Patterns and determinants of plant richness by elevation in a mountain ecosystem in South Korea:area,mid-domain effect, climate and productivity

•

ORIGINAL PAPER

Patterns and determinants of plant richness by elevation in a mountain ecosystem in South Korea:area,mid-domain effect, climate and productivity

Chang-Bae Lee1•Jung-Hwa Chun2

©Northeast Forestry University and Springer-Verlag Berlin Heidelberg 2015

We examined patterns of plant species richness on an elevation gradient and evaluated the effects of climatic variables including mean annual temperature and precipitation,area,the mid-domain effect and productivity on species richness along two transects on Mt.Seorak, South Korea.A total of 235 plant species of 72 families and 161 genera were recorded from 130 plots along the two transects.Two different patterns,monotonic decline and a unimodal shape,were observed for woody plants with the change in elevation along the two transects,whereas multimodal patterns were observed for all plant species considered together and for herbaceous plants.Area and productivity showed signif i cant relationships with total plant richness.Climatic variables were better predictors than other variables for variation by elevation in woody plant richness,whereas productivity was a more important variable for herbaceous plant richness.Although area was an important variable for predicting species richness patterns,the effects differed by transect and plant group.No empirical evidence was linked to the mid-domain effect. Different elevational patterns may characterize different groups in the same taxon and there might be fundamental differences in the mechanisms underlying these richness patterns.

Area⋅Climatic variables⋅Elevational patterns⋅Plant richness⋅Productivity

Introduction

Mountains are generally biologically diverse,support a high proportion of endemic species and harbor most of the world’s protected areas(Khuroo et al.2011).The biodiversity of plants in mountain ecosystems and their geographical distribution have received considerable interest from ecology and biogeography researchers over the last decade because understanding biodiversity patterns is important for biodiversity conservation,sustainable use and natural reserve area planning and management(Grytnes and Vetaas 2002;Sang 2009).The elevational gradient present in mountainous areas provides a useful natural experimental system for clarifying the ecological and evolutionary responses of living organisms to geophysical inf l uences(Ko¨rner 2007;Namgail et al.2012).Elevation is one of the most important physical factors determining biodiversity and species distribution patterns because it inf l uences temperature and precipitation,thus driving the ecological and physiological adaptations of various taxa including plants,mammals,birds and invertebrates(Lomolino 2001).The elevation gradient is often claimed to parallel the latitudinal gradient(Rahbek 1995),and many researchers have recognized that mountains provide the best available test system for understanding the factors that drive the latitudinal effects on biodiversity(Ko¨rner 2000).

The examination of elevational patterns in biodiversity is important for understanding the inf l uence of global change on biodiversity,for determining the broad-scale distribution of species,and for differentiating between alternative diversity hypotheses(Grytnes and McCain 2007;Wang et al.2011).Previous macroecological studies have reported that elevational patterns in species diversity usually occur in either monotonic,unimodal or multimodal patterns(Oommen and Shanker 2005;Sang 2009),of which unimodal patterns are the most common(Rahbek 1995;Colwell and Lees 2000;Grytnes and Vetaas 2002). Although the mechanisms underlying elevational patterns of species diversity remain subject to debate,typical explanations include the inf l uences of factors such as climate,area,geometric constraints or the mid-domain effect, productivity and evolutionary history(McCain 2009).

Climate is the most widely supported predictor of worldwide species diversity(Rowe 2009)because it can directly control species distributions and can indirectly affect photosynthetic activity and biological processes.The area of elevational bands can also explain much of the variation in species richness(McCain 2007),following the species-area relationship in which a larger area provides more diverse habitat that can harbor a larger number of species and an increase in area is accompanied by a decrease in the extinction rate and an increase in speciation or colonization(Rosenzweig 1995).The mid-domain effect (MDE),or geometric constraint,also explains elevational patterns of species richness(McCain 2004;Kluge et al. 2006;McCain 2009).The MDE postulates that geometric constraints on species ranges within a bounded domain will yield a mid-domain peak in richness regardless of ecological factors.The MDE is an abiotic and stochastic hypothesis based on the premise that the spatial distribution of species richness is constrained by the shape of landmasses and by species range size.Under these conditions, random placement of species ranges within a bounded domain creates an overlap of species ranges,and thus a peak of species richness,toward the center of the geographical domain(Colwell and Hurtt 1994;Colwell and Lees 2000).Productivity can also inf l uence species diversity patterns.Although the relationship between species diversity and productivity has been very controversial,with disagreement over whether productivity controls or is controlled by species diversity(Waide et al.1999;Loreau et al.2001),productivity is frequently cited as a fundamental determinant of diversity(Chalcraft et al.2004). Waide et al.(1999)reported four types of relationships between productivity and species diversity:negative,positive,unimodal,and no relationship.

