Species Diversity, Community Structure, and Distribution Patterns in Western Himalayan Alpine Pastures of Kashmir, Pakistan

Abstract Western Himalayan alpine pastures are among the most diverse ecological locations on the globe. Four alpine pastures were investigated to study species distribution patterns, richness, similarity, and community structure in Bagh District, Azad Kashmir, Pakistan. Four communities, Poa–Primula–Sibbaldia, Primula–Caltha–Primula, Poa alpina–Poa pratensis–Scirpus, and Sibbaldia–Poa–Scirpus, were identified on the basis of an importance value index. The average value of species richness was 1.42; Simpson's and Shannon–Wiener's diversity values were 3.13 and 0.91, respectively; the degree of maturity index was 44.1; and species evenness was 0.901. Local alpine flora was dominated by a hemicriptophytic life form with microphyllous leaf spectra. The species–environment correlation was analyzed using canonical correspondence analysis. A negative correlation of both diversity and richness was revealed with altitudinal gradient. Anthropogenic disturbances showed a significant negative impact on distribution of medicinal and palatable species. Unpalatable species dominated the local flora, indicating the heavy grazing pressure in the area. Development and implementation of regional conservation strategies are recommended to protect the threatened Himalayan alpine biodiversity.


Introduction
The Himalayas constitute one of the richest and most unusual ecosystems on Earth (Salick et al 2009). Himalayan alpine vegetation communities retain high ecological significance, because they control the soil stability of their catchment areas, play a major role in ecosystem functioning, and are vital in cultural, ethical and aesthetic aspects (Stirling and Wilsey 2001). The alpines are characterized by low productivity, high intensity of solar radiation, and high degree of resource seasonality because of high ultraviolet (UV) radiation, high wind velocity, blizzards, low temperature, and snowstorms (Nautiyal et al 2004). The vegetation of this fragile zone is adapted to the extreme conditions occurring with sparse populations; is generally dwarfed, stunted, or spiny; and develops a mosaic patch of different forms. Alpines possess an early growth initiation with a short vegetative span ranging from several days to a few months (Tasser and Tappeiner 2002). The community usually exhibits seasonal fluctuations, and its structure and composition are strongly influenced by the extent to which periodic phenomena in the individuals are adjusted to one another (Kershaw 1973).
A linear relationship exists between vegetation attributes such as species richness, diversity, and maturity values and ecological factors such as altitude, aspect, and distance of the site from disturbance stimuli (Schuster and Diekmann 2005). Indigenous flora has a direct relationship with the regional altitude, and it has been studied since the early 19th century (Mani 1978). Species richness is currently the most widely used, simplest, and most easily interpretable indicator of biological diversity determined by a complex of environmental factors (Whittaker 1977;Shrestha and Jha 2009). Numerous studies have examined the relationships among plant species richness, climate, and spatial variables in the area (Nautiyal et al 2001;Kala and Mathur 2002;Panthi et al 2007). Results showed the inverse correlation between altitude and species richness in Himalayan alpines, indicating that for every 100-m increase in altitude, there is a loss of 35-45 species (SCBD 2003). Moisture is also one of the important determinants of species richness and composition (Vetaas 2000).
Himalayan pastures have been grazed intensely for centuries (Miller 1999). They are an important forage resource for local livestock and are grazed from March-April to September-October. Malik (1988) reported that  (Misri 2003) and continued to decrease thereafter (Qureshi et al 2007). It is recommended by vegetation scientists that grazing practices in these fragile alpine communities above 2000 m should be limited and controlled and that they should be carried out in late summer after the full vegetation growth (Altesor et al 2005). But in Kashmir pastures, grazing activities start soon after the snow melts, in early growing season, and continue until it again starts snowing (Gupta 1978). The degraded vegetation is further not able to repair itself because of harsh climatic conditions and a short growing period.
In the western Himalayan mountains, vegetation dynamics have been neglected by researchers because of their remoteness, inaccessibility, danger, and lack of local infrastructure. The community structure and distribution patterns are poorly understood, because they have not been given due attention by plant ecologists to date (Peer et al 2007). The present study made an integrated effort to analyze and discuss ecological community attributes, as well as their correlation with environmental and anthropogenic variables, using multivariate analyses. Objectives also include the evaluation of Raunkier's spectra with the quantitative vegetation data, that is, importance value index (IVI) values, being an index of all quantitative parameters used in studying vegetation dynamics.

Study area
Bagh District lies in the western Himalayas and has subtropical to moist temperate vegetation (Anonymous 2007). Four pastures, Sankh, Garang Plani, Pirkanthi, and Ganga, were selected for the study (Table 1). The study sites are located at the line of control between India and Pakistan in the Pir Panjal subrange of the Himalayas. Sankh faces southeast; Garang Plani, northeast; Pirkanthi faces north; and Ganga, northwest ( Figure 1). Summers are cool, with average temperatures from 9 to 12uC, whereas winters are severe and always below freezing down to 210uC (Anonymous 2009). The area studied remains under snow cover from November to March. Locals, nomads, and shepherds use the area as their summer pastures from April to August.

