Idly casting my still sleep-filled eye over WUWT this Sunday morning (27 June) I was struck and attracted by a most impressive sight – maps of the Northwest USA bedecked in bright hues. These I learned were “Update 3 Northwest USA Heatwave Prediction” published the previous day (26 June) and had been copied from a weather website published by Cliff Mass (who runs a University of Washington high resolution weather prediction system).
What was most striking about the map was the use of vibrant colours, mostly clashing shades reds and yellows for different areas of high temperature predictions. With areas of different colour shaped by the mountains dissected by deep river valleys, the map resembles a blazing fire with bright flames raking the sides of the Olympic Mountains and Cascade Ranges. In other maps the core of the mountains is depicted in brown colours making them resemble burnt-out coals in the centre of the fire.
I immediately thought just how appropriate was the colour scheme predicting a forthcoming heat wave with predicted temperatures expected in places to exceed 110oF.
The only green (=cool) pigments visible were far out in the frigid Pacific. Puget Sound was a solid yellow (= bath water temperatures) hue. I once crossed Puget Sound from Seattle to Victoria in BC by hovercraft. I have never been so scared in my life. Puget Sound was mind-numbingly cold and seemingly full of inquisitive killer whales. It is also notorious for its escaped and waterlogged logs floating just beneath the surface. I imagined the outcome of my hovercraft hitting a 200 footer at speed, and those intelligent killer whales following behind seemed to have motive. But I digress.
I do appreciate a good colour rendition in a map. I was introduced to the art of using high-confidence colours to sell a prospect to the company exploration manager when I worked for an oil company. But I was warned not to overdo it. Managers might well become suspicious if high confidence colours were overused. So was this temperature prediction map overdoing it? Well possibly, but it attracted me and the map’s validity would soon be judged since it was predicting events only a few days hence – so Cliff Mass and his team must be rather confident.
Soon after I had come to appreciate the map’s fizzing colour scheme, my mind suddenly lurched to whether those suffering from colour blindness could appreciate the maps. Then on to the topic of colour blindness and whether climate alarmists suffered from it – unable to recognise the advancing greenness of semi-deserts. But then friend Wikki came to my aid and I discovered that for those with the commonest form it is red apples that appear green, rather than the reverse. Do they perhaps suffer from a much rarer type – trianopia where green coated landscapes appear as cloaked in a purplish red? Something must affect their view of the world.
Cliff Mass threatens a sequel relating the heatwave, after it has past, to climate change. Wonder if Anthony will pick that up?
Here’s the end of June Heat advisory for US
There’s also precipitation
LikeLiked by 1 person
A good coverage of the issues can be found in this paper:
“The misuse of colour in science communication”
However, I don’t think the paper covers the psychology of using colours to engender emotive responses.
Ron, gosh just imagine Cliff Mass’s maps flickering like yours do. The simulation of a raging hot fire would be even more realistic.
John so many thanks for embellishing my morning bit of whimsy with something with scientific weight and relevance.
You say ‘whimsy’ but I say ‘serious business’. As far as I can see, the ‘Update 3 Northwest USA Heatwave Prediction’ is using the common rainbow colour scheme. The appeal of this scheme to the climate alarmed is that it is evocative of the black body spectrum, with things glowing red hot, etc. However, as far as I am concerned, the key statement in the Nature paper is:
“The common rainbow colour map should not be used in data visualisation. There is not a single rainbow colour map with similarly bright colours across the colour bar that comes close to being scientific (e.g., ref. 32).”
Since you mention Cliff Mass, I’ll simply highlight that his attribution argument was rather silly.
The original forecast from the 23rd had, to my eye, a striking resemblance to a hobgoblin:
Or maybe it was yet another psychological quirk of a human mind, this time automatically trying to impose agency on a random pattern (pareidolia).
Gavin Schmidt may very well be justified in saying that Cliff Mass is guilty of ‘naïve accounting’, but I find that the explanation provided by Gavin renders him no less so.
It is essential that anyone who tries to construct a statement of causality should provide an assessment for both components of the causal narrative, i.e. the probability of necessity (PN) and the probability of sufficiency (PS). Despite the fact that Gavin recognizes that there is a role to play for both ‘ultimate cause’ and ‘proximate cause’ he does not seem to be interested in following this through and evaluating how both play their part in the determination of PN and PS. Instead, by restricting himself to matters of fractional attribution, he attempts to construct a causal statement based entirely upon a value for PN. This mistake is commonplace amongst those who comment upon matters of climate attribution, but that does not make it right. In cases where there are several proximate causes strongly influencing the outcome, it is quite possible for a very high figure for PN to be, nevertheless, combined with a low figure for PS. By saying only that ‘If the odds of an event have doubled then climate change is 50% responsible’, Gavin is completely overlooking the sense in which proximate causes can also be said to be responsible. To complete the causal statement he should add something along the lines of ‘But given that the probability of sufficiency is only 0.005, this makes the proximate causes 99.995% responsible’. The apparent paradox here lies in the loose and ambiguous use of the word ‘responsible’. He should really try to avoid such words.
The bottom line is that two people arguing over the ‘contribution of global warming’ without accounting for both PN and PS are both guilty of naïve accounting.
LikeLiked by 1 person
I live in White Rock, British Columbia , near the US-Canada border and the forecasts here have somewhat underestimated the highs. It’s currently 38C even at a station near the ocean. Definitely feels hot as many of us don’t have air conditioning in coastal BC. Of course the media are attributing the “heat dome” 100% to climate change.
I was struck by the Cliff Mass analysis in his blog where he states:
“So without global warming, a location that was 104F would have been 102F. Still a severe heat wave, just slightly less intense.”
