Identification of potential wintering habitat for threatened Atlantic sturgeon Acipenser oxyrinchus oxyrinchus in Saco Bay, Maine, USA

Atlantic sturgeon Acipenser oxyrinchus oxyrinchus are large, anadromous fish, ranging from Florida to Canada, with 5 distinct population segments (DPS). Atlantic sturgeon are endangered in all DPSs except in the Gulf of Maine, where they are listed as threatened under the US Endangered Species Act (NOAA 2012a,b). Although distribution of Atlantic sturgeon in freshwater and estuarine habitats is well documented (ASSRT 2007), the use of marine habitats and migratory corridors is poorly understood, especially on a DPS-specific basis (Erickson et al. 2011). These gaps inhibit management of the species, particularly in winter months when Atlantic sturgeon have emigrated from rivers to the marine environment (Kieffer & Kynard 1993, Fernandes et al. 2010, Erickson et al. 2011), where interactions with commercial fishing gear occur (Stein et al. 2004). The limited studies on wintering behavior indicate that Atlantic sturgeon aggregate near the mouths of large bays and estuaries (Laney et al. 2007), which is consistent with both acoustically tagged fish in the St. Lawrence River (Hatin et al. 2002) and satellite


INTRODUCTION
Atlantic sturgeon Acipenser oxyrinchus oxyrinchus are large, anadromous fish, ranging from Florida to Canada, with 5 distinct population segments (DPS). Atlantic sturgeon are endangered in all DPSs except in the Gulf of Maine, where they are listed as threatened under the US Endangered Species Act (NOAA 2012a,b). Although distribution of Atlantic sturgeon in freshwater and estuarine habitats is well documented (ASSRT 2007), the use of marine habitats and migratory corridors is poorly understood, especially on a DPS-specific basis (Erickson et al. 2011). These gaps inhibit management of the species, particularly in winter months when Atlantic sturgeon have emigrated from rivers to the marine environment (Kieffer & Kynard 1993, Fernandes et al. 2010, Erickson et al. 2011, where interactions with commercial fishing gear occur (Stein et al. 2004).
The limited studies on wintering behavior indicate that Atlantic sturgeon aggregate near the mouths of large bays and estuaries (Laney et al. 2007), which is consistent with both acoustically tagged fish in the St. Lawrence River (Hatin et al. 2002) and satellite tagged fish in the St. John River, Canada (Taylor et al. 2016). Acoustic receivers in coastal marine waters have indicated similar behavior occurring in Gulf of Maine Atlantic sturgeon ); however, no directed studies on wintering behavior have been conducted in this DPS. While the National Marine Fisheries Service (NMFS) has designated the freshwater and estuarine portions of 7 Gulf of Maine rivers as critical habitat for Atlantic sturgeon under the US Endangered Species Act (NOAA 2017), to date, no marine habitats have received protection.
The marine environment outside small coastal systems, such as Saco River Estuary (SRE), Maine, has often been overlooked as habitat essential to the recovery of Atlantic sturgeon in the Gulf of Maine DPS. However, recent studies have identified the SRE as a foraging habitat for Atlantic sturgeon (where a unique diet of American sand lance Ammodytes americanus is sought; Novak et al. 2017), and that it supports a population of late-stage juveniles and non-spawning adults (Wheeler et al. 2016). Longterm acoustic telemetry data also indicate that Atlantic sturgeon occupy the mouth of the SRE from late spring through autumn, and are absent from the system in winter months (Novak et al. 2017). In order to gain insight into the wintering behavior of Atlantic sturgeon in the Gulf of Maine DPS, mark report popoff archival satellite tags (mrPATs) and acoustic tags were affixed to 6 individuals captured in the SRE in November 2016 to document their spatial use of the marine environment.

MATERIALS AND METHODS
Capture, handling, and tagging of Atlantic sturgeon followed recommended protocols described in Kahn & Mohead (2010). All individuals were caught between the jetties at the mouth of the SRE, Maine, on 7 November 2016, and were transported to a 10 000 gallon (37 854 l) tank in the University of New England's Marine Science Center, where sampling protocols (data collection, tagging) as described by Novak et al. (2017) were performed. To track the winter migratory behavior of Atlantic sturgeon in the marine environment over time, individuals received mrPATs with staggered release dates, and acoustic transmitters to determine date of departure and return to the SRE.
In total, 6 individuals were fitted with mrPATs (Wild life Computers; length: 127 mm, diameter: 28 mm, weight: 40 g). Holes (1/4 inch, 6 mm dia meter) were drilled through the ridge of 2 to 3 consecutive dorsal scutes posterior to the pectoral fins, and a small loop of 0.5 cm diameter rubber tubing was inserted through each hole. A 400 lb (ca. 180 kg)-test monofilament leader was looped through the release mechanism of the mrPAT and crimped. The leader was then fed through 7 cm of rubber tubing, 2 metal crimps, the most anterior loop of rubber tubing, and back through the metal crimps to anchor the tag. Additional loops of tubing were closed using monofilament and 2 crimps, and the anchored tag was inserted through these loops to prevent abrasion and entanglement (Fig. 1). All individuals were held for a 24 h recovery period prior to release; no mortalities occurred.
Scheduled pop-off dates were staggered over a period of 8−12 wk to reflect either the progression of linear movement or the occupation of a specific location over time. Tags were active from 9 November until pop-off dates of 15 January (n = 2), 30 January (n = 2), and 15 February (n = 2). While at liberty, mrPATs recorded daily minimum and maximum temperatures (± 0.1°C); these data, along with pop-off location, were transmitted to ARGOS upon release.
In addition, 5 Atlantic sturgeon also received individually coded acoustic tags (model V16; VEMCO; 69 kHz, 16 mm diameter, approximate battery life 10 yr). Tags were surgically implanted into the abdominal cavity of 4 fish following methods of Novak et al. (2017), and externally affixed to the rubber tubing on 1 individual (162182). An array of stationary acoustic receivers (model VR2W; VEMCO) is deployed in the SRE year-round with a maximum detection range of 900−1000 m (see Novak et al. 2017 for details). Maps of pop-off locations were created using ArcMap 10.4.1 (ArcGIS; ESRI), and figures were created using R package 'ggplot2'.

