lable at ScienceDirect

Estuarine, Coastal and Shelf Science 170 (2016) 134e144

Contents lists avai

Estuarine, Coastal and Shelf Science

journal homepage: www.elsevier .com/locate/ecss

Rearing conditions and habitat use of white seabass (Atractoscionnobilis) in the northeastern Pacific based on otolith isotopiccomposition

Alfonsina E. Romo-Curiel a, Sharon Z. Herzka a, *, Chugey A. Sepulveda b,Paula P�erez-Brunius c, Scott A. Aalbers b

a Departamento Oceanografía Biol�ogica, Centro de Investigaci�on Científica y de Educaci�on Superior de Ensenada (CICESE), No 3918 Carretera Tijuana-Ensenada, Ensenada, Baja California, C.P. 22860, M�exicob Pfleger Institute of Environmental Research (PIER), 2110 South Coast Highway Suite F, Oceanside, CA, 92054, USAc Departamento Oceanografía Física, Centro de Investigaci�on Científica y de Educaci�on Superior de Ensenada (CICESE), No 3918 Carretera Tijuana-Ensenada, Ensenada, Baja California, C.P. 22860, M�exico

a r t i c l e i n f o

Article history:Received 14 May 2015Received in revised form17 November 2015Accepted 9 January 2016Available online 13 January 2016

Keywords:White seabassAtractoscion nobilisOtolith aragoniteIsotopic analysisHabitat use

* Corresponding author.E-mail addresses: aeromo@cicese.edu.mx (A.E. R

mx (S.Z. Herzka), chugey@pier.org (C.A. Sepulveda),Brunius), scott@pier.org (S.A. Aalbers).

http://dx.doi.org/10.1016/j.ecss.2016.01.0160272-7714/© 2016 Elsevier Ltd. All rights reserved.

a b s t r a c t

White seabass, Atractoscion nobilis, is an important coastal resource throughout both California and BajaCalifornia, but whether this species comprises a single or multiple subpopulations in the northeasternPacific is not known. The aim of this study was to infer larval rearing habitats and population structure ofwhite seabass by sampling adults from three regions spanning a latitudinal temperature gradient and adistance of over 1000 km, and analyzing the isotopic composition (d18O and d13C) of otolith aragonitecorresponding to the larval, juvenile and adult stages. Otolith cores revealed high isotopic variability andno significant differences among regions, suggesting overlapping rearing conditions during the larvalstage, the potential for long distance dispersal or migration or selective mortality of larvae at highertemperatures. Back-calculated temperatures of aragonite precipitation derived using regional salinity-dwrelationships and local salinity estimates also did not differ significantly. However, there were significantdifferences between the d18O values of the first seasonal growth ring of age 0 fish as well as back-calculated aragonite precipitation temperatures, suggesting the presence of two potentially discretesubpopulations divided by Punta Eugenia (27�N) along the central Baja California peninsula. Thesefindings are consistent with regional oceanographic patterns and are critical for understanding whiteseabass population structure, and provide information needed for the implementation of appropriatemanagement strategies.

© 2016 Elsevier Ltd. All rights reserved.

1. Introduction

Assessing the role of dispersion, migration and the level ofpopulation connectivity is important for managing coastal fishpopulations and identifying ecologically important habitats(Gillanders, 2005; Cowen and Sponaugle, 2009). Sciaenid fisheshave early life stages that can be found in various coastal habitattypes (Cowan and Birdsong, 1985; Gray and McDonall, 1993). Thelarval stage can last several weeks (Donohoe, 1997), and drift

omo-Curiel), sherzka@cicese.brunius@cicese.mx (P. P�erez-

distance is driven by a host of biological and physical processes(Largier, 2003; Cowen and Sponaugle, 2009). Juveniles can spendmonths to years in juvenile habitats (Griffiths and Hecht, 1995), andadults have the potential to migrate tens to hundreds of kilometersto form spawning aggregations (Aalbers and Sepulveda, 2015).Hence, the population structure of sciaenids and other marinespecies with similar life history strategies may reflect the processesinfluencing dispersion as well as migration patterns.

The white seabass, Atractoscion nobilis, is one the largestmembers of the family Sciaenidae, reaching 160 cm total length(TL) and more than 41 kg (Vojkovich and Reed, 1983). It is animportant species for both the recreational and local commercialfisheries along the Pacific coast of California and Baja California(Allen and Franklin, 1992; Allen et al., 2007; Cartamil et al., 2011).

A.E. Romo-Curiel et al. / Estuarine, Coastal and Shelf Science 170 (2016) 134e144 135

Few studies have provided information regarding spawning loca-tions and migration patterns, and there is limited information onlarval and juvenile ecology. The white seabass spawning period isduring the spring and summer months (May to August), andspawning aggregations may be found in the interface between thesandy areas and shallow rocky reefs that support kelp forests(Thomas, 1968; Aalbers and Sepulveda, 2012).

Several studies focusing on different life stages have contributedto the current understanding of white seabass population structure.Moser et al. (1983) found that while the spawning centers arelocated off southern California and central Baja California, a higherabundance of white seabass larvae were caught in the inshore(<50 m from the surf zone) regions of San Sebastian Vizcaino Bayand the Gulf of Ulloa (Fig. 1). Allen and Franklin (1992) studied thedistribution of newly settled juveniles in coastal waters and pro-posed that larvae transported from northern Baja California tosouthern California were a major source juveniles. Previous geneticstudies have suggested a potentially continuous spawning popu-lation in the eastern Pacific (Bartley and Kent, 1990; Franklin, 1997;Coykendall, 2005), but the geographical range, sample size ornumber of genetic markers employed by these studies is limited.Tagging studies suggest limited (ca. <200) seasonal migrationsalong the California coastline (Hervas et al., 2010; Aalbers andSepulveda, 2015). However, the absence of tagging studies in cen-tral and southern Baja California waters precludes the assessmentof connectivity between the northern and southern range of whiteseabass in the Pacific. More recently, regional differences in otolithgrowth rates of age 1 fish have been found between southernCalifornia and southern Baja California, which are likely attributedto differences in environmental conditions (Romo-Curiel et al.,2015). These findings suggest limited mixing, at least during thefirst year of life. In summary, despite the recent work on

Fig. 1. Sampling areas of white seabass, A. nobilis, off southern California and the BajaCalifornia peninsula (rectangles). Dots indicate the pixels considered to calculate theaverage monthly sea surface temperature (SST) using high-resolution estimatesgenerated by the Scripps Photobiology Group at Scripps Institution of Oceanography.

movements and growth of this species, a clear understanding of itspopulation structure remains lacking for this valuable bi-nationalresource. To build upon recent population structure hypothesesand better understand white seabass population dynamics, thisstudy used otolith isotopic composition to assess populationconnectivity.

1.1. Isotopic analyses

The isotopic composition of otolith aragonite (d18O and d13C) hasbeen used as a natural tracer to evaluate habitat use, populationstructure, migration, connectivity and mixing of fish populations,and for reconstructing their temperature and salinity history(Patterson et al., 1993; Valle and Herzka, 2008). Isotopic ratios arepermanently imprinted in the otolith carbonate, recording ambientconditions experienced by an individual fish at the time of arago-nite precipitation (Campana, 1999). The stable isotopes of oxygen(18O/16O) in otolith aragonite are deposited near equilibrium withthe isotopic composition of the ambient water (dw). However, thereis isotope discrimination against 18O as a function of temperature,resulting in a negative relationship between d18O values and tem-perature (Epstein et al., 1953; Iacumin et al., 1992; Patterson et al.,1993). It is possible to estimate the temperature of the water at thetime of aragonite deposition using d18O values and estimates of dw(Campana, 1999; Darnuade et al., 2014), and to thereby discrimi-nate between fish living under different environmental conditions(Iacumin et al., 1992; Godisken et al., 2012). Regional salinity-dwrelationships can be used to estimate the isotopic composition ofseawater (Craig and Gordon, 1965; Paul et al., 1999), as predict d18Ovalues for particular regions (Rowell et al., 2005; Valle and Herzka,2008).

