I wonder whether Cliff Mass can comprehend what he himself has done?
He wants to persuade people that the upward trend of May-through-July temperature in his home state of Washington is “really very little.” How does he do that? Not by analysis, not by any objective criterion of what’s “really very little,” just by saying so:
What about temperature? Here is a plot form the NOAA “Climate at a Glance” website showing May-July temps over Washington State. 2015 is the warmest, although 1958 was close. But there is really very little upward trend.
Still, it seems that the trend is a bit too much for him to be comfortable with, which brings out the truly “cherry” in Cliff:
Let me show highlight by getting rid of 2015 and before 1925 when the record was not reliable. You see much trend? Not really.
Yeah, you got that right. After concluding that the trend is “really very little” from no more analysis than saying so, he reinforces his belief by deliberately removing the data he doesn’t like.
You can’t get much more cherry than that.
I suppose I could point out the the linear trend in May-July temperature in Washington isn’t just upward, it’s statistically significant. I could mention that the trend isn’t actually linear, and that although it’s quite uncertain, a good estimate of the warming rate right now is about 4 deg.F per century. We could argue whether or not it’s really that high (in which case I’d point out that it could in fact be a good bit higher). We could argue whether or not that’s “really very little.”
But, removing the data you don’t like? Getting rid of the earlier, colder stuff with a lame excuse? Getting rid of 2015, apparently because it’s the hottest May-July on record?
That’s classic cherry-picking. It’s a textbook example of cherry-picking. Look it up in the dictionary, Cliff.
Open Mind 
I started a few comments but dumped them all… now I must let the Onion speak – (today!) http://www.theonion.com/article/climate-change-deniers-present-graphic-description-51129
Surprised that it actually has relevance …
“Climate Change Deniers Present Graphic Description Of What Earth Must Look Like For Them To Believe”
“…Additionally, Davis said that for the community to begin believing a single word of any scientific journal article corroborating climate change, every one of Earth’s glaciers would have to retreat at a rate exceeding 20 miles per year, and each of the skeptics, individually, would have to go a decade without seeing naturally occurring ice anywhere….”
The Onion isn’t joking this time.
I’ve often wondered if the average person on the street who seems to be sort of sanguine about climate change might be having difficulty in grasping that 2C (3.6F) of warming is a really big deal. After all, daily temps may fluctuate between 5,10, even 15 degrees or more sometimes, and the world doesn’t end…right?
I’m not being snarky about this either. I really do have a gut instinct that it just doesn’t sound like something that someone who doesn’t follow climate science would see as a significant threat. Has anyone seen any research about this? If it’s true, we need to find a better way of translating that seemingly small number into credible effects that people can understand.
To skepticmac57, one has to remember that we’re talking about an average global temperature, not a day to day fluctuation that when averaged out over a season or a year or a decade generally centers near an average that doesn’t change much baring any climate forcing. However, an increase in the average global temperature of 2C means that an enormous amount of energy has been added to the system in order to get the global average to go up that much. Conversely, if the average goes down 2C, an enormous amount of energy has been removed from the system. For a situation where the global average temperature goes up that extra energy is available to do many things that it didn’t do before the increase occurred. These things would include increasing the amount of water vapor that the atmosphere can hold, increasing the amount of drying out of soils that occurs, increasing the amounts of water evaporated from the oceans, changing the general circulation of the atmosphere as well as altering regional and even local circulations which in turn change the patterns and frequency of weather systems that effect any particular location.
Of course, the added energy also makes heat waves more frequent and extreme by moving the means and the extremes within a typical variation for a location in the warmer direction while it also makes cold waves less frequent and less extreme for the same reason. In addition, a warmer atmosphere tends to increase the transport of tropical air toward the poles which warms the poles and which in turn lessens the temperature gradient that drives the whole general circulation to begin with. A lesser temperature gradient between tropics and poles means a weaker jet stream which in turn means a more sluggish jet stream which has a greater tendency to form blocking systems which get stuck in place for much longer than they would with a more vigorous jet. Stuck blocking systems in turn lead to stuck weather systems which just keep dumping precipitation or extending drought conditions over the same place over and over again, which in turn leads to extreme flooding or snowfalls on one hand or killer droughts on the other depending on which side of the block a certain location happens to be when it sets up.
Anyone who has been paying attention to weather patterns globally for the last few decades would have to conclude that such things are in fact happening!
Ed, yes I understand all of that, since I have been following what climate scientists have to say for the last 8 years since I first got interested.
