Comments on: Another minimum? What about an optimum?

By: Graham Young

Graham Young — Thu, 03 Jan 2008 06:54:21 +0000

Yes, but youre method is wrong. No-one is arguing that there hasn’t been an increase in temperature during the last century, or even the last half century. We’re not talking about trends, we’re talking about whether there has been a hotter temperature since 1998.
Trend analysis can’t define a peak.
And it wasn’t invented by Gauss to eliminate the problem of starting and end points but to predict future values based on past values in a linear model. It was used to predict the paths of celestial bodies in the first place without solving the more difficult equations to do this exactly.
I thought you were using NASA, which is probably the only data-set to show 2005 this high. According to NASA, because of the sampling error, 2005 and 1998 are actually “dead heats”. They admit to some pretty interesting techniques which lead to the higher potential measuring error, such as estimating Arctic temperature up to 1200 kms away from temperature stations! That’s the distance from Brisbane to Cairns.
I think you need to do a bit more googling in these areas to get your head around some fundamental science concepts like measurement error. BTW, it’s normal to state temperatures as an anomaly against the average of the period 1961 to 1990. It doesn’t make any difference to your figures, but just further confirms that whatever training you might have is not in this area.
I’m going to terminate this discussion. It’s completely pointless and taking up too much of my time. If you want to get some better understanding of some of the underlying concepts by arguing with others you’ll have to use someone else’s blog at least until I post something else on global warming. All that appears to be happening here is that you’re trying to distract from real points by arguing about red herrings. No-one is arguing that trends aren’t valuable, or that the world hasn’t got hotter. The question posed was what are the climate risks to mankind, how does CO2 play a part, and to what level should we reduce it, or might it be better even to have a little more of it. The fact that 1998 was the hottest year yet was not really particularly germaine to the whole argument.

By: JM

JM — Thu, 03 Jan 2008 03:45:43 +0000

>using every value in the array it is not an appropriate way to say when a peak was reached.
In your method, yes, but not in the method I proposed which uses a moving 10 year set, thereby avoiding the rookie mistake. You use every value, I only use the most recent ones so I can see if there have been changes in the trend. The technique is perfectly applicable, yours is not.*
> the two cool years have a disproportionate effect on the result.
I agree, but only if you’re using fake data, and making the rookie mistake – which you are – of never moving the start point. If you use a shifting window of 10 years like I did, those “cool years” drop out and you get a realistic picture of behaviour over the last 10 years.
Which is your point isn’t it? That recently the warming has stopped? We have to exclude older data if we are to see any recent cooling. I’ve done exactly that so I can be fair to your argument. (“Give it a fair shake rather than artificially rigging it my way”, you might say.)
But guess what? We don’t. The warming trend has not stopped.
And if you don’t like 10 years as a window, try 5 (at the cost of greater error) – it’s still shows increasing temperature. Hell, let’s really push out the boat – try 3 if you must (I wouldn’t), 2 would be ridiculous. Every option – long and accurate, or short and error-prone – shows increasing temperature, not a stand-still and not a decrease.
> If 2005 was warmer than 1998 we wouldn’t be having this argument, but it wasn’t.
It was. 1998 was not a peak.
>So what’s your source?
NASA. They call it “GISStemp” because it comes from the NASA Goddard Institute for Space Studies, and is the “GLOBAL Land-Ocean Temperature Index”. It measures average global temperature.
But you really didn’t have to ask, I told you in my previous comment:- “The NASA data is available here: http://data.giss.nasa.gov/gistemp/tabledata/GLB.Ts+dSST.txt (It’s a small file)”. Nobody disputes this dataset, and nearly everybody uses it. (Some prefer the Hadley Center data which is only triflingly different.).
I even extracted a subset and posted it in a comment so you could see it, and check for yourself. I assume you did that? Right?
Your argument is in shreds Graham. Concede.
* But if you want to argue against the whole mathematical, scientific, public policy and business community – everybody really – that a standard technique invented by Gauss, one of the greatest mathematicians who ever lived, precisely to get rid of the problem of starting and ending points, and used in almost every data analysis of any kind for every purpose since the early 19th century, is wrong? Be my guest. I wouldn’t even have time to buy popcorn, that’s how long you’d last.

