Content uploaded by Jurij Diaci
Author content
All content in this area was uploaded by Jurij Diaci on May 28, 2015
Content may be subject to copyright.
Gap regeneration patterns in
relationship to light heterogeneity
in two old-growth beech – fi r forest
reserves in South East Europe
ROZENBERGAR DUSAN
1 * , MIKAC STJEPAN
2 , ANI Ć IGOR
2 and
DIACI JURIJ
1
1 Biotechnical Faculty, Department of Forestry and Renewable Forest Resources, University of Ljubljana,
Vecna pot 83, 1000 Ljubljana, Slovenia
2 Faculty of Forestry, University of Zagreb, Croatia
* Corresponding author. E-mail: dusan.rozenbergar@bf.uni-lj.si
Summary
Gap regeneration in two old-growth forest reserves in Slovenia (Rajhenavski Rog) and Croatia
( Č orkova Uvala) was analysed in relation to within gap light heterogeneity. Both reserves were
located in the Dinaric mountain range in south-central Europe and were dominated by beech
( Fagus sylvatica L.) – silver fi r ( Abies alba Mill.) forest communities with similar growing conditions.
In total, the two largest gaps (700 – 2000 m
2 ) in each reserve were included in the study ( n = 4), plus
a further four gaps in Rajhenavski Rog and three in Č orkova Uvala (200 – 500 m
2 ). All the gaps
were ~ 10 years old and originated from one or two successive events, mostly due to a combination
of fungi attack and windthrow. Consequently, all gaps had complex geometry and were covered
by a well-developed regeneration layer. Each gap was mapped, followed by establishment of a N – S
oriented 5 × 5 m grid within and around the area of the canopy opening. At the grid intersections
773, 1.5 × 1.5 m plots were established. On each plot, the coverage of tree regeneration and ground
vegetation, seedling density in different height classes and browsing damage were recorded. In
addition, the total stretched length and last growing season increment of the leading stem was
measured on selected seedlings in each plot. Relative diffuse and direct radiation were estimated
using digital hemispherical photographs. All plots were classifi ed into four microsites according to
direct and diffuse radiation levels, and microsites were tested for differences in regeneration height
and density. While there was more woody regeneration and an almost fi vefold higher (6.2 vs 1.3
m
− 2 ) average total regeneration density in Rajhenavski Rog, mainly due to a high (5.5 vs 0.6 m
− 2 )
density of beech, there was more ground vegetation and a higher density of silver fi r seedlings in
Č orkova Uvala. The within-gap regeneration patterns proved to be similar in both forest reserves
and showed that 1-year-old seedlings of beech and silver fi r and small beech seedlings preferably
recruit on microsites under closed canopy or close to gap edges with lower levels of direct and
diffuse radiation. There was no signifi cant difference in density of large-beech seedlings among the
microsites, yet height and height increment were higher on microsites receiving the highest levels of
direct and diffuse radiation. Within-gap heterogeneity in light conditions appears to signifi cantly
© Institute of Chartered Foresters, 2007. All rights reserved. Forestry, doi:10.1093/forestry/cpm037
For Permissions, please email: journals.permissions@oxfordjournals.org
Forestry Advance Access published September 19, 2007
FORESTRY2 of 14
Introduction
The importance of tree-fall gap disturbances for
forest regeneration processes has long been rec-
ognized ( Watt, 1925 ; Platt and Strong, 1989 ).
One of the important characteristics of gaps that
affects the establishment and growth of tree seed-
lings is within- and around-gap environmental
heterogeneity ( Beatty, 1984 ; Lawton and Putz,
1988 ; Veblen, 1989 ; Gray and Spies, 1996 ).
Within-gap heterogeneity may result from many
factors, such as variation in microtopography
or microsites (i.e. nurse logs and root pits and
mounds). Another important factor includes het-
erogeneity in light regimes, especially if we parti-
tion light into direct and diffuse components. For
example, the amount of diffuse and direct radia-
tion varies within gaps, from the centre to the
edges and along N – S and E – W gradients ( Poulson
and Platt, 1989 ; Canham et al. , 1990 ; Wayne and
Bazzaz, 1993 ; Ritter et al. , 2005 ). There is a sub-
stantial difference in photon fl ux density between
the direct and diffuse components of radiation
and reaction of different plant species to certain
combinations of both components vary signifi -
cantly ( Larcher, 1983 ). Abiotic factors like inci-
dence to frost, soil and air temperature humidity
as well as biotic factors such as insect abundance
and crown and root competition are related to
the distribution radiation components ( Krecmer,
1966 ; Aussenac, 2000 ; Schütz, 2004 ; Brang et al. ,
2005 ). Knowledge about the distribution of both
components can therefore be useful for explaining
environmental heterogeneity within canopy gaps
( Poulson and Platt, 1989 ; Diaci and Thormann,
2002 ), which may have important consequences
for patterns of tree regeneration ( Galhidy et al. ,
2006 ; Mountford et al. , 2006 ).
In the beech – fi r dominated forests in south-cen-
tral Europe, small- to intermediate-scale canopy
gaps are the dominant disturbance process driving
infl uence general patterns of beech and silver fi r regeneration, while the differences in total average
density, regeneration and ground vegetation abundance and spatial distribution cannot solely
be explained by light. Since climate, parent material, soil, exposition, relief and stand conditions
were very similar between both reserves, it appears that these differences are a result of different
densities of large herbivores, which are reported to be much higher in Rajhenavski Rog. This was
also confi rmed in our research, as much more browsing damage was found in Rajhenavski Rog
compared with Č orkova Uvala. Implications for silvicultural practice are outlined.
forest dynamics ( Diaci et al. , 2003 ; Zeibig et al. ,
2003 ; Nagel and Diaci, 2006 ; Nagel et al. , 2006 ).