Most of the mechanisms proposed to explain the relationship between diversity and elevation relate to largescale unimodal patterns and do not explain why monotonic and multimodal patterns are also occasionally observed over smaller scales such as local slopes(Oommen and Shanker 2005).The nature of richness patterns can change with spatial grain and scale(Rahbek 2005),and there is clearly a need to explore small-scale patterns(Storch et al. 2005).The dearth of such analyses is partly due to constraints such as the lack of data for smaller areas and reliance on methodologies commonly used in macroecology(especially dependence on secondary distribution data),as well as the multiplicity of proposed mechanisms (Oommen and Shanker 2005).

In this study,we described the elevational patterns of plant species richness and evaluated a suite of variables for their ability to explain patterns of richness along two transects on Mt.Seorak in South Korea.Using f i eld surveys of plant data,we focused on(1)describing the elevational patterns of species richness for all,woody and herbaceous plants;(2)evaluating the effects of mean annual temperature and precipitation,area,the MDE and productivity on the elevational patterns of plant species richness;and(3) determining whether the elevational patterns of plant richness differed among plant groups and,if so,identifying the primary driver of these differences.

Materials and methods

Study area

This study was conducted between May and July 2011 along two transects on Mt.Seorak in South Korea(Fig.1). Mt.Seorak is the third highest mountain in South Korea and is located in the northern part of the Baekdudaegan Mountains,the major mountain chain and biodiversity hotspot of the Korean peninsula.It was designated as a nature reserve by the South Korean government in 1965 and as a biosphere reserve by UNESCO in 1982.It was also the f i rst South Korean National Park to be designated under the National Park Law in 1970.The nature reserve covers an area of 163.6 km2and contains many peaks over 1200 m asl.including the highest peak,Daecheongbong (1708 m asl.).The nature reserve contains mountain ranges of dissected granite and gneiss and is characterized by rocky hills and ridges.Mt.Seorak contains 28 peaks,58 valleys,two hot springs,two mineral springs and numerous stunning rock formations.Mean annual temperature and precipitation are 13.2°C and 1342 mm,respectively(Han et al.2006).

More than 822 vascular plant species and a total of 1013 plant species have been recorded in the region.Pine trees are abundant on the southern slopes,whereas the northern slopes are characterized by oaks and other deciduous trees. The vegetation on Mt.Seorak is divided into four majorzones along an elevational gradient,viz.(1)temperate (montane)deciduous broad-leaved and pine forest (＜500 m asl)dominated byPinus densif l oraandRhododendron mucronulatumvar.mucronulatum,(2)temperate deciduous broad-leaved and coniferous mixed forest (500–1100 m asl.)dominated byQuercus mongolica,Betula schmidtii,Magnolia sieboldii,Pinus koraiensisandAbiesholophylla,(3)sub-alpine coniferous forest (1100–1500 m asl.)dominated byTaxus cuspidata,Thuja koraiensisandAbies nephrolepis,and f i nally(4)alpine forest(＞1500 m asl.)dominated byBetula ermaniiandPinus pumila(Kong 2008).

Fig.1 Location and topography of the study areas,Namgyori and Osaek transects,on Mt.Seorak,South Korea

Plant data

For f i eld surveys,two 100 m-wide transects were established along elevational gradients using the Namgyori and Osaek trails to Daecheongbon Peak,the longest (～22.4 km)and shortest(～5.3 km)routes to access Daecheongbong Peak,respectively.The two transects were divided into fourteen elevational bands of 100 m widths from 300 to 1700 m asl.on the Namgyori transect and from 500 to 1700 m asl.on the Osaek transect.Plant data were recorded within each 100 m elevational band from May to July 2011.In each band,we standardized sampling effort and area by sampling f i ve 400 m2plots.Within each plot, plants were exhaustively surveyed for 1–2 h depending on the species richness of the plot according to the method of Braun-Blanquet(1965).Plant data were obtained from 70 and 60 plots along the Namgyori and Osaek transects, respectively.

Explanatory variables

Two climatic variables,mean annual temperature(MAT) and precipitation(MAP),were investigated with respect to species richness.We used digital climate maps produced by the Korea Meteorological Administration and National Center of Agrometeorology to extract the meteorological parameters for each elevational band(Yun 2010).MAT data were dated from 1971 to 2008 and MAP data were dated from 1981 to 2009.The spatial resolution of the raster data was 30 m for MAT and 270 m for MAP.MAT and MAP were calculated for each elevational band in the 100 m-wide transects along the two transects.

The two spatial variables used in this study were area and MDE.To test species-area relationships,we calculated the area of each elevational band along the two 100 mwide transects.Calculations were performed using a Digital Elevation Model(DEM)with the 3D Analyst extension in ArcGIS.The MDE null model was used to test the inf l uence of geometric constraints on the spatial patterns ofspecies richness along an elevational gradient.We used a novel,discrete MDE model based on Colwell and Hurtt’s (1994)continuous Model 2,which does not necessitate the use of interpolated ranges(Fu et al.2006).Range Model Software Version 5(Colwell 2006)was used for simulation.The simulation process was repeated 5000 times,and we used expected mean richness to assess the effects of geometric constraints on the elevational gradient of species richness for each plant group(i.e.,total,woody and herbaceous species).

