Arguments
Recent Comments
Prev 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 Next
Comments 25051 to 25100:
-
bozzza at 18:09 PM on 18 March 2016Lots of global warming since 1998
Let's not get rid of the term 'statistical signicance' because it will mean the deniers keep trying to undermine the science by simply saying uncertainty exists rather than providing their own predictions!Contrast draws the eye and the wolves in sheets clothing will remain easier to spot be keeping the term 'statistically significant' by my rationale... That's what you call a cunning strategy: "..shhhhhhhhh!" -
Glenn Tamblyn at 17:40 PM on 18 March 2016Volcanoes emit more CO2 than humans
Ybnvs
Just a reaction to several things you have said:
"My interest in this topic came from hearing environmentalists voice the absurd notion that mankind is responsible for climate change."
Why absurd? This sounds like an 'Argument from Incredulity'. Surely whether humans are responsible for climate change depends on two factors, both ultimately quantitaive.- How big does some influence have to be to impact Climate?
- How big an influence does humanity have?
Both these questions can be explored through measurement, observation and quantification. Just using the label 'absurd' is a cop out.
"Decades ago I noticed that the hockey stick graph actually showed temperature increasing before CO2 increased, but it was only recently that I heard someone else point that out"
Then you need to check your sources. The 'temperature increased before CO2' argument applies to the ice cores that cover time scales 10 to 80 times longer than the 'hockey stick'. The 'hockey stick' does not reference CO2 levels at all. And the ice core ecord is more complex than that. Different ice cores show different raltionships.
'and I'm suspicious of demagogues whose solution to a perceived problem is monetary. '
So what does this have to do with the science? If the science says that we need to reduce CO2 emissions, that is not as as such a monetary question. Implementing it might be but it is a logical fallacy to say that 'Problem A requires a solution that looks like B and since I don't like B, A is not real'.
"I understand that my thoughts will be criticised as simple (Occums razor) but my assertion that we can't establish cause and effect is rock solid."
Sorry, this is illogical. If your assertion is rock solid then you are claiming that you have established an alternative cause and effect relationship. Also there is a fundamental distinction between saying we 'haven't' established a cause and effect relationship and saying we 'can't'. Finally, if you think a cause and effect relationship hasn't yet been established, you need to back that up. -
TonyW at 17:35 PM on 18 March 2016Tracking the 2°C Limit - February 2016
Regarding Mann's 2C call, I believe it is referring to the northern hemisphere only. As a whole, Mann calculated NH Feb as 1.95C above pre-industrial though I've seen references to a period during Feb when NH was above 2C. Sorry, I don't have the links to hand. -
One Planet Only Forever at 15:23 PM on 18 March 2016Lots of global warming since 1998
barry,
There is a difference between 'being particular in an effort to help others better understand something' and 'being particular in an attempt to misrepresent something'.
I am sure you understand that. Misleading political marketing point makers understand how to make 'their points' to drum up popular support for something they understand does not deserve to be popular.
-
Ian Forrester at 15:08 PM on 18 March 2016Lots of global warming since 1998
The problem is that people who have no knowledge of statistics believe that "not statistically significant" means "no warming". That was shown with the interview with Phil Jones a few years ago. Just use a p value, knowledegable peope know what it means and others can ask. Let's get rid of the term "statistical significance" since it allows deniers to mislead.
-
barry1487 at 13:22 PM on 18 March 2016Lots of global warming since 1998
As for being a stickler, blame Tamino for my insistence. He's my stats 'guru.' And comments above kind of make my point. We criticise skeptics in fine detail based on statistical analysis, including stat sig. It surprises me when we start suggesting that this is pedanticism.
-
barry1487 at 13:20 PM on 18 March 2016Lots of global warming since 1998
Ian Forrester,
However, I do get upset when people start quibbling about “statistical significance" when I'm not sure if they know what it means. Firstly, it does not mean that there is no trend if it does not meet statistical significance criteria.
I wrote above:
"RSS - No stat sig warming since 1992: 24 yrs
UAH - No stat sig warming since 1995: 21 yrs
This doesn't 'prove' no warming, of course. That's not how the null hypothesis works. But it does mean you cannot say there has definitely been warming. Not according to those data sets. A broader lok at the climate system tells a different story."
As others have pointed out, starting from 1998 is a cherry-pick. But that was the cherry-pick raised in the OP, and the claim that followed re satellite records was statistically illegitimate.
-
barry1487 at 13:13 PM on 18 March 2016Lots of global warming since 1998
Tom,
If you think there is a contradiction between the claim that "the likelihood that 2010 was warmer than 2014 is greater than 0" and the claim that "the probability that 2014 was the warmest year on record (as of 2014) was 1...
Once again you are presenting a different argument to the one I made.
