ICES Journal of Marine Science: Journal du Conseil Advance Access originally published online on February 5, 2008
ICES Journal of Marine Science: Journal du Conseil 2008 65(2):174-190; doi:10.1093/icesjms/fsn001
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Natural variability in 
18O values of otoliths of young Pacific sardine captured in Mexican waters indicates subpopulation mixing within the first year of life
Department of Biological Oceanography, Center for Scientific Research and Higher Education of Ensenada (CICESE), km 107 Carretera Tijuana-Ensenada, CP 22860 Ensenada, Baja California, Mexico, and PO Box 434844, San Diego, CA 92143, USA
Correspondence to S. Z. Herzka: tel: +52 646 175 0500; fax: +52 646 1750587; e-mail: sherzka{at}cicese.mx
Valle, S. R., and Herzka, S. Z. 2008. Natural variability inOxygen stable isotopes were measured in whole sagittae of young Pacific sardine (18O values of otoliths of young Pacific sardine captured in Mexican waters indicates subpopulation mixing within the first year of life. – ICES Journal of Marine Science, 65: 174–190.
18Ooto) collected throughout their range in the Mexican Pacific to quantify natural variability, to reconstruct temperature histories, and to infer whether fish mix at a population or subpopulation level. Isotopic values and derived temperature estimates (Toto) of sardine captured simultaneously showed high variability (up to 2.0
and 10°C at a given location). Given limited variations in salinity, this implies differences in thermal history and the prevalence of subpopulation-level mixing processes. We tested the null hypothesis of local residence by comparing 18Ooto values with predicted isotopic values on a location-specific basis, and age- and location-specific average sea surface temperatures (SSTs) with derived Toto. Some fish exhibited values outside the local range of predicted oxygen isotope values and SSTs, suggesting that they were not permanent residents. Using an otolith growth model, we show that otolith growth and age differences cannot fully account for the variability in 18Ooto values. The absence of significant differences in 18Ooto values between the Pacific and Gulf indicates that oxygen isotope ratios cannot be used to examine population structure or migration among these regions. However, they can be used to infer mixing within and among subpopulations.
Keywords: Mexico, otolith 18O, Sardinops sagax caeruleus, temperature reconstruction
Received 12 August 2007; accepted 19 December 2007; advance access publication 5 February 2008.
 |
Introduction |
|---|
 Top Introduction Material and methodsResultsDiscussionConclusionsReferences |
|---|
The Pacific sardine (Sardinops sagax caeruleus) is currently distributed from Alaska, USA, to the southern extreme of the peninsula of Baja California and within the Gulf of California, Mexico. Given its commercial importance since the early 20th century, several studies have sought to use various natural and artificial markers to establish natal origin, the extent of mixing among subpopulations, migration patterns, and spatial stock structure. Evidence collected before the collapse of the sardine fishery in the 1940s and 1950s and after its recovery in the 1980s suggests the presence of at least three distinct subpopulations within the range (two along the Pacific coast in the California Current system, and one within the Gulf of California; see Smith, 2005, for review). However, the boundaries among subpopulations and the extent of mixing among them have yet to be established. This may be due to the broad distribution and pelagic habitat of the species, its marked spatial and temporal fluctuations in abundance, and high migratory potential. Additionally, the age at which sardine begin to migrate and the extent of mixing within schools and regions is currently unknown.
Substantial research effort has attempted to link spatial and temporal variability in the population dynamics of the Pacific sardine with environmental conditions (Marr, 1960; Smith, 1981; Hammann et al., 1988; McFarlane et al., 2002; Félix-Uraga et al., 2004). Temperature seems to have an important influence on geographical population structure of the Pacific sardine, either directly or indirectly. Throughout its present-day distribution in the Mexican Pacific, sardine are rarely caught at sea surface temperatures (SSTs) <14°C (Félix-Uraga et al., 2004). The lower thermal limit for the Pacific sardine larvae is ca. 12°C (Lasker, 1964), at least for the Pacific. Smith (1981) proposed the existence of three subpopulations based on temperature and the timing of spawning: a northern subpopulation spawning in spring from Punta Eugenia, Baja California, to San Francisco, USA, a southern subpopulation spawning in late summer from Punta Eugenia to the southern extent of the Baja California peninsula, and a third spawning within the Gulf of California in winter. Félix-Uraga et al. (2004) related catch statistics to sea surface SST and found evidence to support the presence of three subpopulations.
Studies based on measurements of the relative abundance of the stable isotopes of oxygen (18O) in the calcium carbonate of biogenic structures have demonstrated their utility as recorders of environmental variation and thermal history (Wefer and Berger, 1991). In fish, otolith isotopic values (18Ooto) have been successfully used as natural recorders of temperature during growth, to lend insight into geographical population structure and to infer patterns of habitat utilization (Kalish, 1991a, b; Iacumin et al., 1992; Edmonds and Fletcher, 1997; Campana, 1999; Begg and Weidman, 2001; Stephenson et al., 2001; Ayvazian et al., 2004). Otoliths are acellular (i.e. metabolically inert) structures consisting primarily of calcium carbonate as aragonite. The continuous and periodic nature of their growth allows them to be used for reconstructing individual growth rates and estimating age (Campana and Neilson, 1985).
Oxygen stable isotopes (18O and 16O) incorporated into otolith carbonate are deposited at or very near equilibrium with their relative abundance in seawater (see review in Campana, 1999). Laboratory experiments indicate that the incorporation of 18O and 16O is not influenced by kinetic effects or metabolic processes (Kalish, 1991a; Thorrold et al., 1997; Høie et al., 2003). However, thermodynamic considerations show that there is greater discrimination against 18O than 16O during CaCO3 precipitation at higher temperatures. This results in a predictable negative relationship between temperature and carbonate 18O values, with a slope of 
0.2 °C–1 (Grossman and Ku, 1986; Kim and O'Neil, 1997; Campana, 1999).
Temperature can be estimated from 18Ooto values if the oxygen isotope composition of seawater (w) in which a fish precipitated its otolith aragonite is measured directly (Thorrold et al., 1997; Campana, 1999). Alternatively, if data are not available, salinity can be used to calculate w (Paul et al., 1999; Weidman and Millner, 2000; LeGrande and Schmidt, 2006). As w varies as a function of evaporation and precipitation (Dansgaard, 1964), and these same processes control the salinity of sewater, there is a positive linear relationship between w and salinity. However, the slope and intercept of their relationship can vary with latitude, among water masses, and as a function of depth attributable to fractionation of oxygen isotopes during evaporation and precipitation as well as the formation of sea ice (Dansgaard, 1964; Craig and Gordon, 1965; Paul et al., 1999). For north Pacific surface waters, Craig and Gordon (1965) reported the following relationship: w = –18.5 + 0.54 Salinity.
Isotopic ratios measured in whole otoliths represent an integrated measure of the temperature (and salinity) history experienced by individual fish. Variations in carbonate accretion rates throughout life, which may be at least partially mediated by temperature (Høie et al., 2004a; Fey, 2006) may further influence whole-otolith 18O values. Therefore, different 18O values found in a single locality can be interpreted in terms of variability of temperature and salinity throughout the life of a fish. More specifically, if the oxygen isotope composition of conspecifics of similar age and size is highly variable, and salinity exhibits limited variation, this implies that changes in temperature exert the main control on otolith 18O values.
Recent technological developments have led to successful otolith subsampling and analysis (Dettman and Lohmann, 1995; Gao, 1999), allowing for the reconstruction of seasonal growth patterns at the level of an individual fish that are driven mainly by temperature (Weidman and Millner, 2000; Gao et al., 2004; Høie et al., 2004b; Jamieson et al., 2004; Høie and Folkvord, 2006). However, most studies published to date have focused on species that have much larger otoliths than Pacific sardine (e.g. Atlantic cod, Gadus morhua, and sablefish, Anoplopoma fimbria; Schwarcz et al., 1998; Weidman and Millner, 2000; Gao et al., 2004). Although technical limitations associated with microdrilling of small otoliths still exist (see Gao et al., 2001, for an example), recent technological advances should increasingly allow for otolith subsampling at increasingly greater spatial (and hence temporal) resolution (Wurster et al., 1999). Nevertheless, the resolution is still limited to the time interval integrated over the subsample obtained (Andrus et al., 2002).
To date, only a limited number of studies on wild-caught fish have attempted to reconstruct the temperature at which individuals grew based on isotopic measurements of otoliths (e.g. Devereux, 1967; Degens et al., 1969; Kalish, 1991a; Iacumin et al., 1992; Gao et al., 2001; Andrus et al., 2002). Even fewer have attempted to gain insight into population structure and mixing based on both 18O values and the geographical distribution of w (e.g. Ashford and Jones, 2007). In palaeoceanography, in contrast, reconstruction of historical temperature series based on 18O values of biogenic carbonates, w and T–18O relationships has proven invaluable to understanding the relationship between climate variability, ocean temperature, and productivity (Wefer and Berger, 1991; Sharp, 2006). Here, we show that applying a geochemical approach to the interpretation of oxygen isotope values of otolith carbonate of pelagic species can yield additional insights to that obtained from using 18O values as tracers of population structure.
