ORIGINAL ARTICLE
Influence of oceanographic conditions on abundance and
distribution of post-larval and juvenile carangid fishes in the
northern Gulf of Mexico
John A. Mohan
1
|Tracey T. Sutton
2
|April B. Cook
2
|Kevin M. Boswell
3
|
R. J. David Wells
1,4
1
Department of Marine Biology, Texas
A&M University at Galveston, Galveston,
TX, USA
2
Department of Marine and Environmental
Sciences, Nova Southeastern University,
Dania Beach, FL, USA
3
Department of Biological Sciences, Florida
International University, North Miami, FL,
USA
4
Department of Wildlife and Fisheries
Sciences, Texas A&M University, College
Station, TX, USA
Correspondence
John A. Mohan
Email: [email protected]
Funding information
Gulf of Mexico Research Initiative; National
Oceanic and Atmospheric Administration
Abstract
Relationships between abundance of post-larval and juvenile carangid (jacks) fishes
and physical oceanographic conditions were examined in the northern Gulf of Mex-
ico (GoM) in 2011 with high freshwater input from the Mississippi River. General-
ized additive models (GAMs) were used to explore complex relationships between
carangid abundance and physical oceanographic data of sea surface temperature
(SST), sea surface height anomaly (SSHA) and salinity. The five most abundant car-
angid species collected were:Selene setapinnis(34%);Caranx crysos(30%);Caranx
hippos(10%);Chloroscombrus chrysurus(9%) andTrachurus lathami(8%). Post-larval
carangids (median standard length [SL]=10 mm) were less abundant during the
spring and early summer, but more abundant during the late summer and fall, sug-
gesting summer to fall spawning for most species. Juvenile carangid (median
SL=23 mm) abundance also increased between the mid-summer and early fall.
Most species showed increased abundance at lower salinities and higher tempera-
tures, suggesting entrainment of post-larval fishes or feeding aggregations of juve-
niles at frontal convergence zones between the expansive river plume and dynamic
mesoscale eddy water masses. However, responses were species- and life-stage
specific, which may indicate fine-scale habitat partitioning between species. Ordina-
tion methods also revealed higher carangid abundances at lower salinities for both
post-larval and juvenile life stages, with species- and life-stage specific responses to
SST and SSHA, further suggesting habitat separation between species. Results indi-
cate strong links between physical oceanographic features and carangid distributions
in the dynamic northern GoM.
KEYWORDS
Carangidae, generalized additive models, Gulf of Mexico, large midwater trawl, Mississippi
River, MOC
1|INTRODUCTION
Physical oceanographic features structure marine communities
through bottom-up trophic interactions and passive concentration
mechanisms (Godø et al., 2012; Lima, Olson, & Doney, 2002; Lindo-
Atichati et al., 2012; Meekan et al., 2006; Williams, Mcinnes, Roo-
ker, & Quigg, 2015). Oceanic currents and eddies occur over diverse
spatial scales and interact over various temporal scales (Hamilton,
1992; Vukovich, 2007). Hydrodynamic convergence and turbulent
mixing supply nutrients in otherwise oligotrophic waters, fueling
Received: 19 August 2016|
Accepted: 26 January 2017
DOI: 10.1111/fog.12214
Fisheries Oceanography.2017;1–16. wileyonlinelibrary.com/journal/fog ©2017 John Wiley & Sons Ltd|
1

primary production and transferring energy to higher trophic levels
(Bakun, 2006). In the northern hemisphere, anticyclonic warm core
eddies spin clockwise causing downwelling and are considered nutri-
ent-depleted (Biggs, 1992). Cyclonic cold core eddies spin counter-
clockwise, resulting in upwelling that introduces new nitrogen into
the base of the mixed layer (Biggs, 1992; Seki et al., 2001). Frontal
zones occur at the confluence of anticyclone/cyclone eddy pairs and
are often associated with increased primary and secondary produc-
tion that may be advected to remote offshore locations (Toner et al.,
2003).
The Gulf of Mexico (GoM) is a semi-enclosed intercontinental
sea with a unique circulation that is controlled by the intrusion of
the Loop Current (LC) comprising warm water from the Caribbean
Sea that enters through the Yucatan Strait and exits through the
Straits of Florida. As the LC extends into the eastern GoM, anticy-
clonic warm-core mesoscale eddies are shed and transported west-
ward into the central and western GoM (Biggs, 1992; Vukovich &
Crissman, 1986). Frictional interactions of warm-core eddies with
the steep topography of the continental slope form cyclonic/anticy-
clonic eddy pairs (Biggs & M€uller-Karger, 1994; Hamilton, 1992). The
physical characteristics of mesoscale features and fronts including
sea surface height anomaly (SSHA), sea surface temperature (SST),
salinity and nutrient gradients influence the distribution of primary
producers and subsequent secondary consumers including larval and
juvenile fishes (Grimes & Finucane, 1991; Rooker et al., 2013). Addi-
tionally, discharge from the Mississippi River, which drains two-thirds
of the US continent, delivers low salinity, nutrient-rich water into
the northern GoM that enhances primary and secondary productivity
in nearshore waters (Chesney, Baltz, & Thomas, 2000; Grimes,
2001). Increased nitrate and chlorophyll concentrations resulting
from mesoscale circulation, are thought to support enhanced zoo-
plankton and nekton biomass (Zimmerman & Biggs, 1999).
In the northern GoM, the Mississippi River plume is a dominant
feature that can extend over 400 km to offshore oceanic regions
(Del Castillo et al., 2001). The transport of terrestrial enriched river
discharge with high levels of“new”nitrate (Dagg & Breed, 2003)
results in high fishery production in the northern GoM (Chesney
et al., 2000; Grimes, 2001). In 2011, record flooding of the Missis-
sippi River with a peak discharge of 40,000 m
3
s
!1
in May that was
twice the 60 year mean (Walker, Wiseman, Rouse, & Babin, 2005),
resulted in an expansive plume that was identifiable with satellite
measurements (Falcini et al., 2012; Gierach, Vazquez-Cuervo, Lee, &
Tsontos, 2013). Nitrate levels recorded in the far-field plume of the
Mississippi River ranged from 5 to 10lM (Dagg & Breed, 2003;
Lohrenz et al., 1999) and are comparable to nitrate concentrations
measured in cyclonic eddies in the GoM at 100 m depth ranging
from 11 to 15lM (Biggs & M€uller-Karger, 1994; Zimmerman &
Biggs, 1999). Thus, cyclonic eddies and the river plume features may
support similar levels of primary and secondary production and pro-
vide enhanced food for larval and juvenile fishes.
The Carangidae (jack) family of fishes occurs worldwide through-
out tropical and temperate waters, primarily occupying the epipelagic
(upper 200 m) water column. Most fishes in the family Carangidae
form large schools that prey upon shrimps, squids, and other fishes
while at the same time supporting the diets of large predators such
as tunas, sharks, and dolphins (Kiszka, M"endez-Fernandez, Heithaus,
& Ridoux, 2014; Shimose & Wells, 2015; Torres-Rojas, Hern"andez-
Herrera, Galv"an-Maga~na, & Alatorre-Ram"ırez, 2010). This important
family of fishes supports 5% of the annual marine finfish landings
owing to its value as bait, sportfish, and food in many recreational
and commercial fisheries (Ditty, Shaw, & Cope, 2004; Leak, 1981).
Despite the importance of the Carangidae family, few studies have
attempted to link oceanographic and environmental conditions to
the distribution and abundance patterns of post-larvae and juvenile
carangids in the oceanic GoM. Ditty et al. (2004) found carangid lar-
vae were concentrated in areas of abundant zooplankton prey in the
northern GoM, and suggested that dynamic frontal areas served as
nurseries. Grimes and Finucane (1991) reported that carangids were
the most abundant genera at the river plume and shelf-sampling sta-
tions in the northern GoM and hypothesized that increased feeding
and growth would promote occupancy associated with river plume
features. Carangids were also the most abundant species inside a
river plume at the Great Barrier Reef, Australia (Thorrold & McKin-
non, 1995). Given the sparse information on such an important epi-
pelagic fish family in the GoM, the aim of the present study was to
examine the role of oceanic conditions on the abundance and distri-
bution of carangid fishes to better predict important areas such as
spawning, nursery, and feeding grounds. In addition, conditions in
the northern GoM in 2011 were influenced by natural gradients of
salinity due to record freshwater discharge, providing an interesting
setting in which to examine carangid responses. The objectives of
this study were to (i) investigate carangid abundance and distribution
in relation to sea surface temperature, salinity, and sea surface
height anomaly in the GoM; (ii) compare the responses of post-larval
and juvenile carangid life stages that were collected using two gear
types across spring, summer and fall seasons; (iii) explore differences
among the five most abundant carangid species collected. This infor-
mation may provide important data for evaluating carangid
responses to natural climate variability.
2|METHODS
2.1|Study area and collection methods
Collections occurred during four research cruise series totaling
160 days in the northern GoM in 2011:Meg SkansiA from 21 April–
29 June (N=44 stations);Meg SkansiB from 20 July–28 September
(N=44 stations);PiscesA from 23 June–12 July (N=12 stations);
andPiscesB from 8 September–26 September (N=13 stations).
These research expeditions were part of the larger Deepwater Hori-
zon Natural Resource Damage Assessment (NRDA) conducted in the
northern GoM (http://www.gulfspillrestoration.noaa.gov). Samples
were collected aboard the R/VMeg Skansiusing a 10-m
2
Multiple
Opening and Closing Net and Environmental Sensing System (MOC)
net (Wiebe et al., 1985; details below). The R/VPiscesdeployed a
large, dual-warp midwater trawl (LMT) net (Judkins, Vecchione,
2|
MOHAN ET AL.