In contrast, carbon isotope ratios reflect both dietary sources viametabolic processes as well as the dissolved inorganic carbon (DIC)pool (Thorrold et al., 1997; Hoie et al., 2003). The contribution ofcarbon deposited in the otoliths from metabolic sources varies andhas been estimated at 20e30% of the total (Kalish, 1991; Weidmanand Millner, 2000), and the remainder presumably reflects theisotopic composition of DIC (Campana, 1999). Nevertheless,geographic variations the carbon isotope ratios of the otolitharagonite have been successfully used as natural tags to differen-tiate between fish subpopulations (Newman et al., 2010; Steer et al.,2010; Correia et al., 2011).

Data from the otolith core (the central portion of otoliths that isdeposited during early life) can provide information regarding natalorigin and spawning areas, while data corresponding to specificages may indicate changes in fish habitat over time (Gao et al.,2004, 2013; Steer et al., 2010; Gao et al., 2013). The isotopic anal-ysis of subsamples from specific sections of adult otoliths corre-sponding to specific time periods enables the reconstruction ofeach individual's environmental history as a function of age, dateand other life history considerations (Gillanders, 2005, Weidmanand Millner, 2000). Central to the robust interpretation of the iso-topic composition of biogenic carbonates and the calculation ofrobust temperature estimates based on d18O values is the explicitconsideration of the processes that underlie its variability (Thorroldet al., 1997; Campana, 1999; Elsdon and Gillanders, 2002). Thus,oceanographic data should incorporated into the analysis andinterpretation of isotope ratios (Ashford and Jones, 2007).

The isotopic composition of the otolith core of adult whiteseabass collected from three regions of the northeastern Pacific,encompassing a distance of over 1000 km, was measured to inferlarval rearing habitats and population structure. We also analyzedthe isotopic composition of subsamples from seasonal translucentand opaque otolith growth rings corresponding to the juvenile (age0.5e2 yr) and adult (age 8e10 yr) stages to infer ontogenetic

A.E. Romo-Curiel et al. / Estuarine, Coastal and Shelf Science 170 (2016) 134e144136

movement patterns and the level of mixing between different lifestages. Back-calculated aragonite precipitation temperatures werecompared to regional coastal SST's and in situ measurement oftemperature and salinity to evaluate whether the d18O values ofotolith core and growth rings were consistent with regionaloceanographic conditions. Differences in the oxygen and carbonisotopes values in otolith core and growth rings of adult whiteseabass from different locations would indicate that adult fish werereared under different environmental conditions, and would sug-gest limited connectivity between subpopulations. Further, giventhat there is a sharp latitudinal gradient in SST throughout thedistribution of white seabass in Pacific waters, we hypothesizedthat fish sampled in more southern waters would present d18Ovalues consistent with higher aragonite precipitation temperatures.

2. Methods

2.1. Sample collection

Otoliths were obtained opportunistically from commercial andrecreational catches of white seabass from April through August of2009e2011. Commercial fishing activities were conducted aboardsmall (5e7m), outboard-powered fishing boats, within 5e20 km ofcoastal fishing camps along Baja California. Samples were also ob-tained from recreational fishing tournaments targeting white sea-bass within southern California, as well as from PIER researchvessels (SCP ID NUMBER: 002471). A total of 117 samples werecollected in three regions: southern California (SC; n ¼ 40), USA,San Sebastian Vizcaino Bay (VB; n ¼ 36) and the Gulf of Ulloa (GU;n ¼ 41) off the central and southern Baja California peninsula,Mexico, respectively (Fig. 1).

2.2. Isotopic analysis

One sagittal otolith from each fish was embedded in epoxy resinand sectioned transversely using a low speed diamond saw(Buehler, ISOMET™) to cut through the otolith core. The sectioncontaining the core was mounted on a microscope slide usingepoxy resin and polished using abrasive paper of decreasing gritsize. Age estimates of all otoliths sections were evaluated in aprevious study (Romo-Curiel et al., 2015), and the spawning yearfor each fish was back-calculated based on catch date. White sea-bass born between 1999 and 2006 (age range 4e12 yrs; n ¼ 32 forSC, n ¼ 36 for VB and n ¼ 32 for GU) were selected for isotopicanalysis of the otolith core. This period excludes El Ni~no years,which has been shown to influence the isotopic composition of fishotoliths in the California Current region due to the abnormally highwater temperatures (Dorval et al., 2011).

Carbonate was extracted from three different regions along thesectioned otolith surface with a high precision microdrill (ESI NewWaveMicromill) at theMarine Science Institute, University of Texasin Austin and at the Center for Scientific Research and Higher Ed-ucation of Ensenada (CICESE). Each otolith core sample was ob-tained from a 300 � 300 mm square centered on the core using araster pattern with parallel lines separated by 100 mm and with adepth setting of 100 mm. The raster area was designed to coincidewith the dimensions of the otolith core (Fig. 2). The limited di-mensions of the area from which aragonite was extracted in theotolith core was designed to reduce the potential for contaminationwith aragonite precipitated during other life stages (e.g. Begg et al.,1999). Fishes of at least 10 years of age (mean age of 13 yrs) wereselected for isotopic analysis of annual growth rings. Isotopicanalysis of annual growth rings was only conducted for whiteseabass samples collected from the two distal regions (SC and GU)with the most contrasting environmental conditions. The otolith

core and four growth rings corresponding to the juvenile stagewere sampled: the first and second summer growth seasons(opaque rings corresponding to the first and second year of life,hereafter referred as 0.5 and 1.5 yrs) and the first and secondwinters (translucent rings, hereafter age 1 and 2 yrs). Likewise, fivegrowth rings corresponding to the adult stage were sampled (8.5,9.5, and 10.5 yrs, corresponding to the summer opaque growthrings and 9 and 10 yr, corresponding to the winter translucentrings). Each growth ring was sampled by drilling a single curve (650long � 150 mmwide and 100 mm set depth). After the collection ofeach sample, the microdrill bit and work surface were cleaned withcompressed air.

Between 30 and 60 mg of aragonite were weighed on a micro-balance. Analysis of otolith powder was performed in the StableIsotope Laboratory of the University of California at Santa Cruzusing a Kiel IV carbonate device coupled to a Thermo ScientificMAT253 Isotope Ratio Mass Spectrometer under certified standards(CM12, Atlantis II, NBS-18 and NBS-19). Precision measurements(SD) of d18O values were 0.03‰ for CM12, and 0.02‰ for Atlantis II,NBS-18 and NBS-19. For d13C, the precision was 0.06‰ for CM12and Atlantis II, 0.02‰ for NBS-18 and 0.05‰ for NBS-19. All isotopicanalysis andmeasurements are reported in standard d notation (‰)using VPDB (Vienna Peedee Belemnite) as the internationalstandard:

d ¼��

Rsample

.Rstandard

�� 1

�� 1000 (1)

where R is the abundance of the heavy to light isotope ratio of thesample or standard.

The time period integrated by each core sample was estimatedbased on the analysis of young-of-the-year white seabass(4e12 cm TL; n ¼ 20) otolith samples collected during a trawlsurvey along the coast off Camp Pendleton, California on July19e21, 2010. Each sagittal otolith from young-of-the-year whiteseabass was mounted sulcus-side up on a glass slide and polishedusing abrasive paper of decreasing grit and 0.05-micron alumina toenhance the visibility of daily growth increments. Polished otolithswere viewed with a microscope (Leica DMLS; 10e40� magnifica-tion) and photographedwith a digital camera (Leica DFC 450). Dailygrowth increments were counted twice for each individual by twoindependent readers, starting from the focus of the core and corediameter was measured using the Image J software package (Na-tional Institutes of Health, Bethesda, MD, USA).

2.3. Regional SST and salinity

To obtain temperature estimates representative of white sea-bass larval habitat, sea surface temperatures (SST) during thespawning season (AprileSeptember) were obtained by down-loading oceanographic data from the Scripps Photobiology Groupof Scripps Institute of Oceanography, University of California SanDiego (http://spg.ucsd.edu/Satellite_data/California_Current/viewed on October 2014). The SSTs represent an integration ofdata generated with the AquaMODIS, Terra and SeaWiFS satellitesprocessed by Mati Kahru, and are available with high resolution(approx. 1 km2). Monthly mean SST data were compiled for near-coastal waters from April 2000eSeptember 2006 (Fig. 1). Salin-ities at 10 m were obtained from reports and publications derivedfrom in situmeasurements conducted during quarterly CalCOFI andInvestigaciones Mexicanas de la Corriente de California (IMECO-CAL) cruises off the coast of the Baja California peninsula (http://www.calcofi.org/ccpublications/ccreports.html). Salinities wereobtained for two periods: summer (April to September) and winter(October to March) for the same regions used to estimate SST

Fig. 2. Image of a transverse section of an adult white seabass (A. nobilis) otolith. The core is at the center with the raster sampling pattern of the automated micro-milling (dottedlines). White lines represent the single curve for each sampled growth ring of the juvenile and adult stage.