The problem that I run in to when discussing what I’ve learned with people that I know, is that even when they do accept AGW, they still see it as some vague, far off, abstract concept, and since the media endlessly keeps repeating that “We can’t say for sure that (name your disaster) can be attributed to global warming…” then they just go on about their lives with a shrug. And that’s the people who do accept AGW. The ones that don’t will grasp at any argument that supports their bias, and think the rest of us are a bunch of mindless dupes.
But both camps aren’t really interested in doing anything about it, or learning what the credible climate scientists are actually saying. Denial or complacency…pick your poison.
skepticmac57, I get that all the time, too, and it sometimes makes me want to bang my head against a wall. All I can do is try to explain about the added energy and that the effects of that energy have been gradually showing up with more frequency over the last 30 or 40 years (I’m a professional meteorologist who’s been watching this stuff since the late 1960s). Some people will never get it since it goes against their carefully constructed world view that seems to be burned into their brains. Then we have those who say they’re concerned about things happening 50 or a hundred years in the future (like Cliff Mass); they’re probably a worse problem than outright deniers since they have some professional credibility that can easily make the average person complacent and uncomprehending of what we really face. But we do what we can, I guess.
A useful fact to put a perspective on global temperature fluctuations is that in the Last Glacial Maximum (which was a big deal) the global average temperature was 4-5 degC lower than now.
I posted a link to this article on Cliff Mass’s blog.
I don’t approve of the attendant celebration of p-values, but perhaps Mr Mass might benefit from the “Hack your way to scientific glory” at http://fivethirtyeight.com/features/science-isnt-broken/
Tamino, mush as I usually enjoy and learn from your posts in general, in this case I think you have fallen well below your usual standards. To begin with, while Cliff Mass’s post contained no analysis, neither does yours. The closest you come is mentioning that “the warming rate right now is about 4 deg.F per century” without defining what is meant by “right now”.
Because if this, I did my own analysis of the temperature series, and found that for the full temperature record (1895-2015) the linear trend is just 0.52 deg F per century. Even though that is statistically significant (2.13 St Dev), I think such a low warming trend matches Cliff Mass’s claim that there is “really very little upward trend”. Worse, if we remove the 2015 data, the trend becomes 0.35 deg.F per century, and is not statistically significant. While I agree with you that removing the data from 1895-1924 is a cherry pick, Mass’s claim is that 2015 is an unusually hot year for reasons not related to the cause of the overall trend. That being the case, to test that claim it is appropriate to determine the trend absent 2015 data. Regardless of whether or not you agree with that, however, the claim that 2015 is not unusually warm due to the trend is justified by the full data in any event.
I find the trend from 1975 (used because you have previously identified that year as significant from a global break point analysis) to be 3.67 deg.F per century (2.36 deg.F per century to 2014). The result is statistically significant whether the trend 1975-2015 or 1975-2014 is taken. However, using that short trend, unless explicitly defended, is as much a cherry pick as Mass’s truncation of the early data.
Finally, Cliff Mass wrote:
Those are not the words of a denier. Indeed, while 2.4 deg.F is a low temperature increase estimate for the end of century, Mass has repeatedly argued on his blog that the Pacific North West will experience unusually low temperature increases relative to global averages. Consequently there is nothing in the blog post you criticize which disagrees with the IPCC consensus. From my reading of his blog, Cliff Mass is one of those climate scientists whose estimate of climate sensitivity and impacts are probably in the lower half of the IPCC ranges. If the IPCC has done its job correctly, approximately 50% of climate scientists are (including James Annan and John Nielsen-Gammon). That is a perfectly respectable intellectual position, and people ought not to be vilified for holding, or expressing that position.
[Response: Listen to yourself: “if we remove the 2015 data…” How does one justify that? Also, I can’t for the life of me figure out how you got 0.52 deg.F/century. Using May-July temperature I get over 1 deg.F/century, using all months it’s closer to 1.5, figures 2 to 3 times as high as you quote. And let’s face facts: using a *linear trend since 1895* to estimate the warming rate in order to minimize future effects (which Cliff Mass does) is misleading.
As for his admission of global warming, phrases like “that is in our future” and “but we will have decades to do so” fit in with his “no hurry, no worry” agenda. That is the approach of someone in denial.]
Tom…
If you remove the 2015 results for being an “unusually hot year for reasons not related to the cause of the overall trend”, shouldn’t you remove ALL such years? 1957-8 was, you may not know, a strong el Nino year (http://ggweather.com/enso/oni.htm). Shouldn’t we remove it for being an “unusually hot year for reasons not related to the cause of the overall trend”?
Actually, the procedure you suggest IF carried out correctly isn’t such a bad idea. What do you think the results of removing ALL exogenouos variables on Washington temps (e.g., el Nino, volcanic aerosols, and solar variations a la Foster & Rahmstorf 2011) would be?
