Content uploaded by Håkan Hytteborn
Author content
All content in this area was uploaded by Håkan Hytteborn on Aug 11, 2019
Content may be subject to copyright.
International Association of Vegetation Science
Gap Structure, Disturbance and Regeneration in a Primeval Picea abies Forest
Author(s): Liu Qinghong and Håkan Hytteborn
Reviewed work(s):
Source:
Journal of Vegetation Science,
Vol. 2, No. 3 (Jun., 1991), pp. 391-402
Published by: Wiley-Blackwell
Stable URL: http://www.jstor.org/stable/3235932 .
Accessed: 10/10/2012 07:36
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
Wiley-Blackwell and International Association of Vegetation Science are collaborating with JSTOR to digitize,
preserve and extend access to Journal of Vegetation Science.
http://www.jstor.org
Journal of Vegetation
Science 2: 391-402, 1991
? IAVS;
Opulus
Press Uppsala Printed in the United States of America
Gap structure, disturbance and regeneration
in a primeval Picea abies forest
Liu Qinghong & Hytteborn, Hakan
Department of Ecological Botany, Uppsala University, Box 559, S 751 22 Uppsala, Sweden;
Tel. +46 18 182850; Fax +46 18 553419; E-mail LIU@PAX.UU.SE
Abstract. We investigated
patterns
of disturbance and recov-
ery in Fiby urskog, a primeval spruce (Picea abies) forest,
situated south of the border between the Boreo-nemoral
and
Boreal regions in East-central Sweden. The main types of
disturbances
are storm damage, fungal infection and insect
attacks. The response
of the different
tree species varied and
the mode of tree-fall depended
on the different
combinations
of disturbance
agents. The DBH distributions of gap creators
and
gap-border
trees were almost
the same. There was a high
age diversity
(100-240
yr)
among
the fallen trees. We concluded
that all canopy trees (DBH > 20 cm) had the same probability
of being felled by storms,
irrespective
of their
age and DBH.
According
to an estimate along transect
lines, gaps made
up 31% of the spruce
forest area. Individual
gap sizes ranged
from 9 m2 to 360 m2, but 83% of the gaps were < 150 m2. The
varied age structure of logs in individual
gaps indicated that
gap enlargements
were common. 96 tree-falls were observed
on four days with an hourly
mean wind speed > 12.0 m/s; all
trees fell in the direction of the wind. However, when we
consult the 30-yr record
(1959-1989) of the mean
hourly
wind
speed
> 12.0 m/s, it is clear that the pattern
of storm-directions
does not match
the pattern
of orientation of fallen logs.
The present
disturbance
regime and the predominance
of
small
gaps
were more
favourable for the
regeneration
of Picea
abies than of light-demanding
tree species. In one large,
2900
m2
gap, not crossed
by the transects,
all the major
tree
species
had established within 7 yr, suggesting that
classical succes-
sion in the sense of complete species replacement
or 'relay
floristics' did not occur. Our observations
seem rather
to fit
the 'initial
floristic' model. Estimates of turnover
time ranged
from 170 to 228 yr, depending
on the method used.
Keywords: Disturbance
regime; Fiby; Gap creator;
Storm-
gap; Succession; Sweden;
Tree-fall.
Nomenclature' Lid (1985).
Introduction
Gaps are considered to be important for both the
regeneration, dynamics and diversity of several types of
forest ecosystem. Gaps were recognized by Cooper
(1913) on Isle Royale and nearby small islands in the
boreo-nemoral forest zone (sensu Sjors 1963). Cooper
noticed that the regeneration of two of the most common
species, Abies balsamea and Betula papyrifera, took
place in wind-fall areas. Balsam firs growing in groups
were usually of similar age. This was interpreted as a
result of regeneration in gaps.
Semander (1936) formulated a theory about storm-
gap dynamics and regeneration in primeval spruce, Picea
abies, forests in central Sweden. He emphasized that the
mosaic structure of the forest was the result of repeated
treefalls and he called this the 'storm-gap structure'
(1936, p. 220). Semander regarded storm-felling in
spruce forests as a periodic phenomenon. He particularly
emphasized the role of 'dwarf trees' and measured the
height increment of these 'dwarf trees' in connection
with gap formation. However, no quantitative studies of
disturbance regime, gap structure and tree regeneration
were carried out.
The importance of gaps has been described in many
forest types, e.g. nemoral forests in the eastern United
States (Runkle 1981, 1982, 1984; Runkle & Yetter
1987); nemoral Fagus forests in Japan (e.g. Nakashi-
zuka 1984; Yamamoto 1989); high altitude conifer for-
ests in the eastern United States (White, MacKenzie &
Busing 1985; Foster & Reiners 1986) and Japan (Kanzaki
1984); boreal forests in Sweden (e.g. Hyttebom, Packham
& Verwijst 1987; Bradshaw & Zackrisson 1990); ever-
green broadleaved forests in Japan (Naka 1982; Naka &
Yoneda 1984); temperate rainforests in New Zealand
(Veblen & Stewart 1982); beech/hardwood forest in
New Zealand (Stewart 1986); humid tropical forests
(see reviews by Brokaw 1985; Denslow 1987; Swaine,
Lieberman & Putz 1987).
The general conclusion drawn from these studies is
that the micro-environment of the forest is changed after
gap formation, especially insolation on the forest
floor. The species composition of the regeneration phases
is largely determined by the size of the gaps; the
391
Liu Qinghong & Hyttebom,
H.
regeneration
of light-demanding species is expected
to
occur in big gaps. Except for the classical studies by
Semander and
Cooper
few gap
studies of boreo-nemoral
and boreal regions are known to us. Our study is an
addition to the literature on boreo-nemoral
forests; it
deals with Fiby urskog,
a primeval
forest situated in the
boreo-nemoral
region (Sj6rs 1963) ca. 100 km south
of
the border with the boreal region, the forest where
Semander
(1936) developed his theory.
In this study we aim to answer the following ques-
tions about forest
dynamics: 1. Is gap-structure
charac-
teristic of this forest? If so, what are the quantitative
features
of gap structure? 2. What are the gap-forming
processes? 3. Are the gaps of any importance
for re-
generation?
If so, what is the relationship
between gap
structure and regeneration?
4. What is the overall dis-
turbance
regime?
Material and methods
Investigation
area
Fiby urskog is a nature reserve in central Sweden
(59053' N, 17021' E) consisting largely of primeval
Picea abies forest. The general
features and the history
of the area have been described by Semander (1918,
1936), Hesselman (1935), Berg & Lundqvist (1964),
Hyttebom
& Packham
(1985,1987) and Leemans
(1986).
The area measures ca. 64 ha and is surrounded
by
managed coniferous forests and agricultural
land. In
addition to the Picea abies forest, Pinus sylvestris and
Betula forest occur. The main part of the reserve is
Picea abies forest, with which this study is concerned.
