Effect of spatio-temporal shifts in salinity combined with other environmental variables on the ecological processes provided by Zostera noltei meadows

The present study aims to assess the plastic response of Zostera noltei meadows traits under spatio-temporal shifts in salinity combined with sediment environmental variables (temperature; pH; loss-on-ignition (LOI); carbon (C) and nitrogen (N) pools (top 5 cm)). Z. noltei biomass, C and N pools, leaf photosynthetic performance and esterified fatty acid (FA) profile were assessed within a temperate coastal lagoon during winter and late spring, along sites spatially distributed. None of the surveyed traits for Z. noltei displayed a clear spatial trend. Z. noltei proved to be euryhaline, whose biology was only slightly affected within this salinity range, in each season (14–39 in winter; 33–41 in late spring). Seasonal differences in salinity and environmental parameters explain the differences recorded in Z. noltei traits (aboveground biomass, N and C pools; photosynthetic performance). Spatio-temporal salinity shifts did not significantly affect the pool of FA present in Z. noltei. Overall, within the salinity range surveyed, the ecological processes studied and regulating Z. noltei meadows do not appear to be at risk. This work reinforces the plasticity of Z. noltei to salinity shifts within the studied range, with this finding being particularly relevant in the context of extreme weather events (e.g., winter freshwater floods, summer droughts).

a wide salinity range 18 and is commonly classified as euryhaline 19 . Z. noltei has the ability to morphologically and physiologically acclimate to different environmental conditions including fluctuations in physical (e.g., wave exposure, sediment resuspension, turbidity, and sediment particle size) and chemical parameters (salinity changes, nutrient and light availability) [20][21][22] .
Despite these features, shifts in salinity have been shown to impact and structure seagrass communities. Hypo-or hypersaline conditions can affect several seagrass species by triggering shifts in physiological processes, such as photosynthesis and respiration, osmoregulatory processes, carbon metabolism, and nutrient uptake (see ref. 23 for a review; ref. 24). Moreover, fatty acid (FA) synthesis in seagrasses might be affected by salinity shifts, as recorded for seaweeds 25 . Other environmental parameters such as temperature 26,27 and light 28 , are also known to influence FA synthesis. FA signatures reflect physiological changes in seagrasses and may reveal stress conditions due to diverse environmental settings. Namely, linoleic (18:2n-6) and alpha-linolenic acid (18:3n-3), recognized as seagrass (and vascular plants) biomarkers 29,30 , can reflect environmental variations by changing their unsaturation level and relative content.
Therefore, shifts in salinity may affect seagrass biology and physiology. In turn, this will impact the ecological processes and functions supporting the ES provided by these meadows. Particularly, primary production (through photosynthesis), nutrient cycling (through nitrogen (N) and phosphorous (P) uptake) and C cycle regulation (through C sequestration and storage), as well as FA synthesis, are some of the physiological processes and functions likely to be affected by salinity shifts. Other studies have shown that seedling germination of Z. noltei increases at low salinity values 31 and that shoots are more tolerant to lower salinities than to hypersalinity 22 . Seawater salinity (35) has also been shown to trigger mortality in Z. noltei populations 18 . With the expected increase in the frequency of extreme weather events promoted by global climate change 32,33 , salinity in winter can drop to zero due to heavy freshwater floods, while dry hot summers may drive salinity to reach values exceeding those of seawater (e.g. in mesotidal system during low tide).
The main objective of the present study was to assess to what extent spatio-temporal variation in salinity, combined with sediment temperature, pH, LOI (loss-on-ignition) and the sediment pool of C and N on the top 5 cm, affect ecological processes supporting ES provided by Z. noltei meadows. To address this objective authors formulated the following null (H 0 ) hypotheses: H 0 1 -high salinity does not affect photosynthetic performance, C metabolism and nutrient uptake; H 0 2 -ecological functions and ES provided by Z. noltei meadows, namely biomass provisioning (primary production), nutrient cycling (e.g. storage) and climate regulation (through C fixation and storage) are not affected by salinity shifts and other environmental variables; and H 0 3 -FA unsaturation level is not affected under salt stress and temperature decrease.
To address this topic we measured the spatio-temporal shifts in: above-and belowground biomass, N and C pools, photosynthetic performance and FA composition in Z. noltei. Sampling was performed following the Z. noltei distribution along the Mira channel, part of a shallow coastal lagoon (Ria de Aveiro, Portugal). Z. noltei traits were used as a proxy of the status of ecological processes supporting regulation and maintenance ES provided by these meadows. Given the wide geographical distribution of this seagrass species, these results will be applicable to other systems with similar environmental ranges and consequently contribute to a better understanding of Z. noltei meadows' dynamics.

