Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Compound-Specific δ15N Amino Acid Measurements in Littoral Mussels in the California Upwelling Ecosystem: A New Approach to Generating Baseline δ15N Isoscapes for Coastal Ecosystems

Compound-Specific δ15N Amino Acid Measurements in Littoral Mussels in the California Upwelling Ecosystem: A New Approach to Generating Baseline δ15N Isoscapes for Coastal Ecosystems

  • Natasha L. Vokhshoori, 
  • Matthew D. McCarthy


We explored δ15N compound-specific amino acid isotope data (CSI-AA) in filter-feeding intertidal mussels (Mytilus californianus) as a new approach to construct integrated isoscapes of coastal primary production. We examined spatial δ15N gradients in the California Upwelling Ecosystem (CUE), determining bulk δ15N values of mussel tissue from 28 sites between Port Orford, Oregon and La Jolla, California, and applying CSI-AA at selected sites to decouple trophic effects from isotopic values at the base of the food web. Bulk δ15N values showed a strong linear trend with latitude, increasing from North to South (from ∼7‰ to ∼12‰, R2 = 0.759). In contrast, CSI-AA trophic position estimates showed no correlation with latitude. The δ15N trend is therefore most consistent with a baseline δ15N gradient, likely due to the mixing of two source waters: low δ15N nitrate from the southward flowing surface California Current, and the northward transport of the California Undercurrent (CUC), with15N-enriched nitrate. This interpretation is strongly supported by a similar linear gradient in δ15N values of phenylalanine (δ15NPhe), the best AA proxy for baseline δ15N values. We hypothesize δ15NPhe values in intertidal mussels can approximate annual integrated δ15N values of coastal phytoplankton primary production. We therefore used δ15NPhe values to generate the first compound-specific nitrogen isoscape for the coastal Northeast Pacific, which indicates a remarkably linear gradient in coastal primary production δ15N values. We propose that δ15NPhe isoscapes derived from filter feeders can directly characterize baseline δ15N values across major biochemical provinces, with potential applications for understanding migratory and feeding patterns of top predators, monitoring effects of climate change, and study of paleo- archives.


Isotope spatial gradients, or Isoscapes, are maps of systematic isotope variation and provide important biogeochemical information. Isoscapes are becoming increasingly important tools to characterize major biogeochemical zones and gradients in the ocean, and have been also used in ecological studies to help constrain animal migration and fish stock patterns (e.g., [1], [2]). Isoscapes of nitrogen (N) stable isotope values (δ15N) can be particularly informative, because such measurements have the potential to identify major ocean transitions between eutrophic/mesotrophic and oligotrophic regions, the balance of fundamental N cycle processes (e.g., N fixation vs. denitrification), and also basic ecological and food web relationships across major habitat zones. For example, water-column denitrification has a very large isotope effect (ε) of 25–30‰ [3] which greatly increases the δ15N value of all organisms in areas where this process is important. However, this could be rapidly changing in many ocean regions, linked to oceanographic climate events associated with a shifting climate (e.g., [4] Detailed isoscapes can ultimately provide a link between biogeochemical process and larger food webs, a key for understanding marine ecosystems. This is especially critical at a time when both natural and anthropogenic perturbations may be rapidly shifting fundamental biogeochemical processes (e.g., [5]), and potentially entire food web structures [6], [7].

However, the information potential inherent in δ15N values also presents significant challenges for interpretation of bulk δ15N values of organic matter. First, isoscapes are typically constructed from measurements in secondary or higher consumers. This approach provides a temporally integrated measurement; however, by definition, it also results in measured δ15N values being offset from “baseline” δ15Nvalues of primary production, since the 15N content of a consumer increases substantially with each trophic transfer. An average trophic enrichment factor (TEF) of ∼3.4‰ is often assumed [8][10], however it has been shown that the TEF values in fact vary substantially: not only between species, but also depending on tissue type, life stage, growth rate, and a host of other factors [10][12]. Further, for many oceanographic applications, such as understanding shifting gradients in primary production or N cycle processes, it is really the “baseline” δ15N value that is of primary interest (i.e., the δ15N value of primary production or N sources at the base of food webs). Because the bulk isotope value in a consumer is the combined signal of the baseline value and subsequent trophic effects, it is extremely difficult to isolate either factor with confidence.

Compound-specific isotope analysis of amino acids (CSI-AA) is a rapidly evolving technique that can address many inherent issues with bulk isotope data. For δ15N values, a seminal study by McClelland & Montoya (2002) demonstrated strong differential 15N enrichment of different groups of amino acids (AA) with trophic transfer. One group of AA has strongly elevated δ15N values with each trophic transfer (∼4–8‰), and are now termed the “Trophic AAs.” A second group of AA, now termed the “Source AAs,” in contrast has relatively constant δ15N values with trophic transfer, and so largely preserves δ15N values from the base of the food web. This pattern of AA differential enrichment has now been verified across a wide range of photoautotrophs and primary consumers [13], [14] and also in higher trophic organisms [15][19].

Most δ15N CSI-AA studies to date have focused on nitrogen isotopic values of two main AAs: phenylalanine (Phe), as the best indicator of baseline δ15N value, and glutamic acid (Glu), as the best indicator for relative trophic transfer. The relative predictability of 15N offsets between Glu and Phe with trophic transfer has also led to an explicit equation now used widely to calculate CSI-AA based trophic position (see methods). Based on these findings, CSI-AA patterns (δ 15NAA) have now been used to not only estimate trophic position (TP) [16], [20], [21] and to trace source or microbial re-working of organic N sources [4], [19], [22][24], and animal movement across broad ocean basins [19], [25]. Taken together, CSI-AA work to date strongly suggests that if δ15NAA is applied in appropriate heterotrophic organisms, the source AA should be able to indicate baseline δ15N isoscapes, decoupled from influence of trophic transfer.

The California mussel (Mytilus californianus) is a sessile resident of intertidal zones, which continuously filters particulate organic matter (POM). As such, mussels correspond closely to an ideal “baseline indicator” organism (i.e., a long-lived primary consumer; [10]). Because mussels temporally integrate filtered POM into their tissues and shells over annual to decadal timescales, they have been widely used as both sentinel organisms for marine pollutants (e.g., Mussel Watch Project:, as well as to attempt reconstruction of ocean water composition and conditions [6][10], [26][29]. In contrast, many other organism types have important drawbacks for constructing representative isoscapes. For example, highly mobile top predators may rapidly transit distinct biogeochemical zones (e.g. [8][15], [18], [19], [30]), and thus attenuate isotopic variability. In contrast, short-lived organisms (such as zooplankton) can be assumed to not move widely, but because of relatively fast growth rates and rapid N turnover times may not integrate variation, but rather are subject to short temporal isotopic changes in the environment (e.g. [16]). Because of their sessile nature, cosmopolitan distribution, and continuous integration of water column food sources, mussels have major advantages as a potential basis for coastal isoscapes.

