Introduction
Originally used to assess microbial communities of ocean sediments (Ogram, Sayler, & Barkay, 1987), the use of environmental DNA (eDNA) applications have broadened significantly in recent decades to include the detection and monitoring of a wide range of species in marine and freshwater ecosystems (Martellini, Payment, & Villemur, 2005; Ficetola, Miaud, Pompanon, & Taberlet, 2008; Jerde, Mahon, Chadderton, & Lodge, 2011; Dejean, Valentini, Miquel, & Taberlet al., 2012; Thomsen, Kielgast, et al., 2012a). The approach has also increasingly targeted terrestrial species using eDNA deposited within natural or artificial water bodies (e.g. Ushio et al., 2017; Williams, Huyvaert, Vercauteren, Davis, & Piaggio, 2018), or deposited in soils (e.g. Buxton, Groombridge, & Griffiths, 2018; Kucherenko, Herman, III, & Urakawa, 2018; Leempoel, Hebert, & Hadly, 2019; Sales et al., 2019; Walker et al., 2017). More recently, a number of novel techniques to collect eDNA deposited on substrates found in aboveground terrestrial settings (e.g. vegetation surfaces, crops, and spider webs) have broadened the application of eDNA methods to include deployment of monitoring protocols designed to survey terrestrial species and communities (Nichols, Koenigsson, Danell, & Spong, 2012; Valentin, Fonseca, Nielsen, Leskey, & Lockwood, 2018; Valentin, Fonseca, Gable, Kyle, et al., 2020; Xu, Yen, Bowman, & Turner, 2015). However, while the state, transport, and fate (i.e. the ‘ecology’) of eDNA in aquatic ecosystems has been thoroughly explored (Barnes and Turner, 2016) it is not well understood in terrestrial ecosystems, leaving key questions surrounding sampling design and detection rates unanswered.
Understanding the state, transport, and fate of eDNA is critical to the design and interpretation of eDNA surveys (Barnes and Turner, 2016). eDNA can be present in multiple states; either in intracellular, intraorganelle, or extracellular form (Turner, Barnes, Xu, Jones, et al., 2014). Past eDNA surveys have collected a mixture of different states through direct substrate testing (i.e. DNA extractions directly from soil, or fecal material – Kucherenko et al., 2018; Martellini, et al., 2005), or targeted specific states through chemical isolation (e.g. Minamoto, Yamanaka, Takahara, Honjo, et al., 2011; Taberlet, Prud’Homme, Campione, Roy, et al., 2012) or differential size selection via filtration or centrifugation (e.g. Turner et al., 2014; Martellini et al., 2005). Identifying the eDNA state(s) most common within the environment being surveyed, or relevant to the question being addressed, and using appropriate isolation methods to capture the desired state(s), is key to designing protocols that maximize the probability of species detection (Turner et al., 2014). Capture of specific eDNA states in suspension is typically accomplished via direct processing of water, tissue centrifugation, filtration, or a combination thereof (e.g. Martellini, et al., 2005; Goldberg, Pilliod, Arkle, & Waits, 2011; Jerde et al., 2011; Minamoto et al., 2011); with filtration being the most common approach at present. However, the existing literature guiding filtration of eDNA states via specific filter pore sizes (e.g. Turner et al., 2014; Wilcox, McKelvey, Young, Lowe, et al., 2015; Moushomi, Wilgar, Carvalho, Creer et al., 2019) does not represent the full range of pore sizes that may influence optimal capture of intracellular eDNA (i.e., trade-offs between maximum water filtration and DNA yield). Similarly, understanding how environmental conditions affect the decay of each eDNA state over time informs the interpretation of positive or negative detection results.
For instance, if a captured eDNA state persists in the environment for long periods (i.e. months or years; e.g. Andersen, Bird, Rasmussen, Haile, et al., 2012; Turner, Uy, & Everhart, 2015; Strickler, Fremier, & Goldberg, 2015) it is unknown if a positive species detection indicates a recent presence or one over a relatively long time frame. Conversely, eDNA states that degrade quickly after deposition (i.e. hours, days, or weeks; e.g. Zhu, 2006; Thomsen, Kielgast, Iversen, Wiuf, et al., 2012b; Thomsen et al., 2012a) may indicate species presence within the very recent past, or may break down beyond detectability and produce false negative survey results (Schultz & Lance, 2015). Most existing knowledge about the fate of eDNA comes from experiments conducted within water or soil, finding that biotic and abiotic factors such as pH, microbial load, temperature, and enzymatic fragmentation influence the decay rates of eDNA (Barnes and Turner, 2016; Levy-Booth et al., 2007; Nielsen, Johnsen, Bensasson, & Daffonchio, 2007). However, these biotic and abiotic factors are unlikely to determine the fate of eDNA within aboveground terrestrial systems, since eDNA in said systems likely dries shortly after deposition and is thus likely influenced predominantly, if not entirely, by solar radiation (Figure 1 ).
eDNA transport in aquatic environments is facilitated by omnidirectional diffusion, precipitation through the water column, and directional movement via currents or thermal mixing, which can redistribute eDNA meters to kilometers away from the point of original deposition (Eichmiller, Bajer, & Sorensen, 2014; Deiner, Fronhofer, Mächler, Walser, & Altermatt, 2016; Thomsen, Kielgast, et al., 2012a). These processes can increase the availability of eDNA for capture and elevate detection probability, or dilute the available eDNA beyond detectability and reduce detection probability (Schultz & Lance, 2015). Transportation of eDNA deposited within soil is not as well understood beyond recognition that eDNA is unlikely to move laterally through the soil substrate (Taberlet, Bonin, Coissac, & Zinger, 2018). Therefore, the mechanisms that influence eDNA transport in water are not applicable to aboveground terrestrial eDNA. We posit that eDNA deposited on aboveground terrestrial surfaces will be predominately transported by weather events like rainfall, transferring it to the soil where it may percolate through the soil column for unknown distances (Figure 1 ). Given the unknown nature of eDNA transport in soil, surveying for species above the soil substrate necessitates the collection of eDNA from aboveground terrestrial substrates to ensure detection. Thus, the use of terrestrial eDNA aggregation techniques, which pool eDNA from a wide geographic area into a single reservoir (Valentin et al., 2018), will become invaluable for surveys of aboveground terrestrial environments. Aggregation of aboveground terrestrial eDNA has thus far been executed in two ways: by directly collecting substrates and submersing them into a centralized container filled with solution to be sampled later (i.e. vat aggregation – Valentin et al., 2018), or actively sampling and pooling eDNA from the substrate’s surface by physically removing it (Valentin et al., 2020).
For collection of aboveground terrestrial eDNA en masse via aggregation from surface substrates to move into regular use across a variety of survey designs, including rare, threatened, or invasive species detection and community level assessments, further investigation of the ‘ecology’ of aboveground terrestrial eDNA is required. Here, we conduct a series of experiments to investigate (1) the optimal filter pore sizes for isolating extracellular aboveground terrestrial eDNA; (2) how rain events limit eDNA retention on vegetative surfaces; and (3) the rate of degradation due to time of air exposure and ultraviolet (UV) solar radiation.