Natural regeneration and tree species composi-
tion of these ecosystems has been well studied in
the past ( Mlin š ek, 1967 ; Prpi ć , 1972 ; Mayer and
Neumann, 1981 ; Boncina, 2000 ; Bon č ina et al. ,
2003 ), but there is a lack of knowledge concern-
ing the regeneration response to different regimes
of direct and diffuse light in beech-fi r forests in
the Dinaric region ( Mati ć , 1983 ). While seed-
lings of beech and fi r can establish and grow in
the relatively low light levels beneath the forest
canopy ( Madsen, 1995 ; Szwagrzyk et al. , 2001 ;
Stancioiu and O’Hara, 2006b ), the increased light
levels in canopy gaps may be important for the re-
cruitment of saplings for both species ( Collet and
Chenost, 2006 ). However, little is known about
how light asymmetry within gaps infl uences the
density and height growth of regeneration of
these two species.
Data from previous inventories ( Hartman,
1987 ; Prpi ć and Seletkovi ć , 1996 ; Ro ž enbergar,
2000 ) in old-growth Dinaric beech – fi r forest
show an alternation in the dominance of silver fi r
and beech. Once silver fi r-dominated forests are,
especially in the last three decades, turning into
forests dominated by beech in the total living vol-
ume and regeneration. This trend has been much
more emphasized in Slovenia, where a low level
of recruitment success for silver fi r has been docu-
mented ( Debeljak, 1997 ; Bon č ina et al. , 2003 ). In
contrast, no problems with silver fi r regeneration
were reported in similar forest communities on
comparable site conditions in Croatia ( Prpi ć and
Seletkovi ć , 1996 ). Therefore, one of the aims of
this study was to understand if this difference can
be explained with differences in light conditions
and to discuss other factors which could also be
involved.
In this research, we examined regeneration of
beech and silver fi r in relation to light regimes
GAP REGENERATION PATTERNS IN RELATIONSHIP TO LIGHT HETEROGENEITY 3 of 14
in and around a range of different-sized canopy
gaps in two old-growth forest stands in the
Dinaric mountain range. The specifi c objectives
were to (1) examine variation in diffuse and di-
rect radiation within and around canopy gaps
across a range of gap sizes; (2) to determine if
variation in direct and diffuse radiation is related
to the presence, density and growth of silver fi r
and beech regeneration, and (3) to discuss dif-
ferences in regeneration success between the two
forest reserves with contrasting amounts of fi r
regeneration. Finally, we hope to provide use-
ful information about the infl uence of direct and
diffuse radiation on regeneration of silver fi r and
beech, which could help managers in creating ap-
propriate light conditions for successful recruit-
ment of both tree species, using different-sized
gaps in space and time.
Materials and methods
Research sites
The research was performed in two natural old-
growth forest reserves, Rajhenavski Rog in Slove-
nia and Č orkova Uvala in Croatia ( Figure 1 ). The
site conditions are similar in both forest reserves,
except the elevation is a bit higher in Č orkova
Uvala ( Table 1 ). Both forests are dominated by
beech – fi r communities, which are typically lo-
cated between 700 and 1200 m in the Dinaric
mountain range. Other less abundant species are
also present in both forest stands, including syca-
more maple ( Acer pseudoplatanus L.), wych elm
( Ulmus glabra Huds.), spruce [ Picea abies (L.)
Karsten], common ash ( Fraxinus excelsior L.)
and large-leaved lime ( Tilia platyphyllos Scop.).
Previous research has documented a 10 per cent
decrease of silver fi r living volume between 1976
and 1995 in Rajhenavski Rog. In Č orkova Uvala,
silver fi r volume has also declined, but only by
2 per cent from 1970 to 1987. The volume of
beech, in contrast, has increased by >10 per cent
in both reserves during the same periods ( Figure 2 ).
According to the latest inventory, these trends are
continuing in the same direction in both regions
( Ficko and Boncina, 2006 ).
Both study sites occur on a typical karst land-
scape with limestone parent material and free
draining and patchy rendzina soils. Soil depth and
macro- and microtopography are highly variable,
changing the site conditions signifi cantly over
very small spatial scales. In addition, karst phe-
nomena such as sinkholes and rock outcrops at
or close to the surface are common on both sites.
Field methods
Field surveys of both forest reserves were carried
out to fi nd the largest existing canopy gaps with
a regeneration layer of approximately the same
age. The latter was defi ned by the presence of
seedlings and saplings generally not exceeding
2 m in height. Thus, all the gaps chosen in this
study were relatively young (around 10 years
old). Except for very recently created gaps less
than a few years old, where the regeneration re-
sponse was still poor, most gaps in both reserves
had a well-developed regeneration layer in differ-
ent developmental stages depending on the gap
age. Gaps were only chosen if the adjacent can-
opy in the neighbourhood of the gap was closed
in order to avoid the infl uence of gaps in close
surroundings. The canopy around the gaps was
composed of beech and silver fi r in approximately
equal proportion and was ~ 40 m high. The two
largest gaps (700 – 2000 m
2 ) found in each reserve
were included ( n = 4), plus an additional four
gaps in Rajhenavski Rog and three in Č orkova
Uvala (200 – 500 m
2 ) with site conditions (slope,
exposition, soil) and gap age similar to those in
Figure 1 . The location of the old-growth forest
reserves Rajhenavski Rog, Slovenia and Č orkova
Uvala, Croatia.
FORESTRY4 of 14
the large gaps. All selected gaps were diverse in
shape, around 10 years old and created in one or
two successive events mostly due to a combina-
tion of fungi attack and windthrow.
The results of this study cannot be held as rep-
resentative for the whole area of both reserves,
since they do not include information on highly
shaded areas under closed canopy. However, in
both forest reserves, the research plots were cre-
ated in open and partly shaded parts of the for-
est stands and were therefore representative for
the areas where intensive regeneration processes
were taking place, which was the main focus of
this study.
All gaps were mapped to analyse gap shape
and area. The gap edge was defi ned as the verti-
cal projection of the crowns of the canopy trees
surrounding the gap. Between 2001 and 2004,
under each gap and a short distance (5 – 15 m)
under the adjacent canopy, a N – S oriented 5 × 5
m grid was established and 773, 1.5 × 1.5 m
plots were established on the grid intersections
(see Table 1 ). For each plot, we measured the light
conditions and tree regeneration. The per cent
cover of tree regeneration and ground vegetation
was visually estimated to the nearest 1 per cent.
