Linking spatiotemporal disturbance history with tree regeneration and diversity in an old-growth forest in northern Japan

Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13

Contents lists available at ScienceDirect

Perspectives in Plant Ecology, Evolution and Systematics journal homepage: www.elsevier.com/locate/ppees

Research article

Linking spatiotemporal disturbance history with tree regeneration and diversity in an old-growth forest in northern Japan Jan Altman a,b,∗ , Pavel Fibich a,b , Jan Leps b , Shigeru Uemura c , Toshihiko Hara d , Jiri Dolezal a,b a

Institute of Botany of the Czech Academy of Sciences, Zámek 1, 25243 Pr˚ uhonice, Czech Republic ˇ Faculty of Science, Department of Botany, University of South Bohemia, Na Zlaté Stoce 1, 37005 Ceské Budˇejovice, Czech Republic c Forest Research Station, Field Science Center for Northern Biosphere, Hokkaido University, Nayoro 096-0071, Japan d Institute of Low Temperature Science, Hokkaido University, Kita-ku, Sapporo 060-0819, Japan b

a r t i c l e

i n f o

Article history: Received 17 March 2015 Received in revised form 11 April 2016 Accepted 13 April 2016 Available online 3 May 2016 Keywords: Tree rings Disturbance Spatial pattern Forest regeneration Forest dynamics Dwarf bamboo Tropical cyclones

a b s t r a c t Knowledge of long-term spatiotemporal effects of disturbances on forest structure, tree regeneration and species composition is key for understanding forest dynamics and predicting future forest responses to climate change. Here, we explore the spatiotemporal impact of disturbances of different severities on tree recruitment and diversity in species-rich oak-fir-maple forest in Hokkaido, a typhoon-prone area in northern Japan, over the past 230 years. The forest disturbance history was reconstructed by growthrelease analysis from more than 45,500 tree rings of 385 trees belonging to 15 species. A mixed severity disturbance regime was prevalent over the study period. Altogether, 310 major and 293 moderate growth releases were identified. These were both temporally and spatially localized, with 80% of events detected in only four time periods: 1775–1784, 1815–1839, 1880–1909 and 1950–1979. Disturbances were followed by major recruitment pulses, each lasting around 30 years. Dendrochronological reconstructions alone indicate that severe (i.e. high proportion of releases), infrequent disturbances control tree regeneration and forest development (from oak-dominated forests to mixed-stand with higher proportion of shade-tolerant tree species). However, a combination of temporal and spatial pattern analysis revealed that less severe disturbances, creating small gaps, promote higher density and diversity of recruitment (altogether 19 tree species recorded) compared with severe disturbances. The latter create large forest gaps which became overgrown by dwarf bamboo and suppress tree regeneration. These results provide evidence that severe disturbances interacting with a strong biotic understory filter (as dwarf bamboo), can disrupt forest ecosystem dynamics by significantly reducing the extent and diversity of tree recruitment. Our findings are important as most climate models predict an elevated intensity of typhoons in Northeast Asia. We conclude that a combination of temporal and spatial analyses, as presented here, is necessary to disentangle the complex drivers of long-term forest dynamics. © 2016 Elsevier GmbH. All rights reserved.

1. Introduction Natural disturbances play an essential role in shaping forest structure and species composition across the globe (Altman et al., 2013a; Fraver et al., 2009; Chambers et al., 2013; Papaik and Canham, 2006). It is essential for the understanding of forest dynamics to reconstruct historical disturbances, determine disturbance regimes and characterize their spatiotemporal effect in different forest types and regions. Knowledge of natural distur-

∗ Corresponding author at: Institute of Botany of the Czech Academy of Sciences, ˚ Czech Republic. Zámek 1, 25243 Pruhonice, E-mail address: [email protected] (J. Altman). http://dx.doi.org/10.1016/j.ppees.2016.04.003 1433-8319/© 2016 Elsevier GmbH. All rights reserved.

bance dynamics, and recognition of their spatiotemporal changes, are important to both applied and theoretical ecology (Altman et al., 2013b; Fraver et al., 2009). The intensity and frequency of disturbance influences the size and spatiotemporal distribution of the forest gaps they create, which consequently affect the subsequent growth and regeneration of trees. Furthermore, the structure and composition of a forest has an important role in determining the consequences of a disturbance of a given intensity. The subsequent regeneration and composition of plant species that colonize an area beneath the disturbed canopy is controlled both by abiotic and biotic factors (Everham and Brokaw, 1996). There is ongoing discussion about the effects of disturbances of different severities on forest structure and

2

J. Altman et al. / Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13

the “optimum” gap size (i.e. most suitable) for tree regeneration and tree diversity (e.g. Albanesi et al., 2008; Kern et al., 2013; Lorimer, 1989; Romme et al., 1998). Some studies describe lower diversity or forest regeneration in large gaps (Kern et al., 2013; Qin et al., 2011), while others showed the opposite trend (Naaf and Wulf, 2007). However, several studies have recently shown that mixed severity disturbance regimes are common in various forest ecosystems (e.g. Khakimulina et al., 2015; Odion et al., 2014). The variation in disturbance severity enables the coexistence of species with different ecological strategies (Loehle, 2000). Some species can grow very slowly for extended periods beneath the canopy (shade-tolerant species) whereas other species can grow beneath the canopy for only limited periods (light-demanding species) (Mori and Takeda, 2004; Rentch et al., 2003). The former group can persist for a long time beneath the canopy and can wait for a disturbance event to attain canopy status. Light-demanding tree species usually show rapid initial growth, but after a few years without a disturbance, growth declines and the trees die. However, large gaps (i.e. multitree falls) also present an opportunity for ecological invasion of strong herbaceous competitors and the consequent suppression of forest diversity (Baret et al., 2008; Burnham and Lee, 2010). Different preferences for regeneration niches occur due to various life strategies. The trade-offs between species’ physiologically based life-history traits seems to be the most important factor for the mortality and community abundance of individual species (Kobe and Vriesendorp, 2011). Heterogeneous light requirements and variable growth strategies play a key role in species coexistence (Gravel et al., 2010; Valladares et al., 2012). Most studies exploring the effects of disturbances on forest dynamics have focused on either the temporal or the spatial scale of past disturbances. However, relatively few studies have linked these two aspects, especially at smaller scales, while at broader scales some studies do exist (see e.g. Tepley and Veblen, 2015; Trotsiuk et al., 2014). The missing connection between temporal and spatial aspects is evident in dendrochronological studies, despite the fact that they can provide high temporal and very high spatial resolution data when compared with other methods (Frelich, 2002). Recent progress in the methods of spatial analysis of marked point patterns (Illian et al., 2008; Wiegand and Moloney, 2014) provides an excellent tool for the combination of spatial aspects of forest dynamics with the results of temporal dendroecological reconstruction of past disturbances. Regardless, only a few studies have tried to connect temporal and spatial patterns in forest dynamics (Duncan and Stewart, 1991; Samonil et al., 2013; Shimatani and Kubota, 2011; Splechtna et al., 2005; Zielonka et al., 2010) and these studies focused on one tree species, forests with only a few tree species (especially mountain forests in central Europe), or specific disturbance events. Hence, the results of such studies have limited application; the real potential for forest ecology of connecting these techniques remains largely unrealised. Connection of high temporal resolution with detailed spatial data enables new insight into forest dynamics and can uncover long-term changes affecting small-scale processes (e.g. forest regeneration). One of the best opportunities for studying disturbances is offered by old-growth temperate forest ecosystems because their longevity enables the reconstruction of past disturbances (e.g. Altman et al., 2013a). However, old-growth forests without human impact are relatively rare in the temperate regions of the northern hemisphere (Peterken, 1996). Until now, most of the knowledge about disturbance regimes in temperate regions came from studies in North America and Europe, with Asia remaining, to a certain extent, neglected. Temperate forests of Northeast Asia (Japan, Korea, Russia, and China) cover over 930,000 square kilometres and represent the most diverse temperate forests on Earth (White, 1983).

The temperate forests of Northeast Asia are shaped by tropical cyclones, which are one of the most common disturbances affecting large areas in the northern hemisphere (hurricanes in the North Atlantic Ocean and the Northeast Pacific Ocean or typhoons in the Northwest Pacific Ocean). Recent studies have documented an increase in tropical cyclones over the past few decades (Coumou and Rahmstorf, 2012; Emanuel, 2013; Holland and Bruyere, 2014) and a surge in their future intensity is also predicted (Grinsted et al., 2013; Murakami et al., 2013). Unfortunately, instrumental measurements of tropical cyclones usually only go back a few decades (Park et al., 2011; Wu et al., 2005), with very few going further back in time (Altman et al., 2013a). A key compositional component of temperate forests in Eastern Asia is understory dominance of native dwarf bamboo. Understory bamboo species do not occur exclusively here, but are also common in other subtropical and tropical forests around the world (Giordano et al., 2009; Van Goethem et al., 2013). Dwarf bamboo forms dense clumps by vigorously extending their rhizomes and competes intensively with other understory species (e.g. Abe et al., 2002; Dolezal et al., 2009; Montti et al., 2011; Tabarelli and Mantovani, 2000; Tomimatsu et al., 2011), which influences forest regeneration (e.g. Abe et al., 2002; Dolezal et al., 2009). Specifically, high coverage of dwarf bamboo minimizes light availability for the vegetation beneath it and thus limits tree establishment (Royo and Carson, 2006). Furthermore, canopy disturbances are known to increase dwarf bamboo biomass and limit subsequent tree regeneration (Noguchi and Yoshida, 2005; Wang et al., 2009). The aim of this study was to explore the frequency, severity and spatial distribution of disturbances and their impacts on forest regeneration and diversity in an oak-fir-maple forest, which represents a widespread vegetation type in Northeast Asia. Our aims were to (1) reconstruct the disturbance history by means of tree rings and determine the disturbance regime, (2) link the temporal and spatial pattern of tree recruitment and diversity to disturbance events and their severity, and (3) explore the role of understory bamboo cover in these processes by measuring bamboo coverage beneath three different light conditions (undisturbed, partly open, and removed tree canopy). We hypothesise that forest dynamics are controlled by infrequent, severe typhoons, which cause regeneration pulses and bring about major compositional changes. Furthermore, because forest understory vegetation is dominated by dwarf bamboo, we hypothesise that bamboo will take better advantage of larger gaps. This dynamic will restrict tree regeneration in large gaps created by severe disturbances, while intermediate disturbances that partly reduce both the tree and bamboo canopies will promote treeseedling regeneration and hence forest diversity in smaller gaps.