What you can legitimately say is that, given the error margins, the satellite records from 1998 to 2015 do not conclusively show that it warmed in fact from 1998-2015.
That is exactly what I said. Could you not just have agreed in the first place?
Why do we need to restrict ourselves to just one or two minimally accurate data records?
We don't. A point I also made. You're shadow-boxing. My criticism was very specific. We are, in fact, not in disagreement about any of the above points. Perhaps you think I'm attacking the overall message?
-
ubrew12 at 12:35 PM on 18 March 2016Tracking the 2°C Limit - February 2016
Hank@1. All models are incorrect (as a structural engr I'm sure you already know that). Hence, comparing predictions against reality is bound to lose, the predictions of climate scientists properly should be compared against other predictions, like those of climate deniers. Has your denier friend every offered you such a prediction by a prominant denier? If not, why not? Who told him he is automatically 'off the hook' for such information? This discusses James Hansens 1981 prediction and This discusses Wallace Broecker's 1974 prediction, and how each compared with the subsequent reality. But its not the comparison with reality that matters, but the comparison with the predictions of competing theories. If your friend is unable to field such predictions, let him know that this is actually saying something very significant about the theories he prefers.
-
Tom Curtis at 12:00 PM on 18 March 2016Volcanoes emit more CO2 than humans
Ybnvs @268:
1) It is not simple reasoning (except in a perjorative sense) to take William of Ockham's principle that we should not multiply entities beyond necessity and conclude that volcanic seeps and subsurface volcanism exist far in excess of, not just what has been observed, but what would be expected from surveys of the ocean floor. Rather, it invokes a principle as justification of doing the reverse of what the principle dictates by mulitiplying our estimate of the number of seeps and subsurface volcanoes beyond any necessity justified by the data. (Note it is Ockham in the English spelling, or Occam from the anglicized latin spelling - not Occum.)
2) That a NOVA documentary features a volcanic seep near New Guinea (of which several are known) in no way proves the seep to be newly discovered, or extensive enough to alter in any way estimates of subsurface CO2 emissions. And FYI, there are smaller seeps than those listed at the link above such as those in Milne bay, but again these are well known. It remains the case that you have yet to present any evidence for your claims.
3) While the uncertainty about volcanic emissions is sufficiently large that they may be up to double current estimates, we would need to be underestimating volcanic emissions by a factor of 50 for volcanic emissions to represent even 50% of anthropogenic emissions. That scale of error is simply not on the cards, and for you to be certain that the error in current estimates is even greater than that, as it would need to be for volcanic emissions to be the primary cause of the increased CO2 levels, without having become even superficially familiar with the relevant scientific papers shows that your certainty the the scientists who have dedicated their career to studying this issue (and hence who are well familliar with the facts, as you are not; and well familliar with the relevant arguments, as you are not) represents a breath taking arrogance. The style of reasoning you evidence even has a formal name - invincible ignorance.
4) As PhillipeChantreau alludes to, while there is significant uncertainty as to the actual value of volcanic emissions, regardless, other evidence makes as certain as it is possible to be in science that anthropogenic emissions are the cause of the rapid rise in CO2 levels in the twentieth century.
5) CO2 emissions and concentrations started rising around 1750, and rose rapidly after 1850:
In contrast, temperatures did not start rising significantly until 1910:
Again, whatever your argument with regard to temperatures, it is based on a very selective misinterpretation of the evidence.
-
Hank11198 at 10:12 AM on 18 March 2016Tracking the 2°C Limit - February 2016
Thank you very much Tom. I’m pretty sure I understand what you are talking about now.
-
PhilippeChantreau at 09:43 AM on 18 March 2016Volcanoes emit more CO2 than humans
The most recent I could find is a National Geographic story about a hydrothermal field 1300 ft long located 150 miles East of La Paz, Mexico. Not quite a game changer.
-
PhilippeChantreau at 09:27 AM on 18 March 2016Volcanoes emit more CO2 than humans
Ybnvs says "that's a long winded way of saying that we're still not sure of exactly what's occuring." I do not see how Tom's post can possibly be interpreted this way. In fact, it is exactly the opposite. The atmospheric carbon budget is a well understood part of the climate and there are numerous ways to address it. Newly discovered vents would have to spew out amounts of gas ata rate that would defy the laws of physics to make a significant difference.
The isotopic signatures of volcanic vs organic fossil carbon are well studied. The physics of radiative transfer are also well understood. Occam's razor does not cut the way you seem to think it does in this case. Everything that is well known and easily verifiable about Earth climate indicates that the climate should be changing because of anthropogenic CO2; it would be very strange if observations did not match that expectation. Fortunately, they do.
