the influence of site characteristics on beech growth in
TRANSCRIPT
Volume 17(3), 17- 26, 2013 JOURNAL of Horticulture, Forestry and Biotechnology
www.journal-hfb.usab-tm.ro
17
The influence of site characteristics on beech growth in northeastern Romania Lupescu M.1*, Curca M.1
1University “Ștefan cel Mare”, Faculty of Forestry, Suceava, 13 Universității Street, 720229, Suceava,
Romania *Corresponding author. Email: [email protected]
Abstract The influence of orographic factors on beech growth (Fagus sylvatica L.) was studied in stands located in north - eastern Romania. 52 stands were analyzed. The studied stands range from 45 degrees northern latitude to 48 degrees northern latitude and 25 degrees eastern longitude to 28 degrees eastern longitude. Statistical parameters of growth series such as the correlation between trees in a stand, the average size of annual tree ring, standard deviation and altitude, slope exposure and latitude influence on beech growth were analyzed for the period 1959-2008. Correlation study showed that this is quite low on large surfaces of the study area. The highest values are found in the eastern half of Suceava county. The highest values of radial growth are achieved in areas with average altitude that provide a favorable rainfall treatment, but there are also exceptions such as areas with low altitude from Botosani county where beech found optimal conditions for growth. Standard deviation correlated with altitude. The natural tendency is to achieve smaller variations in annual tree ring size at low altitudes, higher variations in annual tree ring size at average altitudes and lower variations of the annual tree ring size at high altitudes. The analysis of the spatial variation of the analyzed series according to altitude led to the formation of three groups. The first group consists of growth series from high altitudes - 900 m, 1000 m, 1100 m and 1200 m The second group consists of series from middle altitudes that range from 500 m to 800 m. The third group consists of growth series achieved at low altitudes - 200 m, 300 m and 400 m The explanation for series spacing is related to the environmental conditions in which these series were formed. At low and high altitudes, slope exposition has a stronger influence, leading to environmental factors changes with a larger impact on growth. The slope exposition at medium altitudes has a lower influence on growth. Hierarchical analysis (cluster) of the series led to grouping according latitude.
Key words spruce, variability, radial growth, altitudinal gradient
The annual tree ring width varies from year to
year due to the environmental factors influence. Their
variation can be annual or periodical and their
influence may reflect on the growth achieved in one or
several years. The correlation between the annual tree
ring size and environmental factors is influenced by the
species ecological requirements or by factors from the
extreme years 16
.
Climate and the orographic characteristics of
sites are the most important factors of tree growth in
any environments where the forest exist. In the
mountainous areas temperature and precipitation vary
along with altitudinal gradients, so does forest
vegetation. Annual growth rings of trees provide the
information about the response of forests to a variety of
environmental perturbations. Tree ring series contain a
series of signals that represent the sum of the
environmental influences on a tree growth 2. Climate
signal is considered as one of the main controlling
factors for the tree growth. The annual tree ring size is
achieved under the cumulative influence of all factors
that influence growth. In its analysis, all the influences
that appear during the growth and evolution of forestry
vegetation must be considered. The tree ring formation
depends on species and on the place where it grows
and expands.
In this study there were taken into account
environmental factors as altitude, exposition and
latitude in order to underline their influence over
growth on beech (Fagus sylvatica L.)
The objective of this study was to analyze
environmental factors as altitude, exposition and
latitude with the aim to demonstrate their influence
over beech growth.
18
Material and Methods
Beech has an extended growth area. A number
of 52 stands were chosen for study. The studied stands
range from 45 degrees northern latitude to 48 degrees
northern latitude and 25 degrees eastern longitude to 28
degrees eastern longitude. In Figure 1 the distribution
of beech stands is presented.
