Article

Assessing Forest Structure and Composition along theAltitudinal Gradient in the State of Sikkim EasternHimalayas India

Yangchenla Bhutia 12 Ravikanth Gudasalamani 1 Rengaian Ganesan 1 and Somidh Saha 341 Suri Sehgal Centre for Biodiversity and Conservation Ashoka Trust for Research in Ecology and the

Environment Royal Enclave Sriramapura Jakkur Post Bengaluru 560064 India2 Manipal Academy of Higher Education (MAHE) Tiger Circle Road Madhav Nagar Manipal

Karnataka 576104 India3 Institute for Technology Assessment and Systems Analysis Karlsruhe Institute of Technology Karlstr 11

D-76133 Karlsruhe Germany4 Chair of Silviculture Institute of Forest Sciences University of Freiburg Tennenbacherstr 4

D-79085 Freiburg Germany Correspondence yangchenlabhutiaatreeorg (YB) somidhsahakitedu (SS)

Tel +91-8348-144-849 (YB) +49-7216-0824-644 (SS)

Received 9 June 2019 Accepted 24 July 2019 Published 27 July 2019

Abstract Understanding the structure and composition of native forests is a prerequisite in developingan adaptive forest management plan for Himalayan forest ecosystems where climate change is rapidHowever basic information on forest structure and composition are still limited in many places of theEastern Himalayas In this study we aimed to understand the diversity structure and compositionof forests and their variations along an altitudinal gradient in Himalayan forests The study wasconducted in the Indian federal state of Sikkim Eastern Himalayas We carried out a comprehensiveand comparative evaluation of species diversity stand basal area and stem density along thealtitudinal gradient from 900 m asl to 3200 m asl We used stratified random sampling to surveyeighty-three plots each 01 ha in forest communities that occurred along the altitudinal gradient(a) lower (900ndash1700 m) altitude forest (N = 24) (b) mid (1700ndash2500 m) altitude forests (N = 37) and(c) higher (2500ndash3200 m)altitude forests (N = 22) We measured and identified all living trees witha gt3 cm diameter at breast height in each plot We counted 10344 individual plants representing114 woody species belonging to 42 families and 75 genera The family Fagaceae and its speciesLithocarpus pachyphyllus (Kurz) Rehder were reported as the most dominant forest trees with thehighest Importance Value Index The Shannon diversity index was recorded as being the highestfor the low-altitude forests whereas measures of structural diversity varied among forests alongwith altitude the mid-altitude forests recorded the highest stem density and the high-altitude forestsshowed the highest mean stem DBH (diameter at 13 m height) One significant finding of our studywas the disparity of the size class distribution among forests along the altitudinal gradient Overallwe found a reverse J-shape distribution of tree diameter signifying the uneven-agedness Howeverwe showed for the first time a complete lack of large trees (gt93 cm DBH) in the lower altitudeforests Our study highlights conservation concerns for the low-altitude forests that record highspecies diversity although lacked large-diameter trees We anticipate that our study will provide acomprehensive understanding of forest diversity composition and structure along the altitudinalgradient to design conservation and sustainable management strategies

Keywords Eastern Himalaya Sikkim species diversity size class distribution uneven-aged forests

Forests 2019 10 633 doi103390f10080633 wwwmdpicomjournalforests

Forests 2019 10 633 2 of 17

1 Introduction

11 Background

Quantitative assessment of the diversity structure and composition of forest trees are essential notonly to understand forest biodiversity and health but also for designing conservation strategies towardsclimate change [1] which is a significant threat to biodiversity in the Himalayas [2ndash5] The EasternHimalaya region is a confluence of multiple biogeographic origins such as the IndondashMalayan realmPalearctic and the SinondashJapanese region and it is characterized by remarkable biodiversity of majorecological and global significance and represents one of the 34 global biodiversity hotspots [67]It covers an extensive area of 524190 sqkm ranging from east of Nepal to northwest Yunnan Provincein China including the northeastern region of India [2] The Himalayan forests stores abundantcarbon in forest soil and vegetation and thus are of prime importance for the regional and globalcarbon cycles [89] It supplies constant water and is the prime regulator of life [10] Further theforest-based biological resources are the chief source of livelihood for people living in the Himalayas [11]Simultaneously the Himalayas are one of the most fragile systems on earth [12] geologically young(45 million years) tectonically active and highly susceptible to natural hazards [4] Moreover theHimalayas are reported as the prime site for changing climate [13] with forest vegetation notablyvulnerable to climate change [14ndash16] The degrading Himalayan forests not only affect the surroundingregion but have global implications [17] While the ecological fragility and the significance of Himalayanforests are much realized basic knowledge on the structure and composition in Himalayan forests arestill limited in many regions particularly in the remote eastern parts [16]

Forest trees are the dominant structural and functional component of the forest ecosystem [18]Tree species diversity size class distribution stem density and basal area are the essential attributesthat describes a forestrsquos ecosystem [1920] and measuring these attributes are fundamental in designingconservation strategies Species diversity is the most crucial descriptors which not only capturesinformation on species richness (number of species in a community) but the relative abundanceof species in a forest It also provides information on the rarity and the commonness of a speciesThe ability of species diversity to quantify forest composition has provided ecologist with the mostprominent tool often used as the scorecard to preserve and restore a forest Likewise the sizeclass distribution of forest trees is the prime indicator of forest structure and dynamics widely usedto examine the forestrsquos health including regeneration [21] Tree diameter is a statistically provenparameter to measure forest carbon stocks [22] Furthermore stem density and basal area are anexcellent surrogate to estimate forest biomass and carbon [2324] A proper understanding of forestcomposition and structure allow foresters national park rangers and landowners to maximize forestecosystemrsquos goods and services by maintaining or conserving a desired structure and composition ofthe forest at stand and landscape level

12 Past Studies in the Eastern Himalayas

Studies in the Eastern Himalayan forests outside of India were mainly carried out in Nepal [2526]Bhutan [27ndash29] and China [3031] to assess forest structure and composition Within the Indian part ofthe Eastern Himalayas studies were carried out in Arunachal Pradesh [32ndash34] Meghalaya [83536]Darjeeling West Bengal [37] and Assam [3839] The majority of these studies presented the structureand composition of different forests along altitudinal gradients with varying intensities of anthropogenicdisturbance However we did not find similar studies from the Sikkim State of the Eastern Himalayaregion in India Studies on the forest of Sikkim have mostly been carried out as case studies withconcentrated sampling on watershed [40] alpine forests [4142] timberline forest [43] and trekkingcorridors [44] with narrow altitudinal range [4243] Albeit more recently Acharya et al [45] andSharma et al [46] performed research with samples from multiple locations but did not describe thecomposition and size class distribution of trees

Forests 2019 10 633 3 of 17

13 Aims of this Study

In this study we aimed to fill the existing knowledge gap on the diversity structure andcomposition of forest tree species using well-represented samples covering forests along the altitudinalgradient in Sikkim a part of the Eastern Himalayas Our study not only provides an overall assessmenton the diversity composition and structure of forest tree communities but also evaluates how theseattributes vary among forest communities which occur along the altitudinal gradient

2 Materials and Methods

21 Study Area Description and Data Collection

Sikkim (2707prime to 2813prime N and 8801prime to 8892prime E) a northeastern region of India falls within theEastern Himalayan biodiversity hotspot (Figure 1) It has a remarkable altitudinal gradient with only anarea of 7096 km2 ranging from 300 m asl to 8586 m asl with Mount Kanchendzongamdashthe third highestpeak in the world The complex topography and enormous altitude are characterized by 12 majorvegetation types from a tropical warm broad-leaved forest at the lowest altitude to alpine meadowat the highest altitude [47] It has a subtropical to temperate climate with annual rainfall between2700 mm to 3200 mm and the mean annual temperature (MAT) varies from 84 C to 232 C [48]

Forests 2019 10 x FOR PEER REVIEW 3 of 18

In this study we aimed to fill the existing knowledge gap on the diversity structure and composition of forest tree species using well-represented samples covering forests along the altitudinal gradient in Sikkim a part of the Eastern Himalayas Our study not only provides an overall assessment on the diversity composition and structure of forest tree communities but also evaluates how these attributes vary among forest communities which occur along the altitudinal gradient

2 Materials and Methods

21 Study Area Description and Data Collection

Sikkim(27deg07rsquoto 28deg13rsquoN and 88deg01rsquoto 88deg92rsquoE) a northeastern region of India falls within the Eastern Himalayan biodiversity hotspot (Figure 1) It has a remarkable altitudinal gradient with only an area of 7096 km2 ranging from 300 m asl to 8586 m asl with Mount Kanchendzongamdashthe third highest peak in the world The complex topography and enormous altitude are characterized by 12 major vegetation types from a tropical warm broad-leaved forest at the lowest altitude to alpine meadow at the highest altitude [47] It has a subtropical to temperate climate with annual rainfall between 2700 mm to 3200 mm and the mean annual temperature (MAT) varies from 84degC to 232degC [48]

Figure 1 Study area with sampling locations along an altitude gradient in Sikkim India

The field study was conducted over 3 years (2013ndash2015) during the dry winter seasons from the beginning of October to late May to avoid challenges caused by understory growth of thick shrubs and herbs during the monsoon season Stratified random sampling was used to sample plots along the altitudinal gradients from 900 m asl to 3200 m asl The forests were selected because forests within the range has undergone significant loss in forest cover [47] Thus quantitative assessment of these forests is imperative to plan conservation strategies The forests were stratified into three altitudinal zones low-altitude forest (LF) ranging from 900 masl to 1700 m asl the mid-altitude forest (MF) ranging from 1700 m aslto 2500 m asl and the high-altitude forest (HF) ranging from 2500 m asl to 3200 m asl Plots were randomly installed at different locations for each altitude zone Altogether we installed 83 plots of 100 m times 10 m (each 01 ha) covering a total area of 83 ha The number of plots varied across forests and were distributed as LF=24 MF = 37 and HF= 22 plots among three forest communities along the altitudinal gradient Within each plot all living woody trees with gt3 cm diameter at breast height (DBH or trunk diameter at 13 m above ground level) were

Figure 1 Study area with sampling locations along an altitude gradient in Sikkim India

The field study was conducted over 3 years (2013ndash2015) during the dry winter seasons from thebeginning of October to late May to avoid challenges caused by understory growth of thick shrubs andherbs during the monsoon season Stratified random sampling was used to sample plots along thealtitudinal gradients from 900 m asl to 3200 m asl The forests were selected because forests withinthe range has undergone significant loss in forest cover [47] Thus quantitative assessment of theseforests is imperative to plan conservation strategies The forests were stratified into three altitudinalzones low-altitude forest (LF) ranging from 900 m asl to 1700 m asl the mid-altitude forest (MF)ranging from 1700 m asl to 2500 m asl and the high-altitude forest (HF) ranging from 2500 m aslto 3200 m asl Plots were randomly installed at different locations for each altitude zone Altogetherwe installed 83 plots of 100 m times 10 m (each 01 ha) covering a total area of 83 ha The number ofplots varied across forests and were distributed as LF = 24 MF = 37 and HF = 22 plots among threeforest communities along the altitudinal gradient Within each plot all living woody trees with gt3 cmdiameter at breast height (DBH or trunk diameter at 13 m above ground level) were measured andidentified Voucher specimens were collected for individuals challenging to identify in the field andwas used later for further identification

Forests 2019 10 633 4 of 17

22 Data Analysis

The analysis included quantitative data on the abundance DBH taxonomic and geo-coordinatedetails for all sites Prior to the quantification of forest attributes we conducted cluster analysis andordination to examine and validate the presence of ecologically meaningful clusters among our sampledplots designed for the study First we computed the BrayndashCurtis distance matrix using the abundancedata with species in the columns and sites in ascending order of altitude in rows Then we performedhierarchical Wardrsquos minimum variance clustering on the BrayndashCurtis distance matrix This method isbased on the least squares linear model criteria and forms cluster with the least variance [49] Secondlywe performed the Nonmetirc multidimensional scaling method (NMDS) an indirect gradient analysisapproach to ordinate our sampled sites using the same distance matrix Nonmetric multidimensionalscaling represents the pairwise dissimilarity among sites in a low dimensional space as closely aspossible [49] We also estimated the goodness of fit measured as ldquostressrdquo from the Shepard diagram toevaluate the appropriateness of the NMDS result [49]

We plotted the species accumulation curve to determine the sampling effectiveness for the entiresample installed in our study Further we plotted sample-based rarefaction curve for different forestcommunities It computes the average richness by randomly drawing 123n samples representing acommunity without replacement [50] The richness estimated from the sample-based rarefaction curveswas also used to compare the richness for different forest communities with distinct sampling effort

Shannonrsquos diversity index (Hrsquo) was estimated by multiplying the proportion of each species totheir natural log Shannon Index (Hrsquo) = Σpilog(ln)pi where pi is the proportion (nN) of individuals ofa particular species found (n) divided by the total number of individuals recorded (N) ln is the naturallog and Σ is the sum of the calculations

Shannon diversity provides information on both species richness and relative abundance amongplots and thus are sensitive to the sampling effort or the number of individuals sampled To avoidbias we estimated other indices namely the rarefied richness for forest communities and comparedrichness among subsets using a fixed sample size Further we estimated the Fisher alpha a parametricrichness estimate which compares richness among samples with varying individuals Pieloursquos J wasalso estimated to check for differences in evenness among communities with different sample sizeApart from these we estimated observed species richness genus richness and family richness for theoverall forest community and three forest communities along the altitudinal gradient

Additionally we estimated four essential stand structural parameters the median DBH stemdensity (haminus1) basal area (m2 haminus1) and size class distribution of forest trees We mostly usedthe median as a measure of central tendency because variables often did not follow a Gaussiandistribution We calculated the Important Values Index (IVI) both for species(SIVI) and families(FIVI)We followed the equations by Keel et al [51] and Ganesh et al [52] for IVI estimation Diversityanalysis clustering and ordination were conducted using the R package ldquoveganrdquoand ldquoBiodiversityRrdquoV29-2 and the built-in function ldquodecostandrdquo ldquovegdistrdquo ldquohclustrdquo ldquowardD2rdquoand ldquometaNMDSrdquoin RStudio (version 11383)

3 Results

31 Distinct Forest Communities along the Altitude Gradient from 900 m to 3200 m

The cluster analysis revealed that three forest communities could be formed based on thecomposition of tree species at different sites although few plots deflected from its designated categories(Figure 2a) The NMDS result superimposed with a cluster dendrogram also showed similar groups(Figure 2b) The Shepard diagram and the goodness of fit also supported our result with high R2

value (Figure 3) It can be stated that both the cluster analysis and NMDS suggested the presence ofthree ecologically meaningful forest communities along the altitudinal gradient with distinct speciescomposition These three forest communities corresponded to the altitudinal gradient hence reiteratedthe importance of altitude as a covariate to determine the composition of tree species in the Himalayas

Forests 2019 10 633 5 of 17

Based on the Species Importance Value Index and a proportion of stem density and basal area(see Table 3) all forests can be classified as mixed-broadleaf forests but with distinct communities atdifferent altitudes The lower altitude (LF) forests can be classified as the association of Schima wallichiiChoisy Castanopsis tribuloides Smith Engelhardtia spicata Blume and Gynocardia odorata RBr (Figure 4)The mid-altitude (MF) forests can be classified asthe association of Symplocos ramosissima Wall exG Don Castanopsis hystrix Hookf amp Thomson ex ADC Viburnum erubescens Wall and Quercuslamellosa Sm (Figure 5) Whereas high-altitude (HF) forests can be classified as the associationof Lithocarpus pachyphyllus Rhododendron arboreum Sm Magnolia campbellii Hookfamp Thomson andSymplocos heishanensis Hayata (Figure 6)

Forests 2019 10 x FOR PEER REVIEW 5 of 18

distinct species composition These three forest communities corresponded to the altitudinal gradient hence reiterated the importance of altitude as a covariate to determine the composition of tree species in the Himalayas Based on the Species Importance Value Index and a proportion of stem density and basal area (see Table 3)all forests can be classified as mixed-broadleaf forests but with distinct communities at different altitudes The lower altitude (LF) forests can be classified as the association of Schima wallichii ChoisyCastanopsis tribuloides Smith Engelhardtia spicata Blume and Gynocardia odorata RBr(Figure 4) The mid-altitude (MF) forests can be classified asthe association of Symplocos ramosissima Wall ex G DonCastanopsis hystrix Hookf amp Thomson ex ADC Viburnum erubescens Wall and Quercus lamellosa Sm (Figure 5) Whereas high-altitude (HF) forests can be classified as the association of Lithocarpus pachyphyllus Rhododendron arboreum Sm Magnolia campbellii Hookfamp Thomson and Symplocos heishanensis Hayata (Figure 6)

Figure 2 Hierarchically clustered dendrogram and Nonmetic multidimensional scaling (NMDS) ordination on the BrayndashCurtis distance estimated using abundance data (a) Hierarchical clustering was done using Wardrsquos algorithm the y-axis represents the BrayndashCurtis distance shared among each cluster (b) NMDS ordination superimposed with the result of the cluster analysis more similar plots are grouped closer to another The red green and blue dots and lines correspond to the low-altitude forest (LF) mid-altitude forest (MF) and high-altitude forest (HF) forest categories in our study

Figure 2 Hierarchically clustered dendrogram and Nonmetic multidimensional scaling (NMDS)ordination on the BrayndashCurtis distance estimated using abundance data (a) Hierarchical clusteringwas done using Wardrsquos algorithm the y-axis represents the BrayndashCurtis distance shared among eachcluster (b) NMDS ordination superimposed with the result of the cluster analysis more similar plotsare grouped closer to another The red green and blue dots and lines correspond to the low-altitudeforest (LF) mid-altitude forest (MF) and high-altitude forest (HF) forest categories in our studyForests 2019 10 x FOR PEER REVIEW 6 of 18

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure2

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees in flowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis and Schima wallichii trees recorded at 1400 m asl

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure 2

Forests 2019 10 633 6 of 17

Forests 2019 10 x FOR PEER REVIEW 6 of 18

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure2

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees in flowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis and Schima wallichii trees recorded at 1400 m asl

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees inflowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis andSchima wallichii trees recorded at 1400 m aslForests 2019 10 x FOR PEER REVIEW 7 of 18

Figure 5Mid-elevation forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl (b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocos species at 2300 m asl

Figure 6High elevation forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at2860 m asl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 m aslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiern were reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

Figure 5 Mid-altitude forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl(b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocosspecies at 2300 m asl

Forests 2019 10 633 7 of 17

Forests 2019 10 x FOR PEER REVIEW 7 of 18

Figure 5Mid-elevation forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl (b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocos species at 2300 m asl

Figure 6High elevation forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at2860 m asl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 m aslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiern were reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

Figure 6 High altitude forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at 2860 masl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 maslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiernwere reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size)including 87 unidentified individuals with the observed richness of 114 species from 42 families and75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentleslope for the cumulative species richness after the sample size of 70 indicating an adequate number ofplots (Figure 7a) However for different forest communities the species accumulation curve showedan increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest speciesrichness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversityvaried from 171 plusmn 040 to 210 plusmn 035 with the highest value for the LF (Table 1) and the Pieloursquosevenness (J) varied from 046 plusmn 07 to 035 plusmn 010 with the highest value for the LF (Table 1)

Tables 2 and 3 show the results for the Importance Value Index (IVI) estimated for both species(SIVI) and families (FIVI) along with species richness families stem density basal area theirproportion to the contribution for the overall forest community and different forest communitiesWe found Lithocarpus pachyphyllus (930) Symplocos ramosissima (872) Castanopsis hystrix (823)Castanopsis tribuloides (499) and Schima wallichii (446) to be the five most important specieswith the highest Species Importance Value Index (SIVI) Similarly we recorded Fagaceae (2786)Symplocaceae (1226) Theaceae (804) Ericaceae (770) and Lauraceae (662) as the top fivefamilies Lauraceae was also recorded as the most diverse family with 13 species (Table 2)

Forests 2019 10 633 8 of 17

Forests 2019 10 x FOR PEER REVIEW 8 of 18

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size) including 87 unidentified individuals with the observed richness of 114 species from 42 families and 75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentle slope for the cumulative species richness after the sample size of 70 indicating an adequate number of plots (Figure 7a) However for different forest communities the species accumulation curve showed an increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest species richness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20 the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversity varied from 171plusmn040 to 210plusmn035 with the highest value for the LF (Table 1) and the Pieloursquos evenness (J) varied from 046plusmn07 to 035plusmn010 with the highest value for the LF (Table1)

Figure 7Species accumulation curve based on the rarefaction method (a) whole tree population (b)three forests LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities along an altitude gradient The bold figures represent the highest value for each parameter The table provides the median and its standard error since richness showed a right-skewed distribution LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Parameters ALL LF MF HF Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200

Total sampled area (ha) 83 24 37 22 Total species richness 114 68 77 51 Total genus richness 75 52 54 37 Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450

Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351 Median of familiesrsquo richness plusmn

SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281

Total individual counted 10344 2591 5463 2290 Shannon index (H) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fishers alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164

Figure 7 Species accumulation curve based on the rarefaction method (a) whole tree population(b)three forests LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities alongan altitude gradient The table provides the median and its standard error since richness showed aright-skewed distribution LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Parameters ALL LF MF HF

Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200Total sampled area (ha) 83 24 37 22Total species richness 114 68 77 51Total genus richness 75 52 54 37Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351

Median of familiesrsquo richness plusmn SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281Total individual counted 10344 2591 5463 2290Shannon index (Hrsquo) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fisherrsquos alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164Pieloursquos evenness (J) 039 plusmn 039 046 plusmn 07 035 plusmn 010 038 plusmn 013

Dominant family Fagaceae Fagaceae Fagaceae Fagaceae

Dominant species Lithocarpuspachyphyllus Schima wallichii Symplocos

ramosissimaLithocarpus

pachyphyllus

Species such as Schima wallichii (1890) Castanopsis tribuloides (1791) Engelhardtia spicata (672)Gynocardia odorata (605) and Eurya acuminata (483) were recorded as the five most dominant specieswith higher species Important Value Index (SIVI) for LF Similarly Symplocos ramosissima (1509)Castanopsis hystrix (1499) Viburnum erubescens (709) Quercus lamellosa (685) and Eurya acuminata(442) were reported as the most dominant species for the MF The HF was dominated by speciessuch as Lithocarpus pachyphyllus (2437) Rhododendron arboreum (1678) Magnolia campbellii (454)Symplocos heishanensis (414) and Symplocos ramosissima (396) with the highest SIVI (Table 3)

Forests 2019 10 633 9 of 17

Table 2 Species richness density basal area Species Importance Value Index (SIVI) FamiliesImportance Value Index (FIVI) and their proportion to the contribution for the top ten species andfamilies for the overall forest community recorded in the study

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

Lithocarpus pachyphyllus Fagaceae 5120 411 1231 2107 930Symplocos ramosissima Symplocaceae 24807 1919 161 276 872

Castanopsis hystrix Fagaceae 6205 498 991 1697 823Castanopsis tribuloides Fagaceae 6241 501 422 723 499

Schima wallichii Theaceae 7819 627 240 411 446Rhododendron arboreum Ericaceae 8554 686 254 435 433

Viburnum erubescens Adoxaceae 10398 834 038 064 416Eurya acuminata DC Pentaphylacaceae 6096 489 081 139 391

Quercus lamellosa Fagaceae 988 079 495 849 386Symplocos heishanensis Symplocaceae 4940 396 038 065 293

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 9 2088 1675 3361 5719 2786Symplocaceae 4 3355 2692 218 371 1226

Theaceae 2 1390 1116 320 545 804Ericaceae 7 1476 1184 381 648 770Lauraceae 13 463 371 497 845 662Adoxaceae 2 1041 835 038 064 464

Magnoliaceae 4 112 090 209 364 270Betulaceae 4 193 155 099 169 219

Primulaceae 3 228 183 041 023 210Sapindaceae 1 069 055 123 210 189

Table 3 Stem density basal area Family (FIVI) and Species Importance Value Indexes (SIVI) and theirproportion to the contribution of the ten dominant species for three forest communities LF MF and HF

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

LF (900 m to 1700 m asl)

Schima wallichii Theaceae 25917 2401 708 2673 1890Castanopsis tribuloides Fagaceae 20250 1876 768 2900 1791

Engelhardtia spicata Juglandaceae 6042 560 235 886 672Gynocardia odorata Achariaceae 9250 857 123 463 605Euryaa cuminata Pentaphylacaceae 6667 857 076 286 483Betula alnoides

Buch-HamexDDon Betulaceae 2125 618 099 375 257

Macaranga denticulataBlume Euphorbiaceae 1808 197 039 148 243

Albizia lebbeck (L)Benth Leguminosae 1708 158 123 464 231

Alnus nepalensis DDon Betulaceae 1208 066 100 377 221Brassaiopsis hispida

Seem Araliaceae 3625 336 019 053 187

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Theaceae 2 942 3018 784 2960 2254Fagaceae 6 673 2157 823 3109 2017

Juglandaceae 3 180 575 238 899 742Achariaceae 1 267 857 123 463 658Betulaceae 3 104 332 209 789 526

Euphorbiaceae 5 123 394 059 222 434Araliaceae 3 157 502 017 066 353

Leguminosae 3 030 096 127 480 290Anacardiaceae 3 059 189 048 181 276

Lauraceae 6 063 201 009 035 264

Forests 2019 10 633 10 of 17

Table 3 Cont

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

MF (gt1700 m to 2500 m asl)

