Ecological Research

, Volume 25, Issue 4, pp 801–812 | Cite as

Effects of altitude and competition on growth and mortality of the conifer Abies sachalinensis

Original Article

Abstract

Altitudinal gradients are convenient subjects to investigate plant responses to air temperature. Plant growth and mortality are also affected by competition at any altitude. This study investigated the effects of altitude and competition on absolute diameter growth rate (ADGR) and mortality of the conifer Abies sachalinensis by using 13-year data. This study was done at two altitudes (200 and 1,000 m a.s.l.) in northern Japan. Local crowding by conifers and broad-leaved trees reduced ADGR of target trees. ADGR was lower in high altitude than low altitude at any DBH and any degree of local crowding because of the short growing season. Observed size-dependent mortality was a U-shaped pattern against DBH at the two altitudes. Smaller and larger trees tended to die of suppression (standing-dead) and disturbances (stem-broken and uprooting), respectively. Mortality of standing-dead trees was negatively correlated with ADGR, irrespective of altitude, i.e., ADGR was a good predictor of mortality. Thus, mortality of standing-dead trees was estimated to be greater at high altitude than low altitude at any degree of local crowding because ADGR was lower at high altitude than low altitude. By contrast, mortality due to disturbances was slightly greater at low altitude than high altitude. Thus, this study showed that a short growth period decreases growth and increases mortality due to suppression at high altitude. Although global warming may increase growth and survival of individual trees at high altitude by prolonging the growth period, prediction on mortality due to disturbances needs caution because the mortality is largely affected by frequency and intensity of disturbances.

Keywords

Altitude Competition Growth Global warming Local crowding Mortality 

Introduction

Climatic changes due to global warming are assumed to affect plant growth and survival (Aber and Federer 1992; Apple et al. 2000; Lindner et al. 2000). An increase of plant growth also contributes to upward distribution shift of altitudinal and latitudinal plant distribution, associated with increase of seed production and seedling establishment (Kullman 2002; Parmesan and Yohe 2003; Kelly and Goulden 2008). Many dendrochronological studies showed that interannual variations in radial growth of trees positively correlated with air temperature in high altitudes and high latitudes (Gostev et al. 1996; Buckley et al. 1997; Peterson and Peterson 2001; Takahashi et al. 2005b). Experimental studies also showed that growth of alpine and tundra plants was increased by artificial warming due to open top chambers and electronic heating (Fransworth et al. 1995; Arft et al. 1999; Molau 1997; Hollister and Webber 2000; Klanderud and Totland 2005; Takahashi 2005). These previous studies provided useful insight into consideration of effects of global warming on vegetation. However, most previous studies are at the levels of individual plant and shoot. In addition, examined plants were mostly small plants such as herbs and shrubs (Jones et al. 1997; Dorrepaal et al. 2003; Hollister et al. 2005), because of the difficulty of making large experimental equipments. Thus, information on growth responses to temperature is still limited for tall trees, although the tall tree is a major component of forest ecosystems. Furthermore, growth and survival of trees are largely affected by competition between trees. Thus, it is necessary to investigate effects of competition on growth and survival of trees for examination of effects of global warming in forest ecosystems.

Altitudinal gradients are convenient subjects to investigate plant responses to temperature because air temperature changes abruptly across a short distance (Körner 2007). Thus, examination of tree competition along altitudes provides useful information to understand the impacts of climatic changes on growth and survival of trees. The growth period of plants decreases with an increase in altitude. A reduced growth period at high altitude decreases plant growth and stem height (Paulsen et al. 2000). Competition from neighboring plants also reduces plant growth and increases mortality, irrespective of altitudes (Hara et al. 1994; Takahashi 1996; Canham et al. 2004), except in harsh environments such as alpine zone where aggregation of plants facilitates plant growth and survival (Šrůtek and Lepš 1994; Germino et al. 2002; Kikvidze 2002; Takahashi et al. 2005a). Although both altitude and competition influence tree growth, only a few studies have examined altitudinal changes in tree growth and mortality in relation to competition (Coomes and Allen 2007).

It is supposed that there are two modes of competition. Each plant acquires soil water and nutrients in proportion to its size (i.e., two-sided competition or symmetric competition). On the contrary, in terms of competition for light, larger plants acquire greater light resource disproportionately relative to the size (i.e., one-sided competition or asymmetric competition) (Weiner 1990; Hara 1992; Kikuzawa 1999). Takahashi (2003a) also showed that the effect of local crowding by evergreen conifers on the growth of evergreen conifer Abies sachalinensis Masters was much greater than that by deciduous broad-leaved trees because of the difference in leaf phenology between evergreen conifers and deciduous broad-leaved trees. A. sachalinensis under or near the crowns of deciduous broad-leaved trees can receive much light before the leaf emergence of the overhead trees, and so is beneficial for photosynthetic production by understory Abies trees. Therefore, the mode of competition and the life-form (evergreen conifers and deciduous broad-leaved trees) of neighboring trees should be also taken into account for tree competition along altitudes.

Mortality is greater in crowded conditions with many neighboring trees (Duncan 1991; Umeki and Kikuzawa 1999), because of small carbon gain due to shading. Several studies also showed that tree mortality is higher for trees with lower growth rates, and, therefore, growth rate can be used as an integrated measure of whole-plant carbon gain (Kobe et al. 1995; Monserud and Sterba 1999; Caspersen and Kobe 2001; Umeki 2002; Kunstler et al. 2005). If a threshold of carbon gain for survival of trees exists irrespective of altitude, and if tree growth (i.e., whole-plant carbon gain) is lower at higher altitudes, the effect of growth rate on mortality should not differ along altitudes and mortality should be greater at any local crowding in higher altitudes. By contrast, trees are often damaged by external disturbances such as strong winds and typhoons, irrespective of the tree growth. Wind damage may be large for isolated trees than trees surrounding by neighboring trees. Thus, it is expected that local crowding differently affects mortality rates due to suppression and disturbances.

Long-term observation is of great importance for examining growth and mortality of trees because mortality measurement of tree species needs at least several years in large areas. This study examined the growth and mortality of the conifer Abies sachalinensis in northern Japan by using 13-year observation data at two altitudes (200 and 1,000 m a.s.l.). Mode of competition and life-form (evergreen conifers and deciduous broad-leaved trees) of neighboring trees were also analyzed as variables of the growth and mortality models. Specifically, I attempt to answer the following questions.
  1. 1.

    Is the growth rate of A. sachalinensis greater at low altitude than at high altitude at any tree size and any degree of competition expressed as local crowding?

     
  2. 2.

    How do altitude and competition affect mortality rates of A. sachalinensis through suppression and disturbances?

     

Materials and methods

Study sites

This study was done at two sites in Hokkaido, northern Japan. The first study site was an old-growth cool-temperate conifer-hardwood forest in Urahoro (43°1′N, 143°38′E, 200 m a.s.l.). Annual precipitation, recorded at Honbetsu Weather Station (ca. 10 km from the study site, 67 m a.s.l.), was 769 mm from 1979 to 2000. The mean annual maximum depth of snow was 45.1 cm from 1986 to 2008. Mean monthly temperatures of the coldest month of January and the hottest month of August at the study site were estimated as −9.8 and 19.1°C, respectively, with 5.2°C annual mean temperature, recorded at Honbetsu Weather Station using the standard lapse rate of −0.6°C for each +100 m altitude. The forest around the study site was dominated by the conifer A. sachalinensis and deciduous broad-leaved hardwood species, such as Acer mono Maxim., Tilia japonica (Miq.) Simonkai and Prunus ssiori Fr. Schm. A. sachalinensis occupied 31% of the total basal area and density in this stand (K. Takahashi, unpublished manuscript). The forest floor was patchily covered with Sasa nipponica Makino. The forest was well preserved, and therefore the vegetation had no anthropogenic effects.

