Comments 1 to 33:

1. CD at 08:42 AM on 21 June, 2020

   I noticed that one of the proposals of this article is more solar power. Yet solar power decreases the Earth's albedo and thereby increases local surface warming. This warming (even without CO2 greenhouse emisions) can be as high as 1 degC as I note here - climatescienceinvestigations.blogspot.com/2020/06/14-surface-heating.html

   In thermodynamics there is no such thing as a free lunch. All human activity warms the planet. So what is the plan? Go back to prehistoric living and get everyone to live in a cave?

   0 0
   Moderator Response:

   [DB] Sloganeering snipped.  Please comport further comments to comply with the Comments Policy.

2. Tom Dayton at 09:12 AM on 21 June, 2020

   Slarty Bartfast, that is a very old, and long discredited, myth. See, for example, this RealClimate post.

   0 0
3. CD at 11:21 AM on 21 June, 2020

   @ Tom Drayton

   Yes I've read the RealClimate post and it makes the same point I made in my <a href="https://climatescienceinvestigations.blogspot.com/2020/06/14-surface-heating.html">recent blog post</a>, that the waste heat is only about 0.03 W/m^2, or the equivalent of a global temperature rise of about 0.013 K. This sounds trivial until you realize it isn't spread evenly across the Earth's surface. The point I made is that when examined on a country-by-country basis, this heating can become very large, e.g. 1.0 K for Belgium & the Netherlands, 0.66 K for England and 0.5 K for Pennsylvania. That is not trivial and it accounts for almost all the temperature rise seen in these states/countries in the entire 20th Century. I'm surprised that doesn't make you stop and think.

   0 0
4. Eclectic at 21:59 PM on 21 June, 2020

   SlartyB @3 , the combined areas of England Belgium Netherlands & Pennsylvania come to less than 0.07% of global area.  ( And most nations of the world are much less densely populated.)  So your hypothesis of "waste heat" being a major part of Anthropogenic Global Warming . . . is sailing itself into a very stiff headwind !   Perhaps you will be kind enough to "show your working" for your supportive arithmetic?

   ( Off-topic :-  In another thread, MA Rodger has indicated your arithmetical blunder wrt sea level rise from thermal expansion.  Your other assertion ~ that the AGW forcing of 0.6 or 0.9 watts/m2 is far too small to produce significant sea level rise in the coming 100 or 200 years ~ also seems to be in error.   For myself, a rough back-of-envelope calculation shows that only about 1% of the AGW forcing is sufficient to cause 2+ mm/year of sea level rise.  Which fits in with the mainstream science.  And which leaves plenty of scope for an accelerating rise in that near future. )

   0 0
5. Eclectic at 22:02 PM on 21 June, 2020

   Post Script explanation ~ my final comments @4 are referring to icemelt.

   0 0
6. MA Rodger at 23:14 PM on 21 June, 2020

   Eclectic @4,

   I fear the stiffest 'head wind' the grand theorising of Starty Bartfast on Energy Use faces is the history of these primary energy uses he employs so boldly.

   He employs the 2018 Primary Energy Use of UK which is given as 177Mtoe,  equivalent to 0.97Wm^-2 for the UK land area. I'm not sure of the conversion to a +0.42°C temperature rise for the UK, but taking that ratio, the UK would have been subject to a cooling of -0.02°C since 1965 (using OurWorldInData energy numbers) and a cooling of -0.11°C since Primary Energy peaked in 2003 (using the same source as Slartibartfast). The thermometers have evidently not been informed of this as, from the HadCET annual data where we should be the farthest from any outside influence, the annual CET averaged 9.5°C in 1965, rose to 10.5°C in 2003 and by 2018 was 10.7°C, thus giving no sign of any cooling associated with changing levels of UK Primary Energy Use.

   0 0
7. Eclectic at 00:42 AM on 22 June, 2020

   MA Rodger @6 , I gather that the "industrial energy" converted into temperature rise ~ is calculated according to Slarty's own special system.  I have been enjoying reading some of Slarty's blogsite, but I have only dipped my toe into it, so far.  He displays a great number of algebraic equations, which I have (perhaps wrongly) not looked into ~ this is a failing of mine, deriving from my past experience of the reams of equations publicized by Lord Monckton (the Moncktonite mathematics suffer major revision every so many months . . . yet always lead to absurd conclusions).

   Slarty seems to calculate on the basis that any industrial (i.e. Anthropogenic) heat energy produced in [say] England, will remain within the national borders.  No flow of wind or water across those borders, nor any transfers per evaporation/condensation.

   There are other peculiarities in his blog.  He states that the Milankovitch cycle produces a 10 degreeC oscillation of global temperature.  Perhaps he thinks Vostok represents the entire planet.  Also, he seems to feel that the CO2 in the atmosphere produces "Greenhouse" by reflecting infrared radiation back to the Earth's surface.

   There were one or two other points he made which seemed in error, at my first glance at his blog : but I've forgotten what they are, now.  Perhaps I can dig them out later.  Of course, his blog may not be quite as bad as I first gathered ~ I may have been mistaken in my own thoughts, and too hasty in my skimming, and some of the errors may be more a matter of him expressing himself in an odd way or through excessive abbreviation of ideas.  Still, it's always a red-flag worry when the earnest blogger seems to arrive at a different conclusion than the world's scientists.  There's usually some blunder at the bottom of it all.

   0 0
8. nigelj at 09:32 AM on 22 June, 2020

   Slartibartfast appears to think global warming is caused largely by the heat output from industry, transport and electricity generation. A simple google search shows global temperatures were exceptionally high during the first half of 2020, the period of covid 19 lockdowns and reduced heat output from transport, electricity generation and industry. Wheres the cooling trend his theory would predict?

