Plants were collected from a monospecific Zostera marina L. meadow situated in western Ireland at Finavarra (FV) (53.149437N, −9.133993W), southern Galway Bay (Figure 1). This area is classified as a Special Area of Conservation (SAC), Natural Heritage Area (NHA), and described as an “unpolluted” area by the Irish Environmental Protective Area (EPA) (NPWS, 2014) (more details in Beca-Carretero et al., 2019a, 2020).

Mature apical shoots were manually harvested by SCUBA diving or snorkeling at a shallow depth of 2–3 meters at intervals of 5–10 meters along a 50 m transect line to prevent potential resampling of the same genotypes. At this depth range, the meadow is denser and more homogenous than at shallower or deeper areas (Beca-Carretero et al., 2019a). We selected shoots with similar weights (average of 2.7 ± 0.6 g DW shoot^–1) and size (average of 32.3 ± 8.3 cm shoot^–1) and the number of mature rhizome segments (3–4). Shoots were transferred to cooling tanks filled with ambient seawater and directly transported to the laboratory at the National University of Ireland Galway campus (Supplementary Figure 1) within 1 h. The collected samples were carefully cleaned, e.g., any remnants of sediment, epiphytic organisms or dead plant materials were removed from the shoots.

We conducted laboratory-controlled experiments on Z. marina plants across four different times of year (April, July, November 2015 and January 2016), and each time, exposed samples to combined temperature (6, 12, 16, 20 and 24°C) and irradiance (180 μmol photons m^–2 s^–1 and 60 μmol photons m^–2 s^–1) treatments (Figure 2).

The experimental temperature range represented the natural annual range of sea surface temperature (SST) in western Ireland (∼5–17°C) (Supplementary Figure 1); the 20°C treatment represented the predicted increases of 2.6–3.1°C of SST for the study region by the end of 2100 (IPCC, 2014; Hobday et al., 2016; Tinker and Howes, 2020); 24°C temperature represented an elevation by +3–4°C above maximum in situ SST at the collection site (∼ 19−20°C, Stengel et al., 1999) that may be reached during a summer heat-wave events in the Atlantic Ocean (Feudale and Shukla, 2011). The duration of the temperature treatment was selected to simulate the duration of previous warming episodes in temperate regions (Hobday et al., 2016).

For each temperature treatment, two irradiance levels (180 μmol photons m^–2 s^–1 and 60 μmol photons m^–2 s^–1) were tested. Light levels were chosen based on in situ values recorded in September 2014 (maximum annual temperatures of 16 –17°C) during (typical) cloudy days at the location where plants were collected for the experiment. It is noteworthy that in western Ireland 63% of days have some precipitation and more than 96% days have more than 20% cloud cover¹. Therefore, we considered that using light levels recorded during cloudy days were ecologically more relevant for Irish coastal ecosystems than those during sunny days. Irradiances measured using a DIVING-PAM-II² along a 50 m transect long were 174.7 ± 42.4 μmol photons m^–2 s^–1 [n = 8]) at intermediate depths of Irish seagrass meadows (2–2.5 m), which is ∼70% higher than at deeper regions of the meadow (4–5 m; 57.6 ± 25.4 μmol photons m^–2 s^–1 [n = 8]) (Beca-Carretero et al., 2019b). Similar light levels were also previously used in previous Z. marina experiments (e.g., Evans et al., 1986; Höffle et al., 2011; Staehr and Borum, 2011; Beca-Carretero et al., 2018a). A cycle of 12 h: 12 h light (L): dark (D) was chosen to represent an intermediate step between winter and summer daylengths (Table 1 and Supplementary Figure 1).

After collection, and prior to each experimental treatment, 120–150 plants were kept for 5 days in 20 L tanks at temperatures measured in situ at the time of collection (April: 11°C; July: 17°C; November: 11°C; January: 6°C), at a salinity of ∼35 PSU, and at an intermediate experimental irradiance of 120 μmol photons m^–2 s^–1 in a L: D cycle of 12 h: 12 h (Beca-Carretero et al., 2018a; Figure 2; Table 1).

