Accepted Article
Received Date : 21-Dec-2014 Accepted Date : 14-Feb-2015 Article type
: Primary Research Articles
Regional-scale directional changes in abundance of tree species along a temperature gradient in Japan
Running head: Regional change in abundance of tree species
Satoshi N. Suzuki1, 3*, Masae I. Ishihara2, 3, Amane Hidaka3 1
The University of Tokyo Chichibu Forest, Graduate School of Agricultural and Life Sciences,
the University of Tokyo, Chichibu, Japan 2
Graduate School for International Development and Cooperation, Hiroshima University,
Higashi-Hiroshima, Japan 3
Network Center of Forest and Grassland Survey in Monitoring Sites 1000 Project, Japan
Wildlife Research Center, c/o Tomakomai Research Station, Filed Science Center for Northern Biosphere, Hokkaido University, Tomakomai, Hokkaido 053-0035, Japan
* Corresponding author The University of Tokyo Chichibu Forest, Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-49 Hinoda-machi, Chichibu, Saitama 368-0034, Japan Tel: +81-494-220272 Fax: +81-494-239620 E-mail: [email protected] This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/gcb.12911 This article is protected by copyright. All rights reserved.
Accepted Article
Keywords: climate change; abundance; demography; directional change; Japan; Monitoring Sites 1000 Project; succession; disturbance; temperate forest; permanent plots
Type of Paper: Primary Research Article
Abstract Climate changes are assumed to shift the ranges of tree species and forest biomes. Such range shifts result from changes in abundances of tree species or functional types. Owing to global warming, the abundance of a tree species or functional type is expected to increase near the colder edge of its range and decrease near the warmer edge. The present study examined directional changes in abundance and demographic parameters of forest trees along a temperature gradient, as well as a successional gradient, in Japan. Changes in the relative abundance of each of four functional types (evergreen broad-leaved, deciduous broad-leaved, evergreen temperate conifer, and evergreen boreal conifer) and the demography of each species (recruitment rate, mortality, and population growth rate) were analysed in 39 permanent forest plots across the Japanese archipelago. Directional changes in the relative abundance of functional types were detected along the temperature gradient. Relative abundance of evergreen broad-leaved trees increased near their colder range boundaries, especially in secondary forests, coinciding with the decrease in deciduous broad-leaved trees. Similarly, relative abundance of deciduous broad-leaved trees increased near their colder range boundaries, coinciding with the decrease in boreal conifers. These functional type-level changes were mainly due to higher recruitment rates and partly to the lower mortality of individual species at colder sites. This is the first report to show that tree species abundances in temperate forests are changing directionally along a temperature gradient, which might be due to current or past climate changes as well as recovery from past disturbances.
This article is protected by copyright. All rights reserved.
Accepted Article
Introduction
Expansion and contraction of species’ ranges have been reported in relation to climate change (e.g., global warming) for various organisms (Parmesan & Yohe, 2003; Hickling et al., 2006; Chen et al., 2011). However, reports of range shifts of long-lived sessile organisms such as trees are relatively few compared with those of short-lived or more mobile organisms (Kelly & Goulden, 2008; Matías & Jump, 2014); this may be attributed to lags in the expansion and contraction of their ranges caused by the limited rates of dispersal and the prolonged longevity of trees. When the geographic distribution of a species shifts, its local abundance and demography (i.e., mortality, recruitment rate, population growth rate) are also expected to change (Iverson & Prasad, 1998; Thomas, 2010). Changes in species abundances result from changes in the balance between recruitment rate and mortality. By using a long-term monitoring plot, recent studies reported climate-related changes in the demography and community composition of trees (Feeley et al., 2011; Ouédraogo et al., 2013). In addition, studies compiling data from widespread, permanent plots in tropical forests have revealed changes in demography, species composition, and carbon stock; these changes may be due to increased drought stress and carbon fertilization (Phillips, 1998; Laurance et al., 2004; Lewis et al., 2004, 2009; Phillips et al., 2004; Fauset et al., 2012). Drought-induced acceleration of tree mortality has also been reported at the regional scale in North America (van Mantgem et al., 2009; Peng et al., 2011). These studies provide powerful evidence that forest ecosystems are responding to current climate changes. However, it is still unclear whether these climate changes affect the distribution of species and biomes, because most studies have been conducted within single biomes (but see van Mantgem et al., 2009; Bell et al., 2014). Under global warming, the abundance of a species is expected to increase near the colder edge of the This article is protected by copyright. All rights reserved.
Accepted Article
species’ range and to decrease near the warmer edge (Colwell et al., 2008; Morin et al., 2008; Ackerly et al., 2010; Thomas, 2010). Therefore, because changes in species abundance are typically most evident near the boundaries between biomes (Matías & Jump, 2014), studies based on networks of plots covering multiple biomes should be more effective in detecting current changes in species abundance. This study aimed to examine directional changes in tree species abundance along a temperature gradient with a broad latitudinal range covering multiple forest biomes (evergreen broad-leaved, summer-green deciduous broad-leaved, and boreal coniferous forests), by using multiple monitoring plots spread widely across the Japanese archipelago. Here, evergreen broad-leaved forests are distributed in the warm-temperate region, from southern to central lowlands; summer-green deciduous broad-leaved forests are distributed in the cool-temperate region, from central montane to northern lowland regions (Numata, 1974; Miyawaki, 1984). Subalpine forests and boreal evergreen coniferous forests occur at higher altitudes and latitudes, in regions colder than the cool-temperate region. Because the regional climate ranges from mesic to humid, geographic distributions of tree species and forest biomes are generally determined by temperature (Kira, 1977). Therefore, the abundance of each functional tree type (i.e., evergreen broad-leaf, deciduous broad-leaf and boreal evergreen coniferous trees) and tree species may be sensitive to on-going climate warming. Simulation models have shown that deciduous broad-leaved and boreal evergreen coniferous trees will be replaced near their warmer boundaries by evergreen broad-leaved and deciduous broad-leaved trees, respectively (Tanaka et al., 2006; Nakao et al., 2010). When examining directional changes in tree species abundance along a temperature gradient, historical effects should be also considered (Phillips et al., 2002). In particular, past human activity and the resulting ecological succession may also produce compositional changes in temperate forests. A large proportion of Japanese forest has historically been used
This article is protected by copyright. All rights reserved.
Accepted Article
for the production of timber, fuel wood, and other materials, often having been maintained as coppice forest (Totman, 1998). For example, secondary forests in the warm-temperate region are often dominated by summer-green deciduous broad-leaved trees, even though forests in the region were originally dominated by evergreen broad-leaved trees (Numata, 1974; Miyawaki, 1984). These forests have been left unmanaged in past years, and succession towards the original evergreen broad-leaved community is currently expected. Because this type of successional change occurs in the same direction as changes predicted by global warming, studies examining current changes to forest communities should attempt to distinguish the effects of global warming from historical effects (i.e., secondary succession). In this study, we used the tree census data from 39 permanent plots and assessed whether tree species of the Japanese archipelago are currently experiencing directional changes in abundance along a temperature gradient. Specific questions included: (1) Does the abundance of each species and functional type (evergreen broad-leaved, deciduous broad-leaved, and temperate and boreal conifer) increase near the colder boundary and decrease near the warmer boundary of its range? (2) Are directional shifts in distribution explained by demographic responses to temperature? (3) Are current changes in abundance caused by global warming or by historical effects, such as forest succession?
Materials and Methods Data The Japanese archipelago occupies a broad latitudinal range, from 24.05° N to 45.51° N and varies in altitude from sea level to 3776 m above sea level. Typical forest biomes include evergreen broad-leaved, summer-green deciduous broad-leaved and boreal coniferous forests. The tree census dataset used in this study was obtained from 39 permanent forest plots of the Monitoring Sites 1000 Project, which cover the above-mentioned forest types; the project was
This article is protected by copyright. All rights reserved.
Accepted Article
launched by the Ministry of the Environment, Japan (see Ishihara et al. 2011, for details). The full dataset is available from the website of the Biodiversity Center of Japan (http://www.biodic.go.jp/moni1000/index.html, in Japanese), and a subset of the data has been published as a data paper (Ishihara et al., 2011). Forest plots were established at each site, and trees with stem girth at breast height (GBH) greater than 15 cm were measured every one to five years, beginning in 2004. We limited our analysis to plots that were censused at least twice during the period from 2004 to 2012 and to trees with stem diameter at breast height (DBH) greater than 5 cm. The analysed data included 47,133 stems of 312 species in 39 plots (Fig. 1, Table S1). Plot sizes were usually 1 ha and ranged from 0.1 ha to 1.2 ha. The plots represented all major climate zones and biomes of the Japanese archipelago. They were located either in secondary forests, which have experienced anthropogenic disturbances in the past (i.e., partial or clear cutting), or in old-growth forests, which showed no signs of anthropogenic disturbance. Mean annual temperatures ranged from 3.3°C to 21.0°C and annual precipitation ranged from 867 mm to 3677 mm (Table S1).