Sernander
was convinced
that
the forest had suffered as
a result of a notorious storm
in 1795
and that the scattered
old trees with large
crowns and
widely spread
branches
found in the 1930's survived this storm and underwent
enhanced
growth after
having been released from sur-
rounding
trees. Hesselman (1935), on the other hand,
argued
that the structure resulted
from an earlier more
open forest with grazing.
According to recent research (e.g. Hytteborn &
Packham 1985) the spruce forest has been very little
disturbed
by logging or grazing during
the last century,
even though Berg & Lundqvist (1964) found more
stumps from cutting than Sernander was aware of. On
the other hand, these stumps
were certainly
more than
100 yr old and perhaps 200-300 yr old (Berg &
Lundqvist). There are no signs of damage by forest
fires; only a small part of the pine forest in the area
burned
down in 1944 (Berg & Lundqvist
1964).
Definitions
We defined a gap as part
of a forest stand where one
or several trees (or limbs) have created
an empty space
in the canopy
either
through
death,
or through
death and
subsequent
falling, and where regenerating
trees have
not reached
more than 2/3 of the average canopy tree
height (21.6 m, range 10.5 - 31.5 m; n = 703). A gap
can exist only in a matrix
of mature
trees, and with a
recognizable,
more or less continuous border
consisting
of trees. The gap is delimited by the projection
of the
surrounding
mature
canopy trees. The tree which cre-
ated the gap is called the gapmaker.
In practice
we de-
fined canopy
trees
as stems, and
gapmakers
as logs with
a DBH (diameter
at breast
height) > 20 cm.
Besides gaps, two other kinds of openings
are
recog-
nized in the area.
Permanent
openings are found where
substrate conditions do not permit
tree growth,
includ-
ing both wet areas (small fens) and rocky areas with
shallow soil. Crown
openings are openings
not created
by the fall of trees or branch-breakage,
but existing
between tree crowns where these do not touch each
other.
Log age was defined as the age of a log (gapmaker)
since it fell. Fallen tree age is the age of a tree when it
died.
Field methods
Gaps were detected
along six parallel
line transects
covering
the whole forest area and with a total
length
of
3176 m. The transects ran in a N-S direction and the
distance between them was 100 m. Sections of the
transects
crossing rocky pine forest
(in total 910 m) have
been left out of consideration.
Thus, the actual meas-
urement
length
for
Picea abies forest
was 2266 m. Gaps
and permanent openings intersected by the transects
were
investigated,
but not the crown
openings.
Accord-
ing to Runkle
(1982), data
on gaps collected in transect
studies may be biased as large gaps have a higher
probability
of being sampled than smaller ones. The
direction of the transect
may affect the distribution of
gap directions, but not their area; gaps with a N-S
orientation
may have a smaller chance of being hit by
the
transects,
but
if intersected,
they are
longer
than
gaps
with a different
compass orientation. An unusually
big
gap, not crossed by any of the transects,
was investi-
gated
separately
in 1989. This large
gap is not included
in the
results
presented
here,
unless indicated otherwise.
Gap
measurements included
length
along
the
transect,
gap shape and direction
of the long axis, length of the
long axis, and width (at the gap centre). Data on gap-
makers included species identification, DBH, length,
orientation on the surface,
mode of death
and estimation
392
- Gap structure,
disturbance and
regeneration
in a primeval
forest -
Table 1. The decomposition
degree scale used for the estima-
tion of log age. See also text. Old
= six-degree
Semander scale
Part Age Degree Description
(yr) New Old
Crown 1-2 1 [1] Most of the leaves/needles remain
on
branches;
some of them are still green.
3-5 2 [2] Most of the fine twigs remain and some
of them have yellow leaves.
6-10 3 [2] Without
leaves; a few fine twigs left;
most of the bark
remains;
the wood still
hard
and not rotten.
11-20 4 [2] Without fine twigs, but main branches
remain,
supporting
the trunk;
small areas
covered
by bark;
wood has started to rot.
Trunk 21-30 5 [3] Trunk has rotted to 1-2 cm depth
and a knife can penetrate
the wood.
31-40 6 [4] Trunk lies on the ground
or is supported
by very thick
branches,
rotten wood
5-10 cm; a knife can easily penetrate
the wood; the branches disconnect
from
the trunk and
can easily be moved by hand.
Log 41-50 7 [5] Whole log is rotten,
almost no branches;
log still keeps the form of a cylinder.
51-60 8 [6] Whole log is very soft and
has sunk to half
of the original
diameter or more. It may
easily be crushed
with the foot.
61-70 9 [6] Almost
disappeared,
some fragments
could
be found under the bryophyte layer.
of log age. Data on gap border trees included species
identification,
DBH and
height. Seedlings, saplings
and
young
trees in the gaps were recorded
by species, height
(in five height classes) and location in one of four
sections delimited by the long and short axes of the
gaps. The age of 14 newly fallen trees was determined
by coring at the base and
counting
of the annual
rings.
Four
transects
were investigated
in the western
part
of the forest (autumn
of 1988) and the other two in the
eastern
part
(spring
of 1989). The measurements in the
spring 1989 were slightly different from the earlier
ones. In total, 88 gaps were investigated,
seven in the
late autumn
1988 when snow covered the ground;
these
were used only in the calculation of gap area.
Gaps are created either by dead standing
trees or
through
tree-fall or branch-breakage.
At least four dif-
ferent modes can be distinguished:
(1) uprooted
living
or dead trees (not detected in this study), (2) trees
snapped
off while still living, (3) trees snapped
off after
death,
and
(4) dead
but still standing
trees.
Mayer
(1989),
in a review of tree damage in connection with storms,
Table 2. The reaction
of different
species of gapmakers
(DBH
> 20 cm) to the natural disturbances
in 81 gaps (0.6822 ha) in
Fiby urskog sampled
from autumn
1988 to spring 1989.
Dead Snapped Up- Not I
Species standing off rooted determined
Betula spp. 0 9 0 9
Picea abies 11 116 72 2 201
Pinus sylvestris 20 5 2 27
Populus tremula 4 24 4 32
Undetermined 1 1 3
Sum 35 155 79 3 272
distinguished four types of tree-fall: stem breakage,
stock breakage, root breakage, and tree throw. In prac-
tice we found it difficult to distinguish between these
types; hence the two first were combined into 'snapped
off' and the other two into 'uprooted'. Schaetzl et al.
(1989) suggested even more categories of uprooted
trees, but Putz et al. (1983), Yamamoto (1989) and
others considered the three categories used in this study
adequate for the description of a variety of forests. The
distinction between stem breakage of living and dead
trees respectively became obvious only during the work
in the spring of 1989, after the first four transects had
been finished. Differences in the modes of death are
probably important in relation to the decomposition rate
and regeneration on logs. Since it is difficult to distin-
guish reliably between these different modes of death of
old logs, especially between living snapped and dead
snapped logs, only two categories were recognized here,
namely uprooted stems and snapped off stems; how-
ever, 49 recently fallen stems were studied in more
detail. The distribution of modes of tree-fall were com-
pared with similar data shown on maps of five semi-
permanent plots published by Semander (1936).
