A Comparative Study of <scp>iDNA</scp> and <scp>airDNA</scp> for Biodiversity Assessments in Tropical Forest Fragments

ABSTRACT Agroforestry landscapes demand monitoring tools that are fast, repeatable, and spatially explicit. We compared invertebrate‐derived DNA (iDNA; blowflies) and airborne eDNA (airDNA) for vertebrate biodiversity assessment across forest fragments in the Tsinjoarivo–Ambalaomby Protected Area (eastern Madagascar). Using dual‐marker metabarcoding of 22 samples (airDNA n = 14; iDNA n = 7), we found that spatially interpolated Functional Richness scores derived from both approaches captured the expected decline in richness across the gradient, demonstrating their value for monitoring changes in beta diversity. Despite low fly captures due to cold weather, iDNA recovered a broader taxonomic assemblage from fewer samples, whereas airDNA detections were restricted to mammals and sampling was more consistent. Domestic species made up a large proportion of airDNA reads (19.8%) but accounted for a smaller fraction of iDNA reads (5.9%). After filtering, both approaches detected endemic and threatened taxa, including Avahi laniger (iDNA) and Daubentonia madagascariensis (airDNA). Species accumulation modeling suggested iDNA reached richness asymptotes more quickly, while airDNA required greater sampling effort. Functional diversity analyses indicated iDNA samples had higher Functional Richness and Rao's Quadric Entropy. Operationally, airDNA was more reliable under field constraints and we therefore recommend a combined, standardized design (co‐located airDNA + iDNA, increased PCR replication, strengthened domestic species DNA blocking, and trait‐based analyses) for routine agroforestry monitoring to track functional change across fragmented landscapes.