Austral Ecology (2009) 34, 259–271
Environmental controls and patterns of cumulative radial
increment of evergreen tree species in montane, temperate
rainforests of Chiloé Island, southern Chile
CECILIA A. PÉREZ,1* MARTÍN R. CARMONA,2 JUAN CARLOS ARAVENA,3
JOSÉ M. FARIÑA1 AND JUAN J. ARMESTO1,2
1
Center for Advanced Studies in Ecology and Biodiversity, Pontificia Universidad Católica de Chile,
Departamento de Ecología, Alameda 340, Santiago, (Email: [email protected]), 2Instituto de Ecología y
Biodiversidad, Facultad de Ciencias, Universidad de Chile, Santiago, and 3Centro de Estudios
Cuaternarios, Punta Arenas, Chile
Abstract We investigated the local environmental controls on daily fluctuations of cumulative radial increment
and cambial hydration of three dominant, evergreen tree species from montane, Coastal rainforests of Chiloé
Island, Chile (42° 22′ S). During 2 years (1997–1998 and 1998–1999) we recorded hourly cumulative radial
increments using electronic band dendrometers in the long-lived conifer Fitzroya cupressoides (Cupressaceae), the
evergreen broad-leaved Nothofagus nitida (Nothofagaceae), and the narrow-leaved conifer Podocarpus nubigena
(Podocarpaceae). We also measured soil and cambial tissue hydration using capacitance sensors, together with air
and soil temperature and rainfall during the period of the study. In addition, we collected cores of these tree species
to evaluate how dendrometer measurements reflect annual tree ring width. One-year long daily time series of
cumulative radial increments suggests that radial growth of Fitzroya cupressoides was initiated slowly in early spring,
with a maximum in early summer. Multiple regressions showed positive relations between daily precipitation and
radial index (i.e. the difference in cumulative radial increment of two consecutive days) in the three species.
According to path analysis there was a significant direct effect of changes in tree hydration on radial index of the
three focal species. In emergent, pioneer species such as Nothofagus and Fitzroya, radial index was negatively
affected by changes in maximum air temperature and photosynthetically active radiation, probably because of high
evapotranspiration demand on warm sunny days. The shade-tolerant species Podocarpus nubigena was positively
affected by photosynthetically active radiation. Our diel scale findings support the use of tree ring widths for
reconstructing past climate in these southern temperate forests and provide evidence that rainforest trees may be
highly sensitive to future declines in rainfall and temperature increases during summer.
Key words: band dendrometer, environmental variables, Fitzroya cupressoides, Nothofagus nitida, Podocarpus
nubigena, radial increment, tree hydration.
INTRODUCTION
Studies of the patterns of tree radial growth have been
undertaken at broad century to millennial scales to
reconstruct past climates. At these scales, dendrochronological time series based on tree-ring widths provide
evidence of past climatic variability and, therefore,
allow the evaluation of recent climate change in the
context of this variability (Fritts 1976; Schweingruber
1988). In southern South America, several studies
have documented that radial growth of the Andean
tree-line species Nothofagus pumilio correlates positively with air temperatures on a decadal scale (Szeicz
1997; Villalba et al. 1997; Aravena et al. 2002a; Lara
*Corresponding author.
Accepted for publication April 2008.
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
et al. 2005), and this correlation has been used to
make inferences about long-term trends in southern
temperate latitudes. However, in the northern range of
this species, in central Chile, higher temperatures
reduce radial growth, while Nothofagus trees respond
positively to precipitation (Lara et al. 2001, 2005).
These contrasting responses are explained because in
central Chile, lower rainfall rather than air temperature
constrains radial growth (Lara et al. 2001, 2005).
Negative correlations between tree-ring width and
precipitation during the growing season (December–
May) documented in some chronologies from southern temperate rainforests in Chile and Argentina have
been attributed to prolonged snow periods associated
with high precipitation years (Szeicz 1997;
Villalba et al. 1997). However, in the same region,
negative correlations with air temperature and positive
doi:10.1111/j.1442-9993.2009.01927.x
260
C . A . P É R E Z ET AL.
correlations with precipitation during the growing
season have been reported for mid and high elevation
conifer forests (Roig & Boninsegna 1990; Lara &
Villalba 1993; Villalba et al. 1998). These examples
show that radial growth responses to climatic factors
are highly variable on a broad geographical scale and
therefore the interpretation of past climate trends must
take into account the local habitat where the trees are
found (Villalba & Veblen 1997).
That climate affects tree rings has been known for
decades (Douglass 1920). Based on experimental
data, Fritts (1966) graphed the chain of events by
which water stress controls tree ring width; low precipitation and high air temperatures increase stomatal
closure, reducing net photosynthesis and food
reserves, resulting in reduced cambial activity and narrower tree ring widths. Finer resolution studies, which
address radial growth responses on a daily time scale
over a year, can be useful to explore the interplay of the
main physical variables that trigger radial growth,
namely air temperature, precipitation and soil humidity (Mitscherlich 1975; Kozlowzki et al. 1991) and
hence help to clarify long-term patterns in tree-ring
width.
In strongly seasonal climates, such as boreal, temperate and high altitude forests, spring and summer
warming often have positive effects on radial growth
(Norton 1984; Graumlich et al. 1989; Kozlowzki et al.
1991; Jacoby & D’Arrigo 1995; Tardif et al. 2001;
Deslauriers & Morin 2005; Rossi et al. 2007). In contrast, in tropical or subtropical forests radial growth
may be strongly limited by rainless periods and therefore controlled by daily rainfall patterns (Palmer &
Ogden 1983; Worbes 1999; Pélissier & Pascal 2000;
Silva et al. 2002; Baker et al. 2003; Prior et al. 2004;
Vieira et al. 2004; Biondi et al. 2005; Brienen &
Zuidema 2005).
