Global Warming: Limitations of General
Carl J Lydick
=>> Here we have one article where the record of solar activity, that
=>> starts in 1743, is compared with the amount of sea ice around
=>> Island [Iceland] and with the global temperature trend (that starts
=>> in 1861).
=
=>Actually, this isn't quite true. The paper compares the Zurich sun-spot number
=>derived solar cycle length to *northen hemisphere land surface temperature*
=>anomalies.
=
= Come on, here you are really funny. OK, I have oversimplified, but
= at the same time I have invited you to read that paper. Which you
= did, and I am glad of it. Now, do you agree that northen hemisphere
= land surface temperature (that have been chosen because this is the
= longest homogeneous record of temperatures available) is a good
= indicator of global worming - or for you is just "meaningless data"
= to be promptly ignored and junked?
Why are you unable to understand that it may be neither? While the data may
mean something, that does NOT mean that it's a good indicator of global
temperatures. Consider, for example, that according to some models, it's
possible for global warming to result in cooling of Europe due to a shift in
the Gulf Stream.
=>Is that a prediction we can hold you to? If so, care to specify a period and a
=>temperature decrease?
=
= I can't do that, of course, because I cannot predict solar activity.
= But now we are at the highest observed maximum of activity, so I guess
= it will not get higher :-)
Come now, are you really so ignorant as to never have heard of conditional
predictions? You can specify various possible solar activity profiles and
corresponding temperature predictions.
--------------------------------------------------------------------------------
Carl J Lydick | INTERnet: CA...@SOL1.GPS.CALTECH.EDU | NSI/HEPnet: SOL1::CARL
Disclaimer: Hey, I understand VAXen and VMS. That's what I get paid for. My
understanding of astronomy is purely at the amateur level (or below). So
unless what I'm saying is directly related to VAX/VMS, don't hold me or my
organization responsible for it. If it IS related to VAX/VMS, you can try to
hold me responsible for it, but my organization had nothing to do with it.
Carlo Izzo
>In article <41a547$1i...@sat.ipp-garching.mpg.de>, Carlo Izzo <izzo> writes:
>=>> Here we have one article where the record of solar activity, that
>=>> starts in 1743, is compared with the amount of sea ice around
>=>> Island [Iceland] and with the global temperature trend (that starts
>=>> in 1861).
>=
>=>Actually, this isn't quite true. The paper compares the Zurich sun-spot number
>=>derived solar cycle length to *northen hemisphere land surface temperature*
>=>anomalies.
>=
>= Come on, here you are really funny. OK, I have oversimplified, but
>= at the same time I have invited you to read that paper. Which you
>= did, and I am glad of it. Now, do you agree that northen hemisphere
>= land surface temperature (that have been chosen because this is the
>= longest homogeneous record of temperatures available) is a good
>= to be promptly ignored and junked?
>
>Why are you unable to understand that it may be neither? While the data may
>mean something, that does NOT mean that it's a good indicator of global
>temperatures. Consider, for example, that according to some models, it's
>possible for global warming to result in cooling of Europe due to a shift in
>the Gulf Stream.
I see. What I don't see is so much difference between the northern
hemisphere land surface temperature and the other profiles (as for
instance the one shown in Kelly & Wigley's paper (yes, I went to
look at it after your recommendation).
>
>=>Is that a prediction we can hold you to? If so, care to specify a
>=>period and a temperature decrease?
>=
>= I can't do that, of course, because I cannot predict solar activity.
>= But now we are at the highest observed maximum of activity, so I guess
>= it will not get higher :-)
>
>Come now, are you really so ignorant as to never have heard of conditional
>predictions? You can specify various possible solar activity profiles and
>corresponding temperature predictions.
>Carl J Lydick | INTERnet: CA...@SOL1.GPS.CALTECH.EDU | NSI/HEPnet: SOL1::CARL
Why have you removed my prediction from the text? I said that
global temperature will follow the solar activity. But you want
me to be more precise - while you know very well that this is
impossible. Anyway, let me think. There was a maximum of solar
activity about 1770; a minimum about 1810; a max about 1850; a min
about 1895; a max about 1940. The cicle repeats itself in about
80-90 years (the Gleissberg period). A minimum was expected in 1985.
Indeed there was one, but not very deep... and then solar activity
started to rise again. It should keep rising until 2030 - you have
all the time to "prove" that global warming is man-made, and ruin
the world economy. OK?
