Regional Studies in Marine Science 48 (2021) 102011
m
D
K
a
b
c
a
1
(
s
r
v
p
δ
a
i
f
Contents lists available at ScienceDirect
Regional Studies inMarine Science
journal homepage: www.elsevier.com/locate/rsma
Stable isotopes used to infer trophic position of green turtles (Chelonia
ydas) fromDry Tortugas National Park, Gulf ofMexico, United States
avid C. Roche a,∗, Michael S. Cherkiss a, Brian J. Smith b, Derek A. Burkholder c,
risten M. Hart a
Wetland and Aquatic Research Center, U.S. Geological Survey, Davie, FL 33314, USA
Department of Wildland Resources and Ecology Center, Utah State University, Logan, UT 84322, USA
Halmos College of Natural Sciences and Oceanography, Nova Southeastern University, Dania Beach, FL 33004, USA
a r t i c l e
i n f o
Article history:
Received 28 May 2021
Received in revised form 20 August 2021
Accepted 15 September 2021
Available online 20 September 2021
Keywords:
Sea turtle
Carbon
Nitrogen
Stable isotopes
LMM
a b s t r a c t
Evaluating resource use patterns for imperiled species is critical for understanding what supports their
populations. Here we established stable isotope (δ13C, δ15N) values for the endangered green sea turtle
(Chelonia mydas) population found within the boundaries of Dry Tortugas National Park (DRTO), south
Florida, USA. There is little gene flow between turtles sampled at DRTO and in other rookeries in
Florida, underscoring the need to study this distinct population. Between 2008 and 2015 we collected
multiple sample types (skin [homogenized epidermis/dermis], whole blood, red blood cells, plasma,
carapace) from 151 unique green turtles, including 43 nesting females and 108 in-water captures; some
individuals were resampled multiple times across years to evaluate consistency of isotope signatures.
Isotopic ratios ranged from -27.3 to -5.4 for δ13C and 3.7 to 10.6 for δ15N. Using linear mixed models,
we evaluated covariates (sample type, turtle size and year) that best explained the isotope patterns
observed in turtle tissues. Predictions from the top model for δ13C indicated a slight decrease over
time and for δ15N a slight increase in the middle sampling years (2010–2012); results indicated that
turtle size appeared to be the driver behind the range in δ13C and δ15N observed in turtle skin. We
found a pattern in stable carbon isotope values that are indicative of an ontogenetic change from an
omnivorous diet in smaller turtles to a seagrass-based diet in larger turtles. When we compared the
stable carbon and nitrogen isotope values of the samples collected from turtles with that of seagrasses
found in DRTO, we found that turtles > 65 cm SCL had similar stable carbon isotope values to the
seagrass species present. Results of this study suggest stable isotope analysis coupled with data for
available resources can be useful for tracking and detecting future changes in green turtle resource
shifts in DRTO.
© 2021 Published by Elsevier B.V.

1. Introduction
Stable isotope analysis is a common tool used to assess various
spects of animal ecology (Tieszen et al., 1983; Godley et al.,
998; Wiley et al., 2019). The ratios of carbon (δ13C) and nitrogen
δ15N) are often used to assess dietary interactions and trophic
tructure (Post, 2002; Layman et al., 2012). Generally, δ13C values
eflect the primary producer at the base of the food web (marine
s terrestrial, planktonic, detrital, etc.), since different primary
roducers may utilize different photosynthetic pathways, but
13C values are generally unchanged through a food chain (DeNiro
nd Epstein, 1978; Peterson and Fry, 1987; Post, 2002). Nitrogen
sotopic ratios exhibit a trophic enrichment with each step in the
ood chain. The lighter isotope, 14N, is preferentially excreted at
∗ Corresponding author.
E-mail address: droche@usgs.gov (D.C. Roche).
https://doi.org/10.1016/j.rsma.2021.102011
2352-4855/© 2021 Published by Elsevier B.V.
each trophic level. This action results in 15N enrichment (∼2–4‰)
in the bodies of the consumers compared to their diet (DeNiro
and Epstein, 1981), providing the relative trophic position at
which the organism is feeding.
Tissue turnover rates (i.e., the time it takes for 50% of the stable
isotopes in a tissue to be replaced by the stable isotopes in the
diet) differ by tissue type and metabolic pathways (Fry, 2006).
For example, plasma, blood, muscle, and bone reflect different
time periods of resource acquisition. In sea turtles, plasma repre-
sents the most recent snapshot of the green sea turtle’s dietary
information. The soft tissue from the rear flipper has a slower
turnover rate than the plasma and allows an ‘older’ view of the
green sea turtle dietary information. The isotopic incorporation
rate for epidermis in another rapidly growing ectotherm, juvenile
loggerheads, is approximately 4 months, faster than the incor-
poration rate of adults, which have slower growth rates (Reich
et al., 2008). By examining stable isotope values across different

