|
Response by Prof. Mann to File Usage The following response from Prof. Mann has been given to David Appell: The
McIntye and McKitrick ("M&M") Energy and Environment paper
criticizing the work of Mann, Bradley, and Hughes is wrong, says Mann. "...the
authors results are entirely spurious. The mistake made insures that the
estimates, in particular prior to 1600-1700, are meaningless." This
leads, Mann says, to "the use of series that are artificial combinations of
early [e.g. 15th-16th century] and late [e.g. 19th-20th] information
accidentally spliced together" with "no relation" to the proxy
data used by Mann et. al. in their 1998 (and subsequent) work. The spreadsheet file they used was a complete distortion of the actual Mann et. al. proxy data set, and was essentially useless, particularly in the earlier centuries. The authors had access to the full data, which has been available on a public ftp site for nearly two years. When they noticed, as described in their paper, some signs of problems with the Excel spreadsheet version of the data, one might think that they would have bothered to check the data available on our public ftp site. Reply
to Prof. Mann First, the correspondence below shows that FTP references were specifically sought from Prof. Mann personally; that Scott Rutherford provided pcproxy.txt (the "Rutherford File") and efforts were made with both Scott Rutherford and Mann to verify that pcproxy.txt was correct. Secondly, as also shown below, our "corrected and updated" file proxy4.txt contains fresh calculations for series #69-80 and #84-99 using source data from the World Data Center for Paleoclimatology. The collation errors made by Rutherford in the Rutherford file are almost certainly limited to series #69-99 (many of the other series being traceable in WDCP data.) Only series #81-83 from the Rutherford file were used in revised proxy4.txt , which extend back to 1450, 1600 and 1750 respectively and do not affect the early 15th century values. Thirdly, the above response deals only with Audit Issues 2 and 3. The other audit issues remain outstanding. SM Oct. 28, 2003 A. Chronology of Data Correspondence 1. On April 8, a request for an FTP location was made to Mann directly - not to an associate of Prof. Mann. No request was made for an Excel spreadsheet. Dear Dr. Mann, I have been studying MBH98 and 99. I located datasets for the 13 series used in 99 at ftp://eclogite.geo.umass.edu/pub/mann/ONLINE-PREPRINTS/Millennium/DATA/PROXIES/ (the convenience of the ftp: location being excellent) and was intereseted in locating similar information on the 112 proxies referred to in MBH98, as well as listing (the listing at http://www.ngdc.noaa.gov/paleo/ei/data_supp.html is for 390 datasets, and I gather/presume that many of these listed datasets have been condensed into PCs, as mentioned in the paper itself. Thank you for your attention. Yours truly, Stephen McIntyre, Toronto, Canada 2. On April 9, Mann didn't seem to know where the FTP site was (!) and referred the matter to Scott Rutherford. Dear Mr. McIntyre, These data are available on an anonymous ftp site we have set up. I've forgotten the exact location, but I've asked my Colleague Dr. Scott Rutherford if he can provide you with that information. best regards, Mike Mann 3. On April 9, after the referral from Prof. Mann, the request for an FTP location was sent to Scott Rutherford. Dear Dr. Rutherford, Prof. Mann indicated that you would be able to locate the ftp for the MBH proxies and I would appreciate it if you could forward me the information. Thank you for your attention. Yours truly, Steve McIntyre 5. On Apr. 11, Rutherford didn't know where the FTP data was. Steve, The proxies aren't actually all in one ftp site (at least not to my knowledge). I can get them together if you give me a few days. Do you want the raw 300+ proxies or the 112 that were used in the MBH98 reconstruction? Scott 6. On April 11, I replied to Rutherford further showing that no request was being made for data to be organized in spreadsheet. [Scott] 7. On April 22, I sent a reminder to Rutherford. [Scott] Any