Materials and Methods
Study site. Ria de Aveiro, a LTER site (Long Term Ecosystem Research; http://www.lter-europe.net/), is a temperate shallow coastal lagoon located on the western coast of Portugal (40°38°N, 8°44°W). It has a complex geometry, forming four main channels with several branches, islands, inner basins and mudflats. In the south, the lagoon forms an elongated channel (about 20 km long), the Mira channel, which is described as having the characteristics of an estuary itself 34 . This shallow channel is characterized by a salinity gradient during high tide 35,36 ; where salinity stratification is not common 34 , unless there is a high freshwater input either from water runoff or heavy rainfall. However, it can be described as a poikilohaline water body 37 , since water salinity shows considerable spatio-temporal shifts of tidal and seasonal origin 34 . Sampling sites were located along the Mira channel between the two-homoiohaline environments 37 , i.e., from downstream, close to the source of marine water; to upstream towards the freshwater source, the Mira river, following the seagrass distribution (Fig. 1).
Sampling strategy. Z. noltei was sampled along 5.5 km of the Mira channel, following the seagrass distribution ( Fig. 1), in winter (February) and late spring (June) 2013. A total of 80 seagrass samples were collected across ten sites in spring tides during low tide, using Ø8 cm corers (50 mm depth) (four replicates at each site); samples were subsequently taken to the laboratory. Sediment pH and temperature were measured in situ at each sampling site and date, using a WTW-pH 330i/set equipped with SenTix ® 41 (WTW, Weilheim, Germany). Water salinity was also measured in situ at each site along the channel in low water pools using a WTW Conductivity meter 330i/set equipped with Tetracon ® 325 probe (WTW, Weilheim, Germany), using the Practical Salinity Scale.
At the laboratory, Z. noltei roots and rhizomes were separated from rhizosediment. Aboveground (shoots/ leaves) and belowground (roots and rhizomes) parts of the seagrass were separated, rinsed with distilled water and weighed. A small portion was freeze-dried and stored at −80 °C for FA analysis and the remainder was dried at 60 °C until constant weight. Rhizosediment was air dried, ground, and passed through a 0.25 mm mesh. Plant material was also ground and homogenised for subsequent analyses. Rhizosediment was characterized for particle size, interstitial water salinity, dry bulk density, organic matter (OM) content through loss-on-ignition (LOI%), and total C and N content. Sediment particle size (granulometry) was assessed once, as seasonal variations are not expected, by sequential sieving of the 0-50 mm depth sediment cores and classified according to Blott and Pye 38 . Interstitial water was extracted from the sediment using centrifugation and salinity was measured using a refractometer. Dry bulk density was assessed with the equation: W d being the dry weight of the sample (g) and V t the total volume of the sample (cm 3 ) (Vt = volume of the particles + volume of the water). Sediment LOI, as a proxy for OM was quantified after 8 h combustion at 500 °C, in prior dried sediment (105 °C, until constant weight).
Total carbon (C) and nitrogen (N) content in rhizosediment and Z. noltei (aboveground and belowground biomass) were quantified in a CHNS/O analyser (Fisons Instruments Model EA 1108, Beverly, Massachusetts, USA). Z. noltei N and C pools (mg m −2 ) were estimated by multiplying biomass (mg DW m −2 ) per N or C content (mg mg −1 DW).
Photosynthetic performance. Once in the laboratory, Z. noltei samples were immersed in water, collected from each sampling site and maintained overnight at 18-20 °C under a dark:light cycle (12:12 h, L:D). The following day, photosynthetic performance of Z. noltei leaves was assessed by rapid light-response curves (RLCs) of photosystem II (PSII) relative electron transport rate (rETR) using a PAM fluorometer. This included a computer-operated PAM-Control Unit (Walz, Effeltrich, Germany) and a WATER-EDF-Universal emitter-detector unit (Gademann Instruments GmbH, Wurzburg, Germany). The fluorometer light source (measuring, actinic and saturating light) is a modulated blue light (LED-lamp peaking at 450 nm, half-bandwidth of 20 nm) 39 . The light delivered by the fluorometer and the fluorescence emitted by the leaf were conducted by a Ø1.5 mm plastic optical fibre. The optical fibre was maintained at a constant distance (1 mm) from the leaf. For each site, five Z. noltei shoots were selected and light adapted (70 μmol photons m −2 s −1 ) for a minimum of 15 min. RLCs were generated by subjecting the second youngest (innermost) leaf of each shoot (5 cm above the sheath (meristem) to 44 μmol photons m −2 s −1 for 2 min, followed by a sequence of increasing actinic light intensities (100, 134, 181, 261, 371, 515, 705 and 1149 μmol photons m −2 s −1 ). Each actinic light intensity was maintained for 10 s, after which a saturating light pulse was applied and fluorescence levels F s and ′ F m were quantified. The effective quantum yield of PSII 40 was calculated from:  fitted to rETR vs E curves, estimating the maximum photosynthetic efficiency (α; the initial slope of the curve), maximum relative electron transport rate (rETR m ), photoacclimation index (E k ) and optimum irradiance (E opt ).
