Bioclimate and growth history affect beech lifespan in the Italian Alps and Apennines

Bioclimate and growth history affect beech lifespan in theItalian Alps and ApenninesALFREDO D I F I L I PPO * , FRANCO B IOND I † , MAUR IZ IO MAUGER I ‡ , BARTOLOMEO

SCH IRONE * and GIANLUCA PIOVESAN*

*DendrologyLab, Department of Agriculture, Forests, Nature and Energy (DAFNE), Universita degli Studi della Tuscia, I-01100,

Viterbo, Italy, †DendroLab, Department of Geography, MS 154, University of Nevada, Reno, NV 89557, USA, ‡Dipartimento di

Fisica, Universita degli Studi di Milano, I-20133, Milan, Italy

Abstract

When site factors reduce growth rates, tree lifespan tends to increase. This study investigates processes leading to

such inverse relationship in Fagus sylvatica stands distributed along two elevation gradients, with an emphasis on

climatic response, suppression periods, and growth trends. Dendrochronological records from old-growth beech pop-

ulations sampled at different elevations within two different bioclimatic regions (Alps vs. Apennines), were used to

investigate factors that control tree lifespan. Differences between old-growth (12) and nearby managed (15) stands

were used to assess effects of silvicultural practices on maximum age. Logging reduced tree lifespan not only by

removing older trees, but also by reducing the number of years beech individuals spent in the shaded understory.

Tree lifespan and growth rates were affected by climate (spring–summer temperature) and were inversely related to

one another along elevation gradients. The greatest lifespan was observed in old-growth high-mountain populations,

and was related not only to slower growth due to a shorter growing season, but also to multidecadal periods of

growth suppression during the initial development stages in the understory (i.e., slower growth rates at the youngest

cambial ages). Past unfavorable climatic periods (in this case, the Little Ice Age) also helped increase tree lifespan.

Using a linear model, we estimated a reduction in beech lifespan of 23 ± 5 years for each degree of warming. Basal

area increment of trees with the maximum observed lifespan showed an increasing trend over time. Because growth

of old (>300 years) trees has increased in the Alps, while it has recently declined in the Apennines, different biocli-

matic regions can have opposite responses to global climatic change. In the next decades, if warming continues, beech

lifespan could be reduced in the Alps by faster growth and in the Apennines by drought-induced mortality.

Keywords: beech, bioclimate, climate change, conservation biology, elevational gradients, Fagus sylvatica, lifespan, old-growth

forest, tree rings

Received 3 May 2011; revised version received 3 November 2011 and accepted 16 November 2011

Introduction

It was proposed long ago that when site factors reduce

wood growth, individual trees tend to live longer

(Schulman, 1954). A typical example of this phenome-

non is found in Thuja occidentalis, which commonly lives

up to a few centuries, but can exceed 1500 years of age

in shrubby form on rocky cliffs (Larson, 2001). While

the apparent paradox of slower metabolic rates provid-

ing greater longevity occurs frequently in plant species

(Loehle, 1988; Bigler & Veblen, 2009; Issartel & Coiffard,

2011), it may also appear in animal populations,

although not as a general rule (Karl & Fischer, 2009).

Even if tree growth may potentially continue ad infi-

nitum (Flanary & Kletetschka, 2005; Penuelas & Munne-

Bosch, 2010), some common rule(s) governing mortality

rates in ecological communities have been proposed for

organisms as diverse as plants and animals, given their

intrinsic mechanisms constrained by body size and

temperature dependence of metabolism (McCoy &

Gillooly, 2008). The maximum lifespan of plants could

be explained by the Metabolic Theory of Ecology (MTE;

Price et al., 2010), which incorporates the effect of tem-

perature on plant metabolism. Simply stated, a higher

metabolic rate can potentially enhance aging/senes-

cence (i.e., intrinsic mortality causes) at the cellular

level by increasing oxidative stress (Issartel & Coiffard,

2011) or telomere length reduction (Watson & Riha,

2011). These effects, still being debated (Marba et al.,

2007; McCoy & Gillooly, 2008), are of fundamental

importance in light of potential climate warming. How-

ever, in forest ecology the processes that control stand

structure and canopy composition are not necessarily

considered to fall within the realm of MTE (e.g.,

Muller-Landau et al., 2006).Correspondence: Gianluca Piovesan, tel. + 39 761 357 387,

fax + 39 761 357 250, e-mail: [email protected]

960 © 2011 Blackwell Publishing Ltd

Global Change Biology (2012) 18, 960–972, doi: 10.1111/j.1365-2486.2011.02617.x

At the ecosystem level tree mortality is normally

linked to competitive and abiotic/biotic disturbance

processes. These so-called extrinsic mortality causes

(Issartel & Coiffard, 2011) include wildfires, wind-

throws, droughts, insect outbreaks, and pest infesta-

tions. In other words, most tree species have lifespans

that are directly connected with local environmental

conditions (Loehle, 1988, 2000; Penuelas & Munne-

Bosch, 2010), and no universal growth–mortality rela-

tionship has yet been identified (Wunder et al., 2008;

Richardson et al., 2009). An important issue would be

to understand if the intraspecific relationship between

growth rate and lifespan is due to genotype, pheno-

typic plasticity, or some combination of the two (Black

et al., 2008). This would explain if some trees innately

grow slowly and reach great ages, or if it is the environ-

ment that, through heavy competition, poor microsite,

or passage through climatically unfavorable eras, can

enforce suppression that induces slow growth and

extends lifespan.

In temperate climates, tree species with wide geo-

graphical and topographical distribution show reduced

growth rates along altitudinal gradients (e.g., Bolstad

et al., 2001; Dang et al., 2007). Several phenological

changes are clearly linked to elevation, including nee-

dle longevity in conifer species (Reich et al., 1996), but

tree age has shown contrasting relationships with ele-

vation. In fact, late successional species often include

older individuals at higher elevations (Splechtna et al.,

2000; Piovesan et al., 2005; Di Filippo et al., 2007; Wang

et al., 2009), but the occurrence of severe disturbance

(e.g., fire) at the landscape scale can be a confounding

factor, as prolonged periods of catastrophic disturbance

reduce the proportion of slow-growing, long-lived trees

to favor faster-growing, shorter-lived individuals

(Black et al., 2008; see also the case of Douglas fir

reported in Poage et al., 2009). Yet other studies found

no longevity/temperature relationship, using either

mean (Peterson & Peterson, 2001) or maximum tree age

(Bigler & Veblen, 2009). As a consequence, the role of

altitude or latitude (i.e., thermal gradients) in mediating

the geographic variation of lifespan for a tree species

requires further analysis, also for better predicting cli-

mate change impacts on forest ecosystems.

Growth suppression is one of the main factors

enhancing tree longevity in closed forests (Black et al.,

2008). While tree life history in old-growth stands is

often characterized by multiple suppression and release

cycles (Peters, 1997), in managed forests the silvicul-

tural control of intertree competition can shorten the

potential lifespan of large trees. As an example, Euro-

pean beech can tolerate very long periods of growth

suppression (e.g., Piovesan et al., 2010), but it is

unknown until what age suppression can have a signi-

ficant effect on tree lifespan (Black et al., 2008). Given

the need to understand how tree lifespan changes

within old-growth and managed stands in relation to

bioclimate and growth history, we analyzed dendro-

chronological records of a late-successional species,

European beech (Fagus sylvatica L.) within two biocli-

matic regions (Alps vs. Apennines) with distinct pre-

cipitation regimes (Oceanic vs. Mediterranean). Beech

stands in Italy have shown a well-defined bioclimatic

organization of dendroclimatic signals and the exis-

tence of altitudinal gradients in tree growth (Piovesan

et al., 2005, 2011; Di Filippo et al., 2007). In this study,

both old-growth and managed beech stands were sam-

pled along an elevation gradient to test the main

hypothesis that bioclimatic zone, climatic change, and

growth history have separable influences on tree life-

span.

