ArticlePDF Available

Abstract and Figures

How do canopy-understory interactions respond to variation in disturbance severity over extended periods of time? For forests with different disturbance histories, do light availability and understory-cohort densities converge towards a common old-growth structure, or do historical legacies influence populations indefinitely?.
Content may be subject to copyright.
Journal of Vegetation Science 28 (2017) 1128–1139
Long-term responses of canopyunderstorey
interactions to disturbance severity in primary Picea
abies forests
Radek Ba
ce , Jonathan S. Schurman, Marek Brabec, Vojt
Cada, Tiphaine Despr
Pavel Janda, Jana L
a, Martin Mikol
s, Robert C. Morrissey ,HanaMrhalov
Thomas A. Nagel, Mark
eta H. Nov
a, Meelis Seedre, Michal Synek, Volodymyr Trotsiuk &
Miroslav Svoboda
Bark beetle; Canopy openness; Disturbance
regime; Mountain forest; Natural regeneration;
Picea abies; Primary forest; Saplings and poles;
Understorey light availability; Wind storms
at et al. (2002)
Received 23 February 2017
Accepted 31 August 2017
Co-ordinating Editor: Beverly Collins
ce, R. (corresponding author,
Schurman, J.S. (
Brabec, M. (
Cada, V. (
es, T. (
Janda, P. (jandap@fld.czu .cz)
a, J. (
s, M. (
Morrissey,R.C. (
a, H. (
Nagel, T.A. (
a, M.H.
Seedre, M. (
Synek, M. (
Trotsiuk, V. (
Svoboda, M. (s
Department of Forest Ecology, Faculty of
Forestry and Wood science, Czech University
of Life Sciences, Kam
a 1176, CZ-16521
Praha, Czech Republic;
Department of Nonlinear Modeling, Institute
of Computer Science, Academy of Sciences of
the Czech Republic, Pod Vod
arenskou v
CZ-18207 Prague 8, Czech Republic;
Questions: How do canopyunderstorey interactions respond to variation in
disturbance severity over extended periods of time? For forests with different
disturbance histories, do light availability and understorey cohort densities con-
verge towards a common old-growth structure, or do historical legacies
influence populations indefinitely?
Locations: Remnants of primary spruce (Picea abies (L.) Karst.) forests through-
out Germany, Slovakia, Ukraine and Romania.
Methods: A disturbance history of >200 yr was reconstructed from 11 278 tree
cores collected from forest plots (n=520). Understorey tree densities of two size
classes and hemispherical photo-based light availabilities were inventoried and
modelled as functions of the severity of the main disturbance and time since the
Results: Variation in understorey tree densities had a hump-shaped distribution
through time. Stem densities were approximately static in the least disturbed
sites, and declined in relation to disturbance severity over approx. 100 yr. Simi-
lar to patterns of stem densities, initially high understorey light availability also
reached a minimum at 100 yr, which indicated crown closure. Following this,
light availability and stem densities both increased as stands transitioned from
stem exclusion to understorey re-initiation. The effect of disturbance severity on
understorey densities and patchiness in light availability persisted for >200 yr.
Conclusions: Long-term trends in canopyunderstorey interactions validate
current conceptual models of forest development. Furthermore, we empirically
validate that these conceptual models generalized over gradients in disturbance
severity. Higher disturbance sites exhibited a more even-aged character with
more pronounced periods of stem exclusion, canopy closure and understorey
re-initiation; forests with low-severity disturbance histories yielded a more sta-
tionary uneven-aged structure. The model identified the extent of variation in
disturbance severity within which these P.abies forests are able to regenerate
and retain their monospecific character, which is increasingly relevant as distur-
bance regimes continue to shift under global climate change.
Department of Forestry and Renewable
Forest Resources, University of Ljubljana,
Vecna Pot 83, SI-1000 Ljubljana, Slovenia;
Swedish University of Agricultural Sciences
SLU, Southern Swedish Forest Research
Centre, Box 49, 230 53 Alnarp, Sweden
These authors contributed equally.
Journal of Vegetation Science
1128 Doi:10.1111/jvs.12581 ©2017 International Association for Vegetation Science
Canopy cover regulates understorey light availability and,
consequently, the performance of the regenerating sap-
lings that will eventually dominate forest canopies (e.g.
Canham et al. 1990; Caspersen & Kobe 2001; Ba
ce et al.
2015). Disturbances, like wind storms or insect outbreaks,
reduce canopy cover, increase light transmission and can
initiate population cycles whereby understorey cohorts
increase in size to fill canopy gaps and are subsequently
repopulated by incoming recruitment (Oliver 1980; Frank-
lin et al. 2002). The majority of empirical research into
canopyunderstorey interactions deals with the initial
responses of understorey populations to recent distur-
bances (e.g. Lang & Knight 1983; Nagel et al. 2006). How-
ever, developmental dynamics unfold over centuries
(Kashian et al. 2005) and empirical investigations into the
full cycle of understorey response and re-initiation are
consequently rare.
Following a disturbance, forest structure transitions
through a sequence of successional phases defined by
developmental shifts in the constraints on recruitment,
growth and mortality (Oliver 1980; Franklin et al. 2002).
Canopy removal is accompanied by a pulse of seedling
recruitment and increased sapling growth rates. In the
stem exclusion phase, saplings compete to increase in size,
density-dependent thinning becomes an important cause
of mortality, and canopy closure occurs as light penetration
to the forest floor is minimized (Assmann 1970). In the
understorey re-initiation phase, occasional canopy tree
deaths permit increased light penetration and the gradual
repopulation of understorey cohorts (Cumming et al.
2000). The stand reaches the old-growth phase after long
periods of low disturbance, where an approximately sta-
tionary uneven-aged structure develops to reflect continu-
ous gap creation and recruitment (Peet & Christensen
1980; Coomes et al. 2003). Quantitative treatments of the
entire developmental cycle are rare, but several features
are necessary to parameterize key aspects of canopy
understorey interactions: the baseline densities of under-
storey cohorts at equilibrium, the duration of time
between disturbance and canopy closure, and the relation-
ship between post-canopy closure gap dynamics and
understorey re-initiation. Furthermore, the conceptual
model outlines the development of a single cohort in hori-
zontally homogeneous conditions following complete
canopy removal by a stand-replacing event; how long-
term responses of development generalized over gradients
in canopy removal, or disturbance severity, is poorly quan-
tified (Romme et al. 1998; Kashian et al. 2005).
Variation in disturbance severity influences the magni-
tude, patchiness and duration of increased light availability
(Liu & Hytteborn 1991; Nicotra et al. 1999) and,
presumably, long-term understorey population dynamics.
In simple monospecific cases, understorey populations will
fluctuate in correspondence with canopy removal as more
individuals are recruited into adult cohorts. Declining sap-
ling densities due to out-going recruitment will be offset by
incoming recruitment from recent germinants. Anticipat-
ing the outcomes of canopyunderstorey dynamics among
multiple competing species is likely to become more com-
plicated, because gap size variation influences interspecific
variation in growth and survival (Sipe & Bazzaz 1995).
However, regeneration strategies evolve under the regu-
larity of disturbance patterns and insulate populations
against stark compositional transitions (Gutschick &
BassiriRad 2003; Johnstone et al. 2016). For example,
shade-tolerant conifers maintain cohorts of advanced
regeneration that have a significant size advantage over
dispersal-dependent species upon canopy removal
(Messier et al. 1999). The capacity of regeneration strate-
gies to promote self-replacement is limited to a range of
post-disturbance conditions (i.e. gap size), but these thresh-
olds are unspecified for many forest types (Johnstone et al.
2016; Kulakowski et al. 2017).
Our objective was to produce an empirical model of
canopyunderstorey interactions in monospecific Picea
abies (L.) Karst temperate mountain forests that span the
range of developmental phases invoked in standard con-
ceptual models, and can also be generalized to variation in
post-disturbance conditions. Dendroecological methods
were used to reconstruct local disturbance histories in frag-
ments of primary forests distributed throughout Central
and Eastern Europe. Disturbance history reconstruction
allowed sites to be stratified by disturbance severity and
time since disturbance (Lorimer & Frelich 1989), and could
be analysed as a chronosequence. Understorey light condi-
tions, along with sapling and pole stem densities, were
additionally inventoried to address the following ques-
1. How do understorey cohort densities change with time
since disturbance? Do densities of understorey cohorts
with different disturbance histories converge towards an
old-growth phase (Peet & Christensen 1980; Kashian et al.
2005), or do sites with different histories diverge in terms
of population structure?
2. Is the increase in light availability larger in magnitude
over a longer period in more severely disturbed plots? How
do light conditions change after canopy closure to influ-
ence understorey re-initiation?
3. Is there an effect of disturbance severity on the prolifer-
ation of other species, implying a limited ability of spruce
to sustain monodominance? Sorbus aucuparia L. (rowan) is
a common light-demanding understorey tree in the
region, and is known to infiltrate large gaps and arrest the
Journal of Vegetation Science
Doi: 10.1111/jvs.12581©2017 International Association for Vegetation Science
R. Ba
ce et al. Long-term canopyunderstorey responses
development of P.abies stands (Hofgaard 1993; _
et al. 2013).