Cook, & Sutton, 2016; details below). Fish abundance was standard-
ized by dividing the number of fishes collected by the volume of
water sampled (m
3
) for each gear type. From here on each cruise will
be identified by gear type deployed (MOC or LMT) and labeled by
season. For instance,Meg SkansiA=MOC spring/summer;Meg
SkansiB=MOC summer/fall;PiscesA=LMT summer; andPisces
B=LMT fall.
2.2|R/V Meg Skansi: MOC sampling
A 10-m
2
mouth area MOC (3-mm mesh: MOC) net system was used
owing to its capability of taking discrete samples over specific depth
strata. At each station, a Conductivity, Temperature, and Depth (CTD)
sensor array (SBE 911 Plus; Sea-Bird Electronics, Inc.) was cast at
dawn and dusk. The MOC was deployed at either 09.00 or 21.00 hr
such that either solar noon or midnight occurred at the midpoint of
the tow. The MOC sampled from 0 to 1,500 m depth during descent
and then sampled five depth strata discretely during retrieval: 1,500–
1,200, 1,200–1,000, 1,000–600, 600–200, and 200–0 m. The volume
filtered by each net was calculated using an algorithm that incorpo-
rated flowmeter (TSK model) data and net angle (inclinometer), with
the latter used to estimate mouth area perpendicular to tow direction.
Ship speed during net deployment was approximately 1.5 knots
(1 knot=~0.5 m/s). As a result of low numbers of carangid fishes col-
lected in deep depth bins (Figure S1) only collections from the surface
depth bin (0–200 m) were considered for further analysis. There was
no significant difference in carangid abundance between day and night
samples for MOC collections for all species pooled and individual spe-
cies (Figure S2 and Table S1), thus day and night collections were
pooled for each station to correspond with daily environmental mea-
surements from the CTD and satellites.
2.3|R/V Pisces: LMT sampling
The R/VPiscesconducted deep sampling using a large, dual-warp,
high-speed pelagic trawl. The LMT is a commercial four-seam midwa-
ter trawl with a minimal drag that is capable of sampling larger and
more mobile species than the MOC (Judkins et al., 2016). The mesh
size of the LMT decreases along the body of the net from 6.5 m down
to 6 cm at the last panel of webbing before the codend. Net
mensuration sensors and data-loggers were used to actively monitor
the fishing depth of the net during the tow. Data from these sensors
also provided information on wingspread and mouth opening during
the tow, which was then used to calculate approximate net geometry.
The LMT effective mouth area was estimated to be 165.5 m
2
. Volume
filtered was calculated using an algorithm that incorporated mouth
area and the oblique distance traveled by the net. Sampling occurred
over a 24-hr period using oblique tows at a speed of five knots from
the surface to depths ranging from 700 m to a maximum of 1,400 m.
At each station, there were 2 day tows and two night tows; however,
there was no significant difference in carangid abundance between
day and night samples for LMT collections for all species pooled and
individual species (Figure S3), thus day and night samples were pooled
together for this study to correspond with daily environmental mea-
surements from the CTD and satellites.
Once the nets were retrieved on deck, the catch was sorted into
“rough”taxonomic groupings (i.e. fish families). Roughly sorted
groupings were weighed on a motion-compensating scale and then
preserved in 10% buffered formalin. All fishes from each station
were kept and archived; nothing was discarded. Later sample pro-
cessing in the laboratory involved further sorting, identification to
lowest taxonomic level possible, species counts, cumulative species
weights (wet weight, after blotting), and length measurements. Abun-
dance data were standardized by effort (volume filtered, m
3
). For the
10 m
2
MOC, volumes were calculated using flowmeter and net
mouth angle data. The volume sampled using the LMT was two
orders of magnitude larger compared to MOC cruises (Table 1);
therefore, abundance was multiplied by 100,000 for MOC data (ex-
pressed as individuals [ind]/m
3
[910
!5
]) and by 10,000,000 for LMT
data (ind/m
3
[910
!7
]) to make values comparable for plotting pur-
poses by displaying abundances on similar scales. The spatial and
seasonal distribution and abundance of carangid species were exam-
ined by generating contoured heat maps for each cruise using the
Data-Interpolating Variational Analysis (DIVA: Troupin et al., 2012)
gridding option in Ocean Data View (ODVversion 4.5.6).
2.4|Sample processing
All specimens collected in LMT and MOC samples were identified to
the lowest possible taxonomic level, in most cases to the species
TABLE 1 Summary statistics of mean ("SD) and range of physical factors measured with CTD (salinity), and satellite (SST and SSHA) at
each station during each research cruise using the MOC and LMT during the spring, summer and fall of 2011
Physical factor Statistic
Research cruise
MOC spring/summer LMT summer MOC summer/fall LMT fall
Temperature (°C) Mean 26.9 "1.7 29.5 "0.43 30.5 "0.95 28.9 "2.2
Range 24.4 to 29.7 29.1 to 30.5 28.6 to 31.8 27.7 to 29.5
Salinity Mean 36 "1 34.5 "3.3 33.6 "2.7 35.3 "2.2
Range 31.5 to 36.7 24.6 to 36.4 23 to 36.3 28.8 to 36.7
SSHA (cm) Mean 6.6 "12.8 24.4 "15.6 16.4 "11 14.6 "10.8
Range !11.5 to 42.3 3.7 to 46.7 !0.36 to 48.7 3.7 to 38.5
MOHAN ET AL. |3

