This is a new paper which takes a look at the statistical uncertainty of the long term warming trends by Pat Frank. He looks at the uncertainty of the data including that created by non-stationary errors and comes to the reasonable conclusion that global trends for the length of the temp record are statistically indistinguishable from zero. This has particular implications for model verification and especially to the real need for vetting errors in station measurements. Pat asked me to post on it here, and he has written a blog style explanation of his results below.
I want to copy the abstract and a bit of the conclusion here before his post just to help frame the discussion.
ABSTRACT
Sensor measurement uncertainty has never been fully considered in prior appraisals of global average surface air temperature. The estimated average ±0.2 C station error has been incorrectly assessed as random, and the systematic error from uncontrolled variables has been invariably neglected. The systematic errors in measurements from three ideally sited and maintained temperature sensors are calculated herein. Combined with the ±0.2 C average station error, a representative lower-limit uncertainty of ±0.46 C was found for any global annual surface air temperature anomaly. This ±0.46 C reveals that the global surface air temperature anomaly trend from 1880 through 2000 is statistically indistinguishable from 0 C, and represents a lower limit of calibration uncertainty for climate models and for any prospective physically justifiable proxy reconstruction of paleo-temperature. The rate and magnitude of 20th century warming are thus unknowable, and suggestions of an unprecedented trend in 20th century global air temperature are unsustainable.
Pat Frank —
First, my thanks to Jeff for his interest and for posting this essay.
Steve McIntyre’s ClimateAudit (CA) blog has been a terrific stimulus to zillions of its readers, including me. Back in May 2008, Steve posted on Nature’s “discovery” of the bucket-adjustment discontinuity in the Sea Surface Temperature record, three years after it had been discussed in detail at CA.
Brohan, 2006 [1] (B06) came up in the discussion of Steve’s post. Brohan, 2006 was the most recent compilation of the global average surface air temperature, HadCRUT3. HadCRUT3 was a production of the UK Met Office, which had taken responsibility from the Climate Research Unit at University of East Anglia. Warwick Hughes, a real hero in the fight to bring transparency to the global temperature record, discussed this transition, here and here.
The discussion at CA led me to read Brohan, 2006, where I noticed that they had described measurement noise as strictly random and didn’t mention systematic error at all. That seemed doubly peculiar, and that led to the analysis I’m presenting here.
Reading the air temperature literature, it became clear that this double peculiarity typified the approach to error right back through the 1980’s and before.
What I found was that Folland et al, had made a guesstimate back in 2001 [2] that the average measurement error was (+/-)0.2 C. This (+/-)0.2 C was applied by B06 and treated as random and uncorrelated among surface stations. So, following the statistics of random errors, B06 decremented the (+/-)0.2 C as 1/(sqrtN), where N = the number of temperature measurements, and as N got large the error rapidly went to zero. And that was the whole B06 ball of wax for measurement error.
To make the long story short, assessment of error methodology showed that guessing an average error is an explicit admission that you have no real physical knowledge of it. Random error is “stationary,” meaning it is defined as having a constant average magnitude and a mean (average) of zero. When one has to make a guesstimate, one doesn’t really know the magnitude, and doesn’t really know whether the error is stationary.
In short, if one doesn’t know the error is random, then applying the statistics of random error is a mistake.
Guesstimated errors don’t go as 1/(sqrtN). They go as 1/(sqrt[N/(N-1)]). That means at large N, the error rapidly goes to 1´(the original guesstimate). So, that (+/-)0.2 C error from Folland, et al., 2001, has to enter unchanged into the surface air temperature record. Since the guesstimated (+/-)0.2 C error enters both into every measurement, and propagates into the 1961-1990 normal used to calculate the anomaly trend, the total measurement error in every point in the global average temperature anomaly trend is sqrt[(0.2)^2 + (0.2)^2]=0.28 C. And so that correction alone more than doubles the nominal uncertainty in the global air temperature anomaly record.
But it’s worse than that. Kenneth Hubbard and Xiaomao Lin at the University of Nebraska, Lincoln, showed that there is a large amount of systematic error in surface station temperature measurements [3-5]. This systematic error mostly comes from solar loading on the radiation shield and from wind speed effects. These effects cause very significant deviations in measured temperatures.
Under very ideal conditions of siting and maintenance, Hubbard and Lin found that a standard Minimum-Maximum Temperature System (MMTS) sensor produced an average daytime bias of 0.43 C away from the correct temperature, with a standard deviation (the uncertainty width) of (+/-)0.25 C.
Cotton Regional Shelters (CRS) – the usual shelter for the older mercury thermometers – produced twice as much systematic error in high precision resistance thermometers, with an uncertainty width of (+/-)0.53 C. The older mercury thermometers inside CRS shields are likely to be even less accurate and less precise. But mercury thermometers inside CRS shields provide the bulk of the 20th century temperature record.
So, at the end, I estimated a lower limit of uncertainty in the 20th century global surface air temperature anomaly record by combining the guesstimated measurement error and the ideal MMTS systematic error. To do this, one has to also propagate these uncertainties into the temperature normal used to calculate the anomalies. This all produced a lower limit of uncertainty of 1-sigma = (+/-)0.46 C.
The Figure below shows what happens when this 1-sigma lower limit of uncertainty is plotted onto the GISS global temperature anomaly trend.
Legend: (•), the global surface air temperature anomaly series through 18 February 2010, (http://data.giss.nasa.gov/gistemp/graphs/). The grey error bars show the annual anomaly lower-limit uncertainty of (+/-)0.46 C.
The message is clear: including the lower limit of instrumental uncertainty, the trend from 1900 through 2000 is indistinguishable from 0 C, at the 1-sigma level.
“Unprecedented” 20th century temperatures? Hardly. It appears no one really knows the rate and magnitude of warming. Once again, climate alarm appears to be rooted in neglect of uncertainty. As with GCM projections. As with proxy paleotemperature reconstructions (which in any case aren’t even science).
This analysis is now out in Energy and Environment [6], and anyone who’d like a reprint can contact me at pfrank830 AT earthlink DOT net.
References:
1. Brohan, P., Kennedy, J.J., Harris, I., Tett, S.F.B. and Jones, P.D., Uncertainty estimates in regional and global observed temperature changes: A new data set from 1850, J. Geophys. Res., 2006, 111 D12106 1-21; doi:10.1029/2005JD006548; see http://www.cru.uea.ac.uk/cru/info/warming/.
2. Folland, C.K., Rayner, N.A., Brown, S.J., Smith, T.M., Shen, S.S.P., Parker, D.E., Macadam, I., Jones, P.D., Jones, R.N., Nicholls, N. and Sexton, D.M.H., Global Temperature Change and its Uncertainties Since 1861, Geophys. Res. Lett., 2001, 28 (13), 2621-2624.
3. Hubbard, K.G. and Lin, X., Realtime data filtering models for air temperature measurements, Geophys. Res. Lett., 2002, 29 (10), 1425 1-4; doi: 10.1029/2001GL013191.
4. Hubbard, K.G., Lin, X. and Walter-Shea, E.A., The Effectiveness of the ASOS, MMTS, Gill, and CRS Air Temperature Radiation Shields, J. Atmos. Oceanic Technol., 2001, 18 (6), 851-864.
5. Lin, X. and Hubbard, K.G., Sensor and Electronic Biases/Errors in Air Temperature Measurements in Common Weather Station Networks, J. Atmos. Ocean. Technol., 2004, 21 1025-1032.
6. Frank, P., Uncertainty in the Global Average Surface Air Temperature Index: A Representative Lower Limit, Energy & Environment, 2010, 21(8), 969-989.
the Air Vent 
UnHIbitted.
=======
But if the anomaly has gone from “somewhere between -0.75 and 0.25” in 1880 to “somewhere between 0 and 1.0” in 2000, this is clear evidence that global warming since 1880 is somewhere between -0.25 and +1.25, which is ever worse than we thought!
Oops, that 1.25 should be 1.75 which is even worse that I mistyped!
So, that (+/-)0.2 C error from Folland, et al., 2001, has to enter unchanged into the surface air temperature record.
I think you need a bit more explanation there. This can only be true if you assume that this +/-0.2 error has perfect correlation among all stations over time. As in, at any time, if one station deviates by some amount, then all stations deviate by exactly the same amount. Otherwise, if the noise is not fully correlated between stations, you end up with a bunch of independent random variables and the central limit theorem applies, right?
(MMTS) sensor produced an average daytime bias of 0.43 C away from the correct temperature, with a standard deviation (the uncertainty width) of (+/-)0.25 C.
The mean of this bias, being a constant up-shift (if I understand you correctly – couldn’t find the .43C value anywhere in Li’s papers), would be eliminated by the anomaly method. The rest would presumably be uncorrelated between individual stations, right?
Your work is very pertinent. Uncertainty (I think) can be seen as the combination of two parts, precision and accuracy. The IPCC and Hansen discussions focus on precision and ignore accuracy. (I suggest most of the blogging does also.) Hansen’s GISS graph shows +/1 0.2K (?) but presents it as if the temperature anomalies since 1880 are really +/- 0.02K. He speaks of 2010 being the hottests year because it is 1/100th of a degree warmer than the others, as if that means something. The imprecession has been lost.
But it is the inaccuracy that is the worst part and receives the least attention. Recently the TOA insolation was lowered from 1365 to 1360 W/m2, a reduction of 1.15 W/m2 whole Earth average. At the same time Trenberth was worrying in papers about a 0.85 +/- 0.12W/m2 of “missing” heat. The accuracy of previous measurements, the most significant input into our climate system, that of the sun’s output, was 1.4X that which Trenberth was trying to explain away. Even now, at WUWT, there is a theatre of discussion of where this missing heat is. But our ability to recognize such a small amount – to attribute such a small amount – is clearly inadequate. The argument really is of the “angels on the head of a pin” type.
Although the TOA insolation value is supposedly nailed down better now, it is itself a contraction of a significant variation. Due to orbital excentricity, the insolation reaching the earth varies by 6% during the year, with the greatest amount arriving in January, and the least in June. The variation is +/-10.8 W/m2. For a fixed albedo of 0.3, this amounts to +/- 7.56 W/m2 heating the Earth. Were other factors stable, one might think this unimportant at a planetary scale, but they are not.
The albedo of the Earth is said to be 0.296, but 0.3 is said to be used due to the uncertainty as to what it’s average actually is. The difference is 102.15 vs 100.8 W/m2,or 1.4 W/m2, the variation from even a glbal “average” that is useful. Long-term solar insolation variance of 0.01% is neglible, at 0.1 W/m2. But the annual variation in albedo is guesstimated at 15-20%, or +/- 8.8 W/2 (using 17.5% variation). The assumption, again, is that this variation is random and averages out over a short time period. There is no data to say that, and thinking a little deeper brings you to seriously question that assumption: the weather, being a daily expression of the differences in temperatures and air pressures, doesn’t average out except over very long times, as anyone over 35 can attest.
We have a fixed variation of +/- 7.56 W/m2 (solar) of heating from January to June with a VARIABLE of +/- 8.8 W/m2 (albedo). The albedo estimate AS AN AVERAGE from which variation occur, is around +/- 0.7 W/m2.We have the PDO and the AMO moving heat around through the oceans on multi-decadal time-frames. The northern and southern hemispheres have fixed but different land/sea/snow ratios (which is why the world is about 2.2K warmer in June than January, even though the sun is further away). Take this all together and ask yourself: what is the ACCURACY (not the precision) in our calculations of the warming of the world? The total oscillation and noise in the system with just these factors is in excess of +/- 20 W/m2. Some cancel each other, some will add, but when? Climates change through TIME. Time is the important factor.
The Earth is MOSTLY in dynamic equilibrium with changes of +/- 20 W/m2 during the year. The heating of the Earth from the sun is therefore about 238 +/- 20W/m2 at any given time during the year. Can we say that the +/- of 20 W/m2 averages out over a 30 year period to 0.3% or 0.7 W/m2? I’d say we can’t.
The measurement of insolation can change by 1.15 W/m2 in a moment while Trenberth et al work hard to find a “missing” 0.85 W/m2. Our ACCURATE understanding of heat flow from the sun into the Earth is probably less than +/- 0.7 W/m2. Instrument error – the precision part, is around +/- 0.3 W/m2. The IPCC claim that a 0.7K change in 100 years is about 1.4 W/2 (or more, or less). All these items we are worrying about are within, in my opinion, our levels of understanding (accuracy) and measurment (precision).
Hansen’s attempt to “correct” the temperature readings is an attempt to improve accuracy, not precision. D’Aleo et al suggest 0.3K of the total 0.7K is due to an improper UHIE adjustment. On the one hand we are attempting to model something beyond our ability to record accurately as well as precisely, and measure something that requires a correction for accuracy that is 40% of the signal to match the model. What part of this makes any sense to do?
Now throw Urban Warming into the mix.
http://www.nasa.gov/topics/earth/features/heat-island-sprawl.html
No warming for at least 12 years!!!
That’s what this article’s headline should have been. /
INSTEAD
AP article:
UN: 2010 tied for warmest year on record
The warmest year on record is a three-way tie: 2010, 2005 and 1998.
http://news.yahoo.com/s/ap/20110120/ap_on_sc/un_un_warmest_year
Re: Doug Proctor (Jan 20 13:46),
Irresistible opportunity for a pedantic comment here. I just looked that up. It’s actually angels on the point of a fine needle, not the head of a pin. And that’s a parody of the original discussion by Aquinas of whether two angels could both occupy the same space (they couldn’t, he decided). Another was whether angels teleported (he didn’t use that word) when they traveled from A to B or had to pass through all intervening points. It’s all Aristotlean logic too.
Pat,
Isn’t all that just the prediction interval for the data? That says very little about the confidence interval for a trend, which would be significantly narrower.
Pat,
To put it another way, if your null hypothesis is that the data do not have a trend and are stationary about the mean, what’s the probability that a random set of 121 numbers with a mean of -0.044 and a standard deviation of 0.46 will have a run of 45 data points below the mean and another run of 27 data points above the mean? I’m guessing vanishingly small, which rejects the null hypothesis of no trend with high confidence.
Then there’s autocorrelation, integration order and moving average, or ARIMA noise modeling. Unit root tests fail to reject the presence of a unit root at high confidence. That’s a strong indication that the process isn’t stationary by your definition. The minimum unexplained variance model fit for ARIMA is either (3,1,0) or (0,1,2). Note the presence of an integration order of 1, i.e. not stationary. I’m not really a fan of that sort of thing as the analysis assumes any trend is linear, which it clearly isn’t. There’s both an oscillation with a period of ~65 years and a non-linear trend under that.
On this error-bar basis, what evidence have climate hysterics ever had that validly exceeds statistical margins-of-error, assuming that accurate measurement is possible and that precise readings –rather than phony “adjustments”– follow in their train?
However defined, “global temperature” is largely a convenient fiction. Of course, Earth since about 1890 has been cyclically rebounding from a 500-year Little Ice Age (LIA), but lacking instrumentation records in aggregate represent mere interpolation.
Another nail in the AGW coffin, as if any were needed.
Here is some relevant info about Metrology. I sent this to Mr McIntyre in 2008 but he never replied, and eventually got around to sticking it on a blog…
http://pugshoes.blogspot.com/2010/10/metrology.html
To me, there always seemed to be a lot of hand waiving by the Team on this subject. They claimed that the Common Anomaly Method eliminated instrument error. As far as I could tell, they were arguing that the central limit theorem holds that the errors in sampled and averaged measurements map to a normal distribution. Therefore, at very large sample sizes, the density of positive errors equals the density of negative errors resulting in the two essentially cancelling out(the aggregate error approaches 0 as the sample size approaches infinity). As with everything the Team does, this seems wildly simplistic (and totally wrong) to me. Anyone have any thoughts on this?
@Mark Cooper
Very interesting, thanks.
That the compilation of thousands of independent observations should have the same uncertainty as a single observation seems rather odd, to say the least. I’ll leave the more statistically inclined to pick this apart more thoroughly, but this doesn’t really pass the smell test.
Also, shouldn’t there be a way to empirically test this? Take a highly sampled are (e.g. the U.S.), randomly remove half the stations, and see how the trend changes? I recall something Mosher did to that end awhile back…
No, Zeke. Just remove the poorly sited stations, the urban stations, and the GISS “corrections.” Questions may begin to clear up very quickly at that point.
Alice,
You mean like this? http://rankexploits.com/musings/2010/in-search-of-the-uhi-signal/
Zeke,
Your work is excellent, but I think you may be surprised at the result when it is done with more thoroughness.
Toto: no. In order for the CLT to apply, the errors need to be not only independent, but identically distributed. Though not explicitly stated by the CLT, stationarity is also required. These conditions are typically assumed, though I have never seen any justification. In absence of each of these conditions, you have to assume that errors do not cancel as the square root of N.
Basically, this is just a more detailed description of why the LLN is an inappropriate reason to arrive at ridiculous claims of precision and accuracy. A claim, I might add, I make regularly.
This does not even address the concept of averaging a value that is a function of multiple variables (the mapping is not one to one.)
Mark
Jeff ID:
I doubt I’ll be very surprised, at least with respect to its influence on global mean average, assuming it’s done properly.
Zeke:
I agree. This claim is pure nonsense.
#23, The possibility of systematic error is completely lost in uncertainty estimation. It wouldn’t take much.
I can’t edit the blog for some reason. How ironic that it happens in the same hour I’m trying to close it.
Thanks to everyone for your interest.
#4, Toto, the (+/-)0.2 C from Folland, 2001 is a guesstimate, as mentioned in the essay. Guesstimated errors are an implied mean, attach to each and every measurement, cannot be assumed random, and decrement as 1/sqrt[N/(N-1)]. At large N, the original guesstimated uncertainty magnitude is retained. This is discussed in the paper, and if you want more details, please look there. The 0.43 C daytime bias for an MMTS sensor came from a Gaussian fit to published data and is in article Table 1. The (+/-)0.25 C resolution width is systematic error and not random. It can not be assumed to average away.
#5, Doug, I fully agree that neglect of uncertainty is at the heart of AGW alarm.
#9, DeWitt, it’s a lower limit of uncertainty in each annual anomaly, plotted on a graph. I don’t see how trends enter into that.
#10, DeWitt, I don’t have a null hypothesis about the statistics of any trend. I have an interest in discovering the instrumental uncertainty, an empirical quantity. The uncertainty bars don’t obviate any trend. They only say that no one knows any details about a trend within the limits of (+/-)0.46 C.
That is, there are an infinite number of possible trends that one can draw between the years, 1900 and 2000, and it’s impossible to know which one is physically correct. Any arbitrary trend one might draw within the uncertainty limits can be made to have improbabilistic statistics. It’s possible that in the infinity of possible trends, most of them will be individually statistically improbable. So, you’ll have to divide your result by the number of statistically improbable trends that are possible within (+/-)0.46 C.
Given the chaos of the climate, there may be an enormous number of individually improbable trends, one of which had to appear as a mean. The fact that it has unusual statistics doesn’t say anything about its physical meaning. The physical meaning is found in the limits of instrumental uncertainty. That meaning is that we don’t know the trend to within (+/-)0.46 C.
#11, DeWitt, I calculated the uncertainties in the anomalies and the normal as though directly averaging ‘N’ temperature measurements. I wanted to use the most straight-forward method, to try to minimize the uncertainty.
The analysis included no fits to time series and no regression analysis, so autocorrelation of residuals doesn’t play into the calculated uncertainty. Clearly, there are other ways to calculate the anomaly trend for which autocorrelation will play in; see Section 3.2 for an example. But those methods were not used here.
#16, MPaul, you’re right, though the entire field seems to have assumed for decades that the Central Limit Theorem applies. Certainly, they have millions of temperature readings. But in experimental science, assumptions about the data must be tested empirically. Evidently, no one has ever done a statistically valid survey of the variances of surface station temperature sensors to test the assumption of the CLT. Likewise, never a survey of the variances of the measured SST’s (at least since Charles Brooks’ outstanding work in 1926) . So, applying the CLT and assuming all measurement errors just average away is empirically unjustified.
#4, Toto, I forgot to answer the point about CLT in my first reply to you. I trust the above does it.
#22, Mark, dead right on.
#18 and #23, the magnitude of the uncertainty approaches a constant at large N when the uncertainty per measurement is a guesstimated mean. One can’t assume the statistics of random error.
Zeke Hausfather said
January 20, 2011 at 9:43 pm | Reply w/ Link
That the compilation of thousands of independent observations should have the same uncertainty as a single observation seems rather odd, to say the least. I’ll leave the more statistically inclined to pick this apart more thoroughly, but this doesn’t really pass the smell test.
Also, shouldn’t there be a way to empirically test this? Take a highly sampled are (e.g. the U.S.), randomly remove half the stations, and see how the trend changes? I recall something Mosher did to that end awhile back…
############
Ya. First off I’ve had issues with that paragraph in Brohan from the first days we discussed it at CA, i’m not convinced by Brohan, but neither am I convinced by Pat’s take on things. But Yes Zeke, I’ve randomly removed half, I’ve removed all urban, I’ve sampled only the longest Rural, I’ve randomly selected one station per cell, You name it. Heck you can even just look at the 4 longest records.
