Impacts of roads and trails on floral diversity and structure of Ganga-Choti forest in Kashmir Himalayas

Roads and trails bring deteriorating changes in the structure and composition of forests. The present study investigated the impact of roads and trails on floral diversity and structure of Ganga Choti forest, Kashmir. Phytosociological attributes using systematic quadrate sampling were recorded and analyzed using multivariate ordination techniques including Principal Component Analysis and Cluster Analysis. The total number of species found in the area was 101 belonging to 47 plant families. The dominant family was Poaceae followed by Labiateae, Asteraceae and Rosaceae. The dominant species of the area included Pinus wallichiana, Abies pindrow, Viburnum grandiflorum, Sarcococca saligna, Poa alpina, Fragaria nubicola, Onychium japonicum and Poa pratensis. The dominant life form was Hemicryptophyte (31.68 %), whereas leaf spectra class was Microphyll (30.69 %). The average diversity value was 3.39 with a maximum of 3.71 at the control site and 3.08 at the disturbed site. Average tree density was 505/ha with a maximum of 680/ha at the control site and a minimum of 330/ha at the disturbed site. Intense deforestation was indicated by a stump density value of 330/ha. Results revealed that roads and trails have adversely affected the composition and structure of vegetation with high pressure on keystone tree species. Disturbed sites showed lower values of diversity indices, species richness and regeneration synchronized with high deforestation, overgrazing, erosion and increased frequency of invasive species. The study area needs immediate attention of the management for conservation of local forests by minimizing the impacts of roads and trails.


Impacts of roads and trails on floral diversity and structure of Ganga-Choti forest in Kashmir Himalayas
Impactos de caminos y senderos en la diversidad de flora y la estructura del bosque de Ganga-Choti en Cachemira, Himalaya Hamayun Shaheen a *, Amra Riffat a , Maria Salik a , Syeda Sadiqa Firdous a

INTRODUCTION
Forest paths and trails strongly influence plant species composition in natural forests (Godefroid and Koedam 2004).Road infrastructure, and the access that it provides, degrades forest ecosystem by increasing edge habitats, reducing forest cover, increasing agricultural expansion, promoting migration and forestland encroachment (Verburg et al. 2004).Roads affect forest ecosystem by altering the level of available light, water, drainage patterns, soil structure and soil nutrients, and act as a barrier to migration and dispersal (van-Wyk and Smith 2001).Major impact on forest ecosystem regarding function is due to alteration of microclimate and the isolation of ecosystem patches by causing fragmentation and creating edges (Hill and Pickering 2006).
Road constructions also affect the impact on conservation potential of the area with secondary impacts such as people influx, economic development and construction activities.The invasion of alien weeds and invasive plants is another major issue associated with road development (Hua et al. 2013).Roads are the major source of landslides and sediment loss as well (Pietikainen 2006).Flow of surface and ground water and nutrients is affected by trails changing the geomorphic processes and sediment loads.Roads provide an easy passageway to nomads as livestock rearing has been the main source of livelihood in Himalayas for centuries (Kumar et al. 2001).Effects of trampling may exceed up to one meter from the trail edge by causing compaction of leaf litter and soil, reducing the abundance of woody and herbaceous plants (Nilsson 2001).
Population rise has resulted in exceeding forest use beyond the carrying capacity of the forest ecosystem in Kashmir Himalayas.Roads and trails are the easy passageways for the population to access the forest area for their needs, resulting into forest degradation (Shameem et al. 2010).The study area has four link roads and numerous trails used for the movement of locals and their livestock.The current study is based on the hypothesis that roads and trails have significant deteriorating consequences on forest structure and species diversity in the study area.Furthermore, roads and trails are associated with increased anthropogenic forest disturbances such as logging, over-grazing and erosion affecting floral diversity and forest structure.The specific objective of the study is to investigate the effects of roads and trails on floral diversity, species composition, richness, community maturity and evenness of forest stands.It was further aimed at analyzing the impact of anthropogenic pressure regarding deforestation, overgrazing and soil degradation on the phytosociological attributes of forest vegetation.

METHODS
The study area lies in Azad Jammu and Kashmir in Pir Panjal Sub range of western Himalayas.Sudhan Galli is located at an elevation of 2300 m at 34º04ꞌ34.92ꞌꞌN latitude and 73º44ꞌ54.52ꞌꞌE Longitude (figure 1).Topographically the entire area is mountainous, generally sloping from northeast to southwest covered with moist temperate forest vegetation.Annual precipitation is about 1,500 millimeter.May and June are the hottest months with average temperature of 20-25 ºC, while December and January are the coldest months with heavy snowfall and freezing temperature (Pak-Met 2014).The study was carried out during May-August 2016.Four forest sites were selected for the study including two disturbed and two control sites.The altitude, latitude and longitude of all study sites were recorded.The effect of roads and trails on vegetation distribution was investigated by placing quadrates along trails and roads.The quadrats were laid 5-20 m away from the trails and 5-50 m away from the roads.The quadrate size of 1 m × 1 m was used for the herbs, whereas for trees and shrubs quadrats size was 20 m × 20 m and 5 m × 5 m respectively.Vegetation attributes were recorded including density, frequency and cover.Phytosociological attributes Mapa del área de estudio (derecha) e imágenes satelitales de los sitios de estudio (izquierda).
including Importance Value Index, Diversity, Evenness, Richness and maturity were calculated following standard protocols (Cox 1967, Ahmed andShaukat 2012).Visual indicators such as hoof marks, animal excreta, browsed vegetation, eroded area and signs of trampling were used to determine grazing intensity.Deforestation rate was recorded by counting the number of stumps, whereas regeneration rate was calculated by counting the number of seedlings (Shaheen et al. 2011).Results were statistically analyzed by using Multivariate ordination techniques including Principal Component Analysis and Cluster Analysis (ter- Braak and Smilaur 1998).

