Landscheidt (cl ch 1988)

137 views
Skip to first unread message

w.m.co...@null.invalid

unread,
Jul 31, 2000, 3:00:00 AM7/31/00
to
Ho hum. Someone put on my desk one of L's papers: "solar rotation, impulses
of the torque in the suns motion, and climate variation" (climatic change
1988, 12, p265) so I had a look.

Passing over the bits of solar physics that mean little to me, we come to
the climatology. First off, fig 10, he relates "JU-Cm-CS" cycles to
(an apparently arbitrary but at least well defined) function of
german rainfall.

Reading from the figure, there are peaks at 1915-, 1925, 1935-, 1945+,
1955+, 1965, 1975+ and 1985- (the figure is poorly drawn, and +/-
corresponds to about 2-3 years). All these peaks in variance of rainfall
have little arrows sitting right over them, and apparently mark JCC
epochs. The correspondence is remarkable: close to exact. Astonishingly so:
geophysical series are almost never this good.

But: looking earlier, the JCC epochs are given as: 1912, 1921, 1930, 1942,
1951, 1959, 1967, 1970, 1974, 1983 (to the nearest year).

You see the problem? There apears to be little correspondence between the
marked arrows and the JCC epochs.

If anyone out there has access to climatic change, and the patience to
read the relevant article, it would be interesting to know if my by-eye
check can be repeated.

-W.

--
William M Connolley | Disclaimer: I speak for myself

Paul D. Farrar

unread,
Jul 31, 2000, 3:00:00 AM7/31/00
to
In article <3985d...@news.nwl.ac.uk>, 31 Jul 2000 20:38:02 GMT,
w.m.co...@null.invalid writes:

I suppose some people will derive satisfaction from Stephen Schneider
getting suckered.

TL always gives himself a lot of latitude on what constitutes a match.
Anything anywhere near a little past halfway through is declared to
have matched the Major (~0.618..); anything before is declared for the
Minor (~0.382..). If that's not good enough, there's the Major of the
Minor etc., as required. See Fig. 3 of
http://www.microtech.com.au/daly/sun-enso/sun-enso.htm
in which the special numbers, and their match bars, cover the great
majority of the cycle, for what looks like, to me, a flat distribution
(ie no connection at all).

TL's triangle plots also require a very optimistic eye. He arranges
small triangles corresponding to various Golden Section epochs along a
time series and notes matches. See Fig.4 of same. Some triangles hit
peaks; some hit valleys. But then there are often peaks and valleys
with no triangles at all, many of them bigger than the ones with
triangles. What are they? And if his rule changes in the middle of the
series, then he can always find something about the baricenter, or
something, that is responsible.

The statistical tests are meaningless. If you contrive a rule to fit
occurances, then the probability of that rule being followed is 1.

The flexibility and extensibility of his method guarantees a
successful explanation. He has found it to apply not just in solar
dynamics and climate, but economics and business, crime, cancer
treatment, Gauquelin's Mars Effect, circadian cycles, sex
determination, and so on. It's not a falsifiable explanation, because
he can always find some way to make it fit.

This is in contrast to a physical law. Conservation of momentum, for
example, forbids far more than it allows. What is not allowed can be
explicitly specified, and so one can test for the existance of
forbidden outcomes.

I don't see any reason why Tl's methods should not work for any other
number, such as sqrt(2).

--
Paul D. Farrar
http://www.datasync.com/~farrar

na...@my-deja.com

unread,
Aug 1, 2000, 3:00:00 AM8/1/00
to
In article <3985fc5f...@news.datasync.com>,


There's a similar figure (Fig. 20) in his web paper "Solar Activity:
a dominant...". However, I find no such discrepancy here, only, as you
say, a remarkable correspondence. I'm not quite sure what you are
saying...are you saying that he has incorrectly plotted the positions of
solar cycle phases (i.e. the JCC epochs) on the figure so that they
appear to coincide with climate phenomena...?
How could that possibly get past peer-review....!? ;-)


>
> I suppose some people will derive satisfaction from Stephen Schneider
> getting suckered.
>
> TL always gives himself a lot of latitude on what constitutes a match.
> Anything anywhere near a little past halfway through is declared to
> have matched the Major (~0.618..); anything before is declared for the
> Minor (~0.382..). If that's not good enough, there's the Major of the
> Minor etc., as required. See Fig. 3 of
> http://www.microtech.com.au/daly/sun-enso/sun-enso.htm
> in which the special numbers, and their match bars, cover the great
> majority of the cycle, for what looks like, to me, a flat distribution
> (ie no connection at all).
>

I can't agree with you here, Paul. The correlations between the exact
major/minor sections and climate phenomena are clear to see. The Fig.3
that you mentioned is based on a small data set of El Ninos, but
nevertheless, the frequency distribution is clearly not "flat" as you
put it. The "match bars" are nothing to do with the "special numbers".
They simply show the ranges of the sunspot cycle in which the majority
of El Nino events occur. The peak at 0.0 can hardly be put down to pure
chance (as his statistical tests show).


> TL's triangle plots also require a very optimistic eye. He arranges
> small triangles corresponding to various Golden Section epochs along a
> time series and notes matches. See Fig.4 of same. Some triangles hit
> peaks; some hit valleys. But then there are often peaks and valleys
> with no triangles at all, many of them bigger than the ones with
> triangles. What are they? And if his rule changes in the middle of the
> series, then he can always find something about the baricenter, or
> something, that is responsible.
>

I think you've misunderstood here. Fig. 4 shows the correlations between
SOI extrema and subcycles of the sunspot cycle only (i.e. major and
minor of the 11-ish year cycle). Figs. 6 and 7 show the correlations
between SOI extrema and small finger subcycles (related to solar orbital
angular momentum). When all these effects are added together, as shown
in Fig.8, then all major peaks and troughs in the SOI are covered. The
reversal in 1968 corresponds to a well defined phase change in solar
orbital angular momentum (L calls it a big finger start)...it's all
explained in the paper.


