Reliable biodiversity assessment is essential to evaluate conservation effectiveness, yet

individual survey methods often provide partial views of ecological communities. In this study,

we integrated soil environmental DNA (eDNA) metabarcoding and camera-trapping data to

assess mammal communities across 31 sites in the Foreste Casentinesi National Park, Central

Italy. The two methods detected only partially overlapping communities: eDNA captured a

broader range of taxa, particularly small and elusive species, whereas camera trapping primarily

detected medium-to-large mammals. While camera-trap detection probability increased with

species body mass, eDNA detectability was primarily driven by the frequency of recorded

passages. Time since the last detection had a weak effect on eDNA, suggesting that soil samples

reflect cumulative activity rather than recent presence, with an estimated detectability half-time

of approximately 26 days. Integrating both datasets provided a more comprehensive ecological

characterisation, highlighting a community structure driven by both exclusive and shared

species. Our results demonstrate that combining eDNA and camera trapping offers a more

robust assessment of mammal communities than either method alone. This integrated approach

improves taxonomic sampling, supporting the use of multi-method frameworks to better inform

biodiversity monitoring and conservation strategies.