Studies of tree radial growth on short temporal
scales can provide valuable insights on issues such as
environmental triggers of radial growth, differences in
species sensitivity to climatic variables, and potential
limitations to forest productivity and carbon gain in
response to future climate change.
The goal of this study was to investigate the direct
and indirect effects of daily changes in environmental
variables on daily changes in tree hydration and cumulative radial increment of three tree species of contrasting shade-tolerance, in two southern temperate forests
with contrasting nutrient cycling.
METHODS
Study area
The study was conducted in two old-growth, montane
coastal rainforests, located 1 km distant from each
doi:10.1111/j.1442-9993.2009.01927.x
other in Cordillera de Piuchué, Chiloé National Park,
Chiloé Island (42° 22′ S), Chile. Conifer forests occur
on exposed slopes of northeast aspect from 650–
700 m above sea level and are dominated by shadeintolerant conifer Fitzroya cupressoides (Mol.) I.M.
Johnst. (Cupressaceae), which accounts for 56% of
the total basal area. Mixed broadleaved forests occur
from 550–650 m a.s.l. of southwest aspect, and are
co-dominated by two broad-leaved shade-intolerant
evergreen trees, Nothofagus nitida (Phil.) Krasser
(Nothofagaceae) and Drimys winteri J.R. Forst & G.
Forst (Winteraceae), which together account for 72%
of the total basal area. The shade-tolerant conifer
Podocarpus nubigena Lindl. (Podocarpaceae) is also
present in the canopy with about 6% of the basal area.
Conifer forests have a slower nitrogen turnover than
mixed broadleaved forests (Pérez et al. 1998, 2003;
Vann et al. 2002; Joshi et al. 2006).
Mean summer temperature is 10.2°C and mean
winter temperature is 4.2°C (Pérez et al. 1998).
Annual rainfall exceeds 4000 mm, but summer
months (December–March) tend to be drier (20% of
the total annual rainfall) than the rest of the year, due
to a weak Mediterranean-climate influence associated
with the latitudinal position of the southwestern
Pacific Anticyclone.
Cumulative radial increment and
environmental variables
Cumulative radial increment was measured in both
forests using 20 electronic band dendrometers (Agricultural Electronic Corporation,Tucson, AZ, USA). In
January 1997, we installed one dendrometer at breast
height (1.3 m) around the trunks of selected trees.
In the conifer forest, measurements were made
in 10 individuals of Fitzroya cupressoides (Table 1).
Maximum distance among trees was 27 m. In the
mixed broadleaved forest (Table 1), dendrometers
were installed on five individuals of Nothofagus nitida
and five trees of Podocarpus nubigena. Maximum distance among trees was 42 m. We recorded hourly
changes in electronic band displacement (resolution;
1 mV = 3.7 mm of band displacement) for each tree
during 2 years. Band displacements were transformed
to cumulative radial increment (radial change
(mm) = band displacement 2p-1). Cumulative radial
increment measurements with band dendrometers
also includes reversible changes of shrinking and swelling associated to water balance of trees (Kozlowzki
et al. 1991). The band of Invar 36 has a thermal
expansion-contraction factor of 1.26 mm m-1 1°C-1,
which makes the effect of temperature negligible on
the measurements. In order to check the accuracy of
dendrometer signals, the dendrometer bands were
periodically manually displaced and contrasted with
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
C U M U L AT I V E R A D I A L I N C R E M E N T O F C H I L E A N T R E E S
261
Table 1. Average, SD and maximum and minimum values for: tree age, two-decade average of tree ring width (1980–1999),
and tree ring width and cumulative radial increment obtained from band dendrometers during the growing seasons 1997–1998
and 1998–1999
Two-decade average
of tree ring width
(mm year-1)
Tree ring width
(mm year-1)
Cumulative radial increment
(mm year-1)
Tree age
(years)
1980–1999
1997–98
1998–99
1997–98
1998–99
Fitzroya cupressoides
Average
362
n
10
SD
53
Range
257–432
0.27
16
0.05
0.16–0.37
0.17
16
0.08
0.01–0.29
0.22
16
0.12
0.02–0.38
0.091
10
0.075
–0.084–0.192
0.202
10
0.170
0.012–0.518
Nothofagus nitida
Average
91
n
5
SD
30
Range
60–138
0.96
4
0.26
0.65–1.48
1.29
4
0.9
0.24–2.12
1.29
4
1.01
0.17–2.64
0.967
5
0.774
0.169–2.084
0.779†
5
0.697
0.025–1.806
Podocarpus nubigena
Average
114
N
5
SD
39
Range
83–181
0.69
7
0.15
0.51–1.08
0.94
7
0.63
0.37–2.18
1.08
7
0.73
0.26–2.48
0.560
5
0.259
0.302–0.958
0.917
5
0.249
0.631–1.172
Tree species
†
This value may be an underestimate because dendrometer band failure of the fastest growing Nothofagus (Fig. 4). The value
of cumulative radial increment was the cumulative value obtained at the end of each growing season. n, number of trees. Data
are for three tree species in temperate montane forests of Chiloé Island, Chile.
the instantly recorded and stored data. These values
were then deleted. Before installing the dendrometers,
the outer bark was peeled off and the inner most
recently formed one was left undisturbed. Additionally, in March 1997, we implanted either two or three
phytograms (Agricultural Electronic Corporation)
beneath the bark of each tree, which measure the water
storage capacity of cambial tissue (Gensler 1997;
Meinzer et al. 2004). This type of measurement was
necessary as a certain degree of hydration is required
for cell enlargement and division during xylem formation (Kozlowzki et al. 1991). Phytograms (27 sensors
per forest) were installed around the trunk near the
position of dendrometers. Capacitance values are
given in microfarads units, which were transformed to
a hydration percentage of the maximum value
obtained per individual tree during the sampling
period. In order to obtain well replicated records, phytogram data were averaged to one value per tree
species. The following environmental variables were
recorded simultaneously during the period of the
study in each forest: soil temperature (three sensors,
only in the conifer forest) and moisture (three sensors)
under the canopy; air temperature (one sensor) and
photosynthetically active radiation (PAR) outside the
forest canopy. Daily rainfall records were obtained
from a Belfort rain gauge installed in an open area, less
than 1 km from the study sites. Hourly records were
obtained from phytograms and environmental sensors.