Functional-type-level analyses We classified tree species into four functional types: evergreen broad-leaved (EB), deciduous broad-leaved (DB), temperate conifer (TC), and boreal conifer (BC). The two types of conifers (temperate and boreal) were analysed separately as they were clearly distinguishable by distribution along the temperature gradient. For each of the four functional types, relative abundance (R) and change in relative abundance from the first to the most recent census (ΔR) were calculated for each plot, as follows: R0ij = N0ij / N0i ΔRij = (Ntij / Nti - N0ij / N0i)/t = (Rtij - R0ij) / t where t is years since the first census; N0ij and Ntij are the number of stems of functional type j
This article is protected by copyright. All rights reserved.
Accepted Article
in plot i in the first (year 0) and last (year t) census, respectively, and N0i and Nti are the total number of stems in plot i at the time of the first and last census, respectively. To detect compositional changes along the temperature gradient, response of R0ij to mean annual temperature (MAT) was analysed by a generalized linear mixed model assuming binomial error with logit as the link function and plot identity as a random factor, using the glmer function of the lme4 library in version 3.0.0 of the statistical programming language R (R Core Team, 2014). Potential explanatory variables were MAT, square of MAT, and forest successional stage (secondary or old-growth forest), and the best models were selected by comparing the full model with reduced models based on Akaike’s Information Criteria (AIC). Degrees of freedom were calculated as sum of the number of fixed variables and that of variance parameters. To test whether the abundance of each functional type increased at the colder boundary and decreased at the warmer boundary of its range, responses of ΔRij to MAT were analysed by a linear model (glm function in R, with normal distribution error and identity as the link function). As variance of ΔRij is expected to be larger for sites with smaller numbers of stems of a given functional type (j), the number of stems at the first census, N0ij, was used as a weight for dispersions in the model-fitting process, wherein dispersions are assumed to be inversely proportional to N0ij. Potential explanatory variables were MAT and forest successional stage, and the best model was selected based on AIC. Square of MAT was not included as an explanatory variable because we wanted to test the directional change in ΔRij along the temperature gradient.
Species-level analyses To evaluate the demographic changes occurring at the species level, we analysed 214 species that occurred in two or more plots. The data included 46,084 stems in 39 plots (Table S2).
This article is protected by copyright. All rights reserved.
Accepted Article
Recruitment rate (rik), mortality (mik), and net population growth rate (λik) of species k in plot i were analysed by a hierarchical Bayesian model (see Appendix S1 for details). These were modelled as follows: Btik ~ Pois (Ntik (1 - (1 - rik)t)) Dtik ~ Binom (N0ik, 1 - (1 - mik)t) Ntik ~ Pois (N0ik (λik)t) where Btik and Dtik are the observed numbers of recruited and dead stems, respectively, of species k in plot i from the first census to the last census t years later; Ntik and N0ik represent the number of stems of species k in plot i at the time of the last and first census, respectively; y ~ Binom (n, x) indicates that y has a binomial distribution with probability x and sample size n; and y ~ Pois (x) indicates that y has a Poisson distribution with mean x. The formulas for recruitment rate and mortality were derived from those of Sheil (2000). Mortality, recruitment rate, and population growth rate were modelled as functions of mean annual temperature Ti and successional stage Si (0 for old-growth forest and 1 for secondary forest), as follows:
g(yik) = βIk + βTk Ti + βSk Si + φi
where yik is the focal demographic parameter (i.e. mik, rik, or λik); g(•) is an appropriate link function: logit for mortality rate and log for recruitment rate and population growth rate; βIk, βTk, and βSk are parameters for intercept, slope against MAT, and effect of successional stage for species k, respectively; φi is the random effect for plot i. βIk, βTk, and βSk are assumed to normally distribute around family-level means μIf, μTf, and μSf, respectively. The family-level means μIf, μTf, and μSf, in turn, are assumed to normally distribute around functional type-level means μIj, μTj, and μSj, respectively. As few trees of type BC occurred in secondary forests, the effect of successional stage was omitted for this functional type (i.e., βSk, μSf, and μSj were not
This article is protected by copyright. All rights reserved.
Accepted Article
included in the model). We used a random-walk Metropolis-Hastings algorithm, which is a type of Markov Chain Monte Carlo (MCMC) method, to characterize the posterior distributions of the model parameters. We ran the MCMC algorithm for three chains of 5 × 107 iterations with a burn in of 2 × 107 iterations and a thinning rate of 3 × 104 iterations. The MCMC program was written in C and compiled to a library for R. The 50% and 95% credible intervals (CIs) of the posterior distributions of the parameters were calculated from the 50% and 95% highest posterior probabilities, respectively, by using the HPDintervals function from the coda library in R (Plummer et al., 2006).
Results Response of functional types to temperature and succession Relative abundances of evergreen broadleaf (EB) and deciduous broadleaf (DB) trees clearly varied with MAT (Fig. 2). AICs of the models are shown in Table S3. Relative abundance of EB was clearly lower when MAT decreased from 15 to 10 °C (Fig. 2a), coinciding with a rise in the relative abundance of DB (Fig. 2b). For mid-temperature sites, relative abundance of EB was lower in secondary forest than in old-growth forest, while this trend was reversed for DB. Relative abundance of temperate conifers (TC) tended to be highest at mid-temperatures (Fig. 2c), although the best model included only the intercept. Relative abundance of boreal conifers (BC) increased sharply when MAT decreased from 8 to 3 °C (Fig. 2d). For EB and DB, changes in relative abundance (ΔRij) were lower for warmer sites than for colder sites. For EB, ΔRij was mostly positive across the temperature gradient and higher in secondary forest than in old-growth forest (Fig. 2e). For DB, ΔRij tended to be positive for colder sites, indicating an increase in abundance, and negative for warmer sites, indicating a decrease in abundance (Fig. 2f). For TC, ΔRij was lower in secondary forest than
This article is protected by copyright. All rights reserved.
Accepted Article
in old-growth forest (Fig. 2g). For BC, ΔRij was generally negative and correlated neither with MAT nor with successional stage (Fig. 2h).
Response of species demography to temperature and succession Recruitment rates of evergreen and deciduous broadleaf species were generally higher for colder sites (Fig. 3a, b); the posterior distributions of the slope against MAT, βTk, deviated from 0 in the negative direction, and the 95% CI of βTk was less than zero for 30 of 72 EB species and for 15 of 128 DB species. The mean βTk for recruitment rate among species within a functional type (μTj) deviated in the negative direction, and its’ 95% CIs were negative for EB and DB (Fig. 4a). The 95% CI of βTk included zero for all TC species but the 50% CI was positive for two species (Fig. 3c). The 50% CI of μTj for TC included zero (Fig 4a). The 95% CI of βTk for all BC species and of μTj for BC were positive (Fig. 3d, Fig. 4a). Mortality of EB and DB species tended to be higher for the warmer sites, but the 95% CIs of βTk were positive for only two EB species and for none of the DB species (Fig. 3e, 3f). Although the 95% CIs of μTj included zero for all functional types, the 50% CIs were positive for EB and TC (Fig. 4b). Population growth rates of EB and DB species were generally higher for the colder sites (Fig. 3i, j). The 95% CIs of βTk for four EB species and two DB species were negative. The 95 CI of μTj for DB and the 50% CI for EB were also negative (Fig. 4c). The mean within a functional type for effects of successional stage (μSj) for recruitment rate and population growth rate deviated in the positive direction for EB, and the 95% CIs for recruitment rate and 50% CIs for population growth rate were both positive (Fig. 4d, f). The 50% CI of μSj for mortality was negative for EB (Fig. 4e). In contrast, for TC, the 95% CI of μSj for recruitment rate was positive, and the 50% CI for mortality was negative (Fig. 4d, 4e). In DB, the 50% CIs of μSj included zero for all demographic parameters.
This article is protected by copyright. All rights reserved.
Accepted Article
Discussion Our results indicate on-going, region-wide directional changes in tree species abundances in Japan. Major changes include: (1) an increase in evergreen broad-leaved trees in the mid-temperature region (5-15 °C in MAT), especially in secondary forests, coinciding with a decrease in deciduous broad-leaved trees and temperate conifers and (2) an increase in deciduous broad-leaved trees increased in cold regions (MAT < 5 °C), coinciding with a decrease in boreal conifers. These functional-type changes can be explained by the demography of the composing species: recruitment rate of the composing species was generally higher and mortality rate tended to be lower toward a colder temperature. This is the first report of relative abundances of tree species changing directionally along a temperature gradient, at a regional scale; distributions of forest biomes and tree species in Japan seem to be moving towards higher altitudes and northern latitudes. Besides temperature, other climatic factors such as precipitation are known to cause change in abundance and distribution of tree species in many regions (Feeley et al., 2011; Fauset et al., 2012). In Japan, the duration of snow cover and/or maximum snow depth are known to affect the distributions of some trees (Shidei, 1979; Gransert, 2004). However, precipitation and snowfall are largely related with temperature across Japan; warmer sites tend to have higher precipitation and shallower snow cover (Table S1). Furthermore, due to sufficient precipitation in Japan, distributions of biomes and tree species are largely dependent on temperature rather than precipitation (Kira, 1977) and biogeographical turnover of most tree species across the Japanese archipelago is strongly controlled by mean annual temperature (Kubota et al., 2014). Therefore, although the observed changes may be partly attributed to other climatic factors such as duration of snow cover, the changes are likely to be basically temperature related.