   0 0
9. Lawrie at 11:34 AM on 22 June, 2020

   Any process that absorbs visible light decreases albedo. Photosysnthesis, the process that produced fossil fuels, would seem to be way ahead of solar power. Following on from Slarty's logic then the quickest way to reduce global warming would be to clear fell all the earth's forests and replace them with reflective concrete

   0 0
10. Eclectic at 20:06 PM on 22 June, 2020

    Lawrie @9 , Slarty Bartfast maintains that there is no global warming of any significance at a statistical level or at a physical planetary level.   So to him, albedo is irrelevant.

    Being more than 24 hours since his last posting, it seems unlikely that Slarty will return to attempt rebuttal of criticisms of against his many positions.  But we can hope he will return, to give a grand explication of his apparent errors and inconsistencies.

    In order to save the valuable time of SkS readers, I have looked further into Slarty's blog of May / June 2020 , and I have pulled out some points of interest.   Slarty's statistical/mathematical skills are (IMO) far exceeding his climate science knowledge  . . .  and somehow I am reminded of the very emeritus & climatically-challenged Ivar Giaever !

    I have taken some care not to misrepresent or quote-mine Slarty.   And please note that Slarty, in his blog, describes himself as: physicist / socialist / evironmentalist.

    1.   Sea level rise cannot be more than slight , because there is no CO2-AGW or CO2-led Greenhouse effect.  And so our coastal cities have zero danger of submersion.

    2.   What little CO2-greenhouse effect is present now, is produced by CO2 reflecting IR back to the planetary surface.

    3.   Weather stations fail to give valid planetary data because they are far too few, and (just as importantly) they are not evenly spaced.

    4.   "temperature records just aren't long enough ... to discern a definite trend ... you need at least 50 years."

    5.   "[land ice] In Antarctica (and Greenland) this is virtually all at altitude (above 1000 m) where the mean temperature is below -20 C, and the mean monthly temperature NEVER gets above zero, even in summer.  Consequently, the likelihood of any of this ice melting is negligible."

    6.   AGW forcing does not supply enough heat to melt ice at the poles [he seems to include the Arctic, too].

    7.   The Arctic is not warming.  [Presumably news to those alarmist Inuit who live there.]

    8.   Berkeley Earth Study repeats the sins of Hadley/ NOAA / etc but in a more transparent way ~ and BEST generates a falsely-positive warming trend through its misuse of Breakpoint Adjustments (rather than using raw data).

    9.   Slarty's oceanic thermal expansion calculations are wrong [as pointed out by MA Rodger].

    And there's more !

    0 0
11. CD at 00:28 AM on 23 June, 2020

    @7 Eclectic

    "Slarty seems to calculate on the basis that any industrial (i.e. Anthropogenic) heat energy produced in [say] England, will remain within the national borders."

    If that were true then the heat would never escape. It would build up day upon day, and the temperature would be rising by far more than 0.66 °C per century. If the heat was just retained in the air over the UK, then the air temperature would increase by nearly 3 °C per year. In reality this heat will slowly diffuse to the rest of the planet and then escape, but by the time it has done so it will be replaced by new heat production. Therefore there will be a steady state temperature gradient created between the heat producing areas and the colder areas. There will not be a uniform temperature rise everywhere.

    By the way, thanks for pointing out the error in the thermal expansion coefficient. I used the wrong one by mistake. That blog post has been corrected. But this doesn't change my overall point, that sea level rise is miniscule and unmeasurable and a long way from what most media stories would imply. It is not going to submerge major cities in the next 100 years.

    0 0
12. CD at 01:57 AM on 23 June, 2020

    @10 Eclectic

    You claim that I said on my blog: "7. The Arctic is not warming"

    I don't believe I did. I did say the South Pole is not warming. In fact the temperature record for Amundsen-Scott shows a slight cooling since 1957. Even Berkeley Earth agree on that (sort of).

    What I did say about the Arctic is that we don't know what the temperature trend is because there are no long term weather stations within 1000 km of the North Pole, and there never have been.

    By the way, the other thing you omitted from your precis was Post 9 - Fooled By Randomness, where I demonstrated that changes in temperature of up to 1 °C per century are entirely possible due to natural variations resulting from chaotic behaviour within the climate system.

    0 0
13. MA Rodger at 05:51 AM on 23 June, 2020

    Slarty Bartfast @12,

    To be exact as to what you wrote on your error-filled blog site, regarding the North Pole you wrote "there is no evidence of warming at the poles ... there is no weather data within 1000 miles of the North Pole and never has been."

    If you consider that to be a factual statement then you are a bigger fool that I thought.

    There is directly measured evidence of warming of the North Pole as we have satellite data showing it. Additionally there is much more indirect evidence, not least the dozens of met stations which operated within 1,000 miles of the pole and which include those located within 1,000km of the pole.

    And I am curious why you say that a global averaged temperature trend of +1 °C per century is "entirely possible due to natural variations resulting from chaotic behaviour within the climate system". Indeed on your blogsite you write "most of what you see in the smoothed and averaged temperature data is noise not systemic change (i.e. warming)" [my bold].

    So my question, Slarty Bartfast, concerns the likelihood of  temperature records not measuring what everybody else says they do. You present a crazy tale which you say proves that a random chaotic source is "entirely possible" to be what is being measured as being a global warming signal and then, a big leap here, you assert this situation "is" actual and not merely "possible". So I ask, is this actual situation dependent on you pressing the 'go' button on your Infinite Improbability Drive? I ask because your slap-dash and ridiculous thesis does stretch credulity to breaking point.

    0 0
14. nigelj at 06:47 AM on 23 June, 2020

    Slarty Bartfast @12

    "But this doesn't change my overall point, that sea level rise is miniscule and unmeasurable and a long way from what most media stories would imply. It is not going to submerge major cities in the next 100 years."

    That is a very bold assertion somewhat lacking in any evidence. And tell that to people living in Florida as discussed here. Nothing miniscule in these sea level rise numbers. People are already having to jack up their houses in some places to alleviate the problem. And yes there are multiple causes of sea level rise in Florida, but climate change is by far the main one as stated.
    