A.E. Romo-Curiel et al. / Estuarine, Coastal and Shelf Science 170 (2016) 134e144 137

values. To evaluate the degree of similarity between monthlyaverage SSTs and salinity very near the coast (not covered by Cal-COFI or IMECOCAL cruises), the SST and salinity records werecompared with in situ measurements made at the Scripps Institu-tion of Oceanography Pier in southern California (http://www.sccoos.org/data/piers). For the waters off the Baja Californiapeninsula regions, the comparisons were made with the seasonalclimatology provided by Durazo et al. (2010).

2.4. Isotopic composition of seawater (dw) and back-calculatedtemperature

The isotopic composition of seawater (dw) was calculated fromregional salinity estimates using the Craig and Gordon (1965)equation for north Pacific surface waters:

dw ¼ �18:5þ 0:54 Salinity (2)

There is an absence of detailed regional data on the isotopiccomposition of seawater, particularly for the waters off central andsouthern Baja California. The processes that contribute to changesin the salinity and dw of surface waters are rainfall, freshwaterinflow, melting ice, evaporation and freezing, as well as verticalmixing and horizontal advection (Schmidt, 1998; Sharp, 2007).However, the waters off southern California and Baja California arenot subject to freezing and are limited in freshwater inflow due tothe region's arid nature (Wilkinson et al., 2009). The primary fac-tors driving variations in salinity in the coastal waters inhabited bywhite seabass are therefore upwelling events, evaporation and themixing of different water masses. The slope of Craig and Gordon(1965) equation is similar to the one reported by Paul et al.(1999) for mid-latitudes and the one derived from GEOSECS data(dw ¼ �17.95 þ 0.50 Salinity; see also Valle and Herzka, 2008),supporting the use of relationship for deriving estimates of dw. Theisotopic composition of the water was corrected (�0.22‰) to ac-count for the difference in the standards used to report oxygenisotope ratios of water and biogenic carbonates (Sharp, 2007).

Back-calculated temperatures were estimated from oxygenisotope ratios (d18Ooto) at the otolith core from SC, VB and GU andalong juvenile growth rings of fish caught in SC and GU by applyingthe aragonite equation proposed by Campana (1999) in an exten-sive review of the literature:

d18Ooto ¼ dw þ 3:71� 0:206 T�C (3)

where d18Ooto is the isotopic composition measured from otolithsubsamples. The isotopic composition of the water was estimatedbased on mean salinity estimates of the region where the corre-sponding adult was caught (see results).

Predicted d18Ooto values for aragonite for temperatures and sa-linities of each region were calculated based on dw estimatesderived using the Craig and Gordon (1965) equation for north Pa-cific surface waters and Campana (1999) aragonite equation.Measured d18Ooto values were compared to predicted d18Ooto valuesto evaluate whether there was overlap in the environmental con-ditions typical of each sampling region and the stage-specific ox-ygen isotopic composition of fish otoliths.

2.5. Statistical analysis

Using the Quartile Method based on median distribution values(Zar, 1999), all outliers were identified and excluded from statisticalanalysis. One-way analysis of variance (ANOVA) was used to testwhether d13C and d18O values of otolith cores differed betweensampling regions. The non-parametric KolmogoroveSmirnov test(KeS) was used to compare the distribution of isotopic valuesamong regions. Back-calculated temperatures were compared be-tween regions using a one-way ANOVA. The post-hoc multiplecomparisons between isotopic values among regions were per-formed using Scheffe's test. Age-specific d13C and d18O values fromjuvenile and adult growth rings were compared between regionswith a two-way repeated measures analysis of variance (RMANOVA); the data corresponding to juvenile and adult growth ringswere treated separately. Holm-Sidak post-hoc tests for pairwisecomparisons were used to identify whether there were differencesin isotopic values between ages and regions. All data complied withparametric assumptions of normality (ShapiroeWilks test) andhomoscedasticity (Levene's test). Data were analyzed using theSystat 13 and plotted using Sigma Plot 12.5 software packages.

3. Results

3.1. Isotopic analysis of otolith cores

Stable isotope values were obtained for 100 white seabassotolith cores (Fig. 3). Two outliers were identified in two differentGU samples with values of �5.37‰ for d13C and �1.88‰ for d18O;both fish were excluded from further analysis. The average d13Cvaluewas�2.66 ± 0.60 for SC (mean ± SD; range�3.73 to�1.54‰),�3.75 ± 0.78‰ for VB (�4.74 to �1.92‰) and �2.81 ± 0.64 for theGU (range �4.17 to �1.28‰). The d18O values were also variable,ranging from �1.52 to �0.02‰ for SC (mean ± SD ¼ �0.61 ±0.31‰),�1.57 to 0.01‰ for VB (�0.75 ± 0.35‰) and �1.49‰to �0.09 (�0.71 ± 0.34) for GU.

There was a significant difference in d13C values between re-gions (ANOVA F(2,95) ¼ 27.47, P < 0.001); the mean carbon isotopecomposition of VB differed from SC and GU (Scheffe's test,

Fig. 3. Carbon and oxygen isotope ratios measured in the otolith cores of WhiteSeabass (A. nobilis) sampled off Southern California (SC), Vizcaino Bay (VB) and in theGulf of Ulloa (GU). White circles indicate outliers that were excluded from statisticalanalyses.

A.E. Romo-Curiel et al. / Estuarine, Coastal and Shelf Science 170 (2016) 134e144138

MS ¼ 0.506, P < 0.001). The frequency distributions of d13C valuesbetween localities differed significantly between VB and thenorthern and southern regions (KeS, P < 0.001). In contrast,average d18O values were not significantly different between re-gions (ANOVA F(2,95) ¼ 1.41, P ¼ 0.249), nor were there differencesin the frequency distributions of isotopic values (KeS, P > 0.05). Therange of d13C and d18O values as a function of spawning year

Fig. 4. d13C (a) and d18O (b) values of otoliths cores of White Seabass plotted as afunction of their spawning year and by sampling region: Southern California (SC),Vizcaino Bay (VB) and Gulf of Ulloa (GU). The numbers above the boxplot represent thesample size for each year.

showed a high level of dispersal, even when the number of in-dividuals analyzed for a particular year was limited (Fig. 4).

3.2. Regional SST and salinity

Comparison of mean monthly SSTs between regions indicated amarked latitudinal gradient (Supplementary Fig. 1), and therewas asimilar pattern in SSTs during the spawning period (AprileSep-tember) from 1999 to 2006. Monthly average temperatures fromApril to September were similar between SC and VB (ranging from15.8 to 21.0 �C and 16.3e21.8 �C, respectively). Higher temperatureswere observed for GU (17.3e26.6 �C). During the summer, SSTs atGUwere 3 to 6 �C higher than at SC and 4 to 5 �C greater than VB. InSC the highest monthly SSTwas observed during August (21.0 �C onaverage). In both VB and the southern region the highest SSToccurred during September (21.8 �C and 26.6 �C, respectively, onaverage).

Regional salinity at 10 m depth during the spawning season(AprileSeptember) and winter (OctobereMarch) months of1999e2006 also showed a latitudinal gradient (Table 1). Given thelinear relationship between salinity and dw, estimates of the iso-topic composition of seawater in GU exhibited lighter values thanthat from SC and VB (Supplementary Fig. 2). Within a region,average dw values showed limited temporal variability, varying by<0.02‰ between seasons in SC and VB, and by 0.25‰ in GU.