The second study site (43°21′N, 143°9′E, 1,000 m a.s.l.) was an old-growth subalpine coniferous forest on Mt. Onsen in Nukabira. This site was ca. 50 km distant from the first study site. Meteorological data were recorded at a station 2 km from the study site (Nukabira Weather Station, 540 m a.s.l.). The annual precipitation was about 1,298 mm from 1979 to 2000. Mean monthly temperatures in January and August at the study site were estimated as −13.7 and 14.7°C, respectively, with 0.7°C annual mean temperature, recorded at Nukabira Weather Station using the standard lapse rate of −0.6°C for each +100 m altitude. The mean annual maximum depth of snow was 102.1 cm from 1986 to 2008. The area was dominated by two conifers A. sachalinensis and Picea glehnii Masters. The other component species were deciduous broad-leaved tree species, such as Betula ermanii Cham., Sorbus commixta Hedland and Acer ukurunduense Trautv et Meyer. Two conifers, A. sachalinensis and P. glehnii, occupied about 80% total density and basal area of this stand. A more detailed description of the forest is available in Takahashi (1994).

The warmth index (WI) has often been used to explain relationships between climatic zones and vegetation patterns along altitudinal and latitudinal gradients (Kira 1948). The WI is calculated as ∑(mt − 5), where mt is the mean monthly temperature above 5°C. The WI expresses the effective heat for plant growth. The WIs at Nukabira and Urahoro were 28.9 and 53°C months, respectively, within the WI ranges of subalpine coniferous forest zone (15–45) and conifer–hardwood forest zone (45–60), respectively. The timberline was located at about 1,300 m a.s.l. around the second study site (Nukabira).

Field methods

Three plots of size 40 × 100 m, 30 × 50 m, and 60 × 100 m (total 1.15 ha) close to each other (ca. 50 m apart) were established at Urahoro in 1994. Each plot was divided into contiguous 5 × 5 m quadrats. All trees ≥1.5 cm diameter at breast height (DBH) were tagged. Their DBH was measured in 1994, and again in 1996, 1997, 1998, 1999, 2004, and 2007. Dead trees were recorded during the census period and were classified as standing-dead, stem-broken, and uprooted trees. Trunk height was also measured for ca. 150 trees randomly selected. Standing-dead trees were assumed to have died of suppression or senescence, dead trees that included stem-broken and uprooted trees were assumed to have died of disturbances (Nakashizuka et al. 1992). Although stem-broken and uprooted trees often occur after being standing-dead trees, this classification (standing-dead trees or not) is a convenient measure of suppression or senescence for tree mortality.

A 150 × 150 m plot was established at Nukabira in 1991 and was divided into contiguous 5 × 5 m quadrats. All trees ≥2 cm DBH were tagged. Their DBH was measured in 1991 and again in 1994, 1997, and 2007. Trunk height was also measured for ca. 650 trees randomly selected. Dead trees were recorded during the census period and were classified as standing-dead, stem-broken, and uprooted trees. The forest floor was heterogeneous in this plot. The quadrats were classified into three types [soil (SL), rock-soil (RS), and rock-moss (RM)], based on the soil thickness on the bedrock (Takahashi 1994). The soil thicknesses of SL, RS, and RM were about 30–100, 20–30, and 0–20 cm, respectively, and the areas were 1.37, 0.6, and 0.28 ha. Dwarf bamboo Sasa senanensis Rehder dominated the forest floor of SL, but was scarce on RS and RM. A thick layer of mosses was on the rocks of RM. The shrub layer of RS and RM was dominated by Ericaceous species such as Menziesia pentandra Maxim. and Ledum palustre var. diversipiosum Hara. The soil nutrient condition was not examined in this study. However, shallow soil with a developed moss community consists primarily of thick accumulation of litter with almost no mineral horizon because of shallow lithic contacts (Kojima 1991). Nutritionally, shallow soil is indicated by low pH, electric conductivity, amount of calcium and magnesium, and base saturation. Predominance of Ericaceous species indicates a poor nutrient condition. Therefore, RS and RM should have lower nutrient availability than SL. To exclude the effect of nutrient availability on growth and mortality (cf. Kobe et al. 1995), this study used data of SL only for the Nukabira plot. Stand basal area of SL (38 m2 ha−1) was smaller than that of Urahoro (49.5 m2 ha−1).

Data analyses

In this study, growth and mortality of A. sachalinensis were examined at the two altitudes [Urahoro (200 m a.s.l.) and Nukabira (1,000 m a.s.l.)], because only A. sachalinensis was a common dominant species at the two altitudes. Interannual variations in climatic conditions often affect growth and mortality of plants (Takahashi et al. 2003, 2005b; Tardif et al. 2006; van der Werf et al. 2007). To avoid the effects of interannual variations in climatic conditions on growth and mortality of A. sachalinensis, growth and mortality were compared between the two altitudes for the same measurement period (1997–2007).

Absolute diameter growth rate (ADGR) of A. sachalinensis was analyzed in relation to altitude, tree size (DBH), local crowding of neighboring trees. Many other studies often defined neighborhood area as a 10 × 10 m quadrat in which the target tree is located (Kohyama 1992, 1993; Kubota and Hara 1995; Nakashizuka and Kohyama 1995; Takahashi and Kohyama 1999). However, this definition of neighborhood area sometimes brings about a crude measure of local crowding if target trees are near the edge of a quadrat, i.e., a target tree near the edge of a quadrat will be affected by trees at the adjacent quadrat but not much by trees at the other end of the same quadrat. In this study, spatial distribution of trees was recorded at the resolution of 5 × 5 m quadrat. Since most trees are close to target trees in a 5 × 5 m quadrat, a 5 × 5 m quadrat more precisely expresses local crowding for target trees compared with a 10 × 10 m quadrat. However, the size of 5 × 5 m quadrat is too small to express the local crowding because of shading from tall neighboring trees outside the quadrat (canopy height of the stands was about 25 m). Thus, the adjacent eight 5 × 5 m quadrats were also included as neighborhood area in addition to the own 5 × 5 m quadrat in which the target tree is located (i.e., neighborhood area was 225 m2). If neighborhood area is widened to the next adjacent quadrats of 5 × 5 m (ex. 25 quadrats, 625 m2), it is too large to estimate the local crowding of the target tree. Thus, neighborhood area of nine quadrats of 5 × 5 m is appropriate to express local crowding for target trees in this study.

This study examined two modes of competition (i.e., one-sided and two-sided competition). In the case of two-sided competition, all trees (DBH > 2.0 cm) within the neighborhood area were treated as neighboring trees. In the case of one-sided competition, only trees with DBH larger than the target tree were treated as neighboring trees. Local crowding was calculated for each target tree as the sum of the basal area of conifer and broad-leaved neighbors. Target trees in the outermost 5 × 5 m quadrats in each plot were not used for statistical analysis.

