A spatially-explicit empirical model for assessing conservation values of conifer plantations

A spatially-explicit empirical model for assessing conservation values of conifer plantations

Forest Ecology and Management 444 (2019) 393–404 Contents lists available at ScienceDirect Forest Ecology and Management journal homepage: www.elsev...

4MB Sizes 0 Downloads 8 Views

Forest Ecology and Management 444 (2019) 393–404

Contents lists available at ScienceDirect

Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco

A spatially-explicit empirical model for assessing conservation values of conifer plantations

T

Yuichi Yamauraa,b,c, , David Lindenmayerb, Yusuke Yamadad, Hao Gongd, Toshiya Matsuurad, Yasushi Mitsudae, Takashi Masakia ⁎

a

Department of Forest Vegetation, Forestry and Forest Products Research Institute, 1 Matsunosato, Tsukuba, Ibaraki 305-8687, Japan Fenner School of Environment and Society, Australian National University, Canberra, ACT 2601, Australia c Shikoku Research Center, Forestry and Forest Products Research Institute, 2-915 Asakuranishi, Kochi, Kochi 780-8077, Japan d Department of Forest Management, Forestry and Forest Products Research Institute, 1 Matsunosato, Tsukuba, Ibaraki 305-8687, Japan e Faculty of Agriculture, University of Miyazaki, 1-1 Gakuen Kibanadai Nishi, Miyazaki, Miyazaki 889-2192, Japan b

ARTICLE INFO

ABSTRACT

Keywords: Indicator Mapping Native broad-leaved tree Naturalness index Plantation species Planted tree density

Plantations are expanding globally and dominate landscapes in many parts of the world. Biodiversity conservation within plantations is becoming an important issue and developing indicators of conservation values is valuable. Although plantations support less biodiversity than natural forests, older plantations can provide habitat for some native trees and associated biota. The amount (basal area) of native trees can be a simple indicator of conservation value of plantations. Various factors are likely to affect rates of increase in native trees in plantations with stand age. We developed an empirical model to predict the amount of broad-leaved trees as an indicator of conservation value of conifer plantations. We quantified relationships between rates of increase in the amount of broad-leaved trees and plantation tree species, density of planted trees, climate, topography and landscape cover. We used a hierarchical modeling framework based on extensive snapshot plot data (n = 3265 plots) from the national forest inventory in Japan. Our results showed that plantation tree species had the largest effect on the rate of increase in broad-leaved trees. Japanese cedar Cryptomeria japonica and hinoki cypress Chamaecyparis obtuse, which are two primary plantation species in Japan (both from the cypress family), had low rates of increase. In plantations of other species (red pine, larch, fir and spruce from the pine family), broad-leaved trees started to increase in amount after 20 years. We found that 50-year-old plantations of the pine family supported 10–20% of the amount of broad-leaved trees typically found in old-growth natural forests. Planted tree density also had important, but nonlinear relationships with the rate of increase in broad-leaved trees. Stands with fewer planted trees had higher rates of increase in broad-leaved trees. Rates of increase also were associated with snow depth, temperature, slope angle and the amount of natural forest in the surroundings. Our results suggest that management practices related to stand age, selection of plantation species and the density of planted trees can contribute to the conservation value of plantations. As our model is based on broadly available covariates and accommodates stand age, it may be applicable to other regions under different management regimes.

1. Introduction Plantations are areas supporting trees established through deliberate planting or deliberate seeding (FAO, 2006). Plantations have expanded by more than 110 million ha since 1990, and now cover 290 million ha worldwide while natural forests cover 3713 million ha (FAO, 2015). In 2012, 46.3% of global industrial round wood was derived from plantations (Payn et al., 2015). Although more than half of the world’s plantations (56%) occur in the temperate zone (Payn et al.,



2015), plantations are also expanding in subtropical and tropical zones, notably by replacing natural forests (Puyravaud et al., 2010; Hua et al., 2018). Indeed, 65% of industrial round wood comes from plantations in tropical and subtropical zones (Payn et al., 2015). Plantations are sometimes called “tree farms” and are typically composed of single tree species that are regularly spaced and even-aged to efficiently produce high wood volumes (Paquette and Messier, 2010). Their simplified tree species composition and stand structure, as well as impoverished biodiversity (Newbold et al., 2015), has led to

Corresponding author at: Shikoku Research Center, Forestry and Forest Products Research Institute, 2-915 Asakuranishi, Kochi, Kochi 780-8077, Japan. E-mail address: [email protected] (Y. Yamaura).

https://doi.org/10.1016/j.foreco.2019.04.038 Received 18 February 2019; Received in revised form 17 April 2019; Accepted 20 April 2019 Available online 07 May 2019 0378-1127/ © 2019 Elsevier B.V. All rights reserved.

Forest Ecology and Management 444 (2019) 393–404

Y. Yamaura, et al.

Table 1 Possible effects of modeled covariates on rates of increase in BA of broad-leaved trees in conifer plantations. Covariate Climate Snow depth Warmth index (growth period) Topography Slope angle Terrain openness (positive openness) Catchment area Solar radiation Landscape Proportion of natural forests Planted trees Density of planted trees Identity of planted species

Possible effects (reason)

Refs.

(as snow depth increases, growth periods shorten) • Negative (as snow depth increases, mortality of planted trees increases [latter is especially the • Positive case for Japanese cedar]) (as the temperature increases, growth period increases) • Positive (many tree species have a common optimal climate when the sampled environmental • Unimodal conditions vary widely) (steep slope induces soil movement and tree fall) • Negative (concave terrain [small openness] induces soil movement and tree fall, whereas • Unimodal convex terrain [high openness] induces water and nutrient deficiencies. Flat terrain has a high forest growth rate due to high soil moisure and nutrients)

(lower part of the slope has high tree growth rate due to high soil moisture and • Positive nutrients) • Negative (greater solar radiation induces water deficiency in the temperate zone) (distance-dependent seed rain from natural forests enhance the recruitment of native • Positive trees)

(non-planted trees in plantations compete for light with planted trees or grow slowly • Negative or disappear due to suppression by planted trees) and cypress species have denser foliage than pine species and therefore limit • Cedar regeneration of native species by little light penetleration into the floor

Peterson and Peterson (2001); Masaki et al. (2004) McKenney and Pedlar (2003); Nothdurft et al. (2012)

