Fagus sylvatica L. - University of Nevada, Reno

Journal of Biogeography (J. Biogeogr.) (2007) 34, 1873–1892

ORIGINAL ARTICLE

Bioclimatology of beech (Fagus sylvatica L.) in the Eastern Alps: spatial and altitudinal climatic signals identified through a tree-ring network Alfredo Di Filippo1*, Franco Biondi2, Katarina Cˇufar3, Martıń de Luis4, Michael Grabner5, Maurizio Maugeri6, Emanuele Presutti Saba1, Bartolomeo Schirone1 and Gianluca Piovesan1

1

DAF, Universita` degli Studi della Tuscia, Viterbo, Italy, 2DendroLab, Department of Geography, University of Nevada, Reno, NV, USA, 3Department of Wood Science and Technology, Biotechnical Faculty, University of Ljubljana, Ljubljana, Slovenia, 4Departamento de Geografıá y Ordenacioń del Territorio, Universidad de Zaragoza, Zaragoza, Spain, 5 University of Natural Resources and Applied Life Sciences, Vienna, Austria, 6Istituto di Fisica Generale Applicata, Milan, Italy

ABSTRACT

Aim To identify the dominant spatial patterns of Fagus sylvatica radial growth in the Eastern Alps, and to understand their relationships to climate variation and bioclimatic gradients. Location Fourteen beech stands in the Eastern Alps, growing between 200 and 1500 m a.s.l. in Italy, Slovenia and Austria. Methods At each site, trees were sampled using increment borers or by taking discs from felled trees. Cores and discs were processed by measuring and crossdating ring width. Ring width series were standardized, averaged, and prewhitened to obtain site chronologies. Hierarchical Cluster Analysis (HCA) and Principal Components Analysis of prewhitened site chronologies were used to identify spatial and altitudinal growth patterns, related to the bioclimatic position of each stand. Bootstrap correlation and response functions were computed between monthly climatic variables and either principal component scores or composite chronologies from stands associated by HCA. The stability of dendroclimatic signals was analyzed by moving correlation functions (MCF). Correlation analysis (teleconnections) based on a data base of 37 Italian and Slovenian beech tree-ring chronologies revealed the spatial extent of principal component scores. Results Sampled trees were 200–400 years old, representing the oldest beech trees that have been crossdated for the Alps to date. Maximum age was directly related to altitude and to the presence of historical forms of conservation. Treering parameters varied according to geographic patterns and the age of sampled trees. Stands were bioclimatically organized according to their location, and with reference to their elevation and distance from the Adriatic Sea. A direct response to winter temperature was found at all elevations. The altitudinal gradient ranged from low-elevation stands, characterized by a Mediterranean-type, late spring– summer drought signal, to mountain and high-elevation stands, characterized by a direct response to growing season temperature plus an inverse response to the previous year’s July temperature. The mountain and high-elevation signal was evident in Austria, the Central Alps and Slovenia, while the low-elevation signal was confined to mountains adjacent to the Adriatic Sea. MCF revealed trends in the response to climatic factors affecting tree-ring formation in mountain and high-mountain stands linked to climatic warming.

*Correspondence: Alfredo Di Filippo, DAF, Facolta` di Agraria, Universita` degli Studi della Tuscia, Via SC de’ Lellis snc, 01100, Viterbo, Italy. E-mail: [email protected].

Main conclusions Dendroclimatic networks can be used for bioclimatic studies of tree populations. A biogeographical separation emerged between the Alps and the Apennines at the upper elevations, while different degrees of mediterraneity distinguished sites at lower elevations. This information will be useful in assessing

ª 2007 The Authors Journal compilation ª 2007 Blackwell Publishing Ltd

www.blackwellpublishing.com/jbi doi:10.1111/j.1365-2699.2007.01747.x

1873

A. Di Filippo et al.

any future climate-related bioclimatic shifts, especially for forests at ecotones and along altitudinal gradients. Keywords Alps, altitudinal gradient, bioclimatology, dendroclimatology, ecological gradient, ecotone, Fagus, old-growth forests, tree growth, tree-ring analysis.

INTRODUCTION The biogeographical study of plant–climate relationships has been an important field of research since the 19th century, and aims to explain vegetation patterns based on climate variation (Woodward, 1987). In particular, the effect of climate on species physiology and geographic range has provided the causal mechanism linking climate with vegetation and biome distribution (Walter, 1985). With respect to forest ecosystems, dendroecology can contribute to bioclimatic studies by improving the analysis of tree growth response to environmental gradients, thereby refining the classifications that are based on climate–vegetation interactions. This approach considers tree-ring parameters as bioindicators that integrate the environmental factors controlling forest growth. Tree-ring records can show growth–climate relationships over space (from a forest stand to a whole hemisphere) and time (from seasons to centuries) (Fritts, 1976). The effect of climate forcing on tree growth has been studied at local (Douglass, 1920), regional (Meko et al., 1993) and hemispheric (Briffa et al., 2002) spatial scales. In this context, sampling stands of the same species (e.g. beech) at different locations and altitudes can provide an objective bioclimatic classification of tree populations (Piovesan et al., 2005a). A number of bioclimatic classification systems are currently available (e.g. Walter, 1985; Bailey, 1996), with each system defining bioclimatic units (e.g. ecoregions and zonobiomes) based on different methods and types of data (e.g. Thompson et al., 2005). Recently, the use of statistical techniques, digital data bases and advanced spatial analytical approaches have allowed the numerical classification of environmental variables to clarify their role in affecting plant distribution (Laurent et al., 2004; Metzger et al., 2005). However, especially at regional scales, there is a need to integrate patterns of species distribution with ecological processes, paying special attention to the effects of climate variability (Whittaker et al., 2005). Horizontal and vertical gradients in tree–climate relationships provide the basis for defining bioclimatic units in terms of the leading dendroclimatic signals. Such studies generate the basic information necessary to perform climatic reconstructions from tree rings (Frank & Esper, 2005), and can be the starting point to define simulation models of plant community response to a changing climate (Cook et al., 2001). In addition,

1874

this way of classifying woody vegetation into bioclimatic units offers a valuable tool for a science based management of forest ecosystem, linking climate fluctuations to forest productivity (Biondi, 1999). Finally, if long instrumental records are available, it is possible to explore the temporal stability of the observed climate–growth relationships (Biondi, 2000), and formulate hypotheses about the potential effects of climatic change on plant functioning and community dynamics (Jump & Pen˜uelas, 2005; Whittaker et al., 2005). Dendroecologists have used European beech (Fagus sylvatica L.) extensively during the last two decades (Eckstein & Frisse, 1982; Gutierrez, 1988; Rozas, 2001; Dittmar et al., 2003; Lebourgeois et al., 2005), taking advantage of the widespread distribution, sensitivity to climate and longevity of beech (Bourquin-Mignot & Girardclos, 2001; Piovesan et al., 2005b). In Italy, investigations were carried out at local (e.g. Biondi, 1993; Bernabei et al., 1996; Piutti & Cescatti, 1997; Piovesan et al., 2003) and regional (Biondi, 1992; Biondi & Visani, 1996; Piovesan et al., 2005a) scales. European beech approaches the southern edge of its geographic range in Italy, where its altitudinal range extends more than 1500 m, from about 300–400 to 2000–2100 m a.s.l. in central and southern Italy, and from 200–300 to 1500–1600 m in the Alps. Our main hypothesis is that, because beech is present in both the Alps and the Apennines, its tree-ring records can be used to detect ecological differences between these two mountain ranges. If a large enough number of chronologies are available, it should also be possible to distinguish dendroclimatic responses along ecological gradients, such as elevation. To date, dendroclimatic networks in the Alps have focused on conifer species growing at high elevation or at the tree line (e.g. Kienast et al., 1987; Urbinati et al., 1997; Rolland, 2002; Frank & Esper, 2005). In this study we tested our hypotheses using a transnational network of 14 European beech stands sampled at different elevations in the Eastern Alps (Italy, Slovenia and Austria), which is a focus region for bioclimatic studies (e.g. Walter, 1985; Ellenberg, 1988). Our main objectives were (1) to identify the dominant spatial patterns of radial growth in Fagus sylvatica, (2) to investigate how such patterns relate to climatic variability and geographic location, and (3) to improve the definition of bioclimatic gradients, and refine the classification systems that rely on them.

Journal of Biogeography 34, 1873–1892 ª 2007 The Authors. Journal compilation ª 2007 Blackwell Publishing Ltd

Bioclimatology of beech in the Eastern Alps (a)

(b)

Figure 1 (a) Geographic distribution of Fagus sylvatica (von Wuehlisch, 2006), with outline (rectangle) of the study area. (b) Enlargement of the study area showing the location of sampled sites.

METHODS Study areas A total of 14 European beech (Fagus sylvatica L.) forests were sampled in the following ranges: Julian Alps in Italy and Slovenia, Carnic Alps in Italy, and northern Alps in Austria (Fig. 1). Sampled sites were located from 46.19 to 48.33 N latitude, and from 12.75 to 15.43 E longitude, covering an altitudinal range of 1300 m, from 200 to 1500 m a.s.l. (Table 1). This region occupies the central-southern portion of the geographic range of beech distribution, where this species has a noticeable spread in altitude (Ellenberg, 1988).

According to the classification of Ko¨ppen–Trewartha (Trewartha, 1968), the Carnic and the Austrian ranges pertain to temperate climates (D), while the Julian mountains are at the boundary between temperate and subtropical dry summer (Cs) climates. A recent environmental classification of Europe (Metzger et al., 2005) defined the central bulk of our network as the Alpine Environmental Zone, the southernmost Julian Alps as the boundary with the Mediterranean Mountains Zone, and the northernmost Dunkelsteinerwald as within the Continental Zone. Almost all stands were managed as high forests where beech was the dominant species; only the lower elevation sites (MOT, NIM, PEC, TOA and TOB; see Table 1 for acronyms)


1875

A. Di Filippo et al. Table 1 Geographical and structural features of the sampled beech sites.

Site

Alpine range

Country

Code

Latitude (N)

Longitude (E)

Elevation* (m a.s.l.)