We used three methods to estimate gap age:
1. Semander's (1936) log decay scale, originally with
six categories, as revised by Hyttebom & Packham
(1987) and extended here to nine degrees (Table 1).
2. Year-of-release of the successors as detected from the
increased length of the annual shoots.
3. Age determination of saplings developing in the gap.
The degree of decay of a log depends not only on the
time of falling but also on log size and species, its
position (rests on the ground or on broken branches) and
the status at the time of the fall (dead or alive). Different
parts of the same log can decompose at uneven rates.
Since Picea abies was the predominant species in the
log population (Table 2), we used Picea logs to build up
a log decay scale and estimate gap age. Semander (1936)
estimated that the time taken by a log to decompose
393
Liu Qinghong & Hytteborn,
H.
Disturbance combination The response of the trees
Fungi
and/or insect attack ---1 Dead
standing
tree,
rotten or not |
-- ' Dead
snapped,
rotten or
not
Fungi
and/or insect
attack | Living snapped
with
rotten wood J
followed
by
storm L
K--" Living
uprooted
with
rotten wood
| Living snapped
without l
K
rotten wood
| Storm |L | Living
uprooted
without
}torm |--' rotten
wood
Fig. 1. Scheme of combinations
of disturbances and the re-
sponses of the trees, based on observations
in Fiby urskog.
completely is ca. 90 yr. Hyttebom & Packham
(1987),
after having re-analysed one of Sernander's semi-
permanent plots, reduced this estimate to under 70 yr
(cf. Hyttebom, Liu & Verwijst in press). Semander
(1936) reported
that
stages 1 and 2 would last 4 and 20
yr respectively,
while he had no direct
observations
for
the later
stages. The estimate
of 70 yr for total decay is
probably
valid only for large trunks.
Age determination
by the 'year-of-release'
method
and age determination of saplings are based on the
assumption
that fast growth and regeneration
occurred
in direct connection with gap formation This is not
always
true, however;
the growth
might
start after
a later
enlargement
of an already
created
gap. So, these esti-
mates must be considered as minimum values.
Other
researchers,
including Semander,
cored
stems
of released trees or gap border trees to estimate
gap age
(Runkle
1982; White,
MacKenzie & Busing 1985;
Fos-
ter & Reiners 1986). When trying this method we did
not find a clear increase
in the radial increment.
In cases where the age determinations
obtained
by
means of the three methods
gave rather
different
results,
gap age was determined
by the degree of log decay.
49 new logs which fell in the autumn of 1988 and
spring of 1989 were investigated in belts, 30 m wide,
along the first four transects. These logs appeared
after
the first four transects had been finished and were not
included
in the investigated
gaps. They were measured
and cored at their base. Only 14 cores could be counted;
the remaining logs were largely
rotten.
Calculations
The area of each gap was calculated using the
formula
for an
ellipse. For the calculation of the quotient
between the gap diameter and mean height of border
trees of each individual
gap,
the diameter of a circle with
the same area as the ellipse was used. Gap area as %
of
the total area was estimated as the ratio between the
length of the transect
intercepted
by gaps and the total
transect
(cf. Greig-Smith
1983). For comparison,
three
0.25 ha plots were also used to estimate
gap area.
In boreal forests disturbed
by fire, turnover
time is
usually estimated
using fire scars combined
with stand
ages (Heinselman 1973). Runkle (1982) used gap area
and
gap age
distribution
and
White,
MacKenzie
& Busing
(1985) estimated disturbance
regimes in several differ-
ent ways based
on tree
age and
on gap area. We used the
mean
age of trees at the time of falling, and
gap creation
rate to estimate turnover time.
Correlation
analysis (SAS 1985) was used to calcu-
late correlation between
relative
gap size (D/H) and the
number of tree seedlings and saplings per gap and per
m2.
A semivariogram
was used to detect
spatial
trends in
gap formation (Burrough 1987), using 2 m transect-
segments as sample units and proportion
of gaps as
variable.
Wind
speed
Hourly mean wind speed (HMS) data were used
from the Marsta
meteorological
station ca. 15 km NNE
of Fiby urskog and some of these data were correlated
with data
on tree-falls.
Results
Mode of death
The main natural disturbances
occurring in Fiby
urskog
are storm-fellings, fungi and insect attacks
and
various combination
of these agents. Observations
of
49 logs which fell in the winter of 1988 and spring
of
1989 showed that
the mode of tree-fall was sensitive to
differences
in disturbance
agents. Among these 49 logs,
living snapped
and
uprooted
trees made
up 57 %
and 41
%
of the total logs, respectively.
86%
of living snapped
trees, but only 15 % of the uprooted
trees were rotten.
These differences were significant (Fisher's
exact test;
p< 0.001).
On the
basis of direct observations
of logs which had
only just or recently
fallen, and coring
of trees/logs, a
scheme was made which summarised combinations of
disturbance
agents and tree responses (Fig. 1). This
scheme
has not yet been applied
to other
boreal
forests.
Uprooted
trees result from the combination of storm
and shallow soil, living snapped trees from fungus
infections and insect attacks followed by storms. Dead
standing
trees, which form a nutrient
gap in the soil and
394
- Gap structure,
disturbance
and
regeneration
in a primeval
forest -
a less pronounced
light gap,
result
only from
insect and/
or fungus
attacks. Other natural forces such as lightning
are not important
in the area.
The mode of death of gapmakers
(DBH > 20 cm)
varied among tree species (p < 0.001) (Table 2). Pinus
sylvestris
had a high percentage
of dead standing
trees
and the two major
deciduous
trees Betula spp. (pendula
and pubescens) and Populus tremula had a high per-
centage of snapped-off stems. In Picea abies a low
proportion
of the stems were
dead
standing
and rather
a
high proportion uprooted.
The mode of death was not
significantly
affected
by the
diameter
in either the whole
log population
(x2,
p > 0.5) or the Picea logs only (p >
0.6).
There was a significant
statistical interaction
between
the two investigation periods,
1935 (Semander,
Table
3)
and 1988-1989 (Table 2); stem breakage was more
frequent today than
50 yr ago (X2,
p < 0.002).
50 ' r? -- Border tree
40 - 0
_ ? Gap creator
30 "
20
"
10 "
0 i i
10 20 30 40 50 60 70
DBH class (cm)
Fig. 2. Comparison between the relative frequency of diam-
eter classes (DBH > 20 cm) of border trees (n = 761), and of
gapmakers (n = 272).
Gapmakers
The DBH distributions of the gapmakers and of the
border trees were not significantly different (V2,p > 0.07)
(Fig. 2). However, the species composition between the
gapmakers and the border trees differed (X2, p < 0.001,
Fig. 3), with more Populus tremula and less Pinus
sylvestris among the logs than expected.
