Contents lists available at ScienceDirect [629556]

Contents lists available at ScienceDirect
Agricultural and Forest Meteorology
journal homepage: www.elsevier.com/locate/agrformet
Impact of Climate Change and Adaptive Genetic Potential of Norway Spruce
at the South –eastern Range of Species Distribution
Georgeta MIHAIa,⁎, Maria TEODOSIUb, Marius-Victor BIRSANc, Alin-Madalin ALEXANDRUa,
Ionel MIRANCEAa, Ecaterina-Nicoleta APOSTOLa, Paula GARBACEAa, Lucia IONITAa
aDepartment of Forest Genetics and Tree Breeding, « Marin Dracea » National Institute for Research and Development in Forestry, Eroilor 128, Voluntari, 077190,
Romania
bC-lung Moldovenesc Station, « Marin Dracea » National Institute for Research and Development in Forestry, Calea Bucovinei 73bis, Campulung Moldovenesc, Suceava,
Romania
cMeteo Romania (National Meteorological Administration), Department of Research and Meteo Infrastructure Projects, Sos. Bucuresti-Ploiesti 97, Bucharest, 013686,
Romania
1. Introduction
Norway spruce ( Picea abies L. Karst) is the most widespread conifer
species in Europe and has high economic and ecological value. Its
natural distribution covers approximately 30,000,000 ha in Europe,
and more than 20% of the current distribution represents its cultivated
range ( Klimo et al. 2000 ). Since the middle of the 18th century, Norway
spruce has been extensively planted both within and outside its naturalrange, mainly at low altitudes in the broadleaf vegetation layer. The
transfer of forest reproductive material to long distances and the use of
non –local seeds was common practice ( Jansen et al. 2017), making it
highly di fficult to separate local and non –local populations today.
Further, damage due to storms, snow, drought and bark insects has
occurred more frequently in Norway spruce forests compared to other
species.
In the last centuries, more than ever, environmental deterioration
and global warming are further potential impact factors a ffecting the
diversity and distribution of forest ecosystems. The most recent climatechange scenarios ( Meinshausen et al. 2011 ) project increases in mean
annual temperature of 1 –4°C by the end of this century. According to
Cheval et al. (2017) , the average annual air temperature over Romania
is estimated to increase by 1.2°C during 2021– 2050 compared to
1991 –2020, and by more than 2°C in 100 years ( Bojariu et al. 2015 ).
Recent climatic changes show statistically signi ficant increases in
warming –related thermal extremes, extreme precipitation events and
diminishing wind speed ( Birsan et al. 2013; Manea et al. 2016;
Busuioc et al. 2016).
Forest species are particularly sensitive to climate change because
the long life span of trees does not allow for rapid adaptation to en-vironmental changes ( Lindner et al. 2010). The changing climate is
expected to contribute to an increased loss of timber, reduction in forestadaptability and signi ficantly altered species distribution
(Beaulieu et al. 2004 ,Mihai et al. 2018 ;Sidor et al. 2019 ). Tree species
adapt over time via natural selection, migration to new habitats or
phenotypic plasticity ( Price et al. 2003 ). The tree species’ capacity to
survive under a changing climate depends on its intraspeci fic adaptive
genetic variation ( Schueler et al. 2013 ). Furthermore, the genetic di-
versity as the ultimate evolutionary basis of many species and popu-lations, guarantees the survival and adaptation of the forest species in a
changing environment.
The genetic variation and local adaptation of Norway spruce have
been highlighted in many studies. Thus, the geographic pattern of its
genetic variation exhibits highly strong latitudinal and altitudinal clines
for bud phenology, cold hardiness and growth traits ( Beuker et al. 1998;
Johnsen et al. 2005; Savolainen et al. 2007; Gomöry et al. 2011;
Kapeller
et al. 2012 ). Studies based on isozyme and DNA markers at
both regional and entire distribution range levels indicated a high de-gree of within –population diversity and particularly low genetic dif-
ferentiation among populations ( Lagercrantz and Ryman 1990;
Vendramin et al. 2000 ;Meloni et al. 2007 ;Konnert 2009;
Nowakowska 2009; Tollesfrud et al. 2009; Tsuda et al. 2016;
Machova et al. 2018 ,Westergren et al. 2018; Stojnic et al. 2019 ). More
recently, the genome –wide polymorphism was used in order to in-
vestigate the patterns of population structure and the link between past
demography and local adaptation of Norway spruce. ( Chen et al. 2019;
Milesi et al. 2019 ;Wang et al. 2020).
Norway spruce is a key European mountainous forest ecosystem
species. As such, understanding its intraspeci fic response to climate
change is becoming increasingly important ( Burczyk and Giertych
1991 ;Bouriaud et al. 2005 ;Suvantoa et al. 2016 ;Ununger et al. 1988;
Pakalaj et al. 2002; Kapeller et al. 2012 ;Ulbrichová et al. 2013 ). The
results of existing studies revealed some contradictory results; for
https://doi.org/10.1016/j.agrformet.2020.108040
Received 27 May 2019; Received in revised form 6 May 2020; Accepted 18 May 2020⁎Corresponding author: Georgeta MIHAI ”Marin Dracea ”National Institute for Research and Development in Forestry, Departament of Forest Genetics and Tree
Breeding. B-dul Eroilor 128, 077190, Voluntari, Ilfov, Romania
E-mail address: gmihai_2008@yahoo.com (G. MIHAI).Agricultural and Forest Meteorology 291 (2020) 108040
Available online 17 June 2020
0168-1923/ © 2020 Elsevier B.V. All rights reserved.
T

example, Trujillo –Moya et al. (2018) andGeorge et al. (2019) reported
that Norway spruce holds a high level of adaptive variation, including
the potential to adjust to the predicted water de ficit in Central Europe.
Other studies have shown that, among European conifers, the Norwayspruce has a higher sensitivity to prolonged drought and extreme heat
(Lévesque et al. 2013; Zang et al. 2014 ).
Inter –population genetic variation of Norway spruce from the
South –eastern Carpathians has been the subject of numerous publica-
tions ( Curtu et al. 2009 ;Mihai 2009; Teodosiu 2011 ;Radu et al. 2014 ).
However, the adaptive genetic potential of Norway spruce populations
from the south –eastern part of its distribution area remains an open
question. More studies have revealed that the Romanian Carpathiansconstitute the most important diversity hotspot for Norway spruce and
have critical importance as long –term stores of genetic diversity and
foci of speciation for this species in Europe ( Hampe and Petit 2005;
Tollefsrud et al. 2008 ;Gomöry et al. 2010).
New knowledge and multidisciplinary approaches are urgently re-
quired to support the conservation of genetic resources and manage-
ment of this species. The integration of quantitative genetics studies in
long –term provenance experiments, analyses of highly informative
DNA markers in natural populations and climatic projections is essen-tial for the sustainable management of Norway spruce ecosystems.
The studies in long –term trials of Norway spruce allow us to identify
the climatic variables that exert strong selective pressure on tested
populations and to develop models that can be used to predict species ’
response to future climate ( Rehfeldt et al. 1999a ;Wang et al. 2006 ).
The studies of neutral variation that reveal genetic diversity and spatial
genetic structure are important for analysing adaptative capacity and
making informed decisions in forest management
(Schueler et al. 2013 ).
Starting from the hypothesis that global warming will reduce the
distribution area and growth of Norway spruce signi ficantly in
South –eastern Europe, the objectives of this study were to 1) assess the
adaptive genetic capacity of Norway spruce provenances, 2) evaluatethe genetic diversity and spatial structure as the genetic basis of
adaptation, 3) determine the climate change impact on species dis-
tribution and 4) provide practical information for the sustainable
management.
Through both quantitative analyses in long –term trials and genetic
analyses in additional natural populations, we will obtain a more
comprehensive picture of adaptive genetic capacity of Norway spruce
populations at the south –eastern distribution limit in Europe. The
growth response models will allow to predict what will be the impact ofclimate change on the growth and current distribution of the species in
this region. By identifying which populations are most vulnerable under
future climatic scenarios we will be able to provide practical informa-
tion for the sustainable management of Norway spruce ecosystems.
2. Materials and Methods2.1. Trial site and plant material
Sixfield
trials with 4– year old seedlings from 33 autochthonous
Norway spruce provenances were established in 1980 ( Enescu 1988 ).
The trials were located in di fferent geographic regions throughout the
entire Romanian Carpathian chain. Four trials were established outside
the natural distribution zone of the Norway spruce, in European beech
(Fagus sylvatica L.) and sessile oak ( Quercus petraea Mattsch. Liebl.)
layers at altitudes of 380 –750 m above sea level (a.s.l.), while two trials
were located within the natural range ( Table 1). At each site, the field
design was a randomised square lattice, type 6×6, with three repeti-
tions and 49 trees per plot planted at 2 × 2 m. For all trials, the soils are
brown forest of high or medium productivity. By this age, the trials
were not thinned or subject to any silvicultural treatments. The rate of
survival ranged from 56 to 70%.
