Omega-3 Pathways in Upwelling Systems: The Link to Nitrogen Supply

Omega-3 long-chain polyunsaturated fatty acids (hereafter, omega-3), including eicosapentaenoic-acid (EPA) and docosahexaenoic-acid (DHA), are essential nutritional compounds for humans, providing several benefits related to cardiovascular and neural health. Human intake of omega-3 occurs mostly via seafood, particularly fish. The primary source of omega-3 in aquatic systems is represented by primary producers, from which omega-3 are transferred throughout the food web. Nitrogen is an essential nutrient for primary producers and can be supplied to surface waters as nitrate upwelled from below, or as ammonium and other regenerated nitrogen forms recycled in situ. Eastern Boundary Upwelling Systems (EBUS) are the most productive marine systems on Earth, together covering only 2% of the ocean’s surface area but supporting 25% of the global fish catch, thereby providing food for humans. In EBUS, nitrate and other nutrients are advected to the surface to support the proliferation of a phytoplankton community dominated by known omega-3 producers (i.e., diatoms). Given current climate change-related projections of ocean warming, acidification, deoxygenation, and increased upwelling intensity, phytoplankton community composition in EBUS may change. Additionally, the global production of EPA + DHA is expected to decrease by up to 30%, rendering its supply for human consumption insufficient by 2050. Here we discuss the state of knowledge related to omega-3 transfer from phytoplankton to small pelagic fish in EBUS, including factors that can influence omega-3 production, links to nitrogen cycling, climate change implications for the omega-3 supply to humans, and suggestions for future research directions to improve our understanding of omega-3 in the ocean.


INTRODUCTION
Omega-3 long chain polyunsaturated fatty acids (hereafter, omega-3) such as eicosapentaenoicacid (20:5n-3, EPA) and docosahexaenoic-acid (22:6n-3, DHA) are essential compounds for marine consumers and human health (Ruess and Müller-Navarra, 2019), and are acquired through dietary intake (Arts et al., 2001). The main producers of omega-3 in the ocean are phytoplankton, with the amount of omega-3 contained in consumers being determined by the type of phytoplankton at the base of the food chain, and the transfer efficiency of omega-3 across trophic levels (Dalsgaard et al., 2003). In this context, zooplankton represent a key trophic link between the producers of omega-3 and small pelagic fish (e.g., sardine, anchovy, and herring; Bi and Sommer, 2020), the latter being a major source of omega-3 for human consumption (Hicks et al., 2019), as well as a main constituent of aquaculture feeds (Tacon and Metian, 2009). The supply of omega-3 is expected to change in the near future as a consequence of climate change (e.g., Hixson and Arts, 2016;Colombo et al., 2020), with implications for meeting the omega-3 requirements of a growing human population that is predicted to reach 11 billion by 2100 (Roser, 2013). Major climate-driven factors relevant to omega-3 production include warming (Hixson and Arts, 2016), ocean acidification (Dörner et al., 2020), and variability in upwelling intensity (Puccinelli et al., 2016a).
Coastal upwelling areas are amongst the most productive ocean regions owing to the wind-driven advection of nutrientrich waters to the surface that promote primary and secondary production (Peterson et al., 1988;Vargas et al., 2007;Flynn et al., 2018). In these regions, nitrate (NO 3 − ) upwelled from depth is the major source of nitrogen (N) for primary production (Waldron and Probyn, 1992;Messié et al., 2009;González-Galisteo et al., 2019), often supporting >>50% of phytoplankton growth Messié et al., 2009). At the cellular level, N is required for the synthesis of nucleic acids, amino acids, and chlorophyll (Dagenais-Bellefeuille and Morse, 2013), thus affecting omega-3 production (Yongmanitchai and Ward, 1991;Griffiths et al., 2012). Eastern Boundary Upwelling Systems (EBUS) occupy 2% of ocean surface but support 25% of the global fish catch and 8% of global primary production (Pauly and Christensen, 1995). These systems facilitate broadscale fertilization of the upper water column and sustain many of the world's important fisheries (Cury and Roy, 1989;Ward et al., 2006), including those centered around small pelagic fish, which provide up to 50% of annual human protein intake in some regions (e.g., West Africa; FAO, 2014). This is particularly relevant for the EBUS located off the African and South American continents, upon which millions of people rely for economic and food security. For instance, the Peruvian anchovy fishery accounts for 65% of global anchovy production and contributes 90% to the Peruvian fishery sector's Gross Domestic Product (Aronés et al., 2019).
Here we provide an overview of the current state of knowledge on omega-3 transfer from phytoplankton to small pelagic fish in EBUS, including factors that can influence omega-3 production, links to the N cycle, climate change implications for omega-3 availability to humans, current gaps in our understanding and suggestions for future developments needed to increase our knowledge of omega-3 in the ocean. As an essential compound for human health (Ruess and Müller-Navarra, 2019), omega-3 could aid in ameliorating malnutrition, thus playing a fundamental role in food security. Additionally, understanding the role that upwelling plays in omega-3 production and how this may vary with climate change, is relevant to the following Sustainable Development Goals (SDGs): zero hunger (SDG2), climate action (SDG13), and life under the sea (SDG14).