level (93% of specimens). For groups that contained only a few spec-
imens, all specimens were identified to the lowest taxonomic classifi-
cation possible and measured to the nearest millimeter (mm)
standard length (SL). For larger catches, a subset of 25 individuals
were measured. The percent occurrence of each species was calcu-
lated as the number of stations a species was present divided by the
total number of stations sampled.
2.5|Environmental data
A CTD array was deployed at each site to record environmental vari-
ables including temperature, salinity, dissolved oxygen (DO) and fluo-
rescence. DO and fluorescence data were not available for every
site and therefore not included in further analysis. However, the lim-
ited fluorescence data was strongly correlated to salinity (Figure S4).
Measurements from the upper 1–3 m and day and night were aver-
aged together for daily measurements to correspond to daily satellite
surface measurements. CTD salinity data were unavailable for 16%
of the stations; therefore, the model estimated salinity data were
used to fill in data gaps (explained below), and SST and SSHA was
obtained exclusively from satellite data. SST measurements from the
CTD were strongly correlated with remotely sensed SST values
(Pearson’sr
2
=.92,p<.0001,N=95). Remotely sensed data for
each station were obtained using the Marine Geospatial Ecology
Tools (MGET) in ArcGIS (v10.2) (Roberts, Best, Dunn, Treml, & Hal-
pin, 2010). SSHA measurements were obtained from Aviso DUAC
2014 gridded products from merged satellites at 1/3-degree resolu-
tion. SST estimates were gathered from the NASA JPL PO.DAAC
MODIS Aqua satellites at 1/24-degree resolution. The HYCOM &
NCODA models were used to estimate surface salinity at 1/25-
degree resolution. All remotely sensed data from the LMT cruises
were downloaded as mean statistics using cumulative climatology
bins over the dates of each cruise that encompassed less than
2 weeks. For MOC cruise series that spanned over 3 months each,
remotely sensed data were downloaded as mean statistics using a
monthly climatological bin type. To examine the spatial and temporal
variability in oceanographic conditions, surface layer maps of salinity
(primarily measured with CTD), SST and SSHA (satellite measure-
ments) were created inODV(version 4.5.6) using the weighted aver-
age gridded data display option. The weighted average gridding
option was chosen as it represents discreet values of conditions
measured at each station where carangid abundance was quantified,
and these paired measurements were used in GAM and RDA
analysis.
2.6|Statistical analysis
GAMs were used to explore relationships between carangid abun-
dance (dependent variable) and physical oceanographic data, includ-
ing salinity (CTD measurements), SST and SSHA (satellite
measurements) as continuous explanatory variables and season as a
categorical factor. GAMs are versions of Generalized Linear Models
that permit complex nonlinear relationships between explanatory
and response variables to be explored (Hastie & Tibshirani, 1986).
Abundance estimates were rounded to the nearest integer for mod-
eling purposes. The general GAM model follows the equation:
E½y$¼g
!1
b
0þ
X
k
skðxkÞ
!
whereE[y]=the expected values of the response variable,g=the
link function,b0=the intercept,xrepresents one ofkexplanatory
variables, ands
k=the smoothing function for each explanatory vari-
able.
Owing to the differences in size selectivity between MOC and
LMT gear types, separate models were run for each dataset.
Collinearity of explanatory variables was examined with variance
inflation factors (VIF) in theusdmpackage inRversion 3.0.2. The
VIFs for all explanatory variables were≤5 so all variables were used
in the GAM models. Logarithmic links with cubic regression splines
were fit with the software packagemgcvin R. A negative binomial
distribution was used because of the high abundance of zeros in the
data set (Drexler & Ainsworth, 2013). All models employed four
degrees of freedom for each variable to prevent over fitting and
reduce the risk of generating ecologically unrealistic responses (Leh-
mann, Overton, & Leathwick, 2002). To explore potential effects of
increased degrees of freedom, k was increased to 6, 8, and 10; how-
ever, the general shape of fish-environment relationships did not
change. Response plots were generated for those physical variables
that were deemed to have a significant influence (a=.05) on the
abundance of carangid fishes; non-significant variables were not
plotted. To examine overall model fit, percent deviance explained
(DE) was calculated for each model (([null deviance–residual
deviance]/null deviance)9100).
Ordination methods were used to further examine relationships
between environmental conditions and the abundance of each spe-
cies. Constrained linear Redundancy Analysis (RDA) was performed
inCANOCO(version 5.04). The RDAs were run separately for each
gear type to see if differences would be apparent between post-lar-
val (MOC) and juvenile (LMT) life stages.
3|RESULTS
3.1|Oceanographic conditions
During MOC–spring/summer, the temperature was lower (25–26°C)
in northern sites with negative (–10 to 0 cm) SSHA, but higher
(>28°C) in southern central locations displaying positive (10–45 cm)
SSHA (Figure 1a and e). The greatly increased SSHA (>30 cm) and
increased temperature (>29°C) suggested an extension of the LC or
an anticyclonic warm core eddy that traveled in a westerly direction
between MOC–spring/summer and MOC–summer/fall (Figure 1e
and f). Salinity was generally homogeneous (~36) during MOC–
spring/summer except for a region of low salinity (~32) at a southern
station with the highest SSHA (Figure 1c). Decreased salinities
(24–32) were evident in northern regions of MOC–summer/fall
4|
MOHAN ET AL.

indicating a southward extension of the Mississippi River plume (Fig-
ure 1d). During MOC–summer/fall, temperatures were approximately
5°C warmer compared with MOC–spring/summer in northern and
western regions, but were~2°C cooler in eastern regions that had
displayed lower SSHA (Figure 1b and f). Similar patterns were exhib-
ited during LMT–summer and LMT–fall cruises: lower salinities (26–
32) in northern sites identified the river plume (Figure 2c and d) and
positive SSHAs (35–45 cm) from the LC extension/anticyclonic eddy
shifted in a westerly direction between the LMT–summer and
LMT–fall (Figure 2e and f). For LMT–summer, warmer temperatures
(30–31°C) occurred at the northern stations (Figure 2a), whereas for
LMT–fall warmer temperatures (~30°C) were associated with positive
SSHA (~35 cm) at western stations (Figure 2b and f). Descriptive
statistics (mean"standard deviation [SD]) and range for environ-
mental data are presented amongst gear types and seasons
(Table 1).
3.2|Carangid abundance
A total of 8,436 carangid fishes were collected and identified to the
family level or below (Table 2). The majority of carangids were col-
lected during the LMT sampling, comprising 26% and 60% of the
total catch in the summer and fall, respectively. Fewer carangids
were collected during the MOC–summer/fall (12%) and MOC–
spring/summer (2%) sampling. The two most abundant species were
Selene setapinnis(34%) andCaranx crysos(30%), followed byCaranx
hippos(10%),Chloroscombrus chrysurus(9%),andTrachurus lathami
(8%).The generaCaranx(5%) andSelene(1%) were next in the order
of the abundance, and these specimens were only identified to gen-
era-specific taxonomic levels. Eleven additional species from the
Carangidae family were also identified, but these comprised<2.8%
of total abundance and thus were not included in the analysis owing
to low sample sizes.
Carangid species displayed a higher frequency of occurrence in
the LMT sampling compared to the MOC sampling (Table 3). For the
LMT,S. setapinnis,C. crysos,C. hippos,andT. lathamiall occurred in
>75% of samples, whereasC. chrysurusoccurred less frequently in
only 26% of samples. Frequencies of occurrence of all carangid spe-
cies from the MOC were generally low, ranging from 7%–35%, with
C. crysosoccurring most frequently (35%) andC. chrysurusoccurring
least frequently (7%) (Table 3). Length frequency histograms revealed
that smaller carangids were collected during MOC compared with
LMT (Figure 3). The median SL for MOC was 10 mm while the
median SL for LMT was 23 mm. The range of carangid lengths
overlapped between cruises (range MOC: 3–55 mm; range LMT:
9–149 mm), but larger fish were collected using the LMT (mean
SL"SD=27.02"15.94 mm) compared with the MOC (mean
SL"SD=12.11"6.62 mm) and this pattern was consistent for
the top five most abundant species examined (Table 4).
3.3|Carangid abundance and distribution
Carangid abundance heat maps were created for individual species
and seasonal cruises to compare abundance and distribution
throughout the northern GoM (Figures 4 and 5). For the MOC sam-
plingS. setapinniswas absent in spring/summer except for 1 individ-
ual (Figure 4a), but in the summer/fallS. setapinnisdisplayed
centralized high abundance (>60 ind/m
3
910
!5
) with zero abun-
dance at western sites and increased abundance ( ~20 ind/
m
3
910
!5
) at eastern sites (Figure 4b). Similarly,C. crysosabun-
dance was low in the spring/summer in south-central sites and
increased in the summer/fall with moderate abundance (20–40 ind/
m
3
910
!5
) at single west and central sites, with higher abundance
(20–80 ind/m
3
910
!5
) observed in the south eastern region (Fig-
ure 4c and d).Caranx hipposdisplayed low patchy abundance
(2–4 ind/m
3
910
!5
) for MOC in the spring/summer in the central
(a) (b)
(c) (d)
(e) (f)
FIGURE 1 Physical oceanographic
conditions of temperature (SST) (a, b),
salinity (c, d), and sea surface height
anomaly (SSHA) (e, f) present during MOC
sampling in the spring/early summer and
late summer/fall of 2011. Colors represent
weighted average gridding in ODV
MOHAN ET AL. |5