Generalized linear model with Gaussian distribution was used to analyze the effects of altitude (200 and 1,000 m a.s.l.), tree size (natural logarithm of DBH), local crowding by evergreen conifers and deciduous broad-leaved trees on ADGR (mm year−1) of A. sachalinensis. The model including all variables is:
$$ {\text{ADGR}} = a_{0} + a_{1} \,{\text{Altitude}} + a_{2} \ln \,{\text{DBH}} + a_{3} \sum {\text{BA}}_{{{\text{C}}_{i} }} + a_{4} \sum {\text{BA}}_{{{\text{B}}_{i} }} $$
where coefficient a0, a1, a2, a3 and a4 are constants, DBH is the initial DBH (cm) in 1997, \( \sum {\text{BA}}_{{{\text{C}}_{i} }} \) (cm2 m−2) is total basal area of coniferous neighbors, \( \sum {\text{BA}}_{{{\text{B}}_{i} }} \) (cm2 m−2) is total basal area of broad-leaved neighbors. A subscript i of \( \sum {\text{BA}}_{{{\text{C}}_{i} }} \) and \( \sum {\text{BA}}_{{{\text{B}}_{i} }} \) represents one-sided competition (i = 1) or two-sided competition (i = 2). “Altitude” was treated as a categorical variable. Values of 0 and 1 were assigned for low altitude (Urahoro, 200 m a.s.l.) and high altitude (Nukabira, 1,000 m a.s.l.), respectively. The others were continuous variables. The model calculation was done for four cases of \( {\text{BA}}_{{{\text{C}}_{i} }} \) and \( {\text{BA}}_{{{\text{B}}_{i} }} \) combinations, i.e., \( ({\text{BA}}_{{{\text{C}}_{1} }} \) and \( {\text{BA}}_{{{\text{B}}_{1} }}), \)\( ({\text{BA}}_{{{\text{C}}_{1} }} \) and \( {\text{BA}}_{{{\text{B}}_{2} }} \)), \( ({\text{BA}}_{{{\text{C}}_{2} }} \) and \( {\text{BA}}_{{{\text{B}}_{ 1} }} \)) and \( ({\text{BA}}_{{{\text{C}}_{2} }} \) and \( {\text{BA}}_{{{\text{B}}_{2} }} \)). Stepwise function was used to choose variables for each case, based on Akaike information criteria (AIC). Furthermore, final models were compared among the four cases, based on AIC. A model with the lowest value of AIC is essentially the best model. However, any model is substantially same with the best model with the lowest value of AIC if the difference in AIC value (ΔAIC) of the model of interest is smaller than 2 from the best model (Burnham and Anderson 2002).
Mortality during the census period (1997–2007) was calculated by the following equation (Sheil and May 1996):
$$ {\text{Mortality }}({\text{year}}^{ - 1} ) = \frac{1}{10}\ln \left({{\frac{{N_{\text{i}} }}{{N_{\text{S}} }}}} \right) $$
where Ni is the initial number of trees in 1997 and Ns is the number of survived trees during the census period (1997–2007). Mortality was calculated for four size classes 2.0–9.9, 10.0–19.9, 20.0–29.9, and ≥30.0 cm DBH at each altitude. A significant difference in mortality between the two altitudes for each size class was tested by 95% confidence intervals generated using a 1,000-iteration bootstrap technique (Crowley 1992).
Mortality model was constructed. Tree mortality is a discrete event. A datum can have only the value 0 (live) or 1 (dead). Therefore, I analyzed the probability of mortality by generalized linear model with binomial distribution:
$$ M = {\frac{1}{1 + \exp (- bX)}} $$
where M (10 year−1) is the probability of decadal mortality from 1997 to 2007, bX is a linear combination of parameter b and independent variable X. This study examined the effects of altitude, DBH, growth rate and local crowding by conifers and broad-leaved trees on tree mortality. Size-dependent mortality sometimes shows a U-shaped pattern with higher mortality of saplings and large canopy trees because of suppression of small trees and because of senescence and killing by disturbances of large trees (Monserud and Sterba 1999; Takahashi and Kohyama 1999). Such a U-shaped mortality can be expressed by a quadric function of size. Degree of local crowding by neighboring trees can be expressed as total basal area of neighbors. Growth rate can be used as an integrated measure of whole-plant carbon gain (Kobe et al. 1995; Monserud and Sterba 1999; Caspersen and Kobe 2001; Umeki 2002; Kunstler et al. 2005), and therefore a growth rate is an important element to estimate mortality. Combining these effects, the hypothesized mortality model (M, 10 years−1) is:
$$ M = {\frac{1}{{1 + \exp [ - (b_{0} + b_{1} {\text{Altitude}} + b_{2} {\text{DBH}} + b_{3} {\text{DBH}}^{2} + b_{4} {\text{ADGR}} + b_{5} \sum {\text{BA}}_{{{\text{C}}_{i} }} + b_{6} \sum {\text{BA}}_{{{\text{B}}_{i} }} )]}}} $$
where coefficient b0, b1, b2, b3, b4, b5 and b6 are constants, DBH is the initial DBH (cm) in 1997. \( \sum {\text{BA}}_{{{\text{C}}_{i} }} \) and \( \sum {\text{BA}}_{{{\text{B}}_{i} }} \) are total basal area (cm2 m−2) of coniferous neighbors and that of broad-leaved neighbors, respectively, where i = 1 (one-sided competition) and i = 2 (two-sided competition). ADGR (mm year−1) is the growth rate from 1994 to 1997 before the mortality measurement (1997–2007). “Altitude” was treated as a categorical variable. A value 0 and 1 were assigned for low altitude (Urahoro, 200 m a.s.l.) and high altitude (Nukabira, 1,000 m a.s.l.), respectively. Four cases of \( {\text{BA}}_{{{\text{C}}_{i} }} \) and \( {\text{BA}}_{{{\text{B}}_{i} }} \) combinations were tested and compared, based on AIC, like as in the growth model.

Standing-dead trees die of suppression or senescence rather than disturbances. Suppressed trees grow slowly and then die if a positive carbon balance cannot be maintained. By contrast, trees are killed by disturbances, irrespective of growth rates. The effect of growth rate on mortality should be greater for standing-dead trees than for the other form of dead trees (stem-broken and uprooted trees). Thus, mortality was analyzed separately for standing-dead trees only and the other dead trees.

Model determination and calculation of the AIC values for the growth and mortality models were performed with the free statistical software R 2.9.0 (R development core team 2009).

Results

The allometry between trunk height and diameter in broad-leaved trees was similar between the two altitudes (200 and 1,000 m a.s.l.) (Fig. 1b). Although the number of conifers measured was only 15 trees and there was no tree larger than 30 cm DBH at low altitude, the allometry between trunk height and diameter (<30 cm) was similar between the two altitudes (Fig. 1a). DBH 10, 20, and 30 cm of conifers corresponded to trunk height ca. 7, 13, and 18 m, respectively (Fig. 1a). Thus, it is tentatively supposed that trees with same DBH had similar shading effects on growth of target trees at the two altitudes for each life-form (conifer and broad-leaved trees) as a precondition for growth and mortality models in this study.
Fig. 1

Relationship between trunk height and DBH for conifers (a) and broad-leaved trees (b) in Hokkaido, northern Japan. Solid circles and crosses represent 1,000 and 200 m a.s.l., respectively