> The statistical tests are meaningless. If you contrive a rule to fit
> occurances, then the probability of that rule being followed is 1.
>

But he has not contrived a rule to fit occurances. His model is based on
the chaotic motion of the sun about the solar system barricentre. This
has to be computed using orbital integration programs. Therefore the
resulting orbital angular momentum subcycles are not of a constant
duration (they are determined by a chaotic process). Therefore the
location of the major or minor of a subcycle cannot be found by linear
interpolation at any arbitrary date. It must be computed. If you notice,
(for example on Fig.8), the time interval between successive yellow
circles or green diamonds is not a constant.
As far as I understand, the SOI oscillation is also thought to be
caused by a chaotic process...? If so, how can one explain an apparent
correlation between two supposedly independant, chaotic phenomena...?


> The flexibility and extensibility of his method guarantees a
> successful explanation. He has found it to apply not just in solar
> dynamics and climate, but economics and business, crime, cancer
> treatment, Gauquelin's Mars Effect, circadian cycles, sex
> determination, and so on. It's not a falsifiable explanation, because
> he can always find some way to make it fit.
>

Well, you're getting to the nitty-gritty now, and here I'm just as
confused as you are. He seems to think that nature (i.e. everything!)
is based on fractals (and Fibonacci numbers, to which the golden section
is closely related). I haven't looked into this in much detail yet, but
what I have read is very interesting...


> This is in contrast to a physical law. Conservation of momentum, for
> example, forbids far more than it allows. What is not allowed can be
> explicitly specified, and so one can test for the existance of
> forbidden outcomes.
>
> I don't see any reason why Tl's methods should not work for any other
> number, such as sqrt(2).
>
> --
> Paul D. Farrar
> http://www.datasync.com/~farrar
>

It's easy to see that they wouldn't work....e.g. the maximum of the
solar cycle falls at the minor of the cycle, not sqrt(2) of the cycle.
Similar examples can be given for other solar subcycles. It's not a
random number.

Regards,
Nathan Harris.


Sent via Deja.com http://www.deja.com/
Before you buy.

w...@null.invalid

unread,
Aug 2, 2000, 3:00:00 AM8/2/00
to
na...@my-deja.com wrote:
>In article <3985fc5f...@news.datasync.com>,

>> In article <3985d...@news.nwl.ac.uk>, 31 Jul 2000 20:38:02 GMT,
>> w.m.co...@null.invalid writes:

>> >Reading from the figure, there are peaks at 1915-, 1925, 1935-,
>1945+,
>> >1955+, 1965, 1975+ and 1985- (the figure is poorly drawn, and +/-
>> >corresponds to about 2-3 years). All these peaks in variance of
>rainfall
>> >have little arrows sitting right over them, and apparently mark JCC
>> >epochs. The correspondence is remarkable: close to exact.
>Astonishingly so:
>> >geophysical series are almost never this good.
>> >
>> >But: looking earlier, the JCC epochs are given as: 1912, 1921, 1930,
>1942,
>> >1951, 1959, 1967, 1970, 1974, 1983 (to the nearest year).

>I'm not quite sure what you are


>saying...are you saying that he has incorrectly plotted the positions of
>solar cycle phases (i.e. the JCC epochs) on the figure so that they
>appear to coincide with climate phenomena...?
>How could that possibly get past peer-review....!? ;-)

Well, try looking at the figure yourself and see what you think. I think
the plotted positions don't correspond to the ones listed earlier.

-W

Thomas Palm

unread,
Aug 2, 2000, 3:00:00 AM8/2/00
to
na...@my-deja.com wrote:
> I'm not quite sure what you are
> saying...are you saying that he has incorrectly plotted the positions of
> solar cycle phases (i.e. the JCC epochs) on the figure so that they
> appear to coincide with climate phenomena...?
> How could that possibly get past peer-review....!? ;-)

One weakness of most scientists is that while they may easily believe
their colleagues are incompetent they tend to be bad at detecting
outright lies. Perhaps this is a consequence of studying nature,
which is basically honest, rather than people.

This has repeatedly been used, e.g., by psychics, to fool initially
sceptic scientists (as proved in an experiment by James Randi.)

I have no trouble believing Landscheid could get a falsified figure
through the review process. After all, Pons & Fleischman could get
a clearly false neutron spectrum published despite the controversy
about cold fusion.

na...@my-deja.com

unread,
Aug 2, 2000, 3:00:00 AM8/2/00
to
In article <39881FD1...@chello.se>,
Thomas Palm <thoma...@chello.se> wrote:

> na...@my-deja.com wrote:
> > I'm not quite sure what you are
> > saying...are you saying that he has incorrectly plotted the
positions of
> > solar cycle phases (i.e. the JCC epochs) on the figure so that they
> > appear to coincide with climate phenomena...?
> > How could that possibly get past peer-review....!? ;-)
>
> One weakness of most scientists is that while they may easily believe
> their colleagues are incompetent they tend to be bad at detecting
> outright lies. Perhaps this is a consequence of studying nature,
> which is basically honest, rather than people.
>
> This has repeatedly been used, e.g., by psychics, to fool initially
> sceptic scientists (as proved in an experiment by James Randi.)
>
> I have no trouble believing Landscheid could get a falsified figure
> through the review process. After all, Pons & Fleischman could get
> a clearly false neutron spectrum published despite the controversy
> about cold fusion.
>