Methodology
Field expeditions were made to the selected pastures from April-August 2008 and 2009. A random sampling technique was used to study and analyze the vegetation dynamics, as well as to collect the primary data for statistical analyses. Quadrates were laid in selected sites that had the best representation of floral biodiversity and geographic extent of the area. Sites were selected according to the altitude, topography, slope, and moisture to get an accurate image of whole-area vegetation. Vegetation attributes, including frequency, density, cover, and richness, were recorded, along with environmental coordinates such as latitude, longitude, altitude, and slope, using a global positioning system (Garmin model 2000). Plants from the isolated vegetation patches were also collected to record the maximum number of species and their distribution patterns. A total of 120 quadrates, 1  3 1 m each, (30 at each of 4 sites) were taken to record general vegetation attributes. The index of diversity was calculated after Simpson (1948) and Shannon and Wiener (1949). Simpson's index of diversity gives the probability that two individuals selected at random will belong to the same species. It is calculated as where D is the diversity index, n is the number of individuals of a particular species, and N is the number of individual of all species. Shannon index is ''a measure of the average degree of uncertainty in predicting what species an individual chosen at random from a sample will belong to'' (Shannon 1948). If p i is the proportion of individuals (from the sample total) of species i, then diversity (H9) is Species evenness was calculated using the Shannon evenness index: J9 5 H9/l n (S), where H9 is the Shannon-Wiener diversity index and S is the number of species (Pielou 1975). The Shannon evenness index ranges from 0 (when one species is dominant) to 1 (when all species are equally abundant).
Species richness was calculated after Menhinick (1964) where d is species richness, S is the total number of species in a community, and N is the total number of individuals of all species in a community. The maturity index was recorded after Pichi-Sermolli's method (1948). Maturity is a dynamic concept and is related to structural complexity and organization (Margalef 1968).

Degree of maturity~F requency of all species in a stand
Total no: of all species The index of similarity was calculated after Sorensen (1948) by using importance values. Sorensen's test is used to measure the ecological similarities and differences among the compared vegetation communities based on their presence or absence and associations. The lowest values in two communities were considered for comparison.

|100
where C is the total importance value for all species common in two communities, A is the total importance value in community A, B is the total importance value in community B, and IS s is the Sorensen index, with the index of dissimilarity 5 100 2 IS s . Leaf spectra and life form classification was carried out according to Raunkier (1956) to determine the biological spectrum of local alpine flora.
Data about medicinally important plant species were acquired by direct interviews with 97 locals. Respondents consisted of 15% of the total population and were selected randomly from all 4 study sites. Village elders, especially women, were preferred because of their strong and continuous relation with medicinal plants.
CCA was used to determine and analyze the relationship between species composition and underlying environmental and anthropogenic factors, including altitude, slope, grazing intensity, and distance from settlements. Mostly, it was observed that a single or a few environmental gradients acted as limiting factors and constructed the whole structure of species distribution pattern (Zahran et al 1990;. CCA is simpler and more efficient than orthodox statistical analyses, requiring much less linear data and giving precise species-environment correlation (Jongman et al 1987). The identified vegetation communities characterized by the phytosociological attributes (Table 2) were compared to get a precise picture of overall vegetation dynamics of the area.

Results
A total of 69 plant species belonging to 28 families were recorded from the area. Predominant plant families in the area included Asteraceae with 10 species, followed by Polygonaceae, Ranunculaceae, Labiateae, Rosaceae, and Poaceae with 5 species each. The biological spectrum of the local alpine flora revealed the dominance of a hemicryptophytic life form (40%) with microphyllous leaf spectra ( Figure 2).

Sibbaldia-Poa-Scirpus community (Ganga site)
The Sibbaldia-Poa-Scirpus community was found on northwestern slopes with altitudes of 2900 m.  (Tables 2 and S1, Supplemental data; http:// dx.doi.org/10.1659/MRD-JOURNAL-D-10-00091.S1). The IVI is the most precise measure of species attributes that we could achieve in this study. On comparing average IVIs for all species from the 4 study sites, we found that S. cunneata was the most dominant species, with an average IVI of 28.98 represented at all 4 sites. It was followed by P. alpina (23.66) at 3 sites and P. denticulata (17.8) at 4 sites.