Guariguata (1990) Hunter and Parker (1993); Chen et al. (1998); Curt et al. (2001) Chen et al. (1998); Curt et al. (2001) Chen et al. (1998); Mitsuda et al. (2007) Brunet (2007); Utsugi et al. (2006) Potvin and Dutilleul (2009); Seiwa et al. (2012) Fimbel and Fimbel (1996); Saito (1989); Kiyono (1990)

associated fauna (Ohsawa and Shimokawa, 2011). Humphrey (2005) suggested that British spruce plantations attain some features seen in old-growth natural forests after 80–100 years of plantation growth. However, as growth rates of plantations are influenced by climate and topography (Mitsuda et al., 2007; Nothdurft et al., 2012), rates of increase in the amount of native trees will likely depend on environmental conditions. Rates of increase in native trees may also differ among plantation tree species. For example, although tropical plantations that are less than 40 years old can support non-planted native trees, the amount of native trees varies among plantation species (Fimbel and Fimbel, 1996; Harrington and Ewel, 1997). Native trees also regenerate after thinning of plantations (Seiwa et al., 2012) and their abundance can be high in plantations with fewer planted trees (Harrington and Ewel, 1997). Surrounding landscape structure is another factor affecting the occurrence of native trees in plantations since adjacent natural forests can be a source of seeds (Utsugi et al., 2006; Brunet, 2007). To our knowledge, no studies have simultaneously quantified the effects of stand attributes, landscape structure as well as topography and climate on the amount of non-planted native trees in plantations. We developed an empirical model to predict the amount of broadleaved trees as an indicator of conservation value of conifer plantations. Using a nation-wide tree plot data (n = 3265 plots) for Japan, we quantified relationships between the rates of increase in broad-leaved trees as the function of plantation tree species, the density of planted trees, climate, topography and landscape attributes. We then examined the relative influence of these covariates on the rates of increase in broad-leaved trees in plantations. We tested a series of hypotheses (Table 1) to answer the question: How do environmental covariates and stand properties affect rates of increase in broad-leaved trees in conifer plantations? Our model is based on readily available geographical covariates, making it broadly applicable to other regions supporting plantations around the world.

them being labelled “green deserts” (Koh and Gardner, 2010). Conversely, it has been suggested that establishment of plantations with high levels of wood production may enable natural forests to be spared from logging (Sedjo and Botkin, 1997). Such spatial separation of wood production and conservation has been successful in New Zealand (Brockerhoff et al., 2008). Nevertheless, biodiversity conservation in plantations can be important for various reasons: (i) plantations are often established in productive environments which potentially also support large amounts of biodiversity (Franklin, 1993; Yamaura et al. unpublished manuscript), (ii) many landscapes are already dominated by plantations (Hartley, 2002), (iii) establishment of plantations may not reduce logging pressure in natural forests without concerted efforts to directly protect natural forests (Paquette and Messier, 2010), (iv) biodiversity conservation can sometimes be enhanced in plantations at little additional cost to management (Norton, 1998). In addition, even a small increase in tree species diversity in plantations may increase their tree volume productivity (annual increment in stem volume) via positive biodiversity-productivity relationships (Liang et al., 2016), enhance their economic values (Messier et al., 2015), and promote social (sum of market and non-market) values of plantations via improved biodiversity (Yamaura et al., 2016). Therefore, biodiversity conservation within plantations is becoming increasingly important globally (Yamaura et al., 2012a; Demarais et al., 2017). In these contexts, developing indicators of the conservation value of plantations is valuable to evaluate and promote biodiversity conservation in plantations (Spellerberg and Sawyer, 1996; Coote et al., 2013). Several studies have shown that plantations that have a well-developed native plant understory and support some native trees can provide habitat for a range of faunal species (Lindenmayer and Hobbs, 2004; Nájera and Simonetti, 2010). For example, increasing basal area (BA) of native broad-leaved trees promotes the diversity of beetles and birds in conifer plantations (Ohsawa, 2007; Yamaura et al., 2008; Lindbladh et al., 2017). Therefore, the amount of non-planted native trees may be a simple, albeit crude indicator of the conservation value of plantations. Plantation age is a key factor influencing the amount of native trees in plantations (Brockerhoff et al., 2008; Paquette and Messier, 2010). As plantations mature, spaces develop between the crowns of planted trees, which enhances the germination and growth of native trees (Fujimori, 2001). Old plantations can therefore support native trees and

2. Materials and methods 2.1. Japan’s plantation, plot data and modeling scheme Most plantations globally were established after the 1980s and are therefore less than 30 years old (FAO, 2006). However, many 394

Forest Ecology and Management 444 (2019) 393–404

Y. Yamaura, et al.

plantations in Japan were established by replacing natural forests and grasslands between the 1950s and 1980s (Yamaura et al., 2012a). Plantations now cover 28% of the land surface (natural forests cover 40% of the land), and are approaching the age of 50 years (Yamaura et al., 2012a; Forestry Agency, 2014). More than 95% of Japanese plantations are conifer plantations (Forestry Agency, 2017a) as in other temperate and boreal regions, especially in Europe and North America (FAO, 2006). A single conifer species mostly planted all at the same time. We constructed our empirical model using data from Japan’s National Forest Inventory (NFI). The NFI is composed of more than 13,000 permanent plots at 4 km-grid points across the country (Japanese Forestry Agency owns the data). Surveys of the NFI plots record species and DBH for trees with > 18 cm, > 5 cm, and > 1 cm DBH in 0.06 ha, 0.03 ha, and 0.01 ha plots, respectively, with smaller plots nested within larger plots (Hirata et al., 2016). We used plot data from plantations of the six major species in Japan (Japanese cedar Cryptomeria japonica, hinoki cypress Chamaecyparis obtuse, Japanese larch Larix kaempferi, Japanese red pine Pinus densiflora, Sakhalin fir Abies sachalinensis, and Sakhalin spruce Picea glehnii) collected between 2009 and 2013 for this study (3349 plots: see Appendix S1 for details of the treatment of NFI data). We excluded six planted species of conifer trees from the plot data and calculated BA of non-planted trees (mainly broad-leaved trees) for individual plots (Appendix S1-2). Japan’s natural forests are comprised primarily of deciduous and evergreen broad-leaved trees (cool- and warm-temperate forests: Kira, 1991). Deciduous broad-leaved forests include such species as Japanese beech Fagus crenata, deciduous oak (e.g., Quercus crispula), and other hardwood species (e.g., Tilia, Acer spp.) while evergreen oak (Quercus spp.), Castanopsis and Lithocarpus spp. comprise evergreen broad-leaved forests (Kira, 1991). We used BA of broad-leaved trees rather than the species richness of broad-leaved trees. This was because tree species richness exhibits regional variation within Japan (it increases in southern area: Kira, 1991). The density of broad-leaved trees is another candidate indicator of conservation value but its use can be problematic as tree density usually decreases with stand age (Nyland, 2016). In our plot data from conifer plantations, the top four non-planted species with the largest BA were deciduous broadleaved trees: Q. crispula, Q. serrata, Castanea crenata and Magnolia obovate. NFI data contains information about current stand structure and