Materials and methods

We used 12 chronologies from old-growth forests in our Ital-

ian network of European beech (F. sylvatica L.) tree-ring sites

(Piovesan et al., 2005; Di Filippo et al., 2007). Four of these sites

came from the eastern Alps (LAT, TIM, GRA, CLE in Piovesan

et al., 2010) and eight from the central Apennines (VCH, COP,

REG, CIM in Piovesan et al., 2010 and ORI, VEN, TER, VCL in

Piovesan et al., 2005). Old-growth forests are defined here as

stands where dominant trees die naturally, allowing new

cohorts to occupy the gaps and enter the canopy (Piovesan

et al., 2010). According to this process-based definition, we

considered as old-growth forests only those entering the last

stages of stand development (i.e., the demographic transition

and multicohort stages of Frelich, 2002). These natural ecosys-

tems are currently very rare throughout the European land-

scape (Ziaco et al., 2012a). As in old-growth forests (according

to our definition) tree death is occurring after natural small-

scale (gap) disturbance, the maximum age of dominant trees

provides insights on beech longevity as a measure of mini-

mum lifespan at that site. Maximum age of dominant trees

served as a surrogate of tree realized longevity, because the

low wood durability of F. sylvatica [ECS (European Committee

for Standardization), 1995], shown by frequent wood rot and

fast decay rates, makes it impossible to count tree-rings in

declining and dead trees (i.e., the actual measure of tree real-

ized lifespan). Old-growth forests were divided into primary,

that is, with no signs of past harvesting, and secondary, that

is, with structures and dynamics still influenced by past

human activities (Frelich, 2002). The high-elevation beech

forest of high Valle Cervara (VCH) was the only primary

old-growth forest in our network (Piovesan et al., 2008), while

11 other stands could be classified as secondary old-growth

forests.

Both the Alps and the Apennines beech populations were

arranged in three altitudinal bioclimatic zones (Table 1),

covering different elevation ranges: low-elevation, mountain

and high mountain (Piovesan et al., 2005; Di Filippo et al.,

2007). While in the eastern Alps beech does not extend beyond

© 2011 Blackwell Publishing Ltd, Global Change Biology, 18, 960–972

BIOCLIMATE, TREE GROWTH, AND LIFESPAN 961

elevations up to 1500 m a.s.l., in the central Apennines it

reaches treeline (~1900 m a.s.l.; Table 1). In the Alps, variabil-

ity of beech radial growth with elevation is influenced more

by air temperature than precipitation (Di Filippo et al., 2007),

while in the Apennines summertime drought plays a domi-

nant role (Piovesan et al., 2005). In addition, fewer canopy spe-

cies exist in beech forests of the Apennines compared with

those of the Alps, where conifers, such as larch (Larix decidua

Miller) and spruce (Picea abies Karst.), are present.

Our field collections were made by selecting at least 20

among the largest dominant or codominant trees located

below timberline, where lifespan can be reduced by extreme

pedoclimatic stress (Biondi, 2001). One wood increment core

was then collected from the lower bole at breast height (along

the contour line whenever possible). All wood samples were

prepared for dendrochronological analysis using standard

methods (Stokes & Smiley, 1996; Grissino-Mayer, 2001). We

used only crossdated ages in the analysis, and this implied a

focus on living, healthy trees, as the frequent occurrence of

wood rot in declining and dead beech trees does not allow

crossdating and, above all, limits the detection of their entire

lifespan. For each sampled beech population, lifespan was

represented by the age of the oldest tree (Agemax) and by the

mean age of the five oldest trees (Age5; Piovesan et al., 2010).

While the maximum observed age (Agemax) is normally linked

to particular microtopographical conditions, Age5 should

more precisely represent average site conditions.

We used ring width to quantify diameter increment, and

ring areas to quantify basal area increment (BAI; Piovesan

et al., 2008). BAI chronologies, obtained by averaging individ-

ual BAI series, were smoothed by cubic splines with a 50%

frequency-response cutoff at 50-year periods to represent

long-term productivity trends (Bunn & Biondi, 2010).

Maximum growth rate, calculated as the 99th percentile of the

distribution of growth rates (either ring width or BAI) avail-

able from that tree population, was used as a proxy of maxi-

mum metabolic rate and, thus, site productivity. While

maximum growth rate for a site should be mainly controlled

by site fertility, the average growth rate of a tree is considered

to represent different biological and ecological factors acting

on an individual plant, such as suppression length (SUP).

The relationship between lifespan and either site elevation,

vegetative season (May–September) mean temperature, sup-

pression period, average or maximum growth rate, was

investigated using correlations and regressions, with or with-

out prior log-transformation of the predictand. To assess

effects of silvicultural practices on tree lifespan, relationships

identified for old-growth forests were compared with those

for managed beech forests (six from the eastern Alps and

nine from the central Apennines) that had not been logged

during the last several decades. To quantify suppression, we

used three different approaches. We first considered mean

ring width during the first 70 years of life (juvenile growth

in the shaded understory) for the five oldest trees at a site,

similar to what was done by Bigler & Veblen (2009). Then,

given that the longest suppression period observed in beech

does not exceed two centuries (Peters, 1997), we computed

SUP during the first 200 years of a tree’s life as the interval

with 4 or more years with ring width less than

0.5 mm year�1 and no sequences of 3 or more years with

growth above the 0.5 mm year�1 threshold (Canham, 1990).

This last approach allowed us to ‘correct’ the age of the five

oldest trees in each stand (Age5) by subtracting the number

of years spent in the understory (‘corrected-Age5). As our

dataset was made of populations with different degree of

old-growthness, excluding the effect of suppression on life-

span helped to better assess the causal relationships between

lifespan and the remaining explanatory variables. Finally, the

Table 1 Main features of old-growth beech populations analyzed within the two bioclimatic regions

Region

Altitudinal

zone Sites

Elevation

range

(m a.s.l.)

Oldest trees All trees

Age DBH

RW BAI

DBH*

(cm)

Oldest5*

(years)

Oldest

tree

(years)

Oldest5*

(cm)

Eastern Alps High mountain 2/3 1200–1500 340 ± 9a 379 61 ± 3a 85 81 54 ± 2

Mountain 2/4 800–1200 253 ± 14b 318 53 ± 3a 55 53 52 ± 2

Low elevation 0/3 200–800 – – – – – –

Central

Apennines

High mountain 3/6 1500–1900 395 ± 17a 503 60 ± 4a 100 92 59 ± 2

Mountain 3/8 900–1500 267 ± 29b 434 76 ± 6a 125 119 68 ± 2

Low elevation 2/3 400–900 114 ± 6c 136 64 ± 6a 114 112 59 ± 1

Sites indicate number of old-growth forests/number of sampled forests per altitudinal zone; DBH indicates diameter at breast

height; superscript letters refer to significant differences (P < 0.05) according to t-tests performed within each column (separately

for the two regions); Oldest5 indicates five oldest trees in the stand; RW indicates available ring width series; BAI indicates avail-

able basal area increment series.* indicates mean ± standard error.

