ArticlePDF Available

Abstract and Figures

Under predicted climate change, native silver fir (Abies alba) and European beech (Fagus sylvatica) are the most likely replacement species for the Norway spruce (Picea abies) monocultures planted across large parts of continental Europe. Our current understanding of the adaptation potential of fir-beech mixed forests to climate change is limited because long-term responses of the two species to environmental changes have not yet been comprehensively quantified. We compiled and analysed tree-ring width (TRW) series from 2855 dominant, co-dominant, sub-dominant and suppressed fir and beech trees sampled in 17 managed and unmanaged mixed beech-fir forest sites across Continental Europe, including Bosnia and Herzegovina, Germany, Italy, Romania and Slovakia. Dendroecological techniques that combine various detrending methods were used to investigate variation in radial growth of co-occurring fir and beech trees. Coincidental with peak SO2 emissions, the growth of silver fir declined between 1950 and 1980 at most sites, whereas beech growth increased during this period. Correspondent to a significant warming trend from 1990–2010, average beech growth declined, but silver fir growth increased. Long-term growth patterns and growth-climate sensitivity of fir and beech trees did not significantly differ between managed and unmanaged forests. Multi-decadal changes in the growth rate of all vertical tree classes were similar. In contrast to previous indications of limited drought susceptibility of beech mixed stands, this study suggests that the mixture of tree species in forest stands does not necessarily prevent growth depressions induced by long-term environmental change. Our results further imply that forest management does not necessarily alter their sensitivity to environmental changes.
Content may be subject to copyright.
Contrasting effects of environmental change on the radial growth of
co-occurring beech and r trees across Europe
Michal Bosela
a,b,
,MartinLukac
b,c
, Daniele Castagneri
d
, Róbert Sedmák
a,b
, Peter Biber
e
, Marco Carrer
d
,
Bohdan Konôpka
f,b
, Paola Nola
g
,ThomasA.Nagel
h,b
, Ionel Popa
i,j
, Catalin Constantin Roibu
k
,
Miroslav Svoboda
b
, Volodymyr Trotsiuk
b
, Ulf Büntgen
l,m,n,o
a
Faculty of Forestry, Technical University in Zvolen, Zvolen, Slovakia
b
Faculty of Forestry and Wood Sciences, Czech University of Life Sciences Prague, Czech Republic
c
School of Agriculture, Policy and Development, University of Reading, Reading RG6 6AR, UK
d
Department TeSAF, Università degli Studi di Padova, Padua, Italy
e
Forest Growth and Yield Science, Technical University of Munich (TUM), Freising, Germany
f
National Forest Centre, Forest Research Institute, Zvolen, Slovakia
g
Department of Earth and Environmental Sciences, Università degli Studi di Pavia, Pavia, Italy
h
Biotechnical Faculty, Universityof Ljubljana, Slovenia
i
National Research and Development Institute for Silviculture, Forest Research Station for Norway Spruce Silviculture, Campulung Moldovenesc, Romania
j
INCE - Mountain Economy Center CE-MONT Vatra Dornei, Romania
k
Forest Biometrics Laboratory, Stefan cel MareUniversity of Suceava, Romania
l
Department of Geography, University of Cambridge, CB2 3EN, UK
m
Swiss Federal Research Institute WSL, Zürcherstrasse 111, 8903 Birmensdorf, Switzerland
n
CzechGlobe, Global Change Research Institute CAS, Brno, Czech Republic
o
Masaryk University, Kotřská 2, 61137 Brno, Czech Republic
HIGHLIGHTS
European r and beech growth acceler-
ated during the last century.
Beech growth declined in northern
Europe since 2000.
Fir growth rates increased over most of
Europe since 1980.
Growth-climate responses were similar
for most tree social classes.
Climate sensitivity of both species was
not affected by forest management.
GRAPHICAL ABSTRACT
abstractarticle info
Article history:
Received 11 August 2017
Received in revised form 9 September 2017
Accepted 10 September 2017
Available online 19 October 2017
Editor: Elena Paoletti
Under predicted climate change, native silver r(Abies alba)andEuropeanbeech(Fagus sylvatica) are the most
likely replacement species for the Norway spruce (Picea abies) monocultures planted across large parts of conti-
nental Europe. Our current understanding of the adaptation potential of r-beech mixed forests to climate
change is limited because long-term responses of the two species to environmental changes have not yet been
comprehensively quantied. We compiled and analysed tree-ring width (TRW) series from 2855 dominant,
co-dominant, sub-dominant and suppressed r and beech trees sampled in 17 managed and unmanaged
mixed beech-r forest sites across Continental Europe, including Bosnia and Herzegovina, Germany, Italy,
Science of the Total Environment 615 (2018) 14601469
Corresponding author at: T.G. Masaryka 24, 960 53 Zvolen, Slovakia.
E-mail address: michal.bosela@tuzvo.sk (M. Bosela).
https://doi.org/10.1016/j.scitotenv.2017.09.092
0048-9697/© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Contents lists available at ScienceDirect
Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv
Romania and Slovakia. Dendroecological techniques that combine various detrending methods were used to in-
vestigate variation in radial growth of co-occurring r and beechtrees. Coincidental with peak SO
2
emissions, the
growth of silver r declined between 1950 and 1980 at most sites, whereas beech growth increased during this
period.Correspondent to a signicant warming trend from 19902010, average beech growth declined, but silver
r growthincreased. Long-term growth patterns andgrowth-climate sensitivityof r and beech trees did not sig-
nicantly differ between managed and unmanaged forests. Multi-decadal changes in the growth rate of all ver-
tical tree classes were similar. In contrast to previous indications of limited drought susceptibility of beech mixed
stands, this study suggests that the mixture of tree species in forest stands does not necessarily prevent growth
depressions inducedby long-term environmental change. Our results further imply thatforest managementdoes
not necessarily alter their sensitivity to environmental changes.
© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Keywords:
Dendroecology
Climate change
Growth sensitivity
Mixed forests
Plantclimate interactions
Tree rings
1. Introduction
Beech (Fagus sylvatica L.) is Europe's most abundant forest species
(Ellenberg, 1996). Naturally distributed across most of continental
Europe, it either grows in pure broad leaved forests or in mixtures
with conifer species. Beech has been increasingly used to convert coni-
fer monocultures to mixed stands, reecting the current transition of
forest management strategies to accommodate non-productive forest
functions and adapt to climate change (Knoke et al., 2008; Tarp et al.,
2000). Silver r(Abies alba Mill.) is a coniferous species native to
Europe, with a geographical distribution similar to that of beech, but
largely limited to the Alpine and the Carpathian arcs. There are indica-
tions that silver r might be more suitable to future European climate
as it grew well under warmer-than-present conditions during the
mid-Holocene (Tinner et al., 2013; Ruosch et al., 2016). Today, mixed
r-beech forests represent an important forest ecosystem forming an
essential part of central and south-eastern European landscapes (EEA,
2006).
Since the mid-19th century or even earlier, a large proportion of
European beech-r forests was converted to conifer monocultures of
Norway spruce (Picea abies L. Karst.) (Spiecker et al., 2004). However,
considerable areas of Norway spruce forests in central Europe suffered
from acid deposition during the second half of the 20th century
(Ulrich, 1995). Today, in many locations it is clear that Norway spruce
is becoming increasingly susceptible to the more frequent summer
droughts induced by climate change (Lévesque et al., 2013; Zang et al.,
2014), as well as the devastating effects of severe windstorms and sub-
sequent bark beetle outbreaks (Hlásny and Turčáni, 2013; Jönsson et al.,
2009). Therefore, the introduction of appropriate replacement species,
such as beech and r, has become a key task (IPCC, 2014).
Despite recent publications describing the growth of silver r and
European beech (Bosela et al., 2016b; Büntgen et al., 2014; Cavin and
Jump, 2016; Dittmar et al., 2003; Gazol et al., 2015; Pretzsch et al.,
2014), relatively little is known about how these species react to envi-
ronmental changes when growing in mixed stands (Vitali et al., 2017).
Recent evidence shows that beech (Bosela et al., 2015), but also rto
some extent (Toïgo et al., 2015), may reach higher stem growth produc-
tivity in mixed stands. There are suggestions that growth sensitivity of
silver r and European beech to summer droughts decreases when
growing in mixed stands (Lebourgeois et al., 2013; Metz et al., 2016;
Vitali et al., 2017), but this has neither been tested nor conrmed
under varying natural conditions across different parts of Europe. A
higher diversity of tree species has been shown to potentially stimulate
radial stem growth by better niche utilisation, but also via improved re-
sistance and resilience at a forest ecosystem level (Gazol et al., 2016;
Jucker et al., 2014; Metz et al., 2016; Paquette and Messier, 2011;
Ruiz-Benito et al., 2014; Vitali et al., 2017). Nevertheless, we still lack
unequivocal evidence supporting the notion that growing in mixed
stands mitigates some of the negative long-term effects of rising tem-
peratures, often associated with increased frequency and/or severity
of droughts. A detailed understanding of the relationship between
diversity and ecosystem productivity and stability is indeed crucial for
advising the policy-forming processes at national and international
levels.
To further complicate the picture, most dendroecological data that
describe growth-climate relationships in European forests originate
from western Europe and tend to consider only dominant and co-
dominant trees (Büntgen et al., 2014; Cavin and Jump, 2016;
Nehrbass-Ahles et al., 2014; Pretzsch et al., 2014). Yet, it is possible
that populations of the same species inhabiting the eastern part of
their distributional range possess different sensitivity levels to environ-
mental factors due to genetic variation (Bosela et al., 2016a).