stand age (Fig. 1). This means that even if different plots have the same rates of increase in broad-leaved trees, they can support different amounts of broad-leaved trees due to differences in stand age, and vice versa. Therefore, an analysis that fails to incorporate stand age can produce misleading results. We used hierarchical Bayesian analysis (Gelman and Hill, 2007; Royle and Dorazio, 2008) to infer unobserved increases in the amount of broad-leaved trees through which the observed states appeared as latent (hidden) processes. This is an important feature of hierarchical Bayesian methods. 2.2. Climate, topography and landscape covariates We used maximum depth of snow cover and a warmth index as climatic covariates in our modeling (Appendix S3-4), both of which are known to affect tree species composition across Japan (Kira, 1991; Nakashizuka and Iida, 1995). The warmth index corresponds to the period of plant growth, and was calculated by summing monthly mean temperatures higher than 5 °C (Kira, 1991). We used a 20-m digital elevation model (DEM) to generate data on terrain (topographical) covariates (slope angle, terrain openness, catchment area, and solar radiation: Appendix S3-4), which influence forest growth and structure (e.g., Guariguata, 1990; Mitsuda et al., 2007). Although other local factors (e.g., soil type, soil nutrient/wetness) can affect the growth of plantations (Klinka and Carter, 1990; Curt et al., 2001), these factors are influenced by topography and can be represented by DEM (Moore et al., 1993; de Castilho et al., 2006). We also note that in our preliminary analyses, we grouped soils and geology of the plots into eight soil and four rock types, respectively, which explained only the small amount of variations in BA of broad-leaved trees. We therefore did not further consider these covariates. We also quantified the proportion of natural forest within 100 m radius from each plot as a landscape covariate (Appendix S3-4). We used ArcGIS ver. 10.3 with Spatial Analyst Extension (ESRI, 2014) and SAGA GIS (Conrad et al., 2015) for geospatial analyses. 2.3. Statistical analysis 2.3.1. Model structure Our model comprised two parts. The first part deals with the density of planted trees, which decreases with stand age (Nyland, 2016). We allowed the rates of decline in planted tree density to depend on Fig. 1. Relevant stand development processes and modeling scheme. (a) (Left panel) Density of planted trees decreases with stand age but rates of decrease can differ among the sites depending on the environment and planted species. (Right panel) The amounts of broad-leaved trees (converted into ‘naturalness index’ ranging 0–1) increase with stand age while the rates of increase depend on the environment, planted tree density, and planted species. Hypothetical four tree plots (A–D) are shown; they have different stand age and stand structure while they are surveyed at the same time. (b) Model structure showing the relationships between environmental covariates and increase in broad-leaved trees. Rates of decrease in planted tree density and rates of increase in broad-leaved trees determine current stand structure given the specific stand age (stored in the NFI data). We reconstructed these hidden processes using hierarchical Bayesian method.

395

Forest Ecology and Management 444 (2019) 393–404

Y. Yamaura, et al.

environmental covariates. Since NFI does not systematically contain information on management history, we could not consider the effects of thinning on planted tree density. We next modeled the amount of broad-leaved trees as a function of stand age. We estimated the rate of increase in the amount as a function of the environment, density of planted trees, and the identity of plantation tree species. We developed our model for the amount on the basis of the framework of old-growth index provided by Larson et al. (2008). We first describe the latter submodel, and then that we developed for planted trees. We obtained an index of the amount of broad-leaved trees (that we term a ‘naturalness index’) in plantation plots as follows:

INAT, i =

xi x old

0

+ coeff _NATi × agei + eNAT, i

1, species [i]

+

k

(

k,1 xk , i

+

2 k ,2 xk , i )

(2)

cnf , i )

=

0, species [i]

+ coeff _densi × log(agei ) + ecnf , i

cnf 50, i / cnf , i

(5) (6) (7) (8)

2.3.3. Relative effects of nine covariates We quantified relative effects of the individual covariates on the rates of increase in BA of broad-leaved trees based on parameter estimates. We estimated rates of increase (Eq. (3)) for six planted tree species using species-specific intercepts (β1,species[i]) while other covariates were held constant at mean values. For each of the other continuous covariates, rates of increase for six equally spaced values within 95% percentiles were obtained (other continuous covariates were again held constant at the mean value) while plantation species was assumed to be cedar (the most common species). Although some environmental covariates were not included in Eq. (3), their effects were considered and quantified through the effects of planted tree density (Eqs. (4)–(6)). For the effects of planted tree density, values at six levels within the 95% percentile of the model-based prediction at 50 years old (adjNcnf50 in Eq. (8)) were obtained and used. We then obtained SDs of six rates of increase and treated them as the measures of relative effects of covariates.

(3)

where β1,species[i] is an intercept and can take different values depending on plantation tree species (with subscripts of species[i]). This term can express the possible suppression effects on the increase associated with the identity of plantation tree species. xk,i is kth covariates of ith site, and βk,1 and βk,2 are coefficients of xk,i and its squared term. Using a quadratic model, we considered possible non-linear effects of covariates (Appendix S6). Since the amount of broad-leaved trees was generally small in conifer plantations (more than 25% plots had ‘zero’ BA of broad-leaved trees), to simplify the model, we excluded three covariates (terrain openness, catchment area, and solar radiation) whose simple and quadratic effects were not significant in Eq. (3) (Appendix S3). Eq. (3) also includes the density of planted trees to consider their suppression (crowding) effects (Fig. 1). However, since planted tree densities (Ncnf,i) decrease with stand age, it was necessary to model this age-dependence given the variation in the stand age of surveyed plots. We assumed that Ncnf,i followed a Poisson distribution (Ncnf,i ∼ Poisson [λcnf,i]), and used the series of equations as follow:

log(

+ coeff _densi × log(50) + ecnf , i

=

+

2.3.2. Model fitting We estimated parameters in Eqs. (1)–(8) (e.g., α and β) using Markov Chain Monte Carlo with conventional vague priors (see Appendix S7 for detailed procedure). We standardized each covariate to enhance convergence (Kéry and Royle, 2016) and stabilize estimates, especially second-order terms (Schielzeth, 2010). Multi-collinearity did not qualitatively change parameter estimates in our model (Appendix S3). We also tested model performance by cross-validation. We randomly chose 90% of the data for model training and tested the model using the 10% of remaining data. We repeated this procedure 100 times and obtained coefficient of determination (R2) for INAT at logit scale at each repetition.

where agei is stand age of ith plot, β0 is an intercept, and coeff_NATi is a coefficient of stand age (can be specific to ith plot), which dictates rates of increase (Larson et al., 2008). The final term (eNAT,i) is an unexplained error term with a normal distribution. We also added or subtracted the minimum non-zero values, called ε, to the possible minimum or maximum values (Appendix S5). We examined the effects of environmental covariates on the rates of increase in BA of broad-leaved trees by modeling coeff_NATi as a function of covariates:

coeff _NATi =

0, species [i]

cnf 50, i )

k ,1, species [i] xk , i

Based on these formulations, we included the expected density of planted trees at 50 years old (adjNcnf50,i) as one of the five covariates in Eq. (3). Covariates for densities (Eq. (5)) comprised seven climatic, topographic, and landscape covariates. These formulations indicate that increases in the amount of broad-leaved trees with stand age can be mediated by the environment directly and indirectly via densities of planted trees. As densities of planted trees can be influenced by the environment differently among species, we used species-specific parameters in these functions dealing with densities of planted trees (Appendix S7). Since stands immediately after planting cannot have planted trees with more than 5 cm DBH (assuming that densities of planted trees monotonically decrease with the age), we used only those plots older than 15 years old (giving 3265 plots after excluding 84 plots). Our final dataset comprised 41 spruce, 119 pine, 349 larch, 1544 cedar, 272 fir and 940 cypress-dominated plots.