We measured the oxygen isotope composition of whole otoliths of relatively young sardine (6–18 months) captured throughout their distribution in the Mexican Pacific to quantify natural variability, to determine whether oxygen isotopes can be used to discriminate among potential subpopulations, and to make inferences about mixing within and among subpopulations and patterns of habitat utilization. We chose to focus on young sardine to minimize the amount of time integrated by isotopic signals.
To infer habitat utilization, we tested the null hypothesis that young sardine collected in various areas of the Mexican Pacific had remained near their capture location using two parallel approaches. First, we compared 18Ooto values of sardine with oxygen isotope ratios predicted from local temperature and salinity. Isotopic values outside the range of local predicted isotope ratios would indicate an individual did not permanently inhabit a given location. Second, we derived the integrated temperature to which individual fish were exposed (Toto) using literature w values, and compared them with an integration of SSTs calculated from the start of otolith growth until capture (hereafter referred to as "local" SSTs). Cases in which Toto estimates and local SSTs do not agree suggests that individuals were not permanent residents of a given area. The sensitivity of Toto to local variations in salinity was examined by recalculating temperatures for a range of w values reflective of local variations in salinity. We also examined the sensitivity of local SST estimates to sardine age as well as to the area selected, to generate "local" averages. Lastly, we used an otolith growth model constructed from literature-derived empirical data to assess the potential influence of growth rate and age differences on whole otolith 18Ooto values.
|
Material and methods |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionConclusionsReferences |
|---|
Study area
In Mexican waters, sardine are distributed off the Pacific coast of Baja California (including the Bahía Magdalena bay system) and within the Gulf of California (Figure 1). Fisheries generally operate within 15 km of the coast in both regions. There is limited inflow of fresh water, and the oxygen isotope composition of precipitation is relatively uniform (Rozanski et al., 1993). Salinity exhibits limited variation off the Pacific coast of Baja California, within Bahía Magdalena and in the central Gulf of California, although there are differences of up to 2 units between these areas (Álvarez-Borrego et al., 1975; Bray, 1988; Goericke et al., 2004, 2005).
|
The salinity off Baja California typically ranges from 33.5 to 34.5 (Goericke et al., 2004, 2005), which implies a variation of ca. 0.5
in w (Craig and Gordon, 1965). The Bahía Magdalena system is located at the southern extent of the California Current and comprises two main bays, Bahía Magdalena proper and Bahía Almejas, as well as smaller lagoons. It is a shallow (maximum depth 40 m) system that usually exhibits inverse estuarine conditions (Álvarez-Borrego et al., 1975). Sardine are captured mostly within Bahía Almejas. The bay is a semi-enclosed system with virtually no fresh-water inflow. However, the salinity in the area in which sardine are fished usually exhibits limited temporal and spatial variation because of the extensive exchange of bay waters with coastal waters (between 34 and 35; Álvarez-Borrego et al., 1975). These salinity values are 1 unit higher than those reported farther north in the Mexican Pacific (Goericke et al., 2005).
The Gulf of California is considered to be the only evaporative basin in the Pacific, and is characterized by a relatively higher surface salinity (Roden, 1958). The average salinity in the central Gulf (between Guaymas and Santa Rosalía) is 35.5 (Bray, 1988; Romero-Centeno, 1995). Mixing and circulation processes are considered more important in controlling w variations in the region than precipitation or fresh-water inflow (Juillet-Leclerc and Schrader, 1987; Bernal and Carriquiry, 2001). Vertical variations in salinity average 0.5–0.8 between the surface and a depth of 200 m (Romero-Centeno, 1995), which corresponds to a w of 0.5.
Sample collection and processing
In all, 12 groups of young sardine were collected within 1 year from various locations along the western coast of the Baja California peninsula and within the Gulf of California (Figure 1, Table 1). Fresh sardine were obtained from fishery landings from Bahía Magdalena (Puerto San Carlos, Baja California Sur), Ensenada (Puerto El Sauzal, Baja California), and Tastiota and Isla Pájaros (Puerto de Guaymas, Sonora). Other samples were obtained by taking advantage of research cruises or local small-scale fishery activities. Samples from locations in the Pacific were collected twice within a 1-year period to examine temporal variation; unfortunately, we were unable to collect multiple samples at a given location in the Gulf (Table 1).
|
Immediately following sardine collection, standard length (SL) was measured and both sagittal otoliths were removed and placed in vials. In the laboratory, each otolith was washed in 15% H2O2 buffered with NaOH, and sonicated for 4–5 min to remove organic matter and surface contaminants. Otoliths were briefly immersed in dilute (1%) HNO3 and rinsed twice with double deionized water in preparation for isotopic analysis and trace element determinations. Finally, otoliths were dried in a laminar flow hood and stored dry. Because previous studies have not found significant differences in
18Ooto values among right and left otoliths (Iacumin et al., 1992; Thorrold et al., 1997), otoliths were selected randomly for isotopic analysis.
Age estimates
The age of each sardine reflects the amount of time each fish has been exposed to the environment, and hence the time over which whole otolith values of 18O integrate their composition. The otolith that was not used for isotopic analysis was fixed to a slide using Krazy Glue, and gently polished using fine-grit (14.5 and 6.5 µm) sandpaper. The age of each fish was estimated based on the pattern of deposition of the annual growth increments observed along the posterior margin. An otolith with a single complete opaque ring represents a fish that is at least 6 months old, a complete opaque and hyaline ring represents a fish 1 year old, and otoliths with two opaque and two hyaline rings reflect fish with 2 years of growth (Barnes and Foreman, 1994; Yaremko, 1996; Quiñonez-Velásquez et al., 2002). Daily increment counts were not performed because they cannot be distinguished readily in sardine that have overwintered and/or are older than 8–10 months (M. Takahashi, pers. comm.).
18O determinations
Otoliths were sent to Iso-Analytical Ltd in the UK for analysis (www.iso-analytical.com). Mean otolith weight was 0.58 mg (±0.32 mg s.d.). Each otolith was cleaned in pure (99.9%) helium, covered with pure phosphoric acid, and allowed to react at 85°C for 1.5 h to release CO2. The oxygen isotope composition of CO2 was analysed in a continuous flow isotope ratio mass spectrometer. Secondary standards, including limestone NBS-19 and IA-R022, were also analysed during sample runs. Precision (±1 s.d.) of secondary standards was 0.05 (n = 4) and 0.19 (n = 19), respectively. Assuming a –0.2 change per °C (Campana, 1999), temperature estimates derived from 18Ooto values should be accurate to ±1°C. Isotopic ratios are reported using standard notation:
|
|
where is the isotopic ratio and R the ratio of the heavy to light isotope (18O/16O) relative to V-PDB.
One-way analysis of variance (ANOVA) was used to evaluate whether there were differences in 18Ooto values among Pacific and Gulf samples and within each of these regions. The assumptions of normality and homoscedacity were evaluated using the Kolgomorov–Smirnov and Cochran's test, respectively. Post hoc multiple comparisons were performed using Scheffe's test. Isotopic data from the Gulf that did not comply with the assumption of homogenous variance were analysed using the non-parametric Kruskal–Wallis and Newman–Keuls multiple comparison tests. Data were analysed using Statistica 6.0.
Calculation of predicted carbonate 18O values
There are limited measurements of w values within our study area (Craig and Gordon, 1965; Rozanski et al., 1993; Schmidt et al., 1999). Hence, we used salinity as a proxy for w values to calculate the theoretical isotope composition of sardine otoliths for each area. For Ensenada, Isla de Cedros, and the coastal waters just north of Bahía Magdalena, we generated seasonal average temperature and salinity vertical profiles using CTD data collected during eight cruises made between January 2003 and October 2004 (Figure 1). The cruises were performed quarterly (winter, spring, summer, and autumn) by the Investigaciones Mexicanas de la Corriente de California (IMECOCAL) program (imecocal.cicese.mx; see Goericke et al., 2004, for the standard cruise plan). As otoliths were collected from sardine captured between May 2004 and February 2005, the cruise data overlapped with the lifespan of the sardine. The CTD stations included in the analysis for the Ensenada area lie within the area 31°11'N 117°47'W–31°41'N 116°47'W and 30°01'N 117°01'W–30°28'N 116°09'W, whereas for Isla de Cedros, the area included in the analysis is delimited by 28°18'N 115°55'W–28°48'N 114°56'W and 27°09'N 115°11'W–27°15'N 114°59'W (Figure 1). To examine the latitudinal pattern in temperature and salinity, we also derived average vertical profiles for the southern area of IMECOCAL cruise plan, which is located just north of Bahía Magdalena (25°25'N 114°05'W–25°55'N 113°08'W and 24°50'N 113°44'W–25°19'N 112°46'W). Salinity and temperature depth profiles were generated for the top 200 m by averaging 9–12 stations within each area.
For Bahía Magdalena, we used seasonal salinity and temperature measurements reported by Robinson et al. (2000), based on measurements made in 1996 (range 34.2–35.0 and 16–23°C). However, our SST series indicated temperatures of up to 25°C during the study period (see below), and predicted oxygen isotope ratios were calculated using that maximum temperature.
Detailed seasonal temperature and salinity data were unavailable for the Gulf of California for the period of our study. Hence, average temperature and salinity vertical profiles were calculated based on Romero-Centeno (1995), who analysed data collected between 1939 and 1990 (a period of 51 years). Anomalous years in which El Niño and La Niña conditions prevailed were excluded from the analysis. Vertical profiles were derived seasonally and centred on February, May, August, and November. Following Romero-Centeno (1995), vertical profiles were plotted for the Canal de Ballenas area, the area east of Canal de Ballenas, and the southern area of the central Gulf, which lies between Santa Rosalía and Guaymas (Figure 1).