OMEGA-3 PRODUCTION AND ROLE IN THE OCEAN
In the ocean, omega-3 are mostly synthesized by phytoplankton, and phytoplankton community composition is a major factor regulating the amount of omega-3 produced in a particular system (Galloway and Winder, 2015). High food-quality species are characterized by high levels of EPA and/or DHA [e.g., diatoms and dinoflagellates, EPA and/or DHA > 25% of total fatty acids (FA); Hixson and Arts, 2016;Jónasdóttir, 2019] and low food-quality species by lower concentrations of EPA + DHA [e.g., cyanobacteria, which produce mostly shortchain polyunsaturated FA (PUFA), i.e., ≤C18 PUFA; Parrish, 2013]. Specifically, diatoms appear to be a major producer of EPA (Dalsgaard et al., 2003), while dinophytes and haptophytes are the main sources of DHA in the ocean (Figure 1; Remize et al., 2020). FA production can be affected by abiotic factors including temperature, which plays a key role in controlling the level of FA unsaturation, influencing the fluidity of cell membranes (Rousch et al., 2003); light availability, which affects phytoplankton species composition (Harrison et al., 1990); and salinity, which influences membrane permeability (Chen et al., 2008).
Omega-3 are a limiting factor for zooplankton and higher trophic levels production (Bi and Sommer, 2020), as these organisms cannot synthesize these FA in sufficient quantities to support their nutritional requirements and must thus acquire them through their diet (Dalsgaard et al., 2003;Litzow et al., 2006). Omega-3 are precursors to eicosanoid hormones and a major component of cell membranes (Parrish, 2013), playing a fundamental role in maintaining membrane functioning, which can be modified by changing the length and the level of unsaturation of the FA carbon chains (Sinensky, 1974;Guschina and Harwood, 2006). The essential role of omega-3 in marine organism fitness is widely known. For instance, zooplankton growth, development and reproduction output are dependent on the amount of omega-3 available in their food (Müller-Navarra et al., 2000;Brett et al., 2006;Ravet et al., 2010). Similarly, fish brain and optic tectum development and oocyte quality depend on the DHA content of their prey (Bell et al., 1995;Garrido et al., 2007), and fish larvae with a higher EPA + DHA content present enhanced swimming ability, schooling behavior, and higher survival rates after hatching (Ishizaki et al., 2001;Mourente, 2003).