    There are plenty of examples from earths past where sea level has risen 2 or 3 meters per century at similar warming rates to presently, eg melt water pulse 1a after the last ice age. We are at risk of triggering a similar event but in shorter time frames.

    I suspect Slartys  rejection of anthropogenic global warming is because he is afraid that climate mitigation costs will hurt poor people. He does say hes a socialist. So tell me Slarty , are you worried about the costs of climate mitigation on poor people?

    IMHO theres nothing wrong with concern for poor people per se, or some light form of socialism, just that its very wrong to think climate mitigation has to hurt poor people. For example, there are simple and obvious ways of structuring things like carbon taxes to avoid this.

    0 0
15. Doug Bostrom at 06:56 AM on 23 June, 2020

    Waste heat on Skeptical Science: https://skepticalscience.com/waste-heat-global-warming.htm

    197 back-and-forth instances there in discussion.

    But why warm the air further? It's not waste heat. 

    [I believe it's still the policy here not to create a tossed salad in discussion threads. Every avenue Slarty is probing is already well developed in existing discussions here. Reduce, reuse, recycle.]

    0 0
16. nigelj at 07:34 AM on 23 June, 2020

    Slarty Bartfast @12

    "What I did say about the Arctic is that we don't know what the temperature trend is because there are no long term weather stations within 1000 km of the North Pole, and there never have been."

    History of climate monitoring in the arctic here. According to the article land based observing stations around the arctic circle were established in the 1880's, giving data on greenland and northern russia and the various islands etc. There were multiple land based and drift stations over the open ocean, including close to the north pole, established from 1960 - 1990. Since that period there have been fewer weather station, and more reliance on satellite data.

    I guess it depends on how you define "long term data" but the article shows there is good data for the whole of the arctic from 1960 - 2020 a fairly long period of time.

    0 0
17. Eclectic at 12:00 PM on 23 June, 2020

    Doug @15 , I take your "tossed salad" point that individual topics of discussion should belong in their own threads, where each can be discussed in depth (and with the historical input from comments of earlier months/years).   That is the excellent SkS system, which works moderately well.   The alternative is a chaotic repetitious churning of multiple topics & distractions, month after month on every page ~ which sabotages the educational purpose of the SkS website.

    OTOH, the SkS Weekly News Roundup threads are somewhat to the side of the standard indexable SkS system.   The Roundups are each a potential hodge-podge of comments . . . which are quickly swept away into oblivion (and this ephemeral nature allows for looseness of topic, and even permits a political tinge at times).

    The commenter Slarty Bartfast has brought his blog to SkS, in effect to promote it and also (just possibly) to solicit comments & criticisms of it.

    In a way, Slarty's blog is suited to a one-week Roundup.   His blog contains so many errors of science and logic ~ each error being so plainly obvious, that it merely needs pointing out rather than detailed rebuttal.

    Possibly that may have a salutary effect on Slarty's thinking, and he will make the effort to educate himself about climate science (unlike Ivar Giaever).   Or possibly it won't ~ if he is unable/unwilling to disentangle himself from his prejudiced Denialist mindset.   [ Slarty, my apologies if that comes across with a patronizing tone . . . but in the circumstances, such a tone is difficult to avoid entirely.]

    IMO, Slarty is an intelligent guy: but we all know of many intelligent people who let their emotional bias override their intellect (especially with the climate science deniers ! )    Slarty, I wish you well, with your internal struggles for objectivism & insight.

    #  In other words, Doug, it could be desirable to corral all of Slarty's ideas into a single location [here] ~ where they can be "lightly cauterized".   And then move on to weightier matters.

    0 0
18. Eclectic at 19:13 PM on 23 June, 2020

    And while the Weekly Roundup iron is still hot, I will give some more points extracted from Slarty Bartfast's blog :-

    Slarty, please note that you are most welcome to correct me wherever you think I have made an error.   I have quoted some of your phrases verbatim , but mostly I give what I believe is an honest gist of your blog messages.

    For instance, you will see (above) in post #10 Point 7.  where I wrote The Arctic is not warming ~ and yes, those were not your verbatim words : but they are your exact meaning [see confirmation by MA Rodger @13 ].    Likewise with my other comments, I give the gist of your messages (and if you look closely, you will recognize some "re-cycling" of some of your own phrasings and word-choices).

    Slarty , let us proceed !   You may find it uncomfortable.  But all publicity for your blog is good publicity . . . as the saying goes, eh.   And for convenience of style, I will refer to you in the Third Person.

    A**   "[the AGW] that climate scientists think they are measuring is probably all just low frequency noise resulting from the random fluctuations of a chaotic non-linear system."  

    [ The catchy phrase of climate being "a chaotic non-linear system" ~ was quote-mined from an IPCC report.  The phrase is a half-truth, and is a misrepresentation often quoted in science-denier blogs . . . where most of the readers are clueless about its precise meaning in climate physics. ]

    B**   Modern global warming is largely just the result of a non-anthropogenic 150-year oscillation in global surface temperatures.

    [ But then again, the warming is "not there" anyway ~ because the temperature records fail statistical significance, it seems?? ]

    0 0
19. Eclectic at 20:33 PM on 23 June, 2020

    . . . . . continuation of post 18.

    C**   It is interesting  that for someone whose blog complains that climatologists fail to appreciate the salient importance of noise in records of temperature, Slarty nevertheless seeks to disprove global warming by citing a very noisy & limited record (the historical tornado record for that section of the Earth known as the USA. )

    D**   Slarty makes handwavy cherry-picking of 4 glaciers in New Zealand.

    [which have very dubious support for his assertions]

    E**   "CO2 is a greenhouse gas ... [but] this does not mean that increasing CO2 levels must lead to an increase in temperature."