2. Materials and methods 2.1. Study area The study was conducted in a natural, conifer-hardwood mixed forest in the Uryu Experimental Forest of Hokkaido University (44◦ 20 N, 142◦ 15 E, 380 m a.s.l.; Fig. S1) in Hokkaido, northern Japan. During 1956–2010, the mean annual temperature and precipitation was around 3.3 ◦ C and 1409 mm, respectively (from Meteorological Report at Moshiri Observatory at Uryu Experimental Forest, Hokkaido University). The mean daily temperature during the coldest months (January and February) was −11.5 ◦ C and for the warmest month (August) +18.4 ◦ C. The snow-cover season extends over half the year, with snow depth around 2–3 m. The climate of Northeast Asia is strongly influenced by cold air masses from Siberia in winter and monsoons and tropical storms or typhoons from the northern Pacific Ocean in summer. The main

J. Altman et al. / Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13

disturbance agents in the study area (and all of East Asia) are typhoons (e.g. Fischer et al., 2013; Nakashizuka and Iida, 1995; Tanaka and Nakashizuka, 1997; Umeki and Kikuzawa, 1999), which occur mainly in August and September. Hokkaido is located on the edge of the typhoon track (Fig. S1) and they occur here less frequently than in areas located southward (Park et al., 2014), which makes this region a suitable place for disturbance reconstruction as the impact of individual events can be delineated. Other disturbance agents such as fires (see Masaka et al., 2000) or insect outbreaks (see Kamata, 2002) are also documented from Hokkaido, although both of these occur predominantly in conifer forests and human-made environments (Furuta, 1989; Kamata, 2002). The forest is composed of deciduous broad-leaved species dominated by Quercus mongolica subsp. crispula, Acer mono, Betula ermanii, Kalopanax pictus, Tilia japonica and Acer japonicum and the conifers Abies sachalinensis and Picea glehnii. The understory is dominated by a dense cover of dwarf bamboo (Sasa senanensis). Large areas of Northeast Asia (Japan, China, Korea and Russia) are dominated by this type of mixed forest (Kuennecke, 2008; Olson et al., 2001). This study was performed in a permanent 1 ha (100 × 100 m) plot, established in 1982 and re-measured in 1992 and 1998. The plot was chosen to be as homogeneous as possible, lying on flat land without any obvious gradient. The soil consists of histosols developed on tertiary andesite. There is currently no forest management applied in our study site and we have no evidence indicating any past forest management. 2.2. Data collection and analysis 2.2.1. Tree-ring analysis Core samples from all stems in the plot >6 cm in diameter at 1.3 m height (DBH) were collected using a steel borer (Mora, Sweden) during 1997–1998. All cores were dried, mounted, sanded and inspected for damage, reaction wood and other aberrant features. Ring widths were measured from pith to bark to the nearest 0.01 mm using the TimeTable measuring device and PAST4 software (http://www.sciem.com). Ring-sequences were crossdated visually using the pattern of wide and narrow rings, and verified using the PAST4 program by the percentage of parallel variation, i.e. Gleichläufigkeit (Eckstein and Bauch, 1969). For results of crossdating see Table S1. 2.2.2. Disturbance reconstruction For the reconstruction of past disturbances, the method of radial-growth averaging criteria, as presented by Nowacki and Abrams (1997), was used. This is one of the most common techniques used for the detection of release events, i.e. an abrupt and sustained increase in the radial growth after the death of neighbouring tree(s), which induce improved light conditions. In this approach, the average radial growth over the preceding 10-year period, M1 (including the target year), and the average radial growth over the subsequent 10-year period, M2 (excluding the target year) are calculated. The percentage growth change (%GC) is obtained by: %GC = [(M2 –M1 )/M1 ] * 100. Releases were subdivided into “moderate” and “major” categories, to better distinguish mild disturbances from severe ones (Black and Abrams, 2003). The minimum thresholds applied for releases were 25% GC for moderate and ≥50% for major release. This technique was selected over newer methods (e.g. Black and Abrams, 2003; Fraver and White, 2005) because of its broad applicability, even for a smaller number of increment cores for individual species (for details see Appendix S1 ‘Rationale behind the selected method for release detection’). Furthermore, to test the robustness of the method used, we compared the pattern of disturbance chronology with the results gained by the more conservative criteria (15-year window; thresholds 50% GC moderate and ≥100% GC major release) as proposed by Lorimer

3

and Frelich (1989). Finally, results of disturbance reconstruction were summarized to 5-year intervals, which minimized the bias caused by missing and partial rings and the lag of tree response to disturbance events (Lorimer and Frelich, 1989). Disturbance reconstruction was made using the R package TRADER (Altman et al., 2014). To disentangle which climate factors triggered disturbances, we compare our reconstructed release chronology with various climatic variables. Specifically, we used 1) the monthly maximum, minimum and mean annual temperature (1885–1979), 2) monthly maximum, minimum and mean annual precipitation (1880–1979), 3) annual maximum, minimum and mean Palmer Drought Severity Index (PDSI) (1765–1979), and 4) maximum wind speed during typhoons (1945–1979). As our reconstruction was made at a 5year resolution, we used the same resolution for climatic variables. Although we have climatic data for the period after 1979, we decided to truncate the dataset as release reconstruction needs a minimum number of rings for release detection (see Nowacki and Abrams, 1997). Temperature and precipitation data are from Sapporo (150 km from study area); PDSI data were accessed from Cook et al. (2010) (http://iridl.ldeo.columbia.edu/SOURCES/.LDEO/.TRL/. MADA/.pdsi/) and data describing typhoon intensity were gained from KNMI climate explorer (http://climexp.knmi.nl/selectfield obs2.cgi?id=someone@somewhere). 2.2.3. Tree establishment reconstruction For establishment reconstruction, trees were divided into four groups: (1) core reaches the pith; (2) core with no pith, but with an arc for missing-ring estimation; (3) core with no pith or arc for missing ring estimation; and (4) trees recorded only during the re-measurement of the plot (i.e. not cored). The estimation of the establishment date was made for each group separately as follows. The first measured year was considered as the recruitment age for the cores that hit the pith (n = 170). For the cores with no pith (n = 168), the number of missing rings to the centre was estimated by the mean growth rate of the 5 rings adjacent to the largest visible arc on the core (for details see Rozas, 2003). Estimation of the cambial age by linear regression based on DBH (see Table S2) of the previous two groups (n = 338) was made for the trees with rotted centres or with absent arcs for missing ring estimation (n = 47), and trees which were not cored (n = 285). Trees recorded only during the re-measurement of the plot were considered as recruits with their first year 1985 or later (n = 202). Establishment reconstruction was estimated at the coring height of 1.3 m. To minimize the lag of establishment reconstruction date compared to the real year of tree germination (caused by missing the years needed for the seedling to reach the coring height), a field experiment was performed. Specifically, the annual height growth rate was measured in 525 seedlings of 23 species located in the vicinity of the permanent plot during 1999–2007. Seedlings were measured both in the dwarf bamboo stand as well as in the plots where bamboo was initially removed (for details of experimental setup see Dolezal et al., 2009). Calculation of mean value per species were based on repeated height measurements. The establishment date (ED) for individual trees was estimated by: ED = YE − (CH/HGs ), where YE = Year of Establishment (derived from the age at coring height), CH = coring height, HGs = mean annual height growth of seedling (s = specific for individual species). This calculation enabled a more accurate date of tree establishment to be obtained. The purpose of this multi-step analysis was to minimize the lag in the establishment reconstruction, not to determine the exact year of tree germination. Currently there is no universal method enabling the precise reconstruction due to environmental and physiological differences between individual seedlings and their different growth rates. To test the robustness of this multi-step approach for different types of samples, a compari-