I tried to locate a recent NOVA documentary on deep sea vents off New Guinea. I located several short segments intended for teachers, the most recent dated from 2005, 11 years ago. Other ranged between 1977 and 1999. If you are to contribute anything here, you should link the specific new information that you think invalidates some really well established knowledge. Extraordinary claims require extraordinary evidence.
-
Ybnvs at 08:49 AM on 18 March 2016Volcanoes emit more CO2 than humans
Well Tom that's a long winded way of saying that we're still not sure of exactly what's occuring. Occum's razor has it's place in discussions such as this and we have to be careful not to get caught up in the lure of excessive and obscure data, after all... figures can lie and liars can figure. Sometimes it's not as difficult to find meaningful information as we tend to make it, and learning of the CO2 emitting vents near New Guinea can be as easy as watching a documentary on Nova. In regard to climate change one thing that has remained constant is change. The planet isn't as it was and won't be as it is. My interest in this topic came from hearing environmentalists voice the absurd notion that mankind is responsible for climate change. Decades ago I noticed that the hockey stick graph actually showed temperature increasing before CO2 increased, but it was only recently that I heard someone else point that out, I wonder why it took so long. Anyway... I don't know anyone who wants dirty air or water and I'm suspicious of demagogues whose solution to a perceived problem is monetary. I understand that my thoughts will be criticised as simple (Occums razor) but my assertion that we can't establish cause and effect is rock solid. Time is the best underwriter and in time I'm sure I will be vindicated. Peace.
-
Tom Curtis at 07:59 AM on 18 March 2016Tracking the 2°C Limit - February 2016
Hank, CMIP stands for Coupled Model Intercomparison Project. By 'Coupled Model' they mean climate models including both an atmosphere model and an ocean model run together, ie, what is generally referred to as a climate model (although there are other types of climate models). The project has run a series of phases in which models participating were required to meet progressively more realistic standards, and given the option of running a set of defined standard experiments, thereby facilitating comparisons between the models. You will often see references to CMIP3 which was the phase run explicitly to meet the needs of the IPCC Assessment Report 4 (AR4), while CMIP5 was the phase run to meet the needs of IPCC Assessment Report 5 (AR5). That is, in short, CMIP5 models are the models used for IPCC AR5.
I don't believe the standards for the next CMIP project have yet been defined, but am open for correction.
-
Hank11198 at 06:30 AM on 18 March 2016Tracking the 2°C Limit - February 2016
Thank you Tom and Charles. I have actually seen the one’s you posted Tom but I don’t know what CMIP5 Ensemble means. I understand that models are run considering different scenarios of CO2 projection levels but I don’t understand if the charts are showing the average of all these different scenarios or what.
I read the Climatology Versus Pseudoscience book by Dana Nuccitelli which was very helpful. Although I mainly use the Skepticalscience website for information I’m looking for a book or website that explains the graphs in more detail. -
PhilippeChantreau at 02:16 AM on 18 March 2016Lots of global warming since 1998
One should add to Ryland's comment that the myth of no warming since 1998 was born indeed of cherry picking the only date that skeptics could find to attempt making the point that there was no recent warming. Because this had not been true of late for most measures, they had to double up on the dishonesty by also cherry picking a given satellite record, plagued by more problems than they are even aware of. So BBHY is in fact correct, anyone who attempted to select 1998 of all years to try to make a point was trying to take you for a ride and should have been ridiculed. Dana's point in the OP is that even such a grotesque cherry pick can no longer be defended because it has become inaccurate on its face. Barry may have had a minor technical point, briefly, because the March anomaly could make the whole conversation moot.
In any case, cherry picking 1998 was always a crock of you know what, it has just become more so, so a point that's rather obvious. There never was a real pause in the either the surface temp or ocean heat content.
-
Tom Curtis at 23:49 PM on 17 March 2016Tracking the 2°C Limit - February 2016
CBDunkerson @4, I have previously caclulated that using the Mann 2008 EIV reconstruction and the 1736-1765 mean value as the 'preindustrial' benchmark', the gives a preindustrial temperature 0.12 C lower than using the GISTEMP 1880-1909 mean. Based on that 1880-1909 baseline, Feb 2016 had a 1.587 C anomaly. Therefore, based on Mann's currently most recently published best estimate of preindustrial temperatures, February 2016 was 1.71 C above the preindustrial average.
Michael Mann may have an as yet unpublished estimate that increases that value (or I may have missed a published estimate that supercedes Mann 2008). Alternatively, he may have used some other baseline terminating in 1749 or 1750. Without knowing his method of determining the value he claims, I cannot comment on whether I think it reasonable.
As a side, note, the 1.35 C anomaly is relative to the GISTEMP baseline of 1951-1980. As noted above, it is substantially greater relative to 1880-1909.