Ca. 62 % of the logs were lying in the directions SW,
S and SE (Fig. 4a). A similar distribution was found by
Semander (1936) (Fig. 4b). There was no interaction
between mode of tree-fall and log orientation in our data
(2, p >0.1).
The age determinations indicated that tree-fall oc-
curred in trees aged 100 to 250 yr. Most of the trunks
were between 120 and 200 yr of age (Fig. 5).
80
70
60
50
40
30
20
10
0 Bes Pia Pot Pis
Species
Fig. 3. Comparison
between the relative frequency of tree
species of border
trees(n
=761) and of gapmakers
(n = 272).
Bes = Betula spp., Pia = Picea abies, Pot = Populus tremula,
Pis = Pinus sylvestris.
Table 3. Frequency
of logs created
by trunk
breakage
(sn) and
by uprooting
(upr)
in five mapped
plots, according
to Semander
1936 (mapping performed
in 1935).
Plot Betula Picea Populus I
No. Size sn upr sn upr sn upr sn upr sum
(ha)
1 0.250 - - 25 19 - - 25 19 44
2 0.375 7 2 12 27 8 1 27 30 57
3 0.750 - - 16 44 6 - 22 44 66
4 0.500 - - 17 17 3 - 20 17 37
5 0.400 - - 18 15 10 1 28 16 44
b 2.275 7 2 88 122 27 2 122 126 248
N
N
b
Fig. 4. (a) Relative frequency
of log orientation in 16 direc-
tions, data from 1988-1989. (b) Relative frequency of log
orientation in eight directions,
data from Semander
(1936).
395
Liu Qinghong & Hyttebom, H.
Freq.
7 -
5
4
3
2
1
0 60 100 140 180 220 260
Age class (years)
Fig. 5. Age distribution
of the fallen trees which fell during
the
autumn
of 1988, and
winter
and
spring
of 1989. The values are
given as midpoints
of age classes.
Freq.
12 '
10
8
6
4
2
0 0
ULJI
L
30 35
5 10 15 20 25
Gap size (10 m )
Fig. 6. Frequency
distribution
of gap sizes, 1 = 0-10 m2, 2 =
11-20 m2, etc.
Freq.
20
15
10
5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Relative gap size (D/H)
Fig. 7. Distribution of relative
gap size calculated as the ratios
between the gap diameter
(D) and the mean height of the gap
border
canopy trees (H).
Table 4. Length of transects,
the length
and
the % of transects
intercepted
by gaps and permanent openings; the number of
gaps and permanent openings in the spruce
forest canopy in
Fiby urskog.
Transect Gaps Permanent
openings
No. length of line % number of line % number
(m) (m) (m)
1 235 75.3 32.0 10 8.2 3.5 1
2 294 122.5 41.7 13 1.5 0.5 1
3 335 114.2 34.1 12 0 0 0
5 410 122.8 30.0 18 13.6 3.3 2
6 570 121.3 21.3 19 5.5 1.0 1
7 422 139.0 32.9 16 41.1 9.7 6
Sum 2266 695.1 88 69.9 11
Mean 30.7 3.0
Table 5. Gap
area estimated
by plots
in the
spruce
forest of Fiby
urskog,
autumn
1988.
Plot No. Plot area Number Gap area*
(m2) of gaps m2 %
1-1 2500 8 673.8 27.0
2-1 2500 9 758.0 30.3
2-2 2500 10 842.2 33.7
* The gap area was estimated
by using the mean
gap area
(84.2 m2)
of
the 81 gaps along the transects
x the number of gaps in
the plots.
Characteristics of the gaps
As estimated by the transects, ca. 31% of the total
spruce forest area was gap (Table 4). A semivariogram
was used to test the spatial variance of gaps. The
semivariance levelled out at about 10 m, suggesting that
there was no large-scale spatial trend in the distribution
of gaps. The number of gaps and the gap percentage
between the six transects showed no significant differ-
ences (no. of gaps/line: X2,
p > 0.5; gap %: X2,
p > 0.1).
The investigation of three plots gave the same gap area
and an estimate of 36 gaps per ha (Table 5).
The individual gap areas ranged from 9 m2 to 370 m2
(Fig. 6), excluding a large gap of 2900 m2 outside the
transects. The mean gap area in transects was 84.2 m2
and the median gap was 63 m2. Gaps smaller than 250
m2 and 150 m2 made up 98% and 83% of the total
number of gaps, respectively. Small gaps were the most
common ones and these were probably under-represented
in the transects (Runkle 1982). Only five gaps had a
396
- Gap structure,
disturbance
and regeneration
in a primeval
forest -
Table 6. Number of seedlings and
saplings
in different
height
classes (cm) in 81 investigated gaps (0.6822 ha) along six
transects in Fiby urskog,
autumn 1988 - spring 1989.
Species < 20 20-49 50-99 100-199 > 200 Sum
Alnus
glutlnosa 0 3 0 1 0 4
Betula spp. 88 208 67 11 26 400
Juniperus
communis 0 2 0 0 0 2
Picea abies 5045 1438 240 183 356 7262
Pinus sylvestris 260 101 5 1 1 368
Populus tremula 131 239 32 0 6 408
Quercus
robur 0 0 0 0 1 1
Sahx caprea 0 0 0 0 2 2
Tilia cordata 0 0 11 10 14 35
Sum 5524 1991 355 206 406 8482
Mean
m-2 0.810 0.292 0.052 0.030 0.060 1.243
ratio between gap diameter and tree height above 0.75,
which is the threshold
beyond which a gap gets direct
sunlight
in mid-summer
(Fig.
7). The
biggest gap, which
was investigated separately,
was formed by the fall of
47 trees, most likely on one occasion in late autumn
1982 (pers. comm. Christina
Skarpe),
which combined
several
small,
old gaps.
The
gap
was enlarged
afterwards
by 16 new dead
trees,
and the enlargement
continues.
In
the autumn of 1989 it contained
a total of 124 logs of
different
age classes.
Most
gaps
contained several
logs, up
to 10
in number,
and about 20% contained
only one gapmaker
(Fig. 8).
However,
most gaps were initially
created
by the fall of
one tree,
and
were gradually enlarged
afterwards. Ca. 65
% of the gaps
contained more
than
one age-class of logs,
indicating
the occasions of enlargement (Fig. 9).
The long axis of the gaps showed no significant
trend in direction
(X2,
p > 0.5).
Regeneration
The most common seedling-species (< 50 cm) and
sapling species (50-200 cm) found in the 81 gaps was
Picea abies (Table 6). About the same numbers of
Pinus,
Betula and
Populus were found. Most Picea and
Pinus saplings were in the smallest class (< 20 cm),
most Betula and
Populus saplings
were in the next class
(20-49 cm). The few individuals
of Tilia cordata and
Quercus
robur we found occurred in the same gap.
