Arguments
Recent Comments
Prev 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 Next
Comments 24951 to 25000:
-
h4x354x0r at 14:34 PM on 20 March 2016CO2 measurements are suspect
Regarding the issue of accuracy or agreement of various sampling methods or ice cores: Dendrochronology (tree rings) provide another surprisingly rich, near continuous data source that includes decipherable clues about CO2 levels. Data from dendrochronology agrees, very strongly, with data from the ice cores.
-
h4x354x0r at 14:19 PM on 20 March 2016CO2 measurements are suspect
The Keeling Curve seems almost uniquely consistent to me, compared to almost every other type of measurement in the natural world, which tends to be highly variable, regardless of long-term trends. The Keeling Curve, by comparison, is almost eerily rigid in trajectory, especially compared to variability of total carbon releases over time. It's like it's got some kind of governor that simply keeps the increase and accelleration from going any faster, no matter how much more carbon we pump into the atmosphere.
Can anyone help me understand how the Keeling Curve is so "tight", while so many other measurements have so much higher variability? Is it just because the other sinks - ocean, land, and plants - have so much more capacity that they can absorb whatever else the atmosphere can't? Even if that's a reasonable simple explanation, I still feel like I'm missing something from the 'limiter' aspect of the phenomenon.
I mean don't get me wrong, I'm figuratively poop-my-pants scared about climate change, particularly that "unprecidented rate" part. I've just recently been struck by the relatively tight natural regulation of the Keeling Curve, compared to a much greater variability I observe in almost every other aspect of nature.
Appreciate anything that might help a mostly layman level understanding of things dude wrap his head around that. Thanks! -
Tom Curtis at 12:57 PM on 20 March 2016Global Warming Basics: What Has Changed?
Just for Billev's (@1 and @3) benefit, here are the NASA GISTEMP annual values detrended to make the trend from 1880-1945 zero, and with the 1974-2002 trend line of the adjusted data shown:
First, you can see quite plainly that while 7 out of 13 years after 2002 do indeed fall below the 1974-2002 trendline, that is not because of any evident reduction in the trend. There is no more basis for assuming a reduction in the trend then there would have been in the 13 years after 1984. Indeed, with an odd number of years in the interval, of necessity the number of years below the trend cannot have equalled those above the trend. If we included 2002 so that we had an even number, we would have equal number under and over the trend. Further, with the monthly value for January and February of 2016, it is almost certain that the 2016 temperature will be at least near that of 2015, and hence over the trend, again leading to equal numbers over and below.
Further, the 1974-2002 trend is exagerated by the near record breaking La Nina of 1974/5, and the near record breaking El Nino of 1997/98. In contrast the period after 2002 has (until now) not experienced a strong El Nino and has experienced near record breaking La Nina's in 2008 and 2012/13. The slight deviation from the 1974-2002 trend in nearly fully accounted for by short term ENSO fluctuation. It is fully accounted for when changes in forcing are included. In contrast, Billev's nascent theory would predict the data, having been adjusted for the trend over a full cycle, should show a strongly negative trend after 2002.
Second, it should be noted that the data, having been adjusted for the trend over a full cycle, does not show significant similarity between phases of the purported cycle. Neither the slope nor temperature pattern from 1880-1910 matches that from 1945-1974. Nor do slope or pattern match between 1910 to 1945 and 1974-2002. Nor to the periods of the half cycles match , with respective half cycle lengths of 30, 35, 29, and 28 years. If you look at longer temperature records you also find the purported cycle disappears prior to 1880, or at best halves its period.
In all, Billev's ad hoc theory has nothing to support it. It is based on an interpretation of only 1 and a half cycles, which is a very statistically tenuous projection; is falsified in its first actually predictive interval; does not have a clear cycle and has no physical basis. Even the half periods are ad hoc, as can be seen by the fact that the 1974-2002 trend fits the data quite will for several years before 1974, which is (as previously mentioned) a near record La Nina year.
-
Digby Scorgie at 11:08 AM on 20 March 2016Global Warming Basics: What Has Changed?
Trends don't have to be linear. When I look at the graphs for shortish periods, the trend seems linear. When I look at the graphs for really long periods, the trend looks to me to be following a curve with gradually increasing slope.
-
billev at 10:22 AM on 20 March 2016Global Warming Basics: What Has Changed?
In the second chart labled NASA a line is drawn from the 1970's to the present. If the line were drawn from the temperature at 1880 to the present the line would show a warming trend through the period it covered. It would also show that there was not constant warming throughout that same perod. If you alter the line shown in the second NASA graph to pass through the temperature for 2002 (what I believe was the end of the warming period that started around 1974) you will observe that most of the temperatures after 2002 fall below the altered line.
