Disturbance, Diversity, Regeneration and Composition in Temperate Forests of Western Himalaya, India

Abstract

We have investigated the impact of anthropogenic and natural disturbances on regeneration, composition and diversity in some temperate forests of Bhagirathi Catchment Area of Garhwal Himalaya. The forests were categorized on the basis of canopy cover and magnitude of disturbance into highly, moderately and least disturbed classes. The dominant tree species at lower elevation were Pinus roxburghii and Quercus leucotrichophora, while Abies pindrow, Q. semecarpifolia and Rhododenron arboreum were the dominant species at the upper elevational forests. Cythula tomentosa and Indegophera heterentha were the dominant shrub species present in all the forests. Similarly, Circium wallichii and Oxalis corniculata were the dominant herb species found in all forests (except Q. leucotrichophora forest), whereas Thalictrum foliolosum and Viola pilosa were noticed in each forest (except P. roxburghii forest). The tree density values oscillated between $400{\pm}10\;trees\;ha^{-1}$ to $750{\pm}89.1\;trees\;ha^{-1}$ which generally decreased from lower to higher disturbance regimes however, the total basal cover value was highest ($88.1{\pm}23.6m^2\;ha^{-1}$) in highly disturbed forest and lowest ($25.8{\pm}2.2m^2\;ha^{-1}$) in moderately disturbed forest. The shrub and herb densities were maximum in least disturbed forest, while the young regenerating individuals i.e., sapling and seedling were observed increasing from high to low disturbed forests which reflected that the forest fragmentation adversely affected the regeneration. However, A. pindrow and P. roxburghii were found invariably encroaching the habitats of R. arboreum and Q. leucotrichophora at various altitudes, respectively. The Canonical Correspondence Analysis clearly indicated that the elevation and lopping intensity have more impact on trees, while shrub and herbs were more influenced by elevation, canopy cover, light attenuation and soil erosion. Pinus roxburghii was the only species which was affected by heavy litter removal and forest fire.

Introduction

Deforestation is a global phenomenon resulting in loss of 10.6 million hectare of forest every year (Food and Agriculture Organization of the United Nations 2015) and is also responsible for degradation of floral and faunal diversity (Davidar et al. 2010). Both anthropogenic and natural disturbances affect ecosystem dynamics in many ways at local and regional scales (Sapkota et al. 2009), which also affect the regeneration potential of trees at a larger scale (Lawes et al. 2007). Many studies (Lawes et al. 2007) have focused on the relationship between disturbance and species richness and reflected how stand structure, composition and regeneration of tree species varied along disturbance gradient (Sapkota et al. 2009). Murphy and Lugo (1986) have suggested that dry tropical forests are among the most vulnerable ecosystems of the world and reflect slow vegetation recovery due to complex interactions among sites, propagules and climatic conditions. Several authors (Chaturvedi et al. 2011; Sagar et al. 2012; Bhadouria et al. 2016; Bhadouria et al. 2017) have studied the structure and functioning of Indian dry tropical forests. The quantification of tree species diversity is an important aspect as it provides habitat for many other species (Cannon et al. 1998). Compositional changes in vegetation are primarily determined by the elevational gradient (Baruch 1984), which constitutes a complex set of environmental conditions including temperature, precipitation and moisture (Körner 2007).

Fredericksen and Mastacedo (2000) have described that disturbance can alter environmental conditions by altering light availability and soil conditions. Pausas et al. (2006) have shown that disturbance favours the plant invasion because it provides a pulse of resources and pave the way for seedling establishment and growth. Therefore, it can be stated that the structure and composition of the forests are largely the results of past activities (Motta and Edouard 2005). Kennard et al. (2002) and Zhu et al. (2007) have opined that post-disturbance regimes cause variation in regeneration dynamics, dominance-diversity and composition.