(1)

INAT, i]) =

log(

adjNcnf 50, i = ratei × Ncnf , i

where INAT,i is the naturalness index (scaled BA of broad-leaved trees) of ith plot, xi is the BA of broad-leaved trees at ith plot, xyoung is a median value of BA of broad-leaved trees for young plantations derived from the plot data, and xold is a median value of BA for qualified old-growth natural forests in Japan. We used median values of BA rather than mean values to reduce the effects of outliers. We set xyoung using plantation plots ≤10 years old (Appendix S2). To obtain xold, we used 18 permanent plots registered as old-growth deciduous and evergreen broadleaved forests (more than 150 years old) in Japan (Ishihara et al., 2010). This simplification enabled us to analyze the national dataset using a single criterion (median value) of xold as a reference value (Appendix S2). We assigned xyoung and xold for the values of xi less than xyoung and larger than xold, respectively, and delimited INAT from 0 to 1. We then examined the effects of environmental covariates on the rates of increase in INAT along with stand age. Following Larson et al. (2008) and Warton and Hui (2011), we logit-transformed the proportional data of INAT, and regressed stand age (explanatory variable) on these logit-transformed data (as response variable):

log(INAT, i /[1

+

ratei =

x young x young

1, species [i]

k

(

2 k ,2, species [i] xk , i )

coeff _densi =

2.3.4. Spatial mapping We created spatial maps of the predicted values for the naturalness index (INAT: scaled BA of broad-leaved trees) for plantations in the northern Ibaraki prefecture, central Japan (that includes the cities of Kitaibaraki and Takahagi). This montane area has large areas of conifer plantations (Yamaura et al., 2009). We compiled the relevant environment covariates based on the above described methods for every 20-m resolution grid. Stand covariates other than density of planted trees were derived from forest registers maintained by the offices of prefecture and national forests. We then produced a map of predicted values (given the current stand age) of INAT for conifer plantations.

(4) 396

Forest Ecology and Management 444 (2019) 393–404

Y. Yamaura, et al.

Fig. 2. Modeled dependency of decreasing rates of planted tree density in conifer plantations on seven environmental covariates. Rates of decrease in planted tree (based on Eq. (5)) are separately shown for individual plantation species, and only 95% percentiles for their ranges are shown. Other covariates were held constant at mean values.

3. Results

been planted in the warmest areas of Japan while fir and spruce in the coolest areas of the country (Fig. 2b). Larch also has been planted in cool areas, and pine is planted in mid-temperature locations. The density of plantation trees declined with stand age (indicated by coeff_densi in Eq. (4)) but density varied among plantation tree species

3.1. Modeling density of planted trees An assessment of the NFI data showed that cedar and cypress have 397

Forest Ecology and Management 444 (2019) 393–404

Y. Yamaura, et al.

Fig. 3. Inferred decreases in planted tree densities in conifer plantations against stand age. Curves with different covariate values (within 95% percentiles for individual species) are depicted, and other covariates were held constant at mean values. (h) Individual plots are differently shown by their planted tree species (see Appendix S8 for semi-log plots).

(Figs. 2 and 3). For example, along with stand age, cypress plantations maintained a relatively high density of planted trees while density quickly declined over time in spruce plantations (Fig. 3). Although cedar density declined with increasing snow depth (Figs. 2a, 3a), other covariate effects on other species were generally weak (Figs. 2 and 3). Many plots that were 40–50 years old had high densities (more than 1000 trees/ha), that were not well explained by the model (Fig. 3; Appendix S8).

extremes of temperature (not intermediate values), higher slope angle, larger amount of natural forest in the surrounding area, and lower densities of planted trees (Fig. 4). There were negative effects of snow depth on planted tree density, and when we considered the influence of covariates on the density of planted trees, the effects of snow depth on broad-leaved trees were intensified (Fig. 5). Inferred logistic curves for the naturalness index (the scaled amount of broad-leaved trees) also highlighted the importance of these covariates (Fig. 6). R2 was 0.20 and cross-validation also produced similar values (mean ± SD: 0.18 ± 0.04). Plots that were 40–50 years old had high values for the naturalness index (more than 0.2), which were not well explained by the model (Fig. 6; Appendix S8).

3.2. Modeling the amount of broad-leaved trees Among the nine covariates, plantation tree species had the strongest effect on the rates of increase in the amount of broad-leaved trees (Table 2); cedar and cypress were characterized by the low rates of increase, indicating that broad-leaved trees increased slowly with stand age in these plantations (Fig. 4). Density of planted tree density also had an important influence on the amount of broad-leaved trees, as did snow depth, warmth index, slope angle and landscape structure (the amount of natural forest in the surrounding area). Effects of these five covariates were non-linear. Specifically, rates of increase in broadleaved trees were higher in plots with deep snow, high and low

3.3. Spatial mapping Spatial predictions showed that values of the naturalness index were relatively high in coastal areas where pine plantations (which support high rates of increase in the amount of broad-leaved trees) were established (Fig. 7). Some inland areas also supported areas with relatively high values for the naturalness index, although in a spatially heterogeneous pattern. These plantations were comprised of pine and larch which had high rates of increase in the amount of broad-leaved trees (Fig. 7; Appendix S9).

Table 2 Relative effects of nine covariates to the rates of increase in BA of broad-leaved trees. Covariate

Effect

Snow depth Warmth index Slope angle Terrain openness Catchment area Solar radiation Landscape structure Planted tree density Plantation species

0.011 0.011 0.006 0.002 0.001 0.001 0.004 0.013 0.021

4. Discussion Although biodiversity conservation in plantations is becoming increasingly important globally, to the best of our knowledge, there are no broadly available indicators of how the conservation values of plantations can vary in space and time. We suggest that the amount of native trees is a simple and albeit crude indicator of conservation value, which is ecologically meaningful, easily understandable, measurable, and accommodates management practices. We developed an empirical model as a function of stand properties, landscape, topography and climate using nationwide tree plot data. Our results showed that plantation species had the highest impacts on the rate of increase in native trees with stand age, followed by the density of planted trees and climate.

Effects of individual covariates were measured by SDs of rates of increase (Eq. (3)) for six different values. 398

Forest Ecology and Management 444 (2019) 393–404

Y. Yamaura, et al.

Fig. 4. Modeled dependency of rates of increase in broad-leaved trees in conifer plantations on five covariates. Rates of increase are based on Eq. (3), and for four covariates other than planted tree density, effects via planted tree density are not included (planted tree density is held constant at the mean value). (e) Range of planted tree density was according to the expected density of planted trees at 50 years old from the NFI data (adjNcnf50 in Eq. (8)).