-
SirCharles at 23:46 PM on 17 March 2016Tracking the 2°C Limit - February 2016
We will soon need an ear for ultrasonic sound.
-
ryland at 23:30 PM on 17 March 2016Lots of global warming since 1998
BBHY@26 I think you may have missed the point of Dana's piece which is entitled "Lots of global warming since 1998". The thrust of the piece, to show there has been warming since 1998, is encapsulated in the first few sentences. These are "The myth of no warming since 1998 was based on the satellite record estimates of the temperature of the atmosphere. However, as discussed in the video below by Peter Sinclair, even that argument is no longer accurate. The satellites show warming since 1998 too".
-
Rob Honeycutt at 23:27 PM on 17 March 2016Tracking the 2°C Limit - February 2016
CDB... If I'm not mistaken, Mike is looking at data that goes back to ~1750, which also includes a lot of land use changes. With all due respect for Dr. Mann (who clearly is far better informed than I am), I tend to think there is likely some natural cooling influence baked into those figures. I've heard that Hansen uses 1880 as preindustrial, so even among experts there's disagreement.
For now I'm going to stick with the preindustrial baseline I have just because it's what I can calculate from the GISS data.
-
SirCharles at 23:20 PM on 17 March 2016Tracking the 2°C Limit - February 2016
@Hank
This latest graph is from Zeke Hausfather, Berkeley Earth:
-
CBDunkerson at 22:45 PM on 17 March 2016Tracking the 2°C Limit - February 2016
Michael Mann has been saying that February was more than 2 C above "pre-industrial". Does that seem plausible? Obviously, the 1880-1909 baseline used to compute the 1.35 C February anomaly in the article above is after the industrial revolution had begun, but does an additional 0.65 C of warming prior to that seem likely or is Mann jumping the gun?
-
BBHY at 20:59 PM on 17 March 2016Lots of global warming since 1998
So, folks are trying to argue that "warming since 1998" doesn't have "statistical signficance"?
News flash: If you have to cherry pick the starting date, then you have already lost the argument. Claiming that statiscial significance has anything to do with it is pure fantasy at that point.
Never assume that someone who uses 1998 as the starting date for their argument is attempting to have a serious discussion with you; they are not. Your best option is to tell them their argument is invalid from the start and refuse to engage in any further discussion. They simply want to waste your time. Do not let them.
-
Tom Curtis at 15:18 PM on 17 March 2016Lots of global warming since 1998
barry:
"Fair points, but not addressing the criticism. The unqualified cliam is;
<i>The satellites show warming since 1998 too.</i>
They do not. Not to statistical significance."
Barry, being a stickler about statistical significance sometimes amounts to deliberate obfustication. We saw that from the deniers when 2014 set a global record annual temperature and they started claiming it was 'statistically indistinguishable' from prior records, and therefore not in fact a record.
In fact, calling the record 'statistically indistinguishable' was a lie. That is because the set of years whose mean estimated global average temperature was within the error margin for 2014 was different to the equivalent set for 2010, or for 1998. Therefore, those years could be distinguished by statistical criteria, ie, were not 'statistically indistinguishable' from each other. More directly the likelihood that 2014 was the warmest year since records began was greater than the likelihood that 2010 was the warmest year since records began, and much greater than the likelihood that 1998 was the warmest year since records began, but the probability that 2014 was the warmest year on record (as of 2014) was 1, because the recorded annual mean for 2014 was greater than the recorded means for any other year (and the interval betweeen error margins was greater than the interval between error margins for any ohter year, and so on).
If you think there is a contradiction between the claim that "the likelihood that 2010 was warmer than 2014 is greater than 0" and the claim that "the probability that 2014 was the warmest year on record (as of 2014) was 1", you are failing to distinguish between the record and the reality. The former is a statement about the reality, while the later is a statement about the record.
Now, by the same token, it is a matter of record that UAH 5.6 shows warming from 1998 to 2015; and likewise a matter of record that the current RSS TMT record shows a warming trend from 1998-2015. That is, 'The satellite [records] show warming since 1998 too'. End of story. You do not refute that claim by expostulating, 'but error margins'.
What you can legitimately say is that, given the error margins, the satellite records from 1998 to 2015 do not conclusively show that it warmed in fact from 1998-2015. But then, so what? Why do we need to restrict ourselves to just one or two minimally accurate data records? When we talk about reality, we should consider all relevant records, and all relevant records taken together show with probablility >99% that it warmed from 1998 to 2015 in reality.
Raising the issure of error margins without looking at all relevant data amounts to obfustication.
This does not mean you should not raise error margins. But when you do you should realize you are greatly reducing the clarity of the discussion for (often) very minimal gain. And if you do raise the error margins, you had better make the distinction between the multiple individual records, and the reality they attempt to describe - and not try to restrict the discussion to just one or two records (and a carefully truncated interval).