Regeneration of trees within a community mainly depends on seed production, seed establishment, survivorship and establishment of seedlings (Cain and Shelton 2001). However, seed production of trees may be limited by various external factors such as resource availability, pollination failure, climatic factors and age and size of trees (Winn and Werner 1987). Seedling establishment and subsequent growth into the saplings is vital initial step towards formation of future canopy. Environmental variation within a small geographical area makes altitudinal gradients ideal for investigating several ecological and biogeographical hypotheses (Körner 1998). Forest composition, structure and diversity patterns are important ecological attributes, which are significantly correlated with prevailing environmental as well as anthropogenic variables (Ahmad et al. 2010). Criddle et al. (2003) have also supported that species diversity along with regeneration defines the nature of future community. Hence it is necessary to understand and evaluate the species richness, structure and regeneration ability of species for proper management and maintenance of forests (Khumbongmayum et al. 2006). Plant regeneration is generally known to be limited in harsh environment. In order to mitigate the ongoing challenges of deforestation and to conserve the various forest types, there is an urgent need to understand the structure and regeneration pattern of forests which are managed under the provisions of protected areas. Singh et al. (1984) and Singh and Singh (1992) have also emphasized that man made disturbances were widespread features in most of the forests in Himalaya. Through this study an attempt has been made to work out the impact of various human induced and natural disturbances on some selected temperate forests/vegetation types in Bhagirathi Catchment Area of Garhwal Himalaya.

Materials and Methods

Study area

The State of Uttarakhand is situated in the northern part of India and shares an international boundary with China in the north and Nepal in the east. It has an area of 53,483 km² and lies between latitudes 28° 43' and 31° 28'N and longitudes 77° 34' and 81° 03'E. The State has a temperate climate except in the plains 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 (Forest Survey of India 2009). The recorded forest area of the State is 24240 km² , which constitutes 45.32% of its geographical area (Forest Survey of India 2015). The present study was carried out in Ghuttu (lower altitudinal zone) and Gangi (higher altitudinal zone) mountain ranges of Tehri district in Bhagirathi Catchment area of Garhwal Himalaya, Uttarakhand. Uttarakhand Action Plan for Climate Change (UAPCC) had categorized the main drainage system of the state into six major catchment areas i.e., Yamuna, Bhagirathi, Alaknanda, Mandakini, Pindar and Kali Catchments and included the river Janhavi and Bhilangna as the main tributaries of Bhagirathi catchment (Uttarakhand Action Plan on Climate Change 2014). A reconnaissance survey of the study area was done in different seasons during the years 2014 to 2016. We have selected seven major forest types(located between 1933- 3038 m asl) in moist temperate region on different slopes and aspects along disturbance gradient. The climate of the study area comprises of three distinct seasons i.e., summer (April to June), monsoon/rainy (June to September) and cold winter (December to March). During the winters, when the surrounding ridges were covered with snow, the reported temperature was observed to fall down significantly and minimum goes below 0°C. In contrast, during summers, the temperature varied between 16°C to 35°C. Overall, the average temperature in the study area was observed between 4.6°C to 36.5°C (Aziem et al. 2016). The annual rainfall ranges from 1409 mm to 2000 mm with a mean of 1600 mm. The relative humidity in the region oscillated between 36% and 75% throughout the year. Lying in the upper Himalayas, these regions contain varying geographic environments ranging from snow free valleys to the high peaks with perpetual snow and glaciers. The details of the study area and forest types are given in Fig. 1 and Table 1 respectively

Fig. 1. Map representing the studied sites.

 Table 1. Geographical details of selected forest types

DC, disturbance category; LD, least disturbed; MD, moderately disturbed; HD, highly disturbed; IVI, importance value index.

Sampling and analysis

Disturbance

The various disturbance categories were formulated on the basis of canopy cover and frequency of anthropogenic activities in each forest type. The canopy cover was measured directly in the field within the sampling unit using spherical densiometer (Werner 2009) and expressed in percent ground cover area covered by the canopy. Densiometer contains 24 grids/squares and the reading were taken in four different directions facing North, East, South and West from the middle in each forest (24 squares×4 dots in each square of the grid=total 96 dots), representing small squares. The uncovered dots were counted and converted and multiplied by 1.04 (derived from 100/96) to obtain the percent of overhead area not occupied by canopy. The difference between this percent and 100% is the estimated overhead canopy cover (as per Werner 2009). 