© 2011 Blackwell Publishing Ltd, Global Change Biology, 18, 960–972

962 A. DI FILIPPO e t a l .

oldest trees (age � 300 years) in the entire dataset (11 in the

Alps and 42 in the Apennines) were examined using 50-year

age intervals of each tree’s life up to 300 years, and growth

thresholds from 0.1 to 1 mm year �1 at 0.1 mm year �1

intervals. The number of years (SUP) when growth rate fell

below a threshold was correlated with tree age, and boot-

strapped confidence levels were used to select which pair of

age interval and growth threshold gave the most significant

results.

To evaluate growth history patterns as well as age effects

on beech productivity, we selected two forests (Lateis and

VCH) that included the oldest trees in, respectively, the Alps

and the Apennines. Mean annual BAI, which can represent

productivity (Moore et al., 2006), was computed for nonover-

lapping 100-year intervals of tree age and by averaging across

all trees. Finally, we regressed BAI, calculated for a recent

common period (1991–2000), against age and diameter at

breast height (DBH) for 44 (Lateis) and 53 (VCH) canopy trees

to determine how growth rate was related to life period and

stem size in individual trees.

Instrumental records of temperature and precipitation for

each site were obtained by the anomaly method (Mitchell &

Jones, 2005). Temperature climatologies were obtained from

Brunetti et al. (2009a). Temperature was also expressed in a

format comparable with the MTE formulas (Munch & Salinas,

2009), which use 1/(kT), with k = Boltzmann’s constant and

T = temperature in K. Because of their high spatial heteroge-

neity, precipitation climatologies were derived from station

data if available for the chronology sites, otherwise they were

obtained according to Brunetti et al. (2009b). Anomalies were

computed for each site as a distance-weighted average of

records from neighboring stations presented by Brunetti et al.

(2006) and available from the Italian Air Force (Simolo et al.,

2010). Climatic data refer to the vegetative season, here con-

sidered as the period covering the months from May to Sep-

tember. Mean air temperature for these 5 months, together

with number of years showing growth suppression, were used

in a multiple regression equation to predict beech maximum

lifespan (Age5) in the 12 old-growth sites. Most numerical

analyses were conducted in the R software environment (R

Development Core Team, 2010).

Results

At the lower elevations, beech showed a lifespan

between 100 and 200 years, while in the high Apen-

nines it reached 500 years (Fig. 1a). Although lifespan

increased with elevation, stem size (DBH) of the same

trees did not show any clear trend (Table 1). In both the

Alps and the Apennines, the oldest populations were

found in the high-mountain zones (Fig. 1a and

Table 1). At the same time, radial growth rates

decreased at higher elevations (Fig. 1b), so that maxi-

mum lifespan (defined using Age5) was inversely

related with mean tree-ring width as elevation

increased. Relationships identified using Age5 were

consistent with those obtained using Agemax (Fig. S1),

as these two parameters were very highly correlated

(r = 0.995). Within the same elevation zone, old-growth

forests generally had greater maximum age than man-

aged ones, a pattern more evident in the Alps than in

the Apennines, where some managed stands showed

Agemax values comparable with those of old-growth

stands growing at a similar elevation (Fig. S1).

We found that maximum tree lifespan was inversely

related with mean ring width during the first few dec-

ades of the tree’s life (70 years in this analysis; Fig. 2a).

There was also a direct correlation between length of

Fig. 1 Relationships identified in our tree-ring network from

old-growth beech (Fagus sylvatica L.) forests in the eastern Alps

(solid symbols) and the central Apennines (empty symbols). Tri-

angles, high-mountain sites; circles, mountain sites; squares,

low-elevation sites (see Table 1 for elevation ranges); gray-

shaded circles, calcareous substrate. A straight line was used to

represent the equation shown in the explanatory box; the vari-

ables within parentheses refer to: R2adj, adjusted R2; F, value of

the F statistic; P, significance of the F-test. (a) Mean age of the

five oldest trees (Age5) vs. site elevation for the 12 old-growth

sites; regression developed for central Apennines only. (b) Mean

age of the five oldest trees (Age5) vs. mean ring width of the five

oldest trees for the 12 old-growth sites; values represented using

log-transformed scales; the number below each symbol is site

elevation in metres a.s.l.; regression developed for all sites.

© 2011 Blackwell Publishing Ltd, Global Change Biology, 18, 960–972

BIOCLIMATE, TREE GROWTH, AND LIFESPAN 963

suppression period and lifespan (Fig. 2b). When the

analysis of suppression periods was not limited to the

first 200 years, the effect of suppression on tree age was

best described by the combination of a 300-year win-

dow and a 0.5 mm year�1 threshold (Fig. 2c). Such

growth rate threshold was consistent with the one

empirically chosen by Canham (1990). It is notable that

the oldest trees (i.e., those exceeding 300 years of age)

experienced low growth rates (i.e., number of years

with ring width less than 0.5 mm year�1) even after

they had entered the canopy and up to 300 years of

age. A number of trees older than 400 years, which

came from two sites in the Apennines (VCH and Scan-

give), experienced the longest periods of growth sup-

pression (>100–150 years).

In the Alps and the Apennines, late spring–summer

mean temperature was linked to both lifespan (Fig. 3a)

and maximum growth rates (Fig. 3b). This relationship

was characterized by a greater slope (about 1; Fig. 3a)

than what was predicted by the MTE if tree lifespan

was controlled by temperature regulation of photosyn-

thesis (about 0.3; Allen et al., 2005). Removing years of

growth suppression during the first 200 years of a

tree’s life, and restricting the analysis to sites on calcar-

eous substrates, generated a lower slope (0.75; Fig. S2a).

The extreme longevity (500 years in VCH) of high

elevation beech in the Apennines was not linked to dif-

ferences in growing season temperature compared with

the Alps (Fig. 3a). Site quality indicators, such as maxi-

mum growth rate – either using maximum BAI

(Fig. 3b) or maximum ring width (Fig. S2b), were

related to temperature with signs opposite to that of

beech lifespan. Precipitation showed no relationship

with lifespan/growth rate gradients (data not shown).

The combination of mean May–September tempera-

ture (mM–S_t) and SUP explained about 90%

(R2adj = 0.88, F2,9 = 43, P = 0.00002) of the variability in

maximum tree lifespan (Age5) for the 12 old-growth

forests, as shown by the following equation:

ln ðAge5Þ ¼ 7:161þ 0:007 SUP� 0:118 mM-S t:

Standard error (0.387) (0.003) (0.023)

P-value (<0.001) (0.03) (<0.001)

Comparable results (e.g., R2adj = 0.90, F2,9 = 51, P

= 0.00001) were obtained without the logarithmic trans-

formation of the predictand (see also Munch & Salinas,

2009):

Age5 ¼ 563:5þ 2:5 SUP� 22:5 mM-S t:

Standard error (87.1) (0.6) (5.1)

P-value (<0.001) (<0.01) (<0.01)

Fig. 2 Relationships between growth suppression and lifespan

in the 12 old-growth beech sites of the eastern Alps (solid sym-

bols) and the central Apennines (empty symbols). Triangles,

high-mountain sites; circles, mountain sites; squares, low-eleva-

tion sites (see Table 1 for elevation ranges). For the meaning of

R2adj, F, and P, see caption to Fig. 1. (a) Mean age of the five old-

est trees (Age5) vs. mean ring width of the five oldest trees in

the first 70 years; values represented using log-transformed

scales. (b) Mean age of the five oldest trees (Age5) vs. the mean

number of suppression years in the five oldest trees; values rep-

resented using log-transformed scales. (c) Age of the trees that

are older than 300 years vs. the number of years with sup-

pressed growth (<0.5 mm year�1) in the first 300 years of the

tree’s life.