Symplocos ramosissima Symplocaceae 50189 3399 336 467 1509Castanopsis hystrix Fagaceae 13243 897 2172 3015 1499

Viburnum erubescens Adoxaceae 21189 1435 075 105 709Quercus lamellosa Fagaceae 2054 139 1067 1481 685Eurya acuminata Pentaphylacaceae 8378 567 123 171 442

Lithocarpus pachyphyllus Fagaceae 4892 331 517 718 432Symplocos heishanensis Symplocaceae 7378 500 056 078 407Castanopsis tribuloides Fagaceae 865 059 449 623 278

Quercus glauca Fagaceae 1432 097 309 429 270Elaeocarpus sikkimensis

Mast Elaeocarpaceae 1054 071 221 307 252

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 7 1082 1644 4570 4570 3012Symplocaceae 4 2890 4391 433 433 2014

Lauraceae 11 308 469 960 960 922Adoxaceae 2 946 1437 075 075 807Theaceae 2 395 600 171 171 573

Elaeocarpaceae 2 048 073 223 223 317Ericaceae 5 249 379 049 049 272

Sapindaceae 1 040 060 150 150 260Primulaceae 3 124 189 016 016 260

Magnoliaceae 4 059 090 175 175 243

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

HF (2500 m to 3200 m asl)

Lithocarpus pachyphyllus Fagaceae 11091 1066 788 5456 2437Rhododendron arboreum Ericaceae 31000 2978 747 1308 1678

Magnolia campbellii Magnoliaceae 1682 162 539 662 454Symplocos heishanensis Symplocaceae 6182 594 581 069 414Rhododendron hodgsonii

Hookf Ericaceae 5455 524 290 315 376

Corylus ferox Wall Betulaceae 2682 258 415 133 268Viburnum erubescens

Wall Adoxaceae 3591 345 415 021 260

Acer campbellii Sapindaceae 1091 105 373 298 259Rhododendron falconeri

Hookf Ericaceae 3955 380 207 186 258

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 4 333 9011 4096 5716 7921Ericaceae 7 1182 2963 1347 1879 7216

Symplocaceae 3 464 207 094 131 2602Magnoliaceae 2 048 1090 495 691 1603

Lauraceae 6 092 551 251 350 1524Betulaceae 2 076 216 098 137 991Theaceae 2 053 146 067 093 969

Berberidaceae 2 139 037 017 024 947Adoxaceae 1 095 033 015 021 992

Sapindaceae 1 029 470 214 298 877

33 Forest Structure

Overall we recorded the median stem DBH of 1542 plusmn 111 cm stem density of124627 plusmn 56828 haminus1 and the basal area of 5435 plusmn 444 m2 haminus1 (Figure 8) Among forest communitiesthe HF recorded the highest median stem DBH (2089 plusmn 309 cm) but showed the lowest stem densitySimilarly MF recorded the highest stem density of about 1320 plusmn 11972 haminus1 with the second highest

Forests 2019 10 633 11 of 17

value for the stem DBH and basal area However the LF recorded the second highest value for medianstem density and the minimum value for the median DBH and basal areaForests 2019 10 x FOR PEER REVIEW 12 of 18

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem density and (c) DBH The horizontal line crossing the box in the center represents the median the box represents the 25th and the 75th percentiles The vertical line outside the box represents the minimum and maximum values ALL= total forest community LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure9) with the highest number of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least number of individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution was documented for different forest communities However they differed in the number of DBH classes MF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded in the study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classes and completely lacked trees above 93 cm DBH

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem densityand (c) DBH The horizontal line crossing the box in the center represents the median the box representsthe 25th and the 75th percentiles The vertical line outside the box represents the minimum andmaximum values ALL = total forest community LF = low-altitude forest MF = mid-altitude forestHF = high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure 9) with the highestnumber of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least numberof individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution wasdocumented for different forest communities However they differed in the number of DBH classesMF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded inthe study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classesand completely lacked trees above 93 cm DBH

Forests 2019 10 633 12 of 17

Forests 2019 10 x FOR PEER REVIEW 13 of 18

Figure 9 Class distribution of tress withgt3 cmDBH recorded in the study(a) Overall stems recorded in the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH ranging from gt93 cm to 303 cm LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change of forest tree species recorded in the Himalayan forests like Sikkim The finding could be related to the close association of altitude with climatic variables temperature and precipitation [53] and thus draws attention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communities between 900 masland3200 m asl Moreover according to the Grierson and Long [54]classification our study resembles three forest types ie LF forest as the warm broad-leaved forest MF as the evergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded some differences in species recorded for high-elevation forests For example Grierson and Longrsquos [54] classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall) Boiss as the major trees for high-elevation forests however we did not report similar findings A reason in the change in species composition in these three forest types compared to the 1960s and 1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for its use as scaffolding timber for construction of houses firewood and clearing for forest

Figure 9 Class distribution of tress with gt3 cmDBH recorded in the study (a) Overall stems recordedin the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH rangingfrom gt93 cm to 303 cm LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change offorest tree species recorded in the Himalayan forests like Sikkim The finding could be related to theclose association of altitude with climatic variables temperature and precipitation [53] and thus drawsattention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communitiesbetween 900 masl and 3200 m asl Moreover according to the Grierson and Long [54] classificationour study resembles three forest types ie LF forest as the warm broad-leaved forest MF as theevergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded somedifferences in species recorded for high-elevation forests For example Grierson and Longrsquos [54]classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall)Boiss as the major trees for high-elevation forests however we did not report similar findingsA reason in the change in species composition in these three forest types compared to the 1960s and

Forests 2019 10 633 13 of 17

1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for itsuse as scaffolding timber for construction of houses firewood and clearing for forest for agriculturepractice as well as climate change effects however warrants more investigation Future researchshould focus on documentation of forest trees for a detailed description of the Himalayan forest andrevision of forest type classifications if needed

42 Tree Species Richness and Diversity

Species accumulation curves are especially useful to decide sampling completeness and alsoto compare richness for communities with different sample sizes [50] In the current analysis theaccumulation curves did not record saturation when plotted for different communities which impliesfinding more species with increasing sampling effort Nevertheless it is clear that low-altitude forestsrecord the highest number of species richness and evenness and the disparity between the observedand rarified richness in the MF was the mere result of the sampling size differences The varied speciesrichness between forests covering different altitudes can be associated with numerous factors Broadlywe assumed the widely reported climatic variables namely precipitation temperature and theirinteraction as the prime factor for varied richness along the altitudinal gradient of the Himalayansystem [465556] Another crucial factor could be anthropogenic disturbance and its intensity [2738]Future research could explore these variables to explain the difference in richness recorded in our studyThe LF recorded not only the highest number of species but recorded more uniformly distributed treespecies However the region lacks comparable literature to discuss optimum richness and evennessWe recommend long-term monitoring of permanent plots which is completely lacking in the IndianEastern Himalayas

One of the important findings of our study was the significance of the family Fagaceae It wasrecorded as having the highest FIVI and SIVI and dominated the forest along the altitudinal gradientbetween 900 m and3200 m asl in Sikkim The reasons should be further investigated becauseecological and silvicultural knowledge of Fagaceae and its species are very limited for the EasternHimalayan forests Furthermore we also showed for the first time the existence of huge trees ofLithocarpus pachyphylls (gt15 m DBH) in high number which implies that this species was the remnantof the old growth primary forests in the altitude from 2500 to 3200 in Sikkim Himalayas Theseold-growth trees should be marked registered monitored and conserved as habitat trees by theforest department

However not all forest communities were dominated by the Fagaceae species For example Schimawallichii and Castanopsis tribuloides dominated the lower altitude forests whereas species like Symplocosramosissima and Castanopsis hystrix dominated the mid-altitude forests Different physiological andclimatic adaptation of a species may have caused distinct species to dominate different altituderesulting in three unique clusters of our samples Moreover the Himalayas are renowned for thepronounced variation in climatic topographic and edaphic factors along the altitudinal gradientApart from these the anthropogenic disturbance is a widely discussed factor for the dominance offast-growing pioneer species like the Symplocos racemose [57]Future research could explore the role ofanthropogenic disturbance and different climatic and topographic factors on species dominance andcomposition along the altitudinal gradient

43 Forest Structure Stem Diameter Stem Density and Basal Area

Size class distribution of trees provides the population structure of forest [58] and are extensivelyused to understand regeneration status [59] The present assessment presented a reverse J-shapeddistribution suggesting uneven-aged forests for sustainable reproduction and regeneration [60] with asufficient number of young individuals to replace the old mature stand However between forest typesthe low-altitude forests completely lacked trees above 93 cm diameter a scenario undesirable for asustainable forest The finding could be associated with forest degradation and deforestation reportedby earlier studies [4761] One potential factor for the absence of large trees could be the practice of

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 3 of 17

13 Aims of this Study

In this study we aimed to fill the existing knowledge gap on the diversity structure andcomposition of forest tree species using well-represented samples covering forests along the altitudinalgradient in Sikkim a part of the Eastern Himalayas Our study not only provides an overall assessmenton the diversity composition and structure of forest tree communities but also evaluates how theseattributes vary among forest communities which occur along the altitudinal gradient

2 Materials and Methods

21 Study Area Description and Data Collection

Sikkim (2707prime to 2813prime N and 8801prime to 8892prime E) a northeastern region of India falls within theEastern Himalayan biodiversity hotspot (Figure 1) It has a remarkable altitudinal gradient with only anarea of 7096 km2 ranging from 300 m asl to 8586 m asl with Mount Kanchendzongamdashthe third highestpeak in the world The complex topography and enormous altitude are characterized by 12 majorvegetation types from a tropical warm broad-leaved forest at the lowest altitude to alpine meadowat the highest altitude [47] It has a subtropical to temperate climate with annual rainfall between2700 mm to 3200 mm and the mean annual temperature (MAT) varies from 84 C to 232 C [48]

Forests 2019 10 x FOR PEER REVIEW 3 of 18

In this study we aimed to fill the existing knowledge gap on the diversity structure and composition of forest tree species using well-represented samples covering forests along the altitudinal gradient in Sikkim a part of the Eastern Himalayas Our study not only provides an overall assessment on the diversity composition and structure of forest tree communities but also evaluates how these attributes vary among forest communities which occur along the altitudinal gradient

2 Materials and Methods

21 Study Area Description and Data Collection

Sikkim(27deg07rsquoto 28deg13rsquoN and 88deg01rsquoto 88deg92rsquoE) a northeastern region of India falls within the Eastern Himalayan biodiversity hotspot (Figure 1) It has a remarkable altitudinal gradient with only an area of 7096 km2 ranging from 300 m asl to 8586 m asl with Mount Kanchendzongamdashthe third highest peak in the world The complex topography and enormous altitude are characterized by 12 major vegetation types from a tropical warm broad-leaved forest at the lowest altitude to alpine meadow at the highest altitude [47] It has a subtropical to temperate climate with annual rainfall between 2700 mm to 3200 mm and the mean annual temperature (MAT) varies from 84degC to 232degC [48]

Figure 1 Study area with sampling locations along an altitude gradient in Sikkim India

The field study was conducted over 3 years (2013ndash2015) during the dry winter seasons from the beginning of October to late May to avoid challenges caused by understory growth of thick shrubs and herbs during the monsoon season Stratified random sampling was used to sample plots along the altitudinal gradients from 900 m asl to 3200 m asl The forests were selected because forests within the range has undergone significant loss in forest cover [47] Thus quantitative assessment of these forests is imperative to plan conservation strategies The forests were stratified into three altitudinal zones low-altitude forest (LF) ranging from 900 masl to 1700 m asl the mid-altitude forest (MF) ranging from 1700 m aslto 2500 m asl and the high-altitude forest (HF) ranging from 2500 m asl to 3200 m asl Plots were randomly installed at different locations for each altitude zone Altogether we installed 83 plots of 100 m times 10 m (each 01 ha) covering a total area of 83 ha The number of plots varied across forests and were distributed as LF=24 MF = 37 and HF= 22 plots among three forest communities along the altitudinal gradient Within each plot all living woody trees with gt3 cm diameter at breast height (DBH or trunk diameter at 13 m above ground level) were

Figure 1 Study area with sampling locations along an altitude gradient in Sikkim India

The field study was conducted over 3 years (2013ndash2015) during the dry winter seasons from thebeginning of October to late May to avoid challenges caused by understory growth of thick shrubs andherbs during the monsoon season Stratified random sampling was used to sample plots along thealtitudinal gradients from 900 m asl to 3200 m asl The forests were selected because forests withinthe range has undergone significant loss in forest cover [47] Thus quantitative assessment of theseforests is imperative to plan conservation strategies The forests were stratified into three altitudinalzones low-altitude forest (LF) ranging from 900 m asl to 1700 m asl the mid-altitude forest (MF)ranging from 1700 m asl to 2500 m asl and the high-altitude forest (HF) ranging from 2500 m aslto 3200 m asl Plots were randomly installed at different locations for each altitude zone Altogetherwe installed 83 plots of 100 m times 10 m (each 01 ha) covering a total area of 83 ha The number ofplots varied across forests and were distributed as LF = 24 MF = 37 and HF = 22 plots among threeforest communities along the altitudinal gradient Within each plot all living woody trees with gt3 cmdiameter at breast height (DBH or trunk diameter at 13 m above ground level) were measured andidentified Voucher specimens were collected for individuals challenging to identify in the field andwas used later for further identification

Forests 2019 10 633 4 of 17

22 Data Analysis

The analysis included quantitative data on the abundance DBH taxonomic and geo-coordinatedetails for all sites Prior to the quantification of forest attributes we conducted cluster analysis andordination to examine and validate the presence of ecologically meaningful clusters among our sampledplots designed for the study First we computed the BrayndashCurtis distance matrix using the abundancedata with species in the columns and sites in ascending order of altitude in rows Then we performedhierarchical Wardrsquos minimum variance clustering on the BrayndashCurtis distance matrix This method isbased on the least squares linear model criteria and forms cluster with the least variance [49] Secondlywe performed the Nonmetirc multidimensional scaling method (NMDS) an indirect gradient analysisapproach to ordinate our sampled sites using the same distance matrix Nonmetric multidimensionalscaling represents the pairwise dissimilarity among sites in a low dimensional space as closely aspossible [49] We also estimated the goodness of fit measured as ldquostressrdquo from the Shepard diagram toevaluate the appropriateness of the NMDS result [49]

We plotted the species accumulation curve to determine the sampling effectiveness for the entiresample installed in our study Further we plotted sample-based rarefaction curve for different forestcommunities It computes the average richness by randomly drawing 123n samples representing acommunity without replacement [50] The richness estimated from the sample-based rarefaction curveswas also used to compare the richness for different forest communities with distinct sampling effort

Shannonrsquos diversity index (Hrsquo) was estimated by multiplying the proportion of each species totheir natural log Shannon Index (Hrsquo) = Σpilog(ln)pi where pi is the proportion (nN) of individuals ofa particular species found (n) divided by the total number of individuals recorded (N) ln is the naturallog and Σ is the sum of the calculations

Shannon diversity provides information on both species richness and relative abundance amongplots and thus are sensitive to the sampling effort or the number of individuals sampled To avoidbias we estimated other indices namely the rarefied richness for forest communities and comparedrichness among subsets using a fixed sample size Further we estimated the Fisher alpha a parametricrichness estimate which compares richness among samples with varying individuals Pieloursquos J wasalso estimated to check for differences in evenness among communities with different sample sizeApart from these we estimated observed species richness genus richness and family richness for theoverall forest community and three forest communities along the altitudinal gradient

Additionally we estimated four essential stand structural parameters the median DBH stemdensity (haminus1) basal area (m2 haminus1) and size class distribution of forest trees We mostly usedthe median as a measure of central tendency because variables often did not follow a Gaussiandistribution We calculated the Important Values Index (IVI) both for species(SIVI) and families(FIVI)We followed the equations by Keel et al [51] and Ganesh et al [52] for IVI estimation Diversityanalysis clustering and ordination were conducted using the R package ldquoveganrdquoand ldquoBiodiversityRrdquoV29-2 and the built-in function ldquodecostandrdquo ldquovegdistrdquo ldquohclustrdquo ldquowardD2rdquoand ldquometaNMDSrdquoin RStudio (version 11383)

3 Results

31 Distinct Forest Communities along the Altitude Gradient from 900 m to 3200 m

The cluster analysis revealed that three forest communities could be formed based on thecomposition of tree species at different sites although few plots deflected from its designated categories(Figure 2a) The NMDS result superimposed with a cluster dendrogram also showed similar groups(Figure 2b) The Shepard diagram and the goodness of fit also supported our result with high R2

value (Figure 3) It can be stated that both the cluster analysis and NMDS suggested the presence ofthree ecologically meaningful forest communities along the altitudinal gradient with distinct speciescomposition These three forest communities corresponded to the altitudinal gradient hence reiteratedthe importance of altitude as a covariate to determine the composition of tree species in the Himalayas

Forests 2019 10 633 5 of 17

Based on the Species Importance Value Index and a proportion of stem density and basal area(see Table 3) all forests can be classified as mixed-broadleaf forests but with distinct communities atdifferent altitudes The lower altitude (LF) forests can be classified as the association of Schima wallichiiChoisy Castanopsis tribuloides Smith Engelhardtia spicata Blume and Gynocardia odorata RBr (Figure 4)The mid-altitude (MF) forests can be classified asthe association of Symplocos ramosissima Wall exG Don Castanopsis hystrix Hookf amp Thomson ex ADC Viburnum erubescens Wall and Quercuslamellosa Sm (Figure 5) Whereas high-altitude (HF) forests can be classified as the associationof Lithocarpus pachyphyllus Rhododendron arboreum Sm Magnolia campbellii Hookfamp Thomson andSymplocos heishanensis Hayata (Figure 6)

Forests 2019 10 x FOR PEER REVIEW 5 of 18

distinct species composition These three forest communities corresponded to the altitudinal gradient hence reiterated the importance of altitude as a covariate to determine the composition of tree species in the Himalayas Based on the Species Importance Value Index and a proportion of stem density and basal area (see Table 3)all forests can be classified as mixed-broadleaf forests but with distinct communities at different altitudes The lower altitude (LF) forests can be classified as the association of Schima wallichii ChoisyCastanopsis tribuloides Smith Engelhardtia spicata Blume and Gynocardia odorata RBr(Figure 4) The mid-altitude (MF) forests can be classified asthe association of Symplocos ramosissima Wall ex G DonCastanopsis hystrix Hookf amp Thomson ex ADC Viburnum erubescens Wall and Quercus lamellosa Sm (Figure 5) Whereas high-altitude (HF) forests can be classified as the association of Lithocarpus pachyphyllus Rhododendron arboreum Sm Magnolia campbellii Hookfamp Thomson and Symplocos heishanensis Hayata (Figure 6)

Figure 2 Hierarchically clustered dendrogram and Nonmetic multidimensional scaling (NMDS) ordination on the BrayndashCurtis distance estimated using abundance data (a) Hierarchical clustering was done using Wardrsquos algorithm the y-axis represents the BrayndashCurtis distance shared among each cluster (b) NMDS ordination superimposed with the result of the cluster analysis more similar plots are grouped closer to another The red green and blue dots and lines correspond to the low-altitude forest (LF) mid-altitude forest (MF) and high-altitude forest (HF) forest categories in our study

Figure 2 Hierarchically clustered dendrogram and Nonmetic multidimensional scaling (NMDS)ordination on the BrayndashCurtis distance estimated using abundance data (a) Hierarchical clusteringwas done using Wardrsquos algorithm the y-axis represents the BrayndashCurtis distance shared among eachcluster (b) NMDS ordination superimposed with the result of the cluster analysis more similar plotsare grouped closer to another The red green and blue dots and lines correspond to the low-altitudeforest (LF) mid-altitude forest (MF) and high-altitude forest (HF) forest categories in our studyForests 2019 10 x FOR PEER REVIEW 6 of 18

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure2

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees in flowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis and Schima wallichii trees recorded at 1400 m asl

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure 2

Forests 2019 10 633 6 of 17

Forests 2019 10 x FOR PEER REVIEW 6 of 18

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure2

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees in flowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis and Schima wallichii trees recorded at 1400 m asl

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees inflowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis andSchima wallichii trees recorded at 1400 m aslForests 2019 10 x FOR PEER REVIEW 7 of 18

Figure 5Mid-elevation forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl (b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocos species at 2300 m asl

Figure 6High elevation forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at2860 m asl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 m aslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiern were reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

Figure 5 Mid-altitude forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl(b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocosspecies at 2300 m asl

Forests 2019 10 633 7 of 17

Forests 2019 10 x FOR PEER REVIEW 7 of 18

Figure 5Mid-elevation forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl (b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocos species at 2300 m asl

Figure 6High elevation forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at2860 m asl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 m aslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiern were reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

Figure 6 High altitude forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at 2860 masl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 maslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiernwere reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size)including 87 unidentified individuals with the observed richness of 114 species from 42 families and75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentleslope for the cumulative species richness after the sample size of 70 indicating an adequate number ofplots (Figure 7a) However for different forest communities the species accumulation curve showedan increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest speciesrichness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversityvaried from 171 plusmn 040 to 210 plusmn 035 with the highest value for the LF (Table 1) and the Pieloursquosevenness (J) varied from 046 plusmn 07 to 035 plusmn 010 with the highest value for the LF (Table 1)

Tables 2 and 3 show the results for the Importance Value Index (IVI) estimated for both species(SIVI) and families (FIVI) along with species richness families stem density basal area theirproportion to the contribution for the overall forest community and different forest communitiesWe found Lithocarpus pachyphyllus (930) Symplocos ramosissima (872) Castanopsis hystrix (823)Castanopsis tribuloides (499) and Schima wallichii (446) to be the five most important specieswith the highest Species Importance Value Index (SIVI) Similarly we recorded Fagaceae (2786)Symplocaceae (1226) Theaceae (804) Ericaceae (770) and Lauraceae (662) as the top fivefamilies Lauraceae was also recorded as the most diverse family with 13 species (Table 2)

Forests 2019 10 633 8 of 17

Forests 2019 10 x FOR PEER REVIEW 8 of 18

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size) including 87 unidentified individuals with the observed richness of 114 species from 42 families and 75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentle slope for the cumulative species richness after the sample size of 70 indicating an adequate number of plots (Figure 7a) However for different forest communities the species accumulation curve showed an increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest species richness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20 the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversity varied from 171plusmn040 to 210plusmn035 with the highest value for the LF (Table 1) and the Pieloursquos evenness (J) varied from 046plusmn07 to 035plusmn010 with the highest value for the LF (Table1)

Figure 7Species accumulation curve based on the rarefaction method (a) whole tree population (b)three forests LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities along an altitude gradient The bold figures represent the highest value for each parameter The table provides the median and its standard error since richness showed a right-skewed distribution LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Parameters ALL LF MF HF Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200

Total sampled area (ha) 83 24 37 22 Total species richness 114 68 77 51 Total genus richness 75 52 54 37 Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450

Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351 Median of familiesrsquo richness plusmn

SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281

Total individual counted 10344 2591 5463 2290 Shannon index (H) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fishers alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164

Figure 7 Species accumulation curve based on the rarefaction method (a) whole tree population(b)three forests LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities alongan altitude gradient The table provides the median and its standard error since richness showed aright-skewed distribution LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Parameters ALL LF MF HF

Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200Total sampled area (ha) 83 24 37 22Total species richness 114 68 77 51Total genus richness 75 52 54 37Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351

Median of familiesrsquo richness plusmn SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281Total individual counted 10344 2591 5463 2290Shannon index (Hrsquo) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fisherrsquos alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164Pieloursquos evenness (J) 039 plusmn 039 046 plusmn 07 035 plusmn 010 038 plusmn 013

Dominant family Fagaceae Fagaceae Fagaceae Fagaceae

Dominant species Lithocarpuspachyphyllus Schima wallichii Symplocos

ramosissimaLithocarpus

pachyphyllus

Species such as Schima wallichii (1890) Castanopsis tribuloides (1791) Engelhardtia spicata (672)Gynocardia odorata (605) and Eurya acuminata (483) were recorded as the five most dominant specieswith higher species Important Value Index (SIVI) for LF Similarly Symplocos ramosissima (1509)Castanopsis hystrix (1499) Viburnum erubescens (709) Quercus lamellosa (685) and Eurya acuminata(442) were reported as the most dominant species for the MF The HF was dominated by speciessuch as Lithocarpus pachyphyllus (2437) Rhododendron arboreum (1678) Magnolia campbellii (454)Symplocos heishanensis (414) and Symplocos ramosissima (396) with the highest SIVI (Table 3)

Forests 2019 10 633 9 of 17

Table 2 Species richness density basal area Species Importance Value Index (SIVI) FamiliesImportance Value Index (FIVI) and their proportion to the contribution for the top ten species andfamilies for the overall forest community recorded in the study

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

Lithocarpus pachyphyllus Fagaceae 5120 411 1231 2107 930Symplocos ramosissima Symplocaceae 24807 1919 161 276 872

Castanopsis hystrix Fagaceae 6205 498 991 1697 823Castanopsis tribuloides Fagaceae 6241 501 422 723 499

Schima wallichii Theaceae 7819 627 240 411 446Rhododendron arboreum Ericaceae 8554 686 254 435 433

Viburnum erubescens Adoxaceae 10398 834 038 064 416Eurya acuminata DC Pentaphylacaceae 6096 489 081 139 391