Study site
Inventory plots were placed in remnants of primary for-
est throughout the Romanian, Slovakian and Ukrai-
nian Carpathians and in the National Park Harz,
Germany. Prior efforts have used tree-ring methods to
empirically reconstruct the natural disturbance regime
of these inventoried locations (e.g. Svoboda et al. 2014;
Trotsiuk et al. 2014; Janda et al. 2017; Meigs et al.
2017). Study sites ranged in altitude from 1000 m (Ger-
many) to 1700 m a.s.l. (Romania; Fig. 1). The average
annual precipitation ranges from 800 to 1400 mm, and
the annual temperature ranges from 2 to 4 °C(UNEP
2007). More detailed information about the study area
is presented in Appendix S1. Satellite imagery, first-
hand knowledge of local people and on-site inspections
focused on signs of former human influence (e.g. cut
stumps), were used to delineate polygons of primary
forest fragments (15300 ha); they were primarily in
remote areas with minor human influence, such as
where steep slopes restrict access for logging or grazing.
More detailed information regarding the selection of
primary forest fragments is presented in Appendix S2.
In each polygon (henceforth stand), we placed six to
60, 1000 m
(or 500 m
in structurally homogenous
stands with high stem densities; n=90) circular plots
using a stratified random design (Svoboda et al. 2014).
In total, we established 520 plots in 31 stands. Norway
spruce dominated tree species composition in the study
area (99.6% of trees 10 cm DBH and 80.0% for regen-
eration 50 cm in height and <10 cm DBH), while
rowan was also relatively abundant in the understorey
layer (0.2% of trees 10 cm DBH and 17.9% for regen-
eration). Other rarer species included Pinus cembra L.,
Abies alba Mill., Acer pseudoplatanus L. and Betula spp.
(in total 0.2% for trees 10 cm DBH and 2.1% for
Data collection
Understorey tree densities were comprehensively invento-
ried from 2010 to 2014. All individuals 0.5 m in height
within forest plots were identified to species, and densities
were recorded for two height classes: saplings (height 0.5
1.3 m) and poles (height ˃1.3 m and DBH <10 cm). At
each plot, six hemispherical photographs (one in the plot
centre, and five distributed 12.1 m from the plot centre at
72°intervals around the plot) were taken 1.3 m above the
ground using a circular fisheye lens (Sigma 4.5 mm F2.8
EX DC). The amount and variability of understorey
light was analysed using WinSCANOPY software (Regent
Instruments, Quebec, CA) using the ‘openness’ variable.
The mean, maximum and coefficient of variation of open-
ness were calculated for each plot, and they were used as
response variables to detect mean, maximum and relative
variability of light distribution in the understorey. The rela-
tive variability of light distribution is hereafter referred to
as patchiness of light.
Fig. 1. Distribution of inventory plots at regional (a), mountain (b) and stand scales(c).
Journal of Vegetation Science
1130 Doi:10.1111/jvs.12581 ©2017 International Association for Vegetation Science
Long-term canopyunderstorey responses R. Ba
ce et al.
Dendrochronology and disturbance reconstruction
We randomly selected 15 or 25 (for 500-m
or 1000-m
plots, respectively) non-suppressed, live trees (i.e. trees
with a significant portion of their crown projection receiv-
ing direct sunlight from above; Lorimer & Frelich 1989)
per plot for the radial growth analysis and age determina-
tion. In rare cases, fewer trees were cored because less than
1525 individuals were present. Suppressed trees were
avoided because their growth patterns may lack informa-
tion important for disturbance history reconstruction
(Splechtna et al. 2005). All trees were cored at a height of
1 m above the ground. In total, 11 278 cores were used to
reconstruct plot histories.
Increment cores were dried, cut with a core microtome,
measured and cross-dated using standard dendrochrono-
logical methods (Speer 2010). Annual rings were mea-
sured to the nearest 0.1 mm using a stereomicroscope and
sliding-stage measuring device in conjunction
software (www.rinntech.ds). Cores
were initially cross-dated visually using the marker year
approach (Speer 2010), and were then verified with PAST4
software ( and, subsequently, COFECHA
Individual tree growth patterns were investigated to
identify when individual trees were released from canopy
suppression and recruited into the adult cohort. Two
modes of canopy recruitment were considered: open-
canopy recruitment and growth release. Open-canopy
recruitment was indicated in trees with high initial growth
rates. Mean growth rates of individuals from the time they
were 4 cm DBH and over the following 5 yr were com-
pared to a growth rate threshold. This threshold was deter-
mined by a logistic regression comparing tree ring-derived
sapling growth rates, mean over 5 yr following 4 cm DBH,
under closed canopies and among gaps of 500 m
=46, N
not suppressed
=41). Individuals with a
mean growth rate exceeding 1.7 mmyr
were catego-
rized as open-canopy recruits (Svoboda et al. 2014).
To identify the release of individuals from canopy sup-
pression, a set of growth rate changes was compiled and
compared to a boundary line that conditions responses on
formed growth rates (Black & Abrams 2003). Ten-year
running means were estimated along ring-width series.
The differences between adjacent 10-yr means were then
calculated. The boundary line is an exponential function
that models the upper limit of variation of percentage
growth changes as a function of prior growth rates. This
method scales the release criteria to pre-disturbance
growth rates, which reduces the number of falsely detected
releases (Black & Abrams 2003). We used a boundary line
function calibrated for P.abies from trees throughout Cen-
tral European spruce forests by Splechtna et al. (2005).
Growth rate changes exceeding 20% of the boundary line
value were treated as gap releases. We also decided that
the detected growth pattern had to be sustained for at least
7 of the 10 yr used to calculate adjacent means (Fraver
et al. 2009). Gap releases were only searched for during
the interval when individuals were <25 cm DBH; trees
>25 cm DBH were assumed to be in the canopy stratum
(Lorimer & Frelich 1989).
Both releases types were converted to total canopy area
disturbed in each decade (Lorimer & Frelich 1989). Crown
radius was estimated for each cored tree and a regression
was fit between crown area and DBH (R
P<0.001, crown area =0.4631 9DBH +0.8948; Svo-
boda et al. 2014). Crown areas of released trees were
summed for the decade of relese, and this quantity was
divided by current plot-level crown areas to produce an
estimate of proportion of crown area released in each dec-
ade (Lorimer & Frelich 1989). For a thorough description
of disturbance history reconstructions, see Svoboda et al.
To account for the protracted nature of some distur-
bances and release tree responses, disturbance severity was
further integrated over a running window of three dec-
ades, and local maxima in this smoothed disturbance pro-
file were used as point estimates of disturbance severity;
the largest local maximum was treated as a plot main
disturbance (Svoboda et al. 2014). Years since the main
disturbance were calculated as the year of data collection
(20102014) minus the year of maximum severity. For
recently disturbed plots, where the current canopy area
disturbed was larger than dendrochronologically detected
maximum disturbance severities, the severity was expressed
by current canopy openness. Current canopy openness
was calculated as the difference between mean canopy clo-
sure of the whole data set and current canopy closure of a
given plot. Distributions of time since disturbance and dis-
turbance severity across countries and stands are shown in
Appendices S3 and S4.
Statistical analyses
Temporal patterns in stem density variation were investi-
gated by modelling the coefficients of variation in log-
transformed stem densities among time-since-disturbance
classes (20 yr) using spline regressions. Independent spa-
tial regression models were fitted to assess how spruce and
rowan densities in the two different height classes, mean,
maximum and variation of understorey light availability
respond to both disturbance severity and the number of
years since the main disturbance. We used a GAMM to fit
a linear combination of smoothed functions of several pre-
dictor variables (plus one smooth interaction of time and
severity), while additionally considering random spatial
Journal of Vegetation Science
Doi: 10.1111/jvs.12581©2017 International Association for Vegetation Science
R. Ba
ce et al. Long-term canopyunderstorey responses
effects, which allowed us to account for systematic vari-
ability in space (Wood 2006). Altogether, four full models
for regeneration densities (two species and two height
categories) and three models for mean, maximum and
variation of understorey light availability were used. For
the i-th stand and j-th plot, we used the following model:
Yij NBðlij:Pij ;hÞ;for stem density
Yij Nðlij:Pij ;r2
uÞ;for understorey light characteristics
logðlijÞ¼b0þbiþslocðxij ;yijÞþstimeðtimeij Þ
Expected values of either log-transformed light or sam-
pling quantities (log(u
)) were considered a function of
several factors. A set of spatially variable functions
(smoothed, penalized splines), including s
,reflects the
smooth spatial trend; s
reflects the (smooth) marginal
effect of time since disturbance; s
reflects the (smooth)
marginal effect of disturbance severity; and s
reflects the (smooth) interaction of time and severity (ob-
tained as a tensor product spline). The vector b
is a set of
independent random effects for individual stands and b
the model intercept; r2
bis the variance among random
effects. Observed densities were modelled as a negative
binomial (NB), with a rate of u
and probability of success h.
Observed light values were modelled as a normal distribu-
tion with mean u
and variance r2
u. All parameters were
estimated based on the optimization of the penalized likeli-
hood function. P
is an offset applied to correct for variation
in plot sizes. Means and standard deviations of variables
used in the analyses are displayed in Appendix S5.