A Europe-wide investigation of species-specicgrowthdynamicsin
mixed forest stands affected by contrasting environmental factors is so
far lacking. To ll this gap, we compiled a database of tree-ring width
(TRW) samples from managed and unmanaged European mixed
beech-r forests covering most of the regions where these two species
co-occur. Our aim was to test the following hypotheses: (H1) do radial
growth patterns of beech and r in mixed forests vary geographically,
(H2) does tree status within the canopy or (H3) forest management in-
terventions affect long-term variation in annual growth, and nally
(H4) was the growth behaviour of co-occurring beech and rsimilarly
affected by the 19701990 period of heavy pollution and the warming
during recent decades?
2. Material and methods
2.1. Tree-ring sampling
A network of 17 beech-r mixed forest stands located in ve
European countries was used to compile 2855 core samples (Fig. 1).
Taking in Slovakia, Romania, Bosnia & Herzegovina, Italy and
Germany, the network covers a large part of Europe's natural distribu-
tion range of mixed beech-r forests. The sites were allocated to 8 re-
gional groups, reecting differences in post-glacial recolonization from
different forest refugia, but also current climatic conditions (Tables 1
and S1). The followingselection criteria were followed atall sites to en-
sure reasonable comparability of observations: 1) growing conditions
were characterised by mesic soils and mean annual precipitation in ex-
cess of 800 mm year
1
, 2) no forest management interventions were
carried out for at least 30 years prior to sampling, 3) specimens of
both species were present in all four social classes, as described by
Oliver & Larson (1996; dominant, co-dominant, sub-dominant and sup-
pressed), 4) all trees above a registration threshold within a xed-area
plot were marked and then either all or a randomly selected subset
were cored. A single location in Slovakia where managed beech-r
stands are found in close vicinity to unmanaged stands was used to
investigate the impact of forest management on growth-climate
sensitivity.
1461M. Bosela et al. / Science of the Total Environment 615 (2018) 14601469
2.2. Tree-ring standardisation
Four standardisation techniques were applied toremove age-related
trends from raw TRW measurements, all aim to preserve the effects of
environmental signals at inter-annual to multi-decadal timescales. The
following detrending methods were used and compared in this study;
a) Modied Exponential Function (MEF, Fritts, 2001). The MEF method
was rst applied to individual tree TRW series, followed by the cal-
culation of ring-width indices (RWI) dened as the ratio between
raw measurements and corresponding MEF function values. Mean
RWI chronologies for each site were calculated as bi-weight robust
means of all individual tree series (Cook and Kairiukstis, 2013)
b) cubic Smoothing Spline with 50% frequency response cut-off at
100 years (SS; Cook, 1985). This setting was used to preserve the
inter-annual to multi-decadal growth uctuations (Büntgen et al.,
2008). The SS standardisation was applied in the same manner as
the MEF approach, the only variation was the use of a different
function.
c) Regional Curve Standardisation (RCS; Briffa and Melvin, 2011). RCS
detrending was applied to individual tree data, but only after the ex-
clusion of all partial-length TRW series where it was not possible to
estimate the number of rings from the beginning of the core to the
pith (Table 1). For example, this includes data series where the be-
ginning of the core was too far from the pith and no ring arc was vis-
ible (Briffa and Melvin, 2011). Raw TRW series were aligned by
cambial age (ring number from bark to pith), followed by mean
TRW calculation for each series. A smoothing spline with a 50% fre-
quency response cut-off at 10% maximum cambial age curve wave-
length (i.e. the Regional Curve (RC)) was then tted to mean TRW
series. RWIs were calculated as ratios between individual series
and the smoothed RC. Individual RWI series were re-aligned by
calendar dates. The nal site chronology was then developed by
using bi-weight robust means.
d) Korf growth function (Korf; Korf, 1939). Multi-decadal growth
changes may be underestimated by common detrending methods
in even-aged forests (Briffa and Melvin, 2011). To counter this ten-
dency we applied a growth function developed by Korf (1939).
Raw TRW series were aligned by age andthe mean curve was calcu-
lated. Then, in contrast to the RCS method, regression parameters of
the Korf function were estimated from the rst 50 years of the mean
curve only. An extrapolation to the full length of the curve was then
carried out, assuming that the rst 50 years indicate both growth
culmination and decline. RWI series at each site were then calculat-
ed as the ratio between raw TRW and estimated Korf's function.
Bi-weight robust mean was used to develop a mean site chronology
as in previous methods, and Korf standardisation was used only on
sites where growth culmination was clearly visible.
Most dendrochronological studies report site chronologies devel-
oped from data describing the dominant trees only (Nehrbass-Ahles
et al., 2014), an approach which may not capture the growth history
of the whole stand. To shed light on this issue, at each site we developed
a mean site chronology for both species from a sample of the 15 largest
trees only (hereafter dominant trees) and compared it to that cover-
ing all measured individuals.
2.3. Growth variability and sensitivity assessments
Growth trends describing several predened periods were
compared to identify effects of known environmental factors. The
predened periods, such as the decades characterised by heavy air
Fig. 1. a) Sampling sites used in the study, with corre sponding Climate Research Unit (CRU) climate data points; b) current distribution of Abies alba and c) current distribution of Fagus sylvatica.
(Source: EUFORGEN).
1462 M. Bosela et al. / Science of the Total Environment 615 (2018) 14601469
pollution in Europe (19501990) and the unprecedented climate
warming (19802000, resp. 19902010), were selected on the basis of
published descriptions of major environmental issues and trends
(Bosela et al., 2016a, 2016b; Büntgen et al., 2014; Gazol et al., 2015;
Jump et al., 2006; Linares and Camarero, 2012). Simple linear regression
was used to describe growth trends within selected periods; regression
coefcients denoting the slope were then used to compare tree growth
between the periods and between regional groups of sites. In this com-
parison,we applied generalized additive models using gammfunction
in mgcvR package (Wood, 2011)tolter out inter-annual high
frequency variation and preserve multi-decadal growth trends. The
gammfunction used Generalized Cross Validation (GCV) to estimate
the smoothing parameter. Populations of regression parameters de-
scribing individual site RWI series were assessed for differences be-
tween species, regional groups, time periods and detrending methods
by ANOVA. All populations were tested for normality of distribution
and equality of variance; no conversion of data was necessary.
Bonferroni correction was used in pair-wise comparisons, and differ-
ences were considered signicant at p b0.05.
Monthly temperature means, precipitation totals and drought indi-
ces (scPDSI self-calibrated Palmer Drought Severity Index) were ob-
tained from gridded CRU TS 3.10 database (http://www.cru.uea.ac.uk/
data/) via Climate Explorer (http://climexp.knmi.nl/)(Dai, 2011;
Harris et al., 2014; Mitchell and Jones, 2005). Data from half-degree
lat/lon grid points nearest to each study site were used to analyse the re-
lationship betweengrowth of both species and climate variation at each
site (Fig. 1). Radial growth responses to climate (standardised mean
chronologies) were quantied by Pearson's correlation coefcients
computed over 31-yr moving window segments to investigate tempo-
ral changes in the climate-growth relationships (Büntgen et al., 2010;
Wilson and Elling, 2004). Bootstrapping was applied to calculate 95%
condence intervals of the correlation coefcients using the bootRes
R package (Zang and Biondi, 2013). The signicance of the correlation
was then tested using 95% percentile range method (Dixon, 2002).
Correlation coefcients describing the relationship between site
chronologies and climate variables were analysed by principal compo-
nent analysis (PCA). PCA was performed with the hclustfunction in
the R Stats package (R Development Core Team, 2008). The distance
matrix was computed using the Euclidian measure. The dendextent
package (Galili, 2015) was used to visualise PCA results. Ward Hierar-
chical Clustering was used to interpret growth trend responses to cli-
mate. Findings from the Ward clustering were then compared with
PCA grouping in a 2-dimmensional space.
3. Results
Regional growth trends identied by the four detrending methods
applied in this study show a wide variation of growth between sites,
species and time periods across Europe (Fig. 2), yet a certain amount
of generalisation is possible (Table 2). Beech growth accelerated be-
tween the 1950s and 1980s in the more northernforests, loosely de-
ned as those above 47°N parallel. Growth acceleration in the north
was followed by a slow decline in annual ring width, with the exception
of the Bavarian forest where beech growth accelerated continuously
since about the 1940s (Fig. 2,Table 2).
At the southernsites, mid-century acceleration has slowed down or
even decreased. At the Cansiglio site in northern Italy, we observed the
earliest onset of this decline starting in the 1950s. While forming the
same forest stands in a mixture with beech, r has shown very different
growth patterns during the last century (H1). A decline of r growth
rate was observed between the1950s and 1980s in most locations
apart from the populations in the southern Carpathians, which
Table 1
Brief description of site location, total number of tree-ring width series used per site (N1) and the number of tree-ring width series that intercepted the pith or were sufciently close to
estimatethe number of ringsto the ptith (N2). Timespan is the minimumand maximum calendaryear of the site chronology. B&Hdenotes Bosnia andHerzegovina, Altdenotes altitudein
m a.s.l.