© 2011 Blackwell Publishing Ltd, Global Change Biology, 18, 960–972

964 A. DI FILIPPO e t a l .

Based on these results, using the last linear model it

was possible to predict a reduction in beech lifespan of

23 ± 5 years for each degree of vegetative season

warming.

Mean annual BAI at Lateis (Alps) and VCH (Apen-

nines) increased over time for all cambial ages (Fig. 4).

For any 100-year age class, the oldest trees had the low-

est growth rates, and trees in the Apennines grew faster

than trees in the Alps (Fig. 4). DBH of trees older than

200 years was relatively uniform (data not shown).

Using data from the two most developed old-growth

forests (Table 2), tree productivity (expressed as BAI

calculated for the common period 1991–2000) was

directly related to stem size (DBH) and inversely

related to age. The highest BAI was found in very large

trees (DBH of 70–100 cm) in both the Apennines and

Alps.

Old (>300 years) beech trees growing at high eleva-

tion in old-growth forests of the Alps and Apennines,

despite having similar age and DBH (box 2 in Fig. 5) at

the time of sampling, presented diverging productivity

trends in the last decades (Fig. 5). After about 250 years

of overlapping growth rates (from the mid-1600s to the

late 1800s, coincident with the Little Ice Age), trees in

the Apennines grew faster than in the Alps for about a

century, until the late 1900s. In the mid-20th century,

beech growth in the Alps rapidly increased, becoming

considerably larger than that in the Apennines during

recent times (Fig. 5).

Discussion

Linking lifespan to growth rate and metabolism

Maximum tree lifespan was inversely proportional to

stem growth rate (a proxy for tree metabolism), in

agreement with other studies (e.g., Johnson & Abrams,

Fig. 3 Relationships between vegetative season (May–Septem-

ber) mean temperature and either tree lifespan or site produc-

tivity in the 12 old-growth beech sites. Straight lines (solid for

all sites; dotted for calcareous sites only) were used to represent

the simple linear regression equations shown in the explanatory

box. For the meaning of R2adj, F, and P, see caption to Fig. 1. (a)

Mean age of the five oldest trees (Age5) vs. temperature; the

predictand values were log-transformed and the predictor was

expressed in °C (upper axis) or by 1/(kT) with k being Boltz-

mann’s constant and T measured in K (lower axis). (b) Maxi-

mum growth rate (Max BAI = 99th percentile of the distribution

of basal area increment) vs. temperature; the predictand values

were log-transformed and the predictor was expressed in °C

(upper axis) or by 1/(kT) with k being Boltzmann’s constant

and T measured in K (lower axis).

Fig. 4 Growth dynamics (annual basal area increment

smoothed using a loess function) by nonoverlapping 100-year

age classes of dominant and codominant beech trees in the two

sites that include the oldest trees in the eastern Alps (Lateis)

and the central Apennines (high Valle Cervara). LAT, Lateis

(solid lines) and VCH, high Valle Cervara (circles). The number

of trees (N) included in each age class is also shown in the

explanatory boxes. Only the full replicated portion of each age

class curve was represented.

© 2011 Blackwell Publishing Ltd, Global Change Biology, 18, 960–972

BIOCLIMATE, TREE GROWTH, AND LIFESPAN 965

2009). This finding also corroborates the longevity para-

dox in forest trees (Larson, 2001) and the general idea

that slowing down metabolism can extend lifespan.

Intrinsic mortality factors, such as cellular senescence,

are probably linked with extrinsic ones, such as size-

mediated physiological efficiency or predisposition to

disturbance. In fact, recent productivity of dominant

trees in the Alps and Apennines was directly related to

their size and inversely to their age (Table 2). A slower

metabolism can increase longevity per se by limiting

oxidative stress or the reduction of telomere length,

or instead by producing smaller dimension of the

organism, thus reducing maintenance and repair costs,

while maximizing durability and strength and mini-

mizing water transport limitations (e.g., Mencuccini

et al., 2005; Watson & Riha, 2011). At the same time,

slower metabolism could extend lifespan as smaller

trees are less affected by extrinsic mortality when prob-

ability of death depends on dimensional thresholds.

The MTE predicts that plant mortality has a much

weaker relationship with temperature than animal mor-

tality, so that a 30 K decrease in temperature would

cause an effect four times greater in animals than in

plants (a 16-fold mortality decrease vs. a 4-fold one;

Table 2 Results of the multivariate linear regression model predicting mean annual basal area increment (BAI) in 1991–2000 from

diameter at breast height (DBH) and age at the time of sampling of dominant and codominant trees at Lateis in the Alps and Valle

Cervara high in the Apennines

N Intercept DBH Age Adjusted R2 F P-value

Lateis 44 �2.71ns 0.80*** �0.08*** 0.53 26 6.2E-08

Valle Cervara High 53 �0.40ns 0.40*** �0.03** 0.50 28 6.6E-09

N, number of trees used in the analysis; ns, not significant; F, F-statistic value; P-value, probability associated to the F-statistic.*P < 0.05; **P < 0.01; ***P < 0.001.

Fig. 5 Basal area increment (BAI, smoothed using a cubic smoothing spline with a 50-year period) of old (>300 years) beech trees sam-

pled in the central Apennines and the eastern Alps (shaded area: mean ± standard error). Horizontal bars at the bottom of the figure

indicate the period with full replication (gray) or covered by at least three trees (white) for each study region. Box 1: BAI splines vs.

high-mountain central Apennines summer temperature 50-year spline (period: 1950–2010). Box 2: boxplots of age and diameter at

breast height by region (t-test in both cases had P > 0.05).

© 2011 Blackwell Publishing Ltd, Global Change Biology, 18, 960–972

966 A. DI FILIPPO e t a l .

Allen et al., 2005). According to Munch & Salinas

(2009), the MTE can be expressed as a function where

log-lifespan is the predictand while log-mass and

inverse temperature are the predictors. This function

can be valid even without including organism mass,

and the MTE would be supported if regression slopes

between log-lifespan and 1/(kT) ranged between 0.2

and 1.2 (Munch & Salinas, 2009). Our data, which show

slopes of 0.8–1.1, for the first time indicate that the tem-

perature link with lifespan, if viewed from the MTE

perspective, is consistent with a metabolic rate control

(Gillooly et al., 2001). In fact, the rate of living (ROL)

theory of aging proposes that longevity is inversely

proportional to metabolic rate, and such theory has

been recently proposed to explain the different life-

spans of stems vs. leaves (Issartel & Coiffard, 2011).

The role of body size in determining tree lifespan still

needs to be reconciled with the MTE, as the oldest trees

often live, with relatively smaller sizes, at the ecological

limits of a species distribution (Penuelas & Munne-Bos-

ch, 2010). In the oldest beech forests, a typical example

of temperate late successional ecosystems, there is

essentially no connection between size and age for

dominant trees (e.g., Piovesan et al., 2010), as the same

DBH can be attained at various ages depending on

individual growth history (see Fig. 4). In fact, it has

been pointed out that structural development processes

in primary old-growth forests do not follow the MTE

predictions (Wang et al., 2009), especially for what con-

cerns the frequency and growth history of larger trees

(Russo et al., 2007; Enquist et al., 2009). Nonetheless,

when sites escape severe exogenous disturbance (fires,

windstorms, etc.), the lifespan/metabolism rate rela-

tionship holds, as in the case of biomass accumulation

during secondary succession (Anderson et al., 2006).