This article is protected by copyright. All rights reserved.
Accepted Article
Our results were consistent with expectations of how plant species respond to global warming: local abundance will increase near colder range boundaries and decrease near warmer boundaries (Iverson & Prasad, 1998; Morin et al., 2008; Walther, 2010). However, as the results were based on relatively short-term observations (less than 10-years), it is unclear whether the tree species are responding to current or past climate changes. As tree species comprise long-lived, sessile organisms, their ability to track climate changes is limited (Svenning et al., 2008). Therefore, there will be a delay in species abundance and distribution to reach the potential equilibrium after the past climate changes. It is reported that many European tree species are not recovering their climatic potential ranges, suggesting post-glacial migration lag (Svenning & Skov, 2004; Normand et al., 2011). Distributions of tree species may have also been affected by the Little Ice Age (LIA) 200-300 years ago, during which Japan experienced a cold and rainy climate (Maejima & Tagami, 1983; Bradley & Jones, 1993; Yamaguchi et al., 2008). For example, it is hypothesized that the beech species Fagus crenata increased in abundance in the south eastern Japanese archipelago during the LIA, and has been decreasing due to a lack of regeneration after the end of the LIA (Shimano, 2002). Evergreen broad-leaved trees showed a greater increase in secondary forests than in old-growth forests, which suggests that the shift to evergreen broad-leaved forest can be attributed in part to forest succession after past anthropogenic disturbance. Evergreen broad-leaved shrubs and small trees, such as Cleyera japonica and Eurya japonica, as well as evergreen broad-leaved canopy trees, such as Castanopsis sieboldii, increased at those sites (see Fig. S1). These species are thought to have originally covered a wide area of the mid-temperature region before the settlement of ancient Japanese people; human activity is thought to have greatly reduced them to their current distribution and abundance (Numata, 1974; Miyawaki, 1984). As managed forests have been abandoned, vegetation succession
This article is protected by copyright. All rights reserved.
Accepted Article
precipitation, frequency of super typhoon strikes) will accelerate or decelerate the process. Long-term and regional-scale network monitoring systems will be required to predict precisely how future climate change will affect the distribution and abundance of tree species.
Acknowledgements The data used here was provided by the Ministry of the Environment Monitoring Sites 1000 Project (TreeDataPackage2012ver1.zip, downloaded from http://www.biodic.go.jp/moni1000/findings/data/index.html on April 22, 2014). We thank its participant organizations and their staffs for data collection and site management efforts. We are also grateful to Prof. Tsutom Hiura and Dr. Shigeru Niwa for their helpful comments and Dr. Ikuyo Saeki for her detailed comments on an earlier version of this manuscript. This study was partially funded by the Environmental Research and Technology Development Fund (S-9-3) of the Ministry of the Environment, Japan.
References Ackerly DD, Loarie SR, Cornwell WK, Weiss SB, Hamilton H, Branciforte R, Kraft NJB (2010) The geography of climate change: implications for conservation biogeography. Diversity and Distributions, 16, 476–487. Bell DM, Bradford JB, Lauenroth WK (2014) Early indicators of change: divergent climate envelopes between tree life stages imply range shifts in the western United States. Global Ecology and Biogeography, 23, 168–180. Bradley RS, Jones PD (1993) “Little Ice Age” summer temperature variations: their nature and relevance to recent global warming trends. Holocene, 3-4, 367–376. Chen I-C, Hill JK, Ohlemüller R, Roy DB, Thomas CD (2011) Rapid range shifts of species associated with high levels of climate warming. Science, 333, 1024–1026.
This article is protected by copyright. All rights reserved.
Accepted Article
precipitation, frequency of super typhoon strikes) will accelerate or decelerate the process. Long-term and regional-scale network monitoring systems will be required to predict precisely how future climate change will affect the distribution and abundance of tree species.
Acknowledgements The data used here was provided by the Ministry of the Environment Monitoring Sites 1000 Project (TreeDataPackage2012ver1.zip, downloaded from http://www.biodic.go.jp/moni1000/findings/data/index.html on April 22, 2014). We thank its participant organizations and their staffs for data collection and site management efforts. We are also grateful to Prof. Tsutom Hiura and Dr. Shigeru Niwa for their helpful comments and Dr. Ikuyo Saeki for her detailed comments on an earlier version of this manuscript. This study was partially funded by the Environmental Research and Technology Development Fund (S-9-3) of the Ministry of the Environment, Japan.
References Ackerly DD, Loarie SR, Cornwell WK, Weiss SB, Hamilton H, Branciforte R, Kraft NJB (2010) The geography of climate change: implications for conservation biogeography. Diversity and Distributions, 16, 476–487. Bell DM, Bradford JB, Lauenroth WK (2014) Early indicators of change: divergent climate envelopes between tree life stages imply range shifts in the western United States. Global Ecology and Biogeography, 23, 168–180. Bradley RS, Jones PD (1993) “Little Ice Age” summer temperature variations: their nature and relevance to recent global warming trends. Holocene, 3-4, 367–376. Chen I-C, Hill JK, Ohlemüller R, Roy DB, Thomas CD (2011) Rapid range shifts of species associated with high levels of climate warming. Science, 333, 1024–1026.
This article is protected by copyright. All rights reserved.
Accepted Article
Colwell RK, Brehm G, Cardelús CL, Gilman AC, Longino JT (2008) Global warming, elevational range shifts, and lowland biotic attrition in the wet tropics. Science, 322, 258–261. Fauset S, Baker TR, Lewis SL et al. (2012) Drought-induced shifts in the floristic and functional composition of tropical forests in Ghana. Ecology Letters, 15, 1120–1129. Feeley KJ, Davies SJ, Perez R, Hubbell STP, Foster RB (2011) Directional changes in the species composition of a tropical forest. Ecology, 92, 871–882. Gansert D (2004) Treelines of the Japanese Alps - altitudinal distribution and species composition under contrasting winter climates. Flora, 199, 143–156. Henbo Y, Itaya A, Nishimura N, Yamamoto S-I (2006) Long-term canopy dynamics analyzed by aerial photographs and digital elevation data in a subalpine old-growth coniferous forest. Ecoscience, 13, 451–458. Hickling R, Roy DB, Hill JK, Fox R, Thomas CD (2006) The distributions of a wide range of taxonomic groups are expanding polewards. Global Change Biology, 12, 450–455. Hiroki S (2001) Invation of Quercus variabilis and Castanopsis cuspidata seedlings into secondary forests on the western slope of Mt. Miroku and adjacent hill in Kasugai City, Aichi Prefecture, Japan. Vegetation Science, 18, 31–37. Hoshino D, Nishimura N, Yamamoto S-I (2003) Effects of canopy conditions on the regeneration of major tree species in an old-growth Chamaecyparis obtusa forest in central Japan. Forest Ecology and Management, 175, 141–152. Ishihara MI, Suzuki SN, Nakamura M et al. (2011) Forest stand structure, composition, and dynamics in 34 sites over Japan. Ecological Research, 26, 1007–1008. Iverson LR, Prasad AM (1998) Predicting abundance of 80 tree species following climate change in the eastern United States. Ecological Monographs, 68, 465–485.
This article is protected by copyright. All rights reserved.
Accepted Article
Kamada M, Nakagoshi N (1990) Pattern and processes of secondary vegetation at a farm village in southwestern Japan after the 1960s. Japanese Journal of Ecology, 40, 137–150 (in Japanese with English summary). Kelly AE, Goulden ML (2008) Rapid shifts in plant distribution with recent climate change. Proceedings of the National Academy of Sciences of the United States of America, 105, 11823–11826. Kira T (1977) A climatological interpretation of Japanese vegetation zones. In: Vegetation science and environmental protection (eds. Miyawaki A, Tüxen R), pp. 21–30, Maruzen, Tokyo, Japan. Kubota Y, Hirao T, Fujii SJ, Shiono T, Kusumoto B (2014) Beta diversity of woody plants in the Japanese archipelago: The roles of geohistorical and ecological processes. Journal of Biogeography, 41, 1267–1276. Laurance WF, Oliveira AA, Laurance SG et al. (2004) Pervasive alteration of tree communities in undisturbed Amazonian forests. Nature, 428, 171–175. Lewis SL, Phillips OL, Baker TR et al. (2004) Concerted changes in tropical forest structure and dynamics: evidence from 50 South American long-term plots. Philosophical transactions of the Royal Society of London, Series B, 359, 421–36. Lewis SL, Lloyd J, Sitch S, Mitchard ETA, Laurance WF (2009) Changing ecology of tropical forests: evidence and drivers. Annual Review of Ecology, Evolution, and Systematics, 40, 529–549. Maejima I, Tagami Y (1983) Climate of Little Ice Age in Japan. Geographical Reports of Tokyo Metropolitan University, 18, 91–111. Van Mantgem PJ, Stephenson NL, Byrne JC et al. (2009) Widespread increase of tree mortality rates in the western United States. Science, 323, 521–524.