By: Graham Young

Graham Young — Wed, 02 Jan 2008 22:57:05 +0000

Yep, average was the wrong word to use, but I think you are deliberately missing the point. There were some cool years just after 1998 which cause the trend line to be upwards. Because the line is set using every value in the array it is not an appropriate way to say when a peak was reached. And because of the method of calculating a best fit linear trend line, the two cool years have a disproportionate effect on the result.
Note that you don’t want to deal with the thought experiment because it illustrates that you are using inappropriate techniques, and prefer to deal with the “real” data. In which case, where are you taking your data from? If 2005 was warmer than 1998 we wouldn’t be having this argument, but it wasn’t. So what’s your source?

By: JM

JM — Tue, 01 Jan 2008 03:04:23 +0000

Graham
“… you are effectively confusing averages with absolutes.”
I have no idea what this means. Until you explain it in detail I will treat it as the gibberish that it appears to be. You’re making the claim, you make the case. If I’m wrong, I’ll concede.
“If I put my oven on to warm to …”.
I don’t give a stuff about your oven or an artificial example (from which you draw a fantasy conclusion). I care about the real world. Which has real data I gave you and you’ve ignored in favor of made-up cooking lessons.
Let’s discuss the real data, not made-up stuff that might or might not be “like” the real thing. Back in reality land, things are a bit more serious than that.
Now that you’ve run the numbers I presume you know there is no cooling of the climate, only warming. Thank you for conceding that at least.
Let’s get real:- 2005 was as hot or hotter than 1998. 1998 was not a peak, just a way-point.
The same as all other “peaks” like 1970 (14.03C), 1973 (14.14C), 1981(14.27C), 1990 (14.38C), 1998 (14.57C) and 2005 (14.62C). Note the trend. Run the whole dataset and it’s confirmed:- warming at a consistent rate of 0.017C per year since 1970. Since 1996, warming at 0.019C per year – an acceleration, not a stand-still, and certainly no cooling.
If there is no cooling, as you now concede (because you want to talk about ovens instead), there is no basis for your blog posting. And there is also no point. There can be no trade-off between CO2 and cooling if there is no cooling to begin with. Which is what I said originally.
You could at least be gracious instead of covering your ego.
If you persist in this “cooling” nonsense of yours now that you clearly understand it to be false, you have no integrity. Intellectual, scientific, economic, political or otherwise.

By: Graham Young

Graham Young — Mon, 31 Dec 2007 13:03:43 +0000

JM, you are effectively confusing averages with absolutes. If I put my oven on to warm to 100 degrees, which it takes 10 minutes to do, and then turn it down to 70 degrees for the next 2 minutes and then increase it to 90 for the next 8 minutes you will get a straight line trend for whatever 10 year period you choose that says temperature is increasing.
That won’t take away from the fact that it peaked at 100 degrees. The average temperature will be higher after the peak than before, but the average doesn’t tell you much about the peak.
The reason that the trend line will still be up from 100 degrees is that I dipped temperature below the final point just after it. If I hadn’t done that, then the trend line would be different, and if temperature stays at around the current level for long enough, or goes a little bit longer, and you take a longer term view, then the trend will have actually been down. Illustrating my point about the problems with trend lines and where they start and finish.
And irrespective, the trend line can’t take away from the fact that it was never hotter than 100 degrees, even though the average was higher after 100 than before.
You’re the one who is deceiving yourself! Assuming this is more than an exercise in discovering what Excel has to offer in the way of charting software.