Fig.. 1. The distribution of beech stands
The stands where beech is the predominant
species are as well arranged on two groups: one is
found in Suceava county and Botosani county and the other is found in Neamt county and Bacau county. The stands are part of the National Forestry Inventory with the distance between points being 4 km. The layout of stands in which the beech is predominant species made possible a separation by longitude of two groups. In each stand was registered the geographical coordinates, altitude, exposition, average slope, flank position, production class, station type, class and type
of soil. Stands altitude for beech varies between 233 m
in the stand F112 and 1151 m in the stand F106. The
altitude classes poorly represented are class 200 m and
1200 m class with one stand each. The most frequent
altitude class found is 500 m with ten stands included.
The southern exposition was determined in 24 stands,
while the northern exposition proved to be
predominant with a number of 28 stands.
From the 52 stands with beech have been
processed a total of 580 growth samples.
The growth samples were taken with the
Pressler increment borer and transported in special
plastic supports. The growth samples were fixed on
wooden plates with adhesive and after its drying, the
samples were ground, in different stages with abrasive
tapes of different granulation, starting from those rough
to the fine ones.
The width measurement of the annual tree ring
was done with the help of the programme Coorecorder.
The precision offered by the programme is of 0,01mm.
The data obtained from the measurements were stored
in the form of electronic files for each growth sample.
In the first stage, the measurements were introduced in
the C – dendro programme where the sampling year
was given. The next stage was the cross dating process
for attributing each ring to its formation date. The cross
dating process was accomplished in two steps. In the
step stage a primary cross dating was done with the
help of the Coorecorder programme by comparing each
growth curve with a growth reference curve calculated
from the total of growth curves, for each species
separately. The analysis was done on segments, both
visually as well as with the help of automatically
calculated correlation coefficients. It was specially
insisted on the period proposed for study (2008 –
1959). The second cross dating step was done with the
help of the Cofecha programme. On the base of the
hints offered by the programme, checkouts were done
over the measurements and corrections were made in
19
the case when these proved to be real. The growth
samples which meaurements didn’t reasonably
correlate with the stand (correlation coefficient <0,2),
were excluded from the study.
The processing of growth series was done with
the help of Arstan programme. Through statistical
processing with Arstan3 programme the statistical
characteristics of each stand were obtained. The main
statistical parameters calculated were: the length of the
dendrochronological series, the minimal year of the
dendrochronological series, the minimal width value
of the measured annual tree ring from the
dendrochronological series, the maximum width value
of the measured annual tree ring from the
dendrochronological series, the maximum year from
the dendrochronological series, the average value of
annual tree rings width, variance, the standard
deviation, average sensitivity, first degree
autocorrelation, correlation.
The parameters of the growth series were
calculated with the programme Arstan6. The
standardization of growth series was done with Arstan.
The growth series standardization implies the
procedure of transforming the chronological series,
representing the size of the annual ring, in a series of
data by extracting a certain signal through the method
of relating the size of the annual tree ring to a value
obtained through a function. For standardizing the
growth series it was used the Arstan programme. The
first standardization was applied in order to exclude the
age influence with an exponential function and then a
second standardization was applied through a cubic
spline function with a length of 20 years for excluding
disturbing endogenous and exogenetic factors.
Results and Discussions
The longest series is analyzed between 1672
and 2010, found the stand F151. The average
calculated age for a stand is the highest in F123 with an
average age of 203 years and the stand with the lowest
average age is calculated in F103 with an average age
of 32 years. Beech proved to be the species with the
highest average age of the species analyzed.
The minimum number of growth samples
processed in stands was 10 samples, and the maximum
number of growth samples processed in stands was 16.
The correlation coefficients in the stand ranged
from 0,299 in stand F125 to 0,684 in stand F146.
Average correlation of all stands has a value around
0,451. This value is smaller that of the fir, but higher
that of spruce.
Radial growth accomplish by the beech ranges
from minimum values of 1,356 mm in stand F106 to
maximum values of 3,316 mm in F129 stand. The
standard deviation of the average actual growth is
between the values of 0,293 in F113 stand and 0,933 in
F124 stand.