On the other hand, when deniers misuse error margins to argue that, the error margin on the trend includes zero so therefore it has not warmed simpliciter (as they tacitly, and sometimes explicitly argue) it is perfectly valid to actually discuss what error margins actually mean without it being incumbent on you to introduce error margins into every post you make aimed at a grade 10 reading comprehension. The cases are not alike.
-
Tom Curtis at 14:51 PM on 17 March 2016Tracking the 2°C Limit - February 2016
Hank, you can try these two from Real Climate:
They are, of course, only current to the end of 2014. 2015 was about 0.12 C warmer than 2014 (GISTEMP), which still leaves it below the model mean, but well within the uncertainty.
2016 looks to be warmer again. In fact, its montly value for February was 0.61 C warmer than the annual value for 2014. That means it is literally of this chart. However, later months are likely to be cooler, bringing the mean down. Therefore it is likely 2016 will be above the model mean, but not necessarilly by much. Overall, I think the model mean overpredicts the temperature trend by about 15%.
-
One Planet Only Forever at 11:53 AM on 17 March 2016Lots of global warming since 1998
barry@22,
Your ability to identify and present political marketing points is duly noted.
-
Hank11198 at 11:13 AM on 17 March 2016Tracking the 2°C Limit - February 2016
I should add that we are both structural engineers who use finite element software extensively. As I understand it the climate models are basically very similar to finite element used in other industries. So If I can show the models, or a specific model with the amount of CO2 assumed then I can make a persuasive argument comparing them to how we verify our models.
-
Hank11198 at 11:07 AM on 17 March 2016Tracking the 2°C Limit - February 2016
I’m always debating with my friend who is a skeptic about global warming. Maybe I’m just slow but I cannot find a graph that shows a projected temperature by a model along with the actual temperature data of say NASA up to some time period where I can add the most recent temperature data to show the models are correct. Can anyone point me to that information? I’ve tried googling everything I can think of. In addition I know the base times have to the same and I’m not sure how to calculate that. Any help would be appreciated.
-
swampfoxh at 08:47 AM on 17 March 2016Sanders, Clinton, Rubio, and Kasich answer climate debate questions
As a Conservative Republican I am faced with the first time the Republican Party has nobody I can vote for. Bernie Sanders is the best candidate for thos who understand AGW , but Clinton will be the nominee. Both Parties will retain the support of their "big money" campaign donors who will insist on "business as usual" from Trump and Clinton.
I suggest a small farm, near fresh water, off the grid, away from the ocean and not in the southwestern U.S. where you can hunker down and be the last to meet the extinction Grim Reaper. Forty-Five years ago, we used to kid about "never getting out of this world alive", but it may be that nobody will get out of this century alive.
-
Ian Forrester at 05:26 AM on 17 March 2016Lots of global warming since 1998
I have to start off by saying that I am no statistician. However, I do get upset when people start quibbling about “statistical significance" when I'm not sure if they know what it means. Firstly, it does not mean that there is no trend if it does not meet statistical significance criteria. After all, 95% for statistical significance is just an arbitrary decision. I much prefer p values, since they represent chance. Thus if data have a p value of <0.05 we can claim statistical significance. However what is the real difference between P=0.04 and p=0.06. In real terms there is really no difference just that in the former case we can claim “statistical significance” and in the second case we can’t. Remember that p=0.05 is just an arbitrary choice; there is no real meaning to it, in terms of the data.
-
barry1487 at 01:47 AM on 17 March 2016Lots of global warming since 1998
OPOF,
Though you have tried to explain your preference for phrasing that implies no warming since 1998, you are relying on having identified a few extremes in all of the available information as the basis for the claim
I have no preference for such language. Only accuracy.
I didn't choose the data set/s or the time period. I only commented on how it was described in the OP.
What I am asking you to do is to honestly answer what your preferred trend value wold have predicted for the February 2016 temperature data set values relative to February 1998.
No analysis of such chaotic data could predict a single month's value, but one could give a probabilistic estimate. For the satellite data on such a short time frame, the estimate would be pretty wide. Looking at the whole system, warming has continued, and physics tells us it will continue. A spike like Feb 2016 was bound to happen sometime.
...go back month by month and identify how many previous months you have to be looking into to get a ststistically significant number of months where there has been 'no warming between the recent monthly value and the monthly value 18 years prior'.
It's not possible, as far as I'm aware, to determine 'no trend' to statistical significance - at least just by applying a linear regression; they're not made to detect 'no trend'. But one can make a reasonable assumption of no trend after a long time with a small trend. Eg, 100 years of temps may be 0.0002C/decade +/- 0.0004. The trend is negligible, as is the uncertainty, so you'd be on solid ground saying there was no trend.