I suppose if one were tricky one could create synthetic data using a Model and then look at the distribution of trends from those stations and then see if the distribution of actual trends matched the synthetically created ones. put another way, if the error really was that large I would expect you see a good number of long term stations with negative trends. but you dont. ?make sense?
I will say this. The handling of measurement error could use a going over.
Not sure but a look at CRN (triple redundant) and comparing it to neighbors doesnt support Pat’s contention. Further, if the error size is what pat calculates, then what would you expect to see in terms of correlation between two sites near to each other? My sense is that if pat is right correlations would be a mess. But they are not.
Pat,
Maybe you don’t have a null hypothesis of no trend, but the header statement by Jeff says that you come to the conclusion that the trend is indistinguishable from zero. Is that a correct statement of your position? If it is, then my comment about run length above and below the mean still applies. There is a trend and it is distinguishable from zero.
Think of it as a Bernoulli trial. The probability that a measurement will be above the mean if there is no trend is 0.5. This is true for all values of the uncertainty of measurement. In other words, it’s a coin toss. What’s the probability that when tossing the coin 121 times, you will start out with a run of 45 heads and end with a run of 27 tails? It’s vanishingly small. So your statement that an error of +/- 0.46C means that any trend in the data is indistinguishable from a trend of zero is wrong. A positive trend in the data is about as certain as it can get.
To reduce the uncertainty by averaging [sqrt(n)] you need to be measuring the same thing. But temperature changes. Wind changes. Clouds change. Solar angle changes. So you don’t have independent measurements of the same thing (like the length of a steel bar in a constant temp environment). And of course if you use different instruments at the same place/time they need calibration and you need to see how well they track. etc. etc. etc.
BTW WUWT started because A. Watts did some experiments and found all kinds of interesting things. And none of them good in terms of measurement. He put a number of temp sensors around his yard and came up with all kinds of numbers. And they did not all vary in the same direction. Some up some down i.e. different trends at different times. Divergence.
It is impossible to say if moving a site 10 ft will give the same result. So what is the Real™ temp? What is the Real™ trend?
Measuring the real world is hard. What is the diameter of a rock? How do you define the axis of measurement? Measured in the day or night? Wet or dry?
DeWitt Payne said
January 21, 2011 at 4:30 pm
Sure you get a positive trend. Now is it real or statistical variation?
Re: M. Simon (Jan 21 18:38),
Now who’s having problems with reading comprehension.
If the probability of something happening by chance is vanishingly small, that generally means it should be considered real. If I take just the 1970-2010 data to minimize non-linearity and calculate a confidence interval for the trend using a standard error of 0.46 degrees instead of the calculated 0.1 degrees, a trend of zero is still outside the trend confidence interval.
#28, Steve, if much of the sensor error is systematic, due to uncontrolled environmental variables, and if environmental variables are correlated regionally, how would that impact overall data correlation among regional surface stations?
#29, DeWitt, I do say in the paper that the trend between 1900 and 2000 is indistinguishable from zero. The conclusion is based primarily on the physical uncertainty at the instrument. No amount of statistical logic about trends is going to fix basic instrumental error.
The mean line is not an arbitrary set of numbers, the meaning of which ends with its statistical properties. The mean line represents a claimed physical phenomenon. However, the measurement uncertainty envelope covers the magnitude over the entire century, and there’s no statistical way around that.
Your argument focuses on the mean we observe, without taking into account that if the systematic instrumental errors were different over the years, the mean line would very likely look different. But it would still exhibit a vanishingly small statistical probability, and a purely statistical argument would insist it must therefore be real.
But it wouldn’t be real, because the systematic error remains large and statistics fail with systematic errors. The problem must be viewed within physics, specifically within measurement theory as applied to instrumental error, and not from within statistics.
Let me add that statistics says nothing about causality. Nevertheless, you’re arguing deterministic causality from the statistics of a two-valued random process. That’s not a good model. The only conclusion true to your analysis is a statistical one, that the mean line exhibits a statistically low probability if the magnitude sequences in observable trends are governed by random processes. But that says nothing about the physical likelihood of a given deterministic trend.
The only way for the observed trend to be physically real is if the set of systematic errors happens to self-cancel. However, the only way to know this is empirically, by monitoring the systematic effects and showing that they’re normally distributed at the large number limit. But there’s no known way to resurrect the systematic effects that impacted surface station sensors over the past century, so there’s no escape.
#31, M. Simon, in my paper I go through three statistical cases where error can decrease as 1/(sqrtN). Two of them involve variable magnitudes. The only real requirement for 1/(sqrtN) error reduction is that the noise be stationary (noise mean = 0, average magnitude = constant).
Systematic error is not stationary, however. It needn’t even repeat. Systematic error due to local meteorology changes literally from minute to minute. The only way I know of, to currently to deal with it, is to use the real-time filtering method described by K. Hubbard and X. Lin. in “Realtime data filtering models for air temperature measurements” Geophys. Res. Lett. (2002) 29(10), 1425.
Also, in Brohan 2006, the guesstimated average (+/-)0.2 C station error isn’t known to be random, either. So decrementing it as 1/(sqrtN) is a mistake.
Re: Carrick (Jan 21 00:20), Zeke:
If there are independant measurements of the same thing, then we can improve the accuracy. But not if they are not measuring the same thing. How does knowing the temperature in Moscow help us refine our estimate of the temperature in New York? Nor does it help to say we are talking about global average, IMO – in that case, the same data that is used to create the average is used to refine the average???
mpaul said
January 20, 2011 at 8:44 pm
Therefore, at very large sample sizes, the density of positive errors equals the density of negative errors resulting in the two essentially cancelling out(the aggregate error approaches 0 as the sample size approaches infinity). As with everything the Team does, this seems wildly simplistic (and totally wrong) to me. Anyone have any thoughts on this?
Yep. As the sampling size approaches infinity, so do The Team’s errors.
Ugh. Blew the tags. The last sentence is mine, not part of the quote.
What I take Pat Frank to be saying here is that one cannot assume the instrumental error in measuring “official” temperatures to be random and therefore the statistics commonly used to estimate uncertainty are incorrect. He also states that the source of the past instrumental errors cannot be resurrected for analysis. The essence of the disagreement here would then appear to be to refute Frank’s claim on using other avenues to determine the randomness of the instrumental error.
The official keepers of the temperature data sets have made efforts (that are certainly open to criticism as being incomplete) to make adjustments to the record and show evidence that the adjusted temperatures contain random errors. Certainly the keepers of the data sets realize the problem to which Frank refers here and it is certain that the net effect of the adjustments has been in one direction, i.e. the unadjusted temperatures contained errors that were not random. Therefore, one must deal with the issue whether the adjustments are correct and complete or that independent methods are available to show that the errors were and are mostly random.
We do have the satellite measured temperatures since 1979 and the agreement with the surface based temperature data sets has been good over that period of time. I have to assume that the satellite measurements are independent of the surface ones since no one has ever showed me where these measurements were dependent. I think that many make the assumption that since the agreement exists from 1979 to present that it well must have existed previously. With that assumption I do not agree.
A good example of where the assumption could fail is the analyses of the USHCN station trends versus the CRN rating placed on them by the Watts team. A recent Menne paper looked at CRN12 versus CRN345 ratings from the Watts team for the period of the 1980s to present (I do not recall the exact year in the 1980s) and found no statistically significant differences. The first problem was that unless you use CRN123 versus CRN45 you do not obtain sufficient sampling of “good” stations for a statistical comparison since variations from station to station within a given CRN rating are large. The bigger issue here, however, is that the changes that most greatly affected the micro climate of the down rated stations could well have occurred before 1980, like for example paving around a station. If I then compare the satellite data to the surface measurements with regards to trends 1980 to present and the surface stations have more or less stable micro climates during this period of time (although some would continue to have poor CRN ratings for past transgressions) I would expect to see similar trends. In my own limited analyses (with RomanMs showing me the correct statistical direction) using CRN123 versus CRN45 and going back to 1920 I found differences between the grouped CRN ratings that were significant.
If any of you have worked with the historic surface temperature data you soon realize how sparse both spatially and temporally the data was as you go back in time. In order to get regional and global mean temperatures requires that a lot of space and time needs infilling and that in turn depends on establishing both spatial and temporal correlations. This infilling then has a measure of uncertainty attached to it that is not a simple as it might appear to estimate. As a fun winter project for this retired guy I am looking at these space/time correlations with proxy, station, surface grid, satellite grid and climate model grid data. There certainly are differences between these correlations from these various sources.
Pat,
Either the temperature data actually measure temperature or they’re made up out of whole cloth. By assigning a finite error value, even if it is much larger than the assumed error, you concede that the data are, in fact, measurements of temperature. In that case the Bernoulli trial argument is still valid and there is still a statistically significant positive trend in the twentieth century data.
Frank has tackled only one of the very many false statistical assumptions of the “Mean Global Surface Te,perature Anomaly record”
They include: very poor, unrepresentative sample. A believable average could not be estimated from such a sample Also the number, location, and reliability of the sample changes every hour, day, week, decade. Averaging is made with skewed samples, so each one produces bias. Watts has shown that the best US sites cannot measuree better than a degree or two, today. Older and remoter sites are bounmd to be worse.Sea samples are even worse. Missing records and regions have to be guessed.
The total error must be several times that sown by Frank
40, DeWitt, the data are temperature measurements contaminated with systematic error and with a station error that is guesstimated to average (+/-)0.2 C and which is not known to be random. Is error-contaminated temperature data statistically identical to uncontaminated temperature data?
For the sake of this conversation, let’s agree that a two-valued Bernoulli model yields a statistically significant trend. Why is a two-valued statistical model physically appropriate to measurements of a chaotic process that need not at all hew to random, and that are further excursed by uncontrolled physical variables?
If it is not, then your statistical conclusion is both true and physically unimportant.
Further, as pointed out above, if the systematic error, in particular, had been different over time, one would have obtained a different trend with a similar statistical improbability. There is a possibly large number of such trends, and your statistical conclusion has to be renormalized in light of them.
In the large array of statistically improbable trends driven by systematic error, one of them must finally appear. Analyzing the finally realized trend in isolation, ignoring the large array of unrealized possibilities, gives a false significance to the statistical conclusion. That the realized trend may individually be statistically significant is particularly impressive. Taken individually, they’d likely all be statistically significant.
The physical meaning of the trend is what we’re interested in. That meaning is conveyed by the uncertainty limits reflecting physical error. Those limits say the trend is physically meaningless.
#41, Vincent, I agree with your points. Local meteorology alone variably impacts the accuracy of surface station temperature sensors. And Anthony’s amazing work, with his volunteer network, has shown the USHCN, reckoned among the world’s best, to be rife with large biases. So, yes, one would think the actual uncertainty would be much larger. Finding out how much larger would be an enormous amount of work, which is why I chose to find a way to a lower limit. 🙂
I’d like to add that though I’ve offered here at tAV to send reprints on request, I’d ask that those of you who have access to academic accounts, or who have fine personal incomes, to please purchase the article from Energy and Environment, here.
Energy and Environment has proved to be one of the few remaining journals in climate science where one can anticipate a uniformly dispassionate and even constructive critical review. In the principled stand of its editor, Dr. Sonja Boehmer-Christiansen and its Multi-Science publisher, Dr. William Hughes, E&E has been thoroughly in support of open and transparent science when so many have abridged it.
The journal merits support, and deserves to recover their fair profit for publishing my article.
Thanks very much,
Pat
#42 should have read, “That the realized trend may individually be statistically significant is notparticularly impressive.”
Interesting discussion, and many thanks for all the work that’s gone into the representations. I’d be interested in a closer analysis of the possibility of the instrumental error being random or not. To me it seems likely that the errors will be random. The alternative is that there is a common class of errors in a significant subset of measurements leading the data to be skewed this way or that, and I would have thought such a common error would be easy to spot?
from K Frisch #39:
“If any of you have worked with the historic surface temperature data you soon realize how sparse both spatially and temporally the data was as you go back in time. In order to get regional and global mean temperatures requires that a lot of space and time needs infilling and that in turn depends on establishing both spatial and temporal correlations. This infilling then has a measure of uncertainty attached to it that is not a simple as it might appear to estimate. ”
I have a lot of philosophical problems with “infilling” data. While it may be useful for helping to test a mathematical model, it has nothing to do with experimental science. The measured numbers are the data available. Rather than use statistical methods to fabricate data, make the best models you can based on the actual data. What happens is that the data is so sparse early in the record that any modeling of a global temperature would be subject to such a wide error estimate that it would be useless. Even today, with the USHCN estimating arctic temperatures as far as 1200 km. the results don’t agree with actual measurements. Nothing beats having adequate data. When you don’t have it you just have to admit it and not try to massage it into a trend.
I’m your typical “alarmist” but open for now to Mr. Frank’s conclusion that the surface air temperature record is not sufficiently precise to show conclusive warming. Challenging the science is absolutely necessary to ensure that it’s robust.
However, as I read Mr. Frank’s graph, no warming, or perhaps even a slight cooling is just as likely an extreme warming trend of about 0.35° C per decade. Any curve within the error bars is equally possible, no?
What, then, to do? Let’s say I’m a teacher leading a group of students on a hike through the woods and we get lost. We find an old path and come upon a bridge. It looks safe enough, but someone, I suspect the property owner who’s afraid of being sued, has put up a sign reading “Danger, keep off”. In the event the bridge fails, though, whoever is on it will die. There is a safer option, but it’s significantly less convenient.
While I’m standing there pondering all this, Mr. Frank magically falls out of the sky with an analysis that shows we can’t be certain that the bridge is safe.
As a bridge-failure alarmist, I’m going around. You?
#49;
Argument from analogy is always problematic, but let’s play your game.
The longer route has a virtual statistical certainty that ~20% of your students will die of hypothermia and starvation before you get to the other side of the river, and most of the rest will have lost most if not all of their hiking supplies.
Now which route are you going to take, Mr. Cautious?
Brian H, #50: Analogy is of limited usefulness, I agree, but can be illustrative, so thanks for humoring me.
Of course there is a cost associated with any action. Logically, you seem to suggest that faced with an unknown risk of catastrophic consequences, the response should be determined by weighing the costs associated with the alternatives. The interesting thing is that we’re no longer talking about the bridge. What if, instead of being reckless, we figured out a better way over the river?
The Precautionary Principle is a Paean to Ignorance. Yes, let’s know more before we act. Sure you’re lost; primun non nocere. Why cross the river in the first place? Did you cross it before you got lost? I’m tellin’ ya’, if a metaphor gets lost in the woods, where are the storm warriors to rescue it?
=============
I can identify three error sources in manual temperature observations.
Siteing error, where the location is not representative.
Measurement error, the thermometer is not calibrated or accurate.
Reading error.
I come from a survey background, so I wish to put forward my observation that humans generally read incremented measurement markers with a positive bias. Translated to mercury thermometers, I’ll bet there is a natural positive bias from reading by eye.
A little search by me on the survey correction technique called Traverse Close, found this little gem. http://keithdennison.com/surveying/traverses.pdf
Here they observe:
There are various methods used to adjust surveys.
No method will improve the accuracy of the observation but only the consistency of geometry.
Adjustment can generate a false confidence in the quality of a survey.
Repeat the Survey; as no amount of statistical manipulation can improve poor observations.
Analogies can be sound, but the difficulty is getting the various elements of an analogy to line up with the real event being compared. In the bridge analogy, the question is, is going the long way as dangerous as the worst case scenario that could occur going over the bridge. I would suggest not. Going the long way, involves an improvement in the efficiency of renewable energy production (including biomass which produces co2) to the point that it can replace fossil fuels. The main groups who stand to lose out here are the fossil fuel industry, and tax payers. The fossil fuel industry will be replaced by another industry revolving around renewable energy production. In due course it will balance out economically and in term of jobs.
So the question in my mind is; are we the tax payer, prepared to fund a movement to a cleaner environment and to a renewable energy based economy and society, with cheap and abundant power (not initially, but after the investment and technology has matured)?
My personal answer is YES. And I believe this is an appropriate path to take regardless of whether global warming is happening or not.
Above where I say including biomass, I of course mean “excluding”!
You are incorrect in you assumption about “the fossil fuel industry.” They are actually just part of “the energy industry” and will find a way to profit off of any new energy sources that present themselves. Indeed, they are hedging their bets investing in the most likely possibilities. Their is a silly myth that “big oil” is behind the skepticism. This is just plain silly – oil companies stand to make enormous profits when market forces eventually drive prices through the roof, particularly if governments force consumers into alternatives that are scarcer. Think about why that may be…
Mark
The problem with your bridge failure analogy, Marcus is that a) you know what the results will be if you cross and the bridge fails and b) it’s all bad. With the global warming, you can only guess at possible outcomes and not all are bad.
Analogies are only useful if the premises and relationships have similar scope and relative magnitude.
Mark
Marc, your analogy is incomplete.
You need to add two engineers arguing passionately beside the bridge, about whether it’s sound or not. But their argument is esoteric and you don’t really understand it.
Occasionally one (the reassure-ist) turns toward you and says, “Look, the bridge is fine, you can go across.” But the other one (the collapse-ist) says, “Don’t listen to him. He’s paid by the construction industry! The bridge is unstable and could collapse with another misstep!”
Nearby, a member of the Progressive NGO, BridgePeace, calls to you, “That collapse-denialist is a Capitalist lacky. Don’t trust him. The collapse-ist engineer is motivated only by altruistic concern!
Meanwhile, you notice that the farmer’s cows are wandering across the bridge, and the bridge doesn’t seem bothered at all (i.e., Earth’s climate isn’t behaving in any unusual way).
The BridgePeace guy notices and says, “Those cows don’t mean a thing! They’re genetically modified, so their weight affects the bridge differently than would the weight of a pure authentic natural auroch.
He goes on, “There’s a shallow swamp predicted to be twenty miles downriver, which you and the dear children should be able to cross safely. And don’t believe the doubt-makers’ rumors about snakes and quicksand. They’re unfounded. No one’s actually ever been to the swamp.”
What would you do?
Of course there’s always the dripping with sarcasm viewpoint…
Mark
@Mark, I accept your distinction about the Fossil Fuel industry being part of the energy industry, but it doesn’t invalidate the point I’m making, which is that movement to a renewable energy based economy is something we should strive for regardless of the global warming scenario. Whether we are talking about the current players in the energy industry, a mix of old and new players, or all new players, is really beside the point. (btw, nowhere in my post was I intending to demonise the oil industry, so if you read any bias into what I said based on that assumption you can disregard that bias)
I disagree with you that the Energy industry is ready to make a quid out of renewable energy – yes they will hedge their bets, but they will also ensure they can milk fossil fuels for all they’re worth in the meantime, even if it means we are paying more at the bowser than we would need to if we had access to electric cars charged by renewable energy sourced electricity. I’m not demonising the energy industry, but I think it would be naive to believe them to be altruistic. There is a place for govt to use environmental legislation (aimed at reducing real pollution, not co2) to pressure car companies and energy companies to start looking at alternatives NOW and not in 50 years when we are paying a much higher proportion of disposable incomes on energy. There is also a place for govts to invest in research and development, and to encourage research and development into alternatives through tax breaks and grants for industry.
@pat, we can banter back and forth with each our own versions of the analogy all day long, but it misses the point taht I think was being made: that we don’t have to cross the bridge and find out if it’s going to collapse or not. There are other paths (designated as safe for the purpose of the analogy) that can be taken that avoid having to test the theory at all. There is nothing dangerous about reducing carbon output, it’s doable and in the long run beneficial in terms of the efficient renewable energy sources that will arise. btw I reject outright such silly measures as sulphur in the atmosphere or gant reflectors in space to artifically cool the globe! Those DO sound dangerous!
50: The problem with *today’s* “renewable” energy sources is the high cost to you and me, of subsidizing their existence. Tripling the cost of energy to subsidize projects using technology that has NO apparent path to long-term feasibility is simply an unneeded burden on a country and its citizens. At least that goes for the intermittent sources, wind and solar. I’m all for nukes and hydro, which *do* have a future, if we get past the politics.
The point is that we will get to that based on market forces irrespective of the global warming issue. No need to rush, it will happen through necessity anyway.
But there is… unless you don’t count the economic impact as dangerous. Whether or not it is doable, physically or otherwise, is certainly arguable.
Mark
#58: As sarcasm goes, Pat, that’s good. Thanks for the laugh. I hope you’ve got a couple of cows up your sleeve.
But assuming you’re right about the surface temperature record errors, is an extreme warming trend not just as possible as no warming trend? If that’s a fair statement, then to my alarmist mind, caution is warranted. That’s all.
Apropos analogies, there’s a wonderful segment of a BBC interview at http://www.youtube.com/watch?v=0wmuhKzYp4s.
Max, if you’re talking about nuclear power, you’re right. But there’s no present viable option for wind, solar, or biomass without a simultaneous return to early 19th century living standards. And that includes much of medicine and drugs, and may include abandoning water chlorination. Making chlorine on large scale takes lots of electricity. Expect a decrease in life expectancy and an increase in misery.
I wasn’t being sarcastic in my analogy, by the way. There’s plenty of evidence supporting the scene in the social-political history of AGW.
#64, Marc, the uncertainty limits don’t say anything about the likelihood of air temperatures. They only say something about our state of knowledge.