Quercus lamellosa Fagaceae 988 079 495 849 386Symplocos heishanensis Symplocaceae 4940 396 038 065 293

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 9 2088 1675 3361 5719 2786Symplocaceae 4 3355 2692 218 371 1226

Theaceae 2 1390 1116 320 545 804Ericaceae 7 1476 1184 381 648 770Lauraceae 13 463 371 497 845 662Adoxaceae 2 1041 835 038 064 464

Magnoliaceae 4 112 090 209 364 270Betulaceae 4 193 155 099 169 219

Primulaceae 3 228 183 041 023 210Sapindaceae 1 069 055 123 210 189

Table 3 Stem density basal area Family (FIVI) and Species Importance Value Indexes (SIVI) and theirproportion to the contribution of the ten dominant species for three forest communities LF MF and HF

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

LF (900 m to 1700 m asl)

Schima wallichii Theaceae 25917 2401 708 2673 1890Castanopsis tribuloides Fagaceae 20250 1876 768 2900 1791

Engelhardtia spicata Juglandaceae 6042 560 235 886 672Gynocardia odorata Achariaceae 9250 857 123 463 605Euryaa cuminata Pentaphylacaceae 6667 857 076 286 483Betula alnoides

Buch-HamexDDon Betulaceae 2125 618 099 375 257

Macaranga denticulataBlume Euphorbiaceae 1808 197 039 148 243

Albizia lebbeck (L)Benth Leguminosae 1708 158 123 464 231

Alnus nepalensis DDon Betulaceae 1208 066 100 377 221Brassaiopsis hispida

Seem Araliaceae 3625 336 019 053 187

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Theaceae 2 942 3018 784 2960 2254Fagaceae 6 673 2157 823 3109 2017

Juglandaceae 3 180 575 238 899 742Achariaceae 1 267 857 123 463 658Betulaceae 3 104 332 209 789 526

Euphorbiaceae 5 123 394 059 222 434Araliaceae 3 157 502 017 066 353

Leguminosae 3 030 096 127 480 290Anacardiaceae 3 059 189 048 181 276

Lauraceae 6 063 201 009 035 264

Forests 2019 10 633 10 of 17

Table 3 Cont

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

MF (gt1700 m to 2500 m asl)

Symplocos ramosissima Symplocaceae 50189 3399 336 467 1509Castanopsis hystrix Fagaceae 13243 897 2172 3015 1499

Viburnum erubescens Adoxaceae 21189 1435 075 105 709Quercus lamellosa Fagaceae 2054 139 1067 1481 685Eurya acuminata Pentaphylacaceae 8378 567 123 171 442

Lithocarpus pachyphyllus Fagaceae 4892 331 517 718 432Symplocos heishanensis Symplocaceae 7378 500 056 078 407Castanopsis tribuloides Fagaceae 865 059 449 623 278

Quercus glauca Fagaceae 1432 097 309 429 270Elaeocarpus sikkimensis

Mast Elaeocarpaceae 1054 071 221 307 252

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 7 1082 1644 4570 4570 3012Symplocaceae 4 2890 4391 433 433 2014

Lauraceae 11 308 469 960 960 922Adoxaceae 2 946 1437 075 075 807Theaceae 2 395 600 171 171 573

Elaeocarpaceae 2 048 073 223 223 317Ericaceae 5 249 379 049 049 272

Sapindaceae 1 040 060 150 150 260Primulaceae 3 124 189 016 016 260

Magnoliaceae 4 059 090 175 175 243

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

HF (2500 m to 3200 m asl)

Lithocarpus pachyphyllus Fagaceae 11091 1066 788 5456 2437Rhododendron arboreum Ericaceae 31000 2978 747 1308 1678

Magnolia campbellii Magnoliaceae 1682 162 539 662 454Symplocos heishanensis Symplocaceae 6182 594 581 069 414Rhododendron hodgsonii

Hookf Ericaceae 5455 524 290 315 376

Corylus ferox Wall Betulaceae 2682 258 415 133 268Viburnum erubescens

Wall Adoxaceae 3591 345 415 021 260

Acer campbellii Sapindaceae 1091 105 373 298 259Rhododendron falconeri

Hookf Ericaceae 3955 380 207 186 258

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 4 333 9011 4096 5716 7921Ericaceae 7 1182 2963 1347 1879 7216

Symplocaceae 3 464 207 094 131 2602Magnoliaceae 2 048 1090 495 691 1603

Lauraceae 6 092 551 251 350 1524Betulaceae 2 076 216 098 137 991Theaceae 2 053 146 067 093 969

Berberidaceae 2 139 037 017 024 947Adoxaceae 1 095 033 015 021 992

Sapindaceae 1 029 470 214 298 877

33 Forest Structure

Overall we recorded the median stem DBH of 1542 plusmn 111 cm stem density of124627 plusmn 56828 haminus1 and the basal area of 5435 plusmn 444 m2 haminus1 (Figure 8) Among forest communitiesthe HF recorded the highest median stem DBH (2089 plusmn 309 cm) but showed the lowest stem densitySimilarly MF recorded the highest stem density of about 1320 plusmn 11972 haminus1 with the second highest

Forests 2019 10 633 11 of 17

value for the stem DBH and basal area However the LF recorded the second highest value for medianstem density and the minimum value for the median DBH and basal areaForests 2019 10 x FOR PEER REVIEW 12 of 18

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem density and (c) DBH The horizontal line crossing the box in the center represents the median the box represents the 25th and the 75th percentiles The vertical line outside the box represents the minimum and maximum values ALL= total forest community LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure9) with the highest number of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least number of individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution was documented for different forest communities However they differed in the number of DBH classes MF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded in the study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classes and completely lacked trees above 93 cm DBH

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem densityand (c) DBH The horizontal line crossing the box in the center represents the median the box representsthe 25th and the 75th percentiles The vertical line outside the box represents the minimum andmaximum values ALL = total forest community LF = low-altitude forest MF = mid-altitude forestHF = high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure 9) with the highestnumber of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least numberof individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution wasdocumented for different forest communities However they differed in the number of DBH classesMF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded inthe study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classesand completely lacked trees above 93 cm DBH

Forests 2019 10 633 12 of 17

Forests 2019 10 x FOR PEER REVIEW 13 of 18

Figure 9 Class distribution of tress withgt3 cmDBH recorded in the study(a) Overall stems recorded in the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH ranging from gt93 cm to 303 cm LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change of forest tree species recorded in the Himalayan forests like Sikkim The finding could be related to the close association of altitude with climatic variables temperature and precipitation [53] and thus draws attention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communities between 900 masland3200 m asl Moreover according to the Grierson and Long [54]classification our study resembles three forest types ie LF forest as the warm broad-leaved forest MF as the evergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded some differences in species recorded for high-elevation forests For example Grierson and Longrsquos [54] classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall) Boiss as the major trees for high-elevation forests however we did not report similar findings A reason in the change in species composition in these three forest types compared to the 1960s and 1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for its use as scaffolding timber for construction of houses firewood and clearing for forest

Figure 9 Class distribution of tress with gt3 cmDBH recorded in the study (a) Overall stems recordedin the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH rangingfrom gt93 cm to 303 cm LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change offorest tree species recorded in the Himalayan forests like Sikkim The finding could be related to theclose association of altitude with climatic variables temperature and precipitation [53] and thus drawsattention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communitiesbetween 900 masl and 3200 m asl Moreover according to the Grierson and Long [54] classificationour study resembles three forest types ie LF forest as the warm broad-leaved forest MF as theevergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded somedifferences in species recorded for high-elevation forests For example Grierson and Longrsquos [54]classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall)Boiss as the major trees for high-elevation forests however we did not report similar findingsA reason in the change in species composition in these three forest types compared to the 1960s and

Forests 2019 10 633 13 of 17

1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for itsuse as scaffolding timber for construction of houses firewood and clearing for forest for agriculturepractice as well as climate change effects however warrants more investigation Future researchshould focus on documentation of forest trees for a detailed description of the Himalayan forest andrevision of forest type classifications if needed

42 Tree Species Richness and Diversity

Species accumulation curves are especially useful to decide sampling completeness and alsoto compare richness for communities with different sample sizes [50] In the current analysis theaccumulation curves did not record saturation when plotted for different communities which impliesfinding more species with increasing sampling effort Nevertheless it is clear that low-altitude forestsrecord the highest number of species richness and evenness and the disparity between the observedand rarified richness in the MF was the mere result of the sampling size differences The varied speciesrichness between forests covering different altitudes can be associated with numerous factors Broadlywe assumed the widely reported climatic variables namely precipitation temperature and theirinteraction as the prime factor for varied richness along the altitudinal gradient of the Himalayansystem [465556] Another crucial factor could be anthropogenic disturbance and its intensity [2738]Future research could explore these variables to explain the difference in richness recorded in our studyThe LF recorded not only the highest number of species but recorded more uniformly distributed treespecies However the region lacks comparable literature to discuss optimum richness and evennessWe recommend long-term monitoring of permanent plots which is completely lacking in the IndianEastern Himalayas

One of the important findings of our study was the significance of the family Fagaceae It wasrecorded as having the highest FIVI and SIVI and dominated the forest along the altitudinal gradientbetween 900 m and3200 m asl in Sikkim The reasons should be further investigated becauseecological and silvicultural knowledge of Fagaceae and its species are very limited for the EasternHimalayan forests Furthermore we also showed for the first time the existence of huge trees ofLithocarpus pachyphylls (gt15 m DBH) in high number which implies that this species was the remnantof the old growth primary forests in the altitude from 2500 to 3200 in Sikkim Himalayas Theseold-growth trees should be marked registered monitored and conserved as habitat trees by theforest department

However not all forest communities were dominated by the Fagaceae species For example Schimawallichii and Castanopsis tribuloides dominated the lower altitude forests whereas species like Symplocosramosissima and Castanopsis hystrix dominated the mid-altitude forests Different physiological andclimatic adaptation of a species may have caused distinct species to dominate different altituderesulting in three unique clusters of our samples Moreover the Himalayas are renowned for thepronounced variation in climatic topographic and edaphic factors along the altitudinal gradientApart from these the anthropogenic disturbance is a widely discussed factor for the dominance offast-growing pioneer species like the Symplocos racemose [57]Future research could explore the role ofanthropogenic disturbance and different climatic and topographic factors on species dominance andcomposition along the altitudinal gradient

43 Forest Structure Stem Diameter Stem Density and Basal Area

Size class distribution of trees provides the population structure of forest [58] and are extensivelyused to understand regeneration status [59] The present assessment presented a reverse J-shapeddistribution suggesting uneven-aged forests for sustainable reproduction and regeneration [60] with asufficient number of young individuals to replace the old mature stand However between forest typesthe low-altitude forests completely lacked trees above 93 cm diameter a scenario undesirable for asustainable forest The finding could be associated with forest degradation and deforestation reportedby earlier studies [4761] One potential factor for the absence of large trees could be the practice of

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 4 of 17

22 Data Analysis

The analysis included quantitative data on the abundance DBH taxonomic and geo-coordinatedetails for all sites Prior to the quantification of forest attributes we conducted cluster analysis andordination to examine and validate the presence of ecologically meaningful clusters among our sampledplots designed for the study First we computed the BrayndashCurtis distance matrix using the abundancedata with species in the columns and sites in ascending order of altitude in rows Then we performedhierarchical Wardrsquos minimum variance clustering on the BrayndashCurtis distance matrix This method isbased on the least squares linear model criteria and forms cluster with the least variance [49] Secondlywe performed the Nonmetirc multidimensional scaling method (NMDS) an indirect gradient analysisapproach to ordinate our sampled sites using the same distance matrix Nonmetric multidimensionalscaling represents the pairwise dissimilarity among sites in a low dimensional space as closely aspossible [49] We also estimated the goodness of fit measured as ldquostressrdquo from the Shepard diagram toevaluate the appropriateness of the NMDS result [49]

We plotted the species accumulation curve to determine the sampling effectiveness for the entiresample installed in our study Further we plotted sample-based rarefaction curve for different forestcommunities It computes the average richness by randomly drawing 123n samples representing acommunity without replacement [50] The richness estimated from the sample-based rarefaction curveswas also used to compare the richness for different forest communities with distinct sampling effort

Shannonrsquos diversity index (Hrsquo) was estimated by multiplying the proportion of each species totheir natural log Shannon Index (Hrsquo) = Σpilog(ln)pi where pi is the proportion (nN) of individuals ofa particular species found (n) divided by the total number of individuals recorded (N) ln is the naturallog and Σ is the sum of the calculations

Shannon diversity provides information on both species richness and relative abundance amongplots and thus are sensitive to the sampling effort or the number of individuals sampled To avoidbias we estimated other indices namely the rarefied richness for forest communities and comparedrichness among subsets using a fixed sample size Further we estimated the Fisher alpha a parametricrichness estimate which compares richness among samples with varying individuals Pieloursquos J wasalso estimated to check for differences in evenness among communities with different sample sizeApart from these we estimated observed species richness genus richness and family richness for theoverall forest community and three forest communities along the altitudinal gradient

Additionally we estimated four essential stand structural parameters the median DBH stemdensity (haminus1) basal area (m2 haminus1) and size class distribution of forest trees We mostly usedthe median as a measure of central tendency because variables often did not follow a Gaussiandistribution We calculated the Important Values Index (IVI) both for species(SIVI) and families(FIVI)We followed the equations by Keel et al [51] and Ganesh et al [52] for IVI estimation Diversityanalysis clustering and ordination were conducted using the R package ldquoveganrdquoand ldquoBiodiversityRrdquoV29-2 and the built-in function ldquodecostandrdquo ldquovegdistrdquo ldquohclustrdquo ldquowardD2rdquoand ldquometaNMDSrdquoin RStudio (version 11383)

3 Results

31 Distinct Forest Communities along the Altitude Gradient from 900 m to 3200 m

The cluster analysis revealed that three forest communities could be formed based on thecomposition of tree species at different sites although few plots deflected from its designated categories(Figure 2a) The NMDS result superimposed with a cluster dendrogram also showed similar groups(Figure 2b) The Shepard diagram and the goodness of fit also supported our result with high R2

value (Figure 3) It can be stated that both the cluster analysis and NMDS suggested the presence ofthree ecologically meaningful forest communities along the altitudinal gradient with distinct speciescomposition These three forest communities corresponded to the altitudinal gradient hence reiteratedthe importance of altitude as a covariate to determine the composition of tree species in the Himalayas

Forests 2019 10 633 5 of 17

Based on the Species Importance Value Index and a proportion of stem density and basal area(see Table 3) all forests can be classified as mixed-broadleaf forests but with distinct communities atdifferent altitudes The lower altitude (LF) forests can be classified as the association of Schima wallichiiChoisy Castanopsis tribuloides Smith Engelhardtia spicata Blume and Gynocardia odorata RBr (Figure 4)The mid-altitude (MF) forests can be classified asthe association of Symplocos ramosissima Wall exG Don Castanopsis hystrix Hookf amp Thomson ex ADC Viburnum erubescens Wall and Quercuslamellosa Sm (Figure 5) Whereas high-altitude (HF) forests can be classified as the associationof Lithocarpus pachyphyllus Rhododendron arboreum Sm Magnolia campbellii Hookfamp Thomson andSymplocos heishanensis Hayata (Figure 6)

Forests 2019 10 x FOR PEER REVIEW 5 of 18

distinct species composition These three forest communities corresponded to the altitudinal gradient hence reiterated the importance of altitude as a covariate to determine the composition of tree species in the Himalayas Based on the Species Importance Value Index and a proportion of stem density and basal area (see Table 3)all forests can be classified as mixed-broadleaf forests but with distinct communities at different altitudes The lower altitude (LF) forests can be classified as the association of Schima wallichii ChoisyCastanopsis tribuloides Smith Engelhardtia spicata Blume and Gynocardia odorata RBr(Figure 4) The mid-altitude (MF) forests can be classified asthe association of Symplocos ramosissima Wall ex G DonCastanopsis hystrix Hookf amp Thomson ex ADC Viburnum erubescens Wall and Quercus lamellosa Sm (Figure 5) Whereas high-altitude (HF) forests can be classified as the association of Lithocarpus pachyphyllus Rhododendron arboreum Sm Magnolia campbellii Hookfamp Thomson and Symplocos heishanensis Hayata (Figure 6)

Figure 2 Hierarchically clustered dendrogram and Nonmetic multidimensional scaling (NMDS) ordination on the BrayndashCurtis distance estimated using abundance data (a) Hierarchical clustering was done using Wardrsquos algorithm the y-axis represents the BrayndashCurtis distance shared among each cluster (b) NMDS ordination superimposed with the result of the cluster analysis more similar plots are grouped closer to another The red green and blue dots and lines correspond to the low-altitude forest (LF) mid-altitude forest (MF) and high-altitude forest (HF) forest categories in our study

Figure 2 Hierarchically clustered dendrogram and Nonmetic multidimensional scaling (NMDS)ordination on the BrayndashCurtis distance estimated using abundance data (a) Hierarchical clusteringwas done using Wardrsquos algorithm the y-axis represents the BrayndashCurtis distance shared among eachcluster (b) NMDS ordination superimposed with the result of the cluster analysis more similar plotsare grouped closer to another The red green and blue dots and lines correspond to the low-altitudeforest (LF) mid-altitude forest (MF) and high-altitude forest (HF) forest categories in our studyForests 2019 10 x FOR PEER REVIEW 6 of 18

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure2

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees in flowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis and Schima wallichii trees recorded at 1400 m asl

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure 2

Forests 2019 10 633 6 of 17

Forests 2019 10 x FOR PEER REVIEW 6 of 18

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure2

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees in flowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis and Schima wallichii trees recorded at 1400 m asl

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees inflowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis andSchima wallichii trees recorded at 1400 m aslForests 2019 10 x FOR PEER REVIEW 7 of 18

Figure 5Mid-elevation forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl (b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocos species at 2300 m asl

Figure 6High elevation forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at2860 m asl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 m aslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiern were reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

Figure 5 Mid-altitude forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl(b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocosspecies at 2300 m asl

Forests 2019 10 633 7 of 17

Forests 2019 10 x FOR PEER REVIEW 7 of 18

Figure 5Mid-elevation forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl (b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocos species at 2300 m asl

Figure 6High elevation forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at2860 m asl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 m aslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiern were reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

Figure 6 High altitude forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at 2860 masl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 maslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiernwere reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size)including 87 unidentified individuals with the observed richness of 114 species from 42 families and75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentleslope for the cumulative species richness after the sample size of 70 indicating an adequate number ofplots (Figure 7a) However for different forest communities the species accumulation curve showedan increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest speciesrichness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversityvaried from 171 plusmn 040 to 210 plusmn 035 with the highest value for the LF (Table 1) and the Pieloursquosevenness (J) varied from 046 plusmn 07 to 035 plusmn 010 with the highest value for the LF (Table 1)

Tables 2 and 3 show the results for the Importance Value Index (IVI) estimated for both species(SIVI) and families (FIVI) along with species richness families stem density basal area theirproportion to the contribution for the overall forest community and different forest communitiesWe found Lithocarpus pachyphyllus (930) Symplocos ramosissima (872) Castanopsis hystrix (823)Castanopsis tribuloides (499) and Schima wallichii (446) to be the five most important specieswith the highest Species Importance Value Index (SIVI) Similarly we recorded Fagaceae (2786)Symplocaceae (1226) Theaceae (804) Ericaceae (770) and Lauraceae (662) as the top fivefamilies Lauraceae was also recorded as the most diverse family with 13 species (Table 2)

Forests 2019 10 633 8 of 17

Forests 2019 10 x FOR PEER REVIEW 8 of 18

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size) including 87 unidentified individuals with the observed richness of 114 species from 42 families and 75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentle slope for the cumulative species richness after the sample size of 70 indicating an adequate number of plots (Figure 7a) However for different forest communities the species accumulation curve showed an increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest species richness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20 the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversity varied from 171plusmn040 to 210plusmn035 with the highest value for the LF (Table 1) and the Pieloursquos evenness (J) varied from 046plusmn07 to 035plusmn010 with the highest value for the LF (Table1)

Figure 7Species accumulation curve based on the rarefaction method (a) whole tree population (b)three forests LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities along an altitude gradient The bold figures represent the highest value for each parameter The table provides the median and its standard error since richness showed a right-skewed distribution LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Parameters ALL LF MF HF Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200

Total sampled area (ha) 83 24 37 22 Total species richness 114 68 77 51 Total genus richness 75 52 54 37 Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450

Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351 Median of familiesrsquo richness plusmn

SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281

Total individual counted 10344 2591 5463 2290 Shannon index (H) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fishers alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164

Figure 7 Species accumulation curve based on the rarefaction method (a) whole tree population(b)three forests LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities alongan altitude gradient The table provides the median and its standard error since richness showed aright-skewed distribution LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Parameters ALL LF MF HF

Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200Total sampled area (ha) 83 24 37 22Total species richness 114 68 77 51Total genus richness 75 52 54 37Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351

Median of familiesrsquo richness plusmn SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281Total individual counted 10344 2591 5463 2290Shannon index (Hrsquo) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fisherrsquos alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164Pieloursquos evenness (J) 039 plusmn 039 046 plusmn 07 035 plusmn 010 038 plusmn 013

Dominant family Fagaceae Fagaceae Fagaceae Fagaceae

Dominant species Lithocarpuspachyphyllus Schima wallichii Symplocos

ramosissimaLithocarpus

pachyphyllus

Species such as Schima wallichii (1890) Castanopsis tribuloides (1791) Engelhardtia spicata (672)Gynocardia odorata (605) and Eurya acuminata (483) were recorded as the five most dominant specieswith higher species Important Value Index (SIVI) for LF Similarly Symplocos ramosissima (1509)Castanopsis hystrix (1499) Viburnum erubescens (709) Quercus lamellosa (685) and Eurya acuminata(442) were reported as the most dominant species for the MF The HF was dominated by speciessuch as Lithocarpus pachyphyllus (2437) Rhododendron arboreum (1678) Magnolia campbellii (454)Symplocos heishanensis (414) and Symplocos ramosissima (396) with the highest SIVI (Table 3)

Forests 2019 10 633 9 of 17

Table 2 Species richness density basal area Species Importance Value Index (SIVI) FamiliesImportance Value Index (FIVI) and their proportion to the contribution for the top ten species andfamilies for the overall forest community recorded in the study

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

Lithocarpus pachyphyllus Fagaceae 5120 411 1231 2107 930Symplocos ramosissima Symplocaceae 24807 1919 161 276 872

Castanopsis hystrix Fagaceae 6205 498 991 1697 823Castanopsis tribuloides Fagaceae 6241 501 422 723 499

Schima wallichii Theaceae 7819 627 240 411 446Rhododendron arboreum Ericaceae 8554 686 254 435 433

Viburnum erubescens Adoxaceae 10398 834 038 064 416Eurya acuminata DC Pentaphylacaceae 6096 489 081 139 391

Quercus lamellosa Fagaceae 988 079 495 849 386Symplocos heishanensis Symplocaceae 4940 396 038 065 293

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 9 2088 1675 3361 5719 2786Symplocaceae 4 3355 2692 218 371 1226

Theaceae 2 1390 1116 320 545 804Ericaceae 7 1476 1184 381 648 770Lauraceae 13 463 371 497 845 662Adoxaceae 2 1041 835 038 064 464

Magnoliaceae 4 112 090 209 364 270Betulaceae 4 193 155 099 169 219

Primulaceae 3 228 183 041 023 210Sapindaceae 1 069 055 123 210 189

Table 3 Stem density basal area Family (FIVI) and Species Importance Value Indexes (SIVI) and theirproportion to the contribution of the ten dominant species for three forest communities LF MF and HF

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

LF (900 m to 1700 m asl)

Schima wallichii Theaceae 25917 2401 708 2673 1890Castanopsis tribuloides Fagaceae 20250 1876 768 2900 1791

Engelhardtia spicata Juglandaceae 6042 560 235 886 672Gynocardia odorata Achariaceae 9250 857 123 463 605Euryaa cuminata Pentaphylacaceae 6667 857 076 286 483Betula alnoides

Buch-HamexDDon Betulaceae 2125 618 099 375 257

Macaranga denticulataBlume Euphorbiaceae 1808 197 039 148 243

Albizia lebbeck (L)Benth Leguminosae 1708 158 123 464 231

Alnus nepalensis DDon Betulaceae 1208 066 100 377 221Brassaiopsis hispida

Seem Araliaceae 3625 336 019 053 187

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Theaceae 2 942 3018 784 2960 2254Fagaceae 6 673 2157 823 3109 2017

Juglandaceae 3 180 575 238 899 742Achariaceae 1 267 857 123 463 658Betulaceae 3 104 332 209 789 526

Euphorbiaceae 5 123 394 059 222 434Araliaceae 3 157 502 017 066 353

Leguminosae 3 030 096 127 480 290Anacardiaceae 3 059 189 048 181 276

Lauraceae 6 063 201 009 035 264

Forests 2019 10 633 10 of 17

Table 3 Cont

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

MF (gt1700 m to 2500 m asl)

Symplocos ramosissima Symplocaceae 50189 3399 336 467 1509Castanopsis hystrix Fagaceae 13243 897 2172 3015 1499

Viburnum erubescens Adoxaceae 21189 1435 075 105 709Quercus lamellosa Fagaceae 2054 139 1067 1481 685Eurya acuminata Pentaphylacaceae 8378 567 123 171 442