-
Leto at 07:37 AM on 20 March 2016Lots of global warming since 1998
Bob@45,
That's great. I hadn't seen them, though I have seen people chases mirages based on variations of the third and fourth graphs - I once was silly enough to start a PhD with some scientists (bad ones, obviously) who had not looked at their raw data in a scatterplot. In the end, they were basing a major project on something like the 4th graph, and the whole thing was spurious.
-
Bob Loblaw at 06:41 AM on 20 March 2016Lots of global warming since 1998
Whenever discussion of linear regression and related statistics comes up, I always like to refer to Anscombe's Quartet. Although the four datasets result in very similar statitical results (of the ones listed on the wiki page - follow the link), it is abundantly obvious from the graphs that the four data sets are not at all similar.
I won't embed the graphs (spoiler), so look at the wiki page.
-
Hank11198 at 23:45 PM on 19 March 2016Tracking the 2°C Limit - February 2016
Ubrew12 I completely agree. Are you a structural engineer? I started engineering long before computers when everything was done with hand calculations using basic assumptions. I have trained several engineers and with the latter ones coming right out of college I’ve become very concerned that they so completely depend on the computer. I’ve had them bring me results that I could see in 2 seconds was not even close. It’s like the proverbial pulling eye teeth to get them to do a rough hand calculation.
Unfortunately some of my models are complicated plate and solid models which can be difficult to do classical hand calculations but I try to at least do exactly as you have said, run different meshes to see if they converge on a number.
-
barry1487 at 22:57 PM on 19 March 2016Lots of global warming since 1998
Kevin C,
Nice tester!
I've read the answers - I would have picked Green as having the most uncertainty, though it actually has the least.
Couple of queries/clatifications:
Blue: 0.2 +/- 0.0
Green: 0.2 +/- 0.07
Red: 0.2 +/- 0.15
Cyan: 0.2 +/- 0.12...We can see the trend in the red line very clearly, and yet it isn't significant.
1. Red is (statistically) significant isn't it? (I'm red green colour blind, so I assume the order of the values matches the order of the charts).
2. Anomalising the seasonal data would decrease the uncertainty, right?
BTW, 2nd chart looks like satellite temp anomalies, 3rd looks like Antarctic, 4th looks like global sea ice mean (+trend), though they are artificial.
...testing the statistical significance of a trend may be helpful, but it can't settle an argument.
Technically yes, though I think it can if the cofidence interval is mutually agreed upon as a robust degree of uncertainty. Ie, a centennial temp trend of 2.3C +/-0.04, 95% confidence interval, would settle that there is a warming trend for all but the most reclacitrant (eg Lubos Motl).
-
Paul D at 21:08 PM on 19 March 2016Global Warming Basics: What Has Changed?
Sigh...
Climate isn't the stock market and economics (both only constructs in the human mind), climate science is real billev. -
barry1487 at 19:18 PM on 19 March 2016Lots of global warming since 1998
Tom,
Had Dana said something more nuanced regarding 1998 sat trends I almost certainly would not have commented. As it is the statement is unambiguous, whatever the intention.
Why, in fact, in a popular discourse, are you discussing a purely academic question.
SkS straddles the line between the academic and the popular. Maybe I set too much store in what the words in the blog title actually mean (no sarc intended).
Stephen Schneider's famous quote is apropos:
"Each of us has to decide what the right balance is between being effective and being honest. I hope that means being both."
I put in my 2 cents with that idea in mind. The ongoing has been a bit more expensive!
[I read up on Carter's article about 4 years ago (written in 2006, not 2005, I believe). There's plenty of dross just like it around the skeptiverse]
-
SirCharles at 17:26 PM on 19 March 2016Tracking the 2°C Limit - February 2016
Re: Climate models
-
Kevin C at 16:59 PM on 19 March 2016Lots of global warming since 1998
The fourth curve is actually chaotic - it's a cubic spline fit through a logistic equation.
There's an error in my analysis though, because I only checked for short range autocorrelation. If I'd corrected for longer autocorrelations (specifically 12 months), then the autocorrelation correction would have corrected for the annual cycle, and given a somewhat lower trend uncertainty for the red curve in line with out expectations.
I should have used a pseudo-periodic function where the period varies over time. The trend would still be plainly apparent to the human eye, but would still give a large uncertainty even with long range autocorrelation correction.
I've got another example where the trend uncertainty tells us the wrong thing which doesn't depend on uncertainty - I'll try and work up some pictures later.