In each quadrat natural and anthropogenic disturbances were noted by their presence (+) and absence (-). The frequency of occurrence was converted into percent to get an idea of various types of prevalent disturbances in each forest. The dung piles were directly observed and counted to know the frequency of cattle encroachment inside the forest zone. Further, the different disturbance categories were named viz., (i) highly disturbed (HD) {if frequency of human activities is high with open canopy (＜40%) i.e. lopping, grazing, litter removal, fire}, (ii) moderately disturbed (MD) {if the human activities are low with moderate canopy cover (40-75%)} and (iii) least disturbed (LD) {if human activities are almost nil or people used these forest in a particular time of the year with closed canopy (＞ 75%)}. In this study, the following parameters were recorded to quantify the woodcutting for timber, fuel wood and fodder lopping pressure as: (i) number of cut stumps in each quadrat, (ii) number of trees with lopping signs in terms of percentage, and (iii) the girth was measured at the cutting point of each cut stump. On the basis of available data of cut stumps and their TBC values the disturbance indices based on the density (DI D) and total basal cover (DI TBC) of cut stumps were also estimated following Murali et al. (1996).

DI D=

\({\text Density of cut stumps per hectare \over \text Total density (standing stems+cut stems) per hectare} \times 100\)

DI TBC=

\({\text TBC of cut stumps per hectare \over \text Total basal cover (standing stems+cut stems ) per hectare} \times 100\)

The light attenuation (L) within each forest type was measured with a digital lux meter (Lutron LX 101 LuxMeter). It was calculated following Chaturvedi et al. (2011) as follows:

Light attenuation=

\({\text Light intensity in open-Light intensity under canopy \over \text Light intensity in open} \times 100\)

Forest composition and diversity

A total of seven forest types having various forms and frequencies of disturbances along with variation in altitude, slope aspect and species composition were observed for the study (Table 1), which were categorized according to the forest composition by the dominant tree species according to Parkash (1986) viz., ≥75% as pure forest; 50-75% as mainly single species dominant forest; 25-50% as mixed forest, and ＜25% miscellaneous forest. Different physiographic factors, such as elevation and aspect across different forest cover types were measured by GPS (Garmin, Rino-130). The composition of the forest types were analysed by nested quadrat methods as per Kent and Coker (1992).The adult tree vegetation in each forest were analysed using 10 m×10 m sized quadrats (07 forest types× 10 quadrats=total 70 quadrats), shrubs and saplings by 5 m×5 m sized quadrats (07 forest types×10 quadrats×05 subquadrats=total 350 quadrats) and herbs by 1 m×1 m sized quadrats (07 forest types×10 quadrats×05 subquadrats=total 350 quadrats) as proposed by Curtis and McIntosh (1950) and Phillips (1959). A transect was established both horizontally and vertically within a 10 m× 10 m sized quadrat in which 2 m×2 m sized quadrat (07 forest types×10 quadrats×10 subquadrats=total 700 quadrats) were formulated to enumerate the tree seedlings. Species richness was simply taken as a count of the total number of species in that particular forest cover type. Additionally, for each species frequency, density, abundance and Importance Value Index (IVI) were also calculated following Cottam and Curtis (1956). The basal cover was calculated by dividing the square of Cbh (circumference at breast height, i.e., 1.37 m) by 4π. The basal cover was multiplied with respective densities of the species to obtain total tree basal cover. The importance value index (IVI) was calculated by summing up relative density, relative frequency, and relative dominance, following Phillips (1959). On the basis of data so obtained the Shannon-Wiener diversity Index (Shannon and Weaver 1963), Simpson dominance Index (Simpson 1949), equitability (Pielou 1966), Margalef Index of species richness (Margalef 1958), Menheink index of species richness (Whittaker 1977) were also calculated. The identification of plants was done with the help of existing taxonomic literature, described in Flora of Gangotri National Park (Pusalkar and Singh 2012), Flora of the district Garhwal North West Himalaya (Gaur 1999) and Herbarium of H.N.B. Garhwal University Srinagar (GUH).

Regeneration dynamics

The density of trees, saplings and seedlings was calculated in 100 m² area of each species to justify the regeneration status. Different regeneration categories (Fair, Good, New, Not and Poor regeneration) were created to know the regeneration dynamics of each species with various frequencies of disturbances in each forest cover type. Regeneration of woody 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

Canonical Corresponding Analysis (CCA) was used to identify the habitat sharing and impact of recorded disturbance parameters on each vegetation layer i.e., trees, shrubs and herbs. In order to assure objectivity of the gradient analysis, the species abundance was processed using the software CANOCO-5 (Microcomputer Power, Ithaca NY). Carl-Pearson Correlation coefficient was calculated between canonical axes of ordination plot (eigenvalues and explained variance) and disturbance parameters using XLStat.