luck with this? 8. On April 22, Rutherford apologized for the delay. Steve, I was just going to send you an e-mail. Honest. I haven't forgotten about you. Bear with me a little longer. Your request is just about at the top of the stack. Regards, Scott 9. Around April 23, Rutherford. sent the text files "pcproxy.txt" and weights.txt. The exact file sent by Rutherford is shown here. Mann now says that "An error was made in preparing this Excel file, in which the early series were successively overprinted by later and later series." The file is not an Excel file. This leads, Mann says, to "the use of series that are artificial combinations of early [e.g. 15th-16th century] and late [e.g. 19th-20th] information accidentally spliced together" with "no relation" to the proxy data used by Mann et. al. in their 1998 (and subsequent) work." If errors were made by Scott Rutherford of the type described by Mann, then they affect series #69-99 in the file pcproxy.txt and do not affect any of the other files. In our reconstruction, we have completely recalculated series #69-80 and series #84-99 from original (and reliable) World Data Center for Paleoclimatology data. Accordingly, our recalculation only contains the alleged Scott Rutherford data for series #81-83, which extend back to 1450, 1600 and 1750 respectively and do not affect the early 15th century calculations. We are in the process of reconciling this file against the additional FTP references now provided by Mann. 10. On Sept. 9, 2003, I sent the following question to Prof. Mann raising questions about his temperature principal component calculations and received no reply. Dear Prof. Mann,
I have tried
diligently to reconstuct your termperature principal components as described in
MBH98, but without success and would appreciate some assistance. I downloaded
hadcrut2.dat from CRU (July 2003 edition), truncated the data to 1902-1995 and
further truncated it to the 1082 cells at gridpoints.loc and arranged as 1082
time-series with 1128 monthly readings. This step was successful as I could
match your map of cell locations. I standardized each series to mean 0 and sd 1
for the period 1902-95. In MBH98,
you say that you carried out “conventional” PCA, but there is so much
missing data that conventional PCA failed when I tried. In particular, 4 cells
had no values at all and I don’t see why they were included in your selection.
Most PCA algorithms balk at missing data or exclude it. How did you deal
with the extensive missing data? I downloaded the
EOFs, PCs and eigenvector loadings from ftp://eclogite.geo.umass.edu/pub/mann/MANNETAL98/EIGENVECTORS/ .
I spliced the EOFs into a 16x1082 matrix and the PCs (pc01.out, etc.)
into a 92x16 matrix. I made a diagonal of the first 16 values in column 2 of “tpca-eigenvals.out",
which look like eigenvalues, and carried out an expansion.
I then deducted the grid-box values generated from this expansion from
the Jones data as above; calculated variance for each year across available
cells and made a sum, comparing this to the variance similarly calculated in the
standardized Jones data. I
obtained very low/much lower explained variance from this than you got.
I also tried some experiments and it also doesn’t seem to me that the
first 16 EOFs maximize explained variance, as they should.
I would appreciate any assistance or clarification which you could give. Regards, Steve
McIntyre 11. On Sep. 10, 2003, I raised questions to Rutherford about potential collation errors in the file pcproxy.txt . Dear Scott, In the file pcproxy.txt which you sent me a while ago, a disporoportionate number of series start in the years *99 or *49, for example, series 73-83, 86-92 and 106-107. Is this intentional? Is this also the case in the underlying data? Thanks, Steve 12. On. Sep. 11, 2003, Rutherford answered as follows. Rutherford's reply suggests that he was not the one who collated pcproxy.txt - that it was "before his time". Steve, These would be the principal components (PCs) of dense tree ring networks. As trees drop out of the network the PCs are recalculated every 50 or 100 years. So I think what is going on is that once the tree network is too sparse, Mike stopped calculating the PCs at the last step where there was a decent network. (This was before my time, but I'm pretty sure that's story.) Scott 13. On Sept. 25, I sent the following inquiry to Prof. Mann on pcproxy.txt, seeking, among other questions, specific verification that pcproxy.txt was the correct file. Dear Prof Mann 14. On Sep. 25, Prof. Mann failed to provide such verification and stated that he would not respond to further inquiries. Dear Mr. McIntyre, The results,
incidentally, are insensitive to this step; essentially the same reconstruction
is achieved if a calibration period terminating in 1970 (prior to the
termination of any of the proxy series) was used instead. |