Scientifica Acta 1, No.

1, 51 53 (2007)

Experimental Ecology and Geobotany

Spatio-temporal analysis of tree-ring growth variability in
high-altitude forests of the Alps in relation to climate change
Maurizia Gandini
Dipartimento di Ecologia del Territorio, Universit di Pavia, Via S. Epifanio, 14, 27100 Pavia Italy
mauriziagandini@gmail.com

1 Introduction
Since 1980, several studies have underlined the importance of the influence of Global Warming on biological systems and especially on plants. This phenomenon turned out to be the main factor responsible
for growth variation (with a positive trend) and tree mortality [13]. Recent research suggests that there is
no common pattern of tree growth between the different sites sampled in Europe, and thus emphasizes the
need to study climate-growth relationship in different areas [4].
During the 1990s, the scientific community shifted the attention to another important question: Is Climate Change really modifying plant response to climate? And if so, how?
With this study we try to make a significant contribution towards understanding these problems for
high-altitude forests in the Alps, generally known as suitable for dendroclimatic research [5].

2 Methods
2.1 Tree-ring and climatic data
This study focused on two tree species, commonly used for dendroclimatic investigations, Picea abies (L.)
Karst. and Pinus cembra L, in the subalpine forest and at timberline.
The Database of Dendroecology Laboratory of the University of Pavia kindly provided dendrochronologic data, collected in every part of the Alps. They gather various chronologies from other institutions,
with whom a permanent collaboration has been established (the University of Padova, the University of
Torino and the University of Marseille).
Site chronologies were built by taking two increment cores for approximately 15 of the largest and
healthier trees from each investigated area; cores were then fixed to wooden supports and prepared using
razor blades or by sanding until optimal surface resolution for measurements was achieved. Ring width was
measured to within 0.01 mm under a microscope at 1640 x magnification using the Computer Controlled
Tree Ring Device [6]. Cross-dating, which ensures that the correct year is assigned to each corresponding annual ring, was accomplished by visually checking the curves and calculating t-values relating to
coefficient of correlation [7] and gleichlaufigkeit [8]. All the ring width data used in the dendroclimatic
analysis were standardized by ARSTAN program (ARS40CWIN version), in order to remove age-trend,
stand dynamics and long-term variations.
Temperature and precipitation data were kindly provided by R. Bhm [9]. Instrumental data from a
wide number of stations located in the Alpine region were homogenised, tested with spatial and temporal
analysis and then gridded on a 1 x 1 degree resolution grid as anomalies [10, 11]. The grid shows 112
grid-points covering an area from 4 E to 19 E longitude and from 43 N to 49 N latitude.
As for temperature data sets, two kinds of grid are available: a low-elevation grid (below 1400 m a.s.l.)
which covers a wide area with 112 grid-points, and a high-elevation grid (above 1400 m a.s.l.) with 30
grid-points.
c 2007 Universit degli Studi di Pavia

52

Scientifica Acta 1, No. 1 (2007)

bootstrapped correlation

4.0
3.0
2.0
1.0
0.0
-1.0
-2.0

T Sep

T Jul

T Aug

T Jun

T Apr

T May

T Mar

T Jan

T Feb

T Dec

T Oct

T Nov

P Sep

P Jul

P Aug

P Jun

P Apr

P May

P Mar

P Jan

P Feb

P Dec

P Oct

-4.0

P Nov

-3.0

monthly climatic variables (P = precipitations; T = temperatures)

Fig. 1: Bootsptrapped correlation functions for Swiss stone pine western regional chronology; statistically significant
correlation values represented by filled bars (threshold at the thick lines, 2).

2.2 Climate-growth investigations

The research project was set up following a spatio-temporal point of view: on one side we studied forest
dynamics, in order to observe forests evolution in time, and to understand the climate impact on their
growth. On the other side, for each species, we studied tree response to temperature and precipitation
along a spatial gradient. The resulting climate-growth relationships were then tested for their temporal
stability.
We structured this research using two approaches:
1. Regional Analysis: dendroclimatic analysis was performed by bootstrapped response and correlation
functions [12, 13] for the period 18301995, on Norway spruce (14 sites) and Swiss stone pine (31 sites)
mean site chronologies, selected along a longitudinal gradient on the Alps and on regional chronologies divided in 3 groups: Western, Central and Eastern chronologies for both species. The resulting relationships
were then investigated by means of moving interval bootstrapped correlation functions in order to test their
temporal stability [13].
2. Population Analysis: this study has been carried out in two different areas of the Alps, Paneveggio
(Trentine) and Bormio (Lombardy), where we set permanent plots (100 x 100 m) along an altitudinal
gradient, from subalpine forest to treeline. In these sites we examined forest dynamics by means of tree
age structure. In Paneveggio mean site chronologies were established for each plot and then used to analyze
climate-growth relationships and to test their temporal stability during the interval 18301995.