BTW, while the present el Nino may well develop into a very large influence, to date it has not yet developed into even a moderate el Nino let alone a strong or very strong one.
jgnfld, 1957-8 was not a strong El Nino year. In particular the six month lagged inverted SOI for 1958 was just 5.65 – just 29th out of the 1877-2015 record (as lagged). That compares to the 14.28 for 1998 (ranked 5th), 20.24 for 1983 (ranked 1st), and 7.78 for 2015 (ranked 22). Whatever caused the extreme summer temperatures in Washington state, it was probably not El Nino. It is instead likely to be associated with the pool of warm water (the blob) that has been lying of the west coast of North America for quite some time now.
And I did not remove 2015 because it was unusually hot.
NOAA disagrees with you about the 1957-58 el Nino. I’ll go with NOAA.
“You” is a generic pronoun. Mass actually said that as you yourself note in your post: “Mass’s claim is that 2015 is an unusually hot year for reasons not related to the cause of the overall trend.” I was speaking to that notion.
What evidence does Mass have that the heat in Washington this year is not caused by AGW? That is the basis of many of his claims. We do not expect heating to be a monotonic rise. Perhaps Washington has had less heating by luck for that past few decades and it has now heated more. It appears to me that his claim is not supported by analysis of data. Can a peer reviewed paper be cited that shows the heat in Washington is not caused by AGW?
I still don’t see any defence of the removal of 2015. It’s certainly true that “2015 is an unusually hot year for reasons not related to the cause of the overall trend.” but the whole data set is full of years that are unusually hot or cold for reasons not related to the cause of the overall trend. Saying that you’re removing one for that reason is obviously nonsense.
I think that the philosophical confusion I pointed out the last time we discussed a Mass post applies here: the ‘trend’ is not a physical cause. Rather, it is a statistical measure of the weather events that occur.
Hence, arguing that statistical outliers are obviously to be disregarded for purposes of attribution is bass-ackward: they do their bit (and more than the ‘average’ bit at that) to create the trend.
Ex post facto, the statistics should square up, of course. But there’s something deeply off-kilter about the expectation that Mass seems to hold about trend and statistical outliers: I suspect that his intuition would incorrectly predict what a full accounting of variance would show. Or so says my intuition, anyway; such an accounting is well beyond my capabilities. Still, I’m pretty sure I don’t have the causation here backwards.
While long term linear fits are the simplest example of “trend” one can posit, they don’t necessarily make for a good description. Using something which is arguably one of the best, yet still simple models, a level-step Kalman smoother, one which admits the possibility that even trend changes by year, the least recent per century rate of change I got, globally, was in excess of a degree per century:
Note this reports a (varying) slope of degrees per year, so you need to multiply them by 100.
(That’s from the page https://johncarlosbaez.wordpress.com/2014/06/05/warming-slowdown-part-2/, which also discusses ways of trending.) This uses only up through 2012 data. Moreover, the same figure suggests an extrapolation might be between 1.05 and 1.7 degrees per century, and does not include priors which embody geophysical knowledge of where consensus estimates put warming anomalies, conditional upon information that is simply not in the data, like expected CO2> emissions profiles.
(Ugh, botched the post. I forgot images cannot be simply linked. Sorry.)
While long term linear fits are the simplest example of “trend” one can posit, they don’t necessarily make for a good description. Using something which is arguably one of the best, yet still simple models, a level-step Kalman smoother, one which admits the possibility that even trend changes by year, the least recent per century rate of change I got, globally, was in excess of a degree per century:
Note this reports a (varying) slope of degrees per year, so you need to multiply them by 100.
(That’s from the page https://johncarlosbaez.wordpress.com/2014/06/05/warming-slowdown-part-2/, which also discusses ways of trending.) This uses only up through 2012 data. Moreover, the same figure suggests an extrapolation might be between 1.05 and 1.7 degrees per century, and does not include priors which embody geophysical knowledge of where consensus estimates put warming anomalies, conditional upon information that is simply not in the data, like expected CO2 emissions profiles.
Along with what Neven says here, add this classic denier line that Mass uses:
“Regarding the eastern Washington wildfires, there is another issue. The eastern slopes of the Cascades have burned for millennia.”
Allow me to reframe – “well this dead body in a pool of blood in the street has a knife sticking out of its back, but people have died of natural causes in the past, therefore this person died of natural causes”
What caused the 1958 spike?
tamino: Kevin Trenberth:This applies to a now closed article where you told me: “At least it’s a bit amusing that you seem to think one learns one’s statistics from the physics department.]”
