Regeneration dynamics, population structure, and forest composition in some ridge forests of the Western Himalaya, India

ABSTRACT The present study aimed to analyze the regeneration dynamics, population structure, and forest composition in some ridge forests of the Western Himalaya to elucidate the impact of climate change. The highest species richness (trees, 17; saplings, 16; seedlings, 16) was recorded in mixed Rhododendron arboreum forest. The maximum tree density (737 ± 25.93 trees ha−1; mixed Quercus floribunda); sapling density (5342 ± 587.54 saplings ha−1; mixed Quercus semecarpifolia), and seedling density (71,429 ± 26,632.29 seedlings ha−1; mixed Cedrus deodara) were recorded in forests of higher altitudes. The mean basal cover values in these forests oscillated between 39.21 ± 1.53 m2 ha−1 (mixed Quercus leucotrichophora) to 87.65 ± 15.45 m2 ha−1 (mixed Abies pindrow). The highest Shannon index value (0.51 ± 0.05) and species evenness (0.36 ± 0.02) for trees were recorded in mixed Rhododendron arboreum forest. Some species like Betula utilis, Myrica esculenta, Ficus rumphii, Ilex dipyrena, Quercus floribunda, Litsea elongata, Symplocos paniculata, and Abies spectabilis were noticed as new recruiters (recent invaders) in new habitats, which may change the future forest composition. This study revealed that ridge forests of the Bhagirathi catchment area were in transition phase at middle altitudes. Mixed Quercus floribunda forest was observed to be the youngest regenerating forest; however, mixed Abies pindrow forest was the most mature old-growth forest.


Introduction
Himalayan mountain forests are one of the most fascinating and characteristic entities among forests of the world because of their unique ecology, having a temperate climate within a tropical zone. As mountain regions cover about 24% of total global land area (UNEP-WCMC 2002), and there have been reports of rapid climate change in mountain regions during the past few decades (IPCC 2007), understanding the changes in structural and functional attributes of forests on ridge tops across a wide range of elevational gradients will help us better predict species migration and future productivity of the forests. Several studies have attributed widespread changes in plant growth or mortality to climate change, but these efforts have focused on general trends within a biome rather than identifying spatially coherent distribution patterns (Pauli et al. 2007;Engler et al. 2009).
Himalayan ridge top ecosystems are considered to be more sensitive to climate change as they are characterized by uniform sunlight exposure and low human interferences, but variation in altitude give birth to climatic differences and subsequent vegetation changes, and hence are perfect places for monitoring and comparing the effects of climate change and predicting future changes in species composition. Furthermore, it is supposed that, in the event of a rise in the temperature at lower elevations, the movement/migration of vegetation would be towards upper elevational ridge tops. This is understandable because of ecological relationships, alteration in plant life history, and general upward shift in the species distribu-tional ranges (McKone et al. 1998;Klanderund 2005;Jurasinski and Kreyling 2007;Pauli et al. 2012).
Species association and composition on ridge top forests in Himalayan ranges depend more on climatic factors, which are not well studied. Plant life at higher elevations is mostly governed by abiotic factors such as temperature and snowfall; therefore, ridge top vegetation composition is greatly affected by these factors (Kammer and Mohl 2002). Consequently, changes in climate are projected to cause changes in vegetation distribution ranges. The lack of evidence of widespread plant range shifts may reflect the limited dispersal of plants or it may simply indicate the paucity of baseline records of plant distributions (Criddle et al. 2003).
Climatic variations on ridge tops may result either in adaptation of species to newer conditions or the migration of species uphill, leading to gradual changes in forest composition (Sharma et al. 2014(Sharma et al. , 2016a. Therefore, continuous monitoring of the population structure of various species in ridge top forests will pave the way for the development and management of forest composition in the near future. In the Western Himalayan region along elevational transects changes in forest composition are evident, but they need to be properly monitored/measured (Chitale et al. 2014). Therefore, the aim of this study was to describe and analyze the forest structure, composition, and distribution patterns on the ridge tops along elevational gradients in the Bhagirathi catchment area in order to explain the changes in forest composition and character in response to changing climate.

Study area
The state of Uttarakhand is situated in the northern part of India and shares international boundaries with China to the north and Nepal to the east. It has an area of 53,483 km 2 and lies between latitude 28 43 0 and 31 28 0 N and longitude 77 34 0 and 81 03 0 E. The state has a temperate climate except in the plains area, where the climate is tropical. Of the total geographical area of the state, about 19% is under permanent snow cover having glaciers and steep slopes where tree growth is not possible due to climatic and physical limitations (FSI 2009). The recorded forest cover of the state is 24,240 km 2 , which constitutes 45.32% of its geographical area (FSI 2015).
