Lance Endersbee (1925-2009): Civil Engineer, Academic, Scientific Sceptic, Mentor
Posted by jennifer, October 5th, 2009 - under News.
Tags: People
I NEVER met Professor Endersbee, but we corresponded by email.
He contacted me about six years ago when I was working on the Murray River and water issues. He expressed concern about Australia’s great artesian basin and over extraction of what he considered a finite resource.
We later corresponded over climate change issue. Lance believed we must try harder to understand the causes of natural climate change instead of assuming anthropogenic global warming. He was particularly interested in the oceans as a source of carbon dioxide. On June 24, 2009 he wrote:
“The relationship between CO2 and ocean temperature is ordained by the solubility relationship. I attach [see above] a chart showing my experience curve for the only reliable temperature records we have. It is difficult to argue against a correlation of 0.99.
I also attach [see below] my experience curve for Global average sea surface temperature and earth rotation. Note the correlation of 0.99.
The reason why SST, CO2 and LOD correlate so well is that they are all dependent variables. The external driving force is the independent variable. It is the electromagnetic field imposed on the earth which causes motion (rotation) and electromagnetically induced heating. The driving force is nominally from the Sun.
However, I do not believe that the Sun causes its own sunspots. I suspect the source is a pulsating emf from galactic sources. As health and time permit I will contemplate further.”
Lance died last week, on Thursday October 1. He was 84 years old.
Since his dead I have learnt that before his official retirement Lance was a world authority on rock behaviour and tunnelling, a former president of the Institution of Engineers Australia, and a recipient of its highest award, the Peter Nicol Russell Memorial Medal.
He had worked on the Snowy Mountain hydroelectricity scheme and was once Pro-Vice Chancellor at Melbourne’s Monash University.
Lance did not consider nature or climate benign and man rather puny in the scheme of things… unless we harness science and ideas in practical ways to protect and nurture civilization. Thus Lance championed big infrastructure projects.
He encouraged me at different times, suggesting that it was important to stay “candid and thoughtful”.
Lance Endersbee died last week, but his ideas will live on.
******************
Links and Notes
Lance Endersbee:
Emeritus Professor, AO, FTSE. ME, Hon FIEAust., Hon Mem Eng Inst Canada, F.ASCE. Former Pro Vice Chancellor, Monash University, 1988-9, Dean, Faculty of Engineering, 1976-88.
Extensive career in hydropower engineering and water resource projects 1950-1976 in Australia and overseas with Snowy Mountains Authority, Hydroelectric Commission of Tasmania, and the United Nations.
Vice Pres. Int Soc Rock Mechanics 1966-70. Pres. IEAust 1980, Cr 1967-80.
Recipient: Peter Nicol Russell award 1986, Chapman Medal 1967, Warren Prize 1963.
Book, A Voyage of Discovery, a history of ideas about the earth, with a new understanding of the global resources of water and petroleum, and the problems of climate change. 2005.
Blog posts:
http://jennifermarohasy.com/blog/2008/02/carbon-dioxide-versus-temperature/
On Line Opinion Author:
http://www.onlineopinion.com.au/author.asp?id=261
The Great Artesian Basin and Plutonic water http://www.onlineopinion.com.au/view.asp?article=1215
For pity’s sake Cohenite, read my post again. Most especially, read the penultimate paragraph.
I said nothing about “future outcomes”, except in the context of where the “future” represents time as an independent variable. In this case, regressions from times past cannot validly be used to predict future dependent variable values.
Nevertheless, I would argue that your quoted definition is not a tight one. I for one would not use it in teaching any of my undergrads. Read the last sentence in your quote, and compare with my statement that “[a coefficient of] regression indicates the proportion of the variability in the dependent variable that may be explained by the variability in the independent variable”.
Can you pick the not particularly subtle, but very important, differences? Moreover, can you discern what the implications are for using regressions?
BJ, why on Earth would you want to know anything other than the [future] proportion of the variability in the dependent variable that may be explained by the variability in the dependent variable? That’s what the R2 [model] does; what other future outcome do you think I was talking about? Why isn’t that a predictive correlation? As usual you have nitpicked yourself into the corner of infinite logical obscurity.
Furthermore BJ, this;
“one of the first lessons that a high school student is taught when learning about regressions is that they cannot be used to predict the path of the trajectory, or of any isolated y values, outside of the range bounded by the minimum and maximum x values used in the performance of the regression”
is as good a reason why the AGW alarmism about ‘tipping points’ and ‘runnaway heating’ should be treated with scorn.
Sod wrote:
Once again, I remind you that we are not measuring GMST at several diverse locations, we are measuring the actual temperature at those locations and using that data to create our GMST number. And while it’s certainly possible to state such an average with a higher resolution than the original measurements, such resolutions are not representative of our true knowledge of the average temperature, which is limited by the resolution of the our measurements! The temperature at Moscow does not help me refine the average of the temperature in Sydney, even if they show some correlation and even if it’s possible to use one to predict the other – simply because they are not measurements of the same thing.
In the context of global temperature averaging, Neil wrote:
“The precision of that kind of average cannot be any better than the precision of the least precise measurement you take.”