(a) (b)
(c) (d)
(e) (f)
FIGURE 2 Physical oceanographic
conditions of temperature (SST) (a, b),
salinity (c, d), and sea surface height
anomaly (SSHA) (e, f) present during LMT
sampling in the summer and fall of 2011.
Colors represent weighted average
gridding in ODV
TABLE 2 Total numbers of carangid fishes (by genera and species) collected in the northern GoM using MOC and LMT gear types during
the spring, summer and fall seasons. Total sum and average volume of water column sampled during each cruise also presented. Carangids
listed highest to lowest based on total abundance. *Specimens only identified to the Genus level owing to morphological damage.
Species
Research cruise
Grand total
MOC
spring/summer
LMT
summer
MOC
summer/fall LMT fall
Selene setapinnis 1 425 258 2,143 2,827
Caranx crysos 2 663 252 1,611 2,528
Caranx hippos 11 409 20 411 851
Chloroscombrus chrysurus 130 357 254 741
Trachurus lathami 5 118 6 518 647
Caranxsp.* 96 314 42 8 460
Selenesp.* 30 78 1 109
Caranx bartholomaei 31 3 30 64
Selar crumenophthalmus 1 43 4 2 50
Decapterussp.* 3 27 10 40
Selene brownii 36 36
Decapterus macarellus 15 12 1 28
Decapterus tabl 6 12 2 20
Caranx ruber 12 12
Decapterus punctatus 99
Alectis ciliaris 41 5
Selene vomer 31 4
Uraspis secunda 314
Pseudocaranx dentex 11
Grand total 170 2,202 1,035 5,029 8,436
Total sum of volume sampled 1,378,606 110,127,577 1,927,156 113,497,582
Mean volume sampled 49,236 9,177,298 52,085 8,730,583
6|
MOHAN ET AL.

region and high abundance in the south eastern region during the
summer/fall (Figure 4e and f).Chloroscombrus chrysuruswas absent
from spring/summer, but highly abundant (>300 ind/m
3
910
!5
) at
one central station in the summer/fall (Figure 4g and h).Trachurus
lathamiexhibited low abundance (2–8 ind/m
3
910
!5
) that shifted
from the east in the spring/summer to central and western sites in
the summer/fall (Figure 4i and j).
For LMT carangid abundances were much higher in both the
summer and fall compared to MOC (Figure 5). In the summer,S. se-
tapinniswas most abundant (100–150 ind/m
3
910
!7
) in the north-
central sites, but the abundances shifted to the east sites (400 ind/
m
3
910
!7
) during the fall but abundances remained high in the
north-central (200 ind/m
3
910
!7
) (Figure 5a and b). Abundances of
C. crysoswere increased (150 ind/m
3
910
!7
) at north-central sites
in the summer and heavily concentrated (600 ind/m
3
910
!7
) at a
single northern site in the fall (Figure 5c and d).Caranx hipposdis-
played high abundance (100 ind/m
3
910
!7
) at northern sites in the
summer that shifted to western sites in the fall (Figure 5e and f).
Abundances ofC. chrysuruswere increased (100 ind/m
3
910
!7
) at
the northern site (Figure 5g and h) during both the summer and fall.
Trachurus lathamiexhibited moderate abundance (50 ind/
m
3
910
!7
) at the northern site in the fall that increased and shifted
to the west (200 ind/m
3
910
!7
) during the fall sampling (Figure 5i
and j).
3.4|Generalized additive models
Salinity was the only variable that exhibited a significant relationship
to abundance for every carangid species (Table 5). However, those
relationships varied among species and between gear types. In gen-
eral, the deviance explained was high (DE>50) for each species and
gear type and ranged from 45% to 96% (Table 5). The season of col-
lection was also a significant factor for most species, except for
C. chysurusandT. lathamiLMT collections (Table 5).Selene setapinnis
TABLE 3 Frequency of occurrence for
each of the most common carangid
species collected using MOC and LMT
gear
Cruise
Carangid frequency of occurrence (%)
Selene
setapinnis
Caranx
crysos
Caranx
hippos
Chloroscombrus
chrysurus
Trachurus
lathami
Species
pooled
MOC 26 35 10 7 8 70
LMT 76 88 76 28 76 100
FIGURE 3 Size frequency distribution
for each carangid species collected using
MOC (red) and LMT (blue) gear types
MOHAN ET AL. |7

displayed increased abundance at high SSHA (>10 cm) and low salin-
ities (<32) for MOC; however, the effect of temperature was not
clear (Figure 6). For LMT, higher abundance ofS. setapinniswas
related to increased SST (>29.5°C) and decreased salinity with a
dome-shaped peak at salinity=29 (Figure 6). Salinity was the only
factor significantly related toC. crysosabundance and the
Cruise Species N Minimum Maximum Median Mean "SD
MOC Selene setapinnis 248 3 43 9 9.54 "3.98
Caranx crysos 231 4 55 11 12.91 "6.55
Caranx hippos 31 4 20 8 8.77 "2.98
Chloroscombrus chrysurus73 6 54 17 19.27 "8.97
Trachurus lathami 13 6 24 16 14.54 "4.86
Combined 596 3 55 10 12.11 "6.62
LMT Selene setapinnis 1,104 9 66 20 22.15 "8.56
Caranx crysos 949 10 142 26 32.27 "21.6
Caranx hippos 605 11 92 21 21.31 "6.79
Chloroscombrus chrysurus142 13 67 27 29.21 "10.26
Trachurus lathami 298 17 149 35 38.88 "18.89
Combined 3,098 9 149 23 27.02 "15.94
N, sample size of measured fish;SD, standard deviation.
Size-frequency distribution presented in Figure 3.
TABLE 4 Summary of standard length
(SL) measurements (mm) for each carangid
species collected using MOC of LMT gear
(a) (b)
(c) (d)
(e) (f)
(g) (h)
(i) (j)
FIGURE 4 Density distribution heat
maps (ind/m
3
10
!5
) forSelene setapinnis
(a, b),Caranx crysos(c, d),Caranx hippos
(e, f),Chloroscombrus chrysurus(g, h) and
Trachurus lathami(i, j) collected during the
MOC sampling in the spring/summer and
summer/fall. Colors represent DIVA
gridding in ODV; note difference in sample
size (N) and scale bar for each plot
8|
MOHAN ET AL.

relationship shifted between MOC and LMT, with a peak at 27 for
MOC and a peak at 32 for LMT (Figure 7). For MOC, higher
C. crysosabundance occurred at salinities of 26–28, and for LMT
there were higher abundances at salinities 30–34 (Figure 7). SST and
salinity were significantly related toC. hipposabundance for both
MOC and LMT, however, the shape and direction of the relation-
ships differed between the gears (Figure 8). For MOC, higherC. hip-
posabundance was related to decreased SST (25–28°C) and
(a) (b)
(c) (d)
(e) (f)
(g) (h)
(i) (j)
FIGURE 5 Density distribution heat
maps (ind/m
3
10
!7
) forSelene setapinnis(a,
b),Caranx crysos(c, d),Caranx hippos(e, f),
Chloroscombrus chrysurus(g, h) and
Trachurus lathami(i, j) collected during the
LMT sampling in the summer and fall.
Colors represent DIVA gridding in ODV;
note difference in sample size (N) and scale
bar for each plot
TABLE 5 Generalized Additive Model
(GAM) results demonstrating the influence
of season and physical factors on the five
most abundant carangid species collected
during MOC and LMT sampling
Cruise Species
Factor
DE (%)Season SSHA SST Salinity
MOC Selene setapinnis <0.0001 0.0005 <0.0001 <0.0001 69
Caranx crysos <0.0001 0.2191 0.0516 0.0092 53
Caranx hippos 0.0007 0.0471 0.0092 0.0086 53
Chloroscombrus chrysurus/ 0.0001 <0.0001 <0.0001 96
Trachurus lathami 0.0007 0.1926 0.0064 0.5338 51
LMT Selene setapinnis <0.0001 0.1840 0.0000 <0.0001 63
Caranx crysos 0.0001 0.0926 0.0745 0.0099 45
Caranx hippos 0.0343 0.6130 0.0000 <0.0001 51
Chloroscombrus chrysurus0.2950 0.2322 0.1142 0.0322 95
Trachurus lathami 0.0784 0.0002 0.0002 <0.0001 71
Significant variables (p<.05) in bold and percent deviance explained (DE) for each model is
presented.
MOHAN ET AL. |9

