1 Introduction

Natural disturbance s and life-history characteristics are key factors influencing the coexistence of tree species (White 1979; Loehle 2000). Disturbances in riparian areas are dynamic and vary more widely in type, frequency, and magnitude compared with those on hillsides. Various disturbance regimes in riparian areas lead to heterogeneous topography (Gregory et al. 1991; Kovalchik and Chitwood 1990) due to repeated destruction and regeneration of riparian forests.

Fraxinus platypoda, Pterocarya rhoifolia, and Cercidiphyllum japonicum (Fig. 7.1) are the dominant tree species in the riparian forests of the Chichibu Mountains in central Japan (Maeda and Yoshioka 1952; Sakio 1997). One such riparian forest, along the Ooyamazawa stream, is an old-growth forest whose high-elevation trees have not been affected by human impact s, e.g., logging or erosion control work s (Chap. 1), and are therefore valuable. The dominant canopy tree species in this forest is F. platypoda, followed by P. rhoifolia and C. japonicum.

Fig. 7.1
figure 1

The three dominant canopy tree species in the Ooyamazawa riparian forest of the Chichibu Mountains in central Japan

The life-history characteristics and regeneration process es of these three canopy tree species were explained in detail in Chaps. 23 and 4. In this chapter, we clarify the coexistence mechanism of riparian tree species in terms of disturbance regime s, tree life-history strategies, and responses to environment factors.

2 Seed Production

All three dominant canopy species in the Ooyamazawa riparian forest produce wind-dispersed winged achenes (Fig. 7.2). The average dry weights (mean ± standard deviation [SD], n = 20) of the fruit of F. platypoda, P. rhoifolia, and C. japonicum are 144 ± 24, 90 ± 11, and 0.82 ± 0.15 mg, respectively. The dry weights (mean ± SD, n = 20) of their seeds are 80 ± 17, 70 ± 8, and 0.58 ± 0.14 mg, respectively (Sakio et al. 2002).

Fig. 7.2
figure 2

Seeds of the three riparian species

The seed s of these three species are released according to different schedules, from autumn to winter (Fig. 7.3).The seeds of P. rhoifolia are released in October, whereas those of F. platypoda are released in November, during leaf fall. C. japonicum seeds are released after leaf fall, from November until spring of the following year. The dispersal distances of P. rhoifolia and F. platypoda are similar, extending to several tens of meters. In contrast, the significantly lighter seeds of C. japonicum (Welch’s t-test, P < 0.001) are dispersed over hundreds of meters, with the maximum dispersal distance recorded being 302 m (Sato et al. 2006).

Fig. 7.3
figure 3

Seed dispersal of the three riparian species and leaf fall in 2002. Leaf fall includes all tree species

Secondary dispersal of F. platypoda seeds can also occur via water, implying an adaptation for waterborne fruit dispersal. We have observed many F. platypoda seeds being transported by the flow of mountain streams during mast seed year s (Chap. 2); this process allows the establishment of many seedling s on the gravel banks of mountain streams in the following year (Fig. 7.4).

Fig. 7.4
figure 4

Current-year F. platypoda seedlings germinated at the stream edge. Some seedlings are submerged

Annual fluctuation in seed production varies among tree species. Generally, late successional species have large seeds and irregular fruiting behavior , whereas pioneer species have small seeds and regular fruiting. F. platypoda had non-mast years in 1997, 2001, 2003, 2005, and 2015, and mast years in 1996, 1998, 2002, 2004, 2006, and 2016 (Fig. 7.5). Thus, this species exhibits irregular fruiting behavior, as has also been observed in Fraxinus excelsior, which is native throughout mainland Europe (Tapper 1992, 1996). Seed production also fluctuates annually in P. rhoifolia, which had non-mast years in 1996, 2006, and 2008, and mast years in 2005, 2007, and 2011. In contrast, C. japonicum produces regular amounts of seeds every year, with only slight fluctuation. Among these three species, F. platypoda had the largest coefficient of variation (CV) of seed production (0.97), followed by P. rhoifolia (0.87) and C. japonicum (0.48).