Esterified fatty acid analyses. Esterified fatty acids (FA) were extracted from Z. noltei above-and belowground biomass, from sites 1, 2, 5, 7 and 10. The chosen sampling sites are geographically widespread along the sampling area and include the most upstream and the most downstream sites.
FA were extracted and converted into fatty acid methyl esters (FAMEs) through transesterification, following the method described by Aued-Pimentel et al. 43 . Freeze-dried Z. noltei samples (100 mg) were dispersed in 1 mL of n-hexane, to extract the FA, containing 0.375 g L −1 of heneicosanoic acid methyl ester (Sigma-Aldrich, Steinheim, Germany), as internal standard. A methanolic solution of 2 M KOH (0.2 mL) and saturated NaCl solution (2 mL) were added to the mixture. After intense vortexing and centrifugation for 5 min (2000 rpm), the supernatant was collected and dried in centrifugal evaporator (SpeedVac). Then, FAMEs were dissolved in n-hexane and 2 μL of this solution were analysed by gas chromatography (GC) equipped with flame ionization detector (FID) on a Perkin Elmer Clarus 400 equipment. The GC oven was programmed from an initial temperature of 50 °C, standing at this temperature for 3 min and following a linear increase to 180 °C at 25 °C min −1 , a linear increase at 40 °C min −1 to 260 °C and maintain this temperature for 3 min. Hydrogen was the carrier gas at a flow rate of 1.7 mL min −1 . The column used was DB1 with 30 m length, internal diameter of 0.25 mm and 0.10 μm film thickness (J&W Scientific, Folsom, CA). Data acquisition and analysis were carried out with TotalChrom Navigator Software. Supelco ® 37 component FAME mix (Sigma-Aldrich, Steinheim, Germany) was used as standard for the identification by retention time and quantification of FAMEs by obtaining calibration curves in relation to the internal standard.
Some Z. noltei samples were run through a gas chromatography column coupled with a mass spectrometer (GC-MS) analyser for confirmation of FAMEs identification. An Agilent Technologies 6890 N Network (Santa Clara. CA) equipped with a DB-1 column (30 m length, 0.25 mm internal diameter and 0.1 μm film thickness) was used. The samples were injected in the injection port at 250 °C lined with a 4.0 mm i.d. splitless glass liner. The detector starts to operate after 5 min of injection (solvent delay). The GC-MS was connected to an Agilent 5973 Network Mass Selective Detector operating with an electron impact mode at 70 eV and scanning the range m/z 40-500 in a 1 s cycle in a full scan mode acquisition. The oven temperature programme was the same used in the GC-FID equipment. Helium was used as carrier gas with a column head pressure of 12 psi. The injector ion source and the transfer line were kept at 230 °C. Each FAME peak was integrated using the equipment's software, and identified considering the retention time and mass spectrum of each FAME when compared to the Supelco ® 37 component FAME mix (Sigma-Aldrich, Steinheim, Germany). The FAMEs identification was confirmed by comparison with the chemical database Wiley and the spectral library "The AOCS Lipid Library".
The concentration of each FA was calculated based on the concentration of the internal standard (heneicosanoic acid methyl ester). Values are expressed as net FA concentration (μg g −1 or mg g −1 ) and as relative percentages of the total pool of FA. FA profile is shown as individual FA, and as three FA classes: SFA (saturated fatty acids), MUFA (monounsaturated fatty acids) and PUFA (polyunsaturated fatty acids -FA with ≥ 2 double bonds).

Statistical analyses.