While we found that the MTE is useful to describe tem-

perature-dependent growth processes in old-growth

beech forests, future genetic analyses could help with

understanding if there is a genetically coded ‘slower

auxology’ in the oldest trees (Black et al., 2008).

Linking lifespan to bioclimate and tree suppression

Tree-ring networks are ideal databases to examine how

a forest species responds to climate at different eleva-

tions, especially when populations are not impacted by

recurring severe disturbance events. In our beech net-

work, lifespan increased from roughly 100–150 to

500 years as temperature decreased with elevation

from 21 to 11 °C. When temperature of the vegetative

season was combined with suppression in a multivari-

ate statistical analysis, it was possible to estimate a

reduction in beech lifespan of 23 ± 5 years for each

degree of warming. The 10° reduction in air tempera-

ture with elevation was also associated with a

100 cm2 year�1 decrease in maximum BAI (i.e., produc-

tivity). A similar reduction in net primary productivity

(Bolstad et al., 2001) and forest turnover rates (Stephen-

son & van Mantgem, 2005) with increasing elevation

(hence decreasing temperature) were previously

reported for temperate forests of North America.

The 200-year maximum lifespan often associated

with deciduous broadleaf species (Loehle, 1988) was

observed only at the lowest elevations of our network.

Since wood durability of F. sylvatica is among the low-

est (ECS, 1995), causing fast decay rates and susceptibil-

ity to windstorm damage, other age-boosting life

history traits, such as tolerance of long suppression

cycles (e.g., Bigler & Veblen, 2009), and the delayed

sexual maturity that results (another trait associated to

tree longevity; Loehle, 1988), must be able to compen-

sate for this deficiency.

While our main findings suggest that geographical

variation in tree growth and lifespan has a common

physiological link with temperature, the oldest trees

were found in high-mountain forests of the Apennines,

in a thermal regime comparable with the high-moun-

tain Alpine ones. Other factors may thus account for

the difference (more than 100 years) in the realized

maximum age between these two bioclimatic regions.

In general, climate differs consistently in the two biocli-

matic zones: high-mountain stands in the Alps experi-

ence colder winters and no water stress during the

growing season, while in the Apennines they are

affected by summer water stress and spring coldness

(Fig. 6). Summer drought in the Apennines may

require beech to develop a more extensive and devel-

oped root structure (Zianis & Mencuccini, 2005), which

in turn decreases the likelihood of tree death by wind-

throw. In the Alps, old-growth forests occur in smaller

patches and with lower specific indicators (such as

deadwood) than in the Apennines, mostly because of a

greater human impact and closer proximity with man-

aged forests or pastures (Piovesan et al., 2010). Greater

anthropic use (i.e., repeated logging) in the Alps may

therefore have reduced tree lifespan by removing older

trees or the long suppression periods, which are found

in primary forests (Parish & Antos, 2006; Piovesan

et al., 2010). According to our results, logging can effec-

tively reduce maximum tree lifespan by reducing com-

petitive interactions and increasing growth rates of

surviving trees.

As the oldest trees came from high-elevation sites on

calcareous substrates (gray-shaded points in Fig. 1) in

both the Alps and the Apennines, long SUP periods

were not only caused by competitive interactions but

may be also influenced by low site fertility. Calcareous

substrates (i.e., low fertility sites) also supported the

© 2011 Blackwell Publishing Ltd, Global Change Biology, 18, 960–972

BIOCLIMATE, TREE GROWTH, AND LIFESPAN 967

oldest conifer populations of the western United States

(Schulman, 1954). In Picea mariana lifespan and the time

needed to reach break-up stage (i.e., secondary old-

growth status) were negatively linked with site quality,

but positively with tree suppression (Robichaud &

Methven, 1993), demonstrating that greater tree age can

be allowed by a reduction in resource availability.

At the lowest elevations in central Italy, beech can

survive the Mediterranean climate only on deep and

fertile (i.e., volcanic) soils. Together with warmer tem-

peratures, these soils increase beech growth rates,

thereby reducing tree lifespan. Low-elevation sites are

also more easily accessible to people, causing more

intense forest management throughout these forests’

history (e.g., Castagneri et al., 2010). Using values calcu-

lated without the suppression period (‘Corrected Age5’

in Fig. S2) allowed the identification of lifespan–tem-

perature relationships across the entire elevational gra-

dient. Furthermore, mean growth of suppressed

individuals in worldwide beech forests (data from

Peters, 1997) showed a correlation with site tempera-

ture (Fig. 7). A change in tree lifespan potentially

driven by climatic variability could therefore begin dur-

ing the suppression period.

When conservation targets favor natural forest

dynamics, it is advisable, based on our results, to avoid

repeated thinning, and logging in general, to allow

individual trees to reach greater ages. In general,

growth and structural development processes slow

down with elevation (e.g., Seynave et al., 2006), imply-

ing not only that trees will get older but also that

old-growth restoration takes longer at high altitude.

Managed low-elevation beech forests are usually more

productive so that, when left to natural development,

they can reach old-growth status relatively quickly

compared with higher bioclimatic zones (see Ziaco

et al., 2012b).

Growth history and climatic change

It is generally assumed that the transition from matu-

rity into senescence can be marked by a decline in BAI

(Duchesne et al., 2003). In our study, multicentury-old

(>300 year) beech trees, both in the Alps and the

Fig. 6 Climate of the high-elevation beech populations on the eastern Alps and the central Apennines. (a, b) Walter–Lieth climatic dia-

grams. (c) Summer precipitation minus potential evapotranspiration (PET) as a function of time.

© 2011 Blackwell Publishing Ltd, Global Change Biology, 18, 960–972

968 A. DI FILIPPO e t a l .

Apennines, showed BAI decline, when present, only

for a few decades in comparison with their entire life-

span. Large canopy trees in beech old-growth forests of

the Alps and Apennines have the highest BAI and, as

trees with DBH > 100 cm are seldom found at these

sites, mortality could occur without long periods of

growth decline. In other European tree species, abrupt

growth reduction can be a signal of impending death in

dominant trees (Linares et al., 2009; Levanic et al.,

2011).

In the last century, BAI trends in the Apennines were

initially greater but then declined starting in the 1970s

compared with those in the Alps, where growth has

continued to increase, especially after the 1950s. The

different climatic regimes (Oceanic vs Mediterranean;

Fig. 6) between the two bioclimatic zones are most

likely involved in these recent divergent growth trends.

In fact, while growing season drought increased in the

central Apennines, the Alps maintained a highly posi-

tive water balance thanks to their abundant vegetative

season precipitation (Fig. 6c).

Overall, BAI has increased over the last decades in

the Alps (Motta & Nola, 2001), and in central Europe

this phenomenon has been attributed to positive global

change effects (Kahle et al., 2008; Bontemps et al., 2011).

Such increasing growth trend was uncovered in all the

mountain beech sites sampled to date in the eastern

Alps (see Fig. 3 in Piovesan et al., 2011) and it could be

linked to synergistic effects of multiple environmental

changes, such as nitrogen deposition, CO2 fertilization,

and warming. As high-frequency radial growth vari-

ability of high-mountain beech populations in the east-

ern Alps (Fig. S3) mainly relates to summer

temperature (Di Filippo et al., 2007), the spring–sum-

mer warming of the last decades (Brunetti et al., 2006)

is an important factor for increased growth in the Alps.