4

J. Altman et al. / Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13

son between groups of trees whose core reaches the pith and trees with extrapolated age was performed. Finally, results of establishment reconstruction were summarized to 5-year intervals, which mitigate any potential inaccuracy in this reconstruction. 2.2.4. Spatial analysis To reconstruct the spatial distribution of disturbance events and subsequent tree recruitment, all trees taller than 2 m were tagged and their spatial coordinates were recorded inside the plot in 1982 (for this purpose, the plot was divided into 400 5 m × 5 m quadrats). All trees were re-measured by using a laser rangefinder with incorporated clinometer in 1992 and 1998 (n = 760 and n = 786, respectively; see Fig. S2), together with dead trees and those new recruits that reached a height above 2 m. Altogether, 872 stems of 19 woody species were recorded. The spatial patterns of trees were described by the pair correlation functions (Law et al., 2009; Perry et al., 2006) defined as: g(r) = K (r)/2␲r, where K (r) is the derivative of Ripley’s K-function, corresponding to the average number of other points found within a distance r from the typical point (Law et al., 2009; Ripley, 1987). The pair correlation function avoids the cumulative effect of Ripley’s K-function (Illian et al., 2008). Values g(r) = 1, g(r) < 1 or g(r) > 1 suggest a random point process (or so-called “complete spatial randomness”), regularity (uniform pattern) or clumping (aggregated, patchy pattern), respectively. The Mann–Whitney–Wilcox test was used for comparison of tree size (DBH) between release categories. To study the spatial association (e.g. attraction, repulsion or neutral) of two types of trees (e.g. released trees and newly established trees), a cross pair correlation function was applied (Illian et al., 2008). This considers one type of tree as focal (e.g. a released tree) and investigates the association of the second type of tree (e.g. a newly established tree) with it. The cross K function was used to show differences between the number of newly established trees around the released and non-released trees. Possible large-scale inhomogeneity of the density of individuals was reflected by an inhomogeneous pair correlation function with heterogeneous Poisson null model. Here, a density gradient of individuals is non-parametrically estimated by a Gaussian smoothing function with a standard deviation (bandwidth) of 35 m (Law et al., 2009). Following the separation of scales concept, this removes virtual aggregation (e.g. due to some environmental gradient in the plot, expected to operate on larger spatial scales, say 35 m and more), but preserves the effect of neighbor–neighbor interactions (often operating up to 10–20 m) (Illian et al., 2008). For the cross pair correlation function (Illian et al., 2008), the null model assumed that the focal type of tree (“from points”, e.g. released trees) was not changed and heterogeneous Poisson null model (described above) was applied to the second type of tree (“to points”, e.g. newly established trees). To compare the number of species (SN) and species diversity (SD) of trees on different scales in the plot, a “spatialDiversity” function was applied (Fibich et al., 2016). This function can compute an individual species-area relationship (ISAR) function (Wiegand et al., 2007) that is able to predict SN and SD of surrounding species for chosen individuals (e.g. a relationship of SN around individuals of one species and distance from these individuals). Moreover, this function is also designed to analyse the overall spatial diversity, similar to the distance-dependent Simpson index ␣(r) (Shimatani, 2001) defined as the probability that a randomly selected pair of trees within a distance r belong to different species. The “spatialDiversity” function was also used to determine if there were differences between release classes (major, moderate and no release) in the SN and SD of newly established trees in their vicinity. To compare SN and SD around different release classes, the null model was applied with random switching between release classes; together with leaving all positions and marks of newly

established trees untouched (Goreaud and Pélissier, 2003). SN and SD were computed separately for each release class under one pattern realization (e.g. to have SN and SD of newly established trees around trees with moderate and major release separately). Then, differences of SN or SD between the classes were computed across spatial scales (r = 0–25 m with 30 spatial steps) and used as a spatial characteristic (e.g. in a similar way to the K-function). For example, if the difference of SN of moderate minus major release is higher than zero, SN was higher around the trees with moderate releases rather than around the trees with major ones. To assess the significance level of non-random observed patterns, the summary statistics were compared with Monte Carlo simulations producing the confidence envelopes around the expected values for various null models (Diggle, 2003; Illian et al., 2008). Here, 199 random simulations with the same number of points as the original data under null model assumptions were conducted; the fifth highest and the fifth lowest values were taken as the boundaries of the 95% confidence (simulation) envelope. Moreover, the observed patterns were tested against null models (random simulations) over the whole 0–25 m range (suggested range for our 100 × 100 m plot; see Baddeley and Turner (2005)) by the goodness-of-fit test developed by Loosmore and Ford (2006). For species level analysis, only species with more than 19 individuals were chosen. All spatial analyses were performed in R version 3.0.3 (R Core Team, 2015) with packages “spatstat” version 1.31–3 (Baddeley and Turner, 2005) for spatial pattern analyses, and “vegan” version 2.0–10 (Oksanen et al., 2013) for diversity index calculation (i.e. Simpson diversity index). All spatial analyses assumed isotropic (Ripley) edge correction, except “spatialDiversity” that for analyses on the given scale considers just trees, positions of which (± given scale) did not reach the border of the plot. The Epanechnikov kernel function was used to smooth the pair correlation functions (Illian et al., 2008). 2.2.5. Understory bamboo Since dwarf bamboo dominates the understory of the studied forest and has a substantial effect on tree regeneration (Dolezal et al., 2009; Takahashi et al., 2003), field measurement was performed in order to establish its variation in coverage under different light conditions. Specifically, dwarf bamboo coverage beneath three different canopy types was measured: (1) under closed undisturbed canopy; (2) under partly disturbed/opened canopy; and (3) under large gaps (>500 m2 ) where tree canopy was removed by severe disturbance. Altogether, 21 sampling sites (7 in each of the three canopy types) were set up during the spring of 1999 in the close vicinity of the permanent plot and the above-ground biomass of bamboo was cut and removed from 1 m2 . The removed fresh bamboo leaf blades were measured by a LI-COR area meter (LICOR 3100, Lincoln, Nebraska, USA) to calculate leaf area index. The Mann-Whitney-Wilcox test was used for comparison of bamboo coverage (expressed by leaf area index) between the three different light conditions. 3. Results 3.1. Disturbance reconstruction In total, 603 releases (293 moderate, 310 major) were identified with radial-growth averaging criteria (Nowacki and Abrams, 1997) from more than 45,500 tree rings of 385 trees belonging to 15 tree species; the average number of release events per tree was 1.6 in the period 1760–1989. A disturbance chronology was constructed for the period from 1760 to 1989 at 5-year intervals (Fig. 1). Releases occurred in 41 of the 46 intervals and the per-

J. Altman et al. / Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13

5

Fig. 1. Disturbance chronology of all species showing the percentage of trees showing release and the number of newly established trees (pooled across all species) at 5-year intervals. Periods comprise values exceeding 10% of trees showing release (red dashed line) and/or more than 10 newly established trees (green dashed line). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 1 Relationship between reconstructed disturbance chronology (i.e. sum of all detected releases) and climate variables. R = the Pearson correlation coefficient between proportion of trees showing release and individual climatic variables; p = significance level of individual correlations, significant relationships in bold. Releases vs.

R

P

min PDSI max PDSI mean temprature min temperature max temperature mean temprature Min precipitation Max precipitation Mean precipitation Max wind speed

−0.11 −0.04 −0.04 0.10 0.15 0.04 −0.33 −0.20 −0.23 0.78

0.50 0.79 0.51 0.67 0.53 0.85 0.15 0.40 0.32 0.04

centage of trees showing release varied substantially over time. A higher proportion of release events was identified for four periods: 1775–1784, 1815–1839, 1880–1909 and 1950–1979. Altogether, 80% of all release events were detected in these four periods, covering less than half (42%) of the investigated five-year intervals. A comparison between the conservative and more liberal criteria showed very similar patterns (R = 0.8) of release reconstruction (Appendix S2). Moreover, 56% of moderate releases (Table S3) belong to canopy dominants of Quercus mongolica, which are less sensitive to disturbances (see Nowacki and Abrams, 1997). Based on this, the results of more liberal radial growth averaging criteria (Nowacki and Abrams, 1997) were used for further analyses, which enabled the reconstruction of releases for less sensitive canopy trees and, at the same time, did not identify different disturbance patterns when compared with the conservative criteria. 3.2. Typhoon intensity induces releases Correlation between disturbance chronology (i.e. release reconstruction) and maximum wind speed during the typhoons (Table 1, Fig. S3) showed a high, significant relationship (R = 0.78, p = 0.04). All other climate variables (PDSI, temperature, precipitation) showed insignificant relationships with disturbance chronology (Table 1).