That’s not quite right. As you know, there are 2 types of error: methodological error and random error. Precision is about the latter. Precision is an index of how well repeated measurements of the same thing agree with one another. A fairly common measure for precision of *individual* measurements is the estimated standard deviation, or some multiple thereof.
Of course, your best estimate for the “true value” of whatever it is that you’re trying to measure takes *all* of the measurements into account–the mean, for example. If you use the mean value, then the estimated standard error of the mean (ESEM) is a good measure of precision. As you know, if you quadruple the sample size, you halve the ESEM, and double your precision. However that’s not consistent with what you said.
The whole point of temperature-averaging is to nail down surface temperature trends for the planet as a whole. Suppose that you use good sampling technique, and take 10 trillion measurements every hour. Further suppose that they all decrease (or increase) monotonically over a 20-year period. Then it would be reasonable to conclude that there’s a trend. Unfortunately, Nature has a sense of humor, and she ain’t likely to hand you the monotonic scenario on a silver platter .
Nevertheless it’s still possible to tease out trends using temperature-averaging, even though that’s distasteful to some people. In that case, you use garden-variety statistical concepts to analyze your data. But you need to be mindful about the distinction between precision and accuracy, which reflects the former as well as methodological error.
Larry Fields wrote:
Indeed. Note the fourth sentence. In this case (GMST) we are not making repeated measurements of the same thing, are we? The temperature in Moscow is not the same thing as the temperature in London or Sydney or anywhere else. Does measuring oranges tell you anything meaningful about the average size of apples? Does measuring the size of apples on your tree several weeks later help you refine the average you took earlier? In both cases the answer is a resounding “no!” because it’s not the same thing you are measuring! And while the second is somewhat more subtle, it’s definately not the same thing – although some people might believe otherwise and attempt to do this, once it’s pointed out that apples generally increease in size over several weeks, their arguement that it “matters” is hardly convincing, is it? And if we are making “corrections” to allow for the growth in between, then that would increase our uncertainty, not refine it, would it not? Yet this is what appears to be happening with GMST – we are adding orange size and correcting for growth based on estimates, not direct measurements, and coming up with numbers that certainly look impressive, but are somewhat “rubbery” and a long way from definitive. They may indeed be indicative, but are hardly what I’d call “proof”.
BJ, why on Earth would you want to know anything other than the [future] proportion of the variability in the dependent variable that may be explained by the variability in the dependent variable? That’s what the R2 [model] does; what other future outcome do you think I was talking about? Why isn’t that a predictive correlation? As usual you have nitpicked yourself into the corner of infinite logical obscurity.
logical obscurity is a pretty nice thing to mention, cohenite. here is a link that you did provide on page 2 of this discussion. as always, you cherry picked start and end date of your trends to be extreme points of the graph.
http://woodfortrees.org/plot/hadcrut3gl/from:1850/to:1880/trend/plot/hadcrut3gl/from:1850/to:1880/plot/hadcrut3gl/from:1880/to:1910/trend/plot/hadcrut3gl/from:1880/to:1910/plot/hadcrut3gl/from:1910/to:1940/trend/plot/hadcrut3gl/from:1910/to:1940/plot/hadcrut3gl/from:1940/to:1976/trend/plot/hadcrut3gl/from:1940/to:1976/plot/hadcrut3gl/from:1976/to:1998/trend/plot/hadcrut3gl/from:1976/to:1998/plot/hadcrut3gl/from:1998/to:2010/trend/plot/hadcrut3gl/from:1998/to:2010
why don t you educate us: what “[future] proportion of the variability in the dependent variable” do your trend lines suggest?
and what did really happen?
ps: and i noticed that you have posted again, without bringing up any correlation or proof of your wild claim about base periods…
Indeed. Note the fourth sentence. In this case (GMST) we are not making repeated measurements of the same thing, are we? The temperature in Moscow is not the same thing as the temperature in London or Sydney or anywhere else. Does measuring oranges tell you anything meaningful about the average size of apples? Does measuring the size of apples on your tree several weeks later help you refine the average you took earlier? In both cases the answer is a resounding “no!” because it’s not the same thing you are measuring
look Neil, your original statement was simply false, it would be a good start for further discussion, if you could simply admit that.
but as Larry pointed out, basically everything that you wrote afterwards is false as well. a simple example are tests, that don t allow multiple instruments. like test crash dummies. when you want to know what will typically happen to the dummy, you will do multiple tests and (shock!!!) AVERAGE the result.
according to you, a single test gives the best result…
Indeed. Note the fourth sentence. In this case (GMST) we are not making repeated measurements of the same thing, are we? The temperature in Moscow is not the same thing as the temperature in London or Sydney or anywhere else. Does measuring oranges tell you anything meaningful about the average size of apples?
If the sun were to double it’s output, would all the temperatures at the sites you named go up? If it were to halve it’s output, would all the temperatures at the sites you named go down?
Sod wrote:
No, you are suggesting that we can average over different manufacturers cars – that we can take an average of different things and that such an average is somehow more exact. How many different ways do I have to say it – WE ARE NOT MEASURING THE SAME THING! The basic premise of the central limit theory is that you are, so you cannot apply it!