Absolute diameter growth rate of A. sachalinensis was positively correlated with DBH at the two altitudes, although the variation was large (Fig. 2). Effects of altitude, DBH and mode of competition expressed as local crowding of each life-form on ADGR were analyzed by generalized linear model with stepwise function for the four cases of local crowding (Table 1). AIC values were almost same between (C1–B1, Eq. 1) and (C2–B1, Eq. 3) and between (C1–B2, Eq. 2) and (C2–B2, Eq. 4), where Ci and Bi are subscripts of \( \sum {\text{BA}}_{{{\text{C}}_{i} }} \) and \( \sum {\text{BA}}_{{{\mathbf{B}}_{i} }} \) (i = 1 or 2), respectively (Table 1). AIC values of Eqs. 1 and 3 were smaller than those of Eqs. 2 and 4 (i.e., ΔAIC > 2). Thus, equations including one-sided competition by broad-leaved trees were good predictors, irrespective of the mode of competition by conifers (Eqs. 1, 3; Table 1).
Fig. 2

Relationship between ADGR and DBH of Abies sachalinensis during a decade (1997–2007) at 1,000 m a.s.l. (a) and 200 m a.s.l. (b) in Hokkaido, northern Japan. Pearson correlation coefficient is R = 0.473 (p < 0.001, n = 548) for Nukabira (1,000 m a.s.l.) and R = 0.521 (p < 0.001, n = 263) for Urahoro (200 m a.s.l.)

Table 1

Selected equations for growth of Abies sachalinensis by generalized linear models with stepwise function

 

Combinationa

a0

a1

a2

a3

a4

AIC

Eq. 1

C1

B1

0.780***

−0.649***

0.672***

−0.0176***

−0.0170***

2338.9

Eq. 2

C1

B2

0.749***

−0.660***

0.696***

−0.0180***

−0.0164***

2342.4

Eq. 3

C2

B1

0.617**

−0.598***

0.760***

−0.0188***

−0.0173***

2339.2

Eq. 4

C2

B2

0.577**

−0.608***

0.786***

−0.0191***

−0.0165***

2343.1

Growth model (n = 810): \( {\text{ADGR}} = a_{0} + a_{1} \,{\text{Altitude}} + a_{2} \ln \,{\text{DBH}} + a_{3} \sum {\text{BA}}_{{{\text{C}}_{i} }} + a_{4} \sum {\text{BA}}_{{{\text{B}}_{i} }} \)

For the subscript i of the equations, 1 indicates one-sided competition, and 2 indicates two-sided competition

a These represent subscripts of \( {\text{BA}}_{{\text{C}_{i} }} \) and \( {\text{BA}}_{{\text{B}_{i} }} \) of equations

C conifers, B broad-leaved trees, 1 one-sided competition, 2 two-sided competition

p < 0.05, **p < 0.01, ***p < 0.001

Local crowding by taller neighbors than the target tree was highly correlated with that by all neighbors for conifers (R2 = 0.83) and broad-leaved hardwood trees (R2 = 0.98). Nevertheless, lower AIC values of one-sided competition than that of two-sided competition for broad-leaved trees indicates that broad-leaved trees affected the growth of target trees through one-sided competition. On the contrary, similar AIC values between one-sided and two-sided competition for conifers indicates that effect of two-sided competition was more prominent in conifers as compared to broad-leaved trees. Thus, ADGR of A. sachalinensis was estimated by Eq. 3 assuming two-sided competition by conifers and one-sided competition by broad-leaved trees. ADGR was greater in larger trees, and was reduced by local crowding of neighboring trees (Fig. 3). ADGR was lower at high altitude than low altitude at any DBH and at any local crowding conditions (Fig. 3).
Fig. 3

Estimated ADGR of Abies sachalinensis against DBH at four local crowding conditions. Solid and broken lines indicate 1,000 m a.s.l. and 200 m a.s.l., respectively. Mode of competition was assumed as two-sided competition by conifers and one-sided competition by broad-leaved trees (Eq. 3; Table 1)

Annual mortality was calculated for four DBH classes. The size-dependent mortality was similar between the two altitudes for all dead trees, standing-dead trees, stem-broken and uprooted trees (Fig. 4). Confidence intervals of 95% overlapped between the two altitudes at each size class for each form of mortality (Fig. 4). Mortality was a U-shaped pattern against DBH with two peaks at small and large size classes for all dead trees that include standing-dead, stem-broken, and uprooted trees (Fig. 4a). However, this pattern was not evident for mortality calculated for standing-dead trees only (Fig. 4b). Mortality of standing-dead trees was highest in the small size class. Larger trees tended to die in the form of stem-broken and uprooting at the two altitudes (Fig. 4c). Although there were no statistical differences in mortality between the two altitudes, mean mortality of stem-broken and uprooting was slightly greater at low altitude than at high altitude.
Fig. 4

Relationship between annual mortality and DBH at 1,000 m a.s.l. (solid circles) and 200 m a.s.l. (open circles) in Hokkaido, northern Japan. Mortality was calculated for size classes 2.0–9.9, 10.0–19.9, 20.0–29.9, ≥30.0 cm DBH at each altitude for all dead trees (a), standing-dead trees (b), stem-broken and uprooted trees (c). Vertical bars indicate 95% confidence intervals generated using a 1,000-iteration bootstrap technique (a, b, c)

Generalized linear model showed that neither \( {\text{BA}}_{{{\text{C}}_{i} }} \) (b5) nor \( {\text{BA}}_{{{\text{B}}_{i} }} \) (b6) was selected for mortality of standing-dead trees in any combination of local crowding. Thus, the final equation became same among the four cases (Eq. 5; Table 2). By contrast, local crowding by broad-leaved trees (b6) had negative effects on mortality of stem-broken and uprooted trees for both one-sided and two-sided competition. Local crowding by conifers was not selected for both cases of one-sided and two-sided competition (Eqs. 6, 7; Table 2). Although equation including the one-sided competition by broad-leaved trees (Eq. 6) is substantially same with the best model including the two-sided competition (Eq. 6) (ΔAIC was 1.9), Eq. 7 is used for further analyses hereafter because of the best model.
Table 2

Selected equations for mortality of Abies sachalinensis by generalized linear models with stepwise function

 

Combinationa

b0

b1

b2

b3

b4

b5

b6

AIC

For standing-dead trees

Eq. 5

C1-2

B1-2

0.513

 

−0.140***

0.00261**

−2.235***

  

515.6

For stem-broken and uprooted trees

Eq. 6

C1-2

B1

−2.936***

−0.915*

0.115**

−0.00139

−0.203

 

−0.0362*

430.2

Eq. 7

C1-2

B2

−2.748***

−1.115**

0.116**

−0.00133

−0.204

 

−0.0408**

428.3

For all dead trees

Eq. 8

C1-2

B1

−0.354

 

−0.0682**

0.00187***

−0.837***

  

835.1

Eq. 9

C1-2

B2

0.189

−0.489

−0.0628*

0.00172**

−0.870***

 

−0.0156

835.6

Mortality model (n = 993): \( M = {\frac{1}{{1 + \exp [ - (b_{0} + b_{1} {\text{Altitude}} + b_{2} {\text{DBH}} + b_{3} {\text{DBH}}^{2} + b_{ 4} {\text{ADGR}} + b_{ 5} \sum {\text{BA}}_{{{\text{C}}_{i} }} + b_{6} \sum {\text{BA}}_{{{\text{B}}_{i} }} )]}}} \)

For the subscript i of equations, 1 indicates one-sided competition, and 2 indicates two-sided competition

a These represent subscripts of \( {\text{BA}}_{{{\text{C}}_{i} }} \) and \( {\text{BA}}_{{{\text{B}}_{i} }} \) of equations

