Contents lists available at ScienceDirect
Biological Conservation
journal homepage: www.elsevier.com/locate/biocon
Satellite tracking of hawksbill turtles nesting at Buck Island Reef National
Monument, US Virgin Islands: Inter-nesting and foraging period movements
and migrations
Kristen M. Harta,, Autumn R. Iversonb, Allison M. Benscotera, Ikuko Fujisakic,
Michael S. Cherkissa, Clayton Pollockd, Ian Lundgrend,1, Zandy Hillis-Starrd
aU.S. Geological Survey, Wetland and Aquatic Research Center, 3321 College Ave., Davie, FL 33314, USA
b CNT, contracted to U.S. Geological Survey, Wetland and Aquatic Research Center, 3321 College Ave., Davie, FL 33314, USA
cUniversity of Florida, Ft. Lauderdale Research and Education Center, 3205 College Ave., Davie, FL 33314, USA
dNational Park Service, Buck Island Reef National Monument, 2100 Church St. #100 Christiansted, US Virgin Islands, USA
A R T I C L E I N F O
Keywords:
Eretmochelys imbricata
Inter-nesting
Home range analysis
Switching state-space model
Kernel density estimation
Minimum convex polygon
A B S T R A C T
To conserve imperiled marine species, an understanding of high-density use zones is necessary prior to designing
and evaluating management strategies that improve their survival. We satellite-tracked turtles captured after
nesting at Buck Island Reef National Monument (BIRNM), St. Croix, US Virgin Islands to determine habitat-use
patterns of endangered adult female hawksbills (Eretmochelys imbricata). For 31 turtles captured between 2011
and 2014, switching state-space modeling and home range analyses showed that inter-nesting (IN) core-use areas
(i.e., 50% kernel density estimates [KDEs]) were 9.6 to 77.7 km2 in area, occupied for 21 to 85 days, and in
shallow water (21 of 26 centroids>10m). The IN zones overlapped with areas both within the protected
borders of BIRNM, and outside BIRNM (32% of turtle-tracking days outside during IN). Turtles migrated to their
foraging grounds between July and October with path lengths ranging from 52 to 3524 km; foraging areas
included 14 countries. Core-use foraging areas (50% KDEs) where turtles took up residence were 6.3 to 95.4 km2,
occupied for 22 to 490 days, with mean centroid depth 66m. Our results show previously unknown habitat-
use patterns and highlight concentrated areas of use both within and adjacent to a US protected area during the
breeding season. Further, our results clearly demonstrate the need for international conservation to protect
hawksbills, as migrating turtles crossed between two and eight different jurisdictions. Our results provide critical
spatial and temporal information for managers charged with designing strategies to minimize human impact to
and maximize survival for this globally imperiled species.
1. Introduction
Marine Protected Areas (MPAs) are important for managing and
sustaining ocean biodiversity (Agardy, 1994), however< 3% of the
world's oceans are within MPAs and<1% are regulated as no-take
(Costello and Ballantine, 2015). Further, global studies of MPA effec-
tiveness indicate successful MPAs include four or five key features (no
take, enforcement, old, large, and isolated), but most MPAs only have
one or two of these features, making them indistinguishable from un-
protected areas (Edgar et al., 2014).
Assessing the effectiveness of MPAs and implementing appropriate
management strategies at a local level requires understanding the
spatial ecology patterns for species of interest. However, with highly
mobile species that travel globally across geopolitical boundaries, such
as sea turtles, it can be difficult to assess these patterns. Remote
tracking through satellite telemetry allows researchers to address in-
creasingly complex questions on habitat-use and movement (Godley
et al., 2008; Hart and Hyrenbach, 2009; Hazen et al., 2012) and gain
understanding of spatial use during breeding, foraging, and migration
for sea turtles (e.g. Hart et al., 2014; Fossette et al., 2010; Schofield
et al., 2010; Shillinger et al., 2010).
Although global MPA coverage is small, studies indicate they are
https://doi.org/10.1016/j.biocon.2018.11.011
Received 30 May 2018; Received in revised form 2 October 2018; Accepted 7 November 2018
Corresponding author at: U.S. Geological Survey, Wetland and Aquatic Research Center, 3321 College Ave., Davie, FL 33314, USA.
E-mail addresses: kristen_hart@usgs.gov (K.M. Hart), ariverson@usgs.gov (A.R. Iverson), abenscoter@usgs.gov (A.M. Benscoter), ikuko@ufl.edu (I. Fujisaki),
mcherkiss@usgs.gov (M.S. Cherkiss), clayton_pollock@nps.gov (C. Pollock), ian_lundgren@nps.gov, ian.lundgren@noaa.gov (I. Lundgren),
zandy_hillis-starr@nps.gov (Z. Hillis-Starr).
1 Present address: NOAA Inouye Regional Center (IRC), NMFS/PIRO, 1845 Wasp Blvd., Building 176, Honolulu, HI 96818 USA.
Biological Conservation 229 (2019) 113
0006-3207/ 2018 Published by Elsevier Ltd.
T
important for sea turtles. Satellite-tracked green sea turtles (Chelonia
mydas) aggregate in MPAs during foraging (Hart et al., 2013; Scott
et al., 2012) and in some cases degrade seagrass habitat within them,
owing to high turtle concentration (Christianen et al., 2014). Similarly,
nine (83%) of the hawksbill sea turtles (Eretmochelys imbricata) tracked
in the Dominican Republic were predominantly within the local MPA
during inter-nesting (IN; Revuelta et al., 2015) and over half (55%) of
post-nesting female hawksbills were tracked to foraging areas that
overlapped with MPA locations in Brazil (Marcovaldi et al., 2012).
Telemetry data can therefore be useful for evaluating the effectiveness
of current MPA boundaries or creating new protected areas (Dawson
et al., 2017; Maxwell et al., 2011).
Hawksbill
sea
turtles primarily
inhabit coral
reef habitats
throughout the Caribbean (Carr et al., 1966; Mortimer and Donnelly,
2008), but are also recorded in mangrove estuaries in the eastern Pa-
cific (Gaos et al., 2011). Endangered in all parts of their range (NMFS
and USFWS, 1993, 1998), the International Union for Conservation of
Nature (IUCN) Red List of Threatened Species listed hawksbills as cri-
tically endangered in 1996, owing to extensive population declines
(Mortimer and Donnelly, 2008). Hawksbill nesting occurs in few places
in the Caribbean (NMFS and USFWS, 1993), and is much reduced from
historic numbers (McClenachan et al., 2006). However, recent increases
in nesting numbers at monitored locations in Antigua (Richardson
et al., 2006), Barbados (Beggs et al., 2007), Guadeloupe (Kamel and
Delcroix, 2009) and Puerto Rico (Van Dam et al., 2008) are promising.
Previous studies on movement patterns of both sexes of adult
hawksbills in the Caribbean were conducted in Puerto Rico (Van Dam
et al., 2008), Barbados (Horrocks et al., 2001; Walcott et al., 2012),
Lesser Antilles (Esteban et al., 2015), Cuba (Moncada et al., 2012),
Costa Rica (Trong et al., 2005), the Dominican Republic (Hawkes
et al., 2012; Revuelta et al., 2015) and the US Virgin Islands (USVI;
Starbird et al., 1999). Meylan (1999) summarized hawksbill flipper tag
returns to infer foraging areas, and more recent satellite tracking (see
studies cited above) has revealed both local and distant foraging sites in
many countries including Nicaragua, Honduras, Venezuela, Bahamas,
St. Eustatius, St. Maarten, British Virgin Islands, St. Barthlemy, Co-
lumbia, and Mexico, highlighting the importance of understanding the
broad scale spatial ecology of sea turtles during different movement
stages (IN, migration, foraging) for adequate protection. Still, the spa-
tial ecology and movements of hawksbills nesting in the USVI are not
well understood, and there is a paucity of information available for
managers on hawksbill distribution (see Godley et al., 2008 for review).
The habitat-use patterns of adults, including the space-use overlap
between individuals, is useful for managers charged with protecting
reproductively active individuals, which is necessary for population
recovery. The US Hawksbill Recovery Plan calls for more demographic
information on all life stages of hawksbills. This includes, but is not
limited to, their distribution, abundance, and seasonal movements
(NMFS and USFWS, 1993). Buck Island Reef National Monument
(BIRNM) is a protected area in the USVI and the only fully protected
area in the Caribbean where hawksbills both forage and nest (NMFS
and USFWS, 1993). Although BIRNM is an important nesting and
foraging area for hawksbills where a saturation tagging program has
been ongoing for 30 years, little information is available on the spatial
ecology of turtles in this protected area. Only one study has delineated
IN habitat use of hawksbills at BIRNM, through radio-tagging seven
females in 1991, which identified that all IN areas were within 3 km of
Buck Island and in water depths from 9 to 20m (Starbird et al.,
1999). In addition, a case study by Sartain-Iverson et al. (2016) tracked
one hawksbill turtle through IN, migration, and foraging. The long-term
movement patterns during IN, migration, and foraging, with a greater
number of individuals and more advanced telemetry will aid in clar-
ifying the spatial distribution of breeding hawksbills at this important
site.