This article is protected by copyright. All rights reserved.
Accepted Article
Matías L, Jump AS (2014) Asymmetric changes of growth and reproductive investment herald altitudinal and latitudinal range shifts of two woody species. Global Change Biology, early view articles, doi: 10.1111/gcb.12683. Miyawaki A (1984) A vegetation-ecological view of the Japanese Archipelago. Bulletin of Institute of Environmental Science and Technology, Yokohama National University, 11, 85–101. Mori AS (2011) Ecosystem management based on natural disturbances: hierarchical context and non-equilibrium paradigm. Journal of Applied Ecology, 48, 280–292. Morin X, Viner D, Chuine I (2008) Tree species range shifts at a continental scale: new predictive insights from a process-based model. Journal of Ecology, 96, 784–794. Nakao K, Matsui T, Horikawa M, Tsuyama I, Tanaka N (2010) Assessing the impact of land use and climate change on the evergreen broad-leaved species of Quercus acuta in Japan. Plant Ecology, 212, 229–243. Normand S, Ricklefs RE, Skov F, Bladt J, Tackenberg O, Svenning J-C (2011) Postglacial migration supplements climate in determining plant species ranges in Europe. Proceedings of the Royal Society, Series B, 278, 3644–53. Numata M (ed.) (1974) The Flora and Vegetation of Japan. Kodansha, Tokyo. Ouédraogo D-Y, Mortier F, Gourlet-Fleury S, Freycon V, Picard N (2013) Slow-growing species cope best with drought: evidence from long-term measurements in a tropical semi-deciduous moist forest of Central Africa. Journal of Ecology, 101, 1459–1470. Parmesan C, Yohe G (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature, 421, 37–42. Peng C, Ma Z, Lei X et al. (2011) A drought-induced pervasive increase in tree mortality across Canada’s boreal forests. Nature Climate Change, 1, 467–471.
This article is protected by copyright. All rights reserved.
Accepted Article
Perry GLW (2002) Landscapes, space and equilibrium: shifting viewpoints. Progress in Physical Geography, 26, 339–359. Phillips OL (1998) Changes in the carbon balance of tropical forests: evidence from long-term plots. Science, 282, 439–442. Phillips OL, Malhi Y, Vinceti B et al. (2002) Changes in growth of tropical forests: evaluating potential biases. Ecological Applications, 12, 576–587. Phillips OL, Baker TR, Arroyo L et al. (2004) Pattern and process in Amazon tree turnover, 1976-2001. Philosophical transactions of the Royal Society of London, Series B, 359, 381–407. Plummer M, Best N, Cowles K, Vines K (2006) CODA: Convergence Diagnosis and Output Analysis for MCMC. R News, 6, 7–11. R Core Team (2014) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.r-project.org/. Saito O, Hoshino Y, Tsuji S, Kanno A (2003) Changes in species richness and species composition of Quercus serrata secondary forests after more than 20 years in Kanto region, Japan. Vegetation Science, 20, 83–96 (in Japanese with English summary). Sheil D, Jennings S, Savill P (2000) Long term permanent plot observations of vegetation dynamics in Budongo, a Ugandan rain forest. Journal of Tropical Ecology, 16, 765–800. Shidei T (1979) Distribution of special forest vegetation in heavy snowy regions in Japan. In: Vegetation und Landschaft Japans (eds Miyawaki A, Okuda S), PP. 93–99. The Yokohama Phytosociological Society, Yokohama. Shimano K (2002) Regeneration dynamics, causal factors, and characteristics of Pacific Ocean-type beech (Fagus crenata) forests in Japan: a review. Folia Geobotanica, 37, 275–296.
This article is protected by copyright. All rights reserved.
Accepted Article
Sprugel DG (1991) Disturbance, equilibrium, and environmental variability: What is “natural” vegetation in a changing environment ? Biological Conservation, 58, 1–18. Suzuki SN, Kachi N, Suzuki J-I (2013) Spatial variation of local stand structure in an Abies forest, 45 years after a large disturbance by the Isewan typhoon. Journal of Forest Research, 18, 139–148. Svenning J-C, Skov F (2004) Limited filling of the potential range in European tree species. Ecology Letters, 7, 565–573. Svenning J, Normand S, Skov F (2008) Postglacial dispersal limitation of widespread forest plant species in nemoral Europe. Ecography, 31, 316–326. Tamate S, Kashiyama T, Sasanuma T, Takahashi K, Matsuoka H (1977) On the distribution map of forest wind damage by typhoon No 15, 1954 in Hokkaido. Bulletin of the Government Forestry Experiment Station, 289, 43–67 (in Japanese with English summary). Tanaka N, Matsui T, Yagihashi T, Taoda H (2006) Climatic controls on natural forest distribution and predicting the impact of climate warming: Especially referring to Buna (Fagus crenata) Forests. Global Environmental Research, 10, 151–160. Thomas CD (2010) Climate, climate change and range boundaries. Diversity and Distributions, 16, 488–495. Totman C (1998) The Green Archipelago: Forestry in Preindustrial Japan. Ohio University Press, Athens. Walther G-R (2010) Community and ecosystem responses to recent climate change. Philosophical transactions of the Royal Society of London, Series B, 365, 2019–2024. Yamaguchi S, Naruse R, Shiraiwa T (2008) Climate reconstruction since the Little Ice Age by modelling Koryto glacier, Kamchatka Peninsula, Russia. Journal of Glaciology, 54, 125–130.
This article is protected by copyright. All rights reserved.
Accepted Article
Supporting Information Appendix S1: Structure of the hierarchical Bayesian model Table S1: Details of study sites Table S2: Analysed species family names and functional types. Table S3: AICs of models for functional-type–level analysis Figure S1: Observed recruitment rate and mortality of species
Figure legends Fig. 1 Locations of study sites in Japanese archipelago.
Fig. 2 Relative abundance and change in relative abundance for four tree functional types along mean annual temperature. For each functional type, only plots in which at least 20 stems were present at first census are included in the change in relative abundance data (e–f). Circle, old-growth forest; triangle, secondary forest. Regression curves estimated by the best model are shown; solid and dashed lines indicate old-growth and secondary forests, respectively, when forest successional stage was selected as an explanatory variable. When successional stage was not selected as an explanatory variable, regression curve common for old-growth and secondary forests is shown.
Fig. 3 Estimated recruitment rate, mortality, and population growth rate of evergreen broad-leaved, deciduous broad-leaved, temperate conifer, and boreal conifer species along mean annual temperature. Each line indicates the response curve for a species to temperature, with an estimated posterior median of parameters for old-growth forests (S = 0). Black line, response curve for the species whose 95% credible interval of the slope parameter (βTk) did not include zero; dark gray line, curve for the species whose 50% credible interval did not
This article is protected by copyright. All rights reserved.
Accepted Article
include zero; light gray line, curve for the species whose 50% credible interval included zero.
Fig. 4 Posterior density of functional-type means of the effect of mean annual temperature (μTj) and forest successional stage (μSj) on (a, d) recruitment rate, (b, e) mortality, and (c, f) population growth rate of species. Means are those among species within a functional type. Black solid line, 95% credible intervals did not include zero; gray solid line, 50% credible intervals did not include zero, gray dashed line, 50% credible intervals included zero. EB, evergreen broad-leaved; DB, deciduous broad-leaved; TC, temperate conifer; BC, boreal conifer.
This article is protected by copyright. All rights reserved.
Recruitment rate Mortality rate
Temperate conifer
Deciduous broadleaves
(a)
Boreal conifer
(c)
(b)
(d)
0.10
0.10
0.10
0.10
0.05
0.05
0.05
0.05
0.00
0.00
0.00 5
10
15
20
5
10
15
(e)
0.00 5
20
10
15
20
5
0.10
0.10
0.05
0.05
0.05
0.05
15
20
5
10
15
(i)
5
20
10
15
5
20
1.05
1.05
1.00
1.00
1.00
1.00
15
20
5
10
15
20
20
0.95
0.95
0.95 10
15
(l)
1.05
5
10
(k)
(j)
1.05
0.95
20
0.00
0.00
0.00 10
15
(h)
0.10
5
10
(g)
(f)
0.10
0.00
Population growth rate
Accepted Article
Evergreen broadleaves
5
10
Mean Annual Temperature (°C)
This article is protected by copyright. All rights reserved.
15
20
5
10
15
20
Accepted Article This article is protected by copyright. All rights reserved.