He has taken the average regional warming of 1-2 F over the past 50 years and added it to the forecast highs. This seems much too simplistic to me as the average warming has nothing to do with an extreme weather anomaly. Hence his conclusion:
“Let me end with the golden rule of temperature extremes: the bigger the temperature extreme the SMALLER the contribution of global warming. Think about that.” Think indeed!
Potentilla, I have been to White Rock back in the 1980s, looking for a house. But nothing came of my move and eventually I moved back to the U.K.
You have my sympathies regarding the current weather, I suspect trying to sleep at night when temperatures are in the twenties (centigrade) and dew points are high must be the worst. I’ve only experienced those temperature conditions at night in Regina but then humidities were low.
Obviously ATTP doesn’t think much of Cliff Mass
I have now read Cliff Mass’s blog post, in which the offending map appears, and I was astonished to read the following:
“The situation is so extreme that I had the colors altered to better define high temperatures.”
This is like saying ‘The situation is so extreme I had the data altered to over emphasise how extreme it was.”
Anyone who has read the Nature paper I cited earlier (“The misuse of colour in science communication”) would know that Cliff is breaking every rule in the book. And yet, at the end of his blog post he proclaims that he is “trying to communicate the best science”.
Yes, but in the least scientific way.
I am having so much fun with this that I now, belatedly, realise that I should have registered the value of Alan’s article by liking it.
LikeLiked by 1 person
And having now read the Cliff Mass post, I am even more certain that he is simply analysing the situation from the perspective of probability of sufficiency. What he actually says is:
“So without global warming, a location that was 104F would have been 102F. Still a severe heat wave, just slightly less intense. Let me end with the golden rule of temperature extremes: the bigger the temperature extreme the SMALLER the contribution of global warming. Think about that.”
What he should be saying is:
“Let me end with the golden rule of temperature extremes: the bigger the temperature extreme the SMALLER the probability of sufficiency. Think about that.”
Neither Gavin or Cliff appear to have the required grasp of causal analysis theory to properly express themselves, and so they argue past each other, using words such as ‘responsible’ and ‘contribution’. This is a perfect example of the misunderstandings that I alluded to in my article ‘A Brief Primer in Causation’, when I said:
“The fact that causality has this duality (probability of necessary cause and probability of sufficient cause) leads to many differences of opinion when attribution statements are discussed, with the alarmed usually focusing upon PN and sceptics focusing upon PS. Worse still, the individuals concerned are often unaware that this is the true nature of their dispute.”
However, it is also worth noting that climate scientists can make the mistake of thinking that a probability of necessity is all you need for a full statement of causation, and no one blinks an eyelid. And yet as soon as someone starts to emphasize the perspective afforded by probability of sufficiency you get the likes of Gavin Schmidt dismissing them as ‘naïve’ and the likes of Ken Rice calling them ‘silly’.
Speaking of Dr Rice, where did he go?
LikeLiked by 1 person
John just a minor point but from one point of view the colour scheme used in the maps might be considered more appropriate than a scientifically more accurate graded colour scheme. If Cliff and his team wished to convey the potential danger of the high temperatures to a general public, then I for one consider their efforts reasonable and appropriate. Hobgoblins aside, the similarity to a hot fire was what drew my attention and I’m sure that of many in the affected area.
Better things to do.
Except, this misses the main point. If we assume that the distribution of events remains normal with the same standard deviation, then a shift of the mean might imply – in a simple sense – that every event shifts by an amount equal to the shift of the mean. However, if for example, you shift the mean by one standard deviation, what was a 4 sigma event becomes a 3 sigma event, or a 3 sigma event becomes a 2 sigma event. That’s a big change in the probability of that event occuring. So, an event that may never have occurred over the timescale considered, could become an event that does occur.
According to Cliff Mass, this would suggest that this event was not influenced by the shift of the distribution, which seems like a rather bizarre conclusion.
I only sort of agree with you. It blurs the boundary between science and activism and I guess a lot of the discussion here at Cliscep reflects a concern for that boundary.
“According to Cliff Mass, this would suggest that this event was not influenced by the shift of the distribution, which seems like a rather bizarre conclusion.”
No, no and thrice no. He is not saying that at all. He is making a case for the probability of sufficiency and you and Gavin insist on judging him as though he is arguing about the probability of necessity. That is the main point.
To go back to your comment about Gavin’s point, I don’t think it’s correct to regard the ultimate cause and the proximate cause as two separate influences. Every climate event will have a set of proximate causes (whatever physical atmospheric processes that operated to produce that event). Understanding the proximate causes will typically not tell you if climate change played some kind of role. To do that you’d need to understand the impact that climate change has had (the ultimate cause). So, the suggestion is that one way to assess the impact of climate change is to consider the probability of that event in an unchanged climate and then compare that to the probability of that event in a changed climate.
Hence, if we shift the distribution so much that an event occurs that had an extremely low probability of occuring in an unchanged climate, one might conclude that climate change played a significant, or even dominant, role. That doesn’t mean that the proximate cause would be atmospheric processes that had never occurred before. It would just mean that climate change has shifted the underlying climate state to the point where it becomes possible for an event to occur that had never occurred before.
There may well be other ways to assess this, but this seems like a reasonable way to estimate the impact of climate change.
Firstly, I’m going to take issue with the use of the term ‘ultimate cause’ in this instance. If one is performing a root cause analysis, in which a succession of events lead to a particular outcome, one can point to the initiating cause as being necessary for what followed. That does not, however, address the question as to whether the initiating event was sufficient for the final outcome. There may be several intervening events that were equally or more necessary and it may not even be possible to discern a single initiating event anyway. By referring to climate change as the ultimate cause of a specific weather event, one affords it a status that is unwarranted. It may provide the backdrop against which extreme weather events may become more likely, but that doesn’t make it an ultimate cause in the root cause analysis sense.
