Novel multilayer network analysis to assess variation in the spatial co-occurrences of close kin in wild caribou populations

Understanding how individuals within populations are connected genetically and through shared space-use is critical to understanding the demographic patterns of at-risk populations. In recent years, non-invasive genetic sampling methods have allowed us to begin addressing these questions. Using a network analysis framework, we examined the spatial co-occurrence and genetic relatedness of boreal and central mountain caribou (Rangifer tarandus) in populations varying in sizes and trend from across western Canada. Using 15 microsatellites loci and a sex-specific marker, we developed pedigree networks for each population, including first to third order relationships. We constructed networks of individual spatial co-occurrence, and using multilayer network analysis, we determined the degree to which caribou that co-occurred spatially were related. Caribou populations varied in the extent and classes of familial relationships which co-occurred in space. In most populations, spatial co-occurrences were greater than expected only for parent-offspring relationships. A sex-specific analysis of the longest sampled population revealed that this was driven by a significant overlap between mothers and their offspring. Whereas in populations where very few parent-offspring relationship were detected, no significant overlap was found between any relationship categories. The lack of overlap in these populations may be an indication of poor calf survival. On the other hand, the higher degree of overlap with more distant relations (i.e., grandparents) seen in some populations may indicate a population that lacks mobility. Our research presents a new approach to explore both spatial and genetic overlap at the individual level in support of the conservation of threatened populations.