分子レベルでの肉食獣糞便調査の検証：気候、糞便の経年、および調査員の経験が識別成功に及ぼす影響

Non-invasive genetic sampling has become an essential tool for monitoring carnivores; however, the success of molecular identifications from scats varies widely across taxa, environments, and observers. Field-based assignments are also prone to misclassification, particularly when species are sympatric and produce similar scats. Understanding the determinants of molecular success and the concordance between field and genetic identifications is therefore critical for designing reliable surveys. We analysed 2,073 carnivore scats collected across five protected areas in Spain. A binomial GLMM showed that scat age, climatic conditions, observer identity, and year significantly influenced the probability of successful genetic identification. Fresh scats had nearly double the odds of yielding a valid genetic result compared with medium-aged scats, whereas higher precipitation and temperatures reduced success. Observer differences were also evident, and interannual variation suggested the presence of additional environmental effects. Of the 1,835 scats successfully identified, the overall match rate between field and molecular assignments was 75.9% (Cohen's κ = 0.68), increasing to 80.0% (κ = 0.74) when genus-rank identifications (Canis sp., Felis sp.) were considered correct. High-confidence field identifications achieved>90% agreement, but congeneric species, such as Martes foina and M. martes, were frequently misclassified. Our findings demonstrate that the interaction of environmental and human factors influences the success of identification. Field-based identifications, although often reliable, can lead to systematic biases in species-rich assemblages. We recommend incorporating molecular validation into carnivore surveys whenever species-level precision is required, especially for rare or threatened taxa.