As a proxy for above-ground net primary productivity, we used the enhanced vegetation index(EVI),which has been preferred over the Normalized Difference Vegetation Index(NDVI)because EVI is not sensitive to soil or atmospheric effects and adjusts the red wavelength as a function of the blue wavelength to minimize brightnessrelated soil effects(Adhikari et al.2012).MODIS-driven EVI images,composited at 16-day intervals,were downloaded in tiles for the period between January 2004 and December 2009 and mosaicked using the MODIS reprojection tool.Averaged monthly EVIs were used to assess the relationship between plant richness and productivity.Correlation analyses were preliminarily performed between species richness and the 12 monthly EVIs,and we selected the September and August EVIs as explanatory variables to interpret the signif i cance of productivity on richness patterns along the Namgyori and Osaek transects,respectively.Thus,we presented results based only on the September and August EVIs as surrogates for productivity.

Data analysis

Plant data from plots within the same elevational band were pooled,and the number of species observed in each band was considered to be a measure of richness(Acharya et al.2011).Because several studies have documented that different elevational patterns may be observed among different plant groups and there can be fundamental differences in the underlying mechanisms even within similar elevational diversity patterns among different plant groups (Bhattarai and Vetaas 2003;Watkins et al.2006;Lee et al. 2013),wedivided plant species into three groups based on life–form,viz.total,woody and herbaceous plants and all analysis were conducted for each of these three life-form groups.To determine the relationship between species richness and elevation,we used a generalized additive model(GAM;Hastie and Tibshirani 1990).The GAM is well-suited to studying ecological responses because of the nonparametric characteristics of such responses and is especially useful for data exploration as it makes no a priori assumptions about the type of relationship being modeled. GAM response curves are data-driven and def i ned by smoothing functions that can take any shape and do not assume a linear or quadratic relationship between the dependent and independent variables(Lehmann et al. 2002).Within the GAM framework,we used a cubic smoothing spline method with four degrees of freedom, and species richness was assumed to follow a Poisson distribution.We also created generalized linear models (GLMs)for comparison with the results of GAM.

Table 1 Observed species richness values for different elevational bands and the species richness of all bands pooled for each site for total, woody and herbaceous plants along the two transects on Mt. Seorak,South Korea

Fig.2 Relationships between elevation and the explanatory variables along the two transects on Mt.Seorak:a mean annual temperature(MAT), b mean annual precipitation(MAP),c area and d the enhanced vegetation index(EVI)

The relationships between plant species richness and the explanatory variables were calculated for each individual variable using a simple linear regression.Such a linear model tests only for linear relationships between the potential explanatory variables and species richness,but there are several plausible scenarios under which a unimodal model is actually more biologically reasonable (Kluge et al.2006).We also examined a polynomial regression model to detect curvilinear relationships by including a quadratic term into the regression function. Redundancy analysis(RDA)was also applied to quantify and test the effects of explanatory variables on plant species richness along elevational gradients.The signif i cant explanatory variables(p＜0.05)were selected with forward selection using an unrestricted Monte Carlo permutation test based on 999 random permutations.Finally,we used forward stepwise multiple regression models to establish the relative importance of MAT,MAP,area, MDE and EVI as explanatory variables of observed species richness.Forward stepwise multiple regressions were used to f i nd a set of independent variables that together provided the best f i t for observed species richness by minimizing the sum of squared residuals.All linear and quadratic terms of the explanatory variables were used in forward stepwise multiple regression.We used S-PLUS version 8.0 for GAM,GLM and forward stepwise multiple regression and CANOCO version 4.5 for RDA.

Results

General richness patterns and explanatory variables

A total of 235 plant species of 72 families and 161 genera were recorded from 130 plots along the two transects.More than half of these species were herbaceous(57%;42 families,101 genera,and 135 species),and woody species accounted for 43%(38 families,66 genera,and 100 species).On the Namgyori transect,we recorded 211 plant species of 71 families and 147 genera with 91 and 120 species of woody and herbaceous plants,respectively.On the Osaek transect we recorded 164 plant species of 59 families and 123 genera with 72 and 92 species of woody and herbaceous plants,respectively(Table 1).

MATs ranged with elevation from 2.95 to 7.59°C on the Namgyori transect and from 2.73 to 9.48°C on the Osaek transect.MATs across the entire elevational ranges were 5.87 and 6.17°C on the Namgyori and Osaek transects,respectively(Fig.2a).MAPs ranged with elevation from 1390 to 2065 mm on the Namgyori transect and from 1468 to 2125 mm on the Osaek transect.MAPs across the entire elevational ranges were 1844 and 1825 mm on the Namgyori and Osaek transects,respectively(Fig.2b).The mean areas of the elevational bands were 15.3 and 3.6 ha on the Namgyori and Osaek transects,respectively.The areas of the elevational bands ranged from 3.5 to 51.8 haon the Namgyori transect and from 1.8 to 8.4 ha on the Osaek transect(Fig.2c).Although the EVIs also showed different patterns between the transects,the minimum EVI values were at 800–1000 m for both transects(Fig.2d). The MDE null model showed deviations of observed from simulated species richness along both transects(Fig.3).On the Namgyori transect 36–50%of the data points for the three life-form groups occurred outside the 95%conf idence interval of the MDE null model and deviations of 17–50%were observed on the Osaek transect.

