Uncertainty, Emergence, and Statistics in Dendrochronology - Dendroclimatology - page 78

4.1 Introduction
The science of dendrochronology resides in a physical and biological world dom-
inated by uncertainty and emergence . The relative importance of each will vary
considerably from study to study, but neither uncertainty nor emergence can be
assumed to be absent from most, if not all, tree-ring studies, especially those that
seek to identify and extract environmental signals from tree rings. The closest
thing we have to absolute certainty in dendrochronology is the assignment of cal-
endar year dates to annual tree rings by an experienced tree-ring scientist using
some accepted method of crossdating (e.g., Huber 1943 ; Douglass 1946 ; Ghent
1952 ; Stokes and Smiley 1968 ; Baillie and Pilcher 1973 ; Heikkenen 1984 ; Wigley
et al. 1987 ; Schweingruber et al. 1990 ; Yamaguchi 1991 ; Yamaguchi and Allen
1992 ; Fowler 1998 ) . Without this foundation, dendrochronology ceases to exist as a
legitimate science.
Yet even here, dendrochronologists have not been immune from accusations of
less dating certainty than is claimed (i.e., zero dating uncertainty), hence the impor-
tance of the paper by LaMarche and Harlan ( 1973 ) that documented the accuracy
of the tree-ring dating of semiarid site bristlecone pine in California for calibration
of the radiocarbon ( 14 C) timescale. A similar study by Pilcher et al. ( 1984 ) further
illustrated the power of crossdating for absolute dating of tree rings in their doc-
umentation of the dating accuracy of the long European oak tree-ring chronology.
In turn, high-precision 14 C measurements of independently crossdated bristlecone
pine and European oak wood over the same time period have yielded very similar
long-term 14 C variations (Linick et al. 1985 ) . This and the confirmation of the radio-
carbon Suess ( 1965 ) 'wiggles' in dendrochronologically dated wood (de Jong et al.
1979 ) have independently validated the robustness of crossdating as a geochrono-
logical dating tool. These studies have collectively established crossdating as the
most precise and accurate dating method in geochronology.
Since that time, quality control programs like COFECHA (Holmes 1983 ) have
become indispensable tools for objectively testing the quality of crossdating and
the correct assignment of calendar year dates to tree rings. These computer-assisted
tools have convinced all but the most intransigent critics of dendrochronology that
crossdating works as claimed when it is applied by properly trained individuals.
This being the case, it is useful to briefly define and discuss now both uncertainty
and emergence in generic terms before proceeding to the heart of this chapter.
Simply put, uncertainty means that we do not know as much as we would like
about the information contained in the tree rings we are studying, and emergence
means that those same tree rings are likely to contain totally unexpected and inher-
ently unpredictable information about environmental effects on tree growth. The
effects of these principles on dendrochronological research will be examined in
the context of statistically identifying and extracting climatic signals from tree
rings, but they truly span the entire science of dendrochronology and its myriad
applications.