Data on cumulative daily radial increment presented
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
here correspond to the periods when band dendrometer records were more continuous over the 2 years of
study. It was not possible to keep permanent records
during the entire period of 2 years because of occasional interruptions due to battery failure during
winter, mice chewing on cables, or sensor failure.
Missing data are recorded as gaps in the time series.
Tree cores
In 2001 several tree cores were collected in both
forests, including the trees with band dendrometers
(Table 1). Core samples at 1.3 m height were obtained
with increment borers from Podocarpus nubigena (n = 7
trees), Nothofagus nitida (n = 4) and Fitzroya cupressoides (n = 16). Cores were mounted, and sanded using
sand paper of increasingly finer grit. Annual rings were
visually cross-dated assigning a calendar year to each
tree ring, according to the latest ring date (Fritts
1976). Tree ring widths were measured to the nearest
0.01 mm with an increment-measuring device and
recorded in a computer. The computer program
Cofecha was used to detect measurement and crossdating errors (Holmes 1983). Cores were used to estimate an average annual tree-ring growth for tree rings
formed in spring–summer of both growing seasons
1997–1998 and 1998–1999. In addition, an average
value of tree ring width was obtained for the period
1980–1999 from all of the sampled cores.
doi:10.1111/j.1442-9993.2009.01927.x
262
C . A . P É R E Z ET AL.
Data transformation and statistical analysis
We averaged hourly fluctuations of radial increment,
tree hydration and environmental variables to obtain
daily values for analyses. To evaluate daily changes in
mean cumulative radial increment, we calculated the
difference between values for two consecutive days.
This difference was defined as ‘radial index’ (Tardif
et al. 2001), which was obtained from the first order
difference to the mean daily cumulative radial increment time series. A PAR index was constructed by
counting the number of hours with PAR above 0 for
each day during the study period.
We used multiple regressions as the simplest model
to express the level of dependence of tree radial index
(dependent variable) on tree hydration and environmental variables (independent variables). Additionally,
we used path analysis to explore causal relationships,
as we had a priori information about the direction of
possible effects of environmental variables on tree
hydration and radial index (Mitchell 1993; Sokal &
Rohlf 1995). In this way we assessed the direct, indirect and total effects of environmental variables on
daily changes in radial index and hydration during two
growing seasons. The statistical package LISREL 8.7
for Windows (Mels 2004) was used to test the hierarchical model composed of exogenous (only outgoing
arrows), intermediate (incoming and outgoing arrows)
and dependent variables (only incoming arrows). It is
worth mentioning that the word ‘effect’ is used here to
express a correlation between variables. The overall
path model fit was assessed by means of the generalized squared multiple correlation coefficient R 2m
(Schumacker & Lomax 1996):
p
Rm2 = 1 − ∏ (1 − Ri2 ) ,
i =1
where Ri2 is the square multiple correlation coefficient
from each individual regression equation within the
model and p is the total number of regression equations (or the total number of dependent variables
within the path model).
Only data for the growing season (December–May
1997–1998 and 1998–1999) were incorporated in the
path analysis, as we were not interested in shrinking
and swelling responses during the winter season. We
did not separate the growing season into shorter
periods, because our final goal was to identify the net
effects of local environmental variables and tree hydration in the formation of annual tree rings. In the
conifer forest, the average radial index and cambial
hydration were obtained from all sampled trees of
Fitzroya. In the mixed broadleaved forest, the average
radial index per species was obtained from the five
sampled trees of both Nothofagus and Podocarpus.
Environmental variables used in the statistical analyses
are those that are not correlated by definition.To make
doi:10.1111/j.1442-9993.2009.01927.x
samples independent from each other, the differences
between the mean value of two consecutive days of tree
hydration and environmental variables (intermediate
variables in the path analysis model) were used. Daily
rainfall and PAR index data (exogenous variables in
the model) were not transformed.
A Gaussian model was fit to the time series of
average tree hydration of Fitzroya (n = 10 trees), the
only species with continuous records for one whole
year. Tree hydration model was then compared with
the time series of cumulative radial increments. The
degree of association in radial index time series of the
three species was assessed by Pearson’s correlation
coefficients.
RESULTS
Patterns of cumulative radial increment
The time series of cumulative daily radial increment
of Fitzroya cupressoides for year one fluctuated around
zero during late autumn (May) and winter (June–
August) (Fig. 1a,b). During this cooler period of the
year, the radial index was noisy, slightly variable,
reflecting short (daily to weekly) cycles of tree shrinking and swelling. At the onset of the austral spring
(September, indicated by the left arrow in the x axis),
most trees entered a period of a positive cumulative
radial increment; although some trees fluctuated more
broadly in radius than others (Fig. 1a,b). During
summer (December–March), three Fitzroya trees continued to show a positive cumulative radial increment
(Fig. 1a), but seven trees showed a pronounced midsummer shrinkage during February 1998 (shown by
dashed arrows in Fig. 1b). Shrinking was followed by a
swelling trend in late summer (March). Pronounced
shrinking in the study period was associated with
peaks of air and soil temperatures and a prolonged
rainless period of 26 days, during an El Niño year
1997–1998 (shown by dashed arrows in Fig. 1d,e).
