Introduction
Benthic macrofauna are abundant in marine and freshwater ecosystems and play vital roles in nutrient cycling and secondary production (Snelgrove, 1997, 1998). For example, macrofauna decompose organic inputs by feeding directly on organic matter, and they promote metabolism of pollutants by bacteria through biological transport and by increasing the oxygen supply in sediments (Pant, Negi, & Kumar, 2017; Sanz-Lázaro & Marín, 2011; Snelgrove, 1998). Because macrofauna are important to ecosystem processes, can be identified by microscopy, and have differing tolerances to enrichment gradients, they have been used widely to assess environmental impacts associated with human activities, including marine finfish farming (Keeley, Forrest, Crawford, & Macleod, 2012; Sanz-Lázaro & Marín, 2011), oil-drilling (Denoyelle, Jorissen, Martin, Galgani, & Mine, 2010), and mining (Josefson, Hansen, Asmund, & Johansen, 2008). Macrofauna assessments have also been used to monitor environmental recovery from these activities (Daan, Booij, Mulder, & Weerlee, 1996; Machado et al., 2018; Zhulay, Reiss, & Reiss, 2015).
To date, characterization of macrofauna communities has been done by traditional morpho-taxonomic identification, which is very time- and labour-consuming and relies heavily on highly specialized taxonomists. These factors intrinsically limit the speed and magnitude of sample throughput that can be achieved by traditional taxonomic approaches (Bourlat et al., 2013; Danovaro et al., 2016; Goodwin et al., 2017; Ji et al., 2013; Pawlowski et al., 2018). Furthermore, individuals that are damaged during mechanical collection and sieving processes as well as with poorly described life stages may be misidentified by visual identification (Danovaro et al., 2016; Goodwin et al., 2017). Exacerbating these challenges is that species that are difficult to distinguish by morpho-taxonomy are commonly present in these communities (Knowlton, 1993), and if improperly identified could lead to underestimation of biodiversity and/or misleading interpretations about indicator species used for environmental assessment (Dean, 2008; Nygren, 2014).
DNA barcoding alleviates the reliance on morphological characters for taxonomic identification and can offer several benefits including higher taxonomic resolution of closely-related or cryptic species and the ability to verify morphologically-identified taxa (Nygren, 2014; Westheide & Schmidt, 2003; Witt, Threloff, & Hebert, 2006). It can also be used to characterize the taxonomic composition of opportunistic indicator complexes (Silva et al., 2017). However, it is still relatively time- and cost-intensive due to the one-by-one nature of the workflow: samples need to be manually sorted into individual specimens and the cost of Sanger DNA sequencing.
Environmental DNA (eDNA) metabarcoding is a high-throughput, rapid, cost-effective, and relatively new method of characterizing species composition in complex environmental samples (Shokralla, Spall, Gibson, & Hajibabaei, 2012; Taberlet, Bonin, Zinger, & Coissac, 2018; Taberlet, Coissac, Hajibabaei, & Rieseberg, 2012). It has been proven to be useful for detecting invasive (Klymus, Marshall, & Stepien, 2017) and rare species (Pikitch, 2018), assessing biodiversity for conservation (Lacoursière‐Roussel et al., 2018; Stat et al., 2019), and monitoring environmental health status (Laroche et al., 2016; Pawlowski, Esling, Lejzerowicz, Cedhagen, & Wilding, 2014). Environmental DNA metabarcoding has been used to characterize benthic macrofauna responses to human activities associated with marine finfish aquaculture (Lejzerowicz et al., 2015), offshore oil-drilling (Lanzén, Lekang, Jonassen, Thompson, & Troedsson, 2016), oil spills (Xie et al., 2018), and chemical contamination (Xie et al., 2017). It has the potential to supplement traditional morpho-taxonomic analysis of macrofauna from sediments for environmental assessment and biomonitoring (Aylagas, Borja, & Rodríguez-Ezpeleta, 2014; Pawlowski et al., 2018), and to increase the spatial and temporal scales at which biomonitoring can realistically be done (Gibson et al., 2015). Available studies comparing eDNA metabarcoding with traditional morpho-taxonomy have only used visual morpho-taxonomy of macrofaunal individuals without the support of DNA barcoding for resolving morphologically similar (or cryptic) species. This makes it difficult to compare macrofauna inventories among studies, particularly for taxonomic groups prone to cryptic speciation including Capitellidae, a family that is especially relevant to benthic biomonitoring. In the absence of an objective means of classifying species based on morphology, there will inevitably be variation among studies as to how taxa were classified when morpho-taxonomy alone is used.
Benthic organic enrichment is a significant environmental issue associated with marine finfish aquaculture (Forrest et al., 2007; Wilson, Magill, & Black, 2009). It can influence faunal diversity and abundance, causing population shifts towards opportunistic taxa (Pearson & Rosenberg, 1978) thereby eliminating a diverse community and the functions originally provided by it within the ecosystem (Snelgrove, 1997, 1998). As such, benthic impacts of finfish farming are subject to regulatory control and routine monitoring in many countries (Wilson et al., 2009). In New Zealand, Norway, and Scotland, biological indices based on morpho-taxonomic macrofauna data are implemented for regulatory purposes (Keeley et al., 2012; Maurer, Nguyen, Robertson, & Gerlinger, 1999). To date, a regulatory biological index for evaluating benthic impacts has not been implemented in Canada. Geochemical indicators (e.g. sediment pore-water sulphide concentration) are used for routine regulatory benthic soft sediment monitoring (Hargrave, 2010; Hargrave, Holmer, & Newcombe, 2008), and macrofauna are analyzed only under certain benthic conditions (AAR, 2016; Wildish, Pohle, Hargrave, Sutherland, & Anderson, 2005). The addition of a biotic measure to routine assessments of soft sediments to complement information obtained by geochemistry would advance the state of the science of benthic monitoring in Canada, providing regulators additional information for decision-making. This could come in the form of eDNA-based data, either by metabarcoding or assays targeting particular taxa once eDNA indicators are identified.
Here we advance existing knowledge on the responses of benthic metazoans to organic enrichment with the goal of testing the potential application of eDNA metabarcoding to benthic impact assessments of fish farms. At seven stations along a distance gradient at each of six salmon farms in British Columbia, Canada, we collected sediment samples for eDNA metabarcoding analysis of metazoans, geochemistry measurements, and morpho-taxonomy combined with DNA barcoding analyses. For morpho-taxonomy and DNA barcoding, we focused on Polychaeta as this was the most abundant group of species observed and taxa therein are known to be good indicators of benthic impacts (Dean, 2008; Tomassetti & Porrello, 2005), especially the Capitella capitatacomplex (Dean, 2008; Wildish et al., 2005). The specific objectives of this study were to: (1) compare biotic signals associated with benthic impacts of salmon farming between morpho-taxonomic polychaete and eDNA metabarcoding data; and (2) identify potential eDNA indicators that could form the basis for development of eDNA-based monitoring methods in Canada. We analyzed biotic signals associated with benthic impacts of fish farms in eDNA metabarcoding data and macrofaunal polychaetes data, and used multiple methods to analyze the relationship between operational taxonomic units (OTUs) and organic enrichment.