Climate models aren’t any good at predicting small changes in Earth’s climate, Marc, and so all the predictions about hot futures from CO2 are completely unreliable.
That means we’re operating in ignorance. As I pointed out in my Skeptic article, the precautionary principle is useless when applied in ignorance. That’s because with no information to bring to bear, free probability says that bad outcomes are just as likely as good outcomes.
That means, without any evidence to help us make an informed choice, curtailing CO2 is just as likely to produce a good or a bad outcome as continuing to release CO2.
So, your alarmist mind has just as much to worry about by cutting CO2 as it does in releasing CO2.
We do have two pieces of information, though. We have ice-core data showing that for the last million years CO2 always followed air temperature, and never drove air temperature. We also know that recently released CO2 has caused the ecosphere to green up.
What is there to be alarmed about, again?
They are taring out the hydro dams here, all four will be removed by 2012.
the river will be “natural” what they will do with the silt in the bottom of the channel i don’t know, how many fish will die in the two lakes? i don’t know.
I don’t get it. It is not the CO2 it is something else.
Tim
#65: So I’m to accept that 150 years worth of thermometer readings don’t say anything about changes (or lack thereof) of air temperatures? That’s too much of a stretch for me, sorry.
Your critique of precaution applied in ignorance makes fine sense. It brought to mind stories of tourists getting caught suddenly in thick fog a few yards from shore on sandbars on the west coast of Denmark. Some of them got disoriented, walked out to sea, and drowned. Seems like the perfect illustration.
But we’ve got plenty of information to inform our choice. While temperature may lead, carbon dioxide and temperature have always danced together. That C02 lags temperature is an excellent talking point if you want to mislead someone who can’t handle credit card debt, but it’s lame from a physical standpoint. What else? Well civilization developed under fairly stable climate conditions which included a C02 concentration under 300 ppm and we’re conducting a giant chemistry experiment on the atmosphere and oceans.
Greening of the biosphere? As in Russian wheat exports halted? As in grain reserves dwindling? As in pine forests dying all over the West. That greening of the biosphere?
The tourists in Denmark, not knowing which way to go, should have stayed where they were.
Marc, 150 years of thermometer readings do not give you any information about the future. That’s the point you are either willfully ignoring or just flatly ignorant of. What those readings do tell us, however, is that nothing bad can be tied to them. The future is not only unknown, in spite of hyperventilation by the likes of you, but likely unknowable. The one thing we DO KNOW is that the measures required to cut CO2 will cause severe economic harm. When it is all said and done, the truth regarding the morality of your plan will be widely known. Personally, that’s not something i’d like hanging over my head through the end of my life. You choice.
Mark
Marc, the arrow of time would suggest that temperature causes CO2 change and not the other way around. Perhaps we are testing something that has not happened before, but do note that CO2 trails temps on the way down, too. CO2 has not been able stop a cooling.
I now mention the Minoan, the Roman, and the Medieval Warm Periods. And the Mayan civilization. Your ‘fairly stable climate conditions’ don’t seem stable to me. Furthermore, were biological niches not to evolve, species evolution would be retarded.
Given the concatenation of cooling phases of the oceanic oscillations, and given the Cheshire Cat Sunspots, and the lack of evidence that CO2 is anything but a weak climate agent, the chances of cooling over the next century are greater than warming. Yes, despite a historically unprecedented release of anthropogenic CO2.
Surely you jest about biomass? The greening is measurable and considered commensurate with the increase in CO2 in the range of 20-20%. You really do have blinders on, don’t you?
Mark T points out your choice. The opportunity costs lost to future generation from this sad scientific mistake are already huge. Choose wisely.
==================
Heh, 20-20%. Now I can claim to be right whatever the truth is. Mebbe I better go look. Certainly Marc should have.
===============
#68: No, Mark, I don’t think I am saying the future is knowable. That’s the problem. Pat’s analysis makes statements about our knowledge of the past and present to the effect that there is no discernible temperature trend. You, however, claim certainty about the economic consequences of measures to cut CO2. You can’t predict physical processes but you can predict economic ones?
#69: Yes, Kim, we all know that warming causes the oceans to release co2 into the atmosphere. And we all know about feedback, right? You may pack your arrows of time back in your quiver and shoot them at someone else.
But hey, just for fun, if you don’t know what the temperature is, how do you know the increase precedes CO2? (Sorry, couldn’t resist.)
Oh, I forgot to address biomass. Yes, we all know, well, most of us know, well, anyway, a few of us know plants take carbon out of the air and convert it to plant fiber and oxygen. We also know that plants require water to grow and thrive under certain temperature conditions. And lest we forget, there are bugs, some of which do better in warmer temperatures, that like to eat plants before we do. In a lab with plenty of water, controlled temperature, and no pests you might well get 20 percent more growth by increasing co2. But in the real world, no, we’re seeing raising food prices, falling grain reserves, drier soil, more pests, and more forest fires. If there is data that shows that the biosphere is becoming a larger carbon sink, I’d sincerely like to see it.
Oh, sure, don’t believe the 800 year lag. The precision of the ice cores is poor, so you have that.
Your cognitive filters about biomass deceive you. The question is not whether biomass is increasing, but rather whether its increase is outstripping the CO2 increase or not.
================
Hmm, I’m beginning to think you are too illogical and ignorant to choose wisely. Please, though, don’t just grab the pretty one; think awhile first. You have not answered satisfactorily Pat Frank’s two points @ 3:51 PM on 1/25, nor my recapitulation thereof.
==================
Oh, gad, it gets worse. I can predict some of the economic consequences of raising the price of energy.
=======
See, you have got right at the problem, though. We cannot predict future climate, and we do not know the consequence of this release of anthropogenic CO2. We can predict the effect of raising the price of energy. The solution to this inequality was the Copenhagen result, which could have been predicted. In fact, I did predict the failure of Copenhagen a year and a half before it failed, and it was precisely because the poor of this earth to not share the precious conceit of a western elite, that guilt over industrialization is worth committing suicide over.
====================
er, ‘the poor of this earth do not share’. And ‘precious’ and ‘conceit’ are both used with archaic meaning.
=========
Pat
As I lack a fine personal income (the cheques from Big oil never seem to arrive) could I have a reprint of your paper please?
tonyATclimatereason.com
tonyb
Kim (73, 74),
Excuse the cognitive filters. I admit to confusion over biomass, I have no idea what you’re saying. I wasn’t very clear in my response, sorry. I’m saying that poorer growing conditions are probably slowing the uptake of carbon in the biosphere faster than the increased carbon dioxide is greening it up. As indicators, I point to falling grain reserves, falling grain exports, and increasing food prices. I don’t have better evidence on hand, so I’d be happy to look at the quantitative evidence of biomass increase in the real world that you base your 20% statement on.
As to being “too illogical and ignorant to choose wisely”, if you go back to #64, Pat’s saying we’re all too ignorant to choose wisely. In that case, we get to choose whatever we like. I like jobs and prosperity, so I’ll choose investing in infinite sources of domestically produced energy so I don’t have to import a billion dollars a day worth of oil. If you’re going to toss out precaution on Pat’s basis, you don’t get to claim dangers going the other way. That would be illogical. So would telling me not to worry about climate change because the Mayans were wiped out by what again?
Marc, the greening seems to be commensurate with the rise in CO2. The evidence is that increased uptake by the biosphere is right in line with the increased CO2. Your cognitive filters see crops and prices as evidence. They are poor evidence.
Oh yes I can claim known dangers from demonizing CO2 and raising the price of energy. And no, you can’t claim known dangers from rising CO2.
Yes, the Mayans were wiped out by natural climate change. I used them as an example to counter your assertion that civilization developed in a ‘relatively stable climate’. Yep, I like your logic. It’s way fun.
===========
#67, Marc, your personal incredulity about a quantitative result (climate station thermometers are inaccurate) is not a rebuttal. Try again.
You wrote: “That C02 lags temperature is an excellent talking point if you want to mislead someone who can’t handle credit card debt, but it’s lame from a physical standpoint.”
Temperature led CO2 by an average of 800 years through all seven ice ages spanning nearly 1 million years. That’s not a “talking point,” that’s a fact. Nor is it “lame from a physical standpoint.” As a physical fact, it refutes the idea that CO2 is a primary driver of air temperature. As Kim pointed out in #69, seven times global temperatures fell into ice ages while atmospheric CO2 was high. It’s only by special pleading that people argue today’s CO2 is different. That’s not science.
Plant stomata show that atmospheric CO2 was significantly higher than 300 ppmv many times in the recent past. WUWT has a nice science-oriented essay about it here. Ice core records suppress the record of variation in atmospheric CO2 because of the 100 years required for ice firn to close.
The greening of the biosphere under increased CO2 is mentioned in my Skeptic paper linked above, with references, and further discussed in Supporting Information Section 6. But, here’s another paper you can look at directly.
The only question important to your concerns, Marc, is whether climate models (GCMs) make reliable predictions about future climates. We know they do not. Therefore, predictions of torrid dooms are unreliable, and all your hand-wringing about insects and forests are AGW pointless. GCMs are unable to say anything valid about the effect of added CO2 on climate. There’s no evidence whatever that increased CO2 is affecting global climate, or global insects.
In the absence of any evidence, your last line applies: “The tourists in Denmark, not knowing which way to go, should have stayed where they were.”
Have the courage to take your own advice, Marc. Not knowing which way to go, we should stay where we are.
The only thing we should be doing is preparing for the inevitable, but unpredictable, natural disasters. We won’t do that worrying about climate boogymen. If anything, AGW alarm is a dangerous diversion of attention.
Re: Marc (Jan 26 12:47),
So turning massive amounts of US corn into ethanol instead of food has no effect on grain reserves, exports and food prices? It’s all due to poor growing conditions. Not!
marc said
“So would telling me not to worry about climate change because the Mayans were wiped out by what again?”
Natural climate variability Marc,not the supposed man made variety. It illustrates that we need a plan A and a plan B. Those would cover natural warming AND natural cooling. Our way of life is not very resilient to disasters so cold would be much more serious than warmth. A good book to read on that subject would be CHILL by Peter Taylor.
Both he and I are agreed that we also need lots of other plans for other more pressing concerns than CAGW. From my perspective an urgent preventative plan to circumvent a man made cyber attack is top of my list and this would also serve to combat a natural and devastating Carrington event.
Those latter two concerns are infinitely more likely and likely to be more catastrophic than a still hypothetical man made climate calamity.
Tonyb
#79 “Pat’s saying we’re all too ignorant to choose wisely.”
Very tendentious misreading of what I wrote, Marc. My point about the “Precautionary Principle” is that total ignorance makes it totally useless.
The wise choice is what you offered and immediately proceeded to ignore, but with which I agree: do nothing peremptory (about CO2).
Getting away from middle eastern oil, on the other hand, is a great idea. Likewise Venezuelan oil. It seems to me that the wisest option is a thoughtful and all deliberate speed move into nuclear power.
Re: Pat Frank (Jan 26 16:01),
Indeed. What else is going to generate electricity to power those million electric cars we’re supposed to be driving by 2050 or whenever? Windmills or solar? Not likely. At least not without the equivalent battery capacity of several million more cars plugged into those windmills and solar arrays. Nuclear is dirt cheap by comparison. Waste storage is easy if you let go of the Carter era illusion that reprocessing leads to weapons proliferation. The proliferation horse left the barn long ago.
84.Pat Frank said
January 26, 2011 at 4:01 pm
Unfortunately, given that oil is a globally traded commodity, getting away from any one source of oil is not only impractical, but likely unrealistic as an option.
Mark
#86, Mark T;
Indeed. The term is “fungible”; it’s a common pool: any avoidance of a particular supplier merely swaps customers from one or more to others.
Less fungible, but still medium-term substitutable, is the huge natural gas supply (with some associated light oils) being found all over the US (and the world):
http://www.forbes.com/forbes/2010/1108/opinions-steve-forbes-fact-comment-energy-crisis-over.html
Among the countries that are likely to become energy-independent:
China, India, Israel. Maybe Japan (methyl hydrates). Big changes afoot.
#83:
You’re right Pat, my incredulity was not a rebuttal. I misunderstood you to be saying that the uncertainty limits set no bounds for possible temperature trends. There must be some bounds, so that didn’t make sense to me, but I think I get it now.
I’m surprised that the co2 lag issue is troublesome. Yes, you’re correct about the 800 year delay, but it’s still lame physics. Assuming for a moment that co2 is a greenhouse gas, it’s a trivial matter to construct an amplifying (or dampening) feedback mechanism. As intuitively appealing as it may be to gullible thinkers, the fact that co2 lags temperature proves only that something else got the ball rolling. You can argue that co2 is not a greenhouse gas if you like, but not with that logic.
Furthermore, their is a research to the effect that the ice age temperature swings are too large to be due to orbital and albedo forcings alone; carbon dioxide is both necessary and sufficient to cause the observed temperature changes.
I stand corrected on the biosphere. Grain prices are a lousy proxy and there is some evidence to support “greening up” at current co2 concentrations.
There’s overwhelming evidence that co2 causes global warming. Here’s the shortest possible presentation of the evidence I find convincing.
1. Paleogeological research strongly suggests that co2 was the primary driver of large temperature swings over millions of years. Multiple, independent co2 reconstructions provide evidence of high co2 levels during warm periods and low co2 levels during warm periods. Carbon was cycled through vulcanism, weathering, and subduction. Nothing explains the temperature changes better than co2.
2. Even if we toss out surface air temperature records as inaccurate, other observations of climate change, including satellite temperature data, indicate warming concurrent with modern rapid co2 increase. There may be other suspects, but they’ve got good alibis.
3. Basic physics explains and predicts warming even if GCM models are not accurate enough to project future warming with confidence. There’s more than one way to arrive at a temperature sensitivity of 3 deg. C per doubling.
Is that a bulletproof, open and shut case? No. But it’s solid evidence and becoming more robust as we learn more. Until the cows cross the bridge, I’ll hold on to my conviction that unabated use of fossil fuels is a very risky path, especially when the technology exists to stabilize and return to 350 ppm.
Thanks for engaging.
1. no.
2. no.
3. yes
====
Oops, reread #3. It’s a no too.
===========
Alright, I’ll give you a little more than that. But since we destroyed your previous statements of faith, why do you think you’ll do any better with these?
1. You are confusing correlation with causation. The arrow of time that you sneered at previously strongly suggests that temperature causes CO2 level, and not the other way around.
2. The only time that temperature changes corresponded well to CO2 level rise lately was in the last quarter of the last century, Not before and not since. There has been a steady rise in temperature since the end of the Little Ice Age, well before CO2 started rising, for which we do not know the cause.
3. CO2 has a radiative effect in the laboratory. Neither you nor anyone else knows what its net effect in the atmosphere, after feedbacks we haven’t even thought of yet have had their influence. Sure there’s more than one way to guess at a sensitivity. That’s why the range of guesses is so huge.
And sorry, the evidence for CO2 causing catastrophic global warming is apparently less robust every year. I say ‘apparently’ because it was never robust to begin with.
=================
‘Technology exists to stabilise and return to 350ppm’. Show me, son. Not the dream, but the actual scenario, complete with human effects.
===================
#91: Yes, Kim, warming causes c02 increase. I don’t dispute that, Al Gore doesn’t dispute that.
But, humor a poor, foolish alarmist for just a moment. What would happen to temperature if, in some hypothetical alternate universe, c02 were able to cause warming? Temp. goes up, co2 goes up. Then what? Could you get a relationship that looks anything like the correlated temperature / co2 record of the last 800,000 years? If that doesn’t make any sense, please sit and learn a little more. There are plenty of challenges to AGW theory, but this just ain’t one of them.
I shot an arrow into the sky. It fell to earth, I know not why.
================
Re: kim (Jan 27 05:34),
Strictly speaking the technology does exist to stabilize CO2, it’s called nuclear power. The problem is in the implementation. It simply isn’t possible to build nuclear power generating plants fast enough. Ignoring the political problems, the infrastructure for building them doesn’t exist and would have to be built. At the most basic level, it’s not at all clear that enough steel and cement plants exist to supply the basic materials of construction. Then the plants would have to be designed, sited and permitted. You can’t do that overnight. By comparison, converting automobile factories into tank factories in WWII was trivial. Then there’s the problem that all that takes a lot of labor and capital. Where’s that going to come from?
D’accord.
=====
Uh, you do realize that the relationship you just pointed out, marc, is the opposite of what you are claiming, right? This is actually a very diificult hurdle for AGW even if you choose not to believe so. AGW theory requires a reversal of cause and effect to work. It also requires that we blindly assume that the several hundred year resolution of the record is sufficient to capture decadal level changes similar to what we are seeing now (this ia a sampling theory problem.)
mark
Mark, sorry, I’m not sure what you’re saying. A reversal of cause and effect? As in the effect (increased C02) causes the cause (warming). Why, yes, that’s just a positive feedback.
I can’t say it very well so I’ll quote Wikipedia. “A system exhibiting positive feedback, in response to perturbation, acts to increase the magnitude of the perturbation. That is, “A produces more of B which in turn produces more of A”.
So, in other words, my position is that it’s not logical to say that, as several people have, that since A causes B, B cannot cause A.
(Apologies if you know all this, I’m responding to what I think the problem is, but I’m not 100% sure. I admit to cognitive filters and my ego has some stake in my being right. Unfortunately, that interferes with my critical thought all the time. I applaud anyone who manages to recognize and overcome that.)
I don’t understand the sampling theory problem so I’ll let someone more knowledgeable address it, and I’d rather not stray off into any other objections to AGW theory at the moment.
<a href="https://noconsensus.wordpress.com/2011/01/20/what-evidence-for-unprecedented-warming/#comment-45656"#88, Marc, no one is arguing that CO2 is not a greenhouse gas. So please put that aside.
That said, CO2 trailed both rising temperature and falling temperature seven times in sequential ice ages and their recoveries. That’s 14 independent instances of lag. High CO2 did not prevent falling temperatures seven times in a row. There isn’t a single known deep-historical example of CO2 leading temperature. That’s not “lame physics” no matter how often you might claim so. It’s a fact.
Your alternative claim that “something else got the ball rolling” for both rising and falling temperatures is just another instance of special pleading. That is lame physics.
If the ice age temperature swings are too large for orbital or albedo forcings, then the rational conclusion is that no one knows what caused the ice ages.
CO2 is neither sufficient nor necessary, because the water vapor feedback that grants a large effect to CO2 is an ad hoc assumption in the models. It’s not a deduced result from a complete physical theory of climate.
Further, Roy Spencer has pointed out that the usual suspects have almost certainly got the temperature-cloud interaction backwards, mistakenly putting cause (clouds) after the effect (air temperature).
So, you really have no leg to stand on there.
You wrote that, “There’s overwhelming evidence that co2 causes global warming.” Let’s be clear: what we’re talking about is recent warming, not the predicted growth of greenhouse warming as CO2 is increased from zero ppmv.
With that clarification, the meaning of a fact in science is granted only by its explanation within a predictively falsifiable physical theory.
The only basis for your claim, therefore, is from the predictive results of climate models. But climate models are known to be predictively unreliable.
Furthermore, no one has ever propagated the parameter uncertainties through climate models, and so their physical uncertainty limits are entirely unknown. Therefore, they are in no condition to give physical meaning to the rise in atmospheric CO2.
That in turn means there is no evidence whatever that the recent rise of atmospheric CO2 has caused any of the recent warming.
Regarding your 1: here is a plot of atmospheric CO2 and temperature over the last 600 million years. Where is your “primary driver” effect?
Regarding recent trends, maybe nothing explains the CO2 changes better than temperature. How do you know which is which? Not from climate models, that’s for sure.
Your 2: So, for you good science is correlation = causation. Life expectancy has been steadily climbing over the 20th century, too. Maybe longer lives are causing higher global temperatures. Why not? Your science finds causality in high correlations.
Your 3: Basic physics does nothing of the kind. Radiation physics only applies to energy. It says nothing about air temperature within a climate system.
Only a complete theory of climate will predict how the minor increase in energy deposited into the atmosphere by higher CO2 will be distributed into the various climate modes. Radiation physics says nothing about that.
If the increased energy goes into a slightly more vigorous tropical hydrology, there may be no detectable change in air temperature.
At the end, you have no evidence and no valid theoretical base. You can believe in AGW as you like, but don’t delude yourself that you’re being rational.
Pat Frank;
Thanks for that. The accounting for the “work” component of energy input-output models and cartoons has been very slapdash. One way or another, everything on Earth is in variable motion.
#84: Pat, I understand and accept your point about applying the precautionary principle under total ignorance. I think it’s extremely rare that total ignorance is the case. We have information, not perfect, about changing the atmospheric concentration of carbon dioxide. If you’re telling me the precautionary principle does not apply, I understand you to be saying that we have no information to bring to bear, i.e., we’re ignorant. Of course you don’t really believe that, but you can’t have it both ways. You can’t say we have information to support your path, but no information to support my path . My statement is tendentious perhaps, but points out the conflict.
How good does information have to be? There’s no easy answer for that. It depends on the scale of the potential consequences. To avoid action because we can’t know the outcome with certainty is an argument that might be used to justify just about any action and thus undermines ethical choice. We can’t be sure of the consequences, so why worry? Ignorance is bliss, in other words, which is a paean if ever there was one.´Is that how you see our choice on this issue?