C conifers, B broad-leaved trees, 1 one-sided competition, 2 two-sided competition. C1-2 (or B1-2) means that final equation is same between the one-sided and two-sided competition of conifers (or broad-leaved trees)

p < 0.05, **p < 0.01, ***p < 0.001

Mortality correlated well with DBH, squared DBH, and ADGR (1994–1997) before the mortality measurement period (1997–2007) for standing-dead trees (Eq. 5; Table 2). In the mortality model for standing-dead trees (Eq. 5), trees with low growth rates tended to die, i.e., negative coefficients (b4) of ADGR (Table 2; Fig. 5a). The coefficient (b4) of ADGR was more negative in the mortality model of standing-dead trees (−2.235 of Eq. 5) than that of stem-broken and uprooted trees (−0.204 of Eq. 7) (Table 2), indicating that mortality due to suppression and senescence (standing-dead trees) is more regulated by ADGR compared to mortality due to disturbances (stem-broken and uprooted trees). Although it is expected that disturbances kill trees irrespective of the growth rates, the weak correlation of stem-broken and uprooted trees probably reflected the inclusion of the stem-broken and uprooted trees after being standing dead. Mortality of trees with ADGR lower than 1 mm sharply increased in standing-dead trees compared to stem-broken and uprooted trees (Fig. 5). Altitude did not affect mortality of standing-dead trees directly because altitude was not selected by stepwise function (b1 of Eq. 5; Table 2). By contrast, mortality of stem-broken and uprooted trees was greater at any size at low altitude than at high altitude (b1 of Eq. 7; Table 2; Fig. 5b, c).
Fig. 5

Estimated decadal mortality (1997–2007) of Abies sachalinensis in relation to ADGR during the 3 years (1994–1997) before the mortality measurement for standing-dead trees (a), and stem-broken and uprooted trees at 1,000 m a.s.l. (b) and at 200 m a.s.l. (c). Mortality was estimated at four DBH by the generalized linear model listed in Table 2 (Eqs. 5, 7), i.e., DBH 5 cm (solid circles), 10 cm (open circles), 20 cm (solid triangles), and 40 cm (open triangles)

Size-dependent mortality was estimated at four conditions of local crowding by the generalized linear model with binomial distribution (Eqs. 5, 7; Table 2). Mortality of standing-dead trees decreased with the increase of DBH, and mortality of small trees (<10–20 cm DBH) was greater at high altitude than low altitude at any conditions of local crowding (Fig. 6). Although altitude was not selected in the mortality model of standing-dead trees (Eq. 5), high mortality at high altitude is ascribed to the low growth rate because mortality of standing-dead trees was strongly regulated by ADGR (Eq. 5; Table 2). Mortality of standing-dead trees at the two altitudes increased with increasing local crowding, especially for trees smaller than 20 cm DBH (Fig. 6). Unlike mortality of standing-dead trees, mortality of stem-broken and uprooted trees decreased in crowded conditions at the two altitudes (Fig. 7). In addition, mortality of stem-broken and uprooted trees increased with DBH (b2 of Eq. 7), and was greater at low altitude than high altitude (b1 of Eq. 7; Fig. 7). These patterns were similar to the actual pattern (Fig. 4c). However, the mortality at low altitude seems to be overestimated than the actual mortality (Figs. 4c, 7).
Fig. 6

Estimated decadal mortality (1997–2007) of Abies sachalinensis for standing-dead trees at 1,000 m a.s.l. (solid line) and 200 m a.s.l. (broken line) in Hokkaido, northern Japan. Mortality was estimated for four local crowding conditions by the generalized linear model (Eq. 5; Table 2). Local crowding is \( \sum {\text{BA}}_{{{\text{C}}_{2} }} \) = \( \sum {\text{BA}}_{{{\text{B}}_{1} }} \) = 0 cm2 m−2 (a), \( \sum {\text{BA}}_{{{\text{C}}_{2} }} \) = \( \sum {\text{BA}}_{{{\text{B}}_{1} }} \) = 10 cm2 m−2 (b), \( \sum {\text{BA}}_{{{\text{C}}_{2} }} \) = \( \sum {\text{BA}}_{{{\text{B}}_{1} }} \) = 20 cm2 m−2 (c), and \( \sum {\text{BA}}_{{{\text{C}}_{2} }} \) = \( \sum {\text{BA}}_{{{\text{B}}_{1} }} \) = 40 cm2 m−2 (d). ADGR was calculated with Eq. (3)

Fig. 7

Estimated decadal mortality (1997–2007) of Abies sachalinensis for stem-broken and uprooted trees at 1,000 m a.s.l. (solid line) and 200 m a.s.l. (broken line) in Hokkaido, northern Japan. Mortality was estimated for four local crowding conditions by the generalized linear model (Eq. 7; Table 2). Local crowding is \( \sum {\text{BA}}_{{{\text{C}}_{2} }} \) = \( \sum {\text{BA}}_{{{\text{B}}_{1} }} \) = 0 cm2 m−2 (a), \( \sum {\text{BA}}_{{{\text{C}}_{2} }} \) = \( \sum {\text{BA}}_{{{\text{B}}_{1} }} \) = 10 cm2 m−2 (b), \( \sum {\text{BA}}_{{{\text{C}}_{2} }} \) = \( \sum {\text{BA}}_{{{\text{B}}_{1} }} \) = 20 cm2 m−2 (c), and \( \sum {\text{BA}}_{{{\text{C}}_{2} }} \) = \( \sum {\text{BA}}_{{{\text{B}}_{1} }} \) = 40 cm2 m−2 (d). ADGR was calculated with Eq. (3)

Discussion

Influences of altitude and competition expressed as local crowding on growth and mortality of A. sachalinensis are summarized based on the results of this study (Fig. 8). While broad-leaved trees reduced growth of A. sachalinensis through one-sided competition (\( {\text{BA}}_{{{\text{B}}_{i} }} \)), the effect of one-sided competition (\( {\text{BA}}_{{C_{i} }} \)) was substantially the same as that of two-sided competition (\( {\text{BA}}_{{{\text{C}}_{2} }} \)) in conifers, according to the AIC values (Table 1; Fig. 8). This suggests that mode of competition by conifers is closer to two-sided competition compared to deciduous broad-leaved trees. One-sided and two-sided competition were also observed in other broad-leaved and coniferous forests, respectively (Kubota and Hara 1995; Kikuzawa and Umeki 1996; Umeki 2001; Nishimura et al. 2002, 2005). Kikuzawa and Umeki (1996) reported that growth of target trees was reduced not only by neighbors larger than target trees but also by neighbors smaller than target trees in a stand dominated by a conifer Picea abies (L.) Karst. because the lower part of the deep crown of P. abies was shaded by smaller neighbors. By contrast, broad-leaved trees with shallow crowns showed one-sided competition (Kikuzawa and Umeki 1996). Their result suggests that the difference in canopy structure caused the difference in the mode of competition between conifers and broad-leaved trees. Thus, the result of this study supports their result.
Fig. 8

Influences of altitude and competition on growth and mortality of Abies sachalinensis, based on the results of this study (Tables 1, 2). Competition was expressed as local crowding \( ({\text{BA}}_{{{\text{C}}_{i} }} \) and \( {\text{BA}}_{{{\text{B}}_{i} }}, \) where i = 1 or 2). All arrows indicate negative influences as denoted by minus symbols