D.C. Roche, M.S. Cherkiss, B.J. Smith et al.
Regional Studies in Marine Science 48 (2021) 102011
t
s
(
i
s
m
2
f
B
n
c
a
a
e
d
f
(
1
(
a
v
b
W
M
w
g
i
d
2
F
t
a
b
a
l
f
s
t
s
c
c
c
a
s
2
2
M
W
S
(
s
2
(
a
o
o
t
t
c
6
W
(
c
t
2
l
t
c
d
c
i
a
h
d
s
t
2
t
a
a
d
w
b
g
2
p
3
s
E
a
a
(
F
n
1
i
o
s
t
issue types, it is possible to examine variation in individual diet
pecialization through repeat sampling of the same individual
Seminoff et al., 2006; Burgett et al., 2018). In addition, stable
sotopes have been used as intrinsic markers and applied to
tudying migratory pathways in sea turtles and a host of other
igratory animals (Hobson et al., 1997; Vander Zanden et al.,
015; Ceriani et al., 2012; Bird et al., 2018; Haywood et al., 2019)
Chelonia mydas (Linnaeus 1758) are primarily herbivores that
eed on seagrasses and algae (Randall, 1965; Mortimer, 1981;
jorndal, 2017); however, previous studies suggest that as juve-
iles, they are omnivorous and feed on tunicates, jellyfish and
tenophores, crustaceans, and mollusks and continue omnivory
s adults (Hiethaus et al. 2002, Hatase et al., 2006; Amorocho
nd Reina, 2007; Cardona et al., 2009; Parker et al., 2011). For
xample, green turtles in Moreton Bay, Australia, exhibit a more
iverse diet of primarily algae and seagrass while occasionally
eeding on mangrove leaves and propagules, as well as jellyfish
Brand-Gardner et al., 1999; Limpus and Coffee, 2019).
Commercial exploitation of green sea turtles prior to the
900’s depleted Atlantic populations by up to 90 percent
Van Houtan and Pimm, 2007). Currently, all species of sea turtles
re under pressure from various threats including urban de-
elopment/habitat degradation, light pollution, plastic pollution,
ycatch in fisheries, and direct harvest (Lutcavage et al., 1997;
itherington, 1992; Seminoff et al., 2015). Located in the Gulf of
exico, Dry Tortugas National Park (DRTO) is a protected area
here important developmental habitat, foraging and nesting
rounds for green turtles exist (Bryan and Ault, 2018). Turtles
n the mixed-species foraging aggregation at DRTO were highly
ifferentiated frommost other Atlantic groups (Naro-Maciel et al.,
017). In 2015, a status review conducted by National Marine
isheries Service (NMFS) of green sea turtles reviewed each dis-
inct population segment (DPS) for its risk of extinction. Florida
nd the Caribbean were included in the North Atlantic DPS and
ecause of factors including abundance, population growth rate,
nd conservation efforts, this DPS was ranked as being a relatively
ow risk for extinction (Seminoff et al., 2015), thus its downlisting
rom ‘‘Endangered’’ to ‘‘Threatened’’ in 2016. As such, this study
ite is of continued importance as an area for recovering this
hreatened species
In this study, our goal was to establish the carbon and nitrogen
table isotope values of the DRTO green turtle population and
ompare them across sample types (whole blood [WB], red blood
ell [RBC], plasma, skin [homogenized epidermis/dermis], and
arapace). We sought to assess whether the population follows
traditional ontogenetic change from an omnivorous diet in
maller turtles to a seagrass-based diet in larger turtles.
. Materials & methods
.1. Study site and sampling
Dry Tortugas National Park located in the southeast Gulf of
exico (24.628141◦N, −82.873070◦W), 70 miles west of Key
est, FL, features extensive seagrass beds of Thalassia testudinum,
yringodium filiforme, Halodule wrightii,
and Halophila
sp.
Fourqurean et al., 2010) that cover between 15% and 30% of the
eafloor in the park (Davis, 1979; Waara et al., 2011).
Sampling trips occurred between May −August annually from
008−2015. Capture methods included hand capture via boat
e.g., dip netting, turtle-jumping, see Ehrhart and Ogren, 1999)
nd interception on the beach after nesting or non-nesting events
n East and Loggerhead Keys in DRTO (Fig. 1). We held each turtle
nboard the vessel for hand captures and corralled each female
urtle encountered on the beach to confine them for workup. Each
urtle was individually marked by inserting a passive integrated
t
2
transponder (PIT) tag in the right flipper and affixing individually
numbered flipper tags to each trailing-edge front flipper. We
collected standard morphometric data including carapace mea-
surements (curved (CCL) and straight (SCL) carapace lengths and
widths) according to methods used by NMFS (2008) on sea turtle
research techniques. We opportunistically re-sampled individ-
ual turtles recaptured after their initial tagging event; recapture
events ranged from 1 to 3 years post initial capture. We parsed
data by size classes as defined by Bresette et al. (2010) for green
turtles as juvenile <65 cm SCL, subadults 65−90 cm SCL, and
adults >90 cm SCL. Additional justification for this size division
omes from two previous studies that documented diet shifts at
2 and 65 cm CCL SCL (Arthur et al., 2008; Cardona et al., 2010).
e determined sex following methods described by Fujisaki et al.
2016), i.e., we externally assessed tail length of each animal and
ategorized them as male, female, or unknown, with males having
ails with cloaca-tip lengths of ≥5.5 cm.
.2. Sample collection and preparation
We collected multiple sample types from each individual fol-
owing established protocols (NMFS, 2008), i.e., skin samples from
he soft portion of the inside trailing edge of a rear flipper and
arapace from the third lateral scute on the right side (Van-
er Zanden et al., 2010). We collected blood from the dorsal
ervical sinus (Owens and Ruiz, 1980) and placed 2 ml samples
nto individually labeled plastic Corning Cryovials. We selected
subset of turtles for 8 ml blood draws collected in a non-
eparinized vacutainer and we immediately spun those samples
own in a centrifuge to separate plasma and RBC. We stored all
amples on ice/cooler packs in the field and then transferred them
o a −20 ◦C freezer for storage until later sample processing.
.3. Laboratory procedure
In the lab, we thawed, rinsed with distilled water, and dried
he skin samples at ∼60 ◦C for up to, but no more than 48 h
nd then pulverized each one to a fine powder using a mortar
nd pestle. We rinsed carapace samples with distilled water,
ried at ∼60 ◦C for up to 48 h, cut each one into smaller pieces