The saplings
were mainly
found
in the northern
part
of the gaps, especially Betula and Populus. The differ-
ence was significant for Betula (92, p < 0.001), Picea
abies (p < 0.001) and
Populus tremula
(p < 0.001), but
Freq.
20 -
15 -
10 -
5-
0 1 2 3 4 5 6 7 8 9 10
No. of gap creators / gap.
Fig. 8. The frequency distribution of the number of logs
(gapmakers)
/ gap.
Freq.
30 -
25 -
20 -
15
10
5
0 2 3 4 5 6 7 8
No. of log age classes / gap.
Fig. 9. Frequency
distribution of log-age classes (determined
by decay degree) per gap. This is an indicator of occasions of
gap enlargement.
not for Pinus sylvestris (p > 0.2).
The numbers of seedlings and saplings of Picea
abies and Betula spp. and seedlings of Pinus sylvestris
increased with relative gap size (D/H) (r = 0.33 - 0.49,
p < 0.001). The seedling densities of Picea abies, Pinus
sylvestris and Populus tremula decreased slightly with
D/H (r = - 0.12 to - 0.25), p < 0.004), while the density
of Betula spp. saplings slightly increased with D/H (r =
0.27, p < 0.004). Few Populus tremula and Pinus
sylvestris saplings occur in the recent gaps. The regen-
eration of Populus tremula was almost uncorrelated with
D/H. Picea seedlings and saplings were found at rather
a low relative gap size, about 10% (the lowest value),
397
Liu Qinghong & Hyttebom, H.
N
a
b
b
Fig. 10. Relative frequency of wind direction; a) wind at
hourly
mean
speed (HMS) > 12.0
m/s; b) wind at HMS
> 14.0
m/s. Data
from the Marsta
Meteorological
station 1959-1989.
Table 7. Number of seedlings and
saplings in different
height
classes (cm) in the biggest gap (0.29 ha) created in the winter
of 1982/1983 in Fiby urskog.
Species < 20 20-49 50-99 100-199 > 200 Total
Betula spp. 0 88 148 120 36 392
Picea abies 664 708 216 108 56 1752
Pinus sylvestris 16 32 0 0 0 48
Populus tremula 0 20 12 0 0 32
Salix caprea 0 12 20 20 20 72
Sum 680 860 396 248 112 2296
Mean/100
m2 24 30 14 9 4 80
Table 8. Frequency
(F) of days with hourly
mean
wind-speed
above 10
m/s, and the number of gap formations
(Gf) observed
(- = no observations).
Witness of
10.1-12 12.1-14 14.1-16 16.1-18 treefall
F Gf F Gf F Gf F Gf
1981 19 - 5 - 2 2 0 - Hyttebom
1982 10 - 3 - 2 1 0 - Skarpe
1983 15 - 8 - I - 0 -
1984 11 - I - 0 - 1 -
1985 6 - I - 0 - 0 -
1986 11 - 9 - 0 - I -
1987 11 - I - I I 0 - Leemans
1988 7 - 4 1 1 1 0 - Liu
1989 11 - 6 - 3 1 1 1 Llu
Betula seedlings and saplings can be found at a relative
gap size of ca. 30%.
Picea saplings also had the highest density in the
large gap (Table 7). The height distributions of Picea
and Pinus in the big gap differed from that in the small
gaps, with more individuals in the height class 20-49 cm
than in the class 0-19 cm. Most individuals of Betula
were between 0.5 and 2 m high (Table 7).
Table 9. Mean annual number
of days with different
hourly
mean wind-speed (HMS) above 10 m/s, in different seasons
(Marsta,
1959-1988).
Season 10.1-12 12.1-14 14.1-16 16.1-18 Total
m/s m/s m/s m/s
Winter
(Dec. - Feb.) 5.23 1.45 0.74 0.10 7.52
Spring
(Mar.
- May) 4.90 1.65 0.19 0.07 6.81
Summer
(Jun.
- Aug.) 2.06 0.32 0.10 0.00 2.48
Autumn
(Sep. - Nov.) 4.16 0.87 0.19 0.03 5.25
Year 16.35 4.29 1.22 0.20 22.0(
Frequency of tree-felling
wind, wind direction and log
orientation
Several
tree-fall events were observed in the field at
an hourly mean wind-speed (HMS) between 14.1 and
16 m/s. In one case, tree-falls were also found at 12.1-
14 m/s (HMS) (Table 8). Storms with HMS above 14.0
m/s occur with
a frequency
of ca. 1.4/yr,
and are
concen-
trated
in the winter season (Table
9). We found in our
four tree-fall events that the trees fell in the same direc-
tion as the wind was blowing (recorded
also as hourly
mean values; 97 logs were measured);
however, the
pattern
of dominant storm
directions from 1959 to 1989
does not match that of the main direction
of the whole
log population
(cf. Fig. 4a and
Fig. 10). The direction
of
prevailing
storms was from WSW, but the main direc-
tion of the logs was towards SSE.
Turnover
time
The turnover time can be estimated in several ways.
The simplest estimate may be the age of the trees at
death.
This gave a mean value of 171 yr and a median
value of 170 yr. This estimate is based on the inappro-
priate assumption
that the regeneration
occurred
directly
after the
gap formation,
and
ignores
advanced-
and
post-
regeneration.
Gap-filling
rate will be discussed in a later
publication.
According to Hyttebom & Packham
(1987), a log
will disappear after 60-70 yr. As only gaps with
gapmaker(s)
were counted, the oldest gaps recorded
were at most 70 yr old. The gaps occupied 31% of the
spruce
forest, which gives a gap creation rate of 51.2 -
43.9 m2. ha-
I yrI (for
60 and 70 yr,
respectively),
and
a
turnover
time of 195 - 228 yr.
398
- Gap structure,
disturbance
and
regeneration
in a primeval
forest -
Discussion
Gaps
The forest is still characterized
by the gap-structure
Semander
(1936) described in 1935. Total gap area is
unusually high. A great variation of gap areas is re-
ported in the literature, recently summarized by
Yamamoto
(1989). Published
figures for total gap area
seem higher
for temperate
forests than for tropical
for-
ests. A gap percentage
of over 30% is known from
only
two other studies
(Foster
& Reiners 1986; Nakashizuka
& Numata 1982). Yamamoto
(1989) measured
a varia-
tion of gap area between 4 and 21%, while Runkle
(1982) found
values between 3.2 and 24.2 %
(mean
9.5
%). However, such figures are not always comparable,
since the lower limit of gap size, average height of
foliage above ground and gap boundary
may follow
different
definitions.
Kapos
et al. (1990) argued
that
gap
area
may not be an important
parameter
for the differ-
entiation between forest types.
The gap size distribution
is skewed, with many
small-sized gaps, as most such distributions are (cf.