We used the empirical 18O–salinity relationship of Craig and Gordon (1965) for north Pacific surface waters (w = –18.5+0.54 Salinity) to estimate w. This equation has a similar slope to that reported by Paul et al. (1999) for mid-latitudes (0.50) and LeGrande and Schmidt (2006) for the North Pacific (0.44). Campana's (1999) equation was then used to calculate equilibrium 18O values: 18O–w = 3.71–0.206 T (°C). This equation is derived from the empirical relationship documented for inorganic calcite by Kim and O'Neil (1997), and includes a correction for the difference in fractionation observed during the formation of calcite and aragonite (aragonite is 0.6 more enriched than calcite relative to the water). The slope of this relationship does not differ from those obtained in controlled laboratory studies for a variety of fish species (Radtke et al., 1996; Thorrold et al., 1997; Høie et al., 2004a). However, small variations in the value of the intercept have been reported, which may be attributed to species-specific fractionation (Høie et al., 2004a). Given that experimentally derived fractionation factors are not available for Pacific sardine, we decided to use Campana's (1999) equation.
Because w () is expressed relative to Standard Mean Ocean Water (SMOW), w values reported in the literature were corrected (–0.2) to account for the difference in standards used to report oxygen isotopic values of water and biogenic carbonates (Epstein et al., 1953). We designated 12°C as a conservative lower temperature limit for the vertical distribution of Pacific sardine, and focused on predicted isotopic values for otoliths formed above the 12°C isotherm.
Estimates of SST
SST values were obtained by analysing satellite images acquired through the National Oceanic and Atmospheric Administration's (NOAA) CoastWatch West Coast Regional Node (http://coastwatch.pfel.noaa.gov/sst_comp_low.html). Estimates are based on data gathered with Advanced Very High Resolution Radiometers (AVHRR) on NOAA polar-orbiting weather satellites calibrated with in situ measurements, and are indicative of the temperature in the top 0.3–1 m (Martin, 2004). Composite monthly temperatures were compiled for the period October 2002–May 2005. For Ensenada, Isla de Cedros, and Bahía Magdalena, SST estimates were derived using data obtained with High Resolution imaging (1.1 km per pixel). Data for Gulf of California localities were available in lower spatial resolution (2.5 and 5.0 km per pixel) for the same period. It is likely that Pacific sardine can migrate during their first year of life (Clark and Janssen, 1945), although the extent of such migrations has not been determined. Hence, we chose to estimate age-specific average local SSTs for a quadrant around each capture location. Given the absence of information regarding the potential for movement in young sardine, we chose quadrant sizes based on local oceanographic characteristics and considering the distance between locations in which sardine were caught. However, we also calculated age-specific average SSTs for a single point (the capture coordinates for each group of Pacific sardine, not shown) to assess the magnitude of the difference between local and point SST estimates. This allowed us to examine the sensitivity of age-specific SSTs to the area used. Comparison of local SST estimates with point estimates were performed separately for the Pacific and Gulf, using correlation and residual analysis.
For Ensenada and Isla de Cedros, a 2° x 2° quadrant around each capture location was arbitrarily selected as representative of the potential area inhabited by the sardine and the average monthly temperature. For Ensenada, mean monthly SSTs were obtained within the area 30–32°N and 116–118°W. For Isla de Cedros, we used the area 27–29°N and 114–116°W. In the specific case of Bahía Magdalena, the average SST for the entire bay system was calculated (23–25°N and 111–113°W). For the samples collected in the Gulf of California, a 1° x 1° quadrant was selected around each of the capture locations. All sites were in the central region of the Gulf (27–30°N and 110–114°W; Figure 1).
Temperature reconstruction and comparison with local SSTs
We assumed a constant salinity and used literature-derived average values of w to calculate Toto using Campana's (1999) equation (18O–w = 3.71–0.206 T), where T is in °C. The value of w applied to samples collected in the Pacific (–0.31, relative to VPDB) was obtained from Grossman and Ku (1986), from water determinations near Santa Barbara Island (33°28'20''N 119°02'00''W, 77 m deep). This area is within the California Current System, which extends south along the Baja California peninsula. A value of w of –0.31 is similar to values reported previously for the area (Craig and Gordon, 1965; Schmidt et al., 1999). Data were not available for Bahía Magdalena, so we used a value obtained from Grossman and Ku (1986) as a first approximation. For the central Gulf of California, a w of 0.04 (relative to VPDB) was derived by averaging literature values. Surface water values were available for Golfo de Santa Clara and Puerto Peñasco, which are in the northeastern Gulf (Payne et al., 1978), and off the coast of the Bahía de Kino area, located near Guaymas along the eastern side of the central Gulf (Rangel-Medina et al., 2002). The values we used are similar to those reported by Juillet-Leclerc and Schrader (1987) for surface waters in the Guaymas Basin (0.00 relative to VPDB). The w value applied to the Gulf is 0.35 more enriched than the value used for the Pacific. A 1 difference in w between the Pacific and Gulf would have been predicted based on salinity. However, the relationship between w and salinity has yet to be examined for the central Gulf of California. We therefore decided to use published isotopic values rather than inferring w from salinity measurements to estimate Toto. Lastly, the sensitivity of Toto to variations in salinity (and by inference w) within our study was examined by re-calculating temperatures for a range of salinity representative of the variation in the area.
To infer whether there was evidence of migration to the quadrants used to calculate local SSTs, we assumed that SST measurements were reflective of the temperature of the mixed layer and the vertical habitat of the Pacific sardine (Jacobson et al., 2005), and compared local SSTs with Toto values. As we did not know the precise age of each sardine, SSTs were calculated for a range of ages (see Results). Finally, linear regressions were used to examine the relationship between age-specific SST estimates and 18Ooto values. Data for the Pacific and Gulf of California were analysed separately.
Otolith growth model
To examine the potential effect of monthly variations in otolith growth rates on 18Ooto values (and hence our back-calculated temperature estimates), we constructed an otolith growth model. Otolith and somatic growth rates of Pacific sardine are highly variable, even among fish of the same cohort (Butler et al., 1996; Quiñonez-Velásquez et al., 2000, 2002). This precludes application of a single otolith growth model to Pacific sardine in general. Rather, our intention was to develop a realistic model to examine the magnitude of the effect of growth rate variations and seasonal temperature fluctuations on 18Ooto. We used the model in conjunction with average monthly SST series to calculate 18Ooto values with and without weighting for monthly growth rate variations.
Carbonate accretion rates have not been measured for Pacific sardine. Therefore, the radius (focus to posterior margin) was used as a proxy for carbonate weight by assuming a direct relationship between these variables. We created a virtual otolith that grew in 1 µm increments (a single growth step), and calculated the corresponding SL (in mm). Butler (1987) examined the relationship between the radius and SL for sardine collected off southern California, and reported an allometric relationship. Early juvenile fish (ca. >34.5 mm SL) exhibited slower otolith growth relative to SL than smaller sardine. Hence, Butler (1987) presented two linear models that we used to calculate SL as a function of the otolith radius:
|
|
|
|
To estimate the time elapsed between successive growth steps, we used size-specific growth rates (mm d–1) reported by Butler (1987) for Pacific sardine up to 85 mm SL, which were derived from daily increment counts. Growth rates were visually estimated at intervals of 5 mm SL from Butler's (1987) best-fit curve. The values reported by Butler (1987) are within the range reported for young sardine in other studies (Quiñonez-Velázquez et al., 2000). For sardine >85 mm SL and up to 1 year of age (140 mm SL in our model), we used the von Bertalanffy growth equation reported by Butler et al. (1996):
|
|
where Lt is length-at-age, L
size at infinity (205.4 mm SL), K the growth coefficient (–1.19 year–1), t the age (years), and t0 the estimated size at age 0 (set to 0). As mentioned by Butler et al. (1996), L underestimates the maximum size of Pacific sardine. However, the value of K is consistent with the fast growth rate exhibited by young sardine. The percentage monthly increase in otolith radius relative to the radius at age 1 was calculated by assuming that each month lasted 30.5 d.
To evaluate the potential effect of seasonal temperature fluctuations and age on whole otolith 18O values, monthly SST data for a 12-month period (January–December 2004) were selected, and monthly and annual isotope ratios were calculated. Series from both the Pacific (Ensenada) and Gulf of California (Cabo Tepoca) were chosen, because the seasonal amplitude of SSTs differs widely among regions. For each monthly SST, Campana's (1999) equation was used to predict 18Ooto. Salinity was assumed to be constant, and the w values applied to the back-calculation of Toto were used. Whole otolith annual 18Ooto values were calculated by (i) averaging monthly isotopic values (unweighted mean), and (ii) by weighting monthly 18Ooto values by the estimated percentage monthly increase in otolith radius. Weighted and unweighted means were calculated for fish ranging from 6 to 12 months old. The final age was always set to December; fish born in January were 12 months old, and those born in July were 6 months old.
|
Results |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionConclusionsReferences |
|---|
The average SL of the sardine captured in the Pacific ranged between 128 and 163 mm (for samples from Isla de Cedros in February 2004 and Ensenada in December 2004, respectively), and for the Gulf of California between 121 and 152 mm (for samples from Santa Rosalía and Tastiota, respectively; Table 1). Sardine captured in the Pacific exhibited both an opaque and a hyaline ring, which indicated a minimum age of 12 months. However, a third ring of various widths was visible, which indicates an age between 12 and 18 months (Table 1). All sardine from the Gulf of California exhibited one complete opaque ring corresponding to the first sixth months of life, and a second ring of varying width. A third ring was not observed, so the age interval assigned to sardine captured in the Gulf was between 6 and 12 months (Table 1).