As a matter of fact, I did take my first probability and statics course form the physics department at Carnegie-Mellon University in 1965, Call the registrar if you want. I got my BS degree in physics at CMU in 1968. It was a different course from the Prob&Stat that the math department taught. I don’t seem to think. I know. Don’t be insulting.
More information if you want it is available.
EG–I’ll be darned! I majored in physics at CMU from 1978 to 1980, then switched to comp sci at Pitt. Didn’t know you’d been in Pittsburgh (or perhaps still are?)
As I’m sure fools no one, the aggregation period of May-Jul ignoring the rest of the year and even fire season maximizes the flat appearance as well should you examine nearby and shorter/longer aggregation periods. Since we are talking wildfires, the graphic should cover fire season, one would think. According to the state fire marshal (http://www.wsp.wa.gov/fire/wildland.htm) “The wildland fire season in Washington usually begins in early July and typically runs through late September, although drought, snow pack, and weather conditions can expand the length of the fire season.”
I don’t know how, or if I can, insert a graphic here. In any case, here is the Jul-Sep (fire season) graph since 1900 which does not even include this year, of course…
Further examination shows the cherrypick heavily involves an extremely warm May in 1958 which is well before normal fire season. Jun and Jul 1958 were still warm, but not so extreme relative to the rest of the record as May. Wonder why May was included? Wonder why Aug-Sep excluded?
Seeing an honest graphic from a denier is well nigh impossible. It just does not happen.
I’m embarrassed to say it fooled me and in the previous FIRE post I used the May-Jul period chosen by Mass. In my defense the game of uncovering all their tricks and dodges gets tiring after a while.
The 30-year trend for the statewide average temperature for Jul-Sep from 1985-2014 is +0.6°F/decade. Hardly negligible.
http://www.ncdc.noaa.gov/cag/time-series/us/45/00/tavg/3/09/1895-2015?base_prd=true&firstbaseyear=1895&lastbaseyear=2014&trend=true&trend_base=10&firsttrendyear=1985&lasttrendyear=2014&filter=true&filterType=loess
The trick with denier graphics is to assume they are dishonest, or at best ignorant, first. Because that is what they are. Then check for what common tricks they have employed. Cherrypicking, employing irrelevant variables, and fiddling with axes/labels are the most common tricks used as far as I can tell. Just checking those 3 will uncover the vast majority of the tricks.
+1
Plus, there is no excuse for Mass asking “You see much trend?” when he could have added a trend line and a smooth with a couple of clicks.
jgnfld, I hardly think it a reasonable criticism of Mass that he has not checked the July-Sept temperature data in mid-August.
That is not the criticism and you really ought to know that if you read my post.
The criticism is to use May data to talk about the fire season for many decades, not just one year, when May doesn’t even reflect the fire season by a goodly margin. If you are going to talk about avg temps in the fire season, perhaps you should consider using avg temps in the fire season.
That very much _is_ a “reasonable criticism”.
Mass seems not to know how valuable the fruit industry is to Washington, or how small a change in average temperature it takes to spoil a crop.
Tamino:
My error. When downloading the May-July average temperatures from NOAA’s Climate at a Glance page, I used cut and paste from the table, then used a find and replace function to turn it into a csv format. In doing so, I accidentally eliminated all negative signs.
Correcting for the error I find a trend to 2014 of 0.89 deg.F per century, and to 2015 of 1.09 deg.F per century. That is 0.49 and 0.6 deg.C per century respectively. These are low trends, but do not match Mass’s description of “very little warming trend” so I withdraw that part of my claim.
One justifies it very simply by noting that trends are used to predict potential temperatures. A prediction of the 2015 data that includes the 2015 data is no prediction at all. Therefore we are always justified in removing the final years data in testing whether or not that years data is what we would have expected given the prior record. More generally, we are always justified in removing the data for the years we are retrodicting from the dataset used to retrodict those years.
Treating the analysis as a prediction, then we expected 2015 to have a temperature anomaly of 0.54 +/- 2.81 deg.F. Treating it as a prediction, including 2015 in the data set is bad practice. It gives us access to data we would not have had in a true prediction, and therefore should not be included. Treating the analysis as a prediction, however, we note that the 2015 temperature lies 3.53 standard deviations above the mean prediction, and hence was not predicted using an OLS + normal distribution model.
I am certain you can improve on the model, both by using a loess smooth (as you prefer) and by improving the model of the residuals (which are almost certainly not normally distributed). But however you improve it, 2015 will almost certainly not be in the prediction interval, and including the 2015 data to make the prediction would be a cheat (and still almost certainly not include 2015 in the prediction interval).