Two square matrices were imported into R (R Core Team 2013) using the read.table() function and containing data on Z. noltei traits (above-and belowground biomass; above-and belowground C pool, above-and belowground N pool; and photosynthetic parameters) and measured environmental variables (salinity (interstitial water), sediment LOI, pH, temperature; sediment C pool; sediment N pool). The square matrix of species traits was first log e (x + 1) transformed (in order to normalize the distribution of the data) and three distance matrices constructed using the Euclidean index with the vegdist() function in the vegan package 44 in R. The distance matrices contained 1. all data, 2. data from February only and 3. data from June only. After controlling for normality with a Shapiro test and homogeneity of variance with a Bartlett test in R (most variables deviated significantly for both measures), we tested for significant differences in Z. noltei aboveground biomass (AbB), aboveground C pool (AbC), aboveground N pool (AbN), belowground biomass (BeB), belowground C pool (BeC), belowground N pool (BeN), maximum photosynthetic efficiency (α = Pal), maximum relative electron transport rate (rETR m = Pmx), photoacclimation index (E k = Pek), optimum irradiance (E opt = Pop), sediment C pool (SdC) and sediment N pool (SdN) between seasons using a Friedman test with the friedman.test function in R. We subsequently assessed the relationship between temperature, pH and salinity with AbB, AbC, AbN, BeB, BeC, BeN, Pal, Pmx, Pek, Pop. We used the lm function in R to test for linear relationships between these variables and temperature, pH and salinity with separate tests performed for each sampling event (winter (February) and late spring (June)). Variation among sample sites was assessed with Principal Coordinates Analysis (PCO) using the cmdscale() function in R with the Euclidean distance matrices as input. Weighted averages scores were computed for Z. noltei traits on the first two PCO axes using the wascores() function in the vegan package. Measured variables, namely sediment temperature, LOI, pH, sediment pools of C and N at the top 5 cm, and salinity were fit onto the PCO ordinations using the envfit() function in vegan. Using the envfit() function, we also tested for significant relationships between these variables and the PCO ordinations of Z. noltei traits using 999 permutations; all other arguments in the function were left as default. Detailed descriptions of the functions used here can be found in R (e.g., ?cmdscale) and online in the reference manuals (e.g., http://cran.r-project. org/web/packages/vegan/index.html; checked 2014 09 21). A t-test was performed for differences on C:N ratio between seasons (winter vs late spring) in aboveground and belowground Z. noltei, using the Statistics software package.
Z. noltei esterified fatty acid (FA) variation among different sites and seasons was tested through PERMDISP (tests the homogeneity of multivariate dispersions), which was performed after a log (x + 1) transformation to the FA net concentrations and after building a resemblance matrix considering Bray-Curtis similarity coefficients. PERMDISP was run on the basis of distances to centroids. A SIMPER (Similarity Percentage) analysis was also run once differences were obtained among sites, in order to identify which FA mostly explained dissimilarities among sites in Z. noltei FA profiles. All FA analyses were performed using PRIMER v.6.0 and PERMANOVA + add on (PRIMER-E Ltd., Plymouth, UK).

Results
Sediment characterization. During winter, a typical estuarine salinity gradient was observed in the sediment interstitial water along the Mira channel. However, this gradient was disturbed by spatial shifts, due to freshwater input at some sites (Table 1). In late spring, the salinity range was much lower than that recorded in the winter (ranging from 33.0 ± 1.4 to 40.5 ± 1.0 and 14.0 ± 1.0 to 39.0 ± 2.0, respectively). Reflecting the freshwater influence pH in the rhizosediment was lower in late spring than winter (Table 1). Sediment particle size (granulometry) ranged from sandy-mud [clay + silt (<63 μm) >50%; sand (>63 μm) <50%] to muddy-sand [sand >50%; clay + silt <50%] and did not show a clear spatial trend. Sites 2, 3 and 10 showed coarser sediments and sites 4 and 5 had a finer granulometry (Table 1). Total C and N rhizosediment pools in the top 5 cm were higher in winter when compared to late spring (Table 1).
The relative contribution of belowground material to the total biomass in winter ranged from 31 to 45% (average 37%) and from 23% to 47% (average 33%) in late spring. Most sites displayed more than 70% of aboveground biomass in late spring. There were no differences on C:N ratio in the aboveground biomass in winter (9.4 ± 0.5) and late spring (9.9 ± 0.9) (t-test, t = −1.636, p > 0.05). The C:N ratio of the belowground biomass increased from winter (18.2 ± 1.0) to late spring (23.5 ± 4.1) (t-test, t = −3.936, p < 0.001). Z. noltei C and N pools were calculated considering both the biomass and C or N content. Thus, C and N pools reflected the seasonal and spatial pattern of these parameters.