Before starting the experiment, shoots were progressively acclimated from in situ temperature to the target experimental temperature (6, 12, 16, 20, and 24°C) by gradually increasing or reducing chamber temperatures by 1°C every 24–48 h (Figure 2). Also, plants were progressively acclimated to the target irradiance applied (60 or 180 μmol photons m^–2 s^–1) by increasing or reducing ∼30 μmol photons m^–2 s^–1 every 5–7 days. The experiment was started when all treatments reached their target temperatures and irradiance to ensure that all plants were pre-acclimated to lab-conditions for the same number of days (10–14 days).

For each temperature and irradiance treatment, three individual plants were incubated in each of the three individual transparent cylinders (n = 3) consisting of cylindrical perspex bottles (ID 10 cm, height 35 cm) with a volume of ∼3.5 L. Shoots were loosely tied to a weighed-down plastic net at the bottom of each container to maintain them in a vertical orientation. Air was supplied through pumps ensuring constant mixing and CO₂ supply. 80% of the seawater in the cylinders was replaced every two to 3 days. Irradiance levels (180 and 60 μmol photons m^–2 s^–1) were measured inside the cylinder at the upper position of the seagrass leaves; the light was provided by Lumilux cool daylight fluorescent lamps (OSRAM L18W/865, Germany) (Beca-Carretero et al., 2018a). We continuously recorded water temperature using HOBO loggers (UA-002-64, Onset) installed in one cylinder per temperature and irradiance treatment. After 15–16 days, we measured a set of response parameters (see following).

Once the Z. marina shoots were acclimated to the target temperatures and before starting the experiment, all individual shoots were morphologically characterized. We measured: shoot fresh weight (g FW), total length (cm shoot^–1), number of leaves (shoot^–1) and the length of the rhizomes (cm). We standardized leaves and rhizome length across the all experiments to avoid potential impacts of morphology in plant responses by adjusting all shoots to two-three mature rhizome segments (2–3 cm total rhizome length) and 3–4 healthy leaves. At the end of each experimental treatment, the number of new leaves and new rhizome segments were counted, and rhizome and leaf elongation rates were assessed. To measure leaf growth (elongation) rates over time, two holes were pierced into leaves, one above the other with a distance of 2 mm above the basal meristem of the plant with a hypodermic needle (Sand-Jensen, 1975; Short and Duarte, 2001). We quantified the leaf formation rate (leaves shoot^–1 d^–1) by identifying the new number of leaves without punched holes, divided by the number of incubation days. Leaf elongation rate (cm shoot^–1 d^–1) was calculated as the length of new leaf material produced during the incubation time (in days), measured (i) from the base of the meristem to the punched holes, and (ii) the length of the newly produced leaves (Short and Duarte, 2001). The relative growth rate (RGR) was calculated according to:

where Bi is the initial weight and Bf, the final weight of the shoot and t is the incubation period in days. We haphazardly selected a group of 10 Z. marina shoots with similar size at the time of collection from the same seagrass meadow to assess the ratio of initial fresh weight (FW): dry weight (DW). Moreover, we evaluated the shoot mortality rate by the following equation:

Nf is the final shoot population (n = 9; 3 shoots per cylinder); Ni is the initial shoot population and t is the duration of the experiment (in days).

After 15–16 days of incubation, only healthy tissues (avoiding epiphytes or damaged parts) of the youngest and second youngest leaves were selected for biochemical analysis. Selected biomass was cleaned with distilled water prior to processing. Samples were frozen at −20°C and 48 h later freeze-dried. The samples were then kept at −20°C. 24 h before conducting the experiments samples were again freeze-dried to remove the potential humidity.

A thorough literature survey evaluated previous seagrass studies on biochemical components with a potential nutritional value in response to controlled temperature exposure. Any data from field experiments or in situ observations were excluded. For the literature review selected compounds include lipids, fatty acids, proteins, carbohydrates, pigments and carbon and nitrogen composition among others. We included reports on necrosis as a parameter that may affect nutritional value. Table 2 contains details of leaf compounds which varied more than ± 5% when seagrasses were grown under predicted warming conditions compared to in situ summer temperature. In addition, Supplementary Tables 13–14 (Excel file) include all results regardless of the degree of variation reported.