4.1. Effects of plantation tree species

larch may be planted in warm areas even though their growth rates can be lowered. Nakajima et al. (1965) reported strong growth of planted larch in high elevation sites in southwestern Japan (Shikoku Island). Another option may be to simply retain (not harvest) plantations that enter an old-growth stage (long-rotation), which is a management option recommended in the United Kingdom (Kerr, 1999). However, as described above, a long time (more than 100 years) would be needed for cedar and cypress plantations to develop reasonable levels of broadleaved trees (e.g., INAT ∼ 0.1). In cedar and cypress plantations, the rates of increase in broadleaved trees were quite low, which may be partially explained by regional variations in silvicultural practices. In fir plantations established in Hokkaido, northern Japan, weeding only occurs around planted trees (Yamaura et al., 2018). Non-planted native trees can regenerate naturally in non-weeded areas (Appendix S10). Indeed, planting at wide spaces has been recommended as a conservation practice (Moore and Allen, 1999). Modification of silvicultural practices, including intensive thinning (see below), would promote conservation values of plantations, including those dominated by cedar and cypress.

Cedar and cypress are two primary plantation species in Japan and are from the cypress family. The rates of increase in broad-leaved trees within plantations of cedar and cypress were quite low; even 80–100 years would be too short to support 5–10% amount of broadleaved trees found in old-growth forests (Fig. 6). Dense foliage of these two plantation tree species is likely to suppress light penetration to the ground and hinder natural regeneration of broad-leaved trees (e.g., Fimbel and Fimbel, 1996). However, for other plantation tree species (from the pine family), 50-year-old plantations can support 10–20% amounts of native broad-leaved trees typically found in old-growth natural forests in Japan (Figs. 6 and 7). Other studies have shown that such levels of broad-leaved trees promote the diversity of beetle and bird communities in conifer plantations (Ohsawa, 2007; Lindbladh et al., 2017). Since Japanese plantations are generally harvested at ∼50 years old (Forestry Agency, 2017b), plantations of species from the pine family would attain certain levels of conservation value before harvesting (INAT ∼ 0.1–0.2). In Japan, plantation tree species have been selected based on the temperature of the locality; for example, cedar and cypress have been planted in warm areas (Fig. 2b). When biodiversity conservation is considered in plantation forestry, the inherent rates of increase in native trees may be considered prior to stand establishment. For example,

4.2. Effects of planted tree density Plantations with a reduced density of planted trees supported higher amounts of broad-leaved trees, and this result was consistent with our 399

Forest Ecology and Management 444 (2019) 393–404

Y. Yamaura, et al.

Fig. 5. Modeled dependency of rates of increase in broad-leaved trees in conifer plantations on eight covariates. Rates of increase are based on Eq. (3), and for covariates other than planted tree density, effects via planted tree density are included. That is, planted tree density is changed with the covariates following Eqs. (4) and (5) (without random site effects [ecnf]), and λcnf50 rather than adjNcnf50 was used in Eq. (3). (h) Panel of planted tree density is the same in Fig. 4e and shown for the comparison with other panels.

prediction at the outset of this study (Table 1). Functional forms of these effects took a concave shape (Fig. 4); given a reduction of 500 planted trees/ha (at 50 years old), the reduction from 1000 to 500 would promote the rates of increase in the amount of broad-leaved trees

greater than that from 2000 to 1500 trees (Fig. 4). As vacant areas created by intensive plantation thinning enhances the regeneration and development of native trees (Ishii et al., 2008; Seiwa et al., 2012), greatly reducing planted trees and creating open spaces for native trees 400

Forest Ecology and Management 444 (2019) 393–404

Y. Yamaura, et al.

Fig. 6. Inferred logistic curves for the amount of broad-leaved trees in conifer plantations. Curves with different covariate values (within 95% percentiles for individual species) are depicted. Although other covariates are held constant at mean values, effects via planted tree density are considered. (h) Planted tree density was not the function of environmental covariates, and its range was according to the expected density of planted trees at 50 years old from the NFI data. See Figs. 3 and 4 for the colors of six planted species.

covariates were limited, indicating that planted trees can be well maintained by silvicultural practices under various environmental conditions. Conversely, there was large amounts of unexplained variation in planted tree densities (Fig. 3). As thinning is common in plantations, considering management practices would be an important next step in refining our empirical model.

would have conservation benefits. 4.3. Environmental effects on amount of broad-leaved trees We found that the amount of broad-leaved trees was higher on sites characterized by deep snow, cool temperatures, and steep slopes. These areas are known to have low growth rates of plantations (Table 1) and therefore our results were unexpected. Extensive (low intensity) plantation management such as retaining non-planted trees in those areas may have enhanced the increase in broad-leaved trees. That is, these environmental conditions likely indicate management intensity rather than growth of plantations. For example, Japanese natural forests, especially old growth forests, are now restricted primarily to harsh environments with steep slopes and deep snow (Yamaura et al., 2011, unpublished manuscript). We detected positive effects of the amount of surrounding natural forests on the increase of broad-leaved trees, which is also consistent with our predictions at the outset of this study (Table 1). There would be colonization opportunities for native trees in plantations adjacent to natural forests. Notably, such landscape effects were non-linear, and small areas of adjacent natural forest could have a non-trivial influence on the amount of broad-leaved trees occurring within plantations. Retention of small patches of native trees or natural forests throughout plantation landscapes may promote conservation values of large areas of plantations (Lindenmayer et al., 2002, in press; Brudvig et al., 2009), and small remnant natural forests in plantation landscapes may have larger values than formerly presumed (Fischer and Lindenmayer, 2002; Fahrig, 2017).

4.5. Management of broad-leaved trees in conifer plantations Broad-leaved trees can regenerate in clear-cuts depending on the kind of logging operations employed during the harvesting (Deal et al., 2004; Hanley et al., 2006). Japanese plantation forestry involves clearcutting, site preparation, planting single conifer species, and weeding. These intensive practices intentionally remove advanced and regenerating broad-leaved trees. However, as shown in this study, broad-leaved trees can occur in mature plantations of species from the pine family (Figs. 5 and 6), and thinning practices can enhance recruitment of broad-leaved trees even for cedar and cypress plantations (Utsugi et al., 2006). Retention of broad-leaved trees during harvesting, including clearcutting would therefore be a useful option to promote conservation values in plantations (Yamaura et al., 2018). Some broad-leaved trees have been retained in plantations since plantations were established by replacing natural broad-leaved forests more than 30 years ago in Japan (Yoshida et al., 2005; Ohsawa, 2007). Such retained trees can continue to be retained when plantations are harvested (Yamaura et al., 2018). Planting broad-leaved and conifer trees together (mixed plantations) can be another management option although few studies on mixed plantations have been conducted in Japan (Nagaike, 2012) and there are uncertainties about the successful species mixtures (Paquette and Messier, 2010). Therefore, valuing broad-leaved trees in conifer plantations and ensuring their retention is important. We suggest that these issues are relevant with other parts of the world since conifer plantations are common in Europe and North America (FAO, 2006), and conifer plantations with broad-leaved trees also support diverse faunal communities beyond Japan (Bibby et al., 1989; Lindbladh et al., 2017).