On the other hand, as summer drought controls beech

growth in central Italy (Piovesan et al., 2005), BAI

decline in the Apennines can be explained by water

stress, rather than aging (box 1 in Fig. 5; see also Fig. 6

in Piovesan et al., 2008). In Mediterranean beech forests

the ameliorating effects of increasing CO2 and N fertil-

ization cannot compensate for the effects of increasing

aridity on gas exchange (Penuelas et al., 2008; Linares

et al., 2009). Increased water-use efficiency during the

20th century did not translate into enhanced tree

growth at sites where growth is mostly limited by

drought (Penuelas et al., 2011).

The existence of such divergent trends between the

Alps and the Apennines could thus point to diversified,

even opposite, climate change effects in different biocli-

matic contexts of the same region (Beck et al., 2011).

With regard to European beech, reduced productivity

has now been found within the continent in mid-eleva-

tion forests of northeastern France (Charru et al., 2010).

Forest ecosystems of very different climates, from bor-

eal to tropical, have been reported to suffer growth

decline due to warming-induced stress (Barber et al.,

2000; Adams & Piovesan, 2005; Lloyd & Bunn, 2007;

Piovesan et al., 2008; Clark et al., 2010; Silva et al., 2010;

Nock et al., 2011). One can conclude that growth

decline in old-growth forests is linked to environmental

limitations, whose removal allows even old trees to

benefit from favorable external conditions (Phillips

et al., 2008; Johnson & Abrams, 2009). Overall, our

results closely align with Johnson & Abrams (2009)

finding that trees with the greatest lifespan are charac-

terized by low growth rates that keep increasing

throughout the tree’s life. This view is in contrast to a

sigmoidal model of growth, and instead emphasizes

the connection among tree life history, site quality, and

growth rate.

If forest species with a wide distribution show

changes in lifespan related to the bioclimatic zone they

occupy, then climatic changes can either increase or

decrease the maximum age an organism can attain.

Long-term climatic patterns may have influenced beech

lifespan, for instance, reduced growth rates during the

Little Ice Age (Grove, 1988) would have resulted in a

longer tree lifespan. Current climatic changes may

also modify life history traits (Bigler & Veblen, 2009).

Fig. 7 Relationship between average tree-ring width (RW) of

suppressed trees and July mean air temperature in a worldwide

north-hemisphere beech network (data from Peters, 1997); a

straight line was used to represent the equation shown in the

explanatory box. For the meaning of R2adj, F, and P, see caption

to Fig. 1. A similar regression is obtained by using annual mean

temperature.

© 2011 Blackwell Publishing Ltd, Global Change Biology, 18, 960–972

BIOCLIMATE, TREE GROWTH, AND LIFESPAN 969

Climate changes could have contributed to the disap-

pearance from European forests of 500-year-old beech

trees, whose existence was reported at the beginning of

the 20th century by Schlich (1910), whereas nowadays

beech trees of such age seemingly exist only in the

Apennines (Piovesan et al., 2011). Increased radial

growth has been reported in temperate and boreal for-

ests (e.g., Kahle et al., 2008), although sometimes only

at very specific elevations (Salzer et al., 2009), and it is

possible that these trends could affect lifespan.

Carbon cycling in forest ecosystems can be impacted

as well, given that lower longevity translates in faster

turnover of dominant species, thereby potentially can-

celing any fertilizing effect associated with increasing

atmospheric CO2 levels in temperate forests (Bugmann

& Bigler, 2011). In the Mediterranean region, drought

could either increase longevity as it reduces growth

rates (Piovesan et al., 2008), or cause mortality of older

trees as it has been suggested for Californian (van

Mantgem et al., 2009), Spanish (Linares & Tıscar, 2010)

and African (Allen et al., 2010) conifers. In the next dec-

ades, if warming will continue, beech maximum life-

span could be reduced in the Alps by faster growth

(‘greening’) and in the Apennines by drought-induced

mortality (‘browning’).

Acknowledgements

A. Di Filippo, G. Piovesan and B. Schirone were funded, in part,by the 2007AZFFAK PRIN project: ‘Climate change and forests– Dendroecological and ecophysiological responses, productiv-ity and carbon balance on the Italian network of old-growthbeech forests’. F. Biondi was funded, in part, by the Universityof Nevada, Reno. We thank the numerous foresters and stu-dents who helped us during the field surveys. The comments ofthree anonymous reviewers helped us improve an earlier ver-sion of this manuscript.

References

Adams JM, Piovesan G (2005) Long series relationships between global interannual

CO2 increment and climate: evidence for stability and change in role of the tropical

and boreal-temperate zones. Chemosphere, 59, 1595–1612.

Allen AP, Gillooly JF, Brown JH (2005) Linking the global carbon cycle to individual

metabolism. Functional Ecology, 19, 202–213.

Allen CD, Macalady AK, Chenchouni H et al. (2010) A global overview of drought

and heat-induced tree mortality reveals emerging climate change risks for forests.

Forest Ecology and Management, 259, 660–684.

Anderson KJ, Allen AP, Gillooly JF, Brown JH (2006) Temperature-dependence of

biomass accumulation rates during secondary succession. Ecology Letters, 9, 673–

682.

Barber VA, Juday GP, Finney BP (2000) Reduced growth of Alaskan white spruce in

the twentieth century from temperature-induced drought stress. Nature, 405, 668–

673.

Beck PSA, Juday GP, Alix C et al. (2011) Changes in forest productivity across Alaska

consistent with biome shift. Ecology Letters, 14, 373–379.

Bigler C, Veblen TT (2009) Increased early growth rates decrease longevities of coni-

fers in subalpine forests. Oikos, 118, 1130–1138.

Biondi F (2001) A 400-year tree-ring chronology from the tropical treeline of North

America. Ambio, 30, 162–166.

Black BA, Colbert JJ, Pederson N (2008) Relationships between radial growth rates

and lifespan within North American tree species. Ecoscience, 15, 349–357.

Bolstad PV, Vose JM, McNulty SG (2001) Forest productivity, leaf area, and terrain in

southern Appalachian deciduous forests. Forest Science, 47, 419–427.

Bontemps J-D, Herve J-C, Leban J-M, Dhote J-F (2011) Nitrogen footprint in a

long-term observation of forest growth over the twentieth century. Trees, 25,

237–251.

Brunetti M, Maugeri M, Monti F, Nanni T (2006) Temperature and precipitation vari-

ability in Italy in the last two centuries from homogenised instrumental time ser-

ies. International Journal of Climatology, 26, 345–381.

Brunetti M, Lentini G, Maugeri M, Nanni T, Simolo C, Spinoni J (2009a) Estimating

local records for Northern and Central Italy from a sparse secular temperature

network and from 1961–1990 climatologies. Advances in Science and Research, 3,

63–71.

Brunetti M, Lentini G, Maugeri M, Nanni T, Simolo C, Spinoni J (2009b) 1961–90

high-resolution Northern and Central Italy monthly precipitation climatologies.

Advances in Science and Research, 3, 73–78.

Bugmann H, Bigler C (2011) Will the CO2 fertilization effect in forests be offset by

reduced tree longevity? Oecologia, 165, 533–544.