Marc: The PP invariably comes down to the value of a human life. Should the best efforts to quell CO2 be made, the results in terms of lives lost are predictable with some certainty, as the destruction of wealth and resources, and the availability of cheap energy (upon which life in developing countries depends critically) is curtailed. Some estimates run into the billions. This, without ANY suggestion that there would be ANY difference in the amount of CO2 emitted, and truly NO evidence to support ANY calculation of improvement to world population. Except that it would be reduced. Which seems to be a primary aim of AGW alarmists.
The mass-murder label isn’t applied as part of the PP scenario, but it should be, and should be publicized widely. No pressure, of course.
CO2 isn’t the primary cause or probably even the major cause of the temperature swing from glacial to interglacial and back again. It can amplify the range of the swing, though. Some people think the amplification may be on the order of 50%, but that would require the long term sensitivity to be at least 5 K/doubling or 1.5 K m2/W. 30% at most would be more like it. The planetary average temperature is estimated to change by about 5 K from glacial to interglacial. A change of 180-280 ppm gives a forcing of about 1.6 W/m2. At 0.3 K m2/W, that’s 0.5 degrees or ~10%, not including any other feedback.
On the mechanism of amplification with lag that doesn’t violate causality, see my post “Explaining Ice Core CO2 Lag“
#103: Thanks, DeWitt. At first glance it looks like a very interesting post that can explain the physics at issue much better than I can. If I read the conclusion right, the repeated glaciation cycles neither prove nor disprove the effect of co2 because we don’t know what else is at play.
#102, Mark: Yeah, that was a terrible video. Look, I know from my own experience that ascribing awful motives gets in the way of critical thought. If I believe your shilling on behalf of a big, bad oil company, then anything that seems to invalidate your position is going to come with a much bigger emotional payoff for me than if I think you’re trying to work with me towards a solution that works for you, me, and our kids. And then I’m impossible to reason with and anything you say is likely to further entrench my views. So I try to remember that we’re all pretty much the same and want the same things. So, yes, lets try to avoid mass genocide. I’m for that.
Marc, I’m only all in an uproar about this because I fear that CO2 is too weak a greenhouse gas to prevent the next century’s cooling from a new Grand or Lesser Solar Minimum. So yes, it’s about the children for me, too.
===========
Kim;
Yep, we’re reellly gonna wish the Magic CO2 Global Thermostat was for real!
It will be one of the most grand tragic ironies ever.
============================
#99, Pat’s comments in italics.
CO2 trailed both rising temperature and falling temperature seven times in sequential ice ages and their recoveries. That’s 14 independent instances of lag. High CO2 did not prevent falling temperatures seven times in a row.
Lag is characteristic of feedback. DeWitt has demonstrated a model that works: Co2 can cause warming even though warming causes co2. I’ve never claimed that this proves the theory — it doesn’t — only that it’s invalid to state that causality precludes effect.
There isn’t a single known deep-historical example of CO2 leading temperature.
Alley and others theorize that co2 is the most likely cause of paleogeologic climate change. They have a plausible mechanism and find evidence of a good correlation between temperature and co2 levels. But you’re right, they can’t say which came first. Paleogeologists have looked at other explanations, but nothing makes as much sense as co2.
Your alternative claim that “something else got the ball rolling” for both rising and falling temperatures is just another instance of special pleading. That is lame physics.
No, that’s just avoiding spelling Milankovitch.
If the ice age temperature swings are too large for orbital or albedo forcings, then the rational conclusion is that no one knows what caused the ice ages.
That’s right. Then you ask well, what might have caused it? What’s the best explanation we can think of? The thoery that it’s C02 has been gathering strength for 150 years.
CO2 is neither sufficient nor necessary, because the water vapor feedback that grants a large effect to CO2 is an ad hoc assumption in the models. It’s not a deduced result from a complete physical theory of climate.
Irrelevant. By sufficient, all I mean is that if true, it would explain the phenomenon. By necessary, I mean some forcing is required, not that that forcing has to be co2. I want to read what DeWitt has to say about this, but from the conclusion I gather the physics are consistent with the theory.
Further, Roy Spencer has pointed out that the usual suspects have almost certainly got the temperature-cloud interaction backwards, mistakenly putting cause (clouds) after the effect (air temperature).
I agree there’s work to be done yet on clouds. I expect we’ll be surprised by some of the ways the climate changes as temperature increases. This recent cold snap is a case in point.
…the meaning of a fact in science is granted only by its explanation within a predictively falsifiable physical theory. The only basis for your claim, therefore, is from the predictive results of climate models.
You’re conflating fact with evidence.
But climate models are known to be predictively unreliable. Furthermore, no one has ever propagated the parameter uncertainties through climate models, and so their physical uncertainty limits are entirely unknown. Therefore, they are in no condition to give physical meaning to the rise in atmospheric CO2. That in turn means there is no evidence whatever that the recent rise of atmospheric CO2 has caused any of the recent warming.
Anyone who claims that models are crystal balls is being silly. But as you know, the only way to test the hypothesis is to pump co2 into the atmosphere. If this were a science experiment, it would never be conducted. But it’s not an experiment, its a consequence of the fact that fossil fuels were once the best available source of energy. Now that they’ve become so heavily entrenched in our economy, it will take a hell of push to move on, but our political system discounts future costs and favors the status quo. We should have started 30 years ago, but I digress.
At some level, the objection comes down to this: the hypothesis can’t be validated because we can’t test it. That’s true. You win.
Regarding your 1: here is a plot of atmospheric CO2 and temperature over the last 600 million years. Where is your “primary driver” effect?
There’s no primary driver effect in this plot, but hum a few error bars and I’ll play it. No, seriously, I have to evaluate how authoritative it is. The temperatures differ quite a bit from plots I’ve seen. Besides, all I’m looking for in the deep past is correlation and the necessary and sufficient conditions.
Regarding recent trends, maybe nothing explains the CO2 changes better than temperature. How do you know which is which? Not from climate models, that’s for sure.
Fair question. But if CO2 doesn’t cause the increase, it must be something else. Thermodynamics and all that. So, until you show that co2 does not cause the predicted warming and/or come up with a plausible alternative, I believe it’s the best explanation.
Your 2: So, for you good science is correlation = causation. Life expectancy has been steadily climbing over the 20th century, too. Maybe longer lives are causing higher global temperatures. Why not? Your science finds causality in high correlations.
So, for you, good logic is effect precludes causation. Maybe stocks can’t go up when people buy them because people buy them when they go up.
Your 3: …If the increased energy goes into a slightly more vigorous tropical hydrology, there may be no detectable change in air temperature.
The energy will heat something up. So come up with all the negative feedbacks you can imagine. It doesn’t matter, we can’t think of them all. Now, lets gather all the theories and plot the predicted temperature sensitivity. There’s a range from half a degree C per doubling on up, right? They center on about 3 deg. C per doubling. Until there’s something more complete, it’s not unreasonable to believe the truth lies somewhere in the middle. In addition, we know the models have yet to take slow feedbacks into account, including the release of carbon from the arctic as it warms.
At the end, you have no evidence and no valid theoretical base. You can believe in AGW as you like, but don’t delude yourself that you’re being rational.
The way I read Lindzen, science agrees on the basics and has questions about the degree. So I look at the range of predicted outcomes, trim the ends and look in the middle. If you want to believe it’s on the low end, fine, but in the absence of facts, that’s wishful thinking.
A 4+ degree swing in global average temperature must also have an effect on absolute humidity, resulting in additional amplification.
As far as ghg’s driving warming rather than lagging, the evidence from the PETM is pretty convincing to me. A sudden massive injection of methane is a quite reasonable hypothesis.
To follow up on the “greening up” point, I found more recent information (Aug. 2010) on carbon flux in the biosphere. Satellite data from 80’s and 90’s showed increases in net primary production in the biosphere, and a slight decrease in the 10 years from 2000 to 2009. NASA did one of their videos summarizing the findings. http://www.youtube.com/watch?v=xjsvL23Sw9Q
#108, Marc:
1. In a feedback, the lagged observable is the driven system. CO2 lagged the rise in T and it lagged the fall. There’s no way CO2 could have been the driver of T, nor could it have been the maintainer.
2. When the mechanism is an assumption of the theory, finding that mechanism in the theoretical output is called circular reasoning. You continue to fall for it.
3. No one must predict a negative. You’ve merely settled on the explanation you like, but one with no deduction from independent theory.
4. That one caused a laugh, Marc. Facts are evidence. In science at any rate. I understand in your world facts are apparently inconvenient and in need of explaining away.
5. The models make no reliable predictions. Therefore, they are non-scientific. The CO2 hypothesis they make is also, therefore, non-scientific. You are crediting a non-scientific hypothesis and expect everyone else to take that seriously.
There is no evidence that the climate of Earth is behaving in any way anomalously, after a 30% increase in atmospheric CO2. You come along with a hypothesis equivalent to a warning that god will punish us for our sins and that the end is near. We are all then required to don sack-cloth and ashes to make you happy. Dream on…
6. Even perfect logic is subject to the assumptions that precede it, Marc. Your assumptions are wrong, and so your impeccable logic of correlation = causation led inevitably to a wrong conclusion.
7. The truth doesn’t lie between two poles of unsubstantiated reasoning. In this case, your logic is fatally deficient. Earth climate has been warmer than now, and there was no catastrophe. Your inchoate personal fears are not predictive.
8. In the absence of facts and theory, and in the presence of no evident problem at all, your position is entirely fanciful.
And on the basis of fanciful imaginings like yours, science has been corrosively subverted, scientific bodies and journals have become corrupt, character assassination has assumed the patina of high morality, careers have been distorted, huge monies have been squandered, carbon market frauds have siphoned off billion$, and solar and wind farms excavate their profits from tax subsidies, all driven by misled, scientifically incompetent, but well-meaning bureaucrats and politicians.
All caused because of people who have your fanciful fears and coupled ideological longings, feel their higher cause justifies whatever it takes.
Pat, I’m not making myself clear about the logic problem, so I’ll remove all the carbon and lame humor and try again.
Posit an imaginary system in which A is a function of 3 variables: B, C, and D. Furthermore, C and D vary as a function of A.
In such an imaginary system, a change in any of the variables (B,C,D) results in a change in A. This, in turn, causes C and D to change. We’ve got a feedback loop in other words.
In this system, a change in B can trigger changes in A, and consequently C and D. That in turn causes further changes to A.
If such a model is possible, it would be illogical to argue that changes in D, when triggered by B, cannot cannot cause a further change in A.
Where does my logic break down?
I’d add to the above that such a mechanism could explain how one of the factors influencing temperature could follow it both up and down. Again, it’s an imaginary model used to validate logic, not to prove the world is coming to an end.
<a href="https://noconsensus.wordpress.com/2011/01/20/what-evidence-for-unprecedented-warming/#comment-45742"#112, your logic is a denovo construction, Marc. It has no grounding in the model of CO2 and climate.
The logic of AGW is that CO2 forcing directly influences temperature. Your new logic has it hidden behind a subsequent chain of downstream events.
So, suppose you explicate your new theory of the CO@ climate driver, by revealing the intermediate climate systems it influences, that go on to eventually produce an increase in air temperature by way of a non-CO2 proximate cause.
Be sure and decorate your theory with some rigorously derived predictions, and present the observational evidence.
Otherwise, you’re just spinning convenient air-bound complexities to save your opinion.
#113, we’re not talking logic, Marc. Logic is merely a tool. We’re talking science.
Nothing new, I assure you. I tried to illustrate a Milankovitch-led warming / cooling episode in overly simplistic terms with temperature (A), solar forcing (B), albedo (c), and co2 concentration (D) to continue an element of our discussion. I suspect you’re smart enough to see where I was going.
We’re not talking logic, Pat. It’s too bad too, but it was still worth the effort. I enjoyed it and learned a thing or two. Again, thanks for sharing your thoughts.
We’re not talking logic. Logic is merely a tool. We’re talking science. Pat Frank
To Jeff and anyone else who may have followed my exchange with Pat:
While we didn’t get anywhere on global warming, it been an instructive exercise in critical thinking. I challenged Pat on his logic and we argued the point back and forth, never reaching agreement. The answer to who’s thinking critically lies in the answer a simple question. If A causes B, can B cause A? I’ll walk you back through it.
Back in #65, Pat said I didn’t have to worry about increasing co2 because warming has always followed temperature increases. We have ice-core data showing that for the last million years CO2 always followed air temperature, and never drove air temperature.
In #67, I wrote, While temperature may lead, carbon dioxide and temperature have always danced together. That C02 lags temperature is an excellent talking point if you want to mislead someone who can’t handle credit card debt, but it’s lame from a physical standpoint.
The reference to credit card debt refers to the fact that interest causes a balance to increase even though interest lags the initial charge by 30 days. He either missed that or doesn’t handle credit well, so I went on to try to convince Pat of my position through the use of logic.
In #69, Kim jumped in. The arrow of time would suggest that temperature causes CO2 change and not the other way around. Perhaps we are testing something that has not happened before, but do note that CO2 trails temps on the way down, too. CO2 has not been able stop a cooling. That’s all well and good, but does not preclude a different explanation based on feedbacks.
I responded to Kim in #71. Yes, Kim, we all know that warming causes the oceans to release co2 into the atmosphere. And we all know about feedback, right? You may pack your arrows of time back in your quiver and shoot them at someone else.
Very simply, the theory is that warming (for any reason) causes oceans to warm, warmer ocean releases co2 into atmosphere, increased co2 in atmosphere enhances greenhouse effect.
Pat jumped back in at #81. Temperature led CO2 by an average of 800 years through all seven ice ages spanning nearly 1 million years. That’s not a “talking point,” that’s a fact. Nor is it “lame from a physical standpoint.” As a physical fact, it refutes the idea that CO2 is a primary driver of air temperature. As Kim pointed out in #69, seven times global temperatures fell into ice ages while atmospheric CO2 was high. It’s only by special pleading that people argue today’s CO2 is different. That’s not science.
You can’t mistake what Pat’s saying: C02 can’t drive temperature because it came second. You can set that up as B (co2) cannot cause A (temp. change), because A caused B first. Logically speaking, that is not a valid argument.
As Pat will state later, logic is just a tool. He’s right. It’s a tool that helps us evaluate statements and think critically. If an statement fails on logical grounds, it’s not critical thinking. The stakes are getting high now, as I’m accused regularly of not being rational.
Somewhere in there, I said that raising grain prices contradicted a separate argument about the biosphere. DeWitt called me on it, I thought about it, and confessed I was wrong. I also admitted that I was wrong about the flux of carbon in the biosphere.
I replied to Pat in #88. I’m surprised that the co2 lag issue is troublesome. Yes, you’re correct about the 800 year delay, but it’s still lame physics. Assuming for a moment that co2 is a greenhouse gas, it’s a trivial matter to construct an amplifying (or dampening) feedback mechanism. As intuitively appealing as it may be to gullible thinkers, the fact that co2 lags temperature proves only that something else got the ball rolling. You can argue that co2 is not a greenhouse gas if you like, but not with that logic.
Pat responded in #99. …CO2 trailed both rising temperature and falling temperature seven times in sequential ice ages and their recoveries. That’s 14 independent instances of lag. High CO2 did not prevent falling temperatures seven times in a row. That’s not “lame physics” no matter how often you might claim so.
Pretty much the same argument.
In #103, DeWitt supported my position with a model that demonstrates that the postulated feedback mechanism works. It does not prove anything on its own, but it shows I’m on solid ground and Pat isn’t.
I replied to Pat in #108. Lag is characteristic of feedback. DeWitt has demonstrated a model that works: Co2 can cause warming even though warming causes co2. I’ve never claimed that this proves the theory — it doesn’t — only that it’s invalid to state that causality precludes effect.
Pat sticks to his guns in #111. In a feedback, the lagged observable is the driven system. CO2 lagged the rise in T and it lagged the fall. There’s no way CO2 could have been the driver of T, nor could it have been the maintainer.
C02 could easily have driven temp if temperature is sufficiently sensitive to CO2. That’s trivial! The question is not which came first, the question is, what’s the sensitivity? I tried once more in #112 to illustrate my point by building a simplified symbolic model of a feedback system that would rise and fall with a lag. I asked Pat to show me where the logic broke down.
When you’re about to lose a game of chess, one option is to throw the board off the table. In #114, Pat responded. …your logic is a denovo construction, Marc. It has no grounding in the model of CO2 and climate. The logic of AGW is that CO2 forcing directly influences temperature. Your new logic has it hidden behind a subsequent chain of downstream events.
He would not show me where my logic failed because, I believe, he could not. If he can see that I’m right on the logic, he could reconsider his position or throw a tantrum. I’d call that a little tantrum.
Doesn’t matter how I see it though. If you’re able to think critically, you can see for yourself who is “spinning convenient air-bound complexities to save their opinion.” And if you’re not able to think critically, you’re only able to follow and evaluate science through intuition. Your intuition will inevitable lead you to conclude Pat’s right because there is nothing intuitive about a trace gas driving catastrophic climate change.
Finally, Pat said, we’re not talking logic, Marc. Logic is merely a tool. We’re talking science.
What is science without logic? It’s intuition. If you’re not going to use logic, don’t expect me to listen.
I haven’t read the exchange but the problem with CO2 lagging temp in ice cores is the magnitude of the lag. ~ 700 years if I recall. We’re now claiming a comparatively near instantaneous response with models. I have always been unconvinced with the hand waiving argument that CO2 made it worse. It looks much more like CO2 had little/nothing to do with it and in addition the evil warming tipping points simply don’t exist. IMO nobody has made an argument which has been very convincing in associating ice core temps to CO2.
Marc, you have incorrectly cited my #69. Nor have you responded to what I did say.
I’ve begun to think you are hopeless. Like many true believers, you have a good heart. Google ‘Livingston and Penn’ and consider the consequences of a Grand Solar Minimum. For the children.
===================
My apologies Kim. I misplaced the “cite” tag, which I can see now should have been an “i” tag. The paragraph should read:
—————
In #69, Kim jumped in. The arrow of time would suggest that temperature causes CO2 change and not the other way around. Perhaps we are testing something that has not happened before, but do note that CO2 trails temps on the way down, too. CO2 has not been able stop a cooling.
That’s all well and good, but does not preclude a different explanation based on feedbacks.
—————
As to your other points, I chose not to respond in order to stay focused. I’m curious about your hypothesis, but I’ve already hijacked a thread about Pat’s research and didn’t want to further damage. Follow up with Pat is fair, going off on a tangent with you wouldn’t be.
I haven’t read the exchange but the problem with CO2 lagging temp in ice cores is the magnitude of the lag. ~ 700 years if I recall. We’re now claiming a comparatively near instantaneous response with models.
Well, you’re talking about two different effects.
First case: temperatures rise first due to an external forcing (presumably involving the Sun). CO2 increases in response.
Second case: CO2 rises first, due to naked apes burning stuff. Temperatures increase in response.
There’s no obvious reason why these two phenomena (CO2 growth from independent warming, and warming from independent CO2 growth) should occur on the same timescales. One involves biology and geology, the other is mostly radiative physics.
Your logic in 112 isn’t really “logic.” I suggest you read up on formal logical arguments for clarification. I also suggest you research how feedback works and how physical systems operate. Your interpretation of what you have created in your argument is incorrect.
Mark
Kim, I’d love to be wrong about this. I’ll joyfully concede that I’m wrong, hopeless even, when someone demonstrates the flaw in my logic. Surely, if I’m as irrational as Pat says, that will be easy.
My underlying point is that much objection to AGW theory is based on intuition. Intuition is often right; I generally trust mine. But intuition is not critical thinking and it sometimes fails us. The earth is not flat, the sun does not revolve around the earth, and a heavier object does not fall faster than a lighter one if air resistance is the same. Though it makes no intuitive sense, a trace gas can drive temperature, co2 can amplify warming even though it lags it, and more snow can fall in a warmer climate.
I checked Livingston and Penn. Very interesting, thanks. We’re coming out of a deep solar minimum. My intuitive response is that we would expect to see a cooling trend over the last decade. Since 2010 was reported to be the hottest year since 1860, the cooling effect doesn’t seem to be a big problem. Indeed, I suspect the cooling is masking warming from other sources. As sunspots increase over the next few years, I would expect to see even higher temperatures as the masking effect wears off.
122.Marc said
February 1, 2011 at 2:03 pm
Again, I suggest you make an attempt to better understand the basic “logic” of your argument here. Feedback control theory and the principle of causality, in particular. Neither is “intuitive,” which is obvious in your responses.
Clearly you do not understand.
Mark
MarkT, you’ve never heard of parametric forcing?
Parametric forcing is an apparent lag from effect to cause, not a real one. That’s going to be difficult to demonstrate with the connection Marc is attempting to explain, particularly since the lag has “switched” (in theory) to simultaneous occurence.
His “logic” in post 112 is actually a circular argument, which is fine for some mathematical proofs, but otherwise a fallacy. If he is so bent on using logic, he should apply it to his own arguments.
Mark
I haven’t read the exchange but the problem with CO2 lagging temp in ice cores is the magnitude of the lag. ~ 700 years if I recall. We’re now claiming a comparatively near instantaneous response with models.
Well, you’re talking about two different processes here.