RESULTS
The Atlantic sturgeon captured in this study were considered adults (Wheeler et al. 2016), ranging from 134.6−172.5 cm fork length (FL) ( Table 1). Individuals with acoustic transmitters were detected near the mouth of the river until emigration, which ranged from date of release on 9 November to 14 December (Table 1). All mrPATs reported popoff locations to ARGOS on their programmed release dates except 162185, which released 1 d late.
A total of 5 mrPATs popped off in Saco Bay, Maine, within 5 km of shore between 10 and 30 m isobaths (Fig. 2a)   Jan uary between isobaths of 100 and 150 m (Fig. 2a), over mud/silt substrate (Fig. 2b). During the first month of deployment, mean daily temperatures varied widely (as much as 3.75°C) between the 5 inshore individuals (Fig. 3). After 18 December, however, mean daily temperature variation among these individuals remained within 1.5°C for the remainder of tag life (Fig. 3b). Mean daily temperatures at liberty for the individual located offshore (162184) were higher than those experienced by the 5 individuals located inshore on their respective pop-off dates (Fig. 3b). All sturgeon with acoustic tags returned to the SRE the following spring (April− June) or summer (July−August) and the average individual was de-tected for 52 d (± 20 d SE) in the SRE through the fall. One individual (162185) was recaptured at the mouth of the SRE on 15 September 2017 in good condition.

DISCUSSION
The unique double tagging method, using acoustic transmitters and mrPATs, allowed wintering locations to be observed during specific temporal periods in the marine environment where no acoustic array was deployed. Data from mrPATs suggest 5 of the 6 fish settled in marine habitat just outside of the SRE in Saco Bay, Maine, from mid-December through duration of tag life in mid-January to mid-February. Unfortunately, low resolution data precluded the determination of accurate daily locations of individuals. However, the considerable inter-individual variation in mean daily temperature from November− December may indicate that individuals were exploring different habitats, perhaps related to foraging behavior. The sudden decrease in variation of mean daily temperature among individuals in mid-December may thus indicate a convergence in habitat occupancy, as would be expected for individuals seeking an over-wintering site. All individuals whose mrPAT deployed in Saco Bay were located in < 30 m depths over sand, the preferred substrate of Atlantic sturgeon and their prey, American sand lance Ammo dytes americanus, in the SRE (Stein et al. 2004, Dzaugis 2013, McLean et al. 2013, Novak et al. 2017. Thus, while a larger sample size is necessary before any conclusions can be drawn, data presented herein suggest that Saco Bay may provide a suitable overwintering habitat for this sturgeon species. The lone individual (162184) located 33 km offshore on 30 January generally occupied higher water temperatures, and may have occupied greater maximum depths (100−150 m) and different substrate than those that remained in Saco Bay. After returning to the SRE in late April, 162184 spent June and July in the Kennebec River spawning sites during the spawning period (G. Wippelhauser pers. comm.). Immediately following this event, 162184 returned to the SRE and remained in and around the mouth of the system through autumn. The movement of this individual fits the recurrent pattern of presumed spawning behavior, where Atlantic sturgeon use the SRE to forage on American sand lance (Novak et al. 2017) before venturing to the Kennebec River system to spawn , returning short ly afterwards to the SRE to resume foraging behavior. While 162184 was not directly assessed for reproductive status, the alternate wintering behavior exhibited by this fish suggests individuals preparing to spawn may have different overwintering requirements than those that are not. However, further re search on reproductive status and higher sample sizes are needed before conclusions can be drawn. Evidence of tagged Atlantic sturgeon remaining in Saco Bay during winter months highlights the need for year-round monitoring of the coastal marine environment outside of rivers which exhibit sturgeon activity. Continued research via long-term satellite tags or a marine acoustic array is needed to determine if Atlantic sturgeon are wintering in Saco Bay annually, and for what portion of the winter. Increased monitoring over winter months may show that the high annual return rate of Atlantic sturgeon to the SRE ) may also correlate to wintering site fidelity, as exhibited in other species of sturgeon such as Gulf sturgeon Acipenser oxyrinchus desotoi (Parauka et al. 2011). At present, only rivers with potential spawning habitat are designated as critical habitat for Atlantic sturgeon in the Gulf of Maine (NOAA 2017). Since the SRE does not have the physical characteristics that allow spawning, it is not deemed critical habitat. Given the availability of a unique prey source (Novak et al. 2017), the high annual return rate of individuals , and the identification of a potential wintering site, the SRE may indeed be critical to supporting the life cycle of Atlantic sturgeon in the Gulf of Maine. The documentation of wintering behavior in Saco Bay emphasizes the pressing need for similar studies in river systems with large resident populations, such as the Kennebec, Penobscot, and Merrimack rivers, to determine if similar winter aggregations are occurring and to identify locations where Atlantic sturgeon may be at risk of bycatch mortality. The identification of such aggregations is crucial to the conservation and recovery of Atlantic sturgeon stocks.