Neil,
You’re missing the point. And sod isn’t smart enough to be wrong 100% of the time.
Let’s keep our eyes on the ball. We’re interested in global surface temperature *change*. Instantaneous global average temperature is not a particularly interesting concept in its own right. It’s a red herring for those who haven’t mastered the concept of subtraction, as in T2 – T1. For rapid humongous changes–above and beyond the piddling warming phase that we experienced back in the 80s and 90s–a single well sited temperature station *may* be able to pick it up. In an extreme case, that may answer the IF question. In a less extreme case, a sample size of 1 would be grossly inadequate.
And there are the equally interesting HOW MUCH and WHERE questions. S Sweden was affected more strongly by that particular round of global warming than was Sacramento County, where I live. Warming effects appear to be partly a function of *latitude*. Ditto for cooling effects.
Having a large network of *well-sited* and well-attended temperature stations would provide much more *useful* quantitative info than a single site.
Example 1: In the short term, it could be helpful for farmers trying to grow a given cash crop in a given marginal location. In contrast, the signal from a single temperature station in that area would be useful, but much noisier.
Example 2: In the long term, such a network may give us some early warning about the next major advance of continental glaciers. Then we’d have more time to weigh the suite of available options: absorbing climate refugees from high latitudes (or shooting them as they try to cut through the barbed wire), shifting to high-yield albeit less-tasty crops in the decreased arable land, cultivating a taste for rat-burgers, and even geo-engineering.
If we bury our heads in the sand, spout pious platitudes about apples and oranges, and pretend that a sample size of 1 gives more useful information about global temperature *change* than a network of *well sited* and well attended temperature stations, we’ll be shooting ourselves in the foot.
Preemptive comment. Yes, I agree that we should get rid of the garbage-data-generating temperature stations in parking lots, adjacent to barbecue pits in public parks, in the exhaust streams of AC vents, etc.
Bernard J,
Do you ever mention this to your undergrads?
What is the current trend in the mean global temperature anomaly?
Global cooling by 0.71 deg C from 1878 to 1911, for 33 years.
Global warming by 0.53 deg C from 1911 to 1944, for 33 years.
Global cooling by 0.48 deg C from 1944 to 1976, for 32 years.
Global warming by 0.67 deg C from 1976 to 1998, for 22 years.
http://www.woodfortrees.org/plot/hadcrut3vgl/compress:12/detrend:0.706/offset:0.52/plot/hadcrut3vgl/trend/detrend:0.706/offset:0.52/plot/hadcrut3vgl/trend/detrend:0.706/offset:0.97/plot/hadcrut3vgl/trend/detrend:0.706/offset:0.07
No, you are suggesting that we can average over different manufacturers cars – that we can take an average of different things and that such an average is somehow more exact. How many different ways do I have to say it – WE ARE NOT MEASURING THE SAME THING! The basic premise of the central limit theory is that you are, so you cannot apply it!
you are still moving the goal post. (there was no central limit theory in your first post. it was about trends, and “resolution of the measurements”)
and “average over different manufacturers cars” is going a little too far. but yes, the next test will most likely be on the next (improved) version of the same chassis or on a different version of the car (the slighly heavier diesel version?)
the average or the trend of results will still have a meaning and will still improve the results!
in practise, you might use the saem instrument multiple times. or different instruments on slightly different objects.
on temperature, we combine those techniques, with multiple measurements with the same instrument at the same location at different times, and more emasurements with different instruments at other locations. (and some of those locations are closer together than London and Sydney..)
the Trend over the avearges of those measurements will be always better than the “resolution of the measurements”. will you finally admit that?
Cohenite.
It’s simple, really.
It is considered inappropriate to use a regression to determine dependent variable values from independent variable values that lie outside of the range of independent variable values that were used to obtain the regression equation parameters.
When time (ie, “year”) is the independent variable, parameters from a regression based on data up to the present should not be used to project into the future. A simple linear regression does not automatically, or even reliably, inform us of the proportion of the future variability in the dependent variable that may be explained by the variability in the independent* variable; and to repeat, nor should such a regression be used to predict values of the dependent variable in the future.
Common sense should tell you this. Take some simple physical facts…
Temperature is predicted to increase by a unit increment for each doubling of CO2. For a constant rate (over time) of CO2 increase, the rate of temperature increase (over the same period of time) would ‘flatten’ – that is, deviate from linearity. Currently the rate of emissions is more exponential than constant (linear), so the projected rate of temperature increase is consequently affected, and depending on the final sensitivity value it might describe a trajectory anywhere from a ‘less flat’ asymptotic path to a temporarily exponential path.
In this range of alternatives a linear future trajectory is certainly one possibility, but without knowing a priori what the sensitivity actually is, it would be rash to cavalierly assume a linear continuation: much more factoring than a mere simple regression is called for.
Additionally, there are many other factors that positively and negatively force temperature over time. These factors are both natural, and human-induced, and the latter include modifiers of emissions such as emissions control and financial crises. How are such stochastic and emergent events accounted for by a simple regression?
