| Introduction |
Austrocedrus
chilensis is a widely distributed arid site conifer of the forest-steppe
transition zone in Argentina and Chile (from latitude 32oS to
43oS). A. chilensis has proven to be an excellent
species for dendroclimatic applications (Schulman, 1956; Villalba, 1994)
especially for the reconstruction of precipitation. The species can be
found on a variety of site types and exposures.
We hypothesized that the
best crossdating and climatic signal would be found in A. chilensis
growing on the most extreme arid sites. We also hypothesized that mortality
events mainly represent past drought episodes. We also compared the
A. chilensis chronology developed for this study with a regional network
of existing sites to assess regional growth patterns that may relate to
larger scale climate variability.
|
| Site
Description |
We selected two A. chilensis
stands growing on two contrasting sites found on a west facing slope near
San Martin de los Andes to test this hypothesis.
Site 1. A basaltic
cliff face with thin soils, little herbaceous vegetation and no fire evidence.
Dwarf A. chilensis trees were present and exhibited characteristic
signs of old growth including strip bark, dieback, twisted and gnarled
branches.
Site 2. At the
base of the cliff face and on other flat areas where volcanic soils were
deep, the A. chilensis trees were larger and abundant herbaceous
vegetation covered the forest floor. These deep soil positions were
relatively more mesic compared with the rocky site and included various
species of Nothofagus. They appear to have been disturbed by logging
and fire in the recent past.
|
| Methods |
Crossdating
We collected core samples
of living A. chilensis on the rocky and deep soil sites to test
the hypothesized difference in crossdating qualities and climatic response.
We also collected cross-sections from dead A. chilensis to determine
their death dates and test the hypothesis that mortality occurs most often
during extended drought events.
Regional Climate Signal
Do tree-ring chronologies
in northern Patagonia have a regional climate signal? Does the San Martin
de los Andes chronology exhibit that signal?
For this analysis, we used
two sources of data: a set of 25 tree-ring chronologies from Villalba and
Veblen (1997) and two chronologies from San Martin de los Andes. We combined
the two San Martin chronologies into a single composite chronology, for
a total of 26 chronologies from northern Patagonia. All chronologies were
detrended and standardized with a -2/3 spline. We employed principal components
analysis (PCA) to detect any modes of variation in the distribution of
loadings among the chronologies. Finally, we used Varimax rotation after
the initial PCA to redistribute variance more accurately.
|
| Results |
Crossdating
Figure 1 demonstrates that
ring-width index series from both sites show drought stress, but the soil
chronology seems be more coherent, while the scatter is larger in the rock
chronology showing more between tree variation in growth patterns. There
are three possible explanations:
1) The trees
on the rocky cliff are often damaged by rock fall.
2) The rock
site actually includes a range of microsite conditions: some trees
have very little available soil and are dwarfed, while trees on ledges
grow on deeper soil and experience better growing conditions.
3) Overall,
the rock site is extremely harsh, marginal for tree growth. The rock site
shows a wider variation in growth than the deep soil site.
Typically the raw ring width
chronology should exhibit a decrease in ring size associated with the increasing
age of trees. While this is true for our rock site, the soil site shows
a period of increasing ring width circa 1920 associated with a sample bias
due the recruitment of young trees during this period.
 |
 |
|
Figure 1. Ring-width index series
for soil and rock sites.
|
 |
|
Figure 2. Chronologies
for rock and soil sites
|
In mesic western forests
of South America, earthquakes result in massive recruitment result of trees
on landslide sites. Within our research area, the site evidence suggests
that fire is the most pervasive disturbance creating extensive even-aged
stands of Austrocedrus chilensis (Veblen et al., 1992). The increase
in young trees on the soil site may be associated with an increase in the
fire regime due to European colonial settlement at the turn of the 19th
century (Veblen et al., 1992). While these fires killed many of the young
trees on the deeper soil, the trees on the rock site were protected.
Figure 2 shows that drought
extremes appear to impact tree growth more severely on the rocky site.
Notice the many times when the rocky low growth extremes exceed the severity
of low growth on the soil site. The arrow indicates low tree growth and
mortality of A. chilensis associated with a very severe dry event
(c. 1955). |
|
Regional Climate Signal
We compared the 2 factors
derived from the 26 chronologies with latitude and longitude to determine
the geographic pattern associated with the distribution of the PCA loadings.
Loadings for the first PCA factor were centered roughly in the center of
the geographic distribution (40º), near the locale for the San Martin
de los Andes chronology. The first factor is therefore clearly not associated
with a geographical gradient. Loadings for the second factor, however,
were distributed along a latitudinal gradient. A N-S gradient from
37º to 42º south latitude causes detectable differentiation
in tree rings across the observed range of A. chilensis (Figure
1). The highest loadings were from chronologies in the south-westernmost
portion of the observed range of Austrocedrus near the Rio Futaleufu whereas
the lowest loadings were from the north-easternmost portion of northern
Patagonia near Huinigango. These results suggest that differential site
responses to environmental change gradients along the north-south axis
of the range of Austrocedrus are likely. Additionally, the San Martin chronology,
which is located near the center of both factor distributions, cannot provide
much information about extremes in gradients Austrocedrus responds to.
However, the San Martin chronology exhibits the same regional signal as
most of the other Austrocedrus chronologies. Thus, general patterns of
regional climate forcings produce similar chronologies in most sites.
In conclusion, the San Martin
de los Andes chronology exhibits the regional climate signal demonstrated
in Villalba and Veblen (1997). Principle components analysis suggests that
tree ring response to climate is consistent throughout northern Patagonia.
|
| Discussion |
Our first hypothesis was
that the rock site would be more strongly affected by climate than the
soil site. Drought extremes do appear to impact tree growth more severely
on the rock site, as shown by the number of times that the low growth extremes
from trees on the rock site exceed the severity of low growth on the deep
soil site.
Our second hypothesis was
that tree mortality occurs in dry years. We were able to crossdate 4 trees
all of which died between 1955 and 1957. These dates correspond with
one of the most severe dry periods of the last one hundred years. This
is the same tree mortality event identified by Villalba and Veblen (1998).
We conclude that drought
is a factor in forest dynamics of A. chilensis on rock and deep
soil sites. However on the deep soil sites, fire is also a major
factor as shown by massive regeneration of c. 1900s (and potentially in
the late 18th century) in the more densely wooded sites in northern Patagonia.
|
| References |
Fritts, H.C. 1976. Tree-rings
and climate. Academic Press, London, U.K.
Schulman, E. 1954. Longevity
under adversity in conifers. Science, 119:1396-1399.
Schulman, E. 1956. Dendroclimatic
change in semiarid America. University of Arizona Press, Tucson, Arizona.
Veblen, T.T., and Lorenz,
D.C. 1988. Recent vegetation changes along the forest/steppe ecotone in
northern Patagonia. Annals of the Association of American Geographers,
78: 93-111.
Veblen, T.T., Kitzberger,
T., and Lara, A. 1992. Disturbance and vegetation dynamics along
a transect from rainforest to Patagonian shrublands. Journal of Vegetation
Science, 3: 507-520.
Villalba, R. 1994. Tree-ring
and glacial evidence for the medieval Warm Epoch and the Little Ice Age
in southern South America. Climate Change, 26: 183-197.
Villalba, R., and Veblen,
T.T. 1998. Influences of large-scale climatic variability on episodes of
tree mortality in northern Patagonia. Ecology, 79(8): 2624-2640.
|