In this study, we used satellite tracking to delineate IN habitats,
migratory routes, and foraging areas for adult female hawksbills post-
nesting at BIRNM. We assessed habitat-use both within and outside the
protected area boundary and examined remotely sensed depth data to
describe characteristics of selected habitat. As the protected waters
within BIRNM are limited, we also determined the extent of habitat-use
overlap between different hawksbill individuals. Further, we char-
acterized timing of migration, and individual turtle site-loyalty and
occupancy patterns. Finally, we determined protected status of waters
containing foraging areas for conservation relevance.
Fig. 1. Study Area. Buck Island Reef National
Monument (BIRNM, US Virgin Islands), where adult
female hawksbills (Eretmochelys imbricata) were sa-
tellite-tagged from 2011 to 2015. The black line
around Buck Island indicates the current boundary of
BIRNM; the prior BIRNM boundary is represented by
the narrow black outline filled with diagonal lines.
Dotted lines indicate 1000m bathymetry depth con-
tours.
K.M. Hart et al.
Biological Conservation 229 (2019) 113
2
2. Materials & methods
2.1. Study site
Tagging and sampling occurred at BIRNM, which includes 76.3 km2
of submerged lands and the 0.7 km2 uninhabited Buck Island, located
on the shallow St. Croix shelf (approx. 15 to 20m depth), 2.4 km
northeast of St. Croix island in the USVI (Fig. 1). BIRNM is a nesting and
foraging area for loggerhead (Caretta caretta), green (Chelonia mydas),
leatherback (Dermochelys coriacea), and hawksbill sea turtles. Ap-
proximately 5080 individual hawksbill females lay nests annually at
BIRNM.
2.2. Turtle capture and transmitter deployment
Nightly surveys were conducted from 19:00 to 05:00 h from 15 July
through 30 September 20112015. Thirty-two satellite transmitters
were used to monitor movements of 31 post-nesting hawksbill turtles
over a 5 yr period from 2011 to 2015 (Table 1). Turtles were outfitted
with transmitters using established protocols (NMFS-SEFSC, 2008),
following methods in Hart et al. (2017). Briefly, we intercepted nesting
hawksbill females after they finished egg-laying on the beach. We used
PTTs from Wildlife Computers (Redmond, WA, USA; SPOT5 [n=27]
and SPLASH [n= 5] models; dimensions [lengthwidthheight]:
SPOT5725624mm, SPLASH10200725530mm). We
streamlined attachment materials to minimize buoyancy or drag effects
on the turtle's swimming ability and limited the epoxy (two part Su-
perbond epoxy) footprint. Each tag was set as active for 24 h d1 with
duty-cycling implemented in 20122014 December through May (every
3rd day) to conserve battery life.
2.3. Sea turtle tracking and switching state-space modeling
We used Satellite Tracking and Analysis Tool (STAT; Coyne and
Godley, 2005) available on www.seaturtle.org to retrieve location data
(see Appendix A for details on Location Class [LC] accuracy and Argos
location processing); LCs 3, 2, 1, 0, A, and B were used to reconstruct
routes.
We applied switching state-space modeling (SSM; Jonsen et al.,
2003; Patterson et al., 2008) as described in Jonsen et al. (2005) to
determine the beginning and end date of the IN period for each
hawksbill. SSM methods follow our previous studies (see Hart et al.,
2013, 2014, 2015; Shaver et al., 2013, 2016, and Appendix A for in-
formation on this technique). Earlier applications defined a binary be-
havioral mode with 'foraging' and 'migration' (Jonsen et al., 2005,
2007); however, since we tagged animals during the nesting season, our
behavioral mode definitions were 'foraging and/or IN' and 'migration'.
From the 'IN and/or foraging' mode, a period was defined as 'IN' if
points occurred before migration away from the nesting beach. We
summarized data until the transmitters stopped sending information or
until the time of data synthesis: 13 January 2015.
2.4. Migration
We used the SSM approach to determine the beginning and end date
of migration mode for each turtle following Hart et al. (2012). We
present migration periods representing movement away from the IN
area to the foraging grounds (i.e., brief [28 days] movements within IN
Table 1
Size and satellite-tracking dates for hawksbill (Eretmochelys imbricata) nesters tagged at Buck Island, 20112014. N=neophyte, R= remigrant, EEZ= the number of
exclusive economic zones the turtle traveled through (calculated over entire tracking period). Only the filtered migration points outside of Buck Island Reef National
Monument area were used for whether the migratory path crossed a protected area.
Turtle
Size (CCL-tip,
cm)
Tracking start (days) N/R
Migration
Migration Period (days)
Path Total Distance (km) Mean Depth (m)
Cross protected area? (% of filtered
points)
EEZ
1
82.4
8/23/11 (61)
R
9/199/26 (8)
282.9
889.8
Y (5%)
2
2
84.5
8/25/11 (536)
N
10/610/9 (4)
NA
1812.0
N
8
3
85.0
8/25/11 (71)
R
10/1510/25 (11)
425.2
501.4
Y (2%)
2
4
88.2
8/26/11 (382)
R
9/269/27 (2)
70.6
420.9
Y (9%)
3
5
86.7
8/26/11 (600)
R
10/1810/20 (3)
167.1
484.3
N
3
6
83.5
8/27/11 (418)
R
9/119/15 (5)
221.1
870.8
Y (2%)
7
7
83.6
8/27/11 (257)
R
10/510/8 (4)
255.6
774.9
Y (24%)
6
8
84.9
8/28/11 (390)
R
10/510/11 (7)
200.9
1007.8
Y (5%)
2
9
92.6
8/28/11 (582)
R
8/289/1 (4)
154.9
1370.4
Y (3%)
3
10
90.5
7/30/12 (497)
R
9/139/13 (1)
51.9
1938.7
N
3
11
95.0
7/30/12 (310)
R
9/1410/26 (43)
2588.6
1997.4
Y (3%)
7
12
87.3
8/1/12 (384)
N
9/89/21 (14)
589.6
635.7
Y (22%)
8
13
84.7
8/1/12 (247)
R
10/1110/21 (11)
602.2
571.5
Y (12%)
7
14
86.4
8/2/12 (336)
R
9/149/21 (8)
173.4
413.6
Y (2%)
2
15
88.6
8/2/12 (445)
R
10/310/8 (6)
199.9
783.7
N
2
16
88.0
8/3/12 (96)
N
9/239/24 (2)
84.7
530.8
N
2
17
89.9
8/3/12 (495)
R
8/48/8 (5)
197.0
979.8
N
3
18
87.0
8/4/12 (120)
N
9/911/16 (69)
3523.8
1280.4
Y (12%)
7
19
91.8
8/7/12 (480)
N
10/1511/13 (27)
1658.7
1871.6
Y (3%)
6
20
85.2
8/8/12 (91)
N
9/259/29 (5)
238.8
1367.8
N
4
21
86.8
8/9/12 (269)
N
10/511/24 (40)
1533.8
1585.5
Y (7%)
4
22
92.7
7/29/13 (135)
R
9/169/18 (3)
137.3
834.0
Y (23%)
2
23
92.3
7/29/13 (135)
R
8/299/11 (14)
901.7
1640.6
N
4
24
85.1
7/29/13 (109)
N
10/210/5 (4)
202.2
590.8
Y (3%)
4
25
96.3
7/30/13 (134)
R
10/1411/18 (33)
1991.7
1307.2
Y (12%)
5
26
86.9
7/30/13 (134)
N
7/318/4 (5)
213.5
961.9
Y (13%)
3
27
95.3
7/31/13 (133)
R
9/129/15 (4)
189.4
738.7
N
2
28
83.1
8/2/13 (462)
R
none
NA
NA
NA
3
29
86.2
8/3/13 (129)
R
9/149/22 (9)
443.2
284.0
Y (10%)
5
30
83.5
8/5/14 (106)
R
9/3010/4 (5)
180.6
481.8
Y (8%)
2
31
96.5
8/5/14 (162)
R
10/2911/28 (31)
2432.7
3006.6
N
8
32
87.7
8/7/14 (104)
R
8/78/9 (3)
95.2
481.2
N
2
K.M. Hart et al.