Actually, I would have thought that the fact that simple regressions do not reliably inform about trajectories outside of their independent variable range is a good reason why ‘tipping points’ should be treated with greater care than is afforded them by denialists, and that more careful and appropriate analyses should be used – a regression has nothing to say at all about potential future tipping points.
As to “runaway heating”, that discounted meme is just hysteria promoted by the media, or taken up the nutcase fringe of AGW proponents, and fixated upon by denialists who are hell-bent on stirring the apparent (but not scientifically supported) exaggerations from the AGW camp. It is a strawman, and it has nothing to do with the subject at hand.
Note that just because simple linear regressions based on past data cannot inform us reliably about the future, does not mean that other (more comprehensive) models are not predictive. It’s all a matter of knowing what one’s statistical tools actually do, and of using one’s statistical tools appropriately.
Whilst the currently available data may approximately fit a linear trajectory (and it does not do so, well), at some point sooner or later in the future the regression parameters would fail to inform about changes in trajectory. That is why appropriate predictive tools should be used, and why the coefficient of a simple linear regression is not a “predictive correlation”.
(* just to point it out, when you quoted me here, you mixed up your dependent and independent variable references.)
What is the current trend in the mean global temperature anomaly?
….
http://www.woodfortrees.org/plot/hadcrut3vgl/compress:12/detrend:0.706/offset:0.52/plot/hadcrut3vgl/trend/detrend:0.706/offset:0.52/plot/hadcrut3vgl/trend/detrend:0.706/offset:0.97/plot/hadcrut3vgl/trend/detrend:0.706/offset:0.07
Drongo, you used the DETREND function to remove the linear trend from that data, and then plotted linear trendlines?
that is a pretty insane approach.
SJT,
“If the sun were to double it’s output, would all the temperatures at the sites you named go up? If it were to halve it’s output, would all the temperatures at the sites you named go down?”
Going UP?? NO!!!!!
We take night and day temps. The measurements on the night side are still going down even as that extra energy starts to warm sites on the day side.
Are you claiming that it is all the Sun now???
HAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHA
Old SOD,
“when my pool thermometer reads 25°C for 70 days and 26°C for the next 30 days, i get a meaningful trend.”
Did you record whether you averaged up or down with each reading??? For instance, if you averaged down for 60 days averaged up for 10 days, averaged down for 10 days and averaged up for 20 days you know the trend won’t be the same!!! So, was it really meaningful just averaging those whole numbers???
If you took multiple readings of the pool temp each day at the same time instead of one, you could average them and increase your accuracy for that measurement. You then might be able to use the daily averages for a more meaningful trend!!! (providing something in the thermometer or procedure wasn’t BIASING the measurement. For instance you were measuring closer to the heater output for that last 30 days or you were not viewing the thermometer at the same angle!!)
A global average temperature of hundreds or thousands of places and thermometers is outside the application of this theory.
Let me try a different direction to attempt to show why this particular stretching of theory breaks down for our intended use, proving or disproving AGW.
You measure 15c at Denver with a humidity of 20%.
You measure 15c at Death Valley with a humidity of 20%.
Is the amount of energy in a cubic foot of air the same with these 2 measurements??
Probably not, but we also need the air pressure!!!
Yes you are measuring the temperature of the air at those thermometers. Unfortunately, what Global Warming is all about isn’t simple air temperature, it is about the amount of ENERGY in the system!!! In other words, we measure temperature for a PROXY of Global Warming. Unfortunately it is a poor proxy.
Without knowing the DENSITY of that cubic foot you have little idea of the energy content and therefore, AVERAGING THE TEMPS IS AN EXERCISE IN POINTLESS ENDEAVOURS IN RELATION TO AGW!!!!!
As Neil has been trying to tell you. IT AIN’T THE SAME THING!!!!
BJ; something we can get out of the way is this assertion by you;
“As to “runaway heating”, that discounted meme is just hysteria promoted by the media, or taken up the nutcase fringe of AGW proponents, and fixated upon by denialists who are hell-bent on stirring the apparent (but not scientifically supported) exaggerations from the AGW camp. It is a strawman, and it has nothing to do with the subject at hand”
http://climatechangepsychology.blogspot.com/2008/12/james-hansens-agu-presentation-venus.html
In addition read something from tipping point Glikson or anything by Gore; the notion of runnaway is entirely a meme of the AGW crowd enough of whom are on record as justifying gross exaggeration to get their ‘message’ across; so drop that garbage if you want to be taken seriously.
Now you say; “When time (ie, “year”) is the independent variable, parameters from a regression based on data up to the present should not be used to project into the future. A simple linear regression does not automatically, or even reliably, inform us of the proportion of the future variability in the dependent variable that may be explained by the variability in the independent* variable; and to repeat, nor should such a regression be used to predict values of the dependent variable in the future.”
That being the case what is your opinion on this;
http://landshape.org/enm/a-semi-empirical-approach-to-sea-level-rise/#more-2618
Cohenite,
Thanks for reference to Climate Change:The Next Generation.
