Utilization distribution

Utilization distribution provides a convenient global representation of spatial use patterns by defining the relative frequency of occurrence of animals (Benhamou & Riotte-Lambert, 2012). We calculated utilization distributions using the nonparametric kernel local convex hull (LoCoH) method to assess spatial use by the studied cranes (Getz & Wilmers, 2004; Getz et al., 2007). This method is more appropriate than parametric kernel methods for constructing utilization distributions and can capture hard boundaries (e.g., rivers and cliff edges) and process large sample sizes (Getz et al., 2007). This method is also very powerful in processing aggregated and clustered data (Getz & Wilmers, 2004) at the population level (Liu et al., 2010). Thus, we constructed kernels with the fixed radius local convex hull (r-LoCoH) method (available at http://locoh.cnr.berkeley.edu) using flock location data within a fixed 500 m radius, which was sufficient to distinguish flocks of the two-crane species. The obtained shapefiles were imported into ArcGIS 9.3 to construct utilization distribution maps. We considered 90% instead of 100% isopleths as the overall crane distribution range by omitting outlying points representing exploratory animal movement rather than that necessary for survival. The 90% utilization distribution isopleths can faithfully reflect actual spatial distribution patterns of animals (Börger et al., 2006). For protection management, 70% and 50% isopleths of utilization distribution are usually recognized as the ordinary and kernel distribution range of wildlife. Thus, we considered 90%, 70%, and 50% utilization distribution isopleths in the current study to determine areas of high conservation priority.