You're right. It's important to get over (or trip up on) this initial
obstacle to progress with regard to Landscheidt's work. We can't confirm
or deny this possibility until we have studied the article in question
(which I don't have yet) and the data that he used to plot the figure.
However, based on his other papers that I have studied and the
fact that his predictions (several years in advance) of climate change
events made using the same theory have been successful (El Ninos 1993,
1995 and 1998 correctly predicted. 90% success rate at predicting strong
solar eruptions over a six year period, and various global temperature
and rainfall predictions contained in his papers), I find it difficult
to entertain the idea that he is falsifying his data to correlate with
climatic phenomena.
But I could be wrong...

na...@my-deja.com

unread,
Aug 3, 2000, 3:00:00 AM8/3/00
to
In article <8ma32v$epp$1...@nnrp1.deja.com>,
na...@my-deja.com wrote:

p.s. further predictions made using Landscheidt's solar model:

1. next ENSOI "El Nino" peak: 2002.9 (+/- 0.5)
2. next "Little Ice Age" around 2030 (discounting anthropogenic CO2
forcing being dominant, of course! ;-)

Sorry, no forecasts of global catastrophe.....hardly in Nostradamus's
class!

I suppose only time will tell....

na...@my-deja.com

unread,
Aug 18, 2000, 3:00:00 AM8/18/00
to
In article <3987e...@news.nwl.ac.uk>,

w...@null.invalid wrote:
> na...@my-deja.com wrote:
> >In article <3985fc5f...@news.datasync.com>,
> >> In article <3985d...@news.nwl.ac.uk>, 31 Jul 2000 20:38:02 GMT,
> >> w.m.co...@null.invalid writes:
>
> >> >Reading from the figure, there are peaks at 1915-, 1925, 1935-,
> >1945+,
> >> >1955+, 1965, 1975+ and 1985- (the figure is poorly drawn, and +/-
> >> >corresponds to about 2-3 years). All these peaks in variance of
> >rainfall
> >> >have little arrows sitting right over them, and apparently mark
JCC
> >> >epochs. The correspondence is remarkable: close to exact.
> >Astonishingly so:
> >> >geophysical series are almost never this good.
> >> >
> >> >But: looking earlier, the JCC epochs are given as: 1912, 1921,
1930,
> >1942,
> >> >1951, 1959, 1967, 1970, 1974, 1983 (to the nearest year).
>
> >I'm not quite sure what you are
> >saying...are you saying that he has incorrectly plotted the positions
of
> >solar cycle phases (i.e. the JCC epochs) on the figure so that they
> >appear to coincide with climate phenomena...?
> >How could that possibly get past peer-review....!? ;-)
>
> Well, try looking at the figure yourself and see what you think. I
think
> the plotted positions don't correspond to the ones listed earlier.
>
> -W
>
> --
> William M Connolley | Disclaimer: I speak for myself
>

I have discussed the paper with Landscheidt and he has sent me a better
quality copy of the Figure 10 and the data points plotted. His initial
response to your query was:

"Figure 10 in the Climate change paper is very rough, but it is
psychologically interesting that climatologist William Connelly imagined
such large discrepanies. I shall send you a finer version of Figure 10
via snail mail so that you can see that JCC epochs and extrema in the
rainfall curve are rather close to each other."
[snipped the rest]

The figure is almost identical to Figure 20 in his paper on "Solar
Activity: a dominant factor in climate dynamics" on Daly's web site and
soon to come out in "Energy and Environment". The open arrows correspond
to particular epochs in the solar orbital angular momentum cycle.
As I suspected, your analysis of the figure was probably hastily
done and therefore incorrect (no offense intended) (the individual ticks
on the Time axis, for example, correspond to 4 years, not 5). A more
accurate look at the original figure and data gives the plotted arrow
positions (in the 20th century) as (to the nearest year):
1912, 1921, 1930, 1943, 1951, 1959, 1970 and 1975 which correspond
exactly with those stated earlier in the text.

So I think we can safely conclude that Landscheidt has not fabricated
his results and that he is onto something very exciting here.

These results were corroborated by rainfall data from England, Wales,
USA and India as well as by similar investigations into temperature and
SOI changes (in his other papers).

Why hasn't anyone in climatology circles seriously addressed these
startling results...? (to simply de-bunk them or otherwise).

na...@my-deja.com

unread,
Aug 20, 2000, 3:00:00 AM8/20/00
to


any comments....?

w...@bas.ac.uk

unread,
Aug 20, 2000, 3:00:00 AM8/20/00
to
Somehow I missed your original post.

na...@my-deja.com wrote:

>As I suspected, your analysis of the figure was probably hastily done

No, it wasn't

> and therefore incorrect (no offense intended) (the individual
> ticks on the Time axis, for example, correspond to 4 years, not 5).

If you think that, you're not looking at the published figure, which is
indeed very rough (I would have rejected it for quality had I been a ref)
but has tick marks at 10 year intervals (not 4 or 5).

>A more accurate look at the original figure
>and data gives the plotted arrow positions (in the 20th century) as (to the nearest year):
>1912, 1921, 1930, 1943, 1951, 1959, 1970 and 1975 which correspond
> exactly with those stated earlier in the text.

No. You are quite wrong here, so wrong that you cannot be looking at the
J. Clim figure (you are a touch unclear as to what you have looked at,
can you confirm that you have seen the original J Clim figure (not
a subsequent redrawing). If you haven't seen the original, I suggest
you get to see one).

The tickmarks are at 10-year intervals, so its hard to be precise as to
some of the positions, but there is clearly no arrow, for example,
between 1935 and 1945. Nor is there one in 1970.



>These results were corroborated by rainfall data from England, Wales,
>USA and India as well as by similar
>investigations into temperature and SOI changes (in his other papers).