Country Species Locality (abr.) N1 N2 Time span Mountain range Long Lat Alt
B & H Abies Lom 158 158 15832005 Western Dinaric Mts 16.47 44.45 1350
B & H Fagus (bh_lom) 440 440 16252005
B & H Abies Perucica b1 44 43 17862006 Eastern Dinaric Mts 18.71 43.3 1200
B & H Fagus (bh_perb1) 81 75 15952006
B & H Abies Perucica b2 76 69 16632007
B & H Fagus (bh_perb2) 70 63 15092007
B & H Abies Perucica f1 118 109 16862006
B & H Fagus (bh_perf1) 35 27 16142006
B & H Abies Perucica f2 82 71 17022007
B & H Fagus (bh_perf2) 64 60 17032007
Germany Abies Bodenmais 28 18201995 Bavarian Forest 13.1 49.09 800
Germany Fagus (de) 21 18211995
Italy Abies Cansiglio 140 140 19312012 Southern Alps 12.42 46.1 1100
Italy Fagus (it) 205 184 18562012
Romania Abies Botiza 54 25 17742013 Eastern Carpathians 24.09 47.61 1050
Romania Fagus (ro_bo) 67 23 16142013
Romania Abies Sinca 281 184 16652013 Southern Carpathians 25.17 45.67 1140
Romania Fagus (ro_si) 163 101 15562013
Slovakia Abies Polana 22 22 18602010 Western Carpathians 19.57 48.62 760
Slovakia Fagus (sk_p) 58 58 18672010
Slovakia Abies Spis S1 20 20 18962010 20.73 48.77 760
Slovakia Fagus (sk_s1) 19 19 18812010
Slovakia Abies Spis S2 36 36 17942010 20.73 48.77 760
Slovakia Fagus (sk_s2) 33 33 18112010
Slovakia Abies Spis S3 59 59 18202011 20.67 48.79 760
Slovakia Fagus (sk_s3) 14 14 18212011
Slovakia Abies Spis S4 24 24 18482010 20.72 48.76 830
Slovakia Fagus (sk_s4) 26 26 18982010
Slovakia Abies OBR 25 21 18052013 Western Carpathians 19.47 48.88 887
Slovakia Fagus (sk_obr) 146 120 17402014
Slovakia Abies SRA 57 57 17832013 19.11 49.19 1048
Slovakia Fagus (sk_sra) 133 133 17172013
Slovakia Abies SUT 29 28 18142013 19.09 49.18 1029
Slovakia Fagus (sk_sut) 27 24 17612013
1463M. Bosela et al. / Science of the Total Environment 615 (2018) 14601469
exhibited a slight acceleration of growth during this period (Fig. 2). This
period of r decline was followed by a rapid acceleration of growth in
the north, and a steep decline in the south, again with the exception
of the southern Carpathians.
A comparison of growth trends created from dominant trees only or
all trees above a DBH threshold (H2) did not show any effect of canopy
position (Fig. S5). Regression parameters denoting the slope of the t
were not affected by tree social status in any of the time periods
under consideration, nor for beech (p = 0.128 to 0.516) or r(p=
0.336 to 0.990). Similarly, we did not nd any difference in annual
growth between managed stands and old-growth unmanaged forests
in the western Carpathians when comparing mean RWI growth trends
within each of the time periods under consideration (H3, p = 0.063 to
0.441), indicating that factors other than forest management affect
long-term growth trends in beech and r. We made use of the pre-
dened periods of environmental stress to explore whether beech and
r respond differentially to acid deposition (19501980) and accelerat-
ing climate change (19902010). We found a strong interaction be-
tween species and time period (H4; p b0.001) when comparing mean
growth rates in these two periods. Considering all sites used in this
study (Fig. 3a), rRWIy
1
was smaller than thatof beech in the period
19501980 but the growth trends of these two species reversed by
19902010. An interesting observation emerged when considering the
more northern and southern sites separately; the relationship between
Fig. 2. Mean ring-width index (RWI) chronologies of silver r and European beech after modied exponential (exp) and Regional Curve Standardisation (RCS) techniques for aggregate
sites (west_carp_man: managed stands in western Carpathians, west_carp_unm unmanaged forests in western Carpathians, east_carp: eastern Carpathians, south_carp: southern
Carpathians, east_dinaric:eastern Dinaric,west_dinaric,south_alps: southernAlps, BavarianForest). Generalized Additive Model (GAM)was applied to lter outthe inter-annual variation
and preserve multi-decadal changes. The shaded bands denote 95% condence intervals.
Table 2
A comparison of detrending methodsapplied to tree-ringwidth data describing radialgrowth of Europeanbeech (Fagus sylvatica) and Silverr(Abies alba)intwotime
periods characterisedby different environmental conditions. Linear regressionwas tted to data detrended by Modied exponential function(MEF), smoothing spline
(SS), regional curvestandardisation (RCS) andKorf growth function(Korf). Colouredcell backgrounds denotea negative regression trend.Stars denote signicance level
of regression tat:***-b0.001, ** - b0.01, * - b0.05. Empty cells represent a non-signicant t, and a dash is used in cases where regression was not possible to t.
European beech Silver fir
MEF SS RCS Korf MEF SS RCS Korf
Locality
1950-80
1990-10
1950-80
1990-10
1950-80
1990-10
1950-80
1990-10
1950-80
1990-10
1950-80
1990-10
1950-80
1990-10
1950-80
1990-10
de *** - - - - - - *-*** - - - - -
sk_p *** ** *** *** *** *** ** * *** ***
sk_s1 ** ** * **** *** *** *** ** *** ***
sk_s2 *** *** *** *** *** *** *** *** *** *** *** *** *** ** ***
sk_s3 *** *** * *** *** * *** *** *** *** *** ***
sk_s4 *** ** ** ** *** *** *** *** *** *** ***
sk_obr *** ** ** *** ** *** *** ***
sk_sra *** * *** - - ** *** *** *** ** *** ** ***
sk_sut * ***** - - *** *** *** *** ** *** *** ***
ro_bo * ** - - *** *** *** *** ***
ro_si * ** * * *** ****** ** ** **
1464 M. Bosela et al. / Science of the Total Environment 615 (2018) 14601469
environmental factors and the rate of radial growth was not affected by
species in thesouth (p = 0.359; Fig. 3c), but there was a strong effect in
the north (p b0.001, Fig. 3b). There was no difference between
the trends identied by the four standardisation methods in 195080
(p = 0.249) and only the Korf standardisation method differed from
the other three in 19902010 (p = 0.004).
Differential response of r and beech growth to environmental
change was conrmed by a PCA testing as well as by a cluster analysis
for the strength of relationship between previous- and current-year
temperature, precipitation and PDSI and tree growth (Fig. 4). Fir growth
at the southern sites was mainly inuenced by drought (scPDSI),
whereas temperature was the most dominant driver of tree ring
width at the northern sites. There was considerable variation between
individual sites in observed effects of climate on r and beech growth,
however a very similar grouping of tree populations and sites was
achieved by cluster analysis (Fig. 5).
In a further attempt to contrast the growth behaviour of the
two species within the sampled range of sites, we calculated correlation
coefcients between RWI and current-year summer temperature and
drought over the century-long time period (Fig. 6). Fig. S6 shows a
large temporal variation of correlation coefcients over the past centu-
ry. On average, a negative correlation between RWI and drought domi-
nates in the south, while a positive correlation with temperature is
present in the north.
4. Discussion
4.1. European beech
Published literature describing the dendroecology of European
beech at European and regional scales does not paint a clear picture;
there is evidence of (i) either increased or decreased growth rates of
beech in the last two decades in Central Europe (Dittmar et al., 2003;
Pretzsch et al., 2014), and (ii) either increased or decreased radial
growth of beech at the southern edge of distribution (Jump et al.,
2006; Tegel et al., 2014). Our study conrms that beech growth rates
have increased during the period between 1950 and 1980 across
Europe, an observation which is in line with published measurements
(Bosela et al., 2016b; Hlásny et al., 2011; Pretzsch et al., 2014)and
model simulations (Hlásny et al., 2011). However, our results contradict
those of Dittmar et al. (2003), who in a Europe-wide study found no de-
tectable increase of beech growth in Europe, but documented a decline
of the rate of growth at high altitudes in central Europe. Interestingly,
the authors found that high summer temperatures favoured radial
growth at the expense of vertical growth. Existing studies and observa-
tions presented in this paper suggest a positive effect of increasing sum-
mer temperature at higher latitudes or altitudes on beechgrowth. Thus,
given the summer period warming observed in the last century
(Büntgen et al., 2011; Luterbacher, 2004) and predicted warming across
Europe (IPCC), it seems reasonable to expect further acceleration of
beech growth at the northern edge of its distribution. A recent
Europe-wide study documents an increased basal area increment in
the last decades in beech forests in temperate and continental core re-
gions of the species distribution range (Cavin and Jump, 2016). Our
study adds evidence suggesting that the growth decline in the southern
localities started in mid-20th century and continues until today. How-
ever, a recent increase in beech radial increments in some Mediterra-
nean beech-rforests(
Tegel et al., 2014) suggests strong regional
differences, probably related to regional climate or site productivity
(Aertsenet al., 2014; Bosela et al., 2016b), which limit any broad extrap-
olation of our results.
Hacket-Pain et al. (2016) found no clear spatial pattern in the
drought sensitivity of European beech, indicating that the populations
from the southern and northern range edges respond to summer
drought equally. In contrast, we found a strong spatial pattern in the
growth responses to summer temperature and to drought. While radial
growth of the species generally did not respond to summer drought in
central Europe (Germany, Slovakia and Romania), it became highly re-
sponsive in the Balkan Peninsula (Bosnia & Herzegovina, except for
the Lom site). Our study thus supports existing observations showing
that some southern European beech populations are increasingly suffer-
ing from summer drought (Linares and Camarero, 2012; Piovesan et al.,
2008). Not only can we conrm the same trends, but we are also able to
pinpoint the onset and the severity of the decline allowing for investiga-
tion of site-specic reasons for growth modication. For example,
unlike beech trees in Perucica, which were sensitive to drought, the
population in the Lom site in the same climatic zone was non-
responsive to sc-PDSI. Trees at the Lom site were substantially younger
than those at Perucica (Fig. S1), growth plasticity of a younger popula-
tion of beech might thus explain the difference as outlined by the
age-related climate response hypothesis (Carrer and Urbinati, 2004;
Primicia et al., 2015).