Jan 1998 - Feb 2016 has a trend of 0.02C +/- 0.185 in the RSS data set. Claiming this is a warming, flat or cool trend is statistically illegitimate. With the uncertainty, the trend could be anywhere between -0.183 to 0.187C/decade (95% confidence level).
That's exactly why I commented in the first place.
Somewhat similar to your suggestion, I checked the longest period for a non-statistically significant tend (UAH/RSS) in a post above.
Add one more year to those results and you have statistically significant warming derived from the ARMA (1,1) model used the the SkS app. So, for the satellite data sets, you need to start a regression in the early '90s before you can say there is a legitimate warming trend. Not from 1998.
My criticism is of the 'analysis' (or lack) of the satellite data sets in this post (and there's more to it, like blurring the line between rcently revised mid-tropospheric trends and the lower tropospheric data), not of the bigger picture. I don't think it serves the message well to fudge these details.
Reckon I've said my piece long enough. Have a good one.
-
One Planet Only Forever at 00:28 AM on 17 March 2016Lots of global warming since 1998
barry,
Please comment on the predictive accuracy of the 'statistical' trend you have been preferring to promote (no significant warming since 1998) regarding the expected values of the temperature data in the past months.
What I am asking you to do is to honestly answer what your preferred trend value wold have predicted for the February 2016 temperature data set values relative to February 1998. And then go back month by month and identify how many previous months you have to be looking into to get a ststistically significant number of months where there has been 'no warming between the recent monthly value and the monthly value 18 years prior'. You can even go the other way and investigate how many years you need to reduce your range to to find a sstistically significantly number of prior year months that are warmer than the current monthly values.
And while you are at it, please comment on the likelihod that the 'no warming trend' will correctly predict that March 2016 will not be warmer than March 1998 (or any other March value).
-
One Planet Only Forever at 00:00 AM on 17 March 2016Lots of global warming since 1998
barry@12,
I understand that when you push the data values at the begining of a time period up to the maximum of the uncertainty band and push the data values at the end of the time period down to the bottom of the uncertainty band you may be able to claim that the result is a trend line that is not increasing.
However, understanding that those potential 'no increase' trend values are 'less likely among the total range of possible trend values' means it would be inappropriate for a person who actually understands things to declare that finding an extreme value of 'no warming' means that it is more correct to state that the data set does not indicate (with statistical significance or whatever other term is wished to be use) that 'warming has occurred during that period'.
Of course there will always be some people who will resist better understanding something when they sense that understanding it better will be contrary to their personal interest (that type of sense is clearly far too common in much of the USA and other places in the so called 'advanced or developed nations' like Alberta). And it is highly likely that such minds will look hard to find extreme examples that justify what they prefer to belive and desire to have others believe. But any success of that type of extremely irrational unreasonable type of thinking is clearly a political marketing problem, not a scientific presentation issue.
Though you have tried to explain your preference for phrasing that implies no warming since 1998, you are relying on having identified a few extremes in all of the available information as the basis for the claim, or said anoyther way, you are wanting to make a statement that you actually could understand is highly unlikely to be correct.
-
Tristan at 23:23 PM on 16 March 2016Lots of global warming since 1998
<i>The satellites show warming since 1998 too.</i>
They do not. Not to statistical significance.
Depends what you're controlling for and what published data you're using. The language should be more precise. It isn't however, merely wrong.
-
SirCharles at 22:37 PM on 16 March 2016Lots of global warming since 1998
Gary and Keihm 1991 showed that natural variability in only 10 years of UAH data was so large that the UAH temperature trend was statistically indistinguishable from that predicted by climate models.
http://science.sciencemag.org/content/251/4991/316.shortHurrell and Trenberth 1997 found that UAH merged different satellite records incorrectly, which resulted in a spurious cooling trend.
http://www.cgd.ucar.edu/cas/abstracts/files/Hurrell1997_1.htmlWentz and Schabel 1998 found that UAH didn’t account for orbital decay of the satellites, which resulted in a spurious cooling trend.
http://www.nature.com/nature/journal/v394/n6694/abs/394661a0.htmlFu et al. 2004 found that stratospheric cooling (which is also a result of greenhouse gas forcing) had contaminated the UAH analysis, which resulted in a spurious cooling trend.
http://www.atmos.washington.edu/~qfu/Publications/nature.fu.2004a.pdfMears and Wentz 2005 found that UAH didn’t account for drifts in the time of measurement each day, which resulted in a spurious cooling trend.
http://science.sciencemag.org/content/309/5740/1548.abstract -
SirCharles at 22:29 PM on 16 March 2016Lots of global warming since 1998
An eye-opening article by Dan Satterfield at the American Geophysical Union blog:
=> NASA- February Temperatures Hottest Ever
It's not only the 5th month in a row that is topping temperature records by a big margin. Dan is also cleaning up some myths.