18O values of sardine otoliths
The 18Ooto values of individual fish captured at each location were highly variable (Figure 2, Table 1). In the Pacific, we observed a spread of 2.03 in isotopic values of otoliths of sardine collected at a single sampling location. For the Gulf, the maximum spread was 1.65. The full range of isotopic values for all localities in the Pacific was –2.45 to 1.10 (an absolute spread of 3.55) and in the Gulf between –1.51 and 0.84 (absolute spread 2.35).
|
The maximum average
18Ooto value for samples collected in the Pacific was 0.31 for samples collected in Ensenada in June 2004. The minimum average value of –1.66 corresponded to sardine captured in Bahía Magdalena in May 2004 (Table 1). The range in average values among sampling locations within the Gulf was smaller (ca. 0.7); the maximum value (average 0.16) was for fish collected in Canal de Ballenas in May 2004, and the minimum value (average –0.57) for samples collected at Tastiota in May 2005. There were no significant differences between isotopic values from the Pacific and the Gulf of California (Kruskal–Wallis H = 1.26, p = 0.289). However, there were significant differences among samples captured in different locations within each region (ANOVA F = 17.25, p < 0.0001 for Pacific samples, and Kruskal–Wallis H = 21.99, p = 0.0005 for Gulf samples). No significant differences were found in 18Ooto values of the samples collected 4–6 months apart in Ensenada and Isla de Cedros. In contrast, there were significant differences in the isotopic values of sardine collected in Bahía Magdalena in May and October 2004 (Table 2).
|
Measured vs. predicted
18Ooto valuesThe seasonal vertical temperature structure of the water column in the Pacific differs substantially from the Gulf of California. In the Pacific, the absolute temperature differences between the surface and the 12°C isotherm was
4–6°C in winter and 8–12°C in summer (the difference was greater at lower latitudes; Figure 3). In the Gulf, this range is ca. 3°C in winter and up to 14–16°C in summer and early autumn (Figure 4). In the Pacific, the 12°C isotherm was located at a maximum depth of 100 m off the coast of Ensenada, and at up to 150 m deep off the southern extent of the peninsula. The 12°C isotherm in the central Gulf of California extended to depths of 300 m.
|
|
In the Pacific, vertical and seasonal differences in salinities above the 12°C isototherm were
ca. 0.4 (Figure 3). The latitudinal range off the coast of Baja California was ca. 0.5 of salinity (maximum 0.8), and the maximum salinity was 34.4. Average salinities in Bahía Magdalena (34.2–35.0) were either similar or marginally higher than those found off the southern Baja California peninsula (Álvarez-Borrego et al., 1975; Robinson et al., 2000). For the Gulf, the average seasonal and vertical salinity gradient above the 12°C isotherm was usually 0.5 (Figure 4).
Comparison of the total range of location-specific predicted otolith 18O values with 18Ooto measurements indicated that, except some samples from Bahía Magdalena, most values were within the range of predicted values (Figure 5). However, two otoliths from sardine captured in Ensenada and one from Isla de Cedros exhibited isotopic values outside the range of predicted values. Interestingly, the isotopic values of several samples collected in Bahía Magdalena in May lay outside the range of predicted 18O values for all locations in the Mexican Pacific, including Bahía Magdalena. This was not the case for the samples collected within the bay in October.
|
SST estimates
In the Pacific, there was a marked latitudinal gradient in SST (Figure 6); the absolute difference between Ensenada and Bahía Magdalena was as high as 7°C. In the Gulf of California, seasonal fluctuations were stronger than in the Pacific (up to 16°C), and monthly SSTs were more similar among locations in the Gulf than in the Pacific (Figure 6).
|
Local SSTs exhibited greater variability in the Gulf than the Pacific because of the greater amplitude in monthly SST (Figure 7). For the Pacific, local SST estimates ranged between 15.5°C and 16.3°C for the samples collected in Ensenada in June 2004, and between 20.6°C and 21.8°C for sardine captured in Bahía Magdalena in October 2004. In general, local SST estimates for sardine captured in the Gulf of California were higher than for the Pacific (Figure 7).
|
Comparison of local SSTs with point estimates for the Pacific (using the full range of ages and all capture locations) yielded a good correlation (y = 1.21 x – 3.39; r = 0.99; n = 34). The standard deviation and range of the residuals (0.38 and –0.55°C to 0.95°C, respectively) showed that the difference between the two estimates was limited. However, the slope was not equal to 1; point estimates were ca. 1.5°C higher than local SSTs at temperatures >20°C. For the Gulf, the correlation was also good (y = 1.12 x – 0.80; r = 0.94; n = 41). However, the standard deviation and range of the residuals showed larger differences between local and point estimates than in the Pacific (s.d. = 0.91, range –1.51°C to 1.70°C). As indicated by the value of the intercept, local SSTs were higher than point-based estimates.
There was a significant negative relationship between 18Ooto and local SST (average) estimates for the samples collected in the Pacific off Baja California (18Ooto = 3.17–0.20 SST; r = –0.50; p < 0.0001; Figure 8). Although the slope and intercept of the relationship was similar to that predicted by Campana's (1999) equation (0.206 and 3.71, respectively), the isotopic values of otoliths of sardine collected in Bahía Magdalena in May and October 2004 did not fall close to the relationship expected based on thermodynamic considerations. This was the case even when Campana's (1999) equation was plotted using a higher w value (to reflect the higher salinity of Bahía Magdalena). The least squares analysis of samples from the Gulf yielded a lower correlation coefficient than for the Pacific, although the relationship was significant (18Ooto = 1.94–0.10 SST; r = –0.33; p = 0.0137). More importantly, the slope and intercept of the relationship was different from that predicted based on thermodynamic considerations (Figure 8), and the temperatures predicted on the basis of isotopic values were generally lower than local SST estimates.
|
Back-calculation of Toto and comparison with local SST
Because of the high level of variability in
18Ooto values of individual sardine captured at a single location, Toto estimates were highly variable. The absolute difference between individual maximum and minimum values for any one location ranged from 4°C to 10°C (Figure 7). Salinity variations of ±0.25 units (a 0.5 range, see salinities above the 12°C isotherm in Figures 3 and 4) resulted in ±0.66°C in back-calculated temperature estimates. Comparison of the range of local SSTs with values of Toto on a location-specific basis indicated that some sardine fell outside the range of expected values, even when temperature estimates were re-calculated to take into account a larger range of salinities. Average Toto values for Ensenada were 15°C and 17°C for sardine captured in June and December 2004, respectively. For Isla de Cedros, average Toto estimates were higher than for Ensenada (19°C and 20°C for fish caught in February 2005 and October 2004, respectively). There was a large difference between Toto estimates of sardine captured in Bahía Magdalena during October and May 2004 (mean values 18°C and 25°C, respectively). For the Gulf of California, Toto estimates indicated average temperatures of 17.4°C for the samples from the Canal de Ballenas on the western side to ca. 21°C for locations along the eastern Gulf. None of the Toto estimates derived from isotopic analyses indicated temperatures lower than 12°C, which is consistent with the lower temperature limit of Pacific sardine larvae (Lasker, 1964).
The ranges of local SSTs for sardine collected in Ensenada and Isla de Cedros were within the range of Toto estimates (Figure 7). However, Toto exhibited substantially more variability than SSTs, even though the latter capture a range of ages. There was no correspondence between Toto and SSTs for sardine collected in Bahía Magdalena. Toto values of sardine collected in May 2004 were mostly higher than local SSTs, whereas those from October 2004 were much lower (Figure 7).
The seasonal variability in monthly SSTs in the Gulf of California led to a broader range of local SSTs than for the Pacific (Figure 7). Although Toto estimates of many sardine were generally within the range of local SST (except for Punta Borrascosa), many sardine had lower Toto values.
Otolith growth model
Size-specific growth rates (in mm d–1) were highest during the early juvenile stage (ca. 30–60 mm SL) and decreased after 60 mm SL (Figure 9a). The relationship between otolith radius and estimated age (in days) was relatively continuous, despite the merging of two datasets (Figure 9b). The radius of the otolith at age 1 was 1.260 mm. The percentage monthly increase in otolith radius was greatest between 2 and 4 months old (14–18% per month; Figure 9c), reflecting the fast growth rates documented by Butler (1987) during the early juvenile stage. Finally, the percentage monthly increase in otolith radius was relatively consistent between ages of 6 and 12 months (ca. 4–6% per month).
|
Comparison of weighted and unweighted
18O values calculated using the SST series for the Pacific and Gulf indicated a maximum difference of –0.69 and –0.37, respectively (Figure 9e and f). For both simulations, the maximum difference was at 12 months. In fish 6 months old, failing to account for monthly differences in otolith growth yielded differences of 0.32 and –0.45 in whole otolith isotopic values (Pacific and Gulf, respectively).