RESULTS
A total of 101 plants were recorded from the study area belonging to 46 families.The dominant family was Poaceae with 11 species followed by Lamiaceae or Labiateae, Asteraceae and Rosaceae with nine members each.Pteridaceae displayed five species, whereas Pinaceae and Polygonaceae showed four species each.The dominant species of the area included Pinus wallichiana and Abies pindrow with IVI values of 79.44 and 67.9, respectively, followed by Viburnum grandiflorum (36.83),Sarcococca saligna (19.8),Poa alpina, Fragaria nubicola, Onychium japonicum, Poa pratensis and Oxalis corniculata.Hemicryptophytes was the dominated life form class having a percentage of 31.68 followed by Phanerophytes (18.8 %) and Geophytes (16.8 %).Microphylls were the dominant leaf spectrum class with a percentage of 30.69 followed by Mesophyll (28.7 %) and Nanophylls (26.7 %) (table 1).
The number of species was higher (65) in the control sites range as compared to disturbed sites (43).Density, frequency and cover values recorded for the disturbed sites were also lower than those of control sites (table 2).The Average Shannon diversity value in the area was 3.39 with a maximum of 3.89 recorded at control site 2, whereas a minimum of 2.93 at disturbed site 1.The disturbed sites showed lower values (0.97) as compared with the control sites (3.71).Simpson diversity was calculated to be 0.97.Average evenness was 0.86 with a maximum of 0.94 at control site 2, whereas a minimum of 0.78 at dis- The average maturity value recorded from the area was 30.17 with a minimum of 23.13 at disturbed site 1, whereas a maximum of 42.22 at control site 2. None of the forest sites was found to be mature (> 60 %) as the average maturity index values for the investigated area was 30.18.Relatively higher values were reported at control sites (34.6) than those reported for disturbed sites (25.75) revealing the negative impact of roads and trails on the forest (table 2).
Forest stands showed tree density of 505/ha with a maximum of 760/ha at control site 1, whereas a minimum of 200/ha at disturbed site 2. The disturbed sites showed lower density values (330/ha) as compared with 680/ha at control sites.Intense deforestation and logging were indicated at the disturbed sites with a stump density value of 520/ha as compared with the 150/ha at control sites.Average stump density recorded from the area was 335/ha with maximum stump density of 600/ha at disturbed site 2, whereas a minimum of 100/ha at control site 2. Average stem/stump value recorded from the area was 2.78 with a maximum value of 6 at control site 2, whereas a minimum value of 0.3 at disturbed site 2. Regeneration value in the area was 290 seedling/ha with a maximum of 500/ha at control site 2, whereas a minimum of 0/ha at disturbed site 2. The disturbed sites showed lower regeneration rate (130/ha), compared with 430/ha at control sites (table 3).Disturbed sites showed higher grazing intensity placed in grazing classes 2 and 3 showing maximum hoof marks, animal excreta and browsed vegetation.Control sites showed relatively lower grazing pressure.Disturbed site 1 was highly eroded and placed in class 3, while disturbed site 2 was moderately eroded and placed in class 2, within erosion classes.No visible effect of erosion was observed at the control sites.
PCA axis explained 94 % variance in the data indicating the statistical strength of the test.PCA Biplot showed dominant species of the area separated away along the Xaxis; i.e., Pinus wallichiana and Abies pindrow representing their ultra-dominance.Viburnum grandiflorum, Poa alpina and Trifolium repens were significantly correlated with the disturbed sites.Cirsium vulgare, Nepeta erecta, Polygonum aviculare, Rumex nepalensis, Poa pratensis, Urtica dioica, Onychium japonicum, Skimmia laureola, Bistorta amplexicaulis were also found clumped near the disturbed sites.On the other hand, Fragaria nubicola, Sarcococca saligna, Dactylis glomerata and Stellaria media showed strong correlation with the control site (figure 2).The cluster analysis dendrogram revealed 104  Dendrograma del análisis de conglomerados de los datos de especies basado en la distancia euclidiana.species arranged in five major associations grouped based on correlation and dominance in communities.Abies pindrow, Pinus wallichiana and Viburnum grandiflorum were branched out at the1st cut level, having maximum relative abundance at all sites.The 2 nd association comprised 9 co-dominant species, whereas the 3 rd association had 24 species abundantly found at disturbed sites.The species of control sites were clustered in the 4 th and 5 th groups with 24 and 15 species respectively (figure 3).