Aspect

Slope (%)

d.b.h. (cm)

Ht.à (m)

La Motta Tolmin A Nimis Pechinie Tolmin B Tolmin C Gracco Cleulis Tre Confini Timau Paularo Lateis Dunkelsteinerwald Hallstatt

J J J J J J C C C C C C N N

I S I I S S I I I I I I A A

MOT TOA NIM PEC TOB TOL GRA CLE TRE TIM PAU LAT DSW HSA

46.1886 46.2000 46.2006 46.2807 46.2167 46.2333 46.5514 46.5584 46.5039 46.5817 46.5297 46.4594 48.3333 47.8333

13.3153 13.7333 13.2769 13.2027 13.7500 13.7666 12.8519 13.0006 13.5831 13.0050 13.1172 12.7489 15.4333 13.7500

250 355 560 670 821 1328 825 930 1100 1160 1385 1450 650 1400

SW SW SE NE SW S S NE – S W S E SE

10–40 40–80 45 10–30 40–80 40–80 70–90 20–60 20–60 70–90 70–90 75–110 40–60 100–120

39–65 39–50 34–65 36–62 39–53 33–55 42–72 35–70 35–55 40–95 52–100 33–70 30–50 15–36

20 24 27 20 27 22 21 24 30 25 27 28 22 12

(200-300) (290-420)

(797-845) (1240-1415) (750-900)

(825-1500) (1275-1500) (1370-1530)

C, Carnic Alps; J, Julian Alps; N, Northern Austrian Alps. I, Italy; S, Slovenia; A, Austria. *Mean with range in parentheses. Diameter at breast height range of sampled trees. àMean height of the three to four tallest trees in the stand.

included species (e.g. Acer spp., Fraxinus spp. and Ostrya carpinifolia Scop.) that were managed as coppice-with-standards, and even at those sites beech maintained a monocormic stem. These low-elevation beech stands, located at the altitudinal limit of the species, cover a few hectares each, and are found within mixed deciduous forests. Low-elevation forests on the Italian side of the Julian Alps and in Slovenia (MOT, NIM, PEC, TOA and TOB) were privately owned. The Slovenian sites were strongly affected by the battles of World War I, and later on they were over-exploited by the local population. Therefore well-preserved forests are only found in remote areas that are difficult to access. All stands sampled in

the Carnic Alps were publicly owned, being situated above towns for protection against avalanches and rock falls (‘protection forests’). Some of these forests (TIM, CLE, LAT and GRA), called ‘boschi banditi’ (Fig. 2), were maintained by the Republic of Venice during the 16th century to produce timber (especially masts) for ships, or set aside by local inhabitants to provide shade and food (beech nuts and understorey species) for livestock (Paiero et al., 1975). As a final note, at several sites included in this network specifically because of their relatively undisturbed conditions and oldgrowth characteristics, land managers stipulated that only one core per tree could be taken so as to minimize damage. The

Figure 2 View of the Gracco ‘protection forest’ in Carnia, where beech age exceeds 300 years (Photograph by G. Piovesan).

1876


Bioclimatology of beech in the Eastern Alps same constraint was imposed in privately owned forests, including most of the low-elevation beech stands. Sampling and chronology building In each stand, tree selection focused on dominant or co-dominant trees, either isolated or grouped, with the best combination of old age and trunk health. Trees were cored at breast height (1.3 m from the ground) using an increment borer, taking one or two cores per tree. At the Slovenian sites, all samples were cross-sections taken from trees already felled for other purposes. Tree-ring chronologies were developed from wood samples using standard dendrochronological procedures (Stokes & Smiley, 1996). After surfacing and preliminary visual crossdating, tree-ring widths were measured to the nearest 0.01 mm using the system CCTRMD (Aniol, 1987) and the program CATRAS (Aniol, 1983) or the LINTAB measuring table and TSAP/X programme (Rinn, 1996). Treering series were visually and statistically compared with each other to ensure accuracy of crossdating and measurement (Holmes, 1983; Grissino-Mayer, 2001). Locally absent rings (LAR), when detected, were given ring width equal to zero. The percentage of locally absent rings was calculated on the entire length of the tree-ring series. Dendrochronological parameters (mean ring width, standard deviation, mean sensitivity and first-order autocorrelation) were computed for all measured ring-width series. The same parameters were then computed on raw site chronologies, obtained by arithmetically averaging the ring-width series by site. The years 1942–2001 were a common period that included at least three ring-width series at each site. Site location and tree age were then used to explain the variability of dendrochronological parameters, computed for the 1942–2001 period using the raw site chronologies. Standardized tree-ring chronologies were produced for each site using the following formula: Pi¼nt 0:5 It ¼ i¼1 ðw yÞit þ cit nt with It = chronology value at year t; nt = number of samples for year t, with nt ‡ 3; w = crossdated ring width of sample i in year t; y = value of sample i in year t computed by fitting a modified negative exponential with asymptote ‡ 0 or a straight line with slope £ 0 to the ith ring-width series; cit = constant added to sample i in year t so that the standardized chronology has mean equal to 1. Although ‘standardized indices’ used in dendrochronology are often calculated as ratios between the measurement and the fitted curve value, there is evidence that variance-stabilized residuals should be preferred (Cook & Peters, 1997; Biondi, 1999; Helama et al., 2004). Each standardized chronology was prewhitened by fitting autoregressive (AR) models (Biondi & Swetnam, 1987) to remove any biological trend and enhance the climatic signal (Cook et al., 2001). The linear correlation between prewhitened site chronologies was computed for the common period 1942–2001. Finally,

composite chronologies were obtained by pooling together all ring-width series from sites whose prewhitened chronologies were associated by multivariate analysis (see next paragraph). The same procedure used to compute prewhitened site chronologies was used to compute prewhitened composite chronologies. Chronology confidence was evaluated by computing the expressed population signal (EPS) (Wigley et al., 1984). EPS values were computed for the period 1942–2001, 1952–2001 and 1962–2001 on the prewhitened site chronologies. These three periods were chosen to show how the increase in the number of available samples affects the EPS statistic. The limited extension of the low elevation sites, together with the restrictions imposed on the number of samples per tree, were responsible for a reduced sample depth in the early part of these site chronologies. For prewhitened composite chronologies (see Table S4 in Supplementary Material), EPS values were calculated over the entire length of the chronology using 50-year moving windows with a 40-year overlap. Multivariate analysis The period 1942–2001, common to the 14 prewhitened beech chronologies, was considered for multivariate analysis. Hierarchical Cluster Analysis (HCA) and Principal Components Analysis (PCA) were based on the correlation matrix between chronologies. HCA was used for the first detection of groupings among the 14 chronologies (Ludwig & Reynolds, 1988). Distance between variables was based on (1 – r), with r = Pearson’s product–moment correlation coefficient, while clusters were identified by means of the average distance between all pairs of variables contained in two groups (Stenson & Wilkinson, 2004). Because clusters were generated according to the degree of growth affinity between chronologies, the HCA dendrogram, when interpreted with consideration to the spatial and altitudinal location of each forest, could reveal the bioclimatic organization of the network (Piovesan et al., 2005a). The main modes of common growth variability among stands were represented by Principal Component (PC) scores, or amplitudes (Piovesan et al., 2005a). Component loadings (eigenvectors), which display the pattern of association of chronologies with each component, were employed to detect groupings in the tree-ring network. Selection of PCs was guided by Kaiser’s Rule (Kaiser, 1992). The combination of HCA and PCA is essential because HCA produces a clear-cut bioclimatic classification of the network, and PCA allows the break down (thus the description) of the dominant climatic signals responsible for the observed classification. The spatial extent of the common signals was investigated by teleconnection analysis (Fritts, 1976), performed by correlating PC scores for the period 1942–88 with beech chronologies developed for Austria, Italy and Slovenia, standardized and prewhitened as in our site chronologies. Graphical representations of correlation maps were produced with the software GMT (Wessel & Smith, 1998).


1877

A. Di Filippo et al.

Figure 3 Graphs of the 14 prewhitened site chronologies and of the number of samples per year. Plots were arranged according to chronology length.

The climate–growth relationship Dendroclimatic correlation and response functions (Biondi & Waikul, 2004) were calculated between monthly climate variables and two types of tree-ring bioindicators. One type was the first and second principal component scores of prewhitened site chronologies; the other type was the prewhitened composite chronologies. The bootstrap method 1878

(Efron & Tibshirani, 1986; Guiot, 1991) was used for significance testing. Explanatory climate variables spanned a 17-month window, from October of the current growth year to June of the previous year. Climatic data for the Italian and Slovenian sites were developed under the research project CLIMAGRI (Brunetti et al., 2006), and were organized in a grid of 1 · 1 cells, with each cell containing the monthly values of minimum and maximum temperature anomalies


Bioclimatology of beech in the Eastern Alps and of precipitation ratios. Climate data for Italy and Slovenia were obtained by averaging the two grid cells (46 N, 13 E and 46 N, 14 E) that included our network sites. In Austria, monthly total precipitation and mean air temperature were available for the period 1942–2001 from the meteorological station of Kremsmu¨nster (48.052 N, 14.127 E, 403 m a.s.l.) for the Hallstatt site, and from the station of Hohe Warte (48.299 N, 16.356 E, 203 m a.s.l.) for the Dunkelsteinerwald site. As climatic series for Italy and Slovenia reached the beginning of the 19th century, we calculated moving correlation functions (MCF) (Biondi, 1997) with the longest and best replicated composite chronologies, i.e. those for mountain and high mountain sites (see Table S4 in Supplementary Material), using a 70-year window. By doing so, we investigated the temporal stability of climatic signals identified for those bioclimatic units. RESULTS Characteristics of the tree-ring network A total of 248 cores from 188 trees were used in dendrochronological analyses (Table 2a). Several old trees, often exceeding 200–300 years of age, were identified (Fig. 3). Mean ring width (MW) was negatively correlated with elevation (Table 3), with annual increments ranging from >4 mm in the lowlands to c. 1 mm at the highest elevations (Table 2a). The maximum age (Nmax) at each stand had a positive correlation with altitude (r = 0.75, P < 0.001; this linear relationship is graphically represented in Fig. 4). The main features of raw and prewhitened site chronologies (plotted in Fig. 3) are summarized in Table 2b.

MW was also negatively correlated with latitude (Table 3). Latitude may affect MW because the southern stands, which are closer to the Adriatic Sea, experience a milder climate than the northern, more continental sites. Finally, MW was negatively correlated with tree age, as old-growth beech forests are characterized by trees showing long periods of reduced growth, occurring when they occupy the suppressed layers of the canopy (Piovesan et al., 2005b). MW values were also negatively correlated with LAR percentage (r = )0.50, P < 0.05), possibly because LARs were frequent in old-growth stands, especially during the suppressed growth periods. As most LARs were present in the younger portion of measured cores during periods of slow growth, it is likely that their positive correlation with altitude can be related to the presence of older trees at higher elevations. The standard deviation (SD) had mean values varying between 0.29 and 1.75 (Table 2a), and was negatively correlated with altitude as well as to age of a site (Length) and of individual trees (Nmax) (Table 3). This may be due to the positive correlation between MW and SD (r = 0.82, P < 0.001). The mean sensitivity (MS), a measure of year-toyear variability, was comprised between 0.19 and 0.36, with minimum values at CLE and at low elevation sites on the Julian Alps (Table 2a). MS was correlated positively with latitude (Table 3), most likely because both Austrian chronologies had MS values above 0.3 (Table 2a). The first-order autocorrelation coefficient (A1), a measure of the persistence in time series, varied between 0.61 and 0.79 (Table 2a), and decreased with latitude, longitude and altitude (Table 3), but correlations with longitude and altitude were immediately below common significance thresholds. It is interesting to note that A1 was inversely correlated with MS (r = )0.62, P < 0.05). The order (p) of the autoregressive model used for

Table 2(a) Summary of the ring width series for each site.