4.2. Silver r
Silver r experienced a severe growth decline in Europe during
19701990, driven by sulphur dioxide emissions (Büntgen et al.,
2014), an event often referred to as Europe-wide r dieback (Cramer,
1984; Larsen, 1986; Meyer, 1957). Our study provides tree-ring width
evidence of this event; however, it is not possible to exclude the possi-
bility that lower summer temperatures during this period might have
contributed to the growth depression (Fig.S2). In contrast to other stud-
ies, our investigation shows that the growth of silver r did not decline
in the eastern part of its distribution range during this period (southern
Carpathians and partly eastern Dinaric Mts.). It has been suggested that
the greater genetic diversity of the Balkan populations helps the species
mitigate effects of changing environmental conditions (Bosela et al.,
2016a), and this study supports that suggestion. Increased genetic di-
versity, but also a greater functional diversity of forest stands where it
occurs, have been shown to increase the capacity of silver r to tolerate
drought (Gazol et al., 2016). The latter relationship is indicative of the
need to understand the implications of ecosystem diversity for species
performance and production stability.
Fig. 3. Linear regression trends of silver r and European beech ring-width indices (RWI)
in two distincttime periods of acid deposition (19501980) and climate warming (1990
2010). Dots indicate average RWI change per annum representing the whole range of
stands considered in this study (A) or in two sub-sets according to geographical location
(B and C). Error bars denote standard deviation.
1465M. Bosela et al. / Science of the Total Environment 615 (2018) 14601469
Followingthe period of growth decline, silver r experienced a rapid
recovery reaching unprecedented levels across most of its distributional
range, coincidental with successful pan-European effort to limit acid de-
position. Even at the Cansiglio site (Italy) characterised by warm and
dry conditions, the growth pattern almost exactly followed that seen
in Slovakia until around 19952000, when at Cansiglio it turned to an-
other decline. This nding brings a new angle to a recent Europe-wide
study of silver r growth throughout the Holocene. Büntgen et al.
(2014) showed an increasing radial increment of r trees growing in
Italian Alps and Apennines until 2000, but did not indicate growth
trends after that year. Our observations from the Cansiglio site indicate
that r populations in the southern parts of the Alps may have recently
experienced a drought-stress related growth decline (Fig. 6a).
4.3. Long-term patterns of radial growth
For the rst time, to our knowledge, this study compares growth
pattern response to climate between managed and unmanaged forests.
We were able to make this comparison only for a series of sites in the
western Carpathians, however a pattern typical for many other sites
where these two species co-exist emerges. There was no discernible dif-
ference between tree growth in managed and unmanaged forests, but
beech RWIs were positive during 19501980 while those of r were
negative. By 19902010, the pattern reversed; RWIs describing r
growth were positive as the species recovered both in managed and un-
managed forests, but those depicting beech growth declined to negative
values. Again, this observation underlines the effect of long-term
Fig. 4. Principal componentanalysis (PCA) of growth-climate responses of silverr and European beech across Europe: a) dissimilarities between the study localities and species and in-
dication of northern and southern sites in the ordination space; b) ordination of contributing climate variables. The matrix of the correlation coefcients between site chronologies (pro-
duced by smoothing spline with 70% cut-off at 10-year segments) and climate variables (temperature, precipitation and sc-Palmer Drought Index) was used as an input to the PCA.
Abbreviations in labels: bh Bosnia & Herzegovina, it Italy, ro Romania, sk Slovakia, per Perucica; a silver randbEuropean beech; temp monthly temperature, ptemp
monthly temperature in the previous year, prec monthly precipitation, pprec monthly precipitation in the previous year, pdsi sc-Palmer drought index, ppdsi monthly sc-Palmer
drought index in the previous year.
Fig. 5. Dissimilaritiesin climate responses of silver r and European beech across Europeansites using hierarchical clustering. The analysis showsa differentiation between northern (red
and dashed-line rectangle) andsouthern (blueand solid-line rectangle) populationsof r and beech. The southernsites that were clusteredwithin the northerncluster are highlighted by a
blue rectangle around the labels. Abbreviations in the x-axislabel: bh Bosnia & Herzegovina, it Italy, ro Romania, sk Slovakia, per Perucica; a silver randbEuropean beech.
1466 M. Bosela et al. / Science of the Total Environment 615 (2018) 14601469
environmental conditions on tree-ring width, which seems to override
even the effect of forest management designed to stimulate bole wood
productivity.
While our observations of European beech show a wide variation of
growth patterns driven by severalfactors, the resultsfor European silver
r are quite consistent. A number of recent studies provides evidence
that diversity, whether species or functional, has a positive effect on
tree growth (Toïgo et al., 2015; Zhang et al., 2012). Higher diversity is
also believed to mitigate the negative impacts of extreme climate events
through higher growth resistance and resilience (e.g. Jucker et al., 2014;
Gazol et al., 2016; Metz et al., 2016). Although our study cannot directly
estimate the benets of growing ina diverse stand, theresults clearly in-
dicate that growing in a mixture does not shield the two species from
impacts of long-term changes in environmental conditions. For exam-
ple, we show that beech growth has been declining over the last two de-
cades in both managed and unmanaged forests and across a range of
conditions in Europe, regardless of the species composition and forest
structure. The same holds true for r's unusual radial increment pat-
terns, whereby long-term changes of environmental conditions seem
to prevailover local ecology.In this context, anycalculation of resistance
and resilience indices based on RWI must take into account multi-
decadal trends as these form the backgroundagainst which tree
growth must be considered.
5. Conclusions
Our study shows that state-of-the-art dendroecologicial techniques
can unravel complex environmental factors that inuence species-
specic tree growth trends. Although growing under the same condi-
tions, European beech and silver r exhibited remarkably different
growth patterns over the last half a century. While r responded posi-
tively to the recent warming, beech growth has declined across our
range of sites, suggesting that r is less susceptible to warmer and
drier conditions than beech. A comparison of growth patterns between
managed and unmanaged mixed beech-r forests revealed that the
long-term growth patterns were the same, suggesting only a limited
scope for tree growth stimulation by active forest management. There
is some support for the use of mixed forests as an adaptation strategy
to climate change. We show that a higher tree species diversity might
help mitigate the effects of short-term climatic events such as drought
and acidication, but may not prevent mixed forests from the long-
term consequences of climate change. Thus, any effort to convert
Norway spruce monocultures to preserve long-term growth at the for-
est ecosystem level should consider a purposeful decision to utilise
both beech and r as replacement species. Further scienticeffort
should be directed towards investigating effects of various management
interventions designed to aid the adaptation of beech-r ecosystems to
future climate change.
Acknowledgements
The work was fully supported by the Slovak Research and Develop-
ment Agency (SRDA) under the contract no. APVV-15-0265. BK was also
supported by the SRDA under the contract no. APVV-14-0086. MS and
VT were supported by the projects CIGA no. 20154316 and COST CZ
LD14074 (The Ministry of Education, Youth and Sports of the Czech
Republic). UB received funding from the Ministry of Education, Youth
and Sports of Czech Republic within the National Sustainability Program
I (NPU I; grant number LO1415).
Authors' contributions
MB conceived the ideasand MB, ML, DC and RS designed methodol-
ogy. MB andML analysed data. MB, ML and UB wrote the manuscript. All
Fig. 6. Correlation between the sitering-width index(RWI) chronologies and meantemperature (a)and sc-Palmer Drought Index(PDSI) (b) duringthe period of Juneto August in silverr
(green dashed lines) and European beech (red solid lines) in different localities across Europe (west_carp_man: managed stands in western Carpathians, west_carp_unm unmanaged
forests in western Carpathians, east_carp: easternCarpathians, south_carp: southern Carpathians, east_dinaric: eastern Dinaric,west_dinaric, south_alps: southern Alps, Bavarian Forest).
RWI were obtained after detrending by modied negative exponential function; bootstrapped correlation was computed over the whole period of about 110 years; values shown repre-
sent mean and 95% condence intervals.
1467M. Bosela et al. / Science of the Total Environment 615 (2018) 14601469
authors collected data, contributed critically to the drafts and gave nal
approval for publication.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.scitotenv.2017.09.092.
References
Aertsen, W., Janssen, E., Kint, V., Bontemps, J.-D., Van Orshoven, J., Muys, B., 2014. Long-
term growth changes of common beech (Fagus sylvatica L.) are less pronounced on
highly producti ve sites. For. Eco l. Manag. 312:252259. https://d oi.org/10.1016/
j.foreco.2013.09.034.
Bosela, M., Tobin, B., Šebeň, V., Petráš, R., Larocque, G.R., Šeben, V., Petráš, R., Larocque,
G.R., 2015. Different mixtures of Norway spruce, silver r, and European beech mod-
ify competitive interactions in central European mature mixed forests.Can. J. For. Res.
45, 15771586.
Bosela, M., Popa, I., Gömöry, D., Longauer, R., Tobin, B., Kyncl, J., Kyncl, T., Nechita, C.,
Petráš, R., Sidor, C.G., Šebeň, V., Büntgen, U., 2016a. Effects of postglacial phylogeny
and genetic diversity on the growth variability and climate sensitivity of European
silver r. J. Ecol. 104:716724. https://doi.org/10.1111/1365-2745.12561.