It was England/Wales actually, not 2 separate regions.

But anyway: the results for those regions were only presented in
summary form, in a way that was impossible to check. The only one that
could be checked, was fig 10, which failed the test.

-W.

w...@bas.ac.uk

unread,
Aug 20, 2000, 3:00:00 AM8/20/00
to
w...@bas.ac.uk wrote:
>natjo:

>>A more accurate look at the original figure
>>and data gives the plotted arrow positions (in the 20th century) as (to the nearest year):
>>1912, 1921, 1930, 1943, 1951, 1959, 1970 and 1975 which correspond
>> exactly with those stated earlier in the text.

To make things easy: the original figure is at
http://www.wmc.care4free.net/sci/Dscn1570-l-f10.jpg

-W.

na...@my-deja.com

unread,
Aug 21, 2000, 3:00:00 AM8/21/00
to
In article <39a04...@news.nwl.ac.uk>,
w...@bas.ac.uk wrote:

> If you think that, you're not looking at the published figure, which
is
> indeed very rough (I would have rejected it for quality had I been a
ref)
> but has tick marks at 10 year intervals (not 4 or 5).
>

It seems that he's sent me the Fig. 10 from another paper. I will get
back to him and remind him of the exact paper reference.
However I have had a look at it on the web address you supplied.

>
> The tickmarks are at 10-year intervals, so its hard to be precise as
to
> some of the positions, but there is clearly no arrow, for example,
> between 1935 and 1945. Nor is there one in 1970.
>

Yes, but there are "open circles" that "designate phases pi radians of
respective JU-CM-CS cycles". Does the text state that the *arrows*
should correspond to the dates, or the *open circles*...? There must be
some explanation in the main body of the paper.


> >These results were corroborated by rainfall data from England, Wales,
> >USA and India as well as by similar
> >investigations into temperature and SOI changes (in his other
papers).
>
> It was England/Wales actually, not 2 separate regions.
>
> But anyway: the results for those regions were only presented in
> summary form, in a way that was impossible to check. The only one that
> could be checked, was fig 10, which failed the test.
>
> -W.
>
> William M Connolley | Disclaimer: I speak for myself
>

...This is supposed to be the referees job.
Have you looked at any of Landscheidt's other results (or other figures
in the same paper)...?

More later...

w...@bas.ac.uk

unread,
Aug 21, 2000, 3:00:00 AM8/21/00
to
na...@my-deja.com wrote:
> w...@bas.ac.uk wrote:

>However I have had a look at it on the web address you supplied.

>Yes, but there are "open circles" that "designate phases pi radians of


>respective JU-CM-CS cycles". Does the text state that the *arrows*
>should correspond to the dates, or the *open circles*...? There must be
>some explanation in the main body of the paper.

The text says that the arrows are epochs and the circles pi radians from these.

The text says that epochs are maxima and pi radians are minima.

You really should get a copy of the actuall paper to work with.

>> But anyway: the results for those regions were only presented in
>> summary form, in a way that was impossible to check. The only one that
>> could be checked, was fig 10, which failed the test.

>...This is supposed to be the referees job.

No, not really. The data supplied in the paper is not sufficient to allow
his stats to be checked.

>Have you looked at any of Landscheidt's other results (or other figures
>in the same paper)...?

Fig 10 is essentially the only piece of testable meteorology
in the paper. Actually, theres fig 11 too, but lets sort out fig 10 first.

-W.

--
William M Connolley | w...@bas.ac.uk | http://www.nerc-bas.ac.uk/icd/wmc/
Climate Modeller, British Antarctic Survey | Disclaimer: I speak for myself

na...@my-deja.com

unread,
Aug 21, 2000, 3:00:00 AM8/21/00
to
In article <39a18...@news.nwl.ac.uk>,

w...@bas.ac.uk wrote:
> na...@my-deja.com wrote:
> > w...@bas.ac.uk wrote:
>
> >However I have had a look at it on the web address you supplied.
>
> >Yes, but there are "open circles" that "designate phases pi radians
> >of respective JU-CM-CS cycles". Does the text state that the *arrows*
> >should correspond to the dates, or the *open circles*...? There must
> >be some explanation in the main body of the paper.
>
> The text says that the arrows are epochs and the circles pi radians
> from these.
>
> The text says that epochs are maxima and pi radians are minima.
>
> You really should get a copy of the actuall paper to work with.
>

I haven't got access to "Climatic Change" at the moment. I'll try to get
it from another colleague...


> >> But anyway: the results for those regions were only presented in
> >> summary form, in a way that was impossible to check. The only one
that
> >> could be checked, was fig 10, which failed the test.
>
> >...This is supposed to be the referees job.
>
> No, not really. The data supplied in the paper is not sufficient to
> allow his stats to be checked.
>
> >Have you looked at any of Landscheidt's other results (or other
> >figures in the same paper)...?
>
> Fig 10 is essentially the only piece of testable meteorology
> in the paper. Actually, theres fig 11 too, but lets sort out fig 10
first.
>
> -W.
>
> --
> William M Connolley | w...@bas.ac.uk |
http://www.nerc-bas.ac.uk/icd/wmc/
> Climate Modeller, British Antarctic Survey | Disclaimer: I speak for
myself
>

Okay. I've written back to Landscheidt with some searching
questions...hopefully he will come up with something helpful this time.

na...@my-deja.com

unread,
Aug 25, 2000, 3:00:00 AM8/25/00
to
In article <8ns9r8$gpk$1...@nnrp1.deja.com>,

na...@my-deja.com wrote:
>
> Okay. I've written back to Landscheidt with some searching
> questions...hopefully he will come up with something helpful this
> time.

