Metabarcoding airborne pollen from subtropical and temperate eastern Australia over multiple years reveals pollen aerobiome diversity and complexity

eDNA metabarcoding is an emergent tool to inform aerobiome complexity, but few studies have applied this technology with real-world environmental pollen monitoring samples. Here we apply eDNA metabarcoding to assess seasonal and regional differences in the composition of airborne pollen from routine samples collected across successive years. Airborne pollen concentrations over two sampling periods were determined using a continuous flow volumetric impaction air sampler in sub-tropical (Mutdapilly and Rocklea) and temperate (Macquarie Park and Richmond), sites of Australia. eDNA metabarcoding was applied to daily pollen samples collected once per week using the rbcL amplicon. Composition and redundancy analysis of the sequence read counts were examined. The dominant pollen families were mostly consistent between consecutive years but there was some heterogeneity between sites and years for month of peak pollen release. Many more families were detected by eDNA than counted by light microscopy with 211 to 399 operational taxonomic units assigned to family per site from October to May. There were 216 unique and 119 taxa shared between subtropics (27°S) and temperate (33°S) latitudes, with, for example, Poaceae, Myrtaceae and Causurinaceae being shared, and Manihot, Vigna and Aristida being in subtropical, and Ceratodon and Cerastium being in temperate sites. Certain genera were observed within the same location and season over the two years; Chloris at Rocklea in autumn of 2017–18 (0.625, p ≤ 0.004) and 2018–19 (0.55, p ≤ 0.001), and Pinus and Plantago at Macquarie Park in summer of 2017–18 (0.58, p ≤ 0.001 and 0.53, p ≤ 0.003, respectively), and 2018–19 (0.8, p ≤ 0.003 and 0.8, p ≤ 0.003, respectively). eDNA metabarcoding is a powerful tool to survey the complexity of pollen aerobiology and distinguish spatial and temporal profiles of local pollen to a far deeper level than traditional counting methods. However, further research is required to optimise the metabarcode target to enable reliable detection of pollen to genus and species level.