ith scissors, and then ground them to a fine powder. Whole
lood, RBC, and plasma samples were thawed, poured out over
lassware to expedite drying, and dried at ∼60 ◦C for at least
4 h but no more than 48 h, scraped off the glassware, and then
ulverized with a mortar and pestle to a fine powder.
We weighed tissue samples (between 0.60 to 0.70 mg) into
.3 × 5.0 mm individual tin boats and sent them for analysis of
table-carbon and stable-nitrogen isotope ratios at the Southeast
nvironmental Research Center (SERC) Stable Isotope Laboratory
t Florida International University (FIU). The Stable Isotope Lab
t FIU uses a continuous flow isotope ratio mass-spectrometer
IRMS) machine coupled to elemental analyzers, specifically, a
innigan Delta C EA-IRMS. Results are expressed in standardized
otation of δ13C and δ15N (DeNiro and Epstein, 1978, 1981) as
follows:
δ
heavy
light
X =
(
heavy X
light X
)
sample
(
heavy X
light X
)
standard
− 1
heavyX/lightX are the ratios of heavy to light isotopes (13C:12C,
5N:14N) in the sample and standard, respectively. Carbon stable
sotope ratios are reported relative to the international standards
f Pee Dee Belemnite (PDB) or the equivalent Vienna PDB (VPDB)
tandard. Nitrogen stable isotope ratios are reported relative to
he standards of atmospheric nitrogen (AIR). Standard error for
his study was based on internal glycine standards of ± 0.18‰

D.C. Roche, M.S. Cherkiss, B.J. Smith et al.
Regional Studies in Marine Science 48 (2021) 102011
c
f
6
I
δ
f
P
e
i
s
t
t
2
i
F
2
w
t
v
t
c
=
t
a
t
a
s
s
i
w
g
i
i
(
Fig. 1. Green turtle (Chelonia mydas) capture locations at Dry Tortugas National Park with mainland Florida, USA in the inset. Turtles < 65 SCL cm were primarily
aptured at Garden Key. Turtles > 65 SCL cm were caught at East Key (nesting females) and Pulaski Shoal (in-water captures).

or δ15N and ± 0.10‰ for δ13C. Internal standards were run every
to 8 experimental samples to ensure proper system calibration.
n lieu of lipid extraction, we used the equation (δ13Cnormalized =
13Cuntreated − 3.32 + 0.99 x C:N) developed by Post et al. (2007)
or samples with > 3.5 C to N ratio, as in Hall et al. (2015).
revious studies by Burkholder et al. (2011), Vander Zanden
t al. (2012, 2014) did not find significant differences of stable
sotope ratios in lipid and non-lipid extracted tissues. Recent
tudies (Haywood et al., 2020) have foregone lipid extraction if
he samples’ C:N ratio fell within an acceptable range.
We used isotope values from seagrass samples collected by
he FIU Seagrass Ecosystems Research Lab (SERL) during 2011–
016. Isotopic analysis of seagrass samples was also performed
n the same lab using the same methods SERL-FIU (Anderson and
ourqurean, 2003; Campbell and Fourqurean, 2009).
.4. Data analysis
To quantify the variation in turtle tissue stable isotope ratios,
e used linear mixed models (LMMs) to identify the covariates
hat best explained the patterns in turtle tissues. We selected four
ariables that we hypothesized were likely to have influenced the
urtle isotopic ratios: sample type (WB, RBC, plasma, skin and
arapace), turtle size (SCL cm), sex (male, female, or unknown
juvenile) and the year of sampling (2008–2015). We included
hese variables due to different turnover rates between tissues
nd metabolic processes, ontogenetic diet and habitat shifts for
he species, behavioral differences between the sexes/age classes,
nd variable baselines across study years. Because we sampled
everal individual turtles multiple times over the course of the
tudy, we accounted for autocorrelation by including turtle ID
n the model as a random effect. Using R (R Core Team, 2019),
e built independent model sets for both δ13C and δ15N, be-
inning each model set with a null model containing only an
ntercept and the random effect. Then we fit a full model, which
ncluded sample type (categorical), turtle size (continuous), sex
categorical), and sampling year (continuous), along with the
3
random effect. The full model included second-order polynomials
for both continuous predictors (i.e., a quadratic and linear term),
allowing us to potentially estimate a parabolic relationship. The
full model also included an interaction between sex and size.
We created a candidate suite of models with different combina-
tions of predictor variables using backward selection (Kéry and
Royle, 2015). We used sample size corrected Akaike informa-
tion criterion (AICc) for model selection, whereupon we selected
the model with the lowest AICc score as the top model. We
fit the LMMs using the function ‘lmer()’ from the R package
‘lme4’ (Bates et al., 2015), and we used the function ‘model.sel()’
from the package ‘MuMIn’ to perform model selection (Bartón,
2020). We assessed goodness-of-fit for the top model in each
set by calculating pseudo-R2 using the method of Nakagawa and
Schielzeth (2013). We used the function ‘r.squaredGLMM()’ from
the package ‘MuMIn’ for this calculation, and we reported both
the marginal (fixed effects) and conditional (fixed and random
effects) R2 [R2GLMM(m) and R2GLMM(c), respectively] .
We used the top model from the model set for each isotope
to draw inference about variation in that isotope. To understand
the patterns the model predicted, we examined each predic-
tor separately. To do this, we held the other predictors in the
model constant, while varying the predictor at-hand through its
observed range. Then we used bootstrapping to generate 95%
confidence intervals for the model predictions. We used the func-
tion ‘predict()’ and the function ‘bootMer()’ from the package
‘lme4’ to generate model predictions and perform the bootstrap-
ping, respectively. Finally, we plotted the predictions to visualize
the modeled relationship between the isotope ratios and the
predictor variables.
To understand how seagrasses were contributing to the turtle
diet, we created a biplot with both turtle isotope samples and
seagrass samples (Table 5). Mixing models are unreliable in sit-
uations where an isotopic endmember is missing (Brett, 2014;
Phillips et al., 2014); isotopic endmembers are the resources that
completely bound the consumer’s isotopic niche space, which
contribute to mixing model geometry to determine source con-
tribution to consumers. Visual examination of this biplot revealed