This is what I wrote to him:

> Dear Dr. Landscheidt,
> Thankyou for sending me the figure. Unfortunately it is not
the
> original Figure 10 from your Climatic Change (1988) paper:


> "solar rotation, impulses of the torque in the suns motion, and
climate

> variation" (climatic change 1988, 12, p265).
> The original figure can be accessed at:
> http://www.wmc.care4free.net/sci/Dscn1570-l-f10.jpg
> (courtesy of William Connelly. It's a scan from the original paper).
> Unfortunately I still haven't been able to find a copy of the paper.>
> The arrows in the figure that you sent me correspond almost exactly
with the
> stated dates of JU-CM-CS epochs (and with peaks in the rainfall data).
> However, on the original Fig.10, they do not.
> The JU-CM-CS epochs are given in the text (according to Connelly) as:


1912,
> 1921, 1930, 1942, 1951, 1959, 1967, 1970, 1974, 1983 (to the nearest
year).

> The tickmarks on the figure are at 10-year intervals, so it's hard to
be
> precise as to some of the positions, but there are
> clearly no arrows (or peaks in the rainfall data) corresponding to
these
> dates, apart from 1912 and 1983. Having said that, there does seem to
be a
> correspondence between the open circles and some of these dates.
> Please could you answer the following questions in detail:>
> 1. What do the listed dates (above) correspond to and what does JU-CM-
CS
> actually mean? (I assume it's something to do with the solar orbital
angular
> momentum cycles, Jupiter-Centre of Mass-Centre of Sun...?).
> 2. Are the listed dates supposed to correspond to the arrows or the
open
> circles in the figure (or both?) ?
> 3. Why is the original Fig.10 (particularly the temporal positions of
the
> arrows and rainfall peaks) different from the figure that you sent me
and
> the one featured (Fig.20) in your "solar activity" paper on Daly's
website?
> (the rainfall data looks identical in profile in all 3 figures except
that
> it has been shifted in time in the latter 2 figures: eg. the highest
peak in
> rainfall occurs in 1965 in the original fig.10, but it occurs in 1959
in the
> other 2 figures).
> I hope that you can clear this up for us once and for all...
> Many thanks in advance,> Yours sincerely,
> Nathan Harris.

Here is the reply:
Part 1:

Dear Dr. Harris,
The copy I sent you is based on the same rainfall data as the figure in
the Climate Change. The data for both of the figures are identical. You
can see this in other publications.
I shall answer your questions in detail soon, but at the moment I have
not got enough time. I am engaged in preparing 2 papers I shall have to
present in September at climate conferences at the Rice Univerity in
Houston and at the ESA conference in Teneriffa.

Part 2:

It just occurred to me that you have got a rather precise description
of the rainfall investigation in my paper "Global Warming or Little Ice
Age?" (pp. 379-380) I sent you. There you will find the answers to your
questions.
Obviously, the time scale in Figure 10 is not correct. In the
printing process someone replaced my numbers for the years with bigger
ones and put them in the wrong place. This happened after I had
corrected the galley proofs. Figure 11 in Climate Change, which shows
unsmoothed values of the same data, gives the correct scale.
JU-CM-CS refers to epochs when Jupiter, center of mass, and the
Sun's center are in line. Conjunctions or oppositions of Jupiter and CM
occur at this time as well as SFSs.

Kind regards,

Theodor Landscheidt

So, there you go. It seems that the original figure has an incorrect
time scale. The correct scale is featured in figure 11 in the same
paper (according to Landscheidt) and in his other papers (I have
checked these out and they are correct...the arrows above the peaks
appear at the correct dates as listed in the text...the rainfall data
is identical in all cases).

I am satisfied with this explanation and I hope that you are too.
Perhaps we could now go onto Figure 11...?

w...@bas.ac.uk

unread,
Aug 25, 2000, 3:00:00 AM8/25/00
to
na...@my-deja.com wrote:
> na...@my-deja.com wrote:
>> The JU-CM-CS epochs are given in the text (according to Connelly) as:

>>1912, 1921, 1930, 1942, 1951, 1959, 1967, 1970, 1974, 1983 (to the nearest
>year).

> Obviously, the time scale in Figure 10 is not correct. In the


>printing process someone replaced my numbers for the years with bigger
>ones and put them in the wrong place. This happened after I had
>corrected the galley proofs. Figure 11 in Climate Change, which shows
>unsmoothed values of the same data, gives the correct scale.

...


>I am satisfied with this explanation and I hope that you are too.
>Perhaps we could now go onto Figure 11...?

I'm not so happy with moving the timescale like this, but lets assume
this for the sake of argument. That makes the tick marks on fig 10
correspond to xxx1 rather than xxx5 as published.

Then: the data range in fig 10 is 0-25. The data range in fig 11
is 0-14. Fig 11 is a plot of the data of fig 11, smoothed
"subjected to a low-pass gaussian filter" in the ambiguous language
of the text. That means the max of fig 10 should be less than the max of
fig 11. But it isn't. The text on the y axes of figs 10,11 is
clearly original and not changed by the journal.

As you say, this leads to fewer discrepancies - but there still are some.
There is an arrow at 1981 (which should have one) but certainly not one at
1983 (where there should be one).

I'd like to see the original data from which L drew his figure.

-W.

--
William M Connolley | Disclaimer: I speak for myself

na...@my-deja.com

unread,
Aug 26, 2000, 3:00:00 AM8/26/00
to
In article <39a6d...@news.nwl.ac.uk>,
w...@bas.ac.uk wrote:

> I'm not so happy with moving the timescale like this, but lets assume
> this for the sake of argument. That makes the tick marks on fig 10
> correspond to xxx1 rather than xxx5 as published.
>

Is this the same as on Fig. 11...? i.e. do the tick marks on Fig. 11
correspond to xxx1...?