Fig. 7.5
figure 5

Annual fluctuation of seed production in the three riparian species from 1995 to 2016

3 Seedling Germination and Growth

3.1 Germination Sites

Seeds of the three species germinate from mid-May to early summer of the year following seed production . Buried seeds in Ooyamazawa riparian forest soil germinated no F. platypoda seeds, one P. rhoifolia seed, and 12 C. japonicum seeds in 30 L of soil (Kubo et al. 2008). Seeds of P. rhoifolia and C. japonicum have exhibited dormancy in nursery seedling tests; in particular, P. rhoifolia seeds have successfully germinated 2 years after sowing.

Germination site s differ among the three species (Fig. 7.6). Figure 7.6 shows the results of a survey conducted in a year following a poor year for F. platypoda seed production (all seedlings aged >1 year). But, in the following year, a mast year, F. platypoda seeds germinated in all environments. F. platypoda can germinate in forest floor environments varying in light, substrate, and water conditions, including at water edges, under herb cover, and on steep slopes and gravel. Most F. platypoda seedling s that germinate in the litter layer die due to fungal damage and lack of moisture; however, some individuals survive for several years (Chap. 2). Current-year seedling s have a very high survival rate on the gravel banks of mountain streams, where herbaceous vegetation is rare. F. platypoda seedlings are highly shade-tolerant and tend to concentrate in former stream channels and small gravel deposit s (Sakio 1997). When a canopy gap forms, thus improving the light environment, F. platypoda seedlings begin to grow into canopy trees.

Fig. 7.6
figure 6

Seed germination sites of the three riparian species (Kubo et al. 2000)

P. rhoifolia can also germinate in any environment, except on steep slopes due to unstable substrates. P. rhoifolia germination has been observed under closed canopies, and in gravel deposits and the litter layer; however, seedling s that germinate in the litter layer die within 1–2 months of germination (Sakio et al. 2002). Surviving individuals develop true leaves, but their growth is greatly affected by light availability. P. rhoifolia seedlings are affected by herbal pressure, even in canopy gap s; they require a brighter environment for survival than F. platypoda and may show arrested tree growth in low-light environments (Kisanuki et al. 1995).

In contrast, C. japonicum seeds germinate in a limited range of environments. Most C. japonicum seedlings have been found in fine mineral soil and on fallen logs. Seedling emergence of small-seeded species is generally reduced in litter deposits (Seiwa and Kikuzawa 1996); the germination and establishment of small-seeded species, such as birch (Betula platyphylla var. japonica), is promoted in fine-grained soil (Koyama 1998). These factors also appear to affect C. japonicum seed germination. Germination begins in mid-May; however, seedlings can be washed away by surface sediment movement during the rainy and typhoon seasons, with a survival rate of <10% in late October. This rate is higher at sites with high illuminance, and seedlings found in such environments have been reported to be significantly larger than those in low-light environments (Kubo et al. 2000). Similar results were obtained in a germination experiment examining the effects of soil and relative photosynthetic effective photon flux density on germination in a nursery (Kubo et al. 2004).

3.2 Seedling Shade Tolerance

The light environment is among the most important factors for plant growth. Plant responses to light differ among plant species. Pioneer species require more light than late successional species . In forests, canopy gap formation strongly affects light conditions. Improvement of the light environment by gap formation is an important factor for growth from seedling to canopy tree (Suzuki 1980, 1981; Nakashizuka and Numata 1982a, b; Nakashizuka 1983, 1984). For seedlings that have established under the canopy, a lack of canopy gap formation within a few years will result in decreased growth and, eventually, death. If an individual establishes under a small canopy gap, and that gap closes due to branch extension within the surrounding canopy, the individual will be unable to survive.

We compared the effects of light condition s on the growth of seedling s of the three Ooyamazawa riparian species in a nursery. P. rhoifolia had the fastest growth rate, growing to 30-cm seedlings within 1 year, whereas F. platypoda and C. japonicum seedlings measured about 10 cm (Fig. 7.7). However, since C. japonicum seeds are much smaller than those of F. platypoda (Fig. 7.2), its relative growth rate was very high. P. rhoifolia seedling growth decreased sharply under nursery light conditions (<20% of outdoor sunlight), and failed to survive in 1% light. Similarly, C. japonicum seedlings did not survive at 1% light. In contrast, all F. platypoda seedlings survived for 1 year at 1% light (Fig. 7.7; Sakio 2008). A study conducted at the Ooyamazawa riparian forest research site showed that branch growth was dramatically faster in 1-m-tall P. rhoifolia sapling s than in F. platypoda saplings of the same height beneath a canopy gap (Sakio 1993). At a seedling size of <1 m, F. platypoda grew in lower light environments than P. rhoifolia and C. japonicum at the same research site (Kubo et al. 2000).