D.C. Roche, M.S. Cherkiss, B.J. Smith et al.
Regional Studies in Marine Science 48 (2021) 102011
t
s
i
f
3
i
2
9
i
i
7
c
8
s
s
t
p
g
a
s
f
t
i
w
b
m
p
t
T
i
t
c
w
R
a
i
d
s
f
c
a
h
<
s
t
−
S
f
s
c
a
4
t
g
u
hat endmembers were clearly missing, i.e., we did not have re-
ources completely bounding the turtle samples, therefore we felt
t was unwarranted to try to fit mixing models and discrimination
actors. Thus, we simply evaluated this plot visually.
. Results
We captured and sampled 151 unique green turtles, includ-
ng 43 nesting females and 108 in-water turtles (Tables 1 and
). Nesting turtles ranged in size from 86.6–110 (mean ± SD,
8.4 ± 5.3) cm straight carapace length (SCL), and turtles capture
n the water ranged from 22.3 to 111.7 (55.3 ± 26.6) cm SCL,
nclusive of both immature and mature turtles. We captured
8 female turtles and 18 male turtles. From those turtles, we
ollected 159 skin samples (including 8 resampled juveniles, and
resampled nesters), 24 whole blood samples, 19 red blood cell
amples, 33 plasma samples (1 resampled turtle), and 61 scute
amples (Table 1, Fig. 2). The ranges of δ13C and δ15N values of
hese sample types varied (Table 1) but were consistently within
revious values published in the literature for green turtles in the
reater Caribbean.
The top model in both the δ13C and δ15N model sets contained
ll the variables of interest – sample type, turtle size, sex, and
ampling year – which in both cases included the quadratic term
or turtle size but only the linear term for year. For both isotopes,
he full model was a close competitor (∆AICc < 2; Tables 3 and 4),
ndicating mixed evidence for the quadratic term for year. Since
e included year to control for varying baselines, but not for
iological inference per se, we drew inference from the simpler
odel in each set. In the carbon model set, the full model out-
erformed the third model by 24.4 ∆AICc and cumulatively with
he top model received > 99.9% of the model weight (Table 3).
he top model had R2GLMM(m) = 0.757 and R2GLMM(c) = 0.837,
ndicating a strong goodness-of-fit. In the nitrogen model set,
he full model outperformed the third model by 5.9 ∆AICc and
umulatively with the top model received 96.7% of the model
eight (Table 4). The top model had R2GLMM(m) = 0.417 and
2
GLMM(c) = 0.800, indicating a moderate goodness-of-fit, but
high proportion of variance explained by the random effect,
.e., individual variation.
Predictions from the top model for carbon showed a slight
ecrease over time in δ13C; a complex relationship between SCL,
ex, and δ13C; and variation among sample types. The top model
or nitrogen showed a slight increase in δ15N with sampling year,
omplex relationship between SCL, sex, and δ15N, and variation
mong sample types (Fig. 3).
The δ13C − δ15N biplot showed that turtles > 65 cm SCL
ad similar δ13C values to the seagrass species (Table 5). Turtles
65 cm SCL were clearly also incorporating a different carbon
ource (Fig. 4) and exhibited a larger carbon range (21.88) than
urtles with SCLs > 65 cm (15.73).
For resampled individuals, δ13C values ranged from −7.42–
15.20; δ15N values ranged from 6.14–10.53 for turtles <65 cm
CL. δ13C values ranged from −6.60–−11.79; δ15N values ranged
rom 6.03–9.53, for turtles > 65 cm SCL. Turtles > 65 cm SCL
howed less variability in their isotopic signatures indicating
onsistency in their diet than smaller turtles which would suggest
more variable diet (Fig. 5).
. Discussion
This work presents the first known isotopic values for green
urtles in DRTO, establishing the baseline stable carbon and nitro-
en isotope values of this population. Furthermore, it adds to the
nderstanding of foraging patterns as we also detected omnivory
4
in smaller size turtles, with a shift towards more specialized her-
bivorous resource use in larger turtles, for this recovering species
in a marine protected area. We found differences in isotopic ratios
among several sample types (blood, plasma, whole blood, tissue,
scute), trends in isotopic ratios by size, and temporal consistency
of green turtle stable isotope values over a span of years.
4.1. Isotope comparisons
Dry Tortugas green turtle skin tissue δ13C and δ15N values from
this study were similar to those found from other Caribbean green
turtle populations. For example, Vander Zanden et al. (2013)
found that skin δ13C values of juvenile green turtles from two
sites in the Bahamas and the northwest Gulf of Mexico (St. Joe
Bay, FL) ranged from −12.2‰ to −4.5‰ and −15.7‰ to −9.0‰,
respectively. Burgett et al. (2018) sampled juvenile green sea
turtles in Bermuda and found δ15N values ranging from 2.4‰ to
12.6‰. Here we found similar values of juvenile green turtle skin
tissue δ13C and δ15N with similar ranges (Fig. 4). The observed
range of Dry Tortugas green turtle skin tissue δ15N values, 3.7‰
to 10.6‰, suggests that the aggregation may occupy more than
one trophic level or it could be a manifestation of different ni-
trogen sources (Ishikawa, 2018); this result could be explained
by the size ranges that were sampled or the variation in foraging
grounds at DRTO.
Considering larger green turtles, Vander Zanden et al. (2013)
sampled adults from two sites in Nicaragua and a nesting beach
in Tortuguero; turtles at these sites were found to have skin
tissue δ13C values of −14.7‰ to −7.3‰ and −17.0‰ to −5.3‰,
respectively. The authors attributed their wide range of values
observed to differences in the biogeochemistry of foraging areas
rather than differences in trophic position. Hart et al. (2013)
examined the benthic habitat at Pulaski Shoal and found seagrass
habitat with greater than 75% coverage at 21.9% of the overall
study site increasing to 42% in the ‘hotspot’ where multiple turtle
activity centers overlapped which is where the majority of the
in-water capture efforts for larger green turtles in this study
occurred. This provided evidence that adult green turtles are
seeking out areas with higher seagrass density for their foraging
sites and that would likely be reflected in their isotopic signa-
tures. In contrast, turtles < 65 cm SCL had enriched δ15N values
of skin compared to turtles larger than > 65 cm SCL, (x̄ = 8.2‰
± 1.1‰ vs. 7.0‰ ± 1.0‰). These enriched δ15N values suggest
a diet shift from a more omnivorous diet as juveniles to a more
seagrass-dominated diet in larger animals (Fig. 4).
The shallow habitat adjacent to Garden Key in DRTO includes
benthic habitats of seagrass and macroalgae and is where we
catch the majority of the juvenile green turtles. Perhaps the
size class partitioning we witness in the Dry Tortugas is due to
predator avoidance. In Shark Bay, Australia, Heithaus et al. (2005)
found that juvenile green turtles used shallow water habitat as
refuge from tiger sharks (Galeocerdo cuvier). Ault et al. (2002)
noted a paucity of sharks around the Dry Tortugas. Anecdotally,
the authors have yet to see a tiger shark near the juvenile or adult
habitat, but 12 satellite tagged tiger sharks from the west coast of
Florida and the Florida Keys showed a higher number of tracking
days near the lower keys and Dry Tortugas (Hammerschlag et al.,
2012).
4.2. Repeat captures
Many of the adult turtles we sampled are resident in DRTO,
indicated by capture–recapture records since 2009 (Hart, unpubl.
data, Roche et al., 2019) and supported by the δ15N values. The
decreasing δ15N values indicating the turtles are feeding at a
lower trophic level as turtle size increases is consistent with