3.3. Back-calculated temperature of otolith cores

The back-calculated temperatures derived from d18Ooto values ofthe otolith cores of white seabass ranged from14 �C to 23 �C (Fig. 5).The back-calculated temperatures for fish captured in each regionwere highly variable, reflecting the variability in d18Ooto values. ForSC, the average back-calculated temperature was 17.8 �C (range15.3e21.9 �C),18.6 �C for VB (range 14.9e22.5 �C) and 18.9 �C for GU(range 16.0e22.2 �C). Despite the overlapping back-calculatedtemperatures between regions, there were significant differencesin d18Ooto values (F (2, 95) ¼ 3.979, P ¼ 0.026); the mean back-calculated temperature in GU was significantly higher than SC(Scheffe's test, MS ¼ 2.572, P ¼ 0.032). There were no significantdifferences between the mean back-calculated temperature for SCvs VB and between VB vs GU (Scheffe's test, P > 0.05).

3.4. Comparison of predicted vs. measured d18O values

The range of predicted aragonite d18O values obtained byconsidering regional summer SSTs and salinities had a broaderrange for the GU than SC and VB, and showed overlap betweenregions (Table 1; Fig. 6). For SC the predicted d18O values rangedfrom �0.10‰ and �1.40‰; the range for VB was similar(from �0.15‰ to �1.50‰). In contrast, the range of predicted d18Ovalues was greatest for GU (�0.17‰ to �2.30‰). Comparison of therange of predicted d18O values with d18Ooto measurements indi-cated that almost all d18Ooto measures were within the range of thepredicted isotopic values for a given region. For SC and VB, therange of predicted d18O values (rectangles) is almost the same asthe measured d18Ooto. For GU, the range of predicted d18O valueswas greater than the measured d18Ooto values; none of the otolithcores had isotope ratios <�1.6‰.

3.5. Isotopic analysis of juvenile and adult growth rings

The d13C values of individual juvenile and adult white seabassotolith growth rings collected in SC and GU exhibited an increasinglevel of variability as a function of age; the range of values increasedfrom 2 to 3‰ during the juvenile stage to ca 4.5‰ for the adult

Table 1Regional salinities (minimum,maximum and average), estimated oxygen isotopic composition of thewater (dw), and summer SSTs obtained from satellite images for SouthernCalifornia (SC), Vizcaino Bay (VB) and Gulf of Ulloa (GU).

Region Salinity dw (‰) SST (�C)

Summer Winter Summer Winter Summer

Min Max Mean Min Max Mean Min Max Min Max Min Max

SC 33.2 33.6 33.4 33.2 33.9 33.6 �0.79 �0.58 �0.79 �0.41 15.76 21.06VB 33.3 33.7 33.6 33.3 33.6 33.5 �0.74 �0.52 �0.74 �0.58 16.32 21.81GU 33.6 34.1 33.9 33.7 34.5 34.1 �0.58 �0.31 �0.52 �0.09 17.28 26.58

Fig. 5. Frequency distribution of back-calculated temperatures obtained from d18Ooto

values of otolith cores of White Seabass, A. nobilis, adults collected in Southern Cali-fornia (SC), Vizcaino Bay (VB) and Gulf of Ulloa (GU).

Fig. 6. Isotopic map of predicted d18O values for aragonite as a function of salinity andtemperature. Rectangles represent the range of predicted d18O values for three regionsof the California current system in which white seabass were sampled, based onregional sea surface temperature and salinities during their summer spawning season.Diagonal lines indicate the full range of isotopic values measured in the otolith coresfor each region: Southern California (line with dots), Vizcaino Bay (dashed line) andGulf of Ulloa (dots).

A.E. Romo-Curiel et al. / Estuarine, Coastal and Shelf Science 170 (2016) 134e144 139

stage (Supplementary Fig. 3a and c). In contrast, the d18O values ofjuvenile growth rings showed a consistent pattern between opaque(summer growth, more depleted isotope ratios) and translucent(winter growth, more enriched isotope ratios) rings and a muchnarrower range than d13C values. However, the seasonal patternwas no longer observed during the adult stage (SupplementaryFig. 3b and d).

The mean carbon and oxygen isotope ratios for white seabassotolith growth rings showed distinct patterns when comparing thejuvenile and adult life stages (Fig. 7). While mean isotope ratiosshowed distinct seasonal shifts during the juvenile stage, thesewere not evident during the adult stage. Two-way repeated mea-sures ANOVA indicated that there were no significant differences inmean d13C and d18O values between regions during either the ju-venile or adult stage (Supplementary Table 1). However, for thejuvenile stage, mean d13C and d18O values differed significantlybetween ages for fishes sampled in both regions (Supplementary

Fig. 7. Mean d13C and d18O values (±standard error) measured in seasonal otolithgrowth rings of juvenile and adult of white seabass, A. nobilis. Samples were collectedin Southern California (SC; white circles dots) and Gulf of Ulloa (GU; black circles).N ¼ otolith core; S ¼ summer growth ring; W ¼ winter growth ring; subscriptscorrespond to the estimated age of each growth ring.

A.E. Romo-Curiel et al. / Estuarine, Coastal and Shelf Science 170 (2016) 134e144140

Table 1). The Holm-Sidak post-hoc pairwise comparison test indi-cated significant differences in d18O values between summer (S)and winter (W) growth rings of juveniles (Fig. 7; SupplementaryTable 2). The only significant differences in d18O values betweenages was for S0.5, which corresponds to summer growth during thefirst year of life (the difference between mean isotope ratios was0.65‰), with significant differences at that age among regions (RMANOVA, t ¼ 2.663, P ¼ 0.010).

3.6. Back-calculated temperature for core and seasonal growthrings of juveniles

The back-calculated temperatures for the seasonal growth ringssampled during the juvenile stage showed a broad range andoverlapped between ages and seasons (Fig. 8). The limited vari-ability in salinity (and hence dw estimates) led to temperaturedifferences of <2 �C for individual growth rings. Higher tempera-tures coincided with the summer aragonite precipitation and lowertemperatures corresponded to the winter rings. Back-calculatedtemperatures were higher in fish captured in GU than those fromSC, with a difference around 1 �C for the same age among regions.However, the temperature of the first summer opaque ring (S0.5)was higher in the GU than SC by about 4.5 �C on average.

Fig. 8. Back-calculated temperatures for the core and seasonal growth rings of whiteseabass sampled off southern California (a) and the Gulf of Ulloa (b). Minimum (blackcircles) and maximum temperatures (white circles) were calculated using minimumand maximum regional salinity estimates. In X axis: N ¼ otolith core, W ¼ winter,S ¼ summer, subscript ¼ age of fishes in years.

4. Discussion

Findings of this study provide evidence to support the presenceof two potentially discrete subpopulations of white seabass dividedby Punta Eugenia in the Baja California peninsula. The isotopicanalysis of the first otolith growth ring indicated that white seabasscollected in southern California must consistently experiencedifferent environmental conditions than individuals reared withinthe Gulf of Ulloa. These results agree with the differences in growthrates reported by Romo-Curiel et al. (2015) for the first year of life inadult white seabass from the same regions, and have significantmanagement implications.

4.1. Isotopic composition of otolith cores and back-calculated larvalrearing temperatures

The estimated time interval represented by otolith core sub-samples was 18e25 days, consistent with the 2e3 week duration ofthe white seabass larval period reported by Donohoe (1997). Thed13C and d18O values of the otolith cores of adult white seabasscaptured within each sampling region were highly variable acrossspawning years and age classes (up to 2‰; Fig. 3). The variability inisotopic ratios were persistent through time and thus representa-tive of the white seabass population, allowing for robust inferencesinto its population structure.

The d13C values of the carbon precipitated in fish otoliths is afunction of the dissolved inorganic carbon (DIC) as well as thebyproducts of an individual's metabolism (Kalish, 1991; Weidmanand Millner, 2000; McMahon et al., 2013). The d13C value of theDIC in surface waters of the Pacific off North America is approxi-mately 0.8e0.9‰ as a result of upwelling events along the coast,and those values do not change substantially at spatial scalesdominated by isolated physical processes (Kroopnick, 1985;McMahon et al., 2013). In contrast, the relative contribution ofmetabolically derived carbon to the d13C values of otolith carbonatevaries between species and even among individuals of the samespecies (Iacumin et al., 1992; Weidman and Millner, 2000; Nelsonet al., 2011). Estimating the percent contribution of each source ofcarbon to the d13C values of white seabass otoliths was beyond thescope of this study. However, the d13C values measured in otolithcore of white seabass from each sampled region were highly vari-able and lighter (ca �1.5 to 5‰) than regional DIC values (ca. 1‰;Fig. 3a), suggesting that the contribution of metabolic carbon islikely important, albeit variable. The difference in average d13C inVB relative to those from SC and GU suggests that larval whiteseabass from VB probably relied on a different food source, or thatthe relative contribution of food-based carbon varied betweenregions.

