Massive green branch removal and damage to trees can still be observed, however (Fig. 2), since the removal of deadwood is allowed. Currently, nine permanent villages and more than a hundred secondary and herding settlements are present in the Park (Stevens, 2013), with 6221 local residents and 1892 head of livestock

(Salerno et al., 2010) (Table 1). We collected data on forest structure and species composition in 173 sample plots during two field campaigns in 2010 and 2011. The plots were randomly distributed ABT-263 solubility dmso within the forest areas in a GIS and then mapped in the field. To detect forest areas, we used a land cover map obtained from a classification of a Terra Aster satellite image taken in February 2006 (Bajracharya et al., 2010). We then used square plots of 20 m × 20 m for the tree (Diameter at the Breast Height – DBH ≥ 5 cm) layer survey, and square subplots of 5 m × 5 m were randomly located within the tree plot for the regeneration (DBH < 5 cm and height > 10 cm) and shrub layers. For all trees, we recorded species, total height, DBH, and species

and density for regeneration and shrubs. The following stand descriptors EGFR inhibitor review were computed for each survey plot to be used in the analyses: tree density, basal area, average DBH, maximum DBH, tree diameter diversity index (Marzano et al., 2012 and Rouvinen and Kuuluvainen, 2005), and Shannon species diversity index (Table 2). Topographic variables

such as elevation, slope, and heat-load index were derived from the NASA/METI ASTER Global Terrain Model, with a geometric resolution of 30 m and vertical root mean square error (RMSE) of about 9 m. We calculated heat-load index (McCune and Keon, 2002) in a GIS and used it as a proxy variable for solar radiation. Anthropogenic variables (forest proximity to buildings, trails, and tourist lodges) were derived Dichloromethane dehalogenase from thematic maps (Bajracharya et al., 2010) and computed using horizontal-Euclidean distance, slope distance and accessibility time, in order to assess possible effects of topographic features. Accessibility time was estimated by dividing the DEM-computed slope distance by the average walking speed (Tobler, 1993). These data allowed estimation of the effect of forest, understory vegetation, and terrain roughness in reducing off-trail walking speed for wood gathering. We gathered summary statistics on tourism activities and fuelwood consumption from previous studies on the Khumbu valley (Salerno et al., 2010) for multivariate statistical analyses. These tests examined the relationships among environmental variables (topographic and anthropogenic) and forest structure and species composition. Three data sets were central for ordination analyses: (i) forest structure (6 variables × 167 plots); (ii) species composition (22 species × 173 plots); (iii) environmental variables (12 variables × 173 plots).