All seedlings of each tree species were counted,
scored for browsing damage (seedlings were cat-
egorized as browsed when the leading stem was
damaged) and categorized within several height
classes, which include 1-year-old seedlings,
small seedlings (<20 cm tall) and large seedlings
(>20 cm tall and <5 cm d.b.h.). Thus, the large
seedling category also includes individuals >1.3 m
in height (generally termed saplings). In addi-
tion, the total stretched length and increment of
the leading stem in the last growing season were
recorded to the nearest 1 cm for the fi ve tallest
dominant beech seedlings in each plot.
Relative per cent diffuse (FDIF) and per cent
direct (FDIR) radiation were estimated from
digital hemispherical photographs taken in com-
pletely overcast sky conditions at 1.3 m height
with a Nikon 995 digital camera and calibrated
Figure 2 . Beech and silver fi r living volume in the forest reserve Rajhenavski Rog between 1957 and 1995
and Č orkova Uvala between 1975 and 1987, based on all trees attaining 5 cm diameter.
Table 1 : Study site characteristics of the two Dinaric beech – fi r old-growth forest reserves
Research site Rajhenavski Rog – SLO Č orkova Uvala – CRO
Area 52.1 ha 79.5 ha
Location 45° 66 ′ N, 15° 01 ′ E 44° 55 ′ N, 15° 32 ′ E
Altitude 740 – 880 m 860 – 1030 m
Annual precipitation 1500 – 1900 mm 1600 – 1700 mm
Average annual temperature 6 – 8°C 7°C
Number of gaps studied 6 5
Surrounding stand height Beech 39 m, s. fi r 41 m Beech 38 m, s. fi r 42 m
Number of plots recorded 435 338
Total living volume 747 m
3 ha
− 1 (2007) 660 m
3 ha
− 1 (2006)
GAP REGENERATION PATTERNS IN RELATIONSHIP TO LIGHT HETEROGENEITY 5 of 14
fi sh-eye lens from Regent WinScanopy accesso-
ries. Light intensity parameters were processed
with WinScanopy pro-d software ( Regent,
2003 ).
Data analyses
Gap geometry was analysed by examining gap
area – perimeter relationships. An index was cal-
culated to show how much the perimeter of the
gap deviates from that predicted for a circle at a
certain gap size. An increasing index value indi-
cates an increase in the irregularity of gap shape
( Lertzman and Krebs, 1991 ). A comparison of
light conditions among gaps was only performed
for plots located inside the gaps without any cov-
erage of the upper canopy.
Due to non-normal data, non-parametric
Mann – Whitney U tests were used to test for
differences in density and browsing damage for
both main tree species between the two sites ( Zar,
1999 ). Binary logistic regression procedures were
used to test the relationship between direct and
diffuse light and the appearance of 1-year-old,
small and large beech and silver fi r seedlings per
square metre, where the presence of silver fi r and
beech was used as a dependent variable. For this
analysis, all plots were classifi ed according to the
occurrence of silver fi r or beech seedlings (0, no
silver fi r or beech seedlings and 1, one or more
silver fi r or beech seedlings). The analysis was
performed in each forest reserve separately for
FDIR and FDIF as predictor variables. The odds
ratio and associated 95 per cent confi dence inter-
val were used for interpretation. The odds ratio
represents the change in odds when the indepen-
dent variable increases by 1 unit. In our case,
an odds ratio >1 indicates that a seedling has a
higher predicted probability of being present
with increasing radiation values, while an odds
ratio <1 indicates a lower predicted probability
of being present with increasing radiation values
( Hosmer and Lemeshow, 2000 ).
To examine the infl uence of direct and diffuse
relative radiation on certain measured parame-
ters, the plots were classifi ed into four microsite
types (A, B, C and D) according to the prevail-
ing combinations of both radiation components
( Diaci, 2002 ). The median radiation values were
used as thresholds for the four types as follows:
A, high FDIF and low FDIR; B, high FDIF and
high FDIR; C, low FDIF and low FDIR, and D,
low FDIF and high FDIR. Regardless of irregu-
lar gap shape, the plot types are typically located
in certain areas of the gaps ( Figure 3 ). The lo-
cations of the plot types indicate that there is
a connection between groups defi ned by radia-
tion and other ecological factors, such as pre-
cipitation, temperature, humus decomposition
rate and soil moisture, as they all partly follow
gap geometry ( Diaci, 2002 ; Diaci et al. , 2003 ).
Kruskall – Wallis tests ( Zar, 1999 ) were then
performed separately for both forest reserves
to examine differences in several variables be-
tween the four microsites. Signifi cant differences
between pairs of microsites were tested using
Nemenyi and Dunn post-hoc tests ( Zar, 1999 ).
The variables tested were woody regeneration
coverage, ground vegetation coverage, density
of 1-year-old beech and fi r seedlings per square
metre, density of small and large beech and fi r
seedlings per square metre and average length
and length increment of fi ve dominant beech
seedlings per plot.
Figure 3 . Typical spatial distribution of plot types
(A, B, C and D) according to received diffuse (FDIF)
and direct (FDIR) radiation values in a gap in an
old-growth beech – fi r forest (A, high FDIF and low
FDIR; B, high FDIF and high FDIR; C, low FDIF
and low FDIR, and D, low FDIF and high FDIR).
One of the research gaps in Rajhenavski Rog with
an area of 794 m
2 and an area – perimeter index of
3.1 is presented. The bold line represents the edge of
a gap (e.g. crowns of surrounding trees) and + signs
show the locations of sampling plots.
FORESTRY6 of 14
Results
The average gap size in Rajhenavski Rog ( N = 6)
was larger (640 ± 166 m
2 ) compared with Č orkova
Uvala ( N = 5) (487 ± 91 m
2 ), which was due to
one larger gap in Rajhenavski Rog (1317 m
2 ).