The study was carried out in various moist temperate mixed ridge forests of Bhagirathi catchment area in the Western Himalaya along an altitudinal gradient 1450-3540 m above sea level (a.s.l.). The selected ridge tops in the study area were comprised of raised hills and ridges, which were different in topography and were more prominent in the upper reaches of the Bhagirathi catchment area. These ridge tops have different tree composition at different altitudes.
The study area lies in two different districts (Uttarkashi and Tehri) of Uttarakhand state, including various mountain ranges: Dayara-Gidara (DG); Kanatal-Dhanolti (KD); Thang-Harunta (TH); and Harshil-Dharali (HD). We assessed 60 ridge top forests in which the following were predominantly distributed: (i) mixed Quercus semecarpifolia forest (MQSF); (ii) mixed Q. floribunda forest (MQFF); (iii) mixed Q. leucotrichophora forest (MQLF); (iv) mixed Rhododendron arboreum forest (MRAF); (v) mixed Abies pindrow forest (MAPF); and (vi) mixed Cedrus deodara forest (MCDF). The geographical details of these forests are given in Figure 1 and Table 1. The climate of the entire study area is moist temperate type with mean minimum monthly temperatures ranging from 7.12 C (January) to 23.20 C (July) and mean maximum monthly temperatures from 17.56 C (January) to 33.35 C (July); mean annual rainfall is 2000 mm ). There are three main distinct seasons: (i) cool and relatively dry winter (December to March); (ii) warm and dry summer (mid-April to June); and (iii) monsoon or rainy season, a warm and wet period (July to mid-September).

Vegetation sampling and data analysis
The vegetation was sampled under three different growth phases, namely: adult tree (30 cm circumference at breast height [CBH]); sapling (10 cm to 30 cm CBH); and seedlings (30 cm height, but <10 cm CBH) following Saxena et al. (1984). For analysis of forest vegetation, population structure, and regeneration pattern in different moist temperate mixed ridge forests, 10 sample plots of 0.1 ha each were randomly laid out in each forest type (10 ridge tops each £ six forest types = 60 sample plots). The sample plots of 0.1 ha were laid out by determining the plot center in such a way as to minimize the slope area on the selected ridge site. But in certain ridge tops where sample plots were located on a slope of >10%, the slope angle was measured using a clinometer and an adjustment was made to the plot area at the time of analysis. The true horizontal distance for those arms going against the slope was calculated using the formula: where L is the true horizontal plot distance, Ls is the standard distance measured in the field along the slope, S is the slope in degrees and cos is the cosine of the angle. The actual area of the sample plot was then calculated as: Within each 0.1 ha plot, 5 m £ 5 m sized quadrats were laid out randomly to analyze the saplings. However, 1 m £ 1 m sized quadrats were used for analyzing tree seedlings (Curtis and McIntosh 1950;Phillips 1959). The measurement was done on individual seedlings, saplings, and trees and the CBH (1.37m from the ground) was used for the determination of tree total basal cover (TBC). To know the complete structure and composition of the forest, the quantitative parameters frequency, density, TBC, and importance value index (IVI) were calculated following Cottam and Curtis (1956). On the basis of data so obtained the Shannon-Wiener diversity index (Shannon and Weaver 1963), Simpson dominance index (Simpson 1949), and evenness (Pielou 1966) for each of the forest types were also determined. To unravel the population structure and regeneration pattern in each forest type diameter at breast height (DBH) classes were established on the basis of range of available data. The density of trees, saplings, and seedlings was calculated for 100 m 2 area of each forest to verify the regeneration status. Different regeneration categories (fair, good, not, new, and poor) were created to assess the regeneration potential of each species. Regeneration of tree species was determined based on population size of seedlings, saplings, and adults and categorized as per Dutta and Devi (2013) modified from Khan et al. (1987), Shankar (2001, and Khumbongmayum et al. (2006). The categories identified were: (1) Good regeneration: if seedlings > saplings > adults (2) Fair regeneration: if seedlings > or saplings adults if seedlings saplings > adults if seedlings saplings and the species had no adults (3) Poor regeneration: if a species survives only in the sapling stage, but no seedlings (though saplings may be <, >, or = adults) (4) No regeneration: if it is absent both in seedling and sapling stages but found only in adults (5) New regeneration: if the species has no adult, but only saplings or seedlings.