Results

Disturbance

Forest canopy cover is one of the significant parameters to measure the disturbance of any geographical area. The mean canopy cover in highly disturbed forest was 32.4% whereas, 61.85% in moderately disturbed and 82.7% in least disturbed forests. The frequency of natural disturbance was higher in mainly A. pindrow forest, pure Q. semecarpifolia and pure P. roxburghii forests. The maximum (29%) forest fire intensity was recorded in P. roxburghii forest, which mainly occurred at lower altitudinal zones and is the causal factors for spreading the fire in nearby forest zones. However, the mainly A. pindrow and pure Q. semecarpifolia forests have shown high frequency of anthropogenic disturbances. The pure R. arboreum forest was the least disturbed forest in the study area with closed canopy (82.7%) and low light availability (16.7%). Human settlement around higher altitudes of Bhagirathi Catchment Area lacked livelihood resources and thereby depends mainly on forests not only for wood fuel and fodder, but also for heating rooms, boiling water and making of agricultural implements. The highly disturbed forests reflected high cattle grazing/movement and consequent habitat fragmentation. The detailed information is presented in Table 2.

 Table 2. Various disturbances in studied forest types

Floral diversity and composition

A total of 143 species [18 trees (adult, sapling and seedling), 40 shrubs and 85 herbs] belonging to 112 genera and 60 families were recorded in all studied forest types. There was a diverse pattern of distribution of various trees, shrubs and herbs in lower and upper Bhagirathi Catchment Area. Fagaceae (03 species) was the dominant family in tree layer, followed by Ericaceae, Pinaceae and Rosaceae (02 species each). Similarly in shrub layer, Rosaceae (10 species) was the dominant family, followed by Amaranthaceae, Caprifoliaccceae and Fabaceae (03 species each). However, Asteraceae (16 species) was the dominant family in herb layer, whereas Lamiaceae (06 species), Polygonanceae (05 species) and Acaranthaceae, Rubiaceae (04 species each) were the co-dominant families. The moderately disturbed forests exhibited higher species richness as compared to other disturbance categories. The highest species richness (tree, 11; Shrub, 30 and herbs, 41) was recorded in FT5 (moderately disturbed forest). Table 3 shows the pattern of diversity along disturbance gradient, indicating that species richness, diversity and evenness decreased as the frequency of disturbance increased. The Shannon index value for all forest types oscillated from (a) 0±0 (pure P. roxburghii) to 1.41±0.1 (mixed broad-leaved forest) for trees; (b) 0±0 (FT7) to 1.22±0.13 (FT2) for saplings; (c) 0.24±0.24 (FT7) to 1.22±0.06 (FT6) for seedlings; (d) 2.28±0.08 (FT7) to 2.73±0.05 (FT5) for shrubs; and (e) 2.68±0.11 (FT1) to 2.90±0.19 (FT3) for herbs (Table 3).

 Table 3. Diversity indices in studied forest types along disturbance gradient

The mean tree density oscillated from 400±10 trees ha^-1 in moderately disturbed forest of pure P. roxburghii at lower altitude to 750±89.1 trees ha^-1 in least disturbed forest of pure R. arboreum at higher altitude. However, the total tree basal cover values ranged between 25.8±2.2 m² ha^-1 (FT6) and 88.1±23.6 m² ha^-1 (FT4) (Fig. 2). The species like A. pindrow, L. ovalifolia, Q. semecarpifolia and R. arboreum were widely distributed from moderate to highly disturbed forest associations (Table 1 and Table 4). The Taxus wallichiana was the codominant species at higher elevation with A. pindrow, R. arboreum and with some mixed broad-leaved tree species along disturbance gradient. Thereby, some late successional species were facing enormous anthropogenic pressure because they are being used as fodder, wood fuel and in making agricultural implements. The sapling density was recorded maximum (2160±233.7 saplings ha^-1) in moderately disturbed forest (FT3) whereas, minimum (680±32.7 saplings ha^-1) in highly disturbed forest (FT2). However, in seedling layer the highest density (4500± 440.2 seedlings ha^-1) was recorded in least disturbed forest (FT1), and lowest (2000±223.6 seedling ha^-1) in highly disturbed forest (FT2) (Fig. 2). For understorey shrub layer, the density values existed between 5040±82.6 shrubs ha^-1 (FT6) to 8640±826.9 shrubs ha^-1 (FT1). Moreover, the maximum herb density value (546000±13964.2 hers ha^-1) was encountered in least disturbed forest (FT1) while, minimum (332000±894.4 herbs ha^-1) in highly disturbed forest (FT2) (Fig. 3).