The 33 provenances have been tested in all experiments andcovered, generally, the optimum vegetation conditions for Norway
spruce in Romania, considering that it occurs naturally from approxi-
mately 650 –1800 m a.s.l. ( Sofletea and Curtu 2001 ) and its optimum
climatic area is characterised by an annual mean temperature of4.0 –7.0°C and precipitation of 800 –1200 mm ( Donita et al. 1990 ). Ex-
ceptions are the provenances from low altitude (numbers 7, 14, 20, 25,29 and 30) (Table A.1, Appendix A), which are plantations established
at the beginning of the last century. Thus, the provenances tested in
these trials come from both natural and arti ficial stands situated at
altitudes of 219 –1522 m a.s.l. (Table A.1, Appendix A). The prove-
nances are located in 7 geographical regions and 11 sub –regions, de-
fined as provenance regions for Norway spruce in Romania
(Parnuta et al., 2010 ). The locations of the field trials and provenances
are presented in Fig. 1. Five of the provenances were also used in
analyses of molecular markers.
2.2. Phenotypic and climate data
In the present study, the height growth (H34) and diameter at breast
height (D34) of trees at age 34 were considered the response variables.
Ten trees per repetition from each provenance were measured, for a
total of 990 trees per trial. The respective trees were chosen so as to
belong to the average diameter per plot.
Independent variables, including the climatic conditions of trial
sites as well as origin location of provenances, were represented by
eight climate variables, namely: mean annual temperature (TMA);
mean temperature of the growing season (April to September) (TM
VEG);
mean temperature for January (TM JAN) and July (TM JUL) (i.e., the
coldest and the warmest months, respectively); mean annual pre-
cipitation amount (PMA); mean precipitation amount during the
growing season (PM VEG); mean precipitation of the coldest (PM JAN) and
the warmest (PM JUL) months. The De Martonne index was also com-
puted for each trial site as an indicator of the degree of climate dryness(Marcu 1983 ). Geographical coordinates in terms of the latitude (LAT),
longitude (LONG) and altitude (ALT) of the trial sites and provenanceorigins were determined in the GIS System.
The climatic data used in the present study consisted of monthly
values of precipitation and air temperature. Trial site climatic variables
were obtained by averaging the monthly observations between 1980
and 2013. Climatic variables of provenance locations were calculated
over a 20– year period (1961 –1980). These were calculated using an
upgraded version of ROCADA ( Dumitrescu and Birsan 2015), a daily
gridded climatic dataset covering the Romanian territory. The dataset
used herein consists of a high spatial resolution (1 km × 1 km) and was
made using state –of–the –art interpolation techniques
(Dumitrescu et al. 2016 ,2017) for improved reproduction of the cli-
matic spatial variability. For climate projection data, we used the
output of the CLMcom –CCLM4 –8–17 Regional Climate Model (RCM),
forced by the MPI –M–MPI –ESM –LR run1) Global Circulation Model
(GCM) (mpimet.mpg.de/en/science/models/mpi-esm/) following the
Representative Concentration Pathway (RCP) 4.5 Scenario
(Thomson et al. 2011 ), available at 0.11 ° × 0.11 ° spatial resolution.
The RCP 4.5 scenario assumes moderate changes in air temperature andprecipitation, and stabilising radiative forcing at 4.5 W/m
2in
2100 –without ever exceeding that threshold. This study focuses on the
2021 –2050 and 2071– 2100 time intervals.
2.3. Data analysis
For each trial site and among sites, analyses of variance were per-
formed using the GLM procedure (SPSS v19). The total amount of
variation was divided into provenances, provenance regions and site
components and the interaction between them. Apart from the trial
location, which was considered fixed, all e ffects were considered
random. The assumptions of the model were checked by Shapiro andWilk test for normality and by Levene's test for homogeneity.G. MIHAI, et al. Agricultural and Forest Meteorology 291 (2020) 108040
2

ANOVA was performed as described in the following mixed model:
=+ + + + + + +μ Y RS B P S P B Pe i j k l n kl ijl ji ji j k l n
where: Y ijkln= response (H34 or D34), μ= the overall mean, R k,Sl,Bi,
Pj,S P lj,B P ji, and e ijklnare effect due to the kthregion, lthsite, ithre-
petition (block), jthprovenance, interaction due to lthsite and jthpro-
venance, interaction due to ithrepetition and jthprovenance, and
random error associated with the ijklnthtrees.
For each trait, the provenances phenotypic plasticity was evaluated
by computing the ecovalence, which represents the contribution of each
provenance to the total provenance x site interaction sum of squares.
Therefore, provenances with low ecovalence values have reduced de-
viation from the mean across environments and are more stable.
Ecovalences were computed using the formula of Wricke (1962) :
=− − +W S( X X. X . X .. ) ij i j ij2
where W i= the ecovalence of provenance i, X ij= the mean of prove-
nance i in environment j, X i.= the mean of provenance i across en-
vironments, X. j= the environment mean and X..= the overall mean.
Relationships between growth traits and the environment of pro-
venances locations were investigated by correlations and regressionanalysis. The Pearson correlations based on provenance means werecomputed between the growth traits and the geographical coordinates
of the provenances ’locations for each experiment.
Then, populations transfer functions were developed using H34 and
D34 as dependent variables and the di fference between the climate of
each provenance location and of the trial site as predictors: D_TMA,
D_TM
IAN, D_TM JUL, D_PMA, D_PM IAN, D_PM JUL, etc. ( Mátyás 1994;
Rehfeldt et al. 1999b ;Andalo et al. 2005 ). Because climatic parameters
differ signi ficantly among trial sites, we calculated the transfer func-
tions for each trial site ( Wang et al. 2010 ). To demonstrate the variation
among sites for climatic parameters, a preliminary analysis was per-
formed using t –tests. This approach demonstrated that the variance of
each parameter di ffered from zero (p < 0.001), indicating that the
parameters di ffered signi ficantly among sites. The responses of prove-
nances to climatic transfer were modelled as a quadratic model. Re-gression analysis was constructed based on temperature, precipitation,
and both temperature and precipitation, in accordance with
Mihai et al. (2018) . For selecting the climatic variables, we used the
stepwise selection method of SPSS. The best models were chosen basedon the R
2coefficient.
The growth response functions were also developed to account forTable 1
Geographic and climatic variables for the Norway spruce trial sites.
Trials Vegetation layer Altitude m TMA0CT M JAN0CT M JUL0C PMA mm PM JANmm PM JULmm De Martonne index De Martonne index SEP
Bretcu Spruce stand 1150 4.25 14.38 -5.81 686 101 28 48 34
Nehoiu Spruce stand 1000 4.84 14.79 -4.88 712 103 30 48 34
Bautar Beech stand 750 6.36 16.05 -3.54 881 104 57 54 42
Avrig Beech-sessil oak stand 630 7.68 17.88 -3.25 749 105 34 42 32
Campina Sessil-oak stand 550 8.75 19.43 -1.70 744 100 35 40 29
Tg.Lapus Sessile-oak stand 380 8.04 18.68 -3.39 774 74 64 43 32
TMA – mean annual temperature, TM JAN- mean temperature of the coldest month (January), TM JUL- mean temperature of the warmest month (July), PMA – mean
annual precipitation amount, PM JAN- mean precipitation of the coldest month (January), PM JUL- mean precipitation of the warmest month (July), De Martonne
index –annual De Martonne index, De Martonne index SEP- De Martonne index in September
Fig. 1. Location of the Norway spruce provenances and trial sites. The numbers from 1 to 20 are the codes of the tested provenances, 3-letter codes represent the
populations sampled for DNA analysis, the squares are the trial sites.G. MIHAI, et al. Agricultural and Forest Meteorology 291 (2020) 108040
3

the climate impact on growth performance of the provenances at trial
sites. The climate response of provenances was analysed using a
quadratic model, an approach that was considered to be more appro-
priate ( Schmidtling 1994; Rehfeldt et al. 1999b ;Andalo et al. 2005;
Wang et al. 2010 ). Thus, the provenances ’means for H34 and D34 were
regressed on climate variables of the trial sites.
To assess the potential impact of climate change on the growth and
spatial distribution of Norway spruce, the RCP4.5 scenario was used
over two periods (2021 –2050 and 2071– 2100). First, we used a map
with the current natural distribution range of Norway spruce(Donita et al. 2008). Second, the present mapping of Norway spruce
was validated with respect to air temperature and precipitation rangessuitable for the species. Third, the precipitation and temperature
thresholds were used as constraining factors in order to delineate the
projected areas suitable for Norway spruce over the 2021 –2050 and
2071 –2100 periods. The computations were done using the R language
and ArcGIS software.
2.4. Populations sampling and genotyping
To provide a wider sampling and to cover a broad spectrum of its
natural distribution, at both upper and lower altitudes, 20 Norway
spruce populations were additionally sampled across the Romanian
Carpathians ( Table 2, Fig. 1). Based on data recorded in forest man-
agement plans, these populations were naturally regenerated and arenative to the Romanian Carpathians. They were grouped as CORE
(populations growing in the optimum vegetation conditions) and per-
ipheral populations, located at either upper (UA) or lower (LA) alti-
tudes. Five of these populations have been tested in provenance trials
too.
Bark disks with cambium were collected using a leather punch from
24 to 30 mature trees from each population, with each sampled tree
separated by a distance of minimum 30 meters. The samples were then
dried in silica gel and frozen at -80°C prior to DNA extraction. Total
DNA was extracted from cambium tissue according toDumolin et al. (1995) .