Site

Code

Trees

Cores

MW* (mm year)1)

SD* (mm year)1)

MS*



19 9 13 12 5 10 20 23 12 21 20 24 14 11

19 9 13 13 5 10 20 23 12 27 20 28 28 21

4.05 2.60 3.09 3.11 3.28 1.61 1.43 1.37 1.64 1.28 2.11 1.05 1.27 0.54

1.75 1.00 1.44 1.40 1.18 0.69 0.70 0.50 0.71 0.72 0.80 0.49 0.59 0.29

0.26 0.22 0.26 0.22 0.20 0.25 0.24 0.19 0.24 0.26 0.27 0.25 0.32 0.36

(1.70–5.44) (1.71–3.50) (1.46–4.34) (2.11–5.77) (2.34–4.78) (0.74–2.63) (0.74–2.59) (0.40–3.19) (0.85–2.81) (0.70–3.79) (0.88–4.08) (0.50–2.08) (0.79–2.06) (0.25–0.73)

(1.10–2.49) (0.59–1.41) (0.75–1.89) (0.62–2.47) (0.88–1.50) (0.45–1.00) (0.41–1.29) (0.19–0.90) (0.41–1.00) (0.36–1.33) (0.47–1.44) (0.23–1.00) (0.36–0.86) (0.15–0.42)

A1* (0.18–0.44) (0.15–0.29) (0.20–0.34) (0.13–0.40) (0.14–0.24) (0.20–0.33) (0.19–0.29) (0.13–0.26) (0.16–0.30) (0.16–0.33) (0.16–0.35) (0.16–0.36) (0.24–0.37) (0.29–0.43)

0.70 0.76 0.76 0.72 0.77 0.74 0.78 0.79 0.74 0.77 0.61 0.77 0.65 0.67

(0.44–0.94) (0.57–0.86) (0.53–0.88) (0.39–0.91) (0.66–0.84) (0.56–0.87) (0.52–0.87) (0.60–0.91) (0.58–0.86) (0.23–0.96) (0.14–0.85) (0.48–0.89) (0.25–0.86) (0.40–0.85)

Nmax (years)

Period

LAR (%)

77 122 120 79 83 271 318 260 172 348 261 380 177 252

1928–2004 1880–2001 1883–2003 1926–2004 1919–2001 1731–2001 1685–2002 1744–2003 1831–2002 1655–2002 1742–2002 1625–2004 1827–2003 1751–2002

0 0 0 0 0 0 0.099 0 0.195 0.162 0.110 0.187 0 –

MW, mean ring-width; SD, standard deviation; MS, mean sensitivity; A1, first-order autocorrelation; Nmax, maximum number of rings counted on a single core; Period, period covered by at least one sample; LAR, percentage of locally absent rings. *Mean values with range in parentheses. Journal of Biogeography 34, 1873–1892 ª 2007 The Authors. Journal compilation ª 2007 Blackwell Publishing Ltd

1879

A. Di Filippo et al. Table 2(b) Summary of the raw and of the prewhitened site chronologies. Prewhitened chronology

Raw chronology

Entire length

Site

Code Trees Cores Period

SD MW (mm (mm Length year)1) year)1) MS A1



65 92 101 75 68 205 289 200 153 327 179 379 176 250

19 9 13 12 5 10 20 23 12 21 20 24 14 11

19 9 13 13 5 10 20 23 12 27 20 28 28 21

1940–2004 1910–2001 1903–2003 1930–2004 1934–2001 1797–2001 1714–2002 1804–2003 1850–2002 1676–2002 1824–2002 1626–2004 1828–2003 1753–2002

2.99 2.00 2.64 2.64 2.93 1.25 1.15 1.08 1.47 1.16 1.57 0.85 1.23 0.57

1.66 1.01 1.08 1.34 1.06 0.57 0.39 0.41 0.42 0.65 0.78 0.36 0.34 0.22

0.21 0.17 0.18 0.16 0.19 0.22 0.16 0.16 0.17 0.19 0.23 0.18 0.21 0.28

0.93 0.92 0.86 0.92 0.85 0.85 0.83 0.90 0.70 0.83 0.83 0.87 0.58 0.58

Common period (1942– 2001)

Entire length

SD MW (mm (mm year)1) year)1) MS A1

p MW SD

MS A1

3.50 2.84 3.47 3.08 3.45 1.53 1.25 1.41 1.72 1.58 2.08 1.42 1.23 0.57

2 2 2 1 3 4 2 1 1 3 1 5 2 3

0.19 0.15 0.18 0.16 0.16 0.13 0.10 0.09 0.12 0.11 0.14 0.10 0.15 0.13

1.47 0.60 0.69 1.20 0.66 0.36 0.29 0.28 0.42 0.46 0.41 0.28 0.28 0.17

0.16 0.14 0.13 0.14 0.14 0.16 0.13 0.12 0.16 0.13 0.19 0.14 0.22 0.25

0.88 0.65 0.62 0.83 0.58 0.64 0.70 0.66 0.60 0.75 0.32 0.51 0.29 0.49

1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

0.16 0.13 0.16 0.17 0.14 0.11 0.09 0.07 0.12 0.10 0.13 0.09 0.13 0.12

)0.08 )0.03 )0.02 )0.04 )0.03 0.00 )0.01 )0.07 0.00 )0.01 0.03 )0.01 )0.02 0.00

Period, period covered by three or more samples; Length, number of years included in Period; MW, mean ring width; SD, standard deviation; MS, mean sensitivity; A1, first-order autocorrelation; p, autoregressive model order.

prewhitening varied between 1 and 5 (Table 2b), and was positively correlated with chronology length (Table 3). Mountain (three sites) and high-mountain (five sites) chronologies had EPS values >0.85. Of the six low-elevation sites, two had EPS > 0.85, and the other four chronologies approached the EPS value in the most recent decades (see Fig. S1). Given the overall high correlation between all tree-ring chronologies (see Table S1), and the importance of our tree-ring network for its ‘rear edge’ location (Hampe & Petit, 2005) between Eurosiberian and Mediterranean regions, no chro-

nology was discarded based on its EPS value. Latitude and altitude were directly correlated with EPS (Table 3), suggesting a stronger climatic control on tree growth moving upward and northward along the network. The highest correlation was found between EPS and stand age (Table 3), suggesting that oldgrowth forests are characterized by greater synchronicity of radial increment, perhaps because they have also escaped human impacts for decades. For composite chronologies (see Table S4), EPS values exceeded the 0.85 threshold during periods used to quantify climate–tree growth relationships (Fig. 5).

Figure 4 Relationship between maximum age at each sampled forest and average elevation. ‘B’ indicates ‘boschi banditi’, forest stands historically protected from logging.

1880


Bioclimatology of beech in the Eastern Alps Table 3 Correlation of dendrochronological parameters with stand geographical location and age. Bold: P < 0.01; italics: P < 0.05.

Latitude Longitude Altitude Length Nmax

MW

SD

MS

A1

p

LAR

EPS

)0.63 )0.11 )0.68 )0.79 )0.79

)0.47 )0.13 )0.67 )0.68 )0.72

0.81 0.63 0.25 0.01 )0.03

)0.62 )0.49 )0.45 )0.23 )0.26

0.01 )0.05 0.42 0.54 0.47

)0.01 )0.41 0.65 0.69 0.65

0.50 0.09 0.60 0.69 0.72

MW, SD, MS, A1 and EPS calculated on the period 1942–2001; MW, SD, MS and A1 were computed for the raw site chronologies; see Table 2b. p, order of autoregressive model used to obtain the prewhitened chronologies; see Table 2b. LAR: see Table 2. EPS: see Figure S1. Latitude, longitude and elevation: see Table 1; Length and Nmax: see Table 2.

Figure 5 Expressed population signal (EPS) statistics for the (a) low-elevation, (b) mountain and (c) high-mountain composite chronologies (see Table S4 in Supplementary Material for a summary of these chronologies). Running EPS values were calculated over the entire length of the chronology using a 50-year window with a 40-year overlap. The dotted line is the 0.85 threshold (Wigley et al., 1984); the light grey background shows the period studied using moving correlation functions; the grey background indicates the years used to compute correlation and response functions.

Bioclimatic units and growth–climate relationships Multivariate analysis was conducted on the 14 prewhitened site chronologies. Hierarchical cluster analysis (HCA) produced a dendrogram where three main clusters were identified (Fig. 6a). The first cluster is made of beech stands growing between 200 and 800 m a.s.l. on the Julian Alps, named ‘Low Elevation Julian Alps’. In the second cluster there are stands growing between 800 and 1200 m a.s.l. on the Carnic Alps, so

that it could be named ‘Mountain Carnic Alps’. In this context, it is important to notice that TOB and GRA, although having the same elevation and exposure, belong to different bioclimatic units. The last cluster consists of chronologies developed above 1300 m a.s.l., reaching the higher edge of the beech altitudinal range, and are thus referred to as ‘High Mountain’. TRE (near Tarvisio), although located at a relatively low elevation, belongs to the ‘‘High Mountain’’ group, possibly due to its exposure to cold northerly winds, low winter


1881

A. Di Filippo et al. (b)

Loading

(a)

PC1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 200 400

2

R = 0.43

600

800

1000

1200

1400

Altitude (m a.s.l.)

Loading

(c)

1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 -1.0

PC2 R2 = 0.82

200

400

600

800

1000

1200

1400

Altitude (m a.s.l.)