Fig.3 Observed species richness,predicted richness(computed from 5000 randomizations)and the 95%conf i dence intervals for the predicted MDE richness as a function of elevation for total(a–b),woody(c–d),and herbaceous(e–f)plants along the two transects on Mt.Seorak

Elevational species richness patterns

Data f i tting by GLM using a parametric method did not reveal any signif i cant trend of total or herbaceous species richness with elevation but showed signif i cant relationships between elevation and woody species richness along both transects,whereas the optimal f i ttings for total and herbaceous plants were obtained by GAM with a nonparametric smoothing method(Fig.4).The elevational patterns of species richness were similar for total andherbaceous plants and clearly differed between woody and total and between woody and herbaceous plants along both transects because there was a higher correlation coeff i cient between total and herbaceous plant species than between total and woody plant species along the Namgyori(total vs.woody,R2=0.25,p=0.07;total vs. herbaceous,R2=0.81,p＜0.001)and Osaek(total vs. woody,R2=0.52,p=0.008;total vs.herbaceous, R2=0.84,p＜0.001)transects.GAM curves representing elevational patterns showed that total and herbaceous plant richness peaked in the elevational band between 1200 and 1300 m(Fig.4a,e)along the Namgyori transect,whereas species richness steeply increased up to 900–1000 m,then declined to 1100–1200 m and then generally increased up to the peak along the Osaek transect(Fig.4b,f).Woody plant richness along the Namgyori transect monotonically declined with elevation (Fig.4c),whereas,a clear unimodal pattern with a peak at a low elevation was observed along the Osaek transect (Fig.4d).

Fig.4 Relationships between elevation and plant richness along the two transects on Mt.Seorak.The unbroken and broken lines were f i tted by generalized additive models(GAMs)and generalized linear models(GLMs),respectively

Richness patterns with explanatory variables

Based on the simple linear models,total and herbaceous plant richness were signif i cantly correlated with EVI along the Namgyori transect,whereas the climatic variables MAT and MAP were good predictors of woody plant richness (Table 2).Along the Osaek transect,area and EVIwere good predictors of total and herbaceous species richness,whereas area was signif i cantly correlated with woody species richness.The results of the quadratic models were somewhat different from those of the simple linear models(Table 2). Along the Namgyoritransect,area and MAP were signif i cant predictors of total and woody plant richness,respectively. However,no variables had signif i cant relationships with herbaceous plant richness.Area and EVI strongly correlated with total plant richness,woody plant richness was highly correlated with area,MAT and MAP and herbaceous plant richness were signif i cantly correlated only with EVI along the Osaek transect.

RDA diagrams described the relationships between plant species richness and the explanatory variables (Fig.5).In the diagrams,each relationship between species richness and an explanatory variable is represented by an arrow.The longer the arrow,the higher the importance of species richness and the explanatory variable for the distribution of the data.The position of plant species richness relative to an explanatory variable indicates how strongly plant species richness is associated with that variable.Plant species richness had close relationships with EVI and two climatic variables for the Namgyori transect(Fig.5a)and with EVI and area for the Osaek transect(Fig.5b).

The variations in total and woody plant richness along the Namgyori transect were best explained by EVI(39% of variation)and MAP(54%of variation),respectively (Table 3).Area and EVI were the most signif i cant variables and explained 53%of the variation in herbaceous plant richness.Along the Osaek transect,area and EVI accounted for more than half of the variation in total and herbaceous plant richness,respectively.Furthermore,area and MAP explained 80%of the variation in woody plant richness.

Discussion

Elevational species richness patterns

We investigated small-scale elevational patterns in plant richness using primary data,unlike many studies that examine broad,large-scale patterns with secondary data. Primary data at a small scale prove central to understanding within-domain diversity in biogeographic groups,whereas large-scale secondary data are critical to understanding thepattern across larger spatial scales(Oommen and Shanker 2005;Karger et al.2011).Therefore,our study benef i ted from exploring the elevational patterns and mechanisms related to an empirical data source at the local scale.

?

Fig.5 Triplot diagrams of the redundancy analysis(RDA)of species richness with explanatory variables along the a Namgyori and b Osaek transects on Mt.Seorak.Species richness is divided into total plants(Total),woody plants(Woody),trees(t),shrubs(s),woody vines(wv),herbaceous plants(Herbs),forbs(f),non-forbs(nf)and herbaceous vines(hv).Non-forbs are composed of grasses,sedges and ferns.Explanatory variables include mean annual temperature (MAT),mean annual precipitation(MAP),area(Area),the middomain effects for total(TMDE),woody(WMDE)and herbaceous (HMDE)plants,enhanced vegetation index(EVI)and elevation (ELEV)