increased salinity (>35) whereas for LMT there were more abundant
C. hipposat increased SST (>29.5) and decreased salinity (26–32)
(Figure 8).Chloroscombrus chrysuruscollected with MOC exhibited
increased abundance at decreased SSHA (<5 cm), decreased salinity
(26–30), and increased SST (>28°C), whereasC. chysuruscollected
with LMT was more abundant at low salinities (<35) (Figure 9). For
both MOC and LMT, increasedT. lathamiabundance was signifi-
cantly related to increased SST (Figure 10). For LMT, the response
ofT. lathamito salinity was variable, with increased abundance
observed at moderate (27–30) and higher (>35) salinities (Figure 10).
The abundance ofT. lathamiwas increased at moderate SSHA
(0–20 cm) for LMT (Figure 10).
Environmental variables accounted for 27% and 20.1% of the
variation in species for MOC and LMT, respectively (Figure 11). The
RDA plots demonstrated thatS. setapinnis, C. crysosandC. chrysurus
all responded similarly with increased abundance at low salinities for
both post-larval (Figure 11a) and juvenile (Figure 11b) life stages.
However, the post-larval stage ofS. setapinnis, C. crysosandC. chry-
suruswere more abundant at higher temperatures, whereas the juve-
nile stages were more abundant at lower SSHA (Figure 11).Caranx
hipposandT. lathamidisplayed similar spatial arrangement in RDA
plots for both post-larval and juvenile life stages, with higher abun-
dances at higher SST (Figure 11a and b).
4|DISCUSSION
The northern GoM in 2011 displayed sharp gradients of SST, salinity,
and SSHA, which affected the abundance and distribution of post-lar-
val and juvenile carangids. Lower salinities that are characteristic of
the Mississippi River plume tended to result in increased abundance of
most carangids, but responses to SST and SSHA were species- and
life-stage specific. The seasonal sampling that encompassed the spring,
summer and fall captured many gradients in oceanographic conditions
that represent dominant mesoscale features in the GoM including the
river plume, warm core eddies and/or the extension of the LC, and
frontal regions where the eddies and the plume intersect. Additionally,
the use of two gear types with different mesh size and tow speeds
allowed comparison between abundance and distribution of post-lar-
val and juvenile carangid species life stages.
FIGURE 6 Response plots forSelene
setapinnisabundance in relation to sea
surface temperature (SST), sea surface
height anomaly (SSHA) and salinity
determined from GAM models for fish
collected using MOC and LMT. Non-
significant variables not plotted
FIGURE 7 Response plots forCaranx
crysosabundance in relation to sea surface
temperature (SST), sea surface height
anolomy (SSHA) and salinity determined
from GAM models for fish collected using
MOC and LMT. Non-significant variables
not plotted
10|
MOHAN ET AL.

Many of the studies that have examined the influence of abi-
otic factors on the distribution and abundance of marine fishes in
the GoM have focused on larval life stages (Kitchens & Rooker,
2014; Randall, Smith, Cowan, & Rooker, 2015; Rooker et al.,
2012). Larvae exhibit limited mobility and are easy to collect in
towed nets. Additionally, collecting larvae allows inference on
spawning location and season (Ditty et al., 2004; Kitchens & Roo-
ker, 2014; Rooker et al., 2012; Shaw & Drullinger, 1990). Fewer
studies have focused on juvenile and sub-adult life stages of mar-
ine fishes, potentially owing to the high mobility of juvenile life
stages and net avoidance behavior (Leak, 1981). Smaller carangids
were collected during MOC sampling with small mesh size (3 mm)
resulting in a median fish SL of 10 mm, a size suggesting these
fishes were post-larvae in a transitory phase between larvae and
juvenile life stages (Aprieto, 1974). Larger juvenile and sub-adult
fishes with a median SL of 23 mm were collected in LMT that
utilized the larger mesh size (51 mm) with a 16.59greater effec-
tive mouth area; however, both net styles did collect some caran-
gids that were larger than 50 mm, but these subadult/adult
specimens were rare. Larger carangids were collected in the LMT
owing to faster tow speeds of 2.6 m/s compared to the MOC,
which towed at speeds of 0.8 m/s. Leis, Hay, Clark, Chen, and
Shao (2006) calculated in situ swimming speeds in larvae and early
juvenile (8–18 mm SL) of a related carangid, the Giant Trevally
Caranx ignobilis,and determined swimming speeds ranged from 2
to 20 cm/s that was linearly related to SL. Therefore, applying the
linear size-to-swimming speed relationship of Leis et al. (2006), a
carangid juvenile of SL=50 mm could swim approximately
70 cm/s, or near the tow speed of the MOC net and a 100-mm
juvenile would approach speeds of 140 cm/s. Larger carangids
most likely escaped net capture by swimming faster than the net
tow speed or moving vertically or horizontally in the water col-
umn to avoid the approach of the net (Misund, Luyeye, Coetzee,
& Boyer, 1999). Although larval and juvenile carangids prefer pela-
gic habitats, some adult carangids prefer benthic habitats (Clarke
& Aeby, 1998) and thus adults may not have been targeted by
the gear types used here. Most of the carangid species examined
here spawn in neritic coastal waters, where most previous surveys
have focused sampling effort in shallow water<100 m deep (Espi-
nosa-Fuentes & Flores-Coto, 2004; Katsuragawa & Ekau, 2003;
Leak, 1981; Shaw & Drullinger, 1990). In contrast, samples in this
study were collected far offshore in depths ranging from 500 to
3,000 m. Thus, the gear types used here and areas sampled most
likely captured late stage larvae to early/late juveniles that were
FIGURE 8 Response plots forCaranx
hipposabundance in relation to sea surface
temperature (SST), sea surface height
anolomy (SSHA) and salinity determined
from GAM models for fish collected using
MOC and LMT. Non-significant variables
not plotted
FIGURE 9 Response plots for
Chloroscombrus chrysosabundance in
relation to sea surface temperature (SST),
sea surface height anolomy (SSHA) and
salinity determined from GAM models for
fish collected using MOC and LMT. Non-
significant variables not plotted
MOHAN ET AL. |11

either passively entrained in circulation patterns of the expansive
river plume (Grimes & Finucane, 1991; Johns et al., 2014) or
actively engaging in ontogenetic migrations from nearshore to
offshore habitats (da Costa, Albieri, & Ara"ujo, 2005) or aggregating
in the hydrodynamic nutrient rich and productive frontal regions
(Ditty et al., 2004; Raya & Sabates, 2015).
4.1|Abundance ranking among species
The two most abundant species collected in this study were the
S. setapinnis(34%) andC. crysos(30%), which together comprised
64% of all species collected. Few studies have reported a high abun-
dance ofS. setapinnis, ranging from 0.2% to 2% of collections in the
southern Atlantic off the Brazilian coast (Campos, De Castro, &
Bonecker, 2010; de Souza & Junior, 2008). Flores-Coto and San-
chez-Ramirez (1989) foundS. setapinniscomprised 6.1% of collec-
tions and were most abundant in warmer months in the southern
GoM, which was the highest reported abundance ofS. setapinnis
before this study. This is in contrast with results of other studies in
the GoM, which have typically foundC. chrysurusto be the domi-
nant species (Ditty et al., 2004; Flores-Coto & Sanchez-Ramirez,
1989). da Costa et al. (2005) examined carangid distributions in a
semi-enclosed bay in southeastern Brazil and foundC. chrysurus
abundance and biomass was significantly related to decreased salin-
ity and shallow water depths. This relationship was the result of a
high number of juveniles (30–90 mm total length) collected from the
inner bay which exhibited increased water temperature, low water
clarity, and high organic loads that supported increased primary pro-
duction and upper trophic levels (da Costa et al., 2005). Ditty et al.
(2004) sampled carangid larvae in the northern GoM and reported
abundance rankings of:C. chrysurus83%;Decapterus punctatus9%;
C. hippos2.9%;C. crysos1.9%.Chloroscombrus chrysuruswas most
abundant west of the Mississippi River, whereasD. punctatuswas
most abundant in the eastern GoM on the Florida Shelf.Caranx hip-
posandC. crysoshad similar spatial overlap, but different temporal
distributions withC. hipposmore abundant in May-June whereas
C. crysosoccurred more frequently in June to August (Ditty et al.,
2004). Interestingly, we found different salinity preferences for both
post-larval and juvenileC. crysosandC. hippos, suggesting spatial
separation and habitat partitioning between these species. Leak
FIGURE 10 Response plots for
Trachurus lathamiabundance in relation to
sea surface temperature (SST), sea surface
height anolomy (SSHA) and salinity
determined from GAM models for fish
collected using MOC and LMT. Non-
significant variables not plotted
FIGURE 11 Redundancy analysis (RDA) plots exploring
relationships between environmental factors (red arrows) and
species abundance (weighted average=triangles) for MOC (a) and
LMT (b) collected samples
12|
MOHAN ET AL.