Here we examined δ15NAA patterns in California mussels (Mytilus californianus) across 10 degrees of latitude in the coastal zone of the California Upwelling Ecosystem (CUE). The CUE is part of the greater California Current System (CCS), and is a highly dynamic region where we would anticipate not only large potential variation in baseline δ15N values, but also the potential for rapid future change linked to a warming climate [26], [28], [29]. Our overall goal was to explore whether δ15N values of source AA, and in particular Phe (δ15NPhe), may can serve as a new, direct proxy for constructing isoscapes of integrated δ15N values of primary production within highly dynamic coastal regions. We compared bulk δ15N and δ15N AA patterns in mussels to first test dependence of bulk isotopic variability on baseline δ15N values, using CSI-AA to constrain variations in TP. We also compared mussel δ15N values with literature values in more offshore sample types (zooplankton and POM), to examine if our results may also apply to the larger CCS. Our results indicate that source AA values in mussels are likely represent a direct record of variation in baseline δ15N values, and suggest that in the CA coast region isoscapes based on δ15NPhe closely follow variations in nitrate δ15N values.


Sample Collection and Preparation

Ethics Statement.

California mussels (Mytilus californianus) analyzed for this study were collected from 28 different sites between Coos Bay, Oregon and San Diego, California (Table 1), under a permit provided by the California Department of Fish and Wildlife.

Table 1. Collection sites and bulk δ15N values for Mytilus californianus.

Mussels were collected in the winter (Dec – Feb) of 2009–2010. Sites were chosen to be approximately evenly distributed along the CA coastline, with ∼80 km geographic separation between each sampling site. Our main goal here was to sample mussels from a wide geographic range across the CCS, although for observing finer scale local or regional variations, a finer-scale sampling strategy would like be required. Typically 5 individual mussels were collected from each site, all between 30–40 mm maximum shell length, which were immediately placed on dry ice until further preparation. The adductor muscle of each individual was dissected for analysis. This tissue was selected because isotopic values in muscle tissue have shown relatively long turnover times; based on past growth data, mussels of this size would be expected to integrate approximately annual variability in suspended food source isotopic values for each location sampled [31]. The dissected adductor tissue was carefully separated from other tissue types, rinsed with deionized water, refrozen, and then freeze-dried for 48 hrs. Lipids were removed following the methods of Dobush et al. (1985) [32], using petroleum ether in a Dionex Accelerated Solvent Extractor (Bannockburn, IL). Finally, in preparation for CSI-AA, composite samples were made from a subset of 13 collection sites (Fig.1a). For each location chosen for CSI-AA (based on the bulk δ15N record), 1±0.05 mg of lyophilized tissue was weighed and combined for each individual mussel (n = 5). Further CSI-AA preparation proceeded as described below.

Figure 1. Collection sites and bulk δ15N values.

Site-specific bulk δ15N values of Mytilus californianus as a function of latitude, in context of a map of collection sites on the California coast. (A) Filled circles indicate all sampling sites, and map locations correspond directly to bulk analysis values in panel B; open squares represent sites chosen for compound-specific isotope analysis. (B) Filled diamonds indicate average δ15N for five individuals sampled from each site; error bars indicate ±1SD. Regression line and statistics indicate strong linear relationship of δ15N values versus latitude.

Bulk Stable Nitrogen Analysis

Stable nitrogen isotope analyses were conducted using standard protocols in the Stable Isotope Lab at the University of California, Santa Cruz (UCSC-SIL). Briefly, homogenized muscle tissue of each individual was weighed into tin capsules and combusted. Isotope values determined on a Carlo Erba 1108 elemental analyzer (Lakewood, NJ) coupled to a Thermo Finnigan Delta Plus XP isotope ratio mass spectrometer (San Jose, CA) (EA-IRMS). Analytical error associated with this measurement was typically <±0.15‰ based on sample replicates. Stable isotopes are reported using standard delta (δ) notation in parts per thousand (‰):δ15N =  [(Rsample/Rstandard) − 1] ×1,000, where R is the ratio of heavy to light isotope, Rsample is from the sample, and the Rstandard is atmospheric N2 (air) for carbon, as provided by pulses of calibrated CO2 reference gas. For details on correction calculations and normalization to international standards refer to UCSC-SIL website (

Compound-specific amino acid δ15N analysis

Amino acid δ15N values were measured as Trifluoroacetyl isopropyl ester (TFA-IP) AA derivatives, following protocols described in detail elsewhere (e.g.,. Briefly, samples were hydrolyzed (6 N HCl, 20 hr at 110°C) under nitrogen, and TFA derivatives subsequently prepared from free AA using a modified version of the protocol described by Silfer (Silfer et al. 1991): isopropyl esters were made with a 1∶5 mixture of Acetyl Chloride (AcCl):2-propanol (110°C, 60 minutes), and then acylated using a 1∶3 mixture of Dichloromethane:Trifluroacetyl acetate (DCM:TFAA) (100°C, 15 minutes). Derivatized AAs were dissolved in DCM to a final ratio of approximately 4 mg of original tissue to 250 µl DCM.

After derivatization, samples were analyzed by a Varian gas chromatograph coupled to a Finnegan Delta-Plus isotope ratio mass spectrometer (GC-IRMS). AAs were separated using a 50 m, 0.32 ID Hewlett Packard Ultra-1 column with 1 µm film thickness. Under our analytical conditions, δ15N values could be reproducibly measured for alanine (Ala), aspartic acid + asparagine (Asp), glutamic acid + glutamine (Glu), leucine (Leu), isoleucine (Ile), proline (Pro), valine (Val), glycine (Gly), lysine (Lys), serine (Ser), phenylalanine (Phe), threonine (Thr), and tyrosine (Tyr) (Fig. S4). Most AAs were measured with a standard error of <1.0‰ (based on n = 4 injections), and the average mean deviations for individual AA δ15N measurements across all tissue sample replicates was 0.5‰.

Amino Acid Categories and Trophic Position Calculations

In all results and discussion, measured AA are grouped into one of three categories: “Trophic” vs. “Source” (after Popp et al. 2007 [15]) and Thr alone is designated as a “metabolic” AA (after Germain et al. 2013 [18]). The measured Trophic AA (with large expected enrichment in 15N with trophic transfer) were Glu, Asp, Ala, Ile, Leu, Pro and Val. The measured Source AA (with expected little to no change in δ15N at higher trophic levels) were Phe, Gly, Ser, Lys and Tyr. The δ15N values of Thr exhibit an apparent inverses isotopic fractionation with trophic transfer, however are also highly variable with organism type [18], so this AA is considered outside the basic Trophic vs. Source division.

Based on this framework, we used unweighted averages of AA groupings, as well as specific TP estimates, to analyze our data. Average Trophic and Source δ15N values were calculated:(1)(2)

For explicit TP calculations, we used the “canonical” AA's (Glu and Phe) to calculate TP of mussels in the CUE, after Chikraishi et al. 2009:(3)where, δ15NGlu– δ15NPhe are measured values, +3.4 is the assumed isotopic difference between the Glu and Phe in primary producers (also referred to as the β value), and +7.6 is the assumed 15N enrichment in Glu relative to Phe with each trophic transfer from food source to consumer, also called the Δ value [13].

Statistical analyses and calculations

Statistical analyses (e.g., Hierarchical cluster analysis and Analysis of Covariance) were conducted using the JMP statistical software package (SAS Inc., Version 10). We used Arc-GIS Spatial Analyst (version 10.1) to produce visual isoscapes of the CUE. Our first model is based on the line for δ15NPhe values vs. latitude (y = −0.3328x+20.053, R2 = 0.63592). Our second model (Fig. S1) is based on one-dimension of δ15N values along the latitudinal extent of sampling area and interpolates between data points of known δ15N values and to 100km offshore.


