Seagrass meadows provide foundational coastal services—stabilizing sediments, sequestering carbon, and supporting fisheries (Unsworth et al., 2018; Kent et al., 2021; Serrano et al., 2021). Yet many populations persist close to their light limits, making them highly sensitive to turbidity pulses and chronic eutrophication (Ralph et al., 2007; Bulmer et al., 2018; Van Katwijk et al., 2024). Although nutrient supply can alter growth and physiology, restoration trajectories often hinge on maintaining a positive carbon balance, which is primarily governed by irradiance at the canopy (Ralph et al., 2007). Accordingly, disentangling the roles of light and nitrogen remains both a practical and theoretical priority for management and recovery planning (Ralph et al., 2007; Van Katwijk et al., 2024).

Zostera marina is a foundation species widely distributed in temperate coastal ecosystems of the Northern Hemisphere (Green and Short, 2003; Olsen et al., 2016). Importantly, Z. marina has been extensively used as a model seagrass species in studies examining light limitation (Bertelli and Unsworth, 2018) and nutrient enrichment (Van Katwijk et al., 1997), providing a well-established empirical basis for evaluating how irradiance and ammonium availability jointly shape early growth trajectories under controlled conditions relevant to coastal eutrophication.

Light governs both instantaneous photochemistry and whole-plant carbon economics in seagrasses: when irradiance drops below compensation, positive carbon balance cannot be maintained, precipitating growth decline and range contraction (Ralph et al., 2007). Consistent with this framework, short-term shading rapidly depletes rhizome and root non-structural carbohydrates (NSC) and reduces biomass and shoot density, reflecting a shift from carbon accrual to carbon expenditure (Burke et al., 1996; Wong et al., 2021). Early-warning signals of this transition typically include declines in photophysiology (Fv/Fm), suppressed leaf extension, and reductions in above- and below-ground biomass, often accompanied by accelerated tissue turnover (senescent leaf loss and below-ground mass loss) (Burke et al., 1996; Ralph et al., 2007; Wong et al., 2021). At the same time, as excitation pressure wanes, chloroplast redox control follows a ‘push–pull’ behavior in which reactive oxygen species (ROS) generation depends on light-driven reductant supply, so oxidative damage does not necessarily intensify under shading (Foyer and Noctor, 2016).

Nitrogen supply can modulate the light-driven trajectory in two opposing ways. Low–moderate enrichment may alleviate nutrient limitation and transiently stimulate growth, whereas high ammonium (NH₄⁺) or co-stress with low light imposes metabolic costs and can become toxic (Van Katwijk et al., 1997; Touchette and Burkholder, 2000). Mechanistically, low irradiance enhances ammonium toxicity because photosynthesis and the carbon stores required to assimilate NH₄⁺ into amino acids are reduced (Collos and Harrison, 2014), consistent with reports that elevated ammonium and low light form a dangerous synergy for eelgrass (Villazán et al., 2013). Importantly, the magnitude and direction of nitrogen effects depend strongly on dose, exposure duration, and light context, highlighting the need to evaluate nitrogen responses within ecologically realistic concentration ranges rather than assuming uniformly negative effects (Van Katwijk et al., 1997; Touchette and Burkholder, 2000). Critically, nitrogen effects depend on dose, exposure duration and light context; therefore, diagnosing limitation is best achieved with a multi-metric suite spanning morphology (shoot density/biomass), photophysiology (F_v/F_m) and, where feasible, biogeochemistry (NSC, δ-isotopes) rather than reliance on a single indicator (McMahon et al., 2013; Collier et al., 2016; Roca et al., 2016; Statton et al., 2017). Within this framework, antioxidant enzyme activities (e.g. SOD, CAT, GSH-Px, POD) and lipid peroxidation products (MDA) provide complementary indicators of cellular redox balance and photo-oxidative stress, allowing discrimination between carbon limitation driven by light reduction and stress responses associated with excess excitation energy (Foyer and Noctor, 2016; Bertelli and Unsworth, 2018).