LINKS BETWEEN MARINE NITROGEN CYCLING AND OMEGA-3
Phytoplankton community composition depends on several factors including nutrient availability. In the euphotic zone, phytoplankton growth can be supported by NO 3 − supplied from below, for instance via upwelling (yielding "new production"), or by N forms such as ammonium (NH 4 + ) recycled in the FIGURE 1 | Schematic representation of omega-3 pathways in an upwelling system across trophic levels, focusing on the link between nitrogen (N) sources [newly upwelled N (NO 3 − , white) and recycled N (NH 4 + , orange)], select primary producers (diatoms, dinoflagellates, autotrophic bacteria), primary consumers (zooplankton), and higher trophic levels (small and large pelagic fish). The thickness of the white and orange lines indicates the relative use of upwelled NO 3 − and recycled NH 4 + by the different phytoplankton groups. Pie charts show the % of eicosapentaenoic-acid (EPA), docosahexaenoic-acid (DHA), and the sum of the other fatty acids (Other) produced by the three phytoplankton groups. With climate change, upwelling is predicted to intensify, which may increase the upward supply of NO 3 − and drive the ecosystem towards enhanced diatom growth, higher transfer efficiency, and elevated omega-3 production (i.e., the thick white lines will become thicker). Alternately, warming may drive a net decline in upwelled NO 3 − , resulting in enhanced NH 4 + recycling, a shift toward smaller phytoplankton species, less efficient trophic transfer, and a decline in omega-3 production (i.e., the thick orange arrows will become thicker, and the white arrows will become thinner). We note that these scenarios represent just two possibilities, and that there are likely other potential outcomes. Additionally, it is conceivable that the four EBUS will not respond to climate change in the same way.
upper ocean ("regenerated production"; Dugdale and Goering, 1967). When available, NH 4 + is often the preferred N form consumed by phytoplankton because its low oxidation state (-3, the same as organic N) renders its assimilation energetically favorable (Dortch, 1990). However, surface NH 4 + concentrations are typically << 1 µM (Paulot et al., 2015), although its flux through the food web can be large (Gruber, 2008;Altieri et al., 2021). High NO 3 − availability typically leads to growth of large phytoplankton including diatoms (Kudela and Dugdale, 2000;, while smaller phytoplankton such as cyanobacteria tend toward NH 4 + dependence (Probyn and Painting, 1985;. Upwelling events are among the main supply mechanisms of nutrients essential for phytoplankton production (e.g., NO 3 − , silicate, iron; Dugdale and Wilkerson, 1998;Chavez and Messié, 2009). In upwelling systems, the ratio between new and total (i.e., new + regenerated) production is significantly higher than in the open ocean (0.5-0.8 versus ∼0.1; Eppley and Peterson, 1979;Laws, 2004) due to the near-constant supply of NO 3 − , which underpins the elevated productivity of these regions (Messié and Chavez, 2015). Similarly, higher-than-average surface-water iron concentrations favor phytoplankton consumption of upwelled NO 3 − over recycled N (Dugdale and Goering, 1967;Bruland et al., 2001), leading to a phytoplankton community dominated by omega-3-rich diatoms. A change in community from diatoms to dinophytes is commonly observed during an upwelling cycle, especially when silicate, essential for the synthesis of diatom frustules, becomes depleted (Martin-Jézéquel et al., 2000;Tilstone et al., 2000).
Since different phytoplankton contain different amounts of omega-3, changes in community composition can cause variations in omega-3 transfer to consumers, ultimately affecting the omega-3 supply throughout the food web (Bermúdez et al., 2016;Garzke et al., 2016). For instance, it was shown that an increase in pCO 2 caused a shift from nanophytoplankton to picoeukaryotes along with a decline in PUFA in both phytoplankton size fractions, which was mirrored by a PUFA reduction in the copepod species feeding on them (Bermúdez et al., 2016). Similarly, interannual FA variations in small pelagic fish in the California Current was attributed to a shift from a dinoflagellate-to a diatom-dominated food source (Litz et al., 2010). Upwelling food webs are made more complex by the presence of mixotrophic nanoflagellates, organisms that are able to directly consume bacteria and picoplankton (Frias-Lopez et al., 2009), but that can also contribute significantly to primary production, especially during periods of upwelling relaxation (Figueiras et al., 2020). Disentangling omega-3 produced by autotrophic versus mixotrophic organisms in the ocean is challenging, and may lead to an over/underestimation of the omega-3 supply from primary producers.