The unimodal richness pattern that we detected for woody plants along the Osaek transect concurs with much of the existing macroecological research(Colwell and Lees 2000;McCain 2004;Rahbek 2005).It also appears that the pattern of monotonic decline in woody plant richness along the Namgyori transect that we found is consistent with earlier studies of mountain ecosystems(Stevens 1992; Kessler 2002).However,the f i nding that total and herbaceous species exhibited multimodal patterns along the elevational gradients of the two transects was unexpected. This multimodal pattern is extremely rarely observed in nature considering the large number of elevational gradients studied.Rahbek(2005)estimated that～50%of the recorded elevational patterns were unimodal,～25%followed a monotonic decline pattern,and the remaining 25%of the gradients followed other patterns.Our results are consistent with those of a recent study on vascular plants along an elevational gradient on Tianshan in central Xinjiang,China by Sang(2009),who observed that total and herbaceous plant species richness showed a bimodal pattern along an elevational gradient,with two peaks of plant richness located in the transition zones between vegetation types.Sang(2009)also found that this pattern was controlled by climatic and soil factors such as temperature,precipitation,soil water and nutrition.

Although the elevational patterns of species richness and their underlying causes have been controversial issues in ecology and biogeography(Wang et al.2011), these patterns have generally been explained by climatic factors including temperature and precipitation(Kluge et al.2006),spatial factors such as area and MDE(Wang et al.2007),energy-related factors such as net primary productivity and evapotranspiration(Chalcraft et al. 2004),and,rarely,evolutionary history(McCain 2009). The relative inf l uence of each of these determinants on richness may vary among elevational gradients and taxa. In that context,what are the underlying mechanisms controlling the patterns of plant richness along the two transects on Mt.Seorak?The different patterns in richness that we found for the plant groups are most likely to be caused by different mechanisms.Below,we discuss how MAT,MAP,area,MDE and productivity might inf l uence the elevational patterns of plant richness along the elevational gradient on Mt.Seorak.

Table 3 Results of forward stepwise multiple regression models for the explanatory variables including all linear and quadratic terms and the species richness of total,woody, and herbaceous plants along two transects on Mt.Seorak,South Korea

Climatic variables:MAT and MAP

The monotonic decline and unimodal patterns observed for woody plants could be caused by climatic variables or by the interplay between climatic variables and area on Mt. Seorak.From the simple and multiple regression models, the importance of the effects of climatic variables was observed along the Namgyori transect,whereas climatic variables and area were important determinants of the pattern along the Osaek transect.Overall,climatic variables were good predictors of the elevational patterns of woody plants on Mt.Seorak.Similar results have been observed in other parts of the world;for example,climatic variables such as temperature and precipitation have determined the richness patterns of vascular plants in other studies(Zhang et al.2011;Namgail et al.2012).Although the inf l uence of MAT was removed in the multiple regression models,we could not rule out the possibility of the effect of MAT on the richness patterns of woody plants.

Many studies have revealed that temperature and precipitation interact to play a signif i cant role in plant physiology,which in turn inf l uences species distribution and richness(Hawkins et al.2003).In fact,the results of the simple regression models revealed that MAT was also an important variable despite having lower explanatory power than MAP.This apparent contradiction may be due to the special relationship between MAT and MAP.Because the two climatic variables are highly correlated(Namgyori transect;R2=0.53,p=0.003,Osaek transects; R2=0.98,p＜0.001,respectively),the MAT effect may be substituted by MAP in the multiple regression models, at least for the plant species in this study.

Spatial variables:area and MDE

Area was an important predictor for all life-form groups along the Osaek transect but only for total life-forms along the Namgyori transect in the simple regression models.The results of the multiple regression models were similar to those of the simple regression models,that is,area was a good predictor of total and woody species diversity along the Osaek transect and of herbaceous species diversity along the Namgyori transect.Area has long been recognized as a crucial predictor determining elevational species richness patterns,and area is believed to have both indirect and direct effects on species richness(Connor and McCoy 1979;Rahbek 1995).Several mechanisms can explain the effect of area on species richness(Fu et al.2006).In general,a larger area of habitat is more heterogeneous and diverse than a smaller area and thus can support more coexisting species.

Primary productivity

Studies using remote sensing-based vegetation indices as surrogates for primary productivity have found signif i cant productivity-richness relationships,suggesting that such estimates can be used to evaluate biodiversity patterns (Rowe 2009).The relationship between productivity and species richness is generally assumed to be unimodal,but other response shapes have also been observed,including positive and,rarely,negative linear ones(Waide et al. 1999).The position of the optimum productivity for species richness might differ by taxon,ecosystem,scale and/or biogeographical region(Waide et al.1999).Our study found strong support for productivity(as measured by EVI) as a primary driver of elevational patterns of herbaceous species richness.However,the relationships between herbaceous plant richness and productivity were different between our two transects.Productivity was positively correlated with herbaceous plant richness on the Namgyori transect,whereas a negative relationship was observed on the Osaek transect.Why might the relationship between productivity and richness be different between the two transects?Recently,Hurlbert(2004)found a positive relationship between species richness and productivity in grasslands but a negative relationship in forests(as measured by NDVI).To explain this difference in results,he suggested that the effective survey area was smaller in forests than in grasslands,but he did not rule out the possibility that there might be real differences in the way that individual species in the two different habitats respond to increasing productivity.Several studies have also reported that the relationships between productivity and species richness depend on spatial scale(Waide et al.1999; Chalcraft et al.2004).The Namgyori transect(～22.4 km) on Mt.Seorak was approximately four times longer than the Osaek transect(～5.3 km).The different relationships between productivity and herbaceous plant richness between the two transects might be derived from the difference in spatial scales or effective survey areas.

