The last issue of the newsletter came out in December 2000, more than 2 years ago. There have been none since then for the simple reason that a flow of contributions slowed, and finally ceased. However, it is clear from web statistics and e-mail received that the articles available are accessed and found useful.
I have therefore decided to continue, but in a different style. I shall be posting news about new programs and software here on a continuous basis, as and when I get information. If anyone wants to volunteer text, or even write an article in the style of the old newsletter, I will be happy to post that. If anyone has comments about the usefulness of the newsletter, I would also be interested to hear those. KDB, 10th April 2003.
Those of you interested in software for developing palaeoecological transfer functions or plotting stratigraphic diagrams may like to take a look at my new program C2, which is now available for download at:
C2 Version 1.3 is a Windows 95/98/NT/2000/XP program for analysing and
visualising palaeoenvironmental data. The main features are: (1)
develop, diagnose and apply transfer functions using weighted averaging
(WA), partial least squares (PLS), weighted averaging partial least
squares (WAPLS), Imbrie & Kipp factor analysis, correspondence analysis
regression (CAR), principal component analysis regression (PCAR), modern
analogue teachnique (MAT), and maximum likelihood regression and
calibration (ML), all with leave-one-out, bootstrap or n-fold leave-out
cross-validation and sample-specific reconstructions errors, (2) create
x-y scatter, line and bubble plots to visualise species environment
relationships and to diagnose fitted models, (3) plot stratigraphic
diagrams with bar, line, silhouette and symbol graphs, overlay multiple
series on the same axes, each with optional error bars, and the ability
to combine multi-proxy data from different datasets measured at
different time or depth intervals in the same diagram, (4) scale
graphics in absolute units (mm) or fit to page, and cut & paste diagrams
to a favourite drawing package, and (5) import / export data to / from a
range of common formats including Excel, Access, Tilia and Cornell, and
ability to edit, sort, transform and merge datasets. A more detailed
description of the program and a user guide can also be found on my web
pages. The program download consists of a single executable that will
install the program, help files, user guide, tutorial, and example
datasets. C2 replaces my previous programs CALIBRATE, WAPLS, MAT and
PDP.
C2 is free to download and use but the data analysis and graphics
functions are restricted to a maximum number of 75 samples. A license
file that removes this restriction must be purchased to use C2 with
larger datasets. The data editing functions in C2 are not
size-restricted.
[S. Juggins, posted on QUATERNARY listserver, 070403]
psimpoll version 4.00, available from May 2002,
incorporated a major change associated with the handling of
calibrated radiocarbon dates. There is no one-to-one conversion of
radiocarbon ages onto a calibrated calendar year timescale. The
calibration curve is irregular, and the result of mapping a
radiocarbon age, with its symmetrical normal probability
distribution, onto the irregular calibration curve is an irregular
probability distribution, which cannot be treated statistically in
the same way as radiocarbon ages. psimpoll approaches the
problem through the output generated by the BCal calibration system. Other
calibration programs are available, but BCal offers two advantages:
psimpoll version 4.10, available from
http://www.kv.geo.uu.se/psimpoll.html
incorporates the approach described above, together with an additional
zonation method, called CONIIC, which resembles the familiar CONISS, but is
based on Information Content rather than Sums of Squares.
[K.D. Bennett, 10th April 2003]
Bwigg, written by J. Andrés Christen, is an interface to the
general purpose Bayesian calibration software mexcal, using
wiggle-matching A more general (however, not involving wiggle
matching) interface, for Bayesian calibration, may be found in BCal.
More details of Bwigg, and download, from:
http://www.cimat.mx/Bwigg/
Information about Diagram, an extension to the PolPal pollen data
system that allows plotting of pollen diagrams directly from the Clipboard
is available here. Contact the authors directly
about obtaining a copy.
The probability distribution of a calibrated date is irregular.
psimpoll handles this by using either the mode or the weighted
average mean as the value for the calibrated age. The mode has no
standard deviation associated with it, but the weighted average
does. If confidence intervals are requested, these are
obtained by drawing random numbers from the irregular distribution,
if the mode is being used, or from the normal distribution defined
by the weighted mean and its standard deviation.