Bulk δ15N values

Bulk δ15N values in the adductor muscle of Mytilus californianus ranged from 7.4‰ to 11.5‰ (Table 1). Bulk δ15N values were measured on tissue from multiple individual mussels (4–6, but typically 5) collected from each site to gauge intra-site variability in individuals. Standard deviations on average δ15N values for individual mussels from the same sites ranged from 0.1 to 0.5‰ (Table 1). The average standard deviation for all intra-site comparisons, across all locations, was 0.3‰. This value is close to EA instrument error (∼0.2 ‰), and therefore indicates an extremely small degree of variation in individual mussel δ15N values within specific sites, implying instead strong homogeneity of δ15N values for mussel populations. When plotted as a function of latitude, the average bulk δ15N values for mussels from each site have a strong linear trend (Fig. 1B). The average bulk δ15N values increase by 0.41‰ per degree of latitude from north to south (R2 = 0.755 and P<0.0001). We note that all sampling sites were located in exposed waters, and variable habitat type (i.e. rocky, jetty, etc.; see table 1) did not appear to be a major factor in the overall latitudinal trend in isotopic value.

Amino Acid δ15N values

Of the 28 sites measured for bulk analysis, 13 were chosen for CSI-AA (Fig 1A). Samples for CSI-AA were chosen first to obtain relatively even geographic spacing, with specific locations within geographic regions then selected to capture maximum offsets in the north to south bulk δ15N trend (Fig. 1B). Measured δ15N values for individual AAs ranged from −1.0‰ to 16.0‰ (Table 2). In all samples, Thr was distinct, with the lowest δ15N values, the Source AA group always had intermediate values, and the Trophic AAs always had the highest δ15N values (Table 2; Fig. S2). Over the entire data set, the range of the averaged Source AA δ15N (see methods) was 5.1 to 10.3‰ (SD = 1.3‰, n = 69) and the averaged Trophic AA values are 10.3 to 15.9‰ (SD = 0.4‰, n = 91). While precision for individual AA δ15N measurements varied (Table 2), it was typically <1‰, with the average analytical standard error across all AA we measured at all sites as 0.8‰ (n = 160).

Table 2. Compound specific amino acid δ15N values for Mytilus californianus.

We focused on Glu and Phe δ15N values as the best proxies for Trophic and Source AA groups, respectively, as has been indicated by a number of recent papers [19], [33][36]. Glu and Phe δ15N values both correlated significantly with average values for Trophic and Source AA groups respectively (Phe vs. average Source AA, R2 = 0.782; P = 0.0006; Glu versus average Trophic AA's, R2 = 0.546, P = 0.0049), confirming the validity of this approach (see also [25]). Both Phe and Glu δ15N values also tracked changes in bulk δ15N with latitude (Fig. 2). The δ15N values of Phe and bulk adductor muscle had a strong and significant linear relationship with latitude (P = 0.0028 and P = 0.0011, respectively). In contrast, there was more variability in the Glu data. The relationship of Glu vs. bulk δ15N was not significant at 95% confidence (P>0.05), however a Fit Model run of Analysis of Covariance shows that δ15N values of bulk, Phe and Glu all share a common slope (effects test, P = 0.0050); in other words, the slope of δ15N change with latitude for Glu and Phe are not significantly different from the slope of bulk δ15N change with latitude.

Figure 2. Latitudinal trends in Glu and Phe δ15N values compared with bulk δ15N.

Bold solid line indicates regression for bulk δ15N values for all sites sampled (as in Fig. 1), thin solid lines indicate linear regressions for Glu (open circles) and Phe (filled circles) (δ15NGlu = −0.270x+24.211, δ15NBulk = −0.313x+21.527, and δ15NPhe = −0.329x+19.927). ANCOVA analysis indicates that all three share a common slope within error (effects test, P<0.0050). Error bars for Glu and Phe indicate analytical standard deviation for CSI-AA performed on a composite sample for all individual mussels from each site (see methods).

Mussel Trophic Position and Trophic Enrichment Factors

The TP of mussels calculated using Eq. 3 ranged from 1.0 to 1.8 with an average TP of 1.4±0.3 (Table 2). TP had no correlation with latitude (P = 0.706), indicating that despite local variability, mussels' suspended POM food sources had similar average TP in all CA coastal regions. Across all mussel samples analyzed with CSI-AA, the average δ15NGlu – δ15NPhe offset was 6.5‰. Prior work indicates that these mussels feed primarily on microalgae [37][39], coupled with data for δ15NGlu – δ15NPhe offsets in phytoplankton and marine macroalgae [13], [22], this average offset would indicate an average TEFGlu-Phe for Mytilus californianus of 3.1‰ (Fig. S3).


This study investigated if δ15N CSI-AA values measured in mussel “bio-archives” can represent a new approach to understanding baseline δ15N patterns in dynamic coastal regions. We hypothesize that the potential for δ15NAA values to decouple trophic shifts from baseline δ15N values may, for the first time, allow construction of isoscapes specifically for temporally integrated δ15N values for primary production, based tissue samples from heterotrophic organisms. Mussels were chosen for this study because they are sessile, filter-feeding organisms with tissue turnover rates integrating suspended POM food sources on monthly to annual timescales. Other studies have previously CSI-AA or bulk δ15N patterns in primary producers, zooplankton [31], or mobile top predators, to trace oceanographic processes [15]. However, as noted above, in many heterotrophic organisms multiple variables might complicate inferences about baseline isotopic signals. For plankton these include variability caused by shorter biochemical turnover rates, coupled with seasonal change in nutrient availability, light intensity and temperature fluctuate [16], while in higher trophic level animals factors such as migration or a mixed diet may dilute the desired signal in question [30]. Overall, we expect that sessile filter feeders such as mussels are likely to be are among the best organisms for baseline source records in systems where they occur.

δ15N latitudinal gradient in the California Upwelling Ecosystem

We hypothesize that the strong δ15N gradients with latitude are driven by the mixing of two NO3 source waters, coupled by upwelling in the CUE. In this region, northern low-15N water is brought south by surface flow of the main CCS [3], [40]. At the same time, southern source of elevated 15N water originating from the zone of denitrification in the ETNP [33], [34], [36], [41], [42] is brought north via the California Undercurrent (CUC). The source area for the CUC (approximately south of the tip of Baja CA peninsula) is one of the major persistent oxygen minimum zones (OMZ) in the world ocean [3], [40] accounting for 35–45% of global pelagic denitrification [36], [41][43]. Bacterial denitrification in the low-oxygen water columns has a very large positive fractionation factor (ε ∼20–30‰; e.g., [3], [34]), and therefore imprints a distinct signal in surrounding waters, such that subsurface NO3 values for the southern CUC near the tip of Baja can approach +14‰ [36], [43][45]. The CUC then moves northward, with its flow attenuating as it progresses along the CA margin. The core of the CUC is near 150 m, directly in source-depth regions for upwelled waters [34], [43]. The main isotopic endmembers for inorganic nitrogen along our study region are therefore open Pacific nitrate from CCS (∼ 5‰, e.g. Sigman et al., 2009), mixed with the 15N enriched nitrate being carried northward by the CUC (∼9–10 ‰; [44][46]), and brought to the surface locally via upwelling. We note, however, that while all previous literature clearly indicates an expected change in baseline δ15N with latitude, it cannot indicate exact δ15N endmembers for the CA coast region we sampled. This is because of the relative paucity of direct nitrate 15N measurements, and the inherent temporal and geographical variation in these measurements, even within similar regions (e.g., [43], [46][48]).

