Résumé
In aquatic ecosystems, environmental DNA (eDNA) can be collected from water samples to produce species inventories. One method for this is passive eDNA sampling, whose development for aquatic biodiversity monitoring is in its early stages. While several materials have been successfully tested for passive eDNA samplers (PEDS), methodological advances are still needed to explore their versatility as a complement to the more common method of active filtration. This study tested for the first time a PEDS using human‐crafted material in controlled marine mesocosms (1 m3) containing one species, the European seabass (Dicentrarchus labrax) in different conditions of fish density (1, 5, 10, 100 fish per m3) and exposure times (30 min, 2 h, 8 h, 24 h). We then tested the influence of another source of eDNA on the sampler's performance by introducing another species, the Pacific oyster (Magallana gigas). In addition, we compared the efficacy of the method with active filtration. The PEDS we produced consisted of a small electrospun polyacrylonitrile (PAN) membrane encapsulated in a custom‐made 3D‐printed frame. Each sampler is low‐cost, easy to manipulate, highly replicable, and customizable. A specific quantitative polymerase chain reaction‐based assay for the seabass was developed. Results were analyzed with multiscale occupancy modeling and continuous response variable modeling. We found that the PAN‐PEDS efficiently collected eDNA in a large volume (1 m3) of renewed water (1 m3/h), with a clear positive effect of high fish density on fish detection. The introduction of oysters did not significantly influence detection. Regarding exposure times, a range of results were obtained that could be attributed to the unreached equilibrium between eDNA shedding and degradation. While active eDNA collection (30 L) outperformed PAN‐PEDS, the ongoing development of passive methods can provide new insights in aquatic species monitoring when spatio‐temporal eDNA collection is required.