide evidence against the corpuscular idea with the wave-like interference property.

"In real science the status quo is the null hypothesis to be rejected, not confirmed.
Yes, this is true. In science an alternative hypothesis is provided H1 'light is wavelike', and evidence is provided that can either confirm the idea is true in a certain confidence limit (for instance there is a 95% chance that the evidence in support of  H1  does not arise from chance - note in some aspects of modern physics confidence limits tend to be very high, and are quoted in multiples of the standard deviation i.e sigma, a '5 sigma' event corresponds to a likelihood of 99.99994%). If the experiment does not meet the agreed confidence limit then we are forced to revert to the null hypothesis 'default position' as it could be argued that the phenomenon can simply be explained by chance. If the confidence limit is met, then the null hypothesis is rejected, and the alternative hypothesis accepted.

"Never in my life have I seen scientists going out to prove the null hypothesis is true...except in the field of climate "science.
Well in this context I suppose the alternative hypothesis might be 'humans are responsible for at least some (significant amount) of the recent warming over the last few decades' (this is just for illustrative purposes, a much less vague hypothesis would in reality be used). Of course in order to confirm this alternative hypothesis within a certain level of confidence, evidence must be provided. Given that there is a consensus that anthropomorphically induced global warming exists, the implication would be that the standard of evidence needed to reject the null hypothesis has been reached. It is possible to find many scientific papers (although abstracts are usually all that is publically avalible) on basic search engines like Google scholar. For instance here is a link to a relativly recent paper in a high impact journal:

but since the view that humans play some role in climate change is so pervasive it is often more common to find articles like this highly cited one from nature:

which do not discuss the existence of AGW but rather go into specific detail about climate change, while implicitly accepting humans have some role in the main body of the paper. Most papers on climate change don't even express a position on AGW which is what the cook et al paper found:

Going out to prove the null hypothesis is true, would presumably mean showing there is no evidence for AGW. Apart from there being a complete lack of scientific journal articles that do that, isn't that what you are saying we shouldn't do? I am a little confused.

"It is called the "scientific method," something people that blindly accept the man made climate change theory apparently know nothing about.
Any blind acceptance is against the scientific method, regardless of the position. I would examine the evidence for man made climate change before claiming such a position is blind. This paper would be an excellent start because it is a meta analysis, i.e it analyses different research done using different methods to produce a representive idea about what is going on - in this case about climate sensitivity and humans.

"Like medieval inquisitors, supporters of climate change "science" don't debate the issue, they insult, intimidate, smear and ridicule.
So that is your alternative hypothesis then? Can you show within appropriate confidence limits that scientists that hold such a view are represented  by 'insult, intimidate, smear and ridicule'. Of course among individuals you should have no problem confirming the alternative hypothesis, but trying to justify this stereotype is harder. I could equally point to 'climate skeptics' that have done similar things, but I don't think it would gain anything. Anyway, this is an argument about tone; and at the end of the day a poor tone is unprofessional but it is still independent of the evidence and a valid argument.

"Real scientists are by nature skeptical, it is a defining characteristic of science.
Of course. So are you skeptical of a round earth, evolution, or the expanding universe? I am, and I would not accept any of them lightly, and I would dismiss them without evidence. But there is evidence for these three things, just as there is evidence for evolution. Part of being skeptical is weighing up the evidence that exists, an unexplained phenomenon that doesn't fit the evidence would be brought to the table as something to discuss, not ignore. That is why dark matter is such a big topic in physics, because our current understanding of gravity fails to adequately explain the rotation of galaxies. Evidence that is not explained by the AGW model would similarly pose an issue and therefore be readily discussed.

"Somehow in Orwellian fashion being a "skeptic" has become an insult, not a merit is climate "science." Skeptics are called "flat earthers," "deniers," and climate "heritics."

The act of being a skeptic is not the reason 'climate skeptic' has become an insult. The reason for this is because it represents throwing out a working model that explains reality well for no reason, just like throwing out a round earth would. So the analog is reasonable, climate scientists are similarly skeptical about a round earth. The crux of the matter is, there is no elephant in the room - rejecting the notion of a round earth would only be reasonable in light of paradigm shifting evidence. Experimental evidence that may arrive tomorrow, although admittedly highly unlikely.

"Skeptics are to be shunned and ignored, and ironically the ones who don't have science on their side.
On the contrary the opposite is true:

This represents some of the issues with the opening paragraph. I will en-devour to examine the rest of the article when I have time.

Friday, 27 September 2013

Some of my conversations with Goddard.

I commented directly on his blog a few times, on some of his posts. I will present some of them here. Now, I want to make it clear I did make mistakes at times and I was not always 100% correct in my tone or my arguments but I will try and address the mistakes I made here if I can too.