A complimentary forcing for δ15N trends could therefore also be variability of upwelling intensity with latitude. The North American west coast is commonly described in terms of three distinct upwelling regions, characterized by differences in overall annual upwelling intensity: Baja California (21–30°N), continental US (30–48°N) and British Columbia and Alaska (48–60°N) [46], [47]. While winds generally increase northward, annual wind intensity is most consistent year-round south of Pt. Conception, strongest seasonally along central CA coast, and generally weakest north of Cape Blanco [39], [46][48]. Our study site does not cross all three of these main regions, however it seems possible that the transition between the southern and central zones of upwelling intensity could influence the overall latitudinal trend. Overall, while expected variation in average upwelling intensity is consistent with our observations of latitudinal δ15N change, wind forcing alone cannot not explain the clear linear decrease in δ15N values with increasing latitude.

Local denitrification is another process that could also contribute to regional δ15N baseline values. Water column denitrification in the CCS has been documented in borderland basins (such as in the Santa Barbara Basin, SBB; [47]), where water column exchange with the ocean is blocked by basin sills, causing basin water to become O2 deficient [34], [44], [47], [49]. It has also been shown to occur in areas along the Oregon coast, due to advection of oxygen-poor water masses onto continental shelves [50]. However, if local denitrification were a main factor driving relative δ15N values, we would predict far more localized δ15N variability. Therefore, while this cannot be ruled out as contributing to δ15N values for specific locations, it seems highly unlikely as the major forcing for such a regular gradient. Finally, both water temperature and sampling season might be considered as additional factors. As noted above, all mussels were collected in the winter season of 2009–2010. While it is possible that mussel metabolism may change throughout the year (high vs. low feeding seasons), the specific tissue analyzed (adductor muscle, see methods) and mussel size class were specifically selected to isotopically integrate over an approximate yearly time frame. This assumption is supported by preliminary data of samples collected in both summer and winter season of the same year for selected sampling sites, for which no significant effect on the observed latitudinal trend was observed [39], [45], [51]. Water temperature is also a general function of latitude in the CCS at all times of year. Change in water temperature might affect the isotopic gradient either directly via changing mussel metabolism, or indirectly as a proxy for upwelling strength. However, in contrast to the mussel N isoscape, the major temperature changes along the CA coast are not linear, but rather shift more strongly at the boundary of the Southern CA Bight, with temperatures generally much warmer south of Pt. Conception, and consistently much cold temperatures (due to stronger upwelling) in central and northern CA.

Overall, the strong latitudinal δ15N trend recorded in mussel tissues seems most consistent with the endmember mixing outlined above, consistent with both modeling and prior discrete sampling. For example, δ15N values of sediment traps and sediment cores contrasted between central CA vs. the Southern CA bight have indicated δ15N values are generally more enriched in the Southern CA bight vs. Northern CA, consistent with our measurements [34], [44], [49], [52], [53]. In addition, basin-scale modeling of δ15N variation [3], [45], [51], [54] also predict a south to north trend of decreasing baseline δ15N values, driven by the ETNP denitrification endmember. Our study therefore represents perhaps the strongest confirmation to date of both model predictions, and also prior discrete-location sampling results. However, no prior sample set has ever tested CCS latitudinal δ15N variation at such high resolution, based on an archive coupling unambiguous source location with approximately annual signal integration. In particular, the striking linear trend in our data is a novel, and also perhaps a surprising finding. This indicates that the diminution in 15N-enriched nitrate supply via the CUC (or relative mixing with southerly CCS) is remarkably regular in the CUE: across the 10 degrees of latitude that we sampled, δ15N change was remarkably consistent (0.41‰±0.04 per degree). We suggest that the ability to capture this regional trend at such high precision is linked to the integrative property of filter feeding consumers, as well as the longer-turnover tissue that we sampled. Overall, we propose that Fig. 1b indicates the integrated approximately annual gradient in 15N values in coastal CUE waters with latitude. If correct, this also suggests that surveys of costal mussel δ15N values might constitute a powerful new tool for constraining physical mixing and circulation models, since they would show the effective mixing of two source waters in great detail.

While satellite data has documented changing global ocean surface chlorophyll concentrations, leading to predictions of declining primary production in the world's oceans due to increased stratification associated with warming [49], [52], [53], [55], the effects of a warming climate on CCS biogeochemistry remain unclear. Some studies have proposed that some CCS zones are already showing an opposite trend of increasing productivity, linked to increased nutrient supply [3], [22], [28], [54]. If the ocean nitrate endmember were to increase over time, this should result in gradual decrease in δ15N values in the CUE, and potentially also a change in the slope of the clear latitudinal gradient we observe. At the same time, increasing stratification in CCS waters is also proposed as one main consequence of warming, and this has already been documented [13], [49], [56]. If this decreased the effective supply of CUC water and associated nitrate, it could also lead to lower δ15N values. However, the potential effects of natural climatic perturbations (e.g. El Nino Southern Oscillation – ENSO, and Pacific Decadal Oscillation – PDO cycles) are currently very difficult to decouple from longer term trends. Given that our current understanding of physical and biological responses of the CUE to a changing climate remains poor. Repeated sampling of mussels could provide a time and geographically integrated record of baseline isotopic change in this system, revealing longer term trends in the regional δ15N gradient, due to either natural fluctuations or climate change, at near annual resolution.

Do coastal mussel δ15N data also reflect broader California Current δ15N values?

An important question is to what degree δ15N data derived from mussels may reflect isotopic values of broader coastal waters, as opposed to only littoral sources and process. This is likely to be a function of relative time scale: the time frame over which mussels integrate δ15N of primary production, vs. the mixing time scale for littoral water with more seaward coastal water masses. If water mixing is relatively rapid vs. sampled tissue isotopic turnover, then it is possible mussels would reflect isotopic values within the broader CUE, and possibly into shoreward extent of the CCS. In contrast, if upwelling and nutrient utilization in the littoral zones are rapid and strongly localized, then littoral mussel δ15N values could be mostly decoupled from values in more offshore coastal waters. To definitively address this question, an extensive sampling program would likely be required, comparing offshore/onshore POM isotope values with those recorded in mussel tissues.

However, for the well-studied Monterey Bay region, past work offers extensive data sets for both coastal and offshore δ15N values in both organisms and detrital OM. We therefore compiled δ15N values for a range of sample types from the Monterey Bay region, and compared these with δ15N values for mussels at our sampling sites in or near the bay (Fig. 3). Specifically, we compared average δ15N values for 2 herbivorous and 2 carnivorous zooplankton species [22], [45], [57]and also OM in sediment traps (450 m) and surface sediment (950 m) samples[13], [15], [16], [24], [34], [35], [39]. Because many of these samples are not primary consumers, δ15N change due to trophic transfer must be taken into account for any direct comparison. We therefore assumed that mussels and herbivorous zooplankton, as primary consumers would feed at the same TP as mussels (TP = 2), requiring no correction. For carnivorous zooplankton (secondary consumers) we assumed one additional trophic transfer (TP = 3), and therefore adjusted reported δ15N values by 3.4‰ (the most broadly accepted average bulk TEF value). For sediment trap and surface sediment samples, we used recent results from Monterey Bay long-term sediment trap records, which have indicated an average trophic position (determined by CSI-AA) of 1.6 (Sherwood and McCarthy, unpublished data; also similar to TP for POM at Station ALOHA,[20], [33], [45]). We therefore adjusted both trap and surface sediment values for 1.5 TP (+1.7‰).