Ciao, Carlo
William Connolley
>I saw your desription of the study, in which ice formation was
>perfectly correlated with solar activity. Very impressive, except...
>EXCEPT ... if I read you correctly, it was ONE location on Earth.
Izzo, unfortunately, choose to describe the study as correlating solar cycle
variations with ice extent around Iceland and also global surface temperature.
You should have looked and seen the second half of his sentence, or still better
read the paper whereupon you would have discovered that it would be better described
as about global (actually NH) surface temperatures and also a bit of sea ice.
That the sea ice is only at one point is unfortunate too. I assume that is is the
longest/best such record. The sea ice is only introduced into the paper to extend the
temperature record back beyond ?1861? when an adequate sfc network is available.
William Connolley
>> > have a look at the 1991
>> > paper by E. Friis-Christensen and K. Lassen, of the Danish Meteorological
>> > Institute, published on Science (Vol. 254, p. 698) with the title "Length
>> > of the Solar Cycle: An Indicator of Solar Activity Closely Associated with
>> > Climate". I think you may find it interesting, since it demonstrates the
>> > correlation between the global temperature and the solar activity.
Notice the care with which the paper has been titled. Not "Solar Cycle Determines
Global Temperature" or anything like that...
> Here we have one article where the record of solar activity, that
> starts in 1743, is compared with the amount of sea ice around
> Island [Iceland] and with the global temperature trend (that starts in 1861).
Actually, this isn't quite true. The paper compares the Zurich sun-spot number
derived solar cycle length to *northen hemisphere land surface temperature* anomalies.
> The three observed maxima of solar activity in 1770, 1850 and 1940,
> match exactly the three observed minima in the amount of sea ice,
> and in addition to that the variations of solar activity replicate
> with astonishing precision the profile of the global temperature
> variations after 1861. -cut- See, the correlation coefficient is 0.95...
I didn't see a corr coeff in the paper.
> I just think that
> presently the effect of the CO2 increase in the atmosphere has been
> strongly overestimated. Now the solar activity is decreasing, and
> in the next few years we may see the global temperature decreasing
> too.
Is that a prediction we can hold you to? If so, care to specify a period and a
temperature decrease?
But, back to Friis-Christensen. You seem to think that paper has been forgotten. In
which case you presumably haven't read Kelly+Wigley, Nature 1992, p328 (and indeed
another article and commentary in the same issue). They point out that, whilst the
FC correlation is indeed good, there are two problems:
1. The implied sensitivity to radiative forcing of the earth is too high
2. The derivation of the time series of solar cycle length by FC is not unique
Point 1 is more important and is indeed mentioned in FC as area number of other caveats
you have skipped over.
- William Connolley w...@bas.ac.uk "Captain Reactive"
Len Evens
>
>Why have you removed my prediction from the text? I said that
>global temperature will follow the solar activity. But you want
>me to be more precise - while you know very well that this is
>impossible. Anyway, let me think. There was a maximum of solar
>activity about 1770; a minimum about 1810; a max about 1850; a min
>about 1895; a max about 1940. The cicle repeats itself in about
>80-90 years (the Gleissberg period). A minimum was expected in 1985.
>Indeed there was one, but not very deep... and then solar activity
>started to rise again. It should keep rising until 2030 - you have
>all the time to "prove" that global warming is man-made, and ruin
>the world economy. OK?
>
This at least we need not worry about. We now have very accurate
measurements of the solar constant from satellites. Assuming that
work is not defunded by the present crop of budget cutters, we
should be able to keep track of exactly what is happening with
forcing from solar radiation. However, disentangling all the
different forcings will, I'm afraid, require the use of those
climate models you seem to think are superfluous.
Leonard Evens l...@math.nwu.edu 708-491-5537
Dept. of Mathematics, Northwestern Univ., Evanston, IL 60208
Carlo Izzo
>In article 1i...@sat.ipp-garching.mpg.de, Carlo Izzo <izzo> writes:
>>> > have a look at the 1991
>>> > paper by E. Friis-Christensen and K. Lassen, of the Danish Meteorological
>>> > Institute, published on Science (Vol. 254, p. 698) with the title "Length
>>> > of the Solar Cycle: An Indicator of Solar Activity Closely Associated with
>>> > Climate". I think you may find it interesting, since it demonstrates the
>>> > correlation between the global temperature and the solar activity.
