Effects of micro-topographies on stand structure and tree species diversity in an old-growth evergreen broad-leaved forest, southwestern Japan

Stand structure and species diversity were studied in correspondence with micro-topographies in an old-growth forest in southwestern Japan. The study was conducted in a 200 × 200 m 2 permanent plot, which were divided into 400 subplots using grids of 10 m × 10 m. Subplots were categorized to four micro-topographies as crest slope (CS), head hollow (HH), upper slope (US) and lower slope (LS), basing on slope of forest floor and plot position, and to two elevational zones as below 450 m and above 450 m. Tree censuses for all individuals with diameter at breast height (DBH) (cid:62) 5 cm were conducted in 2009 and 2013. The results indicated that CS had subplot means of living stems, dead stems, DBH, basal area (G), and basal area increment ( △ G) significantly higher than that in LS. While, means of recruited stems and Shannon diversity index were significantly lower. Comparing between below and above 450 m elevational zones indicated the significantly higherparametersofstandstructureandspeciesdiversityinabove450melevationalzone. The differences of edaphic conditions led to difference of density of living stems, species density, DBH, G, and △ G among micro-topographies. Therefore, crest slope, upper slope, and higher elevational zones should be encouraged for the purposes of carbon accumulation and storage. While, the lower elevational zones should be used for the purposes of species diversity conservation. © 2015 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND


Introduction
It is becoming increasingly apparent that plant communities change compositionally and structurally in response to biotic and abiotic factors.While, micro-topographies affect on edaphic conditions (e.g.fertility, depth, erosion, instability), light anisotropy/availability, and biological attributes (e.g.biota, microorganism activities), leading to difference of tree's growth, crown asymmetry, and forest stand structure and diversity (Matelson et al., 1995;Clémentine et al., 1998;Clark et al., 1998;Gale, 2000;Robert, 2003;Tateno and Takeda, 2003).Therefore, micro-topography is a synthetic factor in governing stand structure and species diversity.
The regime of land surface disturbance in relation to micro-topographies has been noted as an important factor affecting stand structure and species diversity (Kikuchi and Miura, 1993;Sakai and Ohsawa, 1994;Nagamatsu et al., 2003).A relationship between vegetation characteristics and geomorphic processes was found, particularly at sites with frequent catastrophic disturbances such as land slides (Miles and Swanson, 1986;Van der Sman et al., 1993;Bellingham et al., 1996;Nakamura et al., 1997).Topographies have potential effects on species distribution, diversity, and growth through soil water availability and soil fertility (Becker et al., 1988;Silver et al., 1994;Homeier et al., 2010).Steep topography sites with high rainfall and strong typhoons mean that geomorphic processes are on occasion the critical factors determining vegetation patterns (Sakai and Ohsawa, 1993;Bellingham et al., 1996).Species richness, basal area increment, and stem density were higher in valleys than ridges in southern Ecuadorian Andes forests (Homeier et al., 2010).Enoki et al. (1996) concluded that tree species diversity was higher in lower slope than upper slope in secondary temperate forest, Japan.While, basal area was lower in lower slope than upper slope (Enoki et al., 1996).
Study on stand structure and species diversity may support us a better management strategy for each micro-topography within a climatic region.Therefore, the aims of this study were to examine changes of stand structure and species diversity and to generate the correlation of stand parameters among four micro-topographies basing on 4-year interval tree censuses of a 4-ha permanent plot in an old-growth evergreen broad-leaved forest, southwestern Japan.