Fig. 7.7
figure 7

Relationship between growth and relative light intensity for current-year seedling s grown in nursery beds for 1 year (Sakio 2008)

3.3 Seedling Water Tolerance

Trees in riparian areas are always exposed to a moist environment at the water’s edge, and seedling s that have germinated on the gravel banks of mountain streams are affected by flooding during the rainy and typhoon seasons (Fig. 7.4). Water tolerance thresholds differ greatly among the three riparian species. Submergence experiments, in which 1-year-old seedling s were submerged to the soil surface for 1 year (Sakio 2005) and current-year seedling s were submerged for different periods (Fig. 7.8; Sakio 2008), showed that F. platypoda has much higher water tolerance than do P. rhoifolia and C. japonicum. Dry weight was significantly lower in submerged 1-year-old P. rhoifolia and C. japonicum seedlings than in control individuals; moreover, 80% of F. platypoda current-year seedlings survived 20 days of submergence, whereas only 20% of P. rhoifolia seedlings survived, and all C. japonicum seedlings died. The superior water tolerance of F. platypoda may be one reason explaining its dominance in Ooyamazawa riparian forests.

Fig. 7.8
figure 8

Flooding survival rates of current-year seedlings of the riparian species

4 Sprouting

All three riparian tree species exhibit reproduction by sprouting , with most F. platypoda producing single trees (Fig. 7.9; Sakio et al. 2002). P. rhoifolia produces intermediate numbers of sprout s; C. japonicum produces the most sprouts, with a maximum of 60 observed from one tree. In C. japonicum, sprout number was positively correlated with the diameter at breast height (DBH ) of the main stem. The role of the C. japonicum sprouts is maintenance of the individual; this has also been observed in Euptelea polyandra, which maintains individuals by sprouting in areas frequently disturbed by landslide s (Sakai et al. 1995). However, it remains unknown whether the sprouting mechanism of C. japonicum is related to physical damage to the individual, physiological responses to changes in the light environment , or tree age. In P. rhoifolia, sprouting does not play a role in maintenance of the individual in Ooyamazawa riparian forests, although such maintenance by sprouting has been observed in P. rhoifolia growing in environments characterized by heavy snow (Nakano and Sakio 2017, 2018).

Fig. 7.9
figure 9

Relationship between the diameter at breast height (DBH) of the main stem and sprout number for each individual of the three riparian species (Sakio 2008)

5 Forest Structure

5.1 Size Structure

A survey of all living trees (DBH  ≥ 4 cm) without sprout s in an area of 4.71 ha showed that the frequency distribution of DBH was similar between F. platypoda and P. rhoifolia, being characterized by many small individuals with DBH ≤ 10 cm or less (Fig. 7.10). C. japonicum exhibited a different DBH distribution , with only two sapling s at DBH < 10 cm. Both F. platypoda and P. rhoifolia showed frequency distribution peaks around DBH = 40. Within a core plot (0.54 ha), there were 811 F. platypoda, 192 P. rhoifolia, and one C. japonicum individual with tree height ≥ 1 m and DBH < 4 cm. Many F. platypoda and P. rhoifolia seedling s with DBH < 4 cm were distributed beneath the canopy (Chaps. 2 and 3). These results demonstrate that F. platypoda and P. rhoifolia produce many advanced seedling s that will eventually become canopy tree s.

Fig. 7.10
figure 10

DBH class distributions of the three riparian species (Sakio 2008)

5.2 Spatial Distribution and Age Structure

Figure 7.11 shows the spatial distribution of canopy trees (DBH ≥ 20 cm) and young tree s (4 cm ≤ DBH < 20 cm) within the 4.71-ha research plot. Canopy and young trees of F. platypoda are dominant from upstream to downstream , whereas P. rhoifolia canopy trees are distributed in three patches (A, B, and C). Large patches of canopy tree s can reach 50 m in diameter, each with an average tree age of about 90 years. These P. rhoifolia patches were topographically distributed on large landslide or debris flow paths, suggesting that P. rhoifolia established in these large-scale disturbance sites at the same time (Sakio et al. 2002).