Zostera noltei traits versus spatio-temporal shifts in environmental parameters. The PCO ordination (Fig. 6A) was primarily related to variation between samples' above-and belowground biomass, C and N pools versus photosynthetic traits (49% variation explained, axis 1) (Fig. 6Ab). There was no obvious relationship between this gradient and any of the measured environmental variables (Fig. 6Aa). The second PCO axis in contrast showed a clear seasonal relationship with winter samples having higher sediment C and N pools and pH. In turn, late spring samples had higher temperature, LOI and salinity (33% variation explained, axis 2) (Fig. 6Aa). Z. noltei aboveground biomass, C and N pool and photosynthetic performance parameters (rETR m , E k and E opt ), were higher in late spring. On contrary, belowground biomass, C and N pool, and photosynthetic efficiency (α) were similarly related to both seasons (Fig. 6A). When considering solely the samples collected during winter (Fig. 6B), the first PCO axis (68% variation explained) was again primarily related to variation among above-and belowground biomass, C and N pools and photosynthetic parameters (Fig. 6Bb). This axis appeared to be related to sediment temperature (Fig. 6Ba) with warmer sites in winter having lower Z. noltei biomass, and lower C and N pools. The secondary axis (18% variation explained) was primarily related to variation between sites with high aboveground versus high belowground biomass and respective C and N pools (Fig. 6Bb). When considering only the samples taken in late spring (Fig. 6C), the first PCO axis (74% variation explained) was yet again primarily related to variation between aboveand belowground biomass, and the respective C and N pools versus photosynthetic parameters (Fig. 6Cb). The second axis (22% variation explained) was primarily related to variation of Z. noltei aboveground biomass, C and N pool versus belowground biomass, C and N pool, whose higher values are related to a higher LOI (Fig. 6Ca).
Esterified fatty acid profiles in Zostera noltei. Bearing in mind the main objective of the present work, the sixteen most abundant esterified fatty acids (FA) identified from Z. noltei samples were selected for further analysis. Overall, Z. noltei aboveground biomass showed a higher total FA concentration than belowground biomass, both in winter and late spring (Table S2, from supplementary information). Analysing the FA profiles (net concentration) in Z. noltei aboveground biomass, the multivariate dispersions of the FA were equal among sites (PERMDISP, F = 2.865, p > 0.05) and seasons (PERMDISP, F = 1.242, p > 0.05). On the contrary, significant differences were recorded for FA profiles on belowground biomass among sites (PERMDISP, F = 6.993, p < 0.01). Z. noltei from site 10 showed a FA composition statistically different from all the other sites (1, 2, 5 and 7). These differences are mainly explained by the FA 16:1n-7, 20:4n-6, 20:5n-3, 22:6n-3 and 24:1n-9 (SIMPER analysis; Table S3, from supplementary information). However, the mentioned FA altogether account for a very small proportion (1.0 ± 0.5% in late spring to 5.0 ± 1.1% in winter) of total FA in the belowground biomass. Since there were no differences in the FA profiles among these sites, relevant differences are not expected to occur among the other five sampling sites (3, 4, 6, 8 and 9). Therefore, the latter sites were not analysed for FA profiles.

Discussion
Z. noltei meadows were shown to be, but merely to some extent, affected by their habitat spatial dimensions, namely spatial shifts of salinity, sediment temperature and pH. In late spring, spatial shifts of salinity (ranging from 33 to 41), negatively affected Zostera photosynthetic performance, particularly r ETR m (related to the maximum photosynthetic capacity) and E k (photoacclimation index). However, also in late spring, other photosynthetic parameters (α and E k ) were not affected by spatial salinity differences. Consequently, it reduces the range and impact of negative salinity effects on the Z. noltei photosynthetic performance. In winter, even though a larger salinity range was recorded (from 14 to 39), Z. noltei photosynthetic performance was not affected by this spatial salinity shift. Thus, our hypothesis (H 0 1), stating that high salinity does not affect photosynthetic performance is only to some extent corroborated.
Less important was the positive relationship of sediment temperature in winter (ranging from 12.2 to 15.6) and Z. noltei biomass, aboveground C and N pool. The positive relationship of pH spatial shift (from 6.9 to 7.5) and aboveground biomass in late spring was also less relevant. Besides these effects, most of the Z. noltei traits were not significantly related to spatial shifts in these environmental parameters. Therefore, overall, spatial shifts in salinity, temperature or pH do not (totally) explain differences on Z. noltei ecological processes. Only photosynthetic performance and biomass can be partially explained, and not during both seasons surveyed (winter and late spring).