Autocorrelation varies between 0.478 in F117
stand and 0.834 in F133 stand.
The standardized indices have an averaged
value that is close to the value 1, with standard
deviations of minimum 0,002 in F145 stand and
maximum of 0,142 in F133 stand.
Sensitivity ranges from minimum values of
0,213 in F129 stand to maximum values of 0,422 in
stand F112.
Autocorrelation for standardized values ranges
from a minimum value of 0,032 in stand F138 and
maximum values of 0.327 in F122 stand.
Correlation between trees in a beech stand was
calculated and plotted in the form of thematic map with
contour lines and a color scheme. Map is shown in
Figure 2.
20
Fig. 2. The correlation study depending on geographic position
Correlation study showed that it is quite low in
large areas of the study area. The highest values are
found in the eastern half of the Suceava county and
punctually in Bacau county. The lowest values of
correlation are found in northwestern Neamt county.
The distribution of the correlation values is also
punctual correlation with altitude is low. Correlation
values are given specific characteristics of the local
climate.
The average annual radial growth accomplish
by beech in the period 1959-2008 in the study area was
used to achieve a thematic maps. Map made is shown
in figure 3.
Fig. 3. The radial growth study depending on geographic position
21
Average size of annual ring made by beech in
the last 50 years varies greatly across the study area.
The highest values are found in Suceava county and
Botosani county (south-west part). The higher average
growth are also met punctual in Neamt county and
Bacau county. The highest values are found in areas
with medium altitude that provides a favorable rainfall
regime, but there are exceptions such as the low
altitude stands with beech in Botosani county where he
found the optimal conditions for growth. Stands in that
beech achieved lower average radial growth are
situated either at high altitudes, such as the mountain
part of Bacau county and punctual in Neamt county
and Vaslui county due to specific local climatic
conditions
The variations of the standard deviation of the
standardized series of growth was studied by mapping
the study area with specific software. Map made is
shown in Figure 4, it is shown the variation in the
standard deviation for the standardized series.
Fig. 4.. The standard deviation study depending on geographic position
For standard deviation, correlation with altitude
is reflected by the fact that high standard deviation
stands are located at high altitudes, 900 - 1000 m and at
low altitudes, 200 m, but this is not the uniform in the
study area. Distribution of standard deviation values is
punctual, with areas of minimum and areas of
maximum due to specific local conditions.
Principal component analysis is a statistical
method for determining the spatial variability,
information that is stored in the annual ring. For
principal component analysis were used standardized
series calculated with Arstran program. Analysis of
altitude influence on growth was analyzed for ten
classes of altitude. For each class of altitude was
calculated the average growth indices of all samples
taken at that altitude. The data was entered into
Microsoft Excel and principal component analysis was
performed. The graphical representation is shown in
figure 5.
22
200 m
300 m
400 m
500 m600 m
700 m
800 m
900 m
1000 m1100 m
1200 m
-3
-2
-1
0
1
2
3
-5 0 5
PC
2 (
10.4
%)
PC 1 (66.7%)
Fig. 5. Principal component analysis depending on altitude
Series analyzed were named according to
altitude class they belong. The first two principal
components analysis expressed 77.1 % of the
variability of the growth samples and the first
component explains 66.7% of the variability of the
growth samples. Analysis of the spatial variation of the
series according to altitude determined the formation of
three groups. The first group consists of series growing
at high altitudes - 900 m, 1000 m, 1100 m and 1200 m
The second group consists of middle altitude that
ranges from 500 m to 800 m. Third group consists of
series of growth achieved at low altitudes - 200 m, 300
m and 400 m. The spacing of the series is related to
environmental conditions in which they formed. Each
group of altitude expresses unique environmental
conditions in which the trees that make up the series
grew. At high altitudes, marginal habitat, beech
undergoing limiting environmental factors, mainly
temperature. Series of growth achieved at intermediate
altitudes receive normal to optimal conditions of
growth. Low altitudes, marginal habitat, beech
undergoing limiting environmental factors, mainly
rainfall. The three groups can be subdivided into
smaller groups. The group at high altitudes group the
altitude of 900 m, 1000 m and 1200 m series and 1100
m altitude series distance from the others with specific
growths probably due to specific conditions from local
factors. In the middle altitude group the series formed
at altitudes between 500 m and 600 m are closer due to
similar environmental conditions for growth. Also,
there is an approach between the series of altitude of
700 m and 800 m, still due to similar environmental
conditions of growth. In the low altitude group appears
closer ties between series at altitudes of 300 m and 400
m due to similar environmental conditions for growth
and the series at altitudes of 200 m distances
themselves due the increases of limiting local factors.