The remaining gaps were comparable in size ( Fig-
ure 4 ). In all gaps, the relative radiation values
for FDIF and FDIR were relatively low, and they
did not exceed 23 and 18 per cent, respectively.
Both FDIR and FDIF increased with gap size.
When similar-sized gaps were compared, there
was less radiation in Č orkova Uvala in all cases
( Figure 4 ). No signifi cant difference was observed
in the area – perimeter relationships between gaps
in both forest reserves. All gaps had between
two and four times larger perimeter compared
with that of a circle, indicating complex gap
geometry.
The average total density of all regenera-
tion was nearly fi ve times higher (6.2 m
− 2 ) in
gaps in Rajhenavski Rog compared with gaps
in the Č orkova Uvala (1.3 m
− 2 ) forest reserve
( U = 17 482, P < 0.0001). This was mostly due to
the higher (5.5 vs 0.6 m
− 2 ) density of beech in
Rajhenavski Rog (see Figure 5 , right). However,
the average density per plot of silver fi r seedlings
was greater in Č orkova Uvala (0.6 vs 0.3 m
− 2 ,
U = 62 720, P = 0.0004) and there were no silver
fi r seedlings taller than 50 cm in Rajhenavski
Rog (see Figure 5 , left). The distribution curves
for beech and silver fi r ( Figure 5 ) indicate that
in Rajhenavski Rog there had been a recent, vig-
orous fl ush of beech establishment, whereas in
Č orkova Uvala this had been more limited. Sil-
ver fi r seedlings established in both reserves, but
only recruited into taller seedling height classes in
Č orkova Uvala.
On average, the proportion of heavily browsed
beech seedlings was 26.3 and 13.9 per cent in
Rajhenavski Rog and Č orkova Uvala, respectively
( U = 22774, P < 0.0001). There were more heavily
Figure 4 . Mean values and ±95% confi dence interval for relative (%) direct (FDIR) and diffuse (FDIF)
radiation according to gap size separately for the Rajhenavski Rog and Č orkova Uvala old-growth forest
reserves. Only plots with no upper canopy were analysed.
Figure 5 . Density (number per square metre) of fi r (left) and beech (right) 1 year old (oy) and older seedlings
in different height classes (cm) based on all recorded seedlings in the Rajhenavski Rog and Č orkova Uvala
old-growth forest reserves.
GAP REGENERATION PATTERNS IN RELATIONSHIP TO LIGHT HETEROGENEITY 7 of 14
browsed silver fi r seedlings in Rajhenavski Rog
(31.8 per cent) compared with Č orkova Uvala
(11.3 per cent, U = 9424, P < 0.0001). In gen-
eral, the damage was more severe compared with
beech due to the successive occurrence of brows-
ing on the leading and fi rst two lateral stems.
A signifi cant, but weak relationship between
FDIR and FDIF and the presence of silver fi r and
beech seedlings was confi rmed in both forest re-
serves ( Table 2 ). This was especially the case for
1 year old and small seedlings of beech and silver
fi r in Rajhenavski Rog. In Č orkova Uvala, there
was a signifi cant relationship between radiation
and occurrence of silver fi r seedlings in almost all
height classes, which was not the case for beech.
In all signifi cant cases, except for small silver fi r
seedlings and FDIF, the odds ratio was <1, mean-
ing that the probability of seedling occurrence
decreased with an increase of FDIF and FDIR
( Figure 6 ).
We tested the differences among the four micro-
sites (A, B, C and D) for the selected parameters
separately for the Rajhenavski Rog and Č orkova
Uvala forest reserves. The average regeneration
and ground vegetation cover per plot differed
signifi cantly between the four plot types in both
reserves ( Table 3 ). Regeneration cover was high-
est in plot type B, while ground vegetation cover
was highest in plot type A. The results for 1 year
old and small beech seedlings were consistent
with both methods of analysis (Kruskal – Wallis
test and logistic regression) and in both reserves,
and showed that microsite A had the lowest den-
sity of these seedlings. Larger seedlings, which
were the most numerous in both reserves, did not
show any signifi cant difference in density across
the microsites. However, they were signifi cantly
taller and had larger height increments within
gaps (microsites A and B) in both reserves and
thus occupied larger areas compared with smaller
seedlings.
A similar pattern was also observed for 1-
year-old silver fi r seedlings, yet seedling success
was lowest in microsite B. Here, the results were
signifi cant only for Rajhenavski Rog. The results
of the Kruskal – Wallis test were not signifi cant
for small seedlings, whereas the logistic regres-
sion showed a negative relationship with FDIR
in both reserves. Taller silver fi r seedlings were
only present in Č orkova Uvala and again nega-
tively related to direct light, while the post-hoc
comparison of Kruskal – Wallis pairs was not sig-
nifi cant ( Table 3 ).
Discussion
This study provides useful information about
natural regeneration processes in two old-growth,
fi r-beech forest reserves in the Dinaric Moun-
tains. Major differences in the regeneration den-
sity and height distribution of both dominant
species growing in gaps were detected between
Rajhenavski Rog and Č orkova Uvala. We found
an almost fi ve times higher density of beech re-
generation in Rajhenavski Rog, where the densi-
ties of beech seedlings were higher on the whole
range of height classes. There are several possible
explanations for this fl ush of beech regeneration
in Rajhenavski Rog. One possible cause could be
related to the lower relative radiation found in
Č orkova Uvala, even though gap area – perimeter
relationships and the height of trees surround-
ing the gaps were similar in Rajhenavski Rog.
Other studies have reported a more diverse stand
structure, larger gap fraction and lower levels of
total living volume in Rajhenavski Rog ( Mayer
and Neumann, 1981 ; Hartman, 1987 ; Prpi ć and
Seletkovi ć , 1996 ; Boncina, 2000 ). The more open
canopy in Rajhenavski Rog may be partly related
Table 2 : Odds ratio estimates for logistic regression
models predicting the probability of occurrence of
silver fi r and beech seedlings according to FDIF and
FDIR
FDIF FDIR
Rajhenavski Rog
Beech 1 year old 0.8384** 0.9806
Beech small seedlings 0.8898** 0.9688*
Beech large seedlings 0.9740 1.0128
Fir 1 year old 0.9303** 0.9727*
Fir small seedlings 1.0733** 0.9640*
Č orkova Uvala
Beech 1 year old 0.9427 1.0178
Beech small seedlings 0.9397 0.9074*
Beech large seedlings 1.0609 1.0016
Fir 1 year old 0.9128* 1.0086
Fir small seedlings 1.1091* 0.9154*
Fir large seedlings 0.9792 0.9132**
* P < 0.05, ** P < 0.01.