3.3. Establishment reconstruction The comparison between establishment reconstructions for different groups of trees (whose core reaches the pith and trees with extrapolated age) showed high correlation (Appendix S3). Hence, the multi-step analysis used here for establishment reconstruction was validated and used for all further analyses. The age distribution pattern for all trees confirmed the uneven-aged condition of the forest (Fig. 1, Fig. S2, S4). The oldest tree, Quercus mongolica, established in 1660. There are obvious differences in the intensity of tree recruitment between individual species during the past 340 years (Fig. S4). Quercus mongolica mostly regenerated during the period 1660–1855, peaking in 1775–1800, and ceasing during the second half of the 19th century. Abies sachalinensis appeared in 1768 and regenerated regularly, with two peaks in 1915–1925 and 1955–1965, with no single establishment between 1930 and 1955. There were three relatively abundant maples (Acer mono, A. japonica, and A. pictum) with the oldest A. mono established in 1678, followed by A. japonica established in 1705 and A. pictum in 1823. Acer mono regenerated constantly from 1780 to 1860 and sporadically during 1865–1895, followed by a recruitment peak in 1900–1915, and ceased after 1925. Acer japonica regenerated mainly in two periods, 1900–1940 and 1965–1975, and rarely during the rest of the time. Similarly, Acer pictum regenerated sporadically with the exception of the period when its recruitment peak was reached. The remaining less-frequent species established principally after 1860, with combined recruitment peaks in 1925–1930 and 1970–1980. Taking all species together, the rate of tree recruitment was noticeably higher during three periods: 1775–1814, 1900–1939, 1955–1979 and a smaller peak in recruitment was detected in the period 1830–1834 (Fig. 1). During these periods (i.e. in 48% of the investigated five-year intervals), 85% of the present trees were established. 3.4. Periods with increased growth releases and tree recruitment Four time periods were defined on the basis of the results from release and establishment reconstruction. The periods are related to an increased percentage of detected disturbances and/or an increased number of establishments. Specifically, we defined

6

J. Altman et al. / Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13

Fig. 2. In the release period 1950–1979, cross pair correlation functions describing associations of: (A) trees showing no release to trees showing a major release, (B) trees showing no release to trees showing a moderate release and (C) trees showing a major release to trees showing a moderate release. Observed pattern is shown by solid line, dashed line denotes the mean of confidence envelope and grey area corresponds to confidence envelope (solid line outside confidence envelope denotes a significant difference from the inhomogeneous random pattern). p corresponds to the goodness of fit (Loosmore and Ford, 2006).

the individual periods as an interval with more than 10% of trees showing release and/or more than 10 newly established trees. The duration of these individual periods were: Period 1 (1775–1814), Period 2 (1820–1834), Period 3 (1880–1939) and Period 4 (1950–1979). It was found that the peak in tree recruitment followed the peak in disturbances (expressed as the percentage of trees showing release) with a 15-year delay (1890–1990: R = 0.49; 1760–1990: R = 0.15; Fig. S5). 3.5. Spatial analysis Only the last two periods of release events (i.e. 1880–1939, and 1950–1979) were used for spatial analysis as they had a sufficient number of released and newly-established trees. Trees showing a major and a moderate release were mostly randomly located in the plot (Fig. S7A–D) and trees showing no release were clumped (Fig. S7E and F). Trees showing no release were mostly segregated from the trees showing a major release by up to 15 m in the period 1880–1939 (Fig. S8A) and mostly randomly associated, with strong aggregation on <1 m scale, in the last period 1950–1979 (Fig. 2A). On the contrary, trees without any release were aggregated with

Fig. 3. In the last release period (1950–1979) (A) the mean number of newly established trees up to a given distance from all types of release computed by the cross K function. Cross pair correlation functions describing associations of: (B) newly established trees to trees showing a major release, (C) newly established trees to trees showing a moderate release and (D) newly established trees to trees showing no release. Observed pattern is shown by a solid line, dashed line denotes the mean of confidence envelope and the grey area corresponds to confidence envelope (solid line outside confidence envelope denotes a significant difference from the inhomogeneous random pattern). p corresponds to the goodness of fit (Loosmore and Ford, 2006).

respect to the trees with moderate release for most of the scales in both periods (Fig. 2B, Fig. S8B). There was no significant association between trees showing moderate and major releases (Fig. 2C, Fig. S8C). Trees showing a moderate release had the highest DBH in both periods (Mdn = 49 cm and Mdn = 33 cm in periods 1880–1939 and 1950–1979, respectively), trees with major release had smaller DBH (Mdn = 33 and 19) and trees showing no release had the smallest (Mdn = 5 and 6) DBH (Fig. S9). Establishments were strongly clumped together (Fig. S7G and H). The mean number of newly established trees within a given distance was higher for trees experiencing moderate or no release than for trees showing major release (Fig. 3A, Fig. S10A). The associations of established trees to released trees were also similar for the last two periods (Fig. 3, Fig. S10) and establishment was mostly aggregated to trees without any release and to trees with moderate release. On the other hand, establishment was mostly segregated from trees showing major release in the period 1880–1909 (Fig. S10B) and varying, mostly random, associations were observed in the last period (1950–1984), although a significant segregated pattern was observed for the limited 4–15 m scale (p = 0.02) in the last period too (Fig. 3B). Association of newly established trees of Acer japonica, A. mono, Abies sachalinensis and Acanthopanax sciadophylloides to different release levels were the same as the overall trends

J. Altman et al. / Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13

7

Fig. 4. In the release period 1880–1939, cross pair correlation functions describing associations of newly established trees of individual species (in the rows) to trees showing a major, a moderate and no release (in the columns). Observed pattern is shown by a solid line, dashed line denotes mean of confidence envelope and the grey area corresponds to confidence envelope (solid line outside confidence envelope denotes a significant difference from the inhomogeneous random pattern). p corresponds to the goodness of fit (Loosmore and Ford, 2006).

8

J. Altman et al. / Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13

described above and were similar for the last two periods (Fig. 4, Fig. S11). Establishment of Acer japonica, A.mono, Abies sachalinensis and Acanthopanax sciadophylloides showed strongest associations to individuals not showing release, aggregated or no significant association to trees showing moderate release, and mostly no association or segregation to trees showing major release. Contrary to these species, Betula ermanii dominantly showed no associations to all three release categories (Fig. 4). The SN and SD of newly established trees were mostly higher around the trees showing moderate release or no release than around trees showing major release in both periods (Fig. 5, Fig. S12). Trees without release had similar trends of species number and diversity around them as those showing moderate release. 3.6. Understory bamboo The coverage of dwarf bamboo, as expressed by leaf area index, was significantly higher (p = 0.004, W = 25) in the large gaps than beneath the undisturbed closed canopy (Fig. S13). The lowest

leaf area index was beneath the closed canopy (Mdn = 1.9 m2 /m2 ), followed by the partly disturbed canopy (Mdn = 2.5 m2 /m2 ). In contrast, the coverage of dwarf bamboo in the gap without a tree canopy was the highest (Mdn = 4.2 m2 /m2 ). 4. Discussion 4.1. Disturbance and establishment reconstruction Four distinct disturbance events that resulted in increased radial growth and recruitment for windows of around 25 years were detected. There are no existing long-term instrumental or historical data on past disturbance events in the study area that can be used to validate our long-term disturbance reconstruction. However, comparison of our disturbance chronology with typhoon intensity proved that typhoons are the dominant disturbance agent here and play an essential role in shaping forest structure and species composition of temperate forests in Japan (e.g. Bellingham et al., 1996; Nakashizuka and Iida, 1995). Moreover, the last disturbance

Fig. 5. Difference in the number of species (the first column) and diversity, i.e. 1-Simpson index (the second column), of newly established trees in the last period (1950–1979). (A, B) trees showing a moderate release minus trees showing a major release, (C, D) trees showing a moderate release minus trees showing no release and (E, F) trees showing a major release minus trees showing no release. For example, if the difference between diversity of trees showing a moderate release and trees showing a major release is higher than zero then there was higher diversity around trees with the moderate release than around trees with the major release. Observed difference is shown by a solid line, dashed line denotes the mean of confidence envelope and the grey areas correspond to confidence envelopes (solid line outside the confidence envelope denotes a significant difference from the random pattern). p corresponds to the goodness of fit (Loosmore and Ford, 2006).

J. Altman et al. / Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13

9

Fig. 6. Importance and frequency of typhoon induced disturbances of mixed severities in long-term forest dynamics in mixed forest with understory dwarf bamboo.

peak in the mid-1950s coincided with typhoon Marie, which struck Hokkaido in September 1954 with a maximum wind speed >30 m/s and caused wind-induced damage over 16,108 ha of Uryu Experimental Forest (Ishizuka et al., 1997; Yoshida and Noguchi, 2009), where our study was located. Individual peaks in the reconstructed disturbance events were separated by intervals of ∼50–70 years with lower level of disturbances. The reconstructed peaks indicate that the study area was impacted by relatively infrequent, but severe disturbances (e.g. Altman et al., 2013b; Svoboda et al., 2014; Zielonka et al., 2010). These findings agree with information suggesting that only a few typhoons affect the Hokkaido region each year, and most have a low impact on forests (Toda et al., 2011). Most of the reconstructed severe disturbances were followed by an increased level of tree establishment. In the third period, two subsequent peaks occurred and only the second peak (1900–1904) was followed by an increased level of tree establishment. The first peak (1880–1889) could be associated with the 1883 eruption of Krakatoa, which caused significant cooling in the northern hemisphere (e.g. Briffa et al., 1998; Robock, 2005); significant cooling can lead to forest dieback (Kullman, 1989; Mueller-Dombois, 1987; Rampino and Ambrose, 2000). Differences in the number of newly established trees after individual disturbance events can be explained by variation in lifespans of individual species. This is shown by the higher occurrence of establishment after the last two severe disturbances. In contrast, the first detected severe disturbance (230 years ago) was followed by a lower number of newly established trees, because both the density-dependent mortality has already occurred and short-lived

trees have died. The same process can explain the only slight increase in establishment after the detected severe disturbance which occurred in the first half of 19th century. Short-lived species could not have survived until the present time and any space for the new establishment of long-lived species was mostly occupied from the previous event. Most species showed regeneration peak(s) during the 20th century. The only exception was the recruitment of Mongolian oak (Quercus mongolica), which was concentrated at the end of the 18th century. An obvious decrease in oak regeneration beginning at the second half of 19th century was also found, while its regeneration from 1660 to 1855 was more or less constant. The same decline was found by Abrams et al. (1999) in their study from Hokkaido, around 200 km from this study site. Declining oak dominance (various species) and replacement by shade-tolerant species has been reported in studies from North America and from Europe (e.g. Abrams, 2003; Altman et al., 2013b; Gomez-Aparicio et al., 2008; McEwan et al., 2011). Oak decline is mostly connected with decreased understory light levels, natural disturbance regime changes, traditional management abandonment, increased animal browsing and increasing drought. Specifically, forests dominated by Quercus mongolica, and their regeneration pattern, appear to be influenced by the first three factors (Abrams et al., 1999; Altman et al., 2013a; Barnes et al., 1992). 4.2. Spatial-temporal analyses of forest dynamics Observed compositional turnover is not spatio-temporally random. One of the most important factors affecting tree regeneration

