Climatic Inferences from Dendroecological Reconstructions - Dendroclimatology - page 278

Fig. 9.7 Tree establishment and mortality dates from Austrocedrus chilensis woodlands in north-
ern Argentina driven by fluctuations in effective moisture (Villalba and Veblen 1997a ) . (Reprinted
from Villalba and Veblen, with permission from Journal of Ecology, Wiley-Blackwell)
a certain size class within fixed plots or belt transects. There were also differences
in the establishment date (period) resolution in the datasets, including 5-, 10-, and
20-year periods.
Regional patterns of synchrony emerge in an overall compilation of the 12
time series in a ponderosa pine 'establishment index' (Fig. 9.10 ) . This time series
was derived from the logarithms of the time series values (tree numbers and tree
densities; 1.0 was added to all values so that logarithms could be computed for
years/decades with zero values) and the ratio of each log-transformed value to the
mean of the entire log-transformed series. The regional time series is the average
of all the establishment indices computed for each series (site). Logarithmic trans-
formations are useful here because there is a strong tendency for recent periods
(especially after 1900) to have one to several orders of magnitude more trees estab-
lishing (or surviving) than before this period. Hence, to evaluate relative patterns, the
log transform and ratio index help to standardize all datasets for visual comparisons
and combination in the regional index time series.
Overall, the regional composite (Fig. 9.10 ) shows that ponderosa pine forests
are typically uneven-aged and the age structures are often dominated by episodes
of tree establishment (cohorts) that alternated with multi-decade periods when rel-
atively fewer trees established in the stands. Moreover, the composite suggests that
there was substantial regional synchrony of these episodes of establishment (and
lack of trees establishing), indicating the likely influence of broadscale climate forc-
ing. In general, the largest cohorts correspond with wetter conditions, as is shown
in the comparison with dendroclimatic reconstructions of drought indices (Cook
et al. 2004 ) from the region (Fig. 9.10 ) . In particular, the 1810s-1860s and 1890s-
1930s cohorts coincide with exceptionally wet periods. Decreases in establishment