Identifying spawning sites and other critical habitat in lotic systems using eDNA “snapshots”: A case study using the sea lamprey Petromyzon marinus L.

Abstract Many aquatic species of conservation concern exist at low densities and are inherently difficult to detect or monitor using conventional methods. However, the introduction of environmental (e)DNA has recently transformed our ability to detect these species and enables effective deployment of limited conservation resources. Identifying areas for breeding, as well as the ecological distribution of species, is vital to the survival or recovery of a conservation species (i.e., areas of critical habitat). In many species, spawning events are associated with a higher relative abundance of DNA released within an aquatic system (i.e., gametes, skin cells etc.), making this the ideal time to monitor these species using eDNA techniques. This study aims to examine whether a “snapshot” eDNA sampling approach (i.e., samples taken at fixed points in chronological time) could reveal areas of critical habitat including spawning sites for our target species Petromyzon marinus. We utilized a species‐specific qPCR assay to monitor spatial and temporal patterns in eDNA concentration within two river catchments in Ireland over three consecutive years. We found that eDNA concentration increased at the onset of observed spawning activity and patterns of concentration increased from downstream to upstream over time, suggesting dispersal into the higher reaches as the spawning season progressed. We found P. marinus to be present upstream of several potential barriers to migration, sometimes in significant numbers. Our results also show that the addition of a lamprey‐specific fish pass at an “impassable” weir, although assisting in ascent, did not have any significant impact on eDNA concentration upstream after the pass had been installed. eDNA concentration was also found to be significantly correlated with both the number of fish and the number of nests encountered. The application of snapshot sampling techniques for species monitoring therefore has substantial potential for the management of low‐density species in fast‐moving aquatic systems.

Identifying areas for breeding, as well as the ecological distribution of species, is vital to the survival or recovery of a conservation species (i.e., areas of critical habitat). In many species, spawning events are associated with a higher relative abundance of DNA released within an aquatic system (i.e., gametes, skin cells etc.), making this the ideal time to monitor these species using eDNA techniques. This study aims to examine whether a "snapshot" eDNA sampling approach (i.e., samples taken at fixed points in chronological time) could reveal areas of critical habitat including spawning sites for our target species Petromyzon marinus. We utilized a species-specific qPCR assay to monitor spatial and temporal patterns in eDNA concentration within two river catchments in Ireland over three consecutive years. We found that eDNA concentration increased at the onset of observed spawning activity and patterns of concentration increased from downstream to upstream over time, suggesting dispersal into the higher reaches as the spawning season progressed. We found P. marinus to be present upstream of several potential barriers to migration, sometimes in significant numbers. Our results also show that the addition of a lamprey-specific fish pass at an "impassable" weir, although assisting in ascent, did not have any significant impact on eDNA concentration upstream after the pass had been installed. eDNA concentration was also found to be significantly correlated with both the number of fish and the number of nests encountered. The application of snapshot sampling techniques for species monitoring therefore has substantial potential for the management of low-density species in fast-moving aquatic systems.

K E Y W O R D S
conservation biology, environmental DNA, fish, habitat-use, lamprey, qPCR, wildlife management