(1981) sampled carangid larvae in the eastern GoM during 4 years
and found thatD. punctatuswere over 109more abundant than all
other carangids. Shaw and Drullinger (1990) sampled carangid larvae
in coastal waters of Louisiana during 1982–83 and foundC. chrysu-
ruswas most abundant followed by C. crysos,T. lathami,and
D. punctatusin order of abundance. In the southern GoM below
21°N, Flores-Coto and Sanchez-Ramirez (1989) examined seasonal
carangid densities in 1983–84 and described ranked abundance of
the same species examined here:C. chrysurus54%;D. punctatus
16%;T. lathami12%;S. setapinnis6%;C. hippos0.9% andC. crysos
0.7%. Larvae of these carangids were present year-round, except for
T. lathami, which was only present in the winter and spring. Several
other studies have been conducted off the Brazilian coast and have
found different patterns of carangid abundance (Campos et al.,
2010; de Souza & Junior, 2008; Katsuragawa & Matsuura, 1992).
Katsuragawa and Matsuura (1992) reported abundances ofT. lathami
59%;C. chrysurus15%;D. punctatus12%, whereas Campos et al.
(2010) reported abundances ofDecapterus puntatus57%;C. chrysu-
rus17%;C. crysos8%,T. lathami6% andS. setapinnis0.2%, which
was similar to the findings of de Souza and Junior (2008). The pri-
mary difference between previous work and this study was the high
density of theS. setapinnisexhibited here, and the lack ofD. punta-
tusthat is typically a highly abundant carangid in the GoM and
South Atlantic Ocean.Decapterus puntatusspawn year-round in the
eastern GoM and display more intense spawning at higher tempera-
tures (26–32°C) and increased salinities (36–37) and perhaps were
less abundant in the low salinity conditions of 2011 compared to
other carangid species (Leak, 1981).Decapterus puntatusare also
more concentrated in the eastern GoM on the Florida shelf (Ditty
et al., 2004; Leak, 1981) in an area that was not sampled in this
study. Shaw and Drullinger (1990) sampled carangid larvae in coastal
waters of Louisiana and foundT. lathamiwas restricted to deeper
depths (the mean depth range 221–2,768 m) and high salinities
(mean salinity 36). Spawning ofT. lathamiis known to be associated
with“high amplitude event or gradients”and larvae have been col-
lected from turbulent mixed water between the river plume and
oceanic waters (Shaw & Drullinger, 1990). AlthoughT. lathamiwere
found offshore in agreement with other studies,T. lathamipresence
in warmer waters has not been reported previously.
4.2|Inferred spawning seasons and habitats
The MOC collected post-larval carangids, which displayed consistent
and narrow size ranges indicated by median SL from 8 to 17 mm.
Most of the post-larval carangids displayed low abundance from the
spring/summer samples collected in April and June, except for
T. lathamiwhich had comparable abundances between the seasons
but different distributions.Trachurus lathamiis the only species
thought to spawn in the winter, whereas all the other species spawn
in the summer, which would explain the higher abundance ofS. se-
tapinnis, C. crysos, and C. chrysurusin the summer and fall seasons
(Ditty et al., 2004; Leak, 1981; Shaw & Drullinger, 1990). The post-
larval abundances of three species (S. setapinnis, C. chrysurus, and
T. lathami) displayed complex relationships with SST. In general,
there were higher abundances at increased temperatures (also evi-
dent in the RDA plot), but there was high variability in the GAM
response plots resulting in funnel shaped curves at lower tempera-
tures. In contrast,C. hippospost-larvae displayed a distinct peak at
lower temperatures (25–28°C) and low abundance at high tempera-
tures.Caranx hipposwas also the only species that demonstrated
increased abundance at high salinities>35. These results suggest a
separation of spawning habitat betweenC. hipposand the other car-
angids examined here.Caranx hipposwas also the only species that
displayed contrasting relationships between abundance and SST and
salinity between post-larvae and juveniles. Additionally, the non-
overlapping spatial distributions and different shapes of the salinity
response plot among species may suggest a temporal or spatial suc-
cession of spawning events to reduce inter-species competition for
resources (Raya & Sabates, 2015). For instance,S. setapinnisdis-
played a linear salinity response plot, whereasC. crysosandC. chry-
suruswere domed shaped andC. hipposwas linear but negative.
Species-specific responses of post-larvae abundance to SSHA were
also detected.Chloroscombrus chrysurusabundance was increased at
lower SSHA, whereasS. setapinniswas more abundant at increased
SSHA, providing further evidence of habitat partitioning between
species. A study of billfish spawning habitats in the northern GoM
found higher densities of sailfish and swordfish larvae at low SSHA
(<10 cm), but blue marlin larvae displayed increased density at high
(>20 cm) and low SSHA (<!5 cm) (Rooker et al., 2012). Randall et al.
(2015) reported increased bluntnose flyingfishPrognichthys occiden-
talislarvae at low SSHA (<0 cm) and high salinity (>35) and sug-
gested the expansive river plume in 2011 may have decreased
suitable spawning habitat forP. occidentalis, in contrast to our find-
ings for carangids.
The commonalities and differences in species-environment rela-
tionships were also exhibited in the RDA ordination plots, where
S. setapinnis, C. crysos, andC. chrysuruswere correlated positively
with SST and SSHA, and negatively correlated with salinity whereas
C. hipposwas opposite of those three species. Other studies have
demonstrated that river plume features can be characterized by low
salinity and increased temperature (Johns et al., 2014). The higher
abundance ofS. setapinnis, C. crysos, andC. chrysuruspost larvae at
lower salinities and warmer SST suggests association with the river
plume that could either result from passive entrainment of small
buoyant larvae due to hydrodynamic convergence zones (Bakun,
2006; Govoni, Hoss, & Colby, 1989) or active seeking out of plume
waters for feeding (Govoni & Chester, 1990). Enhanced larval feed-
ing may result from the photic environment of the plume, where
increased suspended sediments may increase the visual contrast and
overall diversity of prey types (Govoni & Chester, 1990). Dagg and
Whitledge (1991) reported strong seasonality of zooplankton pro-
duction in the MS River plume, with highest production occurring in
the summer which concurs with our results of higher larval concen-
trations in the summer/fall compared to the spring/summer MOC
cruises. Increased abundance of larval fishes at productive frontal
zones would also attract predators, which could explain the higher
MOHAN ET AL. |13