Here, we employed intact Z. marina cores from Guzhenkou Bay in a fully crossed 4 × 4 irradiance (180, 90.6, 17, 0 μmol photons m^-2 s^-1) × NH₄Cl addition (0, 25, 50, 100 μmol L^-1) matrix over five weeks. The experimental gradients were designed to represent generalized light and ammonium conditions relevant to coastal eutrophication, rather than to reproduce site-specific environmental conditions at Guzhenkou Bay. The selected ammonium concentrations span background to elevated levels commonly reported in eutrophic coastal waters, while remaining below or near thresholds previously associated with acute toxicity in Z. marina, thereby enabling assessment of dose-dependent responses under contrasting light regimes (Van Katwijk et al., 1997; Villazán et al., 2013). Our aim was to determine whether irradiance or ammonium enrichment exerts the dominant control on early structural, physiological and isotopic responses. We quantified (i) structural metrics (shoot density, leaf length, above- and below-ground biomass, leaf shedding rate and below-ground mass loss), (ii) photophysiological/pigment parameters (F_v/F_m, Rubisco activity, chlorophyll a and b), (iii) antioxidant enzyme activities (SOD, CAT, GSH-Px, POD) and malondialdehyde (MDA), and (iv) elemental/isotopic composition (C, N, δ¹³C, δ¹⁵N). Because each light × nutrient cell lacked replicates, we restricted inference to main effects of irradiance and ammonium, acknowledging that interaction effects are confounded with residual error. We hypothesized that (1) irradiance would dominate responses across hierarchical levels—indicative of carbon limitation under shading, (2) within the tested ammonium range, NH₄Cl alone would induce minimal structural or photophysiological effects but would be traceable isotopically via declining δ¹⁵N, and (3) antioxidant enzyme and MDA responses would reflect reduced photo-oxidative load under low irradiance rather than increased oxidative damage. Our results support these hypotheses and suggest a tiered, field-transferable monitoring framework aligning rapid photophysiology with structural and isotopic indicators, while reaffirming water-clarity (and thus canopy irradiance) as the primary management focus.

Our results indicate that irradiance, rather than NH₄Cl enrichment within the tested range, dominated the observed early responses of Zostera marina across morphological and physiological traits. Importantly, because this fully crossed design lacked within-cell replication, statistical inference is restricted to main effects only; thus, the discussion below interprets patterns as evidence of dominant main effects and does not test or attribute causality to specific light × NH₄Cl combinations.

Pronounced declines in shoot density, AGB, BGB, and F_v/F_m under shading are consistent with the classic light-requirement theory for seagrasses, which states that most species require at least 10–20% of surface irradiance to maintain a positive carbon balance (Duarte, 1991; Dennison et al., 1993; Duarte et al., 2005; Ballesteros et al., 2007). For Z. marina, minimum light thresholds of approximately 10% of surface irradiance have been reported [e.g., 11% SI in a Danish embayment; (Lee et al., 2007)], below which shoot mortality and biomass loss may occur. This response is attributed to the fact that, below the compensation irradiance, respiration exceeds photosynthesis, leading to carbohydrate depletion and eventual die-off (Alcoverro et al., 1999; Ralph et al., 2007; Hintz et al., 2022). In agreement, we observed F_v/F_m values falling below 0.2 and near-total biomass collapse under EL conditions, indicating severe photoinhibition and insufficient photochemical energy to support net carbon gain. Such extremely low Fv/Fm values (~0.2) imply a large fraction of nonfunctional PSII reaction centers (Gorbunov and Falkowski, 2022), a pattern commonly observed in seagrasses that are dead or dying (Biber et al., 2004) when light falls below minimum requirements (Bertelli and Unsworth, 2018). For context, unstressed leaves typically show Fv/Fm ≈ 0.83 (Björkman and Demmig, 1987).

Conversely, NH₄Cl enrichment did not elicit significant stress responses under our high-light conditions, in line with mesocosm evidence showing that ammonium becomes deleterious primarily when combined with low light, whereas no such effects occur under high light; moreover, moderate (<10 µM) DIN can even stimulate growth when nutrient-limited (Villazán et al., 2013). At higher concentrations, ammonium toxicity in Z. marina is well documented (Van Katwijk et al., 1997).

Shading shifted plant carbon balance from accrual to expenditure, consistent with reported NSC depletion and declines in both aboveground and belowground biomass (Burke et al., 1996). In our experiment, the increase in aboveground loss and the decline of belowground mass are consistent with a canopy-thinning strategy that lowers whole-plant respiratory demand, while mobilizing rhizome carbohydrates to sustain metabolism under low light (Wong et al., 2021). This dynamic aligns with seagrass light-stress theory: rapid photophysiological impairment under shading precedes growth and survival losses as net carbon balance turns negative (Ralph et al., 2007; Bertelli and Unsworth, 2018). Consequently, the observed decrease in below-ground carbon content provides a strong signature of carbon limitation.

Antioxidant enzyme responses were variable among treatments. In our study, POD activity increased under reduced irradiance, while SOD, CAT and GSH-Px showed no consistent trends, which is plausible under diminished photo-oxidative pressure (lower excitation energy at PSII/PSI) where ROS generation is limited by reductant availability (Foyer and Noctor, 2016). Such responses are consistent with the view that oxidative stress in seagrasses is context-dependent—shaped by light history, stress duration and co-stressors (Bertelli and Unsworth, 2018; Costa et al., 2021). Importantly, our MDA results—no increase under shading and occasionally higher values under high irradiance—support a mechanism where low light does not exacerbate oxidative damage, but higher ROS generation occurs when photosynthetic electron transport is more active, in line with chloroplast “push–pull” redox theory (Foyer and Noctor, 2016), and with observations that increases in light or temperature do not necessarily elevate MDA levels in seagrasses (Costa et al., 2021).