Figure 6 (a) Dendrogram showing results from the hierarchical cluster analysis made on the 14 prewhitened site chronologies (see Table 2b). In parenthesis are reported the elevation and the alpine range of each site (see Table 1). Triangle = high-mountain forest; circle = mountain forest; square = low-elevation forest. Loadings of Italian and Slovenian prewhitened site chronologies on (b) the first and (c) the second principal component are plotted as a function of site elevation.

temperatures and generally cool summers (Mennella, 1967; Trewartha, 1968). Mountain and high-mountain stands could be further combined into a single group. The northernmost site, DSW (Fig. 1 and Table 1), located at 650 m a.s.l., had low affinity with the rest of the network and remained isolated from the previous clusters. This site is located at the beginning of the Continental Environmental Zone (Metzger et al., 2005), and this may explain its uniqueness. Based on these results, three sites were selected from each elevation range, so that the high-mountain composite chronology incorporated all samples from LAT, PAU, TOL; the mountain composite chronology was developed using TIM, CLE and GRA; and the lowelevation composite chronology consisted of MOT, NIM and TOA. The first two principal components of the 14 prewhitened chronologies explained 50% of the total variance, and were retained for describing the beech network (see Fig. S2). Mountain and high-mountain chronologies have the highest loadings on the first principal component, which explains 33% of the total variance. The second principal component, which accounts for 17% of the total variance, is mostly related to the low-elevation chronologies (see Fig. S2). Considering the uniqueness of the Austrian site DSW, and the fact that gridded climatic data covered the location of the Italian and Slovenian sites, climate–tree growth relationships were analyzed separately for the Austrian sites (see Tables S2 and S3). The first two principal components of the 12 Italian and Slovenian chronologies explained 55% of the total variance, and were retained for describing the beech network (Fig. 6b,c). The first component (PC1, 35.4% of the total variance) was mainly related to mountain and high-mountain chronologies; lowelevation stands were more related to the second component (PC2, 19.3% of the total variance). Groups similar to those 1882

identified by HCA were recognized by plotting the first and second principal component loadings as a function of elevation (Fig. 6b,c). Correlation and response function analysis of principal component scores (Table 4) showed that radial growth in both mountain and high-mountain stands (PC1) was related to May precipitation (negatively), September temperature (positively), January minimum temperature (positively) and previous July temperature (negatively). Radial growth in low-elevation beech stands (PC2) was positively related to precipitation and negatively to temperature in the months of May, July and August; wood formation therefore appears to be droughtlimited during the growing season. Previous September and October minimum temperatures were also negatively correlated with low-elevation beech stands, while previous July and November maximum temperatures appeared positively correlated. Correlation and response functions calculated for the period 1942–2001 using composite chronologies (see Table S4) showed that growth in the high-mountain forests was positively correlated with May temperature and negatively correlated with precipitation (Table 5), most likely because the growing season at this elevation begins at the end of May (Dittmar & Elling, 2005). July and August temperatures were positively correlated with the high-elevation composite, whereas previous July temperature was inversely correlated. No climatic variable for the growing season was significantly correlated with the growth of mountain Carnic Alp stands, so that the most important relationship was that with previous summer temperatures. Thus, mountain and high-mountain chronologies, which were similar in terms of principal component loadings, could be separated based on an elevation-related climatic signal. The low-elevation composite was


Bioclimatology of beech in the Eastern Alps Table 4 Bootstrap (a) correlation and (b) response function coefficients calculated for the period 1942–2001 between the scores of the first two PCs of the 12 prewhitened chronologies from Italy and Slovenia, and the gridded climatic data covering the same area, over a 17-month window. Coefficients with P < 0.05 are in bold. Year preceding growth Jun (a) PC1 P Tmax Tmin PC2 P Tmax Tmin (b) PC1 P Tmax P Tmin PC2 P Tmax P Tmin

Year of growth

Jul

Aug

Sep

Oct

Nov

Dec

Jan

Feb

0.14 0.10 0.09

0.18 )0.33 )0.39

0.12 )0.11 )0.16

0.14 )0.12 )0.04

0.06 )0.08 0.10

0.00 0.06 )0.03

0.06 0.12 0.11

0.00 0.22 0.23

)0.23 0.20 0.12

0.01 )0.07 )0.06

0.07 )0.24 )0.18

)0.02 0.09 0.09

)0.19 0.25 0.26

0.25 0.01 0.24

0.15 )0.23 )0.17

)0.19 0.11 0.07

)0.17 )0.07 )0.14

0.12 0.05 0.14 0.03

0.00 )0.20 0.01 )0.28

0.05 )0.04 0.11 )0.06

0.12 )0.03 0.08 0.03

)0.02 0.04 )0.04 0.08

0.01 0.04 )0.03 )0.03

0.03 0.04 0.03 0.03

)0.02 )0.07 )0.02 )0.05

0.02 )0.14 0.05 )0.11

0.15 )0.02 0.11 0.00

)0.08 0.13 )0.12 0.13

0.16 0.02 0.12 0.10

0.03 )0.25 0.01 )0.19

)0.09 0.10 )0.09 0.09

Mar

Apr

May

Jun

0.04 0.03 0.03

0.06 )0.14 )0.17

)0.27 0.12 0.15

)0.06 )0.02 0.08

)0.06 )0.03 )0.02

0.11 0.14 0.23

0.09 )0.07 )0.05

)0.22 0.34 0.35

0.00 0.19 0.03 0.23

)0.05 0.09 )0.07 0.08

0.09 )0.05 0.12 )0.04

)0.02 )0.10 0.00 )0.09

)0.02 0.01 )0.04 )0.07

0.10 )0.05 0.06 0.04

0.12 0.16 0.07 0.13

)0.05 )0.09 0.02 )0.04

Jul

Aug

Sep

Oct

0.06 0.00 0.04

)0.11 0.21 0.22

)0.06 0.27 0.30

0.11 )0.01 0.10

)0.16 0.18 0.21

)0.40 0.33 0.31

)0.25 0.37 0.36

0.04 0.13 0.04

)0.12 0.11 )0.05

)0.16 0.07 )0.15 0.11

0.01 )0.01 0.03 0.04

0.09 0.02 0.13 0.09

0.04 0.09 0.04 0.13

0.02 0.26 )0.02 0.26

0.01 )0.01 0.06 0.03

)0.12 0.17 )0.11 0.18

)0.08 0.06 )0.09 0.07

)0.28 0.06 )0.24 0.06

)0.10 0.16 )0.11 0.10

0.02 )0.07 0.08 )0.11

)0.06 0.01 )0.08 )0.05

P, monthly precipitation; Tmax, monthly maximum temperature; Tmin, monthly minimum temperature.

correlated positively with June–July precipitation and negatively with June–July temperature, pointing to summer drought as the main climate signal. A positive correlation with precipitation and a negative one with temperature also appeared for the previous September. A positive correlation with January minimum temperatures, which was reported for PC1 (Table 4), was also found for low-elevation and mountain composites (Table 5). Climate–growth relationships for the Austrian sites, DSW and HSA (see Tables S2 and S3), reaffirmed their somewhat separate bioclimatic classification. This separation could be the cause for the low correlations between monthly precipitation recorded at the two meteorological stations in Austria and averaged over the two grid cells for Italy and Slovenia (see Fig. S3). Moving correlation function analysis was carried out on mountain and high-mountain composite chronologies to investigate the temporal stability of their main climatic signals (Fig. 7). For mountain stands, the correlation in growth with July maximum temperature during the previous year has a progressive negative trend during the last 150 years (Fig. 7a), while the correlation with January minimum temperature appears to increase in recent years (Fig. 7b). With respect to high-mountain sites, the inverse correlations with May (Fig. 7c) and August precipitation (data not shown) have emerged only in recent times. The correlation of growth with August temperature has fluctuated through time (Fig. 7d), with a tendency to higher values in the most recent years, when correlation with July

and September temperature also becomes significant (data not shown). Teleconnection analysis Teleconnection analysis of PCA scores served to evaluate the geographic extent of climatic signals (Fig. 8). Correlations between PC1 scores and other beech chronologies spread mainly westward up to the Central Alps. It was less strong but still significant eastward up to the Dinaric Mountains and the hills of south-eastern Slovenia, and southward up to the Northern Apennines (Fig. 8a). There was no significant correlation with any beech chronologies further south, with the only exception of Monte Cimino in Latium (r = 0.50, P < 0.001). This might be due to the singularity of this site, characterized by fertile volcanic soils, frequent fog and direct exposure to northerly cold winds. Of the northernmost sites DSW and HSA, which were excluded from the PC-based dendroclimatic analysis, the former showed a less strong correlation with PC1 scores (r = 0.32, P < 0.05) than the latter (r = 0.47, P < 0.001). PC2 scores correlated eastward to the Dinaric Mountains and south-eastern Slovenia, and southward to low elevation stands in Central Italy (Latium) (Fig. 8b). In that region, teleconnections could extend upward to 1050 m a.s.l., where beech can grow in association with holm oak (Quercus ilex L.) (Bernabei et al., 1996). No other Italian chronologies from stands growing at higher elevations in the Alpine region, the peninsula, or Sicily (Castorina et al., 2005),


1883

A. Di Filippo et al. Table 5 Bootstrap correlation (a) and response (b) function coefficients calculated for the period 1942–2001 between the three composite chronologies (see Table S4 in Supplementary Material), and the gridded climatic data covering the same area, over a 17-month window. Coefficients with P < 0.05 are in bold. Year preceding growth Jun

Jul

(a) Low elevation stands P 0.04 0.06 Tmax 0.11 0.04 Tmin 0.12 )0.03 Mountain stands P 0.22 0.17 Tmax )0.04 )0.45 Tmin 0.01 )0.45 High mountain stands P )0.02 0.19 0.18 )0.26 Tmax Tmin 0.14 )0.31 (b) Low elevation stands P 0.07 0.01 Tmax 0.08 0.02 P 0.07 0.01 0.05 )0.03 Tmin Mountain stands P 0.15 )0.02 Tmax )0.03 )0.33 P 0.19 )0.02 Tmin )0.02 )0.38 High mountain stands P )0.03 0.08 Tmax 0.05 )0.16 P )0.02 0.08 Tmin 0.04 )0.23

Year of growth

Aug

Sep

Oct

Nov

Dec

0.12 )0.14 )0.18

0.26 )0.25 )0.20

)0.16 )0.06 )0.16

)0.21 0.27 0.15

0.19 0.02 0.03

0.21 )0.12 )0.15

0.12 )0.14 )0.07

0.14 )0.03 0.15

)0.13 )0.01 )0.08

)0.03 0.00 )0.04

0.02 0.16 0.22

0.16 )0.08 0.15

)0.06 )0.03 )0.01 )0.06

0.12 )0.11 0.14 )0.07

0.16 0.01 0.19 0.01 0.03 )0.08 0.04 )0.10

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

0.13 0.17 0.23

)0.11 0.12 0.03

)0.03 )0.11 )0.16

)0.12 0.00 )0.08

0.08 )0.22 )0.19

0.28 )0.26 )0.20

0.27 )0.26 )0.21

0.15 )0.17 )0.12

)0.08 0.06 0.15

0.17 )0.11 0.08

0.02 0.13 0.13

0.15 0.26 0.23

)0.21 0.12 0.04

)0.02 0.03 )0.01

0.07 )0.10 )0.16

)0.18 0.04 0.10

0.01 )0.08 0.06

0.01 )0.09 )0.04

)0.02 0.15 0.17

)0.12 0.17 0.19

0.07 )0.02 0.03

0.17 )0.01 )0.12

0.00 0.16 0.15

)0.16 0.16 0.13

)0.21 0.26 0.15

0.19 0.08 0.14

0.09 )0.08 )0.06

)0.40 0.28 0.32

)0.16 0.13 0.17

)0.10 0.26 0.28

)0.27 0.41 0.42

)0.02 0.25 0.25

0.09 )0.04 0.08

)0.14 )0.03 )0.11 )0.08

)0.07 0.24 )0.07 0.15

0.06 )0.05 0.07 )0.06

0.02 0.07 0.04 0.16

)0.13 0.08 )0.11 0.00

)0.04 )0.14 0.02 )0.09

)0.05 0.02 )0.09 )0.04

0.03 )0.07 0.05 )0.06

0.17 )0.10 0.22 )0.07

0.20 )0.04 0.19 0.00

0.09 )0.05 0.11 0.02

0.02 0.18 )0.05 0.23

0.10 0.00 0.14 0.08

0.06 0.00 0.03 0.02

0.09 0.08 0.05 0.12

)0.07 )0.05 )0.12 )0.08

)0.04 0.08 )0.04 0.05

0.17 0.21 0.16 0.21

)0.05 0.00 )0.05 0.05

0.06 )0.06 0.09 )0.07

)0.01 )0.01 0.00 )0.05

)0.13 0.07 )0.11 0.09

0.04 )0.03 0.07 0.04

0.05 )0.03 0.09 0.03

0.08 0.13 0.09 0.15

)0.04 0.15 )0.06 0.16

0.04 )0.04 0.08 0.00

0.09 0.14 0.01 0.19

0.03 0.02 )0.01 0.04

0.05 )0.08 0.02 )0.18

0.02 0.05 0.03 0.08

)0.10 0.16 )0.07 0.16

0.02 0.16 )0.02 0.14

0.22 0.01 0.21 0.04

0.01 )0.11 0.04 )0.04

)0.21 0.11 )0.21 0.14

)0.05 0.02 )0.06 0.05

0.00 0.13 0.04 0.20

0.01 0.12 )0.01 0.15

)0.02 0.18 )0.02 0.18

)0.01 )0.06 0.04 0.01

P, monthly precipitation; Tmax, monthly maximum temperature; Tmin, monthly minimum temperature.