Study site
The study was conducted in an old-growth evergreen broad-leaved forest, Aya town, southwestern Japan.Average annual rainfall and average annual temperature of study site were 2509 mm and 17.4 °C from 1981 to 2010 recorded data, respectively.The bedrock is Mesozoic shale and sandstone (Kumamoto Regional Forestry Office, 1963).A moderately moist Brown Forest soil type predominates.Climate including temperature and precipitation was known as ordinary conditions (Sato et al., 2010) and there were no strong typhoons hit the study site between 2009 and 2013, when tree censuses were conducted.
A square permanent 200×200 m 2 plot (Fig. 1) was established in 1989 for long-term ecological research.The plot locates on a steeply inclined north-northwestern slope (32°03 ′ N, 131°12 ′ E), expanding from 380 to 520 m above sea level.There are two gullies and a shallow valley without surface water in the plot.More details of the permanent 200 × 200 m 2 plot can be found at Sato et al. (1999).

Field surveys and data collection
Most studies of vegetation-geomorphology relationship have been based on topographic elements, such as ridges, slopes, and valleys (Harrison et al., 1989;Basnet, 1992;Ashton et al., 1995;Nagamatsu et al., 2003).Each micro-topography has a different geomorphic process, which refers to the interactions between erosion and deposition of soil materials by overland flow and mass movements.Therefore, slope degree and position are the main parameters used to classify microtopographies (Harrison et al., 1989;Basnet, 1992;Ashton et al., 1995;Yamakura et al., 1995;Nagamatsu et al., 2003).The range of slope degree in each micro-topography is different site by site based on the difference of soil type, rainfall, and soil erosion level.In this study, micro-topography classification was inherited from Ohnuki et al. (1998), who classified the present study plot to four micro-topographies and found the significant difference of amount of soil erosion among microtopographies.Crest slope (CS) has a gentle slope of approximate 0°-25°, locating in uppermost area of slopes.Upper slope (US) locates adjacent to CS and has slope of less than 32°.Head hollow (HH) including both head hollow and head wall is a concave slope of less than 42°.Lowe slope (LS) including both lower slope and flat foot slope has slope of less than 49° ( Ohnuki et al., 1998).CS had lowest soil erosion, which equaled to 20% of that in HH.LS equaled 40% soil erosion in HH.US equaled 48% soil erosion in HH.The highest soil erosion belonged to HH.
A 200 × 200 m 2 permanent plot was divided into 400 subplots by using 10 m × 10 m grids (Fig. 1).Each subplot was classified to one of four micro-topographies.If a subplot locates in more than one micro-topography, it was classified to a micro-topography which has highest area ratio.For analyzing the effects of elevation on stand structure and species diversity, subplots were divided into two groups of above 450 m elevational zone and below 450 m elevational zone.The 4-ha plot expands from 380 to 520 m above sea level with a range of 140 m elevation.Contour 450 m was selected as it is the middle of plot, which divides plot to two equal elevation ranges of 70 m and more or less similar number of 100 m 2 plots for two elevational zones.In Septembers 2009 and 2013, tree censuses were conducted for all stems with diameter at breast height (DBH) ≥ 5 cm.All stems were identified to species, mapped, tagged, and measured for DBH.In addition, recruited and dead stems were recorded in 2013 census.