> Fig 11 is a plot of the data of fig 11, smoothed...

Do you mean "Fig 10 is a plot of the data of fig 11, smoothed..."..?

> "subjected to a low-pass gaussian filter" in the ambiguous language
> of the text. That means the max of fig 10 should be less than the max
> of fig 11. But it isn't. The text on the y axes of figs 10,11 is
> clearly original and not changed by the journal.
>

Sounds reasonable. I'll see what he has to say about it...

> As you say, this leads to fewer discrepancies - but there still are
> some. There is an arrow at 1981 (which should have one) but certainly
> not one at 1983 (where there should be one).
>

Are you talking about Fig. 10 or Fig. 11...? There is no JU-CM-CS epoch
at 1981 so there shouldn't be an arrow there. Besides, as far as I can
make out from similar figures, the rainfall data only goes up to 1978
(the x axis maximum is 1980), so JCC epochs after this shouldn't
feature in the plot. The x axis range in the plots that I have is 1860
to 1980. see:
http://www.microtech.com.au/daly/solar/fig20.gif
In comparing Fig.10 with this figure, the first "o" corresponding to
minima in the rainfall data occurs at 1861 (+/- 0.5 years), and the
maximum peak in rainfall data occurs at 1959 (+/- 0.5 years). Notice
that the first peak in rainfall data in Fig. 10 isn't featured in the
figure above i.e. the start dates are not the same although the end
dates are.
All arrows in the plot above are correctly placed according to the
listed JCC data therefore the time axis must be correct.

> I'd like to see the original data from which L drew his figure.
>
> -W.

So would I...maybe he'll send it.

w...@bas.ac.uk

unread,
Aug 26, 2000, 3:00:00 AM8/26/00
to
na...@my-deja.com wrote:
>wmc:

>Is this the same as on Fig. 11...? i.e. do the tick marks on Fig. 11
>correspond to xxx1...?

Yes.

>> Fig 11 is a plot of the data of fig 11, smoothed...

>Do you mean "Fig 10 is a plot of the data of fig 11, smoothed..."..?

Yes. Oops.

>> "subjected to a low-pass gaussian filter" in the ambiguous language
>> of the text. That means the max of fig 10 should be less than the max
>> of fig 11. But it isn't. The text on the y axes of figs 10,11 is
>> clearly original and not changed by the journal.
>>
>Sounds reasonable. I'll see what he has to say about it...

Good.

>> As you say, this leads to fewer discrepancies - but there still are
>> some. There is an arrow at 1981 (which should have one) but certainly
>> not one at 1983 (where there should be one).
>>
>Are you talking about Fig. 10 or Fig. 11...? There is no JU-CM-CS epoch

11.

>at 1981 so there shouldn't be an arrow there. Besides, as far as I can
>make out from similar figures, the rainfall data only goes up to 1978
>(the x axis maximum is 1980), so JCC epochs after this shouldn't
>feature in the plot. The x axis range in the plots that I have is 1860
>to 1980. see:

No, the data is 1851-1983, and the figures are plotted to this range,
assuming the tickmarks are at xxx1. Of course, since they are running
2-y smoothed data (fig 11), and further smoothed in fig 10, not all of
that range is meaningful, possibly why later figures are chopped at 1980.

>http://www.microtech.com.au/daly/solar/fig20.gif

> In comparing Fig.10 with this figure, the first "o" corresponding to
>minima in the rainfall data occurs at 1861 (+/- 0.5 years), and the
>maximum peak in rainfall data occurs at 1959 (+/- 0.5 years). Notice
>that the first peak in rainfall data in Fig. 10 isn't featured in the
>figure above i.e. the start dates are not the same although the end
>dates are.
> All arrows in the plot above are correctly placed according to the
>listed JCC data therefore the time axis must be correct.

Yes, but notice the last peak in your figure - its at 1975 or close
to. That peak (the shape of the graph is the same) is present in
fig 10, but at 1980-ish (assuming that ticks are xxx1). And both have
arrows on. So, they are inconsistent. Similarly there is an "o"
in your url at or just before 1980, which is clearly at or
after 1981 on fig 10. So, whats up?

>> I'd like to see the original data from which L drew his figure.

>So would I...maybe he'll send it.

Perhaps we cuold draw better graphs than L...

na...@my-deja.com

unread,
Aug 27, 2000, 3:00:00 AM8/27/00
to
In article <39a82...@news.nwl.ac.uk>,
w...@bas.ac.uk wrote:

> No, the data is 1851-1983, and the figures are plotted to this range,
> assuming the tickmarks are at xxx1. Of course, since they are running
> 2-y smoothed data (fig 11), and further smoothed in fig 10, not all of
> that range is meaningful, possibly why later figures are chopped at
> 1980.
>
> >http://www.microtech.com.au/daly/solar/fig20.gif
>
> > In comparing Fig.10 with this figure, the first "o" corresponding
> >to minima in the rainfall data occurs at 1861 (+/- 0.5 years), and
> >the maximum peak in rainfall data occurs at 1959 (+/- 0.5 years).
> >Notice that the first peak in rainfall data in Fig. 10 isn't
> >featured in the figure above i.e. the start dates are not the same
> >although the end dates are.
> > All arrows in the plot above are correctly placed according to the
> >listed JCC data therefore the time axis must be correct.
>
> Yes, but notice the last peak in your figure - its at 1975 or close
> to. That peak (the shape of the graph is the same) is present in
> fig 10, but at 1980-ish (assuming that ticks are xxx1). And both have
> arrows on. So, they are inconsistent. Similarly there is an "o"
> in your url at or just before 1980, which is clearly at or
> after 1981 on fig 10. So, whats up?
>
> >> I'd like to see the original data from which L drew his figure.
>
> >So would I...maybe he'll send it.
>

> Perhaps we could draw better graphs than L...
>
> -W.