| INTRODUC TI ON
Freshwater biodiversity is facing unprecedented levels of threat and has experienced over 120 extinctions worldwide within the last century (Ricciardi & Rasmussen, 1999). More than 4,600 freshwater species are currently in the threatened or endangered category (IUCN Red List, 2013). Many aquatic species of conservation concern exist at low densities and are inherently difficult to detect or monitor using conventional methods. The introduction of environmental (e)DNA sampling techniques, however, has recently transformed our ability to detect low-density species and enables more effective and accurate deployment of resources and allocation of time (Ficetola, Miaud, Pompanon, & Taberlet, 2008;Martellini, Payment, & Villemur, 2005;Thomsen et al., 2012). The collection and analysis of eDNA is now becoming commonplace in the detection of freshwater species and assessing biodiversity in aquatic environments (Bohmann et al., 2014;Lodge et al., 2012;Pilliod, Goldberg, Arkle, & Waits, 2013;Taberlet et al., 2012). The probability of detection, however, can vary from species to species and can be dependent on the biology and behavior of the target organism, for example; the amount of DNA they shed, level of activity during sampling period, species density, life cycle stage, and also the type of water body in which they reside (Rees, Maddison, Middleditch, Patmore, & Gough, 2014). Due to the diversity of water bodies and differing quantities of eDNA present in a system, methods for sample collection can vary greatly for rivers or streams, lakes or lagoons, and seawater and are dependent on the size of the environment under study. A range of sampling approaches has previously been employed which have varied in sample size from, for example, c. 1,000 × 2 L samples from a canal and waterway system (Jerde, Mahon, Chadderton, & Lodge, 2011), to 5 × 15 ml samples from a sea pen (volume of 4 million liters) within a harbor (Foote et al., 2012). Therefore, the sampling approach can also greatly influence the likelihood of target species detection with an aquatic system. Quantifying eDNA to estimate the biomass of a target species in running water is invariably complicated and requires the consideration of many variables including eDNA shedding and degradation rate at time of sampling; water temperature; pH; salinity; flow rate; water volume; hydro-morphology; and the dendritic organization of the habitat (Rees et al., 2014;Roussel, Paillisson, Tréguier, & Petit, 2015;Thomsen et al., 2012). The gathering and utilization of these data are not always possible, or feasible, for the long-term monitoring of populations. However, studies have shown that within running water systems an increase in the abundance or density of a target species can lead to an increase in either eDNA concentration (Lacoursière-Roussel, Côté, Leclerc, & Bernatchez, 2016;Pilliod et al., 2013;Takahara et al., 2012;Thomsen et al., 2012) or eDNA detectability (Mahon et al., 2013). Similarly, it has been confirmed that spawning events are characterized by a higher relative abundance of eDNA (Bylemans et al., 2017) making the spawning season an ideal time to utilize eDNA for biomonitoring within lotic systems.
The anadromous sea lamprey (Petromyzon marinus L.) was chosen as the target species for this study as their populations are declining across Europe and facing the threat of extinction due to overharvesting, habitat destruction, and the loss of spawning and nursery grounds from the construction of anthropogenic barriers (dams and weirs) blocking upstream access (Almeida, Quintella, & Dias, 2002;Igoe et al., 2004;Kelly & King, 2001;Lucas, Bubb, Jang, Ha, & Masters, 2009;Renaud, 1997). P. marinus is anadromous and will migrate back into freshwater to begin their search for suitable spawning grounds. See Maitland (2003) for detailed overview of the life cycle of P. marinus and Dawson, Quintella, Almeida, Treble, and Jolley (2015) for details of the larval stage and metamorphosis.
P. marinus spawn on large graveled areas with fast-flowing water and are thought to identify suitable spawning rivers using pheromones (bile acids) released by larval lampreys residing in the sediment (Li, Sorensen, & Gallaher, 1995;Sorensen, Vrieze, & Fine, 2003). This increases the chances of finding suitable spawning rivers at the end of their long and costly upriver migration.
Lampreys display nest-building behavior as they reach the spawning grounds, moving large stones and gravel using their oral discs to create a depression in which to spawn (Jang & Lucas, 2005). Typically, within the depression, spawning usually commences with the male attaching to the cephalic/branchial region of the female and wrapping the rest of his body around hers forming a loop. Once the tail loop is tightened, and ready to squeeze the eggs out of the female's body, both male and female will then thrash and vibrate their tails for several seconds, resulting in the expulsion of ova and milt (seminal fluid) into the gravel depression from where it is dispersed downstream with sand and silt particles by water currents (Applegate, 1950). They usually spawn in pairs or groups (i.e., polygamous mating) and will disperse their eggs in nests or shallow depressions in the bed material (Jang & Lucas, 2005) with female P. marinus holding up to 114,000-165,000 oocytes (Hardisty, 1970;Hardisty & Huggins, 1970;Maitland, 1980). Spawning may last several days for each female but is dependent on the number of eggs available and numbers of eggs expressed during each spawning act. All lamprey species are semelparous, dying after a single spawning season (Larsen, 1980). Throughout the spawning season of P. marinus, there is consequently a considerable increase in the amount of DNA being released into the environment which is in the form of seminal fluid, ova, sloughed cells from nest building and migratory activity, and necrosing tissue from dead or dying adult lamprey. Yamamoto et al. (2016) determined that eDNA generally provides a "snapshot" of fish distribution and biomass in a large area, and the present study adopted this concept. We employed a strategy to target a lowdensity species during the spawning season by taking "snapshots" of the eDNA, that is, a "snapshot" sample is taken at a fixed point in chronological time. This sampling strategy will target a species throughout a period when there is a higher relative abundance of eDNA within a system and will aim to reveal spatio-temporal trends in eDNA concentration to investigate the distribution of P. marinus within our study river catchments. This study aims also to use these snapshots to identify "critical habitat" for our target species, which here is defined as areas of habitat believed to be essential to the species' conservation (U.S. Endangered Species Act). For the P. marinus, critical habitat would specifically include areas used for spawning, as well as habitat utilized during their upstream spawning migration.
This approach, however, may be applied in the identification of critical habitat for any aquatic species of interest.