Lithocarpus pachyphyllus Fagaceae 4892 331 517 718 432Symplocos heishanensis Symplocaceae 7378 500 056 078 407Castanopsis tribuloides Fagaceae 865 059 449 623 278

Quercus glauca Fagaceae 1432 097 309 429 270Elaeocarpus sikkimensis

Mast Elaeocarpaceae 1054 071 221 307 252

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 7 1082 1644 4570 4570 3012Symplocaceae 4 2890 4391 433 433 2014

Lauraceae 11 308 469 960 960 922Adoxaceae 2 946 1437 075 075 807Theaceae 2 395 600 171 171 573

Elaeocarpaceae 2 048 073 223 223 317Ericaceae 5 249 379 049 049 272

Sapindaceae 1 040 060 150 150 260Primulaceae 3 124 189 016 016 260

Magnoliaceae 4 059 090 175 175 243

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

HF (2500 m to 3200 m asl)

Lithocarpus pachyphyllus Fagaceae 11091 1066 788 5456 2437Rhododendron arboreum Ericaceae 31000 2978 747 1308 1678

Magnolia campbellii Magnoliaceae 1682 162 539 662 454Symplocos heishanensis Symplocaceae 6182 594 581 069 414Rhododendron hodgsonii

Hookf Ericaceae 5455 524 290 315 376

Corylus ferox Wall Betulaceae 2682 258 415 133 268Viburnum erubescens

Wall Adoxaceae 3591 345 415 021 260

Acer campbellii Sapindaceae 1091 105 373 298 259Rhododendron falconeri

Hookf Ericaceae 3955 380 207 186 258

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 4 333 9011 4096 5716 7921Ericaceae 7 1182 2963 1347 1879 7216

Symplocaceae 3 464 207 094 131 2602Magnoliaceae 2 048 1090 495 691 1603

Lauraceae 6 092 551 251 350 1524Betulaceae 2 076 216 098 137 991Theaceae 2 053 146 067 093 969

Berberidaceae 2 139 037 017 024 947Adoxaceae 1 095 033 015 021 992

Sapindaceae 1 029 470 214 298 877

33 Forest Structure

Overall we recorded the median stem DBH of 1542 plusmn 111 cm stem density of124627 plusmn 56828 haminus1 and the basal area of 5435 plusmn 444 m2 haminus1 (Figure 8) Among forest communitiesthe HF recorded the highest median stem DBH (2089 plusmn 309 cm) but showed the lowest stem densitySimilarly MF recorded the highest stem density of about 1320 plusmn 11972 haminus1 with the second highest

Forests 2019 10 633 11 of 17

value for the stem DBH and basal area However the LF recorded the second highest value for medianstem density and the minimum value for the median DBH and basal areaForests 2019 10 x FOR PEER REVIEW 12 of 18

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem density and (c) DBH The horizontal line crossing the box in the center represents the median the box represents the 25th and the 75th percentiles The vertical line outside the box represents the minimum and maximum values ALL= total forest community LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure9) with the highest number of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least number of individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution was documented for different forest communities However they differed in the number of DBH classes MF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded in the study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classes and completely lacked trees above 93 cm DBH

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem densityand (c) DBH The horizontal line crossing the box in the center represents the median the box representsthe 25th and the 75th percentiles The vertical line outside the box represents the minimum andmaximum values ALL = total forest community LF = low-altitude forest MF = mid-altitude forestHF = high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure 9) with the highestnumber of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least numberof individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution wasdocumented for different forest communities However they differed in the number of DBH classesMF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded inthe study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classesand completely lacked trees above 93 cm DBH

Forests 2019 10 633 12 of 17

Forests 2019 10 x FOR PEER REVIEW 13 of 18

Figure 9 Class distribution of tress withgt3 cmDBH recorded in the study(a) Overall stems recorded in the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH ranging from gt93 cm to 303 cm LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change of forest tree species recorded in the Himalayan forests like Sikkim The finding could be related to the close association of altitude with climatic variables temperature and precipitation [53] and thus draws attention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communities between 900 masland3200 m asl Moreover according to the Grierson and Long [54]classification our study resembles three forest types ie LF forest as the warm broad-leaved forest MF as the evergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded some differences in species recorded for high-elevation forests For example Grierson and Longrsquos [54] classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall) Boiss as the major trees for high-elevation forests however we did not report similar findings A reason in the change in species composition in these three forest types compared to the 1960s and 1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for its use as scaffolding timber for construction of houses firewood and clearing for forest

Figure 9 Class distribution of tress with gt3 cmDBH recorded in the study (a) Overall stems recordedin the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH rangingfrom gt93 cm to 303 cm LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change offorest tree species recorded in the Himalayan forests like Sikkim The finding could be related to theclose association of altitude with climatic variables temperature and precipitation [53] and thus drawsattention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communitiesbetween 900 masl and 3200 m asl Moreover according to the Grierson and Long [54] classificationour study resembles three forest types ie LF forest as the warm broad-leaved forest MF as theevergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded somedifferences in species recorded for high-elevation forests For example Grierson and Longrsquos [54]classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall)Boiss as the major trees for high-elevation forests however we did not report similar findingsA reason in the change in species composition in these three forest types compared to the 1960s and

Forests 2019 10 633 13 of 17

1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for itsuse as scaffolding timber for construction of houses firewood and clearing for forest for agriculturepractice as well as climate change effects however warrants more investigation Future researchshould focus on documentation of forest trees for a detailed description of the Himalayan forest andrevision of forest type classifications if needed

42 Tree Species Richness and Diversity

Species accumulation curves are especially useful to decide sampling completeness and alsoto compare richness for communities with different sample sizes [50] In the current analysis theaccumulation curves did not record saturation when plotted for different communities which impliesfinding more species with increasing sampling effort Nevertheless it is clear that low-altitude forestsrecord the highest number of species richness and evenness and the disparity between the observedand rarified richness in the MF was the mere result of the sampling size differences The varied speciesrichness between forests covering different altitudes can be associated with numerous factors Broadlywe assumed the widely reported climatic variables namely precipitation temperature and theirinteraction as the prime factor for varied richness along the altitudinal gradient of the Himalayansystem [465556] Another crucial factor could be anthropogenic disturbance and its intensity [2738]Future research could explore these variables to explain the difference in richness recorded in our studyThe LF recorded not only the highest number of species but recorded more uniformly distributed treespecies However the region lacks comparable literature to discuss optimum richness and evennessWe recommend long-term monitoring of permanent plots which is completely lacking in the IndianEastern Himalayas

One of the important findings of our study was the significance of the family Fagaceae It wasrecorded as having the highest FIVI and SIVI and dominated the forest along the altitudinal gradientbetween 900 m and3200 m asl in Sikkim The reasons should be further investigated becauseecological and silvicultural knowledge of Fagaceae and its species are very limited for the EasternHimalayan forests Furthermore we also showed for the first time the existence of huge trees ofLithocarpus pachyphylls (gt15 m DBH) in high number which implies that this species was the remnantof the old growth primary forests in the altitude from 2500 to 3200 in Sikkim Himalayas Theseold-growth trees should be marked registered monitored and conserved as habitat trees by theforest department

However not all forest communities were dominated by the Fagaceae species For example Schimawallichii and Castanopsis tribuloides dominated the lower altitude forests whereas species like Symplocosramosissima and Castanopsis hystrix dominated the mid-altitude forests Different physiological andclimatic adaptation of a species may have caused distinct species to dominate different altituderesulting in three unique clusters of our samples Moreover the Himalayas are renowned for thepronounced variation in climatic topographic and edaphic factors along the altitudinal gradientApart from these the anthropogenic disturbance is a widely discussed factor for the dominance offast-growing pioneer species like the Symplocos racemose [57]Future research could explore the role ofanthropogenic disturbance and different climatic and topographic factors on species dominance andcomposition along the altitudinal gradient

43 Forest Structure Stem Diameter Stem Density and Basal Area

Size class distribution of trees provides the population structure of forest [58] and are extensivelyused to understand regeneration status [59] The present assessment presented a reverse J-shapeddistribution suggesting uneven-aged forests for sustainable reproduction and regeneration [60] with asufficient number of young individuals to replace the old mature stand However between forest typesthe low-altitude forests completely lacked trees above 93 cm diameter a scenario undesirable for asustainable forest The finding could be associated with forest degradation and deforestation reportedby earlier studies [4761] One potential factor for the absence of large trees could be the practice of

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 5 of 17

Based on the Species Importance Value Index and a proportion of stem density and basal area(see Table 3) all forests can be classified as mixed-broadleaf forests but with distinct communities atdifferent altitudes The lower altitude (LF) forests can be classified as the association of Schima wallichiiChoisy Castanopsis tribuloides Smith Engelhardtia spicata Blume and Gynocardia odorata RBr (Figure 4)The mid-altitude (MF) forests can be classified asthe association of Symplocos ramosissima Wall exG Don Castanopsis hystrix Hookf amp Thomson ex ADC Viburnum erubescens Wall and Quercuslamellosa Sm (Figure 5) Whereas high-altitude (HF) forests can be classified as the associationof Lithocarpus pachyphyllus Rhododendron arboreum Sm Magnolia campbellii Hookfamp Thomson andSymplocos heishanensis Hayata (Figure 6)

Forests 2019 10 x FOR PEER REVIEW 5 of 18

distinct species composition These three forest communities corresponded to the altitudinal gradient hence reiterated the importance of altitude as a covariate to determine the composition of tree species in the Himalayas Based on the Species Importance Value Index and a proportion of stem density and basal area (see Table 3)all forests can be classified as mixed-broadleaf forests but with distinct communities at different altitudes The lower altitude (LF) forests can be classified as the association of Schima wallichii ChoisyCastanopsis tribuloides Smith Engelhardtia spicata Blume and Gynocardia odorata RBr(Figure 4) The mid-altitude (MF) forests can be classified asthe association of Symplocos ramosissima Wall ex G DonCastanopsis hystrix Hookf amp Thomson ex ADC Viburnum erubescens Wall and Quercus lamellosa Sm (Figure 5) Whereas high-altitude (HF) forests can be classified as the association of Lithocarpus pachyphyllus Rhododendron arboreum Sm Magnolia campbellii Hookfamp Thomson and Symplocos heishanensis Hayata (Figure 6)

Figure 2 Hierarchically clustered dendrogram and Nonmetic multidimensional scaling (NMDS) ordination on the BrayndashCurtis distance estimated using abundance data (a) Hierarchical clustering was done using Wardrsquos algorithm the y-axis represents the BrayndashCurtis distance shared among each cluster (b) NMDS ordination superimposed with the result of the cluster analysis more similar plots are grouped closer to another The red green and blue dots and lines correspond to the low-altitude forest (LF) mid-altitude forest (MF) and high-altitude forest (HF) forest categories in our study

Figure 2 Hierarchically clustered dendrogram and Nonmetic multidimensional scaling (NMDS)ordination on the BrayndashCurtis distance estimated using abundance data (a) Hierarchical clusteringwas done using Wardrsquos algorithm the y-axis represents the BrayndashCurtis distance shared among eachcluster (b) NMDS ordination superimposed with the result of the cluster analysis more similar plotsare grouped closer to another The red green and blue dots and lines correspond to the low-altitudeforest (LF) mid-altitude forest (MF) and high-altitude forest (HF) forest categories in our studyForests 2019 10 x FOR PEER REVIEW 6 of 18

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure2

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees in flowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis and Schima wallichii trees recorded at 1400 m asl

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure 2

Forests 2019 10 633 6 of 17

Forests 2019 10 x FOR PEER REVIEW 6 of 18

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure2

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees in flowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis and Schima wallichii trees recorded at 1400 m asl

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees inflowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis andSchima wallichii trees recorded at 1400 m aslForests 2019 10 x FOR PEER REVIEW 7 of 18

Figure 5Mid-elevation forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl (b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocos species at 2300 m asl

Figure 6High elevation forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at2860 m asl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 m aslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiern were reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

Figure 5 Mid-altitude forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl(b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocosspecies at 2300 m asl

Forests 2019 10 633 7 of 17

Forests 2019 10 x FOR PEER REVIEW 7 of 18

Figure 5Mid-elevation forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl (b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocos species at 2300 m asl

Figure 6High elevation forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at2860 m asl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 m aslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiern were reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

Figure 6 High altitude forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at 2860 masl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 maslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiernwere reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size)including 87 unidentified individuals with the observed richness of 114 species from 42 families and75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentleslope for the cumulative species richness after the sample size of 70 indicating an adequate number ofplots (Figure 7a) However for different forest communities the species accumulation curve showedan increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest speciesrichness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversityvaried from 171 plusmn 040 to 210 plusmn 035 with the highest value for the LF (Table 1) and the Pieloursquosevenness (J) varied from 046 plusmn 07 to 035 plusmn 010 with the highest value for the LF (Table 1)

Tables 2 and 3 show the results for the Importance Value Index (IVI) estimated for both species(SIVI) and families (FIVI) along with species richness families stem density basal area theirproportion to the contribution for the overall forest community and different forest communitiesWe found Lithocarpus pachyphyllus (930) Symplocos ramosissima (872) Castanopsis hystrix (823)Castanopsis tribuloides (499) and Schima wallichii (446) to be the five most important specieswith the highest Species Importance Value Index (SIVI) Similarly we recorded Fagaceae (2786)Symplocaceae (1226) Theaceae (804) Ericaceae (770) and Lauraceae (662) as the top fivefamilies Lauraceae was also recorded as the most diverse family with 13 species (Table 2)

Forests 2019 10 633 8 of 17

Forests 2019 10 x FOR PEER REVIEW 8 of 18

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size) including 87 unidentified individuals with the observed richness of 114 species from 42 families and 75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentle slope for the cumulative species richness after the sample size of 70 indicating an adequate number of plots (Figure 7a) However for different forest communities the species accumulation curve showed an increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest species richness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20 the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversity varied from 171plusmn040 to 210plusmn035 with the highest value for the LF (Table 1) and the Pieloursquos evenness (J) varied from 046plusmn07 to 035plusmn010 with the highest value for the LF (Table1)

Figure 7Species accumulation curve based on the rarefaction method (a) whole tree population (b)three forests LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities along an altitude gradient The bold figures represent the highest value for each parameter The table provides the median and its standard error since richness showed a right-skewed distribution LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Parameters ALL LF MF HF Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200

Total sampled area (ha) 83 24 37 22 Total species richness 114 68 77 51 Total genus richness 75 52 54 37 Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450

Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351 Median of familiesrsquo richness plusmn

SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281

Total individual counted 10344 2591 5463 2290 Shannon index (H) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fishers alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164

Figure 7 Species accumulation curve based on the rarefaction method (a) whole tree population(b)three forests LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities alongan altitude gradient The table provides the median and its standard error since richness showed aright-skewed distribution LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Parameters ALL LF MF HF

Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200Total sampled area (ha) 83 24 37 22Total species richness 114 68 77 51Total genus richness 75 52 54 37Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351

Median of familiesrsquo richness plusmn SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281Total individual counted 10344 2591 5463 2290Shannon index (Hrsquo) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fisherrsquos alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164Pieloursquos evenness (J) 039 plusmn 039 046 plusmn 07 035 plusmn 010 038 plusmn 013

Dominant family Fagaceae Fagaceae Fagaceae Fagaceae

Dominant species Lithocarpuspachyphyllus Schima wallichii Symplocos

ramosissimaLithocarpus

pachyphyllus

Species such as Schima wallichii (1890) Castanopsis tribuloides (1791) Engelhardtia spicata (672)Gynocardia odorata (605) and Eurya acuminata (483) were recorded as the five most dominant specieswith higher species Important Value Index (SIVI) for LF Similarly Symplocos ramosissima (1509)Castanopsis hystrix (1499) Viburnum erubescens (709) Quercus lamellosa (685) and Eurya acuminata(442) were reported as the most dominant species for the MF The HF was dominated by speciessuch as Lithocarpus pachyphyllus (2437) Rhododendron arboreum (1678) Magnolia campbellii (454)Symplocos heishanensis (414) and Symplocos ramosissima (396) with the highest SIVI (Table 3)

Forests 2019 10 633 9 of 17

Table 2 Species richness density basal area Species Importance Value Index (SIVI) FamiliesImportance Value Index (FIVI) and their proportion to the contribution for the top ten species andfamilies for the overall forest community recorded in the study

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

Lithocarpus pachyphyllus Fagaceae 5120 411 1231 2107 930Symplocos ramosissima Symplocaceae 24807 1919 161 276 872

Castanopsis hystrix Fagaceae 6205 498 991 1697 823Castanopsis tribuloides Fagaceae 6241 501 422 723 499

Schima wallichii Theaceae 7819 627 240 411 446Rhododendron arboreum Ericaceae 8554 686 254 435 433

Viburnum erubescens Adoxaceae 10398 834 038 064 416Eurya acuminata DC Pentaphylacaceae 6096 489 081 139 391

Quercus lamellosa Fagaceae 988 079 495 849 386Symplocos heishanensis Symplocaceae 4940 396 038 065 293

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 9 2088 1675 3361 5719 2786Symplocaceae 4 3355 2692 218 371 1226

Theaceae 2 1390 1116 320 545 804Ericaceae 7 1476 1184 381 648 770Lauraceae 13 463 371 497 845 662Adoxaceae 2 1041 835 038 064 464

Magnoliaceae 4 112 090 209 364 270Betulaceae 4 193 155 099 169 219

Primulaceae 3 228 183 041 023 210Sapindaceae 1 069 055 123 210 189

Table 3 Stem density basal area Family (FIVI) and Species Importance Value Indexes (SIVI) and theirproportion to the contribution of the ten dominant species for three forest communities LF MF and HF

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

LF (900 m to 1700 m asl)

Schima wallichii Theaceae 25917 2401 708 2673 1890Castanopsis tribuloides Fagaceae 20250 1876 768 2900 1791

Engelhardtia spicata Juglandaceae 6042 560 235 886 672Gynocardia odorata Achariaceae 9250 857 123 463 605Euryaa cuminata Pentaphylacaceae 6667 857 076 286 483Betula alnoides

Buch-HamexDDon Betulaceae 2125 618 099 375 257

Macaranga denticulataBlume Euphorbiaceae 1808 197 039 148 243

Albizia lebbeck (L)Benth Leguminosae 1708 158 123 464 231

Alnus nepalensis DDon Betulaceae 1208 066 100 377 221Brassaiopsis hispida

Seem Araliaceae 3625 336 019 053 187

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Theaceae 2 942 3018 784 2960 2254Fagaceae 6 673 2157 823 3109 2017

Juglandaceae 3 180 575 238 899 742Achariaceae 1 267 857 123 463 658Betulaceae 3 104 332 209 789 526

Euphorbiaceae 5 123 394 059 222 434Araliaceae 3 157 502 017 066 353

Leguminosae 3 030 096 127 480 290Anacardiaceae 3 059 189 048 181 276

Lauraceae 6 063 201 009 035 264

Forests 2019 10 633 10 of 17

Table 3 Cont

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

MF (gt1700 m to 2500 m asl)

Symplocos ramosissima Symplocaceae 50189 3399 336 467 1509Castanopsis hystrix Fagaceae 13243 897 2172 3015 1499

Viburnum erubescens Adoxaceae 21189 1435 075 105 709Quercus lamellosa Fagaceae 2054 139 1067 1481 685Eurya acuminata Pentaphylacaceae 8378 567 123 171 442

Lithocarpus pachyphyllus Fagaceae 4892 331 517 718 432Symplocos heishanensis Symplocaceae 7378 500 056 078 407Castanopsis tribuloides Fagaceae 865 059 449 623 278

Quercus glauca Fagaceae 1432 097 309 429 270Elaeocarpus sikkimensis

Mast Elaeocarpaceae 1054 071 221 307 252

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 7 1082 1644 4570 4570 3012Symplocaceae 4 2890 4391 433 433 2014

Lauraceae 11 308 469 960 960 922Adoxaceae 2 946 1437 075 075 807Theaceae 2 395 600 171 171 573

Elaeocarpaceae 2 048 073 223 223 317Ericaceae 5 249 379 049 049 272

Sapindaceae 1 040 060 150 150 260Primulaceae 3 124 189 016 016 260

Magnoliaceae 4 059 090 175 175 243

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

HF (2500 m to 3200 m asl)

Lithocarpus pachyphyllus Fagaceae 11091 1066 788 5456 2437Rhododendron arboreum Ericaceae 31000 2978 747 1308 1678

Magnolia campbellii Magnoliaceae 1682 162 539 662 454Symplocos heishanensis Symplocaceae 6182 594 581 069 414Rhododendron hodgsonii

Hookf Ericaceae 5455 524 290 315 376

Corylus ferox Wall Betulaceae 2682 258 415 133 268Viburnum erubescens

Wall Adoxaceae 3591 345 415 021 260

Acer campbellii Sapindaceae 1091 105 373 298 259Rhododendron falconeri

Hookf Ericaceae 3955 380 207 186 258

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 4 333 9011 4096 5716 7921Ericaceae 7 1182 2963 1347 1879 7216

Symplocaceae 3 464 207 094 131 2602Magnoliaceae 2 048 1090 495 691 1603

Lauraceae 6 092 551 251 350 1524Betulaceae 2 076 216 098 137 991Theaceae 2 053 146 067 093 969

Berberidaceae 2 139 037 017 024 947Adoxaceae 1 095 033 015 021 992

Sapindaceae 1 029 470 214 298 877

33 Forest Structure

Overall we recorded the median stem DBH of 1542 plusmn 111 cm stem density of124627 plusmn 56828 haminus1 and the basal area of 5435 plusmn 444 m2 haminus1 (Figure 8) Among forest communitiesthe HF recorded the highest median stem DBH (2089 plusmn 309 cm) but showed the lowest stem densitySimilarly MF recorded the highest stem density of about 1320 plusmn 11972 haminus1 with the second highest

Forests 2019 10 633 11 of 17

value for the stem DBH and basal area However the LF recorded the second highest value for medianstem density and the minimum value for the median DBH and basal areaForests 2019 10 x FOR PEER REVIEW 12 of 18

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem density and (c) DBH The horizontal line crossing the box in the center represents the median the box represents the 25th and the 75th percentiles The vertical line outside the box represents the minimum and maximum values ALL= total forest community LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure9) with the highest number of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least number of individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution was documented for different forest communities However they differed in the number of DBH classes MF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded in the study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classes and completely lacked trees above 93 cm DBH

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem densityand (c) DBH The horizontal line crossing the box in the center represents the median the box representsthe 25th and the 75th percentiles The vertical line outside the box represents the minimum andmaximum values ALL = total forest community LF = low-altitude forest MF = mid-altitude forestHF = high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure 9) with the highestnumber of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least numberof individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution wasdocumented for different forest communities However they differed in the number of DBH classesMF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded inthe study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classesand completely lacked trees above 93 cm DBH

Forests 2019 10 633 12 of 17

Forests 2019 10 x FOR PEER REVIEW 13 of 18

Figure 9 Class distribution of tress withgt3 cmDBH recorded in the study(a) Overall stems recorded in the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH ranging from gt93 cm to 303 cm LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change of forest tree species recorded in the Himalayan forests like Sikkim The finding could be related to the close association of altitude with climatic variables temperature and precipitation [53] and thus draws attention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communities between 900 masland3200 m asl Moreover according to the Grierson and Long [54]classification our study resembles three forest types ie LF forest as the warm broad-leaved forest MF as the evergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded some differences in species recorded for high-elevation forests For example Grierson and Longrsquos [54] classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall) Boiss as the major trees for high-elevation forests however we did not report similar findings A reason in the change in species composition in these three forest types compared to the 1960s and 1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for its use as scaffolding timber for construction of houses firewood and clearing for forest

Figure 9 Class distribution of tress with gt3 cmDBH recorded in the study (a) Overall stems recordedin the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH rangingfrom gt93 cm to 303 cm LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change offorest tree species recorded in the Himalayan forests like Sikkim The finding could be related to theclose association of altitude with climatic variables temperature and precipitation [53] and thus drawsattention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communitiesbetween 900 masl and 3200 m asl Moreover according to the Grierson and Long [54] classificationour study resembles three forest types ie LF forest as the warm broad-leaved forest MF as theevergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded somedifferences in species recorded for high-elevation forests For example Grierson and Longrsquos [54]classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall)Boiss as the major trees for high-elevation forests however we did not report similar findingsA reason in the change in species composition in these three forest types compared to the 1960s and

Forests 2019 10 633 13 of 17

1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for itsuse as scaffolding timber for construction of houses firewood and clearing for forest for agriculturepractice as well as climate change effects however warrants more investigation Future researchshould focus on documentation of forest trees for a detailed description of the Himalayan forest andrevision of forest type classifications if needed

42 Tree Species Richness and Diversity

Species accumulation curves are especially useful to decide sampling completeness and alsoto compare richness for communities with different sample sizes [50] In the current analysis theaccumulation curves did not record saturation when plotted for different communities which impliesfinding more species with increasing sampling effort Nevertheless it is clear that low-altitude forestsrecord the highest number of species richness and evenness and the disparity between the observedand rarified richness in the MF was the mere result of the sampling size differences The varied speciesrichness between forests covering different altitudes can be associated with numerous factors Broadlywe assumed the widely reported climatic variables namely precipitation temperature and theirinteraction as the prime factor for varied richness along the altitudinal gradient of the Himalayansystem [465556] Another crucial factor could be anthropogenic disturbance and its intensity [2738]Future research could explore these variables to explain the difference in richness recorded in our studyThe LF recorded not only the highest number of species but recorded more uniformly distributed treespecies However the region lacks comparable literature to discuss optimum richness and evennessWe recommend long-term monitoring of permanent plots which is completely lacking in the IndianEastern Himalayas