Variation in regeneration densities
Variation in regeneration densities among plots increased
during the first 100 yr after disturbance, and decreased
within the next 100 yr for both height classes of spruce
(Fig. 2). This unimodal relationship was confirmed by
spline regression (P=0.001, deviance explained =86.4%
for saplings; P=0.002, deviance explained =86.9% for
Understorey light
The model for mean understorey light availability explained
54.7% of the deviance, with a significant spatial term
(P<0.001), random stand effect (P=0.026) and time since
disturbance effect (P<0.001; Table 1). A spatial term
(P<0.001) and years since disturbance effect (P<0.001)
explained 52.9% of deviance in maximum light availability.
The model for patchiness in light explained 37.4% of the
deviance with a spatial term (P=0.024), stand effect
(P<0.001), years since disturbance (P<0.001), distur-
bance severity (P=0.001) and the years 9disturbance
interaction (P=0.037). The amount of understorey light
decreased since the main disturbance, and remained at a
Fig. 2. Coefficients of variation in log-transformed spruce densities in different time-since-disturbance (20 yr) classes for two height classes of
regeneration. The black line and circles represent height class 50130 cm (saplings) and grey line with triangles represents height class 130700 cm
(poles). The points represent exact values and curves represent the spline regressions.
Journal of Vegetation Science
1132 Doi:10.1111/jvs.12581 ©2017 International Association for Vegetation Science
Long-term canopyunderstorey responses R. Ba
ce et al.
low level until ~150 yr since disturbance, but then began to
rise gradually (Fig. 3; Appendix S6). The patchiness of light
became lower through time following moderate to high
severity disturbance events. The patchiness of light was con-
stantly higher on plots under a low-severity disturbance
regime; it exhibited a U-shaped pattern, declining and then
increasing later during stand development (Fig. 3).
Spruce regeneration densities
Densities of small spruce saplings (50130 cm in height)
were significantly influenced by time since disturbance
(P=0.001), disturbance severity (P=0.016), a spatial
term (P<0.001) and a random stand effect (P=0.001),
which explained 48.0% of the deviance (Table 1).
Spruce poles (130700 cm in height) were influenced by
time since disturbance (P<0.001), disturbance severity
(P<0.001), time x disturbance interaction (P=0.042), a
spatial term (P<0.001) and a random stand effect
(P=0.002), which explained 43.0% of the deviance.
Both sapling and pole densities were most stable after
low-severity disturbances (Fig. 4). Following distur-
bance, sapling densities declined to a minimum at ca
100 yr of development. Pole densities follow a similar
Table 1. Results of GAMMs for influence of time since main disturbance, severity of the main disturbance and their interaction on tree densities and under-
storey light availability characteristics. The response variables for understorey light availability followed a Gaussian distribution, whereas a negative
binomial distribution was appropriate for tree densities. The results displayed are: explanatory variables included in the model, their significance (based on
Chi-squared or F-tests, values of P<0.05 are in bold). Also given are the overall percentages of deviance explained (dev) for the model. Estimated degrees
of freedom (edf) indicate linearity: edf =1 represents a linear relationship, edf 2 representa significant non-linear relationship.
Response Variable Explanatory Variable edf FP dev
Mean Understorey Light Availability Random spatial effect 27.3 5.5 0.000 54.7%
Random effect of stand 4.7 0.3 0.026
Smooth main effect of Years since maindisturbance 3.7 10.0 0.000
Smooth main effect of Disturbance severity 1.0 1.7 0.191
Interaction between main effects 2.3 1.7 0.163
Maximum Understorey Light Availability Random spatial effect 26.7 5.1 0.000 52.9%
Random effect of stand 4.6 0.2 0.065
Smooth main effect of Years since maindisturbance 3.6 10.5 0.000
Smooth main effect of Disturbance severity 1.0 0.0 0.960
Interaction between main effects 1.9 0.8 0.386
Patchiness of Understorey Light (CV) Random spatial effect 2.0 3.7 0.024 37.4%
Random effect of stand 20.1 5.3 0.000
Smooth main effect of Years since maindisturbance 1.2 21.0 0.000
Smooth main effect of Disturbance severity 1.0 10.9 0.001
Interaction between main effects 2.1 3.0 0.037
Response Variable Explanatory Variable edf v
Spruce 50130 cm (Saplings) Random spatial effect 2.0 34.8 0.000 48.0%
Random effect of stand 23.9 148.1 0.000
Smooth main effect of Years since maindisturbance 2.9 15.9 0.001
Smooth main effect of Disturbance severity 1.0 5.9 0.016
Interaction between main effects 3.1 4.4 0.373
Spruce 130700 cm (Poles) Random spatial effect 13.8 99.2 0.000 43.0%
Random effect of stand 8.7 16.0 0.002
Smooth main effect of Years since maindisturbance 3.6 40.7 0.000
Smooth main effect of Disturbance severity 2.4 38.5 0.000
Interaction between main effects 2.7 8.4 0.042
Rowan 50130 cm (Saplings) Random spatial effect 6.4 57.7 0.000 63.7%
Random effect of stand 17.7 73.4 0.000
Smooth main effect of Years since maindisturbance 3.6 29.1 0.000
Smooth main effect of Disturbance severity 1.0 0.2 0.618
Interaction between main effects 1.2 1.2 0.394
Rowan 130700 cm (Poles) Random spatial effect 6.6 57.5 0.000 59.6%
Random effect of stand 16.8 57.7 0.000
Smooth main effect of Years since maindisturbance 3.0 18.3 0.001
Smooth main effect of Disturbance severity 1.2 0.2 0.706
Interaction between main effects 4.2 7.8 0.212
Journal of Vegetation Science
Doi: 10.1111/jvs.12581©2017 International Association for Vegetation Science
R. Ba
ce et al. Long-term canopyunderstorey responses
profile, reaching a minimum at ca 100 yr of develop-
ment, but exhibit a larger fluctuation in stem densities in
response to disturbance severity (Fig. 4; Appendix S7).
Rowan regeneration densities
Rowan sapling densities were significantly influenced by
time since disturbance (P<0.001), a spatial term
(P<0.001) and a stand effect (P<0.001), which
explained 63.7% of the deviance (Table 1). Rowan poles
were significantly influenced by years since disturbance
(P=0.001), a spatial term (P<0.001) and a random stand
effect (P<0.001), which explained 59.6% of the deviance.
Rowan densities were highest shortly after disturbance,
and rapidly decreased thereafter (Fig. 4; Appendix S7).
After reaching a minimum (~100 yr), densities gradually
Investigations of canopyunderstorey interactions tend to
focus on initial responses of understorey populations fol-
lowing a stand-replacing event (e.g. Cooper-Ellis et al.
1999; Romme et al. 2016) or compare dynamics among
various gap sizes in mature or old-growth forests (e.g.
Canham & Marks 1985; Nagel et al. 2006; Fahey &
Puettmann 2008). Our study with a long-term perspective
fills an important knowledge gap by simultaneously
addressing initial responses of understorey cohorts and
their re-establishment many decades later. Understorey
densities on low-disturbance sites were more stationary
through time, providing a baseline to contrast against
high-severity sites. Understorey dynamics in plots recover-
ing from more severe disturbances exhibit a more even-
aged character, where the boundaries between develop-
mental phases were more punctuated. We identified criti-
cal points in the co-development of light availability and
sapling densities, and used these points to integrate our
generalized model with stand development concepts (Oli-
ver 1980; Franklin et al. 2002). A summary of our expec-
tations and observations is further provided (Fig. 5).
Canopy closure and stem exclusion
Canopy closure is a critical event in the process of stand
development (Assmann 1970; Jennings et al. 1999). We
observed a minimization of mean understorey light avail-
ability occurring in correspondence with a maximum in
the variation of sapling densities among plots following
100 yr of development (Fig. 3a,b). Sapling and pole densi-
ties diverged during the first 100 yr following disturbance
(Fig. 2) because stem densities declined in proportion to
disturbance severity (Fig. 4). Conceptually, this corre-
sponds to the transition between stand initiation and stem
exclusion, which becomes more pronounced as
disturbance severity increases.
A reduction in sapling densities implies that the incom-
ing recruitment of smaller individuals into the size
Fig. 3. The influence of time since main disturbance, severity of the main disturbance event (percentage of disturbed canopy) and their interaction on
mean understorey light availability (a), maximum understorey light availability (b) and patchiness of understorey light (coefficient of variation) (c)at1.3m
above the ground.
Journal of Vegetation Science
1134 Doi:10.1111/jvs.12581 ©2017 International Association for Vegetation Science
Long-term canopyunderstorey responses R. Ba
ce et al.
categories we analysed is outpaced by the out-going
recruitment of saplings into larger size classes and mortal-
ity. Sapling densities decline in proportion to disturbance
severity, implying that a larger proportion of understorey
cohorts respond to fill larger canopy openings. As saplings
increase in size, they fill the new gap and understorey light
availability declines. Although spruce seedlings can sur-
vive with no height growth for many years (Messier et al.
1999; Hytteborn & Verwijst 2011), recruitment to the
upper strata is strongly light limited and, as established
individuals replace canopy cover, they inhibit incoming
recruitment (Messier et al. 1999).
Post-disturbance light levels (intercepts on Fig. 3) were
estimated to be highest in the most disturbed sites, but this
signal was statistically not significant. This is counter-intui-
tive, but Canham et al. (1990) concluded that the role of
gap size in determining understorey light availability was
less important than the height of the bordering canopy and
latitude (Canham et al. 1990). More observations from
recently disturbed sites would improve confidence in our
estimates of variation in initial light conditions immedi-
ately following disturbances.