occurrence of juvenile and larger carangid collected with LMT in the
plume and or frontal zones. Thus, perhaps some larger sized caran-
gids (>40 mm) were inhabiting the river plume to forage on other
smaller larval fish that may be entrained passively in the plume.
Juvenile carangids collected with LMT experienced a much nar-
rower SST range (27.7–30.5°C) compared with the larval MOC SST
range (24.4–31.8), which may explain the consistent linear relation-
ship between juvenile abundance and SST forS. setapinnis,C. hippos
andT. lathami. However, SST was not significantly related toC. chry-
surusandC. crysosjuvenile abundance. Interestingly, on the RDA
plot onlyC. hipposand to a lesser extentT. lathamishowed direc-
tional ordination with SST. Similar to the post-larval carangids, higher
abundances of juvenile carangids were generally found at lower
salinities and salinity was a significant variable in the GAM models
for all juvenile (LMT) carangids. However, the GAM response plots
were linear forC. chrysurus, dome-shaped forS. setapinnisand
C. crysos, and S-shaped forC. hipposandT. lathami, suggesting
increased abundance at both medium and high salinities. This differ-
ence was also apparent in the RDA plot, whereC. hipposand
T. lathamipointed in different directions than the other species, sug-
gesting habitat partitioning between the species. The abundance
heat map also identified a unique seasonal pattern forC. hipposand
T. lathami(and to a lesser extentS. setapinnis) where the distribution
shifted from the northern region in the summer to more western
regions in the fall. Perhaps the westerly shifts in abundance were
related to the westerly moving warm core eddy/LC extension that
produced a frontal region with increased production and food avail-
ability at the intersection of the mesoscale eddy and river plume. In
contrast, to the post-larval carangids that may have been passively
entrained in the frontal convergence zones, it is likely the larger
juveniles targeted frontal zones for increased feeding (Bakun, 2006).
Additionally, the schooling behavior of carangid juveniles may have
resulted in patchy concentrated zones (Kwei, 1978) that were evi-
dent from highly localized abundance exhibited on the species heat
maps.
5|CONCLUSION
Relationships between carangid abundance and physical oceano-
graphic features were examined in the northern GoM in 2011, when
the Mississippi River experienced record flooding. MOC and LMT
gear types were used to collect fish and both in situ CTD and satel-
lite measurements were used to characterize physical conditions and
mesoscale features. SST, salinity, and SSHA, were related to carangid
density and varied between species as a product of differences in
life history strategies between post-larval and juveniles. The large
expansion of the Mississippi River plume in the record-flooding year,
created frontal zones with dynamic salinity and temperature regimes
that may have passively entrained post-larval carangids or aggre-
gated foraging juveniles. Additional future studies may focus on
growth measurements via otolith microstructure analyses and dietary
analysis with stomach contents and tissue stable isotope analyses
(Syahailatua, Taylor, & Suthers, 2011) to examine potential resource
partitioning between species over multiple years.
ACKNOWLEDGEMENTS
This manuscript includes both work that was conducted and samples
that were collected as part of the Deepwater Horizon Natural
Resource Damage Assessment being conducted cooperatively among
academic partners, NOAA, other Federal and State Trustees, and BP.
This research was made possible in part by a grant from The Gulf of
Mexico Research Initiative. Data are publicly available through the
Gulf of Mexico Research Initiative Information & Data Cooperative
(GRIIDC) at https://data.gulfresearchinitiative.org (doi: 10.7266/
N7VX0DK2, 10.7266/N7R49NTN). The findings and conclusions in
this manuscript are those of the authors and do not necessarily rep-
resent the view of the National Oceanic and Atmospheric Administra-
tion or of any other natural resource trustee for the BP/Deepwater
Horizon Natural Resource Damage Assessment. This is contribution
#35 from the Marine Education and Research Center in the Institute
for Water and Environment at Florida International University.
REFERENCES
Aprieto, V. L. (1974). Early development of five Carangid fishes of the
Gulf of Mexico and the south Atlantic coast on the United States.
Fishery Bulletin,72, 415–443.
Bakun, A. (2006). Fronts and eddies as key structures in the habitat of
marine fish larvae: Opportunity, adaptive response and competitive
advantage.Scientia Marina,70(S2), 105–122.
Biggs, D. (1992). Nutrients, plankton, and productivity in a awarm-core
ring in the western Gulf of Mexico.Journal of Geophysical Research,
97, 160–163.
Biggs, D. C., & M€uller-Karger, F. E. (1994). Ship and satellite observations
of chlorophyll stocks in interacting cyclone-anticyclone eddy pairs in
the western Gulf of Mexico.Journal of Geophysical Research: Oceans,
99, 7371–7384.
Campos, P. N., De Castro, M. S., & Bonecker, A. C. T. (2010). Occurrence
and distribution of Carangidae larvae (Teleostei, Perciformes) from
the Southwest Atlantic Ocean, Brazil (12-23oS).Journal of Applied
Ichthyology,26, 920–924.
Chesney, E., Baltz, D., & Thomas, R. (2000). Louisiana estuarine and
coastal fisheries and habitats: Perspectives from a fish’s eye view.
Ecological Applications,10, 350–366.
Clarke, T. A., & Aeby, G. S. (1998). The use of small mid-water attraction
devices for investigation of the pelagic juveniles of carangid fishes in
Kaneohe Bay, Hawaii.Bulletin of Marine Science,62, 947–955.
da Costa, M. R., Albieri, R. J., & Ara"ujo, F. G. (2005). Size distribuition of
the jackChloroscombrus chrysurus(Linnaeus) (Actinopterygii, Carangi-
dae) in a tropical bay at southeastern Brazil.Revista Brasileira de
Zoologia,22, 580–586.
Dagg, M. J., & Breed, G. A. (2003). Biological effects of Mississippi River
nitrogen on the northern Gulf of Mexico–A review and synthesis.
Journal of Marine Systems,43, 133–152.
Dagg, M. J., & Whitledge, T. E. (1991). Concentrations of copepod nauplii
associated with the nutrient-rich plume of the Mississippi River.Con-
tinental Shelf Research,11, 1409–1423.
de Souza, C. S., & Junior, P. M. (2008). Distribution and abundance of
Carangidae (Teleostei, Perciformes) associated with oceanographic
factors along the Northeast Brazilian Exclusive Economic Zone.Brazil-
ian Archives of Biology and Technology,51, 1267–1278.
14|
MOHAN ET AL.