One of the important findings of our study was the significance of the family Fagaceae It wasrecorded as having the highest FIVI and SIVI and dominated the forest along the altitudinal gradientbetween 900 m and3200 m asl in Sikkim The reasons should be further investigated becauseecological and silvicultural knowledge of Fagaceae and its species are very limited for the EasternHimalayan forests Furthermore we also showed for the first time the existence of huge trees ofLithocarpus pachyphylls (gt15 m DBH) in high number which implies that this species was the remnantof the old growth primary forests in the altitude from 2500 to 3200 in Sikkim Himalayas Theseold-growth trees should be marked registered monitored and conserved as habitat trees by theforest department

However not all forest communities were dominated by the Fagaceae species For example Schimawallichii and Castanopsis tribuloides dominated the lower altitude forests whereas species like Symplocosramosissima and Castanopsis hystrix dominated the mid-altitude forests Different physiological andclimatic adaptation of a species may have caused distinct species to dominate different altituderesulting in three unique clusters of our samples Moreover the Himalayas are renowned for thepronounced variation in climatic topographic and edaphic factors along the altitudinal gradientApart from these the anthropogenic disturbance is a widely discussed factor for the dominance offast-growing pioneer species like the Symplocos racemose [57]Future research could explore the role ofanthropogenic disturbance and different climatic and topographic factors on species dominance andcomposition along the altitudinal gradient

43 Forest Structure Stem Diameter Stem Density and Basal Area

Size class distribution of trees provides the population structure of forest [58] and are extensivelyused to understand regeneration status [59] The present assessment presented a reverse J-shapeddistribution suggesting uneven-aged forests for sustainable reproduction and regeneration [60] with asufficient number of young individuals to replace the old mature stand However between forest typesthe low-altitude forests completely lacked trees above 93 cm diameter a scenario undesirable for asustainable forest The finding could be associated with forest degradation and deforestation reportedby earlier studies [4761] One potential factor for the absence of large trees could be the practice of

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 6 of 17

Forests 2019 10 x FOR PEER REVIEW 6 of 18

Figure 3 Shepard diagram with the goodness of fit (R2) from the NMDS result presented in Figure2

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees in flowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis and Schima wallichii trees recorded at 1400 m asl

Figure 4 The low-altitude forest (LF) (a) Engelhardtia spicata one of the dominant trees inflowering (b)Schima wallichii Symplocos and Castanopsis mixed forest (cndashd) even sized Castanopsis andSchima wallichii trees recorded at 1400 m aslForests 2019 10 x FOR PEER REVIEW 7 of 18

Figure 5Mid-elevation forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl (b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocos species at 2300 m asl

Figure 6High elevation forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at2860 m asl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 m aslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiern were reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

Figure 5 Mid-altitude forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl(b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocosspecies at 2300 m asl

Forests 2019 10 633 7 of 17

Forests 2019 10 x FOR PEER REVIEW 7 of 18

Figure 5Mid-elevation forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl (b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocos species at 2300 m asl

Figure 6High elevation forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at2860 m asl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 m aslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiern were reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

Figure 6 High altitude forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at 2860 masl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 maslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiernwere reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size)including 87 unidentified individuals with the observed richness of 114 species from 42 families and75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentleslope for the cumulative species richness after the sample size of 70 indicating an adequate number ofplots (Figure 7a) However for different forest communities the species accumulation curve showedan increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest speciesrichness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversityvaried from 171 plusmn 040 to 210 plusmn 035 with the highest value for the LF (Table 1) and the Pieloursquosevenness (J) varied from 046 plusmn 07 to 035 plusmn 010 with the highest value for the LF (Table 1)

Tables 2 and 3 show the results for the Importance Value Index (IVI) estimated for both species(SIVI) and families (FIVI) along with species richness families stem density basal area theirproportion to the contribution for the overall forest community and different forest communitiesWe found Lithocarpus pachyphyllus (930) Symplocos ramosissima (872) Castanopsis hystrix (823)Castanopsis tribuloides (499) and Schima wallichii (446) to be the five most important specieswith the highest Species Importance Value Index (SIVI) Similarly we recorded Fagaceae (2786)Symplocaceae (1226) Theaceae (804) Ericaceae (770) and Lauraceae (662) as the top fivefamilies Lauraceae was also recorded as the most diverse family with 13 species (Table 2)

Forests 2019 10 633 8 of 17

Forests 2019 10 x FOR PEER REVIEW 8 of 18

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size) including 87 unidentified individuals with the observed richness of 114 species from 42 families and 75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentle slope for the cumulative species richness after the sample size of 70 indicating an adequate number of plots (Figure 7a) However for different forest communities the species accumulation curve showed an increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest species richness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20 the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversity varied from 171plusmn040 to 210plusmn035 with the highest value for the LF (Table 1) and the Pieloursquos evenness (J) varied from 046plusmn07 to 035plusmn010 with the highest value for the LF (Table1)

Figure 7Species accumulation curve based on the rarefaction method (a) whole tree population (b)three forests LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities along an altitude gradient The bold figures represent the highest value for each parameter The table provides the median and its standard error since richness showed a right-skewed distribution LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Parameters ALL LF MF HF Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200

Total sampled area (ha) 83 24 37 22 Total species richness 114 68 77 51 Total genus richness 75 52 54 37 Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450

Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351 Median of familiesrsquo richness plusmn

SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281

Total individual counted 10344 2591 5463 2290 Shannon index (H) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fishers alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164

Figure 7 Species accumulation curve based on the rarefaction method (a) whole tree population(b)three forests LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities alongan altitude gradient The table provides the median and its standard error since richness showed aright-skewed distribution LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Parameters ALL LF MF HF

Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200Total sampled area (ha) 83 24 37 22Total species richness 114 68 77 51Total genus richness 75 52 54 37Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351

Median of familiesrsquo richness plusmn SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281Total individual counted 10344 2591 5463 2290Shannon index (Hrsquo) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fisherrsquos alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164Pieloursquos evenness (J) 039 plusmn 039 046 plusmn 07 035 plusmn 010 038 plusmn 013

Dominant family Fagaceae Fagaceae Fagaceae Fagaceae

Dominant species Lithocarpuspachyphyllus Schima wallichii Symplocos

ramosissimaLithocarpus

pachyphyllus

Species such as Schima wallichii (1890) Castanopsis tribuloides (1791) Engelhardtia spicata (672)Gynocardia odorata (605) and Eurya acuminata (483) were recorded as the five most dominant specieswith higher species Important Value Index (SIVI) for LF Similarly Symplocos ramosissima (1509)Castanopsis hystrix (1499) Viburnum erubescens (709) Quercus lamellosa (685) and Eurya acuminata(442) were reported as the most dominant species for the MF The HF was dominated by speciessuch as Lithocarpus pachyphyllus (2437) Rhododendron arboreum (1678) Magnolia campbellii (454)Symplocos heishanensis (414) and Symplocos ramosissima (396) with the highest SIVI (Table 3)

Forests 2019 10 633 9 of 17

Table 2 Species richness density basal area Species Importance Value Index (SIVI) FamiliesImportance Value Index (FIVI) and their proportion to the contribution for the top ten species andfamilies for the overall forest community recorded in the study

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

Lithocarpus pachyphyllus Fagaceae 5120 411 1231 2107 930Symplocos ramosissima Symplocaceae 24807 1919 161 276 872

Castanopsis hystrix Fagaceae 6205 498 991 1697 823Castanopsis tribuloides Fagaceae 6241 501 422 723 499

Schima wallichii Theaceae 7819 627 240 411 446Rhododendron arboreum Ericaceae 8554 686 254 435 433

Viburnum erubescens Adoxaceae 10398 834 038 064 416Eurya acuminata DC Pentaphylacaceae 6096 489 081 139 391

Quercus lamellosa Fagaceae 988 079 495 849 386Symplocos heishanensis Symplocaceae 4940 396 038 065 293

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 9 2088 1675 3361 5719 2786Symplocaceae 4 3355 2692 218 371 1226

Theaceae 2 1390 1116 320 545 804Ericaceae 7 1476 1184 381 648 770Lauraceae 13 463 371 497 845 662Adoxaceae 2 1041 835 038 064 464

Magnoliaceae 4 112 090 209 364 270Betulaceae 4 193 155 099 169 219

Primulaceae 3 228 183 041 023 210Sapindaceae 1 069 055 123 210 189

Table 3 Stem density basal area Family (FIVI) and Species Importance Value Indexes (SIVI) and theirproportion to the contribution of the ten dominant species for three forest communities LF MF and HF

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

LF (900 m to 1700 m asl)

Schima wallichii Theaceae 25917 2401 708 2673 1890Castanopsis tribuloides Fagaceae 20250 1876 768 2900 1791

Engelhardtia spicata Juglandaceae 6042 560 235 886 672Gynocardia odorata Achariaceae 9250 857 123 463 605Euryaa cuminata Pentaphylacaceae 6667 857 076 286 483Betula alnoides

Buch-HamexDDon Betulaceae 2125 618 099 375 257

Macaranga denticulataBlume Euphorbiaceae 1808 197 039 148 243

Albizia lebbeck (L)Benth Leguminosae 1708 158 123 464 231

Alnus nepalensis DDon Betulaceae 1208 066 100 377 221Brassaiopsis hispida

Seem Araliaceae 3625 336 019 053 187

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Theaceae 2 942 3018 784 2960 2254Fagaceae 6 673 2157 823 3109 2017

Juglandaceae 3 180 575 238 899 742Achariaceae 1 267 857 123 463 658Betulaceae 3 104 332 209 789 526

Euphorbiaceae 5 123 394 059 222 434Araliaceae 3 157 502 017 066 353

Leguminosae 3 030 096 127 480 290Anacardiaceae 3 059 189 048 181 276

Lauraceae 6 063 201 009 035 264

Forests 2019 10 633 10 of 17

Table 3 Cont

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

MF (gt1700 m to 2500 m asl)

Symplocos ramosissima Symplocaceae 50189 3399 336 467 1509Castanopsis hystrix Fagaceae 13243 897 2172 3015 1499

Viburnum erubescens Adoxaceae 21189 1435 075 105 709Quercus lamellosa Fagaceae 2054 139 1067 1481 685Eurya acuminata Pentaphylacaceae 8378 567 123 171 442

Lithocarpus pachyphyllus Fagaceae 4892 331 517 718 432Symplocos heishanensis Symplocaceae 7378 500 056 078 407Castanopsis tribuloides Fagaceae 865 059 449 623 278

Quercus glauca Fagaceae 1432 097 309 429 270Elaeocarpus sikkimensis

Mast Elaeocarpaceae 1054 071 221 307 252

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 7 1082 1644 4570 4570 3012Symplocaceae 4 2890 4391 433 433 2014

Lauraceae 11 308 469 960 960 922Adoxaceae 2 946 1437 075 075 807Theaceae 2 395 600 171 171 573

Elaeocarpaceae 2 048 073 223 223 317Ericaceae 5 249 379 049 049 272

Sapindaceae 1 040 060 150 150 260Primulaceae 3 124 189 016 016 260

Magnoliaceae 4 059 090 175 175 243

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

HF (2500 m to 3200 m asl)

Lithocarpus pachyphyllus Fagaceae 11091 1066 788 5456 2437Rhododendron arboreum Ericaceae 31000 2978 747 1308 1678

Magnolia campbellii Magnoliaceae 1682 162 539 662 454Symplocos heishanensis Symplocaceae 6182 594 581 069 414Rhododendron hodgsonii

Hookf Ericaceae 5455 524 290 315 376

Corylus ferox Wall Betulaceae 2682 258 415 133 268Viburnum erubescens

Wall Adoxaceae 3591 345 415 021 260

Acer campbellii Sapindaceae 1091 105 373 298 259Rhododendron falconeri

Hookf Ericaceae 3955 380 207 186 258

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 4 333 9011 4096 5716 7921Ericaceae 7 1182 2963 1347 1879 7216

Symplocaceae 3 464 207 094 131 2602Magnoliaceae 2 048 1090 495 691 1603

Lauraceae 6 092 551 251 350 1524Betulaceae 2 076 216 098 137 991Theaceae 2 053 146 067 093 969

Berberidaceae 2 139 037 017 024 947Adoxaceae 1 095 033 015 021 992

Sapindaceae 1 029 470 214 298 877

33 Forest Structure

Overall we recorded the median stem DBH of 1542 plusmn 111 cm stem density of124627 plusmn 56828 haminus1 and the basal area of 5435 plusmn 444 m2 haminus1 (Figure 8) Among forest communitiesthe HF recorded the highest median stem DBH (2089 plusmn 309 cm) but showed the lowest stem densitySimilarly MF recorded the highest stem density of about 1320 plusmn 11972 haminus1 with the second highest

Forests 2019 10 633 11 of 17

value for the stem DBH and basal area However the LF recorded the second highest value for medianstem density and the minimum value for the median DBH and basal areaForests 2019 10 x FOR PEER REVIEW 12 of 18

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem density and (c) DBH The horizontal line crossing the box in the center represents the median the box represents the 25th and the 75th percentiles The vertical line outside the box represents the minimum and maximum values ALL= total forest community LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure9) with the highest number of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least number of individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution was documented for different forest communities However they differed in the number of DBH classes MF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded in the study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classes and completely lacked trees above 93 cm DBH

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem densityand (c) DBH The horizontal line crossing the box in the center represents the median the box representsthe 25th and the 75th percentiles The vertical line outside the box represents the minimum andmaximum values ALL = total forest community LF = low-altitude forest MF = mid-altitude forestHF = high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure 9) with the highestnumber of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least numberof individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution wasdocumented for different forest communities However they differed in the number of DBH classesMF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded inthe study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classesand completely lacked trees above 93 cm DBH

Forests 2019 10 633 12 of 17

Forests 2019 10 x FOR PEER REVIEW 13 of 18

Figure 9 Class distribution of tress withgt3 cmDBH recorded in the study(a) Overall stems recorded in the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH ranging from gt93 cm to 303 cm LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change of forest tree species recorded in the Himalayan forests like Sikkim The finding could be related to the close association of altitude with climatic variables temperature and precipitation [53] and thus draws attention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communities between 900 masland3200 m asl Moreover according to the Grierson and Long [54]classification our study resembles three forest types ie LF forest as the warm broad-leaved forest MF as the evergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded some differences in species recorded for high-elevation forests For example Grierson and Longrsquos [54] classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall) Boiss as the major trees for high-elevation forests however we did not report similar findings A reason in the change in species composition in these three forest types compared to the 1960s and 1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for its use as scaffolding timber for construction of houses firewood and clearing for forest

Figure 9 Class distribution of tress with gt3 cmDBH recorded in the study (a) Overall stems recordedin the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH rangingfrom gt93 cm to 303 cm LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change offorest tree species recorded in the Himalayan forests like Sikkim The finding could be related to theclose association of altitude with climatic variables temperature and precipitation [53] and thus drawsattention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communitiesbetween 900 masl and 3200 m asl Moreover according to the Grierson and Long [54] classificationour study resembles three forest types ie LF forest as the warm broad-leaved forest MF as theevergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded somedifferences in species recorded for high-elevation forests For example Grierson and Longrsquos [54]classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall)Boiss as the major trees for high-elevation forests however we did not report similar findingsA reason in the change in species composition in these three forest types compared to the 1960s and

Forests 2019 10 633 13 of 17

1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for itsuse as scaffolding timber for construction of houses firewood and clearing for forest for agriculturepractice as well as climate change effects however warrants more investigation Future researchshould focus on documentation of forest trees for a detailed description of the Himalayan forest andrevision of forest type classifications if needed

42 Tree Species Richness and Diversity

Species accumulation curves are especially useful to decide sampling completeness and alsoto compare richness for communities with different sample sizes [50] In the current analysis theaccumulation curves did not record saturation when plotted for different communities which impliesfinding more species with increasing sampling effort Nevertheless it is clear that low-altitude forestsrecord the highest number of species richness and evenness and the disparity between the observedand rarified richness in the MF was the mere result of the sampling size differences The varied speciesrichness between forests covering different altitudes can be associated with numerous factors Broadlywe assumed the widely reported climatic variables namely precipitation temperature and theirinteraction as the prime factor for varied richness along the altitudinal gradient of the Himalayansystem [465556] Another crucial factor could be anthropogenic disturbance and its intensity [2738]Future research could explore these variables to explain the difference in richness recorded in our studyThe LF recorded not only the highest number of species but recorded more uniformly distributed treespecies However the region lacks comparable literature to discuss optimum richness and evennessWe recommend long-term monitoring of permanent plots which is completely lacking in the IndianEastern Himalayas

One of the important findings of our study was the significance of the family Fagaceae It wasrecorded as having the highest FIVI and SIVI and dominated the forest along the altitudinal gradientbetween 900 m and3200 m asl in Sikkim The reasons should be further investigated becauseecological and silvicultural knowledge of Fagaceae and its species are very limited for the EasternHimalayan forests Furthermore we also showed for the first time the existence of huge trees ofLithocarpus pachyphylls (gt15 m DBH) in high number which implies that this species was the remnantof the old growth primary forests in the altitude from 2500 to 3200 in Sikkim Himalayas Theseold-growth trees should be marked registered monitored and conserved as habitat trees by theforest department

However not all forest communities were dominated by the Fagaceae species For example Schimawallichii and Castanopsis tribuloides dominated the lower altitude forests whereas species like Symplocosramosissima and Castanopsis hystrix dominated the mid-altitude forests Different physiological andclimatic adaptation of a species may have caused distinct species to dominate different altituderesulting in three unique clusters of our samples Moreover the Himalayas are renowned for thepronounced variation in climatic topographic and edaphic factors along the altitudinal gradientApart from these the anthropogenic disturbance is a widely discussed factor for the dominance offast-growing pioneer species like the Symplocos racemose [57]Future research could explore the role ofanthropogenic disturbance and different climatic and topographic factors on species dominance andcomposition along the altitudinal gradient

43 Forest Structure Stem Diameter Stem Density and Basal Area

Size class distribution of trees provides the population structure of forest [58] and are extensivelyused to understand regeneration status [59] The present assessment presented a reverse J-shapeddistribution suggesting uneven-aged forests for sustainable reproduction and regeneration [60] with asufficient number of young individuals to replace the old mature stand However between forest typesthe low-altitude forests completely lacked trees above 93 cm diameter a scenario undesirable for asustainable forest The finding could be associated with forest degradation and deforestation reportedby earlier studies [4761] One potential factor for the absence of large trees could be the practice of

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 7 of 17

Forests 2019 10 x FOR PEER REVIEW 7 of 18

Figure 5Mid-elevation forest (MF) (a) Castanopsis hystrix with 82cm DBH recorded at 2100 m asl (b) forests dominated by Castanopsis tribuloides at 2000 m asl (cndashd) forests dominated by Symplocos species at 2300 m asl

Figure 6High elevation forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at2860 m asl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 m aslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiern were reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

Figure 6 High altitude forest (HF) (a) Lithocarpus pachyphyllus with 110 cm DBH recorded at 2860 masl (b) Lithocarpus pachyphyllus and Rhododendron mixed forest 2875 m asl (c) Open forest at 2792 maslwhere Lithocarpus pachyphyllus Magnolia campbellii and Acer campbellii Hooffamp Thomson ex Hiernwere reported as the major tree species (d) a Rhododendron forest stand at 3200 m asl

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size)including 87 unidentified individuals with the observed richness of 114 species from 42 families and75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentleslope for the cumulative species richness after the sample size of 70 indicating an adequate number ofplots (Figure 7a) However for different forest communities the species accumulation curve showedan increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest speciesrichness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversityvaried from 171 plusmn 040 to 210 plusmn 035 with the highest value for the LF (Table 1) and the Pieloursquosevenness (J) varied from 046 plusmn 07 to 035 plusmn 010 with the highest value for the LF (Table 1)

Tables 2 and 3 show the results for the Importance Value Index (IVI) estimated for both species(SIVI) and families (FIVI) along with species richness families stem density basal area theirproportion to the contribution for the overall forest community and different forest communitiesWe found Lithocarpus pachyphyllus (930) Symplocos ramosissima (872) Castanopsis hystrix (823)Castanopsis tribuloides (499) and Schima wallichii (446) to be the five most important specieswith the highest Species Importance Value Index (SIVI) Similarly we recorded Fagaceae (2786)Symplocaceae (1226) Theaceae (804) Ericaceae (770) and Lauraceae (662) as the top fivefamilies Lauraceae was also recorded as the most diverse family with 13 species (Table 2)

Forests 2019 10 633 8 of 17

Forests 2019 10 x FOR PEER REVIEW 8 of 18

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size) including 87 unidentified individuals with the observed richness of 114 species from 42 families and 75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentle slope for the cumulative species richness after the sample size of 70 indicating an adequate number of plots (Figure 7a) However for different forest communities the species accumulation curve showed an increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest species richness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20 the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversity varied from 171plusmn040 to 210plusmn035 with the highest value for the LF (Table 1) and the Pieloursquos evenness (J) varied from 046plusmn07 to 035plusmn010 with the highest value for the LF (Table1)

Figure 7Species accumulation curve based on the rarefaction method (a) whole tree population (b)three forests LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities along an altitude gradient The bold figures represent the highest value for each parameter The table provides the median and its standard error since richness showed a right-skewed distribution LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Parameters ALL LF MF HF Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200

Total sampled area (ha) 83 24 37 22 Total species richness 114 68 77 51 Total genus richness 75 52 54 37 Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450

Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351 Median of familiesrsquo richness plusmn

SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281

Total individual counted 10344 2591 5463 2290 Shannon index (H) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fishers alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164

Figure 7 Species accumulation curve based on the rarefaction method (a) whole tree population(b)three forests LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities alongan altitude gradient The table provides the median and its standard error since richness showed aright-skewed distribution LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Parameters ALL LF MF HF

Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200Total sampled area (ha) 83 24 37 22Total species richness 114 68 77 51Total genus richness 75 52 54 37Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351

Median of familiesrsquo richness plusmn SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281Total individual counted 10344 2591 5463 2290Shannon index (Hrsquo) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fisherrsquos alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164Pieloursquos evenness (J) 039 plusmn 039 046 plusmn 07 035 plusmn 010 038 plusmn 013

Dominant family Fagaceae Fagaceae Fagaceae Fagaceae

Dominant species Lithocarpuspachyphyllus Schima wallichii Symplocos

ramosissimaLithocarpus

pachyphyllus

Species such as Schima wallichii (1890) Castanopsis tribuloides (1791) Engelhardtia spicata (672)Gynocardia odorata (605) and Eurya acuminata (483) were recorded as the five most dominant specieswith higher species Important Value Index (SIVI) for LF Similarly Symplocos ramosissima (1509)Castanopsis hystrix (1499) Viburnum erubescens (709) Quercus lamellosa (685) and Eurya acuminata(442) were reported as the most dominant species for the MF The HF was dominated by speciessuch as Lithocarpus pachyphyllus (2437) Rhododendron arboreum (1678) Magnolia campbellii (454)Symplocos heishanensis (414) and Symplocos ramosissima (396) with the highest SIVI (Table 3)

Forests 2019 10 633 9 of 17

Table 2 Species richness density basal area Species Importance Value Index (SIVI) FamiliesImportance Value Index (FIVI) and their proportion to the contribution for the top ten species andfamilies for the overall forest community recorded in the study

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

Lithocarpus pachyphyllus Fagaceae 5120 411 1231 2107 930Symplocos ramosissima Symplocaceae 24807 1919 161 276 872

Castanopsis hystrix Fagaceae 6205 498 991 1697 823Castanopsis tribuloides Fagaceae 6241 501 422 723 499

Schima wallichii Theaceae 7819 627 240 411 446Rhododendron arboreum Ericaceae 8554 686 254 435 433

Viburnum erubescens Adoxaceae 10398 834 038 064 416Eurya acuminata DC Pentaphylacaceae 6096 489 081 139 391

Quercus lamellosa Fagaceae 988 079 495 849 386Symplocos heishanensis Symplocaceae 4940 396 038 065 293

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 9 2088 1675 3361 5719 2786Symplocaceae 4 3355 2692 218 371 1226

Theaceae 2 1390 1116 320 545 804Ericaceae 7 1476 1184 381 648 770Lauraceae 13 463 371 497 845 662Adoxaceae 2 1041 835 038 064 464

Magnoliaceae 4 112 090 209 364 270Betulaceae 4 193 155 099 169 219

Primulaceae 3 228 183 041 023 210Sapindaceae 1 069 055 123 210 189

Table 3 Stem density basal area Family (FIVI) and Species Importance Value Indexes (SIVI) and theirproportion to the contribution of the ten dominant species for three forest communities LF MF and HF

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

LF (900 m to 1700 m asl)

Schima wallichii Theaceae 25917 2401 708 2673 1890Castanopsis tribuloides Fagaceae 20250 1876 768 2900 1791

Engelhardtia spicata Juglandaceae 6042 560 235 886 672Gynocardia odorata Achariaceae 9250 857 123 463 605Euryaa cuminata Pentaphylacaceae 6667 857 076 286 483Betula alnoides

