Introduction

Nutrient-poor soils represent one of the major limitations of primary succession of plants after deglaciation and volcanic eruption (Chapin et al. 1994; Chiba and Hirose 1993; Masuzawa 1985). Of limiting nutrients, nitrogen is the most important element for plant growth because photosynthetic enzymes are nitrogen compounds and therefore photosynthetic capacity positively correlates with nitrogen concentration in both soils and leaves (Chapin 1980; Evans 1989; Hirose and Werger 1987; Reich et al. 1991). To cope with nitrogen deficiency, plants have evolved symbiotic relations with atmospheric nitrogen-fixing bacteria, which effectively increase nitrogen acquisition under nutrient-poor conditions (Huss-Danell 1997; Tjepkema et al. 1986). For instance, alders (Alnus spp.) often dominate at early stages of primary succession after deglaciation because of its symbiotic relations with nitrogen-fixing bacteria Frankia spp. (Bormann and Sidle 1990; Chapin et al. 1994; Dawson 2008; Jones and del Moral 2005). Frankia spp. live in root nodules of actinorhizal plants (Valdés 2008; Wall and Berry 2008; Weber et al. 1987), thereby contributing to nitrogen acquisition, which in turn enables early-successional shrub species such as alder to colonize early stages of primary succession. Although alders are replaced by late-successional species with the progress of succession because alders are shade-intolerant species (Kikuzawa 1983, 1984), carbon and nitrogen accumulations in soil increase with the development of alder forests (Chapin et al. 1994). Therefore, the growth of alders is important for the ecosystem function during primary succession.

Plants provide photosynthetic carbon compounds for symbiotic relationships with nitrogen-fixing bacteria. High carbon allocation to roots and their symbiotic relations increases the acquisition of soil water and nutrients (Markesteijn and Poorter 2009). Although much carbon support for nitrogen-fixing bacteria can be advantageous for nitrogen acquisition, it simultaneously decreases plant growth because plants allocate much photosynthate to root nodules at the expense of growth. For example, Mahon (1977) showed that 17 g of carbohydrate is consumed per gram of nitrogen fixing in Pisum sativum. The cost of nitrogen fixation by symbiotic and autotrophic nitrogen fixers is high relative to that of absorbing ammonium or nitrate (Chapin et al. 2011; Minchin and Witty 2005; Schulze et al. 1999). Furthermore, less allocation to leaves would decrease the net production per plant, which in turn decreases the subsequent plant growth (Takahashi et al. 2001). Thus, it is expected that there is an ideal carbon allocation to root nodules and leaves for plants to increase nitrogen acquisition and net production, simultaneously. However, there have been no studies examining quantitative characteristics of root nodules in relation to nitrogen acquisition and net production.

Allometric relationships between two organs of plants can be used to quantify biomass allocation patterns (Kohyama 1987; Kohyama and Hotta 1990; Takahashi and Obata 2014; Takahashi et al. 2001). The allocation pattern between two organs can be derived from the slope of the double-log linear regression. If a plant organ increases with the same rate of increase of another organ, the slope of the double-log allometric equation between them should be 1.0, i.e., isometric scaling relation. For example, Takahashi and Kohyama (1997) examined the crown development of two small palm species (Licuala spp.) in a tropical rain forest. The double-log regression slope of lamina mass against petiole mass was smaller than 1.0, indicating that larger palm individuals invested more in petiole than in leaf lamina to avoid self-shading within a crown. Therefore, the allometric analysis provides useful information for patterns of carbon allocation among plant organs. Tobita et al. (2010) showed that root nodule mass per plant exponentially increased with stem diameter in Alnus hirsuta var. sibirica. However, there have been no allometric studies that examined quantitative relationships among leaf mass, root nodule mass, whole-plant nitrogen and net production per plant.

The amount of nitrogen-fixing bacteria is predicted to be proportional to the root nodule volume or mass (Mylona et al. 1995; Pawlowski and Bisseling 1996). Tobita et al. (2013) showed that the rate of nitrogen fixation per root nodule mass did not correlate with plant size in A. hirsuta var. sibirica, indicating that the amount of nitrogen acquisition per plant by nitrogen-fixing bacteria is determined by the amount of nitrogen-fixing bacteria irrespective of plant size. Hooker and Wheeler (1987) also observed a similar pattern in other Alnus species. If leaves and nitrogen-fixing bacteria are related as carbon and nitrogen acquisition organs respectively, it is expected that leaf mass, root nodule mass and whole-plant nitrogen will scale isometrically.