Finally (on this point), explanation is just prediction. Ideally we predict by a physical theory, and if the physical theory makes the prediction it explains the phenomenon, and if not it does not. At need we make statistical predictions, and if the prediction is made the explanation of the preceding data in agregate also explains the result, and if not, not. (This allows, of course, that we can have multiple theories predicting, and thus explaining phenomenon.) It follows that as the statistical data to the extent that you have analyzed it does not predict the 2015 temperature, neither does it explain it. In particular, it is not shown that the cause of the overall trend, or the trend since 1975, explains the very high temperatures in 2015.
Frankly, his “no hurry, no worry” agenda (your words, not his – so they should not have been in quotation marks) is a figment of your imagination. He frequently argues that the Pacific North-West, by which he really means the US north of the Klamath, and west of the Cascades, will be a refuge largely spared the worst climate impacts felt elsewhere in the US. But he also argues that Washington state should be active in mitigating AGW now, including instituting a carbon tax and preventing Washington from providing port facilities for coal exports. He also argues that Washington should be already planning for adaption to climate impacts late this century (despite his belief that they will be relatively mild compared to the rest of the country). His mitigation agenda (carbon tax plus no more coal) very closely matches that of Hansen.
As I said previously, he is (from my reading) in the lower half of the IPCC consensus, but that is a perfectly valid position, is not scientific, and is not reason for insults.
[Reponse: Your “justification” for removing 2015 is total bullshit. It’s not even very good bullshit.
As for “insults,” the only things I see in the post are “cherry-picking” and “denier mode.” I regard him as a denier; you’re entitled to your own opinion about that. You are not entitled to your own opinion about him cherry-picking, that’s just a fact.]
IMO there are too many separate things being mixed up here.
First, prediction versus description (here called “explanation”, apparently): Predictive intervals are different from “confidence intervals”, where, from a Bayesian perspective, the latter are “highest probability density intervals” (or “HPDI”), since the Bayesian posterior can be multimodal. To do a predictive interval, something needs to be done to assess the out-of-sample versus in-sample error. That means that whatever dataset one has in hand is a draw from a sampling frame of all possible such datasets. Not only can points be less probable than a “typical value”, but an entire (temperature) dataset is a member of an ensemble of possible temperature datasets. Earth’s climate is decidedly not a deterministic process.
There are several ways to assess the chance of a dataset and its statistics being atypical. One popular way is cross-validation. It can also be done using applications of the bootstrap, a kind of resampling technique. For various reasons too complicated to get into here, a proper Bayesian assessment does not need cross-validation to calibrate its predictions, or its HPDIs. (It doesn’t hurt, but you’ll always find that the Bayesian result will agree with the validation result, to within statistical variability, and, so, is a waste of effort.) There are different ways of doing cross-validation and so, these are not “objective” assessments either. There is also a generalized cross-validation technique, but that’s beyond the scope of this comment.
Second, the problem of comparing “multiple theories” is way more complicated than, say, seeing which one has the least, say, squared error loss with respect to a dataset. The comparison can’t properly be done unless, first, the space of tunable parameters is somehow standardized. That’s hard to do outside of a Bayesian context, where integrations (often numerical integrations) over spaces weighted by priors do the job. Otherwise, the comparison can only be done using “models” which have exactly the same sets of tunable parameters and domains for those, or if asymptotic measures like the various information theoretic criteria (AIC, AICe, BIC, WAIC) are used. Results using these are provably equivalent to cross-validation, at least asymptotically.
Third, comparisons of models with one another, as well as prediction intervals, really need to consider the serial correlation of the series, even (or especially) if they are (temperature) anomalies, and especially if frequentist methods are used. So, for instance, if a bootstrap is used, it’s important to use something like the Politis and Romano stationary bootstrap. (I don’t like the description of that at Wikipedia. Check the original paper, which is referenced there.) If Bayesian methods are used, it’s worth doing the assessment with dynamic linear models. (For further details, see papers and textbooks by Durbin, Koopman, S&aunl;rkka, West, or Harvey.)
From my personal technical perspective, the reason why Bayesian methods are superior is because it’s clear how to bring to their assessments knowledge from pertinent fields: These can be incorporated in the likelihood models and they can be incorporated in prior probabilities on parameters. There are ways of using regularization to achieve the same for likelihood-based methods, but that always seemed roundabout to me. Of course, then the problem is, if someone (as most students of these problems in geophysics and meteorology, in my experience) goes and does a purely frequentist assessment, there’s no proper way of comparing their result with a fully informed Bayesian one: It’s not that they are necessarily different. But even if they agree numerically, they are not the same thing.
Mass doesn’t say “let’s make a projection”.