Principal component analysis is a statistical
method for determining the spatial variability,
information that is stored in the annual ring. For
principal component analysis were used standardized
index series calculated Arstran program. The influence
of altitude on growth was analyzed at Spruce ten
classes of altitude made for exhibitions and exhibitions
sunny shade.
Study of the exposition influence on the growth
of the beech was conducted for the three groups of
altitudes resulted from principal component analysis
for altitude. Within each altitude group, there were
analyzed series of standardized indices. Notation was
made with numbers and letters, numbers representing
altitude class and letters representing type of exposition
(I - southern, U - northern). The results are the figures
6, 7 and 8.
23
200 m U
300 m I
300 m U
400 m I
400 m U
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
-10 -5 0 5
CP
2 (
13,8
%)
CP 1 (72,7%)
Fi
g. 6. Principal component analysis depending on exposition for lower altitudes
In the first group of altitude, which contains low
altitudes between 200 m and 400 m, the series analysis
were grouped according to the type of exposition. The
series with northern exposition have negative values on
the second axis, the axis of the second main
component. Stands situated on the southern exposition
have positive values and negative values very close to
the value 0 on the second axis. In this case, the first
component expressed 72.7% of the variability in series,
and the second main component expressed 13.8% of
the variability in the series.
500 m I
500 m U
600 m U
700 m I
700 m U
800 m I
800 m U
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
-5 0 5
CP
2 (
6,2
%)
CP 1 (80,8%)
Fig. 7. Principal component analysis depending on exposition for middle altitudes
In the second group of altitude, which contains
classes of altitude between 500 m and 800 m, the
layout of the series does not shows the exposition
being a factor that is determining the variability.
Spatial arrangement of the series is mainly carried out
according to altitude, low altitude classes have lower
values and higher elevation classes have higher values
on the second axis.
24
900 m I
900 m U
1000 m I
1000 m U
1100 m I
1200 m U
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
-5 0 5
CP
2 (
10
,8%
)
CP 1 (65,6%)
Fig. 8. Principal component analysis depending on exposition for higher altitudes
In the case of the third group of altitude, the
analysis determined the clearest layout of the
standardized series according to the type of exposition.
The series with northern exposition have negative
values on the second axle and the ones with southern
exposition have positive values on the second axis.
Grouping was done after the second main component
explaining 10.8% of the spatial variability.
The analyzes performed have shown that the
influence of slope exposition is strongest in the class of
altitude marginal to distribution of beech, the influence
is most of temperature and moisture regime. In the
middle altitude classes, exposition influence on beech
growth is much reduced.
Spatial variability analysis was performed for
the studied stands depending on latitude. In the first
group are found seven stands: F102, F103, F104, F105,
F107, F108 and F109 which are located in the southern
study area. Northern group comprises six stands: F140,
F141, F144, F145, F146 and F147. Analysis of spatial
variability for latitude was done through hierarchical
analysis (Cluster) with Microsoft Excel in witch were
introduced standardized series obtained with Arstan
program for each stand. Graphical representation
obtained is shown in Figure 9.