Data of morphological descriptors, growth rates, FA content and composition and photosynthetic pigments were Ln transformed, checked for homogeneity of variance using Bartlett’s test and for normality applying Kolmogorov–Smirnov test. As data did not meet the criteria, the non-parametric test PERMANOVA analyses based on a similarity matrix created from the Euclidean distances were implemented. We performed two PERMANOVA models. The first model aimed at investigating differences in seagrass descriptors responses to month (April, July, November 2015 and January 2016), temperature (6, 12, 16, 20, and 24°C) and irradiance levels (180 μmol photons m^–2 s^–1 vs. 60 μmol photons m^–2 s^–1). We used “temperature,” “irradiance,” and “month” as fixed factors. A pairwise test was applied to identify the treatments that differed significantly (p < 0.05). The second model was implemented to assess responses of seagrass traits to temperature and irradiance treatments each separate month. In this model “temperature” and “irradiance” were fixed factors. A pairwise test was applied to identify the treatments that differed significantly (p < 0.05). All data treatments and statistical analysis were performed using the PRIMERandPERMANOVA 6 statistical package. Growth rates, FAs and pigment concentrations are reported as means and standard deviation (SD).

Pearson correlation analysis was used to assess potential relationships between FA composition (e.g., PUFA, SFA) and experimental temperatures (6, 12, 16, 20, and 24°C) including both irradiance levels (180 μmol photons m^–2 s^–1 vs. 60 μmol photons m^–2 s^–1).

Temperature treatment significantly affected growth of Z. marina plants in terms of leaf elongation, leaf formation, rhizome elongation, internode formation and RGR (PERMANOVA, p < 0.01). Overall, most growth descriptors displayed bell-shaped patterns, with 16–20°C representing more favorable temperatures, and lowest rates reached at coldest (6°C) and warmest temperatures (24°C) at both irradiance levels (Figure 3 and Supplementary Tables 1–5).

Most biochemical descriptors in Z. marina leaves including fatty acid content and composition and photosynthetic pigments were significantly affected by temperature treatments (Figure 4 and Supplementary Table 1). The average total fatty acid (TFA) content of Z. marina leaves was 1.41 ± 0.32% of DW, however, no clear pattern in response to temperature increase was observed. The most abundant FAs were polyunsaturated fatty acids (PUFAs), which accounted for an average of 57.45 ± 6.6% of TFA across all seasons (April, July, November and January). The most abundant PUFA was α-linolenic acid (ALA, 18:3 n-3), followed by linoleic acid (LA, 18:2 n−6) and hexadecatrienoic acid (HTA, 16:3 n−3). The second most abundant group of FAs were saturated fatty acids (SFAs) which accounted for an average value of 35.1 ± 6.6% of TFA, where palmitic acid (16:0) was the most common SFA, followed by stearic acid (18:0). Finally, monounsaturated fatty acids (MUFAs) accounted for an average of 4.3 ± 1.5% of TFA (Supplementary Tables 6–9).

Common to all seasonal experiments, the observed reductions in PUFAs following high temperature exposure were mostly attributed to changes in n−3 PUFAs (18:3 n−3 and 16:3 n−3), and occurred at both high and low light; lowest percentages were detected at 24°C (Figure 5). For example, in plants incubated at higher irradiances, a significant reduction (Pearson correlation, p < 0.05, n = 30) of n−3 PUFAs from 6 to 24°C was observed with a decrease of 0.9% by 1°C increase. On the other hand, a significant increase in SFA and n−6 PUFA was observed with temperature (Figure 5).

Leaf and rhizome growth rates of plants in low light was significantly reduced in comparison to plants exposed to high light (PERMANOVA, p < 0.01) (Figure 3 and Supplementary Table 2). For instance, plants acclimated at light-limited conditions produced new leaves and rhizome segments 21.1 and 18.7% slower than plant acclimated to high irradiances (PERMANOVA, p < 0.05). At temperatures above the optimum range (16–20°C), light limitation induced a significant, abrupt decline in both leaf and rhizome growth compared to plants grown at high light. Additionally, at high temperatures, mortality rates were higher in plants exposed to low light than in plants incubated at high light; at low light, 30% of the incubated shoots died when grown at 24°C (Supplementary Table 12).