He says “let’s see if there is a trend in May-Jul temps”. Last I looked at the calendar, we’re well into August, meaning the 2015 May-Jul should absolutely be included.
Larry, the specific criticism here is not of Mass’s removal of 2015, but of my removal of 2015 in my secondary analysis of the data. (I also analyzed the data including 2015, for like Raff I believe that a statistical analysis that requires any particular data point to be included to make its case is not robust, and should not be relied upon. Of course, likewise, if the argument requires that any particular data point be excluded, it is equally not robust and should not be relied upon.
One also needs to justify the nature of the trend in the data being used to project foward, and in doing so one also needs to include the physical basis of the parameters underpinning the trends. Your analyses appear to be linear extrapolations, and there is absolutely no basis for applying such across extended time intervals (especially on the centennial scale) – Tamino has previous posts looking at the the time-related distribution of residuals that go into this.
If one is going to be conservatively reductive/simplistic in the statistical technique used to identify trend, one should also be conservative in the concluding expression of consequence extrapolated from the statistics. In Mass’s case – and in your defence of him – I don’t see any justifiable care in the use of the conclusions of the statistics.
Could someone please comment on the appropriateness of Mass’s first graph –
Why use the entire Pacific Northwest? Is info not available just for Washington?
Is the 5-yr running mean the most appropriate way to display this?
NICC breaks out its reports by region: http://www.nifc.gov/nicc/predictive/fuels_fire-danger/Julian_Weekly_Year_to_Date_IMSR%20.pdf
I enjoy reading Cliff Mass. He provides great explanation of NW weather events. But he struggles when he posts on climate and his responses when challenged are often defensive or calculated to obscure. For instance when glacier decline is mentioned he just says they’ve been declining for a long time without mentioning that the decline slowed in mid century but has gone into overdrive since the mid 70s.
Papers and other evidence that don’t support his climate claims (on snow pack or jet stream waviness or forest fires) are either ignored or discredited (often just on his say so rather than by any reference to peer reviewed literature). And his analysis of trends is laughable. Don’t just say “Doesn’t look like much of a trend”. Tell us what the trend is or isn’t.
That said I’d hesitate to call him a “denier”. He accepts the consensus and accepts that we need to do something. To me “denier” means something significantly more egregious that being wrong or misguided about one aspect of climate change
Maybe he couldn’t be described as a denier based on a rigid definition of the term, but how you just described his actions could be interpreted as intellectually dishonest. When someone has to calculate and obscure their responses while being defensive, they are not to be trusted as a reliable source of information.
Whether someone agrees with someone is no criterion for determining whether or not they’ve calculated something right or wrong. “If even the devil states that 2 x 2 = 4, I am going to believe him.” — P. P. Waldenstrom, famous Swedish religious leader of the 19th century
I don’t regard Cliff as a reliable information source on attribution of climate change because he has displayed too much selective blindness on evidence that points to a more immediate impact of climate change in the NW. and too MUCH focus on that subset of the evidence that points to a more mild or delayed impact (in the NW).
That said I can’t regard someone as a “climate denier” who repeatedly and unreservedly states that climate change is a big and important issue that needs to be addressed. And I especially can’t call someone a denier who not infrequently uses his public platform to advocate for action on climate change.
I have a confession to make. When trying to counter “skeptics” I have often argued that when looking at a data series, the removal any individual data point should not affect the interpretation of the data. So for example, when looking at a temperature series that has a huge peak at 1998, I have argued that the “pause” disappears when 1998 is removed (set to the average of 1997,1999). I don’t think I ever convinced anyone, but I believed it. Was I wrong?
First, a denier is a kind of pseudo-scientist. As such, calling somebody a denier when they are not is most definitely an insult.
Deniers do not publicly endorse a $25 per tonne carbon tax. Nor do deniers campaign for such taxes, not just in blogs, but in mass reach media as well. They do not say things like:
Nor things like:
Cliff Mass does.
It is an abuse of language to call Mass a denier, not to mention an entirely unwarranted insult to Cliff Mass himself.
Second, we would rightly be scornful of a method of predicting a value which only worked if the value was included in the data used to make the prediction. If it were the case that the data including 2015 included the 2015 value in the prediction range, while with data only to 2014 it did not, insisting that we include the 2015 data would put your method in that absurd position. It follows that making the analysis excluding 2015 is perfectly valid.
[Response: What fucking “prediction” are you talking about? Yours? Cliff Mass’s hope was to convince people the trend is very small — and to do that, he removed the hottest year (*as well as* everything before 1926 — was that part of this “prediction” too?). His cherry-picking is so egregious you had to invent a “prediction” to justify removing 2015.