4.4. Environmental effects on density of planted trees As we predicted (Table 1), the density of planted cedars declined with snow depth. Taken together with the planted tree density effects on broad-leaved trees, ‘unsuccessful’ plantations with fewer planted trees are expected to support a greater amount of broad-leaved trees and therefore have higher conservation values (Yokoi and Yamaguchi, 1998; Masaki et al., 2004). The effects of other environmental 401

Forest Ecology and Management 444 (2019) 393–404

Y. Yamaura, et al.

Fig. 7. Spatial prediction of naturalness index for plantations. Naturalness index (INAT) and relevant covariates for conifer plantations in the study area (left column) and the central part of the area (right column). INAT was estimated based on the current planted species, stand age, climate and topographic covariates. Forest register data providing planted species and stand age were only partially available. See Appendix S9 for the spatial distribution of the other environmental covariates.

4.6. Conclusions

and time under different management scenarios. Our results suggest that stand age, plantation tree species, and density of planted trees have an important influence on the amount of native trees in plantations, and all of them can be considered in shaping management plans for plantations. Finally, growth rate and management regime vary markedly between plantation tree species and regions. For example, although conifer plantations in temperate and boreal zones are usually harvested every 50 years, the rotation age of broad-leaved plantations (e.g., Acacia and Eucalyptus spp.) can be less than 20 years (FAO, 2006). Rates of increase in non-planted native trees would therefore greatly depend on contexts. However, context-specific models can be developed using the framework developed for this study.

Our model for the conservation value of plantations does not incorporate management history, species composition of native trees, amount and types of standing dead trees, the latter two of which are known to affect animal communities in plantations (e.g., Yamaura et al., 2008). Improvement of our model is therefore an important next step. For example, if management history is systematically incorporated into NFI data, years after last thinning or number of thinning may be included into Eqs. (2)–(3). Reduced planted tree density by thinning could also be considered in Eqs. (4)–(5). Another challenge is to consider the effects of deer browsing that can reduce seedlings of broadleaved trees (e.g., Iijima and Nagaike, 2015). Conversely, the development of remote-sensing techniques may help spatial treatment of the amount of broad-leaved trees in the near future (Shugart et al., 2015). Since early-successional forests immediately after timber harvesting also have conservation values (Swanson et al., 2011; Yamaura et al., 2012b), further models dealing with this phase of stand development may be required (Appendix S11). Our model was based on readily available covariates, including stand age, making it possible to predict conservation values in space

Acknowledgements This study was funded by Research Grant #201502 from the Forestry and Forest Products Research Institute. Y.Y. was also supported by JSPS KAKENHI Grant #JP16KK0176. We thank the people who collected and managed National Forest Inventory (NFI) data. The use of NFI data (third term, ver. 1.0) was permitted by the Japanese Forestry Agency. Two anonymous reviewers, members of community 402

Forest Ecology and Management 444 (2019) 393–404

Y. Yamaura, et al.

dynamics laboratory in FFPRI and M. Takahashi provided useful comments on earlier versions of the manuscript.

Hua, F., Wang, L., Fisher, B., Zheng, X., Wang, X., Yu, D.W., ... Wilcove, D.S., 2018. Tree plantations displacing native forests: the nature and drivers of apparent forest recovery on former croplands in Southwestern China from 2000 to 2015. Biol. Conserv. 222, 113–124. Humphrey, J.W., 2005. Benefits to biodiversity from developing old-growth conditions in British upland spruce plantations: a review and recommendations. Forestry 78, 33–53. Hunter, J.C., Parker, V.T., 1993. The disturbance regime of an old-growth forest in coastal California. J. Veg. Sci. 4, 19–24. Iijima, H., Nagaike, T., 2015. Appropriate vegetation indices for measuring the impacts of deer on forest ecosystems. Ecol. Indic. 48, 457–463. Ishihara, M., Ishida, K., Ida, H., Itoh, A., Enoki, T., Ohkubo, T., Yoshida, T., 2010. An introduction to forest permanent plot data at core and subcore sites of the forest and grassland survey of the monitoring sites 1000 project. Jpn. J. Ecol. 60, 111–123 (in Japanese). Ishii, H., Maleque, M., Taniguchi, S., 2008. Line thinning promotes stand growth and understory diversity in Japanese cedar (Cryptomeria japonica D. Don) plantations. J. For. Res. 13, 73–78. Kerr, G., 1999. The use of silvicultural systems to enhance the biological diversity of plantation forests in Britain. Forestry 72, 191–205. Kéry, M., Royle, J.A., 2016. Applied hierarchical modeling in ecology: analysis of distribution, abundance and species richness using R and BUGS. Volume 1. Prelude and static models. Academic Press, San Diego. Kira, T., 1991. Forest ecosystems of east and southeast Asia in a global perspective. Ecol. Res. 6, 185–200. Kiyono, Y., 1990. Dynamics and control of understories in Chamaecyparis plantations. Bull. Forest. For. Prod. Res. Inst. 359, 1–122 (in Japanese). Klinka, K., Carter, R.E., 1990. Relationships between site index and synoptic environmental factors in immature coastal douglas-fir stands. For. Sci. 36, 815–830. Koh, L.P., Gardner, T.A., 2010. Conservation in human-modified landscapes. In: Sodhi, N.S., Ehrlich, P.R. (Eds.), Conservation biology for all. Oxford University Press, Oxford, pp. 236–261. Larson, A.J., Lutz, J.A., Gersonde, R.F., Franklin, J.F., Hietpas, F.F., 2008. Potential site productivity influences the rate of forest structural development. Ecol. Appl. 18, 899–910. Liang, J., Crowther, T.W., Picard, N., Wiser, S., Zhou, M., Alberti, G., ... Reich, P.B., 2016. Positive biodiversity-productivity relationship predominant in global forests. Science 354, aaf8957. Lindbladh, M., Lindström, Å., Hedwall, P.-O., Felton, A., 2017. Avian diversity in Norway spruce production forests – how variation in structure and composition reveals pathways for improving habitat quality. For. Ecol. Manage. 397, 48–56. Lindenmayer, D.B., Cunningham, R.B., Donnelly, C.F., Nix, H., Lindenmayer, B.D., 2002. Effects of forest fragmentation on bird assemblages in a novel landscape context. Ecol. Monogr. 72, 1–18. Lindenmayer, D.B., Blanchard, W., Westgate, M.J., Foster, C., Banks, S.C., Barton, P., . . . Scheele, B.C., in press. Novel bird responses to successive, large-scale, landscape transformations. Ecol. Monogr., e01362. Lindenmayer, D.B., Hobbs, R.J., 2004. Fauna conservation in Australian plantation forests – a review. Biol. Conserv. 119, 151–168. Masaki, T., Ota, T., Sugita, H., Oohara, H., Otani, T., Nagaike, T., Nakamura, S., 2004. Structure and dynamics of tree populations within unsuccessful conifer plantations near the Shirakami mountains, a snowy region of Japan. For. Ecol. Manage. 194, 389–401. McKenney, D.W., Pedlar, J.H., 2003. Spatial models of site index based on climate and soil properties for two boreal tree species in Ontario. Canada. For. Ecol. Manage. 175, 497–507. Messier, C., Puettmann, K., Chazdon, R., Andersson, K.P., Angers, V.A., Brotons, L., ... Levin, S.A., 2015. From management to stewardship: viewing forests as complex adaptive systems in an uncertain world. Conserv. Lett. 8, 368–377. Mitsuda, Y., Ito, S., Sakamoto, S., 2007. Predicting the site index of sugi plantations from GIS-derived environmental factors in Miyazaki prefecture. J. For. Res. 12, 177–186. Moore, S.E., Allen, H.L., 1999. Plantation forestry. In: Hunter, M.L. (Ed.), Maintaining biodiversity in forest ecosystems. Cambridge University Press, Cambridge, pp. 400–433. Moore, I.D., Norton, T.W., Williams, J.E., 1993. Modelling environmental heterogeneity in forested landscapes. J. Hydrol. 150, 717–747. Nagaike, T., 2012. Current status and perspectives of mixed tree plantations: a review of productivity studies. J. Jpn. For. Soc. 94, 196–202 (in Japanese). Nájera, A., Simonetti, J.A., 2010. Enhancing avifauna in commercial plantations. Conserv. Biol. 24, 319–324. Nakajima, Y., Yamahata, K., Fujimoto, K., Tsujita, A., Masuoka, M., Koida, Y., 1965. Studies on the possibility of making larch forest in Shikoku district: some investigations to the forests of Sumitomo at Besshi region. Bull. Ehime Univ. For. 2, 22–47 (in Japanese). Nakashizuka, T., Iida, S., 1995. Composition, dynamics and disturbance regime of temperate deciduous forests in Monsoon Asia. Vegetatio 121, 23–30. Newbold, T., Hudson, L.N., Hill, S.L.L., Contu, S., Lysenko, I., Senior, R.A., ... Purvis, A., 2015. Global effects of land use on local terrestrial biodiversity. Nature 520, 45–50. Norton, D.A., 1998. Indigenous biodiversity conservation and plantation forestry: options for the future. N. Z. Forestry 43, 34–39. Nothdurft, A., Wolf, T., Ringeler, A., Böhner, J., Saborowski, J., 2012. Spatio-temporal prediction of site index based on forest inventories and climate change scenarios. For. Ecol. Manage. 279, 97–111. Nyland, R.D., 2016. Silviculture: concepts and applications. Waveland Press. Ohsawa, M., 2007. The role of isolated old oak trees in maintaining beetle diversity within larch plantations in the central mountainous region of Japan. For. Ecol.