The isotopic composition of seawater (dw) varied seasonallybetween regions, mirroring regional shifts in surface salinities.Craig and Gordon (1965) reported dw values between �0.5and �0.4‰ for the northeastern Pacific surface waters based onsamples taken between 20�N, 130�W and 32�N, 120�W. For thesummer season during which white seabass spawn, the range of dwvalues reported by Craig and Gordon (1965) is only 0.12 and 0.19‰more enriched in 18O than the mean values applied to back-calculated temperature estimates of fish caught in SC and VB,respectively, and similar to the estimates for the GU (�0.58to �0.31‰).

Limited variability in regional salinity (<1, equivalent to 0.5‰ indw) indicates regional temperature is the most important factorinfluencing d18O values observed inwhite seabass otoliths. The highvariability in d18O values measured in otolith core (range 2‰) wasreflected in a wide range of back-calculated temperatures for eachsampling region, which overlapped. Aalbers (2008) reported an

A.E. Romo-Curiel et al. / Estuarine, Coastal and Shelf Science 170 (2016) 134e144 141

ambient water temperature of 11.3e21.7 �C in coastal waters offsouthern California during spawning events of wild white seabass.The highest frequency of spawning activity was recorded between15 and 18 �C, similar to the back-calculated temperatures estimatedfor otolith cores of adults caught in SC waters. The broad range ofback-calculated temperatures might reflect seasonal changes in theenvironmental conditions prevalent during the long spawningperiod of white seabass (Thomas, 1968; Aalbers, 2008). The differ-ences in average back-calculated temperature between SC and GUlikely reflects the well-documented latitudinal temperaturegradient found along the southern California and Baja Californiapeninsula during summer months (e.g. Lynn and Simpson, 1987;Durazo et al., 2010; Supplementary Fig. 1). Alternatively, the dif-ferences in the estimates of average temperature could be biaseddue to variations in the region-specific salinity estimates. Thehigher temperature found in the GU compared to SC could in fact beoffset by the higher salinity found in more southern waters. How-ever, the salinity differed by a maximum of 0.55, which would leadto a limited impact in back-calculated temperature estimates.

The predicted range of d18O values for each region was calcu-lated based on the full range of local salinity and SST estimates forthe larval rearing period. A broader range of potential aragoniteprecipitation temperatures (including lighter d18O values corre-sponding to warmer temperatures) was predicted for GU than SCand VB; however, there was a high degree of overlap between re-gions. Even though the predicted d18O values for GU includedlighter isotope ratios, the isotopic values measured in fish caught inGU covered roughly the same range as those of SC and VB. This maysuggest that (1) white seabass have a behavioral preference formore temperate waters and avoid the warmer nearshore temper-atures found in GU late in the summer season, (2) larvae spawnedtoward the end of the season didn't survive, and (3) that thespawning period is shorter in GU, and that it occurs prior to thewarming of nearshore waters. In addition, the warmer and moresaline waters found in the southern sampling region yielded asimilar range of predicted d18O values than the more northernwaters that are colder and less saline (Fig. 6). Whatever the cause,the high degree of overlap in the d18O values of otolith cores did notallow for discrimination between regions.

4.2. Causes driving variability in d13C and d18O of otolith core ofadults

The predicted regional d18O values calculated from regionaltemperature and salinity measurements did not yield sufficientdifferences to ensure the conclusive discrimination of larval rearinghabitats (i.e., therewas a high degree of overlap in predicted oxygenisotope ratios). In addition, there was a high variability in d18Ootovalues measured in the cores of adult white seabass collected in SC,VB and GU which could be attributed to different causes. Offsouthern California, white seabass have a long spawning seasonthat extends from March to August (Thomas, 1968; Aalbers, 2008);SST increases by around 5 �C in SC waters, 6 �C in VB and 10 �C inthe GU throughout the season. Hence, themixing of larvae reared atdifferent times during the spawning season could drive the largevariation in core d18Ooto and d13C values. Adult white seabass canmigrate relatively long distances to feeding or spawning grounds,in which case adults may spawn in different regions than thosewhere theywere born (Overstreet, 1983). When otolith samples arecollected opportunistically from adults captured by local fishingvessels, as was done in this study, it is possible that fish hadmigrated from different rearing grounds, contributing to the vari-ability in the isotopic composition of otolith cores. In addition,along protected coastal areas of southern California and Baja Cali-fornia, temperatures are typically higher in shallower waters than

offshore, and some exposed coastlines or points are prone to localintense upwelling (Graham and Largier, 1997; Marin et al., 2003;Tapia et al., 2009). Given that the local temperature regime nearthe coast is highly dynamic and variable, it could generate highvariability in the oxygen isotope composition of larval otoliths.White seabass larvae are also epipelagic during the late larvalphase, which may be related with seeking preferential tempera-tures or their predator-avoidance capabilities (Margulies, 1989). Inaddition, off southern Californiawhite seabass larvae change from apelagic distribution to a demersal environment following the12e21 d larval period (Margulies, 1989; Donohoe, 1997). Selectivemortality of larvae at the higher temperatures found in southernwaters late in the spawning season, may explain the lack of adultswith otolith cores with relatively negative isotope ratios in the Gulfof Ulloa. Unfortunately, there are no ecological studies of whiteseabass larvae or juveniles off the Baja California peninsula.

In summary, the results of this study do not clearly identify theprocesses that produce the high variability in d18Ooto from otolithcores of adult white seabass. However, the high degree of overlap inpredicted and measured d18Ooto values between regions coupledwith the potential for long-distance adult migration indicate thatconnectivity among potential subpopulations could occur duringthe adult stage.

4.3. Isotopic composition of juvenile otolith growth rings

The variation in d13C values of juvenile otolith aragonite be-tween translucent and opaque growth rings may be due to differ-ences in the isotopic composition of food sources and changes inmetabolic rate; the latter is considered the most important factorand varies with temperature (Kalish, 1991; Gauldie et al., 1994).White seabass exhibit significant phenotypic plasticity in terms ofgrowth rates and size-at-age (Romo-Curiel et al., 2015), suggestingthat individuals of the same age may have different food sourcesdue to size-related feeding preferences, as has been documentedfor other sciaenid species (Stickney et al., 1975; Peters andMcMichael, 1987). The most depleted d13C values were measuredin the growth rings laid down during the summer, when growthrates, water temperature, and respiration rates are higher, while themost enriched values corresponded to the winter growth rings,when growth and metabolic rates are reduced (Gauldie, 1996; Hoieet al., 2003; Portner et al., 2008). This suggests that the relativecontribution of metabolically derived carbon to the carbon isotoperatios of the otoliths may be higher during the summer, a period ofincreased growth.

The linear relationship between carbonate d18Ooto values andtemperature has been used to validate the seasonality of ringdeposition in otoliths (e.g. Weidman and Millner, 2000). Samplestaken from the translucent rings (summer growth season) of ju-venile white seabass had the lightest d18O values. Opaque growthrings had more enriched d18O values, consistent with the lowertemperatures of the winter season. Similar to other sciaenids,previous age and growth studies of white seabass suggest that thisspecies has an annual ring deposition rate (Romo-Curiel et al.,2015); however, a rigorous validation of the deposition of annuliremained lacking. The clearly seasonal pattern in d18O values foundin juveniles implies that one opaque and one translucent growthring is deposited every year, further validating the yearly depositionassumed by Romo-Curiel et al. (2015).

There was a significant difference in the isotopic composition ofthe first summer growth ring (S0.5) between SC and GU. The ab-solute difference in mean d18O values of 0.65‰ is equivalent to a4 �C back-calculated temperature difference. Southern Californiaand the Gulf of Ulloa are located toward the northernmost andsouthernmost extent of the distribution of white seabass in Pacific

A.E. Romo-Curiel et al. / Estuarine, Coastal and Shelf Science 170 (2016) 134e144142

waters, and along a well-documented latitudinal temperaturegradient (e.g. Lynn and Simpson, 1987; Durazo et al., 2010). Giventhat in this study the otoliths came from fish collected in SC and GUover the course of several years and belonged to different agegroups, the significant difference in the d18O values of the firstsummer growth ring indicates that there is likely limited mixingbetween the northernmost and southernmost sampledpopulations.