From mid (April) to late autumn (May) a positive but
weaker trend of cumulative radial increments was
observed (Fig. 1a,b). Average radial index for the 10
Fitzroya trees (Fig. 1c) showed broad fluctuations
during summer (December–March) and mid to late
autumn (April–May), but they declined before the
onset of positive cumulative radial increment
(December).
During late spring of the second growing season,
most Fitzroya cupressoides trees presented a slight positive trend of cumulative radial increment (Fig. 2a,b),
followed by a burst of positive radial increment, after
an early summer rainfall pulse. Cumulative radial
increment continued to be positive during the entire
summer period in five trees (Fig. 2a). In two trees it
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
C U M U L AT I V E R A D I A L I N C R E M E N T O F C H I L E A N T R E E S
263
Fig. 1. Daily time series (1 May 1997 to 31 May 1998) of cumulative radial increment (mm) of Fitzroya cupressoides (n = 10
trees): trees with slight shrinking-swelling cycles during summer (a) and trees with strong shrinking-swelling cycles during
summer (b), average radial index (c), average tree and soil hydration and precipitation (d) and average air and soil temperatures
(e). Double arrows along the X axis indicate the length of the growing season and the shaded area shows the drier summer of
1998. Dashed arrows, indicate episodes of bole shrinking in the summer of 1998.
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
doi:10.1111/j.1442-9993.2009.01927.x
264
C . A . P É R E Z ET AL.
Fig. 2. Daily time series (1 December 1998 to 31 May 1999) of cumulative radial increment (mm) of Fitzroya cupressoides (n = 9
trees): trees with slight shrinking-swelling cycles during summer (a) and trees with strong shrinking-swelling cycles during
summer (b), average radial index (c), average tree and soil hydration and precipitation (d) and average air and soil temperatures
(e).
doi:10.1111/j.1442-9993.2009.01927.x
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
C U M U L AT I V E R A D I A L I N C R E M E N T O F C H I L E A N T R E E S
was interrupted again by marked and repeated shrinkage episodes (Fig. 2b), coinciding with drops in soil
hydration and higher air temperatures (Fig. 2d) during
rainless periods, as in the previous year. Fluctuations
in radial index tended to be higher in summer and
lower in autumn (Fig. 2c), and in the same range of
values as in the previous year.
Radial growth rates estimated from tree cores collected in the conifer forest resemble cumulative radial
increments estimated for our 2-year records using
electronic band dendrometers (Table 1). The fact that
these two independent measurements did not differ
suggests that cumulative radial increment from band
dendrometers reflected tree ring formation, namely
radial growth. Tree ring width of the warm-dry
summer of 1998 is below the average obtained from
1980–1999 (Table 1).
Daily time series of cumulative radial increments of
Nothofagus nitida and Podocarpus nubigena exhibited
irreversible positive trends from late spring to late
summer, continuing into autumn in some trees, during
both growing seasons (Figs 3a,4a). Fluctuations of the
average radial index were large during the austral
summer, and followed similar trends for both species
(R = 0.605, P < 0.001), becoming narrower towards
mid to late autumn (Figs 3b,4b). Radial index time
series of Nothofagus nitida and Fitzroya cupressoides
were less significantly correlated (R = 0.178, P <
0.003). Radial index time series of Podocarpus and
Fitzroya were not correlated (R = 0.044, P = 1). Positive slopes of cumulative radial increment from late
spring to summer in both species in the mixed broadleaved forest were in general associated with peaks
in tree hydration, coinciding with rain pulses and
increased soil hydration (Figs 3c,4c). Rainless episodes during summer were not associated with noticeable shrinkage in these two species.
Radial growth rates estimated from tree ring widths
in the mixed broadleaved forest resemble cumulative
radial increments using electronic band dendrometers
during the first year for Nothofagus nitida and during
the second year for Podocarpus nubigena (Table 1).
Average tree ring width for Nothofagus nitida and Pococarpus nubigena during the warm-dry summer of 1998
was similar and below the average obtained from
1980–1999, respectively (Table 1).
Multiple regressions and path analysis
R2 values from multiple regression analysis showed
that daily rainfall (PP) and changes in soil hydration
during the growing season were positively correlated to
radial index of all tree species in both forests. All
environmental variables and changes in tree hydration
were significantly related to radial index only in
Podocarpus nubigena (P < 0.001; Table 2).
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
265
Path diagrams showed the simultaneous effect of
exogenous environmental variables (rainfall and PAR
hours) on the intermediate variables (changes in soil
moisture and temperature, maximum air temperature
and tree hydration), as well as on changes radial index
for the three focal species during the growing season
(Fig. 5). The goodness of fit of this model was
R 2m = 0.493 for Fitzroya cupressoides, R 2m = 0.598 for
Nothofagus nitida and R 2m = 0.654 for Podocarpus
nubigena.
In the conifer forest, PAR hours had a significant
negative direct effect on radial index (RI) of Fitzroya
cupressoides during the growing season (Fig. 5a), which
was confirmed by multiple regression analysis
(Table 2). There was a significant direct effect of
changes in tree hydration on radial index. Additionally,
we found significant and positive direct effects of both
changes in soil moisture and maximum soil temperature on changes in tree hydration for these individuals.
The first two variables had a significantly indirect
effect on radial index of Fitzroya trees via differences in
tree hydration (Fig. 5a).
In the mixed broadleaved forest, the radial index of
both Nothofagus nitida and Podocarpus nubigena was
directly and positively affected by tree hydration levels
(Fig. 5b,c). Changes in soil moisture had a significant
and positive indirect effect via changes in tree hydration on the radial index of both species during the
growing season (Fig. 5b,c). PAR hours had a significant and positive direct effect on radial index of
Podocarpus nubigena (also present in multiple regression analysis, Table 2). Other significant indirect
effects were associated with maximum air temperature
(DMAAT), via changes in tree hydration, negatively
affecting the radial index of Nothofagus nitida, but
positively affecting Podocarpus nubigena (Fig. 5b,c).