Bosela, M., Štefančík, I., Petš, R., Vacek, S., 2016b. The effects of climate warming
on the growth of European be ech forests depend critically on thinning strategy
and site productivity. Agric. F or. Meteorol. 2 22:2131. https://doi.org/10.1016/
j.agrformet.2016.03.005.
Briffa, K.R., Melvin, T.M., 2011. A closer look at regional curve standardization of tree-ring
records: justication of the need, a warning of some pitfalls, and suggested improve-
ments in its application. In: Hughe s, M.K., Diaz, H.F., Swetnam, T.W. (Eds.),
Dendroclimatology: Progress and Rospects. Springer Netherlands, pp. 113145.
Büntgen, U., Frank, D., Wilson, R., Carrer, M., Urbinati, C., Esper, J., 2008. Testing for tree-
ring divergence in the European Alps. Glob. Chang. Biol. 14:24432453. htt ps://
doi.org/10.1111/j.1365-2486.2008.01640.x.
Büntgen, U., Trouet, V., Frank, D.C., Leuschner, H.H., Friedrichs, D., Luterbacher, J., Esper, J.,
2010. Tree-ring indicators of German summer drought over the last millennium.
Quat. Sci. Rev. 29:10051016. https://doi.org/10.1016/j.quascirev.2010.01.003.
Büntgen, U., Tegel, W., Nicolussi, K., McCormick, M., Frank, D., Trouet, V., Kaplan, J.O.,
Herzig, F., Heussner, K.-U., Wanner, H., Luterbacher, J., Esper, J., 2011. 2500 years of
European climate variability and human susc eptibility. Science 331:578 582.
https://doi.org/10.1126/science.1197175 (80-.).
Büntgen,U., Tegel, W., Kaplan, J.O., Schaub, M., Hagedorn, F., Bürgi,M., Brázdil, R., Helle, G.,
Carrer, M., Heussner, K.U., Hofmann, J., Kontic, R., Kyncl, T., Kyncl, J., Camarero, J.J.,
Willy, T., Esper, J., Liebhold, A., 2014. Placing unprecedented recent r growth in a
European-wide and Holocene-long context. Front. Ecol. Environ. 12:100106.
https://doi.org/10.1890/130089.
Carrer, M., Urbinati, C., 2004. Age-dependent tree-ring growth responses to climate in
Larix decidua and Pinus cembra. Ecology 85, 730740.
Cavin, L., Jump, A.S., 2016. Highest drought sensitivity and lowest resistance to growth
suppression are found in the range core of the tree Fagus sylvatica L. not the equato-
rial range edge. Glob. Chang. Biol. 23:362379. https://doi.org/10.1111/gcb.13366.
Cook, E.R.,1985. A Time Series AnalysisApproach to Tree Ring Standardization.University
of Arizona.
Cook, E.R., Kairiukstis, L.A., 2013. Methods of Dendrochronology: Applications in the En-
vironmental Sciences. Springer Netherlands.
Cramer, H.H., 19 84. On the predisposition to disorders of Middle European forests.
Panzenschutz-Nachrichten Bayer 37, 97207.
Dai, A., 2011. Characteristics and trends in various forms of the palmer drought severity
index during 19002008. J. Geophys. Res. 116, D12115. https://doi.org/10.1029/
2010JD015541.
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. For. Ecol. Manag. 173, 6378.
Dixon, P.M., 2002. Bootstrap resampling. In:El-Shaarawi, A.H., Piegorsch, W.W. (Eds.), En-
cyclopedia of Environmetrics:p. 9 https://doi.org/10.1002/9780470057339.vab028.
EEA, 2006. European forest types: categories and types for sustainable forest manage-
ment reporting and policy. EEA Technical Report No 9/2006 (Copenhagen).
Ellenberg, H., 1996. Vegetation Mitteleuropas mit den Alpen. 5th ed. Ulmer, Stuttgard.
Fritts, H.C., 200 1. Tree Rings and Climate. The Black burn Press, New York and San
Francisco.
Galili, T., 2015. Dendextend: an R package for visualizing, adjusting and comparing trees
of hierarchical clustering. Bioinformatics 31:37183720. https://doi.org/10.1093/bio-
informatics/btv428.
Gazol, A., Camarero, J.J., Gutiérrez, E., Popa, I., Andreu-Hayles, L., Motta, R., Nola, P., Ribas,
M., Sangüesa-Barreda, G., Urbinati, C., Carrer, M., 2015. Distinct effects of climate
warming on populations of silver r(Abies alba) across Europe. J. Biogeogr. 42:
11501162. https://doi.org/10.1111/jbi.12512.
Gazol, A., Camarero, J.J., Gomez-Aparicio, L., 2016. Functional diversity enhances silver r
growth resilience to an extreme drought. J. Ecol. 104:10631075. https://doi.org/
10.1111/1365-2745.12575.
Hacket-Pain, A.J., Cavin, L., Friend, A.D., Jump, A.S., 2016. Consistent limitation of growth
by high temperature and low precipitation from range core to southern edge of
European beech indicates widespread vulnerability to changing climate. Eur. J. For.
Res. 135:897909. https://doi.org/10.1007/s10342-016-0982-7.
Harris, I., Jones, P.D., Osborn, T.J., Lister, D.H., 2014. Updated high-resolution grids of
monthly climati c observations - the CRU TS3.10 Dataset. Int. J. Cli matol. 34:
623642. https://doi.org/10.1002/joc.3711.
Hlásny,T., Turčáni, M., 2013. Persisting barkbeetle outbreak indicates the unsustainability
of secondary Norway spruce forests: case study from Central Europe. Ann. For. Sci.
70:481491. https://doi.org/10.1007/s13595-013-0279-7.
Hlásny, T., Barcza, Z., Fabrika, M., Balázs, B., Churkina, G., Pajtík, J., Sedmák, R., Turčáni, M.,
2011. Climate change impacts on growth and carbon balance of forests in Central
Europe. Clim. Res. 47:219236. https://doi.org/10.3354/cr01024.
IPCC, 2014. In: Core Writing Team, Pachauri , R.K., Meyer, L.A. (Eds.), Climate Change
2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth As-
sessment Report of the Intergovernmental Panel on Climate Change. IPCC, Geneva,
Switzerland.
Jönsson, A.M., Appelberg, G., Harding, S., Bärring, L., 2009. Spatio-temporal impact of cli-
mate changeon the activity and voltinism of thespruce bark beetle, Ips typographus.
Glob. Chang. Biol. 15:486499. https://doi.org/10.1111/j.1365-2486.2008.01742.x.
Jucker,T., Bouriaud, O., Avacaritei, D.,Coomes, D.A., 2014.Stabilizing effectsof diversity on
aboveground wood production in forest ecosystems: linking patterns and processes.
Ecol. Lett. 17:15601569. https://doi.org/10.1111/ele.12382.
Jump, A.S., Hunt, J.M., Penuelas, J., 2006. Rapid climate change-related growth decline at
the southern range edge of Fagus sylvatica. Glob. C hang. Biol. 12:21632174.
https://doi.org/10.1111/j.1365-2486.2006.01250.x.
Knoke, T., Ammer, C., Stimm, B., Mosandl, R., 2008. Admixing broadleaved to coniferous
tree species: a review on yield, ecological stability and economics. Eur. J. For. Res.
127:89101. https://doi.org/10.1007/s10342-007-0186-2.
Korf, V., 1939. Příspevěk k matematické formulaci vzrůstového zákona lesních porostů.
[Contribution to mathematical denition of the law of stand volume growth.]. Lesn.
práce. 18 pp. 339379.
Larsen, J.B., 1986. Das Tannensterben: eine neue Hypothese zur Klärung des
Hintergrundes dieser rätselhaften Kompexk rankheit der Weisstanne (Abies alba
Mill.). Forstwissenschaftliche Cent. 105, 381396.
Lebourgeois, F., Gomez, N., Pinto, P., Mérian, P., 2013. Mixed stands reduce Abies alba tree-
ring sensitivity to summer drought in the Vosges mountains, western Europe. For.
Ecol. Manag. 303:6171. https://doi.org/10.1016/j.foreco.2013.04.003.
Lévesque, M., Saurer,M., Siegwolf, R., Eilmann, B., Brang, P., Bugmann, H., Rigling, A., 2013.
Drought response of ve conifer species under contrasting water availabilitysuggests
high vulnerability of Norway spruce and European larch. Glob. Chang. Biol. 19:
31843199. https://doi.org/10.1111/gcb.12268.
Linares, J.C., Camarero, J.J., 2012. From pattern to process: linking intrinsic water-use ef-
ciency to drought-induced forest decline. Glob. Chang. Biol. 18:10001015. https://
doi.org/10.1111/j.1365-2486.2011.02566.x.
Luterbacher, J., 2004. European seasonal and annual temperature variability, trends, and
extremes since 15 00. Science 303 :14991503. https://doi.org/10.1126/sci-
ence.1093877 (80-.).
Metz, J., Annighöfer, P., Schall, P.,Zimmermann, J., Kahl, T., Schulze, E.D., Ammer, C., 2016.
Site-adapted admixed tree species reduce drought susceptibility of mature European
beech. Glob. Chang. Biol. 22:903920. https://doi.org/10.1111/gcb.13113.
Meyer, H., 1957. Beitrag zur Frage der Rückgängigkeitserscheinungen der Weisstanne
(Abies alba Mill.) am Nordrand ihres Naturareals. Arch. für Forstwes 6, 719787.