1- The growth in CO2 resulting from temperature increases.
2- The warming from CO2 growth.
There’s no obvious reason why these two different phenomena should have similar timescales. The former involves biology and geology, the latter is mostly radiative physics.
MarkT, I wasn’t trying to defend Marc or deal with his rhetoric…. the point being that you can have CO2 act as a feedback (consistent with the delay between CO2 response and temperature change) in some cases, and in others, like when CO2 get released via e.g. eruptions or anthropogenic activity, as a parametric forcing .
Saying that it’s one does not exclude it from being the other.
Perhaps this was Marc’s point and he just stated it less than perfectly.
I try to leave and they drag me back in. Marc, consider ‘Livingston and Penn’ a little more. Your mentioning ‘We’re coming out of a deep solar minimum’, by which I presume you mean the recent solar cycle, means you do not understand the meaning of Livingston and Penn’s observations.
Well, no one knows the meaning for sure, but your discussion shows you haven’t even considered it. Do you remember the delaying hazards in touring? I could name a couple, but, after you.
If L&P’s observations persist, the sunspots this cycle or next will become invisible. This is not to say the sun’s dynamo weakens, only that the ‘sunspots’ will leave the visible spectrum, so-called Cheshire Cat Sunspots. The sunspot activity is there, our vision of it is not.
Now, during the last several grand or lesser solar minimums, sunspots probably did the same thing. For instance, during the Maunder Minimum, sunspots became sparse, large, and primarily southern hemispheric. It also cooled during the Maunder and other minimums. However, there was also vulcanism with its attendant albedo changes noted during some of those times, so causality from solar minimum to global cooling is unproven. I believe that isotope data is ambiguous.
And yet, the sunspots move. If their migration out of our visible spectrum does correlate with global cooling then I think that is probably because both phenomena have a common cause. Wouldn’t that be something to know?
I believe the aliquot of anthropogenic CO2 injected will trap more energy on earth. I hope it’s enough.
Honor Gaia and bow to Sol.
==============
Heh, I shot an arrow into the sky, it fell to earth smack into the casing.
Marc, I think you are responding with a machine you don’t quite understand. Go listen to ‘The Sorceror’s Apprentice’.
==========
Toto,
If CO2 goes up dramatically, you would expect an immediate response from temp. While you can make the case that the temp rise could start for other reasons and CO2 should lag, if CO2 is a heavy factor, the peak of the CO2 rise should coincide with the peak of the temp rise. It seems impossible that temps in between ice ages would flatten at a maximum centuries before CO2 would flatten. Think about the meaning of that particular point with CO2 being a strong driver of temp. I’ve considered working some serious math on it myself because it looks to me that this would be a very unlikely circumstance were CO2 really amplifying the temp rise.
Actually, if my comment in 130 makes sense, it may be some of the strongest evidence for error in the magnitude assumptions.
Why couldn’t the counterargument be made that CO2 drives the temperature down?
===============
Jeff, re: 130. Do you think the ice cores are good enough to yield to your maths?
==================
I have very skeptical opinions of the ice core meaning but if we assume they are truly temp and CO2 and AGW is correct, the CO2 curve should not flatten after the temp curve. Temp should continue to rise until CO2 flattens. It’s a subtlety which should be hit us in the head obvious.
Sounds like a good problem to work on Jeff. I pondered how the system might attain equilibrium as well. Have you been through DeWitt’s post? (See #103)
Mark, I’m sorry I’m not making myself clear. I’m not sure what’s missing, so I accept your suggestion to review feedback theory and formal logic.
The argument here is that since co2 increased after warming occurred (during the ice age), co2 is not the cause of warming (then or now). Expressed in a syllogism, that looks something like this:
– During the ice age, co2 followed temperature change.
– An effect must follow the cause.
– CO2 must not (ever) be the cause of warming.
The conclusion is false because it deducts a general rule from a specific event. Ok, 14 specific events, but still, the premise is not sufficient. In #112, I complicated things by trying to demonstrate a feedback mechanism that, if co2 is strong enough, contradicts the general rule. I considered the circularity involved, but I’m falsifying the hypothesis and get to stipulate the condition.
Just a thought: if co2 does not cause warming, one might reasonably expect to see co2 levels precede temperature rise at some point during the ice age.
#127, Carrick, yes, that’s what I’m trying to say. Thanks for stating it more clearly.
And Kim, it’s all very interesting. You’re right, by understanding of their research is very superficial. I’ll come back to it another time.
Marc,
Of course I’ve been through DeWitt’s post, having posted it myself. If you look at the offset/lag he found and attributed to ocean lag for CO2 release it makes a strong case that the models are not matching planetary response. If you look at the point when temps peak and the fact that CO2 takes some time to catch up whereas temps don’t change at all, you have to wonder why that bit of CO2 has no effect whatsoever.
Also IMO, the lag is potentially more due to plant/animal life expansion than ocean but that is also conjecture.
Jeff said;
“I have very skeptical opinions of the ice core meaning but if we assume they are truly temp and CO2 and AGW is correct, the CO2 curve should not flatten after the temp curve. Temp should continue to rise until CO2 flattens. It’s a subtlety which should be hit us in the head obvious.”
I’ve got a brilliant idea-why don’t you set up a blog somewhere then you and others can comment on this sirt of subject? 😉
tonyb
Steig on O’Donnell @ RC.
H/t AngusPangus.
=========
Carrick said
February 1, 2011 at 4:27 pm
Possibly, and I understood what you were getting at. The actual functional relationship changes as well in such cases, something that, while not impossible, has not been demonstrated – recall that climate scientists also push the principle of uniformitarianism (a poor application of the principle, IMO.)
The climate science community is in a bind with such a description: it is highly non-linear making it difficult to assess, yet they are describing it with simple feedback relationships. It’s hard enough to evaluate closed-loop feedback (read: next to impossible) in simple systems without knowledge of the open-loop response. If the poles and zeros are moving around due to non-linearities (which could still be described using linear techniques,) or all of the feedbacks/feedforwards combine non-linearly (requiring non-linear analysis techniques,) the problem becomes ridiculously hard to determine. Claims one way or the other are simply guesses at this point, not even educated ones.
Another possibility, of course, is that there is a third agent causing both, which would imply no cause-effect relationship between them.
I wasn’t really saying that even if it came across that way, just that his “logic” was circular. A->B->C->A sort of thing.
Part of the problem with comparing now (CO2, temp, etc.) to past estimates is that now has very fine resolution in the data. Past estimates have, at best, yearly resolution (tree rings, not that they are worth what alarmists claim,) and at worst, decadal resolution (ice core data, also frought with problems.) When you have a core that has 100 years of history “recorded” in a single band, you have no idea what it did during that time, just what the average/integral (minus losses) looked like. Peaks and valleys get smoothed out (it is akin to running the actual data through a low-pass filter of some sort.) The same goes for tree rings, which are even worse because they only grow for a few months out of the year. This is essentially a sampling problem (Nyquist is worth researching, too.)
That’s not to say we have not added to the CO2 in the atmosphere, just that we don’t really have anything to compare it to historically. The point being, of course, is that we do not know what the short-term “wiggles” looked like in the past. There may have been short-term correlations similar to what we are seeing now even though the long-term correlation is ocean warming->CO2 increase.
Marc said:
Note my comment about sampling above. It may have, we just don’t know due to the resolution in the data. Big problem if anyone is trying to make definitive claims regarding short-term relationships.
Indeed. If CO2 does indeed cause a rise in temperatures, there should be some impact at all levels, though maybe the resolution or non-linearity masks the impact. I have not seen a sufficient exposition to formulate an opinion (beyond the null.)
Mark
Uh, the first two quotes were from Carrick, the third and fifth from Jeff, and the fourth from Marc.
I’m violating Nyquist with my quotes. 😉
Mark
With reservations due to resolution, temperature stabilization, etc., there seems to be implicit agreement that the relationship of co2 to temperature, while it proves nothing, is consistent with AGW theory.
141.Marc said
February 2, 2011 at 2:38 pm
Um, no.
Well, unless you define “everything” as “consistent with AGW theory.” There’s a point where a theory becomes laughably not very theoretical when there is nothing that is NOT consistent with the theory. We have reached that point.
Mark
Mark T;
Ya, we’ve been demanding and begging for a list, however short, of potential observations that would be inconsistent with AGW.
So far, no luck.
Of course not.
Even the pipe that just burst in my basement (outside faucet froze) is “consistent with global warming.” Below zero here in CO Springs going on 48 hours now and not supposed to let up till sometime tomorrow morning. Hopefully my basement does not turn into a skating rink in the mean time (it’s heated, I’m just being facetious out of frustration – I’m on a waiting list for the plumber.)
Mark
I would suggest that in ice cores, the flatlining of temp a thousandish years before the flatlining of CO2 is very much inconsistent with CO2 driven temp.
145 – too simple! 🙂 Happy retirement – thanks again for tAV. Best to you and the family. C.
That’s a valid point, Jeff, and something to look at. Technically speaking that’s overshoot, though, not lag.
Mark, the theory was first postulated by Svante Arrhenius. With stunning foresight, he wrote in 1906 that:
…any doubling of the percentage of carbon dioxide in the air would raise the temperature of the earth’s surface by 4°; and if the carbon dioxide were increased fourfold, the temperature would rise by 8°. Although the sea, by absorbing carbonic acid, acts as a regulator of huge capacity,… the slight percentage of carbonic acid in the atmosphere may by the advances of industry be changed to a noticeable degree in the course of a few centuries. Disruption of normal patterns of the winds will drop the temperature to -17 degrees at 9º N, 105º W on Feb. 1, 2011.
So, yeah, that AGW.
Marc;
HAR-HAR!
-17° at 9°N, 105°W? Yesterday? That’s in the tropical Pacific west of Latin America. Even without knowing what scale he was using, I can guarantee that didn’t happen.
And he later recanted on the 4°/doubling, IIRC. Regardless of scale, that’s way too high.
Whoops. That was supposed to be 39° N. Celsius scale. And sure, 4 C is on the high end.
Jeff ID:
I’d encourage you (in your enormous free time) to look at the actual lag… it isn’t always 1000 years, and sometimes is even negative.
Marc;
That’s Colorado Springs, in the mountains. Zero points.
And +4°C is on the high end by about an order of magnitude. And the sign is probably wrong.
Brian,
That’s Colorado Springs. Zero degrees. Get it? The theory is not only consistent with Mark’s frozen pipes, it predicted it!
#117
Marc, your comment about interest and bank balances obviously misconstrued the situation with climate, which is why I didn’t bother to think about it. Interest doesn’t cause a credit card balance to increase. Interest is merely the observable. The cause is the credit policy of the institution. To make your analogy properly parallel, policy = CO2 and interest = the increment of temperature. The charge analogizes the base temperature. So, even in your analogy, the cause precedes the effect and your own analogy defeats your point.
You wrote this, regarding the ice age – CO2 relationship: “Very simply, the theory is that warming (for any reason) causes oceans to warm, warmer ocean releases co2 into atmosphere, increased co2 in atmosphere enhances greenhouse effect.”
But that’s only half your story, isn’t it, because you’re only looking at the rising temperature side of the trend. There’s the falling side to consider, as well, which you left out.
Applying your logic, when the temperature drops, with CO2 at a high level, your thinking requires that ‘high CO2 diminishes the greenhouse effect‘ and causes the temperatures to drop.
Your logic doesn’t terminate when the conclusion it engenders becomes inconvenient.
You wrote: “You can’t mistake what Pat’s saying: C02 can’t drive temperature because it came second. You can set that up as B (co2) cannot cause A (temp. change), because A caused B first. Logically speaking, that is not a valid argument.”
I didn’t write that temperature caused CO2 to rise. I wrote only of lags. If CO2 invariably lags temperature rise, and fall, there is no reason to assign CO2 as causal. Likewise, to suppose that CO2 enhanced warming, without any evidence to support that relation, is merely special pleading.
It’s peculiar that you’d be so wedded to logic for others all the while claiming a logical exemption for your own argument.
You wrote, “The stakes are getting high now, as I’m accused regularly of not being rational.
And I’m accused of being illogical. There has been no ‘regular accusation’ of irrationality, at least by me. In #99, I suggested you shouldn’t delude yourself that you’re “being rational” by holding on to the AGW idea. That’s not an accusation of irrationality. You’ve misrepresented my words before, I’m sure innocently, but I think you should be more careful.
You restated, “the fact that co2 lags temperature proves only that something else got the ball rolling. You can argue that co2 is not a greenhouse gas if you like, but not with that logic.”
No one has written that CO2 is not a greenhouse gas (although “greenhouse” is an obvious misnomer, we’ll let that pass), and so you’re tendentiously misrepresenting the case again. The question is whether CO2 is a primary air temperature driver. There is no evidence that is the case, and it will remain no evidence no matter how often you try and paper over the issue with diversionary malconstructions.
You wrote, “DeWitt supported my position with a model that demonstrates that the postulated feedback mechanism works. It does not prove anything on its own, but it shows I’m on solid ground and Pat isn’t.”
Notice Sky’s comment.
Anyone can come up with any mechanism they like, building it in a way that it gives an already known result. A mathematician friend of mine used to point out, when we were engaged online with creationists, that he could readily produce a mathematical theory for a causal mechanism of human progress by the influence of angels. All it takes is a set of equations written to produce a snooped result and include assumed physico-angelical properties. QED your logic, Marc.
You’re on no solid ground. You’re on no ground at all, despite your inner certainties and your special pleading. There isn’t any physical evidence for your view, and there isn’t any workable physical theory, either. Tendentious logic notwithstanding.
You wrote: “When you’re about to lose a game of chess, one option is to throw the board off the table. In #114, Pat responded. …your logic is a denovo construction, Marc. It has no grounding in the model of CO2 and climate. The logic of AGW is that CO2 forcing directly influences temperature. Your new logic has it hidden behind a subsequent chain of downstream events.
“He would not show me where my logic failed because, I believe, he could not. If he can see that I’m right on the logic, he could reconsider his position or throw a tantrum. I’d call that a little tantrum.”
Call it what you like, Marc. The fact remains that you can propose self-affirming misconstruals as you like, but they don’t prove anything about climate physics.
And I did show how your logic failed. You changed the rules (denovo construction) — another method of ending a game you’re losing. Your original game was stock AGW: CO2 drives air temperature. When that failed, you changed to CO2 drives C drives D, changes A, which drives CO2. Voila, you have evaded the fact that ice core data shows CO2 does not drive temperature. To give everything a little push, you also rang in a new something-or-other “,em>to get the ball rolling.”
Ad hoc theorizing, Marc. Saving the hypothesis. Let’s also notice that your ABCD amounts to a runaway positive feedback loop. Once started, your analogizing ad hoc theorizing would heat-sterilize Earth’s surface. Maybe you can invent another ball to get things stopped.
You wrote: “Your intuition will inevitable lead you to conclude Pat’s right because there is nothing intuitive about a trace gas driving catastrophic climate change.”
Where is the need for intuition to recognize that climate models are unreliable and that there’s no evidence CO2 drives air temperature? Your entire riposte is merely empty polemics.
You wrote: “Finally, Pat said, we’re not talking logic, Marc. Logic is merely a tool. We’re talking science.
“What is science without logic? It’s intuition. If you’re not going to use logic, don’t expect me to listen.”
Science does not begin and end with logic. I’ve already pointed out that using logic on false premises yields false conclusions. Your arguments are a case in point. Virtually none of them are based in science. They’re all based in some special argument of your own, where one need only make some internally consistent picture to produce a scientifically correct result. If that were true, we’d all be using N-ray imaging and infinitely dilute homeopathic medicines.
For some reason, you have an insistent death grip on AGW despite a complete lack of evidence. You’ve shifted your argument to keep your grip, and call that logic. I don’t. And science, it certainly never will be.
It’s not logic that prevents science from being intuition. It’s the empirical tests of falsifiable theory that do so. Logic alone can happily be a handmaiden of intuition. The content of Theology fully demonstrates that.
#135
“The argument here is that since co2 increased after warming occurred (during the ice age), co2 is not the cause of warming (then or now). Expressed in a syllogism, that looks something like this:
“- During the ice age, co2 followed temperature change.
“- An effect must follow the cause.
“- CO2 must not (ever) be the cause of warming.”
A tendentiously false syllogism. The correct syllogism is:
– During the ice age, change in ppmv CO2 always followed air temperature change.
– An effect must follow the cause.
– CO2 rise did not cause the air temperature change after each ice age.
That syllogism, brought into science and coupled with the lack of a predictive theory of climate, renders unlikely any assertion that CO2 drives air temperature.
I would suggest that in ice cores, the flatlining of temp a thousandish years before the flatlining of CO2 is very much inconsistent with CO2 driven temp.
Not that I’m fond of using WUWT posts as sources, but did you see that:
http://wattsupwiththat.com/2009/01/30/co2-temperatures-and-ice-ages/
Apparently the peaks in CO2 and do not seem to lag appreciably from each other (assuming the poster did things right – he doesn’t seem to realise that his own graphs support the standard intepretation of CO2 as a heat-retaining “blanket”).
Perhaps someone has better data though?
147.Marc said
February 2, 2011 at 11:38 pm
That’s a valid point, Jeff, and something to look at. Technically speaking that’s overshoot, though, not lag.
You aren’t seriously citing Arrhenius, are you? The same guy that later decided his own work was incorrect? Puhleeze. You should use the pseudo sarc tags when posting these [sarc][/sarc]. 😉
It was -17 here, btw. Bitter, bitter cold. The plumber will be here shortly. Then the adjuster, then the general contractor to replace my basement bedroom and, potentially, carpet.
150.Carrick said
February 3, 2011 at 1:13 am
In any correlation calculation you’ll always find short periods in which the apparent direction reverses unless the correlation is exactly one, i.e., perfect correlation.
Mark
Oh, and we keep overheating the steam cleaner we’re using to suck up the water from the carpet. Sucks, er, rather, it doesn’t suck continuously. Must be powered by wind.
Mark
MarkT:
In “any correlation”???
LOL. “In any correlation???”
What me to post data? I have plenty of data (atmospheric measurements too) with a correlation that is less than one, but where the distribution never flip signs (nor gets close to zero for that matter).
You’re correct to suggest whether the the negative correlation is due to measurement error, it might be, but that’s the point of doing a study before making a decision.
But while we’re at it, how about delivering one example where the situation you claim occurs, please, where cause and effect have been previously established.
159.Carrick said
February 3, 2011 at 12:06 pm
MarkT:
Constrain your period, Carrick. You’re comparing apples and oranges – the distribution does not have to flip signs, just a short period over which you calculate your correlation. Two different concepts.
By definition, random noise will have periods of positive, negative, and zero correlation. Adding that noise to an underlying “signal” does not zero out such a possibility, it just makes such periods less and less likely (depending upon signal to noise ratio, not that either can be determined in this case, but there is something that can be called signal and something that can be called “random noise.”)
Cause and effect have not been established. Nor have I said so. Correlation is not causation and all that.
As an example, however, generate two long random sequences drawn from a Gaussian distribution. Look at the correlation over short periods across the length of the sequences. Some will be positive, some will be negative, and some will be very near zero (technically, absolutely zero should have a zero probability but due to quantization, even with floating point numbers, there is a small probability of a zero correlation.)
You can do the same with two contrived signals, one a function of the other, then add random noise to each (i.i.d. is not really an issue since we aren’t testing anything dependent upon it.) The larger the signal level compared to the noise level the shorter your periods will need to be to demonstrate this, but you can always increase the length of the signal sufficiently to find periods in which the correlation apparently flips.
Mark
Mark said:
I should point out that IF there is a causal relationship then the lead “signal” wins out as the cause over the lag “signal.” However, just because there is a high correlation does not mean the two data sets are related cause-effect to each other; there could be a third agent causing both.
Mark
Of course, if there is zero noise this wouldn’t apply, so my use of “any” is incorrect in that sense, but we’re not talking about zero noise cases since they are unphysical (pulsars and atomic clocks come close, but they still have at least SOME noise.)
Mark
Let me start by agreeing with Mark that correlation is not causation, and trying to argue causation on simply that ground (one way or the other) is specious.
It’s tough in a blog where you can’t post mathematical equations to have really precise mathematical langauge usage, but what we should be discussing is the cross-correlation Cxy(t, tau), where t is the center of a window of data being analyzed, and tau is the offset of the y data set relative to that window. For sake of discussion let’s assume that x(t) and y(t) are both broad-band signals (e.g., x(t) is a pulse or swept tone or broad band continuous noise).
From a system viewpoint, at the response of measured signal y(t) to a second measured signal x(t). In all cases I can think of where we would do this, we “know” up front that x(t) is the stimulus and y(t) is the response to that stimulus.
Secondly, when we talk about the lag in the correlation, we usually mean the value of tau that maximizes Cxy(t,tau) at that time t. (Sometimes it is |Cxy(t,tau)|… some processes can flip the sign of the correlation between the cause x and the effect y.)
In the “real world” you almost always have nonzero values of Cxy(t,tau) for tau 0 for the maximum Cxy only applies in signals which have decent signal to noise and for which the underlying process relating y to x is linear. In nonlinear systems it’s definitely possible to find values of tau < 0 where Cxy is non-zero and statistically significant. I can post data if anybody wants to see it.