Maximum air temperatures had a significant and negative direct effect on radial index of Podocarpus nubigena
(Fig. 5c). This relationship is confirmed by multiple
regression analysis (Table 2).
DISCUSSION
Measurements of cumulative daily radial increment
during the growing season reflected the typical
cumulative radial growth of trees in all individuals
of Nothofagus nitida and Podocarpus nubigena, during
both years, and only in three out of 10 individuals in
Fitzroya cupressoides, during the dry summer of 1998.
In the other seven Fitzroya trees, remarkable shrinkage and swelling was detected. The difference
among Fitzroya individuals in this forest was not
related to either tree age or canopy position (data not
shown).
Large and significant differences in radial
growth rates reported here among tree species
doi:10.1111/j.1442-9993.2009.01927.x
266
C . A . P É R E Z ET AL.
Fig. 3. Daily time series (1 December 1997 to 21 May 1998) of cumulative radial increment (mm) of Nothofagus nitida (n = 5
trees) and Podocarpus nubigena (n = 5 trees) (a), average radial index (b), average tree and soil hydration and precipitation (c) and
average air temperatures (d). Shaded area shows the drier summer of 1998.
(Nothofagus>Podocarpus>Fitzroya) agree with reported
ecological differences between conifers and
angiosperms (Mitscherlich 1975; Kozlowzki et al.
1991). A higher cumulative radial increment of Nothofagus and Podocarpus can be associated with faster net
N mineralization, accompanied by higher litterfall and
lower litter C/N ratios, reported in the soils of the
doi:10.1111/j.1442-9993.2009.01927.x
mixed broadleaved forest compared with the conifer
forest in the same geographic area of the present study
(Pérez et al. 1998, 2003; Vann et al. 2002; Joshi et al.
2006). In addition, in the study area the average soil
depth is 75 cm in mixed broadleaved forests, while in
conifer forests it is only 40 cm (Joshi et al. 2006). The
shallower soils of conifer forests would make Fitzroya
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
C U M U L AT I V E R A D I A L I N C R E M E N T O F C H I L E A N T R E E S
267
Fig. 4. Daily time series (1 December 1998 to 16 May 1999) of cumulative radial increment (mm) of Nothofagus nitida (n = 5
trees) and Podocarpus nubigena (n = 5 trees) (a), average radial index (b), average tree and soil hydration and precipitation (c), and
average air temperatures (d).
trees more susceptible to dry summer periods, resulting in a decreased tree ring width. Even though trees in
the mixed broadleaved forest are younger than in the
conifer forest, the average age of 362 years for Fitzroya
(Table 1) may represent a similar physiological stand
age than in the mixed broadleaved forest, because
Fitzroya cupressoides may reach even older ages
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
recorded up to 3620 years (Lara & Villalba 1993).
Therefore tree age as a single factor may be less
significant than site conditions in accounting for
differences in radial growth rate.
The temporal patterns of cumulative radial increment of evergreen tree species reported here coincide
with records of radial growth patterns from other
doi:10.1111/j.1442-9993.2009.01927.x
268
C . A . P É R E Z ET AL.
Table 2. Standardized regression coefficients from multiple regressions of radial index of Fitzroya cupressoides, Nothofagus nitida
and Podocarpus nubigena, as the dependent variable, versus differences in tree hydration and local environmental variables, during
two growing seasons (1997–1998, 1998–1999)
Standardised regression coefficients
Independent variable
Fitzroya cupressoides
PP
PAR
DMAST
DMAAT
DTH
DSH
R2
n
F
P (complete model)
0.252*
-0.254*
-0.062
0.036
0.091
0.181*
0.255
263
17.596
<0.001
Nothofagus nitida
Podocarpus nubigena
0.248*
0.089
0.128*
0.170*
0.027
0.067
0.291*
0.184
337
18.729
<0.001
-0.231*
0.357*
0.243*
0.400
337
55.389
<0.001
Asterisks indicate significant regression coefficients (P < 0.05). DMAAT, daily differences in maximum air temperature;
DMAST, daily differences in maximum soil temperature; DSH, daily differences in average soil hydration; DTH, daily
differences in average tree hydration; n, number of days; PAR, photosynthetically active radiation hours index; PP, daily
precipitation.
southern hemisphere temperate forests. In New
Zealand, Nothofagus and Agathis species initiated their
radial growth late in the austral spring (December)
with the highest radial growth rate recorded during the
early to mid austral summer (Palmer & Ogden 1983;
Norton 1984).
According to our band dendrometer records, several
sampled trees of Fitzroya cupressoides showed pronounced shrinking during dry summer episodes,
which were not recorded by band dendrometers
installed on Nothofagus nitida or Podocarpus nubigena.