Fig. 2. Density and total tree basal cover values in different forest types.

 

Fig. 3. Shrub and Herb densities (ind 100 m^-2) in different forest types

The distribution of Cythula tomentosa and Indigofera heterantha was recorded in all forest types from lower elevational moderately disturbed forest to high elevational least disturbed forests. On the other hand the growth and dispersion of some shrub species such as A. adscendens, C. nepalensis, D. elegans, D. multiflorum, R. brunonii, R. parviflora, S. sauveolens and W. fruticosa were restricted to definite forest associations (Table 4). Similarly, the herbaceous vegetation reflected the wide distribution of some species such as C. wallichii, F. nubicola, H. nepalensis, O. corniculata, R. nepalensis, S. oleraceus, S. media, T. foliolosum, V. biflora and V. pilosa. However, higher individuals of herb vegetation indicated their scattered distribution along disturbance gradient. The density and importance values of species under various vegetation layers and disturbance gradients are presented in Table 4.

 Table 4. Density (100 m²) and Importance value index of each species in different forest types

 Table 4. Continued

 Table 4. Continued

 Table 4. Continued

Regeneration pattern

Densities of young individuals i.e., saplings and seedlings, increased from high to low disturbances but at certain altitudes (FT6 and FT7) the restricted growth was recorded in these individuals (Fig. 2). Mostly we have recorded a fairly good regeneration for the native dominant tree species. The maximum number of species (average 38.53%) exhibited good regeneration, followed by 21.46% species as fair regeneraton in all the studied forest types. However, 18.74% species represented no regeneration. Although all dominant species were encountered in good regeneration stage in least, moderately and highly disturbed forests, but densities of these species varied greatly. Some broadleaved tree species in FT5 i.e., A. indica, J. regia, S. cuspidata and S. paniculata were noticed as not regenerating (54.55%). On the other hand some species like A. pindrow and P. roxburghii were found encroaching the new habitats viz., pure R. arboreum and pure Quercus leucotrichophora forests, respectively. The seedling and sapling growth was high in moderately disturbed forests and low in highly disturbed forests (Table 5). 

  Table 5. Regeneration status of dominant tree species in various forest types along disturbance gradient

Canonical Correspondence Analysis (CCA)

The Canonical Correspondence Analysis (CCA) was applied to investigate the impact of total disturbance on plant vegetation coverage in Bhagirathi Catchment Area. The detailed illustrations are given in Fig. 4, 5 and 6 for trees, shrubs and herbs, respectively. The CCA ordination simultaneously displayed plant species as triangle and disturbance/environmental factors as vector arrows to visualize the correlation amongst different vegetation types. The percent variance explained the relation of tree species-disturbance axes on CCA as 35.46% and 64.65% for axis1 and axis 2, respectively. The eigenvalues were 0.95 and 0.79 for both corresponding axes. Similarly in shrub layer, the explained variance for axis1 and axis2 was 41.16% and 59.59%, while the eigenvalues were 0.50 and 0.22, respectively. However, in herbaceous layer the explained variance of axis1 and axis2 was calculated as 26.11% and 43.75%, and the eigenvalues were 0.51 and 0.34, respectively. It is clear from ordination plot that elevation, lopping and soil erosion were the most influencing factors in tree layer. However, elevation, canopy cover, light attenuation, forest fire and soil erosion had governed the differences in the shrub and herb layers vegetation (Fig. 5, 6). In tree layer, the first axis showed negative correlation (-0.254) with land sliding, while axis2 indicated positive (0.319 and 0.232) correlation with lopping and windthrow, respectively (Table 6a). For shrubs, axis1 was positively (0.302) and negatively (-0.220) correlated with elevation and forest fire, respectively. However, axis2 showed negative (-0.133) and positive (0.273) correlation with canopy cover and light attenuation (Table 6b). Similarly, for herbs, axis1 indicated negative (-0.104, -0.222 and -0.215) correlation with slope gradient, forest fire and land sliding, whereas axis2 was positively (0.301) and negatively (-0.125, -0.073) correlated with elevation, canopy cover and soil erosion, respectively (Table 6c).

