It’s not called the “hokey stick” for nothing.
The e-mail is below.
###
from: “Rob Wilson”
reply-to: “Rob Wilson”
subject: Re: 1816 tropical SSTs
to: “Phil Jones”
Morning Phil,
no the Nature paper is a TR based NH extra-tropical temperature
reconstruction that I wrote
with Rosanne D’Arrigo.
Essentially a substantial update (1250 years in length) to the original
Jacoby and D’Arrigo
work.
The coral paper is a different beast. I am leaning towards JGR for submission
as I don’t
think it has the restrictive word length of GRL.
I can send you the last draft of the coral paper if you like, although I hope
to tighten it
in the next week or so.
For the coral reconstruction, I did not include the New Caledonia data.
I set up a fairly stringent screening procedure for the coral data – i.e.
only those series
that correlated significantly with local annual SSTs using both unfiltered
and 1st
differenced time-series were used for later analysis.
The new Caledonia data did not pass this test. This is, in fact, a problem
with many of the
coral data-sets – the long term trends in these data might be related to
salinity changes
rather than SST. However, due to strong linear trends, there is often a
reasonable
correlation with local SST using unfiltered series, which is purely an
artefact of trend.
Hence my use of 1st difference transforms to identify a ‘pure’ temperature
signal. The new
Caledonia data did not correlate significant with local SSTs at these high
frequencies –
hence, using my empirical approach, I could not rationalise the record to be
a true
temperature proxy.
Although a strict interpretation of the new Caledonia O18 record as a
temperature proxy
would indicate cooler conditions around the 1810-1820 period, I do not think
such an
interpretation is entirely correct. Also, through this period, the lowest
index value is
1814, not 1816. 1885 is an inferred cool year which could tie in nicely with
Krakatoa
however.
It is also interesting to note that many of the coral records utilised by
Mike Mann also
failed my screening procedure.
The attached figure shows normalised series of the 5 coral records that go
back to 1801 in
my ‘SST sensitive’ data-set. The y-axis has been inverted as the series are
negatively
correlated to SSTs.
Only one record (MAL = Malindi: western Indian Ocean – Julia Cole) shows an
inferred cool
year in 1816.
As I state in the paper – lots of potential for corals, but there is simply
not enough data
prior to 1850 to derive particularly robust reconstructions.
just say if you want a copy of the paper
hope this helps
best regards
Rob
This should be credited to Climategate 2.0
http://www.ecowho.com/foia.php?file=1944.txt
Hmm, he doesn’t think it’s correct that there was a cooler period around 1810-1820. Perhaps he should read some history and check up on a thing call “The Year without a Summer”.
This is a problem with proxies. In the case of trees, the tree ring does not record a response to temperature, but rather a response to then prevailing environmental conditions of which temperature is one, but not the only factor. indeed, it may not even be the most important factor. For example in some growing seasons growth will be limited by drought and in any event the growing season is not all year but often only just a few months of the year.
Corals suffer from similar drawbacks. Sea temperature is just one of the prevailing environmemtal conditions that determines growth.
Regrettably, it is extremely difficult to extract just the response to temperature for the more general response to environmental conditions as a whole. This is one reason why little certainty can be placed in proxy evidence. Proxy evidence is notoriously unreliable and if honest, it carries huge error bars.
It was interesting to see the comments to a temperature fall pre Krakatoa. Perhaps there was such a fall. Or perhaps the author was right that the “Caledonia data did not correlate significant with local SSTs at these high frequencies” and should therefore be disregarded. But if the Caledonia data was unreliable, it demonstrates that there is reason to doubt the accuracy of coral data to record response to SST and it therefore implies that particular caution should be had when considering coral proxy data in general.
The climate community seek to play up the role of volcanoes partly because there is no real data on them and so assumptions made as to the extent of volcanic activity is a useful tool in tweaking model response to track the thermometer record.
Interesting to note that there seem to be normal people among Phil’s contacts.