This study showed that the growth rate of A. sachalinensis was greater at low altitude than high altitude at any DBH and local crowding (Fig. 8). This result is reasonable because a long growth period at a low altitude is beneficial for growth of trees at any DBH and local crowding. Furthermore, the photosynthetic rate of plants generally depends on temperature (Cunningham and Read 2002), and warm temperature often increases photosynthetic rates of alpine and subalpine plants (DeLucia and Smith 1987). Thus, the greater growth rate at low altitude than at high altitude is due to the increase of both growth period and air temperature. Although many studies showed that plant growth is greater at lower altitudes, these studies did not show if growth of plants increases at lower altitudes irrespective of plant size (Kajimoto 1993; Oleksyn et al. 1998; Mäkinen et al. 2002; Takahashi 2003b). On the contrary, Li et al. (2003) analyzed the growth trajectory of trunk height and diameter in even-aged stands (tree age at 27 years for Larix deciduas Mill. and 28 years for Picea abies) at three altitudes within the subalpine zone (1,680–1,940 m a.s.l.) in Austria, and showed that height growth of P. abies and L. deciduas was greater at lower altitude as tree size increased. Kurahashi et al. (1995) also examined the growth of Abies sachalinensis saplings grown at various altitudes between 230 m and 1,200 m a.s.l. in northern Japan during 18 years from seeding, and revealed that height growth was greater at lower altitudes as tree size increased, as found in Li et al. (2003). The growth pattern of Li et al. (2003) and Kurahashi et al. (1995) differs from that of this study. However, the size of trees examined is considerably smaller in these two studies (trunk height was up to several meters) than in this study (DBH > 2 cm). Li et al. (2003) also described that the growth of small seedlings was reduced by herbaceous plants, which reduced the altitudinal effects on tree growth at the seedling stage. Trees larger than 2 cm DBH (the minimum size in this study) hardly compete for light with herbaceous plants. Factors affecting tree growth probably change with tree life stages (seedlings, saplings, young and adult trees). Thus, it is possible that a long growth period at a low altitude contributes to greater growth rate of A. sachalinensis (DBH > 2 cm), irrespective of tree size and local crowding, compared to high altitude.

Mortality of standing-dead trees was greater for trees with lower growth rates, and altitude had no direct effect on the mortality (Fig. 8). Previous studies also indicated that sapling mortality increased with a reduction of growth rates (Kobe et al. 1995; Caspersen and Kobe 2001; Kunstler et al. 2005), which is probably due to a shortage of photosynthetic production. Umeki (2002) also reported that slow-growing trees are more likely to die than fast-growing trees in A. sachalinensis greater than 5 cm DBH. Thus, it is suggested that carbon balance is important for mortality of A. sachalinensis in the form of standing-dead trees. Since local crowding reduces the growth rate of A. sachalinensis, the mortality in the form of standing-dead trees would be greater in more crowded conditions.

By contrast, local crowding of broad-leaved trees \( ({\text{BA}}_{{{\text{B}}_{2} }}) \) reduced mortality in the form of stem breaking and uprooting (Fig. 8). Umeki (2002) also found a negative relationship between local crowding and mortality of A. sachalinensis, although he analyzed mortality of all dead trees that included standing-dead, stem-broken, and uprooted trees. He described that neighboring trees act as protectors against strong winds because isolated trees are exposed to strong winds and tend to be fallen. The result of this study supports his speculation because a negative relationship between local crowding and mortality was found only in the mortality due to disturbances (stem breaking and uprooting) not in the mortality due to suppression (standing dead). It is unknown why local crowding of conifers did not have an influence on the mortality in the form of stem breaking and uprooting. However, broad-leaved trees are more resistant to strong winds compared to conifers (Kohyama 1984). Thus, strong wind resistance of the broad-leaved trees may reflect the negative relationship between local crowding and mortality.

Actual stand-level mortality was similar between the two altitudes, nevertheless, estimated mortality of standing-dead trees was greater at high altitude than at low altitude. There are two possible factors for this. First, more trees at low altitude tended to die of disturbances than at high altitude. Secondly, many experimental and observational studies suggested that tree growth was positively correlated with air temperature during growing season at high altitudes and latitudes (Nöjd and Hari 2001; Kirdyanov et al. 2003; Takahashi et al. 2005b). High resource availability results in high stand biomass, which in turn increases competition between plants (Grime 1973; Wilson and Keddy 1986; Reader and Best 1989; Kadmon 1995). Maximum DBH and aboveground biomass of temperate coniferous forests are positively correlated with the WI (the sum of monthly mean temperature above 5°C, WI) (Takyu et al. 2005). The stand basal area was greater at low altitude (high WI) than at high altitude (low WI) in this study (49.5 vs. 38 m2 ha−1). It is possible that high stand basal area at low altitude increased tree competition by increasing local crowding for many trees. Thus, these two factors (disturbance and stand development) possibly resulted in a similar mortality between the two altitudes.

This study concludes that (1) the growth rate of A. sachalinensis is greater at low altitude than at high altitude irrespective of tree size and degree of local crowding, and (2) greater growth rates at low altitude reduce the mortality by suppression compared to high altitude, but more trees tend to die of disturbances at low altitude. Thus, growth and mortality of A. sachalinensis are largely affected by altitude and competition. Although global warming may increase growth and survival of A. sachalinensis at high altitudes by prolonging growth period, prediction on mortality due to disturbances is not simple because it largely depends on the frequency and intensity of disturbances. Thus, it needs caution for usage of altitudinal gradients as subjects of mortality responses to temperature.

Notes

Acknowledgments

This study was partially supported by grants from the Ministry of Education, Culture, Sports, Science and Technology, Japan (Nos. 15710007, 19580168). I thank Dr. T. Masaki and an anonymous reviewer for constructive comments.