Statistical analysis
One way analysis of variance (ANOVA) was used to test the elevation-wise differences (significant/non-significant) of various attributes/parameters of different ridge forests. Tukey's post-hoc test was applied to test differences among means when F-test was significant (P 0.05). The statistical analysis was performed by using SPSS v22 (SPSS Inc., Chicago, IL).
In mixed Q. leucotrichophora forest, 46.15% of the total species were observed in the good regeneration category; however, 38.46% were recorded in the fair category. On the other hand, in mixed Q. floribunda forest (21.43%), mixed C. deodara forest (18.18%), mixed R. arboreum forest (13.64%), and mixed Q. semecarpifolia forest (7.69%), a substantial number of species were recorded under the poor regeneration category. Betula utilis, Myrica esculenta, Ficus rumphii, Ilex dipyrena, Quercus floribunda, Litsea elongata, Symplocos paniculata, and Abies spectabilis were found as new regeneration in their respective forest types. Maximum species (38.46%) were not found regenerating in MQSF, whereas minimum regeneration (13.64%) was observed in MRAF.
The details are given in Table 3.

Discussion
In this study the species richness was highest in mixed Rhododendron arboreum forest (situated 1800-3230 m a.s.l.). Bharali et al. (2011) also reported high species richness (30 species) in mixed Rhododendron arboreum forest of Arunanchal Pradesh (1900-2300 m a.s.l.). The same pattern of species richness for temperate forests of Garhwal Himalaya was recorded by Negi et al. (2008). Generally, species richness declines with increasing altitudes as the atmosphere becomes relatively less dense and because only a few species can survive in the harsh climatic conditions found at higher altitudes. However, in this study we focused only on ridge forest ecosystems in which we found that, due to limited altitudinal distribution of MQLF, the growth of Q. leucotrichophora was restricted to a narrow range resulting in lower species richness. The Banj oak (Quercus leucotrichophora) was not found associated with the species of higher altitudes (cooler regions) such as Acer acuminatum, A. caesium, Abies pindrow, A. spectabilis, Taxus wallichiana, and Picea smithiana, preferring the lower altitudes to grow with many subtropical species. In contrast, R. arboreum was noticed growing over a wide altitudinal range mainly with oak and other broad leaved and coniferous species up to the higher subalpine ecotone (wider adaptability) which might be the reason for higher species richness in MRAF. Kumar and Ram (2005) had also emphasized that in temperate forests the mixed broad leaved forests sustained maximum species richness. The variability in distribution of plant species on ridge tops could be attributed to the effect of cofactors such as microclimate and edaphic factors within the same altitudinal range    Table 1. (Wright 1983;Currie 1991;Chaudhary 1999;Pande et al. 2002;Ellu and Obua 2005;Sharma et al. 2016bSharma et al. , 2017. The highest tree density (737 § 25.93 trees ha ¡1 ) was recorded in mixed Q. floribunda forest (2275-2547 m a.s.l.) due to the younger growing stage (more individuals at lower DBH classes) at these altitudes. This recorded value is higher than the earlier reported values of 240 trees ha ¡1 by Ram et al. (2004) from Kumaun Himalaya, 250-340 trees ha ¡1 by Baduni and Sharma (1999) and 366-466 trees ha ¡1 by Ghildiyal et al. (1998) from Garhwal Himalaya for Q. floribunda forest, but lower than reported by Singh et al. (1994) from Kumaun Himalaya. However, our values are similar to those reported by Bisht et al. (2013) (620 trees ha ¡1 ). The upper reaches of Bhagirathi catchment area are mostly undisturbed or least disturbed, therefore the density values are comparatively higher. The highest TBC value (87.65 § 15.45 m 2 ha ¡1 ) was recorded in mixed A. pindrow forest at higher elevations because these old growth stands were mature. Mature A. pindrow forests are valuable in terms of sustaining higher biomass which consequently enables bigger trees to accumulate more carbon from the atmosphere. Additionally, old growth forests develop ideal conditions/environment for regeneration of important plant species. The dead wood matter (woody debris) on the forest floor supplies enough nutrients to sustain forest diversity. These TBC values are higher than earlier reported values of 41.25 m 2 ha ¡1 by Gairola et al. (2011) and 35.66 m 2 ha ¡1 by Singhal et al. (1986), but similar to the recorded TBC values of 54-124 m 2 ha ¡1 (Baduni 1996) The highest Shannon diversity index values of various growth forms varied greatly in the present study: (i) tree layer (0.51 § 0.05) in MRAF; (ii) sapling layer (0.57 § 0.06) in MQLF; and (iii) seedling layer (0.53 § 0.03) in MQFF. Basically, the diversity in any forest depends on species richness and evenness in equal measure. Species evenness refers to the closeness of each species in their numbers. Consequently, high diversity in the tree layer was recorded in MRAF because the species were closely associated in their numbers. On the other hand this attribute was found to be similar in the seedling layer of MQFF. That is why MQFF exhibited high species diversity in the seedling layer in comparison to MRAF.