Reduction in irradiance at optimal temperatures for growth (16–20°C) generated an increase in FA unsaturation levels in leaves. The n−3/n−6 PUFA ratio in Z. marina exposed to low light significantly increased by 9.4% in comparison to plants exposed to high light (PERMANOVA, p < 0.05) (Figure 4 and Supplementary Table 2). Concentrations of photosynthetic pigments were also higher in low-light than in high light-acclimated plants (Figure 4 and Supplementary Table 10).

At all seasons, temperature increases of 4°C or more above in situ temperatures significantly enhanced growth of both rhizomes and leaf structures. For instance, in the experiment conducted in April, leaf growth at 16°C was 44.9% higher than in plants exposed to the in situ temperature of 12°C (PERMANOVA, p < 0.05, Figure 3). In summer, rhizome segments of plants incubated at 20°C grew 58% faster than those of plants incubated at in situ temperatures (16°C). Similar trends were observed in experiments conducted in January and November (PERMANOVA, p < 0.05, Supplementary Table 2).

Leaf and rhizome descriptors displayed different growth patterns depending on the seasonal acclimatization of the plants. Significantly higher leaf elongation and leaf formation rates were observed when the experiment was conducted in the coldest month, January (in situ temperatures of 5–6°C), compared to April or July (PERMANOVA, p < 0.05, Figure 3 and Supplementary Table 2). By contrast, significantly higher rhizome elongation and internode formation rates were observed in plants incubated in warmer months (April and July) than in plants collected in January and November. For instance, at optimal temperatures for growth (16–20°C), rhizome elongation of plants incubated in July (0.9 ± 0.01 cm d^–1) was 71 and 99% higher than in plants incubated in January and November, respectively. Noticeably, both leaf and rhizome growth descriptors were significantly lower in November in comparison with plants incubated during other months (PERMANOVA, p < 0.05). At 24°C, plants incubated in November and January displayed negative RGRs, but RGRs of plants incubated in April and July were positive. Mortality occurred in experiments conducted in April, November and January at 24°C but all plants survived in all experimental treatments in July (Supplementary Table 11), and was more accentuated in plants incubated at low light in January (HL: 0.039 ± 0.02 dead shoot d^–1; LL: 0.054 ± 0.02 dead shoot d^–1). Interestingly, 22% of the plants incubated in January at 12°C displayed the first stages of reproductive structures with the development of leaf sheaths corresponding to reproductive shoots (Figure 6). There was no evidence of reproductive structures in any other seasonal experiment.

TFA and photosynthetic pigment contents of Z. marina leaves differed significantly across seasons, however, there was no apparent clear pattern in response to temperature treatment (Figure 4) (PERMANOVA, p < 0.05). For instance, in plants incubated in April, the highest TFA contents were observed at lowest temperatures (6 and 12°C) with average values of 2.05 ± 0.1% of DW, and lowest contents at highest experimental temperatures (24°C, 1.3 ± 0.1% of DW) (PERMANOVA, p < 0.05). While in November, a contrary pattern with the largest TFA contents observed at 24°C (1.7 ± 0.2% of DW) and lowest TFA were detected at lowest temperatures (12°C, 1.3 ± 0.05% of DW) (Supplementary Table 10).

Nearly all seasonal results revealed significant interactive effects between temperature, irradiance and time on growth descriptors (Supplementary Table 2) (PERMANOVA, p < 0.05). Particularly, growth responses to temperature treatments were differently modulated by light, depending on season. For example, in January, as temperature increased, low light significantly reduced RGR in comparison to high light; on the contrary, plants incubated in April and July where not affected by light treatment (PERMANOVA, p < 0.05). Additionally, rhizome elongation of plants incubated in April did not display negative responses to warming (20–24°C) at either light treatment (Figure 3). However, a different pattern was observed in July, with low light significantly reducing rhizome elongation.

Results confirm our hypothesis that predicted warming episodes of ∼3–4°C above in situ current conditions (IPCC, 2014), including colder periods, in temperate regions may stimulate annual increases in growth of seagrass populations currently living under temperature-limited conditions (e.g., Krause-Jensen and Duarte, 2014).