Alnus fruticosa Rupr. is a typical early-successional shrub species at young moraines in Kamchatka, Russian Far East (Dolezal et al. 2008). The purpose of this study was to clarify how A. fruticosa increases the net production per plant at a young moraine under nutrient-poor conditions. We investigated the pattern of allocation between plant organs including a relationship between whole-plant nitrogen and root nodule mass to reveal whether leaf mass, root nodule mass, whole-plant nitrogen and net production increase isometrically.

Materials and methods

Study site

This study was carried out at the Koryto Glacier Valley (300 m above sea level, 54°45′N, 161°40′E), the Kronotsky Peninsula, Kamchatka, Russian Far East (Figs. 1, 2). The Kronotsky Peninsula is located on the east coast of Kamchatka as a prominent massif protruding into the Pacific Ocean. The elevation of the Kronotsky Peninsula seldom exceeds 1000 m above sea level, and summit parts alone reach to an elevation as high as 1300 m a.s.l. There are 32 glaciers in this region (Vinogradov 1968), in most of which glacial landforms developed U-shaped valleys, terminal and lateral moraines and draining rivers (Fig. 2). The Koryto Glacier is located in the uppermost part of the Bol’haya Chazhma river, extending from 320 m to 1200 m a.s.l. toward northwest exposure with the total glaciated area of 8.2 km2 (Sawaguchi et al. 1999). Mean monthly temperatures of the coldest month (January) and the hottest month (August) at the study site (300 m a.s.l.) were − 9.6 and 9.8 °C, respectively, during the 1939–1999 period. Mean annual temperature was − 0.3 °C. These data were derived from the nearest weather station Stopozh at the Pacific coast (54°35′01″N, 161°11′10″E) about 50 km from the study site, using the standard lapse rate of − 0.6 °C per 100 m of elevation. The study area receives 2.5–3 m winter snowfall and 0.4 m summer rainfall. Snowpack at accumulation zones is 5–10 m deep (Dolezal et al. 2008). The growing season for plants lasted from July to September, because the mean monthly temperatures exceed 5 °C, the effective temperature for plant growth (Kira 1948), during this period.

Fig. 1
figure 1

Location of the Koryto Glacier (filled circle) in the Kronotsky Peninsula, Kamchatka, Russian Far East

Full size image
Fig. 2
figure 2

Photos of (top) the Koryto Glacier and (bottom) the U-shaped valley in mid-August, 1999

Full size image

During the Little Ice Age, the Koryto Glacier extended about 1200 m from the present glacier terminus (Sawaguchi et al. 1999). Several moraines with different ages were formed after its retreat (Fig. 3). This study was conducted in 1999 at a young moraine that was formed ca. 80 years ago (Fig. 3). Alnus fruticosa invades new moraines only a few decades after the moraine formation, and remains a dominant species at the oldest moraine (270 years old). Betula ermanii is the dominant species on surrounding slopes, but deep snow accumulation prevents the growth of B. ermanii at moraines in the valley bottom (Dolezal et al. 2008). The soil thickness was only 4 cm at the moraine examined in this study. Total nitrogen concentrations in soil were 0.014 and 0.072% in the A horizon (0–2 cm depth) and the B horizon (2–4 cm depth), respectively, and total carbon concentrations were 0.098 and 0.72%, respectively. The gravel layer was recognized under the soil layer (Sawaguchi et al. 1999). The details of glacier and vegetation dynamics can be found in several papers (Dolezal et al. 2008, 2013, 2014; Kodama et al. 1996; Muravyev et al. 1999; Sawaguchi et al. 1999; Shiraiwa et al. 1997; Yamagata et al. 1999).

Fig. 3
figure 3

Distribution of terminal moraines with the formation years at the Koryto Glacier Valley in 2000, modified from Dolezal et al. (2008) and Yamaguchi et al. (2008). This study was conducted at the moraine that was formed in 1920

Full size image

Sampling and measurements

We sampled 27 individuals of shrub Alnus fruticosa (shorter than 70 cm in stem height) at the 80-year-old terminal moraine from 15 to 17 August 1999. Although the sampling was done in mid-August, the current-year growth was completed because shoot elongation of Alnus spp. ends during the middle of the growing season (Kikuzawa 1983, 1984) and because the tree-ring width of the current-year was greater than that of the previous year in our Alnus samples (data not shown).

