3.4. Statistical analysis
The possibility of double counting the same bird was first considered by observers during the survey based on direction and distance from transects of each call. Subsequently, we also used ArcGIS 9.3 program to place each calling bird along each transect to define possible overlaps. All double counts were eliminated before conducting density analysis.
First, we combined all line transects for each national park and stratified the analysis between the two national parks to estimate a density. Second, we analyzed line and point transects from Cat Tien National Park and Yok Don National Park separately. The density at Crocodile Lake in Cat Tien National Park was determined by combining all three point transects and stratifying the analysis for each point transect, and reporting the pooled estimates for a density at Crocodile Lake.
Density was estimated using DISTANCE version 6.0 (Thomas et al., 2010) as we considered a detection probability to provide the best approximation of actual densities for green peafowl in each area. Detections were entered as single objects and only aural detections were analyzed. We excluded visual detection as there were too few and they included both males and females, in contrast to the calls, which were only from males. All key functions of uniform, half-normal, hazard, and negative exponential with all series adjustment of cosine, simple polynomial, and hermite polynomial were examined to select the best model with lowest Akaike’s Information Criterion (AIC), following Buckland et al. (2001).