-
ubrew12 at 12:45 PM on 19 March 2016Tracking the 2°C Limit - February 2016
Hank@20 said "from a structural engineering perspective... [the models] better at least be close or bad things happen" I typically make 3 models of anything I'm tasked with understanding: a full 20k-2000k finite-element or difference model using appropriate software, a 100 node model performed with a spreadsheet, and a 1-5 node model calculated 'back-of-envelope' by hand. Always tie your calculations back to something you did by hand. It confirms the most critical assumptions in your calculation and prevents embarrassment. A 2000k-node model can be right in its particulars, and yet spectacularly wrong in its generals.
"anything can be fit with a Fourier series" John von Neumann: "with four parameters I can fit an elephant, and with five I can make him wiggle his trunk." Earth does not exist outside Physics. If your friend cannot tie his math modelling back to the Physical fundamentals, thats a 'red flag'.
-
billev at 09:51 AM on 19 March 2016Global Warming Basics: What Has Changed?
The trend may be up but global temperature has not risen steadily since 1880. There were pauses in temperature rise from about 1880 until about 1910 and from about 1945 until about 1974 and now since about 2002. If this pattern continues then there will only be a steady rise in warming from about 2030 until 2060 and then again from 2090 until 2099 in the current century. The warming trend will continue, however, until about 2350 if the climatologists are correct in saying the Earth began to experience a 500 year warming period around 1850.
-
Tom Curtis at 07:18 AM on 19 March 2016Lots of global warming since 1998
Kevin, am I correct in assuming the fourth curve involves the superpostion of 3 or 4 sine curves of different amplitudes plus linear trend? Also, is the noise model for the second curve white noise or red noise?
I am surprised at the ordering of error margins (obviously), but not surprised that I guessed wrong given that I was trying to eyeball the 95th percentile range of variation from the trend. Although 2 and 4 have the greater peak amplitude, over much of their range the amplitude of the residual is much less than in 3.
-
Kevin C at 07:04 AM on 19 March 2016Lots of global warming since 1998
Tom: I adjusted for autocorrelation in every case, but I thought that was too much detail!
-
Tom Curtis at 07:03 AM on 19 March 2016Lots of global warming since 1998
Kevin C @36, I would guess that all have the same trend by construction, although the second (green) looks to have had its trend reduced slightly by random variability. Clearly the first is statistically significant, while the following three are not, which look to have approximately the same standard deviation by construction, with the linear trend being approximately 0.75% of the error margins. Not sure if they would have the same error margin useing an autocorrelative model though. If not, order as per rkrolph.
-
Kevin C at 06:59 AM on 19 March 2016Lots of global warming since 1998
Well done for giving it a go! As you may have guessed, the examples were set up to expose a misconception, and so the answers are rather counterintuitive. Here are the actual trends and uncertainties:
Blue: 0.2 +/- 0.0
Green: 0.2 +/- 0.07
Red: 0.2 +/- 0.15
Cyan: 0.2 +/- 0.12So something interesting is going on here. All the trends are the same. You correctly identified the straight line as being most significant. But you identified the trend+seasonal cycle (red) as next most significant when it was least significant. And the trend+noise as least significant, when it was second.
We can see the trend in the red line very clearly, and yet it isn't significant. It's very hard to spot the trend in the green line, and yet it is. What's going on?
It's easy, on the basis of common discussions on the web, to imagine that linear regression is some magical tool which can spit out objective answers concerning the existance or non-existance of a trend. But it isn't.
Linear regression is a model which we can use, and which may help our understanding. However if it is used blindly, then it can also lead to completely invalid conclusions. Most importantly are the answers it can give, which can be characterized roughly as:
- Yes, there appears to be a trend, assuming that the model is correct. Or...
- No trend could be detected, either because there is no trend, or because the data are inadaquate, or because the model is inaqequate.
In the latter case we don't know whether there is no trend, or whether there is a problem in the model or the data. Even in the former case, we may detect a non-existant trend if our model is wrong. So the results are in no way objective, and are contingent on a number of other factors.
In other words, testing the statistical significance of a trend may be helpful, but it can't settle an argument.
So let's go back to the examples and try and understand them some more. Linear regression involves an implicit model: that the data consist of a constant, a linear trend, and noise. The green data fit that model exactly. And the noise is not too large, so the calculation correctly identifies the presence of a trend
In the other two cases, the model isn't really representative of the data. In the case of the seasonal cycle, the deviations from linearity aren't noise-like. If we used a better model which included a seasonal cycle, the trend would have come out clearly. Linear regression tells us it can't detect the linear trend because the data aren't linear, even when the trend is totally clear to us.