The first lot came from here

This was a comment on using stereographic projections and pixel counting to try and estimate extent. It is something I wouldn't really try and do, because the area distortion causes pixels to have different weightings. So a lot of ice could turn out to actually be less than it looks or vice versa. Note NSDIC accounts for this in their daily extent graphs.

Steven was actually using IJIS maps to estimate ice increase. This seemed particularly odd to me as IJIS actually provide the numerical extent values for instance:(

The commenter Chris was making a similar point to me about counting pixels. I don't know what the correction factor is, I did attempt to calculate it but am not confident enough to post anything much about it. Chris may-well be right, but I am not able to verify. However for the interested reader NSDIC provides extensive documentation on how it deals with stereographic projections: (

This below is the next post which I think was written in response to my comment:

My original comment was in reference to pixel counting of arctic sea ice extent maps. Steven and the NSDIC had some alleged email correspondence, and I am referring to Walt Meier here.

Steven is referring to the old 'DMI area' charts which can be seen here: ( However it is noteworthy that DMI does not recommend their usage anymore due to them not taking into costal ice. Newer DMI extent maps (note now 15% cap instead of 30%) are available here:

My point in the next post was a rather simple one. If the numeric data is avalible then why use pixel counting, a likely far cruder method than an algorithm such as IJIS.

The next comments were little more than me trying to get Steven to answer my question. I see little point in posting them, so I will just put them in text form.

Me: I take it from your non-response that you have conceded that I am right.
Steve: I take it that you don’t understand what extent and concentration mean, and are just wasting my time.
Me: Of course I know what concentration and extent are. Now answer my previous post. Stop insulting me and have the backbone to actually address my posts
Steve: I already did, but you are too dense to understand what I said.

The only vaguely useful thing here is the query about extent and concentration. I didn't know what this meant at the time. But now I presume he is talking about the difference between area and extent. The former including areas >30% concentration and the later 15%. I think later comments give a context that suggest the fact that ijis doesn't give 30% concentration figures (that I can find) is the justification.

I made a few more comments trying to get an answer, but I see little point in posting them as I got no response you can view them on his blog if you like.

My first comment was about a previous post Steven had made similar to the first link. I mention here the fact that some low concentration ice in 2009 had been apparently ignored in the Laptev area. This again is evidence for wanting a 30%< number. The comment about DMI is irrelevant to the criticism at hand which is why I made the 2nd comment below. 

Another conversation I had with steven is on this post:

I have commented on using these national geographic maps before. This was me directly addressing Steven about them.

My first comment was about using a national geographic map from 1971 to make claims about sea ice extent. As I have mentioned previously the national geographic map in question only mentioned 'multiyear ice limit'. I assumed here it meant maximum (though I turned out to probably be wrong about this as ill explain soon). Steven had used this map to make an overlay with one from 2013. I don't know what this actually was specifically, he didn't say. I suspect it was the daily arctic sea ice image from NSDIC September 9th. Because the green area in his image looks alot like September 9th sea ice extent, and the funny colored line would be from the NSDIC 1981-2000 average line. Also there are bits of false ice in all the places that the NSDIC daily maps would have. Of course, I cannot prove this. But the jist of this was, apples are being compared against oranges. The extent is not the same as 'multiyear ice limit' so any comparison is useless.

Anyway the commenting continues:

My first comment is a reference to steven last comment in the above paragraph. Steven then writes about ice increasing from '71 to '72. And actually this does seem to be true. Of course it is still not relevant to my point.

The claim that the national geographic map was derived from satellite data may well be true, but then again it might not. A citation was not provided in any case, and this doesn't change the fact we are still talking about MYI.


My comment here was refering directly to the national geographic map 'mutliyear ice limit' label. You can see a zoomed in image of this on an earlier one of my blogposts. Incidentally the 1971 national geographic map can be seen in full here: (

 It turned out we were both wrong (probably) on our interpretations. Although mine was plausible to begin with (i.e consistant with the map label). I think now that the label is actually referring to the minimun multiyear ice limit . The reason I think this is from looking at a similar map from 1983, which had a label to this effect:(

Even now the interpretation is still a little ambiguous. For instance does multiyear ice mean 2+ year old ice  or 3+ year old ice. The former is often called 'second year ice' and 3+ year old ice is called simply 'old ice' a term which is used exchangably with 'multi year ice'. In this context I am still not sure. Secondly does 'minimum' refer to the time when there is least MYI that has an age of precisely 2 years or 3 years (or even 1 year) depending on the definition, or is it like the NSDIC method where ice is aged by 1 year every September - in which case are we looking at the MYI before the aging or after?  These questions all have to be answered before such a comparison is justified.

This is more of the same. The link I provided was just the zoomed in label from before.