Vegetation at this moraine was dominated by Salix arctica and A. fruticosa, and was transition from the initial pioneer stages to Alnus dominated sands (Dolezal et al. 2008). Soil-nutrient conditions are more fertile in older moraines, so the maximum plant size of A. fruticosa is large in old moraines (Dolezal et al. 2008). A. fruticosa grows up to ca. 12 cm in trunk diameter and 2.5 m in trunk height (Dirksen and Dirksen 2007; Dolezal et al. 2008). However, we did not sample large A. fruticosa individuals at the other old moraines in this valley. The purpose of this study was to reveal quantitative relationships between leaf mass, net production and plant nitrogen at a nutrient-poor moraine. If we sampled A. fruticosa individuals from various moraines with different soil-nutrient conditions, it would cause variations of their quantitative relationships. Therefore, we sampled A. fruticosa individuals only from a young moraine. But, our dataset included the full size range of A. fruticosa at the moraine, and enabled us to analyze quantitative relationships between leaf mass, net production and plant nitrogen at the nutrient-poor moraine.

As for roots, we carefully dug up individuals with a small shovel. A. fruticosa individuals were sparsely distributed at this moraine, and all sampled individuals were open-grown individuals, i.e., not shaded by other individuals. All sampled individuals were measured for stem height, root depth, horizontal crown- and root-projection areas. The latter were estimated as an ellipse from two perpendicular diameters including that of the maximum. Sampled individuals were divided into four organs, i.e., stem, root, leaf and root nodule. A. fruticosa forms large multi-lobed root nodule galls, which can be easily detached from roots.

All samples (leaf, root, stem and root nodule) of each individual were brought to the laboratory in Japan, and were weighed after oven-drying at 80 °C for at least 2 days, and were ground into powder. Their total nitrogen concentrations were determined using a C–N analyzer (Vario EL, Elementar Analysensysteme GmbH, Hanau, Germany). The nitrogen content of each organ of each individual was obtained by the nitrogen concentration multiplied by its mass. Then, the whole-plant nitrogen content was obtained by the sum of nitrogen contents of four organs, i.e., stem, root, leaf and root nodule.

Estimation of net production

For age dating and radial growth measurement, the cut surface of a main stem at ground level was sanded to count the annual tree-rings and to measure tree-ring width. We measured distances from the pith to the tree ring of the current year (R99, mm) and to the tree-ring of the previous year (R98, mm) using a stereoscopic microscope. The relative growth rate (RGR, year−1) of stem radius was calculated by Eq. (1):

$${\text{RGR }}={\text{ ln }}{R_{{\text{99}}}} - {\text{ln }}{R_{{\text{98}}}}$$
(1)

where ln is natural logarithm. We assumed that RGR of stem diameter is equal to RGR of stem radius. We estimated the stem diameter of the previous year (D98, mm) by substituting D98 and D99 (stem diameter of the current year, mm) for R98 and R99, respectively, in Eq. (1) as:

$${D_{98}}=\exp (\ln {D_{99}} - {\text{RGR}})~$$
(2)

Net production (NP, g year−1) per plant is the current-year increase in dry mass, i.e., total dry mass of leaves (ML, g) and non-photosynthetic organs (stem, root and root nodule) produced in the current year. Total dry mass of the non-photosynthetic organs produced in the current year was estimated by subtracting the dry mass of the non-photosynthetic organs of the previous year (MSRN98, g) from that of the current year (MSRN99, g). We obtained the following allometric equation of MSRN99 against D99:

$${\text{log }}{M_{{\text{SRN99}}}}={\text{ 2}}.{\text{6}}0{\text{2 log }}{D_{{\text{99}}}} - {\text{1}}.{\text{593 }}\left( {{R^{\text{2}}}={\text{ }}0.{\text{981}},P<{\text{ }}0.00{\text{1}},n={\text{ 26}}} \right)$$
(3)

where log is common logarithm (base 10). Then, we calculated MSRN98 by substituting D98 for D99. NP per plant in the current year (1999) was calculated by Eq. (4):

$${\text{NP }}={M_{\text{L}}}+{M_{{\text{SRN99}}}} - {M_{{\text{SRN98}}}}$$
(4)

Allometric analysis

Warton et al. (2006) recommended to use the least square (LS) regression and the reduced major axis (RMA) regression to predict a Y value at a given X value and to evaluate a slope of a bivariate relationship between X and Y, respectively. Therefore, this study used the LS and RMA regressions for Eq. (3) and the other allometric relationships listed in Table 1, respectively. Whether two variables show isometric scaling or not was determined by the 95% confidence interval (CI) of the slope of the RMA regression equation on the logarithmic scale. If the 95% CI of the slope includes 1.0, the bivariate relation is isometrically scaled. The present study analyzed allometry using the program (S)MATR, version 2 (Warton et al. 2006).