Of course the application to real data arises because we know the real data also contains contributions which are not remotely linear - in particular very strong El Ninos at either end of the period (and indeed significant La Ninas just in from them). Since this non-linearity is not accounted for in the linear model, it inflates the uncertainty such that the underlying trend is not detected. (That's not a stupid - if we didn't happen to have El Ninos at both ends of the study period, they would also distort the trend as well as inflating the uncertainty.)
-
Hank11198 at 06:44 AM on 19 March 2016Tracking the 2°C Limit - February 2016
Ubrew12:
Thanks for the links. There were helpful and interesting.
No my denier friend has never given any projections by skeptics. I’ve even pointed out that they have been predicting a downturn in temperature for 20 years and it hasn’t happened yet. Those emails are ignored. Although he has sent me some curve fitting that was used to show ‘natural change’ (never any definition of what ‘natural’ is) and then subtracted from the actual to show that climate sensitivity is very low. I’m sure you have seen those. A few natural cycle sine curves with the amplitude and frequency adjusted and anything can be fit with a Fourier series.My friend is very intelligent. I’ve even asked him how he picks between which science fields he accepts and which he rejects since he rejects very few but just get a convoluted answer about it has to be based on the evidence of all things! He accepts the temperature is rising and CO2 has some effect but not enough that even if we did something it would matter much or that it’s going to get that bad. I’m pretty sure the problem is his conservative views about the government getting involved in this with taxes. You know the story. I don’t think I will ever convince him in this lifetime but I actually enjoy learning something new and our debates require me to study the subject in order to give factual and rational responses. So that’s the benefit I get from our debates. I’m a long way from knowing a lot about the subject but I can follow the math and science part enough to know I need to be worried about my children and grandchildren.
I will respectively disagree with you on comparing projections with models. I do know that models are incorrect as I use finite element almost every day in my work. However at least from a structural engineering perspective they better at least be close or bad things happen. On occasion I have the opportunity to verify my designs by observations from the field. It happened just a few weeks ago when a structure had to be loaded for testing and we took deflection readings while it was loaded. The deflections were within 10% of what my model predicted so I was ecstatic. It verified my model was created correctly and was also a step towards validating the entire method of the software finite element program I was using. I think projections reasonable close to matching the models is a powerful argument that the science is correct and we need to do something about global warming.
-
Rob Honeycutt at 06:30 AM on 19 March 2016Tracking the 2°C Limit - February 2016
Here's an interesting comparison I just found...
-
Rob Honeycutt at 06:23 AM on 19 March 2016Tracking the 2°C Limit - February 2016
There was a lot of heat coming out of the ocean in the mid latitudes as well...
Whereas I believe the 1998 El Nino was primarily limited to the equitorial Pacific.
-
DennisMyers at 06:12 AM on 19 March 2016How Exxon Overstates the Uncertainty in Climate Science
Maybe these chart makers need to standardize their assumptions and only make charts based on current laws, or current pledges. Showing the difference between what has been pledged, and what has been implemented only.
-
rkrolph at 03:53 AM on 19 March 2016Lots of global warming since 1998
Ok Kevin, I will give my visual assessment. I would say that all datasets show a trend. I would guess that if you put these datasets into an excel spreadsheet and applied linear trend lines to each, the trendlines would all match closely to the slope of the top blue line. In order of most to least statistical significance I would say blue, red, cyan, and green.
-
wili at 03:42 AM on 19 March 2016Tracking the 2°C Limit - February 2016
I keep reading reports about how climate scientists are responding to this February peak with words like ‘shocking,’ ‘stunning,’ and just ‘wow.’
I agree, and I know they expected _some_ increases in an El Nino year. But now I want to know what the science didn’t anticipate that has made this monthly peak so high.
Is it just the super strong El Nino on top of GW? Or aret there carbon, or albedo or other feedbacks kicking in? Are we losing our 'aerosol umbrella'? Perhaps calculations of climate sensitivity have been too low? Do we have any clue at all why the reading for February is so much higher than anticipated?
-
wili at 03:30 AM on 19 March 2016Worst Mediterranean drought in 900 years has human fingerprints all over it
Would it be correct to see this as part of the predicted shift of the Sahara and Arabian Deserts north, just as the tropical zones are shifting poleward?
-
Kevin C at 01:52 AM on 19 March 2016Lots of global warming since 1998
I think there is possibly a more fundamental issue here than the semantics, which relates to the meaning of 'statistical significance'. Statistical significance is not an indicator of whether there is a trend or not, it is an indication of whether a trend can be detected by a particular approach under a given set of assumptions.