The genotypes of 567 out of 586 individuals were determined using
four genomic microsatellites (gSSR) and three expressed sequenced tag
microsatellites (EST– SSR). These seven nuclear microsatellites were
separated into two multiplex combinations, with each reaction con-taining 7.5 μl Qiagen multiplex PCR bu ffer, 1.5 μl primer mix, 30– 50 ngof genomic DNA, and RNase –free water to a final volume of 15 μl. The
PCR reactions were run in a Palm –Cycler (Corbett) using the following
steps: 15 min at 95°C, followed by 27 cycles for 30 sec at 94°C, 1.30 minat 57°C, 30 sec at 72°C and a final extension step for 30 min at 60°C for
EST –SSR loci, WS0022.B15, WS0092.A19 and WS00716.F13
(Rungis et al. 2004 ).
For SpAG2 ( Pfeiffer et al. 1997 ), EATC1E03, EATC2G05 and
EATC2B02 ( Scotti et al. 2002), an ampli fication reaction was performed
using the following protocol: 15 min at 95°C, followed by 28 cycles for30 sec at 94°C, 1.30 min at 58°C, 60 sec at 72°C and a final extension
step for 30 min at 60°C.
Obtained PCR products were analysed using a GeXP Genetic
Analysis
System (Beckman Coulter Inc., USA). Genotypes were scored
using Genomelab software ver. 10.2.3 (Beckman Coulter Inc., USA).
The microsatellite locus WS0092.A19 was excluded due to inconsistent
banding patterns, although the remaining six were used in subsequent
analyses.
2.5. Genetic diversity and population structure
The frequency of null alleles for each locus and population was
verified using the program Micro –Checker version 2.2.3
(Van Oosterhout et al. 2004 ).
To assess within –population genetic diversity, the average number
of alleles (Na), observed heterozygosity (Ho) and unbiased expected
heterozygosity (He) were calculated using GenAlEx 6.503 ( Peakall and
Smouse 2006 ). Allelic richness (AR) was calculated for the minimum
number of 24 individuals using FSTAT ver. 2.9.3 ( Goudet, 1995).
FSTAT was also used to estimate deviation from the Hardy –Weinberg
equilibrium through the inbreeding coe fficient (FIS) and to test the
significance level of FIS deviation from zero.
A hierarchical AMOVA (Arlequin 3.5.2.2) was performed to test for
a possible e ffect of altitude on genetic diversity. Genetic di fferentiation
among Norway spruce populations was estimated via the computation
of pairwise FST ( Weir and Cockerham, 1984 ) using Arlequin 3.5.2.2
software ( Excoffier and Lischer 2010) and the signi ficance of FST values
was tested by 10000 permutations. The software GenAlEx 6.5(Peakal and Smouse, 2006 ) was used in the graphical representation of
principal coordinates analysis (PCoA).
The Bayesian clustering method implemented in the STRUCTURE
software v 2.3.4 ( Pritchard et al. 2000 ) was used to assess the
Table 2
Location of the Norway spruce populations studied by molecular analyses.
Nr.crt. Populations Code LAT LONG ALT m Type of periphery Geographic region
1. Rodna ROD 47o31′ 24o54′ 1550 UA NC
2. Ceahlau CEAH 46o57′ 25o57′ 1600 UA EC
3. Sinaia SIN 45o21′ 25o30′ 1650 UA CC
4. Paltinis PAL 45o21′ 25o30′ 1600 UA SC
5. Otelul Rosu OTR 45o20′ 22o37′ 1580 UA SW
6. Ruscova RUS 47o54′ 24o36′ 900 CORE NC
7. Moldovita * TOM 47o41′ 25o25′ 1000 CORE NC
8. Crucea CRU 47o17′ 25o31′ 1100 CORE NC
9. Tarcau TAR 46o44′ 26o04′ 850 CORE EC
10. Comandau * COM 45o42′ 26o18′ 900 CORE CC
11. Viforata VIF 45o34′ 26o25′ 950 CORE CC
12. Vidraru VID 45o25′ 24o39′ 1100 CORE SC
13. Valea Cibinului VCB 45o42′ 23o44′ 1250 CORE SC
14. Bistra BIS 45o25′ 22o40′ 1200 CORE SW
10. Sudrigiu * SUD 46o33′ 23o15′ 1050 CORE WC
16. Marginea * MAR 47o49′ 25o41′ 650 LA NC
17. Frasin * FRA 45o28′ 25o43′ 670 LA NC
18. Tulnici TUL 45o58′ 26o33′ 690 LA CC
19. Voila VOI 45o40′ 24o47′ 720 LA SC
20. Lupeni LUP 45o18′ 23o02′ 750 LA SC
UA – Upper altitude; CORE – Core; LA – Law altitude; NC – Northern Carpathians; EC –Eastern Carpathians; CC- Curvature Carpathians; SC –Southern Carpathians; SW
–South –western Carpathians; WC –Western Carpathians (Apuseni Mountains)
⁎- Populations tested in provenance trialsG. MIHAI, et al. Agricultural and Forest Meteorology 291 (2020) 108040
4

individual –based genetic structure. For each value of K (1 –10), we ran
10 replicates with 100,000 iterations after a burn –in period of 100,000
iterations with an admixture ancestral and LOCPRIOR model. To de-
termine the best number of di fferent clusters using the ΔK parameter
(Evanno et al. 2005), the STRUCTURE HARVESTER software was used.
3. Results
3.1. Genotype x environment interaction
For each trial site, as well as across sites, the provenance e ffect was
significant for growth traits. Analysis of variance results across sites are
presented in Table 3. The provenance factor accounted for 4– 6% of
total variance. For the studied traits, di fferences among regions of
provenances were not statistically signi ficant. Most variation was at-
tributed to environmental e ffects, which accounted for 51% of H34
total variance and 48% of D34 total variance. Provenance x site inter-
action was highly signi ficant, suggesting that the growth traits of pro-
venances depend on the planting site environment and could change inother site conditions. The interaction term explained 20 –17% of the
total variation in these traits.
The highest growth performance was obtained at the Bautar trial
situated in a beech layer in the Banat Mountains, where the PMA valuewas the highest of the values among tested sites. The weakest growth
performance was obtained at the Campina trial, situated in a sessile oak
layer in the Curvature Carpathians, where TMA had the highest value
among tested sites and where the PMA and De Martonne index were the
lowest values recorded. The De Martonne aridity index was calculated
for the studied period (1980 –2013), and each trial site recorded values
between 40 to 54, classifying them as humid areas. However, if weconsider only the end of the growing period (September), the values of
the De Martonne aridity index decreased to values similar to the sil-
vostepic climate for the Campina and Tg Lapus trials.
3.2. Phenotypic correlations
Generally, very few signi ficant correlations were found between
growth traits and geographic variables at provenances locations. H34was positively correlated with LAT at the Avrig and Bautar trials and
negatively correlated with ALT at the Campina and Nehoiu trials. D34
was positively correlated with LAT at the Bautar trial and negativelycorrelated with LONG at the Nehoiu trial (Table A.2, Appendix A).
Strong variation patterns were detected between growth traits and ALT
of provenances locations; these patterns did not di ffer in magnitude, but
rather in direction.
3.3. Populations transfer functions
The populations ’transfer functions were calculated for each trait at
every trial site because the provenance x site interaction was sig-
nificant. The best models are presented in Tables 4 and5. The climatic
variables involved in these models were D_TM
JULand D_PM JULin all
sites except the Campina and Bretcu trials, where D_TMA and D_TM JAN
were the explanatory temperature variables, respectively. For both H34and D34, non –significant individual transfer models were obtained at
the Tg. Lapus trial only. This trial is situated at the lowest altitude (380m a.s.l.) and located far outside of the natural distribution of Norway
spruce, in the Lapus Depression of the Northern Carpathians. Partial R
2
values indicated that D_PM JULis the most important climate variable
for total height, whereas both the temperature and precipitation of
provenance locations were equally important for diameter at breast
height. The climatic transfer models show that for locations situated at
lower altitudes, outside of the natural distribution, the best transfer
models were from wetter climates but with small di fferences between
the temperature of provenances location and of the trial site. Forplanting sites situated within the species optimum, a moderate –distance
transfer from warmer climates could increase wood production. Athigher altitudes, such as those of the Bretcu trial, populations must be
adapted for climate conditions with more precipitation and lower
winter temperatures.
3.4. Growth response functions
The growth response functions of Norway spruce over the planting
sites are presented in Table 6. The relationships with TMA and PMA of
the trial sites for both traits were highly signi ficant. In the case of H34,Table 3
Analysis of variance and genotype x environment interaction of the growth traits at 34 years old
Source of variation H34 D34
DF s2% from total variation s2% from total variation
Region of provenance (Rg) 9 5.500 2 4.240 2
Provenance (P) 33 5.053⁎⁎⁎6 3.754 * 4
Site (S) 5 299.088⁎⁎⁎51 311.638⁎⁎⁎48
Repetition (R) 2 12.490 1 1.162 1Interaction P x S 160 3.697
⁎⁎⁎20 3.476⁎⁎17
Error 396 1.712 20 2.463 28
The level of signi ficance is represented as follows:
⁎p < 0.05;
⁎⁎p < 0.01;
⁎⁎⁎p < 0.001.