    [curiouser and curiouser! ]

    F**   Slarty gives some old chestnuttery ~  the by-proxy denying of mainstream climate science, by strawmanning with the apocalytic hyperbole coming from the ExtinctionRebellioners and suchlike non-scientists .

    G**   They [alarmists? scientists?] want to "get everyone to live in a cave".

    [Actually a quote from upthread here : but a definite red-flagger emotionally.]

    H**   It is interesting  that a self-described Environmentalist claims that he cannot decide on "the optimum surface temperature of Planet Earth".

    I**   The cognitive dissonance of holding mutually-contradictory positions (or at least, claiming to hold them).  And some of the positions are quite unphysical.

    [[ Note the word "unphysical", Slarty.   That is the rock that sinks the ship of your statistical analyses of the climate situation.   You have failed to grasp what is happening at the level  of molecules / atoms / hadrons / photons.  ]]

    0 0
20. Lawrie at 21:28 PM on 24 June, 2020

    Slarty Bartfast 11 States:  But this doesn't change my overall point, that sea level rise is miniscule and unmeasurable and a long way from what most media stories would imply. It is not going to submerge major cities in the next 100 years.

    According to NOAA - https://www.star.nesdis.noaa.gov/socd/lsa/SeaLevelRise/LSA_SLR_timeseries_global.php . NOAA/NESDIS/STAR provide estimates of sea level rise based on measurements from satellite radar altimeters. Plots and time series are available for TOPEX/Poseidon (T/P), Jason-1, Jason-2, and Jason-3, which have monitored the same ground track since 1992, and for most of the altimeters that have operated since 1991, including T/P, Jason-1, Jason-2, Jason-3, ERS-2, GFO, and Envisat.

    Only altimetry measurements between 66°S and 66°N have been processed. An inverted barometer has been applied to the time series. The estimates of sea level rise do not include glacial isostatic adjustment effects on the geoid, which are modeled to be +0.2 to +0.5 mm/year when globally averaged.

    Since 1992 Sea levels have risen by about 80 mm and there is no evidence that the rate of increase is slowing down. Why do scientists continually need to refute the kind of pseudo-scientific nonsense emanating from people like SB?

    0 0
21. MA Rodger at 21:58 PM on 24 June, 2020

    Eclectic @17, 18 & 19,
    I disagree with your assessment that Slarty Bartfast is "an intelligent guy". In my assessment, he has little-to-no understanding of the scientific concepts he wields, hardily a mark of intelligence. And if you dig deeper, you will find the work of Slarty Bartfast has a 'fractal' property, in that crazy assertions and error are used to support further crazy assertions and error.

    …

    With Slarty Bartfast not appearing eager to explain that most prized of his pronouncements within his grand debunking of AGW, perhaps it is beholden on me to explain to Slarty Bartfast what he has managed to have gone and done (or at least the central blunder thereof). This concerns his grand proof that most of what is seen as AGW is but random noise. As I asserted @13, this is nought but Slarty Bartfast pressing the Infinite Improbability Drive button.

    On his blogpage 'Fooled by Randomness', Slarty Bartfast takes three lengthy NZ met station records of monthly temperature anomalies and calculates the distribution of these anomalies about the mean – that is of course the Standard Deviation and that works out at roughly 1ºC.
    (The exact derivation of his 'mean' is not explained. The data from Xch NZ 1864-2013 shown graphed is the raw station data but the SD graphed is not the simple SD of the 1,796 data points.)
    The monthly-data exercise is then repeated for tri-monthly, half-yearly, annual, bi-annual, 5-year and 10-year averages and these seven SDs are plotted against length-of-data-point on a log-log scale and found to produce a nice straight line (shown below).

    

    This straightness allows Slarty Bartfast to extrapolate the relationship to obtain a value for the SD of century-averaged anomalies which is found to be a sixth the SD of monthly data (thus SD≈0.167ºC). And having derived the size of the SD, the spread of the data will be also a sixth as great. This allows Slarty Bartfast to infer that the scatter/spread of a set of century-averaged data-points would be “almost 1°C,” thus a value pretty-much the same as the level of AGW over the last century. (And it will help to say that for Slarty Bartfast, this 'scatter/spread' results from random “noise”.)
    The next bit is implicit within Slarty Bartfast's argument. With the spread from noise being 1ºC, what if today's century-averaged anomaly is at the top of the noise-range and last century's at the bottom? That would account for all this AGW nonsense!!
    This revelation Slarty Bartfast declares @12 to be “entirely possible” although he went further on his shabby little blogsite by first saying it is “probably” so and then concluding that it actually is the case that "most of what yuo see" as AGW is just random noise.

    This, of course, is eye-bulgingly stupid in so many ways but in terms of the Infinite Improbability driving Slarty Bartfast's anti-AGW missiles, the probability of a particular data point sitting 3SD above the mean is roughly 600:1 and the probability of its preceding data point sitting 3SD below the mean combines to a 350,000:1 likelihood. I would suggest that is essily improbable enough for something as inert as a bowl of petunias to think “Oh, no. Not again.”

    But for Slarty Bartfast, such things are real and actual because all you need to do is believe.

    0 0
22. nigelj at 07:30 AM on 25 June, 2020

    MAR @21, I would disagree that Slarty Bartfast is totally unintelligent. Slarty Bartfast would probably be intelligent in the sense of having an above average IQ given his technical abilities. However he applies these abilities in a very shoddy way (as you yourself imply) to construct his house of cards of AGW denial.

    At the very least he lacks much wisdom or quality control of his own reasoning or there is some deliberate stupidity being applied. The question is why is he so sloppy or deliberately stupid?

    Firstly it is hard to believe he is lazy, given the time he has spent constructing his elaborate house of cards. 