Biological Conservation 229 (2019) 113
3
Table2Hawksbill(Eretmochelysimbricata)homerangeanalysisdetailsforindividualstaggedatBuckIslandReefNationalMonument,St.Croix.Filtlocs=filteredlocations,KDE=kerneldensityestimate,MCP=minimumconvexpolygon,MDLs=meandailylocations,C.=centroid,Prot.Area=ProtectedArea.Inter-nestingProtectedAreasincludeBuckIslandReefNationalMonument(BIRNM)andEastEndMarinePark(EEMP,inSt.Croix).ForagingProtectedAreasincludeAgoaSpeciallyProtectedArea(ASPA;CartagenaConvention,Regional),ArrecifesdeTourmalineNaturalReserve(ATNR;National),ArrecifesdelaCordilleraNaturalReserve(ACNR;National),CayosMiskitosyFranjaCosteraInmediataBiologicalReserve(CMFC;National),CabezasdeSanJuanNaturalReserve(CSJNR;National),andNelson'sDockyardNationalPark(NDNP).ForagingareacountryExclusiveEconomicZones(EEZs)includePuertoRico(PR),USVirginIslands(USVI),BritishVirginIslands(BVI),Saint-Martin(SM),Saba(S),SintEustatius(SE),Nicaragua(Nic),Bahamas(Bah),Guadaloupe(Gua),AntiguaandBarbuda(AB),Anguilla(Ang),St.KittsandNevis(KN),Venezuela(Ven)andDominicanRepublic(DR).Italicsmeansthecentroidwaswithin4km,butnotwithin,theprotectedarea.Acellwith"."denotestherearenovaluesavailable.Inter-nestingForagingTurtleInter-nestingperiod(days)Filtlocs50%KDEor95%MCPArea,km2(MDLs)C.depth,(m)C.toshore,(km)Prot.AreaForgingperiod(days)Filtlocs50%KDEor95%MCPArea,km2(MDLs)C.depth,(m)C.toshore,(km)Country(EEZ)Prot.Area18/279/18(23)42MCP525.725238.7No9/2710/22(26)13MCP898.66249522.4PRNo28/2610/5(41)87KDE59.4(40)31.3BIRNM10/102/10(490)1016KDE33.44(319)182.9SMASPA38/2510/14(51)258KDE17.4(51)20.3BIRNM10/2611/3(9)4MCP0.86156.7PRATNR48/279/25(30)131KDE24.9(30)2042.0BIRNM9/2810/21(24)106KDE49.83(24)241.3USVINo4.......10/259/10(322)1183KDE51.48(243)322.7USVINo58/2610/17(53)246KDE77.7(53)21.5BIRNM10/216/4(228)805KDE40.07(169)422.4USVINo5.......6/284/15(292)1063KDE39.44(220)422.2USVINo68/289/10(13)85MCP392.41132.4BIRNM9/1610/24(39)125KDE95.43(32)243.0SENo6.......11/211/19(18)7MCP1082.4277015.2SNo6.......11/3010/17(323)940KDE62.87(231)31.3SENo78/2810/4(38)213KDE14.4(38)20.6BIRNM10/910/19(11)53MCP149.75606.6SMASPA7.......10/235/9(200)524KDE13.58(118)434.6SMASPA88/2910/4(37)116KDE60.3(35)150.9EEMP10/129/19(344)721KDE6.29(201)50.0PRNo9NoINperiod.....9/29/4(369)1052KDE50.66(272)232.1PRACNR9.......9/89/10(3)11MCP44.59285.5PRACNR9.......9/143/30(198)386KDE44.41(114)221.7PRACNR107/319/12(44)305KDE31.1(44)31.1BIRNM9/1412/8(451)1440KDE26.53(357)440.9BVINo117/319/13(45)302KDE27.6(45)60.9EEMP10/276/3(220)858KDE14.76(150)1920.0NicCMFC128/29/7(37)218KDE41.8(37)230.0BIRNM9/228/18(331)953KDE24.67(185)2632.5GuaASPA138/19/2(33)196KDE12.5(33)30.6BIRNM10/224/3(164)370KDE15.11(82)4715.8SNo139/1210/10(29)104KDE22.7(27)30.7BIRNM........148/39/13(42)177KDE9.6(39)30.1BIRNM10/27/3(275)620KDE13.91(137)303.3PRACNR158/310/2(61)310KDE18.6(60)620.8No10/910/19(376)1248KDE9.19(236)2918.1BVINo168/49/22(50)303KDE10.8(50)30.7BIRNM9/2511/6(43)183KDE14.38(43)388.8USVINo17NoINperiod....NA8/912/10(489)2681KDE10.11(365)262.3AngNo(continuedonnextpage)K.M. Hart et al.
Biological Conservation 229 (2019) 113
4
Table2(continued)Inter-nestingForagingTurtleInter-nestingperiod(days)Filtlocs50%KDEor95%MCPArea,km2(MDLs)C.depth,(m)C.toshore,(km)Prot.AreaForgingperiod(days)Filtlocs50%KDEor95%MCPArea,km2(MDLs)C.depth,(m)C.toshore,(km)Country(EEZ)Prot.Area188/59/8(35)179KDE62.7(35)61.8BIRNM11/2711/30(4)30MCP106.17116.1NicCMFC198/810/14(68)405KDE19.3(64)20.6BIRNM11/1412/1(18)100MCP248.13184.3BahNo19.......12/511/28(359)339KDE25.17(65)186.6BahNo208/99/24(47)185KDE31.6(44)61.5EEMP9/3011/5(37)83KDE44.24(23)382.5KNNo218/910/4(57)377KDE14.1(57)31.0BIRNM11/255/4(161)316KDE23.4(61)719.2DRNo227/309/15(48)251KDE13.7(48)60.3BIRNM9/1910/10(22)93KDE24.42(22)251.7PRACNR22......10/1412/10(58)268KDE22.33(52)291.8PRACNR237/308/28(30)208KDE16.8(30)82.5BIRNM9/1212/10(90)399KDE18.1(84)35523.6VenNo247/3010/1(64)377KDE17.4(51)30.5BIRNM10/611/13(39)198KDE29.59(37)1812.6KNNo257/3110/13(75)399KDE20.5(75)20.9BIRNM11/1912/10(22)92MCP364.31139.5BahNo26NoINperiod......8/512/10(128)759KDE41.77(122)392.5PRCSJNR278/19/11(42)187KDE44.3(42)31.4BIRNM9/1612/10(86)340KDE23.32(78)2102.8BVINo288/29/6(36)58MCP905.390.9BIRNM2/229/7(198)175KDE45.69(54)51.6USVINo298/38/21(19)115MCP92.431.6BIRNM9/2312/9(78)205KDE19.2(68)2661.2ABNDNP298/249/13(21)101KDE12(20)411.4BIRNM........308/69/29(55)340KDE19.2(55)31.7BIRNM10/511/18(45)216KDE20.42(44)253.0PRNo318/510/28(85)364KDE19(82)20.4BIRNM11/291/13(46)191KDE21.82(38)132118.3NicNo32NoINperiod......8/1011/18(101)495KDE36.27(99)282.0PRNoK.M. Hart et al.
Biological Conservation 229 (2019) 113
5
or foraging periods were not included). From the raw satellite data, we
filtered out locations that were on land, very distant (> 120 km from
nearest valid point), or that represented straight-line movement
speeds> 5 km h1. We selected the conservative 5 km h1 speed filter
based on Parker et al. (2009) who reported hawksbill travel speeds
between 0.7 and 1.2 km h1 during transit in a Hawaiian study site, and
on previous application of this speed filter in other hawksbill tracking
studies (Gaos et al., 2012; Luschi et al., 1998; Trong et al., 2005; Van
Dam et al., 2008).
We quantified the mean bathymetry (m) across all filtered points in
the migration track, the straight-line distance between IN and foraging
centroids (km; see Migration to foraging areas), and actual distance
along the migration path (km). For bathymetry, we used the ETOPO1
global relief model (bedrock, cell-registered, 1 arc-minute; Amante and
Eakins, 2009). We also determined the number of exclusive economic
zones (EEZ) crossed by each turtle during the entire tracking period
using an EEZ map (Flanders Marine Institute, 2014).
2.5. High-use areas (IN and foraging)
We filtered home range analysis locations as above for migration,
then quantified core-use areas for IN and foraging using 50% kernel
density estimation (KDE); in the absence of 50% KDE, we calculated
95% minimum convex polygon analysis (MCP) to represent the habitat
area used. To minimize autocorrelation of points, mean daily locations
within each IN and foraging period were generated in the software
program R (R Development Core Team, 2014) using filtered satellite
locations, and the resulting coordinates (mean daily) provided raw data
for 50% KDE analyses (applied to periods with>20 mean daily loca-
tions); the filtered points were used to calculate 95% MCP analyses
(applied to periods with< 20 mean daily locations).
Kernel density is a non-parametric method that uses appropriate
weighting of outlying observations
to
identify areas of dis-
proportionately heavy use within a home range (White and Garrott,
1990; Worton, 1987, 1989). To create KDEs, we used the Home Range
Tools for ArcGIS extension (Rodgers and Kie, 2011) and fixed-kernel
least-squares cross-validation smoothing factor (hcv) (n= 23 turtles;
Seaman and Powell, 1996; Worton, 1995) as well as a custom script in R
using the package 'adehabitatHR' (Calenge, 2006; n= 2 turtles, see
Appendix A). When x and y coordinates had unequal variances, data
were rescaled to select the best bandwidth (Laver and Kelly, 2008;
Seaman and Powell, 1996). We used ArcGIS 9.3 (ESRI, 2007) to cal-
culate the in-water area (km2) within each kernel density contour; the
50% KDEs represent core-use area of activity (Hooge et al., 2001).
Minimum convex polygons (MCPs) were created using ArcMap 9.3
(ESRI, 2007; n= 3 turtles) and a custom script in R using the package
'adehabitatHR' (Calenge, 2006; n= 1 turtle). Following Walcott et al.
(2012), we created MCP polygons using 95% of filtered points, as it is
possible for a proportion of distant filtered locations to represent only
occasional movements outside the home range area (Burt, 1943;
Rodgers and Kie, 2011).
We quantified site-fidelity to IN and foraging areas using the Animal
Movement Analysis Extension for ArcView 3.2. Using Monte Carlo
Random Walk simulations (100 and 200 replicates for IN and foraging,
respectively), we tested for spatial randomness of tracks against ran-
domly generated walks (Hooge et al., 2001). Random walks were bound
from 4500m to 0m bathymetry to encompass all filtered locations
during IN, and 5200 to 0m to encompass all filtered locations during
foraging. Tracks exhibiting site-fidelity signify movements that are
spatially constrained and not randomly dispersed (Hooge et al., 2001).
