Thursday, May 24, 2012


date: Wed Nov 8 12:45:44 2000
from: Tim Osborn <>
subject: New tree-ring density data

Scott & Mike,

Keith and I have created a new gridded tree-ring density dataset, by superimposing (as we discussed before I left, Mike) additional low-frequency temperature variability from the age-banded regional timeseries on to the existing gridded tree-ring density dataset that had been traditionally standardised and was therefore lacking in low frequency variance. I've put the new dataset onto holocene for you to pick up (/users/tosborn/data/schweingruber_mxdabd_grid.dat.gz). Once you have gunzip'd it, you'll see that the format is the same as before: columns are the 115 grid boxes, rows are the 595 years from 1400-1994. These data are actually our calibrated data (deg C anomalies wrt 1961-90), though you should make them dimensionless by normalising with their 1900-1960 mean and standard deviation prior to putting them through the Tapio Schneider regularized EM process. And of course, set all post-1960 to missing. The missing code in the file is -9.99.

Although the two Briffa et al. papers that I left with you are the main references to use for the data set and for the regional-mean reconstructions (the Holocene paper for the standardised ones and the JGR paper for the age banded ones), the gridding, the calibration of the gridded data set and the incorporation of the low-frequencies into the gridded data will all be written up in a different paper. The provisional reference for this is:
Osborn TJ, Briffa KR, Jones PD and Schweingruber FH (2000) Reconstructing summer temperature over the Northern Hemisphere since AD1400 from a tree-ring network. In preparation.
So that's the one to use if you wish to cite the gridded datasets.

How are the imputations going with the standardised gridded data?

Best regards


No comments:

Post a Comment