4.4. Isotopic composition of adult growth rings

While seasonal variations in the carbon and oxygen isotopiccomposition were evident during the juvenile stage, this was lessevident for the growth rings corresponding to the adult stage.Aalbers and Sepulveda (2015) reported that adults tagged insouthern California were found at a mean depth of 30 m during thewinter, at average temperatures between 13 and 14 �C. During thespring and summer months, these same fish were primarily foundin surface waters at depths of less than 10.5 m, and at mean tem-peratures between 16 and 18 �C. Hence, white seabass, like manyother pelagic species, may actively select their thermal distributionthroughout the year (Morita et al., 2010; Aalbers and Sepulveda,2015), which is consistent with the more limited range of isotopicvalues that were measured during the adult stage. In addition, thesmaller ring widths in adult otoliths may lead to a less recognizableseasonal cycle in isotope ratios of carbonate subsamples (Rowellet al., 2005). Specifically, the absence of a more marked seasonalpattern in the carbon and oxygen isotope ratios of growth ringssampled during the adult stage could also be due to a limitation inthe spatial resolution of the milling procedure. The narrowing ofthe growth rings with age increased the possibility of samplingmore than one growth season. Lastly, some studies have shown anuncoupling between seasonal patterns of ring deposition andgrowth in adult fishes which could also dampen the seasonal signalrecorded in the isotope ratios of adult stage growth rings (Hoie andFolkvord, 2006; Sturrock et al., 2015).

4.5. Population structure of white seabass relative to oceanographicconditions

Although past studies of larval white seabass speculated thatdispersal could occur over large distances (over hundreds of kilo-meters, e.g. Allen and Franklin, 1992; Donohoe, 1997), recentstudies suggest that localized larval retention may be more com-mon than previously thought (reviewed by Cowen et al., 2007). Inspecies with nearshore spawning and relatively short larval stages(<month), evidence suggests that larval dispersal occurs over amuch more limited spatial scale and that larval behavior and localretention zones may be important (<100 km, Cowen et al., 2007;see also review by Pineda et al., 2007). The large geographicalseparation between the potential subpopulations analyzed in thisstudy (up to 1000 km) makes it unlikely that connectivity betweenSC and GUmay occur during the larval stage. In addition, SC and theGU tend to show large-scale cyclonic circulation during at least partof the spawning season of white seabass, while the circulation in VBtends to be anticyclonic (e.g. Lynn and Simpson, 1987; Gonz�alez-Rodríguez et al., 2012; Durazo, 2015). This divergent circulationlimits the potential for larval dispersal across the Punta Eugeniaregion, which is the southern limit of distribution for a variety oftemperate species, as well as the northern limit of distribution fortropical species characteristic of the subtropical equatorial Pacific(e.g. Bernardi et al., 2003; Lluch Belda et al., 2003). There istherefore independent evidence that the Punta Eugenia region is anatural barrier to connectivity.

5. Conclusions

In conclusion, the isotopic composition of the otolith core ofadult white seabass vary substantially but did not differ sufficientlyto discriminate between potential subpopulations of larvae rearedin three different potential spawning regions. However, there weredifferences between the oxygen isotopic composition and back-calculated temperature of the first seasonal growth ring of fishcaptured in SC and GU, suggesting the presence of two potentiallydiscrete subpopulations divided by Punta Eugenia along the centralBaja California peninsula (27�N).

Acknowledgments

We thank the many students and local fishermen involved infield sampling. Dr. Benjamin Walther (University of Texas, MarineScience Institute) loaned his facilities during processing of theotolith samples on themicrodrill. Dr. Dyke Andreasen (University ofCalifornia in Santa Cruz, Stable Isotope Laboratory) provided tech-nical assistance with isotopic analysis. Dr. Reginaldo Durazo (Uni-versidad Aut�onoma de Baja California, Facultad de CienciasMarinas) and Drs. Oscar Sosa and Juan Carlos Herguera (CICESE)provided helpful comments and suggestions. Mexico's ConsejoNacional de Ciencia y Tecnología (CONACyT) provided a graduatefellowship to A.E.R.C. Financial support was provided by the GeorgeT. Pfleger Foundation and Offield Family Foundation (Project No.1.47), and CICESE (Project No. 625116).

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.ecss.2016.01.016.

References

Aalbers, S.A., 2008. Seasonal, diel, and lunar spawning periodicities and associatedsound production of white seabass (Atractoscion nobilis). Fish. Bull. 106 (2),143e151.

Aalbers, S.A., Sepulveda, C.A., 2012. The utility of long-term acoustic recordingsystem for detecting white seabass Atractoscion nobilis spawning sounds. J. Fish.Biol. 81 (6), 1859e1870. http://dx.doi.org/10.1111/j.1095-8649.2012.03399.x.

Aalbers, S.A., Sepulveda, C.A., 2015. Seasonal movement patterns and temperatureprofiles of adult white seabass (Atractoscion nobilis) off California. Fish. Bull. 113(1), 1e14. http://dx.doi.org/10.7755/FB.113.1.1.

Allen, L.G., Franklin, M.P., 1992. Abundance, distribution, and settlement of young-of-the-year white seabass, Atractoscion nobilis, in the Southern California Bight,1988e1989. U.S. Fish. Bull. 90 (4), 633e641. http://dx.doi.org/10.1016/j.fishres.2007.07.012.

Allen, L.G., Pondella II, D.J., Shane, M.A., 2007. Fisheries independent assessment of areturning fishery: abundance of juvenile white seabass (Atractoscion nobilis) inthe shallow nearshore waters of the Southern California Bight, 1995e2005. Fish.Res. 88 (1), 24e32. http://dx.doi.org/10.1016/j.fishres.2007.07.012.

Ashford, J., Jones, C., 2007. Oxygen and carbon stable isotopes in otoliths recordspatial isolation of Patagonian toothfish (Dissoctichus eleginoides). Geochim.Cosmochim. Acta 71, 87e94. http://dx.doi.org/10.1016/j.gca.2006.08.030.

Bartley, D.M., Kent, D.B., 1990. Genetic structure of white seabass population fromthe southern California Bight region: applications to hatchery enhancement.CalCOFI Rep. 31, 97e105.

Begg, G.A., Hare, J.A., Sheehan, D.D., 1999. The role of life history parameters asindicator of stock structure. Fish. Res. 43 (1), 141e163. http://dx.doi.org/10.1016/S0165-7836(99)00071-5.

Bernardi, G., Findley, L., Rocha-Olivares, A., 2003. Vicariance and dispersal acrossBaja California in disjunction marine fish populations. Evolution 57 (7),1599e1609. http://dx.doi.org/10.1111/j.0014-3820.2003.tb00367.x.

Campana, S.E., 1999. Chemistry and composition of fish otolith: pathways, mech-anisms and applications. Mar. Ecol. Prog. Ser. 188, 263e297.

Cartamil, D., Santanta-Morales, O., Escobedo-Olvera, M.A., Kacev, D., Castillo-Geniz, J.L., Graham, J.B., Sosa-Nishizaki, O., 2011. The artisanal elasmobranchfishery of the Pacific coast of Baja California, Mexico. Fish. Res. 108 (2), 393e403.http://dx.doi.org/10.1016/j.fishres.2011.01.020.

Correia, A.T., Barros, F., Sial, A.N., 2011. Stock discrimination of European conger eel(Conger conger l.) using otolith stable isotope ratios. Fish. Res. 108 (1), 88e94.http://dx.doi.org/10.1016/j.fishres.2010.12.002.

Cowan, J.J.H., Birdsong, R.S., 1985. Seasonal occurrence of larval and juvenile fishes

A.E. Romo-Curiel et al. / Estuarine, Coastal and Shelf Science 170 (2016) 134e144 143

in a Virginia Atlantic coast estuary with emphasis on drums (family Sciaenidae).Estuaries 8 (1), 48e59. http://dx.doi.org/10.2307/1352121.

Cowen, R.K., Gawarkiewicz, G., Pineda, J., Thorrold, S.R., Werner, F.E., 2007. Popu-lation connectivity in marine systems: an overview. Oceanography 20 (3),14e21. http://dx.doi.org/10.5670/oceanog.2007.26.