10

J. Altman et al. / Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13

and forest dynamics is the size of the disturbed area (Li et al., 2005; Naaf and Wulf, 2007; Pederson et al., 2014) and size of the species pool (Lawes and Obiri, 2003). Large infrequent disturbances comprise areas affected by different disturbance intensities and have the potential to generate more heterogeneity than small disturbances (Foster et al., 1998; Turner and Dale, 1998). Although historical gap sizes cannot be determined directly, our results suggest a mixed severity disturbance regime as large infrequent disturbances can create both larger gaps by blow-down of groups of neighbouring trees and small gaps by single tree falls. This pattern is documented in this study, not only in the detection of moderate and major releases, but also in the spatial pattern analyses. Trees showing moderate release are aggregated with respect to trees unaffected by disturbance, which indicates that there is not a large gap around a tree with moderate release. On the contrary, trees showing major release were segregated from the trees which were not affected by disturbance, which indicates a large gap around them. Similarly, Pederson et al. (2014) identified major releases corresponding with historical large canopy openings. The majority of moderate releases (induced by the formation of small gaps) is formed by severe disturbances, while the small disturbances detected between the four main disturbance events represented only a minor proportion of the small gaps. Similarly, Altman et al. (2013a) found that on the Korean peninsula, up to 85% of the detected moderate releases accompanied major releases, which mirrored the most intensive typhoons. Spatio-temporal analyses of forest dynamics revealed differences in density and diversity of newly established trees in relation to disturbance severity levels (no release vs. major vs. moderate releases) and hence the resulting gap size. Our spatiotemporal reconstruction shows that small gaps and undisturbed areas generally represent the most suitable environment for forest regeneration in terms of both the level of establishments and its diversity (Gray and Spies, 1996; Kern et al., 2013; Wang et al., 2006), although there are some interspecific differences. From the five most common species tested for habitat preference, Acer japonica, A. mono, Abies sachalinensis and Acanthopanax sciadophylloides regenerated highly preferentially beneath the undisturbed canopy. A. mono and A. japonica also regenerated successfully in small gaps (at least in last period). Moreover, A. mono and Acanthopanax sciadophylloides were segregated from trees showing major release, while A. japonica and Abies sachalinensis showed mostly a random pattern. Only Betula ermanii did not show any significant regeneration preference related to canopy openness. Betula ermanii is a light-demanding early-successional species, which invades open lands created by disturbance (Hara et al., 1991). Our results thus confirm previous findings of rapid height growth in this species (Hiura et al., 1996) and overtops other plants in the early stages of succession (Hara et al., 1990). The methodological approach used here thus proves to be a good tool for identification of the lifestrategy for individual species. However, this should be tested in the future across various forest ecosystems with high replication for individual species. Based on our findings it could be assumed that small gaps or forests without stand disturbance are ideal for tree regeneration. This raises the question: Is increased forest regeneration really a consequence of disturbance when trees in closed canopies (more precisely trees without detected disturbance) showed an increased level of establishment? The answer is hidden in the understory of the forest, dominated by dwarf bamboo. Bamboo cover is highest beneath the open canopy and decreases with increasing canopy closure. Bamboo’s ability to dominate forest gaps is observed across different forests from temperate to tropical regions (Larpkern et al., 2011; Montti et al., 2011; Taylor et al., 2004). Bamboo thus acts as a biotic filter, restricting tree-seedling regeneration in large gaps due to extremely dense cover and thick layers of its slowly decay-

ing litter. Hiura et al. (1996) found seedling establishment was restricted (98%) to logs and mounds, where the effect of dwarf bamboo was limited. The importance of logs as sites for regeneration is widely recognized across different types of forest (Hiura et al., 1996; Takahashi, 1994; Taylor, 1990; Taylor and Zisheng, 1988; Veblen et al., 1981). In the forests considered here fallen trees and branches could create conditions for forest regeneration by pushing down bamboo shoots as documented by Wang et al. (2009). Trees showing a moderate release were mostly large trees while trees showing no release were smaller sub-canopy trees. Positive association of these two release (and size) categories implies that establishment in the past was enhanced via fallen branches of large trees and consequent bamboo suppression. Our findings thus indicate that sub-canopy losses (branches) can play an important role in forest dynamics; however, disturbance research requires additional studies to account for the role of this factor in forest dynamics (Asner, 2013; Kellner et al., 2011). It seems that disturbances induced by volcanic eruptions, or more precisely by consequent cooling of surface temperatures across large areas, decrease the competition for surviving trees (i.e. growth release), but do not produce fallen branches or stems and thus did not create conditions for subsequent establishment. Another important feature of most bamboo species relating to forest regeneration is their life cycle. They often flower and then die (i.e. monocarpic species) simultaneously over a wide area after relatively long periods (3–120 years) of vegetative growth (Janzen, 1976). Several studies have confirmed that bamboo dieback induces tree germination, regeneration or synchronous tree seedling establishment (e.g. Abe et al., 2002; Giordano et al., 2009; Taylor et al., 2004). Although most such studies demonstrate a positive effect of bamboo dieback on forest regeneration, no evidence of increased tree establishment unconnected with disturbance events was found here (i.e. regeneration peak without a preceding peak in the disturbance reconstruction). There is only a small probability that dieback is synchronous with a disturbance event and no similar relationship has been documented so far. A more probable scenario is that bamboo dieback provides a narrow window of increased resource (mainly light) availability, and only some already established individuals can take advantage of these increasing resources (Montti et al., 2011). Consequently, disturbance events seem to play a dominant role for forest regeneration in the long-term, while bamboo dieback has a short-term effect. 4.3. Forest perspective The globally observed trend of increasing intensity of tropical cyclones over the last few decades is well documented (e.g. Bender et al., 2010; Elsner et al., 2008; Holland and Bruyere, 2014; Park et al., 2014) and a continuing increase is predicted (Grinsted et al., 2013; Knutson et al., 2010; Murakami et al., 2013). Furthermore, recent research has identified a pronounced poleward migration in the average latitude at which tropical cyclones have achieved their lifetime-maximum intensity (Kossin et al., 2014). Our results suggest that severe disturbances play a fundamental role in forest development and help to maintain tree diversity (Fig. 6). The last severe disturbance generated a predominance of large openings, while in the previous three events, the creation of small and large gaps was equal. The higher occurrence of large gaps can be attributed to the senescent canopy dominated by old oaks, the increasing typhoon intensity or a combination of both. It is suggested here that the increasing intensity of tropical cyclones creates a higher number of large gaps at the expense of smaller ones. Moreover, existing forest gaps of all sizes are predisposed to be enlarged by future severe disturbance (Foster and Reiners, 1986; Jansen et al., 2008; Kubo et al., 1996). These further changes can significantly affect forest structure and species diversity (Catford et al.,