By: JM

JM — Sun, 30 Dec 2007 18:27:44 +0000

“Certainly using a straight line fit is a nonsense as it depends far too much on where you start and end.”
You obviously haven’t tried it. Even if you use 1998 as a base year, you still get an upward trend, not downward.
In any case, the choice of base year doesn’t affect linear regression, only naive, ignorant fits apparently used by the likes of Dr. Whitehouse where you draw a line from the first number to the last and ignore all the ones in between.
Pick any 10 year series since 1970. 1971, 1972, 1977, 1997, 1998 – I don’t care.
You’ll always get an upward trend.
The NASA data is available here: http://data.giss.nasa.gov/gistemp/tabledata/GLB.Ts+dSST.txt (It’s a small file)
I’ve extracted the 1970-2006 annual (January-December) averages, converted to absolute Celcius as outlined at the end of the file, and pasted them to the end of this message. Should make it easier for you.
Cut ‘n paste to Excel, create a line chart and add a linear regression trend line. (Click on the chart, select “Chart->Add Trendline” from the top menu.) Easy, shouldn’t take more than a minute.
1998 was hot because there was a very strong El Nino effect that year. 2005 was just as hot, if not hotter.
Whether you object to me calling it an outlier or not, is something I don’t really care about. The result is never cooling, always warming.
I’ve been very clear on what’s an acceptable level of Co2 – one that doesn’t destroy our economies or our food chains.
The 1C we’ve had since 1900 is threatening both of those. The extra degree or two we have in the pipeline risks much greater damage.
Burning the rest of the oil and getting another 4-6 degrees is very risky.
Come back and tell me when you’ve found the elusive cooling. It isnt’ there. The likes of Dr. Whitehouse are educated and should know better. They are deceiving you.
————————————————————————————————————————————————————-
NASA Giss Land-Ocean Temperature, Annual Degrees Celcius
Year Avg Temp
1970 14.03
1971 13.9
1972 14
1973 14.14
1974 13.92
1975 13.95
1976 13.84
1977 14.13
1978 14.02
1979 14.09
1980 14.18
1981 14.27
1982 14.05
1983 14.26
1984 14.09
1985 14.06
1986 14.13
1987 14.27
1988 14.31
1989 14.19
1990 14.38
1991 14.35
1992 14.12
1993 14.14
1994 14.24
1995 14.38
1996 14.3
1997 14.4
1998 14.57
1999 14.33
2000 14.33
2001 14.48
2002 14.56
2003 14.55
2004 14.49
2005 14.62
2006 14.54

By: Graham Young

Graham Young — Sun, 30 Dec 2007 15:26:50 +0000

JM, I see you’re still trying to avoid the infrared issue. I well understand how the “greenhouse” effect works. The issue is how the sun’s energy gets here, not what happens after it does. You stated categorically that it only got here as infrared, which is garbage.
Not the only thing you have problems with. On what basis can a peak in temperature be a “statistical outlier”? You’re misusing statistical analysis, that’s if you really have any understanding of it. Certainly using a straight line fit is a nonsense as it depends far too much on where you start and end. It could quite easily still show an “up trend” when temperature had been declining for quite some time.
I notice you’ve dodged the issue of what the level of greenhouse gases ought to be. What is the range? And what is the climate that you consider we are “evolved for”? Does that include ice ages?
Or will you just label any intelligent assessment of the issues as “denialist” again? (BTW, my understanding is that Whitehouse believes in the greenhouse effect, so I’m vaguely amused at your continual branding him as a “denialist” because he reports some things that you don’t agree with.)

By: JM

JM — Thu, 27 Dec 2007 18:26:18 +0000

Sorry Graham, you asked me a question and I didnâ€™t answer it.
â€œit’s reasonable to ask what the optimal level [of CO2] ought to be.â€
I think I addressed this a couple of days ago. The optimal level is the level (or range rather) that sustains the climate we are evolved for and in which we have constructed our lives and economies.
The climate appears to go through phase changes (what is popularly called â€œtipping pointsâ€), rather than gradual linear evolution. This shouldnâ€™t be surprising in a complex system.
Therefore we donâ€™t have the luxury of choice. We donâ€™t know where the phase boundaries are, but we are clearly getting near one when you look at the melting of very old ice combined with very rapid climate forcing.
Any idea of choice is like â€œhow many angels can dance on the head of a pinâ€. Intellectually stimulating in some contexts, but entirely irrellevent to the main event.

