Fig 1.
Altitudinal changes in soil physico-chemical properties in Zanskar and Tso Moriri.
Generalized Additive Models were used to analyse altitudinal responses. Redundancy analysis biplots show interrelationships between soil variables and elevation. For units, see Table 1. Explained variation (R2) and estimated Type I errors (P values *<0.05, **<0.01, ***<0.001) are shown in parentheses.
Table 1.
Soil physico-chemical properties of Zanskar and Tso Moriri transects in Indian NW Himalaya.
Mean, minimum and maximum values are shown together with F-test and estimated Type I errors (P values *<0.05, **<0.01, ***<0.001) for differences between two transects.
Table 2.
Summary table of abundance and species richness of soil microfauna in two elevational gradients in NW Himalayas.
Abundance x 1000 ind m-2, mean (SD).
Fig 2.
The CCA ordination of soil microfauna.
The most explaining environmental variables were determined by forward selection (red arrows) for (A) Zanskar and (B) Tso Moriri mountain ranges. Source data are log-transformed, species are represented by blue dots (for full names, see S1 Table). Feeding groups, altitude and other environmental variables are passively projected into the CCA ordination diagram (green arrows). The isolines of species richness of samples (grey lines) are fitted using the Loess smoothing splines.
Table 3.
Explained microfaunal compositional variation by individual environmental variables.
The CCA multivariate analyses and permutation test were applied for Zanskar and Tso Moriri faunal assemblages (both nematoda and rotifera) in Indian NW Himalayas. Explained variation (R2) and estimated Type I errors (P values *<0.05, **<0.01, ***<0.001) are shown. Best predictors selected by CCA forward selection models are depicted in bold.
Fig 3.
Altitudinal changes in abundance and richness of rotifers and nematodes along two gradients (Zanskar and Tso Moriri) in Indian NW Himalaya.
Generalized Additive Models were used to analyse altitudinal responses of soil microfauna. Abundances were log-transformed. Explained variation (R2) and estimated Type I errors (P values *<0.05, **<0.01, ***<0.001) are shown in parenthesis.
Table 4.
Relationship between soil properties and richness and abundance of microfaunal communities.
Explained variation (R2) and estimated Type I errors (P values *<0.05, **<0.01, ***<0.001) from generalized linear model are shown. A parsimonious subset of environmental predictors selected by conditional inference trees are in bold.
Fig 4.
Altitudinal changes in abundance of microfaunal feeding groups along two mountain transects (Zanskar and Tso Moriri) in Indian NW Himalayas.
Generalized Additive Models were used to analyse altitudinal responses of omnivorous nematodes (Omnivores), filter-feeding rotifers (Filter-feeders), fungivorous nematodes (Fungivores), bacterivorous nematodes (Bacterivores), predacious nematodes (Predators) and root-fungal feeding nematodes (Root-fungal feeders). Explained variation (R2) and estimated Type I errors (P values *<0.05, **<0.01, ***<0.001) are shown in parenthesis.
Fig 5.
Conditional inference trees showing a significant effect of environmental factors on abundance of individual microfaunal feeding groups.
In each split of the tree, all predictors are tested and the one that best discriminates between higher and lower values is selected. Each split of the tree is described by the factors associated with the split (ovals), the permutation-based significance of the split (P-value) (ovals) and the level at which the split occurs (line between ovals and boxes). The Box-and-Whisker plot and number of plots (n) is given at each terminal node.