For the Pacific, using the SST series for "older" fish led to weighted isotopic values that were more enriched owing to the lower temperatures early in the year. For the Gulf, the weighted isotopic values for younger fish (6–8 months old) were lower than unweighted values. This is attributed to the higher temperatures in the SST series during the second half of the year. Lastly, the potential effect of age differences among fish was compared using the unweighted SST series. Comparison of the isotopic composition of fish 6 and 12 months old (the most extreme and unlikely age difference) indicates a maximum absolute difference of 0.30 for the Pacific and 0.71 for the Gulf. This corresponds to differences in temperature of 1.5°C and 3.5°C, respectively. Hence, even a 6-month difference in age cannot account for the high level of variation in 18O values.
|
Discussion |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionConclusionsReferences |
|---|
Spatial and temporal variability in
18Ooto valuesThe range of salinities found in the Pacific off Baja California and within the Gulf of California is
1 unit. This is equivalent to a 0.5 shift in w (Craig and Gordon, 1965) and consequently otolith 18O values (2.5°C). There is no evidence to suggest mixing between sardine from the central Gulf of California and the northern part of the Baja California peninsula, although there may be mixing among subpopulations towards the southern extreme of the species distribution (Smith, 2005). Hence, the large spread we observed in 18Ooto values of young sardine captured simultaneously implies the temperature regime to which they were exposed differed widely (Ayvazian et al., 2004). There were no significant differences between the 18Ooto values of sardine collected in the Pacific and Gulf of California, so precluding the use of oxygen isotope ratios to discriminate between sardine from these regions. The spread of isotopic values for a given location in the Pacific was 1.1–2.0, whereas for the Gulf, it was more limited (0.5–1.7). Assuming a constant salinity within each region, this is equivalent to an "integrated" lifetime temperature difference of 5.3–9.7°C for the Pacific and 2.4–8.3°C for the Gulf. This implies considerable mixing within subpopulations during the first year or so of life.
Other studies have also reported high levels of variability in 18Ooto values of conspecifics from a given location (Edmonds and Fletcher, 1997; Begg and Weidman, 2001; Gao et al., 2001). Local variations of up to 0.5 have been reported for the Australian sardine (Sardinops sagax; Edmonds and Fletcher, 1997), 1.4 for the European sardine (Sardina pilchardus; Iacumin et al., 1992), and 1.6 for the Australian herring (Arripis georgiana; Ayvazian et al., 2004). Nevertheless, the absolute range of some of the values measured in this study is somewhat higher than that reported for other, mostly demersal, species (Kalish, 1991a; Iacumin et al., 1992; Newman et al., 2000; Begg and Weidman, 2001).
Population mixing may take place during the highly dispersive larval period, among fish from different schools and among fish of various ages. It may also reflect variations in the pattern of habitat utilization: different depths, inshore/offshore areas, and a variety of latitudes. The processes leading to spatial mixing of individual Pacific sardine during their first year of life have not yet been identified, despite their obvious relevance for estimating the relative contribution of different nursery areas or spawning groups to the production of recruits. However, based on tagging studies, Clark and Janssen (1945) reported that large sardine (>250 mm total length, TL) could migrate from southern California to British Columbia, Canada, in 5 or 6 months, although smaller sardine (<150 mm TL) did not seem to exhibit such rapid and long distance migrations. However, Clark (1947) later suggested the presence of horizontal and latitudinal mixing in sardine <1 year old based on the variability of vertebral counts among fish captured in different areas; the number of vertebrae is a function of temperature during early development. Our data are consistent, although not necessarily exclusively explained by Clark's (1947) proposition.
Differences in vertical habitat utilization could also contribute to the variation in 18Ooto values. However, vertical distribution alone is unlikely to be the source of the observed broad range of isotopic values. A 1.5 difference in isotopic values among sardine from a given location would imply that they would have permanently resided at very different depths. Off Ensenada, for example, this would imply that one fish would have consistently inhabited shallow water (top 20 m), whereas another would have resided at 100 m. Given that sardine exhibit diel vertical migration (Krutzikowsky and Emmett, 2005) and schooling behaviour, we find this to be an unlikely situation.
In whole otoliths, isotopic signals (and hence Toto) will be biased towards the temperature and salinity conditions prevalent during the time when most of the aragonite precipitated. If, for example, most aragonite was deposited under warm conditions as a result of faster somatic and otolith growth rates, whole otolith 18Ooto values would tend to be lower. Variations in the rates of otolith accretion among sardine collected simultaneously could therefore account for some of the intra-location variability in otolith 18Ooto values observed in this study.
The otolith growth model constructed from empirical data suggested substantial variation in monthly otolith growth rates (range 4–18% per month). By an age of 5 months, our model otolith was estimated to have reached 65.5% of the radius at age 1. This is consistent with the fast growth rates of larvae and early juvenile sardine (Watanabe and Saito, 1998), and the observation that the first hyaline seasonal ring, which corresponds to the period of fast summer growth during early life, is wider than the opaque ring, which represents winter growth (Yaremko, 1996; Quiñónez-Velázquez et al., 2002). Hence, a disproportionate amount of otolith growth may take place during the first few months of life.
Despite the monthly variations in otolith growth rate, comparisons of unweighted and weighted 18Ooto annual estimates yielded a maximum absolute difference of 0.69, equivalent to 3.3°C. In addition, some variation in whole otolith 18O values would be expected as a result of age differences (Figure 9e and f).
Our otolith growth model assumed a linear relationship between otolith radius and volume; growth of the latter would logically be more reflective of carbonate accretion rates and would imply a cubic relationship between radius and volume. Further, the relationship between the radius and weight of Pacific sardine otoliths could be allometric (Butler et al., 1996). If carbonate accretion rates are higher during the first few months of life, our results could underestimate the difference between weighted and unweighted 18Ooto annual estimates. Nevertheless, our model suggests that it is unlikely that variations in otolith growth rate (and age) among fish captured simultaneously could be solely responsible for the broad range of 18Ooto values (and hence back-calculated temperature estimates). Further research needs to be conducted to evaluate otolith accretion rates in Pacific sardine in relation to environmental conditions such as individual growth rate and temperature (Fey, 2006).
Evaluation of local residence by young sardine
We tested the null hypothesis that young sardine had resided near their capture location throughout their life using two independent, complementary strategies: (i) comparing 18Ooto values with predictions derived from local temperature and salinity conditions, and (ii) assuming a constant salinity in a given region and comparing back-calculated Toto estimates with age-specific local SST estimates. The predicted oxygen isotope ratios for a given location take into account both seasonal and vertical variations in salinity and temperature, and are hence a conservative estimate that should encompass all possible isotopic signatures exhibited by resident sardine. However, this approach does not strictly couple isotopic measurements with the temperature and salinity experienced by individual sardine, but rather serves only as an envelope of possible values. In contrast, the comparison between Toto and local SSTs is coupled in time, but relies on three assumptions: that SST estimates are representative of the vertical temperature habitat of Pacific sardine, that w is constant, and that whole otolith 18Ooto values reflect the lifetime "average" temperature to which sardine were exposed (i.e. that there is a linear relationship between otolith accretion rate and time, as discussed above).
Comparison of 18Ooto with predicted isotopic values
For most locations, 18Ooto values were within the range of predicted values. This suggests that those sardine could have resided within that area during their entire life, although the possibility of partial residence in nearby areas with similar temperature and salinity regimes cannot be excluded. However, a few fish collected in Ensenada and Isla de Cedros, and most of the samples from Bahía Magdalena collected in May, exhibited isotopic values outside the range of predicted values. This is strongly suggestive of partial residence in other areas.
The relatively low 18Ooto values of sardine collected within Bahía Magdalena in May 2004 are of particular interest. A potential cause could be unusually low values of w attributable to high rainfall (rainwater is typically depleted in 18O relative to seawater). However, Bahía Magdalena is located in an extremely arid region in which there is little or no rainfall, except during tropical cyclones (Hastings and Turner, 1965; Englehart and Douglas, 2001; Farfán and Fogel, 2007). In 2004, tropical activity resulting in substantial rainfall (>10 mm) over the southern peninsula was only recorded in September (Farfán and Fogel, 2007), after we had collected the sardine. In 2003, two tropical cyclones came ashore along the southern extent of the Baja California peninsula: Ignacio (24 August) and Marty (22 September; Farfán and Cortez, 2005). Total rainfall in the region near Bahía Magdalena was 50–100 mm during Hurricane Marty (Farfán and Cortez, 2005) and 200 mm during Hurricane Ignacio (L. M. Farfán, pers. comm.). The relationship between precipitation and fresh-water inflow into Bahía Magdalena is unknown. However, for the w of the water to vary, a tremendous amount of fresh water would have to enter the system. Given the limited duration of tropical cyclone activity (days), the high exchange of water between Bahía Magdalena proper and the coast (Álvarez-Borrego et al., 1975; Lluch-Belda et al., 2000), and the fact that relatively low whole otolith isotope ratios would only occur following prolonged exposure to water depleted of 18O, our observations from Bahía Magdalena May 2004 cannot be attributed to high fresh-water inflow. Those sardine clearly reflect a subpopulation that differs from the other sardine sampled in the Mexican Pacific.