>Notice the care with which the paper has been titled. Not "Solar Cycle Determines
>Global Temperature" or anything like that...
>
>> Here we have one article where the record of solar activity, that
>> starts in 1743, is compared with the amount of sea ice around
>> Island [Iceland] and with the global temperature trend (that starts
>> in 1861).
>Actually, this isn't quite true. The paper compares the Zurich sun-spot number
>derived solar cycle length to *northen hemisphere land surface temperature*
>anomalies.
Come on, here you are really funny. OK, I have oversimplified, but
at the same time I have invited you to read that paper. Which you
did, and I am glad of it. Now, do you agree that northen hemisphere
land surface temperature (that have been chosen because this is the
longest homogeneous record of temperatures available) is a good
indicator of global worming - or for you is just "meaningless data"
to be promptly ignored and junked?
>
>> The three observed maxima of solar activity in 1770, 1850 and 1940,
>> match exactly the three observed minima in the amount of sea ice,
>> and in addition to that the variations of solar activity replicate
>> with astonishing precision the profile of the global temperature
>> variations after 1861. -cut- See, the correlation coefficient is 0.95...
>I didn't see a corr coeff in the paper.
Right, this is not in the paper. It is at page 652 of the same
number of Science (were Friis-Christensen and Lassen paper is
introduced.
>
>> I just think that
>> presently the effect of the CO2 increase in the atmosphere has been
>> strongly overestimated. Now the solar activity is decreasing, and
>> in the next few years we may see the global temperature decreasing
>> too.
>Is that a prediction we can hold you to? If so, care to specify a period and a
>temperature decrease?
I can't do that, of course, because I cannot predict solar activity.
But now we are at the highest observed maximum of activity, so I guess
it will not get higher :-)
My prediction is: temperature will follow the solar cicle length
behavior in the same way as it did in the past 252 years.
>
>But, back to Friis-Christensen. You seem to think that paper has been
>forgotten. In which case you presumably haven't read Kelly+Wigley,
>Nature 1992, p328 (and indeed another article and commentary in the
>same issue). They point out that, whilst the FC correlation is indeed
>good, there are two problems:
>1. The implied sensitivity to radiative forcing of the earth is too high
>2. The derivation of the time series of solar cycle length by FC is not unique
>Point 1 is more important and is indeed mentioned in FC as area number of
>other caveats you have skipped over.
>
>- William Connolley w...@bas.ac.uk "Captain Reactive"
>
You are right that point 1 is (far) more important. OK, the implied
sensitivity to radiative forcing of the earth is too high. Which means
that the correlation of this solar activity indicator with the amount
of sea ice and with the temperature, shown in this paper, is just a
"chance". Is this what you mean? Look at those picture again. There
is even a small lag in the 3 ice minima, *following* the 3 solar activity
maxima, that is exactly what one would expect to see. Wouldn't you
rather doubt that sensitivity to radiative forcing of the earth should
be revised? How this sensitivity should be measured then, if not with
data of this kind?
Carlo Izzo
Carlo Izzo
>In article <41ac1p$1i...@sat.ipp-garching.mpg.de>, Carlo Izzo <izzo> wrote:
>>
>>Why have you removed my prediction from the text? I said that
>>global temperature will follow the solar activity. But you want
>>me to be more precise - while you know very well that this is
>>impossible. Anyway, let me think. There was a maximum of solar
>>activity about 1770; a minimum about 1810; a max about 1850; a min
>>about 1895; a max about 1940. The cicle repeats itself in about
>>80-90 years (the Gleissberg period). A minimum was expected in 1985.
>>Indeed there was one, but not very deep... and then solar activity
>>started to rise again. It should keep rising until 2030 - you have
>>all the time to "prove" that global warming is man-made, and ruin
>>the world economy. OK?
>>
>
>This at least we need not worry about. We now have very accurate
>measurements of the solar constant from satellites.
You are right, this is a very important point. This will not
only allow us to track exactly what is happening with forcing
from solar radiation (as you say below) but it will also improve
the relation between the solar radiation and the solar cycle
length (which is not necessarily linear) allowing us to extend
it with more confidence to the past. But this takes time...
At *least* 11 years. Better 22.