 

Fig. 4. CCA ordination diagram of tree species on the basis of disturbance, ecological parameters and importance values. The abbreviation of disturbance and ecological parameters are: EV, elevation; LP, lopping; SE, soil erosion; LA, lighting affect; FF, forest fire; LR, litter removal; CC, canopy cover; WT, wind thrown; LS, land sliding; SL, slope (first three letters of each genus and species were used in the diagram).

Fig. 5. CCA ordination diagram of shrub species on the basis of disturbance, ecological parameters and importance values. The abbreviation of disturbance and ecological parameters are: EV, elevation; SE, soil erosion; LA, lighting attenuation; FF, forest fire; CC, canopy cover; LS, land sliding; SL, slope (first three letters of each genus and species were used in the diagram).

 

Fig. 6. CCA ordination diagram of herb species on the basis of disturbance, ecological parameters and importance values. The abbreviation of disturbance and ecological parameters are: EV, elevation; SE, soil erosion; LA, lighting attenuation; FF, forest fire; CC, canopy cover; LS, land sliding; SL, slope (first three letters of each genus and species were used in the diagram).

Table 6a. Pearson correlation between CCA axes and disturbance parameters in tree layer

Values in bold are different from 0 with a significance level alpha=0.05.

EV, elevation; SL, slope; CC, canopy cover; LP, lopping; FF, forest fire; LS, land sliding; LA, lighting attenuation; LR, litter removal; SE, soil erosion; WT, wind thrown.

Table 6b. Pearson correlation between CCA axes and disturbance parameters in shrub layer

Values in bold are different from 0 with a significance level alpha=0.05.

EV, elevation; SL, slope; CC, canopy cover; FF, forest fire; LS, land sliding; SE, soil erosion; LA, lighting attenuation.

Table 6c. Pearson correlation between CCA axes and disturbance parameters in herb layer.

Values in bold are different from 0 with a significance level alpha=0.05.

EV, elevation; SL, slope; CC, canopy cover; FF, forest fire; LS, land sliding; SE, soil erosion; LA, lighting attenuation.
 

The tree species such as B. capitata and Q. leucotrichophora were represented on axis-2 and their distribution was influenced and controlled by land sliding and slope aspect, while P. roxburghii was affected by high litter removal and forest fire. Consequently, lopping intensity, soil erosion and elevation were observed as the major factors that regulated the habitat and ecological niches of the species. For understorey shrub and herb vegetation the light attenuation and canopy cover were negatively (-0.891) correlated. Shrub species such as Berberis asiatica, Cyathula capitata, Indigofera heterantha, Hypericum podocarpoides, Rosa macrophyla, Rosa sericea, Rubus niveus, Sarcococa saligna, Skimmia anquetilia with some other dominant species were clustered together along negative arm of axis 2 and their distribution was mainly controlled by elevation. However, species like Boehmeria platyphylla, Desmodium elegans, Rubus ellipticus and Woodfordia fruticosa were found susceptible to forest fire. Similarly most of the herbaceous vegetation was affected by elevation, light attenuation and soil erosion (in highly disturbed forests) however, species such as Arisaema jacquemontii, Ipomoea nil, Peristrophe paniculata and Rubia manjith were positioned at the top of the ordination plot and indicated particular habitat along unique set of environmental gradients.

Discussion

The light attenuation levels in these forests were different because of their presence on different physiographic and disturbance gradients. The canopy composition influenced the understorey plant diversity (Table 2 and 3). The mixed broad leaved forest facilitated higher diversity/richness; this was likely due to available heterogeneous environment whereas moderate disturbance allowed co-existence of common native species. However, the pure forest types were found unable to sustain high floral diversity. With an increase in canopy gaps the species richness generally increased (Kumar 2000), because canopy opening provided abundant light exposures to the forest floor and triggered growth of ground vegetation. In this study, the mainly A. pindrow and pure Q. semecarpifolia forests had high disturbance, because the forests were situated close to human settlements.