Comparison of Toto with local SST
Many authors have reported a strong relationship between SSTs and sardine abundance, catches, and migration patterns (e.g. Lluch-Belda et al., 1991; Hammann et al., 1998; Nevárez-Martínez et al., 2001; Félix-Uraga et al., 2004). However, satellite-derived SSTs only reflect the surface layer, and hence do not fully reflect all available habitats to Pacific sardine. To our knowledge, there is no information regarding the vertical distribution of Pacific sardine ca. 1 year old. Off the coast of southern California, late larvae and early juveniles are much more abundant in the top 50 m than in the 50–200-m depth strata (W. Watson, pers. comm.). Sardine typically feed on phytoplankton and zooplankton (Emmett et al., 2005), and they tend to inhabit waters near the surface (Jacobson et al., 2005). However, like other clupeids, sardine exhibit vertical migration; they are more abundant near the surface at night (Blaxter and Holliday, 1963; Nevárez-Martínez et al., 2001; Krutzikowsky and Emmett, 2005). Pacific sardine are caught mainly at night at depths <35 m (Nevárez-Martínez et al., 2001; Krutzikowsky and Emmett, 2005). One-year old fish probably do not permanently inhabit deeper, colder waters. However, they may temporarily move to deeper water to feed, particularly if the temperature near the surface is very high (M. O. Nevárez-Martínez, pers. comm.).
For most Pacific locations, Toto estimates were both lower and higher than the estimated range of local SSTs. If Toto was primarily acquired from within the thermocline, rather than near the sea surface, then differences in the pattern of vertical habitat utilization should result in back-calculated temperature estimates consistently lower than local SSTs (because temperature decreases as a function of depth). Although differences in vertical habitat utilization may contribute to the observed variation in Toto, they cannot account for the full range of measured isotopic values for a given location. Further, differences in salinity (±0.25 units) had a limited impact on back-calculated estimates of temperature. On the other hand, mixing of fish (inshore/offshore or latitudinal) could account for the extent of variation in the back-calculated estimates of temperature for sardine captured simultaneously.
Our interpretation of the comparisons of Toto and SST comparisons in terms of local residence are sensitive to the quadrant sizes used to generate monthly mean SSTs. Comparison of point quadrant-based estimates indicated differences in monthly SSTs of <1°C for the Pacific and up to 3°C for the Gulf. Classifying an individual fish as a potential resident or non-resident of a given area will therefore depend at least partly on the area selected. Here, we interpret our results strictly in terms of the specific areas we designated for generating local SST estimates. Further, it is important to realize that owing to our relatively small sample sizes, interpretation of our results is not intended to convey quantitative estimates of mixing among subpopulations of Pacific sardine in the Mexican Pacific.
Average local SSTs for sardine collected in Ensenada were similar to average Toto (16°C). However, the great variability in individual Toto (12–18°C for June 2004, 11–21°C for December 2004) indicates that they experienced different thermal histories. The range of Toto values found for sardine collected in Ensenada is consistent with the presence of the so-called northern subpopulation (Smith, 1981, 2005), whose main spawning area extends from San Francisco, USA, to Punta Eugenia, Baja California, and which is typically fished at temperatures between 15°C and 21°C (Félix-Uraga et al., 2004).
The average local SST estimated for fish from Isla de Cedros (18.2°C) was slightly lower than the average Toto (19.3°C). At least a few fish seem to have inhabited areas south of Isla de Cedros, where average temperatures are higher. This is consistent with the findings of Félix-Uraga et al. (2004), who identified a temperate group of sardine that is fished commercially at SSTs of 17–22°C. That group inhabits the coast of Baja California and California, and migrates between 25°N and 33°N (Félix-Uraga et al., 2004).
The average local SSTs estimated for sardine captured in Bahía Magdalena in May 2004 was 20.4°C, whereas the average Toto was ca. 24.0°C. Single temperature estimates were as high as 28°C. If sardine captured within the bay had resided solely within the system, they would have been exposed to winter temperatures of 17–18°C for a substantial part of their life, and we would expect back-calculated temperatures to be lower. This uncoupling between Toto and local SSTs could indicate that (i) the w value used as a first approximation was erroneous, or (ii) those sardine were not permanent residents of the bay. The higher salinities found within Bahía Magdalena than in the Pacific implies higher w values (and hence temperatures). This would accentuate rather than explain the uncoupling between Toto and local SSTs. As discussed previously, the aridity of the region renders it unlikely that the isotopic composition of the water was overestimated. It is therefore likely (based on our observations and current population models) that those sardine inhabited warmer water. In the southwestern Gulf of California, surface waters are very warm during summer and autumn, regularly reaching up to 29°C (range 19–30°C; Soto-Mardones et al., 1999). The ingress of sardine from the Gulf of California into Bahía Magdalena was noted in Clark's (1947) early studies. More recently, Félix-Uraga et al. (2004) identified a "warm" group that migrates to Bahía Magdalena from the southern Gulf.
The average Toto calculated for the samples collected within Bahía Magdalena in October 2004 was 17.9°C, whereas the local SST was 21.1°C. The discrepancy between local SST and Toto again suggests that at least some of those sardine were not permanent residents. The back-calculated temperature estimates are similar to those typically found in the Pacific off the central Baja California peninsula (Figures 3 and 6). Those sardine could form part of the temperate group identified by Félix-Uraga et al. (2004), which migrate south towards the bay between April and June.
For samples collected within the Gulf of California, Toto estimates were generally within the range or lower than local SSTs (Figure 7). The SSTs estimated in this study are similar to those presented in previous reports (e.g. Nevárez-Martínez et al., 2001). The spawning habitat and migration patterns of sardine in the Gulf of California suggest the preferential use of colder water throughout the life cycle (Hammann et al., 1988). For example, because of persistent upwelling, the western side of the central Gulf is characterized by colder water than the eastern side (López et al., 2006). Based on 18Ooto (and hence Toto) values, two groups seem to be present within the Gulf of California: one on the western side (Canal de Ballenas and Santa Rosalía), and the other along the eastern margin (Cabo Tepoca, Isla Pájaros, and Tastiota; Table 2).
Relationship between 18Ooto values and local SST
The negative relationship between otolith oxygen isotopic values and ocean temperature has been well established (Kalish, 1991b; Radtke et al., 1996; Edmonds and Fletcher, 1997; Thorrold et al., 1997; Campana, 1999; Høie et al., 2004a). Our results also showed a negative relationship for samples collected in the Pacific and in the Gulf of California, although the correlation coefficients were somewhat low (r = –0.50 and –0.33, respectively; Figure 8). To evaluate the suitability of applying our equations to future studies on Pacific sardine, we compared our results with relationships presented in the literature. Campana's (1999) equation, and others that include estimates of w (Grossman and Ku, 1986; Radtke et al., 1996; Thorrold et al., 1997; Høie et al., 2004a) have a slope of approximately –0.2 °C–1. Our data from the Pacific yielded a similar slope (–0.19 °C–1). However, the isotopic values of otoliths of sardine from Bahía Magdalena were either lower (May 2004) or higher (October 2004) than those predicted by the best-fit line. For the Gulf of California, the relationship had a shallower slope (–0.10 °C–1). Other studies that derived relationships based on SST estimates and 18Ooto measurements without considering w also yielded slopes that differ from the empirical value of –0.2 °C–1 (Kalish, 1991a; Edmonds and Fletcher, 1997). Other workers have reported the lack of a significant relationship (Ayvazian et al., 2004). This suggests that studies seeking to derive temperature from otolith oxygen isotopes should take into consideration w rather than relying on empirical relationships between oxygen isotope ratios and temperature.
|
Conclusions |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionConclusionsReferences |
|---|
Oxygen isotope ratios (and back-calculated temperature estimates) of young Pacific sardine collected throughout their range in Mexican waters indicate a substantial amount of mixing at a subpopulation level during the first year or so of life. Further studies seeking to identify the causes of the high variability in 18Ooto of sardine from any one location are certainly warranted, as are investigations of the degree and processes underlying mixing within and among subpopulations.
|
Acknowledgements |
|---|
This study was funded by a UC MEXUS–CONACYT Collaborative Research Grant to SZH and P. E. Smith (SIO–USCD). Supplementary support was also received through the IMECOCAL Program (Investigaciones Mexicanas de la Corriente de California, CONACYT Grant No. G325326–T). We thank T. R. Baumgartner, G. Gaxiola, J. C. Herguera, D. Field, C. Quiñónez, P. E. Smith, and R. Vetter, who provided input throughout the course of the study and manuscript preparation. M. Nevárez-Martínez and other members of the staff of the Centro Regional de Investigación Pesquera–Instituto Nacional de la Pesca in Guaymas provided the means by which samples from the Gulf were collected. M. A. Argote kindly provided us with the oceanographic data from the Gulf of California. J. F. Moreno (CICESE) provided valuable assistance during field sampling, and K. Hill of the Southwest Fisheries Science Center, La Jolla, in validating annual age estimates. We also thank the two reviewers who provided comments that increased the clarity and scope of the manuscript.
|
References |
|---|
|
TopIntroductionMaterial and methodsResultsDiscussionConclusionsReferences |
|---|
-
Álvarez-Borrego S., Galindo-Bect L. A., Chee-Barragán A. Características hidroquímicas de Bahía Magdalena, B.C.S. Ciencias Marinas (1975) 2:94–110.