Seasonal changes in salinity and other environmental parameters (sediment pH, temperature, LOI and C and N pools) explain the observed seasonal differences in Z. noltei ecological processes and functions. Its aboveground biomass, C and N pools, and photosynthetic performance demonstrate this. Therefore, the following ecosystem services: 1) biomass provisioning (through primary production), 2) nutrient cycling (through N uptake and storage) and 3) climate regulation (through C fixation and storage) might change seasonally but are not disrupted (meaning that H 0 2 can be accepted).
Essentially, Z. noltei meadows have an important role in biomass provisioning through photosynthesis and in nutrient cycling, namely through the uptake of inorganic N and its incorporation in biomass 45 . The organic matter (including organic N) burial in sediment, which will later be mineralized, is also important in N cycling. In addition, seagrass meadows are known to be important C sinks 3 . Namely, Z. noltei belowground biomass plays  (Table S1).
an important role in carbon storage. Z. noltei C sequestration through incorporation of atmospheric CO 2 (photosynthesis) and its investment in biomass (C storage) 7 are important ecological process. In turn, these processes contribute to climate regulation (ES), by the reduction of greenhouse gas concentrations. Additionally, Z. noltei itself has a mass stabilizing effect (ES), being able to reduce sediment resuspension and erosion rates, minimize current velocity and increase the light availability in water column 4,5 .
Z. noltei aboveground and belowground biomass recorded here is within the range found for this species in other geographical locations ( Table 2). The above:belowground biomass ratio showed a similar trend (>1) to that previously described for Mondego estuary (Portugal) 46 , Ria Formosa (Portugal) 21,47 and for meadows at Palmones River estuary (Spain) 48 (Table 2). In contrast, this ratio was <1 at Ischia, Gulf of Naples (Italy) 49 . In late spring, Z. noltei greatly invested in the aboveground biomass (significantly higher than in winter), which is frequently explained by favourable environmental conditions (higher temperatures and light availability) 48 . In turn, investment in belowground biomass has been attributed to an acclimation strategy of Z. noltei to environmental stress conditions, namely high hydrodynamics alone 50 or combined with saturating light conditions 51 . In our study site (Mira channel, Ria Aveiro), Z. noltei does not appear to be under hydrodynamic stress, which is in line with water current velocities estimated for the study area (0.3 m s −1 at the channel head) 52 .
In the current work, the C:N ratio in belowground material of Z. noltei increased from winter to late spring as a result of a reduction in the N pool, which is in line with data from the Palmones River estuary (Spain) and Gulf of Naples (Italy) ( Table 2). The maximum N content in Z. noltei belowground biomass during winter is likely due to a storage strategy for N that may be used in the following months. Along the surveyed spatial-temporal shifts of Figure 5. Z. noltei photosynthetic performance versus temperature, pH and salinity in winter and late spring. Maximum photosynthetic efficiency (α) versus (a) temperature, (b) pH and (c) salinity; maximum relative electron transport rate (rETR m ) versus (d) temperature, (e) pH and (f) salinity; photoacclimation index (E k ) versus (g) temperature, (h) pH and (i) salinity; optimum irradiance (E opt ) versus (j) temperature, (k) pH and (l) salinity. Statistical significant linear relationships within each season are shown (p < 0.05). All statistical results are shown as Supplementary information (Table S1).
The N content of Z. noltei in the Mira channel in winter and spring is similar to that recorded at Ria Formosa (Portugal), respectively in aboveground biomass 21 and in belowground biomass 20 (Table 2). Moreover, it is similar to that found in the Mondego estuary (Portugal) in above-and belowground biomass 46 , but higher than the N content recorded by Peralta et al. 20 for aboveground biomass. Figueiredo da Silva et al. 54 previously showed a mean N content of 2% (%DW) in the Ria de Aveiro, whereas our study revealed a mean Z. noltei N content of 2.7% (mean of above-and belowground biomass N content in winter and late spring) ( Table 2). The C pool of Z. noltei aboveground biomass is in the same order of magnitude of previous studies 48 , but higher than values reported for both above-and belowground biomass by Peralta et al. 55 .
