Climatic Inferences from Dendroecological Reconstructions - Dendroclimatology - page 264

Networks of well-dated disturbance and demographic chronologies can be aggre-
gated at multiple spatial scales to enhance common patterns, just as is done in
regional networks of ring-width or ring-density chronologies. Climatic influence
emerges as patterns of synchrony (or asynchrony) of events, oscillations, or trends.
Very broadscale synchrony (at >10 4 km 2 scales) is typically related to climate vari-
ability affecting the co-occurrence of ecological events in many places, because
most ecological disturbances or processes are not capable of physically spreading
over such large areas. For example, fires generally can not burn across the enormous
desert valleys and canyons that separate forested mountain ranges in the American
Southwest. Therefore, synchrony of annual fire events (or absence of fire events)
among these widely dispersed mountain ranges has to be related to regional-scale,
fire-inducing (or -suppressing) climatic conditions, such as regional drought (wet)
events (Swetnam and Baisan 1996 ) .
In general, synchrony of disturbance events and demographic processes at
regional or broader scales can be inferred to reflect climatic influences. An apt anal-
ogy is crossdating of tree-ring measurements among trees and sites. In the case
of tree-ring widths or densities, it is understood that numerous unique or local
factors (e.g., soils, competition, genetics, human land uses, etc.) at the scales of
trees and forest stands may affect cambial growth and thereby introduce nonsyn-
chronous, non-climatic signals in tree-ring time series (Cook 1990 ) . However, when
sites and trees are carefully selected to maximize climatic sensitivity and ring series
from many locations are combined, the existence of broadscale crossdating (i.e.,
synchrony) of ring patterns across regions is logically and demonstrably related—
both statistically and mechanistically—to climatic variations (e.g., Fritts 1976 ) .
Likewise, disturbance and demographic events are affected by numerous local fac-
tors, but if common, synchronous patterns emerge across broadscale networks, those
ecological patterns are most probably climatic in origin. Interpretations of these
causal relations are most robust when coupled with mechanistic understanding of
climatic/ecological processes, such as the effects of water balance on fuel productiv-
ity, fuel moisture content and fire occurrence (e.g., Swetnam and Betancourt 1998 ;
Westerling et al. 2006 ) .
Assembly of regional- to continental-scale networks of disturbance and tree
demographic histories is just beginning (see the International Multiproxy Paleofire
Database at: http://www.ncdc.noaa.gov/paleo/impd /). Analogous development of
tree-ring width networks in North America and South America began in the
1950s with Edmund Schulman's work (Schulman 1956 ) , and continued through
the 1970s and 1980s with Harold Fritts' efforts (Fritts 1976 , 1991 ) . Many other
dendrochronologists have assembled tree-ring data networks and contributed these
to publicly accessible databases (see the International Tree-Ring Data Bank
at http://www.ncdc.noaa.gov/paleo/treering.html ). These datasets have now been
applied in continental to global-scale dendroclimatic reconstructions (e.g., Fritts
1991 ; Briffaetal. 1994 ; Mann et al. 1998 , 1999 ; Cook et al. 1999 , 2004 ) . In
addition to providing an example for dendroecology of the power of the network
approach, dendroclimatic networks provide a tremendous source of comparative
data. Replicated time series of reconstructed hydroclimate variables or global