We did not use tracks failing site fidelity in home range analyses.
2.6. Characteristics of high-use areas
We calculated centroids for 50% KDEs and 95% MCPs (both IN and
foraging) following Hart et al. (2017). For each centroid we determined
bathymetry, distance to nearest land, and the MPA status. For bathy-
metry in Caribbean waters, we used the GEBCO_2014 Grid (General
Bathymetric Chart of the Oceans) a 30 arc-second continuous terrain
model of both ocean and land (www.gebco.net; accessed 30 June
2016).
To depict IN locations for all turtles, we calculated the number of
turtle-tracking days in grid cells (2 2 km) around BIRNM, using both
"old" and "new" BIRNM boundaries. Specifically, we counted the
number of days each turtle was observed in each grid cell using filtered
IN points. With this grid, we determined the number of IN days inside
and out of the BIRNM boundary. We also extracted bathymetry values
for high-use cells using the GEBCO_2014 Grid. To ensure independence
of the core-area size and tracking duration, we examined the associa-
tion between the two variables for IN using Spearman's . We also
calculated Spearman's between IN core-area size and turtle size (CCL).
2.7. Core area space-use sharing
We conducted a Kruskal-Wallis non-parametric one-way ANOVA to
determine whether there was a difference in space-use sharing among
turtles with different levels of nesting experience: neophyte-neophyte
(NN), neophyte-remigrant (NR), and remigrant-remigrant (RR) turtle
pairs; where neophytes are first-time nesters and remigrants are repeat
nesters at BIRNM. We calculated the core-use area (50% KDEs) space-
use sharing during IN and at common foraging grounds using the
'adehabitat' package in R (Calenge, 2006; R Development Core Team,
2014). We used the utilization distribution overlap index (UDOI) fol-
lowing our previous work (Hart et al., 2017), which is considered the
most appropriate measure of animal space-use sharing (Fieberg and
Kochanny, 2005). Two UDs with no overlap produce a UDOI value of
zero, whereas uniformly distributed UDs have a UDOI value of 1. The
UDOI value can also be>1 for non-uniformly distributed UDs with a
high degree of overlap, which indicates a higher than normal overlap
relative to uniform space-use (Fieberg and Kochanny, 2005). The UDOI
space-use sharing was calculated for the 25 turtles that had 50% KDEs
during IN (n= 300 pairs), as well as the amount of temporal overlap
(days) that turtle pairs were foraging in the same area.
2.8. Regional hawksbill foraging ranges
We mapped foraging locations of female hawksbills satellite-tagged
on Buck Island, US Virgin Islands (n= 31 turtles) along with female
hawksbills from other studies that were tagged in the Caribbean and
satellite-tracked to foraging grounds (n=33 turtles from other studies;
Esteban et al., 2015 [n= 2 turtles], Horrocks et al., 2001 [n=4 tur-
tles], Moncada et al., 2012 [n= 9 turtles], Revuelta et al., 2015 [n=9
turtles], Trong et al., 2005 [n= 2 turtles],Van Dam et al., 2008 [n=7
turtles]). Some studies provided foraging location XY coordinates
(Horrocks et al., 2001; Van Dam et al., 2008); for the other studies,
figure images were georeferenced in ArcGIS v10.2.2 (ESRI, 2014), and
foraging locations were calculated using the center of post-nesting
movement points (Trong et al., 2005), the last migration or tracking
point (Moncada et al., 2012; Esteban et al., 2015), or by determining
the centroid of a 95% MCP or 50% KDE (Revuelta et al., 2015). In total
we mapped 66 foraging locations (e.g., centroids) across seven studies
(including this one). For each centroid, we calculated the number of
centroids within 30.6 km, the average size of the core-use area (50%
KDE) reported in this study.
3. Results
3.1. Turtles
Turtles (n=31 individuals, one tracked in two different years for
32 total tracks) ranged in size from 82.496.5 cm curved carapace
length (CCL; mean SD=88.1 4.1 cm, Table 1). We tracked
K.M. Hart et al.
Biological Conservation 229 (2019) 113
6
turtles for a total of 8810 days, ranging from 61 to 600 days (mean
SD=275.3 173.3 d).
Turtles were remigrants (n= 22) and neophytes (n=9; see
Table 1). Remigrant histories for two individuals date back to the be-
ginning of the saturation-tagging program at Buck Island in 1989
(25 years, n= 1 turtle) and 1993 (21 years, n= 1 turtle). Most re-
migrants were first encountered at BIRNM between 2001 and 2005
(913 years, n= 12 turtles) and the remainder from 2007 to 2009
(57 years, n= 8 turtles; see Table A1). Remigration intervals varied
between turtles with 15 (68%) having 2 yr, 3 yr, or 23 yr remigration
intervals (see Table A1).
3.2. Inter-nesting
We obtained SSM results for all 31 turtles/32 tracks (Fig. A1 and
Table A2 provide example SSM prediction paths and model para-
meters). Twenty-eight turtles had locations available during IN; several
turtles departed the study area immediately after nesting so thus did not
have IN data. Twenty-five of the 28 turtles had enough mean daily
locations for KDE analysis, and all of these displayed site fidelity
(p > 99.0099 for all turtles) for a total of 26 KDEs (Turtle 13 had two
IN KDEs; Table 2). We obtained 1196 mean daily locations for KDE
periods totaling 1218 days (Table 2). The overall size of core-use areas
(50% KDEs) ranged from 9.677.7 km2 (Table 2). Core-use area size
(=0.29, p=0.16) was not strongly associated with number of
tracking days. Also, turtle size (CCL) was not associated with core-use
area size (=0.02, p=0.94).
When KDE analyses were not possible, we calculated 95% MCPs
(Fig. A2); the four turtles showed site fidelity (p > 99.0099) and had a
total of 300 filtered locations over 92 days for analysis. MCP areas
ranged from 92.4905.3 km2 (Table 2). We calculated centroids for
50% KDEs and 95% MCPs (Fig. 2). The mean distances to the nearest
land from core-use area centroids was 1.0 km (3.4 km for MCPs;
Table 2). Mean bathymetry at these centroid locations was 16.2m
(662.0 m for MCPs; Table 2). Two centroids had erroneous positive
values and most of the remaining centroids were in shallow water (21
of 24 centroids>10m; Table 2).
For all turtles, the number of IN days in the current BIRNM
boundary grid (including intersecting cells) was 3212 (68%; total of
4751 turtle days across all cells), with 1539 (32%) days outside the
boundary (Fig. 2). For comparison, the number of IN days in the "old"
BIRNM boundary (see Proclamation 7392, 2001) in place prior to 2001
was 1409 (30%). The mean bathymetry of grid cells with the highest
number of IN turtle-days (203428 days, red cells in Fig. 2 inset) was
4.9m. The median bathymetry values for turtle-day grid cells
(58428 days) were between 3 and 37m.
3.3. Core area space-use sharing during inter-nesting
We calculated UDOI space-use sharing for 300 turtle pairs during IN
(n= 25 turtles). Across all pairs, UDOI ranged from 0 to 0.23
(mean SD=0.08 0.06; Table A3, Fig. A3), where greater UDOI
indicates greater space-using sharing between turtle pairs. The mean
( SD) UDOI was 0.09 0.05 for NN pairs (n=28), 0.08 0.06 for
NR pairs (n=136), and 0.07 0.06 for RR pairs (n=136). Temporal
overlap
across
all
pairs
ranged
from
0
to
62 days
(mean SD=10.5 17.3). The non-parametric ANOVA (Kruskal-
Wallis) comparing UDOI habitat overlap among NN, NR, and RR pairs
was significant (H=6.91, df= 2, p= 0.03). Although there was evi-
dence for greater space-use sharing for NN turtle pairs compared to RR
pairs, pairwise comparisons were not significant when the alpha value
was corrected for the false discovery rate in conducting multiple com-
parisons (Benjamini and Hochberg, 1995; Garcia, 2004).
3.4. Migration to foraging areas
Of the 31 satellite-tagged turtles, 30 turtles showed migration paths
away from Buck Island to their foraging grounds between July and
October. Some migrations to foraging grounds were split by brief
foraging "stopover" periods along the route, which resulted in 37 mi-
gration paths across the 31 turtles (Fig. 3). The "final" hawksbill fora-
ging grounds were in many geographic areas (Fig. 4, Table A4) in-
cluding the Bahaman (Lucayan) Archipelago, east of Nicaragua, the
Greater Antilles including the waters surrounding Puerto Rico and the
Dominican Republic, the Leeward Islands of the Lesser Antilles (e.g.,
British Virgin Islands, US Virgin Islands, St. Kitts and Nevis, Anguilla),
and one turtle that foraged in Venezuelan waters (see Table A4). Mi-
gration periods lasted 169 days (Table 1, Fig. 3). Average mean
Fig. 2. Inter-nesting (IN) areas. Core-use IN areas
(50% kernel density estimation [KDE] and 95%
minimum convex polygon [MCP]) for 25 nesting fe-
male hawksbills (Eretmochelys imbricata) satellite-
tagged on Buck Island (26 KDEs; red line). The red
KDE line represents the outer boundary of all KDEs.
The black line is the Buck Island Reef National
Monument (BIRNM); the tan area shows the prior
BIRNM boundary, before expansion. The KDE (cir-
cles) and minimum convex polygon (MCP; triangles)
centroids are shown for 28 nesting female hawksbills
(26 KDEs for 25 turtles and 4 MCPs for 4 turtles; one
turtle had both KDE and MCP for 28 total turtles).