Table 4
Climatic transfer models for total height of the Norway spruce provenances at 34 years old.
Trial Climatic transfer model Signif. R² Partial R2
Temp.variable Precip.variable
Avrig 20.693 + 0.001 D_PM2
JUL * 0.154 – –
Bautar 22.402 – 0.028 D_TM JUL- 0.0003 D_PM2
JUL⁎⁎0.307 0.018 0.268
Nehoiu 23.247 – 0.036 D_TM JUL- 0.0009 D_PM2JUL* 0.188 0.002 0.152
Campina 17.037 – 0.009 D_TM JUL2+ 0.009 D_PM JUL * 0.152 0.148 0.117
Bretcu 21.445 + 0.025D_TM2JUL* 0.152 – –
D_TM JULis the di fferences in mean temperature of the warmest month (July) between the location of a given provenance and of the trial site; D_PM JULis the
differences in mean precipitations of the warmest month (July) between the location of a given provenance and of the trial site.G. MIHAI, et al. Agricultural and Forest Meteorology 291 (2020) 108040
5

climate variables had nearly similar e ffects and accounted for 34% and
26% of total variation (partial R2), respectively (Fig. A.1, Appendix A).
However, for D34, the trial site TMA was the most important predictor,
accounting for 63% of the variation.
The transfer of populations to warmer environments but with ade-
quate PMA (Bautar trial) resulted in increased growth. However, TMA
and PMA constitute the limiting factors at low altitudes, and mala-
daptation will occur at some time.
Ecovalence analyses has enabled the identi fication of the prove-
nances with high phenotypic plasticity under environmental variation.Considering that the genotypes with zero ecovalence value are stable
across environments, we classi fied the Norway spruce provenances into
three groups: high stability and high growth, high stability and lowgrowth, and with speci ficity for certain environments. Provenances
exhibiting high plasticity (growth over the experiment mean andecovalence near zero and up to a standard deviation) included 3 –Frasin,
8–Pr Bargaului, 2– Dorna Candreni, 26– Valiug, 15– Comandau,
28–Turda, 1– Cosna, 6– Stulpicani, 22– Bistra, 25– Bozovici for H34
(Fig. 2) and 3– Frasin, 26– Valiug, 7– Nasaud, 25– Bozovici, 29– Beius, and
1–Cosna for D34.
3.5. Potential impact of climate change projections
Growth response models were used to predict the impacts of climate
change on the current distribution of Norway spruce in the RomanianCarpathians. Evidently, warming will a ffect the current distribution of
Norway spruce in Romania. The gained areas suitable for Norwayspruce until 2100 are projected to be up to 2%, whereas the lost area is
estimated to be 8% (5% until 2050). Overall, under the RCP 4.5 sce-
nario, a loss of 6% in suitable area for Norway spruce is projected. Fig. 3
presents the changes in species ’present distribution and outlines the
zones with high vulnerability to climate change.
3.6. Genetic diversity
The six nuclear microsatellites loci investigated in the present study
were polymorphic, with 125 di fferent alleles being detected in 567
Norway spruce trees. Micro –Checker software indicated no evidence of
large allele dropout or scoring error due to stuttering. Assuming
Hardy –Weinberg equilibrium, the estimated frequencies of null alleles
over populations range from 0.205 at locus EATC2B02 in the CEAHpopulation to 0.066 at locus SpAG2 in the CRU population. Fourpopulations (TOM, RUS, SIN and VCB) do not have null alleles, whereasCEAH, CRU, ROD, TAR and VID populations show the presence of null
alleles at two loci.
The genetic diversity parameters of the Norway spruce populations
are summarised in Table 7. The average Ν
Αranged from 10.83 to 13.50,
with an average of 11.88. For rare fied allelic richness based on a
minimum of 24 diploid individuals (A R), values ranged from 10.30
(MAR) to 12.69 (PAL), with an average of 11.43. The observed (H o) and
expected heterozygosity (He) for all populations were 0.746 and 0.789,
respectively. A general trend of higher genetic diversity in the high –-
altitude populations and relatively lower diversity in LA populationswas observed. The inbreeding coe fficient ranged from 0.009 in the PAL
population to 0.176 in the VID population, with an overall mean of0.062. F
ISvalues were signi ficantly di fferent from zero in the majority
of populations, with the highest value being observed in CORE popu-lations.
3.7. Population di fferentiation and genetic structure
Genetic di fferentiation among altitudinal groups (F
CT) was sub-
stantially lower and not statistically signi ficant, indicating there is noTable 5
Climatic transfer models for diameter at 1.30 m of the Norway spruce provenances at 34 years old.
Trial Climatic transfer model Signif. R² Partial R2
Temp.variable Precip.variable
Avrig 19.625 + 0.001 D_PM2
JUL * 0.146 – –
Bautar 22.297 –0.050 D_TM JUL- 0.0003 D_PM2
JUL * 0.246 0.029 0.226
Nehoiu 21.750 + 0.068 D_TM2JUL–0.028 D_PM JUL * 0.214 0.141 0.156
Campina 16.150 –0.021 D_TMA2+ 0.016 D_PM JUL⁎⁎0.375 0.304 0.354
Bretcu 22.127 + 0.079 D_TM2IAN+ 0.011D_PM JUL * 0.205 0.187 0.089
D_TMA is the di fferences in mean annual temperature between the location of a given provenance and of the trial site; D_TM JULis the di fferences in mean temperature
of the warmest month (July) between the location of a given provenance and of the trial site; D_TM IANis the di fferences in mean temperature of the coldest month
(January) between the location of a given provenance and of the trial site; D_PM JULis the di fferences in mean precipitations of the warmest month (July) between the
location of a given provenance and of the trial site.
Table 6
Climatic response models for total height and diameter at 1.30 m of the Norway spruce provenances at 34 years old.
Trait Climatic response model p R2Partial R²
Temp. variable Precip. variable
H34 10.577 –0.061TMA2+ 0.018PMA < 0.001 0.412 0.344 0.265
D34 11.709 –0.092TMA2+ 0.017PMA < 0.001 0.692 0.630 0.370
TMA – mean annual temperature; PMA – mean annual precipitation amount
Fig. 2. Ecovalence for total height of the Norway spruce provenances at 34
years old.G. MIHAI, et al. Agricultural and Forest Meteorology 291 (2020) 108040
6

effect of altitude on the genetic diversity of populations at the macro-
spatial level. However, genetic di fferentiation among populations
within groups (F SC) was signi ficant (p<0.001) but explaining only0.76% of the variation within altitudinal groups (Table A.3, Appendix
A).
Based on pairwise F STvalues ( Weir and Cockerham, 1984), there is
also low genetic di fferentiation among Norway spruce populations (see
Fig. 4). F STvalues ranged from 0.006 (between SIN and PAL popula-
tions, both high –altitude populations) to 0.030 (between RUS popula-
tions from the Northern Carpathians and BIS and OTR populations fromthe Southwest Carpathians).
The Bayesian assignment of populations obtained with STRUCTURE
indicated a lack of genetic structure ( Fig. 5) and no separation of the
populations, either altitudinally or by geographic regions. The mostevident pattern detected via principal coordinates analysis (PCoA) was
a separation along the first axis (Coord. 1, Fig. 6) that accounted for
38.94% of the total variation.
4. Discussions
In the present study, four of the six trials were established far out-
side of the species climate distributions, at dry and warm sites. Climatic
variables calculated using ROCADA gridded climatic dataset
(Dumitrescu and Birsan 2015 ) on 1980– 2013 time interval, have shown
an increase in TMA by 3.2 °C, in the Campina trial and 2.5 °C in TgLapus trials, over the climatic optimum, whereas PMA recorded a de-
crease from 26% to 23% from optimum. The De Martonne aridity index
calculated for the growing season or for the last month of it also clearly
demonstrates the xerophytic character of these locations.
The scenarios of the e ffects of climate change forecast a tendency
towards warmer and drier climatic conditions in the next decades in
Romania. Based on RCP 4.5 scenario, the average annual air tempera-
ture in this region is estimated to increase by 1.4°C during 2021 –2050
compared to 1991 –2020 and by 1.8°C until the end of this century.
Therefore, the temperature increases recorded in the last 34 years in the
studied trials allow us to state that the Norway spruce provenances
have already experienced the magnitude of climate change in thesetrials.
The results of the present study highlight that environmental e ffects
have a much higher impact on growth traits at 34 years old than genetic
effects when provenances were transferred among trial sites. The
transfer of provenances to warmer environments with precipitationwithin optimum limits resulted in increased growth. However,
Fig. 3. Changes in natural distribution of Norway spruce until 2100. Yellow colour: present distribution; striped red: 2021 –2050; striped blue: 2071 –2100.
Table 7
Genetic diversity of the Norway spruce populations based on six polymorphic
nSSR loci.