were correlated with PC2 scores. There was also no correlation between PC2 and the chronologies from the two Austrian sites, DSW and HSA. Correlation between PC1 and PC2 scores for the network presented here with PC1 and PC2 scores of the central Italian network (Piovesan et al., 2005a) indicated that the strongest teleconnection is found between low-elevation beech chronologies (r = 0.48, P < 0.001, PC2 Eastern Alps vs. PC2 Latium–Abruzzi). DISCUSSION Several trees almost four centuries old were discovered in this Alpine region. These are heretofore the oldest beech stands scientifically dated in the Alps (Biondi & Visani, 1996; Piutti & Cescatti, 1997; Dittmar et al., 2003). Because beech longevity can reach and exceed 500 years in the Apennines (Piovesan et al., 2005b) and the Pyrenees (Bourquin-Mignot & Girardclos, 2001), beech trees even older than those reported here might occur in the Alps. Carnia yielded the oldest trees of the network within the so-called ‘boschi banditi’, where the 1884

protection of these forests over the past few centuries has allowed the conservation of old individuals, reaching a maximum age at breast height of 380 years at the Lateis site (Table 2a). Limits imposed on local harvesting or industrial logging, for instance by creating nature reserves, have greatly contributed to the survival of old-growth beech stands (Piovesan et al., 2005b). Maximum tree age increased with altitude, as was also found for other beech forests in the rest of Italy (Piovesan et al., 2005a). Lower growth rates, which were observed at the upper limit of the beech altitudinal range, may be responsible for the increased longevity of these trees, possibly because of enhanced wood durability or reduced maintenance and repair costs (Pen˜uelas, 2005). Mean radial increment and its standard deviation decreased when stand age increased. Older trees are often found in stands where shade-tolerant species can undergo long periods of suppression before canopy attainment (Canham, 1990). In particular, European beech can survive several decades in a suppressed status (Piovesan et al., 2005b), a period during


Bioclimatology of beech in the Eastern Alps

Figure 7 Moving correlation function coefficients calculated between selected climatic variables and composite mountain (a, b) and high-mountain (c, d) beech chronologies. Period: 1803–2001; moving window: 70 years. See Table S4 in Supplementary Material for a summary of the composite chronologies.

which growth is reduced to a minimum and locally absent rings are more frequent in the lower stem of the tree (Lorimer et al., 1999). (a)

Geographic patterns characterized beech radial growth in the Eastern Alps, as was previously reported for Apennine forests (Piovesan et al., 2005a). Latitude was negatively (b)

Figure 8 Correlation map (teleconnections) between first (a) and second (b) principal component scores of 12 prewhitened chronologies with Italian, Austrian and Slovenian beech site chronologies. Symbols relate to bioclimatic position as in Fig. 6, and were based on HCA of different data sets (see Fig. 6a in this study, Fig. 2 in Piovesan et al., 2005a; and Fig. 3 in Di Filippo, 2006). Dimension of each symbol is proportional to correlation value; in solid grey are correlations with P-value < 0.01. Calculations were made on the period 1942–1988. Journal of Biogeography 34, 1873–1892 ª 2007 The Authors. Journal compilation ª 2007 Blackwell Publishing Ltd

1885

A. Di Filippo et al. correlated with mean ring-width, and was positively correlated with mean sensitivity. Proximity to the Adriatic Sea most likely generates a milder climate in the Julian Alps, and these influences seemed to reach the south-facing Carnic valleys as well. Relatively warmer temperatures, together with the exceptional high levels of precipitation in this southern range of the Eastern Alps (>2000–3000 mm year)1, Desiato et al., 2005), create favourable conditions for beech growth, as shown by the low mean sensitivity of these chronologies. The higher mean sensitivity of the Austrian chronologies can be explained by a progressive increase in continentality with latitude. In particular, a shift from a Mediterranean to a Continental Environmental Zone, passing through the Alpine one, can be observed in our network going from south to north (Metzger et al., 2005). Altitude exerts an even stronger control on mean ring width, with values increasing three to four times from high to low elevation sites. This pattern is frequently observed along altitudinal gradients (Monserud & Sterba, 1996; Piovesan et al., 2005a). It can be explained considering that growing season length and ecosystem productivity are closely linked (e.g., White et al., 1999). Dittmar & Elling (2005), observed that in Bavaria, beech growing season decreases by about 2– 3 days for every 100-m increase in elevation. Moreover, beech growth may be negatively impacted by frequent late-frost damage at higher elevations (Dittmar et al., 2006). Multivariate analysis applied to tree-ring series has been demonstrated to detect the effects of environmental gradients on the growth of forest species (e.g. for Europe: Biondi & Visani, 1996; Ma¨kinen et al., 2002; Dittmar et al., 2003; Linderholm et al., 2003; Tardif et al., 2003; Frank & Esper, 2005; Piovesan et al., 2005a). Despite the elevational range of this study, most beech forests in the network showed a common climatic signal, characterized by the importance of conditions at the beginning and the end of the growing period, the risk of winter cold damage and the one-year lag effect of floral induction. Statistical relationships pointed to a negative effect of May precipitation, which could be due to beech sensitivity to soil saturation (Nielsen & Jørgensen, 2003), especially considering the high precipitation levels at the study areas (Desiato et al., 2005) and the fact that snow melt typically occurs in late spring (Mennella, 1967). In addition, cloudiness during May is another factor that could reduce growth through light limitation (e.g. Graham et al., 2003). During September, warmer temperatures can contribute to latewood cell wall thickening (Lebourgeois et al., 2005). Before the beginning of the growing season, minimum temperature in January could be related to tree growth because extremely low winter temperature can cause cold damage, such as freezing embolism (Lemoine et al., 1999). The highest correlations were found with previous July temperature, which may be related to floral induction, stimulated by a hot and dry summer in the year preceding masting (Piovesan & Adams, 2001; Schmidt, 2006). This signal was more pronounced in mountain and high-mountain stands, and less so at lower elevations. Previous summer 1886

conditions were most important for mountain chronologies, while the climate of the growing season gains significance at higher elevations (Dittmar & Elling, 1999). This beech response to previous summer conditions has been found throughout Europe, e.g. in the Pyrenees (Gutierrez, 1988; Dittmar et al., 2003), Cantabria (Rozas, 2001), Apennines (Piovesan & Schirone, 2000), French hills (Lebourgeois et al., 2005) and central Europe (Dittmar et al., 2003). Its widespread occurrence suggests a connection with climate controls on physiological processes involved in resource accumulation (e.g. starch reserves) and bud development (with particular reference to leaf primordia) and differentiation of flower buds (see Nakawatase & Peterson, 2006). Thus, the number of differentiated flower buds should influence the amount of photosynthates assigned to reproduction, rather than to growth, during the following season. Beech mast years are known to have a negative effect on wood production, and are sometimes responsible for the formation of ‘pointer years’ (see Figs 19.112 and 19.114 in Schweingruber, 1996; Piovesan & Bernabei, 1997). Studies that have employed a long-term data set have usually found trade-offs between radial growth, masting and climate (Woodward et al., 1994; Piovesan & Bernabei, 1997; Sela˚s et al., 2002; Monks & Kelly, 2006). Relationships between reproduction and growth in plants are not easily detected (Ban˜uelos & Obeso, 2004). Recently, Monks & Kelly (2006) identified such relationships in Nothofagus, which has a reproductive behaviour very similar to Fagus (Richardson et al., 2005). Other authors had already reported a negative relationship between growth and previous summer temperature in Nothofagus menziesii without providing an explanation for it (Cullen et al., 2001). On the other hand, Fagus crenata does not always show a tree-ring response to mast years (Yasumura et al., 2006). Mechanistic explanations will most likely need to focus on the partitioning of carbon allocation between reproductive and vegetative pathways (see Hoch, 2005; Yasumura et al., 2006). In our study, moving correlations suggested an increased importance of this interaction during recent times, possibly due to the combination of climatic warming (Schmidt, 2006) and the progressive aging and development of sampled trees. Close to the upper edge of its altitudinal range, beech benefits from higher temperature and suffers from excessive precipitation in late spring (May) and during summer. The same behaviour was observed in central European highelevation beech populations (830–1240 m a.s.l.) (Dittmar et al., 2003), and can be related to a thermal limit imposed by altitude, plus the associated negative effect of cloudiness and soil saturation. A similar temperature response during the growing season is typical of conifer species growing at high elevations in the Alps (Frank & Esper, 2005) and Pyrenees (Tardif et al., 2003), or at high latitudes in Europe (e.g., Ma¨kinen et al., 2002; Linderholm et al., 2003). Moving correlations showed a transient response of the main climatic signals, and similar behaviour was recently reported in a dendroclimatic study of Larix decidua in the same region (Carrer & Urbinati, 2006).