Data analysis
Basal area increment ( G) was estimated as where, G living stem i in 2009 and G living stem i in 2013 are basal areas of stem i in 2009 and 2013, respectively.n is number of living stems, which had DBH in 2009 ≥ 5 cm and were still alive in 2013.G stem j at DBH=5 cm is basal area of stem j, when it had DBH equal 5 cm or basal area of 0.00196 m 2 .G recruited stem j in 2013 is basal area of recruited stem j in 2013.m is number of recruited stems, which had DBH in 2009 < 5 cm but ≥5 cm in 2013.
Shannon diversity index (Magurran, 1988, H) was estimated as where, p i is stem proportion of species ith in plot.n is species number in plot.
Jaccard's similarity coefficient or species similarity index (SI) between pairs of micro-topographies and that of elevational zones was calculated as SI = no. of species shared by two micro-topographies or elevational zones total species found in two micro-topographies or elevational zones .
(3) Species/area curve is applied to generate relationship between survey area (A) and number of species (S) in the form of Eq. ( 4).
where, a and b are constant and estimated through linear logarithmic transform as log(S) = log(a) + b * log(A).Randomly picked up subplots (10 × 10 m 2 plot) for species aggregation was used.The process was repeated 30 times.Then mean, upper limit, and lower limit (± SE) of species/area curve was regenerated.
Stand parameters including number of living stems, number of dead stems, number of recruited stems, mean DBH, basal area (G), basal area increment ( G during 2009-2013), species density (number of species/subplot), and Shannon diversity index were calculated for each subplot (10 × 10 m 2 plot) separately.G and number of dead stems were calculated basing on data recorded in 2009 and 2013, while other stand parameters were calculated basing on data recorded in 2013.
The effects of four micro-topographies on stand parameters were assessed by univariate analysis of variance (ANOVA) and post-hoc Tukey's test.Meanwhile, pair-comparison using t-test (p = 0.05) was applied to find out the effect of two elevational zones on stand parameters for combination of all micro-topographies and for each micro-topography separately.
Correlations between pair of stand parameters were generated.It usually results in a correlation matrix showing correlation values of all pairs in the form of table, including significant correlations (R > 0.7) and non-significant correlations.For simplification, only significant correlations were shown in the form of correlation circle.Mathematically, significant correlation indicates the change of a parameter leads to change of other positively or negatively.

Vegetation in a 4-ha plot
A total of 5268 stems (DBH ≥ 5 cm) was recorded in a 4-ha plot in 2013 census.These individuals belong to 55 species of 28 families.There were 13 singleton species, six doubleton species, and four tripleton species (Appendix).Ninety four stems with DBH larger than 70 cm, in which 15 stems had >100 cm DBH, were recorded.Plot had a total basal area (G) of 207.5 m 2 .
Species/area curves indicated that at the same area equal to 23 subplots, US had highest species richness, following by HH, LS, and CS (Fig. 3(a), (b), (c), (d)).LS and HH had highest species similarity index (Table 2), following by pair of CS and LS (0.52), US and LS (0.49).The lowest species similarity belonged to pair of CS and US (0.42).J shape best described the stem proportion/DBH relationship in all micro-topographies (Fig. 4(a), (b), (c), (d)).However, the diameter diversity index (DI) indicated that CS was highest (1.79), following US (1.75), HH (1.67), and LS (1.56).The largest DBH stem found in CS was 101.8 cm, it was 93.1 cm in HH, 110.6 cm in LS, and 133.6 cm in US.

Table 1
General stand structure and species diversity of four micro-topographies and two elevational zones in a 4-ha plot.Indicate significant difference of means for combination.
A Self-archived copy in

Elevational zones
A little smaller area in a 200 × 200 m 2 plot was allocated to below 450 m (193 subplots) elevational zone (Table 1), while 207 subplots were allocated to above 450 m elevational zone.However, species richness was much different.There were 52 species found in below 450 m elevational zone with ten singleton species.While only 39 species were found in above 450 m elevational zone with three singleton species (Appendix).Above 450 m elevational zone had significantly higher means of number of living stems and recruited stems, species density, G, G, and Shannon diversity index than that in below 450 m elevational zone (Table 1).Conversely, above 450 m elevational zone had smaller DBH than that in below 450 m elevational zone.While, the difference of number of dead stems between two elevational zones was not significant (Table 1).Species/area curves indicated that at equal area, below 450 m elevational zone had higher species richness than that in above 450 m elevational zone (Fig. 3(e), (f)).Both elevational zones shared 36 species in total of 55 species found, leading to species similarity index of 0.65.J shape was well describing the stem proportion/DBH relationship in both elevational zones (Fig. 4(e), (f)).However, diameter diversity index was higher (1.85) in below 450 m elevational zones compared to that (1.67) in above 450 m elevational zone.2), where p i is stem proportion of diameter class i. 5 cm DBH intervals were used.