Received Date : 21-Dec-2014 Accepted Date : 14-Feb-2015 Article type
: Primary Research Articles
Regional-scale directional changes in abundance of tree species along a temperature gradient in Japan
Running head: Regional change in abundance of tree species
Satoshi N. Suzuki1, 3*, Masae I. Ishihara2, 3, Amane Hidaka3 1
The University of Tokyo Chichibu Forest, Graduate School of Agricultural and Life Sciences,
the University of Tokyo, Chichibu, Japan 2
Graduate School for International Development and Cooperation, Hiroshima University,
Higashi-Hiroshima, Japan 3
Network Center of Forest and Grassland Survey in Monitoring Sites 1000 Project, Japan
Wildlife Research Center, c/o Tomakomai Research Station, Filed Science Center for Northern Biosphere, Hokkaido University, Tomakomai, Hokkaido 053-0035, Japan
* Corresponding author The University of Tokyo Chichibu Forest, Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-49 Hinoda-machi, Chichibu, Saitama 368-0034, Japan Tel: +81-494-220272 Fax: +81-494-239620 E-mail: [email protected] This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/gcb.12911 This article is protected by copyright. All rights reserved.
Accepted Article
Keywords: climate change; abundance; demography; directional change; Japan; Monitoring Sites 1000 Project; succession; disturbance; temperate forest; permanent plots
Type of Paper: Primary Research Article
Abstract Climate changes are assumed to shift the ranges of tree species and forest biomes. Such range shifts result from changes in abundances of tree species or functional types. Owing to global warming, the abundance of a tree species or functional type is expected to increase near the colder edge of its range and decrease near the warmer edge. The present study examined directional changes in abundance and demographic parameters of forest trees along a temperature gradient, as well as a successional gradient, in Japan. Changes in the relative abundance of each of four functional types (evergreen broad-leaved, deciduous broad-leaved, evergreen temperate conifer, and evergreen boreal conifer) and the demography of each species (recruitment rate, mortality, and population growth rate) were analysed in 39 permanent forest plots across the Japanese archipelago. Directional changes in the relative abundance of functional types were detected along the temperature gradient. Relative abundance of evergreen broad-leaved trees increased near their colder range boundaries, especially in secondary forests, coinciding with the decrease in deciduous broad-leaved trees. Similarly, relative abundance of deciduous broad-leaved trees increased near their colder range boundaries, coinciding with the decrease in boreal conifers. These functional type-level changes were mainly due to higher recruitment rates and partly to the lower mortality of individual species at colder sites. This is the first report to show that tree species abundances in temperate forests are changing directionally along a temperature gradient, which might be due to current or past climate changes as well as recovery from past disturbances.
This article is protected by copyright. All rights reserved.
Accepted Article
Introduction
Expansion and contraction of species’ ranges have been reported in relation to climate change (e.g., global warming) for various organisms (Parmesan & Yohe, 2003; Hickling et al., 2006; Chen et al., 2011). However, reports of range shifts of long-lived sessile organisms such as trees are relatively few compared with those of short-lived or more mobile organisms (Kelly & Goulden, 2008; Matías & Jump, 2014); this may be attributed to lags in the expansion and contraction of their ranges caused by the limited rates of dispersal and the prolonged longevity of trees. When the geographic distribution of a species shifts, its local abundance and demography (i.e., mortality, recruitment rate, population growth rate) are also expected to change (Iverson & Prasad, 1998; Thomas, 2010). Changes in species abundances result from changes in the balance between recruitment rate and mortality. By using a long-term monitoring plot, recent studies reported climate-related changes in the demography and community composition of trees (Feeley et al., 2011; Ouédraogo et al., 2013). In addition, studies compiling data from widespread, permanent plots in tropical forests have revealed changes in demography, species composition, and carbon stock; these changes may be due to increased drought stress and carbon fertilization (Phillips, 1998; Laurance et al., 2004; Lewis et al., 2004, 2009; Phillips et al., 2004; Fauset et al., 2012). Drought-induced acceleration of tree mortality has also been reported at the regional scale in North America (van Mantgem et al., 2009; Peng et al., 2011). These studies provide powerful evidence that forest ecosystems are responding to current climate changes. However, it is still unclear whether these climate changes affect the distribution of species and biomes, because most studies have been conducted within single biomes (but see van Mantgem et al., 2009; Bell et al., 2014). Under global warming, the abundance of a species is expected to increase near the colder edge of the This article is protected by copyright. All rights reserved.
Accepted Article
species’ range and to decrease near the warmer edge (Colwell et al., 2008; Morin et al., 2008; Ackerly et al., 2010; Thomas, 2010). Therefore, because changes in species abundance are typically most evident near the boundaries between biomes (Matías & Jump, 2014), studies based on networks of plots covering multiple biomes should be more effective in detecting current changes in species abundance. This study aimed to examine directional changes in tree species abundance along a temperature gradient with a broad latitudinal range covering multiple forest biomes (evergreen broad-leaved, summer-green deciduous broad-leaved, and boreal coniferous forests), by using multiple monitoring plots spread widely across the Japanese archipelago. Here, evergreen broad-leaved forests are distributed in the warm-temperate region, from southern to central lowlands; summer-green deciduous broad-leaved forests are distributed in the cool-temperate region, from central montane to northern lowland regions (Numata, 1974; Miyawaki, 1984). Subalpine forests and boreal evergreen coniferous forests occur at higher altitudes and latitudes, in regions colder than the cool-temperate region. Because the regional climate ranges from mesic to humid, geographic distributions of tree species and forest biomes are generally determined by temperature (Kira, 1977). Therefore, the abundance of each functional tree type (i.e., evergreen broad-leaf, deciduous broad-leaf and boreal evergreen coniferous trees) and tree species may be sensitive to on-going climate warming. Simulation models have shown that deciduous broad-leaved and boreal evergreen coniferous trees will be replaced near their warmer boundaries by evergreen broad-leaved and deciduous broad-leaved trees, respectively (Tanaka et al., 2006; Nakao et al., 2010). When examining directional changes in tree species abundance along a temperature gradient, historical effects should be also considered (Phillips et al., 2002). In particular, past human activity and the resulting ecological succession may also produce compositional changes in temperate forests. A large proportion of Japanese forest has historically been used
This article is protected by copyright. All rights reserved.
Accepted Article
for the production of timber, fuel wood, and other materials, often having been maintained as coppice forest (Totman, 1998). For example, secondary forests in the warm-temperate region are often dominated by summer-green deciduous broad-leaved trees, even though forests in the region were originally dominated by evergreen broad-leaved trees (Numata, 1974; Miyawaki, 1984). These forests have been left unmanaged in past years, and succession towards the original evergreen broad-leaved community is currently expected. Because this type of successional change occurs in the same direction as changes predicted by global warming, studies examining current changes to forest communities should attempt to distinguish the effects of global warming from historical effects (i.e., secondary succession). In this study, we used the tree census data from 39 permanent plots and assessed whether tree species of the Japanese archipelago are currently experiencing directional changes in abundance along a temperature gradient. Specific questions included: (1) Does the abundance of each species and functional type (evergreen broad-leaved, deciduous broad-leaved, and temperate and boreal conifer) increase near the colder boundary and decrease near the warmer boundary of its range? (2) Are directional shifts in distribution explained by demographic responses to temperature? (3) Are current changes in abundance caused by global warming or by historical effects, such as forest succession?
Materials and Methods Data The Japanese archipelago occupies a broad latitudinal range, from 24.05° N to 45.51° N and varies in altitude from sea level to 3776 m above sea level. Typical forest biomes include evergreen broad-leaved, summer-green deciduous broad-leaved and boreal coniferous forests. The tree census dataset used in this study was obtained from 39 permanent forest plots of the Monitoring Sites 1000 Project, which cover the above-mentioned forest types; the project was
This article is protected by copyright. All rights reserved.
Accepted Article
launched by the Ministry of the Environment, Japan (see Ishihara et al. 2011, for details). The full dataset is available from the website of the Biodiversity Center of Japan (http://www.biodic.go.jp/moni1000/index.html, in Japanese), and a subset of the data has been published as a data paper (Ishihara et al., 2011). Forest plots were established at each site, and trees with stem girth at breast height (GBH) greater than 15 cm were measured every one to five years, beginning in 2004. We limited our analysis to plots that were censused at least twice during the period from 2004 to 2012 and to trees with stem diameter at breast height (DBH) greater than 5 cm. The analysed data included 47,133 stems of 312 species in 39 plots (Fig. 1, Table S1). Plot sizes were usually 1 ha and ranged from 0.1 ha to 1.2 ha. The plots represented all major climate zones and biomes of the Japanese archipelago. They were located either in secondary forests, which have experienced anthropogenic disturbances in the past (i.e., partial or clear cutting), or in old-growth forests, which showed no signs of anthropogenic disturbance. Mean annual temperatures ranged from 3.3°C to 21.0°C and annual precipitation ranged from 867 mm to 3677 mm (Table S1).