Secondly, no one, including myself, or Cliff, is for one moment arguing that proximate causes are more relevant than distal causes. Fractional attributions of risk are equivalent to calculations of the probability of necessity and, as such, address the question of how much more likely the event has been made by climate change. But that is only half of the story when talking about a specific weather event. For any event (Y), there are always two questions to be asked:
a) Given that X and Y have happened, what is the probability that Y wouldn’t have happened if X hadn’t happened?
b) Given that neither X or Y has happened, what would be the probability that Y would have happened if X had happened?
The former is the one you are concentrating upon. The latter question is no less fundamental, however, to the question of causation. It is the probability of sufficiency and it cannot be ignored if you want to fully understand the causation of an event. Unfortunately, however, when X is climate change, anyone who focuses upon the latter question is immediately accused of trying to deny the relevance of X. Nothing could be further from the truth.
I am conscious that we are getting very off-topic as far as Alan’s post is concerned. If we want to continue the discussion, perhaps we should consider choosing a different thread.
John and Ken, please be my guest. I wrote my post on a whim, believing that a few readers might just be interested in my early morning, almost ephemeral, thoughts about a resemblance between a heatwave prediction map and the flames in a fire. From this, the post has been transformed by you into a serious discussion about the rights and wrongs of colour representation of data and, wouldn’t you know it, attribution science. I couldn’t be pleaser.
bit o/t – but relevant – check out – https://www.bbc.co.uk/news/resources/idt-985b9374-596e-4ae6-aa04-7fbcae4cb7ee
By the BBC Visual and Data Journalism team – 31 July 2019 – header – “How much warmer is your city?” – scary stuff
You seem to have gone quiet again. Presumably it’s because you have better things to do than respond to comments that you elicited. Perhaps it would help if I were to be more direct and succinct in my response to your previous comment.
a) You use the term ‘ultimate cause’ inappropriately. It is no more appropriate to suggest that climate change is the ultimate cause of an extreme weather event than it is to say that moving to an area with higher covid prevalence was the ultimate cause of catching covid.
b) You have expended effort explaining to me how event attribution works when it should be obvious from what I have said previously that I already understand. How probability of necessity is calculated with regard to climate is a given. I am talking instead about the need to also consider probability of sufficiency in order to complete the causal statement in respect of a specific weather event.
c) We are having this discussion because you referred to Cliff Mass’s ‘golden rule’ as being ‘silly’. I am suggesting that you only consider it silly because you insist on interpreting it as a statement regarding probability of necessity when it is, in fact, alluding to sufficiency. Do you know how probability of sufficiency is calculated? I’ll cite you the relevant paper, if you show any interest.
LikeLiked by 1 person
Full disclosure: I’ve slightly edited my previous comment. Instead of giving what I now realize was a misleading hint, I have offered instead to cite the relevant paper.
LikeLiked by 1 person
Fascinating commentary today at WUWT about the heatwave in the Pacific Northwest by both Anthony Watts and his commentators. One of the latter points out how local the heat anomaly is – that the average temperature over the whole USA is actually down 1 degree F. Talking of global warming….
>”One of the latter points out how local the heat anomaly is…”
Try telling that to the BBC. You might like to check out what the BBC’s ‘Visual Journalism Team’ managed to conjure up by using a common rainbow colour map, employing deliberately evocative colours and an inventive calibration that de-emphasises the locality of the phenomenon and portrays even those parts shivering in 10-15oC as being in peril of melting:
What was it that the ‘misuse of colour paper’ said? Not even close to scientific!
I can imagine the ‘team’ huddled around the computer, endlessly twiddling knobs until they get the desired effect. Then one shouts from the back, “For God’s sake you morons. Don’t use blue!”
Yes, nowhere near scientific, but very near to journalistic.
I’m not that familiar with the probability of sufficiency, so if you were keen to highlight the relevant paper, please do.
Yes, I am keen to highlight the paper, if only for the following reason:
“We have shown, with simple examples, that it is important to distinguish between necessary and sufficient causality. Such a distinction is, at present, lacking in the conventional event attribution framework. Any time a causal statement is being made about a weather or climate-related event, part of the audience understands it in a necessary causation sense, while another part understands it in a sufficient causation sense, which can give rise to many potential misunderstandings. Introducing the clear distinction may thus clarify discussions.”
The above quote is taken from the paper, and it reflects my own concerns. The paper starts out in a deceptively facile manner but quickly reveals its relevance and, IMHO, profound importance. You will also note that it is not in any way a climate denier’s charter. It simply offers a wider and more insightful way of looking at D&A. Enjoy:
Click to access r451-reprint.pdf
Thanks. I think that is related to what I was suggesting in an earlier comment that there might be multiple ways to do this. For example, if you consider all the possible causal factors (anthropogenic forcings, natural forcings, and internal variability) then if the resulting PN (wrt anthropogenic forcings) is high enough, then the PS may not matter.
However, when consider future planning, both PN and PS become relevant. According to the paper, at least, PN is more relevant when considering taking action (If we reduce the forcing will it reduce the frequency of these events) but PS might be relevant if we’re considering what we might do if we don’t take action (i.e., if we don’t take action, will the frequency of these events be maintained).
So, as the paper says:
which seems perfectly reasonable to me.
The problem I can see with insisting on PNS is that it’s the probability that the event would occur in the present of (in this context) increased anthropogenic forcings, but would not occur in their absence. However, we are still interested in how climate change might impact events that could still occur in the absence of climate change, we’re not only interested in events that would not have occurred at all in the absence of climate change.