Understorey re-initiation
After canopy closure, variation in understorey densities
decreased due to a corresponding increase in both light
availability and sapling densities. As we observed, con-
ceptual models of development anticipate correlated
increases in understorey light and sapling densities fol-
lowing canopy closure. Mortality of canopy trees
increases understorey light availability, which facilitates
the incoming recruitment of seedlings into sapling size
classes (Lutz & Halpern 2006). Even following >200 yr
of development, sapling densities still vary as a function
of past disturbance severity, implying that disturbance
has long-term consequences for understorey dynamics
(Kashian et al. 2005).
Picea abies 50−130 cm Picea abies 130−700 cm
Sorbus aucuparia 50−130 cm Sorbus aucuparia 130−700 cm
10 000
10 000
0 50 100 150 200 0 50 100 150 200
Years since main disturbance
Stem density (ha–1)
Severity of
the main
(% of canop
(Spruce saplings) (Spruce poles)
(Rowan saplings) (Rowan poles)
Fig. 4. The influence of time since main disturbance, severity of the main disturbance event (percentage of disturbed canopy) and their interaction on
density of spruceand rowan saplings (50130 cm) and poles (130700 cm).
Journal of Vegetation Science
Doi: 10.1111/jvs.12581©2017 International Association for Vegetation Science
R. Ba
ce et al. Long-term canopyunderstorey responses
Disturbance severity also has a lasting effect on within-
site variation in light availability (Fig. 3c). Light variability
at the low end of the severity spectrum followed an intu-
itive trajectory: variability was high following partial
canopy removal, but declined as these small gaps are filled
by regeneration, and again increased later in stand devel-
opment as individual canopy tree deaths contributed to
patchiness. Contrary to our expectation that light levels
would approach an equilibrium, independent of distur-
bance history, light variability on heavily disturbed plots
declined continuously as mean light availability increased
(Fig. 5c). Increasing evenness suggests density-dependent
processes, like self-thinning (Kashian et al. 2005; Lutz &
Halpern 2006). Persistent divergence in light variability
over multiple centuries suggests that stand dynamics do
not converge to an equilibrium in less than 250 yr. Canopy
patchiness directly impacts the spatial arrangement of sap-
lings in spruce populations, and likely contributes to the
vertical and horizontal complexity later in development
ce et al. 2015).
Regeneration strategies and resilience
Regeneration strategies of late successional species are pre-
sumed to buffer ecosystems against large fluctuations in
community and ecosystem structure (Messier et al. 1999).
Sustaining cohorts of advanced regeneration in anticipa-
tion of future canopy gaps is common among shade-toler-
ant conifers (Messier et al. 1999), and the persistent
monospecific character of our P.abies study sites exempli-
fies this trait. The specific character of disturbance agents
plays a large role in compositional stability. Central Euro-
pean temperate spruce forests are driven by wind storms
and bark beetle outbreaks, which preferentially remove
the largest individuals and leave cohorts of advanced
regeneration intact (Svoboda et al. 2010). Central Euro-
pean spruce forests can be compared with boreal spruce
forests, where fire destroys advanced regeneration, and
loss of this competitive advantage permits the infiltration
of birch and rowan populations into early successional
communities (Hofgaard 1993).
Our approach does not permit us to state with certainty
that advanced regeneration is predominantly responsible
for filling recently formed canopy gaps; that would require
directly observed plots and tracking individuals before and
after disturbance. However, indirect evidence exists to sup-
port this claim. Understorey population densities at the
incidence of disturbance (intercepts in Fig. 4) are similar
among low-severity sites, but the least stable densities are
on high-severity sites. Saplings thus appear to occur at a
high baseline density across a wide range of canopy condi-
tions, implying that there should be sufficient advanced
regeneration available to respond to canopy removal. The
contribution of post-disturbance recruitment to canopy
closure is likely small, given the size advantage of saplings
occurring at high baseline densities (Macek et al. 2016).
Densities of saplings <50-cm tall rapidly increase after dis-
turbance, but growth remains low, giving a substantial size
advantage to advanced regeneration (Zeppenfeld et al.
2015; Macek et al. 2016). These post-disturbance cohorts
Fig. 5. Graphic summary of hypotheses and significant observed trends.
The results show that density-dependent processes do not dampen
initially high variation (premise i), but rather that disturbance induced
heterogeneity in regeneration conditions, which seeds further divergence
(premise ii) (a). Results did not confirm that severe disturbances cause a
larger, more prolonged increase in light availability (b). Light was more
uniformly distributed within plots after severe disturbance, but did not
converge to patchiness of light found under a low-severity disturbance
regime, as was expected (c). The results did not show any pronounced
effect of disturbance severity on rowan abundance, as was predicted (d).
*Low-severity disturbance removes 20% of the plot-level canopy, while
high severity disturbance removes 100% of the plot canopy. Mean
understorey light availability was measured using a fish eye lens, and
patchiness of light was expressed as the coefficient of variation of light
availability within a plot.
Journal of Vegetation Science
1136 Doi:10.1111/jvs.12581 ©2017 International Association for Vegetation Science
Long-term canopyunderstorey responses R. Ba
ce et al.
will contribute to the increase in sapling densities during
understorey re-initiation.
We also considered whether advanced regeneration was
sufficient to ensure self-replacement of spruce following
disturbances of all severities. A change in composition
from spruce to rowan, a light-demanding understorey spe-
cies, following the highest severity events was considered
one potential outcome (Fig. 5d; e.g. Pajt
ık et al. 2015). We
did not detect a significant effect of disturbance severity on
rowan abundance (Fig. 5d), although rowan was encoun-
tered at high densities in some stands. The poor dispersal
ability of rowan, paired with the competitive advantage
provided by advanced regeneration of spruce, appears to
sustain the suppression of rowan under the current distur-
bance regime ( _
Zywiec et al. 2013).
A retrospective approach was used to construct a model
that specified the time scale of understorey cohort
responses to disturbance. Our dendroecological approach
fills a gap in empirical inquiry into stand dynamics by pro-
ducing a long-term, quantitative representation of stem
exclusion and understorey re-initiation, and largely vali-
dates accepted, but untested, development theory (Oliver
1980; Franklin et al. 2002). In this low-diversity system,
sapling densities at locations differing in disturbance his-
tory converge towards a common baseline, butdisturbance
effects persist for >200 yr.
Our study supports the premise that the persistence of
the monospecific character of temperate spruce forest,
despite variation in disturbance severity, is due to the
strategic advantages of advanced regeneration. Our find-
ings estimated boundaries to the limit of variation in dis-
turbance severity that P.abies can sustain to maintain
monospecific character. Inspection of our model predic-
tions suggests that understorey cohorts are lowest ca
100 yr after the most extreme events. It is likely that at this
point of minimization in stem densities, P.abies popula-
tions are most vulnerable to large shifts in composition.
This study was supported by project GACR no. 15-14840S,
Czech University of Life Sciences project CIGA no.
20164310, and institutional project “EXTEMITK”, No.
CZ.02.1.01/0.0/0.0/15_003/0000433 financed by OP RDE.
We would like to thank all co-workers for assistance with
field data collection and measurements. We thank the nat-
ure conservation authorities and forest owners for admin-
istrative support and access to the study sites. RB, JSS, PJ,
RCM, TAN,VT and MSv conceived the ideas and designed
methodology; RB, VC, PJ, JL, MM, HM, TAN, MSy,VT and
MSv collected the data; RB, JSS, MB, PJ and VT analysed
the data; RB, JSS, TD, TAN, MHN and MSe wrote the
manuscript. All authors contributed critically to the drafts
and gave final approval for publication.
Assmann, E. 1970. The principles of forest yield study.Pergamon
Press, New York, NY, US.
ce, R., Svoboda, M., Janda, P., Morrissey, R.C., Wild, J., Clear,
Cada, V. & Donato, D.C. 2015. Legacy of pre-distur-
bance spatial pattern determines early structural diversity
following severe disturbance in montane spruce forests.PLoS
One 10: e0139214.
Black, B.A. & Abrams, M.D. 2003. Use of boundary-line growth
patterns as a basis for dendroecological release criteria. Eco-
logical Applications 13: 17331749.
Canham, C.D. & Marks, P.L. 1985. The response of woody plants
to disturbance: patterns of establishment and growth. In:
Pickett, S.T.A. & White, P.S. (eds.) The ecology of natural distur-
bances and patch dynamics, pp. 197218. Academic Press, New
York, NY, US.
Canham, C.D., Denslow, J.S., Platt, W.J., Runkle, J.R., Spies,
T.A. & White, P.S. 1990. Light regimes beneath closed cano-
pies and tree-fall gaps in temperate and tropical forests. Cana-
dian Journal of Forest Research 20: 620631.
Caspersen, J.P. & Kobe, R.K. 2001. Interspecific variation in sap-
ling mortality in relation to growth and soil moisture. Oikos
92: 160168.
Coomes, D.A., Duncan, P.P., Allen, R.B. & Truscott, J. 2003. Dis-
turbances prevent stem-density di stributions in natural for-
ests from following scaling relationships. Ecology Letters 6:
Cooper-Ellis, S., Foster, D.R., Carlton, G. & Lezberg, A. 1999.