F147
F141
F109
F105
F108
F104
F102
F144
F145
F107
F140
F146
F103
0.20
Fig. 9. Cluster analyses depending on latitude
25
It was achieved a grouping of the series from
different stands. The first group contained two stands
that share group latitude, the two stands are part of the
southern latitude, although they are located at different
altitudes. The second group includes two stands: F109
and F105. They are part of the southern latitude located
at different altitudes. The third group consists of series
of three stands: F147, F141 and F144. They are part of
the northern latitude located at different altitudes. In
this group is the strongest link between the series of
two stands: stand growing series F147 and stand
growing series F141. We can observe also two larger
groups: a southern group and the northern group. The
first large group is the southern group. It consists of
series from stands F108, F104, F102, F145, F109 and
F105. Of these, only F145 is located in the northern
area, but this analyzed series is formed at a low
altitude: 400 m. Occurs a compensation of
environmental factors resulting in similar increases,
which leads to the introduction of this stand southern
group. The second large group consists of stands F147,
F141, F144, F107, F140 and F146. This is the northern
group, with one exception: the stand F107, located in
the southern study area, at an altitude of 700 m on
southern exhibition. It is possible that this series is
formed in specific local site conditions, which could
express inclusion in this group.
Conclusions
Correlation study showed that, at the extremes
of studied altitude, approximately equal to the limits of
the natural area of distribution, the correlation between
the growth of trees is higher. The average radial growth
for all stands with beech is 2.08 mm. The natural
tendency is to have smaller growth values at lower
altitudes, higher values of the radial growth at middle
altitude and lower growth values at high altitudes.
The standard deviation is related to the size of
the annual ring. The natural tendency is to have smaller
variations of the growth series at lower altitudes,
higher variation at middle altitude and lower growth
variation at high altitudes.
Analysis of the spatial variation of the series
analyzed according to altitude caused the formation of
three groups. The first group consists of series growing
at high altitudes: 900 m, 1000 m, 1100 m and 1200 m
The second group consists of middle altitude classes
that ranges from 500 m and finishing with 800 m. The
third group consists of growth series achieved at low
altitudes: 200 m, 300 m and 400 m. The spacing series
explanation is related to environmental conditions in
which they formed this series. Each group of altitude
expresses unique environmental conditions in which
the trees grew. At high altitudes, marginal habitat,
beech is undergoing limiting environmental factors,
mainly temperature. Series of growth achieved at
middle altitudes receive normal to optimal temperature
and water quantities. Low altitudes, marginal habitat,
beech is undergoing limiting environmental factors,
mainly rainfall. In the case of exposition, the resulting
conclusion is that at low and high altitude exposition
has a strong influence, leading to changes in
environmental factors with larger impact on growth.
Exposition at middle altitude has an lower influence.
Latitude affects beech growth both through
annual climatic variations, which determines stronger
changes in the growth, and by changing temperature
and precipitation regime by moving in the temperate
latitude. A lower quantity of precipitation determines
the development of lower growth, even if the
temperature regime is more favorable to growth.
Acknowledgment
This paper was supported by the project
"Improvement of the doctoral studies quality in
engineering science for development of the knowledge
based society-QDOC” contract no.
POSDRU/107/1.5/S/78534, project co-funded by the
European Social Fund through the Sectorial
Operational Program Human Resources 2007-2013.
Bibliography
1.Čejkova, A., Kolar T., 2009 Extreme radial growth
reaction of Norway spruce along an altitudinal gradient
in the Sumava mountains, Geochronometria, pp 41-47
2.Cook, E.R., Kairiukstis, L.A. (eds.), 1990. Methods
of dendrochronology. Applications in the
environmental sciences. Kluwer Academic Publishers.
3.Cook, E.R., Krusic, P.J., ARSTAN ver.
4.1b,http://www.ldeo.columbia.edu, 2006.