Another possibility,based in competitive exclusion theory,is that at low productivity levels,stress and a lack of resources limit the number of species that can survive, and as productivity increases,species richness rises,but at high productivity levels,competitive exclusion reduces species richness,either because competition is more intense(Grime 1973)or because high productivity leads to a decrease in the heterogeneity of limiting resources (Huston 1979).Although the mean annual EVI of the Osaek transect was higher than that of the Namgyori transect(One-way ANOVA;F=34.12,p＜0.001),signif i cant EVIs were derived from different months between the two transects,and other monthly and mean annual EVIs did not indicate signif i cant relationships with herbaceous species richness.Thus,we cannot determine whether the competitive exclusion theory is applicable to our case.In addition,we cannot rule out the possibility that the difference in EVIs between the two transects sampled in this study could be surrogates for other variables such as habitat heterogeneity,availability or disturbance intensity rather than productivity(Koh et al.2006;Sanders et al.2007).

Other potential variables

From the results of RDAs and multiple regression models, the variation in elevational species richness explained by our set of variables differed between our two transects (Fig.5;Table 3).The variations in richness explained by the set of variables in the Osaek transect were generally higher than in the Namgyori transect from both analyses. Especially in the RDAs,the f i rst two axes explained 57 and 87%of the variations in elevational richness patterns in the Namgyori and Osaek transects,respectively.These differences may be explained by the spatial scale and other environmental variables.Ecological patterns and processes vary across space,and the scale at which a researcher chooses to express spatial variation can have a strong impact on the outcome of analysis(Wiens 1989).Indeed,spatial context can inf l uence the structure of biological communities when the mechanisms involved depend on the spatial scale and species distribution patterns can change with spatial grain and scale(Rahbek 2005).As mentioned earlier,our Namgyori transect was approximately four times longer than our Osaek transect.It is possible that spatial grain and scale were the major causes of differences in variations in richness patterns in this study.

Another possibility is that other factors such as habitatrelated variables(e.g.habitatheterogeneity and complexity), stochastic processes(e.g.f i re regimes),anthropogenic pressures(e.g.direct removal of plants)and unpredictable historical events had differing effects on the two transects. Especially,other potential variables might be more important predictors of the unexplained variation;an increase in the amount of explained variation in richness would be expected by computing additional and more relevant variables for the Namgyori transect than for the Osaek transect.

Conclusion

One of the most interesting f i ndings of our study was that the different life-form groups showed different patterns in richness and different controlling factors along the same environmental gradient.We found that climatic variables had higher explanatory power than other variables for the elevational patterns of woody plants,whereas productivity was more important for the pattern of herbaceous plant diversity.Although area was an important variable for predicting species richness patterns,the effects were different between the two transects and among the life-form groups.Furthermore,no empirical evidence was linked to MDE.Evolutionary history,habitat heterogeneity,and human disturbance may also be important variables and should be considered.In addition to explanatory variables, spatial scales and the data source(i.e.,primary or secondary data)can also affect patterns of species richness. Further study of the many factors involved in these patterns,including habitat-related variables such as heterogeneity and availability,human disturbance,evolutionary history,and climatic,spatial and energy-related variables at different spatial scales with a combination of primary and secondary data complementing each other,should be evaluated to better understand the elevational patterns of plant communities in mountain ecosystems.

AcknowledgmentsWe thank Mr.Keun-Wook Lee,Mr.Sang-HyoukSeo and Mr.Min-Woo Park for their invaluable help during the fi eldwork and data analysis in this study.Thanks are also due to Dr. Joon-Hwan Shin and Dr.Jong-Hwan Lim for their support and encouragement.This paper forms a part of the‘Korea Big Tree Project’funded by the Korea Green Promotion Agency,Korea Forest Service.

Acharya BK,Chettri B,Bijayan L(2011)Distribution pattern of trees along an elevation gradient of Eastern Himalaya,India.Acta Oecol 37:329–336

Adhikari D,Barik SK,Upadhaya K(2012)Habitat distribution modelling for reintroduction ofIlex khasianaPurk.,a critically endangered tree species of northeastern India.Ecol Eng 40:37–43

Bhattarai KR,Vetaas OR(2003)Variation in plant species richness of different life forms along a subtropical elevational gradient in the Himalayas,east Nepal.Glob Ecol Biogeogr 12:327–340

Braun-Blanquet J(1965)Plant sociology:the study of plant communities.Hafner Publishing,New York

Chalcraft DR,Williams JW,Smith MD,Willig MR(2004)Scale dependence in the species-richness-productivity relationship;the role of species turnover.Ecology 85:2701–2708

