86
Cbapter 5: Interpreting Higb-Resolution Proxy Climate Data
the cost might be low statistical signal which would then require high sampIe replication to maintain chronology confidence. Where it is not possible to achieve this, the uncertainty inherent in conservatively standardized
chronologies should be explicitly recognised as contributing to uncertainty in
climate reconstructions produced from them.
These uncertainties, and the general implications of tree-ring standardization, are recognised by dendroclimatologists, but they are often not explicit in
the literature. It is not widely appreciated, outside ofthe field, that published
climate reconstructions do not necessarily represent the complete spectrum
of climate forcing of tree growth. Just because a tree-ring chronology (or
a climate reconstruction derived from it) is hundreds or even thousands of
years long, it does not follow that century- or millennial-timescale variability
is represented in it. The standardization of the primary data may preclude
this.
5.5.3 General Relevance of the "Standardization"
Problem
In tree-ring work the need to standardize measurement series is explicit.
However, in other proxy data fields, it may be necessary to recognize the
modification of long timescale signals by processes unrelated to the forcing
of interest. Lake varve thickness may be moderated slowly by changes in
the surface area of deposition or alteration of the surface run off into the
lake (as could occur because of isostacy or surrounding vegetation change).
Similarly the thiekness of annual layers in iee eores may become thinner at
depth or be affected by iee flow within an ice sheet because of compaetion
or bottom topography. Growth of marine corals might be influenced by
long-term land run-off ehanges that are unrelated to climate change or may
be difficult to interpret because of their eomplex three-dimensional shape.
There are only speculations, but as replieated series of data derived from
such sourees become increasingly available for eenturies and millennia, it will
be incumbent on researchers in these fields to establish the degree to whieh
they may need to be standardized to take account of these (or other factors)
and the degree to which such standardization may limit the timescales of
climate variability that can be subsequently reconstructed from them.
5.6 Quantifying Climate Signals
in Chronologies
The prineipal consideration in any tree sampling strategy is to select sites
where the climate variable of interest is expected to exert a strong limitation
on seasonal growth (La Marche, 1982). Whatever the expectation, however,
it is still necessary to demonstrate the actual strength and character of the
Cbapter 5: Interpreting Higb-Resolution Proxy Climate Data
the cost might be low statistical signal which would then require high sampIe replication to maintain chronology confidence. Where it is not possible to achieve this, the uncertainty inherent in conservatively standardized
chronologies should be explicitly recognised as contributing to uncertainty in
climate reconstructions produced from them.
These uncertainties, and the general implications of tree-ring standardization, are recognised by dendroclimatologists, but they are often not explicit in
the literature. It is not widely appreciated, outside ofthe field, that published
climate reconstructions do not necessarily represent the complete spectrum
of climate forcing of tree growth. Just because a tree-ring chronology (or
a climate reconstruction derived from it) is hundreds or even thousands of
years long, it does not follow that century- or millennial-timescale variability
is represented in it. The standardization of the primary data may preclude
this.
5.5.3 General Relevance of the "Standardization"
Problem
In tree-ring work the need to standardize measurement series is explicit.
However, in other proxy data fields, it may be necessary to recognize the
modification of long timescale signals by processes unrelated to the forcing
of interest. Lake varve thickness may be moderated slowly by changes in
the surface area of deposition or alteration of the surface run off into the
lake (as could occur because of isostacy or surrounding vegetation change).
Similarly the thiekness of annual layers in iee eores may become thinner at
depth or be affected by iee flow within an ice sheet because of compaetion
or bottom topography. Growth of marine corals might be influenced by
long-term land run-off ehanges that are unrelated to climate change or may
be difficult to interpret because of their eomplex three-dimensional shape.
There are only speculations, but as replieated series of data derived from
such sourees become increasingly available for eenturies and millennia, it will
be incumbent on researchers in these fields to establish the degree to whieh
they may need to be standardized to take account of these (or other factors)
and the degree to which such standardization may limit the timescales of
climate variability that can be subsequently reconstructed from them.
5.6 Quantifying Climate Signals
in Chronologies
The prineipal consideration in any tree sampling strategy is to select sites
where the climate variable of interest is expected to exert a strong limitation
on seasonal growth (La Marche, 1982). Whatever the expectation, however,
it is still necessary to demonstrate the actual strength and character of the