Figure 3. δ15N values for littoral mussels compared with offshore sample types in Monterey Bay region.

δ15N values for Mytilus californianus (this study) compared with literature values for sediment trap and bottom sediments (Altabet et al. 1999), primary consumer zooplankton (Eucalanus californicus and Calanus pacificus) and secondary consumer zooplankton (Sagitta euneritica and Sagitta bierii) (Rau et al. 2003) from Monterey Bay. For literature sample types diamonds indicate values corrected for estimated trophic position, open squares are the reported average literature value. Shaded band indicates range of our measured δ15N values for mussels from three Monterey Bay sampling sites (SC, ML, ASI); for literature sample types error bars indicate main range for reported values.

While we acknowledge that this approach can provide only a very general initial comparison, the results are nevertheless quite encouraging (Fig.3). For most sample types, the Monterey Bay adjusted values fall directly within the δ15N range for local mussels. This suggests that littoral mussels may in fact reflect δ15N values more broadly for local coastal waters. Given the high wind stress and mixing characteristic of the Central and Northern CA coasts, this may not be surprising. However, clearly this represents only a preliminary comparison, and rests on a range of assumptions that remain to be fully tested (e.g. that time of sampling is relatively unimportant, or that bulk TEF values are accurately estimated). To fully explore the potential of littoral mussels as integrators of coastal isoscapes, we suggest a synoptic sampling program comparing offshore/onshore mussel isotopic values will be required.

Mussel trophic position

CSI-AA provides a unique opportunity to decouple the effects of trophic transfer from δ15N values at the base of the food web. Based on differential enrichment behavior of the Trophic vs. Source AAs introduced above [13], [15], [16], [21], [24], [35], [39], [52], CSI-AA allows for a direct assessment of the role for TP variation may play in bulk δ15N value trends. Our calculations of mussel TP for an extended population along the entire CA coast represents, to our knowledge, the first wide-ranging CSI-AA survey of any filter feeding mollusk population in nature. The lack of any trend in TP with latitude (Fig. 4) indicates that mussels spanning the entire CA coast feed at a very similar TP, therefore likely on similar food sources, independent of location. The consistency of TP is interesting, given the previously documented localized variation in δ13C values for mussels from these same locations (Vokhshoori et al. in press). The similarity of TP from all locations therefore supports the conclusion that δ13C variation is primarily driven by changes in baseline δ13C values. Since TP does not change with latitude also strongly supports our basic hypothesis that the overall δ15N trend with latitude (Fig. 1b) is also driven by north to south variation in baseline δ15N values, most likely linked to nitrate δ15N values.

Figure 4. CSI-AA based trophic position for mussels from California Coast.

Calculated trophic position (TP) for composites of mussels from the sites selected for CSI-AA (filled squares) and average TP ± 1SD for all mussels sampled along CA coast (grey square)

However, the exact TP values (average TP = 1.4±0.2; Fig. 4a) calculated using the standard Glu-Phe approach (see methods, Eq. 1), also are lower than would be expected. Mussels predominately feed on POM derived from primary production [58], [59], so as primary consumers the mussel TP values should be at least 2. There are at least two possible explanations for the lower TP indicated by CSI-AA. One relates to non-algal food sources that might contribute to mussel AA, for example detrital POM (Vokhshoori et al. in press) or non-algal primary production sources. For example, some primary producers can have different baseline CSI-AA patterns from marine microalgae (e.g., seagrasses; [2], [13], [15], [37], [39], such that substantial contributions from non-microalgal sources would change the calculated TP calculated using the standard equation (see Text S1). However, δ13CAA source fingerprinting applied to these same mussels has verified a dominantly marine microalgal diet [38], [39], in agreement with ecological expectations diet (e.g.,[3], [33]). This suggests that if different source δ13CAA patterns do account for the offset, it is more likely due to variations among different microalgal groups. Given the relatively limited current δ13CAA data on different marine algal lineages, this is possible. An alternate possibly, is that the change in Glu δ15N values with trophic transfer in mussels may be smaller than the TEF factor now most commonly applied (and assumed in Equation 1; see methods). Accumulating evidence now suggests that CSI-AA trophic enrichment factors may be specific for different groups of organisms [18], [34], [36], [40][42], [60], however TEF values offsets have so far been documented only higher TP predators. Ultimately, distinguishing between these possibilities is beyond scope of our current data, however we provide a broader explanation of the underlying issues in Text S1.

Overall, however, it is important to stress that apparently low TP values for mussels do not bear in any significant way on our main observations. Specifically, the constant TP with latitude indicated by CSI-AA data is independent of exact TP estimates. However, TP data does suggest that controlled feeding experiments with filter-feeding mollusks, together with a more extensive survey of variation in the δ15N-AA in different algal types, will be needed to clearly interpret TP values derived from mussel tissue or shells. Such work might also be important for future potential development of CSI-AA patterns in archeological mussel shells as potential paleoceanographic bioarchives.

AA-CSIA: new tool to reconstruct primary production δ15N isoscapes

As noted above, all literature to date has indicated that Phe δ15N values are closely linked to δ15N values at the base of the food web, such that δ15NPhe values in a heterotroph can be used to estimate δ15N values of average primary production sources [3], [5], [15], [16], [34][36], [43], [44]. Given that these mussels feed almost uniquely on microalgae [37][39], [46][48], [59] we therefore hypothesize that mussel δ15NPhe should represent a temporally integrated value for δ15N of coastal phytoplankton production. Sampling δ15NPhe in mussel populations along a coastline should therefore yield, for the first time, a way to construct an integrated isoscape of baseline δ15N values.

A direct comparison between bulk δ15N and δ15NPhe values is one way to test this idea (Fig. 5). If we assume average complete NO3 utilization for this region (at least over ∼annual time frames mussels integrate; e.g., [49], [52]), then the slope of the regression for bulk δ15N values (Fig. 1b) should also represent the gradient in NO3 δ15N values along the CA coast. In this case, the “baseline” δ15N values should also be essentially equivalent with NO3 δ15N values. However, the bulk δ15N relationship of course cannot directly represent baseline values, due trophic transfer enrichment factors, as well as tissue-specific offsets. In Fig. 5 we therefore compare two possible approaches for estimating the baseline δ15N values from measured tissue data. The first relies only on bulk δ15N results, and assumes a standard bulk TEF value of 3.4‰, for mussels feeding at TP = 2.0. This approach predicts baseline δ15N values are far more reasonable vs. the measured bulk tissue values (Fig. 5; open diamonds). However, the results also appear to underestimate δ15N of NO3 along the CA coast (∼4.5 to 8‰ north to south), when compared with previous literature data [3], [22]. This may not be surprising, since by definition this approach cannot take into tissue-specific δ15N offsets, nor the actual TEF values for mussels. In contrast, our proposed CSI-AA approach uses δ15NPhe values as a direct proxy baseline δ15N. The fact that the slope of δ15NPhe vs. latitude is identical within error to bulk δ15N (see results) strongly supports the idea of a constant offset from primary production. The δ15NPhe regression-derived baseline values are universally heavier than the bulk- δ15N derived data discussed above, and are a closer match for reported NO3 values in northern vs. southern CA waters.

