2.2 Field data collection
We established a sample line for the study near Sanguanmiao, the core area of the reserve, where the density of giant pandas is known to be high and where pandas are often scent marked (Nie et al., 2012a). We established six types of sample lines along three types of terrain (Table 1): 12, 7, and 3 each of ridge, valley, and slope samples with animal trails, and 11, 7, and 4 each without animal trails, the former as experimental samples and the latter as control samples, corresponding to roughly equal sample line lengths. Animal trails are identifiable trails traveled by wildlife on repeated trips, evident by worn appearance, compacted soil, and lack of vegetation.
The width of the sample line was 4 m, i.e. 2 m on each side of the road, and within this range we could easily detect marked trees and feces. Data collection methods followed Nie et al., (2012a), but differed (Nie et al., 2012a). Walking along the sample line, the number of giant panda faeces found along the line was recorded, as well as the species and location of all trees with DBH greater than 5 cm and several characteristics that may affect the selectivity of giant panda markers (Table 2), and they were marked as ”available trees”. For the trees with scent marks, additional characteristics such as mark type and orientation were recorded as ”marked trees”. Urine marks causing a greener bark with a musky smell, and AGS marks causing the bark to turn brown. To determine if the marked trees were repeatedly marked, we monitored each marked tree on the sample line every other week. The musky smell of the urine marker disappeared after about 5 days, which allowed us to determine if the same area was repeatedly marked. For AGS marking, we cut a small piece of bark in the middle of the marked area, leaving out the lighter colored bark underneath, and monitored it to see if it was again covered with AGS marks to determine if it was repeatedly marked.
For each scent-marked tree, where feasible, we set up a 5*5m sample square centered on them and measured the variables of vegetation density, vegetation coverage, slope and slope aspect (Table 2). After that, a control sample was set up by walking 200m along the sample line and the same indicators were measured for it. Note that if there are marked trees that happen to be located within the control sample, we need to reposition the sample to the nearest area that does not contain the marked trees in order to form a better control.