Table 1 Double-log allometric equations (log Y = a log X + b, where log is common logarithm) of reduced major axis regression for various dimensions in Alnus fruticosa at a young moraine in the Koryto Glacier Valley, Kamchatka, Russia
Full size table

Results

Stem height and diameter of Alnus fruticosa increased with tree age (Fig. 4a, b). Stem height attained about 40 cm by 10 years after germination. The slope of the double-log allometric equation of root depth against stem height was 0.542 (95% CI 0.413–0.713) (Fig. 5a; Table 1), indicating that smaller/younger plants allocate disproportionally more biomass into root length growth than stem height growth. The regression slope for the root-projection area against crown-projection area included 1.0 (95% CI 1.00–1.35) (Fig. 5b; Table 1), indicating that the root-projection area isometrically increased with the crown-projection area.

Fig. 4
figure 4

Relationships of a stem height and b stem diameter to tree age of Alnus fruticosa

Full size image
Fig. 5
figure 5

Allometric relationships a between root depth and stem height and b between root-projection area and crown-projection area of Alnus fruticosa. Allometric equations are listed in Table 1

Full size image

Slopes of double-log allometric equations were significantly smaller than 1.0 for the allometry of root mass, leaf mass and root nodule mass against stem mass and for the allometry of root nodule mass against root mass, i.e., 95% CI of these slopes did not include 1.0 (Fig. 6a–d; Table 1), indicating that smaller plants invested disproportionally more biomass into resource-acquiring leaf and root tissues than to supportive tissues. One the contrary, root nodule mass increased proportionally with leaf mass on the isometric scaling, i.e., 95% CI (0.874–1.114) of the slope of the allometric equation included 1.0 (Fig. 6e; Table 1).

Fig. 6
figure 6

Allometric relationships of a root mass, b leaf mass and c root nodule mass to stem mass, d root nodule mass to root mass and e root nodule mass to leaf mass of Alnus fruticosa. Allometric equations are listed in Table 1

Full size image

Whole-plant nitrogen and leaf nitrogen per plant increased isometrically with root nodule mass (Fig. 7a, b; Table 1), i.e., 95% CI of these slopes of allometric equations included 1.0 (0.855–1.144 and 0.92–1.15, respectively). Net production per plant increased with stem mass, and the slope of the double-log allometric equation was significantly smaller than 1.0 (95% CI 0.820–0.882, Fig. 8a; Table 1), indicating that net production per plant was smaller relative to the stem mass in larger A. fruticosa. On the contrary, net production per plant isometrically increased with leaf mass and leaf nitrogen per plant, i.e., 95% CI of slopes of double-log allometric equations included 1.0 (Fig. 8b; Table 1). Although 95% CI of the slope of the allometry between the net production per plant and root nodule mass did not include 1.0, the lower limit (1.006) of the 95% CI was almost equal to 1.0, indicating isometric scaling.

Fig. 7
figure 7

Allometric relationships of a whole-plant nitrogen and b leaf nitrogen per plant to root nodule mass of Alnus fruticosa. Allometric equations are listed in Table 1

Full size image
Fig. 8
figure 8

Allometric relationships of net production per plant to a stem mass, b leaf mass, c root nodule mass and d leaf nitrogen per plant of Alnus fruticosa. Allometric equations are listed in Table 1

Full size image

Discussion

It has been often observed that soil nutrients are scarce at young moraines and increase with vegetation development (Bormann and Sidle 1990; Uri et al. 2014). This is also true in the Koryto Glacier Valley. Nitrogen and carbon concentrations in the soil were considerably low at the young moraine examined in this study (about 80 years after the formation), and these concentrations were greater at older moraines (Dolezal et al. 2008). The nitrogen concentration at the moraine examined in this study was almost equal to that at the alpine zone with scarce vegetation in Japan (Masuzawa 1985; Singh et al. 2012). Alnus spp. dominate at young moraines with nutrient-poor soil conditions in other sites (Chapin et al. 1994; Jones and del Moral 2005). Therefore, there is no doubt that A. fruticosa gained dominance at the studied moraine by acquiring much of the nitrogen necessary for growth from the atmosphere rather than from the soil.