-
barry1487 at 22:14 PM on 16 March 2016Lots of global warming since 1998
Fair points, but not addressing the criticism. The unqualified cliam is;
<i>The satellites show warming since 1998 too.</i>
They do not. Not to statistical significance.
We wouldn't give this a free pass if a skeptic used the same period to say no warming "because it's not statistically significant," or if they claimed a 'pause.' We frequently remind them of this ommission in their 'analysis'. I'm doing the same here. Neither of those claims are justfied. Neither is a claim of warming for that period for the satellite data.
It's about integrity, whether in 'messaging' or in accurate science.
-
Cedders at 22:11 PM on 16 March 2016Oceans are cooling
Thanks Tom & Rob for replies. The map is indeed impressive and worrying for coral bleaching, mangrove swamp etc (what's the source & baseline period, please?).
I would still be interested in any simplified model, calculus or references that describe the coupling between the surface and the ocean layers: is the deep ocean warming rate expected to accelerate with further SST warming? How long will the 700-2000m layers take to reach equilibrium? If 88% of the heat imbalance is going into oceans, doesn't that mean the oceans are providing a huge air-con service, but one that will eventually disappear? It has policy implications, eg if Earth system sensitivity is 6 °C, but thermal inertia meant 'only' 60% of that surface warming even a century after reaching 550ppm, we might want to think seriously what human capabilities are going to be in the 22nd century.
I still would like to suggest if I may that this contrarian meme may have become sufficiently prevalent for SkS to examine thoroughly to produce a "basic" clarification for lay people like me. (I think UKIP in UK may have picked it up via WUWT.) Thanks again.
-
Tristan at 21:17 PM on 16 March 2016Lots of global warming since 1998
We can say 'there has, unambiguously been warming'. We just can't say it if we know nothing more than certain temp records. That would be a pointless condition to hold ourselves to.
-
Kevin C at 20:52 PM on 16 March 2016Lots of global warming since 1998
A few details:
Neither of the Karl 2015 trends are to the present - they both omit 2015. Include 2015 and both would be statistically significant.
The period from 1998 has presented several problems:
- The ship/buoy transition which has created a bias. The correction has not been totally nailed down yet. There is more work on this in the pipeline.
- Rapid arctic warming, shown by IR satellites, weather models and buoys. This again leads to a bias particularly in the NOAA, UKMO and JMA records. NOAA and UKMO are both working on global records.
- The changing ice-edge anomaly bias. This is a new result from our recent model comparison paper and is not fully understood, although I know NOAA are looking at it.
I expect the picture to become a lot clearer over the next 12 months at least on the first 2 points, with new versions from NOAA, UKMO and JMA. We've identified the issues, they are being addressed as we speak, and we have a good idea what the changes are going to look like. In the mean time I would be cautious of drawing conclusions from records with known biases.
-
barry1487 at 19:36 PM on 16 March 2016Lots of global warming since 1998
Tom,
"RSS is not explicitly out of date..."
V4 TLT is supposedly coming out in about 6 months. I strongly doubt it will change my point for trends run to Dec 2015. I also doubt we'll have statistical significance in v4 TLT 1998- 2016 (inclusive) trend. The relatively large variability will still be there, which is why the satellite data sets need longer time periods to reach this threshod than the surface records, as you know. -
barry1487 at 19:30 PM on 16 March 2016Lots of global warming since 1998
OPOF,
You can also run the trends/uncertainty there to check for 1997/99 (run it for full years to avoid intereference from annual cycle). Last time I looked, the longest time-period lacking statistical significance was RSS - 23 years. Not that that means too much. Failing to disprove the null doesn't mean that there is no trend, just that you can't demonstrate it statistically.
I'll check just now for the logest period without statistical significane in the sat records - just for curiositiy's sake...
RSS - No stat sig warming since 1992: 24 yrs
UAH - No stat sig warming since 1995: 21 yrs
This doesn't 'prove' no warming, of course. That's not how the null hypothesis works. But it does mean you cannot say there has definitely been warming. Not according to those data sets. A broader lok at the climate system tells a different story.
-
barry1487 at 19:29 PM on 16 March 2016Lots of global warming since 1998
@ nigelj,
you can use the SkS trend calculator to check the trends/uncertainty yourself if you'd like.
www.skepticalscience.com/trend.php
Simply, if the uncertainty (+/-) is larger than the trend, then the trend fails statistical significance (is not statstically distingishable from 0).
-
barry1487 at 19:27 PM on 16 March 2016Lots of global warming since 1998
Hi Tom,
I didn't metion models or IPCC predictions.