As I said, the paper is not a denier’s charter. Actually, the part of the paper that asserts that a high PN renders a low PS irrelevant as far as culpability is concerned is the one part that I am not sure of. I can think of a number of situations where a low PS may mitigate culpability notwithstanding a high PN, and I am unconvinced that the climate change example does not fall into such a category. Certainly this is the case where forest fires are concerned, in which lit matches, cuts in fire services and poor forest management are factors. Anyway, we shall see what attitude the courts take when the litigation starts.
>”However, when consider[ing] future planning, both PN and PS become relevant. According to the paper, at least, PN is more relevant when considering taking action (If we reduce the forcing will it reduce the frequency of these events) but PS might be relevant if we’re considering what we might do if we don’t take action (i.e., if we don’t take action, will the frequency of these events be maintained).”
Actually, I think the important distinction is between avoidance and mitigation. PS is certainly of more interest to insurance companies. Even a low PS may concern them because they have to take the longer view.
>”The problem I can see with insisting on PNS…”
I believe the main problem with insisting upon a high value for PNS is that it sets the bar too high for making a causal case.
To return to my initial point, I still believe you and Gavin were misjudging Cliff Mass, mainly because he was pointing out how the causal case looks from a sufficiency perspective and you were reading this as a denial of the necessity perspective. As he said:
“Is global warming contributing to this heatwave? The answer is certainly yes. Would we have had a record heatwave without global warming. The answer is yes as well.”
The Hannart paper may be aligning with comments you have made subsequently but it does not support your initial criticism of Mass. He was simply trying to take on board the duality of causality. If anyone was being naïve, it was Gavin Schmidt.
If the likes of Gail Tverberg are anyway correct then the industry needs an oil price of around $120 a barrel to make developing new fields profitable. So it seems to me that the oil majors have every reason to be cautious about developing new fields. It also could be that they see defeats in the courts as a face-saving way of withdrawing from increasing difficult and unprofitable markets.
The five major western oil companies only control about 15% of world oil production, the remainder being supplied by national companies. So the Grauniad’s attack on oil seems more like an inter-elite fight than anything more serious. When they start attacking the likes of Aramco, the Saudi oil producer, we will know they are serious about shutting down the oil industry.
I’m not quite sure how Cliff can make this claim:
unless he means that at some point this region would probably have experience a heatwave, even in the absence of climate change. It may well be true, but it doesn’t seem all that informative.
I think there may be two ways of thinking about Cliff’s assertions:
a) There was so much internal variability on display that one has to concede that the PS for the temperatures achieved was very low.
b) There was not enough variability to explain the temperatures achieved, so one has to concede a high PN, but there was still more than enough to have broken the previous records without recourse to the contribution made by global warming, i.e. the PN was high with regard to the actual temperatures achieved but low with respect to those that were necessary to break the previous record.
I think both statements may be true but perhaps the latter is more relevant.
Except, you can’t separate out the various causal factors without some kind of model. Clearly, a complex set of climate factors (or, physical climate processes, whatever term seems most suitable) combined to give this climate event (heatwave). And one can clearly analyse the atmospheric data to determine what these processes were. However, that alone doesn’t tell you if climate change played a role, or how much of a role. To do that, you’d need to somehow compare how likely such a set of events might be in the absence of anthropogenic forcings, and in their presence (and you could presumably use this kind of analysis to determine PN and PS). So, unless Cliff has done something like that, it’s hard to see how he can make a confident statement about the emergence of a record heatwave in the absence of global warming (as I said earlier, this is probably true in the sense that one would probably eventually emerge, but this is not particularly informative).
I think Cliff is just conjecturing. I don’t for a minute think he has performed any D&A on this one – but you can bet your life that someone will, or already has. Let me know if you hear anything. If they publish values for p1 and p0 as well as for the FAR then we will be able to use the Hannart-Pearl equations to calculate PN, PS and PNS. Wouldn’t that be fun?
With our discussion having reached a natural conclusion, I had promised myself that I would let the matter drop. However, having just re-read your responses, it occurs to me now that there are still a couple of points that remain to be made.
Firstly, despite the fact that both you and Gavin criticised Cliff Mass’s attribution argument because of a supposed misunderstanding of how necessity works, and despite the fact that I have demonstrated that Cliff was not actually arguing from the perspective of necessity, you have not, at any point, acknowledged your error.
Secondly, after reading the Hannart et al paper that I’d cited, you drew attention to the authors’ statement that ‘when consider[ing] future planning, both PN and PS become relevant.’
You said that this “seems perfectly reasonable to me”.
Back in June 2020, I told you this on your website:
“You need both PN and PS for a full causal narrative and you need a full causal narrative to decide the best policy for managing the risks.”
It’s such a shame that the idea didn’t strike you as perfectly reasonable back then. In fact, you didn’t pay any attention whatsoever. You only sprang to life when defending Willard’s unwarranted censorship of my complaint regarding his behaviour.
In fact, if you had read and taken on board my ‘Brief Primer on Causation’, posted here on Cliscep back in March 2020, you would already have understood all you need to know about the relevance of PN and PS, since it was in that article that I quoted Hannart:
“PS is the appropriate focus for the planner when assessing the future costs that inaction will imply, but PN is at stake when assessing the future beneﬁts of enforcing strong mitigation actions. Policy elaboration requires both sides of this assessment; thus both PN and PS are of interest here.”
LikeLiked by 1 person
Well, here’s more grist to the Ridgway/aTTP mill:
“Climate change: US-Canada heatwave ‘virtually impossible’ without warming”
“The searing heat that scorched western Canada and the US at the end of June was “virtually impossible” without climate change, say scientists.
In their study, the team of researchers says that the deadly heatwave was a one-in-a-1,000-year event.
But we can expect extreme events such as this to become more common as the world heats up due to climate change.
If humans hadn’t influenced the climate to the extent that they have, the event would have been 150 times less likely.