Fig. 7.11
figure 11

Spatial distribution of young and canopy trees, including at the sites of two large-scale disturbances. P. rhoifolia canopy trees and young C. japonicum trees coexist at the same sites

C. japonicum canopy trees are few in number and scattered randomly throughout the research area, although young trees showed aggregated distributions in and around P. rhoifolia patches (Fig. 7.11, D and E). At approximately 90 years, the ages of young C. japonicum and P. rhoifolia trees were similar; therefore, these two species are likely to have established at the same time.

6 Coexistence Mechanisms of the Three Species

The following coexistence mechanism s of canopy species in the Ooyamazawa riparian forest can be considered (Fig. 7.12).

Fig. 7.12
figure 12

Mechanism for the coexistence of three canopy tree species in the Ooyamazawa riparian area

In the Ooyamazawa riparian forest, F. platypoda regenerated to produce canopy tree s through seedling establishment on large-scale disturbance sites, such as landslide and debris flow paths, as well as through the release of advanced forest floor seedlings upon the formation of small gap s. F. platypoda has higher shade tolerance than P. rhoifolia and C. japonicum, as demonstrated by seedling field studies and nursery experiments. F. platypoda also shows higher tolerance to submergence and flooding than P. rhoifolia and C. japonicum, and has adapted to germinate and grow on gravel stream banks.

The 4.71-ha research plot contained several large patches of P. rhoifolia with DBH of about 50 cm; these were established on the sites of large-scale disturbance s that occurred about 90 years ago. Thus, P. rhoifolia regeneration sites are limited to large-scale disturbance sites where large gap s are formed by landslide or debris flow; the regenerants eventually became canopy trees, forming single-species patches.

Unlike F. platypoda, C. japonicum cannot form large cohorts of advanced seedling s on the forest floor beneath canopy trees. However, C. japonicum seeds can disperse into the large-scale disturbance sites favored by P. rhoifolia. The mass production of small wind-dispersed seeds increases the probability of reaching new germination sites cleared by large disturbances (Harper 1977; Augspurger 1984). Thus, C. japonicum regeneration in large-scale disturbance sites can co-occur with that of P. rhoifolia. In these disturbance sites, organic and inorganic matter such as fallen trees, boulders, soil, and sand are mixed together to form complex substrate s. The diversity of microsites produced during this process ensures seed germination , seedling establishment , and growth of C. japonicum. Fine inorganic soil s, which form in boulders and the cracks of fallen trees, are ideal substrates for C. japonicum germination and establishment. These substrates are less susceptible to erosion due to rainfall, such that seedlings can grow stably; if direct sunlight is weak, seedlings may continue to grow because strong light dries out these soils. Thus, C. japonicum waits for rare regeneration opportunities, which appear at the scale of decades or centuries; as a trade-off , the lifespan of the individual is prolonged. C. japonicum generates many sprout s around the main trunk. In F. platypoda and P. rhoifolia, the death of the trunk means the inevitable death of the individual; however, if the main trunk of C. japonicum dies, one of the many surrounding sprouts will grow to become the main trunk (Fig. 7.13). In this manner, once C. japonicum has established, it will survive for long periods by successively producing new trunks.

Fig. 7.13
figure 13

Relationship between tree age and sprouting. In F. platypoda and P. rhoifolia, the death of the main trunk results in the death of the individual. In contrast, C. japonicum individuals are maintained by sprouting for long periods after the death of the main trunk

Therefore, C. japonicum trees represent a small population within the Ooyamazawa riparian forest, but this species sustainably coexists with both the shade-tolerant dominant species F. platypoda and the pioneer species P. rhoifolia.

7 Conclusion

The coexistence mechanism s of the three Ooyamazawa riparian canopy species involve a combination of niche partitioning and chance. The three species exhibit trade-offs in reproductive characteristic s, e.g., seed size, quantity, and annual variation. Coexistence is generally maintained through niche partitioning, especially in the early life-history stages. Chance can also play an important role in P. rhoifolia and C. japonicum regeneration, through unpredictable large-scale, low-frequency disturbance . In conclusion, F. platypoda, P. rhoifolia, and C. japonicum are well-adapted to disturbances in the Ooyamazawa riparian zone throughout their life histories.