No. Population N N a AR Ho He FIS
1. Rodna (ROD) 24 11.33 11.33 0.694 0.734 0.054
2. Ceahlau
(CEAH)24 11.33 11.25 0.771 0.797 0.025
3. Sinaia (SIN) 32 12.00 11.28 0.781 0.815 0.045
4. Paltinis (PAL) 30 13.50 12.69 0.817 0.810 -0.018
5. Otelul Rosu(OTR)30 12.66 11.98 0.815 0.817 -0.008
Mean UA 12.16 11.70 0.775 0.795 0.019
6. Ruscova(RUS)24 12.00 12.00 0.688 0.719 0.062
7. Moldovita(TOM)30 11.33 10.77 0.728 0.772 0.061
8. Crucea (CRU) 30 12.66 12.00 0.710 0.797 0.117
9. Tarcau (TAR) 30 12.83 12.18 0.760 0.803 0.064
10. Comandau
(CMD)30 12.66 11.81 0.741 0.761 0.008
11. Viforata (VIF) 24 11.33 11.33 0.743 0.787 0.045
12. Vidraru (VID) 30 12.83 12.13 0.697 0.803 0.119
13. Valea
Cibinului(VCB)30 12.33 12.13 0.739 0.776 0.052
14. Bistra (BIS) 24 10.83 10.83 0.765 0.815 0.058
15. Sudrigiu
(SUD)30 10.83 10.32 0.706 0.732 0.023
Mean CORE 11.96 11.55 0.727 0.776 0.060
16. Marginea
(MAR)30 10.83 10.30 0.715 0.735 0.040
17. Frasin (FRA) 30 12.16 11.50 0.758 0.799 0.048
18. Tulnici (TUL) 30 11.16 10.61 0.677 0.719 0.081
19. Voila (VOI) 24 11.83 11.83 0.782 0.785 0.009
20. Lupeni (LUP) 30 11.66 10.98 0.759 0.770 0.006
Mean LA 11.52 11.04 0.738 0.761 0.036
Overall mean 11.88 11.43 0.746 0.777 0.038
N–sample size; Na–average number of alleles; AR–rarefied allelic richness for
24 diploid individuals; Ho–observed heterozygosity; He–expected hetero-
zygosity; FIS–inbreeding coe fficientG. MIHAI, et al. Agricultural and Forest Meteorology 291 (2020) 108040
7

transferring the provenances far beyond the optimum climate to ex-
cessively dry and warm sites (i.e. the Campina trial) resulted in a de-
crease in growth performances of approximately 16% for height and
14% for diameter. Growth response models indicate that the TMA and
PMA of planting sites constitute the climatic drivers of genotype x en-
vironment interaction, particularly at marginal climatic sites.
Similar results have been reported by other studies. Estimates of
height growth loss for a 4 °C increase in temperature caused an 11%
loss in height growth for Norway spruce in Canada ( Schmidtling 1994 )
and a 14% loss for white spruce ( Andalo et al. 2005 ).Bouriaud and
Popa (2009) , and Sofletea et al. (2015) also reported high sensitivity to
temperature and precipitation for the radial growth of Norway sprucein some experimental sites from the Eastern and Southern Carpathians
(Romania).
Moreover, the reaction patterns of provenances di ffer depending on
climatic conditions at each provenance location. The developed transfer
models have highlighted that the di fferences in TM
JULand PM JULbe-
tween the origin location of a given provenance and the trial site werethe main climatic variables explaining the adaptive genetic variation of
Norway spruce provenances, and limit the transfer of forest re-
productive material among provenance regions. Thus, for planting sites
situated within the species optimum, moderate-distance transfer from
warmer climates could increase wood production. For locations si-
tuated at lower altitudes, the best transfer models were from wetter
climates with temperatures close to those of the planting site. At higher
altitudes, populations must also be adapted for colder climates. Similar
results were obtained by Gomöry et al. (2011) for Norway spruce
provenance trials established in Slovakia. The authors demonstratedthat provenances originating from dry sites grow better when planted at
sites with higher precipitation, whereas provenances from wet sitesprefer drier conditions. This tendency of Norway spruce populations to
occupy suboptimal environments was also observed in other species,
including Pinus contorda (Rehfeldt et al. 1999b ), white spruce
(Andalo et al. 2005 ) and Scots pine ( Berlin et al. 2016). Norway spruce
is a wind –pollinated and outcrossing species with little population
differentiation and extensive gene flow. These factors likely reduce the
efficiency of selection and promote the adaptation to newenvironmental conditions.
Correlations with the altitude of provenance locations were nega-
tive for all trial sites (except the Bretcu trial), although results were onlystatistically signi ficant for two locations. In contrast, at the Bretcu
trial –situated at the highest altitude among tested sites –growth ex-
hibited a positive correlation with the altitude of the provenance lo-
cation. Provenances from low altitudes (219 –688 m a.s.l., Table A.1.
Appendix A), which are actually plantations of unknown origin, es-tablished at the beginning of the 20th century, obtained good growth
performances in all trials situated outside of the natural area. Recent
studies, have demonstrated that the day length and temperature during
the embryo formation and maturation can infl uence adaptive traits of
the next tree generation, in particular when the maternal parents weretransferred from cold to warmer conditions. This mechanism is termed
epigenetics and is well known in Norway spruce ( Johnsen et al. 2005;
Skrøppa et al. 2010 ). Compared to selection, which acts slowly, epi-
genetic mechanism is a rapid process that does not involve any changes
to DNA sequences.
Rehfeldt et al. (1999a) suggested that optimising growth and sur-
vival rate depends on matching provenance and planting site climates
for the general temperature regime, the coldness of the winter, the
strength of continental e ffects and the balance between temperature
and precipitation. To avoid the potential maladaptation of forest re-productive material, reforestation programmes must rely on climate –-
based seed transfer guidelines and on more resilient genotypes.
Assessment of the ecovalence allowed us to identify provenances
with high phenotypic plasticity and those more vulnerable to un-predictable environmental fluctuations. Thus, provenances from the
Northern Carpathians, Banat Mountains and some provenances fromthe Apuseni Mountains have obtained the best growth and most stableperformances. There are also some provenances from low elevations
that exhibited high plasticity (29 –Beius and 25– Bozovici). Overall, the
analysis of ecovalence highlighted a moderate phenotypic plasticity forstudied provenances. Price et al. (2003) has shown that a moderate
level of phenotypic plasticity is optimal for population survival andgenetic evolution in a new environment.
The results revealed that Norway spruce is a species sensitive to
Fig. 4. Genetic di fferentiation of the Norway spruce populations.
Fig. 5. Genetic structure of the Norway spruce po-
pulations.G. MIHAI, et al. Agricultural and Forest Meteorology 291 (2020) 108040
8

increasing temperature and water de ficit. These observations are con-
gruent with the results obtained by Bouriaud and Popa (2009) ,
Lévesque et al. (2013) andZang et al. (2014) . Climatic projections for
South –eastern Europe will pose severe pressures for this species. Major
changes in wood production are expected to occur in the climatically
marginal environments of species distribution. Populations in these
areas recorded the lowest genetic diversity within populations and were
most heavily a ffected by disturbance factors. Thus, the most vulnerable
populations will be those from the edges of the Eastern Carpathians,Curvature Carpathians, Banat Mountains and the interior of the Car-
pathian chain. Conversely, warming at higher elevation could increase
productivity, mainly due to improving growing for local populations or
through the transfer of populations from low elevation. The results of
our study demonstrate that populations from UA possessed large
within –genetic diversity and could bu ffer the negative e ffects of climate
change.
The level of Norway spruce genetic diversity obtained in this region
is comparable with other studies ( Meloni et al. 2007;
Nowakowska et al. 2009 ; Tollefsrud et al.2009; Unger et al. 2011;
Schueler et al. 2013 ;Cvjetkovic et al. 2017) and generally shows a high
genetic variation and low degree of genetic di fferentiation. The ex-
pected heterozygosity was higher than the value obtained for 37 northEuropean populations (He = 0.77 compared to He = 0.64) (Tollefsrud
et al.2009) and for populations from Bosnia and Herzegovina
(He = 0.77 compared to He = 0.67) ( Cvjetkovic et al. 2017 ). Similar
results were obtained for 17 populations in the Czech Republic, whereHe = 0.78 ( Máchová et al. 2018). Our data show that high altitude
populations have both allelic richness and expected heterozygosityhigher than CORE and LA populations. Maghuly et al. (2006) reported
the same results in di fferent subpopulations of Norway spruce in Aus-
tria and Ciocîrlan et al. (2017) in European beech from the Romanian
Carpathians.
The genetic di fferentiation between populations is low, but is con-
sisted with the geographical trend in genetic diversity for DNA markers
in Norway spruce, i.e. for the Austrian populations FST = 0.002
(Unger et al. 2011 ), the Czech populations FST is between 0.006 to
0.027 ( Máchová et al. 2018) and the northern spruce populations
FST = 0.026 (Tollefsrud et al.2009). Even for a larger distribution area
of Norway spruce, for instance, the genetic divergence among the
northern and southern domain is only 0.024 (Tsuda et al.2016).
In the present study, the weak di fferentiation found among geo-
graphic regions and the relatively high di fferentiation between geo-
graphically close populations can be explained by the population'shistory; not only by postglacial migration pattern but also human in-
tervention. The principal coordinates analysis revealed that the popu-
lations can be separated into two slightly di fferentiated groups: one
consists of the most north –eastern populations and the other the
southern ones (exceptions are PAL, BIS, TAR and TUL populations).