    Secondly its possible he is a crank. "Crank(person)" on wikipedia has an interesting definition and symptoms, and he fits some but not all of these symptoms. So hmmm. Although you will note there are some similarities to Victor.

    en.wikipedia.org/wiki/Crank_(person)

     

    Thirdly Slarty mentioned that hes a socialist. Some individuals have emerged who stridently oppose the AGW consensus and climate mitigation even although they lean left and liberal politically. This is odd, given those on the left tend to be more accepting of AGW than those on the right, according to various polling studies, eg by Pew Research. But some of them have expressed concern of how climate mitigation would hurt poor people and this surely explains their scepticism of the science, especially as they are mostly educated people and not born ignoramuses.

    I suspect Slarty fits this definition of being worried about how climate mitigation might impact on the poor (wrongly I think because its easy to construct things like carbon taxes so they exclude low income people or give them a rebate of some sort). He has run away and not answered the simple question I posed on the matter, which suggests I'm probably right.

    Or there is a fourth possibility that he has slotted in the labels of 'environmentalist' and 'socialist', merely as a tool to help convince those on the left of the veracity of his crack pot ideas.

    Or fifthly, reading Douglas Adams very entertaining series of novels has addled his mind. 

    I did some psychology at university, so we studied human motivations and they intrigue me, although my main degree is in architecture.

    But I think the most likely possibility is Slartys unjustified concerns over the impacts of climate mitigation on poor people have lead to an attempt to find flaws in the science, and this in turn has pushed him towards the definition of a "crank".

    As to Slarty's views that glaciers haven't melted much in New Zealand, and this is because we have "low population density" and so not much "waste heat" as a result. I should say something given I live in the country.  NZ's  glaciers have in fact lost 33% of their ice mass since the mid 1970s (when monitoring began) which is obviously very substantial, and this is close to the rate in Europe from that same time period. Related article below:

    www.nzherald.co.nz/nz/news/article.cfm?c_id=1&objectid=12269796

    I would say to Slarty "so long and thanks for all the (dead) fish"

    0 0
    Moderator Response:

    [PS] I would admit that moderation has been a bit sleepy on this thread but the time has come. Speculations on commentators intelligence are breaches of comment policy. I would request any further commentary is to point and in conformance with comment policy.

23. Bob Loblaw at 07:41 AM on 25 June, 2020

    Nothing here makes me want to spend time gong to Slarty's web site to read it, but MA Rodger's most recent coment is - shall we say - intruiging.

    Trying to follow through the process that leads to the figure MA has included, leads me to this:

    • On the X-axis, N must be the number of months that are used to derive the stddev, so ln(N)=0 is N=1 is monthly data; ln(N)=1.099 is N=3, so the three-month accumulation, and onward to ln(N)=7.09 is N=1200 months, or ten years.
    • 
      As we average over longer periods, the stddev of the individual periods decreases, so ln(stddev) decreases from the first point (ln(N)=0), where ln(stddev) = about 0.07, so stddev is about 1.07, until we reach the last point where ln(stddev) is about -1.25, so stddev is about 0.28.

    ...but of course, for random data, the longer the averaging period, the smaller the standard deviation, according to 1/sqrt(N). So, in a random system, having stddev=1 for N=1 would lead to stddev=0.03 for 1200 months. (1/sqrt(1200)).

    That the observed standard deviation decreases much more slowly than this, as the averaging period increases, is an indication that the data series is not random.

    Even so, any statistical technique that treats each observation (a one-month anomaly, an annual annomaly, etc.) as an independent value is doomed to failure. The values are not a collection of independent observations - they are a time series. So time-series analyisys is required. And for the longer periods you need to account for serial autocorrelation in the data to get the statistical significance right.

    And lastly, if the "century" stddev is (by extraopolation) of the order of 0.167, then how does Slarty mathturbate that into saying a 1C shift is covered by that 0.167 stddev? Does he claim that a 0.167 stddev implies that monthly anomalies can be 1C, and therefor monthly anomalies of 1C compared to a century ago are just random?

    If that were the case, then anomalies of 1C would appear spread across the century, not all clumped at one end. That clumping at one end tells us something - and that "something" is that climate change is causing global warming, and that temperatures have risen, and that the results are statistically significant.

     

    0 0
24. Eclectic at 10:44 AM on 25 June, 2020

    Lawrie @20

    Quite right .  Your final question touches on the heart of the problem.

    You and I, and every level-headed person, can look at the data and come directly to the bleeding obvious deduction.

    But that's not what happens in Denialist brains ~ they continually spout all kinds of pseudo-scientific nonsense.   Alas, it is the nature of the beast.   They are internally motivated to avoid seeing the Elephant standing in front of them.

    Some of the climate denialists are single-issue crackpots.  Rational in other areas . . . but fixated on the "non-greenhouse nature" of certain GH gasses ~ or have some other crazily contrarian Bee in their bonnet.   Most of these guys [rarely a female] have little or no political axe to grind.

    Other climate-science deniers (the majority) start from an extremist political position which originates in (or perhaps is reinforced by) personality traits of perverseness / anti-authoritarianism / fundamentalist religiosity / toxic libertarianism / delusions of superiority [e.g. Dunning-Krugerism] / or plain simple selfishness & lack of altruism.   These also are usually male.  (The female versions I encounter seem to be "just going along" with a toxic husband/boyfriend, for the sake of a quiet domestic life.  But I have met one exception! )

    This majority is known by their demonstration of rampant Motivated Reasoning.  They proclaim all sorts of excuses showing that "the science" is wrong.   Either :-  a serial of excuses ~ like a heavy frog leaping from one undersized lilypad to another . . . and eventually landing on the Island of Conspiracy, from which they can't be dislodged.

    Alternatively , they blast away with shotgun pellets of all sorts of excuses at once (and of course most of these excuses are doomed to be mutually-contradictory).   Our friend Slarty is clearly of the "shotgun" type.   But he has shown an admirable output of energy in constructing his blog, even though entirely misdirected & largely oblivious to the underlying physical processes of planetary climate.