FORESTRY8 of 14
to the decline of silver fi r in the reserve ( Boncina
et al. , 2002 ), which likely increased the radiation
levels beneath the canopy throughout the entire
stand. In Rajhenavski Rog, there were twofold
less silver fi r, amounting to a density of ~ 0.3 m
− 2 .
This number is reported to be insuffi cient for nor-
mal development and recruitment of this tree spe-
cies in the upper canopy ( Perko, 1977 ; Veseli č ,
1991 ; Jarni et al. , 2005 ). Moreover, there were no
silver fi r seedlings taller than 50 cm, which may
have important implications on the future tree
species composition and natural development of
this forest.
Beech, which can tolerate a broad range of un-
derstory light levels, manages to recruit in a vari-
ety of light conditions in young stages of growth,
as it is capable of decurrent and polycyclic growth
( Nicolini, 2000 ; Collet et al. , 2001 ; Stancioiu and
O’Hara, 2006a ). In our study, the results of the
comparison among microsites indicate differences
Figure 6 . Probability of occurrence of silver fi r and beech seedlings (silver fi r or beech seedling present = 1,
no silver fi r or beech seedling = 0) according to FDIR and FDIF in Rajhenavski Rog and Č orkova Uvala
old-growth forests.
GAP REGENERATION PATTERNS IN RELATIONSHIP TO LIGHT HETEROGENEITY 9 of 14
in beech establishment and further recruitment
niches. Beech 1 year old and small seedlings suc-
cessfully establish on microsites under canopies
as advance regeneration ( Szwagrzyk et al. , 2001 ),
sustaining low light and precipitation levels and
higher root competition from mature trees. These
suppressed individuals can wait for long periods
before experiencing a release event.
The fact that there was the highest coverage of
ground vegetation on microsites A in both for-
est reserves suggests that on this locations inside
gaps ground vegetation is an important competi-
tor to tree seedlings. These microsites receive full
precipitation and less direct radiation, which in-
fl uences soil and air temperatures. There is also
less root competition from older trees. Therefore,
microsites A are relatively humid compared
with microsites B and have less root competition
than microsites C or D. Since ground vegetation is
less drought resistant and can sustain less root com-
petition than seedlings, these microsites are more
favourable for development of ground vegetation
( Helliwell and Harrison, 1979 ; Welander and
Ottoson, 1997 ; Lof, 2000 ).
In regard to within gap regeneration of silver
fi r, the less preferred microsite seems to be B,
Table 3 : Mean and P value results of the Kruskal – Wallis test according to the received radiation (plot types A,
B, C and D, see Figure 3 ) separately for the Rajhenavski Rog and Č orkova Uvala old-growth forest reserves
ABCD
P Post hoc *
N = 78 N = 197 N = 105 N = 55
Rajhenavski Rog
Regeneration % 55.96 66.94 42.42 49.00 <0.0001 BC; BD
Ground vegetation % 19.44 10.33 13.80 13.05 <0.0001 AB; BC
Beech 1 year old 0.01 0.03 0.24 0.15 <0.0001 AC; BC
Beech small 0.66 0.70 2.28 1.57 <0.0001 AC; DC;
BC; BD
Beech large 3.73 5.30 4.24 4.48 0.0893 –
Fir 1 year old 0.55 0.39 0.76 0.44 0.0040 BC
Fir small 0.35 0.26 0.28 0.40 0.0624 –
Fir large – – – – – –
Beech length 150.99 179.66 86.52 115.97 <0.0001 AC; BC; BD
Beech increment 17.91 21.49 10.40 12.63 <0.0001 AB; AC; AD;
BC; BD
Č orkova Uvala N = 65 N = 45 N = 142 N = 86
N 65 45 142 86
Regeneration % 26.94 29.33 13.74 18.14 0.0124 AC
Ground vegetation % 45.29 39.96 39.92 30.62 0.0082 AD; CD
Beech 1 year old 0.10 0.12 0.33 0.13 0.3179 –
Beech small 0.05 0.04 0.12 0.07 0.0317 Not
signifi cant
Beech large 0.74 0.48 0.38 0.42 0.1926 –
Fir 1 year old 0.51 0.36 0.38 0.66 0.3391 –
Fir small 0.25 0.07 0.24 0.22 0.1194 –
Fir large 0.39 0.69 0.29 0.52 0.0133 Not
signifi cant
Beech length 182.84 171.93 110.31 127.39 0.0010 AC
Beech increment 17.15 16.29 9.64 9.57 <0.0001 AC; AD; BC
The variables tested were woody regeneration coverage, ground vegetation coverage, density of 1-year-old beech
and silver fi r seedlings per square metre, density small (<20 cm tall) and large (>20 cm tall) beech and silver fi r
seedlings per square metre and average length and length increment of fi ve dominant beech seedlings per plot.
* Only pairs of signifi cantly different plot types detected using Nemenyi and Dunn post-hoc test are presented.
FORESTRY10 of 14
where the highest direct and diffuse light levels
occur. The highest density of older and taller
silver fi r seedlings appears to be outside the mi-
crosites with higher values of direct light. This
pattern was also documented in previous research
( Grassi et al. , 2004 ; Paluch, 2005 ; Stancioiu and
O’Hara, 2006b ). The difference in silver fi r re-
generation density between size classes across
microsites was not as transparent as with beech.