J. Altman et al. / Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13

2012). Our findings thus may be relevant to other cyclone-affected forested regions of the world with changing tropical cyclone intensity. 5. Conclusions The effects of gap size on long-term forest dynamics have been inferred from snapshot data (e.g. Green, 1996; Runkle, 1981; Yamamoto, 1992) or short-term experiments (e.g. Coomes and Allen, 2007; Gray et al., 2012; Nuttle et al., 2013). Whereas manipulative experiments are the only way to directly demonstrate causal relationships, their spatial and temporal extent is rather limited. Consequently, in this study, time-series data derived from treering reconstruction were combined with spatial pattern analysis to provide new insights on forest dynamics over a few centuries. Such an approach took advantage of the combination of high temporal resolution dendrochronological analyses with high spatial resolution forest maps of individual trees. This linked forest regeneration patterns with reconstructed disturbance severities derived from tree rings. The main findings were that forest regeneration is strongly connected to severe disturbances, which results in the creation of both large and small openings. Only the latter supports diverse forest regeneration. Large openings, on the other hand, suppress regeneration due to the development of an extremely dense bamboo understory. Our results therefore stress the importance of combining temporal (dendrochronological) and spatial (tree positions) analyses. Based only on temporal reconstruction of disturbances and regeneration, it might be concluded that most of the regeneration was related to a few severe disturbances and the creation of large gaps. In contrast, based only on spatial analysis of associations between establishment and old generation trees, it might be concluded that an undisturbed canopy supports most of the regeneration. Hence, using only one of these approaches can lead to erroneous or incomplete inferences about the mechanisms driving forest dynamics. Acknowledgements We thank editors Dr. George L. W. Perry and Dr. Kirk A. Moloney, anonymous reviewers, and Dr. Thomas A. Nagel, Dr. Miroslav Svoboda, Dr. Tomasz Zielonka for thoughtful and constructive reviews of this manuscript. The study was funded by research grants 14-12262S, P504/12/1952, 13-13368S, 14-04258S and 1618022S of the Grant Agency of the Czech Republic, long-term research development project no. RVO 67985939, the Grant for Joint Research Program of the Institute of Low Temperature Science, Hokkaido University. We thank the staff of Forest Research Station, Field Science Center for Northern Biosphere, Hokkaido University, for long-term data collection. Dr. Brian George McMillan kindly improved our English. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ppees.2016.04. 003. References Abe, M., Izaki, J., Miguchi, H., Masaki, T., Makita, A., Nakashizuka, T., 2002. The effects of Sasa and canopy gap formation on tree regeneration in an old beech forest. J. Veg. Sci. 13, 565–574. Abrams, M.D., Copenheaver, C.A., Terazawa, K., Umeki, K., Takiya, M., Akashi, N., 1999. A 370-year dendroecological history of an old-growth Abies-Acer-Quercus forest in Hokkaido northern Japan. Can. J. For. Res. 29, 1891–1899. Abrams, M.D., 2003. Where has all the white oak gone? Bioscience 53, 927–939.

11

Albanesi, E., Gugliotta, O.I., Mercurio, I., Mercurio, R., 2008. Effects of gap size and within-gap position on seedlings establishment in silver fir stands. IFOREST 1, 55–59. Altman, J., Dolezal, J., Cerny, T., Song, J.S., 2013a. Forest response to increasing typhoon activity on the Korean peninsula: evidence from oak tree-rings. Global Change Biol. 19, 498–504. Altman, J., Hedl, R., Szabo, P., Mazurek, P., Riedl, V., Mullerova, J., Kopecky, M., Dolezal, J., 2013b. Tree-rings mirror management legacy: dramatic response of standard oaks to past coppicing in central Europe. PLoS One 8, e55770. Altman, J., Fibich, P., Dolezal, J., Aakala, T., 2014. TRADER: a package for tree ring analysis of disturbance events in R. Dendrochronologia 32, 107–112. Asner, G.P., 2013. Geography of forest disturbance. Proc. Natl. Acad. Sci. U. S. A. 110, 3711–3712. Baddeley, A., Turner, R., 2005. spatstat: an R package for analyzing spatial point patterns. J. Stat. Softw. 12, 1–42. Baret, S., Cournac, L., Thebaud, C., Edwards, P., Strasberg, D., 2008. Effects of canopy gap size on recruitment and invasion of the non-indigenous Rubus alceifolius in lowland tropical rain forest on Reunion. J. Trop. Ecol. 24, 337–345. Barnes, B.V., Xü, Z., Zhao, S., 1992. Forest ecosystems in an old-growth pine–mixed hardwood forest of the Changbai Shan Preserve in northeastern China. Can. J. For. Res 22, 144–160. Bellingham, P.J., Kohyama, T., Aiba, S., 1996. The effects of a typhoon on Japanese warm temperate rainforests. Ecol. Res. 11, 229–247. Bender, M.A., Knutson, T.R., Tuleya, R.E., Sirutis, J.J., Vecchi, G.A., Garner, S.T., Held, I.M., 2010. Modeled impact of anthropogenic warming on the frequency of intense Atlantic hurricanes. Science 327, 454–458. Black, B.A., Abrams, M.D., 2003. Use of boundary-line growth patterns as a basis for dendroecological release criteria. Ecol. Appl. 13, 1733–1749. Briffa, K.R., Jones, P.D., Schweingruber, F.H., Osborn, T.J., 1998. Influence of volcanic eruptions on Northern Hemisphere summer temperature over the past 600 years. Nature 393, 450–455. Burnham, K.M., Lee, T.D., 2010. Canopy gaps facilitate establishment, growth, and reproduction of invasive Frangula alnus in a Tsuga canadensis dominated forest. Biol. Invasions 12, 1509–1520. Catford, J.A., Daehler, C.C., Murphy, H.T., Sheppard, A.W., Hardesty, B.D., Westcott, D.A., Rejmanek, M., Bellingham, P.J., Pergl, J., Horvitz, C.C., Hulme, P.E., 2012. The intermediate disturbance hypothesis and plant invasions: implications for species richness and management. Perspect. Plant Ecol. Evol. Syst. 14, 231–241. Chambers, J.Q., Negron-Juarez, R.I., Marra, D.M., Di Vittorio, A., Tews, J., Roberts, D., Ribeiro, G., Trumbore, S.E., Higuchi, N., 2013. The steady-state mosaic of disturbance and succession across an old-growth Central Amazon forest landscape. Proc. Natl. Acad. Sci. U. S. A. 110, 3949–3954. Cook, E.R., Anchukaitis, K.J., Buckley, B.M., D’Arrigo, R.D., Jacoby, G.C., Wright, W.E., 2010. Asian monsoon failure and megadrought during the last millennium. Science 328, 486–489. Coomes, D.A., Allen, R.B., 2007. Mortality and tree-size distributions in natural mixed-age forests. J. Ecol. 95, 27–40. Coumou, D., Rahmstorf, S., 2012. A decade of weather extremes. Nat. Climate Change 2, 491–496. Diggle, P., 2003. Statistical Analysis of Spatial Point Pattern, 2nd ed. Hodder Arnold, London, UK. Dolezal, J., Matsuki, S., Hara, T., 2009. Effects of dwarf-bamboo understory on tree seedling emergence and survival in a mixed-oak forest in northern Japan: a multi-site experimental study. Community Ecol. 10, 225–235. Duncan, R.P., Stewart, G.H., 1991. The temporal and spatial analysis of tree age distributions. Can. J. For. Res. 21, 1703–1710. Eckstein, D., Bauch, J., 1969. Beitrag zur Rationalisierung eines dendrochronologischen Verfahrens und zur Analyse seiner Aussagesicherheit. Forstwiss. Centralbl. 88, 230–250. Elsner, J.B., Kossin, J.P., Jagger, T.H., 2008. The increasing intensity of the strongest tropical cyclones. Nature 455, 92–95. Emanuel, K.A., 2013. Downscaling CMIP5 climate models shows increased tropical cyclone activity over the 21st century. Proc. Natl. Acad. Sci. U. S. A. 110, 12219–12224. Everham, E.M., Brokaw, N.V.L., 1996. Forest damage and recovery from catastrophic wind. Bot. Rev. 62, 113–185. ´ V., Klimeˇs, P., Tˇesˇ itel, J., Molem, K., Damas, K., Weiblen, Fibich, P., Lepˇs, J., Novotny, G.D., 2016. Spatial patterns of tree species distribution in New Guinea primary and secondary lowland rain forest. J. Veg. Sci. 27, 328–339. Fischer, A., Marshall, P., Camp, A., 2013. Disturbances in deciduous temperate forest ecosystems of the northern hemisphere: their effects on both recent and future forest development. Biodivers. Conserv. 22, 1863–1893. Foster, J.R., Reiners, W.A., 1986. Size distribution and expansion of canopy gaps in a northern Appalachian spruce-fir forest. Vegetation 68, 109–114. Foster, D.R., Knight, D.H., Franklin, J.F., 1998. Landscape patterns and legacies resulting from large, infrequent forest disturbances. Ecosystems 1, 497–510. Fraver, S., White, A.S., 2005. Identifying growth releases in dendrochronological studies of forest disturbance. Can. J. For. Res. 35, 1648–1656. Fraver, S., White, A.S., Seymour, R.S., 2009. Natural Disturbance in an Old-growth Landscape of Northern Maine. U. S. A. J. Ecol. 97, 289–298. Frelich, L.E., 2002. Forest Dynamics and Disturbance Regimes: Studies from Temperate Evergreen-Deciduous Forests. Cambridge University Press, Cambridge. Furuta, K., 1989. A comparison of endemic and epidemic populations of the spruce beetle (Ips typographus japonicus Niijima) in Hokkaido. J. Appl. Entomol. 107, 289–295.