Scientists worry that global heating, largely as a result of burning fossil fuels, is now driving up temperatures faster than models predict.”
I notice that in the run-in to the publication of AR6, Ed Hawkins of Reading University has been very vocal. Looking at his CV, it seems that his proudest achievement is the creation of his ‘warming stripes’, a visualisation of the effects of global warming. In fact, such is the notoriety of his stripes that Wikipedia goes so far as to say that he is a specialist in ‘data visualization graphics’. By all accounts, the stripes have been very successful:
“On 22 May 2018, Hawkins published his warming stripes data visualization graphic, which has been used by meteorologists in Climate Central’s annual #MetsUnite campaign to raise public awareness of global warming during broadcasts on the summer solstice. Hawkins’ similar #ShowYourStripes initiative, in which the public could freely download and share graphics customized to specific countries or localities, was launched on 17 June 2019. The warming stripes graphic is used in the logo of the U.S. House Select Committee on the Climate Crisis (2019—).”
Wow! Not bad for a visualisation that uses rainbow colours with similarly bright colours across the colour bar. Now what was it that was said in that paper on the misuse of colour in science communication? Ah, I remember:
“The common rainbow colour map should not be used in data visualisation. There is not a single rainbow colour map with similarly bright colours across the colour bar that comes close to being scientific.”
If you want to know a gamester, you will know him by his stripes.
LikeLiked by 1 person
If winter in Antarctica was the deepest blue, and summer in Death Valley was the reddest red, would the stripes representing global average temperature over the past two centuries even be recognisable as stripes to the human eye?
LikeLiked by 1 person
I need to be a bit more precise regarding my criticism of Ed Hawkins’ warming stripes. He isn’t actually using a rainbow colour scheme across the board. What he is actually doing is combining two Colorbrewer palettes, one for temperatures below the 1971-2000 average (blue) and another for temperatures above (red). The temperature variations are scientifically represented within the range of each of the two palettes but the transition of colour to represent the crossing of the threshold is not a scientific use of colour. That is the sense in which the bad old habits of the rainbow map come into effect — and that is where the game is being played. When asked what the scientific basis was for the transition of palette at the 1971-2000 temperature threshold, Hawkins replied that the:
“choice of colours was an aesthetic decision — I think they look just right.”
As I said, not even close to being scientific.
LikeLiked by 1 person
John I partly understand your reluctance to accept colour schemes that are deemed unscientific. On the other hand if the main purpose of the illustration is to convey specific information (especially if this is a warning) or to focus upon one part of a map or other complex illustration, then I believe the illustrator need not be bound by a “scientific” palette. Thus when I coloured maps to illustrate potential prospects for my oil company’s exploration manager he didn’t care what colour I used so long as it stood out. Thus when Cliff Mass coloured his projected temperature map, I applauded his choice. It conveyed his warning simply without the user having to consult a detailed key, except for a specific location where the user had a specific interest.
I don’t disagree. So long as it is understood that the individual isn’t using colour to convey a scientific idea but is using it instead to psychologically influence, e.g. to draw attention in order to influence the audience. That, I believe, is what Ed Hawkins has done with his warming stripes and why he feels his choice of colour ‘looks just right’. It isn’t because his colour scheme provides a reasonable representation of scale. It’s because he is using colour to over-emphasise the handle of his hockey stick and to create a contrast between the temperatures on either side of an arbitrary threshold.
Just to labour the point:
Data visualization is the art of communicating information succinctly and with immediacy. It can be used in the furtherance of science communication only if it accurately communicates the data. An inaccurate or misleading representation cannot be said to be scientific. The Colorbrewer palettes were developed to ensure that colour can be used in the visualization of data without misrepresenting what the data is saying (also to avoid the problems of colour-blindness). Despite using two of the Colorbrewer palettes in his warming stripes Ed Hawkins failed to abide by the rules of scientific data visualization in the following important respects:
a) He used two contrasting palettes to suggest a trend that isn’t there in the data. If he had stuck to the one palette then the deepening hues within that one palette could have been an accurate visual representation for the full temperature range, but he instead introduced a psychological element by transferring from a blue (Brrr!) palette to a red (Phew!) palette. He said this was an aesthetic choice. Who is he kidding? What it actually does is psychologically de-emphasize one period of warming (the one with lightening blue) when compared with another (with deepening red). If you look at the data, there is equally significant warming before the red is introduced, but this is not the impression that one is left with when looking at the warming stripes. The red draws the attention.
b) The use of two palettes subtly introduces in the viewer’s mind the idea of two phases of warming: a benign phase of lightening blue and a dangerous phase of deepening red. Upon what basis did he decide that the 1971-2000 average temperature could be used as a transition point between those two phases?
c) He says he selected “baseline periods to ensure equally dark shades of blue and red for aesthetic balance.” Aesthetic balance is not, of course, a necessary element of scientific communication. One would have hoped that baseline periods would have been chosen for a more scientific reason. Does this balance mean, for example, that he chose the 1971-2000 average temperature as a transition point between the two palettes because it represents a mid-point in the temperature range portrayed? It doesn’t look like it when one looks at the actual data. To achieve the ‘aesthetic balance’, did he have to cheat by using a different temperature-hue scale for the two palettes? It’s difficult to tell from the data.
d) The lowest temperatures in the range are represented by the deepest hue in the blue Colorbrewer palette. The highest temperatures in the range are represented by the deepest hue in the red Colorbrewer palette. That means that a variation of 1.20c is portrayed as a transition from freezing cold to boiling hot. Is that a scientific data visualization? It looks more like an emotive representation to me.
I’ll leave it there. I’ve already wasted too much of my life thinking about Ed Hawkins and his warming stripes. Nothing I say will alter the fact that they are hugely influential and have already achieved the status of a cultural icon.