This study showed smaller plants invest disproportionally more biomass into resource-acquiring leaf and root tissues than to supportive stem tissue. Also, smaller/younger A. fruticosa individuals have disproportionally more root nodules per root biomass than larger individuals. These allocation patterns of A. fruticosa contribute to anchoring small individuals and to getting enough carbon and soil nutrients in unstable nutrient-poor soil conditions.

Tobita et al. (2010) showed that the root nodule mass (MN, g) was proportional to the square of trunk diameter (DBH, cm) for Alnus hirsuta var. sibirica (MN = 0.442 DBH2.01). In the present study, the logarithmic allometric equation of root nodule mass against stem diameter (D, cm) was expressed as MN = 1.004 D2.16 (R2 = 0.919). The slope (2.16) of the allometric equation of the present study was almost equal to that (2.01) of Tobita et al. (2010) on the logarithmic graph, although the intercept (1.004) of the present study was greater than that (0.442) of Tobita et al. (2010). The rate of atmospheric nitrogen fixation per root nodule mass did not correlate with plant size (Tobita et al. 2013), indicating that the rate of atmospheric nitrogen fixation per plant was determined mainly by the root nodule mass per plant. Johnsen and Bongarten (1991) showed that the nitrogen fertilization treatment decreased root nodule mass and rates of atmospheric nitrogen fixation. Berninger et al. (2000) also showed, using a model, that the proportion of atmospheric nitrogen acquisition by nitrogen-fixing bacteria to the total nitrogen acquisition per plant decreased when the nitrogen absorption from the soil was large. The alder stand of Tobita et al. (2010) was established after artificial disturbances by road construction in the subalpine zone (1100 m a.s.l.) in Japan. It seems that the study site of Tobita et al. (2010) was not extremely nutrient-poor conditions because vegetation originally developed there before artificial disturbances. Therefore, it is thought that A. fruticosa is more dependent on nitrogen-fixing bacteria for nitrogen acquisition in the present study because of considerable nutrient-poor conditions in the soil, compared with the study site of Tobita et al. (2013).

Several researchers showed positive correlations between photosynthetic rates and rates of atmospheric nitrogen fixation by nitrogen-fixing bacteria (e.g., Lawn and Brun 1974). For example, it has been reported that the amount of nitrogen fixation by nitrogen-fixing bacteria decreased as photosynthetic rate decreased due to shading and defoliation treatments because plants could not supply carbohydrates necessary for nitrogen-fixing bacteria (Denison et al. 1992; Grove and Malajczuk 1987; Ruess et al. 2006; Vance et al. 1979). Williams et al. (1982) also showed that high photosynthetic rate per plant increased the rate of atmospheric nitrogen fixation per plant by increasing root nodule mass. These previous studies showed that the photosynthetic rate per plant, root nodule mass and nitrogen-fixing rate are closely related (Bormann and Gordon 1984; Gordon and Wheeler 1978). Therefore, it is necessary for plants to pay maintenance costs in order to maintain many nitrogen-fixing bacteria in root nodules (Chapin et al. 2011; Minchin and Witty 2005; Schulze et al. 1999), but keeping root nodules above the necessary level prevents the growth of plants. In the present study, the net production of A. fruticosa isometrically increased with leaf mass, root nodule mass and leaf nitrogen content per plant. Nitrogen is necessary to increase leaves, and root nodules are necessary to obtain nitrogen. Since the scaling between leaf mass and root nodule mass was isometric, it is suggested that they are mutually dependent and that the increase in both is important for increasing the net production of A. fruticosa.

This study showed that leaf mass, root nodule mass and total nitrogen content per plant were isometrically scaled with each other. This study is the first one to demonstrate the quantitative relationships of whole-plant nitrogen, root nodule mass, net production per plant from an allometric point of view. Shade-intolerant alders dominate at early stages of primary succession, but are replaced by shade-tolerant late-successional species with the progress of succession, accompanied with the increase in soil nutrients and the decrease in light intensity on the forest floor (Bormann and Sidle 1990; Chapin et al. 1994; Jones and del Moral 2005). The symbiotic relationship with nitrogen-fixing bacteria is difficult for plants in dark conditions because photosynthetic production is not enough to support the nitrogen-fixing bacteria (Chapin et al. 2011). The change of the limiting resource from nitrogen to light causes the successional change of dominant plant species. Therefore, the examination of the carbon allocation, net production and nitrogen acquisition of alders along successional gradients would provide useful information to understand the role of alder invasion at early-successional stages in the ecosystem function from the viewpoint of carbon and nitrogen economy.