Figure 5 Records of red panda sign based on slope (right); records of red panda sign-based distance to water
​The highest eigenvalue obtained showed that seven dimensions were required to capture all the data variations. By visualizing the scree plot and variance, at least five dimensions were required to retain 80% of the variance in the data (Fig.7). The first two dimensions of this space were plotted to examine the association among the variables in the red panda habitat. Dimension 1 accounted for 21.5% of the variance in the data and dimension 2 accounted for 18.8 % of the variance (Fig. 6). Canopy cover, aspect and bamboo cover explained over 60% with other variables having low contribution in the first dimension. Similarly, the fallen log, slope, and elevation had high contribution in the second dimension. The analysis showed that these are the key micro-habitat variables influencing the distribution of red panda.
Table 3 Principle Component scores for PCA of habitat variables
Figure 6 Contribution of variables to Dim.1 and Dim. 2 (left and right)