LikeLiked by 2 people
Influence beyond the data- Humans are so prone to metaphor and substitution practice. Almost everyone one knows that colours lend themselves to symbolic use, gold is associated with rich value, red to feelings of heat or blood or danger, hence fire trucks, the Red Cross or red traffic lights. I read that even a stickleback in a fish tank on a a window ledge was seen to go into attack mode by seeing a passing red bus . Those fiery climate graphs, no way innocent colour choice.
LikeLiked by 2 people
Looks like someone else has jumped on the climate stripes bandwagon:
The proud author (Prof James Cheshire) has come up with a geographical map version of the Ed Hawkins warming stripes that he says combines “the most important lessons from the warming stripes with the intricacies of geographical context.”
Unfortunately, the important lesson he seems to have taken on board the most is the use of contrasting Colorbrewer palettes to create a false impression of accelerated warming subsequent to an epoch of his choosing. He is therefore using colour to create an optical illusion — a sort of chromatic hockey stick. He could have, instead, stuck to the one palette throughout and retained some scientific integrity, but scientific integrity does not seem to be the intention. Otherwise, he would not be crowing about the advantages of hiding the uncertainty:
“Uncertainty is a complex thing to communicate in a single chart. In 2018 the UK-based climate scientist Ed Hawkins chose to omit it altogether when he presented his “warming stripes” graphic to help clearly visualise key trends in climate data. Hawkins explained that the warming stripes were designed to remove all superfluous information, leaving behind only the undeniable scientific evidence of a steadily warming world.”
Yes, folks, let us get rid of that pesky uncertainty at all costs. After all, it only serves to misguide the denier:
“One challenge of understanding the information contained in this hockey stick graph – and this is a gift to climate-change deniers – is the inclusion of the grey fuzz of “uncertainty data”: outlying data points that can be cherry-picked to raise doubts about the mass of evidence supporting a general warming trend.”
I’ve said it many times: I’m fed up to the back teeth with professors lecturing to me about a concept (uncertainty) that they themselves only partially understand, whilst telling me that I am too naïve to be entrusted with seeing it for myself.
LikeLiked by 3 people
I clicked to read the article and Prof James Cheshire’s first link is to Michael Mann’s piece on the AR6 WG1 SPM hockey stick on RealClimate, which in turn links to Mann’s piece in TIME Magazine on the wonders of the new report. I was so convinced by all this that I felt no need to read the rest of the Cheshire. Sorry John.
Fortunately, I avoided the “Hockey stick curve” trap but fell instead for the “uncertainty data” one. This link revealed a rather long and detailed paper on how to deal with gaps in observational datasets. I think it wasn’t meant for reading — one was just supposed to reel back in horror, exclaiming “Goodness, this uncertainty thing really is complicated. I must leave it to the professors”. Instead, I did read it, and all it did for me is confirm that these people have a rather narrow understanding of what uncertainty is. There are also some choice quotes in the paper that I’m sure deniers weren’t supposed to notice.
LikeLiked by 1 person
calm down John – your not a “Expert” like these proff’s are 😦
liked the partial quote from your comment – “the inclusion of the grey fuzz of “uncertainty data”: outlying data points that can be cherry-picked to raise doubts”
now that’s a true scientist/proff speaking – (I’m speechless) – what a nob.
Sufficiency or insufficiency? Uncertainty or Certainty? Sol or Man? Nicola Scafetta, Professor of Oceanography and Atmospheric Physics at the University of Naples Federico II (Italy): “The possible contribution of the sun to the 20th-century global warming greatly depends on the specific solar and climatic records that are adopted for the analysis. The issue is crucial because the current claim of the IPCC that the sun has had a negligible effect on the post-industrial climate warming is only based on global circulation model predictions that are compared against climatic records, which are likely affected by non-climatic warming biases (such as those related to the urbanization), and that are produced using solar forcing functions, which are obtained with total solar irradiance records that present the smallest secular variability (while ignoring the solar studies pointing to a much larger solar variability that show also a different modulation that better correlates with the climatic ones).
LikeLiked by 2 people
>”what a nob.”
Whilst I shouldn’t really approve of such terminology, I can fully understand why you would be moved to use it. Nobody patronises quite like a university professor. But the longer such individuals refuse to engage with the ‘deniers’, the easier it becomes for them to downplay or overlook the substantive elements of our arguments. So, whilst you and I may be more interested in the scientific concerns alluded to in Beth’s recent comment, or possibly the statistical methodologies used to create a hockey stick, it suits the good professor to maintain that we are just looking at the zigs and zags on a graph and choosing to ignore the zigs. What simpletons we all are.
LikeLiked by 1 person
Reading back my comment posted 17 August, 7.44pm, I see that it reads as a criticism of the paper cited by Professor Cheshire. This was not my intention. Instead of saying ‘these people’, I should have said ‘people such as Prof Cheshire’.
In fact, it is actually a rather good paper that highlights many of the concerns I have expressed on this blog over the years. Although it focuses upon observational dataset uncertainty and its sources, it includes the following choice quotes that illustrate the need for a deeper understanding of the concept of uncertainty, or at least an understanding that is deeper than that encountered when listening to some scientists (I’m talking here about those who use phrases such as ‘the grey fuzz of uncertainty data’ and ‘outlying data points that can be cherry-picked’). Let’s start with:
“In general terms (epistemic), uncertainty is a lack of knowledge. Representational uncertainty of a dataset is uncertainty regarding how accurately the dataset represents the phenomenon it aims to measure. Non-representational uncertainties arise because abstract properties of datasets, for example, the resolution and the baseline period in global temperature datasets, are more or less adequate for a specific purpose.”
