Exposure of Whimbrels to offshore wind leases during departure from and arrival to a major mid-Atlantic staging site

. The United States is pursuing a diversified energy portfolio that includes offshore wind with a focus on the Atlantic Outer Continental Shelf (OCS). The Western Atlantic Flyway (WAF) supports one of the largest near-shore movement corridors of birds in the world, including several shorebird species of high conservation concern. We used satellite transmitters to examine orientation of Whimbrels ( Numenius phaeopus ) crossing the OCS and their overlap with two wind energy leases. Birds using a migratory staging site along the Delmarva Peninsula in Virginia crossed the OCS along a southeast-northwest axis. A considerable percentage (42.9%) of tracks intersected with one of the two wind leases. The juxtaposition to the Delmarva Peninsula placed wind leases southeast of the peninsula within both the departure and arrival trajectories of Whimbrels. The satellite transmitters used in this study were not equipped with altitude sensors, so we do not know if birds crossed wind leases within the rotor swept zone. Several species of shorebirds, including hundreds of thousands of individuals, make trans-Atlantic flights from three major staging sites: Delaware Bay, the lower Delmarva Peninsula, and Georgia Bight. All of these sites have wind leases positioned to their southeast. One of the most effective strategies for minimizing conflicts between birds and potential hazards is to place hazards away from critical movement corridors. More information is needed about departure and arrival patterns of shorebirds that cross the OCS to inform future lease placement.


INTRODUCTION
Along with many nations throughout the world, the United States is actively pursuing a diversified energy portfolio that includes a greater reliance on clean, renewable sources of domestic energy. Offshore wind represents a significant component of this renewable energy strategy. The Atlantic Coast offers abundant wind resources and shallow, near-shore waters in close proximity to large electricity load centers with some of the most lucrative and rapidly expanding energy markets in the country (Gilman et al. 2016). Near-shore waters (< 30 m depth) support enough potential wind energy that, if fully exploited, could displace the entire land-based generating capacity of the coastal states from Maine through Maryland (Energy Information Administration 2004, Musial et al. 2016). To facilitate the transition to renewable energy sources, the Bureau of Ocean Energy Management (BOEM) is accelerating access to the Atlantic Outer Continental Shelf (OCS) for the purpose of developing commercial-scale wind energy facilities. Currently, 7073 km² of the Atlantic OCS is under lease, with an additional 11,235 km² in the planning phase. Existing leases are widespread, extending along the coast from South Carolina north to Massachusetts (BOEM 2021). To date, development of offshore wind along the Atlantic Coast is limited The Western Atlantic Flyway (WAF) supports one of the largest near-shore movement corridors of birds in the world, annually hosting hundreds of millions of birds of over 160 species, many of which are of conservation concern (Watts 2010, Robinson Willmott et al. 2013). Waterbird populations using the flyway are drawn from breeding ranges across much of North America, funnel along the coast in a thin veneer over near-shore waters, and then fan out again to broadly dispersed winter ranges. In addition to serving as a movement corridor, portions of the Atlantic Coast serve as strategic staging sites where waterbirds stop for extended periods to rest and refuel before continuing their migration (e.g., Atkinson et al. 2007, Spiegel et al. 2017). Among the most prominent groups of waterbirds using the WAF are shorebirds. Of the 35 shorebird populations with trend data that regularly utilize the WAF, 65% are declining, whereas 11% are increasing (Andres et al. 2012).
Whimbrels (Numenius phaeopus) are large shorebirds that use the WAF during both spring and fall migration (Johnson et al. 2016, Skeel andMallory 2020). Surveys within a spring staging site (Watts and Truitt 2011) and within the primary winter grounds  indicate that Whimbrels using the WAF have experienced significant declines over the past three decades. Whimbrels exhibit delayed recruitment and have low reproductive potential (Watts et al. 2015). A recent assessment suggests that adult mortality rates are elevated above sustainable levels and are likely contributing to ongoing declines (Watts et al. 2019). As a consequence of these patterns and relatively small population sizes, Whimbrels have been assigned high conservation scores by both the United States and Canadian shorebird conservation plans (USCPP 2015, Hope et al. 2019).
Because of the low number of offshore wind facilities that have been built within the WAF, we have very little information about the actual impact of these hazards on shorebird populations. However, on the basis of studies with other waterbirds that have traits in common, shorebirds are likely to experience two types of impact from offshore facilities, collisions and displacement (Exo et al. 2003, Pearce-Higgins et al. 2009, Loss et al. 2013, Marques et al. 2014. Collisions may involve the physical turbine (vertical axis and rotor blades) or may involve the vortex near-wake (Jiguet et al. 2021) as birds fly through an operating turbine. Both of these risks require that birds are "exposed" to the hazard or occupy the same physical space. Collision rate is increased when turbines are placed within areas supporting high bird densities, such as migratory staging sites, and may be influenced by weather (Newton 2008, Loss 2016. Displacement occurs when individuals avoid areas supporting turbines (Pearce-Higgins et al. 2012, Niemuth et al. 2013. With respect to movements such as migration, avoidance behavior may require that individuals alter routes to move around wind farms, potentially costing more time and energy Hüppop 2004, Furness et al. 2013).
Because the greatest volume of shorebird migration within the WAF is believed to occur close to the coast (Loring et al. 2020), populations are expected to have relatively little exposure to wind leases located within the Atlantic OCS (5-320 km offshore). An exception to this pattern may be around major staging sites where birds initiate or end transoceanic flights that cross waters of the Atlantic OCS on their way to or from winter grounds within the Caribbean Basin or South America. Well-known shorebird staging sites that occur along the Atlantic Coast include Delaware Bay (Clark et al. 1993), the Delmarva Peninsula (Watts and Truitt 2001, and the Georgia Bight (Wallover et al. 2015, Lyons et al. 2018). Birds departing from or arriving to these staging sites are known to cross waters of the Atlantic OCS (Burger et al. 2012a, Loring et al. 2020, Watts et al. 2021a). Each of these sites has associated wind leases located within the Atlantic OCS. However, we know very little about the departure and arrival pathways and the extent to which these birds may be exposed to future wind energy development within existing leases. Here, we use satellite transmitters to track Whimbrels flying across the OCS as they depart from and arrive at a migratory staging site to evaluate their flight trajectory and potential exposure to two wind energy leases.