    0 0
25. MA Rodger at 22:48 PM on 25 June, 2020

    Bob Loblaw @23,

    You may be intrigued but the multi-layered errors present within the analysis of Slarty Bartfast make understanding hsi work a tad challenging and correcting it work nigh-on impossible.

    The Christchurch NZ Slarty Bartfast analyses is presumably using this Berkeley Earth raw station data which you can see is flat over the record (the adjustmented data isn't) so an SD can be calculated without de-trending.

    

    But the SD comes out at 1.58ºC which is not the 1.06ºC plotted on Slarty Bartfast's graph shown @21. If the wobbles are extracted by taking the monthly variation form the annual mean, the SD reduces but only to 1.5ºC. (If you do the same de-wobbling on BEST global monthly anomalies SD=0.1ºC and with the unwobbly ocean data removed, for global land SD=0.3ºC.)

    Trying to reproduce Slarty Bartfast's graph shown @21, if the SD of the de-wobbled raw Xch data is then calculated for multiple-month periods, as period-length increases the reduction in SD is much steeper than the graph, a reduction you'd expect for a normally-distributed signal (from monthly SD=1.5ºC down to decadal SD=0.01ºC). And extrapolating to 1,200 months yields SD=0.001ºC. This 1,200 month value is greatly smaller than Slarty Bartfast's SD=0.17ºC and would suggest yet another fundamental error within his analysis. Mind, such frequent error seems characteristic of his work. His SLR analysis showed him unable to read a map and unable to read a table, both errors fundamental to his analysis.

    0 0
26. CD at 03:10 AM on 26 June, 2020

    @25 MA Roger

    The red line on the BE Christchurch data you have posted is the 5-year moving average. Its SD is at least ±0.34 °C, which is obvious from the size of its fluctuations. That is nowhere near the 0.01 °C you claim for a ten year average even though the timeframe differs by only a factor of 2. The actual value for 10-year smoothing is ±0.28 °C. And remember that is BE data and BE smoothing, not mine. By the way the green line is not data: it is the best fit linear trend. That is why it is flat!

    You don't extrapolate to a new SD for a different moving average timescale just by dividing the old SD by the new number of months. You first need to smooth with the new sliding window, then recalculate the SD. And by the way the SD inversely scales as the square root of N, not 1/N.

    @21

    On another point, if the SD is 0.167 °C, then the 95% confidence interval is ±3 SD or 6 SD in total. i.e. 1 °C. This generally approximates to the range of the data. That means that, while there is only a 5% chance that any data point will be 0.5 °C above or below the mean, over a millenium or longer, the probability that there is a fluctuation in the mean temperature between centuries of more than 0.5 °C somewhere in that timeframe gets quite big.

    0 0
27. CD at 05:23 AM on 26 June, 2020

    @23 Bob Loblaw

    It appears you don't understand standard deviations (SD) or scaling behaviour, based on what you have written.

    i) The N is not the number of data points in the SD, it is the number in the moving average. Each moving average has its own SD as shown in the Christchurch data posted by MA Rodger. The SD of the blue curve (1 month average) is clearly much bigger than the SD of the red curve (5yr = 60 month average).

    ii) You said: "but of course, for random data, the longer the averaging period, the smaller the standard deviation, according to 1/sqrt(N)."

    NO !!!!!

    The SD stays the same irrespective of the length of your time series. Doubling the averaging period doubles the total value of the terms being summed. Dividing by 2N just gives the same result. You only reduce the SD by repeating the same measurement and averaging them, not by increasing the number of data points. So averaging multiple station records in a regional trend will reduce the SD, doubling the length of single record will not.

    0 0
28. Bob Loblaw at 06:57 AM on 26 June, 2020

    MA Rodger @25

    When I said "intriguing", I did not mean to imply that Slarty's work was scientifically intriguing - rather that is was psychologically intriguing that he could come up with such bizarre results.

    Slarty @ 26:

    So much wrong in such a short space.

    First, standard deviation is a calculation that is applied to a collection of independent values. Although you can do the math on any collection of values, SD is not a good measure of spread if the values are not independent.

    Moving averages are clearly not independent. If you have 20 years of monthly data, you have

    • 240 monthly values
    • 80 3-month values
    • 40 6-month values
    • 20 annual values
    • 2 decadal (10-yr) values.

     

    If you think that you have more than 2 independent 10-yr averages (e.g., from a moving average), you are wrong.

    Even in the numbers I list above, you need to account for autocorelation to get the proper significance tests working. Warm months and warm years (or cool months and cool years) tend to clump together. If you have a warm year, there is a good chance the following year will also be warm. Why? Physics. The earth doesn't randomly jump temperatures - it takes time to warm up or cool down, becuase you have to add or remove a lot of energy.

    And if you are taking overlapping moving averages, there is seriously bad autocorrelation. For a 10-year moving average (120 monthly values), moving one month along the time line drops one value and adds one - 119 of the values used in the next moving average are exactly the same. You don't seriously think that this is an "independent" result, do you? You do know about the assumption of "independent values" in statistics. don't you? You do know what happens if you violate that assumption, don't you?

    And +/-3 SD is 99% range, not 95%.

    And since a normal distribution is unbounded, if global temperature were truely a random variable then it is still possible to get +6 SD or -8 SD in any 100-year or 1000-year period. It doesn't happen, because Physics.

    Slarty @27

    The "Standard deviation" of means of sample size N is propoerly referred to as the standard error of the estimate of the mean (SE). Yes, the SD for a population is constant (as long as the distribution is not changing). The SE decreases by 1/sqrt(N), for random data, exacly as I have said. Larger sample = mean probably closer to correct value = smaller standard error.

    You are the one that created a graph showing SD changing. Now you are sayong both that is is a constant, and that is can also be decreased. You need to seriously read a good statistics book and get your terminology correct.