D.C. Roche, M.S. Cherkiss, B.J. Smith et al.
Regional Studies in Marine Science 48 (2021) 102011
Table 1
Summary of Dry Tortugas National Park Green turtles (Chelonia mydas) by sample types collected by size (SCL),
capture method, along with ranges, mean, and standard deviation of δ13C and δ15N stable isotope values for collected
sample types.
n
δ13C (h)
δ15N (h)
Range
Mean ± SD
Range
Mean ± SD
In water
<65 SCL cm
Skin
60
−16.57–−7.05
−11.25 ± 2.55
6.14–10.61
8.24 ± 1.05
Whole Blood
12
−22.43–−8.94
−13.86 ± 3.75
5.2–9.07
7.03 ± 1.33
Red Blood Cell
14
−17.72–−8.22
−12.29 ± 3.25
4.62–9.14
6.51 ± 1.29
Plasma
17
−17.02–−6.86
−9.94 ± 2.6
5.21–10.05
6.81 ± 1.29
Scute
22
−9.93–−18.15
−14.32 ± 2.69
6.21–9.29
7.7 ± 0.97
>65 SCL cm
Skin
44
−10.58–−6.21
−7.86 ± 0.95
3.7–9.36
6.92 ± 1.22
Whole Blood
5
−14.93–−7.49
−10.55 ± 2.86
4.06–7
5.84 ± 1.07
Red Blood Cell
2
−8.41–7.44
−7.93 ± 0.49
4.22–6.44
5.33 ± 1.11
Plasma
4
−7.72–6.6
−7.39 ± 0.46
3.45–6.98
5.80 ± 1.39
Scute
12
−14.85–−6.4
−9.39 ± 2.45
4.77–9.69
6.8 ± 1.31
Nester
Skin
54
−13.08–−6.16
−7.94 ± 1.32
5.82–9.53
7.14 ± 0.81
Whole Blood
7
−9.58–−7.24
−10.97 ± 6.71
3.43–5.89
4.49 ± 0.82
Red Blood Cell
3
−8.85–−7.38
−8.26 ± 0.64
3.82–6.24
4.96 ± 0.99
Plasma
12
−9.18–−6.35
−7.53 ± 1.05
3.64–6.78
5.62 ± 1.03
Scute
27
−12.82–7.18
−8.78 ± 1.22
4.37–7.64
5.92 ± 0.78

Table 2
Summary of Dry Tortugas National Park Green turtles (Chelonia mydas) skin stable isotope values by year and size class. Size classes
straight carapace lengths: Juvenile: <65 cm; Sub-adult: 65–90 cm; Adult: >90 cm. SD: Standard deviation.
YEAR
Size class
n
Skin stable isotope values
δ13C (h)
δ15N (h)
Range
Mean ± SD
Range
Mean ± SD (h)
2008
Juvenile
16
−14.57 to −7.05
−9.97 ± 2.02
6.14 to 9.84
7.87 ± 0.88
Sub-adult
N/A
N/A
N/A
N/A
N/A
Adult
N/A
N/A
N/A
N/A
N/A
2009
Juvenile
4
−12.90 to −8.79
−10.89 ± 1.73
6.98 to 10.61
8.71 ± 1.53
Sub-adult
7
−9.03 to −6.74
−7.74 ± 1.01
5.29 to 8.04
6.58 ± 1.01
Adult
6
−9.04 to −6.20
−7.29 ± 1.04
6.80 to 8.02
7.62 ± 0.46
2010
Juvenile
8
−10.76 to −7.42
−8.93 ± 1.20
6.49 to 9.63
7.68 ± 1.17
Sub-adult
3
−9.03 to −6.74
−7.74 ± 1.01
5.29 to 8.04
6.58 ± 1.01
Adult
5
−8.09 to −7.41
−7.72 ± 0.25
6.66 to 9.30
7.95 ± 0.99
2011
Juvenile
9
−14.55 to −7.83
−11.77 ± 2.29
7.27 to 10.53
8.63 ± 1.00
Sub-adult
5
−10.58 to −7.36
−8.61 ± 1.10
5.51 to 8.18
7.20 ± 0.93
Adult
17
−10.25 to −6.37
−8.38 ± 1.08
5.82 to 9.36
7.37 ± 0.99
2012
Juvenile
3
−16.57 to −9.52
−13.09 ± 3.53
8.42 to 10.32
9.35 ± 0.95
Sub-adult
3
−7.53
N/A
7.39
N/A
Adult
5
−13.08 to −7.26
−8.66 ± 1.98
5.62 to 7.85
6.80 ± 0.75
2013
Juvenile
3
−15.27 to −8.28
−13.09 ± 3.53
6.88 to 9.81
7.92 ± 1.64
Sub-adult
N/A
N/A
N/A
N/A
N/A
Adult
11
−9.39 to −6.68
−8.27 ± 0.90
3.70 to 9.53
7.03 ± 1.93
2014
Juvenile
3
−15.20 to −9.17
13.13 ± 3.43
7.53 to 9.15
8.35 ± 0.81
Sub-adult
1
−8.46
N/A
7.24
N/A
Adult
5
−8.99 to −5.38
−7.71 ± 1.48
4.88 to 8.28
6.79 ± 1.36
2015
Juvenile
15
−15.59 to −8.65
−13.00 ± 2.17
6.48 to 10.18
8.33 ± 0.98
Sub-adult
1
−7.58
N/A
5.68
N/A
Adult
30
−10.53 to −6.16
−7.57 ± 1.10
5.41 to 8.22
6.85 ± 0.66

Table 3
Model selection table of δ13C model set for Dry Tortugas National Park Green turtles (Chelonia mydas). We ranked
models using sample size corrected AIC (AICc). All models, including the null, included a random intercept for the
individual turtle. The full model contained the variables sample type (whole blood, red blood cells, plasma, skin and
carapace), an interaction between turtle size (continuous) and sex (male, female, or unknown = juvenile), and the
year of sampling (continuous). Both continuous variables were fitted with quadratic terms to allow for parabolic
relationships. The top two models differ only by the quadratic term for year, and together they receiver > 99.9%
of the model weight (ω).
Model
Parameters
AICc
∆AICc
ω
Tissue + Sex ∗ (SCL + SCL2) + Year
16
731.50
0.00
0.72
Tissue + Sex ∗ (SCL + SCL2) + Year + Year2 [Full model]
17
733.42
1.93
0.28
Tissue + Sex + (SCL + SCL2) + Year + Year2
13
755.88
24.38
0.00
Tissue + Sex ∗ SCL + Year
13
764.81
33.31
0.00
Null
3
949.46
217.96
0.00


5


D.C. Roche, M.S. Cherkiss, B.J. Smith et al.
Regional Studies in Marine Science 48 (2021) 102011
t
c
r
Fig. 2. Sampled tissues (whole blood [WB], red blood cell [RBC], plasma, skin [homogenized epidermis/dermis], and carapace) collected between 2008–2015 distributed
across size for Green turtles (Chelonia mydas) in Dry Tortugas National Park collected 2008–2015.