THE EASTERN BOUNDARY UPWELLING SYSTEMS
The world's four major EBUS are associated with the Benguela, California, Canary, and Humboldt Currents. Rates of primary production are comparable among these regions (>900 gCm −2 y −1 , Table 1), with the exception of the California system where they are two-fold lower (480 gCm −2 y −1 ), probably due to the lower NO 3 − supply to the euphotic zone ). The availability of regenerated N and the mechanisms of its production (i.e., herbivores, heterotrophic bacteria, higher trophic-level organisms) vary among the systems (Dugdale, 1972). For instance, anchovies are responsible for most N regeneration in the Humboldt system (Kämpf and Chapman, 2016), while in the southern Benguela, up to 48% of the subsurface nutrients are regenerated on-shelf by heterotrophic bacteria (Flynn et al., 2020). EBUS fertilization by dust deposition also influences productivity, especially in the Canary and Benguela systems where the Sahara and Namib Deserts supply nutrients that regularly limit phytoplankton growth, including iron and phosphate (Jickells et al., 2005;Pabortsava et al., 2017).
The high primary productivity of EBUS propagates throughout their ecosystems, with a given concentration of phytoplankton supporting approximately 25-and 50-times as much zooplankton biomass and fish production Kämpf, and Chapman, 2016). In the Humboldt system, fish production and zooplankton biomass are an order of magnitude higher than in the other three systems (Table 1), which has been attributed to a high transfer efficiency among trophic levels, resulting in optimal conditions for the proliferation and retention of zooplankton and fish . The same pattern is not evident in omega-3 content, which varies considerably within and among systems. For instance, Benguela phytoplankton contain the lowest amount of omega-3, while omega-3 in zooplankton and small pelagic fish is higher compared to the other EBUS. We note, however, that limited information is available on omega-3 in EBUS, especially for the Canary system, precluding our ability to generalize while also highlighting the need for further investigations.

CLIMATE CHANGE EFFECTS ON UPWELLING SYSTEMS AND N SUPPLY
As a result of climate change, upwelling frequency appears to be increasing in several EBUS (Sydeman et al., 2014) due to increasing wind stress (Narayan et al., 2010). This may increase the nutrient supply to phytoplankton, favoring large diatoms and higher transfer efficiency, yielding enhanced production of omega-3. However, changes to upwelling are likely to be complex (Rykaczewski et al., 2015), and along with warming and/or acidification of surface waters, will modify phytoplankton metabolism and community composition, potentially favoring smaller phytoplankton (e.g., picoeukaryotes) at the expense of the nano-and microphytoplankton (e.g., diatoms, chlorophytes, haptophytes; Pérez et al., 2010;Armbrecht et al., 2014). A decrease in phytoplankton cell size will likely lead to longer food webs and a reduction in trophic transfer efficiency, with a consequent decline in omega-3 transfer to higher trophic levels (Berglund et al., 2007;Bermúdez et al., 2016).
Global warming is predicted to enhance seawater stratification while decreasing ocean mixing, reducing the upward NO 3 − supply (Falkowski et al., 1998;Sarmiento et al., 1998;Capotondi et al., 2012) and favoring the growth of species such as cyanobacteria and mixotrophic nanoflagellates (Peter and Sommer, 2012;Leles et al., 2018). In general, smaller phytoplankton preferentially consume NH 4 + Glibert et al., 2016), such that their growth should be favored under lower-nutrient conditions, in comparison to that of diatoms, which mainly utilize NO 3 − (Kudela and Dugdale, 2000;. In contrast, by increasing the bioavailability of micronutrients such as iron, ocean acidification may promote the growth of phytoplankton reliant on NO 3 − , the assimilation of which has a high iron requirement (Price et al., 1991;Schoffman et al., 2016). That said, increased iron availability is also predicted to favor the growth of N 2 fixers (Wannicke et al., 2018), a group of cyanobacteria (Monteiro et al., 2010) that produce predominantly short-chain PUFA (Parrish, 2013).
consumers have different omega-3 requirements (Litzow et al., 2006), such that their responses to climate change are likely to be variable and difficult to predict.