Yamamoto
1989;
Naka
1982;
Runkle
1982;
Nakashizuka
1984; Foster & Reiners 1986; Brokaw 1985; Denslow
1987). The gap
size distribution
depends
upon
the distri-
bution between single and multiple-tree fall gaps
(Denslow 1987),
the
investigation
method
(Runkle
1982)
and the average
tree size and
gap
enlargements.
The gap
sizes in Fiby were rather small in spite of continuous
enlargement.
With
such a high
gap
percentage
as in Fiby
and with a calculated width of the border
trees of 4 m
around
each gap, every subsequent
tree-felling will be
an enlargement.
The special feature of the primeval
forest in Fiby is thus not enlargement,
but the high gap
percentage,
which is related to slow gap-filling.
Turnover
time
Turnover time is defined as mean time between
disturbances at any point in a forest (White & Pickett
1985); it is a parameter
for describing periodic
changes
in the vegetation.
In practice
the estimation of turnover
time could give rather different
values depending
on the
method
used. Semander
(1936, p. 17) realized that
big
storms occurred
periodically,
but he did not calculate
turnover time from a forest dynamics
point of view.
Hartshorn
(1989) gives an alternative
definition for
turnover
time, as "the number of years to cover a unit
area of forest
(e.g. 1 ha) using the average
annual area in
gaps"
(cf. Hartshorn
1978;
Brokaw
1982). Estimation
of
turnover
time, based on gap creation rate, assumes a
stable
mosaic pattern
of the forest
composed of patches
of different
ages. We have inferred
from the size distri-
butions of trees and
logs and the distribution of gap age
area,
that
over a time period
of 50 to 70 yr the forest
has
been in a steady state.
However,
this period
is only part
of the total turnover
time. The estimation of turnover
time is also based on gap percentage
and
maximum
gap
age. The maximum
gap age is derived
from
re-mapping
of old plots (Hyttebom & Packham 1987; Hyttebom,
Liu & Verwijst in prep.). It now seems, after further
experience,
that
60-70 yr is too long as an average
value
for a log to decompose completely.
Thus
the
estimate of
turnover time should
probably
be reduced.
The turnover
times we arrived
at are at the upper
limit of what have been estimated
for temperate
hard-
wood forests (Runkle 1982, 1985; Nakashizuka
1984;
Yamamoto 1989), and subalpine coniferous forests
(White, MacKenzie & Busing 1985), but lower than
Foster
& Reiners
(1986) estimated
for another
subalpine
coniferous
forest. The
time is also longer
than
Zackrisson
(1977) and Engelmark
(1984) have stated for fire dis-
turbance
in northern
Sweden.
The scale of disturbance
and
forest dynamics
The most common disturbance in a natural
forest
landscape
in the boreo-nemoral/boreal
forest region is
probably
forest fire (Heinselman
(1973). Wind throw,
snow damage,
diseases
(mainly fungi
and insect
attacks),
etc. (cf. White 1979) are probably
of less importance.
The area disturbed
by each single fire is on average
much bigger than the area disturbed
by wind throw.
Clearly, disturbances on different scales give different
mosaic patterns,
coarse-scaled and fine-scaled.
The fine-scaled pattern
created by one or several
tree-falls
was probably
uncommon
even before the time
of heavy
human
influence on the landscape.
Fire
refugia,
however, have existed long enough to create this fine-
scaled pattern
inside the coarser pattern
(Heinselman
1973). Such refugia are found on islands, mineral soil
'islands' in peatlands,
areas close to lakes etc., where
fire occurred
infrequently.
National
parks
and reserves
in the boreal
region
may, in the future,
develop into this
small-scale mosaic. The disturbance
pattern in Fiby
urskog
and also in northern
boreal
Sweden (Hyttebom,
Packham
& Verwijst 1987) shows that
windthrow is an
alternative
rejuvenating
factor
operating
on vegetation.
The fine-scale mosaic promotes
regeneration
of shade
tolerant
species, e.g. Picea abies, and should lead to a
decline of the light-demanding
trees, Populus tremula,
Betula spp. and Pinus sylvestris.
We believe that these
shade-intolerant trees grew up after large gaps were
formed.
However, continuing gap enlargement
and
low
growth rate may result in a more open forest, which
could improve the ecological conditions for establish-
ment of these species.
399
Liu Qinghong & Hyttebor, H.
Mode of death and gapmakers
In the four cases of tree-fall we observed,
the trees
fell in the storm direction. Semander (1936) reported
the same pattern.
However, the climate data, 1959 -
1989, showed that the prevailing storm direction was
almost perpendicular
to the log orientation.
Only a few
storms came from a northerly
direction
(Fig. 10). Pre-
vailing
storm
directions
at different threshold wind
speeds
were the same. Several factors may account for this
discrepancy. Wind intensities correlated with tree-falls
are
probably
not measured
effectively as a mean
hourly
wind-speed, but these were the only available data in
our case. Discrepancies may also result from other fac-
tors,
notably:
(1) tree-fall
is weather-determined
and not
climatic; it occurs accidentally
and not regularly;
(2)
northerly
storms have a larger
variation
in speed; (3) the
wind data are not representative
of our study area;
(4)
trees adjust
their structure as they grow in accordance
with normal wind conditions. If this is so, storms from
unusual directions will most likely topple the trees.
However,
Semander found a correlation between number
of days with wind speeds > 12 m/s as divided between
different
directions,
and log orientation.
According
to Putz et al. (1983), wood properties in
a tropical
forest are most important
in determining
the
mode of death. The sensitivity to storms and fungi is a
characteristic
of the species, with Picea more suscepti-
ble than
Pinus (having a pronounced tap root). The in-
creased
frequency
of snapping
off compared
with 1936
may be a result of an increased frequency of fungus
infection, as it is probable
that the age of the stand is
greater
now than
in 1935. Soil characteristics
may also
be important;
Mayer
(1989) found more stem
breakages
in frozen soil, and
Mayer
(1989) and
Yamamoto
(1989)
found more uprooted
stems in wet soil.
Contrary
to our
results,
tree size has elsewhere
been
reported
to influence the mode of tree-fall, with larger
trees
being subjected
to a higher frequency
of uprooting
(Putz et al. 1983) or for Picea to lower frequency of
uprooting (Mayer, Schenker & Zukrigl 1972). Brewer
& Merritt
(1978) and Naka (1982) found that bigger
trees were more sensitive to wind than smaller trees.
The rather small size variation
of trees
in our study
area
may be the reason that our data do not show any
interactions between size and mode of tree-fall.
The close correspondence
between
log size distribu-
tion and the present
distribution
of border trees can be
explained by the double assumption that tree-fall is
independent
of size over a certain limit and that the size
distribution
in the stand
has not changed during
the last
50-70 yr. Alternatively
the difference between log and
border tree size distributions
may simply not be a sen-
sitive measure of canopy structure
change.