Bunn AG, Biondi F (2010) Dendrochronology in R with the dplR library. In: Abstracts

of WorldDendro2010, The 8th International Conference on Dendrochronology (eds Mie-

likainen K, Makinen H, Timonen M), p. 274. METLA, Finland.

Canham CD (1990) Suppression and release during canopy recruitment in Fagus gran-

difolia. Bulletin of the Torrey Botanical Club, 117, 1–7.

Castagneri D, Garbarino M, Berretti R, Motta R (2010) Site and stand effects on coarse

woody debris in Montane mixed forests of Eastern Italian Alps. Forest Ecology and

Management, 260, 1592–1598.

Charru M, Seynave I, Morneau F, Bontemps J-D (2010) Recent changes in forest

productivity: an analysis of national forest inventory data for common beech

(Fagus sylvatica L.) in north-eastern France. Forest Ecology and Management, 260,

864–874.

Clark DB, Clark DA, Oberbauer SF (2010) Annual wood production in a tropical rain

forest in NE Costa Rica linked to climatic variation but not to increasing CO2. Glo-

bal Change Biology, 16, 747–759.

Dang H, Jiang M, Zhang Q, Zhang Y (2007) Growth responses of subalpine fir (Abies

fargesii) to climate variability in the Qinling Mountain, China. Forest Ecology Man-

agement, 240, 143–150.

Di Filippo A, Biondi F, Cufar K et al. (2007) Bioclimatology of beech (Fagus sylvatica

L.) in the Eastern Alps: spatial and altitudinal climatic signals identified through a

tree-ring network. Journal of Biogeography, 34, 1873–1892.

Duchesne L, Ouimet R, Morneau C (2003) Assessment of sugar maple health based

on basal area growth pattern. Canadian Journal of Forest Research, 33, 2074–2080.

ECS (European Committee for Standardization) (1995) Durability of Wood and

Wood Based Products. Natural Durability of Solid Wood. Part 2: Guide to Natural

Durability and Treatability of Selected Wood Species of Importance in Europe. EN

350-2, Brussels.

Enquist BJ, West GB, Brown JH (2009) Extensions and evaluations of a general quanti-

tative theory of forest structure and dynamics. Proceedings of the National Academy

of Sciences, 106, 7046–7051.

Flanary BE, Kletetschka G (2005) Analysis of telomere length and telomerase activity

in tree species of various life-spans, and with age in the bristlecone pine Pinus lon-

gaeva. Biogerontology, 6, 101–111.

Frelich LE (2002) Forest Dynamics and Disturbance Regimes, Studies from Temperate Ever-

green-Deciduous Forests. Cambridge University Press, Cambridge, UK.

Gillooly JF, Brown JH, West GB, Savage VM, Charnov EL (2001) Effects of size and

temperature on metabolic rate. Science, 293, 2248–2251.

Grissino-Mayer HD (2001) Evaluating crossdating accuracy: a manual and tutorial for

the computer program COFECHA. Tree-Ring Research, 57, 205–221.

Grove JM (1988) The Little Ice Age. Routledge, London.

Issartel J, Coiffard C (2011) Extreme longevity in trees: live slow, die old? Oecologia,

165, 1–5.

Johnson SE, Abrams MD (2009) Age class, longevity and growth rate relationships:

protracted growth increases in old trees in the eastern United States. Tree Physiol-

ogy, 29, 1317–1328.

Kahle HP, Karjalainen T, Schuck A et al. (2008) Causes and Consequences of Forest

Growth Trends in Europe. European Forest Institute Research Report n. 21 – Result

of the RECOGNITION Project. Brill, Leiden.

Karl I, Fischer K (2009) Altitudinal and environmental variation in lifespan in the

Copper butterfly Lycaena tityrus. Functional Ecology, 23, 1132–1138.

Larson DW (2001) The paradox of great longevity in a short-lived tree species. Experi-

mental Gerontology, 36, 651–673.

© 2011 Blackwell Publishing Ltd, Global Change Biology, 18, 960–972

970 A. DI FILIPPO e t a l .

Levanic T, Cate M, McDowell NG (2011) Associations between growth, wood anat-

omy, carbon isotope discrimination and mortality in a Quercus robur forest. Tree

Physiology, 31, 298–308.

Linares JC, Tıscar PA (2010) Climate change impacts and vulnerability of the

southern populations of Pinus nigra subsp. salzmannii. Tree Physiology, 30, 795–806.

Linares JC, Camarero JJ, Carreira JA (2009) Interacting effects of climate and forest-

cover changes on mortality and growth of the southernmost European fir forests.

Global Ecology and Biogeography, 18, 485–516.

Lloyd AH, Bunn AG (2007) Responses of the circumpolar boreal forest to 20th cen-

tury climate variability. Environmental Research Letters, 2, 045013.

Loehle C (1988) Tree life history strategies: the role of defenses. Canadian Journal of

Forest Research, 18, 209–222.

Loehle C (2000) Strategy space and the disturbance spectrum: a life history model for

tree species coexistence. American Naturalist, 156, 14–33.

van Mantgem PJ, Stephenson NL, Byrne JC et al. (2009) Widespread increase of tree

mortality rates in the Western United States. Science, 323, 521–524.

Marba N, Duarte CM, Augustı S (2007) Allometric scaling of plant life history. Pro-

ceedings of the National Academy of Sciences, 104, 15777–15780.

McCoy MW, Gillooly JF (2008) Predicting natural mortality rates of plants and ani-

mals. Ecology Letters, 11, 710–716.

Mencuccini M, Martınez-Vilalta J, Vanderklein D, Hamid HA, Korakaki E, Lee S,

Michiels B (2005) Size-mediated ageing reduces vigour in trees. Ecology Letters, 8,

1183–1190.

Mitchell TD, Jones PD (2005) An improved method of constructing a database of

monthly climate observations and associated high-resolution grids. International

Journal of Climatology, 25, 693–712.

Moore DJP, Aref S, Ho RM, Pippen JS, Hamilton JG, De Lucia EH (2006) Annual basal

area increment and growth duration of Pinus taeda in response to eight years of

free-air carbon dioxide enrichment. Global Change Biology, 12, 1367–1377.

Motta R, Nola P (2001) Growth trends and dynamics in sub-alpine forest stands in

the Varaita Valley (Piedmont, Italy) and their relationships with human activities

and global change. Journal of Vegetation Science, 12, 219–230.

Muller-Landau HC, Condit RS, Harms KE et al. (2006) Comparing tropical forest tree

size distributions with the predictions of metabolic ecology and equilibrium mod-

els. Ecology Letters, 9, 589–602.

Munch SB, Salinas S (2009) Latitudinal variation in lifespan within species is

explained by the metabolic theory of ecology. Proceedings of the National Academy of

Sciences, 106, 13860–13864.

Nock CA, Baker PJ, Wanek W, Leis A, Grabner M, Bunyavejchewin S, Hietz P (2011)

Long-term increases in intrinsic water-use efficiency do not lead to increased stem

growth in a tropical monsoon forest in western Thailand. Global Change Biology, 17,

1049–1063.

Parish R, Antos JA (2006) Slow growth, long-lived trees, and minimal disturbance

characterize the dynamics of an ancient Montane forest in coastal British Colum-

bia. Canadian Journal of Forest Research, 36, 2826–2838.

Penuelas J, Munne-Bosch S (2010) Potentially immortal? New Phytologist, 187, 564–

567.

Penuelas J, Hunt JM, Ogaya R, Jump AS (2008) Twentieth century changes of tree-

ring d13C at the southern range-edge of Fagus sylvatica: increasing water-use effi-

ciency does not avoid the growth decline induced by warming at low altitudes.