Recent palaeoecological studies or based on the combined analysis of
DNA and fossil pollen have shown that important refugia for Norway
spruce existed in the Romanian Carpathians, both in the northern part
of the Eastern Carpathians ( Tollefsrud
et al. 2008 ) or in the Southern
Carpathian ( Feurdean et al. 2007 ). Further investigations are needed
regarding the impact of historical translocation of forest reproductivematerial on genetic diversity of Norway spruce in this region.
Despite the few populations studied in both quantitative genetics
and DNA analysis it can be observed a relationship between the adap-
tive capacity of Norway spruce provenances highlighted in field trials
and the parameters of within population genetic diversity. Thus, theFrasin (3) and Comandau (15) provenances obtained the best growth
and most stable performance, displayed high phenotypic plasticity and
also had the highest values of genetic diversity based on six micro-
satellite markers. Moreover, the Sudrigiu (31) and Marginea (4) po-
pulations, which showed the lowest levels of genetic diversity, also
exhibited weak phenotypic plasticity. Additional researches are cer-
tainly needed by combining genotype and phenotype analyses to un-
derstand the genetic basis of adaptation to climate extremes.
The results of the present study suggest that adaptation strategies
could rely on assisted population migration and the selection of more
resilient genotypes. However, the transfer of forest reproductive ma-
terial should be moderate, from latitudinal –adjacent provenance re-
gions or from lower altitudinal distances. In this manner, gradual andmoderate transfer will allow genotypes to adapt to changing climate
conditions.
5. Conclusions
This study highlights the high adaptive genetic variation of Norway
spruce in South –eastern Europe. Norway spruce populations from this
region hold high genetic diversity, which will allow it to cope with the
negative e ffects of climate warming. The level of genetic diversity and
current genetic structure is primarily the result of species evolutionfollowing the last glaciation, most likely from some refugia located in
the Romanian Carpathians and is closely related to the history of
Europe and of this region.
The transfer to warmer climates resulted in increased provenance
Fig. 6. Results of principal coordinates analysis for the Norway spruce populations. Populations were marked as follow: red –Northern and Eastern Carpathians;
blue –Curvature Carpathians; green –Southern Carpathians; yellow –Southwestern Carpathians; pink –Western Carpathians (Apuseni Mountains)G. MIHAI, et al. Agricultural and Forest Meteorology 291 (2020) 108040
9

growth in locations situated within the ecological optimum. However,
transfer far beyond the optimum climate at excessively dry and warm
sites resulted in a 14% to 16% decrease in growth. TMA and PMA of the
planting site as well as di fferences in TM JULand PM JULfrom the origin
location are the climatic factors associated with risk when transferringforest reproductive materials.
This study emphasises the importance of maintaining and mon-
itoring genetic diversity in forest ecosystems and development of the
association studies between molecular markers, quantitative traits and
climate factors. Finally, the paper provides solutions to some problems
encountered by forest practice, such as selection and management of
the seed stands, elaboration of the transfer rule for forest reproductive
materials and highly vulnerable areas for Norway spruce regeneration.
Contributions
Conception: Georgeta Mihai; methodology: Georgeta Mihai, Maria
Teodosiu, Marius Victor Birsan; acquisition of the data: Maria Teodosiu,
Marius Victor Birsan, Ionel Mirancea, Ecaterina Nicoleta Apostol, Paula
Garbacea; quantitative data analysis: Georgeta Mihai, Alin Madalin
Alexandru, Lucia Ionita; molecular markers analyses: Maria Teodosiu;
climate modelling: Marius Victor Birsan; wrote the paper: Georgeta
Mihai, Maria Teodosiu, Marius Victor Birsan.
Funding
this work was financed by the Executive Agency for Higher
Education, Research, Development and Innovation Funding in the fra-
mework of the projects: PN-II-PC-PCCA-2013-4-0695 (PartnershipProgramme) and PN 19070303 (Nucleu Programme).
Declaration and veri fication
The authors declare that the paper has been not published else-
where, and not include any form of plagiarism. All the authors men-
tioned above have approved the manuscript and have agreed with the
submission of the manuscript to Agricultural and Forest Meteorology.
Declaration of Competing Interest
The authors declare that there is no con flict of interest regarding the
publishing of the paper by Agricultural and Forest Meteorology,Acknowledgements
This research was carried out within the framework of the GENCLIM
project (Evaluating the adaptive genetic potential of the main con-
iferous species for the sustainable forest management in the context of
climate change, grant number PN –II–PC–PCCA –2013– 4–0695) and PN
19070303 project (Reviewing the provenance regions for productionand use of forest reproductive materials in Romania in order to increase
the adaptive capacity of forest ecosystems to climate change), financed
by the Executive Agency for Higher Education, Research, Development,and Innovation Funding in Romania. We would like to thank the editor
and anonymous reviewers for their useful advice, which helped to im-
prove the manuscript.
We acknowledge the World Climate Research Programme's Working
Group on Regional Climate, and the Working Group on Coupled
Modelling. We also thank the climate modelling groups CLMcom
Consortium and the Max– Planck Institute (Climate Service Center,
Hamburg) for producing and making their model output available. We
also acknowledge the Earth System Grid Federation infrastructure, the
European Network for Earth System Modelling and other partners in the
Global Organisation for Earth System Science Portals (GO –ESSP).Supplementary materials
Supplementary data associated with this article can be found, in the
online version, at doi: 10.1016/j.agrformet.2020.108040 .
References
Andalo, C., Beaulieu, J., Bousquet, J., 2005. The impact of climate change on growth of
local white spruce populations in Quebec. Forest Ecol. Manag 205, 169 –182.
Beaulieu, J., Perron, M., Bousquet, J., 2004. Multivariate patterns of adaptive genetic
variation and seed sources transfer in Picea mariana . Can. J. Forest Res 34, 531 –545.
Berlin, M.E., Persson, T., Jansson, G., Haapanen, M., Ruotsalainen, S., Bärring, L.,
Andersson, Gull, 2016. Scots pine transfer e ffect models for growth and survival in
Sweden and Finland. Silva Fenn 50 (3), 1 –21.
Beuker, E., Valtonen, E., Repo, T., 1998. Seasonal variation in the frost hardiness of Scots
pine and Norway spruce in old provenance experiments in Finland. Forest Ecol.
Manag. 107, 87 –98.
Birsan, M.V., Marin, L., Dumitrescu, A., 2013. Seasonal changes in wind speed in
Romania. Rom. Rep. Phys. 65 (4), 1479 –1484 .
Bojariu, R., Birsan, M.V., Cic ă, R., Velea, L., Burcea, S., Dumitrescu, A., Dasc ălu, S.I.,
Gothard, M., Dobrinescu, A., C ărbunaru, F., Marin, L., 2015. Schimb ările climati-
ce–de la bazele fizice la riscuri și adaptare [Climatic changes –from physical basis to
risk and adaptation]. Printech, Bucharest, pp. 200pp .
Bouriaud, O., Popa, I., 2009. Comparative dendroclimatic study of Scots pine, Norway
spruce, and silver fir in the Vrancea Range. Eastern Carpathian Mountains. Trees 23,
95–106.
Bouriaud, O., Leban, J.M., Bert, D., Deleuze, C., 2005. Intra –annual variations in climate
influence growth and wood density of Norway spruce. Tree Physiol 25 (6), 651 –660.
Burczyk, J., Giertych, M., 1991. Response of Norway spruce ( Picea abies (L.) Karst.)
annual increments to drought for various provenances and locations. Silvae Genet 40
(3/4), 146 –152.
Busuioc, A., Birsan, M.V., Carbunaru, D., Baciu, M., Orzan, A., 2016. Changes in the
large –scale thermodynamic instability and connection with rain shower frequency
over Romania. Veri fication of the Clausius– Clapeyron scaling. Int. J. Climatol. 36 (4),
2015 –2034 .
Chen, J., Li, L., Milesi, P., Jansson, G., Berlin, B., Karlsson, B., Aleksic, J., Vendramin,
G.G., Lascoux, M., 2019. Genomic data provides new insights on the demographichistory and the extent of recent material transfers in Norway spruce. Evol.
Applications 8, 1539 –1551 .
Cheval, S., Dumitrescu, A., Birsan, M.V., 2017. Variability of the aridity in the
South –Eastern Europe over 1961 –2050.
Catena 151, 74 –86.
Ciocîrlan, E., So fletea, N., Ducci, F., Curtu, A.L., 2017. Patterns of genetic diversity in
European beech ( Fagus sylvatica L.) at the eastern margins of its distribution range.
iForest 10, 916 –922.
Curtu, L.A., Șofletea, N., Radu, R., Bacea, A., Abrudan, I.V., Butiuk– Keul, A., Farcas, S.,
2009. Allozyme variation of coniferous tree species from Maramures Mountains,
Romania. Not. Bot. Horti Agrobo 37 (2), 245 –251.
Cvjetkovic, B., Konnert, M., Fussi, B., Mataruga, M., Šijačič–Nikoli ć, M., Danič ić, V.,
Lučić, A., 2017. Norway spruce ( Picea abies Karst.) variability in progeny test in
Bosnia and Herzegovina. Genetika 49, 259 –272.