Functional-type-level analyses We classified tree species into four functional types: evergreen broad-leaved (EB), deciduous broad-leaved (DB), temperate conifer (TC), and boreal conifer (BC). The two types of conifers (temperate and boreal) were analysed separately as they were clearly distinguishable by distribution along the temperature gradient. For each of the four functional types, relative abundance (R) and change in relative abundance from the first to the most recent census (ΔR) were calculated for each plot, as follows: R0ij = N0ij / N0i ΔRij = (Ntij / Nti - N0ij / N0i)/t = (Rtij - R0ij) / t where t is years since the first census; N0ij and Ntij are the number of stems of functional type j
This article is protected by copyright. All rights reserved.
Accepted Article
in plot i in the first (year 0) and last (year t) census, respectively, and N0i and Nti are the total number of stems in plot i at the time of the first and last census, respectively. To detect compositional changes along the temperature gradient, response of R0ij to mean annual temperature (MAT) was analysed by a generalized linear mixed model assuming binomial error with logit as the link function and plot identity as a random factor, using the glmer function of the lme4 library in version 3.0.0 of the statistical programming language R (R Core Team, 2014). Potential explanatory variables were MAT, square of MAT, and forest successional stage (secondary or old-growth forest), and the best models were selected by comparing the full model with reduced models based on Akaike’s Information Criteria (AIC). Degrees of freedom were calculated as sum of the number of fixed variables and that of variance parameters. To test whether the abundance of each functional type increased at the colder boundary and decreased at the warmer boundary of its range, responses of ΔRij to MAT were analysed by a linear model (glm function in R, with normal distribution error and identity as the link function). As variance of ΔRij is expected to be larger for sites with smaller numbers of stems of a given functional type (j), the number of stems at the first census, N0ij, was used as a weight for dispersions in the model-fitting process, wherein dispersions are assumed to be inversely proportional to N0ij. Potential explanatory variables were MAT and forest successional stage, and the best model was selected based on AIC. Square of MAT was not included as an explanatory variable because we wanted to test the directional change in ΔRij along the temperature gradient.
Species-level analyses To evaluate the demographic changes occurring at the species level, we analysed 214 species that occurred in two or more plots. The data included 46,084 stems in 39 plots (Table S2).
This article is protected by copyright. All rights reserved.
Accepted Article
Recruitment rate (rik), mortality (mik), and net population growth rate (λik) of species k in plot i were analysed by a hierarchical Bayesian model (see Appendix S1 for details). These were modelled as follows: Btik ~ Pois (Ntik (1 - (1 - rik)t)) Dtik ~ Binom (N0ik, 1 - (1 - mik)t) Ntik ~ Pois (N0ik (λik)t) where Btik and Dtik are the observed numbers of recruited and dead stems, respectively, of species k in plot i from the first census to the last census t years later; Ntik and N0ik represent the number of stems of species k in plot i at the time of the last and first census, respectively; y ~ Binom (n, x) indicates that y has a binomial distribution with probability x and sample size n; and y ~ Pois (x) indicates that y has a Poisson distribution with mean x. The formulas for recruitment rate and mortality were derived from those of Sheil (2000). Mortality, recruitment rate, and population growth rate were modelled as functions of mean annual temperature Ti and successional stage Si (0 for old-growth forest and 1 for secondary forest), as follows:
g(yik) = βIk + βTk Ti + βSk Si + φi
where yik is the focal demographic parameter (i.e. mik, rik, or λik); g(•) is an appropriate link function: logit for mortality rate and log for recruitment rate and population growth rate; βIk, βTk, and βSk are parameters for intercept, slope against MAT, and effect of successional stage for species k, respectively; φi is the random effect for plot i. βIk, βTk, and βSk are assumed to normally distribute around family-level means μIf, μTf, and μSf, respectively. The family-level means μIf, μTf, and μSf, in turn, are assumed to normally distribute around functional type-level means μIj, μTj, and μSj, respectively. As few trees of type BC occurred in secondary forests, the effect of successional stage was omitted for this functional type (i.e., βSk, μSf, and μSj were not
This article is protected by copyright. All rights reserved.
Accepted Article
included in the model). We used a random-walk Metropolis-Hastings algorithm, which is a type of Markov Chain Monte Carlo (MCMC) method, to characterize the posterior distributions of the model parameters. We ran the MCMC algorithm for three chains of 5 × 107 iterations with a burn in of 2 × 107 iterations and a thinning rate of 3 × 104 iterations. The MCMC program was written in C and compiled to a library for R. The 50% and 95% credible intervals (CIs) of the posterior distributions of the parameters were calculated from the 50% and 95% highest posterior probabilities, respectively, by using the HPDintervals function from the coda library in R (Plummer et al., 2006).
Results Response of functional types to temperature and succession Relative abundances of evergreen broadleaf (EB) and deciduous broadleaf (DB) trees clearly varied with MAT (Fig. 2). AICs of the models are shown in Table S3. Relative abundance of EB was clearly lower when MAT decreased from 15 to 10 °C (Fig. 2a), coinciding with a rise in the relative abundance of DB (Fig. 2b). For mid-temperature sites, relative abundance of EB was lower in secondary forest than in old-growth forest, while this trend was reversed for DB. Relative abundance of temperate conifers (TC) tended to be highest at mid-temperatures (Fig. 2c), although the best model included only the intercept. Relative abundance of boreal conifers (BC) increased sharply when MAT decreased from 8 to 3 °C (Fig. 2d). For EB and DB, changes in relative abundance (ΔRij) were lower for warmer sites than for colder sites. For EB, ΔRij was mostly positive across the temperature gradient and higher in secondary forest than in old-growth forest (Fig. 2e). For DB, ΔRij tended to be positive for colder sites, indicating an increase in abundance, and negative for warmer sites, indicating a decrease in abundance (Fig. 2f). For TC, ΔRij was lower in secondary forest than
This article is protected by copyright. All rights reserved.
Accepted Article
in old-growth forest (Fig. 2g). For BC, ΔRij was generally negative and correlated neither with MAT nor with successional stage (Fig. 2h).
Response of species demography to temperature and succession Recruitment rates of evergreen and deciduous broadleaf species were generally higher for colder sites (Fig. 3a, b); the posterior distributions of the slope against MAT, βTk, deviated from 0 in the negative direction, and the 95% CI of βTk was less than zero for 30 of 72 EB species and for 15 of 128 DB species. The mean βTk for recruitment rate among species within a functional type (μTj) deviated in the negative direction, and its’ 95% CIs were negative for EB and DB (Fig. 4a). The 95% CI of βTk included zero for all TC species but the 50% CI was positive for two species (Fig. 3c). The 50% CI of μTj for TC included zero (Fig 4a). The 95% CI of βTk for all BC species and of μTj for BC were positive (Fig. 3d, Fig. 4a). Mortality of EB and DB species tended to be higher for the warmer sites, but the 95% CIs of βTk were positive for only two EB species and for none of the DB species (Fig. 3e, 3f). Although the 95% CIs of μTj included zero for all functional types, the 50% CIs were positive for EB and TC (Fig. 4b). Population growth rates of EB and DB species were generally higher for the colder sites (Fig. 3i, j). The 95% CIs of βTk for four EB species and two DB species were negative. The 95 CI of μTj for DB and the 50% CI for EB were also negative (Fig. 4c). The mean within a functional type for effects of successional stage (μSj) for recruitment rate and population growth rate deviated in the positive direction for EB, and the 95% CIs for recruitment rate and 50% CIs for population growth rate were both positive (Fig. 4d, f). The 50% CI of μSj for mortality was negative for EB (Fig. 4e). In contrast, for TC, the 95% CI of μSj for recruitment rate was positive, and the 50% CI for mortality was negative (Fig. 4d, 4e). In DB, the 50% CIs of μSj included zero for all demographic parameters.
This article is protected by copyright. All rights reserved.
Accepted Article
Discussion Our results indicate on-going, region-wide directional changes in tree species abundances in Japan. Major changes include: (1) an increase in evergreen broad-leaved trees in the mid-temperature region (5-15 °C in MAT), especially in secondary forests, coinciding with a decrease in deciduous broad-leaved trees and temperate conifers and (2) an increase in deciduous broad-leaved trees increased in cold regions (MAT < 5 °C), coinciding with a decrease in boreal conifers. These functional-type changes can be explained by the demography of the composing species: recruitment rate of the composing species was generally higher and mortality rate tended to be lower toward a colder temperature. This is the first report of relative abundances of tree species changing directionally along a temperature gradient, at a regional scale; distributions of forest biomes and tree species in Japan seem to be moving towards higher altitudes and northern latitudes. Besides temperature, other climatic factors such as precipitation are known to cause change in abundance and distribution of tree species in many regions (Feeley et al., 2011; Fauset et al., 2012). In Japan, the duration of snow cover and/or maximum snow depth are known to affect the distributions of some trees (Shidei, 1979; Gransert, 2004). However, precipitation and snowfall are largely related with temperature across Japan; warmer sites tend to have higher precipitation and shallower snow cover (Table S1). Furthermore, due to sufficient precipitation in Japan, distributions of biomes and tree species are largely dependent on temperature rather than precipitation (Kira, 1977) and biogeographical turnover of most tree species across the Japanese archipelago is strongly controlled by mean annual temperature (Kubota et al., 2014). Therefore, although the observed changes may be partly attributed to other climatic factors such as duration of snow cover, the changes are likely to be basically temperature related.