In this study, regardless of light treatment, experimental optimum growth temperatures (between 16 and 20°C) for Z. marina from western Ireland were lower than previously reported for other centrally located populations of the species (Lee et al., 2007; review in Beca-Carretero et al., 2018a). Such different latitudinal responses may be explained by the local adaptations of seagrass thermal tolerance to in situ temperature conditions (Short and Neckles, 1999; Bennett et al., 2015). In the context of climate mitigation potential of seagrasses in temperate regions, the expected increases in productivity can be interpreted as a likely higher capacity to uptake and store carbon in the sediments, therefore reinforcing their role as a carbon sink (Macreadie et al., 2019). Negative effects of thermal stress were observed in most of the seagrass descriptors at 24°C. These results are in accordance with previous studies which correlated increases in respiration, growth inhibition and shoot death with anomalous high temperatures ranging from 20 to 30°C (e.g., review in Lee et al., 2007; Collier and Waycott, 2014; Gao et al., 2017).

The observed warming stimulation in seagrass productivity under lower irradiances can favor their capacity to colonize deeper areas (more than 4–6 m in Irish coasts), as warming enhances photosynthetic efficiency of temperature-limited seagrass populations (Krause-Jensen et al., 2012; Krause-Jensen and Duarte, 2014). In agreement with this, warm-adapted Atlantic, Pacific and Mediterranean Z. marina populations can colonize depths of 10–20 m (Morita et al., 2007; Ondiviela et al., 2014; Boutahar et al., 2020), while the vertical distribution of central and northern Z. marina populations is restricted to 4–10 m (e.g., Phillips, 1974; Krause-Jensen et al., 2011; Beca-Carretero et al., 2020). It is noteworthy that other factors such as greater light penetration can also explain the colonization of deeper areas within its southern distribution range such as in Mediterranean regions. In addition, both marine and freshwater macrophytes were reported to extend their vertical distribution to greater depths in response to increases in temperature (Duarte and Kalff, 1987; Rooney and Kalff, 2000; Jorda et al., 2020). As expected, plants growing at limiting irradiances had lower biomass production in both above and belowground tissues, probably explained by lower photosynthetic activity and carbon acquisition (e.g., Lee and Dunton, 1997; Ruiz and Romero, 2003).

Results confirm that plants living at reduced irradiances or are experiencing episodes of high turbidity in the water column may be more vulnerable to warming episodes (e.g., Longstaff and Dennison, 1999; Krause-Jensen et al., 2005). In our study, at higher temperatures, growth was reduced at low light compared to high light, and there were more visible signs of leaf necrosis alongside higher mortality rates in most seasonal experiments. A recent study reported large-scale declines of Z. marina populations after a combined effect of light reduction due to precipitation events alongside anomalous warming episodes (Johnson et al., 2021). The authors also found a rapid demographic recovery, mostly explained by the high abundance of the seedling. Studies in the Mediterranean Sea show that extreme summer temperatures provoked high mortality in deep-adapted (17–25 m) populations of Posidonia oceanica which were not able to recover in subsequent years, whereas shallow populations were less affected and recovered fully (Marbà and Duarte, 2010; Aoki et al., 2020). Additionally, projected sea-level rise will further reduce the irradiance available to Z. noltei plants distributed at their deeper range (∼2–2.8 m), resulting in more extreme conditions and physiological stress (Ondiviela et al., 2020).

Interestingly, proportions of n−3 PUFA relative to n−6 PUFA in plants receiving less light increased significantly, thus increasing the level of unsaturation of FA in leaves. Similar results were previously observed in the tropical seagrass H. stipulacea (Beca-Carretero et al., 2019a) and also in other marine primary producers (Guihéneuf and Stengel, 2013). This regulation of FA unsaturation levels in photosynthetic structures under light limitation may represent a photoadaptative mechanism enabling optimal photosynthetic and physiological performance (Gombos et al., 1994; Sanina et al., 2008; Solovchenko et al., 2008). Regarding pigmentation, plants exposed to low light accumulated higher contents of chlorophyll than those exposed to higher irradiances. This is a well-described mechanism of seagrasses to favor photosynthetic efficiency in low-light environments (e.g., Olesen et al., 2002; York et al., 2013).