Forest response to catastrophic wind: results from an experi-
mental hurricane. Ecology 80: 26832696.
Cumming, S.G., Schmiegelow, F.K.A. & Burton, P.J. 2000. Gap
dynamics in boreal aspen stands: is the forest older than we
think? Ecological Applications 10: 744759.
Fahey, R.T. & Puettmann, K.J. 2008. Patterns in spatial extent of
gap influence on understory plant communities. Forest Ecol-
ogy and Management 255: 28012810.
Franklin, J.F., Spies, T.A., Van Pelt, R., Carey, A.B., Thornburgh,
D.A., Berg, D.R., Lindenmayer, D.B., Harmon, M.E., Keeton,
W.S. & Shaw, D.C. 2002. Disturbances and structural devel-
opment of natural forest ecosystems with silvicultural impli-
cations, using Douglas-fir forests as an example. Forest
Ecology and Management 155: 399423.
Fraver, S., White, A.S. & Seymour, R.S. 2009. Natural distur-
bance in an old-growth landscape of northern Maine, USA.
Journal of Ecology 97: 289298.
Gutschick, V.P. & BassiriRad,H. 2003. Extreme events as shaping
physiology, ecology, and evolution of plants: toward a uni-
fied definition and evaluation of their consequences. New
Phytologist 160: 2142.
Journal of Vegetation Science
Doi: 10.1111/jvs.12581©2017 International Association for Vegetation Science
R. Ba
ce et al. Long-term canopyunderstorey responses
Hofgaard, A. 1993. Structure and regeneration patterns in a vir-
gin Picea abies forest in northern Sweden. Journal of Vegetation
Science 4: 601608.
Hytteborn, H. & Verwijst, T. 2011. The importance of gaps and
dwarf trees in the regeneration of Swedish spruce forests: the
origin and content of Sernander’s (1936) gap dynamics the-
ory. Scandinavian Journal of Forest Research 26(S10): 316.
Janda, P., Trotsiuk, V., Mikol
s, M., Ba
ce, R., Nagel, T.A., Seidl,
R., Seedre, M., Morrissey, R.C., Kucbel, S. & Jaloviar, P.
2017. The historical disturbance regime of mountain Norway
spruce forests in the Western Carpathians and its influence
on current forest structure and composition. Forest Ecology
and Management 388: 6778.
Jennings, S., Brown, N. & Sheil, D. 1999. Assessing forest cano-
pies and understorey illumination: canopy closure, canopy
cover and other measures. Forestry 72: 5974.
Johnstone, J.F., Allen, C.D., Franklin, J.F., Frelich, L.E., Harvey,
B.J., Higuera, P.E., Mack, M.C., Meentemeyer, R.K., Metz,
M.R. & Perry, G.L. 2016. Changing disturbance regimes, eco-
logical memory, and forest resilience. Frontiers in Ecology and
the Environment 14: 369378.
Kashian, D.M., Turner, M.G., Romme, W.H. & Lorimer, C.G.
2005. Variability and convergence in stand structural devel-
opment on a fire-dominated subalpine landscape. Ecology 86:
Kulakowski, D., Seidl, R., Holeksa, J., Kuuluvainen, T., Nagel,
T.A., Panayotov, M., Svoboda, M., Thorn, S., Vacchiano, G.
& Whitlock, C. 2017. A walk on the wild side: disturbance
dynamics and the conservation and management of Euro-
pean mountain forest ecosystems. Forest Ecology and Manage-
ment 388: 120131.
Lang, G.E. & Knight, D.H. 1983. Tree growth, mortality, recruit-
ment, and canopy gap formation during a 10-year period in
a tropical moist forest. Ecology 64: 0751080.
Liu, Q. & Hytteborn, H. 1991. Gap structure, disturbance and
regeneration in a primeval Picea abies forest. Journal of Vegeta-
tion Science 2: 391402.
Lorimer,C.G. & Frelich, L.E. 1989. A methodologyfor estimating
canopy disturbance frequency and intensity in dense tem-
perate forests. Canadian Journal of Forest Research 19: 651
Lutz, J.A. & Halpern, C.B. 2006. Tree mortality during early for-
est development: a long-term study of rates, causes, and con-
sequences. Ecological Monographs 76: 257275.
Macek, M., Wild, J., Kopeck
y, M.,
Cervenka, J., Svoboda, M.,
a, J., Br
una, J., Mosandl, R. & Fischer, A. 2016.
Life and death of Picea abies after bark-beetle outbreak: eco-
logical processes driving seedling recruitment. Ecological
Applications 27: 156167.
Meigs, G., Morrissey, M., Ba
ce, R.,
Cada, V., Donato, D., Janda,
P., Seedre, M., Mikol
s, M., Nagel, T., (...) & Svoboda, M.
2017. More ways than one: mixed-severity disturbance
regimes foster structural complexity via multiple develop-
mental pathways. Forest Ecology and Management 406: 410
Messier, C., Doucet, R., Ruel, J., Claveau, Y., Kelly, C. & Lechow-
icz, M.J. 1999. Functional ecology of advance regeneration
in relation to light in boreal forests. Canadian Journal of Forest
Research 29: 812823.
Nagel, T.A., Svoboda, M. & Diaci, J. 2006. Regeneration patterns
after intermediate wind disturbance in an old-growth Fagus-
Abies forest in southeastern Slovenia. Forest Ecology and Man-
agement 226: 268278.
Nicotra, A.B., Chazdon, R.L. & Iriarte, S.V. 1999. Spatial hetero-
geneity of light and woody seedling regeneration in tropical
wet forests. Ecology 80: 19081926.
Oliver, C.D. 1980. Forest development in North America follow-
ing major distrubances. Forest Ecology and Management 3:
ık, J., Kon^
opka, B., Bo
sel’a, M.,
n, V. & Ka
stier, P. 2015.
Modelling forage potential for red deer: a case study in post-
disturbance young stands of Rowan. Annals of Forest Research
58: 91107.
Peet, R.K. & Christensen, N.L. 1980. Succession: a population
proces. Vegetatio 43: 131140.
Romme, W.H., Everham, E.H., Frelich, L.E., Moritz, M.A. &
Sparks, R.E. 1998. Are large, infrequent disturbances quali-
tatively different from small, frequent disturbances? Ecosys-
tems 1: 524534.
Romme, W.H., Whitby, T.G., Tinker, D.B. & Turner, M.G. 2016.
Deterministic and stochastic processes lead to divergence in
plant communities 25 years after the 1988 Yellowstone fires.
Ecological Monographs 86: 327351.
Sipe, T. & Bazzaz, F. 1995. Gap partitioning among maples (Acer)
in central New England: survival and growth. Ecology 76:
Speer, J.H. 2010. Fundamentals of tree-ring research. University of
Arizona Press, Tucson, AZ, US.
Splechtna, B.E., Gratzer, G. & Black, B.A. 2005. Disturbance his-
tory of a European old-growth mixed-species forest A spa-
tial dendro-ecological analysis. Journal of Vegetation Science
16: 511522.
Svoboda, M., Fraver, S., Janda, P., Ba
ce, R. & Zen
a, J.
2010. Natural development and regeneration of a Central
European montane spruce forest. Forest Ecology and Manage-
ment 260: 707714.
Svoboda, M., Janda, P., Ba
ce, R., Fraver, S., Nagel, T.A.,
Rejzek, J., Mikol
s, M., Douda, J., Boubl
ık, K. &
P. 2014. Landscape-level variability in historical disturbance
in primary Picea abies mountain forests of the eastern
Carpathians, Romania. Journal of Vegetation Science 25:
Trotsiuk, V., Svoboda, M., Janda, P., Mikolas, M., Bace, R.,
Rejzek, J., Samonil, P., Chaskovskyy, O., Korol, M. &
Myklush, S. 2014. A mixed severity disturbance regime
in the primary Picea abies (L.) Karst. forests of the Ukrai-
nian Carpathians. Forest Ecology and Management 334:
UNEP. 2007. Carpathians environmental outlook 2007.TheUnited
Nations Environment Programme, Geneva, CH.
Journal of Vegetation Science
1138 Doi:10.1111/jvs.12581 ©2017 International Association for Vegetation Science
Long-term canopyunderstorey responses R. Ba
ce et al.
Wood,S.2006.Generalized additive models: an introduction with R.
CRC Press, Boca Raton, FL, US.
Zeppenfeld, T., Svoboda, M., DeRose, R.J., Heurich, M.,
uller, J.,
a, P., Star
y, M., Ba
ce, R. & Donato, D.C.
2015. Response of mountain Picea abies forests to stand-
replacing bark beetle outbreaks: neighbourhood effects
lead to self-replacement. Journal of Applied Ecology 52:
Zywiec, M., Holeksa, J., Wesołowska, M., Szewczyk, J., Zwijacz-
Kozica, T. & Kapusta, P. 2013. Sorbus aucuparia regenera-
tion in a coarse-grained spruce forest a landscape scale.
Journal of Vegetation Science 24: 735743.
Supporting Information
Additional Supporting Information may be found in the
online version of this article:
Appendix S1. List of stands and their environmental
Appendix S2. Description of delineation of primary
forests in each country.
Appendix S3. Distribution of time since disturbance
and disturbance severity across countries.