This could be due to additional factors, namely
deer browsing ( Ammer, 1996 ; Gill and Bear-
dall, 2001 ; Kuiters and Slim, 2002 ; Heuze et al. ,
2005 ). Silver fi r is far more preferred as forage
and much more susceptible to browsing damage
than beech ( Motta, 1996 ; Senn and Suter, 2003 ;
Heuze et al. , 2005 ). Moreover, the reported deer
density is substantially different between the two
areas, amounting to 0.8 and 0.2 per 100 ha for
roe deer and red deer, respectively, in the area
around Č orkova Uvala ( Anonymous, 2007 ), and
0.9 and 6.6 per 100 ha for roe deer and red deer,
respectively, in the area around Rajhenavski Rog
( Jerina, 2006 , 2007 ). This was also supported
in this study, where a higher level of browsing
damage was found in Rajhenavski Rog compared
with Č orkova Uvala.
Major differences in gap fi lling patterns as well
as the analysed parameters were observed be-
tween the two old-growth forest reserves. This
was especially the case for coverage of ground
vegetation and regeneration. The former was sig-
nifi cantly higher across all micosites in Č orkova
Uvala, while the latter was signifi cantly lower.
Since the microsite groups were based on the
same light data pool and therefore have the same
light intervals in both reserves, light cannot ex-
plain this difference. Moreover, other ecological
factors, such as climate, parent material, soil,
exposition, relief and stand conditions were very
similar between both reserves. Therefore, the dif-
ference in the density of large herbivores seems to
be the most likely explanation for the differences
in regeneration and ground vegetation abundance
and spatial distribution.
Our study focused on middle-aged (around 10
years) gaps fi lled with seedlings, while younger
gaps and relatively closed stands were not in-
cluded. Therefore, the results presented here are
only relevant for gap-phase regeneration in these
stands and are not representative of regeneration
patterns under the full range of canopy conditions
found in these forests. In addition, only a short
period of the regeneration cycle was studied, so
that future studies should focus on further devel-
opment of the regeneration including the recruit-
ment of both tree species to the canopy layer.
Implications for forestry
This research suggests that the regeneration pat-
terns and tree species composition cannot simply
be explained using only light conditions in cer-
tain parts of a stand or canopy gap. There are
many other factors that infl uence the regenera-
tion (e.g. browsing, soil moisture, ground vegeta-
tion), which should be taken into account when
discussing the use of different light conditions as
a tool for reaching different silvicultural goals.
Nevertheless, we believe that this study demon-
strates the importance of partial shading from the
upper canopy in regeneration processes in natural
beech – fi r forests. Two independent analyses used
in this study confi rmed, that beech and silver fi r
successfully establish beneath closed canopies.
Further recruitment of beech into taller sapling
stages seems to profi t from both direct and diffuse
light, while silver fi r seems to avoid direct light.
It was demonstrated that the establishment and
further regeneration development of both spe-
cies differs. Taking into account the higher light
demand of beech in later regeneration phases
and especially the tendency toward plagiotropic
growth in low light levels, it seems reasonable
to propose a selection system approach in areas
with more silver fi r, while in areas with prevalent
beech regeneration the tending of already estab-
lished beech stems in gaps seems more appropri-
ate. Foresters in Central Europe have traditionally
used the shelterwood system for regeneration of
beech forests and partly in beech – silver fi r forests
( Matthews, 1989 ). Due to the different develop-
ment niche of older seedlings between silver fi r
and beech, a combination of a selection system
and irregular shelterwood system appears to be a
favourable management solution.
Funding
Governmental agencies in Slovenia (Slovenian Re-
search Agency) and Croatia (Ministry of Science,
GAP REGENERATION PATTERNS IN RELATIONSHIP TO LIGHT HETEROGENEITY 11 of 14
Education and Sports, grant no. 068-0682041-1950),
NAT-MAN Project (Nature-based management of
beech in Europe) and European Community fi fth
Framework Programme (grant no. QLKS-CT-1999-
01349).
Acknowledgements
We wish to thank all four reviewers for their construc-
tive comments, and Tom Nagel for improving an ear-
lier version of the manuscript.
Confl ict of Interest Statement
None declared.
References
Ammer , C. 1996 Impact of ungulates on structure and
dynamics of natural regeneration of mixed mountain
forests in the Bavarian Alps . For. Ecol. Manage. 88 ,
43 – 53 .
Anonymous 2007 Popis brojnog stanja divlja č i u
lova č kim udrugama Krivi javor, Golo trlo, Vrhovine
i Krekova č a . Lova č ke udruge republike Hrvatske .
Aussenac , G. 2000 Interactions between forest stands
and microclimate: ecophysiological aspects and conse-
quences for silviculture . Ann. For. Sci. 57 , 287 – 301 .
Beatty , S.W. 1984 Infl uence of microtopography and
canopy species on spatial patterns of forest under-
story plants . Ecology . 65 , 1406 – 1419 .
Boncina , A. 2000 Comparison of structure and bio-
diversity in the Rajhenav virgin forest remnant and
managed forest in the Dinaric region of Slovenia .
Glob. Ecol. Biogeogr. 9 , 201 – 211 .
Boncina , A. , Diaci , J. and Cencic , L. 2002 Comparison
of the two main types of selection forests in Slovenia:
distribution, site conditions, stand structure, regen-
eration and management . Forestry . 75 , 365 – 373 .
Bon č ina , A. , Diaci , J. and Ga š per š i č , F. 2003 Long-
term changes in tree species composition in the
Dinaric mountain forests of Slovenia . For. Chron.
79 , 227 – 232 .
Brang , P. , Von Felten , S. and Wagner , S. 2005 Morn-
ing, noon, or afternoon: does timing of direct radia-
tion infl uence the growth of Picea abies seedlings in
mountain forests? . Ann. For. Sci. 62 , 697 – 705 .
Canham , C.D. , Denslow , J.S. , Platt , W.J. , Runkle , J.R. ,
Spies , T.A. and White , P.S. 1990 Light regimes be-
neath closed canopies and tree-fall gaps in temperate
and tropical forests . Can. J. For. Res. 20 , 620 – 631 .