The star indicates approximate tagging areas (bea-
ches on south and west shores of Buck Island). Red
and yellow centroids in main panel show centroids
for turtles with more than one IN area (KDE and/or
MCP). Inset: The IN days per 2 km grid cell for 28
nesting
female hawksbills
satellite-tagged post-
nesting on Buck Island. World Exclusive Economic
Zones (EEZs) are shown by the dark blue lines.
Abbreviations for inset: VI = U.S Virgin Islands; BVI
= British Virgin Islands; PR = Puerto Rico. (For in-
terpretation of the references to colour in this figure
legend, the reader is referred to the web version of
this article.)
K.M. Hart et al.
Biological Conservation 229 (2019) 113
7
(caption on next page)
K.M. Hart et al.
Biological Conservation 229 (2019) 113
8
bathymetry across all migration periods was 1045.7 m; Table 1).
Straight-line migration distance ranged from 20.22370.0 km and the
distance along the migration path ranged from 51.93523.8 km
(Table 1, Fig. 3). During the entire tracking period, turtles crossed be-
tween two and eight different EEZ zones (Table 1).
3.5. Foraging areas
All 32 tracks had SSM results with time periods predicted as fora-
ging. Across the 32 tracks, there were 41 distinct foraging periods
(some turtles had multiple foraging periods that were interrupted by
brief movements, as determined by SSM) that exhibited site fidelity. Of
these 41foraging periods, 33 had adequate sample size to conduct 50%
KDE analyses (Table 2), and 95% MCPs were determined for the other 8
(Table 2). The 33 KDE foraging periods totaled 6626 days across all
turtles, and ranged from 22 to 490 d. We obtained 4345 mean daily
locations for KDE analyses and the range of core-use areas (50% KDEs)
was 6.395.4 km2 (Table 2). The eight 95% MCPs totaled 111 days
across all eight turtles and 310 filtered locations. MCP area ranged from
0.91082.4 km2 (Table 2, Fig. A2). Mean distance to the nearest land
from centroids of 50% KDEs was 10.8 km (MCP mean=23.3 m;
Table 2) and mean bathymetry at the 50% KDE centroid locations was
65.8 m (MCP mean=422.6 m; Table 2, Fig. 4).
3.6. Core area space-use sharing during foraging
We calculated the amount of UDOI space-use sharing for 66 turtle
pairs foraging near two common foraging areas in Puerto Rico and the
British Virgin Islands (n= 12 turtles). Across all 66 pairs, UDOI ranged
from 0 to 0.17 (mean SD=0.003 0.02; Table A5), where greater
UDOI indicates greater space-using sharing between turtle pairs.
Temporal overlap across all pairs ranged from 0 to 494 days
(mean SD=74.2 119.8).
3.7. Regional foraging areas
Hawksbills forage in numerous areas across the Caribbean Sea
(Figs. 4 and 5). Furthermore, multiple areas provide foraging sites for
breeding turtles that nest in different locations (Fig. 6); foraging areas
include the waters east of Nicaragua and Honduras, the waters east of
Puerto Rico in the Greater Antilles, and the Leeward Islands of the
Lesser Antilles. Of the 66 foraging locations across the seven studies
(including this one), there were 32 turtles (i.e., centroids) that had at
least one other centroid within 30.6 km, the average size of the core-use
area (50% KDE) in this study. Of these 32 total tracks, 13 had one
centroid within 30.6 km, and the other 19 had 212 centroids within
30.6 km.
4. Discussion
By using satellite tracking technologies alongside advanced spatial
modeling approaches, we delineated important in-water habitats used
by hawksbills during inter-nesting periods, through migration, and at
foraging areas. All turtles were tagged after nesting at BIRNM, a
Caribbean MPA that supports breeding hawksbills that migrate and
forage through waters of multiple countries (Sartain-Iverson et al.,
2016). Migration paths crossed through multiple EEZs as turtles tra-
veled to foraging sites in 14 different countries; these results underscore
the importance of international conservation initiatives for the recovery
of depleted hawksbill populations in the Caribbean basin. Using the
robust method of SSM, we determined not only the size and location of
intensely used areas, but also the time periods when turtles moved
through international waters and arrived at their respective foraging
areas. We also further characterized 'overlap' of individual space-use at
foraging sites, which underscores the importance of these supporting
resources that are critical for turtle survival. This is the first study to
delineate high-use habitats throughout IN, foraging, and migration, for
multiple critically endangered hawksbills nesting at BIRNM.
Fig. 3. Migration paths. Migration paths from IN grounds to foraging grounds of 31 adult female hawksbills (Eretmochelys imbricata) satellite-tagged on Buck Island,
US Virgin Islands (USVI), and migrating to A. the Leeward Islands (n= 9 turtles) and Venezuela (n= 1 turtle), B. the Greater Antilles (n=15 turtles; BVI= British
Virgin Islands), and C. the Dominican Republic, the Bahamas, and Nicaragua. Circles represent centroids of 50% foraging kernel density estimation core-use areas
(50% KDEs) and triangles represent centroids of 95% foraging minimum convex polygons (MCPs). Stars represent tagging location and the origin of migration.
Fig. 4. Foraging Centroids. Foraging locations for
female hawksbills (Eretmochelys imbricata) that were
satellite-tagged in this study (nested on Buck Island
[star], n= 31). Remigrant and neophyte foraging
areas often occurred in similar areas/regions and
foraging occurred in shallow waters (light blue). (For
interpretation of the references to colour in this
figure legend, the reader is referred to the web ver-
sion of this article.)
K.M. Hart et al.
Biological Conservation 229 (2019) 113
9
4.1. Inter-nesting (IN)
We found that IN core-use areas (mean 28.1 km2) were larger than
previously found in this area through radio telemetry (resident areas
within 1.5 km2; Starbird et al., 1999). Satellite telemetry may have a
larger spatial error than radio-telemetry resulting in larger home range
analysis estimates, but satellite tracking has advantages including a
more robust representation of the IN duration. The previous study
tracked 7 turtles for up to 45 days, whereas IN periods defined here for
28 turtles by SSM were up to 85 days; this longer tracking period may
have resulted in more widespread locations. In the Dominican Republic,
mean IN residence areas delineated by satellite tracking were 37.1 km2
(90% KDEs) and 13.2 km2 for core-use areas (50% KDEs; Revuelta et al.,
2015). In Barbados, 23 hawksbill IN residence areas for 17 individual
turtles generated using GPS satellite tags ranged from 0.01 to 0.40 km2
(Walcott et al., 2012). Home range sizes are likely influenced by local
resources as well as analytical methodology. For example, the location
accuracy as well as the number of locations used in analyses can greatly
influence home range area estimates (Thomson et al., 2017). Com-
bining locations received from an existing acoustic array (1st author,
unpublished results) at BIRNM with satellite locations could help clarify
finer-scale IN habitat use patterns for hawksbills in this area. Further,
Fastloc-GPS technology provides locations with high accuracy and
could be used to refine home range analysis estimates and uncover
details on patch-use within core-use areas (Thomson et al., 2017).
We found that both neophytes and remigrants (with nesting records
up to 24 years) used habitat close to the nesting beach in and around
Buck Island during IN (up to 2.5 km [core-use areas] and 8.7 km
[MCPs]). This finding is in line with Starbird et al. (1999) who showed
seven adult females stayed within 3 km of the nesting beach during IN.
These distances are similar to other sites in the Caribbean such as
Barbados (mean 6.1 km; range: 0.721.2 km; Walcott et al., 2012), the
Dominican Republic (mean maximum distance of 39 km, but usually
from 1.4 to 4.3 km; Revuelta et al., 2015), and Costa Rica where one
hawksbill stayed within 30 km of the nesting beach (Trong et al.,
2005). A recent study on the diving behavior of gravid hawksbills from
a nearby site in USVI (on St. Croix) found that turtles rested on the
seafloor and spent most of the IN time at a single depth range, which
could indicate staying within a restricted area (Hill et al., 2017). Re-
maining near the nesting beach during IN may help females conserve
energy as they transit to or from nesting sites.
In addition to being close to shore, IN habitat was also in shallow
water (median bathymetry values in high-use grid were 3 to37m).
This finding is similar to results from tracked hawksbills in the
Dominican Republic (50% KDEs over water>100m; Revuelta et al.,
2015) and Barbados (18 to 41.5 m; Walcott et al., 2012) as well as
for hawksbills tracked at nearby St. Croix (most of time spent at 20 to
30m or less; Hill et al., 2017) and in BIRNM with radio-telemetry
(9 to 20m; Starbird et al., 1999). The grid cells with the highest IN
turtle-days, while in shallow water, were north of Buck Island which
offers closer access to the deep-water shelf to the north. Hill et al.
(2017) found that some St. Croix nesting hawksbills traveled to deeper
depths (up to 95m) during IN. Obtaining breeding season dive in-
formation for hawksbills nesting at BIRNM could help discern whether
these turtles occasionally travel to nearby deep waters during their
reproductive phase.
The IN core-use areas were all in the same general area NE of St.