Donita, N., Chirita, C.D., Stanescu, V., 1990. The types of forest ecosystems. Editura
Tehnica Silvica 496p .
Donita, N., Bandiu, C., Biris, I.A., Gancz, V., Apostol, J., Marcu, C., 2008. Harta padur-
ilor –pe unitati ecosistemice in Romania (scara 1:500 000) [Forest Map of
Romania –on forest ecosystem types, scale 1: 500 000]. Editura Silvica, Bucharest (in
Romanian) .
Dumitrescu, A., Birsan, M.V., 2015. ROCADA: a gridded daily climatic dataset over
Romania (1961 –2013) for nine meteorological variables. Natural Hazards 78 (2),
1045 –1063 .
Dumitrescu, A., Birsan, M.V., Manea, A., 2016. Spatio –temporal interpolation of sub– –
daily (6 –hour) precipitation over Romania for the period 1975 –2010. Int. J. Climatol.
36 (3), 1331 –1343 .
Dumitrescu, A., Birsan, M.V., Nita, I.A., 2017. A Romanian daily high –resolution gridded
dataset of snow depth (2005– 2015). Geo fizika 34 (2), 275 –295.
Dumolin, S., Demersure, B., Petit, R.J., 1995. Inheritance of chloroplast and mitochon-
drial genomes in pedunculate oak investigated with an e fficient PCR method. Theor.
Appl. Genet. 91, 1235 –1256 .
Enescu,
V., 1988. Cercetari privind variabilitatea genetica inter si intrapopulationala la
molid ( Picea abies L. Karst) [Research on inter and intrapopulational genetic varia-
bility of Norway spruce]. Redactia de propaganda tehnica agricola 115p (in
Romanian) .
Evanno, G., Regnaut, S., Goudet, J., 2005. Detecting the number of clusters of individuals
using the software STRUCTURE: a simulation study. Mol. Ecol. 14, 2611 –2620 .
Excoffier, L., Lischer, H.E.L., 2010. Arlequin suite ver 3.5: A new series of programs to
perform population genetics analyses under Linux and Windows. Mol. Ecol. Resour.
10, 564 –567.
Feurdean, A., Wohlfarth, B., Björkman, L., Tantau, I., Bennike, O., Willis, K.J., Farcas, S.,
Robertsson, A.M., 2007. The in fluence of refugial population on Lateglacial and early
Holocene vegetational changes in Romania. Review of Palaeobotany and Palynology
145, 305 –320.
George, J.P., Grabner, M., Campelo, F., Karanitsch –Ackerl, S., Mayer, K., Klumpp, R.T.,
Schueler, S., 2019. Intra –specifi c variation in growth and wood density traits under
water –limited conditions: Long –term, short –term, and sudden responses of four
conifer tree species. Sci. Total Environ. 660, 631 –643.
Gomöry, D., Longauer, R., Hlasny, T., Pacalaj, M., Strmen, S., Krajmerova, D., 2011.
Adaptation to common optimum in di fferent populations of Norway spruce ( Picea
abies Karst.). Eur. J. Forest Res 131, 401 –411.
Gomöry, D., Longauer, R., Paule, L., Krajmerova, D., Schmidtova, J., 2010. Across –speciesG. MIHAI, et al. Agricultural and Forest Meteorology 291 (2020) 108040
10

patterns of genetic variation in forest trees of Central Europe. Biodivers. Conserv 19,
2025 –2038 .
Goudet, J., 1995. FSTAT (version 1.2): a computer program to calculate F –statistics. J.
Hered. 86, 485 –486.
Hampe, A., Petit, R.J., 2005. Conserving biodiversity under climate change: the rear edge
matters. Ecology Letters 8 (5), 461 –467.
Jansen, S., Konrad, H., Geburek, T., 2017. The extent of historic translocation of Norway
spruce forest reproductive material in Europe. Ann. Forest Sci. 74 (56), 17 pp .
Johnsen, Ø., Daehlen, O.G., Ostreng, G., Skrøppa, T., 2005. Day length and temperature
during seed production interactively a ffect adaptive performance of Picea abies pro-
genies. New Phytol 168, 589 –596.
Kapeller, S., Lexer, M.J., Geburek, T., Hiebl, J., Schueler, S., 2012. Intraspeci fic variation
in climate response of Norway spruce in the eastern Alpine range: Selecting appro-
priate provenances for future climate. Forest Ecol. Manag. 271, 46 –57.
Klimo, E., Hager, H., Kulhavý, J., 2000. Spruce monocultures in Central Europe –problems
and prospects. EFI Proc 33, 5 –9.
Konnert, M., 2009. Genetic variation of Picea abies in southern Germany as determined
using isozymes and STS markers. Dendrobiology 61, 131 –136.
Lagercrantz, U., Ryman, N., 1990. Genetic structure of Norway spruce ( Picea abies ):
concordance of morphological and allozymic variation. Evolution 44 (3), 8 –53.
Lévesque, M., Saurer, M., Siegwolf, R., Eilmann, B., Brang, P., Bugmann, H., Rigling, A.,
2013. Drought response of five conifer species under contrasting water availability
suggests high vulnerability of Norway spruce and European larch. Global Change Biol
19 (10), 3184 –3199 .
Lindner, M., Maroschek, M., Netherer, S., Kremer, A., Barbati, A., Garcia –Gonzalo, J.,
Seidl, R., Delzon, S., Corona, P., Kolstrom, M., Lexer, M.J., Marchetti, M., 2010.
Climate change impacts, adaptive capacity and vulnerability of European forestecosystems. Forest Ecol. Manag. 259, 698 –709.
Máchová, P., Tr čková, O., Cvr čková, H., 2018. Use of Nuclear Microsatellites Loci for
Evaluating Genetic Diversity of Selected Populations of Picea abies (L.) Karsten in the
Czech Republic. Forests 9 (92), 1 –15.
Maghuly, F., Pinsker, W., Praznik, W., Fluch, S., 2006. Genetic diversity in managed
subpopulations of Norway spruce ( Picea abies (L.) Karst). For. Ecol. Manag 222,
266 –271.
Manea, A., Birsan, M.V., Tudorache, G., Carbunaru, F., 2016. Changes in the type of
precipitation and associated cloud types in Eastern Romania (1961 –2008). Atmos.
Res. 169, 357 –365.
Marcu,
M., 1983. Meteorologie si Climatologie Forestiera (Forestry Meteorology and
Climatology). Editura Ceres 250 p .
Matyas, C., 1994. Modeling climate change e ffects with provenance test data. Tree
Physiol 14 (7 –9), 797 –804.
Meinshausen, M., Smit, S.J.H., Calvin, K., Daniel, J.S., Kainuma, M.L.T., Lamarque, J.F.,
Matsumoto, K., Montzka, S., Raper, S., Riahi, K., Thomson, A., Velders, G.J.M., van
Vuuren, D.P., 2011. The RCP greenhouse gas concentrations and their extensions
from 1765 to 2300. Climatic Change 109, 213 –241.
Meloni, M., Perini, D., Binelli, G., 2007. The distribution of genetic variation in Norway
spruce ( Picea abies (L.) Karst.) populations in the western Alps. J. Biogeogr. 34,
929 –938.
Mihai, G. (Ed.), 2009. Surse de seminte testate pentru principalele specii de arbori for-
estieri din Romania [Tested seed sources for the main forest species in Romania].
Editura Silvica, Bucharest, pp. 280 p .
Mihai, G., Birsan, M.V., Dumitrescu, A., Alexandru, A., Mirancea, I., Ivanov, P., Stuparu,
E., Teodosiu, M., Daia, M., 2018. Adaptive genetic potential of European silver firi n
Romania in the context of climate change. Ann. Forest Res. 61 (1), 95 –108.
Milesi, P., Berlin, M., Chen, J., Orsucci, M., Li, L., Jansson, G., Karlson, B., Lascoux, M.,
2019. Assessing the potential for assisted gene flow using past introduction of
Norway spruce in Southern Sweden: Local adaptation and genetic basis of quanti-tative traits in trees. Evol. Applications 2, 1946 –1959 .
Nowakowska, J.A., 2009. Mitochondrial and nuclear DNA di fferentiation of Picea abies
populations in Poland. Dendrobiology 61, 119 –129.
Pakalaj, M., Longauer, R., Krajmerova, D., Gomöry, D., 2002. E ffect of site altitude on the
growth and survival of Norway spruce ( Picea abies L.) provenances on the Slovak
plots of IUFRO experiment 1972. J. Forest Sci 48 (1), 16 –26.
Parnuta, G., Lorent, A., Tudoroiu, M., Petrila, M., 2010. Regiunile de provenienta pentru
materialele de baza din care se obtin material forestiere de reproducere din Romania
[The provenance regions for the basic materials established in Romania]. Editura
Silvica 122p .
Peakall, R.O., Smouse, P.E., 2006. Genalex 6: Genetic analysis in Excel. Population ge-
netic software for teaching and research. Mol. Ecol. Notes 6 (1), 288 –295.
Pfeiff er, A., Olivieri, A.M., Morgante, M., 1997. Identi fication and characterization of
microsatellites in Norway spruce ( Picea abies K.). Genome 40, 411 –419.