Declarations of interest None. Appendix A. Supplementary material Supplementary data to this article can be found online at https:// doi.org/10.1016/j.foreco.2019.04.038. References Bibby, C.J., Aston, N., Bellamy, P.E., 1989. Effects of broadleaved trees on birds of upland conifer plantations in north Wales. Biol. Conserv. 49, 17–29. Brockerhoff, E., Jactel, H., Parrotta, J., Quine, C., Sayer, J., 2008. Plantation forests and biodiversity: oxymoron or opportunity? Biodivers. Conserv. 17, 925–951. Brudvig, L.A., Damschen, E.I., Tewksbury, J.J., Haddad, N.M., Levey, D.J., 2009. Landscape connectivity promotes plant biodiversity spillover into non-target habitats. Proc. Natl. Acad. Sci. USA 106, 9328–9332. Brunet, J., 2007. Plant colonization in heterogeneous landscapes: an 80-year perspective on restoration of broadleaved forest vegetation. J. Appl. Ecol. 44, 563–572. Chen, H.Y.H., Klinka, K., Kabzems, R.D., 1998. Site index, site quality, and foliar nutrients of trembling aspen: relationships and predictions. Can. J. For. Res. 28, 1743–1755. Conrad, O., Bechtel, B., Bock, M., Dietrich, H., Fischer, E., Gerlitz, L., ... Böhner, J., 2015. System for automated geoscientific analyses (SAGA) v. 2.1.4. Geosci. Model Dev. 8, 1991–2007. Coote, L., Dietzsch, A.C., Wilson, M.W., Graham, C.T., Fuller, L., Walsh, A.T., ... O’Halloran, J., 2013. Testing indicators of biodiversity for plantation forests. Ecol. Indic. 32, 107–115. Curt, T., Bouchaud, M., Agrech, G., 2001. Predicting site index of Douglas-Fir plantations from ecological variables in the Massif central area of France. For. Ecol. Manage. 149, 61–74. de Castilho, C.V., Magnusson, W.E., de Araújo, R.N.O., Luizão, R.C.C., Luizão, F.J., Lima, A.P., Higuchi, N., 2006. Variation in aboveground tree live biomass in a central Amazonian forest: effects of soil and topography. For. Ecol. Manage. 234, 85–96. Deal, R.L., Hennon, P.E., Orlikowska, E.H., D'Amore, D.V., 2004. Stand dynamics of mixed red alder-conifer forests of southeast Alaska. Can. J. For. Res. 34, 969–980. Demarais, S., Verschuyl, J.P., Roloff, G.J., Miller, D.A., Wigley, T.B., 2017. Terrestrial vertebrate biodiversity and intensive forest management in the U.S. For. Ecol. Manage. 385, 308–330. ESRI, 2014. ArcGIS Desktop. Environmental Systems Research Institute, Redlands. Fahrig, L., 2017. Ecological responses to habitat fragmentation per se. Annu. Rev. Ecol. Evol. Syst. 48, 1–23. FAO, 2006. Global planted forests thematic study: results and analysis. Forestry Department, Food and Agriculture Organization of the United Nations, Rome. FAO, 2015. Global forest resources assessment 2015. Food and Agriculture Organization of the United Nations, Rome. Fimbel, R.A., Fimbel, C.C., 1996. The role of exotic conifer plantations in rehabilitating degraded tropical forest lands: a case study from the Kibale forest in Uganda. For. Ecol. Manage. 81, 215–226. Fischer, J., Lindenmayer, D.B., 2002. Small patches can be valuable for biodiversity conservation: two case studies on birds in southeastern Australia. Biol. Conserv. 106, 129–136. Forestry Agency, 2017b. White paper on forestry and forests, 2016. Japan Forestry Association, Tokyo (in Japanese). Forestry Agency, 2017a. Handbook of statistics on forest and forestry, 2017. Japan Forestry Association, Tokyo (in Japanese). Forestry Agency, 2014. Status of forest resources. Forestry Agency, Ministry of Agriculture, Forestry and Fishery, Japan, Tokyo. http://www.rinya.maff.go.jp/j/ kouhou/toukei/index.html (in Japanese). Franklin, J.F., 1993. Preserving biodiversity: species, ecosystems, or landscapes? Ecol. Appl. 3, 202–205. Fujimori, T., 2001. Ecological and silvicultural strategies for sustainable forest management. Elsevier, Amsterdam. Gelman, A., Hill, J., 2007. Data analysis using regression and multilevel/hierarchical models. Cambridge University Press, Cambridge. Guariguata, M.R., 1990. Landslide disturbance and forest regeneration in the upper Luquillo mountains of Puerto Rico. J. Ecol. 78, 814–832. Hanley, T.A., Deal, R.L., Orlikowska, E.H., 2006. Relations between red alder composition and understory vegetation in young mixed forests of southeast Alaska. Can. J. For. Res. 36, 738–748. Harrington, R.A., Ewel, J.J., 1997. Invasibility of tree plantations by native and nonindigenous plant species in Hawaii. For. Ecol. Manage. 99, 153–162. Hartley, M.J., 2002. Rationale and methods for conserving biodiversity in plantation forests. For. Ecol. Manage. 155, 81–95. Hirata, Y., Hosoda, K., Nishizono, T., Kitahara, F., Nagame, I., Nakamura, M., 2016. Japan. In: Vidal, C., Alberdi, I.A., Hernández Mateo, L., Redmond, J.J. (Eds.), National Forest Inventories: Assessment of Wood Availability and Use. Springer International Publishing, Cham, pp. 507–520.

