date: Tue, 1 Apr 1997 13:26:01 -0500 (EST)
from: ray bradley <rbradley@geo.umass.edu>
subject: standardisation
to: k.briffa@uea.ac.uk

Hiya Keef:
Hope you survived the journey back from Cape Town.  I hear you got hit with
a large bill at the hotel...maybe you covered my bar bill which did not
appear on my statement!!!

Could you look over this section which  I propose to add to my book, dealing
with the problems of standardisation.....& let me know if I got it
wrong...(The proposed Figures 10.9-10.11 are Figs 2- 4 from your Climate
Dynamics paper)...All comments would be appreciated.  I've enjoyed reading
your papers on this & other matters....

".......It is clear that this standardization procedure is not easy to apply
and may actually remove important low-frequency climatic information.  It is
not possible, a priori, to decide if part of the long-term change in ring
width is due to a coincident climatic trend.  The problem is exacerbated if
one is attempting to construct a long-term dendrochronological record, when
only tree fragments or historical timbers are available and the
corresponding growth function may not be apparent.
***added***  
        The consequences of different approaches to standardisation are well
illustrated by the studies of long tree-ring series (Scots pine, Pinus
sylvestris) from northern Fennoscandia by Briffa et al. (1990, 1992).  In
order to produce a long dendroclimatic reconstruction extending over 1500
years, Briffa et al. (1990) constructed a composite chronology made up of
many overlapping cores which varied in their individual length, from less
than 100 to more than 200 years (Figure 10.9).  In the shorter segments, the
growth function is significant over the entire segment length, but in longer
segments the growth factor becomes less significant (cf Figure 10.7, upper
panel).  In Briffa et al. (1990) each segment was standardised individually
(the procedure used in almost all dendroclimatic studies) in this case by
the use of a spline function which retains variance at periods less than
~2/3 of the record length.  Thus, in a 100 year segment, variance at periods
>66 years would be lost whereas in a 300 year segment, variance at periods
up to 200 years would be retained.    All standardised cores were then
averaged together, producing the record shown in Figure 10.10c.  This shows
considerable interannual to decadal scale variability, but little long-term,
low frequency variability.   In fact, since the mean segment length varies
over time (Figure 10. 9) so too will the low frequency variance represented
in the composite series.
         In Briffa et al. (1992) the standardisation procedure was revised
by first aligning all core segments by their relative age, then averaging
them (i.e. all values of the first year in each segment (t1) were averaged,
then all values of t2 etc.... to tn.; this assumes that in each segment
sampled, t1 was at (or very close to) the center [pith?] of the tree).  The
resulting 'regional curve' then provided a target for deriving a mean growth
function which could be applied to all of the individual core segments,
regardless of length (Figure 10.11).  Averaging together the core segments,
standardised in this way by the regional curve, produced the record shown in
Figure 10.10b.  This has far more low frequency information than the record
produced from individually standardised cores (Figure 10.10c) and retains
many of the characteristics seen in the original data.  From this series,
low growth from the late 1500s to the early 1800s is clearly seen,
corresponding to other European records which record a "Little Ice Age"
during this interval.  Also seen is a period of enhanced growth from
~950-1100 A.D., during a period which Lamb (1965) characterised as the
"Medieval Warm Epoch".  It is apparent from a comparison of Figures 10.9b
and 10.9c that any conclusions drawn about which were the warmest or coldest
years and decades of the past can be greatly altered by the standardisation
procedure employed.  All of the high frequency variance of Figure 10.10c is
still represented in the record produced by regional curve standardisation
(Figure 10.10b) but potentially important climatic information at lower
frequencies is also retained.  The problem of extracting low frequency
climatic information from long composite records made up of many individual
short segments is addressed explicitly by Cook et al. (1995) who refer to
this as the 'segment length curse'!  Although it is of particular concern in
dendroclimatology, it is in fact an important problem in all long-term
paleoclimatic reconstructions which utilise limited duration records (e.g.
historical data, composite coral records etc) yet it has received relatively
little attention in the literature so far.******