The comments by Obama heading up the site was revealing……” Because the truth is that promoting science isn’t just about providing resources, it’s about protecting free and open inquiry !!!!!!! (Yeah Right)
It’s about ensuring that the facts and evidence are never twisted or obscured by politics or ideology”!!!!!!!!!!!!!!!!!!!! Barack Obama. Ahahaha
AAhahahahahahaha . Ahahahhahaha.
Isn’t that just a classic. When a salesman wants to sell you a product that he knows is defective in a certain aspect and he knows that you know the defect, he’ll blatantly point out that the product does not have this defect.
Would you buy a used car off this guy?
Cohenite.
Let’s get this out of the way… I challenge you to go to RealClimate and assert that “scientifically-based proponents of anthropogenic global warming promote the idea of a ‘runaway’ warming”. Then we’ll see just what the professional climatology community says about this idea. If your insistence is correct, you will not be challenged on such a claim.
I note though that Hansen’s thoughts (as detailed in the pdf) that runaway warming could occur are predicated on an additional forcing that is “as small as 10-20W/m2″. He thinks that this would occur if ALL coal, plus ALL tar sands and All tar shales, were burned. I admit that I do not know off-hand if such a huge amount of combustion would give the necessary forcing for runaway, but two things are fundamental here – 1) if all such sources of carbon-based energy are combusted, the planet will be screwed from ocean acidification and other pollution and ecological breakings-down effects irrespective of any warming, as well as humanity being screwed in having no more energy-dense carbon fuel, and 2) I harbour a hope that humanity would not be so self-indulgent as to go so far down the path of such total carbon combustion as this.
If you have the specific analysis, based in physics, that contradicts Hansen’s opinion on the forcing resulting from the combustion of all coal, tar sands, and tar shales, then please share them with us.
In addition, whatever the caveats to his thoughts, Hansen himself remains as one single voice in the consensus, so you have not in any way established a case that demonstrates that runaway warming is subscribed to by the majority of climatologists/physicists.
And none of the above distraction changes the fact that you are avoiding a direct comment about the appropriateness of regressions in the context to which you seek to apply them. You have not in any way made a case that is defensible in science or statistics.
As to your landshape link, consider it carefully. It actually proves my point, notwithstanding the fact that a broken (analogue) watch is still correct twice a day.
Did you record whether you averaged up or down with each reading??? For instance, if you averaged down for 60 days averaged up for 10 days, averaged down for 10 days and averaged up for 20 days you know the trend won’t be the same!!! So, was it really meaningful just averaging those whole numbers???
i assume that the person doing the measurement is aware of the basic rounding rules, or that the instrument is already giving full numbers.
the example wasn t perfect. the real results would show the higher temperature slowly moving in, with a period that will have a change between the two numbers.
If you took multiple readings of the pool temp each day at the same time instead of one, you could average them and increase your accuracy for that measurement. You then might be able to use the daily averages for a more meaningful trend!!!
please explain this to Neil. he calls this “comparing apples to oranges”. and claims that, as you are not measuring the same thing, you can t get a more meaningful trend this way….
A global average temperature of hundreds or thousands of places and thermometers is outside the application of this theory.
this is completely false. the oceans are not different from a big pool. the very same principals apply.
You measure 15c at Denver with a humidity of 20%.
You measure 15c at Death Valley with a humidity of 20%.
Is the amount of energy in a cubic foot of air the same with these 2 measurements??
Probably not, but we also need the air pressure!!!
Yes you are measuring the temperature of the air at those thermometers. Unfortunately, what Global Warming is all about isn’t simple air temperature, it is about the amount of ENERGY in the system!!! In other words, we measure temperature for a PROXY of Global Warming. Unfortunately it is a poor proxy.
Without knowing the DENSITY of that cubic foot you have little idea of the energy content and therefore, AVERAGING THE TEMPS IS AN EXERCISE IN POINTLESS ENDEAVOURS IN RELATION TO AGW!!!!!
that is fine. so just provide us with 150 years of data showing the “amount of energy in the system”, and we will stop that thermometer thingy and use your data.
we are using temperature, because we have the data. we are not using an energy balance, because we don t have the data. simple.
for the trend, the difference will be tiny anyway. unless you can show us a massive change in earth energy capacity.
i am waiting…
As Neil has been trying to tell you. IT AIN’T THE SAME THING!!!!
no. what Neil said was: the problem is the “resolution of the measurements”. he moved pretty far from that completely false claim…
Old Sod,
“i assume that the person doing the measurement is aware of the basic rounding rules, or that the instrument is already giving full numbers.”
The question was not about whether the measurement was done correctly, although that is always a good question. The question was to illustrate that the trend can still be wrong. You can NEVER forget the possible errors. The trend could be flat and you would not know it due to the error.