Climate is both noisy and nonlinear so trying to explore "cause and effect" using correlation by itself is certainly worrisome.
However, we do have some facts on our side here in resolving this:
1) We know that a change in temperature will yield a change in atmospheric CO2 content. First there is the dissolution of CO2 from seawater (higher temperatures reduce the solubility of CO2 in water).
2) Secondly, we know the biosphere responds to changes in temperature, but also to other ecological effects such as precipitation or even things like forest fires. Since the biosphere changes can affect the balance of emitted and absorbed CO2. So the biosphere itself can act as a net source or sink of CO2, irrespective
This carbon cycle diagram is rather interesting in that respect. Roughly 200 gigatons of CO2 (GTC) are released and absorbed by land/ocean/atmospheric processes. (Humans account for maybe a net 8 GTC in the most recent data.)
3) Of course natural events like volcanic eruptions can also release large amounts of CO2, these would be a net source for CO2.
4) Controversial to some, but for most of us it is fact, CO2 acts as an atmospheric greenhouse gas: Changes in CO2 concentration affect the net radiation balance for the Earth.
Given these four facts, if you have CO2 concentration change in response to temperature, it will act as a feedback.
If you have CO2 atmospheric concentration changes driven by ecological or geological events, that will change the net radiation balance of the system, and will act as a parametric forcing. It’s easy enough in 1-d models to see that CO2 acts parametrically: In this case, the effect of the change in CO2 concentration directly multiplies the solar irradiance S(t) just as (1-abeldo_coffcient), the net transmittance to the surface, does.
The way I would put it, cause and effect are established not by looking at the correlation Cxy(t,tau) over time, but rather by looking at the underlying physical and biological processes. Periods where the maximum correlation occurs at negative taus might correspond to periods of ecological change, to volcanism or just to measurement error. On the other hand, trying to start with the fact you have a positive (or negative) value of tau, without looking at the4 underlying system relating x to y, is a log like the tail wagging the dog.
One more technicality, then I’m through for today…sorry work demands.
It is often useful to construct the analytic signal x(t) + j xh(t) , where xh(t) is the Hilbert transform of x(t) and then correlate that analytic signal with the real-valued y(t). In this case you’d definitely be looking at |Cxy(t,tau)| over time.
The reason for this is there are processes which can result in a 90° phase shift in the response. You would see zero-real valued correlation for those cases.
(I’ll mention in practice for signal processing aficionados that you don’t have to actually compute xh(t) first. Simply zero the negative frequencies and multiply the convolution integral by two before inverse transforming, and you’ll get the complex-valued correlation function).
Pat,
I stand corrected on the instances of imprecision.
The context here is whether or not increasing co2 concentration in the atmosphere is cause for alarm. In this context, you said: “We have ice-core data showing that for the last million years CO2 always followed air temperature, and never drove air temperature.”
From this statement I inferred this argument: since co2 never (in the last million years) drove air temperature, c02 never drives air temperature. This inference seemed reasonable as it was offered to support your position that I have no reason to be alarmed about the rise in co2. After all, I’m not worried about ice age warming.
I think you will agree that one can’t deduce the conclusion (co2 does not drive air temperature) from the inferred premise (co2 did not drive air temperature during the ice ages). First, the argument draws a general rule from a specific case.
Second, we can suppose conditions in which the premise is true, but the conclusion is not. If we suppose that something else, for example orbital variations, “got the ball rolling” and caused an increase in co2 that in turn drove additional warming, then the conclusion would be false, while the premise still holds.
The argument fails on logic, yet you maintain your position is logical. You must have been arguing something else. So lets be clear: how do the ice core temperature / co2 reconstructions inform the current climate change debate?
I don’t disagree with any of that, Carrick, at least, not from the quick scan on my phone.
I was merely pointing out that you can find any arbitrary correlation between data corrupted by noise. The supposed 800 year lag is over a fairly large period and the correlation is rather large (I do not recall exactly how large) which decreases the likelihood it is do to chance. Unfortunately, this speaks nnothing of causality, particularly when non-linearities are present (feedback is hard enough.)
Mark
Mark T,
Let’s look at ice/albedo. For ice to go away, the temperature first has to rise. So now ice melts and albedo increases leading to more absorption of solar energy and more temperature rise. But wait, that means ice/albedo lags temperature. So, according to you, it can’t possibly amplify the temperature increase. Do you really believe that? Amplification by CO2 works in a similar fashion except that it reduces energy out flow rather than increasing energy in flow. The peak timing thing is only true if CO2 is the only other amplifying factor and the system is heavily damped. Neither is true.
I didn’t say that, nor did I imply that. Quite frankly, your assessment leads me to believe you do not understand what I actually claimed.
Mark
Re: Mark T (Feb 3 16:45),
I’ve just reviewed the entire thread. You never do claim anything other than someone else is wrong. You then hand wave about causality and control theory.
So tell me straight out, can changes in ice/albedo and CO2 caused by temperature changes cause a larger change in global enthalpy, however measured, than the original forcing, presumed to be Milankovitch cycles, or not?
169.DeWitt Payne said
February 3, 2011 at 5:46 pm
Actually, even that’s not true. I pointed out exacly why someone else is wrong. Maybe next time you’ll read a little more carefully before launching into unfounded accusations.
There’s no handwave at all, your insufficient understanding of causality and control theory notwithstanding. I clearly explained exactly what I was claiming as well as the circumstances in which those claims hold.
My opinion on this topic is immaterial. For you to even ask is nothing more than a red herring.
What I can say is that you cannot determine such an effect simply by looking at correlations. I can also say that for any relationship A->B (the -> means implies, or causes,) then a “switch” to B->A (or even some B->A + f(B)) will cause a problem w.r.t. stability in the presence of feedback. I’ll leave it to you to figure out why, but knowledge of feedback control will definitely be a plus. Irrespective of what you would like to believe, feedback will continue to operate under the known physical laws that it is based upon.
FYI, linear feedback mechanisms can be determined by observing both the open-loop and closed-loop responses quite easily, non-linear feedback mechanisms will take significantly more effort than that (if feedback terms add linearly the task is simpler, of course, even if the feedback terms behave non-linearly.) Since we cannot observe the open-loop response of the system we are attempting to describe, the task becomes even more difficult whether the feedbacks are linear or not. Simple physics is not enough. Thought experiments don’t work well, either.
Mark
#154, Pat, in the process of preparing yet another drawn out point for point response, I had a brief, shining moment of clarity. Maybe it will help clear up my position (and yours).
First, let me repeat that I’m not trying to establish causality through the ice core records. My aim is and has been to show that the co2 lag does not preclude causality.
I don’t want to get out my credit card again, so try this: Can one deduce that increasing albedo (due to greater snow and ice cover) does not cause cooling because albedo change lags cooling?
Unfortunately, the ice cores don’t show conclusively which came first. (intended humor)
#170: I’m in over my head, but if co2 changes linearly and temperature responds geometrically, won’t that tend to be stable? Or is that too simple?
Too simple. If I get a chance I will try to post a simple explanation of why. Posting from my phone makes it difficult.
Mark
The need for a brake is obvious. (BTW, that’s why I threw in the extra variable in #112). Maybe the answer lies in the other piece of information Pat provided, the “greening up” of the biosphere. If increased biosphere co2 uptake balanced the increased release of co2 from the ocean, co2 (and temperature) would stabilize.
That might also provide a mechanism to bring temperature back down, i.e., biosphere uptake > ocean release = falling co2. Lag would need to be explained, so that’s not a great candidate, but once the system is at equilibrium, anything that got the ball rolling back down hill would start the process. How certain are we of the cooling lag?
Marc;
You keep trying to get rid of the CO2. The only effect you can even tentatively refer to is “greening up”. FYI, Erl Happ has proposed that after a massive input of CO2 from e.g. a flood basalt, the biosphere eats CO2 right down to its own starvation level, and awaits another mega-belch. We mere fauna are unable to keep up.
Or maybe now we can. Help end the CO2 famine! Maximize your CO2 production! Free coal-generated electricity all ’round, for everyone!
Re: Brian H (Feb 4 02:34), Correction, and apologies to “Chefio”, AKA E.M. Smith.
http://chiefio.wordpress.com/2010/10/10/got-wood/
How ’bout them apples? And phytoplankton, and tomatoes?
Arg. That’s “Chiefio”. My excuse for calling him Chefio is that I hadn’t finished cooking.
P.S.
The “lags” per ice cores etc. are dwarfed by the multimillion year disconnects found throughout paleohistory. Ice ages with 4,000 ppm CO2, Hot Houses (22°C) with levels in the low hundreds, and switches back and forth, with no correlations.
Now we’re down to the short strokes (annual/decadal) and the only correlations are pre-’60s rings in a tree matching temperatures in a completely separate sub-climate regime hundreds of miles away. But only during the “calibration” period; before and afterwards, the rises and declines needed hiding.
It’s all garbage, Marc. Logical, statistical, scientific, mathematical garbage in and out. Enjoy your meal!
Good morning, Brian. The question currently under discussion is whether or not the correlation of temperature and co2 shown in the ice core records of the last 800,000 years precludes the hypothesis that co2 is a significant driver of temperature. Would you like to offer your opinion?
After you do that, you can spread your BS all over the Amazon to test your limiting factor hypothesis. Since you missed the humor in one of my earlier replies, let be clear. CO2 as limiting factor is irrelevant because it’s highly impractical.
Sorry, Brian, I mistook your point, so my objection to it was invalid.
Now that I’ve enjoyed my humble pie, and if you grant that the lag does not in itself preclude co2 from being a significant temperature driver, I’d be interested in your thoughts on how the increased co2 uptake you point out in #175 and #176 might play a limiting role in a hypothetical co2-as-an-amplifying-feedback mechanism.
RE: 172 and 173…
Basically, the “stability” of a feedback system is determined by the poles of a system (in the transfer function between input and output.) You can include zeros, too, because a zero can cancel a pole, but in reality, with continuous time systems, this is an unlikely proposition because they tend to move around a little bit leaving the chance of system runaway (unstable.) With discrete time, quantized systems you can “fix” a pole and a zero to coincide (because you can use exact numbers) thus eliminating the potential for instability. Once they cancel, of course, the resulting transfer function is changed as if neither existed in the first place.
What does this mean?
First, the transfer function is the relationship between input (say x) and output (say y) of a system: y = H * x.
A controls engineer would likely use continuous time variables, e.g., y(t) and x(t), with various derivatives, e.g., y_dot(t) and y_dot_dot(t), and the transfer function would be determined by implementing a Laplace transform on the state equation, yielding H as a function of s = j*w where j is sqrt(-1) and w is radian frequency (note that H is not a function of time unless the system changes over time.)
A signal processing engineer, on the other hand, tends to do things in discrete time, e.g., y(n) and x(n), with various lags, e.g., y(n-1) and y(n-2). In this case, n is a discrete time step, typically chosen to be small enough to accurately represent the highest frequency in the system to avoid Nyquist sampling issues which requires two samples per cycle minimum (if something cycles once per day, for example, you’d need at least n = 1/2 day, preferably even shorter.) Of course, sometimes your n is fixed by the sample rate of the data you’ve been given, leaving you little choice in the matter. The transfer function is determined by taking the Z-transform of the difference equation relating y(n) to x(n) resulting in H as a function of z (or z^-1 more commonly.)
I’m a signal processing engineer, so I see things from the viewpoint of the latter.
A simple first-order transfer function (feedback yielding a single pole) is:
y(n) = x(n) + a*y(n-1)
That is, the current output equals the current input plus a * the previous output. The parameter a is the “gain” in the feedback path. The transfer function is determined by first grouping all of the like terms (y’s and x’s) then taking the Z-transform:
y(n) – a*y(n-1) = x(n)
Y(z) – a*Y(z)*z^-1 = X(z)
Y(z)*(1 – a*z^-1) = X(z)
Y(z) = 1/(1 – a*z^-1) * X(z)
therefore H(z) = 1/(1 – a*z^-1) = z/(z – a)
Stability is first determined by finding the region of convergence for the transfer function (note that z is a radius in the complex plane.) The ROC is the region outside of the radius described by the pole with the largest magnitude. In this case, the only pole has a magnitude of |z| = a. Stability is acheived if all of the poles reside within the unit circle, i.e., |z| = 1. Therefore, stability is acheived if |a| < 1.
This makes sense, of course, because that means the amount fed back must be less than the previous output. If you multiply the previous output by more than one and add it back in, the output will grow without bound even if the input goes to zero. For example, if a = 2, and at n = 0 we have an input of x(0) = 1, then y(0) = x(0) + 2*y(-1) (assume x(-1) = y(-1) = 0,) which yields y(0) = 1 + 0 = 0. Then at time n = 1, even if x(1) = 0 and stays that way, you'll get growth: y(1) = x(1) + 2*y(0) = 0 + 2*1 = 2.
If, however, |a| = 0, then you have y(0) = 1 as before, y(1) = 1 + 0.5*1 = 1.5, y(2) = 1 + 0.5*1.5 = 1.75, y(3) = 1 + 0.5*1.75 = 1.875, etc., until y(n) -> 2.0 as n -> infinity. This means the DC “gain” is 2.0, or 1/(1-a) which makes sense since the conversion is summation of 1/a^n and this only converges for |a| < 1, which is the same as the stability criterion above.
Applying all of this to climate is not as straightforward as a single pole system such as the one I showed here.
Mark
“If, however, |a| = 0” should be “If, however, |a| < 1… then, as an example with a = 0.5"
Mark
180.Marc said
February 4, 2011 at 11:33 am
Note, btw, that the term “amplifying” is actually a misnomer. The input is, technically, energy, and energy cannot be amplified (or multiplied) through the first law of thermodynamics. What happens is that there is increased storage of energy, which appears as amplification but really indicates that at some point energy leaving the system (the earth) is less than energy entering the system.
People often like to use the example of a microphone. There is true amplification in such an example but it fails as an analogy because not only does the microphone have an input (sound from the surrounding areas,) but it also has a power supply (either batteries or a plug into the wall, or similar.) In other words, the gain from feedback is a result of the added power from the power supply.
In the case of the climate system, the input is the sun and there is no “plug in the wall” for true amplification (unless there is some other energy source we have not discovered, which changes everything anyway.)
Mark
Mark,
Thanks for taking the time to put that post together. It helps me to think about gain. I’m sure “not as straightforward” is quite an understatement. And I imagine folks smarter than I have tried to estimate gain from the amplitude of the temperature swings.
Gains are not the issue; there are lags, which are either long enough to make a measurable difference, or are not. The nature of the lagging force is crucial; if it happens to interact with a negative feedback process with a fast response time, the effect will be minimal. H2O’s multiple roles in the weather/climate system make it the most obvious suspect, and there is excellent evidence that it reacts rapidly to limit any temperature changes from lag-induced imbalance in energy levels or flows. These range from simple (its high specific heat) to more complex (heat-pipe transport via phase-change dynamics to, or near, TOA where its fingerprint radiation rather readily escapes.
It is crucial for Warmism to recruit H2O as a positive feedback, but the efforts are very hand-wavy and lame, to date, with no sign of improvement.
The energy and temperature lags, btw, relevant to atmospheric CO2 are measured in milliseconds, not millions or thousands of years. Those are oceanic and biosphere deposition of thousand-foot chalk bed kinds of time periods. Which is why the vast majority of CO2 on the planet is currently entrained in carbonates etc. in sedimentary rock.
The global temperature falls. Oceans become better at dissolving and fixing CO2, especially in the fringes of the very cold and surface-agitated Arctic Ocean and Bering Strait. CO2 falls. The temperature rises for some unrelated reason. Oceans outgas, everything from oxygen to CO2 to methane to nitrogen to sulphur dioxide. Some external forcing driver changes, and the oceans cool. Rewind, with variations.
(#186, 187: Thanks, Brian. Fine points worthy of discussion, but I’m trying to stay on topic, so let me get back to H20.
My question about gain relates to the possibility of a runaway system. My intuition is that it is not, but I’m trying to think it through. That will help me evaluate the credibility of someone who claims we’re going to Venus. Can the gain be greater than 1? Hypothetically, I see no reason why not, but I see no evidence of that either.
The possibility of lag-driven warming was pretty well demonstrated here by DeWitt, so it’s obviously fallacious to argue it’s not possible, or even likely, based solely on the evidence of the correlation. I wanted to look at the feasibility of biosphere co2 uptake as a stabilizing brake because I’ve been accused of ignoring half the story when it’s inconvenient. So lets look at it.
I suspect (if we accept AGW theory), “greening up” might help explain how the temperature reached an upper equilibrium. Does that make sense?
The closed loop “gain” can easily be greater than 1… the path gain, I.e., the feedback coefficient can never be greater than one in a passive system, as I described above.
As a result, true “runaway,” which is caused by instability, is all but impossible. Marginal stability may be possible.
Mark
I appreciate your spending time on this, Mark. I’m still having trouble though. That gain in a passive system must be less than 1 makes sense. But in my mind — and this has nothing to do with AGW — I’ve got it set up as an active system, like the audio example. I see the sun as the power supply and (the storage of) thermal energy as the input signal. The climate system, with a little orbital variation on the side, is the amplifier, the gate that controls energy from the sun (which varies as well, complicating matters). Gain is the sum of climate system responses to changes in the signal. Is that an incorrect way to think about it?
Getting back to Pat and logic, let’s revisit #155 for a moment.
I agree that the 1st syllogism is false. I set it up like that to clarify what I understand Pat’s argument to be. In #154, he wrote: Voila, you have evaded the fact that ice core data shows CO2 does not drive temperature. So, by calling the argument false, Pat is either agreeing that his logic is flawed or I’ve misunderstood his argument. The 3rd possibility is that I’m deluding myself that I’m being rational.
The second syllogism is a small step in the right direction, but still flawed.
Though both of the premises are true, the conclusion is only valid if an effect can have no subsequent effect, that is if mutual, cyclical relationships can not exist. That’s silly of course. All one can say from the ice cores is that CO2 increase did not trigger the warming after each ice age. Claiming anything more is an appeal to intuition and warmists, being warmists, may suspect anyone making this argument is deliberately trying to confuse the issue.
Okay, let’s look at the final conclusion in that post. Pat wrote: That syllogism, brought into science and coupled with the lack of a predictive theory of climate, renders unlikely any assertion that CO2 drives air temperature.
“Unlikely” is better. It acknowledges that we have to use inductive reasoning instead of deductive reasoning. That’s legit, but not as conclusive.
Even so, with a lack of a predictive theory, the only way I can get to Pat’s “unlikely” conclusion is with special pleading and circular reasoning, perhaps with a dash of correlation precludes causality thrown it: everything I’ve been accused of. So help a poor warmist understand, what route DO you take?
Sorry to dwell on this. It’s all well and good to point out legitimate areas of uncertainty, such as in climate models, but invalid appeals to intuition come across as either dishonest or ignorant and are a discredit to legitimate skepticism.
The input is thermal energy, Marc, from the sun. Where else did you think it came from? Natural systems cannot be anything other than passive. You can’t feedback more than you have.
Mark
And, yes, that is an incorrect way to think about it. The first law says you cannot create nor destroy energy which necessarily implies it cannot “amplify” it without additional iinput. The overwhelming source of energy is what we get from the sun. The only way to get an apparent amplification is to create an imbalance between what comes in and what goes out (work inside the system and internal sources play a small role, too.) If you want an active system you need another source besides the sun.
As long as you you stick with this flawed belief, your rationalizations regarding Pat’s comments will remain equally flawed.
Mark
Mark, a basic amplifier circuit consists of an input signal, a source of energy, a gate modulating the energy source in relation to the input, and an output. No? I’ll do my best to stay within the first law, promise.
Yes… the sun can be one or the other (input or supply,) not both. The output, of course, is energy radiated into space. The stored energy, what you incorrectly referred to as the input, is a result of the feedback, it is internal to the system.
Mark
Mark, my train of thought with the amplifier didn’t go anywhere so I won’t pursue it. I wanted to treat the sun as an external power source and consider the system to be active, but I’m just being intuitive and need to read up. I don’t have an argument to make, really just trying to understand the requirements for a stable climate system. Thanks for your help.
No problem.
Mark
Re: Mark T (Feb 7 04:18),
Not so. Problem. That is, I think he’s exposing the “double counting” that underlies all warmist runaway models.
No, Brian, just trying to learn a little. Not all believers in AGW are sinister socialists. Honest.
Uh, I meant as “I don’t mind answering legitimate questions as long as you are willing to learn.” Had he continued, I would have answered differently.
Mark
Pat, since this thread is about your research, it may be useful for future readers to wrap up this lag business.
Back in #172, I asked if one can logically conclude that increasing albedo does not cause cooling because albedo increase lags cooling.
The question is meant to illustrate the problem with saying, as you have, that increasing c02 does not cause warming because co2 increase lags warming.
Marc;
Desperately changing the subject, are we? The forcing driver in AGW is the claim by you Warmists that CO2 is the “forcing driver” (an invented concept, not relevant to real physics) for the Earth’s climate. Now that it is finally being ack’d that it lags the onsets of both Warming and Cooling, y’all want to make it the “forcing amplifier”.
Blech.
The Magic Thermostat is dead. Get over it.
