Climatic Inferences from Dendroecological Reconstructions - Dendroclimatology - page 270

Overall, the Southwest climate and fire comparisons (Figs. 9.3 and 9.4 ) not only
illustrate the close coupling between wet/dry patterns and fire in this region, but
they also provide a strong confirmation of the high fidelity of these two types
of dendrochronological reconstructions. The regional reconstructed climate and
fire time series are based on independently collected and quite different types of
tree-ring data. The climate-sensitive trees used to develop the drought and Niño-3
reconstructions generally come from dry, rocky sites where fire spread is unlikely,
and ring widths are the measured variable. The fire-scar records generally come
from ponderosa pine and mixed conifer forests in the same region, but typically
in sites where grass and pine needle understories could carry widespread fires.
Fire scars are the primary observed variable in these time series. Despite these
differences, the remarkable coincidence of extreme regional dry (wet) years and
high (low) fire occurrence suggests that the two kinds of network-based reconstruc-
tions are accurately identifying ecologically effective climatic conditions at annual
resolutions.
9.2.2 Western Spruce Budworm Outbreaks and Climatic
Entrainment
One of the most widespread forest defoliators in the western United States is
the western spruce budworm ( Choristoneura occidentalis ). During its larval stage,
this Lepidopteran moth is a voracious feeder on the buds and needles of true firs
( Abies species) and Douglas-fir ( Pseudotsuga menziesii ) trees. The name 'spruce'
budworm is a bit of a misnomer, because although spruce trees ( Picea spp.) are
minor hosts, the true firs and Douglas-fir typically suffer much heavier defoli-
ation and higher mortality rates than spruces. Another closely related budworm
species ( Choristoneura fumiferana ) occurs throughout eastern Canada and parts of
the northeastern United States, where it is the most widespread and important defo-
liator of conifers. Numerous tree-ring studies have reconstructed outbreak histories
of eastern and western spruce budworm (e.g., Blais 1981 ; Hadley and Veblen 1993 ;
Morin et al. 1993 ; Swetnam and Lynch 1993 ; Krause 1997 ; Jardon 2001 ; Ryerson
et al. 2003 ) , making this an excellent candidate for development of regional (and
perhaps continental) networks of outbreak chronologies.
The tree-ring basis for developing outbreak chronologies is the observation
of very sharply reduced ring growth in the host species during the defoliation
episode, which typically lasts for a decade or longer (Fig. 9.5 ) . The methods involve
multiple steps for confidently identifying reduced growth periods as outbreaks
(including distinguishing the outbreak signals from potentially confounding climate
effects; e.g., droughts), and compiling the observations into stand-level chronolo-
gies of outbreak events and subregional composites (see Swetnam et al. 1985 ;
Swetnam and Lynch 1993 ; and Ryerson et al. 2003 ) . Specifically, the process we
have followed includes: (1) the separate development of 'host' and 'non-host' tree
species ring-width chronologies from nearby sites, (2) detailed graphical and sta-
tistical comparisons of these chronologies against each other and with independent