You argument is such bullshit, it’s far more insulting to the intelligence of everybody here than anything I even came close to saying.]
If you cannot see that, you have a problem. If you cannot even recognize it as a relevant point, you only show that statistics without understanding of philosophy of science is as marred as statistics without physics.
I am not so disappointed with your post above. People make mistakes. But your inability to recognize that excluding 2015 is not an error (indeed is preferable), and in particular you inability to recognize that Cliff Mass is not a denier despite very clear evidence that he is not are deeply disappointing.
[Response: Your inability to see the utter crap of your arguments is beyond “disappointing.”]
TC: Deniers do not publicly endorse a $25 per tonne carbon tax.
BPL: To be effective a carbon dioxide tax would have to be at least $120 per tonne, which is $33 per tonne. And that figure was calculated several years ago, and should probably be $130 or $140 by now. So it’s possible that a denier could call for an ineffectively low carbon tax.
Perhaps he’s a climate risk denier? As in: there is absolutely no chance that AGW could cause any serious damage.
BPL, the specific proposal supported by Mass is that:
That is the text of the ballot initiative that Mass has endorsed. Given that final value is just 25% less than the value you recommend, it is rather a big call to call him a denier on so slender a basis.
Further, the proposal he is endorsing is not his. He has said that he would prefer the tax to cap at a $5 per gallon price total price on petrol. $100 per tonne of CO2 works out at 90 cents per gallon on petrol, whereas a $5 per gallon represents a $1.50 increase on the price of petrol, suggesting Mass would prefer a higher tax (although that is complicated by the issue of petrol price increases).
Further, focusing just on his views on a carbon tax ignores his proposal for a total ban on coal exports.
NevenA:
Given that he thinks “there will surely be serious climate impacts”, it is hardly reasonable to suggest that he thinks “there is absolutely no chance that AGW could cause any serious damage”.
In that case there is something of a gap between his statements and his actions.
This is a sweeping statement that is not necessarily correct. Deniers of climate change could quite easily endorse a carbon tax because they are concerned about non-radiatively-active consequences of carbon emissions, or because they believe that a finite, limited resource should be husbanded with care. One could still vehemently oppose appropriate pricing of carbon to mitigate against climate change but still want to price it for the other two reasons.
Note that I am not commenting on the nature of Mass’s endorsement of a $25 price. All I’m saying is that it needn’t be inconsistent with denial of science, whether of either the complete or the lukewarm varieties of denial.
Tom Curtis:
“First, a denier is a kind of pseudo-scientist. As such, calling somebody a denier when they are not is most definitely an insult.”
Roy Spencer is not a “pseudo-scientist”. Yet he denies evolution. It’s wrong to call him an evolution denier because he does satellite temperature reconstructions?
dhogaza, if Spencer claims his denial of evolution is scientifically justified, then he is a pseudo-scientist on evolution. Frankly, even in climate science some of his work has been pseudo-scientific (although not all), and it is that work and his even worse public commentary that lead me to call him a denier.
Actually, no – not always. In logic terms yours is a fallacious statement, and I’m surprised that someone with your philosophy background would make such a proposition.
$33 per tonne of carbon, I meant to say.
Thanks, that’s a helpful clarification.
BTW, note that the British Columbia carbon tax is $30 (CDN) per tonne CO2e, and appears to have restrained that province’s emissions measurably, so I wouldn’t call it ‘ineffective’–though I wouldn’t call it ‘sufficient’ either; that will take co-ordinated action across more jurisdictions, which in turn would allow BC to either broaden the base of the tax, or raise rates.
Click to access Carbon_Tax_Review_Topic_Box.pdf
http://blogs.worldbank.org/climatechange/british-columbia-s-carbon-tax-shift-environmental-and-economic-success
Washington state is attempting to follow BC’s carbon tax policy by getting I-732 on the ballot: http://carbonwa.org/
Doc, you may recall I live in BC and may be interested in some opinion on the carbon tax. I agree that it is insufficient (when people readily hop into their vehicle on a Sunday morning to buy some grocery item they didn’t put on their shopping list the day before, you know that energy is cheap). Our government has stated that it will not raise the tax until other governments implement one. We implemented it incrementally over a number of years, so that people could make adjustments. I angers me to hear politicians in other jurisdictions making claims that such a tax is an economy destroyer. This is willful ignorance in the extreme. Fortunately, most people here accept it as good policy. When I meet regular wage earners who are opposed to it, they generally have no knowledge of the actual amount paid on, say, a litre of gasoline or a gigajoule of natural gas, or the fact that their personal income tax rates were reduced when the carbon tax was introduced. In this regard they are like so many ill-informed, innumerate, loud-mouthed climate change deniers.