Table 4
Model selection table of δ15N model set for Dry Tortugas National Park Green turtles (Chelonia mydas). We ranked
models using sample size corrected AIC (AICc). All models, including the null, included a random intercept for the
individual turtle. The full model contained the variables sample type (whole blood, red blood cells, plasma, skin and
carapace), an interaction between turtle size (continuous) and sex (male, female, or unknown = juvenile), and the
year of sampling (continuous). Both continuous variables were fitted with quadratic terms to allow for parabolic
relationships. The top two models differ only by the quadratic term for year, and together they receiver > 96% of
the model weight (ω).
Model
Parameters
AICc
∆AICc
ω
Tissue + Sex ∗ (SCL + SCL2) + Year
16
546.86
0.00
0.53
Tissue + Sex ∗ (SCL + SCL2) + Year + Year2 [Full model]
17
547.27
0.40
0.43
Tissue + Sex ∗ SCL + Year + Year2
14
552.77
5.91
0.03
Tissue + Sex + (SCL + SCL2) + Year + Year2
13
556.11
9.25
0.01
Null
3
644.41
97.55
0.00

Table 5
Summary of seagrass samples.
Year
Species
n
Seagrass stable isotope values
δ13C (h)
δ15N (h)
Range
Mean ± SD
Range
Mean ± SD (h)
2011
H. decipiens
N/A
N/A
N/A
N/A
N/A
H. wrightii
N/A
N/A
N/A
N/A
N/A
S. filiforme
1
−4.96
N/A
0.83
N/A
T. testudinum
2
−7.57 to −7.07
−7.32 ± 0.35
2.54 to 3.02
2.78 ± 0.34
2012
H. decipiens
N/A
N/A
N/A
N/A
N/A
H. wrightii
N/A
N/A
N/A
N/A
N/A
S. filiforme
10
−7.49 to −4.74
−5.59 ± 0.92
0.48 to 3.12
2.38 ± 0.75
T. testudinum
14
−10.01 to −6.1
−7.73 ± 0.97
0.17 to 3.91
2.42 ± 1.00
2013
H. decipiens
N/A
N/A
N/A
N/A
N/A
H. wrightii
N/A
N/A
N/A
N/A
N/A
S. filiforme
6
−7.6 to −5.83
−6.62 ± 0.71
0.93 to 3.44
2.43 ± 0.95
T. testudinum
9
−9.36 to −6.29
−8.19 ± 0.94
1.41 to 4.58
2.63 ± 0.97
2015
H. decipiens
1
−4.78
N/A
3.05
N/A
H. wrightii
1
−3.5
N/A
3.14
N/A
S. filiforme
9
−10.58 to −4.51
−7.93 ± 1.91
0.46 to 4.37
2.30 ± 1.35
T. testudinum
12
−10.68 to −5.62
−7.38 ± 1.70
0.19 to 3.95
2.65 ± 1.24

past studies that showed the classic shift to herbivory as turtles
matured (Arthur et al., 2008; Cardona et al., 2010). In addition,
the seagrass δ13C and δ15N data supports that larger turtles shift
o herbivory, as the isotope values from larger sea turtles more
losely reflects the range of values in seagrass (Table 5).
Stable isotope values of recaptured juveniles became more en-
iched in δ13C over time, with an overall shift towards δ13C values
of larger turtles (Fig. 5A, B). The growth rate of juveniles heavily
influences the isotopic incorporation rates of C and N. Gastric
lavage data collected in 2008 from juveniles in the Dry Tortugas
indicated that all turtles had recently consumed seagrass, with T.
testudinum comprising the majority of the samples. Small jellyfish
6
(Cassiopea sp.) were found in one of the sampled juveniles (K.
Hart unpubl. data). Stomach content studies of Caribbean green
turtles reported that T. testudinum is the primary forage species
(Bjorndal, 1980; Mortimer, 1981). Based on resampling results of
skin, larger DRTO green turtles exhibited fidelity to their feeding
regimes, e.g., had lower variability in δ15N values. (Fig. 5C, D). This
sampled population could be made up of a generalist population
with generalist individuals, where individuals may vary widely
in their resource use or maintain consistent resource use within
a narrow isotopic niche space. However, variation among indi-
viduals results in a wide population isotopic niche, whereas with
specialist individuals both the individual and population isotopic

D.C. Roche, M.S. Cherkiss, B.J. Smith et al.
Regional Studies in Marine Science 48 (2021) 102011
t
c
t
a
Fig. 3. The predicted relationship between δ13C, δ15N, and the predictor variables (Year, Straight Carapace Length (SCL), Sex, and green turtle sample type) for Green
urtle (Chelonia mydas) in Dry Tortugas National Park according to the top model for each isotope. Dashed lines and error bars represent the 95% confidence intervals.
Samples types are as follows (whole blood [WB], red blood cell [RBC], plasma, skin [homogenized epidermis/dermis], and carapace).

niche widths are narrow (Vander Zanden et al., 2010). Future
opportunities for repeat sampling of resident turtles due to high
recapture rates at this long-term study site may help to tease
apart resource use or shifts with serially sampled individuals.
To date, studies comparing aspects of male and female green
turtle trophic ecology have found no differences between the
sexes (Vander Zanden et al., 2013; Prior et al., 2016). However,
males in this study displayed a different pattern in δ15N values
ompared to females. Given that the majority of males sampled
hus far were DRTO residents, the error associated with the curves
re likely reflective of the sample size differential between sexes
7
and of the variability found in the stable isotope values or the
sea grasses present in DRTO. Previous green turtle tracking work
in DRTO revealed site fidelity of nesting females to local and
regional foraging areas (Hart et al., 2013), and some overlap of
high-use zones in the park where in-water captured green turtles
also foraged (Fujisaki et al., 2016). Additional fine-scale behav-
ioral analysis of resident juvenile green turtles showed dive and
resting patterns in a foraging ground in another part of the park
(Hart et al., 2016). Two studies that included analysis of genetic
samples collected from green turtles at DRTO showed that adult
nesting females at DRTO are distinct from nesting subpopulations

D.C. Roche, M.S. Cherkiss, B.J. Smith et al.
Regional Studies in Marine Science 48 (2021) 102011
(
6
D
e
t
d
2
5
t
a
Fig. 4. Isotopic biplot of Dry Tortugas National Park green turtle (Chelonia mydas) homogenized epidermis/dermis (skin) based on turtle straight carapace length
SCL, cm) and seagrass samples (Hd = Halophila decipiens, Hw = Halodule wrightii, Sf = Syringodium filiforme, Tt = Thalassia testudinum).

Fig. 5. Green turtle (Chelonia mydas) isotopic consistency of δ 13C over capture events for skin tissue (homogenized epidermis/dermis) samples from turtles both <
5 cm SCL and > 65 cm SCL (A, B). Isotopic consistency of δ 15N over capture events for skin tissue samples from turtles both < 65 cm SCL and > 65 cm SCL (C,
). Markers indicate capture events, straight carapace length (SCL).
r
t
o
t
u
r
ven 40 km away and on the mainland (Shamblin et al., 2020) and
urtles in the mixed foraging aggregation at DRTO were highly
ifferentiated frommost other Atlantic groups (Naro-Maciel et al.,
017).
. Conclusion
This work presents the first known isotopic values for green
urtles in DRTO, thereby establishing the baseline stable carbon
nd nitrogen isotope values of this population which will allow
8
esearchers to track changes in turtle resource use patterns in
he future. The results from the study add to our understanding
f green turtle foraging patterns, with omnivory in smaller size
urtles and a shift towards more specialized herbivorous resource
se in larger turtles; the LMM’s provides insight into how the
elationship of δ13C, δ15N and several variables manifests in this
population of green sea turtles. Seagrass is an important resource
for larger green turtles, and we suggest additional coincident
sampling of resources and turtles in the future to be able to draw