4.Cropper, J.P., 1979. Tree ring skeleton plotting by
computer. Tree Ring Bulletin 39. pp.47-54
5.Cybis Electronik, 2007. Coorecorder, Cdendro.
http://www.cybis.se
6.Dang, H. et al., 2006 Growth responses of subalpine
fir (Abies fargesii) to climate variability in the Qinling
Mountain, China, Forest Ecology and Management, pp
143–150
7.Davi, N. K. et al., 2002, Boreal temperature
variability inferred from maximum latewood density
and tree-ring width data, Wrangell Mountain region,
Alaska, Quaternary Research, pp 252–262
8.DeRose, R. J., Gardner, R. S., 2010. Technique to
improve visualization of elusive tree ring boundaries
in aspen (Populus tremuloides), Aspen Bibliography.
Paper 3475. pp. 75-78
9.Fan, Z.X. et al., 2009 Growth–climate responses of
high-elevation conifers in the central Hengduan
Mountains, southwestern China, Forest Ecology and
Management , pp 306–313
10.Flocea, M. 1996b. Aplicatii ale dendrocronologiei
in domeniul studiilor de impact. Bucovina Forestiera.
1-2. pp. 31-43
26
11.Giurgiu V., 1972, Metode ale statisticii matematice
aplicate in silvicultură, Ed.Ceres, Bucureşti
12.Grissino-Mayer, H.D., 2003. A manual a tutorial for
the proper use o an increment borer. Tree Ring
Research, Vol. 59. pp. 63-81
13.Gruud, H., Briffa K. R. et al., 2002. The Holocene
12,6 pp. 657–665
14.Guay R.,1992. A new automatic and interactive tree
ring measurement system based on a line scan camera,
The Forestry Chronicle Vol. 68, No. 1, pp. 138-141
15.Holmes, R.J., 1983. Computer assisted quality
control in tree ring dating and measurement. Tree Ring
Bulletin 43. pp.69-78
16.Holtmeier, F. K., 2009 - Montain timberlines
ecology, patchiness, and dynamics. Springer-Verlag
Berlin Heinelberg
17Horodnic, S., 2004, Elemente de biostatistica
forestiera, Ed. Universităţii Stefan cel Mare Suceava,
159p.
18.Hughes, M.K., Swetnam, T.W., 2011.
Dendroclimatology - Progress and Prospects. Springer
Science+Business Media
19.Maxwell, R. S., Wixom, J. A., Hessl, A. E., 2011.A
comparison of two techniques for measuring and
crossdating tree rings, Dendrochronologia 29. pp. 237–
243.
20.Popa, I. 2003. Analiza comparativă a răspunsului
dendroclimatologic al molidului (Picea abies (L.)
Karst.) şi bradului (Abies alba Mill.) din nordul
Carpaţilor Orientali. Bucovina Forestieră 11(2), pp. 3-
14.
21.Popa, I. 2004. Fundamente metodologice şi aplicaţii
de dendrocronologie. Editura Tehnică Silvică,
Staţiunea Experimentală de Cultura Molidului,
Câmpulung Moldovenesc, 200 p.
22.Rigozo, R.N. et al, 2004. An interactive method for
digital tree-ring width measurement, Geofisica
Internacional, vol. 43, nr. 002. pp. 281-285
23.Robinson, W. J., Evans, R., 1980. A
microcomputer-based tree-ring measurement system.
Tree Ring Bulletin 40. pp.59-64
24.Rybnicek, M. et al., 2010 Radial Growth and health
condition of Norway spruce stands relative to climate,
Geochronometria, pp 9-16
25.Tardif, J. et al., 2003 Spatiotemporal variability in
tree growth in the central Pyrenees: climatic and site
influences, Ecological Monographs, 73(2), 2003, pp.
241–257
26.Vaganov, E.A., Hughes, M.K., Shashkin, A.V.,
2006. Growth dynamics of conifer tree rings. Springer-
Verlag Berlin Heinelberg, 354 p.