I keep records of all my expenditures, and I also have my old income tax forms, so I am able to do a comparison. In 2014, I spent $255 on the carbon tax buying gasoline and natural gas for heating my home. But my personal income taxes were $538 lower than what I would have paid if the old rates were still in effect, so I came out ahead $283. The difference would have been greater had I chosen to earn more.
To address the potential regressive aspect of this tax, low income earners receive a direct quarterly payment. Two of my children, early in their careers, are in this category. This year they each receive a little under half of what I paid into it last year. Their actual carbon footprint is a small fraction of mine because they do not operate vehicles and live in a warm part of the province.
James, good to learn of that Washington initiative. The constant price signal will cause people to change their fuel consumption.
As outright denial becomes politically untenable or simply ineffective we see more statements of acceptance by opponents of sufficient and appropriate climate/emissions/energy policy. Outright denial is outsourced rather than engage in directly by canny political players, who will not rebuke or repudiate it in allies and supporters whilst refraining from being explicit about their own position. Even insincere calls for extremely strong action of sorts and at scales that will be widely unacceptable can be thrown up to deliberately set the bar too high – with nothing less being acceptable and proponent of other means cast as obstructors.
Frankly I don’t think any tactic is beneath the staunchest and most self interested opponents of climate action – slander, misinformation and lies are so common that they barely raise eyebrows.
Personally I think eliminating only about a quarter of the data is much less of a cherry pick than the standard denier cherry pick for defining the pause which eliminates over 80% of the data. And if eliminating a small amount of the data eliminates the trend then the trend should not be considered significant. Eliminating a quarter of the data, and doing so in two different spots does stretch the boundaries of what I’d call ‘a small amount’.
But if Washington was the only place on the planet where temperature was measured I would combine the fact that the trend is statistically significant, the fact that it goes away when the data set is reduced, my opinion that statistical significance is not the final judge of true significance and say that the evidence of a significant long term warming trend is inconclusive. But Washington is not the only place on the planet where we measure temperature, and in my opinion by far the biggest cherry pick of this exercise is to ignore what the temperature of the rest of the planet is doing. The significance of the warming trend for the planet as a whole is way beyond anything that I would consider a reasonable doubt.
The global temperature series makes it very clear that the lack of a trend for Washington between 1925 and 2014 is due to a regional variation on top of the global trend. There is no reasonable expectation that this regional variation should continue indefinitely, and it is quite reasonable to suspect that the higher temperature in 2015 is the start of this variation ending. The argument that there is no evidence of a gradual long term trend in the Washington data is therefore worthless.
Michael Hauber, global temperatures do not causally effect fire risk in Washington state. Washinton State temperatures do. That is in itself sufficient justification for using Washington State temperatures rather than global temperatures, and it is no more cherry picking then using temperatures in the Nino 3 region as an ENSO index rather than global temperatures.
Excluding the early data was a cherry pick, though the reason given (unreliable data) is not a “lame excuse” as Tamino would have it. (If he thinks it is, why does he not make the same charge of cherry picking against NASA for excluding temperature data prior to 1880, or the Australian Bureau of Meteorology for excluding temperature data prior to 1910?) It is, however, insufficiently justified, so that the exclusion is a cherry pick.
“…global temperatures do not causally effect fire risk in Washington state.”
Really? Somehow I think that during “snowball Earth” episodes, fire risk went down. Way, way down.
If y’really want to get a handle on how the trend varies in the absence of datapoints, why not use the stationary bootstrap of Politis and Romano? What if the points which are contributors to a distortion are not in any contiguous subset?
TC: Excluding the early data was a cherry pick, though the reason given (unreliable data) is not a “lame excuse” as Tamino would have it.
BPL: Sure it is. There’s no reason on God’s green Earth to pick 1925 as a cutoff. NASA and NCDC use 1880, Hadley Centre uses 1850. They both have extensive arguments to justify it. The poster Tamino is commenting on simply states that the data is unreliable before 1925. No basis at all.
Why do you continue to defend this man’s dishonest thesis?
Do you really think there good, statewide temperature records for the state of Washington for the period 1850-1900?
Wikipedia says that the state’s population in 1880 was 75,000.
Bob
Bob, how many points do you need to get a t-statistic accurate to 95% confidence, assuming no bias?
Do you understand why I’m asking this?
[edit]
I will no longer waste my time on people for whom reason is so obviously only a sporadic guide.
[Response: Nor shall I, so goodbye.]
If you live in the Seattle area and listen to KPLU FM, a station who features Cliff Mass, do like we did and tell them to get rid of Cliff Mass or you will stop giving them financial support.