Steven is wrong here, the burden of proof is on the person that made the claim. I did not claim that 1971 was not of a lesser extent than 2013. I simply demanded evidence to show that it was, and I would not consider the image Steven provided as evidence; fairly reasonable given a few of the problems I have listed.

There were a few more posts then that I made about how the overlays were made and photoshopping. Which I don't think are interesting enough to post. Steven then asks me again to provide evidence that his claim is false. Again, I don't have to do this. One doesn't have to provide evidence that a claim made isn't true in order to ask that reasonable evidence is given in support of said claim. In reality, there is evidence that the sea ice extent in 1971 is much greater than 2013, but this isn't really the point.


This is me trying to get to grips with the interpretation of the label. As you might notice it isn't as lucid as it is here. Steven is actually right that the NSDIC ages sea ice every year. But it still isn't clear that say 1 month ice in October counts as first year ice like the NSDIC definition. I would guess it does, and my comment here probably isn't helpful but, again, this is still a question that needs to be resolved.

So instead I offer NSDICs interpretation of sea ice extent in '71; something I shouldn't really have to do to refute an inappropriate comparison (even if '71 had the same extent as 2013 it wouldn't necessarily make what Steven is doing valid).

So my comments continue in direct response to what Steven says:

The 1990 IPCC report can incidently be found here: (‎)

The paper I linked in, was referenced in the 1990 IPCC report and is relevant here.

The irony is, Steven is right in his last comment. This diversion is interesting but pointless. I don't have to show that '71 had a greater extent than '13 because Steven still hasn't justified the use of the national geographic map to make that claim in the first place.

It turns out it probably did mean minimum after all!

There are more comments of mine on realscience, but this is a fairly decent sample anyway. Now I think it would be quite interesting to compare the 1983 national geographic map with the sea ice minimum in 1983; since both are easily available (the advantages over the 1971 map are that the satellite data was more 'complete' after '79 and the national geographic map label was less ambiguous).

1983 Arctic Ocean Map

Note: of course I am assuming the white area represents the same in this map as in the '71 map. I cannot prove this, although it seems to be a fairly safe assumption. The point is that the fact checking of these labels is the responsibility of the claimant, not me. Incidentally the white area is pretty difficult to see in the Siberian section as it is on the very pale blue of the shallow sea. I recommend zooming into the map. To help, I will say that the most southern extent of the white area (in the siberian part) is about the same as severnya zemyla which is the north island towards the top of the map.

Now compare to the minimun extent of 1983. I am assuming this is roughly september 4th, which is the day where the CT sea ice area was at its lowest. Ill include two images of september 4th and september 12th anyway from 1983.

As expected they are similar. In either case they show significantly different distributions to the white area in the national geographic map. There is sea ice in the Laptev, and southern Siberian seas that completely closes of the N route (if we were to take the national geographic map to mean sea ice minimum then this is contradictory). The Beaufort too is a good place to look, the ice comes right to the coast in the CT maps.

Anyway I think this is at least enough to show that Stevens use of the maps is not enough to show that 2013 had a higher sea ice extent than 1971. A further issue, that I don't think I mentioned in the comments is that the 2013 sea ice extent was not used during its actual minimum.

Tuesday, 10 September 2013

You misunderstood the animation Steven.

Firstly sea ice is no where near 1971 levels; I covered that here: (

And the whole point of this GIF animation:


Is to illustrate the differences between trends and natural variability. Also considering this animation was made last year and the minimum of 2013 hasn't even occurred yet (let alone the whole of september) it should be of no surprise to anyone that the 2013 point is absent. As for the 1979 point being truncated? I don't know, but if this is your case against the point of this graphic; it is a pretty weak one. Add 1979 to this graph if you like, NSDIC has it slightly above 1980, it really is going to make very little difference to the animation, and it makes no difference at all to the point. A point which you keep missing, attested by the continued fixation of 2013 being above 2012 i.e regression to the mean.

Monday, 9 September 2013

Sea ice in 1923

Note that Gavin seems to be doing a pretty good job of defending himself:

Gavin asked whether he thought there was less ice in 1923 or today? Steven replied with a historical weather review article from November the previous year. The article was very interesting and talked mostly about how the sea ice had receded significantly compared to previous years, and of a warming. Observations such as open water to the 81st parallel were made. The problem here is; steven is not answering Gavins question. A sea ice recession in the twenties does in no way prevent sea ice extent from being less now than it was then, and detail from the actual newspaper doesn't help so much either (for example there is open water to the 81st parallel now on the Atlantic side). A direct comparision is needed, and this is difficult, as there is a reliance on first hand observations such as shipping data; unlike the satellite record now the conditions in the arctic are much less well known. However I give two sources that we may at least get an indication of the differences; the first is from the danish meteorological institute (DMI)

The white area is the presumed arctic sea ice extent in August 1923. It is significantly greater extent-wise than 2013 when taken on value. However the red areas are the most important, as they were direct observations taken during that year, so we can be confident they are reasonably accurate. We see a few interesting features like ice in the kara sea north of novaya zeymla and in the chuckini sea that definitely was not present this August. On this evidence alone we might conclude that 2013 does have less ice (in August at least). The longer term time series (some of which I posted in the previous post back this up). Notice that open water is present to 80N, and probably 81N (you can see the boundary clearly north of Svalbard). These days having open water north of 80 degrees is not that much of an achivement.