Could you post Fig.11 as you did for Fig.10 so that I could see the
time axis labels...?

Landscheidt *has* drawn better graphs than this (fig.10). This is an
early piece of work and the plot is very rough (his words: see later).
I don't like this "assuming that ticks are xxx1" business - it seems to
be causing problems and hindering progress more than anything else. I
think that we should concentrate on similar, newer figures plotted
using the same data which we know have correct axis labelling, eg. the
one in the URL above and the ones below.

Anyway, Landscheidt has sent another reply. Here it is:

Dear Dr. Harris,

The variance values on the different y-axes are relative values. I
concede that it would have been better to work with absolute values
that make valid comparisons possible.

I do not understand why William Connolly insists on discussing very
rough plots when figures with much better resolution are available. I
shall send you my original smoothed data via snail mail so that you can
see that the new plots are based on solid data. After so many years I
have the original rainfall data no longer in my computer. I would have
to go back to the original source I quoted in my paper. Connolly may do
this himself. As to Gaussian low pass smoothing the weights are
available in the literature, but I am willing to give them to
Connolly and tell him exactly how I proceeded. He will then find that
his result is the same as in my newer publications. I did not pretend
that the correlation is exact in each single case, but the statistical
result, I described in detail, is beyond doubt, especially as there are
significant replications based on other rainfall data.

There are other climatological results based on Ju-CM-CS, as for
instance Figures 4 and 5 in my new ENSO paper which I sent to you. You
may send him copies, as these figures were published before in another
context.

Best regards,

Theodor Landscheidt

The figures he is referring to are similar to ones already published on
his web-papers. They can be seen at:
Figure 4:
http://www.microtech.com.au/daly/solar/temps.htm

Figure 5:
http://www.microtech.com.au/daly/solar/fig24.jpg

The figure explanations can be found within the accompanying text.

w...@bas.ac.uk

unread,
Aug 27, 2000, 3:00:00 AM8/27/00
to
na...@my-deja.com wrote:
>In article <39a82...@news.nwl.ac.uk>,

>Could you post Fig.11 as you did for Fig.10 so that I could see the
>time axis labels...?

Very well. But it *really* is about time you
got a copy of this paper for yourself or gave up the discussion.
After all - you're in communication with L himself - how come he
can't send you the stuff?

http://www.wmc.care4free.net/sci/Dscn1627.jpg

Now we're onto fig 11, note an extreme oddity: the data in fig 11
is "unsmoothed 2-year running variance of the rainfall data". Now,
look carefully, you see that there is a tendency for peaks at regular
intervals (though of different heights). This interval looks liek
2.5 years to me, though it could be 2-year. That is odd in data
that is a 2-year running mean. Further - there are absolute
zeroes in the data as plotted. How can this be? Look for example
before the 1931 tickmark. There is a sharp peak, then a zero. I
think that is incompatible with a 2-year running mean.

>Anyway, Landscheidt has sent another reply. Here it is:

>The variance values on the different y-axes are relative values. I


>concede that it would have been better to work with absolute values
>that make valid comparisons possible.

Relative to what? There is absolutely no hint in the text that these
are relative not absolute values.

>I do not understand why William Connolly insists on discussing very
>rough plots when figures with much better resolution are available. I

Well, oddly enough, I prefer to discuss peer-reviewed stuff where
available. It would seem that in this case peer review has somewhat
failed (though actually its common to review and say "figures must
be improved" and its then up to the editor to enforce this).

>shall send you my original smoothed data via snail mail so that you can
>see that the new plots are based on solid data. After so many years I
>have the original rainfall data no longer in my computer. I would have

Isn't this curious? Didn't you supply me the URL of a recently-drawn
picture which uses this data?

>to go back to the original source I quoted in my paper. Connolly may do

Hmm, thats enough miss-spellings. Time to ask: did you supply my name
to him misspelt or is this his mistake?

>this himself. As to Gaussian low pass smoothing the weights are
>available in the literature, but I am willing to give them to

This is twaddle. A Gaussian filter means little without knowing the
width, which the paper does not give.

>Connolly and tell him exactly how I proceeded. He will then find that
>his result is the same as in my newer publications. I did not pretend

So... if its the same as in the newer publications, doesn't that mean
he still has the data available?

>that the correlation is exact in each single case, but the statistical
>result, I described in detail, is beyond doubt, especially as there are
>significant replications based on other rainfall data.

The other data is not graphed and so is unverifiable.

-W

Climate Modeller, British Antarctic Survey | Disclaimer: I speak for myself

na...@my-deja.com

unread,
Aug 28, 2000, 3:00:00 AM8/28/00
to
In article <39a97...@news.nwl.ac.uk>,

> w...@bas.ac.uk wrote:
> >na...@my-deja.com wrote:
> >In article <39a82...@news.nwl.ac.uk>,
>
> >Could you post Fig.11 as you did for Fig.10 so that I could see the
> >time axis labels...?
>
> Very well. But it *really* is about time you
> got a copy of this paper for yourself or gave up the discussion.
> After all - you're in communication with L himself - how come he
> can't send you the stuff?
>
> http://www.wmc.care4free.net/sci/Dscn1627.jpg
>
He doesn't have hard copies of the paper anymore, or copies on his
computer...(see below). I've asked a friend at Armagh Observatory if
he'll send a copy of the paper to me...they have an excellent
meteorology library there and it might have L's paper...

I'm not convinced about the x-axis (Time) labelings in Fig.11
either...they look as if they've been changed by the editor too.