Appendix S4. Distribution of time since disturbance
and disturbance severity across stands.
Appendix S5. Means SD of variables used in the
analyses, per stand and country.
Appendix S6. Probability density of maximum
understorey light availability as a function of the influence
of time since main disturbance and severity of the distur-
Appendix S7. Probability density of stem densities as
a function of the influence of time since main disturbance
and severity of the disturbance.
Journal of Vegetation Science
Doi: 10.1111/jvs.12581©2017 International Association for Vegetation Science
R. Ba
ce et al. Long-term canopyunderstorey responses
... The tight link between disturbance and regeneration paired with the sporadic nature of forest disturbances places a major constraint on the study of tree growth during recruitment. Disturbance studies are often limited to the rare opportunities where disturbances impact pre-existing plot networks (e.g., Bače et al., 2017;Drobyshev, 1999;Kneeshaw and Bergeron, 1998;Richards and Hart, 2011;Runkle and Yetter, 1987), and are strongly complemented by advances in remote sensing (Senf et al., 2019). Such cases provide detailed before-after controlimpact style studies, but plot-level studies are generally too limited in spatial scope to assess landscape variability and the time scales of by plot-based and remotely sensed data streams are of insufficiently length to assess differences between past dynamics compared to our contemporary era of rapid environmental change. ...
... Indeed, previous studies have already provided evidence for this facilitating effect stimulating height growth in saplings occupying the same growing space (Cunningham et al., 2006;Küßner et al., 2000). Competition between GRTs of the same size promotes quick radial growth while competition with advanced regeneration (RTs), given the size advantage of this group (Bače et al., 2017;Macek et al., 2017), negatively affects GRTs. ...
... Disturbance legacies have long-lasting effects on understorey growth trends (Bače et al., 2017;Kashian et al., 2005). The magnitude of growth release in GRTs, which declines from 1900 to 1975, could reflect the region's disturbance history. ...
While shifting disturbance rates and climate change have major implications for the structure of contemporary forests through their effects on adult tree mortality, the responses of regenerating trees to disturbances and environmental variation will ultimately determine the structure and functioning of forests in the future. Assessing the resilience of forests to changing conditions requires information on what constrains tree performance during recruitment and whether recruitment dynamics have changed throughout history. We analyzed growth patterns in a large sample of tree cores (n = 14 793) collected from primary Picea forests throughout the Carpathian Mountains. Growth rate anomalies recorded in tree-rings permitted the reconstruction of several key recruitment and disturbance parameters: (1) whether individuals were recruited after a period of competitive suppression (Released Trees; RT; 66% of trees) or immediately following gap formation (Gap Recruited Trees; GRT; 33%), (2) growth rates during recruitment, (3) the duration of recruitment and (4) historical disturbance severity variation. High neighborhood density led to lower growth rates in RTs, but favored a higher growth rate in GRTs. Winter temperatures were positively correlated with Picea growth during recruitment, GRTs were also more sensitive to winter precipitation. Recent increases in growth during recruitment and reductions in recruitment intervals suggest that rates of canopy replacement have increased over recent decades. Assessments of forest resilience must recognize that constraints on tree growth differ during recruitment and interact with disturbance severity. An individual's experience prior to competitive release and factors altering the immediate abiotic conditions of a recruiting individual (competition and disturbance severity) are important determinants of canopy replacement rates; these recruitment parameters will certainly interact with shifting disturbance regimes. Ultimately, increasing growth rates and decreasing recruitment intervals suggest that forest dynamics are accelerating, and are potentially compensating for recent increases in tree mortality rates.
... In boreal forests, Norway spruce is a shade-tolerant tree, and even small gaps resulting from the death of individual trees are quickly filled with young ones that had been suppressed for decades in the seedling and sapling bank under the canopy (Hytteborn et al., 1991;Kuuluvainen and Aakala, 2011;Hytteborn and Verwijst, 2014). In subalpine forests, at high altitudes the shade tolerance of Norway spruce is considerably lower, and although advance regeneration can play a role, direct sunlight constitutes an important factor in seedling establishment (Brang et al., 2005;Baier et al., 2007;Bače et al., 2017). At higher elevations, direct insolation probably counteracts the temperature limitation on tree growth (Körner, 2012). ...
Full-text available
Regeneration of coniferous tree species depends on forest structures related to tree death: canopy gaps, decaying wood, and windthrow mounds and pits. This also applies to Norway spruce (Picea abies), which abundantly regenerates from seeds on dead wood and windthrow mounds, and which benefits from gaps in the forest canopy. Accumulated evidence shows much lower mortality of seedlings on elevated structures than on soil, but for saplings the long-term mortality pattern is poorly known. We studied Norway spruce sapling mortality in a natural old-growth subalpine forest in the Western Carpathians (S Poland), tracking the fate of 1304 saplings growing in different conditions. We found that the 25-year sapling mortality rate significantly differed between three types of microsites, being highest on dead wood (49%), intermediate on undisturbed soil (41%) and lowest on mounds (30%). Sapling mortality decreased with increasing canopy openness. Mortality was highest for medium-sized saplings (2–3 m high), regardless of microsite type and mortality cause. For saplings growing on dead wood, mortality increased with increasing decay class of the wood. For those growing on soil, mortality was higher in patches of Athyrium distentifolium than when surrounded by Vaccinium myrtillus and Dryopteris dilatata. We did not detect differences in mortality between saplings growing next to tree trunks and those growing away from them. The pattern of sapling mortality differed from that of seedlings, but it did not change the regeneration pattern over microsites, which is established at the seedling stage. In particular, we believe that the considerable differences in mortality rates between dead wood and the other microsites at the sapling stage cannot counterbalance increased seedling recruitment and survival on dead wood.
... Moreover, the "complexity challenge" related to old-growth forests is not restricted to the boreal forests of eastern Canada. The complexity of oldgrowth forests has been highlighted in temperate forests in North America [18,56,57], as well as in Europe, in boreal [16,58,59] and temperate forests [60][61][62]. Since old-growth temperate forests in Asia, Oceania, and South America share similar attributes with those in Europe and North America [63], it is also likely that we observe similar dynamic patterns in these regions. ...
Full-text available
Research Highlights: Radial growth patterns of trees growing in old-growth boreal forests in eastern Canada can be grouped into a small number of simple patterns that are specific to different old-growth forest types or successional stages. Background and Objectives: Identifying the main radial growth trends in old-growth forests could help to develop silvicultural treatments that mimic the complex dynamics of old-growth forests. Therefore, this study aimed to identify the main radial growth patterns and determine how their frequencies change during forest succession in old-growth forests, focusing on boreal landscapes in eastern Canada. Materials and Methods: We used dendrochronological data sampled from 21 old-growth stands in the province of Quebec, Canada. Tree-ring chronologies were simplified into chronologies of equal length to retain only primary growth trends. We used k-means clustering to identify individual growth patterns and the difference in growth-pattern frequency within the studied stands. We then used non-parametric analyses of variance to compare tree or stand characteristics among the clusters. Results: We identified six different growth patterns corresponding to four old-growth forest types, from stands at the canopy breakup stage to true old-growth stands (i.e., when all the pioneer cohort had disappeared). Secondary disturbances of low or moderate severity drove these growth patterns. Overall, the growth patterns were relatively simple and could be generally separated into two main phases (e.g., a phase of limited radial increment size due to juvenile suppression and a phase of increased radial increment size following a growth release). Conclusions: The complexity of old-growth forest dynamics was observed mainly at the stand level, not at the tree level. The growth patterns observed in true old-growth forests were similar to those observed following partial or stem-selection cuts in boreal stands; thus, these silvicultural treatments may be effective in mimicking old-growth dynamics.
... Contrary to the fertile climate, manifestations of such negative factors as drought, dust storm, and hot wind have a detrimental effect on agriculture in the Stavropol Territory. All of the above processes significantly worsen the condition of agricultural lands, thereby causing harm to forest plantations, which in turn contribute to protective functions [3]. ...
Full-text available
The development of sustainable management by forest shelterbelts is one of the initial factors in the agricultural and forest sector. A long, productive, economically interesting interrelation of anthropogenesis and forest ecosystems entails a stable management of forest shelterbelts. In order to preserve forest resources and forest areas, a solid foundation is needed to exploit the needs of the future of humanity. The development of indicators of sustainable development of forest ecosystem use is rapidly moving all over the world. Forest shelterbelts have such undeniable features affecting the environmental factor as a unique forest floor, unique herbaceous plants, specific animal world, a rare microclimate, the direction of the soil-forming process, etc. The main economic role of forest plantations is certainly protective. Forest plantations have a positive impact on the state of agricultural land due to the ratio of various forms of land use, as well as the introduction of washed and semi-washed plots into economic circulation. Sometimes, forest belts also get into the risk zone; this can be due to both natural conditions and human exposure.
... In particular, we employed a penalised spline approach with crossvalidated penalty coefficient estimation. This is an approach that has been used successfully previously in various fields, ranging from solar irradiance modelling (Brabec et al., 2014) to climatology to forest development (Bace et al., 2017) and many others. ...