Brian, you’re half right. I’m not changing the subject, but I am changing the forcing. Why? I’m sure that Pat’s plenty logical most of the time, but perhaps he’s so sure that co2 can’t doesn’t drive/amplify/cause global warming that he’s still convinced there must be a flaw in my delusional reasoning. So to illustrate the point I’m (desperately) trying to make clear, I got rid of co2 forcing and put in albedo forcing. The logic is exactly the same and should be obvious even to you.
In a fully interactive system, there is NO “forcing driver”. Because none of the variables are truly independent, and/or their feedbacks are complexly related.
TSI, however defined (with/without UV, etc., etc.) might be close, but it is not sufficiently “forcing” that the 30% increase since the Cool Sun has had much of an effect on the favorite dependent variable, temperature.
Every attempt to isolate and “grab onto” a forcing driver is doomed. The entire AGW-proving & exploiting-project is that of a toy model with vast delusions of grandeur.
I repeat, give it up.
#166, Marc you wrote: “I stand corrected on the instances of imprecision.
“The context here is whether or not increasing co2 concentration in the atmosphere is cause for alarm. In this context, you said: “We have ice-core data showing that for the last million years CO2 always followed air temperature, and never drove air temperature.”
“From this statement I inferred this argument: since co2 never (in the last million years) drove air temperature, c02 never drives air temperature. (bolding added)”
Here’s what I actually wrote in #156: “[Your corrected] syllogism, brought into science and coupled with the lack of a predictive theory of climate, renders unlikely any assertion that CO2 drives air temperature.”
Where do you see “never” in my statement? It’s nowhere to be found. You seem to continually and tendentiously distort the meaning of my statements, recasting them to conform with your own internalized logic. I write as a scientist, Marc, not as a logician. I wrote, “renders unlikely” and that’s all we’ve got as scientists.
Science should be your context, rather than the logical sets into which you improperly leverage your thinking about the physical world.
You wrote: “I think you will agree that one can’t deduce the conclusion (co2 does not drive air temperature) from the inferred premise (co2 did not drive air temperature during the ice ages). First, the argument draws a general rule from a specific case.”
Hume’s fork about factual inference doesn’t apply to science. Physical theory predicts consistent deterministic observables, unless some discontinuous force has appeared. A physical theory of climate says that climate today, where the theory was developed, operates the same way as the climates of the past, unless there is some mechanistic discontinuity.
That discontinuity would not be, e.g., a drift in the Melankovitch cycle, but rather a fundamental change in the way the climate responds to its drivers, or a discontinuous change in drivers, either case requiring a different theory with different physical formulations of the forces.
But climate theory, so far, is not time-wise internally inconsistent. So, in the context of a time-wise extensive theory of climate, the behavior of past climates does tell us something about the behavior of present climate. Your argument, to be accepted, requires discarding science in favor of an inductive logic that allows no predictions at all. Insistence on that course is anti-knowledge (it’s the canonical post-modern cant, actually).
“Second, we can suppose conditions in which the premise is true, but the conclusion is not. If we suppose that something else, for example orbital variations, “got the ball rolling” and caused an increase in co2 that in turn drove additional warming, then the conclusion would be false, while the premise still holds.”
You’d need to demonstrate your “ball” using physical theory to show the significance of the effect plus observational evidence of the presence of the causal impulse. Ringing in some vague but convenient a priori phenomenon after the fact to save your premise is scientifically invalid.
“The argument fails on logic, yet you maintain your position is logical.”
I maintain my position is scientific. It’s you who seems wedded to (a scientifically sterile) logic.
“You must have been arguing something else. So lets be clear: how do the ice core temperature / co2 reconstructions inform the current climate change debate?”
It informs the current debate in the context of physical uniformity.
Here are the basics: Scientific meaning derives from a falsifiable theory tested by replicable observations. GCMs embody our best theory of climate. GCMs produce errors in predicted observables, such as the intensity of upwelling IR at the TOA, that are *much* larger in W/m^2 than the forcings of CO2 they are trying to detect. I.e., the uncertainties within the physical theory are much larger than the observational effect under study.
That means the forcing effects of CO2 predicted by GCMs are physically meaningless.
That, in turn, means we have no idea what added CO2 will do to the climate.
So, we turn to observables to try and get an empirical view: Despite all the studies and straining at statistics, there is *nothing* unusual about recent weather.
The warming trend since 1880 is mild at best, it’s less than 1 C (a very optimistic lower limit of resolution), and is not in any way unprecedented.
Our empirical view of the ice ages says that CO2 did not drive the temperature swings.
So, in the scientific context, what is there to produce *material worry* about the climate impact of doubling atmospheric CO2?
#172, Marc, you wrote: “Can one deduce that increasing albedo (due to greater snow and ice cover) does not cause cooling because albedo change lags cooling? ”
Marc, your question is a kind of analogical non-sequitur with respect to the preceding argument. To be consistent with your argument about the ice-age/CO2 relationship, your question about ice albedo should be, ‘can one deduce that increasing ice albedo did not cause the cooling that preceded it?’
I’m guessing you know the answer to that.
#191, Marc, you wrote: “Getting back to Pat and logic, let’s revisit #155 for a moment.
“I agree that the 1st syllogism is false. I set it up like that to clarify what I understand Pat’s argument to be. In #154, he wrote: Voila, you have evaded the fact that ice core data shows CO2 does not drive temperature. So, by calling the argument false, Pat is either agreeing that his logic is flawed or I’ve misunderstood his argument. The 3rd possibility is that I’m deluding myself that I’m being rational.”
Marc, the context of post 155 was the opportunistic change in your ice core argument. You have now taken my comment out of context and falsely made it into a claim of invariant acausality. I really wish you’d stop re-interpreting what I write to conform to your expectations.
You wrote: “. All one can say from the ice cores is that CO2 increase did not trigger the warming after each ice age. Claiming anything more is an appeal to intuition and warmists, being warmists, may suspect anyone making this argument is deliberately trying to confuse the issue.”
You left out the cooling phases of the ice ages again, Marc. They initiated at high CO2, and continued even as atmospheric CO2 was relatively high. The context of physical theory to derive some meaning from that is not an appeal to intuition. The meaning is that the lag of CO2 behind both rising and falling air temperatures provides no evidence of a CO2 climate driver. Without such evidence, there’s no reason to think that CO2 did drive the changes in air temperature.
You may offer some hypothetical ‘action at a distance’ argument for ice age CO2 as a hidden influence exacerbating warming without influencing subsequent cooling or maybe, even more radically, a hypothesis for CO2 cryptically exacerbating both warming and cooling behind a causal chain, but if you want to be taken seriously you’ll have to formulate a predictive and falsifiable physical theory.
Your interpretation of my argument was wrong, and your dark conclusion about “deliberately trying to confuse the issue. is also wrong. And though decrying intuitive judgments you made one there, and it was an implied assault on character to boot.
You wrote: ““Unlikely” is better. It acknowledges that we have to use inductive reasoning instead of deductive reasoning. That’s legit, but not as conclusive.”
The context is within the framework of deductive physical theory, however, not Humean atheoretical induction.
Following the meaning of Ice Age CO2 evidence as provided by the uniformitarity of causal climate mechanisms provided by physical theory, the follow-up question is whether there is any evidence in ice age air tempertures to think that CO2 will drive modern air temperatures.
What’s the answer, Marc?
“Even so, with a lack of a predictive theory, the only way I can get to Pat’s “unlikely” conclusion is with special pleading and circular reasoning, perhaps with a dash of correlation precludes causality thrown it: everything I’ve been accused of. So help a poor warmist understand, what route DO you take?
That you require circular thinking does not mean that mode drives the thinking of others (and what exactly does “correlation precludes causality” mean, anyway?). The lack of a predictive theory, which you now acknowledge, means that no one knows what they’re talking about with respect to the impact of CO2 on climate. Note this has been my position all along, here, for example, in my opening essay (“no one really knows the rate and magnitude of warming”), and in my Skeptic article.
You wrote: “It’s all well and good to point out legitimate areas of uncertainty, such as in climate models,..”
Do you understand that the large uncertainty in climate models in and of itself disallows any conclusion that human-produced CO2 is at all impacting modern climate?
“… but invalid appeals to intuition come across as either dishonest or ignorant and are a discredit to legitimate skepticism.
I have never even once appealed to intuition. Your failed attempts to show that have rested on incorrect interpretations or tendentious misrepresentations of my posts, e.g..
#201, Marc, you wrote: “The question is meant to illustrate the problem with saying, as you have, that increasing c02 does not cause warming because co2 increase lags warming.”
But I didn’t write (say) that. My point has been that because CO2 lagged both warming and cooling during the ice ages, there’s no evidence that CO2 had a causal role in the changing air temperatures; in 81, for example, and #155.
As a consequence of that evidence, and in the absence of a falsifiable theory of climate, there’s no reason to suppose that CO2 is presently driving air temperatures.
Recall that our physical context is uniform causality.
I’ve now replied directly to the albedo formulation in #206. Sorry for the delayed response.
#203, Marc, you wrote: “Why? I’m sure that Pat’s plenty logical most of the time, but perhaps he’s so sure that co2 can’t doesn’t drive/amplify/cause global warming that he’s still convinced there must be a flaw in my delusional reasoning”
I hope, after posts 205-208, that you’ve been disabused of the notion that I’m “so sure that co2 can’t doesn’t drive/amplify/cause global warming.”
It’s not a question of “sure,” it’s a question of evidence. There’s no evidence that CO2 has driven air temperature, that it drives air temperature, or that it’s responsible for any recent warming of the climate.
There’s no evidence that global weather has recently been unusually variable, heated, or extreme.
There is, therefore, no reason for alarm about a large scale human influence on the climate.
There certainly is no excuse for 20 years of frantic alarm about global heating, for character assassination of skeptical scientists, for the corruption of science, or, since the 2AR, for the intentional and dishonest misrepresentation of the evidence, either as written within the SPM by the IPCC or as pronounced by IPCC members.
Pat, I tried to express in my previous post what I understand you to be saying, not to put words in your mouth. I can see how understanding is tough in this forum, so thank you for the detailed response. Before reply, I’d like to clarify your meaning in #208. You wrote:
I’ve understood you to be arguing that the fact that “there’s no evidence that CO2 had a causal role” — which I agree with — can be taken as positive evidence of the contrary, that co2 did not then, and does not now, have a causal role in warming the planet, which I disagree with.
In the formulation of your argument in #208 quoted above, the way “no evidence” in one sentence becomes “that evidence” in the next sentence suggests this interpretation. Do I have the essence of your position correct?
#210, Marc, you wrote: “I’ve understood you to be arguing that the fact that “there’s no evidence that CO2 had a causal role” — which I agree with — can be taken as positive evidence of the contrary, that co2 did not then, and does not now, have a causal role in warming the planet,…
I am not arguing, and have not argued, past ‘there’s no evidence of a causal role‘, Marc. Why should anyone infer further?
Let me decompose my sentence: “As a consequence of that evidence, and in the absence of a falsifiable theory of climate, there’s no reason to suppose that CO2 is presently driving air temperatures.”
Notice the stressed clause. The first part of the sentence refers to empirical evidence — ice age data. The second part to the absence of applicable theory.
The time lag exhibited by the first does not support a theoretically naive imputation of causation. The absence of adequate theory does not allow an assignment of causality now (or then).
Consequently, there’s no basis for a causal assignment.
But let me complete the analysis from the view of science. The absence of an adequate theory does not allow a rejection of CO2 causality, either. We plain don’t know.
‘Not knowing,’ is why everyone looks for statistical peculiarities in weather trends, and why the argument goes back and forth with such vigor among climate scientists. No one has a valid theory to establish causality, no matter that GCMs are continually misrepresented as such. It all devolves to opinion, which disallows a settled debate.
A big problem with the empirical arguments about statistical weather trends is that we now know for sure that climate phenomena exhibit long term cycles, probably at all time scales. So, 30- or even 50-year linear trends probably don’t really tell us much about unnatural peculiarities in climate.
So, nobody knows whether anything peculiar is happening, but nothing looks peculiar given what we do know. Shouldn’t that be what informs our decisions, and makes them diffident?
Maybe our added CO2 really will detectably warm the climate. So, should we wean ourselves from fossil fuels? Maybe. No one knows. Maybe producing more CO2 now, really is delaying an incipient and catastrophic ice age. No one knows. So, should we wean ourselves from fossil fuels?
We literally have no answers. How should we proceed?
Does the fact that nothing untoward has demonstrably happened to global climate, despite all the looking and all the shouting, allow large scale dislocations anyway?
There are good political reasons to import much less oil. Maybe it would be a good thing to accommodate that in a deliberate, careful, and reasoned manner — allowing our society time to adjust smoothly to the economic and engineering transition. I’d be happy to see that.
But intemperate actions biased by fearful beliefs unsupported by evidence is a recipe for disaster. Note the long and continuing history of sectarian cruelties as an object example.
The frantically accusatory culture that has grown up around CO2 emissions is dangerous and entirely unjustified, and the corruption of science is inexcusable. It’s a case study in the shameful and deliberate subversion of social and scientific ethics.
I hope this finds an accommodation for us, and thanks for your civility.
Pat, I’ll wind this down by pointing out just a few places where I take issue with your recent replies.
Regarding #205: In response to me saying, From this statement I inferred this argument: since co2 never (in the last million years) drove air temperature, c02 never drives air temperature., you wrote:
“Never” comes from your reply at #65, which got the ball rolling, so to speak. You wrote:
Where you say:
I disagree. To invalidate the argument, I hold that all that is required is a plausible case in which the conclusion is false such as the one DeWitt’s demonstrated. But we can agree to disagree on this.
I’ll address a few other points later.
Marc, yyou completely distorted his use of the word never. Pat said “never drove” and you switched it to “never drives,” a completely different statement.
Bad form.
Mark
Perhaps, Mark, but given that the context is concern over future warming, and not an academic exploration of interglacial warming, it seems at the very least reasonable to infer that the evidence of the lag was offered to refute the forcing potential of increased co2 now or in the future. I made it clear, you’ll note, that I had drawn this inference. If it’s incorrect, it’s easy enough to just say so. However, Pat specifically stated that the ice cores, let me grab the quote, “informs the current debate in the context of physical uniformity.”
Mark, I believe my inference that co2 never drives warming is supported by Pat’s argument back in #81.
Seems reasonable to infer from that the argument that if co2 never drove warming, it never drives warming.
No… you are really adept at contorting what people write to fit your world view. That’s not even remotely close to a correct interpretation of what he wrote.
Sorry, but you are no longer worth the effort.
Mark
We need some theology and geometry. Some taste and decency.
H/t J. Ignatius.
==================
Mark, I’ll take “adept” as a compliment. Thank you.
You’ve got a fair point. Pat’s most recent replies, starting at #205, state his argument in a way that differs from my formulation of it. Either he is modifying the original argument, I misunderstood it all along, or I’m misrepresenting it. Your hypothesis is plausible and if that’s what you believe, then I agree there’s no point in continuing. Sorry you feel that way.
I have yet to address Pat’s replies. There may room for some agreement, but I’ve got to look up a few words and make sure I give’s Pat’s words full and fair consideration before I do.
Marc – you are just a bit self-important, aren’t you? Go away and do some more studying.
Mark F;
subtlety and understatement are futile in this instance. Don’t say “just a bit” when you mean the opposite. Say it properly: “excessively”, or some synonym.
220 – agreed on all points.
Marc has conducted a very successful “thread highjack” here. The goal is distraction and obfuscation. Respond accordingly.
Pat, it’s one thing to say that the lag provides no evidence of a causal role, it’s another to say that the lag provides evidence that co2 had (or has) no causal role.
I may have misunderstood you or got your argument wrong. Biases get in the way, so if you tell me I’ve misconstrued your argument, I have to accept that. Thank you for elaborating.
Allow me to clear up my position, because I’ve not done a terribly good job explaining it. Again, it’s my argument, not yours.
I find one of the sillier arguments against AGW to be the intuitive notion that because co2 increase lagged temperature increase during the ice ages, there is evidence that temperature changes causes co2 increase and not vice versa. (I mistook your position, such as back in #81, to be an argument of this type. If you put on my warmist glasses for a moment, you might be able to see how I got there.)
Let’s take two mutually exclusive assertions:
1) There’s (nothing to worry about because there’s) no evidence increasing co2 causes global warming.
2) There’s (plenty to be alarmed about because there’s) evidence that increasing co2 causes global warming.
Now let’s look back (#65) at the statement that I objected to.
Does the lag help us evaluate which assertion is more likely to be true?
No.
Each statement relates to whether increasing co2 causes global warming. True, the lag supports the idea that temperature increase causes co2 increase rather than the other way around.
However, the alarmist premise requires an increase in co2 to cause global warming, whether due to natural or human causes. The ice cores are therefore entirely consistent with the alarmist position.
Because the AGW hypothesis requires a lag, simple logic prevents you from using the lag to refute the hypothesis.
I know that’s trivial, but some people get confused by the arrows of time and it ticks me off that a question so important gets discussed at that level.
Sorry to mix up what you were trying to say about the ice cores, Pat. To respond to what you are indeed arguing, I’d have to refer to Hansen and the evidence from paleoclimate research. That’s a whole ‘nuther kettle of fish. Perhaps Jeff will be so kind as to invite me to do a post on that someday.
‘Nuff said. Thanks again for engaging civilly. — Marc
PS, I’ve got a question about your temperature analysis that I’d like to ask. Yeah, I’d like to get back on topic!
Pat,
.
Getting back to the temperature measurements…
We’re talking about recording temperature from a thermometer. The recording error was guesstimated at +/- 0.2 C. It seems reasonable to assume that the recording error would be constrained to the nearest degree, i.e., +/- 0.5 C. My question is this: why does it matter what the magnitude of the error is if it’s truly random? Given sufficiently large number of measurements (and random error), wouldn’t a low-level trend come through a relatively large measurement error of say +/- 1 C?
Pat, back in #99 you offered this source to make your point about the deep past and a lack of correlation between CO2 and temperature. I got into a discussion somewhere else who pointed to the same trash, so I tracked down the source.
Guess where it came from. Was it :
A) a highly regarded, published academic geologist?
B) an up and coming grad student?
C) a coal mining safety engineer?
D) a high school senior’s science project?
E) a 14 year old girl who decided to debunk global warming in her spare time?
Here’s the original source of the graphic. The name of the source is at the bottom. From there you can find his credentials. The answer, I’m afraid, is neither A nor B.
I thought you’d want to know who you’re citing as authority on this matter of import.
I shot an arrow into the sky. And yet, it flies.
===========
#212, Marc, DeWitt showed some interesting empirical work in his tAV article, but he didn’t produce a physical mechanism.
As regards mechanism, DeWitt included a very cogent statement in his conclusion, which you apparently missed: “While you can’t prove that CO2 didn’t amplify warming unless you know the precise shape of the underlying forcing, you also can’t prove that it did amplify warming or determine the magnitude of the effect…” DeWitt made exactly the correct caveat following an empirical analysis: that is, it may support a hypothesis but does not support any conclusion.
So, there’s no “plausible case,” as you have it.
DeWitt didn’t make a case, he nicely opened the door to a hypothesis, without actually producing that hypothesis. Just to be clear, a hypothesis in science (or engineering) is a predictive, testable, and (typically) mathematics-based elaboration of theory explicating the hypothetical mechanism in detail. The explication is in sufficient detail to make a prediction precise enough that its non-observation constitutes a falsification. A hypothesis is not just words (with one or two notable exceptions), and it’s certainly not about vague inexplicables “[getting] the ball rolling.”
With Mark T, I agree that “never drove” is not “never drives.”
I’ve already explained the difference between empirical observation and deduction from theory in 155, 205 and 207 . It shouldn’t need repeating.
By the way, Mark T, I thought 182 was an excellent post.
214, Marc, the relevance of the unvarying interglacial lag by CO2 is not “academic” in the sense of unimportant to the reality external to a university.
You wrote: “it seems at the very least reasonable to infer that the evidence of the lag was offered to refute the forcing potential of increased co2 now or in the future.” I already discussed the meaning in a scientific context in 99, 155, 156, and 205. In 205, I observed that, “Physical theory predicts consistent deterministic observables, unless some discontinuous force has appeared. A physical theory of climate says that climate today, where the theory was developed, operates the same way as the climates of the past, unless there is some mechanistic discontinuity.”
In the same post, I added that, “Following the meaning of Ice Age CO2 evidence as provided by the uniformitarity of causal climate mechanisms provided by physical theory, the follow-up question is whether there is any evidence in ice age air temperatures to think that CO2 will drive modern air temperatures.”
At the end of 205, I asked you to answer that question. Right now seems to be a good time to ask you again: given the uniformity of physical climate mechanisms, is there anything in the lag of ice age CO2, behind temperature, to support an inference that CO2 is a central driver of modern air temperature?
What’s the answer, Marc?
You wrote, “I made it clear, you’ll note, that I had drawn this inference [that “the evidence of the lag was offered to refute the forcing potential of increased co2 now or in the future”]. If it’s incorrect, it’s easy enough to just say so. However, Pat specifically stated that the ice cores, let me grab the quote, “informs the current debate in the context of physical uniformity.””
Given my past posts, you either haven’t read anything I’ve written prior to your 214 or else you’re being very disingenuous. Say it ain’t so, Marc.