This article is protected by copyright. All rights reserved.
Accepted Article
Our results were consistent with expectations of how plant species respond to global warming: local abundance will increase near colder range boundaries and decrease near warmer boundaries (Iverson & Prasad, 1998; Morin et al., 2008; Walther, 2010). However, as the results were based on relatively short-term observations (less than 10-years), it is unclear whether the tree species are responding to current or past climate changes. As tree species comprise long-lived, sessile organisms, their ability to track climate changes is limited (Svenning et al., 2008). Therefore, there will be a delay in species abundance and distribution to reach the potential equilibrium after the past climate changes. It is reported that many European tree species are not recovering their climatic potential ranges, suggesting post-glacial migration lag (Svenning & Skov, 2004; Normand et al., 2011). Distributions of tree species may have also been affected by the Little Ice Age (LIA) 200-300 years ago, during which Japan experienced a cold and rainy climate (Maejima & Tagami, 1983; Bradley & Jones, 1993; Yamaguchi et al., 2008). For example, it is hypothesized that the beech species Fagus crenata increased in abundance in the south eastern Japanese archipelago during the LIA, and has been decreasing due to a lack of regeneration after the end of the LIA (Shimano, 2002). Evergreen broad-leaved trees showed a greater increase in secondary forests than in old-growth forests, which suggests that the shift to evergreen broad-leaved forest can be attributed in part to forest succession after past anthropogenic disturbance. Evergreen broad-leaved shrubs and small trees, such as Cleyera japonica and Eurya japonica, as well as evergreen broad-leaved canopy trees, such as Castanopsis sieboldii, increased at those sites (see Fig. S1). These species are thought to have originally covered a wide area of the mid-temperature region before the settlement of ancient Japanese people; human activity is thought to have greatly reduced them to their current distribution and abundance (Numata, 1974; Miyawaki, 1984). As managed forests have been abandoned, vegetation succession
This article is protected by copyright. All rights reserved.
Accepted Article
precipitation, frequency of super typhoon strikes) will accelerate or decelerate the process. Long-term and regional-scale network monitoring systems will be required to predict precisely how future climate change will affect the distribution and abundance of tree species.
Acknowledgements The data used here was provided by the Ministry of the Environment Monitoring Sites 1000 Project (TreeDataPackage2012ver1.zip, downloaded from http://www.biodic.go.jp/moni1000/findings/data/index.html on April 22, 2014). We thank its participant organizations and their staffs for data collection and site management efforts. We are also grateful to Prof. Tsutom Hiura and Dr. Shigeru Niwa for their helpful comments and Dr. Ikuyo Saeki for her detailed comments on an earlier version of this manuscript. This study was partially funded by the Environmental Research and Technology Development Fund (S-9-3) of the Ministry of the Environment, Japan.
References Ackerly DD, Loarie SR, Cornwell WK, Weiss SB, Hamilton H, Branciforte R, Kraft NJB (2010) The geography of climate change: implications for conservation biogeography. Diversity and Distributions, 16, 476–487. Bell DM, Bradford JB, Lauenroth WK (2014) Early indicators of change: divergent climate envelopes between tree life stages imply range shifts in the western United States. Global Ecology and Biogeography, 23, 168–180. Bradley RS, Jones PD (1993) “Little Ice Age” summer temperature variations: their nature and relevance to recent global warming trends. Holocene, 3-4, 367–376. Chen I-C, Hill JK, Ohlemüller R, Roy DB, Thomas CD (2011) Rapid range shifts of species associated with high levels of climate warming. Science, 333, 1024–1026.
This article is protected by copyright. All rights reserved.
Accepted Article
precipitation, frequency of super typhoon strikes) will accelerate or decelerate the process. Long-term and regional-scale network monitoring systems will be required to predict precisely how future climate change will affect the distribution and abundance of tree species.
Acknowledgements The data used here was provided by the Ministry of the Environment Monitoring Sites 1000 Project (TreeDataPackage2012ver1.zip, downloaded from http://www.biodic.go.jp/moni1000/findings/data/index.html on April 22, 2014). We thank its participant organizations and their staffs for data collection and site management efforts. We are also grateful to Prof. Tsutom Hiura and Dr. Shigeru Niwa for their helpful comments and Dr. Ikuyo Saeki for her detailed comments on an earlier version of this manuscript. This study was partially funded by the Environmental Research and Technology Development Fund (S-9-3) of the Ministry of the Environment, Japan.
References Ackerly DD, Loarie SR, Cornwell WK, Weiss SB, Hamilton H, Branciforte R, Kraft NJB (2010) The geography of climate change: implications for conservation biogeography. Diversity and Distributions, 16, 476–487. Bell DM, Bradford JB, Lauenroth WK (2014) Early indicators of change: divergent climate envelopes between tree life stages imply range shifts in the western United States. Global Ecology and Biogeography, 23, 168–180. Bradley RS, Jones PD (1993) “Little Ice Age” summer temperature variations: their nature and relevance to recent global warming trends. Holocene, 3-4, 367–376. Chen I-C, Hill JK, Ohlemüller R, Roy DB, Thomas CD (2011) Rapid range shifts of species associated with high levels of climate warming. Science, 333, 1024–1026.
This article is protected by copyright. All rights reserved.
Accepted Article
Colwell RK, Brehm G, Cardelús CL, Gilman AC, Longino JT (2008) Global warming, elevational range shifts, and lowland biotic attrition in the wet tropics. Science, 322, 258–261. Fauset S, Baker TR, Lewis SL et al. (2012) Drought-induced shifts in the floristic and functional composition of tropical forests in Ghana. Ecology Letters, 15, 1120–1129. Feeley KJ, Davies SJ, Perez R, Hubbell STP, Foster RB (2011) Directional changes in the species composition of a tropical forest. Ecology, 92, 871–882. Gansert D (2004) Treelines of the Japanese Alps - altitudinal distribution and species composition under contrasting winter climates. Flora, 199, 143–156. Henbo Y, Itaya A, Nishimura N, Yamamoto S-I (2006) Long-term canopy dynamics analyzed by aerial photographs and digital elevation data in a subalpine old-growth coniferous forest. Ecoscience, 13, 451–458. Hickling R, Roy DB, Hill JK, Fox R, Thomas CD (2006) The distributions of a wide range of taxonomic groups are expanding polewards. Global Change Biology, 12, 450–455. Hiroki S (2001) Invation of Quercus variabilis and Castanopsis cuspidata seedlings into secondary forests on the western slope of Mt. Miroku and adjacent hill in Kasugai City, Aichi Prefecture, Japan. Vegetation Science, 18, 31–37. Hoshino D, Nishimura N, Yamamoto S-I (2003) Effects of canopy conditions on the regeneration of major tree species in an old-growth Chamaecyparis obtusa forest in central Japan. Forest Ecology and Management, 175, 141–152. Ishihara MI, Suzuki SN, Nakamura M et al. (2011) Forest stand structure, composition, and dynamics in 34 sites over Japan. Ecological Research, 26, 1007–1008. Iverson LR, Prasad AM (1998) Predicting abundance of 80 tree species following climate change in the eastern United States. Ecological Monographs, 68, 465–485.
This article is protected by copyright. All rights reserved.
Accepted Article
Kamada M, Nakagoshi N (1990) Pattern and processes of secondary vegetation at a farm village in southwestern Japan after the 1960s. Japanese Journal of Ecology, 40, 137–150 (in Japanese with English summary). Kelly AE, Goulden ML (2008) Rapid shifts in plant distribution with recent climate change. Proceedings of the National Academy of Sciences of the United States of America, 105, 11823–11826. Kira T (1977) A climatological interpretation of Japanese vegetation zones. In: Vegetation science and environmental protection (eds. Miyawaki A, Tüxen R), pp. 21–30, Maruzen, Tokyo, Japan. Kubota Y, Hirao T, Fujii SJ, Shiono T, Kusumoto B (2014) Beta diversity of woody plants in the Japanese archipelago: The roles of geohistorical and ecological processes. Journal of Biogeography, 41, 1267–1276. Laurance WF, Oliveira AA, Laurance SG et al. (2004) Pervasive alteration of tree communities in undisturbed Amazonian forests. Nature, 428, 171–175. Lewis SL, Phillips OL, Baker TR et al. (2004) Concerted changes in tropical forest structure and dynamics: evidence from 50 South American long-term plots. Philosophical transactions of the Royal Society of London, Series B, 359, 421–36. Lewis SL, Lloyd J, Sitch S, Mitchard ETA, Laurance WF (2009) Changing ecology of tropical forests: evidence and drivers. Annual Review of Ecology, Evolution, and Systematics, 40, 529–549. Maejima I, Tagami Y (1983) Climate of Little Ice Age in Japan. Geographical Reports of Tokyo Metropolitan University, 18, 91–111. Van Mantgem PJ, Stephenson NL, Byrne JC et al. (2009) Widespread increase of tree mortality rates in the western United States. Science, 323, 521–524.