Z. noltei has been shown to be able to adapt its biomass partitioning (above:belowground ratio) and morphology to environmental pressure 56 . Indeed, its plasticity was demonstrated at different organizational levels (cell, individual and population) along an intertidal vertical gradient in the Ria Formosa, reinforcing its ability to acclimate to abrupt environmental gradients 21 . Figure 6. Principal Coordinates Analysis (PCO) showing the variation among sampling sites. Z. noltei traits included are: above-(AbB) and belowground biomass (BeB), above-(AbC) and belowground C pool (BeC), above-(AbN) and belowground N pool (BeN); and photosynthetic parameters (photosynthetic efficiency (α = Pal), maximum relative electron transport rate (rETR m = Pmx), photoacclimation index (E k = PEk) and optimum irradiance (E opt = Pop). Measured environmental variables are salinity (interstitial water) (Sal), sediment LOI, pH, temperature (Tem), sediment C pool (SdC) and sediment N pool (SdN). The distance matrices contained (A) all data, (B) data from winter only and (C) data from late spring only.  The photosynthetic performance of Z. noltei, particularly concerning light-saturated photosynthesis (rETR m ), is in line with values reported in the literature 57,58 , suggesting that the specimens surveyed were adapted to the physicochemical conditions along the channel. The photosynthetic performance of Z. noltei revealed a seasonal pattern in photoacclimation state, characterized by an overall optimization of light usage through change in both light-limited and light-saturated photosynthetic activity. Higher light absorption efficiency (α) was observed in winter, allowing Z. noltei to cope with the lower light availability. On the other hand, rETR m increased towards summer, denoting an increased efficiency of the carbon fixation metabolism, with Z. noltei taking advantage of higher light levels and temperature. However, as mentioned before, maximum relative electron transport rate (rETR m ) and photoacclimation index (E k ) were negatively affected by salinity increase in late spring. Of interest, this seasonal change occurred irrespectively of the surveyed spatial shifts of salinity. In winter ranged from 14 to 39; in late spring, ranged from 33 to 41; i.e., essentially the same pattern in all sampling sites. No trend was evidenced along the surveyed spatial shifts of salinity at Mira channel.
A previous experimental study by Fernández-Torquemada and Sánchez-Lizaso 22 , showed that Z. noltei (in Alicante, SE Spain) is more tolerant to lower salinities (e.g., 2) than to hypersalinity (shoot mortality significantly increased at salinities >47). Moreover, it has been shown in situ (Vaccarès lagoon, Southern France) that Z. noltei is not outcompeted by other species, even under long-term exposure to low salinities (e.g., exposed to a salinity of 5 for 3 years) 58 . In laboratorial experiments, Z. noltei germination rate and seedling survival was shown to be higher at low salinity (1) and decrease with increasing salinities (10 to 40) 59 . Overall, the tolerance shown by Z. noltei to spatio-temporal shifts of salinity (ranging between 14 and 41 in interstitial water) may play a key role in its resilience to extreme weather events. Indeed, these events are likely to affect the freshwater flows into this   Table 2. Zostera noltei biomass, carbon (C) and nitrogen (N) dynamics at different geographical locations (AGaboveground; BG -belowground; mean ± standard deviation; DW: dry weight).
coastal lagoon. Considering the cited works, Z. noltei is likely to survive and adapt to salinity changes as a result of extreme droughts or flood events. Thus, the studied ecological processes are not expected to be at risk. Absence of relevant differences on the Z. noltei individual FA profile along the Mira channel and between seasons (winter and late spring) corroborates the plasticity mentioned above for this seagrass, in face of the spatial salinity shifts and the seasonal environmental changes recorded. Only Z. noltei belowground biomass from site 10 (the most upstream location surveyed) showed a contrasting FA profile from those recorded in other sampling sites, which was explained by the FA 16:1n-7, 20:4n-6, 20:5n-3, 22:6n-3 and 24:1n-9. However, consistent with previous results for Z. noltei 30,60,61 , these FA relative content correspond to a very small proportion of total FA (≤5%, altogether). Moreover, palmitoleic (16:1n-7) and eicosapentaenoic acid (20:5n-3) are frequently used as diatom markers, while docosahexaenoic acid (22:6n-3) is a well-established marker for flagellates 62 . Therefore, these FA may likely origin from diatoms and flagellates neighbouring Z. noltei. It reveals that the difference recorded on the FA profile may not be related to seagrass physiology.