Bioclimatology of beech in the Eastern Alps A word of caution is needed on the interpretation of MCF. In fact, it is simplistic to consider these changes in moving correlations as evidence of deviation from the ‘principle of uniformitarianism’ (Camardi, 1999), which assumes that modern natural processes have acted similarly in the past, and is equivalent to the statistical assumption of ‘stationarity’. First, because results are based on empirical relationships, changes in data quality (both dendrochronological and instrumental) can be responsible for different correlation values over time. Second, the ‘uniformity principle’ commonly refers to the fact that limiting factors controlled tree-ring parameters in the past just as they do today, but the role of different factors at a single location or over an entire region could change over time. This possibility has been raised, for example, to explain the ‘divergence’ between temperature and ring parameters (width and maximum latewood density) during the late 20th century (Jacoby & D’Arrigo, 1995; Briffa et al., 1998). In Alaska, recent increases in air temperature are not reflected in tree-ring thickness because water (that is, drought stress) has become the limiting factor (Barber et al., 2000; Lloyd & Fastie, 2002; Wilmking & Juday, 2005). In Siberia, on the other hand, reduced correlation of growth rates with summer temperature has been attributed to increasing winter precipitation, which leads to delayed snowmelt in permafrost environments, thus shortening the tree growing season (Vaganov et al., 1999). The second PC of the 12 site chronologies in the Eastern Alps showed contrasting growth patterns for low- and high-mountain sites. Dendroclimatic results pointed to an environmental gradient of increasing summer drought with low-elevation sites negatively influenced and high-elevation ones impacted positively (Table 5). Reversing tree-ring responses to a certain climatic factor along an altitudinal gradient were previously reported for European beech in the central Apennines (Piovesan et al., 2005a), for Douglas fir and mountain hemlock in western North America (Fagre et al., 2003; Zhang & Hebda, 2004), and for birch in Japan (Takahashi et al., 2005). Environmental differences associated with elevation gradient, such as a decrease in temperature and in the length of the growing season with increase in elevation, may lead to such spatial variation in radial growth. With future warming, we hypothesize a reduction of growth at the lower elevations, whereas high-mountain beech productivity could increase due to a milder and longer vegetative period (see also Nakawatase & Peterson, 2006). At the lower elevations, reduced temperatures in previous summer–early autumn can favour hardening-related processes that reduce the risk of early frost damage, whereas mildness in November could enhance the storage of reserves for the next growing season (Barbaroux & Bre´da, 2002). Teleconnection analysis revealed that the Eastern Alps climatic signal is still found in the Central Alps, spreads northward in Austria, eastward to Slovenia, and southward to the northern Apennines. The northern Apennines occupy an intermediate position, as beech chronologies from that area are correlated with both Alpine and central-southern Apen-

nine chronologies (Piovesan et al., 2005a). This mountain range is placed in a transition zone between the Mediterranean and Temperate climate (e.g. Walter, 1985; Bailey, 1996). From a climatic point of view, the distinction between the Apennines and the Alps is mainly a difference in precipitation regime (Trewartha, 1968), as was confirmed by a recent climatic zonation of Italy (Brunetti et al., 2006). As a possible consequence, beech growth in central-southern Italy is characterized by a dominant response to summer drought (Piovesan et al., 2005a), while in the Alps the response shifts towards thermal factors (Di Filippo, 2006). Considering the autoecology and biogeography of European beech, it is reasonable to expect higher exposure to drought stress toward the southern limit of beech geographic range (Becker, 1981), and an increased importance of temperature moving northward and upward (Dittmar & Elling, 2005; Fang & Lechowicz, 2006). Evidence for the spatial separation between these two bioclimatic zones emerged even in a study of the climatic patterns influencing conifer tree rings across the Northern Hemisphere (Fig. 7 in Briffa et al., 2002). Further confirmation can be found in the bioclimatic classification of Italy proposed by Pignatti (1979) according to vegetation analysis. Low elevation chronologies were correlated through central and northern Italy and in Slovenia, suggesting the existence of common climatic factors controlling growth at these altitudes. Late spring–summer drought emerged as important for tree-ring formation in both central Italy (Piovesan et al., 2005a) and low elevation stands in the Julian Alps. From the teleconnection analysis, a similar influence extended to the Dinaric Mountains and south-east Slovenian hills. In particular, May climate seems to be important at hilly sites, possibly because higher temperature at the beginning of the growing season allows for early crown development, exposing beech to water stress during the summer (Piovesan et al., 2005a). This expansion of the Mediterranean region up to the Pre-Alps is consistent with placing the Julian Alp sites at the boundary between the Mediterranean Mountains and the Alpine Environmental Zones (Metzger et al., 2005; see also Zohary, 1973). Thus, TOB (Julian Alps) and GRA (Carnic Alps), even though located at the same elevation on a southern exposure, belong to, respectively, the low elevation and the mountain range. Additional evidence of Mediterranean influences at these sites comes from the presence upon rocky cliffs of the evergreen holm oak (Quercus ilex L.), and from the cultivation of the olive tree (Olea europea L.). These two species reach the northernmost limit of distribution (the former) and cultivation (the latter) in north-eastern Italy (Pignatti, 1998). In low-elevation Julian stands, the response to drought appears combined with one to early-frost damage. A mixed response to summer drought stress and previous autumn–winter temperature was found in hilly beech treering chronologies of central and eastern Europe (Dittmar et al., 2003). In French beech sites, such responses are generally observed below 600 m a.s.l. (Lebourgeois et al., 2005).


1887

A. Di Filippo et al. Genetic research has shown a different origin of the Apennine and Alpine beech populations (Vettori et al., 2004; Magri et al., 2006). The Apennines and the southern part of the Balkan Peninsula were Mediterranean refuge areas for European beech during the ice ages. These refugia were separated from Central European populations, from which the Alpine Fagus sylvatica stands originated (Magri et al., 2006). Even though beech radial growth is sensitive to summer drought both in the Apennines and at low elevation in the Alps, the Mediterranean populations seem more drought resistant (Nahm et al., 2006). Because summer drought is becoming increasingly important for beech forest dynamics (and eventually for natural selection) in central Europe (Czajkowski et al., 2005) and in the Mediterranean Basin (Jump et al., 2006), it is vital to understand how beech populations with varying genetic material have responded to drought in the past. CONCLUSION The Alps and the Apennines belong to two different bioclimatic zones, both possessing a clear vertical dendroclimatic zonation in altitudinal ranges, each with its own distinctive climatic signal (Dittmar et al., 2003; Piovesan et al., 2005a). In both zones, response to low temperature increases with altitude. However, this phenomenon is more evident in the Alps than the Apennines, probably because at high elevations low temperatures can be limiting even in the middle of the growing season (Dittmar et al., 2003). In the high-mountain Mediterranean environment, temperature response is concentrated in early spring (late-frost damage and/or temperature requirements for growth reactivation). Alpine high-elevation beechwoods have dendroclimatic signals opposite those for low-elevation ones, which correlate with precipitation in May and coolness in summer (Dittmar & Elling, 1999; Dittmar et al., 2003). The cause may be the altitudinal difference in temperature regime and, in particular, an altitude-mediated phenological shift in growing season onset (Dittmar & Elling, 2005). While drought in the Apennines is a limiting factor across all Mediterranean settings, from hilly to tree-line beech stands (Piovesan et al., 2005a), its influence in the Alps remains limited to the low-elevation environments. Thus, despite similar elevational response, summer influence on treering chronologies marks the bioclimatic separation between the Apennines and the Alps. Is this bioclimatic difference stable through time? Moving response functions suggest a recent expansion of the growing season for Alpine high-elevation beechwoods. As for plant phenological changes (e.g. Parmesan & Yohe, 2003), the observed changes can be explained by the positive temperature trend reported for Italy over the last 130 years (Brunetti et al., 2006). This phenomenon is emphasized when considering that the first decades used in moving correlations fall within the Little Ice Age, which ended in central Europe during the second half of the 19th century (Xoplaki et al., 2005). Under a changing climate, bioclimatic 1888

shifts could characterize vegetation arranged along altitudinal gradients or at ecotonal boundaries (e.g. Pen˜uelas & Boada, 2003). The beech old-growth network presented here will therefore be useful to assess any future changes in forest growth related to climate. As dendroecological data sets provide ground-truth information on spatial and temporal variability of climate–forest interactions, they can also benefit the successful implementation of adaptive management strategies. ACKNOWLEDGEMENTS We are grateful to Livio Silverio and to the Forest Service of Friuli–Venezia Giulia for their support during the sampling excursions, and to Anzˇe Rutar for his contribution to beech research at the Tolmin site. FB thanks Stanford University for sabbatical support. KCˇ thanks the Ministry of Education, Science, and Sport of the Republic of Slovenia, Research Program ‘Lesarstvo’, for financial support. The comments of the Handling Editor and two anonymous referees helped in improving on an earlier version of this manuscript. REFERENCES Aniol, R.W. (1983) Tree-ring analysis using CATRAS. Dendrochronologia, 1, 45–53. Aniol, R.W. (1987) A new device for Computer Assisted Measurement of Tree-Ring Widths. Dendrochronologia, 5, 135–141. Bailey, R.G. (1996) Ecosystem geography. Springer Verlag, New York. Ban˜uelos, M.-J. & Obeso, J.-R. (2004) Resource allocation in the dioecious shrub Rhamnus alpinus: the hidden costs of reproduction. Evolutionary Ecology Research, 6, 397–413. Barbaroux, C. & Bre´da, N. (2002) Contrasting distribution and seasonal dynamics of carbohydrate reserves in stem wood of adult ring-porous sessile oak and diffuse porous beech trees. Tree Physiology, 22, 1201–1210. Barber, V.A., Juday, G.P. & Finney, B.P. (2000) Reduced growth of Alaskan white spruce in the twentieth century from temperature-induced drought stress. Nature, 405, 668–673. Becker, M. (1981) Ecologie du hetre et de la hetraie – Caracterisation climatique. Le heˆtre (ed. by E. Teissier du Cros, F. Le Tacon, G. Nepveu, J. Parde`, R. Perrin and J. Timbal), pp. 71–77, INRA, Paris. Bernabei, M., Lo Monaco, A., Piovesan, G. & Romagnoli, M. (1996) Dendrocronologia del faggio (Fagus sylvatica L.) sui Monti Sabini (Rieti). Dendrochronologia, 14, 59–70. Biondi, F. (1992) Development of a tree-ring network for the Italian Peninsula. Tree-Ring Bulletin, 52, 15–29. Biondi, F. (1993) Climatic signals in tree-rings of Fagus sylvatica L. from the central Apennines, Italy. Acta Oecologica, 14, 57–71. Biondi, F. (1997) Evolutionary and moving response functions in dendroclimatology. Dendrochronologia, 15, 139–150.