>Assuming that
>work is not defunded by the present crop of budget cutters, we
>should be able to keep track of exactly what is happening with
>forcing from solar radiation. However, disentangling all the
>different forcings will, I'm afraid, require the use of those
>climate models you seem to think are superfluous.
>
>Leonard Evens l...@math.nwu.edu 708-491-5537
>Dept. of Mathematics, Northwestern Univ., Evanston, IL 60208
I actually feel uncomfortable in saying that those models are
superfluous. What is really disappointing me is not so much the
poorness of those models, but the exaggerated importance that is
given to their results. The uncertainties are enormous - so enormous
that sometimes they are not even reported (as in the article
by P.M.Kelly & T.M.L.Wigley - that somebody of you suggested
me to read). Yes, Leonard, you are right, modeling should be
supported anyway - even if it has been a failure up to now.
$122.6 are not so many, after all.
Carlo Izzo
William Connolley
> Bullshit. I am wonderning, by the way, why all of you concentrate
> on such details, instead of explaining why a correlation that holds
> at long, medium and short time scales, and lasting 252 years, should
> be rejected? The lack of a known physical explanation is NOT a reason.
> Imagine a man, thousands of years ago, noticing the regularity of
> the cycle of seasons. He wonders why it is so. He has no physical
> explanation for it. Should he conclude that, for this reason,
> there is no such thing as a seasonal cycle?
OK, this is the core of the problem. What we have is
1. a good correlation between solar cycle length and NH land temps for the last 100+y
2. a believed-to-be-reasonably-reliable-theory that says that the effects of solar
variation are too small to account for the temperature changes.
What your fictional man thousands of years ago lacked was anything corresponding to
point 2. Len Evens has posted the question "why should the apparently stronger
greenhouse forcing be supposed to have less effect than the solar forcing".
Its a good question.
Note that Kelly+Wigley (Nature 1992) believe that even given the FC data, the
temperature series is best fit by a combination of CO2 and Solar forcing that
gives a greater weight to the CO2 forcing.
Also note, while we're here, the fit of the ice data is not nearly as good as the
fit of the (shorter) temperature series.
Note also: correlations without physical mechanisms are to be somewhat distrusted.
There is always the possibility of a fit by random chance. OK, this seems unlikely
in this case. But don't forget, the usual statistical statement of "the correlation
could not be achieved by chance at the 95% level" only really applies if 2 series
are compared out of the blue. That isn't true in this case - the solar length series
is the end product of a lot of work trying to find a solar-something that fits (this
is not an attack on FC).
In short, I think some explanation of the fit of the FC series is still needed. But
the weight of evidence favours point 1. at the moment, IMHO.
- William Connolley - w...@bas.ac.uk - http://www.nbs.ac.uk/public/icd/wmc
ps - fortunately I'm on holiday next 2 weeks so I'll probably miss sequels to this!
William Connolley
>OK, this is the core of the problem. What we have is
>1. a good correlation between solar cycle length and NH land temps for the last 100+y
>2. a believed-to-be-reasonably-reliable-theory that says that the effects of solar
>variation are too small to account for the temperature changes.
>In short, I think some explanation of the fit of the FC series is still needed. But
>the weight of evidence favours point 1. at the moment, IMHO.
Oh B*gg*r. I meant to say, favours point 2. Arghhh!
- w...@bas.ac.uk
Carlo Izzo
>In article 20...@sat.ipp-garching.mpg.de, Carlo Izzo <izzo> writes:
>> Bullshit. I am wonderning, by the way, why all of you concentrate
>> on such details, instead of explaining why a correlation that holds
>> at long, medium and short time scales, and lasting 252 years, should
>> be rejected? The lack of a known physical explanation is NOT a reason.
>> Imagine a man, thousands of years ago, noticing the regularity of
>> the cycle of seasons. He wonders why it is so. He has no physical
>> explanation for it. Should he conclude that, for this reason,
>> there is no such thing as a seasonal cycle?
>
>1. a good correlation between solar cycle length and NH land temps for the last 100+y
>2. a believed-to-be-reasonably-reliable-theory that says that the effects of solar
>variation are too small to account for the temperature changes.
>
>point 2.
Point 2 for the fictional man thousands of years ago: "The
earth is obviously flat" (i.e. it cannot be that the seasons
are changing because the earth is like a ball spinning around a
tilted rotation axis that is accomplishing one orbit around the
sun in 365.25 days).