Study Area
This study was conducted within a major spring and fall staging site for Whimbrels along the seaward margin of the lower Delmarva Peninsula in Virginia (Watts and Truitt 2011) and included two BOEM wind energy lease sites within the Atlantic OCS (Fig. 1). The lower Delmarva supports thousands to tens of thousands of Whimbrels during both the spring and fall migration periods (Watts and Truitt 2011). Whimbrels stage within the site for approximately 4 weeks before initiating the next leg of migration (Watts et al. 2021a). In addition to Whimbrels, the Delmarva Peninsula supports several other shorebird species of conservation concern during migratory periods (Watts and Truitt 2001 and has been designated as a UNESCO Biosphere Reserve (https://en.unesco.org/biosphere/wnbr) and a Western Hemisphere Shorebird Reserve Site with international status (http://www.whsrn.org), and is the site of a National Science Foundation Long-term Ecological Research site (https:// lternet.edu/site/virginia-coast-reserve-lter/) and the focus of a multi-organizational partnership dedicated to bird conservation.

Field Methods
We captured 10 Whimbrels between 2008 and 2012 on migratory staging sites along the lower Delmarva Peninsula in Virginia, USA (n = 9; 37.398 N, 75.865 W) and along the coast of Georgia, USA (n = 1; 31.148 N, 81.379 W). All birds were aged as adults by plumage (Prater et al. 1977, Pyle 2008) and were banded with We fitted all birds with satellite transmitters called Platform Transmitter Terminals (PTTs) using a modification of the legloop harness (Rappole andTipton 1991, Sanzenbacher et al. 2000). Instead of an elastic cord, we used Teflon ® ribbon (Bally Ribbon Mills, Bally, Pennsylvania, USA) that was fastened with brass rivets or crimps (Watts et al. 2008). We glued transmitters to a larger square of neoprene to elevate it above the body and prevent the bird from preening feathers over the solar panels. The transmitter package was below 3% of body mass measured at the time of deployment (563.9 ± 20.6 g) for all individuals tracked in this study. The PTTs used in this study were 9.5 g PTT-100 solarpowered units produced by Microwave Telemetry, Inc. (Columbia, Maryland, USA).