Forest fire is a significant ecological factor that can produce variable effects on community parameters (Keane et al. 2002). Fire strongly affects species composition within a forest ecosystem (Danthu et al. 2003) by influencing different aspects of growth and development such as flowering, seed dispersal, germination, seedling establishment and plant mortality (De Luis et al. 2005). Annual forest fires are common in most parts of Uttarakhand Himalaya including BCA. Generally, species diversity decreased and dominance increased immediately after fire due to mortality in lower size classes. This study revealed that the lower altitudinal pine forest was greatly affected by frequent fires which suppressed the existence of younger individuals in the study area. Husch et al. (2002) have pointed out that composition and species richness are important indicators for assessing the biodiversity which depend strongly on proper management practices. The presence of 143 plant species in the area shows that the forests are rich in biodiversity (Table 4) as in other Himalayan regions (Singh and Singh 1987). The highest species richness (tree, 11; shrub, 30 and herb, 41) was recorded in mixed broad leaved forests of Bhagirathi Catchment Area. Gairola et al. (2011) have reported high tree, shrub and herb richness in mixed broad leaved forests of Western Himalaya at various altitudes and aspects. Sundriyal and Sharma (1996) have reported 81 tree species in the temperate forests of Sikkim, while Hussain et al. (2008) had reported 63 tree species and 56 shrub species from Kumaon Himalaya. Comparatively lower tree species richness was noticed in BCA, mainly because we have selected pure forest stands. Species richness in canopy gaps was high in moderately disturbed forests. Kumar (2000) has also reported that high species richness was found in higher disturbed sites. The canopy opening within forests provided abundant light exposures to the forests floor and triggered growth of many ground vegetation. Rathor (1993) had reported high species richness in P. roxburghii-mixed broad leaved forest, while in high elevational Pinus roxburghii forests had the lowest.

It is widely accepted that moderate disturbance conditions provides greater opportunity for species turnover, colonization and persistence of high species richness (Mishra et al. 2004; Kumar and Ram 2005). The species diversity pattern showed varying degree of sensitivity for disturbance intensity in the study area. Diversity of vegetation patterns and the spatial variation of Shannon index of plant communities might be simple but effective indicator for predicting the species richness levels (Jiang et al. 2007). The understanding of forest structure and composition is a prerequisite to describe various ecological processes and also to model the functioning and dynamics of forests (Elourard et al. 1997). The maximum Shannon index and evenness in tree and shrub layers were found in mixed broad leaved forest, however in herb layer, it was maximum in mixed Q. semecarpifolia forest. The nature of forest communities largely depends on the ecological characteristics of site, species diversity and regeneration status of tree species. In this study the highest diversity in moderately disturbed forests has indicated that moderate disturbance has given birth to a variety of microclimates in the area. The Shannon index values of this study were lower to the values (3.09 to 3.33) reported by Gairola et al. (2011), but higher than the values (1.06-2.80) given by Khera et al. (2001) for mixed broad-leaved forests form various parts of Himalaya. Whittaker (1972) had observed that the dominance of one layer might affect the diversity of another layer

The density values in tree, shrub and herbs layers varied according to the forest type and disturbance gradient. The highest density was recorded in least disturbed R. arboreum forest due to occurrence of comparatively lower anthropogenic activity (low lopping, cutting and grazing) in this forest. The tree density values of this study are comparable to the previously reported values of 70-380 trees ha^-1 by Srivastava et al. (2005), 130-830 trees ha^-1 by Sharma et al. (2001) and 440-1180 trees ha^-1 by Singh et al. (1994) for Various R. arboreum associations at various altitudes. Bharali et al. (2011) had also reported high range of tree density (707-963 trees ha^-1) for similar forest types. However, many authors had reported much higher density from Himalayan temperate forests; 1570-1785 trees ha^-1 by Bhandari and Tiwari (1997) and 2090-2100 trees ha^-1 by Kunwar and Sharma (2004). Variation in understorey shrub and herb densities may be due to changes in many environmental factors within the forest type.