Shrinking of some Fitzroya trees was similar (in the
second year) or even greater (in the first year) in magnitude than the accumulated annual radial increment,
accounting for up to 600 mm. Daily or even hourly
changes of similar or even twice the magnitude of
annual radial growth have also been reported for
northern temperate conifers such as Picea abies
(Kozlowzki & Winget 1964; Zweifel & Haesler 2001)
and for some tropical trees, with values from 0.1–
2.0 mm (Sheil 2003). This indicates that strong bole
shrinking is a widespread feature of forest trees, and is
presumably related to physiological mechanisms to
overcome temporary water stress. Such shrinking
occurred most notably in Fitzroya trees in response to
warm, rainless episodes, particularly during the austral
summer of 1998. Summer droughts in southern South
America are the consequence of El Niño events
(Veblen et al. 1999; Kitsberger et al. 2001; Montecinos
& Aceituno 2003), with the most severe recorded El
Niño occurring during 1997–1998. In the southern
temperate region, the drier summer of 1998 caused
high mortality of Nothofagus dombeyii trees in the
eastern and drier slopes of the Andes at 40–41°S in
Argentina (Suarez et al. 2004). We report here that
doi:10.1111/j.1442-9993.2009.01927.x
these dry episodes during summer can also affect tree
radial growth in wetter coastal forests west of the
Andes, even in areas receiving more than 4000 mm of
rain per year. In Chiloé, average accumulated radial
increment of Fitzroya trees from band dendrometers
was 0.091 ⫾ 0.075 mm in 1998, in contrast to
0.202 mm ⫾ 0.170 during 1999 (Table 1). For the
conifers Fitzroya and Podocarpus values of average
cumulative radial increment from band dendrometers
for the year 1998 were lower than the average tree
rings width for the previous two decades, whereas this
difference was not present for Nothofagus. This suggests that Fitzroya cupressoides and Podocarpus nubigena
were more sensitive than Nothofagus nitida to the negative effects on radial growth of the dry summer of 1998
in our study sites.
Path diagrams showed similarities and differences
among tree species in the strength of the effect of local
environmental variables on tree hydration and radial
index during the growing season. A common path was
the positive indirect effect of changes in soil hydration
on the radial index of Fitzroya cupressoides Nothofagus
nitida and Podocarpus nubigena via changes in tree
hydration, confirming the fact that cambial hydration
is required for radial growth. However, both species in
the mixed broadleaved forest differed in their response
to other environmental variables. The fact that the
radial index of Podocarpus responded positively to the
direct effect of PAR hours may be because it is a
shade-tolerant species, that develops under the canopy
during early and mid succession (Aravena et al.
2002b). Any increase in the duration of PAR may
positively affect photosynthesis and therefore cumulative radial increment of Podocarpus during the growing season. However the direct negative effect of
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
C U M U L AT I V E R A D I A L I N C R E M E N T O F C H I L E A N T R E E S
269
Fig. 5. Path diagrams describing the dependence of radial index of Fitzroya cupressoides (a), Nothofagus nitida (b) and
Podocarpus nubigena (c), and tree hydration status on local environmental variables during two consecutive growing seasons. The
thickness of the arrows represents the path coefficient. Dashed line = negative path coefficients. Significant paths are indicated
with an asterisk (P < 0.05). DMAAT, daily differences in maximum air temperature; DMAST, daily differences in maximum soil
temperature; DSH, daily differences in average soil hydration; DTH, daily differences in average tree hydration; PAR, photosynthetically active radiation hours index; PP, daily precipitation; RI, radial index; U, residual terms.
maximum air temperature, but positive via changes in
tree hydration suggests that radial index of Podocarpus
will be positive only if changes in maximum air temperature are associated to higher water status of trees
and soil moisture. In the same forest, the negative
indirect effects of maximum air temperatures via tree
hydration on radial index of the emergent and shadeintolerant species Nothofagus nitida can be explained
by the enhancing effect of temperature on stomatal
closure. Stomatal closure during warm-sunny days
would reduce photosynthesis, therefore reducing the
rates of tree radial growth. Negative effects of leaf
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
temperature via vapour pressure deficit and leaf conductance on in situ photosynthesis rates has been
reported for northern temperate deciduous and
conifer tree species (Fritts 1966; Bassow & Bazzaz
1998). In the same way, greater PAR in the conifer
forest may increase evapotranspiration of trees during
prolonged warm, sunny days and hence influence
negatively the radial growth of the canopy emergent
and shade-intolerant Fitzroya. The fact that conifer
forest has a lower index of ‘growing season degree
days’ than mixed broadleaved forest (Joshi et al. 2006)
explains the positive indirect effect of changes in
doi:10.1111/j.1442-9993.2009.01927.x
270
C . A . P É R E Z ET AL.
maximum soil temperature on radial index.The strong
shrinkage detected in Fitzroya during dry summer
periods is explained by the significant effect of changes
in soil hydration via changes in tree hydration.
The Gaussian fit of average tree hydration (n = 10
trees) during an annual cycle showed maximum
cambial hydration in late spring, just before a steady
increase in cumulative radial increment of Fitzroya
trees. This result agrees with previous evidence that
cell turgidity is necessary for cell division in cambial
tissue leading to radial growth (Kozlowzki et al. 1991).
Results from multiple regressions and path analyses,
together with the remarkable shrinking found in most
Fitzroya trees suggest that radial growth of this species
would be slowed by increased evaporative demand
during short rainless episodes and sunny days. Dendrochronological findings showing the negative correlation of Fitzroya cupressoides tree ring width with
air temperature and positive with precipitation on a
regional and decadal scale (Villalba 1990; Villalba
et al. 1990; Lara & Villalba 1993; Neira & Lara 2000)
is the consequence of the short-term response to local
environment documented in the present study.
According to our results, this species appears to be
highly sensitive to dry periods and therefore is a good
indicator of warming and drying trends within the
southern temperate region. If the trend towards
decreasing precipitation in southern South America
suggested by instrumental data (Walther et al. 2002)
continues in the future, we may expect to see a decline
in forest productivity and carbon storage, particularly
in conifer forests dominated by Fitzroya cupressoides,
which is responsible for significant long-term carbon
sequestration due to its extended longevity (Lara &
Villalba 1993).