Fog is a very important and complex atmospheric phenomenon of the utmost importance for the environment and for human society. For practical reasons, fog occurrence is observed regularly at meteorological stations worldwide. Decreasing trends in fog frequency reported from numerous regions have been often associated with either decreasing pollution or climate change, including increasing temperature and changes in atmospheric circulation. We have examined the data on fog occurrence from twelve Czech sites representing different environments (urban, rural, mountain), geographical areas, and altitudes across the country. For our analysis we used long-term records from the time period of 1961-2018, covering both the ambient air's heavily polluted periods of the 1970s and 1980s and the cleaner period, following the adoption of new, more stringent legislation and effective countermeasures after the 1990s. We applied a generalised additive model (GAM) framework as a flexible, semiparametric regression approach to address nonlinear trend shapes in a formalised and unified way. In particular, we employed a penalised spline approach with cross-validated penalty coefficient estimation. Our study confirmed non-linear behaviour for both year-to-year trends and annual seasonality. Our results showed further that over the analysed, almost sixty-year period, fog occurrence has decreased significantly at all the examined sites, though the pattern of the long-term change differed among individual sites. Moreover, we have found significant seasonality in fog occurrence, though it is different at individual sites. Furthermore, apart from the overall annual fog probability change over the years, at some sites the fog's seasonal profile has also deformed substantially over the long term.
... PERMDISP analysis on the matrix of pairwise dissimilarities of plots based on the Bray-Curtis dissimilarity metric calculated on square-root transformed stem counts for each species in each plot. Canopy and understory layers have been found to have differing responses to disturbance (e.g.Bače et al. 2017), so we tested three subsets of data: trees only (≥10cm DBH), shrubs and saplings only (<10cm DBH and >1m tall), and an overall community assessment which combined both subsets. Surveyors did not count all the tree stems in 2000; therefore we omitted that year.To determine which species became more frequent and/or abundant on the landscape 3 years and 19 years after the storm we used an indicator species analysis (ISA). ...
Full-text available
1. When deforestation results in small forest fragments surrounded by a non‐forest matrix, forest stands within these fragments experience changes in structure and community composition. They also continue to experience natural disturbances like hurricanes and ice storms. It is unclear whether the landscape context of forest stands influences plant community response to natural disturbances. 2. Using data from surveys of forested plots in the years immediately following and 19 years after a severe ice storm, we measured changes in woody stem density, species richness, and beta diversity. 3. Plots with greater storm damage had greater gains in stems and species, and greater shifts in community composition. In addition, there were interactions between the degree of storm damage and landscape context. The short‐term effects of storm damage were magnified in plots with less forest on the surrounding landscape and farther from the forest edge. In plots with high damage, a return towards pre‐storm conditions over the long‐term occurred more often in plots farther from the forest edge compared to those close to the edge. 4. Synthesis: Future climate scenarios predict increases in severe weather and accompanying ecosystem disturbance. Our results show that it is important to consider landscape context when assessing the response of forest communities to such disturbances.
Development of primary spruce forests is driven by a series of disturbances, which also have an important influence on the understorey vegetation and its diversity. Early post-disturbance processes have been intensively studied, however, very little is known about the long-term effects of disturbances on the understorey. We quantified disturbance history using dendrochronological methods to investigate its impact on vascular plant diversity and understorey species composition. We sampled 141 plots randomly assigned throughout primary stands located in the zone of natural montane acidophilous forests dominated by Picea abies (L.) Karst. in the Western Carpathians. Dendrochronological, dendrometric, and environmental parameters were related to understorey properties using ordination methods and a Bayesian approach using multilevel linear models (GLMM). Time since the last disturbance (23–260 years ago; mostly windstorms and bark beetle outbreaks) had a significant effect on understorey species composition of the current communities, and it also interacted with disturbance severity to influence species diversity. The effect of disturbances on the understorey was largely mediated by the alteration of stand structure (age, DBH, canopy openness), Vaccinium myrtillus L. cover, and topsoil chemical properties. A period of severe disturbances between 1860 and 1890 resulted in a legacy of our current, relatively homogeneous spruce stands with less diverse sciophilous understorey dominated by V. myrtillus, which is in contrast to heterogeneous stands (in terms of age and spatial structure) driven by small-scale, lower-severity disturbances, which led to an understorey enriched by species with higher demands on light and topsoil quality (higher K concentration and lower C/N ratio). All developmental pathways following disturbances create a unique complex of spatiotemporal understorey variability in the montane spruce forests. Therefore, to preserve their full diversity, disturbances of all severities and sizes should be accepted as natural drivers, both in the field of nature conservation and close-to-nature forestry efforts.
Understanding the processes shaping the composition of assemblages at multiple spatial scales in response to disturbance events is crucial for preventing ongoing biodiversity loss and for improving current forest management policies aimed at mitigating climate change and enhancing forest resilience. Deadwood-inhabiting fungi represent an essential component of forest ecosystems through their association with deadwood decomposition and the cycling of nutrients and carbon. Although we have sufficient evidence for the fundamental role of deadwood availability and variability of decay stages for fungal species diversity, the influence of long-term natural disturbance regimes as the main driver of deadwood quantity and quality has not been sufficiently documented. We used a dendroecological approach to analyse the effect of 250-years of historical natural disturbance and structural habitat elements on local (plot-level) and regional (stand-level) species richness of deadwood-inhabiting fungi. We used data collected from 51 study plots within nine best-preserved primary spruce forest stands distributed across the Western Carpathian Mountains. Historical disturbances shaped the contemporary local and regional species richness of fungi, with contrasting impacts of disturbance regime components at different spatial scales. While local diversity of red-listed species has increased due to higher disturbance frequency, regional diversity of all species has decreased due to higher severity historical disturbances. The volume of deadwood positively influenced the species richness of deadwood-inhabiting fungi while canopy openness had a negative impact. The high number of observed rare species highlights the important role of primary forests for biodiversity conservation. From a landscape perspective, we can conclude that the distribution of species from the regional species pool is-at least to some extent-driven by past spatiotemporal patterns of disturbance events. Natural disturbances occurring at higher frequencies that create a mosaic forest structure are necessary for fungal species-especially for rare and endangered taxa. Thus, both the protection of intact forest landscapes and forest management practises that emulate natural disturbance processes are recommended to support habitats of diverse fungal communities and their associated ecosystem functions.
This book brings together segmental knowledge and creates new insights on the sustainability of agricultural systems, critically analyzing not only individual system components, but also focusing on interactions between them and external environments. This book is primarily devoted to (1) agricultural agribusiness, (2) policies and institutions, and (3) farming systems. The compelling collection of chapters presents critical, comparative, and balanced perspectives on what changes are needed to achieve and maintain sustainability in agricultural systems, actively leading to new ways of thinking about these complex issues. The research presented relies on an array of methods developed within complex systems science, addresses the existing gaps in the scholarship, and uses original data collected on the development of agricultural systems. Finally, the authors provide robust conclusions and recommendations for both scholars and practitioners in the field of studying, constructing, and maintaining sustainable agricultural systems. The special focus of the book is on technologies, policies, and management systems enabling sustainable agricultural development. A rich collection of practical cases could be used to move from theories to reality. The book appeals to both academics and professionals working in the field.
Forest reclamation should be the first step in the development of agricultural land. The formation of sustainable forest management systems is one of the primary tasks of forestry and rural agriculture. Sustainable management is aimed at purposefully, temporarily, and economically advantageous interactions between people and forest systems. Forest resources and forest areas should be used on a straightforward basis to satisfy the current and future needs of people. The indicators for the sustainable use of forest ecosystems are actively developed.
Full-text available
The severity and spatial extent of bark-beetle outbreaks substantially increased in recent decades worldwide. The ongoing controversy about natural forest recovery after these outbreaks highlights the need for individual-based long-term studies, which disentangle processes driving forest regeneration. However, such studies have been lacking. To fill this gap, we followed the fates of 2,552 individual seedlings for 12 years after a large-scale bark-beetle outbreak that caused complete canopy dieback in mountain Norway spruce (Picea abies) forests in SE Germany. Here we explore the contribution of advance, disturbance-related and post-disturbance regeneration to forest recovery. Most seedlings originated directly within the three-year dieback of canopy trees induced by bark-beetle outbreak. After complete canopy dieback, the establishment of new seedlings was minimal. Surprisingly, advance regeneration formed only a minor part of all regeneration. However, because it had the highest survival rate, its importance increased over time. The most important factor influencing the survival of seedlings after disturbance was their height. Survival was further modified by microsite: seedlings established on dead wood survived best, whereas almost all seedlings surrounded by graminoids died. For 5 cm tall seedlings, annual mortality ranged from 20% to 50% according to the rooting microsite. However, for seedlings taller than 50 cm, annual mortality was below 5% at all microsites. While microsite modified seedling mortality, it did not affect seedling height growth. A model of regeneration dynamics based on short-term observations accurately predicts regeneration height growth, but substantially underestimates mortality rate - thus predicting more surviving seedlings than were observed. We found that Picea abies forests were able to regenerate naturally even after severe bark-beetle outbreaks owing to advance and particularly disturbance-related regeneration. This, together with microsite-specific mortality, yields structurally and spatially diverse forests. Our study thus highlights the so far unrecognized importance of disturbance-related regeneration for stand recovery after bark-beetle outbreaks.