Cowen, R.K., Sponaugle, S., 2009. Larval dispersal and marine population connec-tivity. Annu. Rev. Mar. Sci. 1, 443e466. http://dx.doi.org/10.1146/annurev.marine.010908.163757.

Coykendall, D.K., 2005. Population Structure and Dynamics of White Seabass(Atractoscion nobilis) and the Genetic Effect of Hatchery Supplementation onthe Wild Population (PhD thesis). University of California, Davis, CA.

Craig, H., Gordon, L., 1965. In: Proceedings of the Conference on Stable Isotopes inOceanographic Studies and Paleotemperatures. Chapter Deuterium andOxygen-18 Variations in the Ocean and the Marine Atmosphere. Luschi andFigli, Spoleto, Italy, pp. 1e130.

Darnuade, A.M., Sturrock, A., Trueman, C.N., Mouillot, D., EIMF, Campana, S.E.,Hunter, E., 2014. Listening in on the past: what can otolith d18O values reallytell us about the environmental history of fishes? PLoS One 9 (10), e108539.http://dx.doi.org/10.1371/journal.pone.0108539.

Donohoe, C.J., 1997. Age, growth, distribution and food habits of recently settledwhite seabass, Atractoscion nobilis, off San Diego Country, California. U.S. Fish.Bull. 95 (4), 709e721.

Dorval, E., Piner, K., Robertson, L., Reiss, C.S., Javor, B., Vetter, R., 2011. Temperaturerecord in the oxygen stable isotopes of Pacific sardine otoliths: experimental vs.wild stocks from the Southern California Bight. J. Exp. Mar. Biol. Ecol. 397 (2),136e143. http://dx.doi.org/10.1016/j.jembe.2010.11.024.

Durazo, R., 2015. Seasonality of the transitional regi�on of the California currentsystem off Baja California. J. Geophys. Res. 120 (2), 1173e1196. http://dx.doi.org/10.1002/2014JC010405.

Durazo, R., Ramirez-Manguilar, A.M., Miranda, L.E., Soto-Mardones, L.A., 2010. Cli-matología de variables hidrogr�aficas. In: Gaxiola-Castro, G., Durazo, R. (Eds.),Din�amica del ecosistema pel�agico frente a Baja California 1997-2007: Diez a~nosde investigaciones mexicanas de la Corriente de California, pp. 25e57. M�exico.

Elsdon, T.S., Gillanders, B.M., 2002. Interactive effects of temperature and salinity onotolith chemistry: challenges for determining environmental histories of fish.Can. J. Fish. Aquat. Sci. 59 (11), 1796e1808. http://dx.doi.org/10.1139/f02-154.

Epstein, S., Buchsbaum, R., Lowenstam, H., Urey, H.C., 1953. Revised carbonate waterisotopic temperature scale. Bull. Geol. Soc. Am. 64 (11), 1315e1326 doi: 10.1130/0016-7606(1953)64[1315:RCITS]2.0.CO;2.

Franklin, M.P., 1997. An Investigation into the Population Structure of White Seabass(Atractoscion nobilis) in the California and Mexican Waters Using MicrosatelliteDNA Analysis (PhD thesis). University of California, Santa Barbara.

Gao, Y.W., Joner, S.H., Bargmann, G.G., 2013. Stable isotopic composition of otolithsin identification of spawning stocks of pacific herring (Clupea pallasi) in Pugentsound. Can. J. Fish. Aquat. Sci. 58 (1), 2113e2120. http://dx.doi.org/10.1139/f01-146.

Gao, Y.W., Joner, S.H., Svec, R.A., Weinberg, K.L., 2004. Stable isotopic comparison inotoliths of juvenile sablefish (Anoplopoma fimbria) from the waters off theWashington and Oregon coast. Fish. Res. 68 (1), 351e360. http://dx.doi.org/10.1016/j.fishres.2003.11.002.

Gauldie, R.W., 1996. Biological factors controlling the carbon isotope record in fishotoliths: principles and evidence. Comp. Biochem. Physiol. 115B (2), 201e208.http://dx.doi.org/10.1016/0305-0491(96)00077-6.

Gauldie, R.W., Thacker, C.E., Merret, N.R., 1994. Oxygen and carbon isotope variationin the otoliths of Beryx splendens and Coryphaenoides profundicolus. Comp.Biochem. Physiol. 108A (2), 153e159. http://dx.doi.org/10.1016/0300-9629(94)90080-9.

Gillanders, B.M., 2005. Using elemental chemistry of fish otoliths to determineconnectivity between estuarine and coastal habitats. Est. Coast. Shelf Sci. 64 (1),47e57. http://dx.doi.org/10.1016/j.ecss.2005.02.005.

Godisken, J.A., Power, M., Borgstrom, R., Dempson, J.B., Svenning, M.A., 2012.Thermal habitat use and juvenile growth of Svalbard Artic charr: evidence fromotolith stable oxygen isotope analyses. Ecol. Freshw. Fish 21, 134e144. http://dx.doi.org/10.1111/j.1600-0633.2011.00533.x.

Gonz�alez-Rodríguez, E., Trasvi~na-Castro, A., Gaxiola-Castro, G., Zamudio, L., Cer-vantes-Duarte, R., 2012. Net primary productivity, upwelling and coastal cur-rents in the Gulf of Ulloa, Baja California, M�exico. Ocean Sci. 8 (4), 703e711.http://doi.org/10.5194/os-8-703-2012.

Graham, W.M., Largier, J.L., 1997. Upwelling shadows as nearshore retention sites:the example of northern Monterey Bay. Cont. Shelf Res. 17 (5), 509e532. http://dx.doi.org/10.1016/S0278-4343(96)00045-3.

Gray, C.A., McDonall, V.C., 1993. Distribution and growth of juvenile mullowayArgyrosomus hololepidotus (Pisces: Sciaenidae), in the Hawkesbury river, south-eastern Australia. Aust. J. Mar. Fresw. Res. 44 (3), 401e409. http://dx.doi.org/10.1071/MF9930401.

Griffiths, M.H., Hecht, T., 1995. Age and growth of South African dusky kob Argyr-osomus japonicus (Sciaenidae) based on otoliths. S. Afr. J. Mar. Sci. 16 (1),119e128. http://dx.doi.org/10.2989/025776195784156548.

Hervas, S., Lorenzen, K., Shane, M.A., Drawbridge, M.A., 2010. Quantitative assess-ment of white seabass (Atractoscion nobilis) stock enhancement program inCalifornia: post-release dispersal, growth and survival. Fish. Res. 105 (3),237e243. http://dx.doi.org/10.1016/j.fishres.2010.06.001.

Hoie, H., Folkvord, A., 2006. Estimating the timing of growth rings in Atlantic codotoliths using stable oxygen isotopes. J. Fish. Biol. 68, 826e837. http://dx.doi.org/10.1111/j.1095-8649.2006.00957.x.

Hoie, H., Folkvord, A., Otterlei, E., 2003. Effect of somatic and otolith growth rate onstable isotopic composition of early juvenile cod (Gadus morhua l) otoliths.J. Exp. Mar. Biol. Ecol. 4018, 1e18. http://dx.doi.org/10.1016/S0022-0981(03)00034-0.

Iacumin, P., Bianucci, G., Longinelli, A., 1992. Oxygen and carbon isotopic compo-sition of fish otoliths. Mar. Biol. 113 (4), 537e542. http://dx.doi.org/10.1007/BF00349696.

Kalish, J.M., 1991. d13C and d18O isotopic disequilibria in fish otoliths: metabolic andkinetic effects. Mar. Ecol. Prog. Ser. 75 (2), 191e203.

Kroopnick, P.M., 1985. The distribution o 13C of SCO2 in the world ocean. Deep SeaRes. 32 (1), 57e84. http://dx.doi.org/10.1016/0198-0149(85)90017-2.

Largier, J.L., 2003. Considerations in estimating larval dispersal distances fromoceanographic data. Ecol. Appl. 13 (1), 71e89 http://dx.doi.org/10.1890/1051-0761(2003)013[0071:CIELDD]2.0.CO;2.

Lluch Belda, D., Lluch-Cota, D.B., Lluch-Cota, S.E., 2003. Baja California's biologicaltransition zones: refuges for the California sardine. J. Oceanogr. 59 (4), 503e513.http://dx.doi.org/10.1023/A:1025596717470.