The highest sapling density (5342 § 587.54 saplings ha ¡1 ) in this study was recorded in MQSF, while in the seedling layer it was highest (71,429 § 26,632.29 seedlings ha ¡1 ) in MCDF. Gairola et al. (2012) reported 3100 saplings ha ¡1 in mainly Q. semecarpifolia forest between 2650 and 2850 m a.s. l. However, in high mountain ridge tops, gradual changes in vegetation structure and composition are expected as a consequence of changing environmental conditions along increasing altitudes. Low sapling density in MQFF (2275-2547 m a.s.l.) on ridge tops was noticed because of the seedlings' mortality and insufficient numbers at the sapling stage. Dense canopy in forests such as MQLF, MQFF, and MRAF contained lower densities of seedlings at various elevations. According to Harper (1977) and Harcombe (1987) the most drastic and observable change in life history of trees was noticed during the seedling stage. However, Dai et al. (2002) have suggested that soil moisture is a critical factor restricting the germination and survival of seedlings. The topographic position maintains seedling establishment in temperate forests and determines the species composition on ridge tops. Seiwa and Kikuzawa (1996) and Seiwa (1997) have also reported that seedling survival is often inhibited by water deficit and litter accumulation. Low seedling density at mid altitudes (1800-3230 m a.s.l.) in MRAF may be referred to low light intensity on the forest floor due to the dense canopy of broad leaved tree species. Barik et al. (1992) and Tripathi (2002) have also suggested that low light intensity on the ground floor due to dense overhead canopy may check the growth of seedlings. Although the response of tree seedlings to the changing climate has not yet been fully explored, Nautiyal et al. (2004) and Chaturvedi et al. (2007) observe that changes in snow pattern and fluctuating temperature affects the distribution and phenology of some plants.
This study revealed that most of the forests on ridge tops in upper Bhagirathi catchment area were in transition phase (demographic transition) because lower DBH classes of various tree species have grown to form the forest composition. Most of the overstory old growth trees had either died or no longer occupied the top canopy due to past disturbances and fragmentation of habitat. Consequently, this stage contains multiple tree sizes with different species composition and their DBH distribution varied greatly. The average recorded diameter in these forests was 31.38 cm, which fluctuated between 10.56 cm and 209.84 cm. The compositional changes recorded in these forests were due to the introduction of new species. Because of this, these forests sustain lower understory diversity of herbs and shrubs. The regeneration in these forests appear again (although their process of growth may slow) comparatively which was also reported by Oliver and Larson (1996). Although maximum tree density in each selected forest was observed in 21-40 cm DBH class , followed by <20 cm DBH class in MQFF (2275- s.l.) the highest densities were found in higher DBH class (41-60 cm) than <20 cm DBH class. Moreover, MQSF, MAPF, and MCDF were the only forests on ridge tops in which maximum densities lie in higher DBH classes (i.e., 81-100 cm and >100 cm), which clearly indicates that these ridge forests are well established and mature. In these forests mixed A. pindrow forest and mixed C. deodara forest were old-growth forests because they do not have fodder value, situated mostly in inaccessible areas and requiring less rainfall. In mixed Q. semecarpifolia forests any new regeneration has become a problem probably because of available climatic/ environmental conditions and vivipary between 2418 and 3540 m a.s.l. (Table 3). Sharma et al. (2016b) have also suggested that high altitudinal ridge forest ecosystem services are mainly regulated by climatic factors.