Len Evens has posted the question "why should the apparently stronger
>greenhouse forcing be supposed to have less effect than the solar forcing".
>Its a good question.
It was. What did you find wrong in my answer?
>
>Note that Kelly+Wigley (Nature 1992) believe that even given the FC data, the
>temperature series is best fit by a combination of CO2 and Solar forcing that
>gives a greater weight to the CO2 forcing.
How can you possibly trust the results published in that article.
They pick a greenhouse forcing history (history 2, from Wigley and
Raper, Nature 357, p.293), that is the result of a FIT to the
temperature record, and then they REFIT it together with FC to the
SAME temperature record! Of course they find out that "history 2"
weights more than FC.
But even in case you reject this objection (because you may claim
that "history 2" is not the result of a fit - even if it is), did you
see the errors reported for the computed quantities? No? In fact,
there aren't any. They are too shamefully HUGE to be reported. Look
at the table on page 329. 41.2% greenhouse, 19.4% solar... it looks
quite accurate right? Look at what they dare to write: "The variance
explained by these two simulations is 60.8% and 60.6%, respectively.
The results based on the FCL record are similar, although the variance
explained is slightly less". This is ridiculous. Do you have an idea
of the uncertainties on those numbers? 60.8 and 60.6 are IDENTICAL,
when in presence of HUGE uncertainties. You don't say "this is
slightly better than that".
And to make it apparent that the uncertainties are huge: the explained
variances are ranging between 50 and 60 (with the various combinations
of greenhouse and solar cycles records). The best fit is with the
"solar forcing only", that corresponds to a climate sensitivity of 12 K
- which is obviously unlikely because too high. But the best fit is
rather "loose" (given the uncertainties): in fact, forcing a constraint
to the climate sensitivity within the interval 1.5 - 4.5 K (which is
reasonable) the variance explained using "solar forcing only" is still
55% (that is, within the interval 50% - 60% of all the other best
fits - see page 329, on top-right). In other words, "solar forcing
only" simply cannot be ruled out! It still explains the data quite
well (better than greenhouse only, anyway, ah!,ah!).
They FORCED greenhouse in those data, that's it. And actually,
"solar forcing only" explains the past temperature record BETTER
than "greenhouse forcing only". I would say that this is a hint of
a stronger influence of solar forcing with respect to the greenhouse
forcing, and NOT the opposite. This is what you can *really* find in
that paper.
>
>Also note, while we're here, the fit of the ice data is not nearly as good as the
>fit of the (shorter) temperature series.
Ice data can reflect only long time scales variations.
There is some more (thermal) inertia in water than in air.
>
>Note also: correlations without physical mechanisms are to be somewhat distrusted.
>There is always the possibility of a fit by random chance. OK, this seems unlikely
>in this case. But don't forget, the usual statistical statement of "the correlation
>could not be achieved by chance at the 95% level" only really applies if 2 series
>are compared out of the blue. That isn't true in this case - the solar length series
>is the end product of a lot of work trying to find a solar-something that fits (this
>is not an attack on FC).
You are correct. This is really the case, in fact, and I am glad
that you mentioned it. Nevertheless, the correlation is really strict,
right? Even considering your objection, it is still quite hard
for me to reject it.
I was always puzzled by this problem in computing probabilities
of observed results. It's a philosophical problem... Suppose
that two isolated researchers discover independently the same
relation between physical quantities. Suppose that the first
researcher finds the "right" relation after 10 attempts, while
the second researcher - who is perhaps more clever - finds the
relation after just 3 failures. Mind it, it is the same
relation they found. They compute therefore different
probabilities for the *reality* of their correlation. What
is the real probability of the correlation? The one associated
to 6.5 attempts (after all, the relation has been discovered
*twice*)?
Anyway, you are right, the number of trials is working against
the probability of a correlation. But not enough in this case,
I swear!
>
>In short, I think some explanation of the fit of the FC series is still needed. But
>the weight of evidence favours point 1. at the moment, IMHO.
>
OK
>- William Connolley - w...@bas.ac.uk - http://www.nbs.ac.uk/public/icd/wmc
>ps - fortunately I'm on holiday next 2 weeks so I'll probably miss sequels to this!
>
Carlo