Buch-HamexDDon Betulaceae 2125 618 099 375 257

Macaranga denticulataBlume Euphorbiaceae 1808 197 039 148 243

Albizia lebbeck (L)Benth Leguminosae 1708 158 123 464 231

Alnus nepalensis DDon Betulaceae 1208 066 100 377 221Brassaiopsis hispida

Seem Araliaceae 3625 336 019 053 187

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Theaceae 2 942 3018 784 2960 2254Fagaceae 6 673 2157 823 3109 2017

Juglandaceae 3 180 575 238 899 742Achariaceae 1 267 857 123 463 658Betulaceae 3 104 332 209 789 526

Euphorbiaceae 5 123 394 059 222 434Araliaceae 3 157 502 017 066 353

Leguminosae 3 030 096 127 480 290Anacardiaceae 3 059 189 048 181 276

Lauraceae 6 063 201 009 035 264

Forests 2019 10 633 10 of 17

Table 3 Cont

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

MF (gt1700 m to 2500 m asl)

Symplocos ramosissima Symplocaceae 50189 3399 336 467 1509Castanopsis hystrix Fagaceae 13243 897 2172 3015 1499

Viburnum erubescens Adoxaceae 21189 1435 075 105 709Quercus lamellosa Fagaceae 2054 139 1067 1481 685Eurya acuminata Pentaphylacaceae 8378 567 123 171 442

Lithocarpus pachyphyllus Fagaceae 4892 331 517 718 432Symplocos heishanensis Symplocaceae 7378 500 056 078 407Castanopsis tribuloides Fagaceae 865 059 449 623 278

Quercus glauca Fagaceae 1432 097 309 429 270Elaeocarpus sikkimensis

Mast Elaeocarpaceae 1054 071 221 307 252

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 7 1082 1644 4570 4570 3012Symplocaceae 4 2890 4391 433 433 2014

Lauraceae 11 308 469 960 960 922Adoxaceae 2 946 1437 075 075 807Theaceae 2 395 600 171 171 573

Elaeocarpaceae 2 048 073 223 223 317Ericaceae 5 249 379 049 049 272

Sapindaceae 1 040 060 150 150 260Primulaceae 3 124 189 016 016 260

Magnoliaceae 4 059 090 175 175 243

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

HF (2500 m to 3200 m asl)

Lithocarpus pachyphyllus Fagaceae 11091 1066 788 5456 2437Rhododendron arboreum Ericaceae 31000 2978 747 1308 1678

Magnolia campbellii Magnoliaceae 1682 162 539 662 454Symplocos heishanensis Symplocaceae 6182 594 581 069 414Rhododendron hodgsonii

Hookf Ericaceae 5455 524 290 315 376

Corylus ferox Wall Betulaceae 2682 258 415 133 268Viburnum erubescens

Wall Adoxaceae 3591 345 415 021 260

Acer campbellii Sapindaceae 1091 105 373 298 259Rhododendron falconeri

Hookf Ericaceae 3955 380 207 186 258

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 4 333 9011 4096 5716 7921Ericaceae 7 1182 2963 1347 1879 7216

Symplocaceae 3 464 207 094 131 2602Magnoliaceae 2 048 1090 495 691 1603

Lauraceae 6 092 551 251 350 1524Betulaceae 2 076 216 098 137 991Theaceae 2 053 146 067 093 969

Berberidaceae 2 139 037 017 024 947Adoxaceae 1 095 033 015 021 992

Sapindaceae 1 029 470 214 298 877

33 Forest Structure

Overall we recorded the median stem DBH of 1542 plusmn 111 cm stem density of124627 plusmn 56828 haminus1 and the basal area of 5435 plusmn 444 m2 haminus1 (Figure 8) Among forest communitiesthe HF recorded the highest median stem DBH (2089 plusmn 309 cm) but showed the lowest stem densitySimilarly MF recorded the highest stem density of about 1320 plusmn 11972 haminus1 with the second highest

Forests 2019 10 633 11 of 17

value for the stem DBH and basal area However the LF recorded the second highest value for medianstem density and the minimum value for the median DBH and basal areaForests 2019 10 x FOR PEER REVIEW 12 of 18

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem density and (c) DBH The horizontal line crossing the box in the center represents the median the box represents the 25th and the 75th percentiles The vertical line outside the box represents the minimum and maximum values ALL= total forest community LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure9) with the highest number of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least number of individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution was documented for different forest communities However they differed in the number of DBH classes MF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded in the study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classes and completely lacked trees above 93 cm DBH

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem densityand (c) DBH The horizontal line crossing the box in the center represents the median the box representsthe 25th and the 75th percentiles The vertical line outside the box represents the minimum andmaximum values ALL = total forest community LF = low-altitude forest MF = mid-altitude forestHF = high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure 9) with the highestnumber of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least numberof individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution wasdocumented for different forest communities However they differed in the number of DBH classesMF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded inthe study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classesand completely lacked trees above 93 cm DBH

Forests 2019 10 633 12 of 17

Forests 2019 10 x FOR PEER REVIEW 13 of 18

Figure 9 Class distribution of tress withgt3 cmDBH recorded in the study(a) Overall stems recorded in the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH ranging from gt93 cm to 303 cm LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change of forest tree species recorded in the Himalayan forests like Sikkim The finding could be related to the close association of altitude with climatic variables temperature and precipitation [53] and thus draws attention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communities between 900 masland3200 m asl Moreover according to the Grierson and Long [54]classification our study resembles three forest types ie LF forest as the warm broad-leaved forest MF as the evergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded some differences in species recorded for high-elevation forests For example Grierson and Longrsquos [54] classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall) Boiss as the major trees for high-elevation forests however we did not report similar findings A reason in the change in species composition in these three forest types compared to the 1960s and 1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for its use as scaffolding timber for construction of houses firewood and clearing for forest

Figure 9 Class distribution of tress with gt3 cmDBH recorded in the study (a) Overall stems recordedin the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH rangingfrom gt93 cm to 303 cm LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change offorest tree species recorded in the Himalayan forests like Sikkim The finding could be related to theclose association of altitude with climatic variables temperature and precipitation [53] and thus drawsattention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communitiesbetween 900 masl and 3200 m asl Moreover according to the Grierson and Long [54] classificationour study resembles three forest types ie LF forest as the warm broad-leaved forest MF as theevergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded somedifferences in species recorded for high-elevation forests For example Grierson and Longrsquos [54]classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall)Boiss as the major trees for high-elevation forests however we did not report similar findingsA reason in the change in species composition in these three forest types compared to the 1960s and

Forests 2019 10 633 13 of 17

1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for itsuse as scaffolding timber for construction of houses firewood and clearing for forest for agriculturepractice as well as climate change effects however warrants more investigation Future researchshould focus on documentation of forest trees for a detailed description of the Himalayan forest andrevision of forest type classifications if needed

42 Tree Species Richness and Diversity

Species accumulation curves are especially useful to decide sampling completeness and alsoto compare richness for communities with different sample sizes [50] In the current analysis theaccumulation curves did not record saturation when plotted for different communities which impliesfinding more species with increasing sampling effort Nevertheless it is clear that low-altitude forestsrecord the highest number of species richness and evenness and the disparity between the observedand rarified richness in the MF was the mere result of the sampling size differences The varied speciesrichness between forests covering different altitudes can be associated with numerous factors Broadlywe assumed the widely reported climatic variables namely precipitation temperature and theirinteraction as the prime factor for varied richness along the altitudinal gradient of the Himalayansystem [465556] Another crucial factor could be anthropogenic disturbance and its intensity [2738]Future research could explore these variables to explain the difference in richness recorded in our studyThe LF recorded not only the highest number of species but recorded more uniformly distributed treespecies However the region lacks comparable literature to discuss optimum richness and evennessWe recommend long-term monitoring of permanent plots which is completely lacking in the IndianEastern Himalayas

One of the important findings of our study was the significance of the family Fagaceae It wasrecorded as having the highest FIVI and SIVI and dominated the forest along the altitudinal gradientbetween 900 m and3200 m asl in Sikkim The reasons should be further investigated becauseecological and silvicultural knowledge of Fagaceae and its species are very limited for the EasternHimalayan forests Furthermore we also showed for the first time the existence of huge trees ofLithocarpus pachyphylls (gt15 m DBH) in high number which implies that this species was the remnantof the old growth primary forests in the altitude from 2500 to 3200 in Sikkim Himalayas Theseold-growth trees should be marked registered monitored and conserved as habitat trees by theforest department

However not all forest communities were dominated by the Fagaceae species For example Schimawallichii and Castanopsis tribuloides dominated the lower altitude forests whereas species like Symplocosramosissima and Castanopsis hystrix dominated the mid-altitude forests Different physiological andclimatic adaptation of a species may have caused distinct species to dominate different altituderesulting in three unique clusters of our samples Moreover the Himalayas are renowned for thepronounced variation in climatic topographic and edaphic factors along the altitudinal gradientApart from these the anthropogenic disturbance is a widely discussed factor for the dominance offast-growing pioneer species like the Symplocos racemose [57]Future research could explore the role ofanthropogenic disturbance and different climatic and topographic factors on species dominance andcomposition along the altitudinal gradient

43 Forest Structure Stem Diameter Stem Density and Basal Area

Size class distribution of trees provides the population structure of forest [58] and are extensivelyused to understand regeneration status [59] The present assessment presented a reverse J-shapeddistribution suggesting uneven-aged forests for sustainable reproduction and regeneration [60] with asufficient number of young individuals to replace the old mature stand However between forest typesthe low-altitude forests completely lacked trees above 93 cm diameter a scenario undesirable for asustainable forest The finding could be associated with forest degradation and deforestation reportedby earlier studies [4761] One potential factor for the absence of large trees could be the practice of

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 8 of 17

Forests 2019 10 x FOR PEER REVIEW 8 of 18

32 Diversity and Composition of Tree Species in Three Forest Communities

Overall the study recorded 10344 individuals from 83 ha (83 sample plots of each 01 ha in size) including 87 unidentified individuals with the observed richness of 114 species from 42 families and 75 genera (Table 1) The species accumulation curve for the entire tree community showed a gentle slope for the cumulative species richness after the sample size of 70 indicating an adequate number of plots (Figure 7a) However for different forest communities the species accumulation curve showed an increasing asymptote (Figure 7b) Nevertheless the curve clearly recorded the highest species richness (77) for the MF followed by LF (68) and HF(51) However for the same sample size of 20 the LF recorded the highest average species richness of 6536 (Figure 7b) The Shannon (H) diversity varied from 171plusmn040 to 210plusmn035 with the highest value for the LF (Table 1) and the Pieloursquos evenness (J) varied from 046plusmn07 to 035plusmn010 with the highest value for the LF (Table1)

Figure 7Species accumulation curve based on the rarefaction method (a) whole tree population (b)three forests LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities along an altitude gradient The bold figures represent the highest value for each parameter The table provides the median and its standard error since richness showed a right-skewed distribution LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

Parameters ALL LF MF HF Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200

Total sampled area (ha) 83 24 37 22 Total species richness 114 68 77 51 Total genus richness 75 52 54 37 Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450

Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351 Median of familiesrsquo richness plusmn

SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281

Total individual counted 10344 2591 5463 2290 Shannon index (H) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fishers alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164

Figure 7 Species accumulation curve based on the rarefaction method (a) whole tree population(b)three forests LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Table 1 Richness species diversity stem densities and basal area for different forest communities alongan altitude gradient The table provides the median and its standard error since richness showed aright-skewed distribution LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

Parameters ALL LF MF HF

Altitude (m) 900ndash3200 900ndash1700 gt1700ndash2500 gt2500ndash3200Total sampled area (ha) 83 24 37 22Total species richness 114 68 77 51Total genus richness 75 52 54 37Total family richness 42 33 35 25

Median of speciesrsquo richness plusmn SE 14 plusmn 434 16 plusmn 443 14 plusmn 304 10 plusmn 450Median of genusrsquo richness plusmn SE 12 plusmn 396 1550 plusmn 406 12 plusmn 272 10 plusmn 351

Median of familiesrsquo richness plusmn SE 10 plusmn 289 13 plusmn 294 9 plusmn 195 9 plusmn 281Total individual counted 10344 2591 5463 2290Shannon index (Hrsquo) plusmn SD 187 plusmn 044 210 plusmn 035 171 plusmn 040 178 plusmn 048

Fisherrsquos alpha richness 444 plusmn 184 568 plusmn 200 414 plusmn 135 351 plusmn 164Pieloursquos evenness (J) 039 plusmn 039 046 plusmn 07 035 plusmn 010 038 plusmn 013

Dominant family Fagaceae Fagaceae Fagaceae Fagaceae

Dominant species Lithocarpuspachyphyllus Schima wallichii Symplocos

ramosissimaLithocarpus

pachyphyllus

Species such as Schima wallichii (1890) Castanopsis tribuloides (1791) Engelhardtia spicata (672)Gynocardia odorata (605) and Eurya acuminata (483) were recorded as the five most dominant specieswith higher species Important Value Index (SIVI) for LF Similarly Symplocos ramosissima (1509)Castanopsis hystrix (1499) Viburnum erubescens (709) Quercus lamellosa (685) and Eurya acuminata(442) were reported as the most dominant species for the MF The HF was dominated by speciessuch as Lithocarpus pachyphyllus (2437) Rhododendron arboreum (1678) Magnolia campbellii (454)Symplocos heishanensis (414) and Symplocos ramosissima (396) with the highest SIVI (Table 3)

Forests 2019 10 633 9 of 17

Table 2 Species richness density basal area Species Importance Value Index (SIVI) FamiliesImportance Value Index (FIVI) and their proportion to the contribution for the top ten species andfamilies for the overall forest community recorded in the study

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

Lithocarpus pachyphyllus Fagaceae 5120 411 1231 2107 930Symplocos ramosissima Symplocaceae 24807 1919 161 276 872

Castanopsis hystrix Fagaceae 6205 498 991 1697 823Castanopsis tribuloides Fagaceae 6241 501 422 723 499

Schima wallichii Theaceae 7819 627 240 411 446Rhododendron arboreum Ericaceae 8554 686 254 435 433

Viburnum erubescens Adoxaceae 10398 834 038 064 416Eurya acuminata DC Pentaphylacaceae 6096 489 081 139 391

Quercus lamellosa Fagaceae 988 079 495 849 386Symplocos heishanensis Symplocaceae 4940 396 038 065 293

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 9 2088 1675 3361 5719 2786Symplocaceae 4 3355 2692 218 371 1226

Theaceae 2 1390 1116 320 545 804Ericaceae 7 1476 1184 381 648 770Lauraceae 13 463 371 497 845 662Adoxaceae 2 1041 835 038 064 464

Magnoliaceae 4 112 090 209 364 270Betulaceae 4 193 155 099 169 219

Primulaceae 3 228 183 041 023 210Sapindaceae 1 069 055 123 210 189

Table 3 Stem density basal area Family (FIVI) and Species Importance Value Indexes (SIVI) and theirproportion to the contribution of the ten dominant species for three forest communities LF MF and HF

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

LF (900 m to 1700 m asl)

Schima wallichii Theaceae 25917 2401 708 2673 1890Castanopsis tribuloides Fagaceae 20250 1876 768 2900 1791

Engelhardtia spicata Juglandaceae 6042 560 235 886 672Gynocardia odorata Achariaceae 9250 857 123 463 605Euryaa cuminata Pentaphylacaceae 6667 857 076 286 483Betula alnoides

Buch-HamexDDon Betulaceae 2125 618 099 375 257

Macaranga denticulataBlume Euphorbiaceae 1808 197 039 148 243

Albizia lebbeck (L)Benth Leguminosae 1708 158 123 464 231

Alnus nepalensis DDon Betulaceae 1208 066 100 377 221Brassaiopsis hispida

Seem Araliaceae 3625 336 019 053 187

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Theaceae 2 942 3018 784 2960 2254Fagaceae 6 673 2157 823 3109 2017

Juglandaceae 3 180 575 238 899 742Achariaceae 1 267 857 123 463 658Betulaceae 3 104 332 209 789 526

Euphorbiaceae 5 123 394 059 222 434Araliaceae 3 157 502 017 066 353

Leguminosae 3 030 096 127 480 290Anacardiaceae 3 059 189 048 181 276

Lauraceae 6 063 201 009 035 264

Forests 2019 10 633 10 of 17

Table 3 Cont

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

MF (gt1700 m to 2500 m asl)

Symplocos ramosissima Symplocaceae 50189 3399 336 467 1509Castanopsis hystrix Fagaceae 13243 897 2172 3015 1499

Viburnum erubescens Adoxaceae 21189 1435 075 105 709Quercus lamellosa Fagaceae 2054 139 1067 1481 685Eurya acuminata Pentaphylacaceae 8378 567 123 171 442

Lithocarpus pachyphyllus Fagaceae 4892 331 517 718 432Symplocos heishanensis Symplocaceae 7378 500 056 078 407Castanopsis tribuloides Fagaceae 865 059 449 623 278

Quercus glauca Fagaceae 1432 097 309 429 270Elaeocarpus sikkimensis

Mast Elaeocarpaceae 1054 071 221 307 252

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 7 1082 1644 4570 4570 3012Symplocaceae 4 2890 4391 433 433 2014

Lauraceae 11 308 469 960 960 922Adoxaceae 2 946 1437 075 075 807Theaceae 2 395 600 171 171 573

Elaeocarpaceae 2 048 073 223 223 317Ericaceae 5 249 379 049 049 272

Sapindaceae 1 040 060 150 150 260Primulaceae 3 124 189 016 016 260

Magnoliaceae 4 059 090 175 175 243

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

HF (2500 m to 3200 m asl)

Lithocarpus pachyphyllus Fagaceae 11091 1066 788 5456 2437Rhododendron arboreum Ericaceae 31000 2978 747 1308 1678

Magnolia campbellii Magnoliaceae 1682 162 539 662 454Symplocos heishanensis Symplocaceae 6182 594 581 069 414Rhododendron hodgsonii

Hookf Ericaceae 5455 524 290 315 376

Corylus ferox Wall Betulaceae 2682 258 415 133 268Viburnum erubescens

Wall Adoxaceae 3591 345 415 021 260

Acer campbellii Sapindaceae 1091 105 373 298 259Rhododendron falconeri

Hookf Ericaceae 3955 380 207 186 258

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 4 333 9011 4096 5716 7921Ericaceae 7 1182 2963 1347 1879 7216

Symplocaceae 3 464 207 094 131 2602Magnoliaceae 2 048 1090 495 691 1603

Lauraceae 6 092 551 251 350 1524Betulaceae 2 076 216 098 137 991Theaceae 2 053 146 067 093 969

Berberidaceae 2 139 037 017 024 947Adoxaceae 1 095 033 015 021 992

Sapindaceae 1 029 470 214 298 877

33 Forest Structure

Overall we recorded the median stem DBH of 1542 plusmn 111 cm stem density of124627 plusmn 56828 haminus1 and the basal area of 5435 plusmn 444 m2 haminus1 (Figure 8) Among forest communitiesthe HF recorded the highest median stem DBH (2089 plusmn 309 cm) but showed the lowest stem densitySimilarly MF recorded the highest stem density of about 1320 plusmn 11972 haminus1 with the second highest

Forests 2019 10 633 11 of 17

value for the stem DBH and basal area However the LF recorded the second highest value for medianstem density and the minimum value for the median DBH and basal areaForests 2019 10 x FOR PEER REVIEW 12 of 18

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem density and (c) DBH The horizontal line crossing the box in the center represents the median the box represents the 25th and the 75th percentiles The vertical line outside the box represents the minimum and maximum values ALL= total forest community LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure9) with the highest number of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least number of individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution was documented for different forest communities However they differed in the number of DBH classes MF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded in the study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classes and completely lacked trees above 93 cm DBH

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem densityand (c) DBH The horizontal line crossing the box in the center represents the median the box representsthe 25th and the 75th percentiles The vertical line outside the box represents the minimum andmaximum values ALL = total forest community LF = low-altitude forest MF = mid-altitude forestHF = high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure 9) with the highestnumber of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least numberof individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution wasdocumented for different forest communities However they differed in the number of DBH classesMF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded inthe study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classesand completely lacked trees above 93 cm DBH

Forests 2019 10 633 12 of 17

Forests 2019 10 x FOR PEER REVIEW 13 of 18

Figure 9 Class distribution of tress withgt3 cmDBH recorded in the study(a) Overall stems recorded in the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH ranging from gt93 cm to 303 cm LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change of forest tree species recorded in the Himalayan forests like Sikkim The finding could be related to the close association of altitude with climatic variables temperature and precipitation [53] and thus draws attention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communities between 900 masland3200 m asl Moreover according to the Grierson and Long [54]classification our study resembles three forest types ie LF forest as the warm broad-leaved forest MF as the evergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded some differences in species recorded for high-elevation forests For example Grierson and Longrsquos [54] classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall) Boiss as the major trees for high-elevation forests however we did not report similar findings A reason in the change in species composition in these three forest types compared to the 1960s and 1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for its use as scaffolding timber for construction of houses firewood and clearing for forest

Figure 9 Class distribution of tress with gt3 cmDBH recorded in the study (a) Overall stems recordedin the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH rangingfrom gt93 cm to 303 cm LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change offorest tree species recorded in the Himalayan forests like Sikkim The finding could be related to theclose association of altitude with climatic variables temperature and precipitation [53] and thus drawsattention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communitiesbetween 900 masl and 3200 m asl Moreover according to the Grierson and Long [54] classificationour study resembles three forest types ie LF forest as the warm broad-leaved forest MF as theevergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded somedifferences in species recorded for high-elevation forests For example Grierson and Longrsquos [54]classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall)Boiss as the major trees for high-elevation forests however we did not report similar findingsA reason in the change in species composition in these three forest types compared to the 1960s and

Forests 2019 10 633 13 of 17

1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for itsuse as scaffolding timber for construction of houses firewood and clearing for forest for agriculturepractice as well as climate change effects however warrants more investigation Future researchshould focus on documentation of forest trees for a detailed description of the Himalayan forest andrevision of forest type classifications if needed

42 Tree Species Richness and Diversity

Species accumulation curves are especially useful to decide sampling completeness and alsoto compare richness for communities with different sample sizes [50] In the current analysis theaccumulation curves did not record saturation when plotted for different communities which impliesfinding more species with increasing sampling effort Nevertheless it is clear that low-altitude forestsrecord the highest number of species richness and evenness and the disparity between the observedand rarified richness in the MF was the mere result of the sampling size differences The varied speciesrichness between forests covering different altitudes can be associated with numerous factors Broadlywe assumed the widely reported climatic variables namely precipitation temperature and theirinteraction as the prime factor for varied richness along the altitudinal gradient of the Himalayansystem [465556] Another crucial factor could be anthropogenic disturbance and its intensity [2738]Future research could explore these variables to explain the difference in richness recorded in our studyThe LF recorded not only the highest number of species but recorded more uniformly distributed treespecies However the region lacks comparable literature to discuss optimum richness and evennessWe recommend long-term monitoring of permanent plots which is completely lacking in the IndianEastern Himalayas

One of the important findings of our study was the significance of the family Fagaceae It wasrecorded as having the highest FIVI and SIVI and dominated the forest along the altitudinal gradientbetween 900 m and3200 m asl in Sikkim The reasons should be further investigated becauseecological and silvicultural knowledge of Fagaceae and its species are very limited for the EasternHimalayan forests Furthermore we also showed for the first time the existence of huge trees ofLithocarpus pachyphylls (gt15 m DBH) in high number which implies that this species was the remnantof the old growth primary forests in the altitude from 2500 to 3200 in Sikkim Himalayas Theseold-growth trees should be marked registered monitored and conserved as habitat trees by theforest department

However not all forest communities were dominated by the Fagaceae species For example Schimawallichii and Castanopsis tribuloides dominated the lower altitude forests whereas species like Symplocosramosissima and Castanopsis hystrix dominated the mid-altitude forests Different physiological andclimatic adaptation of a species may have caused distinct species to dominate different altituderesulting in three unique clusters of our samples Moreover the Himalayas are renowned for thepronounced variation in climatic topographic and edaphic factors along the altitudinal gradientApart from these the anthropogenic disturbance is a widely discussed factor for the dominance offast-growing pioneer species like the Symplocos racemose [57]Future research could explore the role ofanthropogenic disturbance and different climatic and topographic factors on species dominance andcomposition along the altitudinal gradient

43 Forest Structure Stem Diameter Stem Density and Basal Area

Size class distribution of trees provides the population structure of forest [58] and are extensivelyused to understand regeneration status [59] The present assessment presented a reverse J-shapeddistribution suggesting uneven-aged forests for sustainable reproduction and regeneration [60] with asufficient number of young individuals to replace the old mature stand However between forest typesthe low-altitude forests completely lacked trees above 93 cm diameter a scenario undesirable for asustainable forest The finding could be associated with forest degradation and deforestation reportedby earlier studies [4761] One potential factor for the absence of large trees could be the practice of

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 9 of 17

Table 2 Species richness density basal area Species Importance Value Index (SIVI) FamiliesImportance Value Index (FIVI) and their proportion to the contribution for the top ten species andfamilies for the overall forest community recorded in the study

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

Lithocarpus pachyphyllus Fagaceae 5120 411 1231 2107 930Symplocos ramosissima Symplocaceae 24807 1919 161 276 872