12

J. Altman et al. / Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13

Giordano, C.V., Sanchez, R.A., Austin, A.T., 2009. Gregarious bamboo flowering opens a window of opportunity for regeneration in a temperate forest of Patagonia. New Phytol. 181, 880–889. Gomez-Aparicio, L., Perez-Ramos, I.M., Mendoza, I., Matias, L., Quero, J.L., Castro, J., Zamora, R., Maranon, T., 2008. Oak seedling survival and growth along resource gradients in Mediterranean forests: implications for regeneration in current and future environmental scenarios. Oikos 117, 1683–1699. Goreaud, F., Pélissier, R., 2003. Avoiding misinterpretation of biotic interactions with the intertype K12-function: population independence vs. random labelling hypotheses. J. Veg. Sci. 14, 681–692. Gravel, D., Canham, C.D., Beaudet, M., Messier, C., 2010. Shade tolerance, canopy gaps and mechanisims of coexistence of forest trees. Oikos 119, 475–484. Gray, A.N., Spies, T.A., 1996. Gap size, within-gap position and canopy structure effects on conifer seedling establishment. J. Ecol. 84, 635–645. Gray, A.N., Spies, T.A., Pabst, R.J., 2012. Canopy gaps affect long-term patterns of tree growth and mortality in mature and old-growth forests in the Pacific Northwest. For. Ecol. Manage. 281, 111–120. Green, P.T., 1996. Canopy gaps in rain forest on Christmas Island, Indian ocean: size distribution and methods of measurement. J. Trop. Ecol. 12, 427–434. Grinsted, A., Moore, J.C., Jevrejeva, S., 2013. Projected Atlantic hurricane surge threat from rising temperatures. Proc. Natl. Acad. Sci. U. S. A. 110, 5369–5373. Hara, T., Vanrijnberk, H., During, H., Yokozawa, M., Kikuzawa, K., 1990. Competition process and spatial pattern formation in a Betula ermanii population. Spat. Proc. Plant Com., 127–143. Hara, T., Kimura, M., Kikuzawa, K., 1991. Growth-patterns of tree height and stem diameter in populations of Abies veitchii: A. mariesii and Betula ermanii. J. Ecol. 79, 1085–1098. Hiura, T., Sano, J., Konno, Y., 1996. Age structure and response to fine-scale disturbances of Abies sachalinensis Picea jezoensis, Picea glehnii, and Betula ermanii growing under the influence of a dwarf bamboo understory in northern Japan. Can. J. For. Res. 26, 289–297. Holland, G., Bruyere, C.L., 2014. Recent intense hurricane response to global climate change. Climate Dynam. 42, 617–627. Illian, J., Penttinen, A., Stoyan, H., Stoyan, D., 2008. Statistical Analysis and Modelling of Spatial Point Patterns. John Wiley & Sons, Chichester. Ishizuka, M., Toyooka, H., Osawa, A., Kushima, H., Kanazawa, Y., Sato, A., 1997. Secondary succession following catastrophic windthrow in a boreal forest in Hokkaido. Jpn. J. Sustain. For. 6, 367–388. Jansen, P.A., Van Der Meer, P.J., Bongers, F., 2008. Spatial contagiousness of canopy disturbance in tropical rain forest: an individual-tree-based test. Ecology 89, 3490–3502. Janzen, D.H., 1976. Why bamboos wait so long to flower. Annu. Rev. Ecol. Syst. 7, 347–391. Kamata, N., 2002. Outbreaks of forest defoliating insects in Japan, 1950–2000. Bull. Entomol. Res. 92, 109–117. Kellner, J.R., Asner, G.P., Vitousek, P.M., Tweiten, M.A., Hotchkiss, S., Chadwick, O.A., 2011. Dependence of forest structure and dynamics on substrate age and ecosystem development. Ecosystems 14, 1156–1167. Kern, C.C., D’Arnato, A.W., Strong, T.F., 2013. Diversifying the composition and structure of managed, late-successional forests with harvest gaps: what is the optimal gap size? For. Ecol. Manage. 304, 110–120. Khakimulina, T., Fraver, S., Drobyshev, I., 2015. Mixed-severity natural disturbance regime dominates in an old-growth Norway spruce forest of northwest Russia. J. Veg. Sci. 27, 400–413. Knutson, T.R., McBride, J.L., Chan, J., Emanuel, K., Holland, G., Landsea, C., Held, I., Kossin, J.P., Srivastava, A.K., Sugi, M., 2010. Tropical cyclones and climate change. Nat. Geosci. 3, 157–163. Kobe, R.K., Vriesendorp, C.F., 2011. Conspecific density dependence in seedlings varies with species shade tolerance in a wet tropical forest. Ecol. Lett. 14, 503–510. Kossin, J.P., Emanuel, K.A., Vecchi, G.A., 2014. The poleward migration of the location of tropical cyclone maximum intensity. Nature 509, 349–352. Kubo, T., Iwasa, Y., Furumoto, N., 1996. Forest spatial dynamics with gap expansion: total gap area and gap size distribution. J. Theor. Biol. 180, 229–246. Kuennecke, H.B., 2008. Temperate Forest Biomes. Greenwood Press, Westport. Kullman, L., 1989. Cold-induced dieback of montane spruce forests in the Swedish Scandes—a modern analog of paleoenvironmental processes. New Phytol. 113, 377–389. Larpkern, P., Moe, S.R., Totland, O., 2011. Bamboo dominance reduces tree regeneration in a disturbed tropical forest. Oecologia 165, 161–168. Law, R., Illian, J., Burslem, D.F.R.P., Gratzer, G., Gunatilleke, C.V.S., Gunatilleke, I.A.U.N., 2009. Ecological information from spatial patterns of plants: insights from point process theory. J. Ecol. 97, 616–628. Lawes, M.J., Obiri, J.A.F., 2003. Canopy gaps in subtropical forest in South Africa: size of the species pool and not the number of available niches limits species richness. J. Trop. Ecol. 19, 549–556. Li, Z.Q., Bogaert, J., Nijs, I., 2005. Gap pattern and colonization opportunities in plant communities: effects of species richness, mortality, and spatial aggregation. Ecography 28, 777–790. Loehle, C., 2000. Strategy space and the disturbance spectrum: a life-history model for tree species coexistence. Am. Nat. 156, 14–33. Loosmore, N.B., Ford, E.D., 2006. Statistical inference using the G or K point pattern spatial statistics. Ecology 87, 1925–1931. Lorimer, C.G., Frelich, L.E., 1989. A methodology for estimating canopy disturbance frequency and intensity in dense temperate forests. Can. J. For. Res. 19, 651–663.

Lorimer, C.G., 1989. Relative effects of small and large disturbances on temperate hardwood forest structure. Ecology 70, 565–567. Masaka, K., Ohno, Y., Yamada, K., 2000. Fire tolerance and the fire-related sprouting characteristics of two cool-temperate broad-leaved tree species. Ann. Bot. 85, 137–142. McEwan, R.W., Dyer, J.M., Pederson, N., 2011. Multiple interacting ecosystem drivers: toward an encompassing hypothesis of oak forest dynamics across eastern North America. Ecography 34, 244–256. Montti, L., Campanello, P.I., Genoveva, M.G., Blundo, C., Austin, A.T., Sala, O.E., Goldstein, G., 2011. Understory bamboo flowering provides a very narrow light window of opportunity for canopy-tree recruitment in a neotropical forest of Misiones. Argent. For. Ecol. Manage. 262, 1360–1369. Mori, A., Takeda, H., 2004. Effects of undisturbed canopy structure on population structure and species coexistence in an old-growth subalpine forest in central Japan. For. Ecol. Manage. 200, 89–100. Mueller-Dombois, D., 1987. Natural dieback in forests. Bioscience 37, 575–583. Murakami, H., Wang, B., Li, T., Kitoh, A., 2013. Projected increase in tropical cyclones near Hawaii. Nat. Clim. Change 3, 749–754. Naaf, T., Wulf, M., 2007. Effects of gap size: light and herbivory on the herb layer vegetation in European beech forest gaps. For. Ecol. Manage. 244, 141–149. Nakashizuka, T., Iida, S., 1995. Composition, dynamics and disturbance regime of temperate deciduous forests in Monsoon Asia. Vegetation 121, 23–30. Noguchi, M., Yoshida, T., 2005. Factors influencing the distribution of two co-occurring dwarf bamboo species (Sasa kurilensis and S. senanensis) in a conifer-broad leaved mixed stand in northern Hokkaido. Ecol. Res. 20, 25–30. Nowacki, G.J., Abrams, M.D., 1997. Radial-growth averaging criteria for reconstructing disturbance histories from presettlement-origin oaks. Ecol. Monogr. 67, 225–249. Nuttle, T., Royo, A.A., Adams, M.B., Carson, W.P., 2013. Historic disturbance regimes promote tree diversity only under low browsing regimes in eastern deciduous forest. Ecol. Monogr. 83, 3–17. Odion, D.C., Hanson, C.T., Arsenault, A., Baker, W.L., DellaSala, D.A., Hutto, R.L., Klenner, W., Moritz, M.A., Sherriff, R.L., Veblen, T.T., Williams, M.A., 2014. Examining historical and current mixed-severity fire regimes in ponderosa pine and mixed-conifer forests of western North America. PLoS One 9, e87852. Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R., Simpson, G.L., Solymos, M., Stevens, H.H., Wanger, H., 2013. vegan: Community Ecology Package, R package version 2. 0-10, . Olson, D.M., Dinerstein, E., Wikramanayake, E.D., Burgess, N.D., Powell, G.V.N., Underwood, E.C., D’Amico, J.A., Itoua, I., Strand, H.E., Morrison, J.C., Loucks, C.J., Allnutt, T.F., Ricketts, T.H., Kura, Y., Lamoreux, J.F., Wettengel, W.W., Hedao, P., Kassem, K.R., 2001. Terrestrial ecoregions of the worlds: a new map of life on Earth. Bioscience 51, 933–938. Papaik, M.J., Canham, C.D., 2006. Species resistance and community response to wind disturbance regimes in northern temperate forests. J. Ecol. 94, 1011–1026. Park, D.S.R., Ho, C.H., Kim, J.H., Kim, H.S., 2011. Strong landfall typhoons in Korea and Japan in a recent decade. J. Geophys. Res-Atmos. 116, 11. Park, D.-S.R., Ho, C.-H., Kim, J.-H., 2014. Growing threat of intense tropical cyclones to East Asia over the period 1977–2010. Environ. Res. Lett., 9. Pederson, N., Dyer, J.M., McEwan, R.W., Hessl, A.E., Mock, C.J., Orwig, D.A., Rieder, H.E., Cook, B.I., 2014. The legacy of episodic climatic events in shaping temperate broadleaf forests. Ecol. Monogr. 84, 599–620. Perry, G.L.W., Miller, B.P., Enright, N.J., 2006. A comparison of methods for the statistical analysis of spatial point patterns in plant ecology. Plant Ecol. 187, 59–82. Peterken, G.F., 1996. Natural Woodland: Ecology and Conservation in Northern Temperate Regions. Cambridge University Press, Cambridge, England. Qin, X., Li, G., Wang, D., Liu, R., Yang, G., Feng, Y., Ren, G., 2011. Determinism versus chance in canopy gap herbaceous species assemblages in temperate Abies-Betula forests. For. Ecol. Manage. 262, 1138–1145. R Core Team, 2015. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Rampino, M.R., Ambrose, S.H., 2000. Volcanic winter in the Garden of Eden: the Toba supereruption and the late Pleistocene human population crash. In: McCoy, F.W., Heiken, G. (Eds.), Volcanic Hazards and Disasters in Human Antiquity. , pp. 71–101. Rentch, J.S., Fajvan, M.A., Hicks, R.R., 2003. Oak establishment and canopy accession strategies in five old-growth stands in the central hardwood forest region. For. Ecol. Manage. 184, 285–297. Ripley, B., 1987. Spatial point pattern analysis in ecology. In: Legendre, P., Legendre, L. (Eds.), Development in Numerical Ecology. Springer, Berlin, DE, pp. 407–429. Robock, A., 2005. Cooling following large volcanic eruptions corrected for the effect of diffuse radiation on tree rings. Geophys. Res. Lett. 32. Romme, W.H., Everham, E.H., Frelich, L.E., Moritz, M.A., Sparks, R.E., 1998. Are large, infrequent disturbances qualitatively different from small, frequent disturbances? Ecosystems 1, 524–534. Royo, A.A., Carson, W.P., 2006. On the formation of dense understory layers in forests worldwide: consequences and implications for forest dynamics biodiversity, and succession. Can. J. For. Res. 36, 1345–1362. Rozas, V., 2003. Tree age estimates in Fagus sylvatica and Quercus robur: testing previous and improved methods. Plant Ecol. 167, 193–212. Runkle, J.R., 1981. Gap regeneration in some old-growth forests of the eastern United States. Ecology 62, 1041–1051.