Tracking
Birds were located by using satellites of the National Oceanic and Atmospheric Administration and the European Organization for the Exploitation of Meteorological Satellites with onboard tracking equipment operated by Collecte Localisation Satellites (CLS America, Inc., Largo, Maryland, USA; Fancy et al. 1998). Transmitters were programmed to operate with a duty cycle of 24 h off and 5 h on (n = 9) or 48 h off and 10 h on (n = 1) and collected 1-34 ( = 5.48 ± 0.07) locations per cycle. Locations in latitude and longitude decimal degrees, date, time, and location error were received from CLS America within 24 h of satellite contact with PTTs. Locations were estimated by the Advanced Research and Global Observation Satellite (ARGOS) system (www.Argos-system.org), which uses a Doppler shift in signal frequency and calculates a probability distribution within which the estimate lies. The standard deviation of this distribution gives an estimate of the location accuracy and assigns it to a "location class" (LC): LC3, < 150 m; LC2, 150-350 m; LC1, 350-1000 m; LC0, > 1000 m; LCA, location based on 3 messages and has no accuracy estimate; LCB, location based on 2 messages and has no accuracy estimate; and LCZ, location process failed. We used LC classes 1-3 to determine the last Whimbrel locations before and after flights to and from the Delmarva Peninsula.

Departure and Arrival
We estimated the seasonality of potential exposure to lease areas by using departure and arrival dates, respectively, from and to the lower Delmarva staging site. We used tracking data to determine the dates of departure and arrival. We considered departure or the onset of migration to be when birds made decisive breakout movements away from the site. In order to identify these breakout movements, we first located the first and last Whimbrel location overlaying the Delmarva Peninsula and used these locations to develop centroids for each individual. We considered the first departure movement to be the first location that exceeded 2 standard deviation units beyond the mean of movements around centroids. We considered the dates of breakout movements to be the dates of departure. We considered arrival to be the dates of first locations overlaying the Delmarva Peninsula. In cases when departure and arrival times occurred outside the transmitter's duty cycle, we calculated the speed between the arrival or departure location and the last or next flight location. If the speed was less than 2 SDs below the mean Whimbrel flight speed ( = 14.7 ± 0.3 m/s, n = 45; Watts et al. 2021a), we interpolated arrival and departure times by using the mean Whimbrel flight speed and great circle distance between the two points. Departure and arrival were both abrupt and we recorded no aborted attempts at migration.
We delineated departure and arrival areas within the staging site. We considered the last location prior to departure and first location after arrival to represent locations of departure and arrival. We mapped all departure and arrival locations using a kernel density estimator (KDE) method (Worton 1995) with the "ks" package (Duong 2007) in program R (R Core Team 2020). We used the normal (or Gaussian) kernel and a smooth crossvalidation bandwidth selector (Duong and Hazelton 2005) to map 50% kernel densities. We considered the 50% KDE to be the area of highest departure and arrival activity. We estimated the centroid of KDE polygons for departure and arrival using the package "geosphere" (Hijmans et al. 2021) in Program R (R Core Team 2020).
We used tracking data to delineate migratory pathways to and from the Delmarva Peninsula. We considered pathways to include the route traveled between the location of departure or arrival and the arrival or departure location on the winter grounds. Because of the duty cycle of the transmitters our dataset had temporal gaps in coverage. We filled these gaps using continuoustime correlated random walk (CRAWL) models (Johnson et al. 2008, Johnson et al. 2018 in Program R (R Core Team 2020) that allowed us to interpolate a pathway for each individual. We used the segment of the migratory pathway for each individual that extended from the departure or arrival location to the first point east or south of wind leases to determine the direction (degrees) of travel for birds departing from or arriving to the staging site. We also determined the angle of juxtaposition between departure and arrival areas and wind leases by estimating the bearing between KDE polygon centroids and centroids of wind leases. All bearings were determined in ArcGIS Desktop 10.7.1 (Environmental Systems Research Institute, Inc. 1999-2010).

Exposure to BOEM Leases
We examined exposure of Whimbrels to BOEM wind leases by using tracking data. We overlaid individual tracks on polygons of both lease sites to determine the frequency of overflights during both autumn and spring migrations. We considered tracks to have overflown a lease if the track overlapped any portion of the lease. We projected mean flight lines (± SE) from stopover centroids. We overlaid individuals' flight lines on polygons of wind leases and calculated the proportion (estimated percentage of population) of birds that overflew the polygons as an estimate of population-level exposure.