References

  1. Aber JD, Federer CA (1992) A generalized, lumped-parameter model of photosynthesis, evapotranspiration and net primary production in temperate and boreal ecosystems. Oecologia 92:463–474CrossRefGoogle Scholar
  2. Apple ME, Olszyk DM, Ormrod DP, Lewis J, Southworth D, Tingey D (2000) Morphology and stomatal function of Douglas fir needles exposed to climate change: elevated CO2 and temperature. Int J Plant Sci 161:127–132CrossRefPubMedGoogle Scholar
  3. Arft AM, Walker MD, Gurevitch J, Alatalo JM, Bret-Harte MS, Dale M, Diemer M, Gugerli F, Henry GHR, Jones MH, Hollister RD, Jónsdóttir IS, Laine K, Lévesque E, Marion GM, Molau U, Mølgaard P, Nordenhäll U, Raszhivin V, Robinson CH, Starr G, Stenström A, Totland Ø, Turner PL, Walker LJ, Webber PJ, Welker JM, Wookey PA (1999) Responses of tundra plants to experimental warming: meta-analysis of the international tundra experiment. Ecol Monogr 69:491–511Google Scholar
  4. Buckley BM, Cook ER, Peterson MJ, Barbetti M (1997) A changing temperature response with elevation for Lagarostrobos franklinii in Tasmania, Australia. Clim Change 36:477–498CrossRefGoogle Scholar
  5. Burnham KP, Anderson DR (2002) Model selection and multi-model inference: a practical information theoretic approach. Springer, Berlin Heidelberg New YorkGoogle Scholar
  6. Canham CD, LePage PT, Coates KD (2004) A neighborhood analysis of canopy tree competition: effects of shading versus crowding. Can J For Res 34:778–787CrossRefGoogle Scholar
  7. Caspersen JP, Kobe RK (2001) Interspecific variation in sapling mortality in relation to growth and soil moisture. Oikos 92:160–168CrossRefGoogle Scholar
  8. Coomes DA, Allen RB (2007) Effects of size, competition and altitude on tree growth. J Ecol 95:1084–1097CrossRefGoogle Scholar
  9. Crowley PH (1992) Resampling methods for computation-intensive data analysis in ecology and evolution. Ann Rev Ecol Syst 23:405–447CrossRefGoogle Scholar
  10. Cunningham SC, Read J (2002) Comparison of temperate and tropical rainforest tree species: photosynthetic responses to growth temperature. Oecologia 133:112–119CrossRefGoogle Scholar
  11. DeLucia EH, Smith WK (1987) Air and soil temperature limitations on photosynthesis in Engelmann spruce during summer. Can J For Res 17:527–533CrossRefGoogle Scholar
  12. Dorrepaal E, Aerts R, Cornelissen JHC, Callaghan TV, van Logtestijn RSP (2003) Summer warming and increased winter snow cover affect Sphagnum fuscum growth, structure and production in a sub-arctic bog. Glob Chang Biol 10:93–104CrossRefGoogle Scholar
  13. Duncan RP (1991) Competition and the coexistence of species in a mixed podocarp stand. J Ecol 79:1073–1084CrossRefGoogle Scholar
  14. Fransworth EJ, Nunez-Farfan J, Careaga SA, Bazzaz FA (1995) Phenology and growth of three temperate forest life forms in response to artificial soil warming. J Ecol 83:967–977CrossRefGoogle Scholar
  15. Germino MJ, Smith WK, Resor AC (2002) Conifer seedling distribution and survival in an alpine-treeline ecotone. Plant Ecol 162:157–168CrossRefGoogle Scholar
  16. Gostev M, Wiles G, D’Arrigo R, Jacoby G, Khomentovsky P (1996) Early summer temperatures since 1670 A.D. for Central Kamchatka reconstructed based on a Siberian larch tree-ring width chronology. Can J For Res 26:2048–2052CrossRefGoogle Scholar
  17. Grime JP (1973) Competitive exclusion in herbaceous vegetation. Nature 242:344–347CrossRefGoogle Scholar
  18. Hara T (1992) Effects of the mode of competition on stationary size distribution in plant populations. Ann Bot 69:509–513Google Scholar
  19. Hara T, Yokoishi E, Ishiwata M, Kimura M (1994) Canopy tree competition and species coexistence in an Abies veitchii and A. mariesii mixed subalpine forest, central Japan. Ecoscience 1:239–248Google Scholar
  20. Hollister RD, Webber PJ (2000) Biotic validation of small open-top chambers in a tundra ecosystem. Glob Chang Biol 6:835–842CrossRefGoogle Scholar
  21. Hollister RD, Webber PJ, Bay C (2005) Plant response to temperature in northern Alaska: implications for predicting vegetation change. Ecology 86:1562–1570CrossRefGoogle Scholar
  22. Jones MH, Bay C, Nordenhall U (1997) Effects of experimental warming on arctic willows (Salix spp.): a comparison of responses from the Canadian High Arctic, Alaskan Arctic, and Swedish Subarctic. Glob Chang Biol 3(Suppl 1):55–60CrossRefGoogle Scholar
  23. Kadmon P (1995) Plant competition along soil moisture gradients: a field experiment with the desert annual Stipa capensis. J Ecol 83:253–262CrossRefGoogle Scholar
  24. Kajimoto T (1993) Shoot dynamics of Pinus pumila in relation to altitudinal and wind exposure gradients on the Kiso Mountain Range, central Japan. Tree Physiol 13:41–53PubMedGoogle Scholar
  25. Kelly AE, Goulden ML (2008) Rapid shifts in plant distribution with recent climate change. Proc Natl Acad Sci USA 105:11823–11826CrossRefPubMedGoogle Scholar
  26. Kikuzawa K (1999) Theoretical relationships between mean plant size, size distribution and self thinning under one-sided competition. Ann Bot 83:11–18CrossRefGoogle Scholar
  27. Kikuzawa K, Umeki K (1996) Effect of canopy structure on degree of asymmetry of competition in two forest stands in northern Japan. Ann Bot 77:565–571CrossRefGoogle Scholar
  28. Kikvidze Z (2002) Facilitation and competition in alpine plant communities. Glob Environ Res 16:53–58Google Scholar
  29. Kira T (1948) On the altitudinal arrangement of climatic zones in Japan. Kanti-Nogaku 2:143–173 (in Japanese)Google Scholar
  30. Kirdyanov A, Hughes M, Yaganov E, Schweingruber F, Silkin P (2003) The importance of early summer temperature and date of snow melt for tree growth in the Siberian Subarctic. Trees 17:61–69CrossRefGoogle Scholar
  31. Klanderud K, Totland O (2005) Simulated climate change altered dominance hierarchies and diversity of an alpine biodiversity hotspot. Ecology 86:2047–2054CrossRefGoogle Scholar
  32. Kobe RK, Pacala SW, Silander JA, Canham CD (1995) Juvenile tree survivorship as a component of shade tolerance. Ecol Appl 5:517–532CrossRefGoogle Scholar
  33. Kohyama T (1984) Regeneration and coexistence of two Abies species dominating subalpine forests in central Japan. Oecologia 62:156–161CrossRefGoogle Scholar
  34. Kohyama T (1992) Size-structured multi-species model of rain forest trees. Funct Ecol 6:206–212CrossRefGoogle Scholar
  35. Kohyama T (1993) Size-structured tree populations in gap-dynamic forest: the forest architecture hypothesis for the stable coexistence of species. J Ecol 81:131–143CrossRefGoogle Scholar
  36. Kojima S (1991) Classification and ecological characterization of coniferous forest phytogeocoenoses of Hokkaido, Japan. Vegetatio 96:25–42CrossRefGoogle Scholar
  37. Körner C (2007) The use of ‘altitude’ in ecological research. Trends Ecol Evol 22:569–574CrossRefPubMedGoogle Scholar
  38. Kubota Y, Hara T (1995) Tree competition and species coexistence in a sub-boreal forest, northern Japan. Ann Bot 76:503–512CrossRefGoogle Scholar
  39. Kullman L (2002) Rapid recent range-margin rise of tree and shrub species in the Swedish Scandes. J Ecol 90:68–77CrossRefGoogle Scholar
  40. Kunstler G, Curt T, Bouchaud M, Lepart J (2005) Growth, mortality, and morphological response of European beech and downy oak along a light gradient in sub-Mediterranean forest. Can J For Res 35:1657–1668CrossRefGoogle Scholar