So what is being identified as a 'lack of statistically significant trend' in some datasets does not mean that there is no trend. It means that under a particular set of assumptions using a particular method of diagnosing a trend, the diagnosed trend could not be confidently distinguished from zero. However, the assumptions affect the answer. These include features such as the nature of the noise, the noise model, the variable used, the presence of a residual annual cycle and other factors. And the assumptions in the method have no bearing on whether or not there is a trend in the data, only on whether we diagnose one.
To illustrate the problem, would anyone care to speculate on which of the following datasets contain a trend? Is the trend significant in each case?
-
Rob Honeycutt at 01:08 AM on 19 March 2016Tracking the 2°C Limit - February 2016
Tom Curtis... Mann also tweeted several comments related to this anomaly being 2°C over preindustrial.
-
ranyl at 21:29 PM on 18 March 2016Tracking the 2°C Limit - February 2016
Basically Feb has peaked at ~2C above pre-industrial, a level considerred so dangerous to geopoltical stability that even the politicians have made honest pledges that they don't want to go above it.
What carbon budget??????
Like we can seriously put more GHG into the atmosphere safely, what a load of nonsesne.
2015 was also the highest yearly CO2 rise at ~3ppm.
Wars are raging and Wars cost lots CO2 (manufacturing all those nice weapons) and cause mass environmental damage.
And the highly elite driven system humankind is under isn't intending to change anything any time soon and when pressure falls on the basic necessities and the weather ravages homes and crops, more wars and migrants are inevitable.
The pope has already called the plight of the Syrian refugees who are average kind people escaping tyranny and terror an invasion of a holy nature into Europe.
Antartica is melting quickly, Greenland too, the Arctic sea ice is vanishing rapidly and the permafrost is melting, forest fires raging and peaks lands burning all relaeasing more GHG than expected in all the policy makers models.
???
Isn't time to transform everything we do?
Impossible???
Or does anyone know a place of sanctuary for those who care to escape to?
-
Tom Curtis at 19:29 PM on 18 March 2016Lots of global warming since 1998
barry @29:
"Once again you are presenting a different argument to the one I made."
No! I am presenting my own argument in relationship to the claim that you made. If you were not so busy trying to call strawman or agreement, you might have learnt something - possibly realized that what you already knew but weren't thinking about made your original claims misleading and obfusticatory.
Very briefly, once more, had Dana said:
"The line which has the least sum of the squared differences for the satellite data has a positive slope"
and you objected:
"It does not. Not to statistical significance."
Your claim would be simply false, representing a category error. You would have interpreted a claim about the data as a claim about reality.
Conversely, had Dana said:
"That the line which has the least sum of the squared differences for the satellite data has a positive slope shows temperatures in fact rose over the interval"
and you had objected:
"Not it does not, for the error margin on the slope of the line includes zero"
you would have been correct.
Dana, of course, said nothing so precise as either of those statements. He wrote:
"The satellites show warming since 1998[.]"
The question is then, which of the more formal statements above most closely conveys Dana's meaning. It think it is the first, and clearly the first. That is (firstly), because it was made to a popular audience, and popular audiences do not trouble themselves with error margins. Therefore, it is invalid to interpret the claim in a way that requires understanding of error margins to be understood unless dictated by necessity. It is also because Dana would, if challenged as to whether or not the satellite record by itself shows the warming to be real (ie, statistically significant), be genuinely puzzled as to why you would ask that question to the exclusion of the majority of the data, and to the exclusion of physics. Why, in fact, in a popular discourse, are you discussing a purely academic question.
Perhaps you are confused on this point because the 'no warming since 1998' has gone through several incarnations. It started in 2005 with Bob Carter pointing out that the annual temperature of no year to 2005 was higher than 2005 (totally ignoring trends). Overtime, it started bringing in discussion of statistical significance, but over the last year the argument has reverted to Carter's fraudulent basis. In particular, the deniers have countered the fact that 2014 and 2015 were record setting years in the surface record by pointing out that they were not in the satellite record. And, that simplistic argument no longer works. Indeed, given that Spencer and Christy know about the lag in satellite temperature response to ENSO as well as anybody else, they knew it was only a matter of time before that argument would stop working when they made it. That is, they knew the argument to be deceptive as they spoke.
-
Tom Curtis at 18:53 PM on 18 March 2016Tracking the 2°C Limit - February 2016
It turns out that Michael Mann's claim was made in an email, the rest of which has not been quoted. As such, I can make no informed comment on what he meant. Supposing TonyW @13 is correct, however, then using GISTEMP, the February 2016 anomaly relative to 1951-1980 was 1.9 C, making it 2.18 C greater than the 1880-1909 average.
-
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.
Prev 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 Next