Croix and surrounding BIRNM, and we found spatial overlap for almost
all of the 300 turtle pairs (n= 25 turtles) with temporal overlap from 0
to 62 days. Neophyte pairs had slightly higher mean space-use sharing
than neophyte-remigrant pairs, with remigrant pairs having the lowest
mean space-use sharing. It is possible that the small sample size of
neophytes contributed to the non-significant statistical result (after the
false discovery rate correction). Space-use sharing was also seen in the
Dominican Republic, with hawksbills having large overlaps of IN and
common-use areas (37.9 km2 for 50% KDEs and 212.2 km2 for 90%
KDEs; Revuelta et al., 2015). While the much smaller IN areas of
hawksbills tagged in Barbados with GPS tags did not show overlap
within years (Walcott et al., 2012) these authors did find overlap in
residence areas from females tracked in different years. That remigrant
pairs at BIRNM had lower overlap indicates that they may use their
experience to benefit from sites unoccupied by other nesters.
While in-water habitat-use was very similar across turtles, we did
observe some plasticity in nesting behavior. Two turtles (one neophyte
[turtle 26], one remigrant [turtle 32]) left the area immediately after
tagging. Leaving the nesting beach early could represent either migra-
tion to foraging grounds or movement to a different nesting beach (e.g.,
Revuelta et al., 2015; Esteban et al., 2015). One of our tracked neophyte
turtles (turtle 19) showed nesting variation when after her 2year re-
migration, we received high-quality locations at a beach about 12 km
distant on St. Croix (Sartain-Iverson et al., 2016). Turtles tagged in
another study in the Lesser Antilles showed nesting site-selection
plasticity as well: one hawksbill traveled in a circular pattern over
Fig. 5. Inter-nesting, foraging, and movement timeline for hawksbill turtles
tagged after nesting at on Buck Island, US Virgin Islands. Turtles are organized
by their foraging destinations. Breaks in the timeline indicate modes other than
inter-nesting, migration or foraging, such as short movements between re-
sidency periods.
K.M. Hart et al.
Biological Conservation 229 (2019) 113
10
200 km from the original nesting site, likely nesting in two other places
(Anguilla and St Croix) before returning to forage within 50 km of the
original site (Esteban et al., 2015). Similarly, hawksbills in the Do-
minican Republic nested at beaches up to 190 km apart (n=2; Esteban
et al., 2015). Characteristics of the nesting beach can influence the
proportion of hatchlings to survive (Lee and Hays, 2004), so nesting at
multiple beaches may provide an evolutionary advantage.
Most turtle-tracking-days during IN (68%) were inside the currently
protected area of BIRNM which has reduced human impacts (fishing
restrictions, no light and no point-source pollution from Buck Island,
minimal boat traffic on north side due to shallow complex reef areas).
This is an improvement in comparison to the 30% observed turtle-
tracking days in the previous BIRNM boundary (Fig. 2); the current
boundary of BIRNM was expanded in 2001 from the original 1961
designation, adding 73.4 km2 of submerged lands (Proclamation 7392,
2001). In addition, most human recreation around Buck Island takes
place to the south and is limited to daytime. Future habitat assessments
of the reefs in this area along with fine-scale activity data could help
point to important habitat features such as preferable reef structures for
resting as well as help determine how females allocate their activity
budgets during this energetically expensive time.
4.2. Migration periods
Timing of migration varied by individual and year, but most turtles
traveled through multiple EEZs (range 28, mean= 4, mode= 2) be-
tween July and October. Turtles began migration periods in July
(n=1), August (n=4), September (n=14) and October (n=12),
and individual migration periods ranged from 2 to 69 days, with longer
migrations for those that traveled across the Caribbean or to the
Bahamas (Fig. 4). These results highlight the late summer and early fall
as a critical time period to protect migrating females. However, few of
these paths passed through other protected areas, indicating vulner-
ability to anthropogenic threats such as major shipping lanes during the
migration periods. Depths on migration routes varied and reached up to
approximately 3000m.
4.3. Foraging areas
Turtles arrived at foraging areas in August (n= 3), September
(n= 12), October (n=11), and November (n=5); one turtle (turtle
28) had unclear arrival time owing to a time lapse in transmissions
(Fig. 5). Locations and sizes of foraging areas in our study were similar
to those in previous tracking studies (Fig. 6) but we did not track any
turtles to Cuba (see Moncada et al., 2012) or as far south as Trinidad
and Tobago (see Horrocks et al., 2001). However, additional tracking
efforts may reveal use of those foraging areas and others by BIRNM
nesting hawksbills. The common foraging areas used in Puerto Rico and
off the coast of Nicaragua represent hotspots where multiple turtles
took up residence; such information can be used to prioritize these
areas for conservation. As several of the foraging areas delineated here
are within or adjacent to current MPA boundaries, a fine-scale ex-
amination of habitat-use and movement patterns at those sites is war-
ranted to assess how well current boundaries encompass required core-
use areas.
5. Conclusions
Our results show previously unknown habitat-use patterns and
highlight concentrated areas of use by hawksbills both within and ad-
jacent to a US protected area during the inter-nesting season.
Individuals used areas within the recently expanded boundary, high-
lighting that this additional protected area is beneficial for this im-
periled species. However, our results also clearly demonstrate the need
for international conservation to protect hawksbills, as migrating turtles
crossed between two and eight different national jurisdictions. These
results provide critical spatial and temporal information for managers
charged with designing strategies to minimize human impact to this
globally imperiled species. Although protecting migratory corridors
would be challenging due to the international jurisdictions and remote
nature of open ocean locations used during migration between breeding
and foraging areas, protection of distinct foraging areas designated here
may be possible. As adult female survival rates have an especially
strong effect on population recovery (NMFS and USFWS, 1993), man-
agement strategies could beneficially focus on protecting adult females.
Acknowledgements
We thank NPS interns and USGS employees J. Beauchamp, M.
Denton, A. Daniels, H. Crowell, A. Crowder, and B. Smith for assistance
Fig. 6. Caribbean hawksbill foraging areas. Foraging
locations of female hawksbills (Eretmochelys im-
bricata) that were satellite-tagged in this study
(nested on Buck Island, n=31), compared to other
studies
that
tagged female hawksbills
in
the
Caribbean and satellite-tracked them to foraging
grounds (Esteban et al., 2015 [n=2 turtles],
Horrocks et al., 2001 [n= 4 turtles], Moncada et al.,
2012 [n= 9 turtles], Revuelta et al., 2015 [n=9
turtles], Trong et al., 2005 [n=2 turtles], Van Dam
et al., 2008 [n= 7 turtles]). Inset shows the nest/
tagging locations for each study.
K.M. Hart et al.
Biological Conservation 229 (2019) 113
11
deploying satellite tags in the field. B. Smith helped with editing tables
and creating Figure 5 and E. Connolly-Randazzo assisted with literature
searches. We thank T. Selby for helpful comments on an earlier version
of the manuscript. Permission to tag and sample turtles was given under
various
permits
(BUIS2012SCI0002,
BUIS-2014-SCI-0009,
USGSSESCIACUC 201105, USFWS permit TE38906B-0 issued to I.
Lundgren, Government of the Virgin Islands Department of Planning
and Natural Resources scientific permit #STX-02 issued to J. Tutein).
Funding was provided by the USGS Natural Resources Protection
Program, the National Park Service, and USGS Ecosystems Program.
Any use of trade, product, or firm names is for descriptive purposes only
and does not imply endorsement by the U.S. Government.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.biocon.2018.11.011.
References
Agardy, M.T., 1994. Advances in marine conservation: the role of marine protected areas.
Trends Ecol. Evol. 9, 267270.
Amante, C., Eakins, B.W., 2009. ETOPO1 1 ArcMinute Global Relief Model: Procedures,
Data Sources and Analysis. NOAA Technical Memorandum NESDIS NGDC24.
National Geophysical Data Center, NOAAhttps://doi.org/10.7289/V5C8276M. (8
April 2014).
Beggs, J.A., Horrocks, J.A., Krueger, B.H., 2007. Increase in hawksbill sea turtle
Eretmochelys imbricata nesting in Barbados, West Indies. Endanger. Species Res. 3,
159168. https://doi.org/10.3354/esr003159.
Benjamini, Y., Hochberg, Y., 1995. Controlling the false discovery rate: a practical and
powerful approach to multiple testing. J. R. Stat. Soc. Ser. B 57, 289300.
Burt, W.H., 1943. Territoriality and home range concepts as applied to mammals. J.
Mammal. 24, 346352.
Calenge, C., 2006. The package "adehabitat" for the R software: a tool for the analysis of
space and habitat use by animals. Ecol. Model. 197, 516519. https://doi.org/10.
1016/j.ecolmodel.2006.03.017.
Carr, A., Hirth, H.F., Ogren, L., 1966. The ecology and migrations of sea turtles, 6. In: The
Hawksbill Turtle in the Caribbean Sea. American Museum of Natural History, New
York, New York, USA.
Christianen, M.J.A., Herman, P.M.J., Bouma, T.J., Lamers, L.P.M., van Katwijk, M.M., van
der Heide, T., Mumby, P.J., Silliman, B.R., Engelhard, S.L., van de Kerk, M., Kiswara,
W., van de Koppel, J., 2014. Habitat collapse due to overgrazing threatens turtle
conservation in marine protected areas. Proc. Biol. Sci. 281, 20132890. https://doi.
org/10.1098/rspb.2013.2890.