Price, T.D., Qvarnstrom, A., Irwin, D.E., 2003. The role of phenotypic plasticity in driving
genetic evolution. In: Proceedings of the Royal Society London. 270. pp. 1433– 1440 .
Pritchard, J.K., Stephens, M., Donnelly, P., 2000. Inference of population structure using
multilocus genotype data. Genetics 155, 945 –959.
Radu, R.G., Curtu, L.A., Spârchez, G., Șofletea,
N., 2014. Genetic diversity of Norway
spruce [ Picea abies (L.) Karst.] in Romanian Carpathians. Ann. Forest Res 57 (1),
19–29.
Rehfeldt, G.E., Tchebakova, N.M., Barnhardt, L.K., 1999a. E fficacy of climate transfer
functions: Introduction of Eurasian populations of Larix into Alberta. Can. J. Forest
Res 29, 1660 –1668 .
Rehfeldt, G.E., Ying, C.C., Spittlehouse, D.L., Hamilton, D.A., 1999b. Genetic responses to
climate in Pinus contorda : niche breadth, climate change, and reforestation.
Ecological Monographs 69, 375 –407.
Rungis, D., Berube, Y., Zhang, J., Ralph, S., Ritland, C.E., Ellis, B.E., Douglas, C.,
Bohlmann, J., Ritland, K., 2004. Robust simple sequence repeat markers for spruce (Picea spp.) from expressed sequence tags. Theor. Appl. Genet 109, 1283 –1294 .
Savolainen, O., Pyhajärvi, T., Knurr, T., 2007. Gene flow and local adaptation in trees.
Annu. Rev. Ecol, Evol. S. 38, 595 –619.
Schmidtling, R.C., 1994. Use of Provenance Tests to Predict Response to Climate Change:
Loblolly Pine and Norway Spruce. Tree Physiol 14, 805 –817.
Schueler, S., Kapeller, S., Konrad, H., Geburek, T., Mengl, M., Bozzano, M., Koskela, J.,
Lefevre, F., Hubert, J., Kraigher, H., Longauer, R., Olrik, D., 2013. Adaptive geneticdiversity of trees for forest conservation in a future climate: a case study on Norway
spruce in Austria. Biodivers. Conserv. 22, 1151 –1166 .
Scotti, I., Magni, F., Paglia, G.P., Morgante, M., 2002. Trinucleotide microsatellites in
Norway spruce ( Picea abies ): Their features and the development of molecular
markers. Theor. Appl. Genet 106, 40 –50.
Sidor, C.H., Camarero, J.J., Popa, I., Badea, O., Apostol, E.N., Vlad, R., 2019. Forest
vulnerability to extreme climatic events in Romanian Scots pine forests. Sci. Total
Env. 678, 721 –727.
Skroppa, T., Tollefsrud, M.M., Sperisen, C., Johnsen, O., 2010. Rapid change in adaptive
performance from one generation to the next in Picea abies –Central European trees in
a Nordic environment. Tree Genet. Genomes 6 (1), 93 –99.
Sofletea, N., Curtu, L., 2001. Dendrologie [Dendrology], Editura pentru viata. Brasov
296 pp .
Sofletea, N., Curtu, A.L., Daia, M.L, Budeanu, M., 2015. The dynamics and variability of
radial growth in provenance trials of Norway spruce ( Picea abies L. Karst) within and
beyond the hot margins of its natural range. Not Bot Horti Agrobo 43 (1), 265 –271.
Stojnic, S., Avramidou, E.V., Fussi, B., Westergren, M., Orlovi ć, S., Matovi ć, B., Trudi ć, B.,
Kraigher, H., Aravanopoulos, F.A., Konnert, M., 2019. Assesement of GeneticDiversity and Population Genetic Structure of Norway Spruce ( Picea abies (L.)
Karsten)
at Its Southern Lineage in Europe. Implications for Conservation of Forest
Genetic Resources. Forest 10, 258 .
Suvantoa, S., Nöjda, P., Henttonenb, H.M., Beuker, E., Mäkinena, H., 2016. Geographical
patterns in the radial growth response of Norway spruce provenances to climaticvariation. Agr. Forest Meteorol. 222, 10 –20.
Teodosiu, M., 2011. Cercetari privind variabilitatea genetica in arboretele de molid din
Obcinele Bucovinei [Research regarding genetic variability in Norway spruce stands
from Obcinele Bucovinei]. Editura Silvica 150 pp .
Thomson, A.M., Calvin, K.V., Smith, S.J., Kyle, G.P., Volke, A., Patel, P., Delgado –Arias,
S., Bond– Lamberty, B., Wise, M.A., Clarke, L.E., 2011. RCP4.5: a pathway for stabi-
lization of radiative forcing by 2100. Clim. Change 109, 77 –94.
Tollefsrud, M.M., Kissling, R., Gugerli, F., Johnsen, O., Skrøppa, T., Cheddadi, R., Van der
Knaap, W.O., Lataàowa, M., Terhürne –Berson, R., Litt, T., Geburek, T., Brochmann,
C., Sperisen, C., 2008. Genetic consequences of glacial survival and postglacial co-lonization in Norway spruce: combined analysis of mitochondrial DNA and fossil
pollen. Mol. Ecol. 17, 4134 –4150 .
Tollesfrud, M.M., Sonstebo, J.H., Brochmann, C., Johnsen, Ø., Skroppa, T., Vendramin,
G.G., 2009. Combined analysis of nuclear and mitochondrial markers provides new
insight into the genetic structure of North European Picea abies . Heredity 102,
549 –662.
Trujillo –Moya, C., George, J.P., Fluch, S., Geburek, T., Grabner, M., Karanitsch –Ackerl, S.,
Konrad, H., Mayer, K., Sehr, E.M., Wischnitzki, E., Schueler, S., 2018. Drought
Sensitivity of Norway Spruce at the Species ’Warmest Fringe: Quantitative and
Molecular Analysis Reveals High Genetic Variation Among and Within Provenances.G3–Genes Genom. Genet 8, 1225 –1245 .
Tsuda, Y., Chen, J., Stocks, M., Källman, T., Sønstebø, J.H., Parducci, L., Semerikov, V.,
Sperisen, C., Politov, D., Ronkainen, T., Vā liranta, M., Vendramin, G.G., Tollesfrud,
M.M., Lascoux, M., 2016. The extent and meaning of hybridization and introgressionbetween Siberian spruce ( Picea obovata ) and Norway spruce ( Picea abies ): cryptic
refugia as stepping stones to the west? Molecular Ecology 25, 2773 –2789 .
Ulbrichova, I., Podrazsky, V., Olmez, Z., Beran, F., Prochazka, J., Fulin, M., Kubecek, J.,
Zahradník, D., 2013. Growth performance of Norway spruce in the Czech –German
provenance trial plot Lede č. Sci. Agr. Bohemica 44 (4), 223 –231.
Unger, G.M., Konrad, H., Geburek, T., 2011. Does spatial genetic structure increase with
altitude? An answer from Picea abies in Tyrol. Austria. Plant Syst. Evol. 292, 133 –141.
Ununger, J., Ekberg, I., Kang, H., 1988. Genetic Control and Age –related Changes of
Juvenile Growth Characters in Picea abies . Scand. J. Forest Res 3 (1), 55 –66.
Van Oosterhout, C., Hutchinson, W.F., Willis, D.P.M., Shipley, P., 2004. Micro –checker:
software
for identifying and correcting genotyping errors in microsatellites data. Mol.
Ecol. Notes 4, 535 –538.
Vendramin, G.G., Anzidei, M., Madaghiele, A., Sperisen, C., Bucci, G., 2000. Chloroplast
microsatellite analysis reveals the presence of population subdivision in Norwayspruce ( Picea abies K.). Genome 43, 68 –78.
Wang, T., Hamann, A., Yanchuk, A., O'Neill, G.A., Aitken, S.N., 2006. Use of response
functions in selecting lodgepole pine populations for future climates. Global Change
Biol 12, 2404 –2416 .
Wang, T., O'Neill, G.A., Aitken, S.N., 2010. Integrating environmental and genetic e ffects
to predict responses of tree populations to climate. Ecol. Appl. 20 (1), 153 –163.
Wang, X., Bernhardsson, C., Ingvarsson, P.K., 2020. Demography and Natural Selection
Have Shaped Genetic Variation in the Widely Distributed Conifer Norway Spruce (Picea abies ). Genome Biol. Evol 12 (2), 3803 –3817 .
Weir, B.S., Cockerham, C.C., 1984. Estimating F –statistics for the analysis of population
structure. Evolution 38, 1358 –1370 .
Westergren, M., Bozic, G., Kraigher, H., 2018. Genetic diversity of core vs. peripheral
Norway spruce native populations at a local scale in Slovenia. iForest 11, 104 –110.
Wricke, S.G., 1962. Ube eine mthodezur Erfussung der Okologischen streubreite in
Feldversuchen. Z. P flanzenzuechi 47, 92 –96.
Zang, C., Hartl –Meier, C., Dittmar, C., Rothe, A., Menzel, A., 2014. Patterns of drought
tolerance in major European temperate forest trees: Climatic drivers and levels ofvariability. Global Change Biol 20 (12), 3767 –3779 .G. MIHAI, et al.
Agricultural and Forest Meteorology 291 (2020) 108040
11