Toward repurposing global passive air sampling networks for insect monitoring: Promises and pitfalls of airborne eDNA

Jordi Vilanova , Luis Rodrigo Arce-Valdés, Xianming Zhang, Julia J. Mlynarek, Rassim Khelifa

1. Polyurethane foam passive air samplers (PUF-PAS) are widely deployed to monitor environmental pollutants, yet their capacity to capture biological signals such as environmental DNA (eDNA) remains largely unexplored. Recent advances in airborne eDNA research create a timely opportunity to evaluate PUF-PAS as a tool for biodiversity monitoring and to leverage existing global sampling networks for conservation. Focusing on insects, we assess whether PUF-PAS can provide eDNA-based biodiversity data comparable to those obtained with traditional sampling methods. 2. We compared the taxonomic richness, community composition, and detection biases of PUF-PAS to a conventional method for sampling flying insects, Malaise traps. We deployed both sampling approaches across five sites in Montreal, Canada, during the entire flight season, and collected samples biweekly. Insect communities were characterised at the order level using morphological identification (Malaise traps) and DNA metabarcoding (PUF-PAS). 3. Insect communities collected with PUF-PAS and Malaise traps showed substantial overlap, with 73.3% (n = 11) of insect orders detected by both methods; 20% (n = 3) were unique to Malaise traps and 6.7% (n = 1) to PUF-PAS. On average, estimates of insect abundance from Malaise traps were strongly positively associated with PUF-PAS detection probability and read counts. However, PUF-PAS detection probability and read counts were best explained by Malaise trap counts measured several weeks prior and integrated over extended temporal windows, indicating a delayed biological signal. 4. Our results suggest that PUF-PAS is better suited for interannual monitoring of common insect taxa, particularly for detecting broad temporal trends rather than short-term dynamics. Given the availability of global passive air sampling infrastructure, this approach not only has strong potential for large-scale biodiversity monitoring, but also could foster interdisciplinary collaborations between environmental toxicologists and biodiversity scientists.