Andrus F. C. T., Crowe D. E., Sandweiss D. H., Reitz C. S., Romanek E. J. Otolith 18O record of mid-Holocene sea surface temperatures in Peru. Science (2002) 295:1508–1511.
Ashford J., Jones J. Oxygen and carbon stable isotopes in otolith records of Patagonian toothfish (Dissostichus eleginoides). Geochimica et Cosmochimica Acta (2007) 71:87–94.[CrossRef][Web of Science]
Ayvazian S. G., Bastow T. P., Edmonds J. S., How J., Nowara G. B. Stock structure of Australian herring (Arripis georgiana) in southwestern Australia. Fisheries Research (2004) 67:39–53.[CrossRef][Web of Science]
Barnes J. T., Foreman T. J. Recent evidence for the formation of annual growth increments in the otoliths of Pacific sardines (Sardinops sagax). California Fish and Game (1994) 80:29–35.[Web of Science]
Begg G. A., Weidman C. R. Stable 13C and 18O isotopes in otoliths of haddock Melanogrammus aeglefinus from the northwest Atlantic Ocean. Marine Ecology Progress Series (2001) 216:223–233.[CrossRef][Web of Science]
Bernal G., Carriquiry J. D. Stable isotope paleoenvironmental record of a coral from Cabo Pulmo, entrance to the Gulf of California, Mexico. Ciencias Marinas (2001) 27:155–174.[Web of Science]
Blaxter J. H., Holliday F. G. The behaviour and physiology of herring and other clupeids. Advances in Marine Biology (1963) 1:261–393.[CrossRef]
Bray N. A. Thermohaline circulation in the Gulf of California. Journal of Geophysical Research (1988) 93:4993–5020.[CrossRef]
Butler J. L. Comparisons of the larval and juvenile growth and larval mortality rates of Pacific sardine and northern anchovy and implications for species interactions. In: PhD thesis, University of California (1987) San Diego. 240. pp.
Butler J. L., Granados G., Barnes G. J., Yaremko M., Macewicz B. J. Age composition, growth, and maturation of the Pacific sardine (Sardinos sagax) during 1994. CalCOFI Reports (1996) 37:152–159.
Campana S. E. Chemistry and composition of fish otoliths: pathways, mechanisms and applications. Marine Ecology Progress Series (1999) 188:263–297.[CrossRef][Web of Science]
Campana S. E., Neilson J. D. Microstructure of fish otoliths. Canadian Journal of Fisheries and Aquatic Sciences (1985) 42:1014–1032.
Casas-Valdez M. Distribución en tiempo y abundancia de las especies de sardina y macarela en Bahía Magdalena, BCS, México. Investigaciones Marinas CICIMAR (1987) 3:11–30.
Clark F. N. Analysis of populations of the Pacific sardine on the basis of vertebral counts. California Fish and Game (1947) 65.:26. pp.
Clark F. N., Janssen J. F. Movements and abundance of the sardine as measured by tag returns. California Fish and Game (1945) 61.:42. pp.
Craig H., Gordon L. I. Deuterium and oxygen-18 variations in the ocean and the marine atmosphere. In: Proceedings of the Conference on Stable Isotopes in Oceanographic Studies and Palaeotemperatures—Tongiorgi E., ed. (1965) Spoleto, Italy,: Lischi and Figli. 9–130. and Consiglio Nazional delle Richerche, Pisa.
Dansgaard W. Stable isotopes in precipitation. Tellus (1964) 16:436–468.
Degens E. T., Deuser W. G., Haedrich R. L. Molecular structure and composition of fish otoliths. Marine Biology (1969) 2:105–113.[CrossRef]
Dettman D. L., Lohmann K. C. Approaches to microsampling carbonates for stable isotope and minor element analysis: physical separation of samples on a 20 micrometer scale. Journal of Sedimentary Research (1995) A65:566–569.
Devereux I. Temperature measurements from oxygen isotope ratios of fish otoliths. Science (1967) 155:1684–1685.
Edmonds J. S., Fletcher W. J. Stock discrimination of pilchards, Sardinops sagax, by stable isotope ratio analysis of otolith carbonate. Marine Ecology Progress Series (1997) 152:241–247.[CrossRef][Web of Science]
Emmett R. L., Brodeur R. D., Miller T. W., Pool S. S., Bentley P. J., Krutzikowsky G. K., McCrae J. Pacific sardine (Sardinops sagax) abundance, distribution, and ecological relationships in the Pacific Northwest. CalCOFI Reports (2005) 46:122–143.
Englehart P. J., Douglas A. V. The role of eastern North Pacific tropical storms in the rainfall climatology of western Mexico. International Journal of Climatology (2001) 21:1357–1370.[CrossRef][Web of Science]
Epstein S., Buchsbaum R., Lowenstam H., Urey H. C. Revised carbonate-water isotopic temperature scale. Bulletin of the Geological Society of America (1953) 64:1315–1326.
Farfán L. M., Cortez M. An observational and modeling analysis of the landfall of Hurricane Marty (2003) in Baja California, Mexico. Monthly Weather Review (2005) 133:2069–2090.[CrossRef][Web of Science]
Farfán L. M., Fogel I. Influence of tropical cyclones on humidity patterns over southern Baja California, Mexico. American Meteorological Society (2007) 135:1208–1224.
Félix-Uraga R., Gómez-Muñoz V., García-Franco W., Quiñonez-Velázquez C., Melo-Barrera F. N. On the existence of Pacific sardine groups off the west coast of Baja California and southern California. CalCOFI Reports (2004) 45:146–151.
Fey D. P. Temperature and somatic growth effect on the otolith growth of larval fish. Journal of Fish Biology (2006) 69:794–806.[CrossRef][Web of Science]
Gao Y. W. Microsampling of fish otoliths: a comparison between DM 2800 and Dremel in stable isotope analysis. Environmental Biology of Fishes (1999) 55:443–448.[CrossRef][Web of Science]
Gao Y. W., Joner S., Svec R. A., Weinberg K. L. Stable isotopic comparison in otoliths of juvenile sablefish (Anoplopoma fimbria) from waters off the Washington and Oregon coast. Fisheries Research (2004) 68:351–360.[CrossRef][Web of Science]
Gao Y. W., Joner S. H., Bargmann G. G. Stable isotopic composition of otoliths in identification of spawning stocks of Pacific herring (Clupea pallasi) in Puget Sound. Canadian Journal of Fisheries and Aquatic Sciences (2001) 58:2113–2120.
Goericke R., Bograd S. J., Gaxiola-Castro G., Gómez-Valdés J., Hooff R., Huyer A., Hyrenbach K. D., et al. The state of the California Current, 2003–2004: a rare "normal" year. CalCOFI Reports (2004) 44:27–59.
Goericke R., Venrick E., Mantyla A., Bograd S. J., Schwing F. B., Huyer A., Smith R. L., et al. The state of the California Current, 2004–2005: still cool? CalCOFI Reports (2005) 46:32–71.
Grossman E. L., Ku T. L. Oxygen and carbon isotopic fractionation in biogenic aragonite: temperature effects. Chemical Geology (1986) 59:59–74.[CrossRef][Web of Science]
Hammann G., Baumgartner T., Badan-Dangon A. Coupling of the Pacific sardine (Sardinops sagax caeruleus) life cycle with the Gulf of California pelagic environment. CalCOFI Reports (1988) 29:102–109.
Hammann M. G., Nevárez-Martínez M. O., Green-Ruíz Y. Spawning habitat of the Pacific sardine (Sardinops sagax) in the Gulf of California: egg and larval distribution 1956–1957 and 1971–1991. CalCOFI Reports (1998) 39:169–179.
Hastings J. R., Turner R. M. Seasonal precipitation regimes in Baja California, Mexico. Geografiska Annaler, Series A, Physical Geography (1965) 47:204–223.[CrossRef]
Høie H., Andersson C., Folkvord A., Karlsen O. Precision and accuracy of stable isotope signals in otoliths of pen-reared cod (Gadus morhua) when sampled with a high-resolution micromill. Marine Biology (2004) b 144:1039–1049.[CrossRef]
Høie H., Folkvord A. Estimating the timing of growth rings in Atlantic cod otoliths using stable oxygen isotopes. Journal of Fish Biology (2006) 68:826–837.[CrossRef][Web of Science]
Høie H., Folkvord A., Otterlei E. Effect of somatic and otolith growth rate on stable isotopic composition of early juvenile cod (Gadus morhua L) otoliths. Journal of Experimental Marine Biology and Ecology (2003) 289:41–58.[CrossRef][Web of Science]
Høie H., Otterlei E., Folkvord A. Temperature-dependent fractionation of stable oxygen isotopes in otoliths of juvenile cod (Gadus morhua L.). ICES Journal of Marine Science (2004) a 61:243–251.
Iacumin P., Bianucci G., Longinelli A. Oxygen and carbon isotopic composition of fish otoliths. Marine Biology (1992) 113:537–542.[CrossRef]
Jacobson L. D., Bograd S. J., Parrish R. H., Mendelssohn R., Schwing F. B. An ecosystem-based hypothesis for climatic effects on surplus production in California sardine (Sardinops sagax) and environmentally dependent surplus production models. Canadian Journal of Fisheries and Aquatic Sciences (2005) 62:1782–1796.