Experimental growth rates measured across seasons at respective in situ temperatures (6°C in January, 12°C in April and November and 16°C in August) were highly similar to those measured in the field in shallow seagrass meadows in western Ireland (53.3272°N, 9.6168°W) (Supplementary Table 12). These results confirm the robustness of our experimental conditions and resultant response measurements but also highlight the importance of capturing seasonal responses of temperate seagrasses to in situ conditions.

Unexpectedly, optimal conditions for growth (16–20°C) were consistent across all seasons but critically, seasonality clearly modulated growth patterns of above and belowground compartments. On the one hand, higher rates of rhizome elongation and segment formation were observed in spring and summer when, in situ, plants were exposed to optimal environmental conditions for photosynthesis. This finding may be related to the seasonal capacity of temperate seagrasses to accumulate and store excess of energy produced from photosynthesis as carbohydrates in rhizomes in more suitable environmental conditions (e.g., Hemminga, 1998; Alcoverro et al., 2001). Such high-energy compounds can later support flower and seed production or act as an internal carbon source in less favorable environmental conditions in winter (e.g., Govers et al., 2015). On the other hand, the significantly higher leaf production observed in plants incubated in January-February might represent a mechanism to adjust photosynthetic structures to more suitable environmental conditions in spring. In contrast to this, plants incubated in November displayed the lowest leaf and rhizome growth rates which may represent a physiological acclimation to less suitable growth conditions in winter (e.g., Orth and Moore, 1986; Olesen and Sand-Jensen, 1993). There was evidence that seasonality modulated thermal and irradiance responses. For example, the growth performance of plants collected in April (in situ temperature at 12°C) was less negatively impacted by extreme warming (24°C) than plants collected in summer (in situ temperature 16°C), particularly under low light conditions. Two further aspects suggest that seasonal acclimation modulates the observed experimental thermal responses: (i) highest mortality was observed in January (winter) while in July (summer) all plants survived; (ii) there were contrasting growth responses observed in plants incubated in April and November: in situ environmental conditions at these times of the year are very similar, however, growth rates in autumn were nearly half of those observed during experiments conducted in spring.

The flowering of Z. marina shoots observed in January at 12°C suggests that warming in cold seasons may advance flowering events in temperate seagrasses. Recent field studies reported that flowering in Irish Z. marina populations in situ currently starts in early March-April (11–12°C) and ends in October–November (Beca-Carretero et al., 2019b). Previous studies also reported that anomalous warming can alter reproductive processes by increasing seed production in seagrasses (Ruiz et al., 2018; Qin et al., 2020a). Overall, year-round effects of climate change can lead to an internal decoupling of physiological and biochemical adjustments in marine primary producers (Bokhorst et al., 2009). For example, the accumulation of energy reserves such as non-structural carbohydrates may be affected, limiting their accumulation and impairing plant survival in cold winter periods (Alcoverro et al., 2001). Similarly, changes in temperature can affect flower production, flower fertilization and consequently seed production and germination (Qin et al., 2020a,b).

Detailed experimental studies that have investigated environmental stress on plants across different seasons are scarce and are mostly limited to terrestrial systems. For example, De Boeck et al. (2011) reported different sensitivities and responses to extreme events in terrestrial plants, dependent on the seasonal timing of the experiment. Our findings clearly demonstrate that, for temperate seagrasses, the time of year when temperature exposure occurs contributes to the observed growth, biochemical and survival responses. These results have relevant implications for future experimental design and interpretation of climate change effects on seagrasses. Nevertheless, caution is needed when extrapolating laboratory-derived results to natural habitats due to the complexity and the synergic interaction of several environmental parameters co-occurring in natural environments (De Boeck et al., 2010). Clearly, additional investigations are necessary to capture seasonal responses more comprehensively.

Particular nutritionally valuable compounds in seagrasses comprise essential fatty acids, carbohydrates, proteins or pigments, alongside secondary metabolites like phenolic compounds (e.g., Wang and Zheng, 2001; Pradheeba et al., 2011; Ahmed and Stepp, 2016).