Statistics
We developed descriptive statistics, including means and standard errors, for departure and arrival dates and directions of travel. We tested for patterns in directionality of departure and arrival using the Rayleigh uniformity test (Berens 2009). We compared orientation of autumn departure and spring arrival using Welch's t-test to accommodate unequal sample sizes.

RESULTS
Whimbrel departure and arrival positions were located within the lagoon system of the lower Delmarva Peninsula and were generally consistent between autumn and spring seasons (Fig. 1). Birds departed the staging site in autumn (n = 15) moving along a southeast bearing (Fig. 2a) and exhibited significant directionality (mean bearing = 144 ± 7.9°, = 0.88, p < 0.001). Birds arrived (n = 6) on the staging site moving along a northwest bearing (Fig. 2b) and also exhibited significant directionality (mean bearing = 342 ± 10.9°, back azimuth = 162 ± 10.9°, = 0.91, p = 0.002). Although the orientation of spring arrival was more north-south compared to fall departure, the 20° difference was not statistically significant (t = -1.33, df = 11, p = 0.21). The juxtaposition of the wind leases and the departure/arrival KDEs is in agreement with these bearings (Table 1), which suggests that many Whimbrels departing the Delmarva Peninsula migrate along routes that pass through these wind lease areas. Distances between departure/arrival centroids and centroids of wind leases ranged from 66 to 131 km (Table 1).

Fig. 2. Frequency distribution of orientation (bearings) of
Whimbrels (Numenius phaeopus) tracked departing in the autumn (n = 15) from (2a) and arriving in the spring (n = 6) to (2b) the lower Delmarva Peninsula, Virginia, USA staging site between 2008 and 2013. Arrows represent the mean bearings for autumn departures and spring arrivals. Bearings in the spring represent the back-azimuth of travel to better relate the orientation relative to wind leases. Tracked Whimbrels crossed the Atlantic OCS 21 times, 15 in the autumn and 6 in the spring. A total of 9 (42.9%) of these tracks overflew wind leases, 6 over the Virginia lease and 3 over the North Carolina lease. There is some evidence that Whimbrels departing the staging site (6 of 15 overflew the lease) may have higher exposure to the Virginia wind lease compared to Whimbrels arriving (0 of 6 overflew the lease) in the spring (G-statistic = 2.12, df = 1, p = 0.07). This may be expected because the autumn trajectory has a more east-west component compared to the spring trajectory. However, the small sample size limits our ability to resolve possible lease site by season patterns.
Whimbrels crossed the Atlantic OCS during relatively narrow time windows in early autumn and spring. Birds departed from the lower Delmarva staging site from 23 July through 19 September ( = 24 August ± 4 days SE) during autumn migration, though the earliest departure occurred 13 days prior to the next earliest departure. Birds arrived on site from 5 April through 6 May ( = 17 April ± 5 days), though four arrivals occurred from 5 April to 13 April and two later arrivals occurred on 29 April and 6 May.