Regeneration
Light intensity is thought to play a major role in
forest
regeneration,
and hence a meaningful
measure
of
gap size is the ratio
between
the diameter
(D) of the gap
and the mean
border tree
height
(H). The significance
of
the D/H quotient varies with the latitude, and in the
study area the angle of the sun varies between 53.5? and
30? in the growing season, 53.5? and 6.5? over a year.
Consequently, the gap floor on a horizontal surface
receives direct sunlight at a D/H ratio of 0.74 at mid-
summer and of 1.73 in midwinter. Only five gaps - in
addition to the big one - could get direct sunlight if
the border trees surround the gaps without
interruption
(Fig. 7).
The distribution of gap sizes, and the composition
of
species and size of seedlings/saplings
indicated that the
present
disturbance
regime
is favourable for Picea abies
regeneration.
The difference between performance
of
seedlings/saplings
of Pinus sylvestris and Betula spp.
indicated that Pinus is more dependent
on gap size and
Betula more on substrate,
the latter
species being most
abundant on highly decomposed logs or mineral soil.
In the big gap, the number
and height of Populus
tremula
saplings was low compared
to the Betula spp.
This distribution
implies that
the way Populus tremula
regenerates differs from the other species. Populus
tremula
is the only one of the more common tree
species
with vegetative spreading by root suckers, and it has
very short-lived seeds. We believe that the number of
suckers and their condition of growth was determined
by the mother trees and their locations. The high
number
of Salix caprea probably
also depends on local distri-
bution of mature
individuals.
Betula spp. mainly colonized mounds or pits, and
formed a scattered
pattern
of dense clusters. Picea abies
saplings and young trees grew mainly along old logs.
Pinus
sylvestris saplings
were
randomly
scattered. These
regeneration patterns
were
formed
by both reproductive
strategies
and the properties
of the micro-environment.
The
different
height
distributions between small
gaps
and the single big gap suggested that the most active
regeneration
in the big gap occurred in the first year
after gap creation and that this period, after 7 yr, had
already passed.
The initial conditions of a disturbed site
are important
for colonization and invasion, as Egler
(1954) suggested.
The regeneration
of the boreal forest differs from
temperate, subtropical
and
tropical
forests. The canopy
in the natural
boreal forest is fairly open (Leemans
1986), so the effects of gap formation are less clear.
Secondly, the forest floor is usually covered by a
thick moss layer, which is a disadvantage
for seedling
establishment. Most of the saplings grew on decom-
400
- Gap structure,
disturbance and regeneration
in a primeval
forest -
posed logs, root plates, boulders or in mounds, which
suggests that the properties
of the substrate where the
seedlings
establish, and
competition
from the moss and
field layers
are
very important
(cf. Hyttebom
& Packham
1987).
There
are only few important
tree species, but all of
them
germinated
and established in connection with the
creation of the big gap. There was no sign of succession
in the classical sense of tree replacement
(Connell &
Slatyer
1977). This seems to be typical
for boreal forests
disturbed
by fires (Heinselman
1981) and storms.
Acknowledgements. This study was supported by the World
Wild Life Fund for Nature
(WWF).
We thank
Mark Fulton for
writing the program for calculating semivariance and for
valuable discussions. Thanks are also due to Theo Verwijst
and anonymous referees for comments on the manuscript,
Willy Jungskar
for his help with the computing, and Yang
Xiaohua, Department
of Meteorology, Uppsala University,
who provided
us with the wind data. The first author received
a scholarship
from the Swedish Institute.
References
Berg, S. & Lundqvist,
P.-U. 1964. Vegetationsundersikning
och synpunkter
pa Fibyskogens utvecklingshistoria
och
dess vdrde som naturreservat.
Report, Department of
Ecological Botany, Uppsala.
Bradshaw,
R. H. W. & Zackrisson,
0. 1990. A two thousand
year history
of a northern
Swedish boreal forest stand. J.
Veg. Sci. 1: 519-528.
Brewer, R. & Merritt, P. G. 1978. Wind throw and tree
replacement
in a climax
beech-maple
forest. Oikos 30: 149-
152.
Brokaw, N. V. L. 1982. Treefalls: frequency, timing, and
consequences.
In:
Leigh
Jr.,
E. G., Rand,
A. S. & Windsor,
D. M. (eds.) The ecology of a tropical forest: seasonal
rhythms
and long-term changes,
pp. 101-108. Smithsonian
Institution
Press, Washington
DC.
Brokaw, N. V. L. 1985. Treefalls, regrowth, and commu-
nity structure in tropical forests. In: Pickett, S. T. A. &
White,
P. S. (eds.) The
ecology of natural disturbance and
patch dynamics,
pp. 53-69. Academic Press, Orlando.
Burrough,
P. A 1987. Spatial aspects of ecological data. In:
Jongman,
R. H. G., ter
Braak,
C. J. F. & van Tongeren,
0.
F. R. (eds.) Data analysis in community
and landscape
ecology, pp. 230-237. Pudoc, Wageningen.
Connell, J. H & Slatyer,
R. 0. 1977. Mechanisms of succes-
sion in natural
communities and their role in community
stability
and organization.
Am.
Nat. 111: 1119-1144.
Cooper, W. S. 1913. The climax forest of Isle Royale, Lake
Superior,
and its development.
Bot. Gaz. 15: 1-44, 115-
140, 189-235.
Denslow, J. S. 1987. Tropical
rainforest
gaps and tree species
diversity.
Annu. Rev. Ecol. Syst. 18: 431-451.
Egler,
F. E. 1954. Vegetation
science concepts.
I. Initial floris-
tic composition, a factor in old-field vegetation develop-
ment. Vegetatio 4: 412-417.
Engelmark, 0. 1984. Forest fire in the Muddus National Park
(northern Sweden) during the past 600 yr. Can. J. Bot. 62:
893-898.
Foster, J. R. & Reiners, W. A. 1986. Size distribution and
expansion of canopy gaps in a northern Appalachian spruce-
fir forest. Vegetatio 68: 109-114.
Greig-Smith, P. 1983. Quantitative plant ecology. University
of California Press, Berkeley.
Hartshorn, G. S. 1978. Tree falls and tropical forest dynamics.
In: Tomlinson, P. B. & Zimmerman, M. H. (eds.) Tropical
trees as living systems, pp. 617-638. Cambridge Univer-
sity Press, New York.
Hartshorn, G. S. 1980. Neotropical forest dynamics. Biotropica
12 (Suppl.): 16-23.
Hartshorn, G. S. 1989. An overview of neotropical forest
dynamics. In: Gentry, A. H. (ed.) Four neotropicalforests.
Yale University Press, New Haven CT.
Heinselman, M. L. 1973. Fire in the virgin forests of the
Boundary Waters Canoe Area, Minnesota. Quat. Res. 3:
329-382.
Heinselman, M. L. 1981. Fire and succession in the conifer
forests of northern North America. In: West, D. C., Shugart,
H. H. & Botkin, D. B. (eds.) Forest succession, concepts
and application,
pp.