Combination of micro-topographies and elevational zones
ANOVA analysis indicated that combination of micro-topographies and elevational zones had no significant effect on stand parameters (p > 0.07).Meanwhile, pair-comparison between above 450 m elevational zone and below 450 m elevational zone in each micro-topography indicated significant difference of stand parameters (Table 1).In LS, number of living stems and recruited stems, species density, and Shannon diversity index were significantly higher in above 450 m elevational zone.In HH, number of living stems, G, species density, and Shannon diversity index were significantly higher in above 450 m elevational zone.In CS, number of living stems and recruited stems, and G were significantly higher in above 450 m elevational zone.While, in US the significant difference between two elevational zones was only found for number of living stems and species density.In all micro-topographies, significant difference between two elevational zones was not found for number of dead stems, mean DBH, and basal area (Table 1).

Correlation among stand parameters
Among eight concerned parameters of stand structure and species diversity, number of living stems had positive correlations with two parameters (number of dead stems and species density) and negative correlations with three others (number of recruited stems, DBH, and G).While it had no correlation with basal area and Shannon diversity index (Fig. 5).Number of dead stems, species density, DBH, and G had six significant correlations with others.Number of living stems and number of recruited stems had five correlations with others.Basal area had four correlations with others.Meanwhile, Shannon diversity index had only two significant correlations with basal area and species density (Fig. 5).