By: JM

JM — Thu, 27 Dec 2007 17:36:44 +0000

> I’m sorry you didn’t have anything better to do with your time on Christmas morning
That’s ‘cos I prepare them beforehand, fact check them, and post them when I have a spare minute.
Graham, when I’m wrong I concede, what do you do?
I’ve been wondering what it would take to make you see the error of your thinking. How about a quote from your mainmain Dr. Whitehouse?
“In principle the greenhouse effect is simple. Gases like carbon dioxide present in the atmosphere absorb outgoing infrared radiation from the earthâ€™s surface causing some heat to be retained.
Consequently an increase in the atmospheric concentration of greenhouse gases from human activities such as burning fossil fuels leads to an enhanced greenhouse effect. Thus the world warms, the climate changes and we are in trouble.
The evidence for this hypothesis is the well established physics of the greenhouse effect itself and the correlation of increasing global carbon dioxide concentration with rising global temperature. ” Dr. David Whitehouse, author of ‘Sun: A Biography’, New Statesman, 19 December 2007.
So infrared is responsible after all. Now we have that little contretemps out of the way, lets cut to the chase.
In that same article, Dr. Whitehouse goes off the rails a couple of paragraphs later when he says:
“For the past decade the world has not warmed. Global warming has stopped.â€
[See Iâ€™m fair, I donâ€™t cherry pick quotations or misrepresent peoples views â€“ unlike Dr. Whitehouseâ€™s article in the The Independent and the quote mining he does on Dr. Hathaway.]
This statement is just flat out untrue, and is simple denialist garbage. It also assumes the conclusion that he wants the reader to reach.
It relies on the fact that 1998 was the hottest year on record and subsequent years have been cooler.
It also relies on that very reliable statistical technique (not) of â€œeyeballlingâ€ the data, rather than using all that fancy, namby-pamby mathematical stuff called statistical analysis. I think you were guilty of the same a little while ago when you pointed out the maxima and minima points in the tempreture series since 1998. Maxima and minima donâ€™t matter so much, what matters is the average and the trend.
Unfortunately, things donâ€™t work out quite so well if you do a straight line fit to the data points â€“ ie. use a technique, that although simple, is actually scientifically and mathematically valid. If you do that, the â€œcoolingâ€ dissappears and the continued warming over the last decade is quite clear. 1998 is simply an outlier. (Try it in Excel, the data series are publically available on the web.)
More sophisticated and careful analysis makes it even clearer, and the warming is present regardless of the base year, 1998 or any other. Weâ€™ve not just had a linear increase, weâ€™ve had an acceleration. There are a couple of really good expositions of this at Open Mind (http://tamino.wordpress.com/2007/12/16/wiggles/#more-507 and http://tamino.wordpress.com/2007/08/31/garbage-is-forever/). The good Dr. W even gets a guernsey as some of his other denialist propaganda is quoted .
I suggest you take the time to read them
. I await your concession.

By: Graham Young

Graham Young — Wed, 26 Dec 2007 21:28:38 +0000

JM, just give us all a rest. I’m sorry you didn’t have anything better to do with your time on Christmas morning than post this nonsense.
You claimed way back down in the thread that the only way that the Sun’s heat reached the earth was via infrared. “Heat from the sun is infra red radiation. It becomes temperature when the gases in the earths atmosphere are excited by it.”
That statement is flat wrong.
You try to cover your tracks in the previous post by referring just to warming of the atmosphere. I’m not particularly interested in the atmosphere on its own. The earth is heated by light across the spectrum and the atmosphere captures and re-radiates some of it that is re-radiated from the earth as infrared. You’ve obviously done a google search and boned-up, although not well enough, but it’s a pretty lame debating point to suddenly try and limit yourself to talking about the atmosphere only.
But at least you now seem to accept that it’s not just infrared that heats the earth, so we’re making some progress.
So, as you’re into grudging concessions, perhaps you could concede the point of my post – if you’re going to talk about reducing carbon dioxide levels it’s reasonable to ask what the optimal level ought to be. Now before you go off and misrepresent me again, I should point out that an optimal level of CO2 does not imply that temperature will always be pleasant. I assume you’re not suggesting a zero level of C02 as being the appropriate level, and if that’s the case, then you must have some thoughts on a greater than zero figure.