Dendroclimatology in High-Resolution Paleoclimatology - Dendroclimatology - page 24

per climate parameter). A substantial literature exists in which the strengths and
weaknesses of this empirical-statistical approach are investigated, and many of the
most important issues are further explored by Cook and Pederson ( Chapter 4 ,this
volume). They particularly emphasize the necessity for testing the stability of any
such statistical models, whether they are being used as response functions to esti-
mate tree-ring variables or as transfer functions to estimate past climate from tree
rings.
2.3.2 Identifying Climate Signal—Process-Modeling Approaches
From the inception of dendroclimatology, care has been taken to evaluate the eco-
logical reasonableness of the statistical models used to extract climate signals from
tree-ring data (Fritts 1976 ) . The mechanistic bases for this evaluation are now sum-
marized in simulation models; for example, those focusing on the activity of the
tissue that forms wood, the vascular cambium, as discussed by Vaganov et al.
( Chapter 3 , this volume). Evans et al. ( 2006 ) have shown that a relatively simple
process-based model (the Vaganov-Shashkin, or VS model) achieved skill compa-
rable to that of regression models fitted to each of almost 200 chronologies from a
wide range of conditions in the United States and Russia, even though the simulation
model used a single set of parameters for all of these, and was not tuned chronology
by chronology, unlike the statistical model. Anchukaitis et al. ( 2006 ) demonstrated
that varying only a single model parameter produced a similarly impressive per-
formance of the VS model for a combination of species and region for which it
performed poorly in the Evans et al. ( 2006 ) work, underlining the generality of
its applicability. This result indicates that there is a mechanistic basis for the use
of tree-ring variables as proxy climate records, based primarily on the environ-
mental control of the activity of the vascular cambium. As Vaganov et al. ( 2006 ;
Chapter 3 , this volume) indicate, there is a growing understanding of the ecophysi-
ological, developmental, and even molecular mechanisms of the climatic control of
tree-ring formation.
Recent progress in the observation of the timing of events in the vascular
cambium of conifers in alpine and subarctic conifers has revealed a number of
remarkably consistent common features of the environmental control of xylogen-
esis, the process of wood formation, in cool-region conifers. These features include
the existence of threshold temperatures for the onset of xylogenesis (Rossi et al.
2008 ) , which are most consistent when measurements are made at the stem, rather
than in the air or soil (Rossi et al. 2007 ) , and tend to be the same regardless of
location and species. They also show a tendency for the maximum rate of cell pro-
duction and xylem growth to occur close to the summer solstice, not necessarily at
the warmest time of year (Rossi et al. 2006 ) . In fact, cell division in the cambium
stopped in July or August, when temperatures were still high (Rossi et al. 2008 ) .
Process modeling of the climate control of tree-ring growth can only benefit from
the continuation of this kind of work and the incorporation of its results into existing
and new models.