| Study sites and selection of sampling locations
This study spanned a 3-year period (2015,2016,2017) in two separate catchments in Ireland which vary in both spatial scales and in the relative densities of the target species within these catchments (Table 1, Figure 1). "Target Species Density" (Table 1) refers to previous evidence of adult sea lamprey activity within the relevant catchment, in terms of nest counts and/or of individual fish. In excess of 500 sea lamprey and 136 nests were reported below the weir at Annacotty on the Mulkear by Igoe et al. (2004). Likewise, in excess of 50 adult sea lamprey were taken over a single day on the Mulkear by netting for use in a telemetry study by Rooney, Wightman, Ó'Conchúir, and King (2015). In contrast, King and Linnane (2004) had a total nest count of 65 for a float-over survey using kayaks over a distance in excess of 50 km on the Munster Blackwater (MBW).
Based on these figures, we can ascertain that the Mulkear (MLK) has a relatively high density of P. marinus as compared to the MBW.
The Mulkear River has been documented as an important spawning river for P. marinus (Igoe et al., 2004;Kelly & King, 2001)  the Annacotty weir (220 and 90 m respectively) which is 2.2 km upstream from the confluence with the River Shannon ( Figure 1). All other sampling sites on the Mulkear river are located upstream of Annacotty weir which poses a potential barrier to P. marinus migration (Rooney et al., 2015). However, past assessments have deemed the weir virtually impassable for migrating adult lamprey during flow conditions typical of the spawning season in Ireland (May to July; 3-11 m 3 /s as measured downstream of the weir) (Rooney et al., 2015). In 2016, a lamprey-specific fish pass was re-installed at Annacotty weir (prior to the beginning of the lamprey spawning run) which was originally part of the EU Mulkear Life project (Rooney et al., 2015). This gave us the opportunity to also investigate the effect of this lamprey pass on the ability of this species to move upstream.
There are smaller weirs also present in the Mulkear catchment ( Figure 1) but the dispersal of adult sea lamprey into the catchment points to these features not being a major problem for sea lamprey passage. A gravel trap installed at Blackboys Bridge ( Figure 1) is a vertical impediment to upstream fish migration which has a Deniltype fish pass installed. This trap, as well as the Annacotty weir, was recorded as "impassable" to sea lamprey, in the prevailing conditions at the time of study, in a WFD SNIFFER (2010) fish passage assessment (Barry, Coghlan, Cullagh, Kerr, & King, 2018). The Annacotty weir is the most significant structure that may impede sea lamprey passage in this system, however, within 2 km upstream of this structure, migrants also encounter a crump weir fish counter and the remnants of a weir breached to permit sediment transport and fish passage. There is also a low-level gauging station crump weir structure at MLK02 and a sloped bridge apron at MLK01 which pose potential barriers also.
The MBW is relatively a much larger catchment than the Mulkear with a substantially larger mean volume discharge (Table 1). It is one of Ireland's largest and longest river systems and is also a designated SAC for P. marinus. Twelve sampling sites were designated along the main stem of the river based on ease of access and coverage of system from source to the tidal areas. Two major weirs are present in the lower part of the main stem. Clondulane Weir is located 25 km upstream of the tidal limit (between MBW09 and MBW10;

| Field survey
From late May to early June, increases in water temperature above 15°C correspond with the commencement of P. marinus spawning activity within areas of suitable habitat (Kelly & King, 2001). Site In the MBW catchment, sampling occurred on two sampling dates each year (2015,2016,2017). A cooler blank was included on each sampling date (the cooler blank contained 1 L deionized water, which was brought to the field, and was treated identically to the other water sampling bottles except that it was not opened at the field sites), as was a field blank (a 1 L water sample which should not contain any P. marinus DNA) which was taken at a site upstream of an impassable waterfall in the Mulkear (MLK00; Figure 1) and at the source of the Blackwater (MBW01; Figure 2) where no P. marinus should be present. Water samples were immediately stored in an insulated cooler box filled with ice for transport and stored overnight at 4°C until laboratory-based filtration the following day.
In conjunction with each water sampling event, spawning counts (individual and nest counts) were also conducted.