The other map is from the J walsh monthly series


More information about the J walsh series may be found here:

I think when you look at publications based on long term sea ice history (some have been given in the previous post) I think it is very reasonable to conclude the sea ice extent in 2013 is less than it was in 1923. Not that I think that the contents of this blog post would be enough to meet the standards of evidence, but I think time series constructed in journal articles are pretty rigorous and convincing (e.g m Kinnard et al., 2008).

Steven vs Trends

(quotes will be in italic font type)

A day or so after I published my take on the daily mails very flawed (and that is being generous) article, the guardian did an even better job and can be viewed here:

Steven has attempted to post his own rebuttal to this article. In fact very little of the entire thing was brought up, S only chose to 'debunk' a few points.

Alarmists like Nuttercelli are hiding behind what they call “the long term trend” in Arctic ice, as a way to avoid a serious discussion about the 60% increase in ice this year.
 Actually the article did acknowledge the truth of the 60% increase as can be seen in the following quotation:
While this factoid is technically true, it's also largely irrelevant.
The article makes a point of backing up this claim of irrelevance and that is the whole point of the article. I am not however above the very basic technical fact checking. What is your source of the 60% increase steven? Different algorithms spit out different figures, if you provide no source then it makes it all the more difficult to verify. I gave the daily mail the benefit of the doubt and used the NSDIC August figure, but I am getting tired of having to do this.

 The first thing to note is that NSIDC starts their graph at the century maximum in 1979. In the early 1970s, sea ice was much less extensive than later in the decade. If NSIDC included the entire satellite record in their graph, it would look like a sine wave, not a straight line.

The claim that 1979 is the centuary maximun is extremely dubious. I cannot find a reliable source that can confirm this claim. Steven links back to an early IPCC report, and a sea ice NH extent diagram that only goes back to circa 1972. Indeed 1979 does seem to be at a maximun, although a close inspection shows that this is only true for the winter period, and indeed subsequent years have probably higher maxima in August according to the diagram. The NSDIC also produced a long term chart which I have copied in below (hopefully this should address Stevens final sentence).


appearing to also contradict the claim that 1979 was a century maximum (if this is still not enough here is a further source that appears to disagree with stevens claim:  Even if it was a century maximun (citation still needed), there is a legitimate reason for it being the first point on the graph; that is when the reliable satellite record began so is certainly not arbitrary in stark contrast to taking 1998 as a start year in the global temperature record, a very strong  el nino year.
Another important issue is that plotting linear trends on short sections of a cyclical function – is junk science and mathematics. Even if August ice was at an all-time record high, the post-1979 linear trend would still be down, as shown in the demonstration purposes only graph below.
I think you might have misunderstood Anthony Watt here Steven. I don't mean to be insulting for no reason but you are confusing your mistake here:

with an appropriate place to take a linear trend line. Firstly August monthly sea ice extent averages plotted yearly are not cyclical in the same way that plotting daily sea ice extent are. In the later case there is a well defined period (a year), and amplitude (although it may vary) can quite easily be defined also. Now compare to the graph from the nsdic you are trying to debunk:

ScreenHunter_404 Sep. 09 09.36
No easily defined period, or amplitude. And it certainty does not look like a sine wave. There may be natural 'cycles' hidden in there to some degree, but that is certainly a conclusion that cannot come from looking at this graph alone, and if it is being used as a premise then, again, it should be sourced appropriately.

This quote from Anthony watts (on your article) seems strangely appropriate:
This is always a problem with graphing any cyclical trend, but the short length of the record (8 years) makes it more problematic than what would be seen in a 30 year record.
But whatever way you look at it Steven, the summer sea ice NH trend is still down (as it is incidentally in other months too). You are actually correct that introducing an anachronistic 2013 point above 1979 would not change the overall trend, if this had happened it would be an anomaly and it certainly would raise a few eyebrows. However the only way the long term trend is going to stop pointing down, is several years that continuously make it above the current trend. Otherwise such an anomaly is as meaningless as 2013 being above 2012 as the guardian article pointed out.