A progressive reduction in tree density and TBC values was observed from undisturbed to disturbed forest by Mishra et al. (2004) and Bhuyan et al. (2001). The lower altitudes of Bhagirathi Catchment Area were occupied by immature regenerating forests, while at higher altitudes the mature old growth forests were noticed. Our TBC values (25.8 and 88.1 m² ha^-1) were best fitted with the earlier reported values i.e., 22.43 m² ha^-1 by Sanjeev et al. (2006), 25 m² ha^-1 by Ghildiyal et al. (1998), 32.71 m² ha^-1 by Ralhan et al. (1982), 72.90 m² ha^-1 by Singh and Singh (1986), 63.30 m² ha^-1 by Rawal and Pangtey (1994) and 51.87 m² ha^-1 by Kumar and Ram (2005) for similar forest types. Smiet (1992) has emphasized that basal cover is related to the stand disturbance index.

Lower seedling and sapling growth in highly disturbed forests indicated restricted establishment due to high human intervention and natural disturbances. Some species like B. alnoides, P. roxburghii and Pyrus pashia were recorded as new introduction in moderately disturbed conditions because moderate disturbances creates additional micro-sites that favours the germination and survival of opportunistic species. Species such as A. acuminatum, A. indica, J. regia, S. cuspidata, S. panicualata and T. wallichiana were observed in adult stage only and their seedlings and saplings were absent completely. Benton and Werner (1976) suggested that such type of populations could become extinct if this trend continues for a longer period.

This study demonstrated the effect of timber harvesting, fire and litter removal on the structure and regeneration of mainly A. pindrow and pure Q. semecarpifolia forests. Repeated forest fires have changed the soil temperature and eventually replaced the habitats of Oak at lower altitudes (Kumar and Ram 2003). Due to more solar insulation in forest gaps the relative humidity tends to decrease significantly (Murcia 1995). If the situation persists for longer period, these species will not be able to maintain their fundamental structures (Tabarelli et al. 1999).

The vegetation v/s disturbance/ecological parameters relationship in CCA ordination indicated that the disturbance strongly influenced the abundance of dominant plant species in different layers. Dong et al. (2010) suggested that native forest was the best habitat for most protected/endangered species and primitive species. However, habitat with high human activities may serve as a source for invasion of exotic/alien species into more pristine environment (Hobes 2000). Elevation strongly influences the composition and structure of forest landscapes (Behera and Kushwaha 2007). We have noticed that elevation is a significant factor that controls the dispersion of vegetation. The slope in this study was positively correlated with land sliding (Table 6) and indicated that an increase in slope could lead to higher land sliding. In the same way high light intensity was responsible for greater canopy gaps (Table 2). Higher light availability and bare soil exposure have favoured the invasion by alien plant species in new habitats (Flory and Clay 2006). That is why we have observed vigorous growth and wide distribution pattern of E. adenophorum in the study area (Fig. 5). The patchy distribution of species in CCA under each vegetation layer was due to availability of nutrient contents and effect of disturbance frequencies.

Conclusion

The regeneration of tree species and altitudinal pattern of diversity and composition have exhibited a clear and strong relationship with disturbance regimes in Bhagirathi Catchment Area of Garhwal Himalaya. An increase in anthropogenic disturbances negatively impacted the overall diversity and dispersion of species. All the forests, except R. arboreum, were under increasing biotic pressure. However, the regeneration of important trees i.e., Abies pindrow, Ilex dipyrena, Lyonia ovalifolia, Pinus roxburghii Q. semecarpifolia and R. arboreum was still satisfactory along disturbance gradient. Contrary to this, Acer acuminatum, Aesculus indica, Juglans regia, Sorbus cuspidata, Symplocos paniculata and Taxus wallichiana showed inadequate seedling establishment. Moreover, overexploitation of A. pindrow and Q. semecarpifolia forests for timber, fodder, fuel wood and leaf litter removal by local inhabitants, may lead the reduction in seedling and sapling growth in these forest types. The chronic disturbances have ruined the recovery potential of most of the ecosystems and restricted regeneration dynamics. The Canonical Correspondence Analysis (CCA) has clearly indicated that the P. roxburghii forests (Chir Pine) are susceptible to extreme forest fires (generally pre-monsoon fire dominates the area). Consequently, adverse impact may originate in terms of top soil removal and nutrient loss, which may negate the biodiversity. Therefore, pre-monsoon fire should be prevented to increase more survivability of young individuals.