Our study of patterns of cumulative radial increment using a daily scale of observation confirms that
tree ring widths are an appropriate proxy to infer
past climate trends, and provides evidence of the
physiological mechanisms involved in tree species
responses. Particularly important is the evidence that
the radial growth of rainforest trees may be quite sensitive to short-term declines in precipitation during
summer, and that Fitzroya cupressoides can be negatively affected by extended exposure to photosynthetic active radiation. The contrasting indirect effect
of air temperature in species of the same forest, positive for Podocarpus, but negative for Nothofagus, suggests that the sensitivity of radial activity to climatic
variables is highly dependent on the degree of shade
tolerance of these species. Neither canopy position
nor tree age accounted for the differences in cumulative radial index among Fitzroya individuals, but it
remains an open question if deep rooting may serve
to overcome summer droughts, allowing standing
radial growth in some trees during adverse growing
seasons.
doi:10.1111/j.1442-9993.2009.01927.x
ACKNOWLEDGEMENTS
Funding for this work was provided by A.W. Mellon
Foundation (USA), Cátedra Presidencial en Ciencias
(to J.J. Armesto) and Fondecyt-Fondap 1501-0001
(2001). We are grateful to Corporación Nacional Forestal, Chile for permissions and logistic support. We
dedicate this paper to our friend Luis Cavieres. This is
a contribution to the Cordillera de Piuchué Ecosystem
Studies (CPES) and to the research program of Senda
Darwin Biological Station, Chiloé.
REFERENCES
Aravena J. C., Carmona M. R., Pérez C. A. & Armesto J. J.
(2002b) Changes in tree species richness, stand structure
and soil properties in a successional chronosequence in
northern Chiloé island, Chile. Revista Chilena de Historia
Natural 75, 339–60.
Aravena J. C., Lara A., Wolodarsky A., Villalba R. & Cuq E.
(2002a) Tree ring growth patterns and temperature reconstruction from Nothofagus pumilio (Fagaceae) forests at the
upper tree line of southern Chilean Patagonia. Revista
Chilena de Historia Natural 75, 361–76.
Baker T. R., Burslem D. F. & Swaine M. D. (2003) Associations
between tree growth, soil fertility and water availability at
local and regional scales in Ghanaian tropical rain forest.
J. Trop. Ecol. 19, 109–25.
Bassow S. L. & Bazzaz F. A. (1998) How environmental conditions affect canopy-level photosynthesis in four deciduous
tree species. Ecology 79, 2660–75.
Biondi F., Hartsough P. C. & Galindo I. (2005) Daily weather
and tree growth at the tropical treeline of North America.
Artic Antarctic Alpine Res. 37, 16–24.
Brienen R. J. & Zuidema P. A. (2005) Relating tree growth to
rainfall in Bolivian rainforests: a test for six species using tree
ring analysis. Oecologia 146, 1–12.
Deslauriers A. & Morin H. (2005) Intra-annual tracheid production in balsam fir stems and the effect of meteorological
variables. Trees 19, 402–8.
Douglass A. E. (1920) Evidence of climatic effects in the annual
rings of trees. Ecology 1, 24–32.
Fritts H. C. (1966) Growth-rings of trees: their correlation with
climate. Science 154, 973–9.
Fritts H. C. (1976) Tree Rings and Climate. Academic Press,
London.
Gensler W. (1997) Phytogram/Dendrometry: Method and Interpretation of Diurnal Activity in the Main Stem of Trees. AEC
Manual, Tucson, AZ.
Graumlich L. J., Brubaker L. B. & Grier C. C. (1989) Long term
changes in forest primary productivity: Cascade Mountains,
Washington. Ecology 70, 405–10.
Holmes R. L. (1983) Computer assisted quality control in
tree-ring dating and measurement. Tree Ring Bull. 43,
69–75.
Jacoby G. C. & D’Arrigo R. (1995) Tree ring width and density
evidence of climate and potential forests change in Alaska.
Global Biogeochem. Cycles 9, 227–34.
Joshi A. M., Vann D. R. & Johnson A. H. (2006) Litter quality
and climate decouple nitrogen mineralization and productivity in Chilean temperate rainforests. Soil Sci. Soc. Am. J.
70, 153–62.
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
C U M U L AT I V E R A D I A L I N C R E M E N T O F C H I L E A N T R E E S
Kitsberger T., Swetnam T. W. & Veblen T. T. (2001) Interhemispheric synchrony of forest fires and the El Niñosouthern oscillation. Global Ecol. Biogeogr. 10, 315–26.
Kozlowzki T., Kramer P. & Pallardy S. (1991) The Physiological
Ecology of Woody Plants. Academic Press, New York.
Kozlowzki T. & Winget C. (1964) Diurnal and seasonal variation
in radii of tree stems. Ecology 45, 149–55.
Lara A., Aravena J. C., Villalba R., Wolodarsky-Franke A.,
Luckman B. & Wilson R. (2001) Dendroclimatology of high
elevation Nothofagus pumilio forests at their northern distribution limit in the central Andes of Chile. Can. J. Forest Res.
31, 925–36.
Lara A. & Villalba R. (1993) A 3620- year temperature record
for Fitzroya cupressoides tree rings in southern South
America. Science 260, 1104–6.
Lara A., Villalba R., Wolodarsky A., Aravena J. C., Luckman B.
H. & Cuq E. (2005) Spatial and temporal variation in
Nothofagus pumilio growth at tree line along its latitudinal
range (35°40′–55°S) in the Chilean Andes. J. Biogeogr. 32,
879–93.
Meinzer F. C., James S. A. & Goldstein G. (2004) Dynamic of
transpiration, sap flow and use of stored water in tropical
forest canopy trees. Tree Physiol. 24, 901–9.
Mels G. (2004) Student Edition of LISREL 8.7 for Windows:
Getting Started Guide. Scientific Software International, Inc.,
Lincolnwood.
Mitchell R. J. (1993) Path analysis: pollination. In: Design and
Analysis of Ecological Experiments (eds S. M. Scheiner & J.
Gurevitch) pp. 211–31. Chapman and Hall, Inc., NewYork.
Mitscherlich G. (1975) Wald,Wachstum und Umwelt. J.D. Sauerländers Verlag, Frankfurt.