Bioclimatology of beech in the Eastern Alps Biondi, F. (1999) Comparing tree-ring chronologies and repeated timber inventories as forest monitoring tools. Ecological Applications, 9, 216–227. Biondi, F. (2000) Are climate–tree growth relationships changing in north-central Idaho? Arctic, Antarctic and Alpine Research, 32, 111–116. Biondi, F. & Swetnam, T.W. (1987) Box–Jenkins models of forest interior tree-ring chronologies. Tree-Ring Bulletin, 47, 71–95. Biondi, F. & Visani, S. (1996) Recent developments in the analysis of an Italian tree-ring network with emphasis on European beech (Fagus sylvatica L.). Tree rings, Environment and Humanity (ed. by J.S. Dean, D.M. Meko and T.W. Swetnam), pp. 713–725. Radiocarbon, University of Arizona, Tucson. Biondi, F. & Waikul, K. (2004) DENDROCLIM 2002: a C++ program for statistical calibration of climate signals in treering chronologies. Computers and Geosciences, 30, 303–311. Bourquin-Mignot, C. & Girardclos, O. (2001) Construction d’une longue chronologie de heˆtres au pays basque: la foreˆt d’Iraty et le Petit Age Glaciaire. Revue de Geógraphie des Pyreńeés et du Sud-Ouest Europeén, 11, 59–71. Briffa, K.R., Schweingruber, F.H., Jones, P.D., Osborn, T.J., Shiyatov, S.G. & Vaganov, E.A. (1998) Reduced sensitivity of recent tree-growth to temperature at high northern latitudes. Nature, 391, 678–682. Briffa, K.R., Osborn, T.J., Schweingruber, F.H., Jones, P.D., Shiyatov, S.G. & Vaganov, E.A. (2002) Tree-ring width and density data around the Northern Hemisphere: Part 1, local and regional climate signals. The Holocene, 12, 737–757. Brunetti, M., Maugeri, M., Monti, F. & Nanni, T. (2006) Temperature and precipitation variability in Italy in the last two centuries from homogenised instrumental time series. International Journal of Climatology, 26, 345–381. Camardi, G. (1999) Charles Lyell and the Uniformity Principle. Biology and Philosophy, 14, 537–560. Canham, C.D. (1990) Suppression and release during canopy recruitment in Fagus grandifolia. Bulletin of the Torrey Botanical Club, 117, 1–7. Carrer, M. & Urbinati, C. (2006) Long-term change in the sensitivity of tree-ring growth to climate forcing in Larix decidua. New Phytologist, 170, 861–872. Castorina, R., Di Filippo, A. & Piovesan, G. (2005) Dendroecology of beech (Fagus sylvatica L.) on Mount Etna Natural Park (Sicily, Italy). Abstract book of Eurodendro2005 (ed. by M. Sarlatto, A. Di Filippo, G. Piovesan and M. Romagnoli), pp. 44, Editore Sette Citta`, Viterbo. Cook, E.R. & Peters, K. (1997) Calculating unbiased tree-ring indices for the study of climatic and environmental change. The Holocene, 7, 361–370. Cook, E.R., Glitzenstein, J.S., Krusic, P.J. & Harcombe, P.A. (2001) Identifying functional groups of trees in West Gulf Coast forests (USA): a tree-ring approach. Ecological Applications, 11, 883–903. Cullen, L.E., Palmer, J.G., Duncan, R.P. & Stewart, G.H. (2001) Climate change and tree-ring relationships of Nothofagus

menziesii tree-line forests. Canadian Journal of Forest Research, 31, 1981–1991. Czajkowski, T., Ku¨hling, M. & Bolte, A. (2005) Impact of the 2003 summer drought on growth of beech sapling natural regeneration (Fagus sylvatica L.) in north-eastern Central Europe. Allgemeine Forst- und Jagdzeitung, 176, 133–143 [In German]. Desiato, F., Lena, F., Baffo, F., Suatoni, B. & Toreti, A. (2005) Indicatori del clima in Italia elaborati attraverso il sistema SCIA. APAT, Rome. Di Filippo, A. (2006) Study for a bioclimatic classification of Italian beech forests on a dendroclimatic basis. PhD thesis, Universita` degli Studi della Tuscia, Viterbo (In Italian). Dittmar, C. & Elling, W. (1999) Radial growth of Norway spruce and European beech in relation to weather and altitude. Forstwissenschaftliches Centralblatt, 118, 251–270. Dittmar, C. & Elling, W. (2005) Phenological phases of common beech (Fagus sylvatica L.) and their dependence on region and altitude in Southern Germany. European Journal of Forest Research, 125, 181–188. Dittmar, C., Zech, W. & Elling, W. (2003) Growth variations of common beech (Fagus sylvatica L.) under different climatic and environmental conditions in Europe: a dendroecological study. Forest Ecology and Management, 173, 63–78. Dittmar, C., Fricke, W. & Elling, W. (2006) Impact of late frost events on radial growth of common beech (Fagus sylvatica L.) in Southern Germany. European Journal of Forest Research, 125, 249–259. Douglass, A.E. (1920) Evidence of climatic effects in the annual rings of trees. Ecology, 1, 24–32. Eckstein, D. & Frisse, E. (1982) The influence of temperature and precipitation on vessel area and ring width of oak and beech. Climate from tree rings (ed. by M.K. Hughes, P.M. Kelly, J.R. Pilcher and V.C. LaMarche Jr), p. 12. Cambridge University Press, Cambridge. Efron, B. & Tibshirani, R. (1986) Bootstrap methods for standard errors, confidence intervals, and other measures of statistical accuracy. Statistical Science, 1, 54–75. Ellenberg, H. (1988) Vegetation ecology of Central Europe. Cambridge University Press, Cambridge. Fagre, D.B., Peterson, D.L. & Hessl, A.E. (2003) Taking the pulse of the mountains: ecosystem responses to climatic variability. Climatic Change, 59, 263–282. Fang, J. & Lechowicz, M.J. (2006) Climatic limits for the present distribution of beech (Fagus L.) species in the world. Journal of Biogeography, 33, 1804–1819. Frank, D. & Esper, J. (2005) Characterization and climate response patterns of a high elevation, multi-species tree-ring network in the European Alps. Dendrochronologia, 22, 107–121. Fritts, H.C. (1976) Tree rings and climate. Academic Press, London. Graham, E.A., Mulkey, S.S., Kitajima, K., Phillips, N.G. & Wright, S.J. (2003) Cloud cover limits net CO2 uptake and growth of a rainforest tree during tropical rainy seasons.


1889

A. Di Filippo et al. Proceedings of the National Academy of Sciences USA, 100, 572–576. Grissino-Mayer, H.D. (2001) Evaluating crossdating accuracy: a manual and tutorial for the computer program COFECHA. Tree-Ring Research, 57, 205–221. Guiot, J. (1991) The bootstrapped response function. TreeRing Bulletin, 51, 39–41. Gutierrez, E. (1988) Dendroecological study of Fagus sylvatica L. in the Montseny mountains (Spain). Acta Oecologica, 9, 301–309. Hampe, A. & Petit, R.J. (2005) Conserving biodiversity under climate change: the rear edge matters. Ecology Letters, 8, 461–467. Helama, S., Lindholm, M., Timonen, M. & Eronen, M. (2004) Detection of climate signal in dendrochronological data analysis: a comparison of tree-ring standardization methods. Theoretical and Applied Climatology, 79, 239–254. Hoch, G. (2005) Fruit-bearing branchlets are carbon autonomous in mature broad-leaved temperate forest trees. Plant, Cell and Environment, 28, 651–659. Holmes, R.L. (1983) Computer-assisted quality control in treering dating and measurement. Tree-Ring Bulletin, 43, 69–78. Jacoby, G.C. & D’Arrigo, R.D. (1995) Tree-ring width and density evidence of climatic and potential forest change in Alaska. Global Biogeochemical Cycles, 9, 227–234. Jump, A.S. & Pen˜uelas, J. (2005) Running to stand still: adaptation and the response of plants to rapid climate change. Ecology Letters, 8, 1010–1020. Jump, A.S., Hunt, J.M. & Pen˜uelas, J. (2006) Rapid climate change-related growth decline at the southern range-edge of Fagus sylvatica. Global Change Biology, 12, 2163–2174. Kaiser, H.F. (1992) On Cliff’s formula, the Kaiser–Guttman rule, and the number of factors. Perceptual and Motor Skills, 74, 595–598. Kienast, F., Schweingruber, F.H., Bra¨ker, O.L. & Scha¨r, E. (1987) Tree-ring studies on conifers along ecological gradients and the potential of single-year analyses. Canadian Journal of Forest Research, 17, 683–96. Laurent, J.M., Bar-Hen, A., François, L., Ghislain, M. & Cheddadi, R. (2004) Refining vegetation simulation models: from plant functional types to bioclimatic affinity groups of plants. Journal of Vegetation Science, 15, 739–746. Lebourgeois, F., Bre´da, N., Ulrich, E. & Granier, A. (2005) Climate–tree-growth relationships of European beech (Fagus sylvatica L.) in the French Permanent Plot Network (RENECOFOR). Trees, 19, 385–401. Lemoine, D., Granier, A. & Cochard, H. (1999) Mechanism of freeze induced embolism in Fagus sylvatica L. Trees, 13, 206–210. Linderholm, H.W., Solberg, B.Ø. & Lindholm, M. (2003) Treering records from central Fennoscandia: the relationship between tree growth and climate a west–east transect. The Holocene, 13, 889–897. Lloyd, A.H. & Fastie, C.L. (2002) Spatial and temporal variability in the growth and climate response of treeline trees in Alaska. Climatic Change, 52, 481–509. 1890

Lorimer, C.G., Dahir, S.E. & Singer, M.T. (1999) Frequency of partial and missing rings in Acer saccharum in relation to canopy position and growth rate. Plant Ecology, 143, 189–202. Ludwig, J.A. & Reynolds, J.F. (1988) Statistical ecology. John Wiley and Sons, New York. Magri, D., Vendramin, G.G., Comps, B., Dupanloup, I., Geburek, T., Go¨mo¨ry, D., Latalowa, M., Litt, T., Paule, L., Roure, J.M., Tantau, I., van der Knaap, W.O., Petit, R.J. & de Beaulieu, J.L. (2006) A new scenario for the Quaternary history of European beech populations: palaeobotanical evidence and genetic consequences. New Phytologist, 171, 199– 221. Ma¨kinen, H., No¨jd, P., Kahle, H.-P., Neumann, U. & Tveite, B. (2002) Radial growth variation of Norway spruce (Picea abies (L.) Karst.) across latitudinal and altitudinal gradients in central and northern Europe. Forest Ecology and Management, 171, 243–259. Meko, D.M., Cook, E.R., Stahle, D.W., Stockton, C.W. & Hughes, M. K. (1993) Spatial patterns of tree-growth anomalies in the United States and southeastern Canada. Journal of Climate, 6, 1773–1786. Mennella, C. (1967) Il clima d’Italia. Fratelli Conti Editori, Naples. Metzger, M.J., Bunce, R.G.H., Jongman, R.H.G., Mu¨cher, C.A. & Watkins, J.W. (2005) A climatic stratification of the environment of Europe. Global Ecology and Biogeography, 14, 549–563. Monks, A. & Kelly, D. (2006) Testing the resource-matching hypothesis in the mast seeding tree Nothofagus truncata. Austral Ecology, 31, 366–375. Monserud, R.A. & Sterba, H. (1996) A basal area increment model for individual trees growing in even- and unevenaged forest stands in Austria. Forest Ecology and Management, 80, 57–80. Nahm, M., Radoglou, K., Halyvopoulos, G., Geßler, A., Rennenberg, H. & Fotelli, M.N. (2006) Physiological performance of beech (Fagus sylvatica L.) at its Southeastern distribution limit in Europe: seasonal changes in nitrogen, carbon and water balance. Plant Biology, 8, 52–63. Nakawatase, J.M. & Peterson, D.L. (2006) Spatial variability in forest growth–climate relationships in the Olympic Mountains, Washington. Canadian Journal of Forest Research, 36, 77–91. Nielsen, C.N. & Jørgensen, F.V. (2003) Phenology and diameter increment in seedlings of European beech (Fagus sylvatica L.) as affected by different soil water contents: variation between and within provenances. Forest Ecology and Management, 174, 233–249. Paiero, P., Candidi-Tommasi, R. & Caniglia, G. (1975) Il bosco ‘bandito’ di Cleulis (Paluzza): fustaia di faggio derivata dall’invecchiamento naturale di un ceduo matricinati. Monti e Boschi, 4, 1–13. Parmesan, C & Yohe, G. (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature, 421, 37–42. Pen˜uelas, J. (2005) A big issue for trees. Nature, 437, 965–966.