KNOWLEDGE GAPS AND FUTURE DIRECTIONS
There is broad consensus that upwelling shapes the diets of primary and secondary consumers, from coastal (Puccinelli et al., 2016a,b;Docmac et al., 2017) to open-ocean environments Schukat et al., 2021). Globally, the omega-3 supply from fish is projected to soon become insufficient for the growing human population, although the sign and magnitude of future changes in EBUS omega-3 production are unknown.
Global estimates indicate that phytoplankton and zooplankton EPA + DHA supplies amount to 80 Mt, yet approximately 90% of EPA + DHA is lost between primary producers and higher predators (Hamilton et al., 2020). Humans require 1.4 Mt of EPA + DHA annually to meet the 500 mg daily recommended intake, yet only 0.8 Mt are available, largely from aquaculture and fisheries (Tocher, 2015;Hamilton et al., 2020). Despite the general understanding that the omega-3 supply to humans is limited, information on omega-3 production and transfer in productive upwelling systems is scarce, and almost absent at the level of primary consumers such as zooplankton and small pelagic fish ( Table 1). The available studies are largely laboratory-or mesocosm-based investigations that constrain selected environmental parameters in order to investigate a specific metabolic or physiological pathway (De Troch et al., 2012;Mayor et al., 2015). Meta-analyses have been used as an alternative tool to formulate predictions related to omega-3 supply (e.g., Galloway and Winder, 2015;Hixson and Arts, 2016), but they are usually conducted at a large scale (regional, global) and focused on primary producers. Recently, an algorithm that links satellite chlorophyll-a data and EPA production was developed for the northwestern Atlantic and used for global EPA estimates (Budge et al., 2014). Applying this algorithm to EBUS that collectively support 8% of global primary production yields an EBUS EPA production rate of 19.2 Mt year −1 . Assuming a transfer efficiency of 10% among trophic levels and a three-level food chain, we estimate that 0.19 Mt year −1 of EPA is available in the small pelagic fish of EBUS. While this approach is useful for illustrating the role of EBUS in omega-3 production, it has several limitations, such as assuming that diatoms are the only EPA producers, neglecting DHA production, and assuming a 10% transfer efficiency.
Additionally, while the studies cited above have advanced our knowledge of omega-3 in the ocean, they do not provide reliable estimates of EPA + DHA or consider future changes in omega-3 production. Coupled physical-biogeochemical models are powerful tools for understanding ocean dynamics and their role in biogeochemical cycles at various spatial (from global to regional) and temporal (event to climate) scales (Brasseur et al., 2009;Bachèlery et al., 2016). Biogeochemical models are often used to represent the cycles of carbon and selected nutrients (e.g., N), but they can also simulate energy flow to higher trophic levels (Koné et al., 2005;Chust et al., 2014). These models encompass a variety of complexities and applications, ranging from Nitrate-Phytoplankton-Zooplankton-Detritus to more complex planktonic food webs (Rose et al., 2007;Aumont et al., 2015;Butenschön et al., 2016). They depend, however, on field observations and experimental data to define parameters and constrain assumptions. 3D-physical-biogeochemical models have been developed for the four EBUS to simulate marine productivity and carbon cycling at low trophic levels (i.e., phytoplankton and zooplankton) (Lachkar and Gruber, 2012;Gutknecht et al., 2013;Auger et al., 2016;Espinoza-Morriberón et al., 2017;Glock et al., 2018;Cheresh and Fiechter, 2020). While the models account for the N source(s) available to phytoplankton and the main nutrients ratios (cell-quota formulation), none of them track omega-3 in either the primary or secondary consumers. The only mathematical model to investigate omega-3 transfer between phytoplankton and zooplankton is the mechanistic approach developed by Perhar et al. (2012Perhar et al. ( , 2013, which accounts for omega-3 availability in its calculation of grazing efficiency. This model has been used to investigate the role of N, phosphorus and omega-3 in zooplankton growth and abundances. A combined physicalbiogeochemical model that includes the approach of Perhar et al. (2012Perhar et al. ( , 2013 could be implemented for EBUS to (i) elucidate the mechanisms governing the fate of omega-3 production based on the N source(s) available to phytoplankton, and (ii) quantify the amount of omega-3 available to small pelagic fish, at present and in the future, considering EBUS-specific parameters.

SUMMARY
Omega-3 are essential compounds for all living organisms and their supply is tightly connected to phytoplankton growth and community composition in the ocean, as well as to the efficiency of their transfer through the food web. Human population growth and the likely decline in omega-3 availability linked to global warming underscore the urgency of prioritizing long-term predictions of omega-3 supply in EBUS. In particular, there is a clear need for a predictive model to explore climate-induced changes in omega-3 supply throughout the food web; such a model would be valuable to multiple sectors, including fisheries management and food production.

AUTHOR CONTRIBUTIONS
EP, LP, SF, EM, and PS conceptualization. EP wrote the first manuscript draft with major contribution from FS, SF, and PS. All authors contributed to the article and approved the submitted version.

FUNDING
This work is supported by the ISblue project, Interdisciplinary graduate School for the blue planet (ANR-17-EURE-0015) and co-funded by a grant from the French government under the program "Investissements d'Avenir." This work is a contribution to the SOLAB project (ANR-18-CE32-0009). SF acknowledges the support of a Royal Society/African Academy of Sciences FLAIR fellowship.