Yes, Prof Cheshire, when it comes to hockey sticks, we deniers are asking questions about ‘representational uncertainty’ – we don’t worry about the zigs and zags until they become too large to trust the underlying signal. And, when it comes to warming stripes, we also worry about ‘non-representational uncertainties’.
Then there is this:
“One of the most fundamental distinctions can be drawn between epistemic and aleatory uncertainty, which in the terminology of Walker et al. (2003) concerns the nature of uncertainty. Epistemic uncertainty results from a lack of knowledge about the system under investigation, for example, an imperfect understanding of physical processes, and can thus be reduced by more research. Aleatory uncertainty, in contrast, is a property of the system itself, for example, natural variability in the climate system.”
If anyone reading this didn’t already know about epistemic and aleatory uncertainty, then I have been wasting a lot of my time here on Cliscep. When you talk about ‘grey fuzz’, Prof Cheshire, which class of uncertainty do you have in mind? Then there is this:
“Something can be known with certainty, expressed by means of a probability statement; by means of possibilities, we can know that we are ignorant about certain aspects or we can be completely ignorant about it, so-called unkown unkowns (see Kennedy, 2014). Furthermore, it has been noted that the assessment of the level of uncertainty itself can be subject to substantial uncertainty (for this second-order uncertainty see Knüsel, Baumberger, Zumwald, Bresch, & Knutti, 2020; Parker, 2014; Steel, 2016).”
Okay, then. It’s not just about zigs and zags that create fuzzy graphs, is it Prof Cheshire? Tell me what you know about evidence theories. Finally, there is this:
“While being a consistent and powerful framework [i.e. the Guide to the Expression of Uncertainty in Measurement (GUM)], its application is limited since many uncertainties of more complex datasets, for example, those due to the underdetermination of structural modeling choices, cannot be represented as a probability density function.”
Please try telling that to professors Gleick, Lewandowsky and Rice. I’ve been trying for a while, but I don’t think they are listening to me.
I’d love to be addressing these issues to Professor Cheshire directly, over at The Conversation, but I know that their moderation policy does not allow for such scurrilous naysaying. We just have to leave him to get on with it.
LikeLiked by 2 people
John – I meant “nob” in the nice way – as in “a nod of butter in your baked (in) potato 🙂
everything seems to be “baked in” if you follow the MSM climate/now extended to anything news – so a little “nob” needs to be added IMHO !!!
LikeLiked by 1 person
wonder who started the “baked in” meme ?
only a quick search, found – https://www.sciencedaily.com/releases/2017/07/170731114534.htm
“Two degrees of warming already baked in
Date:July 31, 2017 Source:University of Colorado at Boulder Summary:Even if humans could instantly turn off all our emissions of greenhouse gases, the Earth would continue to heat up about two more degrees Fahrenheit by the turn of the century, according to a sophisticated new analysis.”
the sophisticated new analysis –
“The new assessment by Pincus and lead author Thorsten Mauritsen, from the Max Planck Institute for Meteorology is unique in that it does not rely on computer model simulations, but rather on observations of the climate system to calculate Earth’s climate commitment.”
“Among Pincus’ and Mauritsen’s findings:
•Even if all fossil fuel emissions stopped in 2017, warming by 2100 is very likely to reach about 2.3 F (range: 1.6-4.1) or 1.3 degrees C (range: 0.9-2.3).
•Oceans could reduce that figure a bit. Carbon naturally captured and stored in the deep ocean could cut committed warming by 0.4 degrees F (0.2 C).
•There is some risk that warming this century cannot be kept to 1.5 degrees C beyond pre-industrial temperatures. In fact, there is a 13 percent chance we are already committed to 1.5-C warming by 2100.”
so they do not rely on computer model simulations, but give “(range: 1.6-4.1) or 1.3 degrees C (range: 0.9-2.3) !!!!
@John – may not be best place for this comment so move or delete as you see fit.
LikeLiked by 1 person
The concept of “warming baked in” is actually just clever sounding bs.
There is no science behind the assertion. It is nonsense, not science.
Yes Ed, but your “show your stripes tie” is suspiciously bluish. Does this mean you’re a cool(ist) dude’s ? If you think you are, I would have thought his pre-kingship would have cast you into his deepest, dankest dungeon.
LikeLiked by 2 people
Fiery maps and warming stripes are so last year. What you need to do to properly understand the true peril we are in is to turn the temperature data into music. Step aside Ed Hawkins:
“The statement shirts making climate data fashionable at Cop26
With stripes showing temperature rise in the polar regions, the shirts are designed to start conversations about global heating”
“Cop26 may not be the first place that springs to mind when thinking about fashion, but stripes were all the rage in Glasgow this week, where striking white, blue and red shirts were the latest must-have piece for fashion-forward scientists. These statement shirts may look like some 70s throwback but they have been designed using the latest polar data.
Climate data is not a natural conversation starter but the shirts, made by Cambridge-based company DressCode Shirts, are designed to get people’s attention. They tell the story of two datasets, which show how rapidly polar regions are warming, using soft blues to represent colder temperatures and reds to show hotter ones.
The stripes on the body of the shirt represent 70 years of warming in the Arctic, from 1950 to 2020. On the cuff is a graph depicting 800,000 years of Antarctic temperatures and CO2 data, which shows how closely the two are linked. The single black button represents the ozone layer.
“We’ve had some jaws drop today,” says Pilvi Muschitiello from the British Antarctic Survey (BAS), who helped design the shirts and has been modelling them in the green zone. “A lot of people are very enthused when they see the shirt. People are surprised just looking at the CO2 levels and the historical record, as well as where we are today,” she says.
Scientists from BAS are researching polar regions, looking at a range of issues, from what iceberg carving does to ecosystems to how melting is making them less habitable for polar wildlife. The aim of the clothing line is to help people engage with complex scientific data.”
Yes, stripy shirts – that should do it.