Del Castillo, C. E., Coble, P. G., Conmy, R. N., M€uller-Karger, F. E., Van-
derbloemen, L., & Vargo, G. A. (2001). Multispectral in situ measure-
ments of organic matter and chlorophyll fluorescence in seawater:
Documenting the intrusion of the Mississippi River plume in the
West Florida Shelf.Limnology and Oceanography,46, 1836–1843.
Ditty, J. G., Shaw, R. F., & Cope, J. S. (2004). Distribution of carangid lar-
vae (Teleostei: Carangidae) and concentrations of zooplankton in the
northern Gulf of Mexico, with illustrations of earlyHemicaranx
amblyrhynchusandCaranxspp. larvae.Marine Biology,145, 1001–
1014.
Drexler, M., & Ainsworth, C. H. (2013). Generalized additive models used
to predict species abundance in the Gulf of Mexico: An ecosystem
modeling tool.PLoS One,8, e64458.
Espinosa-Fuentes, M. L., & Flores-Coto, C. (2004). Cross-shelf and verti-
cal structure of ichthyoplankton assemblages in continental shelf
waters of the Southern Gulf of Mexico.Estuarine, Coastal and Shelf
Science,59, 333–352.
Falcini, F., Khan, N. S., Macelloni, L., Horton, B. P., Lutken, C. B., McKee,
K. L....Jerolmack, D. J. (2012). Linking the historic 2011 Mississippi
River flood to coastal wetland sedimentation.Nature Geoscience,5,
803–807.
Flores-Coto, C., & Sanchez-Ramirez, M. (1989). Larval distribution and
abundance of Carangidae (Pisces), from the southern Gulf of Mexico.
1983–1984.Gulf and Caribbean Research,8, 117–128.
Gierach, M. M., Vazquez-Cuervo, J., Lee, T., & Tsontos, V. M. (2013).
Aquarius and SMOS detect effects of an extreme Mississippi River
flooding event in the Gulf of Mexico.Geophysical Research Letters,
40, 5188–5193.
Godø, O. R., Samuelsen, A., Macaulay, G. J., et al. (2012). Mesoscale
eddies are oases for higher trophic marine life.PLoS One,7, e30161.
Govoni, J., & Chester, A. (1990). Diet composition of larvalLeistomus xan-
thurusin and about the Mississippi River plume.Journal of Plankton
Research,12, 819–830.
Govoni, J. J., Hoss, D. E., & Colby, D. R. (1989). The spatial distribution
of larval fishes about the Mississippi River plume.Limnology and
Oceanography,34, 178–187.
Grimes, C. (2001). Fishery production and the Mississippi River discharge.
Fisheries,26, 17–26.
Grimes, C., & Finucane, J. (1991). Spatial distribution and abundance of
larval and juvenile fish, chlorophyll and macrozooplankton around the
Mississippi River discharge plume, and the role of the plume in fish
recruitment.Marine Ecology Progress Series,75, 109–119.
Hamilton, P. (1992). Lower continental slope cyclonic eddies in the cen-
tral Gulf of Mexico.Journal of Geophysical Research,97, 2185–2220.
Hastie, T. J., & Tibshirani, R. (1986). Generalized additive models.Statisti-
cal Science,1, 297–318.
Johns, E. M., Muhling, B. A., Perez, R. C., et al. (2014). Amazon River
water in the northeastern Caribbean Sea and its effect on larval reef
fish assemblages during April 2009.Fisheries Oceanography,23, 472–
494.
Judkins, H., Vecchione, M., Cook, A., & Sutton, T. (2016). Diversity of
midwater cephalopods in the northern Gulf of Mexico: Comparison
of two collecting methods.Marine Biodiversity,1–11. http://link.
springer.com/article/10.1007/s12526-016-0597-8.
Katsuragawa, M., & Ekau, W. (2003). Distribution, growth and mortality
of young rough scad,Trachurus lathami, in the south-eastern Brazilian
Bight.Journal of Applied Ichthyology,19, 21–28.
Katsuragawa, M., & Matsuura, Y. (1992). Distribution and abundance of
carangid larvae in the southeastern Brazilian Bight during 1975–
1981.Boletim do Instituto Oceanogr!afico,40, 55–78.
Kiszka, J. J., M"endez-Fernandez, P., Heithaus, M. R., & Ridoux, V. (2014).
The foraging ecology of coastal bottlenose dolphins based on stable
isotope mixing models and behavioural sampling.Marine Biology,161,
953–961.
Kitchens, L. L., & Rooker, J. R. (2014). Habitat associations of dolphinfish
larvae in the Gulf of Mexico.Fisheries Oceanography,23, 460–471.
Kwei, E. A. (1978). Food and spawning activity ofCaranx hipposoff the
coast of Ghana.Journal of Natural History,12, 195–215.
Leak, J. C. (1981). Distribution and abundance of Carangid fish larvae in
the Eastern Gulf of Mexico, 1971–1974.Biological Oceanography,1,
1–28.
Lehmann, A., Overton, J. M. C., & Leathwick, J. R. (2002). GRASP: Gener-
alized regression analysis and spatial prediction.Ecological Modelling,
157, 189–207.
Leis, J. M., Hay, A. C., Clark, D. L., Chen, I. S., & Shao, K. T. (2006).
Behavioral ontogeny in larvae and early juveniles of the giant trevally
(Caranx ignobilis) (Pisces: Carangidae).Fishery Bulletin,104, 401–414.
Lima, I. D., Olson, D. B., & Doney, S. C. (2002). Biological response to
frontal dynamics and mesoscale variability in oligotrophic environ-
ments: Biological production and community structure.Journal of
Geophysical Research,107,1–21.
Lindo-Atichati, D., Bringas, F., Goni, G., Muhling, B., Muller-Karger, F. E.,
& Habtes, S. (2012). Varying mesoscale structures influence larval fish
distribution in the northern Gulf of Mexico.Marine Ecology Progress
Series,463, 245–257.
Lohrenz, S. E., Fahnenstiel, G. L., Redalje, D. G., Lang, G. A., Dagg, M. J.,
Whitledge, T. E., & Dortch, Q. (1999). Nutrients, irradiance, and mix-
ing as factors regulation primary production in coastal waters
impacted by the Mississippi River plume.Continental Shelf Research,
19, 1113–1141.
Meekan, M. G., Carleton, J. H., Steinberg, C. R., et al. (2006). Turbulent
mixing and mesoscale distributions of late-stage fish larvae on the
NW Shelf of Western Australia.Fisheries Oceanography,15, 44–59.
Misund, O., Luyeye, N., Coetzee, J., & Boyer, D. (1999). Trawl sampling
of small pelagic fish off Angola: Effects of avoidance, towing speed,
tow duration, and time of day.ICES Journal of Marine Science,56,
275–283.
Randall, L. L., Smith, B. L., Cowan, J. H., & Rooker, J. R. (2015). Habitat
characteristics of bluntnose flyingfishPrognichthys occidentalis
(Actinopterygii, Exocoetidae), across mesoscale features in the Gulf
of Mexico.Hydrobiologia,749, 97–111.
Raya, V., & Sabates, A. (2015). Diversity and distribution of early life
stages of carangid fishes in the northwestern Mediterranean:
Responses to environmental drivers.Fisheries Oceanography,24, 118–
134.
Roberts, J. J., Best, B. D., Dunn, D. C., Treml, E. A., & Halpin, P. N.
(2010). Marine Geospatial Ecology Tools: An integrated framework
for ecological geoprocessing with ArcGIS, Python, R, MATLAB, and
C++.Environmental Modelling and Software,25, 1197–1207.
Rooker, J. R., Kitchens, L. L., Dance, M. A., Wells, R. J. D., Falterman, B.,
& Cornic, M. (2013). Spatial, temporal, and habitat-related variation in
abundance of pelagic fishes in the Gulf of Mexico: Potential implica-
tions of the Deepwater Horizon Oil Spill.PLoS One,8, e76080.
Rooker, J. R., Simms, J. R., David Wells, R. J., Holt, S. A., Holt, G. J.,
Graves, J. E., & Furey, N. B. (2012). Distribution and habitat associa-
tions of billfish and swordfish larvae across mesoscale features in the
Gulf of Mexico.PLoS One,7, e34180.
Seki, M. P., Polovina, J. J., Brainard, R. E., Bidigare, R. R., Leonard, C. L.,
& Foley, D. G. (2001). Biological enhancement at cyclonic eddies
tracked with GOES thermal imagery in Hawaiian waters.Geophysical
Research Letters,28, 1583–1586.
Shaw, R. F., & Drullinger, D. L. (1990).Early-Life-History profiles, seasonal
abundance, and distribution of four Species of Carangid larvae from the
Northern Gulf of Mexico, 1982 and 1983, NOAA NMFS Technical
Report No. 89. 37p.
Shimose, T., & Wells, R. (2015). Feeding ecology of bluefin tunas. In T.
Kitagawa, & S. Kumura (Eds.),Biology and ecology of bluefin tunas(pp.
78–97). Boca Raton, FL: CRC Press, Taylor and Francis Group.
MOHAN ET AL. |15

Syahailatua, A., Taylor, M. D., & Suthers, I. M. (2011). Growth variability
and stable isotope composition of two larval carangid fishes in the
East Australian Current: The role of upwelling in the separation zone.
Deep-Sea Research Part II: Topical Studies in Oceanography,58, 691–
698.
Thorrold, S., & McKinnon, A. (1995). Response of larval fish assemblages
to a riverine plume in coastal waters of the central great barrier reef
lagoon.Limnology and Oceanography,40, 177–181.
Toner, M., Kirwan, A., Poje, A., Kantha, L., Muller-Karger, F., & Jones, C.
(2003). Chlorophyll dispersal by eddy-eddy interactions in the Gulf of
Mexico.Journal of Geophysical Research,108,1–12.
Torres-Rojas, Y. E., Hern"andez-Herrera, A., Galv"an-Maga~na, F., & Ala-
torre-Ram"ırez, V. G. (2010). Stomach content analysis of juvenile,
scalloped hammerhead sharkSphyrna lewinicaptured off the coast of
Mazatlan, Mexico.Aquatic Ecology,44, 301–308.
Troupin, C., Barth, A., Sirjacobs, D., et al. (2012). Generation of analysis
and consistent error fields using the Data Interpolating Variational
Analysis (DIVA).Ocean Modelling,52–53, 90–101.
Vukovich, F. M. (2007). Climatology of ocean features in the Gulf of
Mexico using satellite remote sensing data.Journal of Physical
Oceanography,37, 689–707.
Vukovich, F. M., & Crissman, B. W. (1986). Aspects of warm rings in the
Gulf of Mexico.Journal of Geophysical Research,91, 2645–2660.
Walker, N. D., Wiseman, W. J., Rouse, L. J., & Babin, A. (2005). Effects of
river discharge, wind stress, and slope eddies on circulation and the
satellite-observed structure of the Mississippi River Plume.Journal of
Coastal Research,216, 1228–1244.
Wiebe, P., Morton, A., Bradley, A., et al. (1985). New developments in
the MOC, an apparatus for sampling zooplankton and micronekton.
Marine Biology,87, 313–323.
Williams, A. K., Mcinnes, A. S., Rooker, J. R., & Quigg, A. (2015). Changes in
microbial plankton assemblages induced by mesoscale oceanographic
features in the northern Gulf of Mexico.PLoS One,10(9), e0138230.
doi:10.1371/journal.pone.0138230.
Zimmerman, R. A., & Biggs, D. C. (1999). Patterns of distribution of
sound-scattering zooplankton in warm- and cold-core eddies in the
Gulf of Mexico, from a narrowband acoustic Doppler current profiler
survey.Journal of Geophysical Research,104, 5251–5262.
SUPPORTING INFORMATION
Additional Supporting Information may be found online in the sup-
porting information tab for this article.
How to cite this article:Mohan JA, Sutton TT, Cook AB,
Boswell KM, David Wells RJ. Influence of oceanographic
conditions on abundance and distribution of post-larval and
juvenile carangid fishes in the northern Gulf of Mexico.Fish
Oceanogr. 2017;00:1–16. https://doi.org/10.1111/fog.12214
16|
MOHAN ET AL.