Colwell RK(2006)RangeModel:A Monte Carlo simulation tool for assessing geometric constraints on species richness. Version 5.User’s Guide and application published at:http:// viceroy.eeb.uconn.edu/rangemodel

Colwell RK,Hurtt GC(1994)Nonbiological gradients in species richness and a spurious Rapport effect.Am Nat 144:570–595

Colwell RK,Lees DC(2000)The mid-domain effect:geometric constraints on the geography of species richness.Trends Ecol Evol 15:70–76

Connor EF,McCoy ED(1979)The statistics and biology of the species-area relationship.Am Nat 113:791–833

Diniz-Filho JAE,Bini LM,Hawkins BA(2003)Spatial autocorrelation and red herrings in geographical ecology.Glob Ecol Biogeogr 12:53–64

Dunn RR,Colwell RK,Nilsson C(2007)When does diversity f i t null model predictions?Scale and range size mediate the mid-domain effect.Glob Ecol Biogeogr 16:305–312

Fu C,Hua X,Li J,Chang Z,Pu Z,Chen J(2006)Elevational patterns of frog species richness and endemic richness in the Hengduan Mountains,China:geometric constraints,area and climate effects.Ecography 29:919–927

Grime JP(1973)Competitive exclusion in herbaceous vegetation. Nature 242:344–347

Grytnes JA,McCain CM(2007)Elevational trends in biodiversity.In: Levin S(ed)Encyclopedia of biodiversity.Elsevier,Amsterdam, pp 1–8

Grytnes JA,Vetaas OR(2002)Species richness and altitude:a comparison between null models and interpolated plant species richness along the Himalayan altitudinal gradient Nepal.Am Nat 159:294–304

Han SK,Park YJ,Choi SK,Lee JO,Choi JH,Sung JM(2006)Some unrecorded higher fungi of the Seorakson and Odaesan National Parks.Mycobiology 34:56–60

Hastie TJ,Tibshirani RJ(1990)Generalized additive models. Chapman and Hall,New York,p 352

Hawkins BA,Field R,Cornell HV,Currie DJ,Gue´gan JF,Kaufman DM,Kerr JT,Mittelbach GG,Oberdorff T,O’Brien E,Porter EE,Turner JRG(2003)Energy,water,and broad-scale geographic patterns of species richness.Ecology 84:3105–3117

Hurlbert AH(2004)Species-energy relationships and habitat complexity in bird communities.Ecol Lett 7:714–720

Huston MA(1979)A general hypothesis of species diversity.Am Nat 113:81–101

Karger DN,Kluge J,Kro¨mer T,Hemp A,Lehnert M,Kessler M (2011)The effect of area on local and regional elevational patterns of species richness.J Biogeogr 38:1177–1185

Kessler M(2002)The elevational gradient of Andean plant endemism:varying inf l uences of taxon-specif i c traits andtopography at different taxonomic levels.J Biogeogr 29:1159–1165

Khuroo AA,Weber E,Malik AH,Reshi ZA,Dar GH(2011) Altitudinal distribution patterns of the native and alien woody fl ora in Kashmir Himalaya,India.Environ Res 111:967–977

Kluge J,Kessler M,Dunn RR(2006)What drives elevational patterns of diversity?A test of geometric constraints,climate and species pool effects for pteridophytes on an elevational gradient in Costa Rica.Glob Ecol Biogeogr 15:358–371

Koh CN,Lee PF,Lin RS(2006)Bird species richness patterns of northern Taiwan:primary productivity,human population density,and habitat heterogeneity.Divers Distrib 12:546–554

Kong WS(2008)Biogeography of Korea plants.GeoBook Publishing,Seoul,p 320

Ko¨rner C(2000)Why are there global gradients in species richness? Mountains might hold the answer.Trends Ecol Evol 15:513–514

Ko¨rner C(2007)The use of‘altitude’in ecological research.Trends Ecol Evol 22:569–574

Lee CB,Chun JH,Um TW,Cho HJ(2013)Altitudinal patterns and determinants of plant species richness on the Baekdudaegan Mountains,South Korea:common versus rare species.J Ecol Environ 36:193–204

Lehmann A,Overton JM,Austin MP(2002)Regression models for spatial prediction:their role for biodiversity and conservation. Biodivers Conserv 11:2085–2092

Lomolino MV(2001)Elevational gradients of species-density: historical and prospective views.Glob Ecol Biogeogr 10:3–13

Loreau M,Naeem S,Inchausti P,Bengtsson J,Grime JP,Hector A, Hooper DU,Huston MA,Raffaelli D,Schmid B,Tilman D, Wardle DA(2001)Biodiversity and ecosystem functioning: current knowledge and future challenges.Science 294:804–808

McCain CM(2004)The mid-domain effect applied to elevational gradients:species richness of small mammals in Costa Rica. J Biogeogr 31:19–31

McCain CM(2007)Area and mammalian elevational diversity. Ecology 88:76–86

McCain CM(2009)Global analysis of bird elevational diversity. Glob Ecol Biogeogr 18:346–360