DISCUSSION
Whimbrels using the lower Delmarva staging site followed a consistent direction during departure and arrival. This finding is in agreement with numerous other studies that have examined the orientation of departure and arrival flights of shorebirds using major staging sites (e.g., Richardson 1979, Piersma et al. 1990, Battley et al. 2012, Tan et al. 2018. During both seasons, Whimbrels followed a southeast-northwest axis with a more easterly component during autumn departure. This finding is consistent with the position of the lower Delmarva relative to the primary winter grounds along the northern coast of South America and what we know about the orientation of migratory pathways. Whimbrels stage along the lower Delmarva in autumn, make a transoceanic flight to winter grounds along the northern coast of South America, depart the winter grounds in early spring, and make a transoceanic flight back to the lower Delmarva to stage during the spring (Johnson et al. 2016, Watts et al. 2021b). The pattern observed in Whimbrels is consistent with a generalized southeast orientation for shorebirds leaving the Atlantic Coast in fall to make transoceanic flights to winter grounds that has been documented for several species (Loring et al. 2020).
Several authors have emphasized the need to use information on activities and movement of shorebirds to inform the planning of wind facilities in order to minimize potential impacts (O'Connell et al. 2011, Burger et al. 2012b, Howell et al. 2020. Of particular interest are departure and arrival pathways around staging sites where ascent or descent may expose shorebirds to hazards. The location of the Virginia and North Carolina wind leases are southeast of the lower Delmarva staging site and within the flight lines for both departure and arrival. More than 40% of the tracks that crossed the Atlantic OCS flew over one of the leases, suggesting that birds would have had exposure to turbines constructed within the leases. Because of the differences in orientation between autumn and spring it appears that birds may have greater exposure to the Virginia wind lease during departure compared to arrival. Additional information is required to evaluate the possible influence of season on risk. Ninety-five percent of Whimbrels are expected to arrive in spring between 26 March and 12 May and depart in autumn between 1 August and 17 September (Watts et al. 2021a).
Flight altitude is believed to be one of the largest determinants of collision risk for birds crossing wind facilities (Fox et al. 2006, Fijn et al. 2015. The satellite transmitters used in this study were not equipped with altitude sensors because of weight constraints, so we do not know if birds crossed wind leases within the rotor swept zone (RSZ, 25-250 m). Galtbalt et al. (2021) found that Whimbrels using the Australasian migratory pathway fly over ocean at relatively low altitudes (median altitude = 138 m). Piersma et al. (1990) examined climbing rates of eight shorebird species departing from the Banc d' Arguin and found Whimbrels to have the lowest (0.21 ms -1 ) climbing rate. Some birds had not moved beyond the RSZ by the time they could no longer be observed (1.5 km). This pattern is consistent with observations of some Whimbrel flocks leaving the lower Delmarva staging site in spring (Watts et al. 2017) that do not rise above the RSZ before flying out of site (Wilke, unpublished data). Loring et al. (2020) modeled flight altitudes of shorebirds crossing the OCS and found that most were above the RSZ, although values varied greatly, and 24% and 36% were within the RSZ during spring and fall, respectively. Migration altitude is known to be influenced by several factors, including wind direction, precipitation, and time of day (e.g., Eastwood and Rider 1965, Shamoun-Baranes et al. 2006, Lindström et al. 2021. Although additional work is needed to determine the altitude of Whimbrels flying over the Virginia and North Carolina wind lease sites, variation in flight altitudes observed in shorebirds in general suggest that some Whimbrels are likely crossing through the RSZ. The position of the Virginia and North Carolina wind leases southeast of the lower Delmarva staging site is not unique (Loring et al. 2020 Loring et al. 2020). These species likely take trajectories similar to those of Whimbrels during autumn departure and spring arrival and may have significant exposure to wind leases associated with these staging sites.
One of the most effective strategies for mitigating the impact of wind facilities placed within the OCS is to locate facilities away from bird activity centers. Although most bird species migrate along a north-south axis over nearshore waters, some shorebird species that make trans-Atlantic flights between coastal staging sites and winter grounds cross the OCS. Such crossings are concentrated around major staging sites. For Whimbrels using the lower Delmarva site, these flights have a consistent southeastnorthwest orientation that results in overflight of downstream wind leases. A similar juxtaposition occurs with wind leases near other coastal staging sites. Because populations of high conservation concern depend on these sites, consideration should be given to locating wind leases north or south of these movement corridors. For offshore wind facilities that have been approved or built, options to mitigate collision risk are primarily operational, including seasonal or time of day restrictions. The seasonality of spring arrival and autumn departure is known for Whimbrels along the Atlantic Coast (Watts et al. 2021a) and this information could be used to inform the schedule of operations in an effort to reduce collision risk. Whimbrel departure from staging areas has been shown to be highly concentrated within a two-hour window (1.5 to 3.5 h before twilight) of the day (Watts et al. 2017). This information could be used to further refine operations. Although birds did overfly wind leases, we do not know their flight altitude. A research priority is to collect altitude data for birds to improve our understanding of vertical exposure to offshore wind turbines.
Responses to this article can be read online at: https://www.ace-eco.org/issues/responses.php/2312 Author Contributions: Bryan D. Watts -conceptualization, formal analysis, funding acquisition, investigation, methodology, project administration, writing Chance Hines -Formal Analysis, methodology, review Laura Duval -project administration, review Alexandra L. Wilke -Conceptualization, review