Schematic representation of eDNA site-occupancy model framework.

Adapted from Griffin et al. [<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0318414#pone.0318414.ref038" target="_blank">38</a>] of the hierarchical site-occupancy model implemented in the RShiny application [<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0318414#pone.0318414.ref066" target="_blank">66</a>]. Environmental DNA is collected from <i>S</i> independent sites using <i>M</i> independent water samples, analyzed using <i>K</i> independent quantitative polymerase chain reactions (qPCR) replicates. Incidental observations confirming target species presence is denoted as <i>k</i> with a common probability of confirmed detection, <i>π</i>. For every <i>s</i>th site and <i>m</i>th water sample, a positive qPCR result is denoted as <i>y</i>_<i>sm</i>. The occupancy state defines eDNA presence-absence at a site, <i>z</i>, or in a sample, <i>w</i>. The probability of eDNA being present at a site, <i>ѱ</i>, captured in a sample, <i>θ</i>, or detected in an individual qPCR replicate, <i>p</i>, are estimated. More specifically, the model estimates field (stage 1) capture using the probability of eDNA presence in a sample from a site is <i>θ</i>_<i>11</i> if the site was occupied and <i>θ</i>_<i>10</i> if it was unoccupied (with false-negatives and true negatives being <i>1-θ</i>_<i>11</i> and <i>1-θ</i>_<i>10</i>, respectively). Laboratory (stage 2) detection is estimated with positive qPCR replicates from samples containing target eDNA, <i>p</i>_<i>11</i>, or not containing target eDNA, <i>p</i>_<i>10</i> (with false negatives and true negatives being <i>1-p</i>_<i>11</i> and <i>1-p</i>_<i>10</i>, respectively). (DOCX)