Castanopsis hystrix Fagaceae 6205 498 991 1697 823Castanopsis tribuloides Fagaceae 6241 501 422 723 499

Schima wallichii Theaceae 7819 627 240 411 446Rhododendron arboreum Ericaceae 8554 686 254 435 433

Viburnum erubescens Adoxaceae 10398 834 038 064 416Eurya acuminata DC Pentaphylacaceae 6096 489 081 139 391

Quercus lamellosa Fagaceae 988 079 495 849 386Symplocos heishanensis Symplocaceae 4940 396 038 065 293

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 9 2088 1675 3361 5719 2786Symplocaceae 4 3355 2692 218 371 1226

Theaceae 2 1390 1116 320 545 804Ericaceae 7 1476 1184 381 648 770Lauraceae 13 463 371 497 845 662Adoxaceae 2 1041 835 038 064 464

Magnoliaceae 4 112 090 209 364 270Betulaceae 4 193 155 099 169 219

Primulaceae 3 228 183 041 023 210Sapindaceae 1 069 055 123 210 189

Table 3 Stem density basal area Family (FIVI) and Species Importance Value Indexes (SIVI) and theirproportion to the contribution of the ten dominant species for three forest communities LF MF and HF

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

LF (900 m to 1700 m asl)

Schima wallichii Theaceae 25917 2401 708 2673 1890Castanopsis tribuloides Fagaceae 20250 1876 768 2900 1791

Engelhardtia spicata Juglandaceae 6042 560 235 886 672Gynocardia odorata Achariaceae 9250 857 123 463 605Euryaa cuminata Pentaphylacaceae 6667 857 076 286 483Betula alnoides

Buch-HamexDDon Betulaceae 2125 618 099 375 257

Macaranga denticulataBlume Euphorbiaceae 1808 197 039 148 243

Albizia lebbeck (L)Benth Leguminosae 1708 158 123 464 231

Alnus nepalensis DDon Betulaceae 1208 066 100 377 221Brassaiopsis hispida

Seem Araliaceae 3625 336 019 053 187

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Theaceae 2 942 3018 784 2960 2254Fagaceae 6 673 2157 823 3109 2017

Juglandaceae 3 180 575 238 899 742Achariaceae 1 267 857 123 463 658Betulaceae 3 104 332 209 789 526

Euphorbiaceae 5 123 394 059 222 434Araliaceae 3 157 502 017 066 353

Leguminosae 3 030 096 127 480 290Anacardiaceae 3 059 189 048 181 276

Lauraceae 6 063 201 009 035 264

Forests 2019 10 633 10 of 17

Table 3 Cont

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

MF (gt1700 m to 2500 m asl)

Symplocos ramosissima Symplocaceae 50189 3399 336 467 1509Castanopsis hystrix Fagaceae 13243 897 2172 3015 1499

Viburnum erubescens Adoxaceae 21189 1435 075 105 709Quercus lamellosa Fagaceae 2054 139 1067 1481 685Eurya acuminata Pentaphylacaceae 8378 567 123 171 442

Lithocarpus pachyphyllus Fagaceae 4892 331 517 718 432Symplocos heishanensis Symplocaceae 7378 500 056 078 407Castanopsis tribuloides Fagaceae 865 059 449 623 278

Quercus glauca Fagaceae 1432 097 309 429 270Elaeocarpus sikkimensis

Mast Elaeocarpaceae 1054 071 221 307 252

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 7 1082 1644 4570 4570 3012Symplocaceae 4 2890 4391 433 433 2014

Lauraceae 11 308 469 960 960 922Adoxaceae 2 946 1437 075 075 807Theaceae 2 395 600 171 171 573

Elaeocarpaceae 2 048 073 223 223 317Ericaceae 5 249 379 049 049 272

Sapindaceae 1 040 060 150 150 260Primulaceae 3 124 189 016 016 260

Magnoliaceae 4 059 090 175 175 243

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

HF (2500 m to 3200 m asl)

Lithocarpus pachyphyllus Fagaceae 11091 1066 788 5456 2437Rhododendron arboreum Ericaceae 31000 2978 747 1308 1678

Magnolia campbellii Magnoliaceae 1682 162 539 662 454Symplocos heishanensis Symplocaceae 6182 594 581 069 414Rhododendron hodgsonii

Hookf Ericaceae 5455 524 290 315 376

Corylus ferox Wall Betulaceae 2682 258 415 133 268Viburnum erubescens

Wall Adoxaceae 3591 345 415 021 260

Acer campbellii Sapindaceae 1091 105 373 298 259Rhododendron falconeri

Hookf Ericaceae 3955 380 207 186 258

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 4 333 9011 4096 5716 7921Ericaceae 7 1182 2963 1347 1879 7216

Symplocaceae 3 464 207 094 131 2602Magnoliaceae 2 048 1090 495 691 1603

Lauraceae 6 092 551 251 350 1524Betulaceae 2 076 216 098 137 991Theaceae 2 053 146 067 093 969

Berberidaceae 2 139 037 017 024 947Adoxaceae 1 095 033 015 021 992

Sapindaceae 1 029 470 214 298 877

33 Forest Structure

Overall we recorded the median stem DBH of 1542 plusmn 111 cm stem density of124627 plusmn 56828 haminus1 and the basal area of 5435 plusmn 444 m2 haminus1 (Figure 8) Among forest communitiesthe HF recorded the highest median stem DBH (2089 plusmn 309 cm) but showed the lowest stem densitySimilarly MF recorded the highest stem density of about 1320 plusmn 11972 haminus1 with the second highest

Forests 2019 10 633 11 of 17

value for the stem DBH and basal area However the LF recorded the second highest value for medianstem density and the minimum value for the median DBH and basal areaForests 2019 10 x FOR PEER REVIEW 12 of 18

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem density and (c) DBH The horizontal line crossing the box in the center represents the median the box represents the 25th and the 75th percentiles The vertical line outside the box represents the minimum and maximum values ALL= total forest community LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure9) with the highest number of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least number of individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution was documented for different forest communities However they differed in the number of DBH classes MF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded in the study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classes and completely lacked trees above 93 cm DBH

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem densityand (c) DBH The horizontal line crossing the box in the center represents the median the box representsthe 25th and the 75th percentiles The vertical line outside the box represents the minimum andmaximum values ALL = total forest community LF = low-altitude forest MF = mid-altitude forestHF = high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure 9) with the highestnumber of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least numberof individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution wasdocumented for different forest communities However they differed in the number of DBH classesMF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded inthe study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classesand completely lacked trees above 93 cm DBH

Forests 2019 10 633 12 of 17

Forests 2019 10 x FOR PEER REVIEW 13 of 18

Figure 9 Class distribution of tress withgt3 cmDBH recorded in the study(a) Overall stems recorded in the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH ranging from gt93 cm to 303 cm LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change of forest tree species recorded in the Himalayan forests like Sikkim The finding could be related to the close association of altitude with climatic variables temperature and precipitation [53] and thus draws attention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communities between 900 masland3200 m asl Moreover according to the Grierson and Long [54]classification our study resembles three forest types ie LF forest as the warm broad-leaved forest MF as the evergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded some differences in species recorded for high-elevation forests For example Grierson and Longrsquos [54] classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall) Boiss as the major trees for high-elevation forests however we did not report similar findings A reason in the change in species composition in these three forest types compared to the 1960s and 1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for its use as scaffolding timber for construction of houses firewood and clearing for forest

Figure 9 Class distribution of tress with gt3 cmDBH recorded in the study (a) Overall stems recordedin the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH rangingfrom gt93 cm to 303 cm LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change offorest tree species recorded in the Himalayan forests like Sikkim The finding could be related to theclose association of altitude with climatic variables temperature and precipitation [53] and thus drawsattention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communitiesbetween 900 masl and 3200 m asl Moreover according to the Grierson and Long [54] classificationour study resembles three forest types ie LF forest as the warm broad-leaved forest MF as theevergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded somedifferences in species recorded for high-elevation forests For example Grierson and Longrsquos [54]classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall)Boiss as the major trees for high-elevation forests however we did not report similar findingsA reason in the change in species composition in these three forest types compared to the 1960s and

Forests 2019 10 633 13 of 17

1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for itsuse as scaffolding timber for construction of houses firewood and clearing for forest for agriculturepractice as well as climate change effects however warrants more investigation Future researchshould focus on documentation of forest trees for a detailed description of the Himalayan forest andrevision of forest type classifications if needed

42 Tree Species Richness and Diversity

Species accumulation curves are especially useful to decide sampling completeness and alsoto compare richness for communities with different sample sizes [50] In the current analysis theaccumulation curves did not record saturation when plotted for different communities which impliesfinding more species with increasing sampling effort Nevertheless it is clear that low-altitude forestsrecord the highest number of species richness and evenness and the disparity between the observedand rarified richness in the MF was the mere result of the sampling size differences The varied speciesrichness between forests covering different altitudes can be associated with numerous factors Broadlywe assumed the widely reported climatic variables namely precipitation temperature and theirinteraction as the prime factor for varied richness along the altitudinal gradient of the Himalayansystem [465556] Another crucial factor could be anthropogenic disturbance and its intensity [2738]Future research could explore these variables to explain the difference in richness recorded in our studyThe LF recorded not only the highest number of species but recorded more uniformly distributed treespecies However the region lacks comparable literature to discuss optimum richness and evennessWe recommend long-term monitoring of permanent plots which is completely lacking in the IndianEastern Himalayas

One of the important findings of our study was the significance of the family Fagaceae It wasrecorded as having the highest FIVI and SIVI and dominated the forest along the altitudinal gradientbetween 900 m and3200 m asl in Sikkim The reasons should be further investigated becauseecological and silvicultural knowledge of Fagaceae and its species are very limited for the EasternHimalayan forests Furthermore we also showed for the first time the existence of huge trees ofLithocarpus pachyphylls (gt15 m DBH) in high number which implies that this species was the remnantof the old growth primary forests in the altitude from 2500 to 3200 in Sikkim Himalayas Theseold-growth trees should be marked registered monitored and conserved as habitat trees by theforest department

However not all forest communities were dominated by the Fagaceae species For example Schimawallichii and Castanopsis tribuloides dominated the lower altitude forests whereas species like Symplocosramosissima and Castanopsis hystrix dominated the mid-altitude forests Different physiological andclimatic adaptation of a species may have caused distinct species to dominate different altituderesulting in three unique clusters of our samples Moreover the Himalayas are renowned for thepronounced variation in climatic topographic and edaphic factors along the altitudinal gradientApart from these the anthropogenic disturbance is a widely discussed factor for the dominance offast-growing pioneer species like the Symplocos racemose [57]Future research could explore the role ofanthropogenic disturbance and different climatic and topographic factors on species dominance andcomposition along the altitudinal gradient

43 Forest Structure Stem Diameter Stem Density and Basal Area

Size class distribution of trees provides the population structure of forest [58] and are extensivelyused to understand regeneration status [59] The present assessment presented a reverse J-shapeddistribution suggesting uneven-aged forests for sustainable reproduction and regeneration [60] with asufficient number of young individuals to replace the old mature stand However between forest typesthe low-altitude forests completely lacked trees above 93 cm diameter a scenario undesirable for asustainable forest The finding could be associated with forest degradation and deforestation reportedby earlier studies [4761] One potential factor for the absence of large trees could be the practice of

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 10 of 17

Table 3 Cont

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

MF (gt1700 m to 2500 m asl)

Symplocos ramosissima Symplocaceae 50189 3399 336 467 1509Castanopsis hystrix Fagaceae 13243 897 2172 3015 1499

Viburnum erubescens Adoxaceae 21189 1435 075 105 709Quercus lamellosa Fagaceae 2054 139 1067 1481 685Eurya acuminata Pentaphylacaceae 8378 567 123 171 442

Lithocarpus pachyphyllus Fagaceae 4892 331 517 718 432Symplocos heishanensis Symplocaceae 7378 500 056 078 407Castanopsis tribuloides Fagaceae 865 059 449 623 278

Quercus glauca Fagaceae 1432 097 309 429 270Elaeocarpus sikkimensis

Mast Elaeocarpaceae 1054 071 221 307 252

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 7 1082 1644 4570 4570 3012Symplocaceae 4 2890 4391 433 433 2014

Lauraceae 11 308 469 960 960 922Adoxaceae 2 946 1437 075 075 807Theaceae 2 395 600 171 171 573

Elaeocarpaceae 2 048 073 223 223 317Ericaceae 5 249 379 049 049 272

Sapindaceae 1 040 060 150 150 260Primulaceae 3 124 189 016 016 260

Magnoliaceae 4 059 090 175 175 243

Species Family Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea SIVI

HF (2500 m to 3200 m asl)

Lithocarpus pachyphyllus Fagaceae 11091 1066 788 5456 2437Rhododendron arboreum Ericaceae 31000 2978 747 1308 1678

Magnolia campbellii Magnoliaceae 1682 162 539 662 454Symplocos heishanensis Symplocaceae 6182 594 581 069 414Rhododendron hodgsonii

Hookf Ericaceae 5455 524 290 315 376

Corylus ferox Wall Betulaceae 2682 258 415 133 268Viburnum erubescens

Wall Adoxaceae 3591 345 415 021 260

Acer campbellii Sapindaceae 1091 105 373 298 259Rhododendron falconeri

Hookf Ericaceae 3955 380 207 186 258

Families Richness Stem Density(haminus1)

StemDensity

Basal Area(m2 haminus1)

BasalArea FIVI

Fagaceae 4 333 9011 4096 5716 7921Ericaceae 7 1182 2963 1347 1879 7216

Symplocaceae 3 464 207 094 131 2602Magnoliaceae 2 048 1090 495 691 1603

Lauraceae 6 092 551 251 350 1524Betulaceae 2 076 216 098 137 991Theaceae 2 053 146 067 093 969

Berberidaceae 2 139 037 017 024 947Adoxaceae 1 095 033 015 021 992

Sapindaceae 1 029 470 214 298 877

33 Forest Structure

Overall we recorded the median stem DBH of 1542 plusmn 111 cm stem density of124627 plusmn 56828 haminus1 and the basal area of 5435 plusmn 444 m2 haminus1 (Figure 8) Among forest communitiesthe HF recorded the highest median stem DBH (2089 plusmn 309 cm) but showed the lowest stem densitySimilarly MF recorded the highest stem density of about 1320 plusmn 11972 haminus1 with the second highest

Forests 2019 10 633 11 of 17

value for the stem DBH and basal area However the LF recorded the second highest value for medianstem density and the minimum value for the median DBH and basal areaForests 2019 10 x FOR PEER REVIEW 12 of 18

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem density and (c) DBH The horizontal line crossing the box in the center represents the median the box represents the 25th and the 75th percentiles The vertical line outside the box represents the minimum and maximum values ALL= total forest community LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure9) with the highest number of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least number of individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution was documented for different forest communities However they differed in the number of DBH classes MF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded in the study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classes and completely lacked trees above 93 cm DBH

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem densityand (c) DBH The horizontal line crossing the box in the center represents the median the box representsthe 25th and the 75th percentiles The vertical line outside the box represents the minimum andmaximum values ALL = total forest community LF = low-altitude forest MF = mid-altitude forestHF = high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure 9) with the highestnumber of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least numberof individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution wasdocumented for different forest communities However they differed in the number of DBH classesMF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded inthe study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classesand completely lacked trees above 93 cm DBH

Forests 2019 10 633 12 of 17

Forests 2019 10 x FOR PEER REVIEW 13 of 18

Figure 9 Class distribution of tress withgt3 cmDBH recorded in the study(a) Overall stems recorded in the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH ranging from gt93 cm to 303 cm LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change of forest tree species recorded in the Himalayan forests like Sikkim The finding could be related to the close association of altitude with climatic variables temperature and precipitation [53] and thus draws attention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communities between 900 masland3200 m asl Moreover according to the Grierson and Long [54]classification our study resembles three forest types ie LF forest as the warm broad-leaved forest MF as the evergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded some differences in species recorded for high-elevation forests For example Grierson and Longrsquos [54] classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall) Boiss as the major trees for high-elevation forests however we did not report similar findings A reason in the change in species composition in these three forest types compared to the 1960s and 1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for its use as scaffolding timber for construction of houses firewood and clearing for forest

Figure 9 Class distribution of tress with gt3 cmDBH recorded in the study (a) Overall stems recordedin the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH rangingfrom gt93 cm to 303 cm LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change offorest tree species recorded in the Himalayan forests like Sikkim The finding could be related to theclose association of altitude with climatic variables temperature and precipitation [53] and thus drawsattention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communitiesbetween 900 masl and 3200 m asl Moreover according to the Grierson and Long [54] classificationour study resembles three forest types ie LF forest as the warm broad-leaved forest MF as theevergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded somedifferences in species recorded for high-elevation forests For example Grierson and Longrsquos [54]classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall)Boiss as the major trees for high-elevation forests however we did not report similar findingsA reason in the change in species composition in these three forest types compared to the 1960s and

Forests 2019 10 633 13 of 17

1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for itsuse as scaffolding timber for construction of houses firewood and clearing for forest for agriculturepractice as well as climate change effects however warrants more investigation Future researchshould focus on documentation of forest trees for a detailed description of the Himalayan forest andrevision of forest type classifications if needed

42 Tree Species Richness and Diversity

Species accumulation curves are especially useful to decide sampling completeness and alsoto compare richness for communities with different sample sizes [50] In the current analysis theaccumulation curves did not record saturation when plotted for different communities which impliesfinding more species with increasing sampling effort Nevertheless it is clear that low-altitude forestsrecord the highest number of species richness and evenness and the disparity between the observedand rarified richness in the MF was the mere result of the sampling size differences The varied speciesrichness between forests covering different altitudes can be associated with numerous factors Broadlywe assumed the widely reported climatic variables namely precipitation temperature and theirinteraction as the prime factor for varied richness along the altitudinal gradient of the Himalayansystem [465556] Another crucial factor could be anthropogenic disturbance and its intensity [2738]Future research could explore these variables to explain the difference in richness recorded in our studyThe LF recorded not only the highest number of species but recorded more uniformly distributed treespecies However the region lacks comparable literature to discuss optimum richness and evennessWe recommend long-term monitoring of permanent plots which is completely lacking in the IndianEastern Himalayas

One of the important findings of our study was the significance of the family Fagaceae It wasrecorded as having the highest FIVI and SIVI and dominated the forest along the altitudinal gradientbetween 900 m and3200 m asl in Sikkim The reasons should be further investigated becauseecological and silvicultural knowledge of Fagaceae and its species are very limited for the EasternHimalayan forests Furthermore we also showed for the first time the existence of huge trees ofLithocarpus pachyphylls (gt15 m DBH) in high number which implies that this species was the remnantof the old growth primary forests in the altitude from 2500 to 3200 in Sikkim Himalayas Theseold-growth trees should be marked registered monitored and conserved as habitat trees by theforest department

However not all forest communities were dominated by the Fagaceae species For example Schimawallichii and Castanopsis tribuloides dominated the lower altitude forests whereas species like Symplocosramosissima and Castanopsis hystrix dominated the mid-altitude forests Different physiological andclimatic adaptation of a species may have caused distinct species to dominate different altituderesulting in three unique clusters of our samples Moreover the Himalayas are renowned for thepronounced variation in climatic topographic and edaphic factors along the altitudinal gradientApart from these the anthropogenic disturbance is a widely discussed factor for the dominance offast-growing pioneer species like the Symplocos racemose [57]Future research could explore the role ofanthropogenic disturbance and different climatic and topographic factors on species dominance andcomposition along the altitudinal gradient

43 Forest Structure Stem Diameter Stem Density and Basal Area

Size class distribution of trees provides the population structure of forest [58] and are extensivelyused to understand regeneration status [59] The present assessment presented a reverse J-shapeddistribution suggesting uneven-aged forests for sustainable reproduction and regeneration [60] with asufficient number of young individuals to replace the old mature stand However between forest typesthe low-altitude forests completely lacked trees above 93 cm diameter a scenario undesirable for asustainable forest The finding could be associated with forest degradation and deforestation reportedby earlier studies [4761] One potential factor for the absence of large trees could be the practice of

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 11 of 17

value for the stem DBH and basal area However the LF recorded the second highest value for medianstem density and the minimum value for the median DBH and basal areaForests 2019 10 x FOR PEER REVIEW 12 of 18

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem density and (c) DBH The horizontal line crossing the box in the center represents the median the box represents the 25th and the 75th percentiles The vertical line outside the box represents the minimum and maximum values ALL= total forest community LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure9) with the highest number of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least number of individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution was documented for different forest communities However they differed in the number of DBH classes MF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded in the study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classes and completely lacked trees above 93 cm DBH

Figure 8 Boxplot with jittered data points showing the distribution of (a) basal area (b) stem densityand (c) DBH The horizontal line crossing the box in the center represents the median the box representsthe 25th and the 75th percentiles The vertical line outside the box represents the minimum andmaximum values ALL = total forest community LF = low-altitude forest MF = mid-altitude forestHF = high-altitude forest

The overall size class distribution showed a reverse ldquoJrdquo-shaped pattern (Figure 9) with the highestnumber of individuals in the smallest DBH class between 3 to 13 cm (Figure 9a) and the least numberof individuals in the largest DBH class between 293ndash303 cm (Figure 9b) A similar distribution wasdocumented for different forest communities However they differed in the number of DBH classesMF recorded the maximum number of DBH classes with the biggest tree of 295 cm DBH recorded inthe study LF and MF lacked comparable DBH trees The LF recorded the least number of DBH classesand completely lacked trees above 93 cm DBH

Forests 2019 10 633 12 of 17

Forests 2019 10 x FOR PEER REVIEW 13 of 18

Figure 9 Class distribution of tress withgt3 cmDBH recorded in the study(a) Overall stems recorded in the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH ranging from gt93 cm to 303 cm LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change of forest tree species recorded in the Himalayan forests like Sikkim The finding could be related to the close association of altitude with climatic variables temperature and precipitation [53] and thus draws attention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communities between 900 masland3200 m asl Moreover according to the Grierson and Long [54]classification our study resembles three forest types ie LF forest as the warm broad-leaved forest MF as the evergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded some differences in species recorded for high-elevation forests For example Grierson and Longrsquos [54] classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall) Boiss as the major trees for high-elevation forests however we did not report similar findings A reason in the change in species composition in these three forest types compared to the 1960s and 1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for its use as scaffolding timber for construction of houses firewood and clearing for forest

Figure 9 Class distribution of tress with gt3 cmDBH recorded in the study (a) Overall stems recordedin the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH rangingfrom gt93 cm to 303 cm LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change offorest tree species recorded in the Himalayan forests like Sikkim The finding could be related to theclose association of altitude with climatic variables temperature and precipitation [53] and thus drawsattention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communitiesbetween 900 masl and 3200 m asl Moreover according to the Grierson and Long [54] classificationour study resembles three forest types ie LF forest as the warm broad-leaved forest MF as theevergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded somedifferences in species recorded for high-elevation forests For example Grierson and Longrsquos [54]classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall)Boiss as the major trees for high-elevation forests however we did not report similar findingsA reason in the change in species composition in these three forest types compared to the 1960s and

Forests 2019 10 633 13 of 17

1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for itsuse as scaffolding timber for construction of houses firewood and clearing for forest for agriculturepractice as well as climate change effects however warrants more investigation Future researchshould focus on documentation of forest trees for a detailed description of the Himalayan forest andrevision of forest type classifications if needed

42 Tree Species Richness and Diversity

Species accumulation curves are especially useful to decide sampling completeness and alsoto compare richness for communities with different sample sizes [50] In the current analysis theaccumulation curves did not record saturation when plotted for different communities which impliesfinding more species with increasing sampling effort Nevertheless it is clear that low-altitude forestsrecord the highest number of species richness and evenness and the disparity between the observedand rarified richness in the MF was the mere result of the sampling size differences The varied speciesrichness between forests covering different altitudes can be associated with numerous factors Broadlywe assumed the widely reported climatic variables namely precipitation temperature and theirinteraction as the prime factor for varied richness along the altitudinal gradient of the Himalayansystem [465556] Another crucial factor could be anthropogenic disturbance and its intensity [2738]Future research could explore these variables to explain the difference in richness recorded in our studyThe LF recorded not only the highest number of species but recorded more uniformly distributed treespecies However the region lacks comparable literature to discuss optimum richness and evennessWe recommend long-term monitoring of permanent plots which is completely lacking in the IndianEastern Himalayas

One of the important findings of our study was the significance of the family Fagaceae It wasrecorded as having the highest FIVI and SIVI and dominated the forest along the altitudinal gradientbetween 900 m and3200 m asl in Sikkim The reasons should be further investigated becauseecological and silvicultural knowledge of Fagaceae and its species are very limited for the EasternHimalayan forests Furthermore we also showed for the first time the existence of huge trees ofLithocarpus pachyphylls (gt15 m DBH) in high number which implies that this species was the remnantof the old growth primary forests in the altitude from 2500 to 3200 in Sikkim Himalayas Theseold-growth trees should be marked registered monitored and conserved as habitat trees by theforest department

However not all forest communities were dominated by the Fagaceae species For example Schimawallichii and Castanopsis tribuloides dominated the lower altitude forests whereas species like Symplocosramosissima and Castanopsis hystrix dominated the mid-altitude forests Different physiological andclimatic adaptation of a species may have caused distinct species to dominate different altituderesulting in three unique clusters of our samples Moreover the Himalayas are renowned for thepronounced variation in climatic topographic and edaphic factors along the altitudinal gradientApart from these the anthropogenic disturbance is a widely discussed factor for the dominance offast-growing pioneer species like the Symplocos racemose [57]Future research could explore the role ofanthropogenic disturbance and different climatic and topographic factors on species dominance andcomposition along the altitudinal gradient