Mitchell, T.D., Jones, P.D., 2005. An improved method of constructing a datab ase of
monthly climate observations and associated high-resolution grids. Int. J. Climatol.
25:693712. https://doi.org/10.1002/joc.1181.
Nehrbass-Ahles,C.,Babst,F.,Klesse,S.,Nötzli,M.,Bouriaud,O.,Neukom,R.,
Dobbertin, M., Frank, D., 2014. The inuence of sampling design on tree-ring-
based quantication of forest growth. Glob. Chang. Biol. 20:28672885. https://
doi.org/10.1111/gcb.12599.
Oliver, C.D., Larson, B.C., 1996. Forest Stand Dynamics. Wiley, New York.
Paquette, A., Messier, C., 2011. The effect of biodiversity on tree productivity: from tem-
perate to boreal forests. Glob. Ecol. Biogeogr. 20:170180. https://doi.org/10.1111/
j.1466-8238.2010.00592.x.
Piovesan, G., Biondi, F., Di Filippo, A., Alessandrini, A., Maugeri, M., 2008. Drought-
driven growth reduction in old beec h (Fagus sylvatica L.) forests of the central
Apennines, Italy. Glob. Chang. Biol. 14:12651281. https://doi.org/10.1111/j.1365-
2486.2008.01570.x.
Pretzsch, H., Biber, P., Schütze, G., Uhl, E., Rötzer, T., 2014. Forest stand growth dynamics
in Central Europe has accelerated since 1870. Nat. Commun. 5:4967. https://doi.org/
10.1038/ncomms5967.
Primicia,I., Camarero, J.J.,Janda, P., Čada, V., Morrissey, R.C., Trotsiuk, V., Bače, R., Teodosiu,
M., Svoboda, M., 2015. Age, competition, disturbance and elevation effects on tree
and stand growth response of prima ry Picea abies for est to climate. Fo r. Ecol.
Manag. 354:7786. https://doi.org/10.1016/j.foreco.2015.06.034.
R Development Core Team, 2008. R: A Language and Environment for Statistical
Computing.
Ruiz-Benito, P., Gómez-Aparicio, L., Paquette, A., Messier, C., Kattge, J., Zavala, M.A., 2014.
Diversityincreases carbon storage and tree productivity in Spanish forests.Glob. Ecol.
Biogeogr. 23:311322. https://doi.org/10.1111/geb.12126.
Ruosch, M., Spahni, R., Joos, F., Henne, P.D., van der Knaap, W.O., Tinner, W., 2016. Past
and future evolution of Abies alba forests in Europe - comparison of a dynamic veg-
etation model with palaeo data and observations. Glob. Chang. Biol. 22:727740.
https://doi.org/10.1111/gcb.13075.
Spiecker,H., Hansen, J., Klimo,E., Skovsgaard, J.P.,Sterba, H., von Teuffel, K., 2004. Norway
Spruce Conversion - Options and Consequences. Koninklijke Brill NV, Leiden, The
Netherlands.
Tarp, P., Helles, F., Holten-Andersen, P., Bo Larsen, J., Strange, N., 2000. Modelling near-
natural silvicultural regimes for beech - an economic sensitivity analysis. For. Ecol.
Manag. 130:187198. https://doi.org/10.1016/S0378-1127(99)00190-5.
1468 M. Bosela et al. / Science of the Total Environment 615 (2018) 14601469
Tegel, W.,Seim, A., Hakelberg, D., Hoffmann,S., Panev, M., Westphal, T., Büntgen,U., 2014.
A recent growth increase of European beech (Fagus sylvatica L.) at its Mediterranean
distribution limit contradicts drought stress. Eur. J. For. Res. 133 :6171. https://
doi.org/10.1007/s10342-013-0737-7.
Tinner, W., Colombaroli, D., Heiri, O., Henne, P.D., Steinacher, M., Untenecker, J., Vescovi,
E., Allen, J.R.M., Carraro, G., Conedera, M., Joos, F., Lotter, A.F., Luterbacher, J.,
Samartin, S., Valsecchi, V., 2013. Thepast ecology of Abies albaprovides new perspec-
tives on future responses of silver r forests to global warming. Ecol. Monogr. 83:
419439. https://doi.org/10.1890/12-2231.1.
Toïgo, M.,Vallet, P., Perot, T.,Bontemps, J.-D., Piedallu, C., Courbaud, B., 2015. Overyielding
in mixed forests decrease s with site prod uctivity. J. Ecol. 103:502512. https://
doi.org/10.1111/1365-2745.12353.
Ulrich, B., 1995. The history and possible causes of forest decline in central Europe, with
particular attention to the German situation Bernhard. Environ. Rev. 3, 262276.
Vitali, V., Büntgen, U., Bauhus, J., 2017. Silver r and Douglas r are more tolerant to ex-
treme droughts than Norway spruce in south-western Germany. Glob. Chang. Biol.
https://doi.org/10.1111/gcb.13774.
Wilson, R., Elling, W., 2004. Temporal instability in tree-growth/climate respo nse in
the Lower Bavarian Forest region: implications for dendroclimatic reconstruction.
Trees - Struct. Funct. 18:1928. https://doi.org/10.1007/s00468-003-0273-z.
Wood, S.N., 2011. Fast stable restricted maximum likelihood and marginal likelihood es-
timation of semi parametric generalized line ar models. J. R. S tat. Soc. Ser. B Stat
Methodol. 73:336. https://doi.org/10.1111/j.1467-9868.2010.00749.x.
Zang, C., Biondi, F., 2013. Dendroclimatic calibration in R: the bootRes package for re-
sponse and correlation functi on analysis. Dendrochronologia 31:6874. htt ps://
doi.org/10.1016/j.dendro.2012.08.001.
Zang, C., Hartl-Meier,C., Dittmar, C., Rothe,A., Menzel, A., 2014. Patterns of drought toler-
ance in majorEuropean temperateforest trees: climatic drivers andlevels of variabil-
ity. Glob. Chang. Biol. 20:37673779. https://doi.org/10.1111/gcb.12637.
Zhang, Y.,Chen, H.Y.H., Reich,P.B., 2012. Forest productivity increases withevenness, spe-
cies richness and trait variation: a global meta-analysis. J. Ecol. 100:742749. https://
doi.org/10.1111/j.1365-2745.2011.01944.x.
1469M. Bosela et al. / Science of the Total Environment 615 (2018) 14601469
... Despite its functional adaptability and great ecological plasticity, it is affected by drought, as confirmed by studies of its response in southern Europe [7]. It thrives in pure and mixed stands with conifers, especially firs, whose distribution is mainly limited to the area of the Alps and the Carpathians [8]. A long-term decline in radial growth at lower elevations in Central Europe since around the 1980s [9][10][11] suggests that it is sensitive to a warmer and drier climate [12]. ...
... In southwestern Europe, fir is more resilient to climatic extremes compared to other tree species [8]. At the same time, two fir populations have been distinguished in the Carpathian region: the eastern one, which is similar to the Balkan population; and the western one, which is less sensitive to summer droughts [8]. ...
... In southwestern Europe, fir is more resilient to climatic extremes compared to other tree species [8]. At the same time, two fir populations have been distinguished in the Carpathian region: the eastern one, which is similar to the Balkan population; and the western one, which is less sensitive to summer droughts [8]. In the south-exposed areas of Eastern Carpathians, fir was the least sensitive of studied tree species [43]; its growth rate increased continuously and remained at a high level even in old individuals compared to Scots pine or Norway spruce. ...
Article
Full-text available
European forests are becoming increasingly threatened by climate change and more frequent droughts. The likely responses of species to climate change will vary, affecting their competitiveness , their existence, and consequently, forest management decisions and measures. We determined the influence of climate on the radial growth of European beech and silver fir along the Carpathians to find similarities between the two species and the main differences. Along the Carpathian Mountains, seven sites with mature fir-beech stands above 800 m above sea level were selected and analyzed. Our study confirmed different responses depending on species and location. A more pronounced response of tree growth to climate was observed on the eastern side of the Carpathians, while it was less expressed or even absent on the southern sites. Both beech and fir show better radial growth with higher precipitation in July and slower growth with higher average and maximum temperatures in June of the current year. Fir demonstrates a positive correlation between radial growth and temperature in winter, while beech demonstrates a negative correlation between radial growth and temperature in summer. In the 1951-1960 decade, the average tree ring widths in fir and beech were largest at the southern sites compared to the other sites, but since 2011, the southern sites have had the lowest increase while northern sites have had the largest. Both species respond differently to climate and are likely to follow different competitive paths in the future.
... Because of its shade tolerance, beech can survive several decades of suppression and release periods. After reaching the canopy the lifespan usually does not exceed 200 years (Hobi et al., 2015), although on less productive sites, individual beech trees can live even longer than 400-500 years (Di Filippo et al., 2007, Bosela et al., 2018. The developmental cycle of beech forests with an estimated duration of 220-250 years is mostly dominated by multi-layered stands with high structural diversity (Korpeľ, 1995). ...
... Between 1960 and 2000, beech has increased stand productivity exceeding predictions using yield models by 5-77%, with a gradient from east to west (Pretzsch et al., 2014, Bosela et al., 2016 which represents a high variability. According to other studies, beech populations across Europe have experienced a decline in growth rates over recent decades along a latitudinal gradient (Bosela et al., 2018, Martinez del Castillo et al., 2022. ...