Discussion
Twenty canopy species accounted for 46.2% stems and 81.4% basal area in a 4-ha plot (Appendix).Meanwhile, ten of 13 singletons are under-canopy species and accounted for <0.2% basal area.Of 13 singleton species, five had DBH < 10 cm, three had DBH of 10-20 cm, and five had DBH of 20-60 cm.These indicated less importance of singleton species in stand structure.Small size-stems of singleton species are vulnerable by typhoons (Sato et al., 1999), which usually happen in the present study site and cause the large and big trees falling, leading to its disappearance.The falling of big trees may cause singleton species dead.A comparison is that there were 60 species (DBH ≥ 5 cm) recorded in 1991 tree census (Sato et al., 1999), while only 55 were found in 2013 tree census (Appendix).Five species (DBH ≥ 5 cm) disappeared in a duration of 22 years.
The differences of parameters of stand structure and species diversity in the present study (Table 1) resulted from the differences of edaphic conditions, soil erosion degree, and light availability in each micro-topography (Ohnuki et al., 1995;Tanouchi and Yamamoto, 1995;Enoki et al., 1996;Clark et al., 1998;Robert, 2003;Tateno and Takeda, 2003).Ohnuki et al. (1998) indicated in the present 4-ha plot there was lowest soil erosion in CS and highest in HH.In addition, US and HH have porosity in higher suction range, leading to less water retention.The highest density of living stems, dead stems, larger DBH, G, and G (Table 1) in CS in the present study site may result from low soil erosion (Ohnuki et al., 1998) and sunlight availability (Bailey, 1996;Tanouchi and Yamamoto, 1995).The same patterns of highest living stem density and G in CS were observed in an evergreen broad-leaved forest on Amami Ohshima Island, southwestern Japan (Hara et al., 1996), which resulted from the differences of soil disturbance regime.The numerous density of living stems in CS (15 stems 100 m −2 ) may lead to high competition among stems, which resulted in highest number of dead stems in this micro-topography (Table 1).Opposite pattern was found in LS, where there were lowest number of living stems and dead stems (Table 1).Lowest living stem density of canopy species was also reported in LS in the present study plot (Tanouchi and Yamamoto, 1995).Recruited stems were highest in LS among four micro-topographies.Lower number of living stems (Table 1) with smaller size-stems (Fig. 4(c)) in LS led to less competition among individuals.In addition, soil in LS had high water retention (Ohnuki et al., 1998).These favored seed germination and stem recruitment in LS compared to other three micro-topographies (Table 1).Again, less sunlight availability (Tanouchi and Yamamoto, 1995) and poor and/or unstable edaphic condition (Ohnuki et al., 1998) in LS responded for smaller DBH, and lower G and G compared to that in CS, HH, and US.
Mathematically, Fig. 5 shows increase number of living stems leads to increasing number of dead stems but decreasing number of recruited stems; increase mean DBH leads to increasing G and G, but reducing number of recruited stems; increase species density leads to increasing Shannon diversity index, but reducing DBH and G.However, ecologically changes of stand structure and species diversity are mainly governed by climates, edaphic conditions, and fauna itself (Gao et al., 2014;Suratman, 2012).The differences of natural conditions among micro-topographies lead to differences of stand structure and species diversity, and to significant correlations among them.Correlation is not just for only two parameters, it is a matrix of correlations in which parameters have cross-correlations with many others.In the present study, the significant correlations between stand parameters resulted from micro-topographies, which have different soil conditions as soil erosion and water retention (Tanouchi and Yamamoto, 1995;Ohnuki et al., 1998).Therefore, it is obvious that flat land areas, which have homogeneous soil and natural conditions, are preferably selected for timber plantation to ensure fair growth of all stems.While, for purpose of diversity conservation areas with diversity of micro-topographies are usually selected, since those are suitable for co-existence of many species and stem sizes.
In a quite narrow elevation range (380-520 m) in the present 200 × 200 m 2 plot, the difference of stand structure and species diversity between two elevational zones of 380-450 m and 450-520 m was significant (Table 1).Regardless of microtopography, below 450 m elevational zone had much larger DBH but lower number of living stems and recruited stems, basal area (G), G, species density, and Shannon diversity index than that in above 450 m elevational zone.If considering each micro-topography separately, number of living stems, number of recruited stems, species density, and Shannon diversity index in above 450 m elevational zones were also significantly higher than that in below 450 m elevational zone (Table 1).This may be explained by the differences of light availability (Tanouchi and Yamamoto, 1995) and edaphic conditions as soil  Data in 1995 and 1991 were cited from Sato et al. (1999) and only dominant species were listed.While total stems and basal area were of 4-ha plot.Singleton species are species with only one stem in a 4-ha plot.c indicates canopy species (Tanouchi and Yamamoto, 1995).e indicates singleton species.
A Self-archived copy in Kyoto University Research Information Repository https://repository.kulib.kyoto-u.ac.jp erosion, fertility, and water retention (Ohnuki et al., 1998) between two elevational zones.Higher species density (species number 100 m −2 ) in above 450 m elevational zone (Table 1) indicates higher species holding capacity (Rosenzweig, 1995;Mitchell and Ryan, 1998) of this zone compared to below 450 m elevational zone, resulting from higher availability of energy and more stability of soil surface in upper elevational zone (Ohnuki et al., 1998).While, higher species richness as a total of 52 species appeared in below 450 m elevational zone resulted from more diversity of micro-topographies in lower zone.Since there are more number of 100 m 2 plots of HH and CS in below 450 m elevational zone (Table 1), compared to that in above 450 m elevational zone.However, studies on vegetation zonation in the present study site should be conducted for further understanding.

Conclusion
Results from the present study indicated that crest slope, upper slope, and high elevational zone are most suitable for biomass accumulation and carbon storage in forests.While, for the purposes of species conservation lower elevational zones are encouraged since they have much diverse micro-conditions favoring the growth and coexistent of many species.However, the successful forest management for low elevational zones is much challenging as higher vulnerability on species establishment compared to that in crest slope, upper slope, and higher elevational zones.

Fig. 2 .
Fig. 2. Number of dead stems in different diameter classes (5 cm DBH intervals were used).The largest DBH dead stem was 111.7 cm, belonging to species Persea thunbergii.

Fig. 5 .
Fig. 5. Correlation circle of parameters of stand structure and species diversity.G is basal area increment.Negative value of R indicates negative correlation between two stand parameters.Only correlation (R) > 0.7 is shown.

Table 2
Species similarity index among micro-topographies in a 4-ha plot.

Table A .1
List of species, their stems, and basal area in four micro-topographies and two elevational zones in a 4-ha plot.