Collet , C. and Chenost , C. 2006 Using competition and
light estimates to predict diameter and height growth
of naturally regenerated beech seedlings growing
under changing canopy conditions . Forestry . 79 ,
489 – 502 .
Collet , C. , Lanter , O. and Pardos , M. 2001 Effects of
canopy opening on height and diameter growth in
naturally regenerated beech seedlings . Ann. For. Sci.
58 , 127 – 134 .
Debeljak , M. 1997 Jelka ( Abies alba Mill.) v pom-
ladku pragozda Pe č ka v zadnjih tridesetih letih . Zb.
gozdarstva lesarstva . 53 , 29 – 48 .
Diaci , J. 2002 Regeneration dynamics in a Norway
spruce plantation on a silver fi r-beech forest site in
the Slovenian Alps . For. Ecol. Manage. 161 , 27 – 38 .
Diaci , J. , Rozenbergar , D. and Bon č ina , A. 2003 In-
teractions of light and regeneration in Slovenian
Dinaric Alps: patterns in virgin and managed for-
ests . In Natural Forests in the Temperate Zone of
Europe – Values and Utilisation . B. Commarmot
(ed). Swiss Federal Research Institute WSL , Birmens-
dorf , pp. 154 – 160 .
Diaci , J. and Thormann , J.J. 2002 A comparison
of solar radiation estimation methods in natural
beech forests in Slovenia from the point of view of
ecological regeneration . Schweiz. Z. Forstwes. 153 ,
39 – 50 .
Ficko , A. and Boncina , A. 2006 Silver fi r ( Abies alba
Mill.) distribution in Slovenian forests . Res. Rep.
For. Wood Sci. Technol. 79 , 19 – 35 .
Galhidy , L. , Mihok , B. , Hagyo , A. , Rajkai , K. and
Standovar , T. 2006 Effects of gap size and associated
changes in light and soil moisture on the understorey
vegetation of a Hungarian beech forest . Plant Ecol.
183 , 133 – 145 .
Gill , R.M.A. and Beardall , V. 2001 The impact of deer
on woodlands: the effects of browsing and seed dis-
persal on vegetation structure and composition . For-
estry . 74 , 209 – 218 .
Grassi , G. , Minotta , G. , Tonon , G. and Bagnaresi , U.
2004 Dynamics of Norway spruce and silver fi r natu-
ral regeneration in a mixed stand under uneven-aged
management . Can. J. For. Res. 34 , 141 – 149 .
Gray , A.N. and Spies , T.A. 1996 Gap size, within-gap
position and canopy structure effects on conifer seed-
ling establishment . J. Ecol. 84 , 635 – 645 .
Hartman , T. 1987 Gozdni rezervati Slovenije – Pragozd
Rajhenavski Rog . Univerza Edvarda Kardelja v
Ljubljani, Biotehni š ka fakulteta , Ljubljana , pp 99 .
FORESTRY12 of 14
Helliwell , D.R. and Harrison , A.F. 1979 Effects of light
and weed competition on the growth of seedlings of 4
tree species on a range of soils . Q. J. For. 73 , 160 – 171 .
Heuze , P. , Schnitzler , A. and Klein , F. 2005 Is brows-
ing the major factor of silver fi r decline in the Vosges
Mountains of France? . For. Ecol. Manage. 217 ,
219 – 228 .
Hosmer , D.W. and Lemeshow , S. 2000 Applied Logis-
tic Regression . Wiley , New York .
Jarni , K. , Robi č , D. and Bon č ina , A. 2005 Analysis
of the infl uence of ungulates on the regeneration
of Dinaric fi r-beech forests in the research site Trn-
ovec in the Ko č evje forest management region . Zb.
gozdarstva lesarstva . 74 , 141 – 164 .
Jerina , K. 2006 Prostorska razporeditev, obmo č ja
aktivnosti in telesna masa jelenjadi (Cervus elaphus
L.) glede na okoljske dejavnike . University of Lju-
bljana, Biotehni š ka fakulteta , Ljubljana , pp 193 .
Jerina , K. 2007 The effects of habitat structure on red
deer (Cervus elaphus L.) body mass . Zb. gozdarstva
lesarstva . 82 .
Krecmer , V. 1966 Das Mikroklima der Kieferlochkahls
chläge . Wetter und Leben . 18 , 133 – 141 .
Kuiters , A.T. and Slim , P.A. 2002 Regeneration of
mixed deciduous forest in a Dutch forest-heathland,
following a reduction of ungulate densities . Biol.
Conserv. 105 , 65 – 74 .
Larcher , W. 1983 Physiological Plant Ecology .
Springer-Verlag , München, Germany , pp 303 .
Lawton , R.O. and Putz , F.E. 1988 Natural disturbance
and gap-phase regeneration in a wind-exposed tropi-
cal cloud forest . Ecology . 69 , 764 – 777 .
Lertzman , K.P. and Krebs , C.J. 1991 Gap-phase struc-
ture of a sub-alpine old-growth forest . Can. J. For.
Res. 21 , 1730 – 1741 .
Lof , M. 2000 Establishment and growth in seedlings
of Fagus sylvatica and Quercus robur: infl uence of
interference from herbaceous vegetation . Can. J. For.
Res. 30 , 855 – 864 .
Madsen , P. 1995 Effects of soil-water content, fertiliza-
tion, light, weed competition and seedbed type on
natural regeneration of beech (Fagus-Sylvatica) . For.
Ecol. Manage. 72 , 251 – 264 .
Mati ć , S. 1983 The impact of ecological and structural
factors on natural regeneration of fi r and beech selec-
tion forests in Gorski Kotar . Glas. š umske pokuse .
21 , 223 – 400 .
Matthews , J.D. 1989 Silvicultural Systems . Clarendon
Press , Oxford , pp 284 .
Mayer , H. and Neumann , M. 1981 Struktureller und
entwicklugsdynamischer Vergleich der Fichten – Tan-
nen – Buchen – Urwälder Rothwald/Niederösterreich
und Č orkova Uvala/Kroatien . Forstw. Cbl. 100 ,
111 – 132 .