Across traits, NH₄Cl enrichment alone showed minimal main effects, consistent with evidence that low–moderate inorganic N additions, typically reported at concentrations <10 µM may stimulate seagrass growth rather than cause toxicity (Touchette and Burkholder, 2000; Villazán et al., 2013). Ammonium toxicity in Z. marina typically emerges at relatively high water-column concentrations—on the order of ~10² µM—and can be exacerbated by low light (Van Katwijk et al., 1997). Because our highest addition (100 μmol L^-¹) sits close to this lower toxicity bound, we interpret the weak ammonium responses as conditional on the short experimental duration and light context, and anticipate that clearer effects could develop with longer exposure and/or under stronger light limitation. Given that light limitation is a primary proximal stressor for seagrasses, and eutrophication often depresses performance via indirect light reduction (Ralph et al., 2007), our data align with the view that N effects are contingent on concentration and context, whereas light is the dominant immediate driver in this experiment.

Isotopically, both aboveground and belowground δ¹⁵N declined with NH₄Cl addition, consistent with assimilation of ¹⁵N-depleted ammonium and isotope fractionation during uptake/assimilation (Ariz et al., 2011; Foscari et al., 2021). Leaf δ¹⁵N is widely used to trace nitrogen sources in seagrass and coastal producers, with lower values indicating greater reliance on ammonium when source δ¹⁵N and/or uptake fractionation favor ¹⁴N (Kahmen et al., 2008; Jones et al., 2018). In Z. marina, DIN-uptake/δ¹⁵N studies support this interpretation, so the observed δ¹⁵N shift provides an internal tracer corroborating NH₄⁺ assimilation despite minimal whole-plant morphological/physiological responses (Robinson, 2001; Sandoval-Gil et al., 2015).

The present experimental design does not explicitly separate direct ammonium uptake through leaves from indirect effects mediated via sediment processes. Ammonium was added to the water column, and given the relatively short experimental duration, direct uptake by aboveground tissues likely contributed substantially to the observed nitrogen-related responses. This interpretation is supported by previous uptake studies showing that maximum ammonium uptake rates in aboveground tissues of Z. marina are substantially higher than those in belowground tissues (approximately 2.6-fold) (Li et al., 2020). Nevertheless, sediment-mediated processes, including microbial transformation and changes in sediment nitrogen availability, may also contribute to plant nitrogen acquisition, particularly over longer timescales (Herbert, 1999).

Although treatment effects on δ¹³C were not statistically significant, we observed a slight tendency toward less negative values under EL–L (−11.4‰) relative to H (−14 to −16.5‰). We note, however, that most experimental and review studies report δ¹³C enrichment under higher irradiance, as increased photosynthetic demand and boundary-layer drawdown (and/or greater HCO₃^- use) reduce discrimination, yielding less negative δ¹³C (Grice et al., 1996; Hemminga and Mateo, 1996; Lepoint et al., 2004). Conversely, shading typically increases discrimination and drives more negative δ¹³C (Durako and Hall, 1992). Our non-significant pattern should therefore be interpreted cautiously and may reflect transient DIC speciation, site-specific boundary-layer conditions, or tissue turnover effects. One possible explanation is that leaf and rhizome δ¹³C reflect carbon acquisition integrated over tissue turnover times that may exceed the five-week experimental duration, thereby constraining the development of strong isotopic responses under short-term light manipulation.

We acknowledge that this unreplicated factorial design limits formal statistical inference and precludes testing for interaction effects. Consequently, the present results should be interpreted as indicative of dominant main effects during early response stages, rather than as definitive tests of light × nitrogen interactions under the present experimental design.

Our controlled microcosm shows that irradiance, not ammonium enrichment within 0–100 μmol L^-1, governs the early trajectory of Zostera marina. Shading depressed shoot density, AGB, BGB and F_v/F_m, increased tissue turnover, and reduced below-ground carbon—an integrated signature of light-driven carbon limitation. In contrast, NH₄Cl produced minimal main effects on morphology, pigments, F_v/F_m or MDA and yielded isotopic, not structural, evidence of uptake via declining δ¹⁵N. Antioxidant responses were variable (POD increase; other enzymes inconsistent), and MDA did not rise under shading, consistent with reduced excitation pressure at low light. Collectively, these results argue for tiered diagnostics: rapid photophysiology (F_v/F_m) to flag approaching thresholds, structural metrics (shoot density, AGB/BGB, turnover) to quantify functional loss, and integrative biogeochemical markers (below-ground C; δ-isotopes) to assess resource balance. From a management perspective, maintaining water clarity to meet species-specific minimum light requirements remains paramount, which can be operationalized in practice using routinely available optical proxies such as underwater PAR, Secchi depth, or turbidity-based estimates within existing coastal monitoring programs; nitrogen effects are context- and dose-dependent. Methodologically, the absence of within-cell replication precluded formal interaction tests, so subtle light × NH₄⁺ effects may have been missed. Longer exposures and graded NH₄⁺ treatments should explicitly resolve interaction thresholds and link optical conditions to % surface-light targets, while direct NSC measurements will strengthen inferences connecting photophysiology, turnover and survival.