Montecinos A. & Aceituno P. (2003) Seasonality of the ENSOrelated rainfall variability in central Chile and associated
circulation anomalies. J. Clim. 16, 281–96.
Neira E. & Lara A. (2000) Desarrollo de cronologías de ancho
de anillos para alerce (Fitzroya cupressoides) en Contao y
Mirador (Chile). Revista Chilena de Historia Natural. 73,
693–703.
Norton D. A. (1984) Phenological growth characteristics of
Nothofagus solandri trees at three altitudes in the Craigieburn
Range, New Zealand. N. Z. J. Bot. 22, 413–24.
Palmer J. & Ogden J. (1983) A dendrometer band study of the
seasonal pattern of radial increment in kauri (Agathis
australis). N. Z. J. Bot. 21, 121–6.
Pélissier R. & Pascal J. P. (2000) Two-year tree growth patterns
investigated from monthly girth records using dendrometer
bands in a wet evergreen forest in India. J. Trop. Ecol. 16,
429–46.
Pérez C. A., Armesto J. J., Torrealba C. & Carmona M. R. (2003)
Litterfall dynamics and nutrient use efficiency in two evergreen temperate rain forests of southern Chile. Austral Ecol.
28, 291–600.
Pérez C. A., Hedin L. O. & Armesto J. J. (1998) Nitrogen
mineralization in two unpolluted old-growth forests of contrasting biodiversity and dynamics. Ecosystems 1, 361–73.
Prior L. D., Eamus D. & Bowman D. M. (2004) Tree growth
rates in north Australian savanna habitats: seasonal patterns
and correlations with leaf attributes. Aust. J. Bot. 52, 303–
14.
Roig F. A. & Boninsegna J. A. (1990) Environmental factors
affecting growth of Adesmia communities as determined
from tree rings. Dendrochronologia 8, 39–68.
© 2009 The Authors
Journal compilation © 2009 Ecological Society of Australia
271
Rossi S., Deslauriers A., Anfodillo T. & Carraro V. (2007) Evidence of threshold temperatures for xylogenesis in conifers
at high altitudes. Oecologia 152, 1–12.
Schumacker R. E. & Lomax R. G. (1996) A Beginner’s Guide to
Structural Equation Modeling. Lawrence Erlbaum Associates,
Mahwah, NJ.
Schweingruber F. H. (1988) Tree Rings, Basics and Applications of
Dendrochronoly. D. Redley, Boston.
Sheil D. (2003) Growth assessment in tropical trees: large daily
diameter fluctuations and their concealment by dendrometer bands. Can. J. Forest Res. 33, 2027–35.
Silva R. P., Santos J., Tribuzy E., Chambers J., Nakamura S. &
Higuchi N. (2002) Diameter increment and growth patterns
for individual trees growing in central Amazon, Brazil. Forest
Ecol. Manage. 166, 295–301.
Sokal R. R. & Rohlf F. J. (1995) Biometry, 3rd edn. W H
Freeman & Co., New York.
Suarez M. L., Ghermendi L. & Kitsberger T. (2004) Factors
predisposing tree mortality in Nothofagus-site, climatic sensitivity and growth trends. J. Ecol. 92, 954–66.
Szeicz J. M. (1997) Growth trends and climate sensitivity of trees
in southern Patagonian rainforests of Chile. Can. J. Forest
Res. 27, 1003–14.
Tardif J., Flannigan M. & Bergeron Y. (2001) An analysis of the
daily radial activity of 7 boreal tree species, northwestern
Quebec. Environ. Monit. Assess. 67, 141–60.
Vann D. R., Joshi A., Pérez C. et al. (2002) Distribution and
cycling of C, N, Ca, Mg, K and P in three pristine, oldgrowth forests in the Cordillera de Piuchué, Chile.
Biogeochemistry 60, 25–47.
Veblen T. T., Kitzberger T., Villalba R. & Donnegan J. (1999)
Fire history in northern Patagonia: the role of humans and
climatic variation. Ecol Monogr. 69, 47–67.
Vieira S., Barbosa P., Selhorst D. et al. (2004) Forest structure
and carbon dynamics in Amazonian tropical rainforests.
Oecologia 140, 468–79.
Villalba R. (1990) Climatic fluctuations in northern Patagonia
during the last 1000 years as inferred from tree-ring records.
Quaternary Res. 34, 346–60.
Villalba R., Boninsegan J. A., Veblen T. T., Schmelter A. &
Rubulis S. (1997) Recent trends in tree ring record from
high elevation sites in the Andes from northern Patagonia.
Clim. Change 36, 425–54.
Villalba R., Jacoby G. C., D’Arrigo R., Veblen T.T. & Jones P. D.
(1998) Tree ring based reconstructions of northern Patagonia precipitation since AD 1600. Holocene 8, 659–74.
Villalba R., Leiva J. C., Rubulls S., Suarez, J. & Lenzano L.
(1990) Climate, tree ring, and glacial fluctuations in the Río
Frio Valley, Río Negro, Argentina. Artic Alpine Res. 22, 215–
32.
Villalba R. & Veblen T. T. (1997) Spatial and temporal variation
in Austrocedrus growth along the forest–steppe ecotone in
northern Patagonia. Can. J. Forest Res. 27, 580–97.
Walther G. R., Post E., Convey P. et al. (2002) Ecological
responses to recent climate change. Nature 416, 389–95.
Worbes M. (1999) Annual growth rings, rainfall-dependent
growth and long term growth patterns of tropical trees from
the Caparo Forest Reserve in Venezuela. J. Ecol. 87, 391–
403.
Zweifel R. & Haesler R. (2001) Dynamics of water storage in
mature subalpine Picea abis: temporal and spatial patterns of
change in stem radius. Tree Physiol. 21, 561–9.
doi:10.1111/j.1442-9993.2009.01927.x