Global Change Biology, 14, 1076–1088.

Penuelas J, Canadell JG, Ogaya R (2011) Increased water-use efficiency during the

20th century did not translate into enhanced tree growth. Global Ecology and Bioge-

ography, 20, 597–608.

Peters R. (1997) Beech Forests, Geobotany 24. Kluwer Academic Publishers, The Neth-

erlands.

Peterson DW, Peterson DL (2001) Mountain hemlock growth responds to climatic

variability at annual and decadal time scales. Ecology, 82, 3330–3345.

Phillips NG, Buckley TN, Tissue DT (2008) Capacity of old trees to respond to envi-

ronmental change. Journal of Integrative Plant Biology, 50, 1355–1364.

Piovesan G, Biondi F, Bernabei M, Di Filippo A, Schirone B (2005) Spatial and altitu-

dinal bioclimatic zones of the Italian peninsula identified from a beech (Fagus sylv-

atica L.) tree-ring network. Acta Oecologica, 27, 197–210.

Piovesan G, Biondi F, Di Filippo A, Alessandrini A, Maugeri M (2008) Drought-

driven growth reduction in old beech (Fagus sylvatica L.) forests of the central

Apennines, Italy. Global Change Biology, 14, 1265–1281.

Piovesan G, Alessandrini A, Baliva M et al. (2010) Structural patterns, growth

processes, carbon stocks in an Italian network of old-growth beech forests. Italian

Journal of Forest and Mountain Environments, 65, 557–590.

Piovesan G, Alessandrini A, Biondi F, Di Filippo A, Schirone B, Ziaco E (2011)

Bioclimatology, growth processes, longevity and structural attributes in an Italian

network of old-growth beech forests spreading from the Alps to the Apennines.

In: Beech Forests - Joint Natural Heritage of Europe, BfN-Skripten n. 297 (eds Knapp

HD, Fichtner A), pp. 173–192. Bonn-Bad Godesberg, Bonn, Germany.

Poage NJ, Weisberg PJ, Impara PC, Tappeiner JC, Sensenig TS (2009) Influences of cli-

mate, fire, and topography on contemporary age structure patterns of Douglas fir

at 205 old forest sites in western Oregon. Canadian Journal of Forest Research, 39,

1518–1530.

Price CA, Gilooly JF, Allen AP, Weitz JS, Niklas KJ (2010) The metabolic theory of ecol-

ogy: prospects and challenges for plant biology. New Phytologist, 188, 696–710.

R Development Core Team (2010) R: A Language and Environment for Statistical Com-

puting. R Foundation for Statistical Computing, Vienna, Austria. Available at:

http://www.r-project.org/(accessed 13 December 2011).

Reich PB, Oleksyn J, Modrzynski J, Tjoelker MG (1996) Evidence that longer needle

retention of spruce and pine populations at high elevations and high latitudes is

largely a phenotypic response. Tree Physiology, 16, 643–647.

Richardson SJ, Smale MC, Hurst JM et al. (2009) Short communication: large-tree

growth and mortality rates in forests of the central North Island, New Zealand.

New Zealand Journal of Ecology, 33, 208–215.

Robichaud E, Methven IR (1993) The effect of site quality on the timing of stand

breakup, tree longevity, and the maximum attainable height of black spruce. Cana-

dian Journal of Forest Research, 23, 1514–1519.

Russo SE, Wiser SK, Coomes DA (2007) Growth-size scaling relationships of woody

plant species differ from predictions of the Metabolic Ecology Model. Ecology Let-

ters, 10, 889–901.

Salzer MW, Hughes MK, Bunn AG, Kipfmueller KF (2009) Recent unprecedented

tree-ring growth in bristlecone pine at the highest elevations and possible causes.

Proceedings of the National Academy of Sciences, 106, 20348–20353.

Schlich W (1910) Schlich’s Manual of Forestry, Volume II. Bradbury, Agnew and Co.,

London.

Schulman E (1954) Longevity under adversity in conifers. Science, 119, 396–399.

Seynave I, Gegout JC, Herve JC, Dhote JF (2006) Facteurs ecologiques et production

du hetre en France. Foret Entreprise, 167, 41–45.

Silva LCR, Anand M, Leithead MD (2010) Recent widespread tree growth decline

despite increasing atmospheric CO2. PLoS ONE, 5, e11543.

Simolo C, Brunetti M, Maugeri M, Nanni T, Speranza A (2010) Understanding climate

change-induced variations in daily temperature distributions over Italy. Journal of

Geophysical Research – Atmospheres, 115, D22110.

Splechtna BE, Dobry J, Klinka K (2000) Tree-ring characteristics of subalpine fir (Abies

lasiocarpa (Hook.) Nutt.) in relation to elevation and climatic fluctuations. Annals of

Forest Science, 57, 89–100.

Stephenson NL, van Mantgem PJ (2005) Forest turnover rates follow global and

regional patterns of productivity. Ecology Letters, 8, 524–531.

Stokes MA, Smiley TL (1996) An Introduction to Tree-Ring Dating. University of

Arizona Press, Tucson, AZ.

Wang X, Hao Z, Zhang J, Lian J, Li B, Ye J, Yao X (2009) Tree size distributions in an

old-growth temperate forest. Oikos, 118, 25–36.

Watson JM, Riha K (2011) Telomeres, aging, and plants: from weeds to Methuselah.

A mini-review. Gerontology, 57, 129–136.

Wunder J, Brzeziecki B, Zybura H, Reineking B, Bigler C, Bugmann H (2008) Growth

mortality relationships as indicators of life-history strategies: a comparison of nine

tree species in unmanaged European forests. Oikos, 117, 815–828.

Ziaco E, Alessandrini A, Blasi S, Di Filippo A, Dennis S, Piovesan G (2012a) Commu-

nicating old-growth forest through an educational trail. Biodiversity and Conserva-

tion, 21, 131–144.

Ziaco E, Di Filippo A, Alessandrini A, D’Andrea E, Piovesan G (2012b) Old-growth

attributes in a network of Apennines (Italy) beech forests: disentangling the role

past human interferences and biogeoclimate. Plant Biosystems, doi: 10.1080/

11263504.2011.650729.

Zianis D, Mencuccini M (2005) Aboveground net primary productivity of a beech (Fa-

gus moesiaca) forest: a case study of Naousa forest, northern Greece. Tree Physiol-

ogy, 25, 713–722.

© 2011 Blackwell Publishing Ltd, Global Change Biology, 18, 960–972

BIOCLIMATE, TREE GROWTH, AND LIFESPAN 971

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Figure S1. Maximum tree age (AgeMax) vs. site elevation for all the 27 (12 old-growth + 15 managed) beech forests in the easternAlps (solid symbols) and the central Apennines (empty symbols).Figure S2. Linear sample relationships (equations shown in the boxes) between vegetative season (May–September) mean tempera-ture and log-transformed (a) mean age of the five oldest trees, computed after excluding the suppression years and (b) 99th percen-tile of the distribution of ring width (Max RW).Figure S3. Basal area increment (BAI) of old (>300 years) beech trees sampled in the central Apennines and the eastern Alps(above).

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by theauthors. Any queries (other than missing material) should be directed to the corresponding author for the article.

© 2011 Blackwell Publishing Ltd, Global Change Biology, 18, 960–972

972 A. DI FILIPPO e t a l .