374-405. Springer-Verlag,
New York.
Hesselman,
H. 1935.
Fibyskogen
och dess utvecklingshistoria.
(Der Fibywald und seine Entwicklungsgeschichte.) Medd.
Stat. Skogsforsoksanst.
28: 525-570.
Hytteborn,
H. & Packham,
J. R. 1985. Left to nature: Forest
structure and regeneration in Fiby urskog, central Sweden.
Arboric.J. 9: 1-11.
Hytteborn,
H. & Packham,
J. R. 1987. Decay rate of Picea
abies logs and the storm gap theory: a re-examination of
Sernander plot III, Fiby urskog, central Sweden. Arboric.
J. 11:299-311.
Hytteborn, H., Liu Qinghong & Verwijst, T. in press. Natural
disturbance and gap dynamics in a Swedish boreal spruce
forest. In: Nakagoshi, N. & Golley, F. B. (eds.) The ecol-
ogy of pine and spruce forests. Vegetation processes,
population dynamics and productivity. SPB Academic
Publishing, The Hague.
Hytteborn, H., Packham, J. R. & Verwijst, T. 1987. Tree
population dynamics, stand structure and species compo-
sition in the montane virgin forest of Vallibacken, north-
ern Sweden. Vegetatio 72: 3-19.
Kanzaki, M. 1984. Regeneration in subalpine coniferous for-
ests. I. Mosaic structure and regeneration process in a
Tsuga diversifolia forest. Bot. Mag. Tokyo 97: 297-311.
Kapos, V., Pallant, E., Bien, A. & Freskos, S. 1990. Gap
frequencies in lowland rain forest sites on contrasting soils
in Amazonian Ecuador. Biotropica 22: 218-225.
Leemans, R. 1986. Structure of the primeval forest of Fiby. In:
Fanta, J. (ed.) Forest dynamics research in Western and
Central Europe, pp. 221-230. Pudoc, Wageningen.
Lid, J. 1985. Norsk, svensk,finsk flora. Det Norske Samlaget,
Oslo.
Mayer, H. 1989. Windthrow. Phil. Trans. R. Soc. Lond. B
324: 267-281.
Mayer, H., Schenker, S. & Zukrigl, K. 1972. Der Urwaldrest
401
Liu Qinghong & Hyttebom, H.
Neuwald beim Lahnsattel. Centralbl. Ges. Forstwesen
89(3): 147-190.
Naka, K. 1982. Community dynamics
of evergreen
broadleaf
forests in southwestern
Japan.
I. Wind damaged
trees and
canopy gaps in an evergreen
oak forest.
Bot. Mag. Tokyo
95: 385-399.
Naka, K. & Yoneda, T. 1984. Community dynamics
of ever-
green
broadleaf
forests in southwestern
Japan.
III.
Reveg-
etation
in gaps in an
evergreen
oak forest. Bot.
Mag. Tokyo
97: 275-286.
Nakashizuka,
T. 1984. Regeneration process of climax beech
(Fagus crenata
Blume) forests IV. Gap
formation.
Jap. J.
Ecol. 34: 75-85.
Nakashizuka,
T. & Numata,
M. 1982. Regeneration process
of
climax beech forests I. Structure of a beech forest with the
undergrowth
of Sasa. Jap. J. Ecol. 32: 57-67.
Putz, F. E., Coley, P. D., Lu, K., Montalvo, A. & Aiello, A.
1983. Uprooting
and
snapping
of trees: structural determi-
nants and ecological consequences.
Can. J. For. Res. 13:
1011-1020.
Runkle, J. R. 1981. Gap regeneration
in some old-growth
forests of the eastern United States.
Ecology
62:1041-1051.
Runkle,
J. R. 1982. Patterns of disturbance in some old-growth
mesic forests
of eastern north America.
Ecology 63: 1533-
1546.
Runkle, J. R. 1984. Development of woody vegetation in
treefall
gaps in a beech-sugar maple
forest. Holarct. Ecol.
7: 157-164.
Runkle,
J. R. 1985. Disturbance
regimes in temperate
forests:
In: Pickett, S. T. A. & White, P. S. (eds.) The ecology of
natural disturbance and
patch dynamics,
pp. 17-33. Aca-
demic Press, Orlando.
Runkle, J. R. & Yetter, T. C. 1987. Treefalls revisited: gap
dynamics
in the southern
Appalachians. Ecology 68: 417-
424.
SAS 1985. Statistical
analysis system.
5th ed. SAS Institute,
Cary,
NC.
Schaetzl, R. J., Johnson, D. L., Bums, S. F. & Small, T. W.
1989.
Tree
uprooting:
review of terminology,
process,
and
environmental
implications.
Can. J. For. Res. 19: 1-11.
Semander,
R. 1918. Fiby urskog.
Sveriges
Natur 9: 29-37.
Semander,
R. 1936. Granskar
och Fiby urskog. (Summary:
The primitive
forests of Granskiir and
Fiby: a study
of the
part played by storm-gaps
and dwarf
trees
in the regenera-
tion of the Swedish spruce
forest).
Acta
Phytogeogr.
Suec.
8.
Sjors,
H. 1963. Amphi-atlantic
zonation,
nemoral to arctic. In:
Love, A. & Love, D. (eds.) North atlantic biota and their
history,
pp. 109-125. Pergamon
Press, Oxford.
Stewart,
G. H. 1986. Forest dynamics and disturbance
in a
beech/hardwood
forest,
Fiordland,
New Zealand.
Vegetatio
68: 115-126.
Swaine,
M. D., Lieberman,
D. & Putz,
F. E. 1987. The dynam-
ics of tree
populations
in tropical
forest: a review. J. Trop.
Ecol. 3: 359-366.
Veblen, T. T. & Stewart,
G. H. 1982. On the regeneration gap
in New Zealand: the dynamics of Libocedrus bidwillii
stands on South Island. J. Ecol. 70: 413-436.
White,
P. S. 1979. Pattern,
process, and natural disturbance in
vegetation.
Bot. Rev. 45: 229-299.
White,
P. S. & Pickett,
S. T. A. 1985. Natural
disturbance
and
patch dynamics:
an introduction.
In: Pickett, S. T. A. &
White,
P. S. (eds.) The
ecology of natural
disturbance
and
patch dynamics,
pp. 3-13. Academic Press, Orlando.
White,
P. S., MacKenzie,
M. D. & Busing, R. T. 1985.
Natural
disturbance
and gap phase dynamics
in southern
Appala-
chian spruce-fir
forests. Can. J. For. Res. 15: 233-240.
Yamamoto,
S. 1989. Gap dynamics
in climax Fagus crenata
forests. Bot. Mag. Tokyo
102: 93-114.
Zackrisson, 0. 1977. Influence of forest fire on the North
Swedish boreal
forest. Oikos
29: 22-32.
Received 14 December 1990;
Revision received 25 April 1991;
Accepted 10 May 1991.
402