| DNA extraction and filtration
All laboratory work was conducted in a dedicated eDNA laboratory where DNA extractions and PCR procedures were conducted in two separate laminar flow hoods with a UV light to avoid cross contamination of samples. Each water sample was filtered through individual 0⋅45 μm Whatman cellulose nitrate filters within 24 hr of collection.
The filters were dehydrated with 100% ethanol before storage at −20°C, in 2015, but in 2016 and 2017 were frozen directly at −20°C without ethanol. Each filter was subsequently cut into half (half for analysis and half for archival storage) and extracted using Chelex ® Chelating resin using a modified protocol from Estoup et al. (1996).
Briefly, filters were cut into small pieces within a 2-ml tube using some fine forceps. A volume of 500 µl of Chelex ® (10%) and 20 µl proteinase K (0.1 mg/ml) was added to the tubes and left to digest at 56°C for 2 hr whilst shaking. The temperature was then increased to 99°C for a total of 15 min and left to cool before centrifugation (6,000 g for 10 min). The supernatant was then transferred to a separate tube and stored at −20°C until use in qPCR analysis.

| qPCR Amplification and eDNA quantification
Concentrations of eDNA in samples were determined by qPCR using an Applied Biosystems ViiA 7 (Life Technologies, Inc., Applied Biosystems) quantitative thermocycler in combination with a species-specific P. marinus and Salmo trutta assay (Gustavson et al., 2015). Respectively, the sequences for P. marinus and the 30 μl reaction volume. The qPCR cycling condition was as follows: 50 ∘ C for 5 min and 95 ∘ C for 10 min, followed by 40 cycles between 95 ∘ C for 15 s and 60 ∘ C for 1 min. Standard curves for P. marinus (starting concentration 64.5 ng/μl using seven 10:1 serial dilutions) were generated using DNA extracted from tissue and quantified using fluorometric quantitation (Qubit, ThermoFisher).

| "Snapshots" provide an overview of the spatial distribution of the target species within a catchment to reveal spawning aggregations and critical habitat
Snapshot sampling was found to be successful in revealing spawning aggregations and habitat use within the Mulkear and MBW throughout the spawning season for all years sampled. All samples collected over each sampling year were combined (per site), which enabled the identification of relative habitat use and distribution within the catchment each year (Figures 3 and 4). Overall, eDNA concentration within the Mulkear catchment (with relatively higher sea lamprey densities and lower discharge) was noticeably higher than in the MBW catchment (with lower sea lamprey densities and higher discharge) as would be expected.  Figure 6). Unfortunately, no nest/ individual counts could be carried out at Fermoy in the MBW due to severe turbidity and high water flows. Apart from the areas that are traditionally surveyed for lamprey spawning (i.e., Fermoy weir MBW09 and Annacotty MLK08 and MLK10), peaks in eDNA concentration were found in areas previously not identified as important habitat for P. marinus (e.g., MLK05, MLK06, MLK07). Snapshot sampling also showed the extent of the upstream distribution in both catchments (relative to sampling sites) revealing that in the MBW, P. marinus were able to reach as far as MBW06 which is over 100 km upstream from the mouth of the river, and in the Mulkear, P. marinus reached the uppermost sampling sites in all rivers sampled.

| Mulkear
Over the 3-year sampling period, eDNA concentrations ranging from 0 to 831 pg/L were recorded in the Mulkear. All field controls and laboratory controls were found to be negative. In the Mulkear,  (2016) the addition of the lamprey pass at Annacotty weir. This indicates that although the lamprey pass at Annacotty may assist lamprey in ascending the weir, there does not seem to be any significant difference in lamprey eDNA concentration at these sampling sites the year after the lamprey pass has been installed.