DISCUSSION
The current study was based on the assessment of impact of roads and trails on the vegetation of the Ganga Choti forests.The forest structure and composition of the investigated area showed important influence of the impacts of roads and trails.Vegetation parameters fluctuated significantly correlated with the intensity of road and trail disturbance.
Average tree density of the forest stands was calculated to be 505/ha.This value is less than 1828/ha calculated in the temperate forest of Garhwal Himalaya (Devlal and Sharma 2008), 610/ha in Subtropical Zone of Garhwal Himalaya, India (Tiwari et al. 2010), 1028/ha in the temperate forest of Garhwal Himalaya (Kumar et al. 2001), 963/ha in the temperate forest of Arunachal Pradesh, India (Bharali et al. 2011), 820/ha in Gangotri valley, India (Dhaulkhandi et al. 2008) and 602/ha in Himalayas of Pakistan (Ahmed et al. 2006).Low tree density indicates high disturbance due to fuelwood and timber extraction at disturbed areas (Sagar et al. 2003).This is synchronized with high stump density (520/ha) at disturbed sites indicating severe deforestation and pressure.Presence of roads and trails are as low as 1 % of the land area and can degrade the vegetation of an area up to 20 % (Nepstad et al. 2001).Locals access the forest for fuel wood, timber, fodder and non-timber products leading to forest degradation.
Natural regeneration was as poorly represented as 130 seedlings/ha at disturbed sites.Average regeneration rate calculated for the studied area was 290 seedlings/ha, which is less than 5,474/ha in the Sikkim Himalaya (Sundriyal et al. 1994), 5,200/ha in Gangotri, India (Dhaulkhandi et al. 2008) and 2681/ha in Garhwal Himalaya, India (Tiwari et al. 2010).High disturbance prevents the vegetation from reaching the climax community.Simpson's diversity value of 0.98 showed the instability of vegetation communities at the investigated sites due to high intensity of anthropogenic pressure, microclimatic variations and edaphic factors correlated with roads and trails disturbance (Sen et al. 2008).
Average value of specie richness was high at the disturbed site (1.86) as compared with 1.76 at control sites, which showed the invasion of non-native flora at the site.The new land use type favors the non-native species and non-managed alien vegetation (Cardoso et al. 2013).Mild disturbances and high moisture level increase the species richness allowing increased light, water and nutrients in open trail sections as compared with closed tree canopies of control plots (Atik et al. 2011).The percentage of climbers/lianas was very low (4.9 %) in the area, indirectly indicating the lower trees diversity and density (Dhaulkhandi et al. 2008).
Results revealed high soil erosion at disturbed sites due to road cuttings, exposing the soil to slides and erosion.The forest areas with roads are reported to have fivefold increase in landslide intensity, which is also reflected in our study area (Ovando 2008).Slope is also a key factor influencing potential for impacts to soil and vegetation.Exposed trail slopes are subjected to surface erosion, evident from diminished vegetation cover and reduced plant heights along the trail (White et al. 2006).Roads and trails associated activities cause loss of herbaceous understory and compaction of soil resulting in decreased forest productivity and sediment loss leading to loss of freshwater habitats and altered stream hydrology (Pietikainen 2006).

CONCLUSIONS
The investigated forest showed significant deteriorating impacts on forest structure and composition associated with the presence of roads and trails.The disturbed sites with roads and trails were characterized by lower floral diversity, species richness and immature communities.High grazing pressure, deforestation intensity and soil degradation were observed at the disturbed sites correlated with roads and trails presence.There is an immediate need for sustainable management of local forests by minimizing the impacts of roads and trails.It is recommended to incorporate biodiversity conservation into the design and implementation of road projects along with Sound road engineering and design work, minimizing forest fragmentation, maintaining partial tree canopy cover and controlling erosion and runoff.

Figure 1 .
Figure 1.Map of the study area (Right) and satellite imagery of study sites (Left).

Figure 2 .
Figure 2. Principal Component Analysis Biplot of the species and study sites.Gráfico del análisis de componentes principales de las especies y sitios de estudio.

Figure 3 .
Figure 3. Cluster Analysis Dendrogram of the species data based on Euclidean distance.

Table 1 .
Floristic composition, importance values, life form and leaf spectra of studied sites.Composición florística, valores de importancia, forma de vida y espectros foliares en los sitios estudiados.

Table 2 .
Phytosociological parameters of forest communities from the study area.Average richness recorded for the area was 1.81 with a maximum richness value of 2.23 at disturbed site 1, whereas a minimum value of 1.48 at disturbed site 2. The disturbed sites showed lower values of maturity index (25.75)when compared with 34.6 at control sites.

Table 3 .
Structural attributes and disturbance indicators of studied forest stands.Atributos estructurales e indicadores de perturbación de rodales forestales estudiados.