FA composition in seagrasses 63 and seaweeds 27 is known to depend on environmental parameters such as light 28 , temperature 26 , salinity 25 or nutrient availability 64 . Actually, considering that marine macrophytes are ectothermic organisms, membrane fluidity depends on the environmental temperature 65 and seasonal changes in FA composition can occur 26,63 . However, in the present work and supported by the results obtained for the analyses of individual FA profiles, no significant differences were recorded in the Z. noltei FA profiles, when they are grouped as SFA, MUFA and PUFA. Namely, sediment temperature increased from 14 ± 1 °C in winter (mean ± standard deviation, all sites) to 19.3 ± 2.9 °C in late spring, but it did not trigger a significant change on Z. noltei FA profile. Even though net PUFA in aboveground biomass was to some extent higher in winter (12.1 ± 3.6 mg g −1 , mean ± standard deviation, all sites) than in late spring (6.9 ± 1.6 mg g −1 ), evidencing a slight adaptation of Z. noltei to the environmental changes recorded, this difference was not statistically significant. In turn, acclimation of Z. marina to temperature through changes in FA composition was reported by Sanina et al. 63 in the Sea of Japan, when water temperature decreased from 20-23 °C (summer) to 3 °C (winter). Z. marina PUFA/SFA ratio increased from summer to winter due to an increase on n-3/n-6 PUFA's ratio. In tropical brown seaweeds, lower temperatures were also shown to promote a decrease in SFA and an increase in PUFA, so that membrane fluidity is maintained even at lower temperatures 27 . In addition, considering that salt stress in photosynthetic organisms can be perceived through FA desaturation 66 , the spatial and temporal salinity shifts (from 14 to 41) observed did not affect Z. noltei FA synthesis and is therefore easily endured by this seagrass. Overall, environmental changes recorded at Mira channel were not sufficient to trigger and significantly modify the FA composition in Z. noltei (therefore accepting our hypothesis H 0 3). Once again, the obtained results corroborate the resilience of this species to environmental fluctuations.
Major FA profiles (relative content) recorded in this work are in line with previous works performed for Z. noltei at different geographical regions such as Ria de Aveiro 60 and Ria Formosa 61 , in Portugal and Marennes-Oléron Bay, in the French Atlantic coast 30 . Also, the high PUFA relative content in Z. noltei biomass, higher than the sum of all SFA and MUFA present in the FA pool, is in accordance with the above mentioned studies performed on Z. noltei at Ria de Aveiro and Ria Formosa, as well as data recorded for Z. marina at Ria Formosa 61 and Sea of Japan 63 .
Besides the environmental variables addressed in the current work, other environmental variables and stressors can affect the Z. noltei ecological processes studied. For instance, light availability 55, 67 was shown to influence the overall Z. noltei growth, but also the plant development pattern, such as growth of lateral shoots from the main rhizomes 55 . In addition, under increasing light levels, the nitrogen content decreases, particularly in belowground material 55 . Nevertheless, the elevation of Z. noltei sampling sites along Mira channel is similar, meaning that daily hours of emersion (and immersion) for this intertidal seagrass is similar at all sites 68 . Therefore, the effect of light availability on Z. noltei ecological processes is likely to be the same at all sampling sites.
Overall, shifts in salinity and other environmental parameters recorded at Mira channel did not trigger changes in the Z. noltei FA profile, which corroborates the resilience potential of this seagrass to salinity shifts, which are likely to increase in the future.
Current climate change scenarios projected for this system predict an increase in flood area due to mean sea level rise, with Mira channel being likely to experience erosion that may lead to a shoreline retreat, a reduction in sand spit stability and even to a potential opening of a new inlet upstream (and consequent seawater intrusion) 69 . In addition to this, projected river discharges from the catchment to the lagoon, to the mid and end of this century, predict a reduction in average annual water discharges, with freshwater contributions to the lagoon decreasing, especially during summer 70 . Even though the predicted increase in flood area and the potential for seawater intrusion in Ria de Aveiro, the plasticity observed for Z. noltei -at different levels -to salinity shifts may anticipate its ability to cope with those changes, as well to expand to adjacent unvegetated areas in the Mira channel.

Conclusions
Z. noltei proved to be euryhaline, with its biology being only slightly affected within the salinity shifts surveyed in each season (i.e. between 14 and 39 in winter; and 33 and 41 in late spring). Seasonal differences in Z. noltei traits, such as aboveground biomass, aboveground N and C pools, as well as photosynthetic performance, are explained by the pronounced seasonal shifts in salinity and other environmental parameters (temperature, pH, LOI, sediment C and N pool). Fatty acid synthesis in Z. noltei was not affected by the spatio-temporal environmental fluctuations recorded. To sum up, our results show that, within the salinity range surveyed, the ecological processes addressed in our study supporting the ecosystem services provided by Z. noltei meadows do not appear to be at risk. Considering the wide geographical distribution of this seagrass (eastern Atlantic coastline) and the global predictions of increasing extreme weather events, these findings can anticipate the plastic response of the considered traits of Z. noltei to shifts in salinity and potential impacts on the ES provided.