Mlin š ek , D. 1967 Verjüngung und Entwicklung der
Dickungen im Tannen-Buchen Urwald “Rog”
(Slowenien) . Zb. BF . 15 , 7 – 32 .
Motta , R. 1996 Impact of wild ungulates on forest re-
generation and tree composition of mountain forests
in the Western Italian Alps . For. Ecol. Manage. 88 ,
93 – 98 .
Mountford , E.P. , Savill , P.S. and Bebber , D.P. 2006
Patterns of regeneration and ground vegetation as-
sociated with canopy gaps in a managed beechwood
in southern England . Forestry . 79 , 389 – 408 .
Nagel , T.A. and Diaci , J. 2006 Intermediate wind dis-
turbance in an old-growth beech-fi r forest in south-
eastern Slovenia . Can. J. For. Res. 36 , 629 – 638 .
Nagel , T.A. , Svoboda , M. and Diaci , J. 2006 Regen-
eration patterns after intermediate wind disturbance
in an old-growth Fagus-Abies forest in southeastern
Slovenia . For. Ecol. Manage. 226 , 268 – 278 .
Nicolini , E. 2000 New observations on morphology
of beech growth units ( Fagus sylvatica L.) – shoot
symmetry, refl ection of tree vigor . Can. J. Bot. 78 ,
77 – 87 .
Paluch , J.G. 2005 The infl uence of the spatial pattern
of trees on forest fl oor vegetation and silver fi r (Abies
alba Mill.) regeneration in uneven-aged forests . For.
Ecol. Manage. 205 , 283 – 298 .
Perko , F. 1977 The infl uence of game on the regenera-
tion of High Carst Fir-Beech forest . Gozdarski vestn.
35 , 191 – 204 .
Platt , W.J. and Strong , D.R. 1989 Special feature – treefall
gaps and forest dynamics – gaps in forest ecology .
Ecology . 70 , 535 – 575 .
Poulson , L.T. and Platt , W.J. 1989 Gap light regimes in-
fl uence canopy tree diversity . Ecology . 70 , 553 – 555 .
Prpi ć , B. 1972 Neke zna č ajke pra š ume Č orkova uvala .
Š um. List . 9–10 , 325 – 333 .
Prpi ć , B. and Seletkovi ć , Z. 1996 In The research in
Croatian virgin forests and the application of results
to natural forests . In Unapredjenje proizvodnje bio-
mase š umskih ekosustava . B. Mayer (ed). Hrvatsko
š umarsko dru š tvo , Zagreb , pp. 97 – 104 .
Regent , 2003 WinSCANOPY for Hemispherical Image
Analysis – Manual . Regent Instruments Inc., Chemin
Sainte-Foy, pp 104 .
Ritter , E. , Dalsgaard , L. and Einhorn , K.S. 2005 Light,
temperature and soil moisture regimes following gap
formation in a semi-natural beech-dominated forest
in Denmark . For. Ecol. Manage. 206 , 15 – 33 .
GAP REGENERATION PATTERNS IN RELATIONSHIP TO LIGHT HETEROGENEITY 13 of 14
Ro ž enbergar , D. 2000 Stand dynamics of the virgin
forest remnants of Pe č ka and Rajhenavski Rog .
Gozdarski vestn. 58 , 53 – 55 .
Schütz , J.P. 2004 Opportunistic methods of controlling
vegetation, inspired by natural plant succession dy-
namics with special reference to natural outmixing
tendencies in a gap regeneration . Ann. For. Sci. 61 ,
149 – 156 .
Senn , J. and Suter , W. 2003 Ungulate browsing on silver
fi r ( Abies alba ) in the Swiss Alps: beliefs in search of
supporting data . For. Ecol. Manage. 181 , 151 – 164 .
Stancioiu , P.T. and O’Hara , K.L. 2006a Morphological
plasticity of regeneration subject to different levels of
canopy cover in mixed-species, multiaged forests of
the Romanian Carpathians . Tree Struct. Funct. 20 ,
196 – 209 .
Stancioiu , P.T. and O’Hara , K.L. 2006b Regeneration
growth in different light environments of mixed spe-
cies, multiaged, mountainous forests of Romania .
Eur. J. For. Res. 125 , 151 – 162 .
Szwagrzyk , J. , Szewczyk , J. and Bodziarczyk , J. 2001
Dynamics of seedling banks in beech forest: results
of a 10-year study on germination, growth and sur-
vival . For. Ecol. Manage. 141 , 237 – 250 .
Veblen , T.T. 1989 Tree regeneration responses to gaps
along a transandean gradient . Ecology . 70 , 541 – 543 .
Veseli č , 1991 Na Postojnskem pre š tevilna rastlinojeda
divjad š e naprej hudo ogro ž a gozdno mladje . Goz-
darski vestn. 49 , 147 – 157 .
Watt , A.S. 1925 On the ecology of British beechwoods
with special reference to their regeneration: part II,
sections II and III. The development and structure
of beech communities on the Sussex Downs . J. Ecol.
13 , 27 – 73 .
Wayne , P.M. and Bazzaz , F.A. 1993 Morning vs after-
noon sun patches in experimental forest gaps – con-
sequences of temporal incongruency of resources to
birch regeneration . Oecologia . 94 , 235 – 243 .
Welander , N.T. and Ottoson , B. 1997 The infl uence of
shading on growth and morphology in seedlings of
Quercus robur L. and Fagus sylvatica L . For. Ecol.
Manage. 107 , 117 – 126 .
Zar , J.H. 1999 Biostatistical Analysis . Prentice Hall
Corp , Upper Saddle River, NJ , pp 929 .
Zeibig , A. , Diaci , J. and Wagner , S. 2003 In Gap distur-
bance patterns of a Fagus sylvatica virgin forest rem-
nant in the mountain vegetation belt in Slovenia . Nat-
ural Forests in the Temperate Zone of Europe – Values
and Utilisation . B. Commarmot (ed). Swiss Federal
Research Institute WSL , Birmensdorf , pp. 81 – 92 .
Received 16 November 2006