    No, "repeating the same measurement and averaging them" does not reduce the SD - it provides a more reliable estimate of the average (mean) according to the SE.

    Fitted curves (1-month average, 60-month average) don't have Standard Deviations. When you calculate the "standard deviation" (the formula looks the same) of the residuals (I am guessing this is what you are talking about - but that nomenclature problem again), then again you are talking about standard errors.

     

     

    0 0
29. MA Rodger at 20:23 PM on 26 June, 2020

    Slarty Bartfast @26,
    You are a glutton for making yoursef look foolish.

    The BEST graphic does say the red trace is the 5-year rolling average. There is no reason to repeat that message. Now had you read my comment @21 properly, you would note that your methods are not properly explained for me. That deficiency would be something you could have put right rather than your misplaced lecturing. And I hope my explanation of what I did is properly explained @21. If it is not, you should have said.

    So let us use that raw BEST data (which is actually an error) but without the detrending I employed @21.
    If the decadal average values are calculated, they should yield a list like this (labelled by final year):-

    1873 ... ... +0.0307°C
    1883 ... ... -0.0919°C
    1893 ... ... -0.2841°C
    1903 ... ... -0.3403°C
    1913 ... ... -0.2466°C
    1923 ... ... -0.0231°C
    1933 ... ... -0.2009°C
    1943 ... ... -0.1486°C
    1953 ... ... -0.0967°C
    1963 ... ... +0.1220°C
    1973 ... ... +0.1539°C
    1983 ... ... -0.0482°C
    1993 ... ... +0.0990°C
    2003 ... ... +0.0077°C
    2013 ... ... +0.0130°C

    The final decade is a few months short. Including it the SD=0.145°C.
    I have no idea how you manage to obtain the SD values you quote. I am using the raw data from BEST's station 157045. If this is not what you are using to obtain your decadal SD=0.28°C, perhaps you could provide a link to the data.

    Your second lesson on how to read an annotated graph are a bit wasted. Why would the Green line be anything other than what the graphic says it is? (And it is of course the data that determines flatness, not some green horizontal line!!!)
    Ditto your lessons on how not to calculate seasonal, annual, decadal etc SDs from monthly data. Mind, the results I obtained with the detrending described @21 do show SD reducing by 1/N (with the exception of the one I happened to check and that did happen to conform to 1/√N).

    And concerning another of your blunderful statements @26.
    Maybe where you come from it's different but where I come from the values of a normal distribition looks like this:-
    
    The table shows that the SD that includes 95% (so 2 x 0.4975) is SD=+/-2.81. At SD=/-3.0, the percentage is 99.73% (so 2 x 0.4987).
    And while you are correct to say that (in analogy) if you play the lottery long enough, you ticket will eventually win the lottery (although over timespans of millions of years for standard weekly lotteries, not a single millenium for a centenial lottery). Yet this is not what you propose. You suggest it is "probable" when, if you buy a raffle ticket one week (one of 769 sold) and then again the next week, that you will win both times. That 'probability' is actually one chance in 769 x 769 = 1:591,361. (So not quite what I said @21 where, in my haste, I pulled the punch with a little arithmetical error of my own.)

    Finally Slarty Bartfast,  Bob Loblaw @28 is correct. The nature of your mistakes and errors are "intriguing".

    0 0
30. CD at 22:21 PM on 27 June, 2020

    The only mistakes I made were these:

    @27 "You only reduce the SD by repeating the same measurement and averaging them, not by increasing the number of data points."

    Even that is not true. What I meant was, multiple measurements result in a regression to the mean. The SD will be the same for sufficiently large N irrespective of its value. For a temperature series the SD is a measure of the spread of the data about the mean. That does not change with the length of the series.

    @26 : Yes the 95% is 4-sigma not 6-sigma. Thank you all for spotting it. But that still results in fluctuations of up to 0.7 °C minimum. These are still comparable to what is being claimed by climate scientists. It is not trivial. That is the point.

    0 0
31. Bob Loblaw at 03:34 AM on 28 June, 2020

    No, Slarty, you've made lots of errors. You just don't seem to be willing or able to face them.

    I"ll just point out one more, in comment 30. You say:

    That does not change with the length of the series.

    Focus on your own words: "the series". One hundred years of temperature data (local, global, whatever) is not, I repeat not, a random sample of temperatures from a constant distribution. It is a time series.

    The statistics of a climate temperature time series are not constant over the period of the series. There is a daily cycle. There is an annual cycle. There are systematic variations due to factors such as El Nino, solar output (11 and 22-year cycles, and perhaps some longer), atmospheric aerosols, and other physical factors (CO2 amongst them) that cause variations in temperature.

    Those variations can cause changes in the mean. Those variations can cause changes in the spread (SD or other measures).

    Any collection of temperatures values that form a continuous subset of the complete record (a month, a year, a decade) represents a time series that can - no, will - have a different mean and (possibly) SD from the complete series. This is not because of random sampling difference - this is because of physics.

    Within that shorter period, adjacent measurements (two months in a row, two years in a row, etc.) will exhibit autocorrelation. They are not independent. They are not random samples.

    Everything you are doing seems to be in complete ignorance of the fact that elementary statistics of random sampling is not enough for examining time series. You can end that ignorance through learning. But first you have to accept that you still have a lot to learn.

    Here is a hint: any statistical test that you do, that will return exactly the same result if you randomly re-ordered the data (i.e. you arrange them in some order that ignores the time variable), is ignoring the time-series aspect of the data.

    You have it fixed in your mind that there is no change in the time series. That is implicit in the analysis that you do, even if you do not realize it. Concluding, on the basis of your analysis, that temperature is not changing, is not a "conclusion", it is an assumption you started with.

    You've assumed your conclusion.

    0 0
32. nigelj at 10:28 AM on 28 June, 2020

    Slarty Bartfast@30

    Slarty says "The only mistakes I made were these... " (he lists two mistakes and tries to make excuses for them.)