D.C. Roche, M.S. Cherkiss, B.J. Smith et al.
Regional Studies in Marine Science 48 (2021) 102011
m
s
o
s
s
d
y
a
c
B
h
C
t
W
m
B
–
g
e
D
c
t
A
f
e
r
t
C
m
T
2
D
a
f
n
F
C
P
a
R
A
A
A
A
B
ore definitive links between turtles and DRTO resources that
upport green turtles at this protected marine wilderness site.
In this study, resampling of individuals provided a unique
pportunity to generate a baseline understanding of turtle re-
ource use at both the individual- and population-levels. In a
imilar study, Burgett et al. (2018) observed changes in turtle
iet from 12 individuals sampled in Bermuda in two consecutive
ears. Thus, this type of information, when coupled with data for
vailable resources, can serve as a baseline for detecting future
hanges in green turtle resource shifts in areas like DRTO and
ermuda that are often impacted by significant events such as
urricanes and tropical storms.
RediT authorship contribution statement
David C. Roche: Conceptualization, Investigation, Visualiza-
ion, Writing – original draft. Michael S. Cherkiss: Investigation,
riting – review & editing. Brian J. Smith: Investigation, For-
al analysis, Visualization, Writing – review & editing. Derek A.
urkholder: Conceptualization, Writing – original draft, Writing
review & editing. Kristen M. Hart: Conceptualization, Investi-
ation, Resources, Writing – original draft, Writing – review &
diting, Funding acquisition.
eclaration of competing interest
The authors declare that they have no known competing finan-
ial interests or personal relationships that could have appeared
o influence the work reported in this paper.
cknowledgments
We acknowledge past and present employees and volunteers
or assistance in the field and lab, Kelsey Roberts for help with
diting our revised manuscript files, and Will Jenkins for careful
eview of an earlier draft of the manuscript. All work was permit-
ed under the following permits issued to K. Hart: USGS Animal
are permit USGS-SESC-2014-03, NMFS Scientific Research Per-
its 13307, 17381, Florida Marine Turtle Permit 176 and Dry
ortugas Scientific Research Permits DRTO-2008-SCI-0008, DRTO-
010-SCI-0009, DRTO-2012-SCI-0008, DRTO-2014-SCI-0004, and
RTO-2016-SCI-0008. Data generated during this study are avail-
ble as a USGS data release (Roche et al., 2019). Any use of trade,
irm, or product names is for descriptive purposes only and does
ot imply endorsement by the U.S. Government.
inancial support
Funding for portions of this study were provided by the USGS
oastal and Marine Geology Program, USA, USGS Ecosystems
rogram, USA, the USGS Priority Ecosystem Studies Program, USA,
nd the U.S. National Park Service.
eferences
morocho, D.F., Reina, R.D., 2007. Feeding ecology of the East Pacific green
sea turtle chelonia mydas agassizii at Gorgona National Park, Colombia.
Endanger. Species Res. 3, 43–51.
nderson, W.T., Fourqurean, J.W., 2003. Intra- and interannual variability in
seagrass carbon and nitrogen stable isotopes from south Florida, a prelim-
inary study. Org. Geochem. 34, 185–194. http://dx.doi.org/10.1016/S0146-
6380(02)00161-4.
rthur, K.E., Boyle, M.C., Limpus, C.J., 2008. Ontogenetic changes in diet and
habitat use in green sea turtle (Chelonia mydas) life history. Mar. Ecol. Prog.
Ser. 362, 303–311. http://dx.doi.org/10.3354/meps07440.
ult, J.S., Smith, S.G., Meester, G.A., Luo, J., Bohnsack, J.A., Miller, S.L., 2002.
Baseline Multispecies Coral Reef Fish Stock Assessment for the Dry Tortugas.
Fish. Sci. NOAA technical memorandum NMFS-SEFSC ; 487.
artón, K., 2020. MuMIn: Multi-model inference. R Packag. version 1.43.17.
9
Bates, D., Mächler, M., Bolker, B.M., Walker, S.C., 2015. Fitting linear mixed-
effects models using lme4. J. Stat. Softw. http://dx.doi.org/10.18637/jss.v067.
i01.
Bird, C.S., Veríssimo, A., Magozzi, S., Abrantes, K.G., Aguilar, A., Al-Reasi, H.,
Barnett, A., Bethea, D.M., Biais, G., Borrell, A., Bouchoucha, M., Boyle, M.,
Brooks, E.J., Brunnschweiler, J., Bustamante, P., Carlisle, A., Catarino, D.,
Caut, S., Cherel, Y., Chouvelon, T., Churchill, D., Ciancio, J., Claes, J., Cola¸ co, A.,
Courtney, D.L., Cresson, P., Daly, R., de Necker, L., Endo, T., Figueiredo, I.,
Frisch, A.J., Hansen, J.H., Heithaus, M., Hussey, N.E., Iitembu, J., Juanes, F.,
Kinney, M.J., Kiszka, J.J., Klarian, S.A., Kopp, D., Leaf, R., Li, Y., Lorrain, A.,
Madigan, D.J., Maljković, A., Malpica-Cruz, L., Matich, P., Meekan, M.G., Mé-
nard, F., Menezes, G.M., Munroe, S.E.M., Newman, M.C., Papastamatiou, Y.P.,
Pethybridge, H., Plumlee, J.D., Polo-Silva, C., Quaeck-Davies, K., Raoult, V.,
Reum, J., Torres-Rojas, Y.E., Shiffman, D.S., Shipley, O.N., Speed, C.W.,
Staudinger, M.D., Teffer, A.K., Tilley, A., Valls, M., Vaudo, J.J., Wai, T.-C.,
Wells, R.J.D., Wyatt, A.S.J., Yool, A., Trueman, C.N., 2018. A global perspective
on the trophic geography of sharks. Nat. Ecol. Evol. 2, 299–305. http://dx.
doi.org/10.1038/s41559-017-0432-z.
Bjorndal, K.A., 1980. Nutrition and grazing behavior of the green turtle Chelonia
mydas. Mar. Biol. 56, 147–154. http://dx.doi.org/10.1007/BF00397131.
Bjorndal, K.A., 2017. Foraging ecology and nutrition of sea turtles. In: The Biology
of Sea Turtles, Volume I. http://dx.doi.org/10.1201/9780203737088.
Brand-Gardner, S.J., Lanyon, J.M., Limpus, C.J., 1999. Diet selection by imma-
ture green turtles, Chelonia mydas, in subtropical Moreton Bay, south-east
Queensland. Aust. J. Zool. 47, 181. http://dx.doi.org/10.1071/ZO98065.
Bresette, M.J., Witherington, B.E., Herren, R.M., Bagley, D.A., Gorham, J.C.,
Traxler, S.L., Crady, C.K., Hardy, R., 2010. Size-class partitioning and herding
in a foraging group of green turtles Chelonia mydas. Endanger. Species Res.
9, 105–116. http://dx.doi.org/10.3354/esr00245.
Brett, M.T., 2014. Resource polygon geometry predicts Bayesian stable iso-
tope mixing model bias. Mar. Ecol. Prog. Ser. http://dx.doi.org/10.3354/
meps11017.
Bryan, D., Ault, J., 2018. Natural Resource Condition Assessment: Dry Tortugas
National Park. Natural Resource Report NPS/DRTO/NRR—2018/1791.
Burgett, C.M., Burkholder, D.A., Coates, K.A., Fourqurean, V.L., Kenworthy, W.J.,
Manuel, S.A., Outerbridge, M.E., Fourqurean, J.W., 2018. Ontogenetic diet
shifts of green sea turtles (Chelonia mydas) in a mid-ocean developmental
habitat. Mar. Biol. 165, http://dx.doi.org/10.1007/s00227-018-3290-6.
Burkholder, D.A., Heithaus, M.R., Thomson, J.A., Fourqurean, J.W., 2011. Diversity
in trophic interactions of green sea turtles Chelonia mydas on a relatively
pristine coastal foraging ground. Mar. Ecol. Prog. Ser. 439, 277–293. http:
//dx.doi.org/10.3354/meps09313.
Campbell, J.E., Fourqurean, J.W., 2009. Interspecific variation in the elemental
and stable isotope content of seagrasses in South Florida. Mar. Ecol. Prog.
Ser. 387, 109–123. http://dx.doi.org/10.3354/meps08093.
Cardona, L., Aguilar, A., Pazos, L., 2009. Delayed ontogenetic dietary shift and
high levels of omnivory in green turtles (Chelonia mydas) from the NW coast
of Africa. Mar. Biol. 156, 1487–1495. http://dx.doi.org/10.1007/s00227-009-
1188-z.
Cardona, L., Campos, P., Levy, Y., Demetropoulos, A., Margaritoulis, D., 2010.
Asynchrony between dietary and nutritional shifts during the ontogeny of
green turtles (Chelonia mydas) in the Mediterranean. J. Exp. Mar. Bio. Ecol.
393, 83–89. http://dx.doi.org/10.1016/j.jembe.2010.07.004.
Ceriani, S.A., Roth, J.D., Evans, D.R., Weishampel, J.F., Ehrhart, L.M., 2012. Inferring
Foraging Areas of nesting Loggerhead Turtles using satellite telemetry and
stable Isotopes. PLoS One 7, e45335. http://dx.doi.org/10.1371/journal.pone.
0045335.
Davis, G.E., 1979. Outer Continental Shelf Resource Management Map, Coral
Distribution, Fort Jefferson National Monument, the Dry Tortugas. Bureau
of Land Management, OCS Office.
DeNiro, M.J., Epstein, S., 1978. Influence of diet on the distribution of carbon
isotopes in animals. Geochim. Cosmochim. Acta 42, 495–506. http://dx.doi.
org/10.1016/0016-7037(78)90199-0.
DeNiro, M.J., Epstein, S., 1981. Influence of diet on the distribution of nitrogen
isotopes in animals. Geochim. Cosmochim. Acta. 45, 341–351. http://dx.doi.
org/10.1016/0016-7037(81)90244-1.
Ehrhart, L.M., Ogren, L.H., 1999. Studies in foraging habitats: capturing and han-
dling turtles. In: Research and Management Techniques for the Conservation
of Sea Turtles. http://dx.doi.org/10.1037/h0033590.
Fourqurean, J.W., Manuel, S., Coates, K.A., Kenworthy, W.J., Smith, S.R., 2010.
Effects of excluding sea turtle herbivores from a seagrass bed: Overgrazing
may have led to loss of seagrass meadows in Bermuda. Mar. Ecol. Prog. Ser.
419, 223–232. http://dx.doi.org/10.3354/meps08853.
Fry, B., 2006. Stable Isotope Ecology. Springer, New York, http://dx.doi.org/10.
1007/0-387-33745-8.
Fujisaki, I., Hart, K., Sartain-Iverson, A., 2016. Habitat selection by green turtles in
a spatially heterogeneous benthic landscape in Dry Tortugas National Park,
Florida. Aquatic Biol. 24, 185–199. http://dx.doi.org/10.3354/ab00647.
Godley, B.J., Thompson, D.R.D., Waldron, S., Furness, R.R.W., 1998. The trophic
status of marine turtles as determined by stable isotope analysis. Mar. Ecol.
Prog. Ser. 166, 277–284. http://dx.doi.org/10.3354/meps166277.