J. Altman et al. / Perspectives in Plant Ecology, Evolution and Systematics 21 (2016) 1–13 Samonil, P., Dolezelova, P., Vasickova, I., Adam, D., Valtera, M., Kral, K., Janik, D., Sebkova, B., 2013. Individual-based approach to the detection of disturbance history through spatial scales in a natural beech-dominated forest. J. Veg. Sci. 24, 1167–1184. Shimatani, I.K., Kubota, Y., 2011. The spatio-temporal forest patch dynamics inferred from the fine-scale synchronicity in growth chronology. J. Veg. Sci. 22, 334–345. Shimatani, K., 2001. Multivariate point processes and spatial variation of species diversity. For. Ecol. Manage. 142, 215–229. Splechtna, B.E., Gratzer, G., Black, B.A., 2005. Disturbance history of a European old-growth mixed-species forest—a spatial dendro-ecological analysis. J. Veg. Sci. 16, 511–522. Svoboda, M., Janda, P., Bace, R., Fraver, S., Nagel, T.A., Rejzek, J., Mikolas, M., Douda, J., Boublik, K., Samonil, P., Cada, V., Trotsiuk, V., Teodosiu, M., Bouriaud, O., Biris, A.I., Sykora, O., Uzel, P., Zelenka, J., Sedlak, V., Lehejcek, J., 2014. Landscape-level variability in historical disturbance in primary Picea abies mountain forests of the Eastern Carpathians. Rom. J. Veg. Sci. 25, 386–401. Tabarelli, M., Mantovani, W., 2000. Gap-phase regeneration in a tropical montane forest: the effects of gap structure and bamboo species. Plant Ecol. 148, 149–155. Takahashi, K., Mitsuishi, D., Uemura, S., Suzuki, J.I., Hara, T., 2003. Stand structure and dynamics during a 16-year period in a sub-boreal conifer-hardwood mixed forest northern Japan. For. Ecol. Manage. 174, 39–50. Takahashi, K., 1994. Effect of size structure, forest floor type and disturbance regime on tree species composition in a coniferous forest in Japan. J. Ecol. 82, 769–773. Tanaka, H., Nakashizuka, T., 1997. Fifteen years of canopy dynamics analyzed by aerial photographs in a temperate deciduous forest. Jpn. Ecol. 78, 612–620. Taylor, A.H., Zisheng, Q., 1988. Tree replacement patterns in subalpine Abies-Betula forests, wolong natural reserve, China. Vegetation 78, 141–149. Taylor, A.H., Huang, J.Y., Zhou, S.Q., 2004. Canopy tree development and undergrowth bamboo dynamics in old-growth Abies-Betula forests in Southwestern China: a 12-year study. For. Ecol. Manage 200, 347–360. Taylor, A.H., 1990. Disturbance and persistence of sitka spruce (Picea sitchensis (Bong) Carr.) in coastal forests of the Pacific Northwest. N. Am. J. Biogeogr. 17, 47–58. Tepley, A.J., Veblen, T.T., 2015. Spatiotemporal fire dynamics in mixed-conifer and aspen forests in the San Juan Mountains of southwestern Colorado. U. S. A. Ecol. Monogr. 85, 583–603. Toda, M., Kolari, P., Nakai, T., Kodama, Y., Shibata, H., Yoshida, T., Uemura, S., Sumida, A., Kato, K., Ono, K., Hara, T., 2011. Photosynthetic recovery of foliage after wind disturbance activates ecosystem CO2 uptake in cool temperate forests of northern Japan. J. Geophys. Res-Biogeogr. 116(G2.

13

Tomimatsu, H., Yamagishi, H., Tanaka, I., Sato, M., Kondo, R., Konno, Y., 2011. Consequences of forest fragmentation in an understory plant community: extensive range expansion of native dwarf bamboo. Plant Species Biol. 26, 3–12. 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 Ukrainian Carpathians. For. Ecol. Manage. 334, 144–153. Turner, M.G., Dale, V.H., 1998. Comparing large, infrequent disturbances: what have we learned? Ecosystems 1, 493–496. Umeki, K., Kikuzawa, K., 1999. Long-term growth dynamics of natural forests in Hokkaido, northern Japan. J. Veg. Sci. 10, 815–824. Valladares, F., Saldana, A., Gianoli, E., 2012. Costs versus risks: architectural changes with changing light quantity and quality in saplings of temperate rainforest trees of different shade tolerance. Aust. Ecol. 37, 35–43. Van Goethem, D., De Smedt, S., Valcke, R., Potters, G., Samson, R., 2013. Seasonal, diurnal and vertical variation of chlorophyll fluorescence on Phyllostachys humilis in Ireland. PLoS One 8. Veblen, T.T., Donoso, C., Schlegel, F.M., Escobar, B., 1981. Forest dynamics in south-central Chile. J. Biogeogr. 8, 211–247. Wang, W., Franklin, S.B., Ren, Y., Ouellette, J.R., 2006. Growth of bamboo Fargesia qinlingensis and regeneration of trees in a mixed hardwood-conifer forest in the Qinling Mountains. China For. Ecol. Manage. 234, 107–115. Wang, Y.J., Tao, J.P., Zhong, Z.C., 2009. Factors influencing the distribution and growth of dwarf bamboo, Fargesia nitida, in a subalpine forest in Wolong Nature Reserve, southwest China. Ecol. Res. 24, 1013–1021. White, P.S., 1983. Eastern Asian-Eastern North American floristic relations: the plant community level. Ann. Mo. Bot. Gard. 70, 734–747. Wiegand, T., Moloney, A.K., 2014. Handbook of Spatial Point-pattern Analysis in Ecology. Chapman and Hall/CRC Press. Wiegand, T., Gunatilleke, C.V.S., Gunatilleke, I., Huth, A., 2007. How individual species structure diversity in tropical forests. Proc. Natl. Acad. Sci. U. S. A. 104, 19029–19033. Wu, L.G., Wang, B., Geng, S.Q., 2005. Growing typhoon influence on east Asia. Geophys. Res. Lett. 32. Yamamoto, S., 1992. Gap characteristics and gap regeneration in primary evergreen broad-leaved forest of western Japan. Bot. Mag. Tokyo 105, 29–45. Yoshida, T., Noguchi, M., 2009. Vulnerability to strong winds for major tree species in a northern Japanese mixed forest: analyses of historical data. Ecol. Res 24, 909–919. Zielonka, T., Holeksa, J., Fleischer, P., Kapusta, P., 2010. A tree-ring reconstruction of wind disturbances in a forest of the Slovakian Tatra Mountains, Western Carpathians. J. Veg. Sci. 21, 31–42.