3 Results
Our findings from dendroclimatic analysis confirm summers temperatures importance for both species: we
found positive, highly significant and stable climate-growth relationships between June and July temperatures and Swiss stone pine and Norway spruce, respectively. In Fig. 1 we display the results of bootstrapped
correlation functions for Swiss stone pine western regional chronology: in this example we could notice
the relevance of June temperatures and winter precipitations.
Despite of the quiescence period, winter season is essential for trees growth. Heavy precipitations
(for Swiss stone pine) and mild temperatures (for Norway spruce) display positive significant correlation
with growth. Furthermore the influence of winter climatic parameters on trees growth seems to increase
together with longitude, becoming more evident on the Eastern Alps.
Temporal stability has been tested for the most significant findings by specific analysis. Only few parameters were found statistically significant for the entire period: most of change, gain or loss of importance,
seems to have happened during 1900-1950 interval (value 1950 on x axis). For example, the relationship
c 2007 Universit degli Studi di Pavia

Scientifica Acta 1, No. 1 (2007)

53

6.0
bootstrapped correlation

west
5.0

central
east

4.0
3.0
2.0
1.0

1990

1980

1970

1960

1950

1940

1930

1920

1910

1900

1890

1880

0.0

years

Fig. 2: Bootstrapped correlation functions on moving intervals (window span of 50 years) between December precipitations and Norway spruce regional chronologies. The years represented on x axis correspond to the last years of each
50-years intervals used in the analysis. Statistical significance has been tested at 95% (0.05 level), (threshold at the
thick lines, 2).

between Norway spruce and winter precipitations (especially December) becomes more and more relevant
for the Western sites, while it loses significance for those in the Central and Eastern Alps (Fig. 2).
These last findings on response-to-climate variability might be the evidence of decreasing trees sensitivity to temperatures and precipitations, already pointed out in literature (Briffa, 1998; Wilmking et al.
2004; Wilson & Elling, 2004 and Carrer & Urbinati, 2006).
Finally, the temporal viewpoint has been investigated also observing stand dynamics: we found that
since the second half of the XX century young plants started to colonize higher sites, with a remarkable
increase in the number of individuals during 1980s and 1990s especially for Swiss stone pine. The interpretation of those results is though very complex because theyre probably due to several elements interaction,
including anthropogenic factors.
Acknowledgements I am grateful to Paola Nola for the precious advices and the contribution she gave to this research. I am also grateful to Marco Carrer and Jean Louis Edouard for kindly providing dendrochronological data, and
Reinhard Bhm for meteorological data from the ALP-IMP project.

References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]

[11]
[12]
[13]

M. Becker, Canadian Journal of Forest Research, 19, 1110 (1989).

J. L. Innes, Environmental Pollution, 83, 237 (1994).
H. Makinen, P. Nojd, K. Mielikainen, Trees, 15, 177 (2001).
D.Frank, J. Esper, Dendrochronologia, 22, 107 (2005).
F. H. Schweingruber, Tree Rings and Environment. Dendroecology (Paul Haupt Verlag, Berne, 1996), p. 609.
R. W.Aniol, Dendrochronologia, 5, 135 (1987).
M. J. L. Baillie, J. R. Pilcher, Tree-Ring Bulletin, 33, 7 (1973).
F. H. Schweingruber, Tree rings. Basics and applications for Dendrochronology (Kluwer Academic Publishers,
Dordrecht, Boston, London, 1988).
ALP-IMP, Multi-centennial climate variability in the Alps based on instrumental data, model simulation and
proxy data, 2004. http://www.zang.ac.at/alp-imp/main.htm
I. Auer, R. Bohm, A. Jurkovic, A. Orlik, R. Potzmann, W. Schoner, M. Ungersbock, M. Brunetti, T. Nanni, M.
Maugeri, K. R. Briffa, P. Jones, D. Efthymiadis, O. Mestre, J. M. Moisselin, M. Begert, R. Brazdil, O. Bochnicek,
T. Cegnar, M. Gajic Capkaj, K. Zaninovic, Z. Majstorovic, S. Szalai, T. Szentimrey, L. Mercalli, International
Journal of Climatology, 25, 139 (2005).
R. Bohm, I. Auer, M. Brunetti, M. Maugeri, T. Nanni, W. Schoner, International Journal of Climatology, 21,
1779 (2001).
J. Guiot, C. Goeury, Dendrochronologia, 14, 295 (1996).
F. Biondi, K. Waikul, Computers and Geosciences, 30, 303 (2004).
c 2007 Universit degli Studi di Pavia