The detailed review of the literature, together with experimental evidence from the present study, reveal sometimes contradicting trends pointing toward differential responses of specific biochemical compounds to increasing temperature. For instance, in some studies (e.g., Marín-Guirao et al., 2018; Artika et al., 2020), carbohydrates and photosynthetic pigments significantly increased in seagrass leaves at elevated temperatures; however, other studies report a significant decrease (e.g., Egea et al., 2018). Similar contrasting observations are documented for proteins, carbon, nitrogen and phosphorous contents (e.g., Hernán et al., 2017; Nguyen et al., 2020a; Viana et al., 2020; Table 2 and Supplementary Tables 14, 15). These differential patterns are likely to be linked to individual species-specific thermal tolerances, different degrees and duration of warming experiments or experimental time of year; they highlight an important gap in research into the effect of warming on the nutritional value of seagrasses. Common to all studies was the increase in necrotic leaf area under increased temperatures.

Our experimental results clearly demonstrate that warming significantly reduces omega-3 (n−3) PUFA levels, and favors the accumulation of SFA (see also Beca-Carretero et al., 2018b, 2020), as previously reported for other marine primary producers such as micro- and macroalgae (e.g., Gladyshev et al., 2013; Aussant et al., 2018). PUFAs play a key role in regulating the membrane fluidity, electron transport and photosynthetic activity in the thylakoids of the chloroplast; hence, the reported lower levels of FA saturation in photosynthetic structures with increases in temperature could be expected (e.g., Sanina et al., 2008; Gosch et al., 2015; Sijil et al., 2019). Changes in FA levels and composition have previously been implemented as a sensitive indicator of responses of seagrasses, saltmarshes and algae to different environmental settings including temperature, irradiance or salinity (Duarte et al., 2018; Kostetsky et al., 2018; Beca-Carretero et al., 2020).

Climate change is expected to increase metabolic demands of seagrass herbivores (Burnell et al., 2013) and thus affect their nutritional preferences (de los Santos et al., 2012; Hernán et al., 2016; Pagès et al., 2018). Recent studies have demonstrated that changes in the biochemical and nutritional composition of terrestrial and marine primary producers induced by climate change effects are impacting trophic interactions between plants and consumers and their feeding behavior (O’Connor, 2009; Prado et al., 2010; Hoover et al., 2012; Hernán et al., 2016). Herbivore feeding experiments should be undertaken to evaluate the potential effects of diets with reduced n−3 PUFA contents, simulating future warming conditions.

Under global warming scenarios marine heatwaves are predicted to increase in duration, intensity and frequency but also can occur across different seasons including in colder periods (Belkin, 2009; Frölicher et al., 2018). Our experimental results are thus ecologically highly relevant as they indicate that warming will not only affect seagrass productivity and physiology but also their reproductive effort. Conventionally, climate experiments and particularly warming assessment (review in Beca-Carretero et al., 2018b), have been conducted in warmer months, however, our study highlighted the importance of conducting climate change experiments across all seasons. In addition, our outcomes showed the higher vulnerability of plants exposed to low levels of irradiance under warming episodes. This is particularly important in the context of expected increases and variability of extreme events such as storms and extreme warming events associated with global change, resulting in synergetic stresses in shallow marine habitats (Lefcheck et al., 2017). Nevertheless, our experiments only represent short-time simulations such as marine heatwaves, and complementary experiments simulating long-term effects of warming and light stress are necessary to complement such an approach. Additionally, the magnitude of the selected range of temperatures, irradiance treatments and the duration of the exposure would significantly affect seagrass responses to experimental warming (Nguyen et al., 2021). Overall, our results have important ecological implications for the adequate management of seagrass ecosystems to ensure a balance of optimal temperature and irradiance conditions; they also point toward the critical importance to develop all-year monitoring programs in order to capture more comprehensively the potential effects of climate change events across all seasons.

Lastly, our results demonstrated that warming will dismiss the capacity of temperate seagrasses to produce and accumulate the essential omega-3 PUFA. These observations can have critical biological consequences as omega-3 are essential FAs that cannot be synthetized by herbivores and are transferred throughout the trophic chain (Sargent et al., 1999; Tocher, 2003; Parrish, 2009). Besides the direct ecological impact on herbivores, a reduction in omega-3 PUFA in seagrasses in response to climate change may also reduce the nutritional quality of some marine food sources (fishes or crustaceans) for human consumption (e.g., Kang, 2011; Hixson and Arts, 2016).