Even here the conditions could bias the measurement. Did you start measuring at a shady spot and by the end you were measuring in a sunny spot? Is the pool filter on?? Are you measuring at the same depth? Is the pool used heavily some days and not at all others?? Do you not put on a cover every night?? blah blah blah
The theory is based on a LIMITED number of sources for random errors(comparatively). By changing so many of the conditions under which the measurement is made you invalidate the application of the theory. By measuring something DIFFERENT (location) you invalidate the theory. That is, if you infilled missing measurements with a measurement from your neighbors pool you would invalidate the theory. If you measured the temps in 5 pools close to your house and averaged them it would not increase the accuracy of that days measurement of your pool and would not give you a higher accuracy trend for your pool. (are the pools the same volume, do they have the same albedo, do they get the same insolation, are the measurements done in an equivalent fashion…)
You also seem to assume that computing a trend is equivalent to averaging one site’s measurements. It simply isn’t the same thing. You are not doing a simple AVERAGE when you compute a trend.
“please explain this to Neil. he calls this “comparing apples to oranges”. and claims that, as you are not measuring the same thing, you can t get a more meaningful trend this way….”
No, I must continue to try and get across the idea to YOU that multiple measurements of your pool at the same time and in the same location can be averaged for a probable increase in accuracy and a probable valid trend and that it is an appropriate application of the theory. Measuring air temps at different locations, times, and conditions and averaging them is NOT!!!! This theory has nothing to do with computing a trend from single measurements.
As you are only taking one measurement each time you are not increasing the accuracy so you have no basis for claiming sub 1c accuracy on the trend. Once time has passed before the next measurement, the idea that the random errors from one measurement to the next averages out becomes invalid due to too many changing conditions. In other words you have done nothing to reduce the possible random errors and you have increased the probability of bias in the measurements which invalidates the theory in these situations.
Let’s list again the problems involved in accepting that our measurements can be used to obtain a valid global average.
1) different locations
2) different times
3) different air pressure
4) different elevation
4) different humidity
5) different wind conditions
6) different thermometers
7) different TYPES of thermometers
8) merging brightness and thermometers
9) different methods of achieving di-urnal averaging for some stations
10) procedural differences including siting and methods of reading and transcribing data
11) ADJUSTMENTS (whether valid or not)
(feel free to jump in and add anything I missed)
“that is fine. so just provide us with 150 years of data showing the “amount of energy in the system”, and we will stop that thermometer thingy and use your data. ”
Since you do not have appropriate data to evaluate a situation you will continue ASSuming extreme conclusions. Yup, that sounds like AGW allllrighty.
By the way, since the thermometer thingy is PART OF determining the energy in the system, exactly why would you stop doing it?? Do it better maybe but stop?? Oh yeah, you might get data that contradicts your BELIEF SYSTEM!!
HAHAHAHAHAHAHAHAHAHA
“we are using temperature, because we have the data. we are not using an energy balance, because we don t have the data. simple.”
I am using your head to pound this nail because I do not have a hammer!!! I probably won’t get the nail in and I will either seriously damage you or you will pound ME!!
HAHAHAHAHAHAHAHAHAHA
I am not talking about energy BALANCE, only about the actual energy content of the atmosphere at the time of the measurement.
“for the trend, the difference will be tiny anyway. unless you can show us a massive change in earth energy capacity.”
The AGW trend is already tiny. No need for any massive changes in the energy in the system to invalidate it (although the heat capacity difference of high humidity air and low humidity air is pretty massive meaning your error is already potentially massive).
Isn’t that what this whole discussion is really about?? CO2 is supposed to contribute less than half of AGW according to the IPCC. So, we are not talking about even .6-.7c over 30 YEARS, we are only talking about a maximum of .3-.35c OVER 30 YEARS!!!! It is PROBABLY less. As virtually everything has changed in the system, except for the basic physics, during those 30 years, the idea that this particular theory of measurement could apply and give greater accuracy is simply ludicrous.
Sod wrote:
You are confused. There is no goal post movement. I suggested that the resolution was insufficient. You replied that CLT could be sed to improve it. And I replied that it does not apply.
So crash testing a brand new falcon will help me refine my estimate of the likely passenger injuries sustained from a crash in the previous model? Or that a crash test on an E-class Mercedes Benz will help me refine data I collected on an A class Mercedes Benz? You must be joking.
Multiple instruments that measure the same thing would help in this context. And it is not always possible to use exactly the same object – eg destructive testing. Using, eg, several instances of the same product from the same production line (”identical” products) is the only reasonable option in such cases.
And therein lies the problem. It can be raining here and now, but dry only several hundred metres away. It can be 30C here, and 25C 5km away, let alone 1000km away. But even ignoring that for the moment, where can I measure GMST? I cannot. I can create such a construct, and indeed several institutions do – HADCRU, GISS, RSS, UAH for example. These different groups do not produce data that match each other, so clearly methodology and data sources matter, yet you seem to be suggesting that all of these are better estimates of GMST than the magnitude of differences between them. They are all supposed to be estimates of GMST, and all are quoted to within 0.1C, yet the differences between them exceed 0.1C in many cases. They can’t all be right.
Providing we meet certain conditions, I have already agreed to this. My arguement is not that such techniques do not work, but that you cannot apply them because the assumptions on which they are based are not met. You have not yet provided any arguements to refute this – the only time you have come close is in the very post I am replying to, and even then you appear to be suggesting that there is some distance limit that may apply, but it’s very vague exactly at what point we have stopped measuring the “same thing”.