... The reduced air pollution loads in the region since 1990 may have also contributed to the released beech growth (Bosela et al., 2019). While the effect of air pollution on the radial growth was reported for silver fir (Abies alba, Mill.) at regional (Elling et al., 2009) and at European scales , it was not found for European beech when compared with co-occurring silver fir trees on the same sites (Bosela et al., 2018). ...
... The relationship of growth to the scPDSI corroborated strong limitation of spruce growth by the The observed deep growth depression in fir from the early 1960s to the late 1980s corresponds to the findings of other studies from different parts of Europe (Becker et al. 1989;Elling et al. 2009;Gazol et al. 2015;Vitali et al. 2017;Bošela et al. 2018). The growth decline was probably the result of exposure to atmospheric pollution, in particular to sulphur dioxide, to which fir is sensitive (Elling et al. 2009;Mikulenka et al. 2020). ...
... The vitality of silver fir has increased since the end of 1980s in many stands in Central and Western Europe (Büntgen et al. 2014;Gazol et al. 2015;Bošela et al. 2018) due to a decrease in SO 2 emissions, air warming with constant precipitation and an increase in nitrogen deposition (Elling et al. 2009;Büntgen et al. 2014). Our study shows that the increase in fir growth was interrupted by a drought in 2003 (Ciais et al. 2005), after which growth stagnated or decreased. ...
... способность растений повышать термоустойчивость в ответ на воздействие высоких температур [3]. Текущая тенденция к потеплению может как ускорить камбиальный рост растений бука [5] в летний период, так и наоборот, снизить популяцию Fagus sylvatica L. из-за увеличения продолжительности засушливого периода [13]. Для того, чтобы оценить реакцию древесных растений на воздействие высоких температур и выявить состояние здоровья лесных насаждений, необходимо применять методы, позволяющие провести быстрые и точные анализы. ...
Thesis
Full-text available
Was studied the effect of heat shock of different intensities on the leaves of European beech (Fagus sylvatica L.) seedlings, the seeds of which are obtained from different cultivation zones (Ukraine, Moldova, Romania). The amount of damage to cellular structures caused by heat shock was determined by the electrolyte leakage method. The experiment was carried out in the appearance period of 4–5 true leaves of European beech seedlings using a heat shock temperature of 59°C. The leaves of the European beech seedlings from the Suceava zone, compared with the leaves of European beech seedlings from the Hоrjavca and Cernăuţi zones, showed increased resistance to heat shock. The results make it possible to apply the electrolyte leakage method to determine the primary heat resistance of beech seedlings from different cultivation zones with various environmental conditions.
... Due to the unknown past tree mortality and its reasons, the survivor bias may still be present in our data, though (Pretzsch, 2021). The survivor bias may become larger if the forest ecosystem or tree species suffered a higher mortality rate due to the environmental stress in the past, such as air pollution, drought events (Bosela et al., 2018(Bosela et al., , 2021b or pest outbreaks (Canelles et al., 2021;Cooke et al., 2003), as well as in intensively managed forests where the suppressed trees are often removed at thinning. ...
Article
Process-based models and empirical modelling techniques are frequently used to (i) explore the sensitivity of tree growth to environmental variables, and (ii) predict the future growth of trees and forest stands under climate change scenarios. However, modelling approaches substantially influence predictions of the sensitivity of trees to environmental factors. Here, we used tree-ring width (TRW) data from 1630 beech trees from a network of 70 plots established across European mountains to build empirical predictive growth models using various modelling approaches. In addition, we used 3-PG and Biome-BGCMuSo process-based models to compare growth predictions with derived empirical models. Results revealed similar prediction errors (RMSE) across models ranging between 3.71 and 7.54 cm2 of basal area increment (BAI). The models explained most of the variability in BAI ranging from 54 % to 87 %. Selected explanatory variables (despite being statistically highly significant) and the pattern of the growth sensitivity differed between models substantially. We identified only five factors with the same effect and the same sensitivity pattern in all empirical models: tree DBH, competition index, elevation, Gini index of DBH, and soil silt content. However, the sensitivity to most of the climate variables was low and inconsistent among the empirical models. Both empirical and process-based models suggest that beech in European mountains will, on average, likely experience better growth conditions under both 4.5 and 8.5 RCP scenarios. The process-based models indicated that beech may grow better across European mountains by 1.05 to 1.4 times in warmer conditions. The empirical models identified several drivers of tree growth that are not included in the current process-based models (e.g., different nutrients) but may have a substantial effect on final results, particularly if they are limiting factors. Hence, future development of process-based models may build upon our findings to increase their ability to correctly capture ecosystem dynamics.
... In Europe, warmer and more variable climate conditions have been linked to increased forest dieback and tree mortality rates (Neumann et al., 2017;Camarero 2022, Piedallu et al., 2022). Among European tree species, silver fir (Abies alba Mill.) is a key conifer in temperate mountain forests, which has shown different growth patterns and responses to climate across its distribution range (Büntgen et al., 2014;Zang et al., 2014;Gazol et al., 2015;Vitali et al., 2017;Bošela et al., 2018;Vitasse et al., 2019). The distribution of silver fir is expected to shift northwards and upwards according to warmer climate projections, whereas the current range would be substantially reduced in southern Europe in response to warmer-drier conditions (Tinner et al., 2013;Ruosch et al., 2016). ...
Article
Full-text available
Assessing tree growth patterns and deviations from expected climate baselines across wide environmental gradients is fundamental to determine forest vulnerability to drought. This need is particularly compelling for the southernmost limit of the tree species distribution where hot droughts often trigger forest dieback processes. This is the case of some silver fir (Abies alba) populations located in southwestern Europe (Spanish Pyrenees) which present ongoing dieback processes since the 1980s. We sampled 21 silver fir stands showing different dieback intensity, assessed using defoliation levels, quantified their growth patterns and characterized their responses to climate. Then, we assessed growth deviations from climatic predictions using the process-based Vaganov-Shashkin (VS) growth model. The forests showing most intense dieback, i.e. highest defoliation levels, were mainly located in low-elevation sites of the western Pyrenees. Trees in these stands displayed the lowest growth rates and the highest year-to-year variability in growth and their growth was limited by late-summer evaporative demand. In eastern and central Pyrenees, we detected a mild growth limitation by low soil moisture during the late growing season and positive growth recovery in recent years with respect to a climate baseline. Decreasing growth trajectories were the most common pattern, while rising trends were common in stands with low dieback in eastern and central Pyrenees. Our results portend systematic spatial variability of growth trends across the Pyrenean silver fir populations forming the southwestern distribution limit of the species in Europe. Decoupling of growth between eastern and western populations observed in the recent decades suggests contrasting vulnerability to climate change, and more importantly, the decoupling of growth patterns in western clusters could be used as an early-warning signal of impending dieback. Consequently, we foresee future dieback events to have more detrimental effects in the western compared with the eastern Pyrenees.
... Karst.) [1][2][3][5][6][7][8]. ...
Article
Full-text available
Citation: Michalak, M.; Plitta-Michalak, B.P.; Suszka, J.; Naskręt-Barciszewska, M.Z.; Kotlarski, S.; Barciszewski, J.; Chmielarz, P. Identification of DNA Methylation Changes in European Beech Seeds during Desiccation and Storage. Int. J. Mol. Sci. 2023, 24, 3557..pl (B.P.P.-M.) † These authors contributed equally to this work. Abstract: Ageing and deterioration of seeds is a major problem for the maintenance of seed quality and viability during long-term storage. Prediction of early stages of seed deterioration in order to point out the plantlets' regeneration time is a major challenge of successful storage. In preserved seeds, damages accumulate within cells at the rate mainly related to their moisture content and temperature of storage. Current research reveals global alterations in DNA methylation in lipid-rich intermediate seeds during desiccation and storage at various regimes covering nonoptimal and optimal conditions. We show for the first time that monitoring of 5-methylcytosine (m 5 C) level in seeds can be used as a truly universal viability marker regardless of postharvest category of seeds and their composition. For seeds stored up to three years, in varied conditions, moisture content, temperature, and time of storage had significant influence on seedling emergence and DNA methylation (p < 0.05). Similarities among lipid-rich intermediate and orthodox seeds regarding different reactions of embryonic axes and cotyledons to desiccation are newly revealed. Along with previous studies on seeds dramatically different in desiccation tolerance (recalcitrant vs. orthodox), results regarding lipid-rich seeds positioned in-between (intermediate) prove that maintaining global DNA methylation status is crucial for maintaining seed viability.
Article
Full-text available
Climate change’s negative effects, such as rising global temperatures and the disruption of global ecological ecosystems as a direct effect of rising carbon emissions in the atmosphere, are a significant concern for human health, communities, and ecosystems. The condition and presence of forest ecosystems, especially those in peri-urban areas, play an essential role in mitigating the negative effects of climate change on society. They provide direct benefits to the residents of large cities and their surrounding areas, and they must be managed sustainably to protect all their component ecosystems. This research was carried out in the forests of Lunca Muresului Natural Park and Bazos Arboretum, located in the Romanian sector of the Pannonian Plain, near urban agglomerations. The results showed high variability in the stands. Using the height-to-diameter ratio indicator concerning dbh and species, a strong Pearson correlation was registered (between 0.45 and 0.82). These values indicate the high stability of these stands, providing positive human–nature interactions such as recreational or outdoor activities (and a complementary yet indirect use value through attractive landscape views). Protecting these ecosystems offers a so-called insurance policy for the next generations from a climate change standpoint.