Costello, M., Ballantine, B., 2015. Biodiversity conservation should focus on no-take
marine reserves: 94% of marine protected areas allow fishing. Trends Ecol. Evol. 30,
507509.
Coyne, M.S., Godley, B.J., 2005. Satellite Tracking and Analysis Tool (STAT): an in-
tegrated system for archiving, analyzing and mapping animal tracking data. Mar.
Ecol. Prog. Ser. 301, 17. https://doi.org/10.3354/meps301001.
Dawson, T.M., Formia, A., Agambou, P.D., Asseko, G.M., Boussamba, F., Cardiec, F.,
Chartrain, E., Doherty, P.D., Fay, J.M., Godley, B.J., Lambert, F., Mabert, B.D.K.,
Manfoumbi, J.C., Metcalfe, K., Minton, G., Ndanga, I., Nzegoue, J., Oliwina, C.K.K.,
Du Plessis, P., Sounguet, G.-P., Tilley, D., Witt, M.J., Maxwell, S.M., 2017. Informing
marine protected area designation and management for nesting olive ridley sea
turtles using satellite tracking. Front. Mar. Sci. 29, 312. https://doi.org/10.3389/
fmars.2017.00312.
Edgar, G.J., Stuart-Smith, R.D., Willis, T.J., Kininmonth, S., Baker, S.C., Banks, S., Barrett,
N.S., Becerro, M.A., Bernard, A.T.F., Berkhout, J., Buxton, C.D., Campbell, S.J.,
Cooper, A.T., Davey, M., Edgar, S.C., Frsterra, G., Galvn, D.E., Irigoyen, A.J.,
Kushner, D.J., Moura, R., Parnell, P.E., Shears, N.T., Soler, G., Strain, E.M.A.,
Thomson, R.J., 2014. Global conservation outcomes depend on marine protected
areas with five key features. Nature 506, 216220. https://doi.org/10.1038/
nature13022.
ESRI [Environmental Systems Research Institute], 2007. ArcGIS Release 9.3. Redlands,
CA, USA.
ESRI [Environmental Systems Research Institute], 2014. ArcGIS Release 10.2.2. Redlands,
CA, USA.
Esteban, N., Van Dam, R.P., Harrison, E., Herrera, A., Berkel, J., 2015. Green and
hawksbill turtles in the Lesser Antilles demonstrate behavioural plasticity in inter-
nesting behaviour and post-nesting migration. Mar. Biol. 162, 11531163. https://
doi.org/10.1007/s0022701526562.
Fieberg, J., Kochanny, C.O., 2005. Quantifying home-range overlap: the importance of
the utilization distribution. J. Wildl. Manag. 69, 13461359. https://doi.org/10.
2193/0022541X(2005)69[1346:QHOTIO]2.0.CO;2.
Flanders Marine Institute, 2014. Exclusive Economic Zones Boundaries (EEZ). Available
at. http://www.marineregions.org, Accessed date: 21 September 2017.
Fossette, S., Hobson, V.J., Girard, C., Calmettes, B., Gaspar, P., Georges, J.-Y., Hays, G.C.,
2010. Spatio-temporal foraging patterns of a giant zooplanktivore, the leatherback
turtle. J. Mar. Syst. 81, 225234. https://doi.org/10.1016/j.jmarsys.2009.12.002.
Gaos, A.R., Lewison, R.L., Yaez, I.L., Wallace, B.P., Liles, M.J., Nichols, W.J., Baquero,
A., Hasbn, C.R., Vasquez, M., Urteaga, J., Seminoff, J.A., 2011. Shifting the life-
history paradigm: discovery of novel habitat use by hawksbill turtles. Biol. Lett. 8,
5456. https://doi.org/10.1098/rsbl.2011.0603.
Gaos, A.R., Lewison, R.L., Wallace, B.P., Yaez, I.L., Liles, M.J., Nichols, W.J., Baquero,
A., Hasbn, C.R., Vasquez, M., Urteaga, J., Seminoff, J.A., 2012. Spatial ecology of
critically endangered hawksbill turtles Eretmochelys imbricata: implications for man-
agement and conservation. Mar. Ecol. Prog. Ser. 450, 181194.
Garcia, L.V., 2004. Escaping the Boferroni iron claw in ecological studies. Oikos 105,
657663.
Godley, B.J., Blumenthal, J.M., Broderick, A.C., Coyne, M.S., Godfrey, M.H., Hawkes,
L.A., Witt, M.J., 2008. Satellite tracking of sea turtles: where have we been and where
do we go next? Endanger. Species Res. 4, 322. https://doi.org/10.3354/esr00060.
Hart, K., Hyrenbach, K., 2009. Satellite telemetry of marine megavertebrates: the coming
of age of an experimental science. Endanger. Species Res. 10, 920. https://doi.org/
10.3354/esr00238.
Hart, K.M., Lamont, M.M., Fujisaki, I., Tucker, A.D., Carthy, R.R., 2012. Common coastal
foraging areas for loggerheads in the Gulf of Mexico: opportunities for marine con-
servation. Biol. Conserv. 145, 185194.
Hart, K.M., Zawada, D.G., Fujisaki, I., Lidz, B.H., 2013. Habitat use of breeding green
turtles Chelonia mydas tagged in dry Tortugas National Park: making use of local and
regional MPAs. Biol. Conserv. 161, 142154. https://doi.org/10.1016/j.biocon.2013.
03.019.
Hart, K.M., Lamont, M.M., Sartain, A.R., Fujisaki, I., 2014. Migration, foraging, and re-
sidency patterns for northern gulf loggerheads: implications of local threats and in-
ternational movements. PLoS One 9, e103453. https://doi.org/10.1371/journal.
pone.0103453.
Hart, K.M., Sartain, A.R., Fujisaki, I., 2015. Bahamas connection: residence areas selected
by breeding female loggerheads tagged in Dry Tortugas National Park, USA. Anim.
Biotelem. 3, 3. https://doi.org/10.1186/s4031701400192.
Hart, K.M., Iverson, A.R., Benscoter, A.M., Fujisaki, I., Cherkiss, M.S., Pollock, C.,
Lundgren, I., Hillis-Starr, Z., 2017. Resident areas and migrations of female green
turtles nesting at Buck Island Reef National Monument, St. Croix, US Virgin Islands.
Endanger. Species Res. 32, 89101. https://doi.org/10.3354/esr00793.
Hawkes, L.A., Toms, J., Revuelta, O., Len, Y.M., Blumenthal, J.M., Broderick, A.C.,
Fish, M., Raga, J.A., Witt, M.J., Godley, B.J., 2012. Migratory patterns in hawksbill
turtles described by satellite tracking. Mar. Ecol. Prog. Ser. 461, 223232.
Hazen, E.L., Maxwell, S.M., Bailey, H., Bograd, S.J., Hamann, M., Gaspar, P., Godley, B.J.,
Shillinger, G.L., 2012. Ontogeny in marine tagging and tracking science: technologies
and data gaps. Mar. Ecol. Prog. Ser. 457, 221240. https://doi.org/10.3354/
meps09857.
Hill, J.E., Robinson, N.J., King, C.M., Paladino, F.V., 2017. Diving behavior and thermal
habitats of gravid hawksbill turtles at St. Croix, USA. Mar. Biol. 164, 19. https://doi.
org/10.1007/s0022701630504.
Hooge, P.N., Eichenlaub, W., Hooge, E.R., 2001. Animal movement 2.5. Us Geological
Survey, Alaska Biological Science Center, Anchorage, AK, USA.
Horrocks, J.A., Vermeer, L.A., Krueger, B., Coyne, M., Schroeder, B.A., Balazs, G.H., 2001.
Migration routes and destination characteristics of post-nesting hawksbill turtles
satellite-tracked from Barbados, West Indies. Chelonian Conserv. Biol. 4, 107114.
Jonsen, I.D., Myers, R.A., Flemming, J.M., 2003. Meta-analysis of animal movement using
state-space models. Ecology 84, 30553063. https://doi.org/10.1890/020670.
Jonsen, I.D., Flemming, J.M., Myers, R.A., 2005. Robust state-space modeling of animal
movement data. Ecology 86, 28742880. https://doi.org/10.1890/041852.
Jonsen, I.D., Myers, R.A., James, M.C., 2007. Identifying leatherback turtle foraging be-
haviour from satellite telemetry using a switching state-space model. Mar. Ecol. Prog.
Ser. 337, 255264. https://doi.org/10.3354/meps337255.
Kamel, S.J., Delcroix, E., 2009. Nesting ecology of the hawksbill turtle, Eretmochelys im-
bricata, in Guadeloupe, French West Indies from 200007. J. Herpetol. 43, 367376.
https://doi.org/10.1670/07231R2.1.
Laver, P.N., Kelly, M.J., 2008. A critical review of home range studies. J. Wildl. Manag.
72, 290298. https://doi.org/10.2193/2005589.
Lee, P.L.M., Hays, G.C., 2004. Polyandry in a marine turtle: females make the best of a
bad job. Proc. Natl. Acad. Sci. U. S. A. 101, 65306535. https://doi.org/10.1073/
pnas.0307982101.
Luschi, P., Hays, G.C., Del Seppia, C., Marsh, R., Papi, F., 1998. The navigational feats of
green sea turtles migrating from Ascension Island investigated by satellite telemetry.
Proc. Biol. Sci. 265, 22792284.