Our study suggests that Abies pindrow, Abies spectabilis, Acer acuminatum, Acer caesium, Betula alnoides, Cedrus deodara, Euonymus tingens, Ilex dipyrena, Lyonia ovalifolia, Myrica esculenta, Picea smithiana, Pinus roxburghii, Pinus wallichiana, Prunus cornuta, Pyrus pashia, Quercus floribunda, Q. leucotrichophora, Q. semecarpifolia, Rhododendron arboreum, Rhododendron barbatum, Symplocos paniculata, and Taxus wallichiana exist in the good regeneration category across various forest types. In this study we observed fairly good regeneration of various tree species in MQFF (2275-2547 m a.s.l.) and MAPF (2454-2830 m a.s.l.), which indicates the climatic adaptation of these species at these altitudes. In each forest type few tree species (ranging from 13.64% to 38.46%) were represented by tree stage only and their seedlings and saplings were completely absent along the altitudinal zones (Table 3). Benton and Werner (1976) suggested that such types of population could become extinct if this tendency continues. Some tree species like Betula alnoides, Lindera pulcherrima, and Lyonia ovalifolia were poorly regenerating species in MQFF (2275-2547 m a.s.l.); however, B. alnoides was found to be a good regenerating species in MQSF at comparatively higher altitudes (2418-3540 m a.s.l.) indicating that this tree species prefers comparatively higher altitudes and is now sharing habitat with mixed Quercus semecarpifolia forest. Lyonia ovalifolia also preferred higher altitudes but made a good association with Rhododendron arboreum forest (1800-3230 m a.s.l.). Lindera pulcherrima was poorly regenerating/not able to survive in association with Quercus floribunda and avoids comparatively cooler climatic conditions at intermediate altitudinal zone.
Although Acer acuminatum is conventionally a good associate of Q. semecarpifolia, A. pindrow, A. spectabilis, and R. arboreum in the temperate zone, in our study A. acuminatum expressed its poor regeneration in MQSF (2418-3540 m a.s. l.), and MRAF but showed good regeneration in MAPF (2454-2830 m a.s.l.). Gairola et al. (2011) also referred A. acuminatum as the dominant species in A. pindrow forest (2500-2600 m a.s.l.). However, in this study, A. acuminatum was not recorded as the top dominant species due to lower importance value and limited population size which was found growing well with A. pindrow. The average DBH of A. acuminatum was recorded as higher (mean 28.45 cm, ranging from 15.35 cm to 57.32 cm) in MAPF than in MQSF and MRAF. Consequently, the mature population of A. acuminatum was observed in MAPF, which had greater potential for seed production. The association of A. pindrow (less crown expanse species) with A. acuminatum possibly offers an adequate microenvironment including suitable soil nutrients for seed germination, which is responsible for good regeneration of A. acuminatum in MAPF. A few species such as Betula utilis, Myrica esculenta (in MQFF), Ficus rumphii, Ilex dipyrena, and Quercus floribunda (in MRAF), Litsea elongata, Myrica esculenta, and Symplocos paniculata (in MAPF), and Abies spectabilis (in MCDF) were recent invaders, which may gradually increase their populations and ultimately will change the forest composition in the near future.

Conclusions
The information generated through this research will help in establishing ridge tops as a potential site for long term ecological monitoring to evaluate the species range shift and compositional changes in high mountain ranges of the Himalaya. The mixed forest ecosystem is an example of a complete and organized structure of forest where all dominant associates contribute almost equally to constitute the stand structure and may be more helpful than the other forest types to evaluate variations in species composition and structure. Fairly good growth and survival of tree seedlings with adequate conversion into saplings of certain native and important tree species were noticed in the studied ranges. On the other hand, some tree species were unable to reproduce because they were represented by trees only (no seedlings or saplings) on ridge tops, which may be due to the scarcity of available soil moisture content required for seed germination. Forest managers and policy makers should take necessary steps to replant such species artificially (afforestation) so that their populations can be maintained. The other aspect of this study was that we have recorded some new regenerating tree species such as Abies spectabilis, Betula utilis, Ficus rumphii, Ilex dipryena, Litsea elongata, Myrica esculenta, Quercus floribunda, and Symplocos paniculata, which were found encroaching the new habitats on uphill slopes. The naturally grown seedlings and saplings (new introductions) of some broad leaved species in conifer forest stands should be considered as resources for future forest. Mixed Cedrus deodara forest was found as widely adapted forest (temperate to climate sensitive subalpine zone) on ridge tops, particularly in harsh climatic conditions. Species of subalpine zone such as Abies pindrow, Abies spectabilis, Acer acuminatum, Betula utilis, Quercus semecarpifolia, and Rhododendron arboreum were observed to expand their upper limits of growth to alpine meadows. Mixed Quercus floribunda forest was recorded as the youngest regenerating forest (2275-2547 m a.s.l.), followed by mixed Quercus leucotrichophora forest (1910-2482 m a.s.l.). On the other hand, mixed Abies pindrow forest was mature old growth forest (2454-3125 m a. s.l.), followed by mixed Quercus semecarpifolia forest (2418-3540 m a.s.l.) and mixed Cedrus deodara forest (1450-3125 m a.s.l.) in the Bhagirathi catchment area of Garhwal Himalaya.