#215, Marc, do you understand the difference between a causal driver and a primary causal driver?
Here’s an example of the difference, from evolutionary biology: spilled sugar in your kitchen is a driver of ant behavior. Predation is a primary driver of ant behavior.
#223, Marc, your analysis in 223 is not internally consistent. You’re mixing axiomatic logic with science. That’s oil and water. They don’t mix.
It may be an AGW premise that CO2 causes air temperature to warm. That premise does not make it reasonable to conclude that CO2 in fact is responsible for current air temperatures. Even in logic, it is mere triviality to restate the premise as a conclusion.
In the on-going argument, the premise that CO2 is the cause of recent air temperatures is required to be embedded in a physical theory, to make a testable prediction, and to be validated or falsified by reference to observation and/or experiment.
One cannot just stand on the premise and end the argument there.
You wrote that, “True, the lag supports the idea that temperature increase causes co2 increase rather than the other way around.”
That case for support is weaker than your statement implies. The physical theory of climate is so poor, that the two-valued observable of ice age temperature and CO2 do not allow an either-or statement of causality.
It’s not merely that, ‘either T drove CO2 or CO2 drove T.’ There may be a third causal impulse that drove both. That is, the correlation between CO2 and T may just as well reflect a force that drove both of them independently. Their correlation is then deterministic but acausal in the ‘either-or’ context.
An example might be a large scale regime-shift in the PDO, causing the Pacific to warm, driving CO2 into the air and raising the air temp at the same time.
And what might cause the PDO to shift like that? Well, maybe the chaotic internal pseudo-periodic dynamics of energy flux through the climate produced a beat in resonance with the PDO, and drove it into a new regime. No one knows, but the system is extremely complex and disallows linear either-or arguments.
You wrote: “Because the AGW hypothesis requires a lag, simple logic prevents you from using the lag to refute the hypothesis.”
The AGW hypothesis requires the opposite lag than the one observed. AGW requires T to lag CO2. The opposite occurred 14 times. So, you may have been ticked off, but it wasn’t because I had made an arrow of time mistake.
The idea that CO2 amplified T during the ice ages, by the way, also requires that T overshoot CO2 in time. Frank Lansner had an extensive analysis of the T-CO2 relationship on Anthony Watts’ WUWT. From Figure 2 in his analysis, we see that instead, CO2 overshoots T in time.
So, there’s nothing in the ice age CO2-T relationship to support the AGW premise or position.
Don’t look for help in paleoclimate research, Marc. Consider how imprecise is our data on modern climate, where we have instruments actually deployed. The modern data aren’t good enough to calibrate the parameter-driven climate models to within a degree C in air temp.
Our knowledge of ancient climates is far, far poorer and extremely coarse. That state of knowledge disallows any precision conclusions applicable to the very modest rise in air temps we’ve experienced over the last 130 years.
#224, the error isn’t known to be random, Marc. That’s the whole point.
The state of being “random” was merely assigned by fiat in one paper, and then that paper was cited as authority by everyone else. That’s hardly valid error analysis.
Maybe, following AGW logic, they could have premised the error as random, and then concluded that it’s random. 🙂
It’s a little humorous that they even assigned a number. Given the millions of temperature measurements, they could have guesstimated an error of (+/-)10 C and then ‘1/sqrtN’ed it down to zilch, anyway.
Still, I’m grateful to the original authors that they explicated their reasoning in the “B06” paper. Without that in print, I’d never have been alerted to the issue of an assigned random (or a neglect of systematic error) and never have analyzed the air temperature uncertainty. So, kudos to them for being transparent in proper scientific fashion.
#225, the origins of the data are posted right under the graphic, Marc. Did you look at those? If so, what is there to criticize about them?
What difference do the credentials of the actual poster make, if the sources of the data are valid and the data are represented correctly?
Your criticism leads me to respond with this question: do you understand the difference between a red herring and a substantive argument?
But in any case, take a look here, at an article composed about AGW by a working geologist. Scroll down to his first figure. It’s a little complex, but if you look carefully, you’ll notice that 450 million years ago, air CO2 was at about 6500 ppmv, while air temp was about 2 C warmer than now. I.e, an increase of 20x (2000%) in CO2 and an increase of 12.5% in T.
Thanks, Pat. I’d give up if I were you, however. Anything you say will be twisted in an attempt to make it fit with the foregone conclusion. O10 confirms O9, for example.
Mark
Pat Frank, I have a question. The question is about the ice core data. From what I read above, I would assume that CO2 lags temperature and that one cannot find a case where temperature lags CO2 is a main point of yours.
John, my main point concerns 20th century global average surface air temperature, namely that surface station reading error is not known to be random, and that systematic instrumental error has never been appraised, i.e., the E&E study that was the original topic of the essay here.
The stuff about CO2 and ice cores came up in response to Marc’s posts. I haven’t studied the ice core-T data, and I stand by only what I’ve posted.
Pat,
First, thanks for addressing the question about random error, I thought that would be the case. Do you address the extent to which the error is or is not likely to be random? I’d think that would be an interesting bit of research.
Second, no, I didn’t answer your question, but only to spare anyone who might still be paying attention. (Is anyone still paying attention?) Just to restate it, you wrote: At the end of 205, I asked you to answer that question. Right now seems to be a good time to ask you again: given the uniformity of physical climate mechanisms, is there anything in the lag of ice age CO2, behind temperature, to support an inference that CO2 is a central driver of modern air temperature?
What’s the answer, Marc?
No.
The simple fact of a lag, taken alone, no more supports “my” theory than it supports “yours”.
We’d have to consider additional evidence to infer that C02 is a central driver of temperature. Perhaps, given the title of this post, we should dive into it, but I doubt this is the best place. I doubt anyone’s paying attention. Besides, I’ve said all I’m going to about lags. If my argument isn’t convincing, it’s not because it’s not logical.
Pat, about the “beer-reviewed” science: I’m not really surprised that you conclude that the temperature record you analyzed tells us essentially nothing, but you can you’re willing to draw important conclusions from a dated graph that appears to have been created in Paint, by an amateur when more recent, actually published research is available that includes information about the methodology, resolution, error, and other factors that would help us evaluate it’s relevance to the task.
That is not a rebuttal by the way.
I’m still investigating that graph though. I’ll let you know if I learn anything more.
Pat, a couple notes regarding your #230:
Right after referencing a post at WUWT that makes a new argument (I’ll come back to that), you conclude with:
Don’t look for help in paleoclimate research, Marc. Consider how imprecise is our data on modern climate, where we have instruments actually deployed. The modern data aren’t good enough to calibrate the parameter-driven climate models to within a degree C in air temp.
Our knowledge of ancient climates is far, far poorer and extremely coarse. That state of knowledge disallows any precision conclusions applicable to the very modest rise in air temps we’ve experienced over the last 130 years.
So, what, because you’re on the home field you get to use the past to argue your point, but I can’t? Are there any rules?
No, there are no rules, but that’s okay, I’m enjoying this anyway.
Here’s your BTW comment: The idea that CO2 amplified T during the ice ages, by the way, also requires that T overshoot CO2 in time. Frank Lansner had an extensive analysis of the T-CO2 relationship on Anthony Watts’ WUWT. From Figure 2 in his analysis, we see that instead, CO2 overshoots T in time.
Well maybe you see that with your rose-colored skeptical glasses on, but I don’t see the overshoot. When I bring figure 2 into an image editor and drop a vertical rule right on the temperature peak, there’s no co2 overshoot whatsoever. There is a simultaneous peak. Maybe if I squint real hard I can even see a tiny temperature overshoot, but I doubt the data’s good enough to be conclusive. (I’m learning, eh?)
Lansner says the AGW theory works on the rise, but not on the fall:
So if you agree with that, we’re half in agreement, and if you don’t agree with it, why’d you use it? You think you can just you half the graph?
Lansner’s problem isn’t lag per se, it’s the fact that temperature falls faster than co2. In other words you’re arguing sensitivity, not timing. I believe you called that a denovo construction. Which I’m used to by now, so I don’t have a problem with it. Following through to the end and his analysis, he admits, proves nothing. And to his speculation as to why cooling was slower than warming, um, I’d say that albedo has to be considered as well as the constraints on snow-driven ice sheet growth.
As if that’s not enough, I can’t believe you’d put much stock in an explanation like this that doesn’t explain how the graph was derived, that leaves out a major, well-known forcing, and concludes with this caveat:
What else you got?
Pat,
More on the analysis by Lansner over on WUWT. When I zoomed in on figure 2, it was clear that co2 and temp. peak simultaneously. So it’s pretty clear that what he’s done here is index both temperature and co2 to peak temperature and then plot the average temperature to co2 relationship. It’s a fine idea with potential to illuminate; I like it. Obviously, though, there can’t be any overshoot in the graph because the underlying assumption is that the peaks are simultaneous. What it does tell me is that you are sometimes too lazy or too quick to ask whether the data really tells you what you’d like it to tell you.
Now lets tie in your question about primary causal drivers because it relates to the weakness in the co2/temperature analysis. I think I do know, but since it’s evidently important, why don’t you tell me what the primary causal drivers of temperature are.
#236, Marc, the point about random/not-random is that workers in the field were assigning “random” to a set of uncertainties about which no one had, or has, any information at all. The assignment of “random” was specious and empirically unjustifiable. Failing time-travel, no one is likely to ever know whether that uncertainty was random or not.
“No.” Thank-you.
You wrote: “The simple fact of a lag, taken alone, no more supports “my” theory than it supports “yours”.
“We’d have to consider additional evidence to infer that C02 is a central driver of temperature. Perhaps, given the title of this post, we should dive into it, but I doubt this is the best place. I doubt anyone’s paying attention. Besides, I’ve said all I’m going to about lags. If my argument isn’t convincing, it’s not because it’s not logical.”
I quoted you at length here, because this issue recurs in our conversation and I’d like to finish with it. The logic you’re using is that of the classic philosophical ‘black swan’ question. All known swans are white. But seeing only white swans can’t logically rule out the existence of a black swan.
That logic is scientifically irrelevant, Marc. And it tells us nothing about swans.
A proper theory about swans, as physical swans (not as philosophical constructs), is scientific and is the province of Biology. Evolutionary Biology tells us that the “swan” concept, like the species concept, is fluid. There is no such Biological thing as a swan essence that can be asserted, logically manipulated, or violated.
So, correct or not, your logic about the CO2 lag is improperly applied to the scientific question we have been discussing, and tells us nothing about CO2.
A breeding flock of black swanoids, even if they were otherwise morphologically indistinguishable from white swans, would have a genetic relationship to the larger clade of swans that would be described using a species name embodying that association. Philosophy and inductive philosophical logic would have nothing to say about that.
Neither Philosophy, nor Hume’s logic, nor any other sort of logic outside of science, have anything to say about whether or not the lag of ice age CO2 lends any empirical weight to the scientifically rational evaluation of the effect of the modern atmospheric CO2 increase on climate.
#237, Marc, what difference would it make to the correctness of the graph if it were drawn on newspaper with a blunt pencil?
So what if it was made using “Paint”?
The only relevant scientific issue is whether the graph is faithful to its sources. Political discredit is another issue, of course. But we know your interest is the science.
#238 Marc, your notes about #230 ignore 3/4 of the post. Here’s what you opted to not discuss:
1. That it is not correct to mix axiomatic logic with science (which you have repeatedly done).
2. That concluding your premise is not an argument.
3. That either-or, T-or-CO2, does not exhaust the possible physical climate inputs.
4. That the terrestrial climate has “chaotic internal pseudo-periodic dynamics [that disallow] linear either-or arguments.”
These points are enough to make the CO2-T argument pointless. But you have continued it anyway. My entire point all along has only been that the argument about ice-age CO2/T does not support the AGW hypothesis.
Please notice that I’ve never argued that the ice-age CO2/T relationship, as such, tells us anything necessarily important about how the climate actually works. That is the implicit message conveyed by the point about PDO possibly driving both, in terms of a possible happenstantial PDO/energy-flux resonance.
No one knows what drives climate, although the North Atlantic ice-rafted debris fields correlate with the geological solar 10-Be production rates. See Figure 10, here(pdf download; HTML QuickView here)
Where you did begin your reply, you wrote: “Well maybe you see that with your rose-colored skeptical glasses on, but I don’t see the overshoot. When I bring figure 2 into an image editor and drop a vertical rule right on the temperature peak, there’s no co2 overshoot whatsoever. There is a simultaneous peak.”
A simultaneous peak is not the point. The point is that CO2 stays high as temperature falls. CO2 overshoots T. Look at Lansner’s Figure 2 again. CO2 is at about 269 ppm twice. The first time is at about -250 years, when T ~ +1 K. The second time is at about +7,500 years and T ~ -2 K. That’s a difference of ~3 K, at the same CO2 level, and the second instance has CO2 overshooting T. In fact, CO2 overshoots T for the entire +24,000 years.
In your quote from Lansner, that, “the supposed CO2-driven raise of temperatures works ok,” Lansner only means that the qualitative inference is as AGW theory suggests.
He doesn’t mean that AGW theory is predictively correct with respect to the observable trends. But that’s what you imply him to mean (“Lansner says the AGW theory works on the rise, but not on the fall“), and that implication is wrong. AGW theory, meaning climate models, can’t predict the ice ages ab initio, meaning from basic physical theory.
You wrote: “In other words you’re arguing sensitivity, not timing.” No, I’m not. Sensitivity has to do with intensity (degrees K/doubling-of-CO2), not with timing.
Your entire argument is misconceived, and your “What else you got?” is a fine bravura ending, but, as now shown, analytically unsupported.
#239 Marc, what Lansner did was interpolate all six of his reference ice-ages onto a common difference-year scale of -9,000 years to +24,000 years around a local zero point that is common to each CO2-T pair.
For example, he might have chosen year -324,000 as a local year zero, and subtracted all the years between -350,000 to -250,000 from that number. He’d get a relative year scale of -26,000 to +78,000, passing through 0. The T and CO2 data between -350,000 and -250,000 years would now vary +/- about 0 years on the new relative Time (x) scale.
Once Lansner did that for each of the six ice ages, he was able to average together all six ice-age trends in T or in CO2 (each sum of 6 trends divided by 6), and then plot the averaged trends together on the common relative time scale.
He did not do what you suggest he did, namely, “index both temperature and co2 to peak temperature.” Your description supposes that the CO2 trend was artificially shifted in time so that the maximum of CO2 was made to occur at the same relative zero as the maximum T. Lansner did not do that. He merely put the data on a common relative scale, and averaged it all together.
There is no, “underlying assumption … that the peaks are simultaneous.” So, your conclusion following that mistake is also mistaken.
For your last question, I don’t know what “the primary causal drivers of temperature are.” You should have understood my view by now, which has been and is that no one knows the answer to that question.
And so all the public hysteria about CO2 has been and remains entirely misplaced, and the widespread stoop to character assassination by those touting alarm has been entirely unjustified and noxious.
#240: Huh? Pat, you’ve completely lost me with the black swans. I couldn’t follow that train and I don’t know where I supposedly got on it. I suspect you’ve misunderstood me, but since I clearly don’t have a clue what black swans have to do with anything, I don’t really know. Perhaps you could spell out how I’m using black swan logic, if that’s what I’m doing.
#241: Pat, “Paint” really isn’t the issue, though the line width on that chart is probably close to a million years, so it’s pretty tough to show something like the Late Ordovician event, which probably wasn’t much longer than that, and may have been shorter, but that’s a scale problem. The problem is that it is a simple tool that conveys nothing about the nature of the data itself. Since more sophisticated tools are readily available, the use of a simple tool like Paint to present data suggests that the author is not terribly sophisticated. Sure , if the underlying source were adequate to the task of relating co2 and temperature over 600 million years, and that source were fairly recreated, then fine. But I’ve checked the source. Your graph gets the shape right, but misses the scale by a wide margin. The 600 million year curve could easily be recreated with something like 50 data points, and half of those would be within the last 100 million years. So, while there’s no discussion of the data resolution, it could be on the order of 10 – 20 million years. There’s no discussion of the precision of the method either – but climate is recreated from rock types and fossils. And finally, except for the Late Ordovician, the graph you’re using shows no change in temperature for over 200 million years. What does that tell you? So “Paint” is shorthand for simple and poorly presented.
Marc, you were so wedded to the probabilistics from induction in defending the logic of your ice-age/CO2 argument that I was sure you’d be familiar with the classic white swan case that exemplifies your approach.
Science solves Hume’s problem of induction because its predictions project knowledge into the same future-present in which the observed events happen.
As to the rest, my best wishes in figuring out how to think about the meaning of evidence in science.
Pat, great link, thanks.
You’ve done a nice job turning the tables here: …you were so wedded to the probabilistics from induction in defending the logic of your ice-age/CO2 argument that I was sure you’d be familiar with the classic white swan case that exemplifies your approach.
Remember that I was criticizing the soundness of the argument I took you to be making: that the fact that temperature increase preceded co2 increase, as shown by the ice core evidence, undermined, or falsified, the hypothesis that increasing atmospheric co2 will cause temperature to increase.
I now understand your position to be that the ice core data does not support the hypothesis. I don’t have an issue with that, but it seems to me to be an uninteresting position that tells us nothing.
Back in the syllogistic presentation of your argument, a showed how your conclusion, excuse me, the conclusion I took you to be making, could not be arrived at through deductive reasoning (logic). I also criticized a different formulation of yours, the “unlikely” formulation, where you seemed to be reasoning inductively.
So now, it seems, you’ve attacked inductive logic. Fine. I’m only puzzled because I’m not defending my logic. I’m questioning yours.
As to the amateur, non-reviewed, web-published, biased, inadequately referenced, Paint-generated evidence that you present, I find that it falls short of Tufte’s criteria for credibility. Even the author of the WUWT post stated it was not good enough for the quantitative analysis you make. Using it like that is also wrong, even if it’s accurate, because the implicit assumption is that temperature is determined solely by co2. No one has ever made that claim. With even a casual acquaintance with Hansen, you’d know that he attributes to co2 about a third of the ice age climate forcing. So you’re overreaching. That may be logical, but it’s not scientific.
Marc, the sources for the data were given right at below the plot.
Here’s the source of the temperature part of that plot: a professor of Paleogeology.
Here’s the source of the CO2 part: R. A. Berner and Z. Kothavala (2001) “Geocarb III: A Revised Model of Atmospheric CO2 over Phanerozoic Time” Am. J. Sci. 301, 182-204. Journal abstract here. Notice its extensive citation in further publications listed below the abstract, dating right up to the present.
Where’s the professional or scientific inadequacy?
You could have determined that the sources of the Temp-CO2 plot were entirely credible at any time before you chose to make an unfounded criticism.
The rest of your post doesn’t warrant serious attention.
Pat;
Do any of his posts? They’re pure predictable parrotting, as far as I can tell.
Marc tried hard, Brian. Look at it from his point of view: We must be wrong. His job was to show that. So, he cast around and cast around trying to find an angle of entry. One-by-one, the avenues he tried proved to be dead-ends. We can’t fault him for trying.
Marc comes across as an honest guy, but one who doesn’t really understand science. He’s bought the AGW “consensus,” probably assuming that people who share his politics are invariably the more honest. But they’re not.
My own personal enlightenment about politics and honesty came after many years of dealing with the outright and repeated lies of creationists (generally right-wingers, we can agree). Later on, I had reason to start checking how Mr. Noam Chomsky treated his sources. You may know that Mr. Chomsky, a linguist at MIT, is reckoned the top-rank left-wing intellectual. He’s written many dozens of very accusatory books and essays. I have access to the Stanford University Libraries, and discovered that he had systematically misrepresented his sources to make his political targets uniformly appear to be racist, callous, venal, and all-around cold-hearted people. I tracked his work across 35 years and his method was the same — taking quotes out of context, misrepresenting meanings, and even juxtaposing widely separated text to get the effect he wanted. Thirty-five years of character assassination. This gets you top rank among intellectuals, apparently. One wonders if anyone fact-checks anymore.
Anyway, he was as unfairly partisan as any creationist, but just infinitely more subtle.
So, I realized that the problem is ideological commitment. It makes people morally self-righteous. Left-wing or right-wing, they’ll all lie to get their way, justified because the goal is so very wonderful and right.
History shows, though, that ‘their way’ really means getting their hands on the levers of power and violently imposing their views on everyone else.
The lies revealed by Climategate and by Steve McIntyre’s tireless auditing show the AGW folks are clearly partisan, and demonstrate the characteristic trait of ideologically driven polemics. But the subtext is that they want power. AGW is just their lever, like religion is the lever of the far right. Ordinary folk buy into the lever that suits their personal prejudice and end up supporting their flavor of ideologues.
So, a pox on all their houses. The only route to sanity is to remain firmly apolitical and use reason as one’s judge.
Chomsky gained fame and tenure by proposing an inborn grammar parser in the brain, and the theory was popular and dominant for some years. He then decided to transplant his inborn rightness to saving the world. The graft was successful insofar as gaining him a new much larger and politically more ambitious fan base. Ever since, he’s been like a frothing attack dog, repeating the same message (corporations thrive by means of verbal mind control) endlessly.
If the real power-hungry on the Left take over, I give him about 3 months before they put him up against the wall just to shut him up.