D.C. Roche, M.S. Cherkiss, B.J. Smith et al.
Regional Studies in Marine Science 48 (2021) 102011
H
H
H
H
H
H
H
H
H
I
K
L
L
L
M
N
N
N
O
P
R
R
R
R
S
S
S
T
V
V
V
V
V
V
W
W
W
all, Goodman A., Avens, L., Braun McNeill, J., Wallace, B., Goshe, L., 2015.
Inferring long-term foraging trends of individual juvenile loggerhead sea
turtles using stable isotopes. Mar. Ecol. Prog. Ser. 537, 265–276. http://dx.
doi.org/10.3354/meps11452.
ammerschlag, N., Gallagher, A.J., Wester, J., Luo, J., Ault, J.S., 2012. Don’t bite
the hand that feeds: Assessing ecological impacts of provisioning ecotourism
on an apex marine predator. Funct. Ecol. 26, 567–576. http://dx.doi.org/10.
1111/j.1365-2435.2012.01973.x.
art, K.M., White, C.F., Iverson, A.R., Whitney, N., 2016. Trading shallow safety
for deep sleep: Juvenile green turtles select deeper resting sites as they grow.
Endanger. Species Res. http://dx.doi.org/10.3354/esr00750.
art, 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, 142–154. http://dx.doi.
org/10.1016/j.biocon.2013.03.019.
atase, H., Sato, K., Yamaguchi, M., Takahashi, K., Tsukamoto, K., 2006. Individual
variation in feeding habitat use by adult female green sea turtles (Chelonia
mydas): are they obligately neritic herbivores? Oecologia 149, 52–64.
aywood, J.C., Casale, P., Freggi, D., Fuller, W.J., Godley, B.J., Lazar, B., Mar-
garitoulis, D., Rees, Al. F., Shutler,
J.D., Snape, R.T., Swain-Diaz, N.R.,
Widdicombe, S., Broderick, A.C., 2020. Foraging ecology of Mediterranean
juvenile loggerhead turtles: insights from C and N stable isotope ratios. Mar.
Biol. 167, 28. http://dx.doi.org/10.1007/s00227-020-3647-5.
aywood, J.C., Fuller, W.J., Godley, B.J., Shutler, J.D., Widdicombe, S., Broder-
ick, A.C., 2019. Global review and inventory: how stable isotopes are helping
us understand ecology and inform conservation of marine turtles. Mar. Ecol.
Prog. Ser. http://dx.doi.org/10.3354/meps12889.
eithaus, M.R., Frid, A., Wirsing, A.J., Bejder, L., Dill, L.M., 2005. Biology of sea
turtles under risk from tiger sharks at a foraging ground. Mar. Ecol. Prog.
Ser. 288, 285–294. http://dx.doi.org/10.3354/meps288285.
obson, K.A., Gloutney, M.L., Gibbs, H.L., 1997. Preservation of blood and tissue
samples for stable-carbon and stable-nitrogen isotope analysis. Can. J. Zool.
75, 1720–1723. http://dx.doi.org/10.1139/z97-799.
shikawa, N.F., 2018. Use of compound-specific nitrogen isotope analysis of
amino acids in trophic ecology: assumptions, applications, and implications.
Ecol. Res. 33, 825–837. http://dx.doi.org/10.1007/s11284-018-1616-y.
éry, M., Royle, J.A., 2015. Applied Hierarchical Modeling in Ecology: Analysis
of distribution, abundance and species richness in r and BUGS, Applied
Hierarchical Modeling in Ecology: Analysis of distribution, abundance and
species richness in R and BUGS.
ayman, C.A., Araujo, M.S., Boucek, R., Hammerschlag-Peyer, C.M., Harrison, E.,
Jud, Z.R., Matich, P., Rosenblatt, A.E., Vaudo, J.J., Yeager, L.A., Post, D.M.,
Bearhop, S., 2012. Applying stable isotopes to examine food-web structure:
an overview of analytical tools. Biol. Rev. 87, 545–562. http://dx.doi.org/10.
1111/j.1469-185X.2011.00208.x.
impus, C.J., Coffee, O.I., 2019. Marine turtles in Moreton Bay. In: Tibbetts, I.R.,
Rothlisberg, P.C., Neil, D.T., Homburg, T.A., Brewer, D.T., Arthington, A.H.
(Eds.), Moreton Bay Quandamooka & Catchment: Past, present, and future.
In: Advances in Design and Control, The Moreton Bay Foundation, Brisbane,
Australia, Available from: https://moretonbayfoundation.org/.
utcavage, M.E., Plotkin, P., Witherington, B., Lutz, P.L., 1997. Human impacts on
sea turtle survival. In: The Biology of Sea Turtles, Volume I. http://dx.doi.
org/10.1201/9780203737088.
ortimer, J.A., 1981. The Feeding Ecology of the West Caribbean Green Turtle
(Chelonia mydas) in Nicaragua. Biotropica 13, 49. http://dx.doi.org/10.2307/
2387870.
akagawa, S., Schielzeth, H., 2013. A general and simple method for obtaining
R2 from generalized linear mixed-effects models. Methods Ecol. Evol. http:
//dx.doi.org/10.1111/j.2041-210x.2012.00261.x.
aro-Maciel, E., Hart, K.M., Cruciata, R., Putman, N.F., 2017. DNA and dis-
persal models highlight constrained connectivity in a migratory marine
megavertebrate. Ecography (Cop.) http://dx.doi.org/10.1111/ecog.02056.
ational Marine Fisheries Service Southeast Fisheries Science Center, 2008.
Sea Turtle Research Techniques Manual. NOAA Technical Memorandum
NMFS-SEFSC-579, p. 92.
wens, D.W., Ruiz, G.J., 1980. New methods of obtaining blood and cerebrospinal
fluid from Marine turtles. Herpetologica 36, 17–20.
arker, D.M., Dutton, P.H., Balazs, G.H., 2011. Oceanic Diet and Distribution of
Haplotypes for the Green Turtle, Chelonia mydas, in the Central North Pacific.
Pacific Sci. 65, 419–431. http://dx.doi.org/10.2984/65.4.419.
Peterson, B.J., Fry, B., 1987. Stable isotopes in ecosystem studies. Annu. Rev. Ecol.
Syst. 18, 293–320.
10
Phillips, D.L., Inger, R., Bearhop, S., Jackson, A.L., Moore, J.W., Parnell, A.C.,
Semmens, B.X., Ward, E.J., 2014. Best practices for use of stable isotope
mixing models in food-web studies. Can. J. Zool. http://dx.doi.org/10.1139/
cjz-2014-0127.
Post, D.M., 2002. Using stable isotopes to estimate trophic position: models,
methos, and assumptions. Ecology 83, 703–718. http://dx.doi.org/10.2307/
3071875.
Post, D.M., Layman, C.A., Arrington, D.A., Takimoto, G., Quattrochi, J., Mon-
taña, C.G., 2007. Getting to the fat of the matter: Models, methods and
assumptions for dealing with lipids in stable isotope analyses. Oecologia 152,
179–189. http://dx.doi.org/10.1007/s00442-006-0630-x.
Prior, B., Booth, D.T., Limpus, C.J., 2016. Investigating diet and diet switching in
green turtles (Chelonia mydas). Aust. J. Zool. 63, 365–375.
Core Team, 2019. R: A Language and Environment for Statistical Computing.
R Foundation for Statistical Computing, Vienna, Austria, URL https://www.R-
project.org/.
andall, J.E., 1965. Grazing effect on Sea Grasses by Herbivorous Reef Fishes in
the West Indies. Ecology 46, 255–260. http://dx.doi.org/10.2307/1936328.
eich, K.J., Bjorndal, K.A., Martínez Del Rio, C., 2008. Effects of growth and tissue
type on the kinetics of 13C and 15N incorporation in a rapidly growing
ectotherm. Oecologia http://dx.doi.org/10.1007/s00442-007-0949-y.
oche, D.C., Smith, B.J., Burkholder, D.A., Hart, K.M., 2019. Green turtle (Chelonia
mydas) stable isotope signatures from Dry Tortugas National Park, Florida,
sampled 2008–2015: U.S. Geological Survey data release. http://dx.doi.org/
10.5066/P9060E4Q.
eminoff,
J.A., Allen, C.D., Balazs, G.H., Dutton, P.H., Eguchi, T., Haas, H.,
Hargrove, S.A., Jensen, M., Klemm, D.L., Lauritsen, A.M., MacPherson, S.L.,
Opay, P., Possardt, E.E., Pultz, S., Seney, E.E., Van Houtan, K.S., Waples, R.S.,
2015. Status review of the green turtle (Chelonia mydas) under the
Engangered Species Act. NOAA technical memorandum NMFS.
eminoff, J., Jones, T., Eguchi, T., Jones, D., Dutton, P., 2006. Stable isotope
discrimination (δ13C and δ15N) between soft tissues of the green sea
turtle Chelonia mydas and its diet. Mar. Ecol. Prog. Ser. 308, 271–278.
http://dx.doi.org/10.3354/meps308271.
hamblin, B., Hart, K., Martin, K., Ceriani, S., Bagley, D., Mansfield, K., Ehrhart, L.,
Nairn, C., 2020. Green turtle mitochondrial microsatellites indicate finer-scale
natal homing to isolated islands than to continental nesting sites. Mar. Ecol.
Prog. Ser. http://dx.doi.org/10.3354/meps13348.
ieszen, L.L., Boutton, T.W., Tesdahl, K.G., Slade, N.A., 1983. Fractionation and
turnover of stable carbon isotopes in animal tissues: Implications for δ13C
analysis of diet. Oecologia 57, 32–37. http://dx.doi.org/10.1007/BF00379558.
an Houtan, K., Pimm, S., 2007. Assessment of the Dry Tortugas National Park
Sea Turtle Monitoring Program 1982–2006: Ecological Trends & Conservation
Recommendations. Natl. Park Serv., Homestead, FL, USA.
ander Zanden, H.B., Arthur, K.E., Bolten, A.B., Popp, B.N., Lagueux, C.J., Harri-
son, E., Campbell, C.L., Bjorndal, K.A., 2013. Trophic ecology of a green turtle
breeding population. Mar. Ecol. Prog. Ser. 476, 237–249. http://dx.doi.org/10.
3354/meps10185.
ander Zanden, H.B., Bjorndal, K.A., Mustin, W., Ponciano, J.M., Bolten, A.B.,
2012. Inherent variation in Stable Isotope values and discrimination factors
in two life stages of green turtles. Physiol. Biochem. Zool. 85, 431–441.
http://dx.doi.org/10.1086/666902.
ander Zanden, H.B., Bjorndal, K.A., Reich, K.J., Bolten, A.B., 2010. Individual
specialists in a generalist population: results from a long-term stable isotope
series. Biol. Lett. 6, 711–714. http://dx.doi.org/10.1098/rsbl.2010.0124.
ander Zanden, H.B., Tucker, A.D., Bolten, A.B., Reich, K.J., Bjorndal, K.A., 2014.
Stable isotopic comparison between loggerhead sea turtle tissues. Rapid
Commun. Mass Spectrom. 28, 2059–2064. http://dx.doi.org/10.1002/rcm.
6995.
ander Zanden, H.B., Tucker, A.D., Hart, K.M., Lamont, M.M., Fujisaki, I., Addi-
son, D.S., Mansfield, K.L., Phillips, K.F., Wunder, M.B., Bowen, G.J., Pajuelo, M.,
Bolten, A.B., Bjorndal, K.A., 2015. Determining origin in a migratory marine
vertebrate: a novel method to integrate stable isotopes and satellite tracking.
Ecol. Appl. 25, 320–335. http://dx.doi.org/10.1890/14-0581.1.
aara, R.J., Patterson, J.M., Atkinson, a.J., Estep, a.J., National Park Service, U.S.D.
of the I., 2011. Development and policy applications of the 2010 benthic
habitat map for Dry Tortugas National Park.
iley, A.E., Rossman, S., Ostrom, P.H., France, C.A.M., Penniman, J., Bailey, C.,
Duvall, F., Zipkin, E.F., James, H.F., 2019. From ecologically equivalent in-
dividuals to contrasting colonies: quantifying isotopic niche and individual
foraging specialization in an endangered oceanic seabird. Mar. Biol. 166,
1–13. http://dx.doi.org/10.1007/s00227-019-3483-7.
itherington, B.E., 1992. Behavioral responses of nesting sea turtles to artificial
lighting. Herpetologica 48, 31–39.