Marcovaldi, M.A., Lopez, G.G., Soares, L.S., Lopez-Mendilaharsu, M., 2012. Satellite
tracking of hawksbill turtles Eretmochelys imbricata nesting in northern Bahia, Brazil:
turtle movements and foraging destinations. Endanger. Species Res. 17, 123132.
https://doi.org/10.3354/esr00421.
Maxwell, S.M., Breed, G.A., Nickel, B.A., Makanga-Bahouna, J., Pemo-Makaya, E.,
Parnell, R.J., Formia, A., Ngouessono, S., Godley, B.J., Costa, D.P., Witt, M.J., Coyne,
M.S., 2011. Using satellite tracking to optimize protection of long-lived marine spe-
cies: olive ridley sea turtle conservation in Central Africa. PLoS One 6, e19905.
McClenachan, L., Jackson, J.B.C., Newman, M.J.H., 2006. Conservation implications of
historic sea turtle nesting beach loss. Front. Ecol. Environ. 4, 290296. https://doi.
org/10.1890/15409295(2006)4[290:CIOHST]2.0.CO;2.
Meylan, A.B., 1999. International movements of immature and adult hawksbill turtles
(Eretmochelys imbricata) in the Caribbean region. Chelonian Conserv. Biol. 3,
189194.
Moncada, F.G., Hawkes, L.A., Fish, M.R., Godley, B.J., Manolis, S.C., Medina, Y., Nodarse,
G., Webb, G.J.W., 2012. Patterns of dispersal of hawksbill turtles from the Cuban
shelf inform scale of conservation and management. Biol. Conserv. 148, 191199.
https://doi.org/10.1016/j.biocon.2012.01.011.
Mortimer, J.A., Donnelly, M., 2008. Eretmochelys imbricata [WWW Document]. IUCN Red
K.M. Hart et al.
Biological Conservation 229 (2019) 113
12
List Threat. Species 2008. URL. https://doi.org/10.2305/IUCN.UK.2008.RLTS.
T8005A12881238.en, Accessed date: 30 June 2016.
NMFS, USFWS, 1993. Recovery Plan for Hawksbill Turtles in the U.S. Caribbean Sea,
Atlantic Ocean, and Gulf of Mexico. St. Petersburg, FL.
NMFS, USFWS, 1998. Recovery Plan for U.S. Pacific Populations of the Hawksbill Turtle
(Eretmochelys imbricata). Silver Spring, MD.
NMFS-SEFSC, 2008. Sea turtle research techniques manual. In: NOAA Technical
Memorandum NMFS-SEFSC579. Miami, FL, USA.
Parker, D.M., Balazs, G.H., King, C.S., Katahira, L., Gilmartin, W., 2009. Short-range
movements of hawksbill turtles (Eretmochelys imbricata) from nesting to foraging
areas with the Hawaiian Islands. Pac. Sci. 63, 371382.
Patterson, T.A., Thomas, L., Wilcox, C., Ovaskainen, O., Matthiopoulos, J., 2008. State-
space models of individual animal movement. Trends Ecol. Evol. 23, 8794. https://
doi.org/10.1016/j.tree.2007.10.009.
Proclamation 7392, 2001. Boundary Enlargement and Modifications of the Buck Island
Reef National Monument, 66 Fed. Reg. 7333 (January 22, 2001). Federal Register:
The Daily Journal of the United States. Web. 22 January 2001.53.
R Development Core Team, 2014. R: A Language and Environment for Statistical
Computing. R Foundation for Statistical Computing, Vienna, Austria.
Revuelta, O., Hawkes, L., Len, Y.M., Godley, B.J., Raga, J.A., Toms, J., 2015. Evaluating
the importance of marine protected areas for the conservation of hawksbill turtles
Eretmochelys imbricata nesting in the Dominican Republic. Endanger. Species Res. 27,
169180. https://doi.org/10.3354/esr00660.
Richardson, J.I., Hall, D.B., Mason, P.A., Andrews, K.M., Bjorkland, R., Cai, Y., Bell, R.,
2006. Eighteen years of saturation tagging data reveal a significant increase in
nesting hawksbill sea turtles (Eretmochelys imbricata) on Long Island, Antigua. Anim.
Conserv. 9, 302307. https://doi.org/10.1111/j.14691795.2006.00036.x.
Rodgers, A.R., Kie, J.G., 2011. HRT: Home Range Tools for ArGIS-draft August 10, 2011.
Centre for Northern Forest Ecosystem Research, Ontario (32 p).
Sartain-Iverson, A.R., Hart, K., Fujisaki, I., Cherkiss, M.S., Pollock, C., Lundgren, I., Hillis-
Starr, Z.-M., 2016. Hawksbill satellite-tracking case study: implications for remigra-
tion interval and population estimates. Mar. Turt. Newsl. 148, 27.
Schofield, G., Lilley, M.K.S., Bishop, C.M., Brown, P., Katselidis, K.A., Dimopoulos, P.,
Pantis, J.D., Hays, G.C., 2010. Conservation hotspots: implications of intense spatial
area use by breeding male and female loggerheads at the Mediterranean's largest
rookery. Endanger. Species Res. 10, 191202. https://doi.org/10.3354/esr00137.
Scott, R., Hodgson, D.J., Witt, M.J., Coyne, M.S., Adnyana, W., Blumenthal, J.M.,
Broderick, A.C., Canbolat, A.F., Catry, P., Ciccione, S., Delcroix, E., Hitipeuw, C.,
Luschi, P., Pet-Soede, L., Pendoley, K., Richardson, P.B., Rees, A.F., Godley, B.J.,
2012. Global analysis of satellite tracking data shows that adult green turtles are
significantly aggregated in Marine Protected Areas. Glob. Ecol. Biogeogr. 21,
10531061. https://doi.org/10.1111/j.14668238.2011.00757.x.
Seaman, D.E., Powell, R.A., 1996. An evaluation of the accuracy of kernel density esti-
mators for home range analysis. Ecology 77, 20752085. https://doi.org/10.2307/
2265701.
Shaver, D.J., Hart, K.M., Fujisaki, I., Rubio, C., Sartain, A.R., Pea, J., Burchfield, P.M.,
Gamez, D.G., Ortiz, J., 2013. Foraging area fidelity for Kemp's ridleys in the Gulf of
Mexico. Ecol. Evol. 3, 20022012. https://doi.org/10.1002/ece3.594.
Shaver, D.J., Hart, K.M., Fujisaki, I., Rubio, C., Sartain-Iverson, A.R., Pea, J., Gamez,
D.G., de Jesus Gonzales Diaz Miron, R., Burchfield, P.M., Martinez, H.J., Ortiz, J.,
2016. Migratory corridors of adult female Kemp's ridley turtles in the Gulf of Mexico.
Biol. Conserv. 194, 158167. https://doi.org/10.1016/j.biocon.2015.12.014.
Shillinger, G.L., Swithenbank, A.M., Bograd, S.J., Bailey, H., Castelton, M.R., Wallace,
B.P., Spotila, J.R., Paladino, F.V., Piedra, R., Block, B.A., 2010. Identification of high-
use internesting habitats for eastern Pacific leatherback turtles: role of the environ-
ment and implications for conservation. Endanger. Species Res. 10, 215232. https://
doi.org/10.3354/esr00251.
Starbird, C.H., Hillis-Starr, Z., Harvey, J.T., Eckert, S., 1999. Internesting movements and
behavior of hawksbill sea turtles (Eretmochelys imbricata) around Buck Island Reef
National Monument, St. Croix, U.S. Virgin Islands. Chelonian Conserv. Biol. 3,
237243.
Thomson, J.A., Borger, L., Christianen, M.J.A., Esteban, N., Laloe, J.-O., Hays, G.C., 2017.
Implications of location accuracy and data volume for home range estimation and
fine-scale movement analysis: comparing Argos and Fastloc-GPS tracking data. Mar.
Biol. 164, 204. https://doi.org/10.1007/s00227-017-3225-7.
Trong, S., Dutton, P.H., Evans, D., 2005. Migration of hawksbill turtles Eretmochelys
imbricata from Tortuguero, Costa Rica. Ecography (Cop.) 28, 394402. https://doi.
org/10.1111/j.09067590.2005.04110.x.
Van Dam, R.P., Diez, C.E., Balazs, G.H., Coln, L.A., McMillan, W.O., Schroeder, B., 2008.
Sex-specific migration patterns of hawksbill turtles breeding at Mona Island, Puerto
Rico. Endanger. Species Res. 4, 8594. https://doi.org/10.3354/esr00044.
Walcott, J., Eckert, S., Horrocks, J.A., 2012. Tracking hawksbill sea turtles (Eretmochelys
imbricata) during inter-nesting intervals around Barbados. Mar. Biol. 159, 927938.
https://doi.org/10.1007/s0022701118709.
White, G., Garrott, R., 1990. Analysis of Wildlife Radiotracking Data. Academic Press,
New York, New York, USA.
Worton, B.J., 1987. A review of models of home range for animal movement. Ecol. Model.
38, 277298.
Worton, B.J., 1989. Kernel methods for estimating the utilization distribution in home-
range studies. Ecology 70, 164168.
Worton, B.J., 1995. Using Monte Carlo simulation to evaluate kernel-based home range
estimators. J. Wildl. Manag. 59, 794800.
K.M. Hart et al.
Biological Conservation 229 (2019) 113
13