Here's Landscheidt's latest reply:

Dear Dr. Harris,

Please, convey my apologies to William Connolley for misspelling his
last name though I copied exactly what I got in the corresponence, as
for instance "Connolly to Newsgroup". [oops...that was me! - NH]

Connolley intimates several times that I have still got the original
rainfall data in my computer, but nevertheless I tell you and him that
they are no longer available. In other words he alleges that I am a
liar. Scepticism is a valuable science tool. Argumenta ad hominem,
however, go beyond legitimate scepticism. An apology would be in order,
especially as the raw rainfall data are really no longer on my hard
disk. This is plausible as I last worked on the data in the mid-
eighties.

There is no circumstantial evidence to the contrary like "Isn't this


curious? Didn't you supply me the URL of a recently-drawn picture which

uses this data?".
Recently drawn pictures of this kind do not exist. The plot in the
publication on the internet, but also the figures in the Journal of
Coastal Research and in the NASA Conference Publication 3086 are simply
copies of a hardcopy produced in the mid-eighties. I sent you a copy of
this hardcopy. It is easy to see that it was produced by technical
means available in the eighties. Please, send Connolley a copy so that
he can check this. The copy of the original smoothed
data, which is on the way to you, is again an original hardcpopy from
the eighties which in no way indicates that the data are still in my
computer. Please, send a copy of this document to Connolley, too, so
that he can make his own plot and see whether the JU-CM-CS events
match. The Gaussian smoothing procedure is also described in detail. As
this is a document that could be needed as evidence, please send it
back to me when you have made your copies.

The other data which corroborate the discussed result are not graphed,
but not "unverifiable". The data sources were specified.

As to the comments on Fig. 11 Connolley does not consider that we are
not dealing with running means of the rainfall data, but with 2-year
running variance. Obviously, Connolley has never worked with this tool.

I still do not understand why Connolley is eager to discuss poor plots
while much better ones are available. The Journal of Coastal Research
in which a better plot was published is also peer reviewed, so this
should not be an obstacle.

Best regards,

Theodor Landscheidt

I certainly agree that analysis of a more recent peer-reviewed paper
with more accurate plots might be more productive.

Thomas Palm

unread,
Aug 28, 2000, 3:00:00 AM8/28/00
to
na...@my-deja.com wrote:
> He doesn't have hard copies of the paper anymore, or copies on his
> computer...(see below).

But Landscheid does have a computer? Perhaps then someone could
show him how to use newsgroups so that we could avoid this odd
three way conversation. I suspect a disuccsion directly between
Landsheid and Connoley would be more interesting than with
Nathan as a mediator. (The same goes for Courtney too, of course.)

na...@my-deja.com

unread,
Aug 29, 2000, 3:00:00 AM8/29/00
to
In article <39AAB89A...@chello.se>,

There's gratitude for you! ;-)
I have been trying to encourage them to contribute directly to the
newsgroup but, you know, old dogs and new tricks...
At the moment I think 3-way is better than 0-way.

w...@bas.ac.uk

unread,
Aug 29, 2000, 3:00:00 AM8/29/00
to
Thomas Palm <thoma...@chello.se> wrote:
>na...@my-deja.com wrote:
>But Landscheid does have a computer? Perhaps then someone could
>show him how to use newsgroups so that we could avoid this odd
>three way conversation.

you mean other people are listening...?

-W.

w...@bas.ac.uk

unread,
Aug 29, 2000, 3:00:00 AM8/29/00
to
na...@my-deja.com wrote:
>> w...@bas.ac.uk wrote:

>I'm not convinced about the x-axis (Time) labelings in Fig.11
>either...they look as if they've been changed by the editor too.

Yes indeed, it has been relabelled, but probably in the right places.

By the way, do you think that L intends to contact the editor of
Cli Ch and point out the error?

>Here's Landscheidt's latest reply:

>Connolley intimates several times that I have still got the original


>rainfall data in my computer, but nevertheless I tell you and him that
>they are no longer available.

OK, if he says so so explicitly.

>Please, send a copy of this document to Connolley, too...

No, don't bother.

>The other data which corroborate the discussed result are not graphed,
>but not "unverifiable". The data sources were specified.

By unverifiable I meant from within the local context of the paper.

>As to the comments on Fig. 11 Connolley does not consider that we are
>not dealing with running means of the rainfall data, but with 2-year
>running variance. Obviously, Connolley has never worked with this tool.

Yes, that is true. I haven't. In fact, reading the paper closer, its
not really clear what he is doing. The data is yearly total rainfall.
From this he derives "2-year running variances". What are these?
Can he possibly mean var(year, year-1)? That would be odd. Apart
from anything else, it would make the zeroes on fig 11 impossible.

>I still do not understand why Connolley is eager to discuss poor plots
>while much better ones are available. The Journal of Coastal Research
>in which a better plot was published is also peer reviewed, so this
>should not be an obstacle.

Well, as mentioned above, its not the quality of the plot its the
unclarity of what is plotted.

>I certainly agree that analysis of a more recent peer-reviewed paper
>with more accurate plots might be more productive.

Well, if anyone happens to put one on my desk, which is where this began...

-W.

w...@bas.ac.uk

unread,
Aug 30, 2000, 3:00:00 AM8/30/00
to
w...@bas.ac.uk wrote:
>natjo:

>>I certainly agree that analysis of a more recent peer-reviewed paper
>>with more accurate plots might be more productive.

>Well, if anyone happens to put one on my desk, which is where this began...

Malign chance has landed a copy of the coastal research paper on my
desk... you know who you are ;-) Now I'll have to read it...

Reply all
Reply to author
Forward
0 new messages