Namgail T,Rawat GS,Mishra C,van Wieren SE,Prins HHT(2012) Biomass and diversity of dry alpine plant communities along altitudinal gradients in the Himalayas.J Plant Res 125:93–101

Oommen MA,Shanker K(2005)Elevational species richness patterns emerge from multiple local mechanisms in Himalayan woody plants.Ecology 86:3039–3047

Rahbek C(1995)The elevational gradient of species richness:a uniform pattern?Ecography 18:200–205

Rahbek C(2005)The role of spatial scale and the perception of largescale species-richness patterns.Ecol Lett 8:224–239

Rosenzweig ML(1995)Species diversity in space and time. Cambridge University Press,Cambridge,p 460

Rowe RJ(2009)Environmental and geometric drivers of small mammal diversity along elevational gradients in Utah.Ecography 32:411–422

Sanders NJ,Lessard JP,Fitzpatrick MC,Dunn RR(2007)Temperature,but not productivity or geometry,predicts elevational diversity gradients in ants across spatial grains.Glob Ecol Biogeogr 16:640–649

Sang W(2009)Plant diversity patterns and their relationships with soil and climatic factors along an altitudinal gradient in the middle Tianshan Mountain area,Xinjiang,China.Ecol Res 24:303–314

Stevens GC(1992)The elevational gradient in altitudinal range:an extension of Rapoport’s latitudinal rule to altitude.Am Nat 140:893–911

Storch D,Marquet PA,Gaston MJ(2005)Untangling an entangled bank.Science 307:684–686

Waide RB,Willing MR,Steiner CF,Mittelbach G,Gough L,Dodson SI,Juday GP,Parmenter R(1999)The relationship between productivity and species richness.Annu Rev Ecol Syst 30:257–300

Wang Z,Tang Z,Fang J(2007)Altitudinal patterns of seed plant richness in the Gaoligong Mountains,south-east Tibet,China. Divers Distrib 13:845–854

Wang J,Soininen J,Zhang Y,Wang B,Yang X,Shen J(2011) Contrasting patterns in elevational diversity between microorganisms and macroorganisms.J Biogeogr 38:595–603

Watkins JE,Cardelu´s C,Colwell RK,Moran RC(2006)Species richness and distribution of ferns along an elevational gradient in Costa Rica.Am J Bot 93:73–83

Wiens JA(1989)Spatial scaling in ecology.Funct Ecol 3:385–397

Yun JI(2010)Agroclimatic maps augmented by a GIS technology. Korean J Agric For Meteorol 12:63–73

Zapata FA,Gaston KJ,Chown SL(2003)Mid-domain models of species richness gradients:assumptions,methods and evidence. J Anim Ecol 72:677–690

Zhang JT,Xu B,Li M(2011)Diversity of communities dominated byGlycyrrhiza uralensis,an endangered medicinal plant species, along a precipitation gradient in China.Bot Stud 52:493–501

MDE has

increasing attention as a primary driver of species richness along spatial gradients in various domains for the last decade,we found no support for the MDE null models in this study.One possible cause of this result could be the non-interpolation of species data in our study.Unlike many previous studies,we did not use interpolated species richness modif i ed from actual distribution records.The underlying reason for interpolation is that gaps are present in elevational distributions due tounder sampling(Kluge et al.2006).However,recent studies have reported problems with interpolation(Grytnes and Vetaas 2002;Diniz-Filho et al.2003;Kluge et al. 2006).First,interpolation disrupts the crucial control of sampling area and intensity,as species are added that were not,in fact,present in the plots.Second,interpolation might artif i cially increase richness to a higher degree at intermediate elevations because gaps are f i lled only between the lower and upper range limits;this basically assumes that no individuals of a species have been missed beyond the observed range limits but that individuals have been missed at sampling points within the range limits. Third,interpolated species richness values at nearby elevations are more similar than at distant elevations,and the resulting spatial autocorrelation inf l ates Type I errors. Several recent studies also suggest that both the prediction of MDE and its underlying assumptions lack empirical evidence(Zapata et al.2003)and that the explanatory power of MDE can depend on the domain size(Oommen and Shanker 2005;Dunn et al.2007).Furthermore, McCain(2007)suggests that when area is accounted for, MDE is no longer a strong predictor of richness,a f i nding which is not corroborated by our study.The spurious effects of autocorrelation increase when using interpolated distribution data.We suspect that,at least for plant species in this study,the generality of MDE as the primary predictor of elevational patterns of species richness is in doubt.

Received:21 January 2014/Accepted:15 May 2014/Published online:21 July 2015

Project fund:This work was a part of the‘Korea Big Tree Project’funded by the Korea Green Promotion Agency,Korea Forest Service.

The online version is available at http://www.springerlink.com

Corresponding editor:Yu Lei

✉Chang-Bae Lee

fl ora1@kgpa.or.kr

1Korea Green Promotion Agency,121 Dunsanbukro,Seogu, Daejeon 302-831,Republic of Korea

2Division of Forest Ecology,Korea Forest Research Institute, 57,Hoegiro,Dongdaemungu,Seoul 130-712, Republic of Korea