Picturing plant biodiversity from airborne environmental DNA.

While eDNA approaches have gained interest over the past decades all types of organisms have not been addressed evenly. In particular terrestrial plants have been the subject of less attention. Here we address the possibility to represent plant biodiversity from airborne environmental DNA (eDNA) sampling and metabarcoding. We collected air using a biological air sampler in the Botanical Garden of Montpellier (France) and compared the list of revealed plant species to the botanical inventory of the Garden. Ninety-two plant species could be detected from three sampling points across the 4,6 ha of the Garden, after one hour sampling allowing to filter 9 m3 of air. We recorded the plants carrying flowers at the time of the experiment, which allowed us to estimate that plants flowering at the time of the sampling could be detected 10 times more easily than plants that were not, given the number of plants carrying flowers. However, flowering is far from being required as a vast majority of plants still was detected without flowering. We also show that not all species orders are detected with the same probability, tree species being better represented in the sample than herbal plants, given the number of trees present in the garden. Finally using diagnostic species, present only once in the garden, we estimate that the maximum sampling distance allowed by the biological air sampler is at least 110 m. Our study underlines that air sampling is a promising method for monitoring terrestrial plant biodiversity and highlights the parameters that should be adjusted to optimize the approach.