Figure 5. Two approaches for the estimation of baseline CUE δ15N values from mussel isotopic data.

The CSI-AA approach, based on average values for δ15NPhe (filled circles, solid regression line) predicts average baseline δ15N values most consistent with expected NO3 δ15N gradients along the CA coast. An alternate approach is based on measured bulk δ15N values, adjusted for an assumed trophic position (open diamonds, dashed regression line). This approach cannot take into account either TEF or tissue-specific fractionations, and returns lower than expected values in most locations. The regression for measured bulk δ15N values in adductor muscle tissue (heavy dashed line) is provided for reference.

We therefore hypothesize that δ15NPhe is ultimately more accurate representation of baseline δ15N, because it requires no assumptions about TEF values in any specific organism. However, in order to derive precise baseline δ15N predictions based on δ15NPhe, it will also be necessary to have robust calibrations for the offset between δ15NPhe and average algal δ15N values. The close match we observe between the δ15NPhe regression values and expected NO3 δ15N values of this region (Fig.5; [13], [49]) strongly suggests this offset cannot be very large in this system. This would agree with recent analysis of δ15N AA patterns measured in a range of phytoplankton species [22], [45]. However, other work has indicated larger offsets in some macroalgal and also micro-algal species tested in feeding experiments [13], [39]. An alternate approach for the future also could be to derive more broad-based corrections, based on δ15N values of multiple Source AA. Together with indications regarding the importance of representative β values for TP calculations, this further underscores the need for future work aimed at a more robust understanding of δ15NAA patterns across representative algal sources.

Baseline δ15N isoscapes from CSI-AA data

Taken together, these results suggest that CSI-AA has the potential, for the first time, to allow direct reconstruction of δ15N isoscapes of primary production, based on δ15NPhe values measured in consumers. A baseline δ15N isoscape for the CUE (Fig. 6), derived on our δ15NPhe values, represents to our knowledge the first such application. While the general trend of decreasing δ15N values with latitude is similar to broad trends predicted in regional or basin-scale models [45], we suggest that the new potential for CSI-AA to directly produce baseline δ15N isoscapes represents a major advance. Further, in initial data for selected resampling has shown that specific site-to-site offsets in bulk δ15N values have so far been highly reproducible [39]. This suggests that, while our initial CUE baseline isoscape is clearly based on relatively few locations, specific geographic variations are may also be meaningful (Fig. S1). While further sampling will be required to verify this conclusion, if mussel-derived δ15N values can indicate reproducible, fine scale geographic variation in baseline δ15N values, then our results suggest the potential to create highly detailed spatial maps of isotopic baselines, even in complex coastal environments.

Figure 6. δ15NPhe isoscape of the California Upwelling Ecosystem.

Color gradient bar indicates δ15N values. Isoscape is based on the linear relationship of δ15NPhe vs. latitude (R2 = 0.635, P = 0.0011). Squares represent sampling sites chosen for compound-specific isotope analysis; small black dots show all sampling sites for reference.

Overall, a CSI-AA approach for constructing baseline isoscapes could have broad importance in both modern and paleoceanographic studies. While CSI-AA based isoscapes could also be generated from other consumers, we suggest that for coastal zones mussels may be a particularly useful bioarchive. The combination of ubiquitous occurrence in many coastal regions, relatively long-term integration of microalgal isotopic signatures, and unambiguous source locations together would provide strong confidence in geographic patterns. We suggest that the degree to which mussel-derived isoscapes also reflect more offshore coastal waters will be an important topic for future work. If our mussel data indicates broad similarity to near-shore coastal isotopic data as a general result (Fig. 3), then sampling within largely existing shore-based programs might rapidly produce detailed, annualized, baseline isoscapes for the entire CCS. Such data could be invaluable in understanding the changing environmental factors driving spatial variability within the CCS; for example the effects of ENSO and PDO cycle effects on baseline δ15N isoscape gradients, and also provide a more clear understanding of isotopic baselines needed to evaluate possible long-term trends linked to climate change. We also note that isoscapes constructed using CSI-AA from mussels also would not necessarily be limited to coastlines. Mussels frequently attach to the base of fixed moorings located offshore (e.g., and therefore might be used to examine temporal change in isoscapes in many offshore instrumented locations. Finally, our results also suggest potential for paleoceanographic reconstructions. Mussel shells are often the major species found in archeological middens widely distributed from Baja California to Alaska along the US west coast. If source AA δ15N patterns were also well preserved in archeological shell, this could potentially extend the reconstruction of coastal baseline isoscapes back through much of the Holocene.

Supporting Information

Figure S1.

Alternate δ15N Isoscape approach. Alternate δ15N Isocape of the California Upwelling Ecosystem, showing δ15N gradients between sampling stations. As in text Fig. 6 in main text, color gradient indicates δ15N values. However, this isoscape interpolates between δ15N at each specific site. While CSI-AA data coverage was not large in this study, preliminary re-sampling has indicated offsets are reproducible. While clearly additional sampling would be required to verify such variation, this approach directly suggests the potential for high resolution isoscapes that capture finer scale regional patterns. Given the ubiquity of mussels along the CA coast, as well as relative ease of sampling, such high resolution coastal isoscapes of baseline δ15N might be readily constructed.


Figure S2.

δ15NAA patterns in the California Mussel (Mytilus californianus ). δ15N amino acid signatures of Mytilus californianus from 13 sampling sites selected for CSI-AA(values based on n = 4 analytical replicate injections). Absolute δ15N values normalized to the δ15NPhe, so that patterns can be compared. Measured amino acids are categorized into Trophic, Source, and Metabolic (M), based on relative changes with trophic transfer (see main text). Site and amino acid abbreviations are as defined in main text. Overall δ15NAA patterns conform closely to those expected from other heterotrophic organisms, with Trophic AA enriched in 15N vs. Source AA, and Thr strongly depleted in 15N.


Figure S3.

Low CSI-AA based Mussel Trophic Position Results. Relationships between measured δ15NGlu-Phe values vs. expectations for standard CSI-AA TP equations. Measured δ15NGlu-Phe of mussels are plotted vs. latitude (filled diamonds). Shaded bar on average value represents ± 1SD for entire data set. Assumed β values for primary producers are indicated by lower dotted line, (βGlu-Phe, 3.4 per mil). Commonly assumed TEFGlu-Phe for a single trophic transfer for a primary consumer (7.6 per mil ) is represented by upper dashed line. Arrow represents ΔGlu-Phe the theoretical isotopic enrichment from a TP1 to a TP2.


Figure S4.

Representative chromatogram of a GC-IRMS analysis of amino acids. Mussel amino acid gas chromatogram. A representative gas chromatogram of derivatized individual amino acids from Mytilus californianus. Abbreviations: Ala, alanine; Gly, Glycine; Thr, threonine; Ser, serine; Val, valine; Leu, leucine; Ile, isoleucine; Nor, Norleucine (internal standard); Pro, proline; Asp, aspartic acid, Met, Methionine; Glu, glutamic acid; Phe, phenylalanine; Lys, Lysine.