Bioclimatology of beech in the Eastern Alps Pen˜uelas, J. & Boada, M. (2003) A global change-induced biome shift in the Montseny Mountains (NE Spain). Global Change Biology, 9, 131–140. Pignatti, S. (1979) I piani di vegetazione in Italia. Giornale Botanico Italiano, 113, 411–428. Pignatti, S. (1998) I boschi d’Italia. UTET, Turin. Piovesan, G. & Adams, J.M. (2001) Masting behaviour in beech: linking reproduction and climatic variation. Canadian Journal of Botany, 79, 1039–1047. Piovesan, G. & Bernabei, M. (1997) L’influenza delle precipitazioni estive sulla crescita e la riproduzione del faggio (Fagus sylvatica L.) in una stazione meridionale dell’areale. Italia Forestale e Montana, 6, 444–459. Piovesan, G. & Schirone, B. (2000) Winter North Atlantic Oscillation effects on the tree rings of the Italian beech (Fagus sylvatica L.). International Journal of Biometeorology, 44, 121–127. Piovesan, G., Bernabei, M., Di Filippo, A., Romagnoli, M. & Schirone, B. (2003) A long-term tree ring beech chronology from a high-elevation old-growth forest of Central Italy. Dendrochronologia, 21, 1–10. Piovesan, G., Biondi, F., Bernabei, M., Di Filippo, A. & Schirone, B. (2005a) Spatial and altitudinal bioclimatic zones of the Italian Peninsula identified from a beech (Fagus sylvatica L.) tree-ring network. Acta Oecologica, 27, 197–210. Piovesan, G., Di Filippo, A., Alessandrini, A., Biondi, F. & Schirone, B. (2005b) Structure, dynamics and dendroecology of an old-growth Fagus forest in the Apennines. Journal of Vegetation Science, 16, 13–28. Piutti, E. & Cescatti, A. (1997) A quantitative analysis of the interactions between climatic response and intraspecific competition in European beech. Canadian Journal of Forest Research, 27, 277–284. Richardson, S.J., Allen, R.B., Whitehead, D., Carswell, F.E., Ruscoe, W.A. & Platt, K.H. (2005) Climate and net carbon availability determine temporal patterns of seed production by Nothofagus. Ecology, 86, 972–981. Rinn, F. (1996) TSAP: Reference Manual (Version 3.0). Frank Rinn, Heidelberg. Rolland, C. (2002) Decreasing teleconnections with inter-site distance in monthly climatic data and tree-ring width networks in a mountainous Alpine area. Theoretical and Applied Climatology, 71, 63–75. Rozas, V. (2001) Detecting the impact of climate and disturbances on tree-rings of Fagus sylvatica L. and Quercus robur L. in a lowland forest in Cantabria, Northern Spain. Annals of Forest Science, 58, 237–251. Schmidt, W. (2006) Temporal variation in beech masting (Fagus sylvatica L.) in a limestone beech forest (1981–2004). Allgemeine Forst- und Jagdzeitung, 177, 9–19 (In German). Schweingruber, F.H. (1996) Tree rings and environment dendroecology. Swiss Federal Institute of Forest, Snow and Landscape Researches, WSL/FNP, Birmensdorf. Sela˚s, V., Piovesan, G., Adams, J.M. & Bernabei, M. (2002) Climatic factors controlling reproduction and growth of Norway spruce in southern Norway. Canadian Journal of Forest Research, 32, 217–225.

Stenson, H. & Wilkinson, L. (2004) Factor analysis. systat 11 Statistics I (ed. by systat Software, Inc.), pp. 359–377, Richmond, CA. Stokes, M.A. & Smiley, T.L. (1996) An introduction to tree-ring dating. University of Arizona Press, Tucson. Takahashi, K., Tokumitsu, Y. & Yasue, K. (2005) Climatic factors affecting the tree-ring width of Betula ermanii at the timberline on Mount Norikura, central Japan. Ecological Research, 20, 445–451. Tardif, J., Camarero, J.J., Ribas, M. & Gutierrez, E. (2003) Spatiotemporal variability in tree growth in the Central Pyrenees: climatic and site influences. Ecological Monographs, 73, 241–257. Thompson, R.S., Shafer, S.L., Anderson, K.H., Strickland, L.E., Pelltier, R.T., Bartlein, P.J. & Kerwin, M.W. (2005) Topographic, bioclimatic, and vegetation characteristics of three ecoregion classification systems in North America: comparisons along continent-wide transects. Environmental Management, 34, S125–S148. Trewartha, G.T. (1968) An introduction to climate. McGrawHill, New York. Urbinati, C., Carrer, M. & Sodiro, S. (1997) Dendroclimatic response variability of Pinus cembra L. in upper timberline forests of Italian eastern Alps. Dendrochronologia, 15, 101–17. Vaganov, E.A., Hughes, M.K., Kirdyanov, A.V., Schweingruber, F.H. & Silkin, P.P. (1999) Influence of snowfall and melt timing on tree growth in subarctic Eurasia. Nature, 400, 149–151. Vettori, C., Vendramin, G.G., Anzidei, M., Pastorelli, R., Paffetti, D. & Giannini, R. (2004) Geographic distribution of chloroplast of beech (Fagus sylvatica L.). Theoretical and Applied Genetics, 109, 1–9. von Wuehlisch., G. (2006) EUFORGEN Technical Guidelines for genetic conservation and use for European beech (Fagus sylvatica). International Plant Genetic Resources Institute, Rome, Italy. http://www.ipgri.cgiar.org/networks/euforgen/ Euf_Distribution_Maps.asp Walter, H. (1985) Vegetation of the Earth and ecological systems of the geobiosphere. Springer-Verlag, Berlin. Wessel, P. & Smith, W.H.F. (1998) New, improved version of Generic Mapping Tools released. EOS Transactions of the American Geophysical Union, 79, 579. White, M.A., Running, S.W. & Thornton, P.E. (1999) The impact of growing-season length variability on carbon assimilation and evapotranspiration over 88 years in the eastern US deciduous forest. International Journal of Biometeorology, 42, 139–145. Whittaker, R.J., Arau´jo, M.B., Jepson, P., Ladle, R.J., Watson, J.E.M. & Willis, K.J. (2005) Conservation biogeography: assessment and prospect. Diversity and Distributions, 11, 3–23. Wigley, T.M.L., Briffa, K.R. & Jones, P.D. (1984) On the average value of correlated time series, with applications in dendroclimatology and hydrometeorology. Journal of Applied Meteorology, 23, 201–213.


1891

A. Di Filippo et al. Wilmking, M. & Juday, G.P. (2005) Longitudinal variation of radial growth at Alaska’s northern treeline: recent changes and possible scenarios for the 21st century. Global and Planetary Change, 47, 282–300. Woodward, F.I. (1987) Climate and plant distribution. Cambridge University Press, Cambridge. Woodward, A., Silsbee, D.G., Schreiner, E.G. & Means, J.E. (1994) Influence of climate on radial growth and cone production in subalpine fir (Abies lasiocarpa) and mountain hemlock (Tsuga mertensiana). Canadian Journal of Forest Research, 24, 1133–1143. Xoplaki, E., Luterbacher, J., Paeth, H., Dietrich, D., Steiner, N., Grosjean, M. & Wanner, H. (2005) European spring and autumn temperature variability and change of extremes over the last half millennium. Geophysical Research Letters, 32, L15713. Yasumura, Y., Hikosaka, K. & Hirose, T. (2006) Resource allocation to vegetative and reproductive growth in relation to mast seeding in Fagus crenata. Forest Ecology and Management, 229, 228–233. Zhang, Q.-B. & Hebda, R.J. (2004) Variation in radial growth patterns of Pseudotsuga menziesii on the central coast of British Columbia, Canada. Canadian Journal of Forest Research, 34, 1946–1954. Zohary, M. (1973) Geobotanical foundations of the Middle East. Fischer Verlag, Stuttgart. SUPPLEMENTARY MATERIAL The following supplementary material is available for this article online: Table S1 Correlation matrix for the 14 prewhitened beech site chronologies (period 1942–2001). Table S2 Bootstrap correlation (a) and response (b) function coefficients calculated between the Dunkelsteinerwald prewhitened site chronology and the 17-month climatic window for the period 1942–2001. Table S3 Bootstrap correlation (a) and response (b) function coefficients calculated between the Hallstatt prewhitened site chronology and the 17-month climatic window for the period 1942–2001. Table S4 Summary of composite tree-ring chronologies, which were developed using three sites each.

1892

Figure S1 Expressed population signal (EPS) statistics for the 14 prewhitened site chronologies arranged according to their length. Figure S2 Eigenvectors (loadings) of the first two components for the 14 prewhitened site chronologies (a) Eigenvalues (percentage of explained variance) of the first two components (b, total; c, PC1; d, PC2). Figure S3 Correlation between climatic data used in dendroclimatic analyses. This material is available as part of the online article from: http://www.blackwell-synergy.com/doi/abs/10.1111/j. 1365-2699.2007.01747.x Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

BIOSKETCHES Alfredo Di Filippo conducted this research as a part of his PhD thesis. He is currently working at the Universita` della Tuscia (Italy) as a research assistant, conducting investigations in the fields of dendroclimatic networks and dendroecology applied to the study of old-growth forests and disturbance regimes. Franco Biondi directs the DendroLab at the University of Nevada, Reno (USA). His experience and interests are in dendrochronology, climate change at multiple spatial and temporal scales, numerical analysis of proxy climate data, and late-Holocene dynamics in forests, woodlands and mountains. Gianluca Piovesan is professor of Dendrology and Ecological Forest-Landscape Management at the Universita` della Tuscia (Italy). His research interests include old-growth forest, masting, dendroecology, forest response to climatic variability and restoration of forest ecosystems.

Editor: Mark Bush