43 Forest Structure Stem Diameter Stem Density and Basal Area

Size class distribution of trees provides the population structure of forest [58] and are extensivelyused to understand regeneration status [59] The present assessment presented a reverse J-shapeddistribution suggesting uneven-aged forests for sustainable reproduction and regeneration [60] with asufficient number of young individuals to replace the old mature stand However between forest typesthe low-altitude forests completely lacked trees above 93 cm diameter a scenario undesirable for asustainable forest The finding could be associated with forest degradation and deforestation reportedby earlier studies [4761] One potential factor for the absence of large trees could be the practice of

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 12 of 17

Forests 2019 10 x FOR PEER REVIEW 13 of 18

Figure 9 Class distribution of tress withgt3 cmDBH recorded in the study(a) Overall stems recorded in the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH ranging from gt93 cm to 303 cm LF= low-altitude forest MF= mid-altitude forest HF= high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change of forest tree species recorded in the Himalayan forests like Sikkim The finding could be related to the close association of altitude with climatic variables temperature and precipitation [53] and thus draws attention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communities between 900 masland3200 m asl Moreover according to the Grierson and Long [54]classification our study resembles three forest types ie LF forest as the warm broad-leaved forest MF as the evergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded some differences in species recorded for high-elevation forests For example Grierson and Longrsquos [54] classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall) Boiss as the major trees for high-elevation forests however we did not report similar findings A reason in the change in species composition in these three forest types compared to the 1960s and 1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for its use as scaffolding timber for construction of houses firewood and clearing for forest

Figure 9 Class distribution of tress with gt3 cmDBH recorded in the study (a) Overall stems recordedin the study (b) DBH ranging from gt3 to 93 cm among three forest communities and (c) DBH rangingfrom gt93 cm to 303 cm LF = low-altitude forest MF = mid-altitude forest HF = high-altitude forest

4 Discussion

41 Forest Communities and Species Composition

The cluster analysis highlighted the significance of altitude in driving compositional change offorest tree species recorded in the Himalayan forests like Sikkim The finding could be related to theclose association of altitude with climatic variables temperature and precipitation [53] and thus drawsattention towards the susceptibility of Himalayan forests to climate change

Additionally cluster analysis also implied the presence of three distinct forest communitiesbetween 900 masl and 3200 m asl Moreover according to the Grierson and Long [54] classificationour study resembles three forest types ie LF forest as the warm broad-leaved forest MF as theevergreen oak forest and HF as the temperate mixed forests [54] Nevertheless we recorded somedifferences in species recorded for high-elevation forests For example Grierson and Longrsquos [54]classification reported Tsuga dumosa (DDon)Eichler Larix graffithii Hookf and Picea smithiana (Wall)Boiss as the major trees for high-elevation forests however we did not report similar findingsA reason in the change in species composition in these three forest types compared to the 1960s and

Forests 2019 10 633 13 of 17

1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for itsuse as scaffolding timber for construction of houses firewood and clearing for forest for agriculturepractice as well as climate change effects however warrants more investigation Future researchshould focus on documentation of forest trees for a detailed description of the Himalayan forest andrevision of forest type classifications if needed

42 Tree Species Richness and Diversity

Species accumulation curves are especially useful to decide sampling completeness and alsoto compare richness for communities with different sample sizes [50] In the current analysis theaccumulation curves did not record saturation when plotted for different communities which impliesfinding more species with increasing sampling effort Nevertheless it is clear that low-altitude forestsrecord the highest number of species richness and evenness and the disparity between the observedand rarified richness in the MF was the mere result of the sampling size differences The varied speciesrichness between forests covering different altitudes can be associated with numerous factors Broadlywe assumed the widely reported climatic variables namely precipitation temperature and theirinteraction as the prime factor for varied richness along the altitudinal gradient of the Himalayansystem [465556] Another crucial factor could be anthropogenic disturbance and its intensity [2738]Future research could explore these variables to explain the difference in richness recorded in our studyThe LF recorded not only the highest number of species but recorded more uniformly distributed treespecies However the region lacks comparable literature to discuss optimum richness and evennessWe recommend long-term monitoring of permanent plots which is completely lacking in the IndianEastern Himalayas

One of the important findings of our study was the significance of the family Fagaceae It wasrecorded as having the highest FIVI and SIVI and dominated the forest along the altitudinal gradientbetween 900 m and3200 m asl in Sikkim The reasons should be further investigated becauseecological and silvicultural knowledge of Fagaceae and its species are very limited for the EasternHimalayan forests Furthermore we also showed for the first time the existence of huge trees ofLithocarpus pachyphylls (gt15 m DBH) in high number which implies that this species was the remnantof the old growth primary forests in the altitude from 2500 to 3200 in Sikkim Himalayas Theseold-growth trees should be marked registered monitored and conserved as habitat trees by theforest department

However not all forest communities were dominated by the Fagaceae species For example Schimawallichii and Castanopsis tribuloides dominated the lower altitude forests whereas species like Symplocosramosissima and Castanopsis hystrix dominated the mid-altitude forests Different physiological andclimatic adaptation of a species may have caused distinct species to dominate different altituderesulting in three unique clusters of our samples Moreover the Himalayas are renowned for thepronounced variation in climatic topographic and edaphic factors along the altitudinal gradientApart from these the anthropogenic disturbance is a widely discussed factor for the dominance offast-growing pioneer species like the Symplocos racemose [57]Future research could explore the role ofanthropogenic disturbance and different climatic and topographic factors on species dominance andcomposition along the altitudinal gradient

43 Forest Structure Stem Diameter Stem Density and Basal Area

Size class distribution of trees provides the population structure of forest [58] and are extensivelyused to understand regeneration status [59] The present assessment presented a reverse J-shapeddistribution suggesting uneven-aged forests for sustainable reproduction and regeneration [60] with asufficient number of young individuals to replace the old mature stand However between forest typesthe low-altitude forests completely lacked trees above 93 cm diameter a scenario undesirable for asustainable forest The finding could be associated with forest degradation and deforestation reportedby earlier studies [4761] One potential factor for the absence of large trees could be the practice of

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 13 of 17

1970s could be related to increasing anthropogenic pressure such as open grazing tree felling for itsuse as scaffolding timber for construction of houses firewood and clearing for forest for agriculturepractice as well as climate change effects however warrants more investigation Future researchshould focus on documentation of forest trees for a detailed description of the Himalayan forest andrevision of forest type classifications if needed

42 Tree Species Richness and Diversity

Species accumulation curves are especially useful to decide sampling completeness and alsoto compare richness for communities with different sample sizes [50] In the current analysis theaccumulation curves did not record saturation when plotted for different communities which impliesfinding more species with increasing sampling effort Nevertheless it is clear that low-altitude forestsrecord the highest number of species richness and evenness and the disparity between the observedand rarified richness in the MF was the mere result of the sampling size differences The varied speciesrichness between forests covering different altitudes can be associated with numerous factors Broadlywe assumed the widely reported climatic variables namely precipitation temperature and theirinteraction as the prime factor for varied richness along the altitudinal gradient of the Himalayansystem [465556] Another crucial factor could be anthropogenic disturbance and its intensity [2738]Future research could explore these variables to explain the difference in richness recorded in our studyThe LF recorded not only the highest number of species but recorded more uniformly distributed treespecies However the region lacks comparable literature to discuss optimum richness and evennessWe recommend long-term monitoring of permanent plots which is completely lacking in the IndianEastern Himalayas

One of the important findings of our study was the significance of the family Fagaceae It wasrecorded as having the highest FIVI and SIVI and dominated the forest along the altitudinal gradientbetween 900 m and3200 m asl in Sikkim The reasons should be further investigated becauseecological and silvicultural knowledge of Fagaceae and its species are very limited for the EasternHimalayan forests Furthermore we also showed for the first time the existence of huge trees ofLithocarpus pachyphylls (gt15 m DBH) in high number which implies that this species was the remnantof the old growth primary forests in the altitude from 2500 to 3200 in Sikkim Himalayas Theseold-growth trees should be marked registered monitored and conserved as habitat trees by theforest department

However not all forest communities were dominated by the Fagaceae species For example Schimawallichii and Castanopsis tribuloides dominated the lower altitude forests whereas species like Symplocosramosissima and Castanopsis hystrix dominated the mid-altitude forests Different physiological andclimatic adaptation of a species may have caused distinct species to dominate different altituderesulting in three unique clusters of our samples Moreover the Himalayas are renowned for thepronounced variation in climatic topographic and edaphic factors along the altitudinal gradientApart from these the anthropogenic disturbance is a widely discussed factor for the dominance offast-growing pioneer species like the Symplocos racemose [57]Future research could explore the role ofanthropogenic disturbance and different climatic and topographic factors on species dominance andcomposition along the altitudinal gradient

43 Forest Structure Stem Diameter Stem Density and Basal Area

Size class distribution of trees provides the population structure of forest [58] and are extensivelyused to understand regeneration status [59] The present assessment presented a reverse J-shapeddistribution suggesting uneven-aged forests for sustainable reproduction and regeneration [60] with asufficient number of young individuals to replace the old mature stand However between forest typesthe low-altitude forests completely lacked trees above 93 cm diameter a scenario undesirable for asustainable forest The finding could be associated with forest degradation and deforestation reportedby earlier studies [4761] One potential factor for the absence of large trees could be the practice of

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 14 of 17

cardamom cultivation mainly in the low-altitude forest of Sikkim Cardamom is a high-valued cashcrop and it is intensively cultivated in the low- and mid-altitude forests where forests are partiallycleared for its cultivation [61] Furthermore the drying of cardamom demands a substantial supply offuelwood often generated from the nearby forests Moreover until 2004 Sikkim produced the highestamong of large cardamom (Amomum subulatum Roxb) in India with the largest share in the worldrsquosmarket [62] number of trees felled must been profound

5 Conclusions

Our study provides a comprehensive and quantitative understanding of diversity structureand composition of forest trees along the altitudinal gradient ranging from 900 m asl to 3200 masl in the Sikkim Himalayas The study estimated high-species richness in the low-altitude forestoccurring at 900 m asl to 1700 m asl whereas the high-altitude forest covering 2500 m asl to3200 m asl recorded the highest mean stem DBH and stand basal area Furthermore we found thatFagaceae trees are the most significant with a dominant contribution towards the total basal area andstem density Hence Fagaceae trees could prove as having a potential for carbon storage and climatechange mitigation in the Himalayas At the same time large-sized Fagaceae trees may have veryhigh biodiversity values which are yet to be quantified One of the significant findings of our studywas the disproportionate size class distribution within forests at different altitudes The low-altitudeforests completely lacked trees above 93 cm DBH Overall the study highlights the need to conservelow-altitude forests to maintain diversity and improve forest structure We suggest further researchto understand the factors associated with varied diversity composition and uneven forest structurerecorded in the study

Author Contributions YB conceptualized the research concept methodology collected field data conducted allthe statistical analyses and wrote the original draft SS RG (Ravikanth Gudasalamani) and RG (RengaianGanesan) supervised and reviewed the article RG (Rengaian Ganesan) helped with plant identification SSacquired funding for the Article Processing Charge

Funding The Department of Biotechnology Government of India funded the fieldwork and data collection(Grant No BT01NEPSNCBS09) The work was partially funded by the National Mission for HimalayanStudies (NMHS) under the Ministry of Environment Forest and Climate Change Government of India(NMHS2015ndash16HF1111)The International Union of Forest Research Organizations Vienna (IUFRO) awardedthe Young Scientist Award to YB for a short-term visit to the Thuumlnen Institute Germany The Karlsruhe Instituteof Technology Germany funded the Article Processing Charge

Acknowledgments YB acknowledges the support of the Department of Forest Environment and WildlifeManagement Department Government of Sikkim including various officials of the concerned Sanctuary forpermits Further YB is grateful to Janga Bir Subba Director (retd) Khangchendzonga National Park for providingpermits field crew and a forest guest house during field data collection YB also gratefully acknowledges theDepartment of Science and Technology for sharing satellite images during the initial phase of the fieldwork YBalso thanks Robert John Chandran for acquiring the grant from the Department of Biotechnology and guidanceduring fieldwork YB gratefully acknowledges IUFRO for awarding her the IUFRO-EFI Young Scientists Initiativeaward of 2019 Lastly YB thanks Roshan Subba field assistant for the genuine enthusiasm and support in thefield SS acknowledges his colleagues in IUFROrsquos task force on ldquoForest Restoration and Adaptation under GlobalChangerdquo for their helpful suggestions on the scope of forest restoration research in the Himalayas

Conflicts of Interest The authors declare no conflict of interest

References

1 Condit R Sukumar R Hubbell SP Foster RB Predicting population trends from size distributionsA direct test in a tropical tree community Am Nat 1998 152 495ndash509 [CrossRef] [PubMed]

2 Sharma E Tse-ring K Chettri N Shrestha A Kathmandu N Biodiversity in the Himalayasndashtrendsperception and impacts of climate change In Proceedings of the International Mountain BiodiversityConference Kathmandu Nepal 16ndash18 November 2008

3 Chaudhary P Bawa KS Local perceptions of climate change validated by scientific evidence in theHimalayas Biol Lett 2011 7 767ndash770 [CrossRef] [PubMed]

4 Pandit MK The Himalayas must be protected Nat News 2013 501 283 [CrossRef] [PubMed]

Forests 2019 10 633 15 of 17

5 Gautam MR Timilsina GR Acharya K Climate Change in the Himalayas Current State of Knowledge TheWorld Bank Washington DC USA 2013

6 Myers N Mittermeier RA Mittermeier CG Da Fonseca GA Kent J Biodiversity hotspots forconservation priorities Nature 2000 403 853ndash858 [CrossRef] [PubMed]

7 Mittermeier RA Myers N Mittermeier CG Robles G Hotspots Earthrsquos Biologically Richest and MostEndangered Terrestrial Ecoregions CEMEX SA Agrupacioacuten Sierra Madre SC Mexico City Mexico 1999

8 Upadhaya K Structure and floristic composition of subtropical broad-leaved humid forest of Cherapunjeein Meghalaya Northeast India J Biodivers Manag For 2015 4 2 [CrossRef]

9 Brown S Sathaye J Cannell M Kauppi PE Mitigation of carbon emissions to the atmosphere by forestmanagement Commonw For Rev 1996 75 80ndash91

10 Haripriya G Biomass carbon of truncated diameter classes in Indian forests For Ecol Manag 2002 1681ndash13 [CrossRef]

11 Saha D Sundriyal RC Utilization of non-timber forest products in humid tropics Implications formanagement and livelihood For Policy Econ 2012 14 28ndash40 [CrossRef]

12 Diaz HF Grosjean M Graumlich L Climate variability and change in high elevation regions Past presentand future Clim Chang 2003 59 1ndash4 [CrossRef]

13 Williams JW Jackson ST Kutzbach JE Projected distributions of novel and disappearing climates by2100 AD Proc Natl Acad Sci USA 2007 104 5738ndash5742 [CrossRef]

14 Shrestha UB Gautam S Bawa KS Widespread climate change in the Himalayas and associated changesin local ecosystems PLoS ONE 2012 7 e36741 [CrossRef] [PubMed]

15 Krishnaswamy J John R Joseph S Consistent response of vegetation dynamics to recent climate changein tropical mountain regions Glob Chang Biol 2014 20 203ndash215 [CrossRef] [PubMed]

16 Chakraborty A Saha S Sachdeva K Joshi PK Vulnerability of forests in the Himalayan region to climatechange impacts and anthropogenic disturbances A systematic review Reg Environ Chang 2018 181783ndash1799 [CrossRef]

17 Xu J Grumbine RE Shrestha A Eriksson M Yang X Wang Y Wilkes A The melting HimalayasCascading effects of climate change on water biodiversity and livelihoods Conserv Biol 2009 23 520ndash530[CrossRef]

18 Khan ML Menon S Bawa KS Effectiveness of the protected area network in biodiversity conservationA case-study of Meghalaya state Biodivers Conserv 1997 6 853ndash868 [CrossRef]

19 Shaheen H Ullah Z Khan SM Harper DM Species composition and community structure of westernHimalayan moist temperate forests in Kashmir For Ecol Manag 2012 278 138ndash145 [CrossRef]

20 Peet RK The measurement of species diversity Annu Rev Ecol Syst 1974 5 285ndash307 [CrossRef]21 Khan M Rai J Tripathi R Population structure of some tree species in disturbed and protected subtropical

forests of north-east India Acta Oecolog 1987 8 247ndash25522 Brown S Measuring carbon in forests Current status and future challenges Environ Pollut 2002 116

363ndash372 [CrossRef]23 Chave J Andalo C Brown S Cairns MA Chambers J Eamus D Foumllster H Fromard F Higuchi N

Kira T et al Tree allometry and improved estimation of carbon stocks and balance in tropical forestsOecologia 2005 145 87ndash99 [CrossRef]

24 Forrester DI Tachauer IHH Annighoefer P Barbeito I Pretzsch H Ruiz-Peinado R Stark HVacchiano G Zlatanov T Chakraborty T et al Generalized biomass and leaf area allometric equationsfor European tree species incorporating stand structure tree age and climate For Ecol Manag 2017 396160ndash175 [CrossRef]

25 Pandey KP Structure Composition and Diversity of Forest along the Altitudinal Gradient in the HimalayasNepal Appl Ecol Environ Res 2016 14 235ndash251 [CrossRef]

26 Panthi MP Chaudhary RP Vetaas OR Plant species richness and composition in a trans-Himalayaninner valley of Manang district central Nepal Himal J Sci 2007 4 57ndash64 [CrossRef]

27 Tenzin J Hasenauer H Tree species composition and diversity in relation to anthropogenic disturbances inbroad-leaved forests of Bhutan Int J Biodivers Sci Ecosyst Serv Manag 2016 12 274ndash290 [CrossRef]

28 Mukhia PK Wangyal JT Gurung D Floristic composition and species diversity of the chirpine forestecosystem Lobesa Western Bhutan For Nepal 2011 1ndash4

Forests 2019 10 633 16 of 17

29 Wangda P Ohsawa M Structure and regeneration dynamics of dominant tree species along altitudinalgradient in a dry valley slopes of the Bhutan Himalaya For Ecol Manag 2006 230 136ndash150 [CrossRef]

30 Tang CQ Li Y-H Zhang Z-Y Species Diversity Patterns in Natural Secondary Plant Communities andMan-made Forests in a Subtropical Mountainous Karst Area Yunnan SW China Mt Res Dev 2010 30244ndash251 [CrossRef]

31 Li R Dao Z Li H Seed Plant Species Diversity and Conservation in the Northern Gaoligong Mountainsin Western Yunnan China Mt Res Dev 2011 31 160ndash165 [CrossRef]

32 Bharali S Paul A Khan ML Singha LB Species diversity and community structure of a temperatemixed Rhododendron forest along an altitudinal gradient in West Siang District of Arunachal Pradesh IndiaNat Sci 2011 9 125ndash140

33 Nath P Arunachalam A Khan M Arunachalam K Barbhuiya A Vegetation analysis and treepopulation structure of tropical wet evergreen forests in and around Namdapha National Park northeastIndia Biodivers Conserv 2005 14 2109ndash2135 [CrossRef]

34 Yam G Tripathi OP Tree diversity and community characteristics in Talle Wildlife Sanctuary ArunachalPradesh Eastern Himalaya India J Asia-Pac Biodivers 2016 9 160ndash165 [CrossRef]

35 Tripathi O Pandey H Tripathi R Distribution community characteristics and tree population structure ofsubtropical pine forest of Meghalaya northeast India Int J Ecol Environ Sci 2004 29 207ndash214

36 Tripathi OP Tripathi RS Community composition structure and management of subtropical vegetationof forests in Meghalaya State northeast India Int J Biodivers Sci Ecosyst Serv Manag 2011 6 157ndash163[CrossRef]

37 Sinha S Badola HK Chhetri B Gaira KS Lepcha J Dhyani PP Effect of altitude and climate in shapingthe forest compositions of Singalila National Park in Khangchendzonga Landscape Eastern Himalaya IndiaJ Asia-Pac Biodivers 2018 11 267ndash275 [CrossRef]

38 Gogoi A Sahoo UK Impact of anthropogenic disturbance on species diversity and vegetation structure ofa lowland tropical rainforest of eastern Himalaya India J Mt Sci 2018 15 2453ndash2465 [CrossRef]

39 Dutta G Devi A Plant diversity population structure and regeneration status in disturbed tropical forestsin Assam northeast India J For Res 2013 24 715ndash720 [CrossRef]

40 Sundriyal RC Sharma E Rai LK Rao SC Tree structure regeneration and woody biomass removal ina subtropical forest of Mamlay watershed in the Sikkim Himalaya Vegetatio 1994 113 53ndash63 [CrossRef]

41 Singh HB Sundriyal RC Composition Economic Use and Nutrient Contents of Alpine Vegetation in theKhangchendzonga Biosphere Reserve Sikkim Himalaya India Arct Antarct Alp Res 2005 37 591ndash601[CrossRef]

42 Tambe S Rawat GS The Alpine Vegetation of the Khangchendzonga Landscape Sikkim HimalayaMt Res Dev 2010 30 266ndash274 [CrossRef]

43 Pandey A Rai S Kumar D Changes in vegetation attributes along an elevation gradient towards timberlinein Khangchendzonga National Park Sikkim Trop Ecol 2018 59 259ndash271

44 Chettri N Sharma E Deb DC Sundriyal RC Impact of Firewood Extraction on Tree StructureRegeneration and Woody Biomass Productivity in a Trekking Corridor of the Sikkim Himalaya Mt Res Dev2002 22 150ndash158 [CrossRef]

45 Acharya BK Chettri B Vijayan L Distribution pattern of trees along an elevation gradient of EasternHimalaya India Acta Oecolog 2011 37 329ndash336 [CrossRef]

46 Sharma N Behera MD Das AP Panda RM Plant richness pattern in an elevation gradient in theEastern Himalaya Biodivers Conserv 2019 28 2085ndash2104 [CrossRef]

47 Tambe S Arrawatia ML Sharma N Assessing the Priorities for Sustainable Forest Management in theSikkim Himalaya India A Remote Sensing Based Approach J Indian Soc Remote Sens 2011 39 555ndash564[CrossRef]

48 Hijmans RJ Cameron SE Parra JL Jones PG Jarvis A Very high resolution interpolated climatesurfaces for global land areas Int J Climatol 2005 25 1965ndash1978 [CrossRef]

49 Legendre P Legendre LF Numerical Ecology Elsevier Amsterdam The Netherlands 2012 Volume 2450 Gotelli NJ Colwell RK Estimating species richness Biol Divers Front Meas Assess 2011 12 39ndash5451 Keel S Gentry AH Spinzi L Using vegetation analysis to facilitate the selection of conservation sites in

eastern Paraguay Conserv Biol 1993 7 66ndash75 [CrossRef]

Forests 2019 10 633 17 of 17

52 Ganesh T Ganesan R Devy MS Davidar P Bawa KS Assessment of plant biodiversity at a mid-elevationevergreen forest of Kalakad-Mundanthurai Tiger Reserve Western Ghats India Curr Sci 1996 71 379ndash392

53 Korner C The use of lsquoaltitudersquo in ecological research Trends Ecol Evol 2007 22 569ndash574 [CrossRef]54 Grierson AJ Long DG Flora of Bhutan Including a Record of Plants from Sikkim Royal Botanic Garden

Edinburgh UK 198355 Manish K Telwala Y Nautiyal DC Pandit MK Modelling the impacts of future climate change on

plant communities in the Himalaya A case study from Eastern Himalaya India Model Earth Syst Environ2016 2 92 [CrossRef]

56 Kharkwal G Mehrotra P Rawat YS Pangtey YPS Phytodiversity and growth form in relation toaltitudinal gradient in the Central Himalayan (Kumaun) region of India Curr Sci 2005 89 873ndash878

57 Mishra B Tripathi O Tripathi R Pandey H Effects of anthropogenic disturbance on plant diversity andcommunity structure of a sacred grove in Meghalaya northeast India Biodivers Conserv 2004 13 421ndash436[CrossRef]

58 Saxena AK Singh JS Tree population-structure of certain Himalayan forest associations and implicationsconcerning their future composition Vegetatio 1984 58 61ndash69 [CrossRef]

59 Veblen TT Structure and dynamics of nothofagus forests near timberline in south-central Chile Ecology1979 60 937ndash945

60 Vetaas OR The effect of environmental factors on the regeneration of Quercus semecarpifolia Sm in CentralHimalaya Nepal Plant Ecol 2000 146 137ndash144 [CrossRef]

61 Kanade R John R Topographical influence on recent deforestation and degradation in the Sikkim Himalayain India Implications for conservation of East Himalayan broadleaf forest Appl Geogr 2018 92 85ndash93[CrossRef]

62 Gudade B Chhetri P Gupta U Deka T Vijayan A Traditional practices of large cardamom cultivationin Sikkim and Darjeeling Life Sci Leafl 2013 9 62ndash68

copy 2019 by the authors Licensee MDPI Basel Switzerland This article is an open accessarticle distributed under the terms and conditions of the Creative Commons Attribution(CC BY) license (httpcreativecommonsorglicensesby40)