MM2003 criticise MBH1998 on many counts, some related to deficiencies in the description of the data used and possible irregularities in the data itself. These issues have been largely resolved in Mann et al. (2004) [the Corrigendum].
Did Juckes carry out any due diligence in order to make the latter statement? Because I’m not sure what issues were “resolved” in the Corrigendum. Today I’ll mention one amusing issue that definitely wasn’t resolved in the Corrigendum.
In MM03, we reported that the instrumental precipitation record for the New England gridcell used in MBH98 did not match any historical data from the area or from the citation, but did match historical precipitation from Paris, France. (“The rain in Maine falls mainly in the Seine”). In experimenting subsequent to MM03, I determined that the MBH precipitation series assigned to the South Carolina gridcell matched a series from Toulouse, France. I was unable to match the MBH Bombay gridcell series to historical data from Bombay or anywhere else, but it was somewhat similar to the series from Philadelphia.
In our Materials Complaint to Nature, we requested details on the provenance of these series. The answer in the Corrigendum was “NOAA” – nothing more. However the Corrigendum SI stubbornly retained the fiction that MBH98 proxy data included precipitation from New England and South Carolina and Bombay. I asked Nature for further particulars on the actual provenance of this data, but they refused to provide it. Maybe Juckes can resolve this conundrum. How about it, Marty? Where the hell do the MBH98 precipitation series actually come from?
This is an interesting illustration of the teleconnection principle. In Mannian statistics, incorrect geographical locations “don’t matter” because of teleconnections. Rain in Maine or rain in Spain – doesn’t matter, put them in the teleconnection machine. Assume that they are temperature plus noise. See – that was easy.