BJ, Hansen clearly says Earth could end up being like Venus; the cause of the coal and tar burning is neither here nor there [and there are a number of analyses which put the total CO2 concentration from that scenario at ~ 1000ppm; if that concentration was going to do a Venus it would have done so in the past]; but there are plenty of other apocalyptic scientists out there; Glickson;
http://webdiary.com.au/cms/?q=node/2699
The lads at MIT;
http://wattsupwiththat.com/2009/05/26/how-not-to-make-a-climate-photo-op/#more-8038
Stephen Schneider, ANU;
http://www.theage.com.au/environment/climate-threat-to-heritage-sites-20090801-e592.html
Thomas Schelling, Kofi Annan, Clive Hamilton and the media;
http://www.theage.com.au/opinion/stop-wasting-time-and-save-the-planet-mr-rudd-20090613-c6oo.html?page=-1
If pro-AGW scientists were promoting a realistic view don’t you think they should repudiate such bunkum; the fact they don’t makes them culpable and collaborators in this scare-mongering.
Cohenite.
So, how did your challenge at RealClimate go?
And excuse me if I don’t accept media and blog sources as accurate representations of what the climatology community agree on with respect to Venusian runaway heatings. Perhaps you could refer to peer-reviewed, Web of Science indexed professional literature to make your ‘point’?
And this is all beside the real point anyway. I guess that as a lawyer and as the secretary of the Australian Climate Sceptics Party you are bound to defend the innocence of your ‘client’, and when you client happens to be the inapproprite use of simple linear regressions to make claims beyond the scope of regressed data, you are bound to attempt to distract the jury with irrelevant strawmen arguments.
Well, that isn’t going to fly. In case you issed it the first time, I’ll repeat it for the jury’s benefit:
Your use of regressions to predict future temperatures is counter to the statistical guidelines for using regressions, and in any event it does not reflect the methodologies that climatologists use to predict the rise of temperature under different emissions scenarios.
Of course, if you believe that your statement that:
you will use the regression to predict, with “78% probability” of accuracy, the “movement” of annual mean global temperatures for the next ten years.
This should be interesting, for a number of reasons…
BJ, Hansen clearly says Earth could end up being like Venus; the cause of the coal and tar burning is neither here nor there [and there are a number of analyses which put the total CO2 concentration from that scenario at ~ 1000ppm; if that concentration was going to do a Venus it would have done so in the past]; but there are plenty of other apocalyptic scientists out there; Glickson;
And I have talked to one once who said it was very unlikely, since we had much higher levels in the past. There is, of course, a lot of debate and differences of opinion over the extent of AGW, but no argument over the basic facts and the need for action now if we want to prevent serious consequences. As Dirty Harry said, do you feel lucky, punk?
Sod,
What is the current trend in the mean global temperature anomaly?
Global cooling by 0.71 deg C from 1878 to 1911, for 33 years.
Global warming by 0.53 deg C from 1911 to 1944, for 33 years.
Global cooling by 0.48 deg C from 1944 to 1976, for 32 years.
Global warming by 0.67 deg C from 1976 to 1998, for 22 years.
The point you just don’t seem to understand is that the oscillating warming during this period is as natural as the oscillating cooling by a similar amount.
BJ; just drop the RC puerility please; to clarify further; the independent variable here is CO2 and the dependent variable can be temperature, sea level, hurricane intensity etc; you say:
“Your use of regressions to predict future temperatures is counter to the statistical guidelines for using regressions, and in any event it does not reflect the methodologies that climatologists use to predict the rise of temperature under different emissions scenarios”
This from sod and little will’s favourite reference:
“Regression analysis is widely used for prediction (including forecasting of time-series data”
I gave you some links to the Ramsdorf controversy about predictions of sea level increase before; Ramsdorf used regression [all tarted up] to generate those predictions; he was wrong; here is another link; Stokesy and I have an exchange at the end and to assist you I want you to think running means and why a break model is preferable;
http://landshape.org/enm/rahmstorf-2007-discredited/#disqus_thread
SJT;
“As Dirty Harry said, do you feel lucky, punk?”
Actually, I AM lucky PUNK!!!
How many 57 year olds do you know who have been riding sports motorcycles like idiots for 40 years with only a couple of broken bones to show for it?!?!?!?! Those who haven’t been hospitalized, like me, are ALL extremely LUCKY!!!
I am also still quite healthy and run and bicycle regularly!!
YUP, just call me LUCKY!!!!
Of course, with AGW, I don’t NEED to be lucky cause it AIN’T HAPPENIN’ Y’ALL!!
HAHAHAHAHAHAHAHAHAHAHAHAHAHA
Bernard J., although I’m not educated enough to understand much of your arguments, I must say as a skeptic that you have introduced a level of sophistication to the AGW side of the debate here that has been sadly lacking. I’ve also noticed that Sod and SJT have improved in the quality of their arguments recently, perhaps inspired by your standards.
Please continue contributing to this blog so that there can be more robust debate. I’m afraid our side has had it too easy for far too long.
Derek Smith.