403

Forest Ecology and Management 444 (2019) 393–404

Y. Yamaura, et al. Manage. 250, 215–226. Ohsawa, M., Shimokawa, T., 2011. Extending the rotation period in larch plantations increases canopy heterogeneity and promotes species richness and abundance of native beetles: implications for the conservation of biodiversity. Biol. Conserv. 144, 3106–3116. Paquette, A., Messier, C., 2010. The role of plantations in managing the world's forests in the Anthropocene. Front. Ecol. Environ. 8, 27–34. Payn, T., Carnus, J.-M., Freer-Smith, P., Kimberley, M., Kollert, W., Liu, S., ... Wingfield, M.J., 2015. Changes in planted forests and future global implications. For. Ecol. Manage. 352, 57–67. Peterson, D.W., Peterson, D.L., 2001. Mountain hemlock growth responds to climatic variability at annual and decadal time scales. Ecology 82, 3330–3345. Potvin, C., Dutilleul, P., 2009. Neighborhood effects and size-asymmetric competition in a tree plantation varying in diversity. Ecology 90, 321–327. Puyravaud, J.-P., Davidar, P., Laurance, W.F., 2010. Cryptic destruction of India's native forests. Conserv. Lett. 3, 390–394. Royle, J.A., Dorazio, R.M., 2008. Hierarchical modeling and inference in ecology: the analysis of data from populations, metapopulations and communities. Academic Press, Amsterdam. Saito, M., 1989. The relationship between the amount of undergrowth and solar radiation in Cryptomeria plantations. J. Jpn. For. Soc. 71, 276–280 (in Japanese). Schielzeth, H., 2010. Simple means to improve the interpretability of regression coefficients. Methods Ecol. Evol. 1, 103–113. Sedjo, R.A., Botkin, D., 1997. Using forest plantations to spare natural forests. Environment 39, 14–30. Seiwa, K., Etoh, Y., Hisita, M., Masaka, K., Imaji, A., Ueno, N., ... Kimura, M., 2012. Roles of thinning intensity in hardwood recruitment and diversity in a conifer, Criptomeria japonica plantation: a 5-year demographic study. For. Ecol. Manage. 269, 177–187. Shugart, H.H., Asner, G.P., Fischer, R., Huth, A., Knapp, N., Le Toan, T., Shuman, J.K., 2015. Computer and remote-sensing infrastructure to enhance large-scale testing of individual-based forest models. Front. Ecol. Environ. 13, 503–511. Spellerberg, I.F., Sawyer, J.W.D., 1996. Standards for biodiversity: a proposal based on biodiversity standards for forest plantations. Biodivers Conserv 5, 447–459. Swanson, M.E., Franklin, J.F., Beschta, R.L., Crisafulli, C.M., DellaSala, D.A., Hutto, R.L.,

... Swanson, F.J., 2011. The forgotten stage of forest succession: early-successional ecosystems on forest sites. Front. Ecol. Environ. 9, 117–125. Utsugi, E., Kanno, H., Ueno, N., Tomita, M., Saitoh, T., Kimura, M., Seiwa, K., 2006. Hardwood recruitment into conifer plantations in Japan: effects of thinning and distance from neighboring hardwood forests. For. Ecol. Manage. 237, 15–28. Warton, D.I., Hui, F.K.C., 2011. The arcsine is asinine: the analysis of proportions in ecology. Ecology 92, 3–10. Yamaura, Y., Katoh, K., Takahashi, T., 2008. Effects of stand, landscape, and spatial variables on bird communities in larch plantations and deciduous forests in central Japan. Can. J. For. Res. 38, 1223–1243. Yamaura, Y., Ikeno, S., Sano, M., Okabe, K., Ozaki, K., 2009. Bird responses to broadleaved forest patch area in a plantation landscape across seasons. Biol. Conserv. 142, 2155–2165. Yamaura, Y., Amano, T., Kusumoto, Y., Nagata, H., Okabe, K., 2011. Climate and topography drives macroscale biodiversity through land-use change in a human-dominated world. Oikos 120, 427–451. Yamaura, Y., Oka, H., Taki, H., Ozaki, K., Tanaka, H., 2012a. Sustainable management of planted landscapes: lessons from Japan. Biodivers. Conserv. 21, 3107–3129. Yamaura, Y., Royle, J.A., Shimada, N., Asanuma, S., Sato, T., Taki, H., Makino, S., 2012b. Biodiversity of man-made open habitats in an underused country: a class of multispecies abundance models for count data. Biodivers. Conserv. 21, 1365–1380. Yamaura, Y., Shoji, Y., Mitsuda, Y., Utsugi, H., Tsuge, T., Kuriyama, K., Nakamura, F., 2016. How many broadleaved trees are enough in conifer plantations? The economy of land sharing, land sparing and quantitative targets. J. Appl. Ecol. 53, 1117–1126. Yamaura, Y., Akashi, N., Unno, A., Tsushima, T., Nagasaka, A., Nagasaka, Y., Ozaki, K., 2018. Retention experiment for plantation forestry in Sorachi, Hokkaido (REFRESH): a large-scale experiment for retaining broad-leaved trees in conifer plantations. Bull. Forestry For. Prod. Res. Inst. 17, 91–109. Yokoi, S., Yamaguchi, K., 1998. Growing condition of Japanese ceder (Cryptomeria japonica D. DON) and hardwoods in the unsuccessful plantations of cedar in the snowy region. Jpn. J. Forest Environ. 40, 91–96 (in Japanese). Yoshida, T., Hasegawa, M., Taira, H., Noguchi, M., 2005. Stand structure and composition of a 60-year-old larch (Larix kaempferi) plantation with retained hardwoods. J. For. Res. 10, 351–358.

404