Chapter
The need of forestry in Primorsky Krai in planting material of the main forest-forming species for forest renewal is increasing from year to year. The issues related to effective use of large-sized planting material of solid fir are considered. The main problem of this valuable species planting material cultivation is connected with high cost of seeds preparation and rare productive years. Therefore, rational use of planting material of needle fir in nursery farms is one of the important tasks of foresters in the region. Based on the long-term research and analysis of literature data, the main methods of effective use of planting material aimed at modern qualitative restoration of degraded fir forests of the region are considered. The work solved both theoretical and practical problems on the chosen research theme.KeywordsReforestationNeedle firLarge-scale planting materialForest cropsPlantationsLandscapingSchool nursery farm branchSeedling heightDiameter at the seedling root crown
Article
Full-text available
Improving our understanding of the potential of forest adaptation is an urgent task in the light of predicted climate change. Long-term alternatives for susceptible yet economically important tree species such as Norway spruce (Picea abies) are required if the frequency and intensity of summer droughts will continue to increase. Although Silver fir (Abies alba) and Douglas fir (Pseudotsuga menziesii) have both been described as drought-tolerant species, our understanding of their growth responses to drought extremes is still limited. Here, we use a dendroecological approach to assess the resistance, resilience, and recovery of these important central Europe to conifer species the exceptional droughts in 1976 and 2003. A total of 270 trees per species were sampled in 18 managed mixed-species stands along an altitudinal gradient (400-1200 m a.s.l.) at the western slopes of the southern and central Black Forest in southwest Germany. While radial growth in all species responded similarly to the 1976 drought, Norway spruce was least resistant and resilient to the 2003 summer drought. Silver fir showed the overall highest resistance to drought, similarly to Douglas fir, which exhibited the widest growth rings. Silver fir trees from lower elevations were more drought-prone than trees at higher elevations. Douglas fir and Norway spruce, however, revealed lower drought resilience at higher altitudes. Although the 1976 and 2003 drought extremes were quite different, Douglas fir maintained consistently the highest radial growth. Although our study did not examine population level responses, it clearly indicates that Silver fir and Douglas fir are generally more resistant and resilient to previous drought extremes and are therefore suitable alternatives to Norway spruce; Silver fir more so at higher altitudes. Cultivating these species instead of Norway spruce will contribute to maintaining a high level of productivity across many Central European mountain forests under future climate change.
Article
Full-text available
The aim of our study was to determine variation in the response of radial growth in Fagus sylvatica L (European Beech) to climate across the species full geographical distribution and climatic tolerance. We combined new and existing data to build a database of 140 tree-ring chronologies to investigate patterns in growth–climate relationships. Our novel meta-analysis approach has allowed the first investigation of the effect of climate on tree growth across the entire geographical distribution of the species. We identified key climate signals in tree-ring chronologies and then investigated how these varied geographically and according to mean local climate, and by tree age and size. We found that the most important climate variables significantly correlated with growth did not show strong geographical patterns. Growth of trees in the core and at the southern edge of the distribution was reduced by high temperature and low precipitation during the growing season, and by high temperatures in the previous summer. However, growth of trees growing in warmer and drier locations was more frequently significantly correlated with summer precipitation than other populations. Additionally, the growth of older and larger trees was more frequently significantly correlated with previous summer temperature than younger and smaller trees. Trees growing at the south of the species geographical distributions are often considered most at risk from climate change, but our results indicate that radial growth of populations in other areas of the distribution is equally likely to be significantly correlated with summer climate and may also be vulnerable. Additionally, tree-rings from older trees contain particular growth–climate relationships that are rarely found in younger trees. These results have important implications for predicting forest carbon balance, resource use and likely future changes to forest composition across the continent.
Article
Full-text available
Biogeographical and ecological theory suggests that species distributions should be driven to higher altitudes and latitudes as global temperatures rise. Such changes occur as growth improves at the poleward edge of a species distribution and declines at the range edge in the opposite or equatorial direction, mirrored by changes in the establishment of new individuals. A substantial body of evidence demonstrates that such processes are underway for a wide variety of species. Case studies from populations at the equatorial range edge of a variety of woody species have led us to understand that widespread growth decline and distributional shifts are underway. However, in apparent contrast, other studies report high productivity and reproduction in some range edge populations. We sought to assess temporal trends in the growth of the widespread European beech tree (Fagus sylvatica) across its latitudinal range. We explored the stability of populations to major drought events and the implications for predicted widespread growth decline at its equatorial range edge. In contrast to expectations, we found greatest sensitivity and low resistance to drought in the core of the species range, whilst dry range edge populations showed particularly high resistance to drought and little evidence of drought-linked growth decline. We hypothesize that this high range edge resistance to drought is driven primarily by local environmental factors that allow relict populations to persist despite regionally unfavourable climate. The persistence of such populations demonstrates that range-edge decline is not ubiquitous and is likely to be driven by declining population density at the landscape scale rather than sudden and widespread range retraction.
Article
Full-text available
It is expected that extreme climate events such as droughts will increase in both severity and intensity as a consequence of forecasted climate change. Complementarity among tree species in resource acquisition strategy may reduce interspecific competition and increase the occurrence of facilitative interactions, resulting in an improved tree growth and resilience to extreme climatic events. However, the response of individual trees growing in more functionally diverse stands to extreme events is still under debate. We investigate the growth response of silver fir ( Abies alba ) to an extreme drought event which occurred in 1985–1986 in the Central Spanish Pyrenees and how the growth and resilience of trees relate to their neighbourhood functional diversity, stand heterogeneity and intraspecific and interspecific competition. Dendroecological methods were used to reconstruct radial growth. We recorded all tree species living around each sampled tree and calculated indices of intra‐ and interspecific competition among coexisting trees. Functional diversity around each focal tree was calculated by using the Rao's quadratic entropy. Silver fir growth decreased significantly in response to the drought in the vast majority of trees, but most of them showed a growth recovery 3 years later. Despite most trees showed negative growth trends following the drought event (70%), those trees growing in more functionally diverse stands were more resilient, recovered more quickly and displayed greater growth and growth trends. Synthesis . The presented findings support the growing body of evidence that shows a positive impact of diversity on forest function. In addition, we provide a direct linkage between tree growth response to climate at the individual tree scale and a direct quantification of neighbourhood functional diversity and competition. An enhanced functional diversity facilitating a more efficient use of resources at the interspecific level and potentially increasing facilitative interactions contributes to lessening the negative impacts of extreme events as droughts on forest growth.
Article
Full-text available
Growth rates of European silver fir (Abies alba Mill.) rapidly increased in the last century. At the same time, ring widths declined at the species southern distribution limits in the Mediterranean. Such diverse growth trends and responses have largely been attributed to regional climate conditions, but this was prior to considering the species’ postglacial phylogeny. 2.A dendrochronological network composed of 1,961 tree-ring width series (TRW) from 78 silver fir sites between 365 and 1,400 m a.s.l. along the Carpathian Arc was compiled. Spatial differences in the species’ genetic diversity were investigated from genetic data of 69 silver fir populations in the region. Differences in growth variability and climate sensitivity were then related to postglacial phylogeny and genetic diversity. 3.Significant differences in inter-annual and longer-term growth trends and climate responses across the Carpathian Arc were found to coincide with the geographical north-south separation of two postglacial populations from effective refugia originating from the Apennine and Balkan peninsulas. Summer temperature was the main driver of growth in the western (Apennine) lineage, whereas ring widths in the Balkan population from the east were predominantly controlled by summer drought. Fir specimens that originated from the Balkan lineage exhibited higher genetic diversity, more regular growth dynamics, and also appeared to be less sensitive to air pollution during the 1970s. 4.Synthesis. Although the phylogeny of forest trees has largely been neglected in most dendroecological studies, results here indicate the importance of different postglacial histories for the growth sensitivity and adaptability to varying environmental factors. Decision making under future climate warming scenarios (for building resilience through forest management) should therefore consider different phylogenetic origins. This article is protected by copyright. All rights reserved.
Article
Summary. Recent work by Reiss and Ogden provides a theoretical basis for sometimes preferring restricted maximum likelihood (REML) to generalized cross-validation (GCV) for smoothing parameter selection in semiparametric regression. However, existing REML or marginal likelihood (ML) based methods for semiparametric generalized linear models (GLMs) use iterative REML or ML estimation of the smoothing parameters of working linear approximations to the GLM. Such indirect schemes need not converge and fail to do so in a non-negligible proportion of practical analyses. By contrast, very reliable prediction error criteria smoothing parameter selection methods are available, based on direct optimization of GCV, or related criteria, for the GLM itself. Since such methods directly optimize properly defined functions of the smoothing parameters, they have much more reliable convergence properties. The paper develops the first such method for REML or ML estimation of smoothing parameters. A Laplace approximation is used to obtain an approximate REML or ML for any GLM, which is suitable for efficient direct optimization. This REML or ML criterion requires that Newton–Raphson iteration, rather than Fisher scoring, be used for GLM fitting, and a computationally stable approach to this is proposed. The REML or ML criterion itself is optimized by a Newton method, with the derivatives required obtained by a mixture of implicit differentiation and direct methods. The method will cope with numerical rank deficiency in the fitted model and in fact provides a slight improvement in numerical robustness on the earlier method of Wood for prediction error criteria based smoothness selection. Simulation results suggest that the new REML and ML methods offer some improvement in mean-square error performance relative to GCV or Akaike's information criterion in most cases, without the small number of severe undersmoothing failures to which Akaike's information criterion and GCV are prone. This is achieved at the same computational cost as GCV or Akaike's information criterion. The new approach also eliminates the convergence failures of previous REML- or ML-based approaches for penalized GLMs and usually has lower computational cost than these alternatives. Example applications are presented in adaptive smoothing, scalar on function regression and generalized additive model selection.