Jamieson R. E., Schwarcz H. P., Brattey J. Carbon isotopic records from the otoliths of Atlantic cod (Gadus morhua) from eastern Newfoundland. Canadian Journal of Fisheries and Aquatic Sciences (2004) 68:83–97.
Juillet-Leclerc A., Schrader H. Variations in upwelling intensity recorded in varved sediment from the Gulf of California during the past 3,000 years. Nature (1987) 329:146–149.[CrossRef]
Kalish J. M. Oxygen and carbon stable isotopes in the otoliths of wild and laboratory-reared Australian salmon (Arripis trutta). Marine Biology (1991) a 110:37–47.[CrossRef]
Kalish J. M. 13C and 18O isotopic disequilibria in fish otoliths: metabolic and kinetic effects. Marine Ecology Progress Series (1991) b 75:191–203.[Web of Science]
Kim S. T., O'Neil J. R. Equilibrium and nonequilibrium oxygen isotope effects in synthetic carbonates. Geochimica et Cosmochimica Acta (1997) 61:3461–3475.[CrossRef][Web of Science]
Krutzikowsky G. K., Emmett R. L. Diel differences in surface trawl fish catches off Oregon and Washington. Fisheries Research (2005) 71:365–371.[CrossRef][Web of Science]
Lasker R. An experimental study of the effect of temperature on the incubation time, development and growth of the Pacific sardine embryos and larvae. Copeia (1964) 2:399–405.
LeGrande A. N., Schmidt G. A. Global gridded data set of the oxygen isotopic composition in seawater. Geophysical Research Letters (2006) 33:L12604.[CrossRef]
Lluch-Belda D., Hernández-Rivas M. E., Saldierna.Martínez R., Guerrero-Caballero R. Variabilidad en la temperatura superficial del mar en Bahía Magdalena, B.C.S. Oceánides (2000) 15:1–23.
Lluch-Belda D., Lluch-Cota D. B., Hernández-Vazquez S., Salinas-Zavala C. A., Schwartzlose R. Sardine and anchovy spawning as related to temperature and upwelling in the California current system. CalCOFI Reports (1991) 32:105–111.
López M., Candela J., Argote M. L. Why does the Ballenas Channel have the coldest SST in the Gulf of California? Geophysical Research Letters (2006) 33:L11603.[CrossRef]
Marr J. The causes of major variations in the catch of the Pacific sardine, Sardinops caeruleus (Girard). In: Proceedings of the World Scientific meeting on the Biology of Sardines and Related Species—Rosa H., Murphy. G. I., eds. (1960) III. Rome: FAO. 667–791.
Martin S. An Introduction to Ocean Remote Sensing. (2004) Cambridge, UK: Cambridge University Press. 454. pp.
McFarlane G. A., Smith P. E., Baumgartner T. R., Hunter J. R. Climate variability and Pacific sardine populations and fisheries. American Fisheries Society Symposium (2002) 32:195–214.[Web of Science]
Nevárez-Martínez M. O., Lluch-Belda D., Cisneros-Mata M. A., Santos-Molina J. P., Martínez-Zavala M. A., Lluch-Cota S. E. Distribution and abundance of the Pacific sardine (Sardinops sagax) in the Gulf of California and their relation with the environment. Progress in Oceanography (2001) 49:565–580.[CrossRef][Web of Science]
Newman S. J., Steckis R. A., Edmonds J. S., Lloyd J. Stock structure of the goldband snapper Pristipomoides multidens (Pisces: Lutjanidae) from the waters of northern and western Australia by stable isotope ratio analysis of sagittal otolith carbonate. Marine Ecology Progress Series (2000) 198:239–247.[CrossRef][Web of Science]
Paul A., Mulitza S., Patzold J., Wolff T. Simulation of oxygen isotopes in a global ocean model. In: Use of Proxies in Paleoceanography: Examples from the South Atlantic—Fisher G., Wefer. G., eds. (1999) Berlin: Springer. 665–686.
Payne B., Quijano L., Latorre C. A study of the leakage between two aquifers in Hermosillo, Mexico, using environmental isotopes. International Atomic Energy Agency Bulletin (1978) 158:113–130.
Quiñonez-Velázquez C., Alvarado-Castillo R., Félix-Uraga R. Individual growth and population abundance relationship in the Pacific sardine Sardinops caeruleus (Pisces: Clupeidae) (Girard 1856) from Isla de Cedros, Baja California, Mexico. Revista de Biología Marina y Oceanografía (2002) 37:1–8.
Quiñonez-Velázquez C., Nevárez-Martínez M. O., Gluyas-Millán M. G. Growth and hatching dates of juvenile Pacific sardine Sardinops caeruleus in the Gulf of California. Fisheries Research (2000) 48:99–106.[CrossRef][Web of Science]
Radtke R. L., Showers W., Moksness E., Lenz P. Environmental information stored in otoliths: insights from stable isotopes. Marine Biology (1996) 127:161–170.[CrossRef]
Rangel-Medina M., Monreal-Saavedra R., Morales-Montaño M., Castillo-Gurrola J. Vulnerabilidad a la intrusión marina de acuíferos costeros en el Pacifico Norte Mexicano. Un caso, el acuífero costa de Hermosillo, Sonora, México. Revista Latinoamericana de Hidrogeología (2002) 2:31–51.
Robinson C. J., Gómez-Gutiérrez J., Félix-Uraga R., Arenas-Fuentes V. Seasonal hydro-acoustical observations of small pelagic fish behaviour in Bahía Magdalena, Mexico. Aquatic Living Resources (2000) 13:11–18.[CrossRef][Web of Science]
Roden G. I. Oceanographic and meteorological aspects of the Gulf of California. Pacific Science (1958) 12:21–45.[Medline]
Romero-Centeno R. Comportamiento de los campos hidrográficos y flujos de calor y masa en el Canal de Ballenas. MSc thesis, Centro de Investigación Científica y de Educación Superior de Ensenada. Ensenada, Baja California, México. (1995) 126 pp..
Rozanski K., Araguas-Araguas L., Gonfiantini R. Isotopic patterns in modern global precipitation. In: Climate Change in Continental Isotopic Records—Swart P. K., Lohmann K. C., McKenzie J. A., Savin. S., eds. (1993) Washington, DC: American Geophysical Union. 1–36. Geophysical Monograph, 78.
Schmidt G. A., Bigg G. R., Rohling E. J. Global Seawater Oxygen-18 Database. (1999) http://data.giss.nasa.gov/o18data/.
Schwarcz H. P., Gao Y., Campana S., Browne D., Knyf M., Brand U. Stable carbon isotope variations in otoliths of Atlantic cod. Canadian Journal of Fisheries and Aquatic Sciences (1998) 55:1798–1806.
Sharp Z. Principles of Stable Isotope Geochemistry. (2006) London: Pearson/Prentice Hall. 344.
Smith P. E. Fisheries Coastal Pelagic Schooling Fish. In Marine Fish Larvae. In: Morphology, Ecology and Relation to Fisheries—Lasker. R., ed. (1981) Washington: Sea Grant Program Publisher. 1–32.
Smith P. E. A history of proposals for subpopulation structure in the Pacific sardine (Sardinops sagax) population off western North America. CalCOFI Reports (2005) 46:75–82.
Soto-Mardones L., Marinone S. G., Parés-Sierra A. Time and spatial variability of the sea surface temperature in the Gulf of California. Ciencias Marinas (1999) 25:1–30.[Web of Science]
Stephenson P. C., Edmonds J. S., Moran M. J., Caputi N. Analysis of stable isotope ratios to investigate stock structure of red emperor and Rankin cod in northern Western Australia. Journal of Fish Biology (2001) 58:126–144.[CrossRef][Web of Science]
Thorrold S. R., Campana S. E., Jones C. M., Swart P. K. Factors determining 13C and 18O fractionation in aragonitic otoliths of marine fish. Geochimica et Cosmochimica Acta (1997) 61:2909–2919.[CrossRef][Web of Science]
Watanabe Y., Saito H. Feeding and growth of early juvenile Japanese sardines in the Pacific waters off central Japan. Journal of Fish Biology (1998) 52:519–533.[CrossRef][Web of Science]
Wefer G., Berger W. H. Isotope paleontology: growth and composition of extant calcareous species. Marine Geology (1991) 100:207–248.[CrossRef][Web of Science]
Weidman C. R., Millner R. High-resolution stable isotope records from North Atlantic cod. Fisheries Research (2000) 46:327–342.[CrossRef][Web of Science]
Wurster C. M., Patterson W. P., Cheatham M. M. Advances in microdrilling techniques: a new apparatus for acquiring high-resolution oxygen and carbon stable isotope values and major/minor elemental ratios from accretionary carbonate. Computers and Geosciences (1999) 25:1159–1166.[CrossRef]
Yaremko M. L. Age determination in Pacific sardine, Sardinops sagax. US Department of Commerce, NOAA Technical Memorandum, NOAA TM NMFS SWFSC (1996) 223. pp.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Johannessen, L. Nottestad, A. Ferno, L. Langard, and G. Skaret Two components of Northeast Atlantic herring within the same school during spawning: support for the existence of a metapopulation? ICES J. Mar. Sci., September 1, 2009; 66(8): 1740 - 1748. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||