| Munster Blackwater
Over the 3-year sampling period, eDNA concentrations ranging from 0 to 31.6 pg/L were recorded in the MBW which is overall much lower (nearly 27× lower) than that encountered in the Mulkear. Using

| D ISCUSS I ON
The concentration of eDNA at any point in time is dependent on both the rate of production of eDNA (influenced by the level of activity of individuals, their metabolic rate, and behavior such as spawning, fighting etc.) as well as the density of the species within a system and the hydrology of the area. Therefore, the amount of eDNA in an environment will vary seasonally in response to environmental changes and the behavioral ecology of a given species (Barnes et al., 2014;Goldberg, Pilliod, Arkle, & Waits, 2011;Lacoursière-Roussel et al., 2016). The increase in P. marinus biomass during the spawning season caused by the presence of large-bodied adults, gametes, and later their carcasses in the river system greatly increased the chance of de- as compared to other storage methods (Majaneva et al., 2018).
In the Mulkear, site MLK08 (most downstream site in catchment) generally exhibited the highest concentration of eDNA as would be expected due to its location within the largest known spawning site for P. marinus in Ireland (incorporating also MLK10). We observed a positive correlation between spawning activity (measured as number of individuals/nests) and eDNA concentration, which is consistent with other studies (Doi et al., 2017;Pilliod et al., 2013;Takahara et al., 2012;Thomsen et al., 2012). Bylemans et al. (2017)  Snapshot sampling also allowed the identification of peaks in eDNA concentration in areas that were not previously identified as important habitat for P. marinus in Ireland (e.g., MLK05, MLK06, MLK07 and then MLK03 and MLK09 later in the season). Although these peaks in eDNA concentration are indicative of increased P. marinus densities within these areas, without having more information about eDNA degradation rates, and flow rates within these areas, it cannot be ascertained how much eDNA is dispersed from areas upstream of these sites. However, this does outline areas of interest for future spawning surveys. This study has shown that eDNA snapshot sampling can be effectively used to identify areas of critical habitat for low-density species of conservation concern such as P. marinus. The results above have shown that eDNA can be very effective in outlining the general locations of spawning aggregations as well as the upstream extent of migrating individuals relative to potential migration barriers within a catchment.
Not only can specific areas be identified for the focus of future spawning surveys, but the magnitude of a spawning aggregation relative to other sites, or other years, can also be very useful for future management decisions.
Currently, literature dealing with running waters is still ambiguous about the effect on the downstream transportation of DNA (Roussel et al., 2015). In the study conducted by Gingera et al. (2016), water samples that were taken 1-2 km downstream had higher detection frequencies (75%-80%) than those collected at the most upstream site (approximately 50%). This suggests that, for a general management application, the chances of detecting a target species is increased if sampling is performed lower in the watershed, presumably because downstream sampling integrates the eDNA from a larger number of the target organism. Laramie, Pilliod, and Goldberg (2015) quantified the eDNA of Chinook salmon (Oncorhynchus tshawytscha) in relation to stream location and found no consistent relationship between stream distance and eDNA concentration. This would indicate that eDNA is not accumulating in downstream reaches, but is instead being removed through processes such as settling or destruction from physical forces (Piggott, 2016). This hypothesis is further supported by the work of Jane et al. (2015), who found that the distance eDNA travelled from the source was reduced at low flows due to a combination of cell settling, turbulence, and dilution effects. Nonetheless, our results have shown that, without sampling at many locations throughout a catchment, finescale patterns of movement and habitat use may be overlooked.
Arguably, these are some of the most prevalent concerns when considering the best management practices for a conservation species.
The potential power of eDNA as a conservation tool is not fully exploited when only presence and absence are considered. This study has also highlighted that without a prior knowledge of the biology and ecology of a target species, it would be extremely difficult to ascertain if increased eDNA concentration at a point in time reflects higher densities of the target organism, or if there is a behavioral/environmental reason for an increase in eDNA concentration/ detectability. Therefore, prior knowledge of target species' biology/ ecology is crucial in the interpretation of the results of future eDNA studies. A better understanding of the way in which eDNA disperses and persists in a system will also greatly improve future sampling design and maximize the likelihood of detection. However, we have here shown that utilizing knowledge about the ecology of a target species can greatly improve not only the chances of detection, but also improves the complexity of the information discernible from the eDNA samples.

ACK N OWLED G M ENTS
The first author is grateful to the Irish Research Council for supporting her postdoctoral fellowship as well as Inland Fisheries Ireland for supporting this research.

CO N FLI C T O F I NTE R E S T
None declared.