When the skeptics talk about short-term trends, their critics - rightly - scold them for ignoring statistical uncertainty. When Tamino criticises papers on temp trends that lack uncertainty estimates he is right to do so (as he did recently with a paper Mann co-authored, while commending other parts of the paper).
The statement in the OP was clear. "The satellites show warming, too." Not to statistical significance. It's inconsistent to correct 'skeptics' for this oversight and then do the same thing.
"...unless the trend shows a statistically significant difference from 0.165 C/decade (for GISTEMP), the correct assumption is that the trend is continuing..."
As I said to begin with, it would have been better to lay the case out this way."You'd be on firmer ground showing that the temps since 1998 have not deviated from prior warming to statistical significance."
"I do not know if the UAH data used in the trend calculator is vs 5.4"
More likely v5.6. Strongly doubt it's beta6.
-
Tom Curtis at 15:44 PM on 16 March 2016Lots of global warming since 1998
barry @5 & 6, the trends are:
HadCRUT4: 0.107 +/- 0.111 C/decade
NOAA (Karl 2015): 0.107 +/- 0.123 C/decade
GISTEMP: 0.154 +/- 0.113 C/decade
BEST: 0.125 +/- 0.104 C/decade
HadCRUT4 krig: 0.142 +/- 0.117 C/decade
Karl 2015 Global: 0.134 +/- 0.141 C/decade
RSS*: 0.002 +/- 0.185 C/decade
UAH: 0.106 +/- 0.186 C/decade
RSS is not explicitly out of date, with a recent improvement of the diurnal drift adjustment which significantly increaste the 1998-2016 trend having been applied to the TMT channel, but not yet carried through to the TLT channel. I do not know if the UAH data used in the trend calculator is vs 5.4 or the as yet un peer reviewed vs 6.
The key thing to note is that the only dataset in which the approx 0.2 C per decade model predicted lies outside the uncertainty interval is RSS, which is explicitly acknowledged by its author to be less accurate than surface thermometer records, and in its current version, to be biased low by inaccurate correction for diurnal drift. That means there is better statistical grounds to say that recent temperature trends are following the models than there is to say that they have paused, or are in a hiatus. Insisting that not showing a statistically significant difference from zero is important, but not showing a statistically significant difference from model trends is irrelevant amounts to special pleading. I amounts to forcing your preferred story on the data because you like the narrative.
Deniers try to avoid the charge of special pleading by insisting that a trend of zero is the null hypothesis. The correct response to that claim is neither polite, nor in keeping with the comments policy.
Being more polite, the null hypothesis depends on the theory you are testing. If you are testing the theory that the temperatures are following model trends, the null hypothesis (ie, that which you are trying to falsify) is that the trend is that of the models, and only trends showing a statistically significant difference from the model trends are relevant in that only they can contribute towards falsifying the model.
More naively, if we do not have a specific model in mind, the null hypothesis is that nothing interesting will happen, ie, that the trend will continue as it was doing before. More precisely, the trend will remain the same as it has been since the last stastically determined 'breakpoint' in the trend (which happens to be about 1970). In other words, unless the trend shows a statistically significant difference from 0.165 C/decade (for GISTEMP), the correct assumption is that the trend is continuing at 0.165 C per decade. As it happens, the GISTEMP trend differs by about a tenth of the error from the ongoing trend.
-
One Planet Only Forever at 14:33 PM on 16 March 2016Lots of global warming since 1998
barry,
What do 'your' statistics say about the data since 1997 or 1999?
-
bozzza at 14:06 PM on 16 March 2016Sanders, Clinton, Rubio, and Kasich answer climate debate questions
OPOF,
The market place is made of profiteers: that includes the Governments that regulate it.
Ultimately anarchy is not what Governments want but until the tax base stops growing their is little impetus for changing the rules. I suppoe a point of diminishing returns comes into play somewhere however...
-
nigelj at 13:06 PM on 16 March 2016Lots of global warming since 1998
Barry @ 6. I will take your word that the satellite record does not have statistically significant warming since 1998. I suppose it gives you climate change sceptics some straw to hold onto.
I think my mind is pretty made up. There is now a mountain of evidence that we are altering the climate, and at a significant rate.
The surface is warming strongly according to nasa giss. The satellites measure the atmosphere, and this may be heating more slowly, but its the surface that dominates weather patterns and rates of ice melt.
-
barry1487 at 10:42 AM on 16 March 2016Lots of global warming since 1998
Should have checked before I spoke. NOAA and GISS have statstically significant warming since 1998 (to year-end 2015). But not HadCRU4, and definitely not RSS and UAH satellite records, no matter which version you use.
Prev 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 Next