Full-text available
Young, recently burned forests are increasingly widespread throughout western North America, but forest development after large wildfires is not fully understood, especially regarding effects of variable burn severity, environmental heterogeneity, and changes in drivers over time. We followed development of subalpine forests after the 1988 Yellowstone fires by periodically resampling permanent plots established soon after the fires. We asked two questions about patterns and processes over the past 25 years: (1) Are plant species richness and community composition converging or diverging across variation in elevation, soils, burn severity, and post-fire lodgepole pine (Pinus contorta var. latifolia) density? (2) What are the major controls on post-fire species composition, and has the relative importance of controls changed over time? For question 1, we sampled 10-m2 plots (n = 552) distributed among three geographic areas that differ in elevation and substrate; plots spanned the spectrum of fire severities and were resampled periodically from 1991 to 2013. For question 2, we sampled 0.25-ha plots (n = 72), broadly distributed across areas that burned as stand-replacing fire, in 1999 and 2012. Richness and species composition diverged early on between infertile low-elevation areas (lower richness) and more fertile high-elevation areas (greater richness). Richness increased rapidly for the first 5 yr post-fire, then leveled off or increased only slowly thereafter. Only 6% of 227 recorded species were nonnative. Some annuals and species with heat-stimulated soil seed banks were associated with severely burned sites. However, most post-fire species had been present before the fire; many survived as roots or rhizomes and regenerated rapidly by sprouting. Among the 72 plots, substrate, temperature, and precipitation (the abiotic template) were consistently important drivers of community composition in 1999 and 2012. Post-fire lodgepole pine abundance was not significant in 1999 but was the most important driving variable by 2012, with a negative effect on presence of most understory species, especially annuals and shade-intolerant herbs. Burn severity was significant in 1999 but not in 2012, and distance to unburned forest had no influence in either year. The 1988 fires did not fundamentally alter subalpine forest community assemblages in Yellowstone, and ecological memory conferred resilience to high-severity fire.
Mixed-severity disturbance regimes are prevalent in temperate forests worldwide, but key uncertainties remain regarding the variability of disturbance-mediated structural development pathways. This study investigates the influence of disturbance history on current structure in primary, unmanaged Norway spruce (Picea abies) forests throughout the Carpathian Mountains of central and eastern Europe, where windstorms and native bark beetle outbreaks are the dominant natural disturbances. We inventoried forest structure on 453 plots (0.1 ha) spanning a large geographical gradient (\>1,000 km), coring 15–25 canopy trees per plot for disturbance history reconstruction (tree core total n = 11,309). Our specific objectives were to: (1) classify sub-hectare-scale disturbance history based on disturbance timing and severity; (2) classify current forest structure based on tree size distributions (live, dead, standing, downed); (3) characterize structural development pathways as revealed by the association between disturbance history and current forest structural complexity. We used hierarchical cluster analysis for the first two objectives and indicator analysis for the third. The disturbance-based cluster analysis yielded six groups associated with three levels of disturbance severity (low, moderate, and high canopy loss) and two levels of timing (old, recent) over the past 200 years. The structure-based cluster analysis yielded three groups along a gradient of increasing structural complexity. A large majority of plots exhibited relatively high (53\%) or very high (26\%) structural complexity, indicated by abundant large live trees, standing and downed dead trees, and spruce regeneration. Consistent with conventional models of structural development, some disturbance history groups were associated with specific structural complexity groups, particularly low-severity/recent (very high complexity) and high-severity/recent (moderate complexity) disturbances. In other cases, however, the distribution of plots among disturbance history and structural complexity groups indicated either divergent or convergent pathways. For example, multiple disturbance history groups were significantly associated with the high complexity group, demonstrating structural convergence. These results illustrate that complex forest structure – including features nominally associated with old-growth – can be associated as much with disturbance severity as it is with conventional notions of forest age. Because wind and bark beetles are natural disturbance processes that can induce relatively high levels of tree mortality while simultaneously contributing to structural complexity and heterogeneity, we suggest that forest management plans allow for the stochastic occurrence of disturbance and variable post-disturbance development trajectories. Such applications are especially appropriate in conservation areas where biodiversity and forest resilience are management objectives, particularly given projections of increasing disturbance activity under global change.
In order to gauge ongoing and future changes to disturbance regimes, it is necessary to establish a solid baseline of historic disturbance patterns against which to evaluate these changes. Further, understanding how forest structure and composition respond to variation in past disturbances may provide insight into future resilience to climate-driven alterations of disturbance regimes.
Ecological memory is central to how ecosystems respond to disturbance and is maintained by two types of legacies – information and material. Species life-history traits represent an adaptive response to disturbance and are an information legacy; in contrast, the abiotic and biotic structures (such as seeds or nutrients) produced by single disturbance events are material legacies. Disturbance characteristics that support or maintain these legacies enhance ecological resilience and maintain a “safe operating space” for ecosystem recovery. However, legacies can be lost or diminished as disturbance regimes and environmental conditions change, generating a “resilience debt” that manifests only after the system is disturbed. Strong effects of ecological memory on post-disturbance dynamics imply that contingencies (effects that cannot be predicted with certainty) of individual disturbances, interactions among disturbances, and climate variability combine to affect ecosystem resilience. We illustrate these concepts and introduce a novel ecosystem resilience framework with examples of forest disturbances, primarily from North America. Identifying legacies that support resilience in a particular ecosystem can help scientists and resource managers anticipate when disturbances may trigger abrupt shifts in forest ecosystems, and when forests are likely to be resilient.
Mountain forests are among the most important ecosystems in Europe as they support numerous ecological, hydrological, climatic, social, and economic functions. They are unique relatively natural ecosystems consisting of long-lived species in an otherwise densely populated human landscape. Despite this, centuries of intensive forest management in many of these forests have eclipsed evidence of natural processes, especially the role of disturbances in long-term forest dynamics. Recent trends of land abandonment and establishment of protected forests have coincided with a growing interest in managing forests in more natural states. At the same time, the importance of past disturbances highlighted in an emerging body of literature, and recent increasing disturbances due to climate change are challenging long-held views of dynamics in these ecosystems. Here, we synthesize aspects of this Special Issue on the ecology of mountain forest ecosystems in Europe in the context of broader discussions in the field, to present a new perspective on these ecosystems and their natural disturbance regimes. Most mountain forests in Europe, for which long-term data are available, show a strong and long-term effect of not only human land use but also of natural disturbances that vary by orders of magnitude in size and frequency. Although these disturbances may kill many trees, the forests themselves have not been threatened. The relative importance of natural disturbances, land use, and climate change for ecosystem dynamics varies across space and time. Across the continent, changing climate and land use are altering forest cover, forest structure, tree demography, and natural disturbances, including fires, insect outbreaks, avalanches, and wind disturbances. Projected continued increases in forest area and biomass along with continued warming are likely to further promote forest disturbances. Episodic disturbances may foster ecosystem adaptation to the effects of ongoing and future climatic change. Increasing disturbances, along with trends of less intense land use, will promote further increases in coarse woody debris, with cascading positive effects on biodiversity, edaphic conditions, biogeochemical cycles, and increased heterogeneity across a range of spatial scales. Together, this may translate to disturbance-mediated resilience of forest landscapes and increased biodiversity, as long as climate and disturbance regimes remain within the tolerance of relevant species. Understanding ecological variability, even imperfectly, is integral to anticipating vulnerabilities and promoting ecological resilience, especially under growing uncertainty. Allowing some forests to be shaped by natural processes may be congruent with multiple goals of forest management, even in densely settled and developed countries.
Young, recently burned forests are increasingly widespread throughout western North America, but forest development after large wildfires is not fully understood, especially regarding effects of variable burn severity, environmental heterogeneity, and changes in drivers over time. We followed development of subalpine forests after the 1988 Yellowstone fires by periodically resampling permanent plots established soon after the fires. We asked two questions about patterns and processes over the past 25 years: (1) Are plant species richness and community composition converging or diverging across variation in elevation, soils, burn severity, and post-fire lodgepole pine (Pinus contorta var. latifolia) density? (2) What are the major controls on post-fire species composition, and has the relative importance of controls changed over time? For question 1, we sampled 10-m2 plots (n = 552) distributed among three geographic areas that differ in elevation and substrate; plots spanned the spectrum of fire severities and were resampled periodically from 1991 to 2013. For question 2, we sampled 0.25-ha plots (n = 72), broadly distributed across areas that burned as stand-replacing fire, in 1999 and 2012. Richness and species composition diverged early on between infertile low-elevation areas (lower richness) and more fertile high-elevation areas (greater richness). Richness increased rapidly for the first 5 yr post-fire, then leveled off or increased only slowly thereafter. Only 6% of 227 recorded species were nonnative. Some annuals and species with heat-stimulated soil seed banks were associated with severely burned sites. However, most post-fire species had been present before the fire; many survived as roots or rhizomes and regenerated rapidly by sprouting. Among the 72 plots, substrate, temperature, and precipitation (the abiotic template) were consistently important drivers of community composition in 1999 and 2012. Post-fire lodgepole pine abundance was not significant in 1999 but was the most important driving variable by 2012, with a negative effect on presence of most understory species, especially annuals and shade-intolerant herbs. Burn severity was significant in 1999 but not in 2012, and distance to unburned forest had no influence in either year. The 1988 fires did not fundamentally alter subalpine forest community assemblages in Yellowstone, and ecological memory conferred resilience to high-severity fire.