Lynn, R.J., Simpson, J.J., 1987. The California current system: the seasonal variabilityof its physical characteristics. J. Geophys. Res. 92 (C12), 12947e12966. http://dx.doi.org/10.1029/JC092iC12p12947.

McMahon, K.W., Hamady, L.L., Thorrold, S.R., 2013. A review of ecogeochemistryapproaches to estimating movements of marine animals. Limnol. Oceanogr. 58(2), 697e714. http://dx.doi.org/10.4319/lo.2013.58.2.0697.

Margulies, D., 1989. Size-specific vulnerability to predation and sensory systemdevelopment of white seabass, Atractoscion nobilis, larvae. Fish. Bull. 87 (3),537e552.

Marin, V.H., Delgado, L.E., Escribano, R., 2003. Upwelling shadows at Mejillones Bay(northern Chilean coast): a remote sensing in situ analysis. Invest. Mar. Valpso.31 (2), 47e55.

Morita, K., Fukuwaka, M., Tanimata, N., Yamamura, O., 2010. Size-dependent ther-mal preferences in pelagic fish. Oikos 119, 1265e1272. http://dx.doi.org/10.1111/j.1600-0706.2009.18125.x.

Moser, H.G., Ambrose, D.A., Busby, M.S., Butler, J.L., Sandknop, E.M., Sumida, B.Y.,Stevens, E.G., 1983. Description of early stages of white seabass, Atractoscionnobilis, with notes on distribution. CalCOFI Rep. 24, 182e193.

Nelson, J., Hanson, C.W., Koenig, C., Chanton, J., 2011. Influence of the diet on stablecarbon isotope composition in otoliths of juvenile red drum Sciaenops ocellatus.Aquat. Biol. 13, 89e95. http://dx.doi.org/10.3354/ab00354.

Newman, S.J., Allsop, Q., Ballag, A.C., Garret, R.N., Gribble, N., Meeuwing, J.J.,Mitsopoulos, G.E.A., Moore, B.R., Pember, M.B., Rome, B.M., Saunders, T.,Skepper, C.L., Stapley, J.M., van Herwerden, L., Welch, D.J., 2010. Variation instable isotope (d18O and d13C) signatures in the sagittal otolith carbonate of kingthreadfin, Polydactylus macrochir, across northern Australia reveals multifacetedstock structure. J. Exp. Mar. Biol. Ecol. 396 (1), 53e60. http://dx.doi.org/10.1016/j.jembe.2010.09.011.

Overstreet, R.M., 1983. Aspects of the biology of the red drum, Sciaenops ocellatus, inMississippi. Gulf Res. Rep. 1, 45e68.

Patterson, W.P., Smith, G.R., Lohmann, K.C., 1993. Continental paleothermometryand seasonality using the isotopic composition of aragonitic otoliths of fresh-water fishes. In: Swart, P.K., Lohmann, K.C., Mckenzie, J., Savin, S. (Eds.), ClimateChange in Continental Isotopic Records. American Geophysical Union, Wash-ington, D. C. http://dx.doi.org/10.1029/GM078p0191.

Paul, A., Mulitsa, S., Patzold, J., Wolff, T., 1999. Use of Proxies in Paleography: Ex-amples From the Southern Atlantic, Chapter Simulation of Oxygen Isotopes in aGlobal Ocean Model. Springer, Berlin, pp. 665e686.

Peters, K.M., McMichael, R.H., 1987. Early life history of the red drum, Sciaenopsocellatus (Pisces: Sciaenidae), in Tampa Bay, Florida. Estuary 10 (2), 92e107.http://dx.doi.org/10.2307/1352173.

Pineda, J., Hare, J.A., Sponaugle, S., 2007. Larval transport and dispersal in the coastalocean and consequences for population connectivity. Oceanography 20 (3),22e39. http://dx.doi.org/10.5670/oceanog.2007.27.

Portner, H.O., Bock, C., Knust, R., Lanning, G., Lucassen, M., Mark, F.C., Sartoris, F.J.,2008. Cod and climate in a latitudinal cline: physiological analyses of climateeffects in marine fishes. Clim. Res. 37, 253e270. http://dx.doi.org/10.3354/cr00766.

Romo-Curiel, A.E., Herzka, S.Z., Sosa-Nishizaki, O., Sepulveda, C.A., Aalbers, S.A.,2015. Otolith-based growth estimates and insights into population structure ofwhite seabass, Atractoscion nobilis, off the Pacific coast of North America. Fish.Res. 161, 374e383. http://dx.doi.org/10.1016/j.fishres.2014.09.004.

Rowell, K., Flessa, K.W., Dettman, D.L., Rom�an, M., 2005. The importance of ColoradoRiver flow to nursery habitats of the Gulf corvine (Cynoscion othonopterus). Can.J. Fish. Aquat. Sci. 62, 2874e2885. http://dx.doi.org/10.1139/F05-193.

Schmidt, G.A., 1998. Oxygen-18 variation in a global ocean model. Geophys. Res.Lett. 25 (8), 1201e1204. http://dx.doi.org/10.1029/98GL50866.

Sharp, Z., 2007. Principles of Stable Isotope Geochemistry. Pearson, Prentice Hall,University of New Mexico.

Steer, M.A., Halverson, G.P., Fowler, A.J., Gillanders, B.M., 2010. Stock discriminationof southern garfish (Hyporhamphus melanochir) by stable isotope ratio analysisof the otolith aragonite. Environ. Biol. Fish. 89, 369e381. http://dx.doi.org/10.1007/s10641-010-9670-5.

Stickney, R.R., Taylor, G.L., White, D.B., 1975. Food habits of five species of youngsoutheastern United States estuarine Sciaenidae. Chesap. Sci. 16 (2), 104e114.http://dx.doi.org/10.2307/1350687.

Sturrock, A.M., Hunter, E., Milton, J.A., EIMF, Johnson, R.C., Waring, C.P.,Trueman, C.N., 2015. Quantifying physiological influences on otolith

A.E. Romo-Curiel et al. / Estuarine, Coastal and Shelf Science 170 (2016) 134e144144

microchemistry. Methods Ecol. Evol. 6, 806e816.Tapia, F.J., Navarrete, S.A., Castillo, M., Menge, B.A., Castilla, J.C., Largier, J.,

Wieters, E.A., Broitman, B.L., Barth, J.A., 2009. Thermal indices of upwellingeffects on inner-shelf habitats. Prog. Ocean. 83, 278e287. http://dx.doi.org/10.1016/j.pocean.2009.07.035.

Thomas, J.C., 1968. Management of the white seabass (Cynoscion nobilis) in Cali-fornia waters. Calif. Dep. Fish Game Fish Bull. 142, 1e33.

Thorrold, S.R., Campana, S.E., Jones, C.M., Swart, P.K., 1997. Factors determining d13Cand d18O fractionation in aragonitic otoliths of marine fish. Geochim. Cosmo-chim. Acta 14 (61), 2909e2919. http://dx.doi.org/10.1016/S0016-7037(97)00141-5.

Valle, S.R., Herzka, S.Z., 2008. Natural variability in d18O values of otoliths of youngpacific sardine captured in Mexican waters indicates subpopulation mixing

within the first year of life. ICES J. Mar. Sci. 65 (2), 174e190. http://dx.doi.org/10.1093/icesjms/fsn001.

Vojkovich, M., Reed, R.J., 1983. White seabass, Atractoscion nobilis, in California-Mexican waters: status of the fishery. Calif. Coop. Ocean. Fish. Invest. Rep. 24,79e83.

Weidman, C.R., Millner, R., 2000. High-resolution stable isotope records from northAtlantic cod. Fish. Res. 46 (1), 327e342. http://dx.doi.org/10.1016/S0165-7836(00)00157-0.

Wilkinson, T., Wiken, E., Bezaury Creel, J., Hourigan, T., Agardy, T., Hermann, H.,Janishevski, L., Madden, C., Morgan, L., Padilla, M., 2009. Ecorregiones marinasde Am�erica del Norte. Comisi�on para la Cooperaci�on Ambiental, Montreal,p. 200.

Zar, J., 1999. Biostatistical Analysis, fourth ed. Prentices Hall, Upper Saddle River, NJ.