This article is protected by copyright. All rights reserved.
Accepted Article
Matías L, Jump AS (2014) Asymmetric changes of growth and reproductive investment herald altitudinal and latitudinal range shifts of two woody species. Global Change Biology, early view articles, doi: 10.1111/gcb.12683. Miyawaki A (1984) A vegetation-ecological view of the Japanese Archipelago. Bulletin of Institute of Environmental Science and Technology, Yokohama National University, 11, 85–101. Mori AS (2011) Ecosystem management based on natural disturbances: hierarchical context and non-equilibrium paradigm. Journal of Applied Ecology, 48, 280–292. Morin X, Viner D, Chuine I (2008) Tree species range shifts at a continental scale: new predictive insights from a process-based model. Journal of Ecology, 96, 784–794. Nakao K, Matsui T, Horikawa M, Tsuyama I, Tanaka N (2010) Assessing the impact of land use and climate change on the evergreen broad-leaved species of Quercus acuta in Japan. Plant Ecology, 212, 229–243. Normand S, Ricklefs RE, Skov F, Bladt J, Tackenberg O, Svenning J-C (2011) Postglacial migration supplements climate in determining plant species ranges in Europe. Proceedings of the Royal Society, Series B, 278, 3644–53. Numata M (ed.) (1974) The Flora and Vegetation of Japan. Kodansha, Tokyo. Ouédraogo D-Y, Mortier F, Gourlet-Fleury S, Freycon V, Picard N (2013) Slow-growing species cope best with drought: evidence from long-term measurements in a tropical semi-deciduous moist forest of Central Africa. Journal of Ecology, 101, 1459–1470. Parmesan C, Yohe G (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature, 421, 37–42. Peng C, Ma Z, Lei X et al. (2011) A drought-induced pervasive increase in tree mortality across Canada’s boreal forests. Nature Climate Change, 1, 467–471.
This article is protected by copyright. All rights reserved.
Accepted Article
Perry GLW (2002) Landscapes, space and equilibrium: shifting viewpoints. Progress in Physical Geography, 26, 339–359. Phillips OL (1998) Changes in the carbon balance of tropical forests: evidence from long-term plots. Science, 282, 439–442. Phillips OL, Malhi Y, Vinceti B et al. (2002) Changes in growth of tropical forests: evaluating potential biases. Ecological Applications, 12, 576–587. Phillips OL, Baker TR, Arroyo L et al. (2004) Pattern and process in Amazon tree turnover, 1976-2001. Philosophical transactions of the Royal Society of London, Series B, 359, 381–407. Plummer M, Best N, Cowles K, Vines K (2006) CODA: Convergence Diagnosis and Output Analysis for MCMC. R News, 6, 7–11. R Core Team (2014) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.r-project.org/. Saito O, Hoshino Y, Tsuji S, Kanno A (2003) Changes in species richness and species composition of Quercus serrata secondary forests after more than 20 years in Kanto region, Japan. Vegetation Science, 20, 83–96 (in Japanese with English summary). Sheil D, Jennings S, Savill P (2000) Long term permanent plot observations of vegetation dynamics in Budongo, a Ugandan rain forest. Journal of Tropical Ecology, 16, 765–800. Shidei T (1979) Distribution of special forest vegetation in heavy snowy regions in Japan. In: Vegetation und Landschaft Japans (eds Miyawaki A, Okuda S), PP. 93–99. The Yokohama Phytosociological Society, Yokohama. Shimano K (2002) Regeneration dynamics, causal factors, and characteristics of Pacific Ocean-type beech (Fagus crenata) forests in Japan: a review. Folia Geobotanica, 37, 275–296.
This article is protected by copyright. All rights reserved.
Accepted Article
Sprugel DG (1991) Disturbance, equilibrium, and environmental variability: What is “natural” vegetation in a changing environment ? Biological Conservation, 58, 1–18. Suzuki SN, Kachi N, Suzuki J-I (2013) Spatial variation of local stand structure in an Abies forest, 45 years after a large disturbance by the Isewan typhoon. Journal of Forest Research, 18, 139–148. Svenning J-C, Skov F (2004) Limited filling of the potential range in European tree species. Ecology Letters, 7, 565–573. Svenning J, Normand S, Skov F (2008) Postglacial dispersal limitation of widespread forest plant species in nemoral Europe. Ecography, 31, 316–326. Tamate S, Kashiyama T, Sasanuma T, Takahashi K, Matsuoka H (1977) On the distribution map of forest wind damage by typhoon No 15, 1954 in Hokkaido. Bulletin of the Government Forestry Experiment Station, 289, 43–67 (in Japanese with English summary). Tanaka N, Matsui T, Yagihashi T, Taoda H (2006) Climatic controls on natural forest distribution and predicting the impact of climate warming: Especially referring to Buna (Fagus crenata) Forests. Global Environmental Research, 10, 151–160. Thomas CD (2010) Climate, climate change and range boundaries. Diversity and Distributions, 16, 488–495. Totman C (1998) The Green Archipelago: Forestry in Preindustrial Japan. Ohio University Press, Athens. Walther G-R (2010) Community and ecosystem responses to recent climate change. Philosophical transactions of the Royal Society of London, Series B, 365, 2019–2024. Yamaguchi S, Naruse R, Shiraiwa T (2008) Climate reconstruction since the Little Ice Age by modelling Koryto glacier, Kamchatka Peninsula, Russia. Journal of Glaciology, 54, 125–130.
This article is protected by copyright. All rights reserved.
Accepted Article
Supporting Information Appendix S1: Structure of the hierarchical Bayesian model Table S1: Details of study sites Table S2: Analysed species family names and functional types. Table S3: AICs of models for functional-type–level analysis Figure S1: Observed recruitment rate and mortality of species
Figure legends Fig. 1 Locations of study sites in Japanese archipelago.
Fig. 2 Relative abundance and change in relative abundance for four tree functional types along mean annual temperature. For each functional type, only plots in which at least 20 stems were present at first census are included in the change in relative abundance data (e–f). Circle, old-growth forest; triangle, secondary forest. Regression curves estimated by the best model are shown; solid and dashed lines indicate old-growth and secondary forests, respectively, when forest successional stage was selected as an explanatory variable. When successional stage was not selected as an explanatory variable, regression curve common for old-growth and secondary forests is shown.
Fig. 3 Estimated recruitment rate, mortality, and population growth rate of evergreen broad-leaved, deciduous broad-leaved, temperate conifer, and boreal conifer species along mean annual temperature. Each line indicates the response curve for a species to temperature, with an estimated posterior median of parameters for old-growth forests (S = 0). Black line, response curve for the species whose 95% credible interval of the slope parameter (βTk) did not include zero; dark gray line, curve for the species whose 50% credible interval did not
This article is protected by copyright. All rights reserved.
Accepted Article
include zero; light gray line, curve for the species whose 50% credible interval included zero.
Fig. 4 Posterior density of functional-type means of the effect of mean annual temperature (μTj) and forest successional stage (μSj) on (a, d) recruitment rate, (b, e) mortality, and (c, f) population growth rate of species. Means are those among species within a functional type. Black solid line, 95% credible intervals did not include zero; gray solid line, 50% credible intervals did not include zero, gray dashed line, 50% credible intervals included zero. EB, evergreen broad-leaved; DB, deciduous broad-leaved; TC, temperate conifer; BC, boreal conifer.
This article is protected by copyright. All rights reserved.
Recruitment rate Mortality rate
Temperate conifer
Deciduous broadleaves
(a)
Boreal conifer
(c)
(b)
(d)
0.10
0.10
0.10
0.10
0.05
0.05
0.05
0.05
0.00
0.00
0.00 5
10
15
20
5
10
15
(e)
0.00 5
20
10
15
20
5
0.10
0.10
0.05
0.05
0.05
0.05
15
20
5
10
15
(i)
5
20
10
15
5
20
1.05
1.05
1.00
1.00
1.00
1.00
15
20
5
10
15
20
20
0.95
0.95
0.95 10
15
(l)
1.05
5
10
(k)
(j)
1.05
0.95
20
0.00
0.00
0.00 10
15
(h)
0.10
5
10
(g)
(f)
0.10
0.00
Population growth rate
Accepted Article
Evergreen broadleaves
5
10
Mean Annual Temperature (°C)
This article is protected by copyright. All rights reserved.
15
20
5
10
15
20
Accepted Article This article is protected by copyright. All rights reserved.