  41. Kurahashi A, Kisanuki H, Ogasawara S (1995) Variation in growth responses of Saghalien fir (Abies sachalinensis) seedlings associated with altitudinal gradients—results of set of provenance tests at various altitudes in eighteen years. Trans Meet Hokkaido Br Jpn For Soc 43:200–202 (in Japanese)Google Scholar
  42. Li MH, Yang J, Krauchi N (2003) Growth responses of Picea abies and Larix decidua to elevation in subalpine areas of Tyrol, Austria. Can J For Res 33:653–662CrossRefGoogle Scholar
  43. Lindner M, Lasch P, Erhard M (2000) Alternative forest management strategies under climatic change—prospects for gap model applications in risk analyses. Silva Fenn 34:101–111Google Scholar
  44. Mäkinen H, Nöjd P, Kahle HP, Newmann U, Tveite B, Mielikäinen K, Röhle H, Spiecker H (2002) Radial growth variation of Norway spruce (Picea abies (L.) Karst.) across latitudinal and altitudinal gradients in central and northern Europe. For Ecol Manage 171:243–259CrossRefGoogle Scholar
  45. Molau U (1997) Responses to natural climatic variation and experimental warming in two tundra plant species with contrasting life forms: Cassiope tetragona and Ranunculus nivalis. Glob Chang Biol 3(Suppl 1):97–107CrossRefGoogle Scholar
  46. Monserud RA, Sterba H (1999) Modeling individual tree mortality for Austrian forest species. For Ecol Manage 113:109–123CrossRefGoogle Scholar
  47. Nakashizuka T, Kohyama T (1995) The significance of the asymmetric effect of crowding for coexistence in a mixed temperate forest. J Veg Sci 6:509–516CrossRefGoogle Scholar
  48. Nakashizuka T, Iida S, Tanaka H, Shibata M, Abe S, Masaki T, Niiyama K (1992) Community dynamics of Ogawa Forest Reserve, a species rich deciduous forest, central Japan. Vegetatio 103:105–112Google Scholar
  49. Nishimura N, Hara T, Miura M, Manabe T, Yamamoto S (2002) Tree competition and species coexistence in a warm-temperate old-growth evergreen broad-leaved forest in Japan. Plant Ecol 164:235–248CrossRefGoogle Scholar
  50. Nishimura N, Hara T, Kawatani M, Hoshino D, Yamamoto S (2005) Promotion of species co-existence in old-growth coniferous forest through interplay of life-history strategy and tree competition. J Veg Sci 16:549–558CrossRefGoogle Scholar
  51. Nöjd P, Hari P (2001) The effect of temperature on the radial growth of Scots pine in northernmost Fennoscandia. For Ecol Manage 142:65–77CrossRefGoogle Scholar
  52. Oleksyn J, Tjoelker MG, Reich PB (1998) Adaptation to changing environment in Scots pine population across a latitudinal gradient. Silva Fenn 32:129–140Google Scholar
  53. Parmesan C, Yohe G (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature 421:37–42CrossRefPubMedGoogle Scholar
  54. Paulsen J, Weber UM, Körner C (2000) Tree growth near treeline: abrupt or gradual reduction with altitude? Arct Antarc Alp Res 32:14–20CrossRefGoogle Scholar
  55. Peterson DW, Peterson DL (2001) Mountain hemlock growth responds to climatic variability at annual and decadal time scales. Ecology 82:3330–3345CrossRefGoogle Scholar
  56. R development core team (2009) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, http://www.R-project.org
  57. Reader RJ, Best BJ (1989) Variation in competition along an environmental gradient: Hieracium floribundum in an abandoned pasture. J Ecol 77:673–684CrossRefGoogle Scholar
  58. Sheil D, May RM (1996) Mortality and recruitment rate evaluations in heterogeneous tropical forests. J Ecol 84:91–100CrossRefGoogle Scholar
  59. Šrůtek M, Lepš J (1994) Variation in structure of Larix olgensis stands along the altitudinal gradient on Paektu-san, Changbai-shan, North Korea. Arc Alp Res 26:166–173CrossRefGoogle Scholar
  60. Takahashi K (1994) Effect of size structure, forest floor type and disturbance regime on tree species composition in a coniferous forest in Japan. J Ecol 82:769–773CrossRefGoogle Scholar
  61. Takahashi K (1996) Plastic response of crown architecture to crowding in understorey trees of two co-dominating conifers. Ann Bot 77:159–164CrossRefGoogle Scholar
  62. Takahashi K (2003a) Difference between evergreen conifers and deciduous hardwood trees in suppressing the growth of Abies sachalinensis understory trees. Veg Sci 20:65–68Google Scholar
  63. Takahashi K (2003b) Effects of climatic conditions on shoot elongation of alpine dwarf pine (Pinus pumila) at its upper and lower altitudinal limits in central Japan. Arct Antarc Alp Res 35:1–7CrossRefGoogle Scholar
  64. Takahashi K (2005) Effects of artificial warming on shoot elongation of alpine dwarf pine (Pinus pumila) on Mt. Shogigashira, central Japan. Arct Antarc Alp Res 37:620–625CrossRefGoogle Scholar
  65. Takahashi K, Kohyama T (1999) Size-structure dynamics of two conifers in relation to understorey dwarf bamboo: a simulation study. J Veg Sci 10:833–842CrossRefGoogle Scholar
  66. Takahashi K, Azuma H, Yasue K (2003) Effects of climate on the radial growth of tree species in the upper and lower distribution limits of an altitudinal ecotone on Mt. Norikura, central Japan. Ecol Res 18:549–558CrossRefGoogle Scholar
  67. Takahashi K, Nagano S, Maruta E (2005a) Relationships between vegetation cover and seedling distribution of the alpine dwarf pine Pinus pumila on Mt. Norikura, central Japan. Veg Sci 22:147–152Google Scholar
  68. Takahashi K, Tokumitsu Y, Yasue K (2005b) Climatic factors affecting the tree-ring width of Betula ermanii at the timberline on Mount Norikura, central Japan. Ecol Res 20:445–451CrossRefGoogle Scholar
  69. Takyu M, Kubota Y, Aiba S, Seino T, Nishimura T (2005) Pattern of changes in species diversity, structure and dynamics of forest ecosystems along latitudinal gradients in East Asia. Ecol Res 20:287–296CrossRefGoogle Scholar
  70. Tardif JC, Conciatori F, Nantel P, Gagnon D (2006) Radial growth and climate responses of white oak (Quercus alba) and northern red oak (Quercus rubra) at the northern distribution limit of white oak in Quebec, Canada. J Biogeogr 3:1657–1669CrossRefGoogle Scholar
  71. Umeki K (2001) Growth characteristics of six tree species on Hokkaido Island, northern Japan. Ecol Res 16:435–450CrossRefGoogle Scholar
  72. Umeki K (2002) Tree mortality of five major species on Hokkaido Island, northern Japan. Ecol Res 17:575–589CrossRefGoogle Scholar
  73. Umeki K, Kikuzawa K (1999) Long-term growth dynamics of natural forests in Hokkaido, northern Japan. J Veg Sci 10:815–824CrossRefGoogle Scholar
  74. van der Werf GW, Sass-Klaassen UGW, Mohren GMJ (2007) The impact of the 2003 summer drought on the intra-annual growth pattern of beech (Fagus sylvatica L.) and oak (Quercus robur L.) on a dry site in the Netherlands. Dendrochronologia 25:103–112CrossRefGoogle Scholar
  75. Weiner J (1990) Asymmetric competition in plant populations. Trends Ecol Evol 5:360–364CrossRefGoogle Scholar
  76. Wilson SD, Keddy PA (1986) Measuring diffuse competition along an environmental gradient: results from a shoreline plant community. Am Nat 127:862–869CrossRefGoogle Scholar

Copyright information

© The Ecological Society of Japan 2010

Authors and Affiliations

  1. 1.Department of Biology, Faculty of ScienceShinshu UniversityMatsumotoJapan
  2. 2.Institute of Mountain ScienceShinshu UniversityMatsumotoJapan

Personalised recommendations