    But hes made plenty of other mistakes. Here are just a couple of others:

    Slarty @11, "By the way, thanks for pointing out the error in the thermal expansion coefficient. I used the wrong one by mistake"

    Slarty @12 "But this doesn't change my overall point, that sea level rise is miniscule and unmeasurable" This is clearly an error, no matter what way you look at it, as pointed out several times eg me @14.

    Slarty @12 "What I did say about the Arctic is that we don't know what the temperature trend is because there are no long term weather stations within 1000 km of the North Pole, and there never have been."... "You need 50 years to establish a climate trend"

    I pointed out @16 we have at least 60 years of data over the open arctic. And I would add saying you need 50 years of climate data is only Slartys opinion. Every source I've read says 30 years is sufficient, for example the experts on realclimate.org.

    Slartys website blog claimed that climate change is caused by waste heat and quoted New Zealand as evidence where population density and industrial output is low and our glaciers haven't shrunk very much. I pointed out research @22 showing our glaciers have shrunk a great deal at 33% since the 1970s. Anyway there will be some difference between hemispheres because the northern hemisphere is warming faster due to a preponderance of land mass.

    Own you mistakes Slarty. I suspect you have a pre determined conclusion, and when you work like that you tend to twist things to suit the conclusion and mistakes multiply. Nothing personal: you are well educated and know more statistics than me, but I have a razor sharp ability to recognise nonsense in almost any field of study.

    0 0
33. MA Rodger at 19:56 PM on 30 June, 2020

    Given I have criticised Slarty Bartfast for the egregious level of error he achieves, I really should correct an error of my own up-thread.

    Due to a spreadsheet error, the table of decadal averages within my comment @29 is entirely wrong, along with the associated result SD=0.14°C. The spreadsheet calculation corrected, I can now concur with Slarty Bartfast on his calculated decadal SD=0.28°C. (I calculate SD=0.287°C.)

    The same does not go for other of his calculated SD for the Christchurch NZ monthly raw temperature data, these scaled from the graphic @21 above. (And of course, this data shouldn't be used for any reason in their raw state and and the SD shouldn't be calculated without consideration of the trend.) The comparison runs as follows:-
    Av'ged period (months) .. .. My SDs ..... .. His SDs
    1 .. ..... ..... ..... ..... ..... ..... 1.58°C ..... ..... 1.06°C
    3 .. ..... ..... ..... ..... ..... ..... 1.30°C ..... ..... 0.76°C
    6 .. ..... ..... ..... ..... ..... ..... 0.60°C ..... ..... 0.62°C
    12 . ..... ..... ..... ..... ..... ..... 0.49°C ..... ..... 0.51°C
    24 ' .... ..... ...... ..... ..... ..... 0.44°C ..... ..... 0.43°C
    60 . ..... ..... ..... ..... ..... ..... 0.34°C ..... ..... 0.34°C
    120 ...... ..... ..... ..... ..... ..... 0.29°C ..... ..... 0.28°C
    It is evident that the SD calculation does not yield the straight line that Slarty Bartfast insists they do. And there is a pile of other methodological reasons for not attempting to extrapolate the data to provide an SD for an averaged century-long 1200 month period.
    But as Slarty Bartfast does, let us not ignore all those problems.

    Instead consider the situation if, as Slarty Bartfast insists is the case, for 1200 month period there were actually SD=0.18°C. What is then easy to demonstrate are the errors in his argument that a +1.0°C increase in global average temperature between two consecutive centuries is "entirely possible" indeed "probable," this "due to natural variations resulting from chaotic behaviour within the climate system," or "mostly" so.

    Slarty Bartfast @30 asserts that this SD=0.18°C would have a 95% probability of a 4-sigma fluctuation or "0.7 °C minimum". There are however a couple of fundamental errors in this bold assertion.

    Firstly, a 20:1 chance is not what anybody would describe as "probable." Such odds are usually seen as being "improbable."

    Secondly, for a normal distribution, the odds of seeing an increase of +0.7°C, or a positive increase equal in size to 4xSD, is not 20:1 but roughly 800:1. It is actually highly improbably. And we have the data to demonstrate this point.
    The raw monthly data used by Slarty Bartfast can be averaged over those different periods and so the number of actual occurances of a +4xSD fluctuation can be totted-up. So, how many times do we find the increase between two data points is greater than 4x the calculated SD? Despite over 3,000 attempts, we don't find even one. The best we can do is 3.4xSD (which for a normal distribution is about 10x more likely than 4xSD).

    Av'ged period (months) .. .. 4 x SDs ..... .. Dev (& xSD) achieved
    1 .. ..... ..... ..... ..... ..... ..... .. +6.33°C ..... ..... +5.16°C (3.26)
    3 .. ..... ..... ..... ..... ..... ..... .. +5.20°C ..... ..... +4.44°C (3.42)
    6 .. ..... ..... ..... ..... ..... ..... .. +2.40°C ..... ..... +2.06°C (3.44)
    12 . ..... .... ..... ...... ..... ..... .. +1.96°C ..... ..... +1.30°C (2.65)
    24 . ..... ..... ..... ..... ..... ..... .. +1.74°C ..... ..... +1.18°C (2.70)
    60 . ..... ..... ..... ..... ..... ..... .. +1.36°C ..... ..... +0.87°C (2.55)
    120 ...... ..... ..... ..... ..... ..... .. +1.15°C ..... ..... +0.43°C (1.51)

    As would be expected, the longer the averaged period, the fewer the data points, the smaller the probability of consecutive 4xSD fluctuations between consecutive data points.

    So, even if it were correct to use raw data, even if the presence of trends were ignorable, even if the SD for century-long smoothing could be obtained; even if all this were so, we find Slarty Bartfast's chosen data does not show what Slarty Bartfast says.

     

    0 0

You need to be logged in to post a comment. Login via the left margin or if you're new, register here.