Text S1.

Mussel trophic position discussion.



We thank N. Quintana-Krupinsky, F. Batista, E. Gier and D. Andreasen for helping to collect, prepare, and/or process samples, and J. Felis for his support on map construction.

Author Contributions

Conceived and designed the experiments: NLV MDM. Performed the experiments: NLV. Analyzed the data: NLV MDM. Contributed reagents/materials/analysis tools: MDM. Wrote the paper: NLV MDM.


  1. 1. Graham BS, Koch PL, Newsome SD, McMahon KW, Aurioles D (2009) Using Isoscapes to Trace the Movements and Foraging Behavior of Top Predators in Oceanic Ecosystems Dordrecht: Springer Netherlands. pp. 299–318. doi:10.1007/978-90-481-3354-3_14.
  2. 2. McMahon KW, Ling Hamady L, Thorrold SR (2013) A review of ecogeochemistry approaches to estimating movements of marine animals. Limnol Oceangr 58: 697–714
  3. 3. Sigman DM, Kash KL, Casciotti KL (2009) Ocean process tracers: nitrogen isotopes in the ocean. Encyclopedia of ocean science, 2nd edn Elsevier, Amsterdam.
  4. 4. Liu Z, Altabet MA, Herbert TD (2008) Plio-Pleistocene denitrification in the eastern tropical North Pacific: Intensification at 2.1 Ma. Geochem Geophys Geosyst 9: : n/a–n/a. doi:10.1029/2008GC002044.
  5. 5. Sherwood OA, Guilderson TP, Batista FC, Schiff JT, McCarthy MD (2013) Increasing subtropical North Pacific Ocean. Nature: 1–26.
  6. 6. Jackson JBC (2001) Historical Overfishing and the Recent Collapse of Coastal Ecosystems. Science 293: 629–637
  7. 7. Ware DM, Thomson RE (2005) Bottom-up ecosystem trophic dynamics determine fish production in the Northeast Pacific. Science 308: 1280–1284.
  8. 8. DeNiro MJ, Epstein S (1978) Influence of diet on the distribution of carbon isotopes in animals. GEOCHIMICA ET COSMOCHIMICA ACTA 42: 495–506.
  9. 9. Vander Zanden M, Rasmussen JB (2001) Variation in δ15N and δ13C trophic fractionation: implications for aquatic food web studies. Limnol Oceangr 46: 2061–2066.
  10. 10. Post DM (2002) Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 83: 703–718.
  11. 11. McCutchan JH, Lewis WM, Kendall C, McGrath CC (2003) Variation in trophic shift for stable isotope ratios of carbon, nitrogen, and sulfur. Oikos 102: 378–390.
  12. 12. Vanderklift MA, Ponsard S (2003) Sources of variation in consumer-diet ?15N enrichment: a meta-analysis. Oecologia 136: 169–182
  13. 13. Chikaraishi Y, Ogawa NO, Kashiyama Y, Takano Y, Suga H, et al. (2009) Determination of aquatic food-web structure based on compound-specific nitrogen isotopic composition of amino acids. Limnology and Oceanography: Methods 7: 740–750.
  14. 14. Chikaraishi Y, Kashiyama Y, Ogawa NO, Kitazato H, Ohkouchi N (2007) Metabolic control of nitrogen isotope composition of amino acids in macroalgae and gastropods: implications for aquatic food web studies. Mar Ecol Prog Ser 342: 85–90.
  15. 15. Popp BN, Graham BS, Olson RJ, Hannides C, Lott MJ, et al. (2007) Insight into the Trophic Ecology of Yellowfin Tuna, Thunnus albacares, from Compound-Specific Nitrogen Isotope Analysis of Proteinaceous Amino Acids. Terrestrial Ecology 1: 173–190.
  16. 16. Hannides CC, Popp BN, Landry MR, Graham BS (2009) Quantification of zooplankton trophic position in the North Pacific Subtropical Gyre using stable nitrogen isotopes. Limnol Oceangr 54: 50.
  17. 17. Dale JJ, Wallsgrove NJ, Popp BN, Holland KN (2011) Nursery habitat use and foraging ecology of the brown stingray Dasyatis lata determined from stomach contents, bulk and amino acid stable isotopes. Mar Ecol Prog Ser 433: 221–236
  18. 18. Germain LR, Koch PL, Harvey J, McCarthy MD (2013) Nitrogen isotope fractionation in amino acids from harbor seals: implications for compound-specific trophic position calculations. Mar Ecol Prog Ser 482: 265–277.
  19. 19. Ruiz-Cooley RI, Ballance LT, McCarthy MD (2013) Range Expansion of the Jumbo Squid in the NE Pacific: δ15N Decrypts Multiple Origins, Migration and Habitat Use. PLoS ONE 8: e59651
  20. 20. Hannides CCS, Popp BN, Anela Choy C, Drazen JC (2013) Midwater zooplankton and suspended particle dynamics in the North Pacific Subtropical Gyre: A stable isotope perspective. Limnol Oceangr 58: 1931–1946
  21. 21. Choy CA, Davison PC, Drazen JC, Flynn A, Gier EJ, et al. (2012) Global Trophic Position Comparison of Two Dominant Mesopelagic Fish Families (Myctophidae, Stomiidae) Using Amino Acid Nitrogen Isotopic Analyses. PLoS ONE 7: e50133
  22. 22. McCarthy MD, Lehman J, Kudela R (2013) Compound-specific amino acid Î ´15N patterns in marine algae: Tracer potential for cyanobacterial vs. eukaryotic organic nitrogen sources in the ocean. GEOCHIMICA ET COSMOCHIMICA ACTA 103: 104–120
  23. 23. Sigman DM, DiFiore PJ, Hain MP, Deutsch C, Karl DM (2009) Sinking organic matter spreads the nitrogen isotope signal of pelagic denitrification in the North Pacific. 36: L08605
  24. 24. McCarthy MD, Benner R, Lee C, Fogel ML (2007) Amino acid nitrogen isotopic fractionation patterns as indicators of heterotrophy in plankton, particulate, and dissolved organic matter. GEOCHIMICA ET COSMOCHIMICA ACTA 71: 4727–4744
  25. 25. Seminoff JA, Benson SR, Arthur KE, Eguchi T, Dutton PH, et al. (2012) Stable Isotope Tracking of Endangered Sea Turtles: Validation with Satellite Telemetry and δ15N Analysis of Amino Acids. PLoS ONE 7: e37403
  26. 26. Kahru M, Mitchell BG (2000) Influence of the 1997-98 El Nifio on the surfacechlorophyll in. Geophys Res Lett 27: 2937–2940 Available:
  27. 27. Ford HL, Schellenberg SA, Becker BJ, Deutschman DL, Dyck KA, et al. (2010) Evaluating the skeletal chemistry of Mytilus californianusas a temperature proxy: Effects of microenvironment and ontogeny. Paleoceanography 25: PA1203
  28. 28. Rykaczewski RR, Dunne JP (2010) Enhanced nutrient supply to the California Current Ecosystem with global warming and increased stratification in an earth system model. Geophys Res Lett 37: : n/a–n/a. doi:10.1029/2010GL045019.
  29. 29. Rykaczewski RR, Checkley DM (2008) Influence of ocean winds on the pelagic ecosystem in upwelling regions. Proceedings of the National Academy of Sciences 105: 1965–1970.