Medicinally important herbs distributed in studied alpines
Respondents mentioned that 46 herb species were medicinally important and were regularly collected and used by the local communities. A large proportion (69%) of the respondents were involved in the use of medicinal plants for health problems. Some highly valued medicinal herbs included Arisaema jacquemontii Blume, Achillea millefolium L, Artemisia vulgaris L, Bergenia ciliata (Haw) Sternb, Bistorta affinis D Don, C. alba Jacq ex Com, Corydalis govaniana Wall, F. roylei Hook f, G. wallichianum D Don ex Sweet, Gerbera gossypina Royle, Nepeta erecta (Bth) Bth, Onopordum acanthium L, Pimpinella diversifolia DC, Plantago major L, P. macrophylla D Don, Rumex hastatus D Don, Rumex nepalensis Spreng, Taraxacum officinale Weber, Thymus linearis Benth, Valeriana pyrolifera Medik, Valeriana jatamansii Jones, and Viola canescens Wall ex Roxb. These important medicinal herbs were distributed throughout the study area. Various ailments treated by ethnomedicinal remedies were fevers, cough, cold, asthma, rheumatism, stomach disorders, parasitic worms, kidney stones, and body weaknesses.

Canonical correspondence analysis
Canonical correspondence analysis (CCA) results showed a total inertia of 0.938. The first axis alone explained 54.3% of total unexplained variance. Taken together, the first and second axes of the data set explained 80% of total inertia, indicating a high species-environment correlation. The altitude, slope, erosion, and distance from settlements all showed strong positive correlation with the first CCA axis, whereas grazing intensity showed strong correlation along the second CCA axis. Although all environmental variables showed great strength, the most influential limiting factor was altitude. The three relatively lower sites, Ganga, Garang, and Pirkanthi, and their respective vegetation are clustered to the left of the CCA plot, clearly separated from the highest and unique Sankh site (Figure 3). The lower site's vegetation cluster is dominated by alpine grasses, including P. alpina, Scirpus spp, P. vulgaris, and S. cunneata. These dominant unpalatable species showed a strong affiliation with grazing intensity, because the lower sites experienced more grazing pressure. The higher elevation cluster comprises rare and threatened herbs, including P. denticulata, P. macrophylla, Tussilago farfara, Cichorium intybus, B. affinis, and B. ciliata, restricted to most remote, isolated and protected habitat. The results of Sorensen's test showing the striking dissimilarity of the Sankh site to others also support the segregation of its community from the rest.

Discussion
The average species diversity of 3.13 recorded in the present study is comparable to the results of similar investigations in different Himalayan regions: 1.53-2.88 in the western Himalayas (Samant et al 1998;Guar and Joshi 2006), 2.39-4.63 in the Gharwal Himalayas (Nautiyal and Guar 1999), and 2.5-3.10 in the trans-Himalayan alpines of Nepal (Panthi et al 2007). The overall dominance of a few alpine grasses and herbs, including P. alpina, S. cunneata, P. stewartii, P. denticulata, and P. vulgaris, indicates their broad ecological amplitude (Ge et al 2005).
Although the 3 lower pastures showed an overall similarity because of their similar geoclimatic conditions and human interactions, little variation in diversity and richness was observed that can be attributed to the seasonal live stock movements (Bhattarai et al 2004). The similarity in species composition and community structure differed with increasing distance among sites, indicating high beta diversity in the area (Table 3). A negative correlation between altitude and vegetation attributes, including species diversity, richness, and community maturity, was observed in the area. The highest Sankh site exhibited patchy and isolated vegetation distribution with little species overlap, also evidenced by the clear separation along the CCA first axis.
Grazing activity was prevalent throughout the study area, with clear impacts on vegetation structure evidenced by the dominance of rosette herbaceous flora. The dominance of unpalatable species such as Primula, Poa, and Sibbaldia indicates the high grazing pressure in the alpine pastures. Medicinally important rare herbs were observed to be confined to remote and isolated places in the pastures. According to the local elders, the medicinally important plants previously enjoyed abundant distribution in the alpines, but because of overgrazing and overharvesting, their distribution and abundance has been reduced. Unsustainable anthropogenic activities are posing a serious threat to these precious plant resources (Mayer et al 2009).
The studied vegetation communities were characterized by a high degree of species evenness (ie 0.901), indicating that individual species exhibit uniform distribution. Fewer numbers of species in each community decreases interspecific competition, which results in a high degree of species evenness. All communities exhibited low maturity index values (Table 2), which indicated disturbed and heterogeneous community structure (Tiessen and Wu 2002). The high intensity of anthropogenic disturbances and harsh climatic conditions regularly disturb the natural balance of fragile alpine vegetation communities during the short growing season, thus preventing them from reaching a climax stage of community maturity (Saxena and Singh 1982).

Conclusion
The community attributes, including species richness, evenness, diversity, and maturity, responded significantly to environmental, as well as anthropogenic, stimuli. Anthropogenic and biotic disturbances such as uncontrolled heavy seasonal grazing, trampling, and harvesting of medicinal plants were identified as the potential threats causing degradation and retrogression of pastures and affecting the natural diversity and community structure of this fragile alpine biome. There is a need to control the degradation and disturbance processes deteriorating the community structure and plant biodiversity of the western Himalayan alpines. However, this requires a detailed socioeconomic study in the area, which is beyond the scope of the present study. Development and implementation of regional conservation strategies are necessary to protect Himalayan alpine biodiversity.