Mark – stripes are the big thing & everywhere – https://www.bbc.co.uk/news/uk-scotland-59185007
scroll to the bottom (or read if you can be bothered) – stripes with no explanation I can see !!!
having said “read if you can be bothered” it’s worth a read as we now have a new link (to me anyway) “https://scientistrebellion.com/”
partial blurb extract –
“Still, mega-corporations ransack the natural world with support from their servants in public office. Governments who stray from protecting corporate interest in favour of human need are attacked and delegitimised in the billionaire press, face the prospect of international capital flight, and of political or military coups. This corruption of democracy sits at the heart of climate inaction…
read on to find out “Why We Rebel”
There have been a number of press reports recently that refer to the ‘debunking’ of the ‘daft’ suggestion that the new Met Office weather maps are designed to visually accentuate high temperatures. These are based upon Aidan McGivern’s tweets explaining the adoption of the new presentation standards:
He correctly points out that there is a doctored weather map doing the rounds on the internet that misrepresents the use of the new standard. This much is true.
He also points out that the new standard has been adopted to help the colour-blind. Insofar as the new standard employs Colorbrew palettes, this much is also true.
And he says that that’s all there is to it – there is no conspiracy to alarm, and the presence of extremely dark red hues on today’s maps is entirely down to unprecedented temperatures. Unfortunately, this is untrue.
In his own tweet you will note from his example maps that the new standard uses significantly deeper hues to represent the same temperatures. This has nothing to do with the adoption of Colorbrew and the problem of colour-blindness — it must have some other purpose. So the legitimate question that remains is this: When Aidan adopted Colorbrew, why did he also take the opportunity to change the colour scale? Unfortunately, the legitimate adoption of Colorbrew to address colour-blindness and the presence of a doctored map on the internet gives Aidan the ideal material with which to ‘debunk’ a perfectly legitimate accusation.
Personally, I suspect that he has adopted a more dramatic colour scale in order to be consistent with that used by Ed Hawkins. The hype goes on.
LikeLiked by 1 person
I now look forward to a Reality Check from one of the BBC’s intrepid climate disinformation specialists.
LikeLiked by 2 people
McGivern now says that the image wasn’t doctored:
I can’t find the July 2016 Tweet but here is the ‘doctored’ map in a July 2016 article in the Mail:
Need a more credible source? Here’s most of the map in a 2018 article by The Priestley International Centre for Climate:
‘That’s why I know: – The image on the right is doctored’
McGivern has changed his mind about this. I posted a comment a few hours ago with links to a tweet by him and to articles using the ‘doctored’ map. The links seem to have spamhaused me.
Here’s the July 2016 Met Office tweet with the ‘doctored’ map:
And here’s a bigger version of the ‘ridiculous comparison’ that for two days McGivern insisted was based on a fake map and was ‘just the latest example of a vocal minority trying to spread misinformation in response to the Met Office’s science-based weather and climate forecasts’:
Can anyone spot any differences?
The ‘It’s Called Summer’ pic does make a silly comparison but it’s not misinformation. McGivern’s Twitter thread, however, and his interviews in the Indie etc…
Vinnie, I have now dragged you out of the pit. Sorry about that. WordPress things. As you say, it seems not to like the links – even with a commenter like you who has made multiple contributions in the past.
LikeLiked by 1 person
I stand corrected. The ‘doctored’ map appears to be genuine after all. The reason that I accepted the assertion that it is doctored is because it doesn’t appear to be using the latest colour/temperature calibration as exemplified in McGivern’s tweet. Can someone help me out here?
“New style weather map – designed to look like fear and destruction”. So what? Wasn’t the map part of a package of measures including a new red warning and weather warnings on TV, radio and newspapers warning the public about the very real dangers of this particular heatwave. Dangers that were real. I see absolutely no problem with the maps, they performed their function well.
There are comments that the colour palette is not “scientific” but the data on which the map is based are depicting anticipated temperature measurements – hardly high scientific predictions.
Judge the map based upon what actually transpired – uncomfortably high temperatures causing distress and danger for some (many?) and widespread fires. Good predictions in my book.
LikeLiked by 2 people
Previous comments on this thread alluding to ‘unscientific’ use of colour related to the use of a red-green-blue palette and to Hawkins’ reference to an aesthetic basis for calibration. Neither issue is relevant here. What is of greater concern to me is the veracity of McGivern’s explanation and whether it adequately explains the reasoning behind the new standard. I think it fails on both counts. I grant you that the new colour standard achieves the desired effect and is consistent with current levels of risk tolerance.
LikeLiked by 1 person
wonder how countries that have these temps year in year out cope ?
And I wonder what their temperature maps look like.
Of course the stock answer to your question is to say, ‘ah but our infrastructure is not designed for such heat’. The same design argument applies in winter when it snows. It makes you wonder what it was designed for. A drizzly Tuesday afternoon?
LikeLiked by 1 person
John, you’ve answered your own question: infrastructure is not designed for exceptional weather extremes, hot or cold. Most countries become unstuck eventually.
Have you noticed that Britain has transformed itself from being a “Land of Weather” into being a “Land of Climate”.
I have absolutely no idea how I managed to like one of my own posts within This thread. I have been unsuccessful in my attempts to like the work of others for months. Needless to say my own self-aggrandisement was an error.
LikeLiked by 2 people
Agreed, Alan. We build for changeable weather and not for changeable climate, in which weather extremes may evolve. Thus, in answer to the question as to what constitutes the optimum climate, the answer is always the climate we have, because that is what we built for. The recognition that climates do change is tantamount to the acceptance that adaptability must form a major part of our risk management strategy.
LikeLiked by 1 person
John – a song springs to mind.
An interesting parallel from ‘stralia:
LikeLiked by 2 people