I have just posted another piece that seems to have evaporated into the moderation bin. Assuming that it passes muster, and in light of your very temperate words (for which I thank you), I should probably apologise in advance if there is a slightly snarky undertone to what I typed.
It is encouraging when debaters in any issue are actually open to opposing ideas, and it is especially encouraging that someone here has been willing to consider that perhaps I trying to make a valid point.
Truly sceptical arguments from both sides of the fence benefit from properly informed participants. If cohenite is going to demonstrate where climatologists, physicists and statisticians have it wrong, he and his colleagues need to do it using the appropriate analytical tools, and to use them in appropriate contexts.
If he can do so, and show that there really is more defensible scientifically-based doubt about the state of the body of AGW understanding that, then I for one would be at the front of the pack to hear about it.
Bernard J wrote:
Couldn’t agree more. Only problem is, I’m a “luke-warmer” – I expect that we should see some warming from increased CO2, as this is basic physics. However, what data we do have does not appear to suggest – to me anyway – that we can expect to see the positive feedbacks that are apparent in the models. I would have no problem with the climate modelling community if they would simply acknowledge that it appears they have previously over-estimated positive feedbacks – hey, they’re doing the best they can, and it’s certainly one of, if not the, most complex system anyone has ever tried to model, so mistakes this early in the piece are to be expected. I cringe when I hear the proclaimations of impending doom based on such models when there is never any mention of uncertainty and so on. Mayhap it’s not their fault that this happens, but certainly they play their part in making it so. Nor am I much impressed when the forcing (anthro CO2)is greater than anticipated, the resultant (GMST) is lower than anticipated, and yet we hear that “things are worse than we thought” – arrrgh! The ducking and weaving over the recent “stalling” of AGW – “it’s in the oceans”, “err, well, no it’s not in the oceans it may have been radiated back to space” etc etc – is equally cringe-worthy. As is the almost religeous dismissal of people like McIntyre who find mistakes – that such people did not learn from the hockey-stick affair shows appalling hubris and arrogance (witness SM and RC “slogging it out” over the latest kerfuffle – use of “upside-down” proxies – even when the originator of the dataset in question says it has been used improperly, they still keep up with “it doesn’t matter”!).
Over-selling of AGW is much more alarming than the AGW we have so far experienced itself – when we finally do have a good handle on what’s happening, these people will be ignored because they cried wolf once too often. And that really could be fatal to humanity. Yet whenever anyone points this out, they are labelled as a “denier” or in the pay of “big oil/coal” and their concerns dismissed without discussion. Hey, the alarmists may be right even though it doesn’t look like it as yet, but with so much unknown about climate and such short (in climate terms) reliable records, to make “90% certain” projections is, well, risky at best and reckless at worst. IMO, of course.
Bernard J wrote:
I would be interested in hearing your thoughts on Spencer’s latest WRT cloud coverage as a forcing rather than a feedback.
IMO, Spencer is one of the few that questions everything. Sure, he has to “back-pedal” often, but at least he’s trying to poke holes in things rather than just accepting them as gospel – we always learn something from such people, and seeing the treatment he gets from the AGW “faithful” makes me want to shake them and ask “Do you have no knowledge of history? Can you not see that this is the very essence of the scientific method? Can you not see that this is a reflection of the Royal Society’s motto?” Arrgh! Enough…
Bernard J., thanks for the comment and I look forward to reading your other response when it comes through. Now if someone from the AGW scientific fraternity could just stop people like Tim Costello from claiming that the recent earthquakes and tsunamis were of climate change origin(Adelaide Advertiser), there would be less huffing and puffing on the skeptics side.
Cheers.
Well said Neil, my position is very similar.
The mention of temperature and pressure measurements reminded me of a question I have asked on blogs before but not received any answer, as the temperature on Venus at one bar is close to 37o C at 67n ( Magellan Probe, Jenkins)and is in the ballpark for earth temps considering increased insolation because of its position, and internal heat supply, what happened to runaway Greenhouse because of its preponderant CO2 atmosphere?
Julian,
Do you have a link for that Venusian temperature measurement? It sounds a bit odd. The insolation there is roughly 4X what we Earthlings are accustomed to. (Earth is roughly twice as far from the sun, and relative insolation is approximated by an inverse-square law.)
You did specify 1 bar pressure. My limited understanding is that the Venusian atmosphere is considerably denser than that of our fair planet. If so, that 1 bar would be high in that atmosphere, rather than at the toasty surface. Or are there hidden correction factors at play here?
Testing, 2 replies to Larry didn’t get through
Larry,
For some reason links are not coming through, Google Venus temperature pressure and it is link one and two , data sync and nov stanford edu
with four times TSI temperatures should increase by a quarter of that according to Lief Svalgaard which is still way bellow the observed at one Bar which is around 59km.
Either Co2 is cooling, I think there has been at least one paper on that, or sulphates are doing the job. There is an article in today’s Australian, “Heat is on Global Warming ” that has a cheap solution to warming if it does happen.
Larry,
got a bit muddled with the third attempt, the observed temperatures are way below what they should be, given the increased TSI, yes?