  30. 30. Olson RJ, Popp BN, Graham BS, López-Ibarra GA, Galván-Magaña F, et al. (2010) Food-web inferences of stable isotope spatial patterns in copepods and yellowfin tuna in the pelagic eastern Pacific Ocean. Progress in Oceanography 86: 124–138
  31. 31. Raikow DF, Hamilton SK (2001) Bivalve diets in a midwestern U.S. stream: A stable isotope enrichment study. Limnol Oceangr 46: 514–522.
  32. 32. Dobush GR, Ankney CD, Krementz DG (1985) The effect of apparatus, extraction time, and solvent type on lipid extractions of snow geese. Canadian Journal of Zoology 63: 1917–1920.
  33. 33. Asmus RM, Asmus H (1991) Mussel beds: limiting or promoting phytoplankton? Journal of Experimental Marine Biology and Ecology 148: 215–232.
  34. 34. Altabet MA, Pilskaln C, Thunell R, Pride C, Sigman D, et al. (1999) The nitrogen isotope biogeochemistry of sinking particles from the margin of the Eastern North Pacific. Deep-Sea Research Part I 46: 655–679.
  35. 35. Décima M, Landry MR, Popp BN (2013) Environmental perturbation effects on baseline δ15N values and zooplankton trophic flexibility in the southern California Current Ecosystem. Limnol Oceangr 58: 624–634
  36. 36. Voss M, Dippner JW, Montoya JP (2001) Nitrogen isotope patterns in the oxygen-deficient waters of the Eastern Tropical North Pacific Ocean. Deep-Sea Research Part I 48: 1905–1921.
  37. 37. Larsen T, Ventura M, Andersen N, O'Brien DM, Piatkowski U, et al. (2013) Tracing Carbon Sources through Aquatic and Terrestrial Food Webs Using Amino Acid Stable Isotope Fingerprinting. PLoS ONE 8: e73441
  38. 38. Vokhshoori NL, Larsen T, McCarthy MD (2014) Reconstructing δ13C isoscapes of phytoplankton production in a coastal upwelling system with amino acid isotope values of littoral mussels. Mar Ecol Prog Ser. doi:10.3354/meps10746.
  39. 39. Vokhshoori N (2013) Using compound-specific amino acid isotope analysis in littoral mussels to generate baseline δ15N and δ13C isoscapes for the California Upwelling Ecosystem. Master's Thesis. Santa Cruz: University of California, Santa Cruz. 125 pp.
  40. 40. Gruber N, Sarmiento JL (1997) Global patterns of marine nitrogen fixation and denitrification. Global Biogeochem Cycles 11: 235–266.
  41. 41. Cline JD, Richards FA (1972) Oxygen deficient conditions and nitrate reduction in the eastern tropical North Pacific Ocean. Limnol Oceangr: 885–900.
  42. 42. Codispoti LA, Richards FA (1976) An analysis of the horizontal regime of denitrification in the eastern tropical North Pacific. Limnol Oceanogr 21: 379–388.
  43. 43. White AE, Foster RA, Benitez-Nelson CR, Masqué P, Verdeny E, et al. (2013) Nitrogen fixation in the Gulf of California and the Eastern Tropical North Pacific. Progress in Oceanography 109: 1–17 Available:
  44. 44. Sigman DM, Granger J, DiFiore PJ, Lehmann MM, Ho R, et al.. (2005) Coupled nitrogen and oxygen isotope measurements of nitrate along the eastern North Pacific margin. Global Biogeochem Cycles 19 : n/a–n/a. doi:10.1029/2005GB002458.
  45. 45. Somes CJ, Schmittner A, Altabet MA (2010) Nitrogen isotope simulations show the importance of atmospheric iron deposition for nitrogen fixation across the Pacific Ocean. Geophys Res Lett 37..
  46. 46. Schwing FB, O'Farrell M, Steger JM, Baltz K (1996) Coastal Upwelling indices west coast of North America. NOAA Tech Rep NMFS SWFSC 231: 144.
  47. 47. Sigman DM, Robinson R, Knapp AN, van Geen A, McCorkle DC, et al.. (2003) Distinguishing between water column and sedimentary denitrification in the Santa Barbara Basin using the stable isotopes of nitrate. Geochem Geophys Geosyst 4 : n/a–n/a. doi:10.1029/2002GC000384.
  48. 48. Checkley DM Jr, Barth JA (2009) Patterns and processes in the California Current System. Progress in Oceanography 83: 49–64
  49. 49. Kienast SS, Calvert SE, Pedersen TF (2002) Nitrogen isotope and productivity variations along the northeast Pacific margin over the last 120 kyr: Surface and subsurface paleoceanography. Paleoceanography 17: 7–1–7–17
  50. 50. Grantham BA, Chan F, Nielsen KJ, Fox DS, Barth JA, et al. (2004) Upwelling-driven nearshore hypoxia signals ecosystem and oceanographic changes in the northeast Pacific. Nature 429: 749–754.
  51. 51. Brandes JA, Devol AH, Deutsch C (2007) New Developments in the Marine Nitrogen Cycle. Chem Rev 107: 577–589
  52. 52. Waser N, Harrison PJ, Nielsen B, Calvert SE, Turpin DH (2010) Nitrogen isotope fractionation during the uptake and assimilation of nitrate, nitrite, ammonium, and urea by a marine diatom. Limnol Oceangr 43: 215–224 Available:
  53. 53. Behrenfeld MJ, O'Malley RT, Siegel DA, McClain CR, Sarmiento JL, et al. (2006) Climate-driven trends in contemporary ocean productivity. Nature 444: 752–755
  54. 54. Aksnes DL, Ohman MD (2009) Multi-decadal shoaling of the euphotic zone in the southern sector of the California Current System. Limnol Oceangr 54: 1272.
  55. 55. Polovina JJ, Howell EA, Abecassis M (2008) Ocean's least productive waters are expanding. Geophys Res Lett 35: L03618
  56. 56. McKinnell SM, Dagg MJ (2010) Marine ecosystems of the North Pacific Ocean, 2003-2008. PICES Special Publication 4: 393.
  57. 57. Rau GH, Ohman MD, Pierrot-Bults A (2003) Linking nitrogen dynamics to climate variability off central California: a 51 year record based on 15N/14N in CalCOFI zooplankton. Deep Sea Research Part II: Topical Studies in Oceanography 50: 2431–2447
  58. 58. Duggins DO, Simenstad CA, ESTES JA (1989) Magnification of secondary production by kelp detritus in coastal marine ecosystems. Science 245: 170–173.
  59. 59. Bracken ME, Menge BA, Foley MM, Sorte CJ, Lubchenco J, et al. (2012) Mussel selectivity for high-quality food drives carbon inputs into open-coast intertidal ecosystems. Mar Ecol Prog Ser 459: 53–62
  60. 60. Lorrain A, Graham B, Ménard F, Popp B, Bouillon S, et al. (2009) Nitrogen and carbon isotope values of individual amino acids: a tool to study foraging ecology of penguins in the Southern Ocean. Mar Ecol Prog Ser 391: 293–306