Marine Macrophytes as Carbon Sinks: Comparison Between Seagrasses and the Non-native Alga Halimeda incrassata in the Western Mediterranean (Mallorca)

Seagrass species play a critical role in the mitigation of climate change by acting as valuable carbon sinks and storage sites. Another important ecosystem service of this coastal vegetation is nutrient removal. However, coastal ecosystems are under increasing pressure of global warming and associated establishment of invasive species. To elucidate the respective contributions of seagrass species Posidonia oceanica and Cymodocea nodosa and the non-native macroalga Halimeda incrassata as primary producers and nutrient sinks in coastal habitats we conducted in-situ incubations in the North-western Mediterranean Sea. Measured metabolic activity and nutrient removal as well as calcification rates in these habitats over a 24 h period in spring and summer confirmed that the endemic seagrass P. oceanica represents a valuable ecosystem with high O2 production and considerable carbon capture. The documented regression of P. oceanica meadows with higher temperatures and decline in autotrophy as measured here causes concern for the continuity of ecosystem services rendered by this habitat throughout the Mediterranean Sea with progressing climate warming. In contrast, the enhanced performance of C. nodosa and the calcifying alga H. incrassata with increasing temperatures, under expected rates of future warming is uncertain to mitigate loss of productivity in case of a potential shift in marine vegetation. This could ultimately lead to a decline in ecosystem services, decreased carbon storage and mitigation of climate change. Furthermore, this study provides a first estimate for the growth rate of H. incrassata in the Mediterranean Sea, supporting evidence for the mechanism of its rapid extension.

Seagrass species play a critical role in the mitigation of climate change by acting as valuable carbon sinks and storage sites. Another important ecosystem service of this coastal vegetation is nutrient removal. However, coastal ecosystems are under increasing pressure of global warming and associated establishment of invasive species. To elucidate the respective contributions of seagrass species Posidonia oceanica and Cymodocea nodosa and the non-native macroalga Halimeda incrassata as primary producers and nutrient sinks in coastal habitats we conducted in-situ incubations in the North-western Mediterranean Sea. Measured metabolic activity and nutrient removal as well as calcification rates in these habitats over a 24 h period in spring and summer confirmed that the endemic seagrass P. oceanica represents a valuable ecosystem with high O 2 production and considerable carbon capture. The documented regression of P. oceanica meadows with higher temperatures and decline in autotrophy as measured here causes concern for the continuity of ecosystem services rendered by this habitat throughout the Mediterranean Sea with progressing climate warming. In contrast, the enhanced performance of C. nodosa and the calcifying alga H. incrassata with increasing temperatures, under expected rates of future warming is uncertain to mitigate loss of productivity in case of a potential shift in marine vegetation. This could ultimately lead to a decline in ecosystem services, decreased carbon storage and mitigation of climate change. Furthermore, this study provides a first estimate for the growth rate of H. incrassata in the Mediterranean Sea, supporting evidence for the mechanism of its rapid extension.

INTRODUCTION
Ever since the 18th century, humanity has caused raised and still increasing emission of greenhouse gases (GHGs), mainly carbon dioxide (CO 2 ) with annual averages of 410 ppm in 2019 (IPCC, 2021) and predicted CO 2 levels four times higher compared to pre-industrial (Wigley, 1983) levels (IPCC, 2021). Global warming, including the rise of ocean temperature, is the most direct consequence of increasing CO 2 levels in the atmosphere. The ocean plays a relevant role mitigating the CO 2 rise in the atmosphere through the absorption of around a third of the anthropogenic emissions (Gruber et al., 2019). This CO 2 uptake impacts the chemical balance of seawater by increasing bicarbonate ion concentration (HCO 3 − ) and reducing the availability of carbonate ions (CO 2 3− ) by 60% (Feely et al., 2004). This shift in water chemistry results in a decrease of seawater pH, a process referred to as Ocean Acidification (OA; Doney et al., 2009). In fact, sea surface pH is expected to decrease from 0.3 to 0.5 units by the end of this century (Caldeira and Wickett, 2003). Ocean acidification is a global phenomenon, impacting calcification and metabolism of organisms, which has been documented in many marine regions (Kroeker et al., 2013).
The Mediterranean Sea is strongly affected by climate change and is among the marine regions with the highest rates of warming, from two to four times higher than in other ocean regions (Vargas-Yáñez et al., 2010;Burrows et al., 2011). Despite its high productivity, the semi-enclosed Mediterranean Sea is most likely to be more impacted by climate change than other seas (Richon et al., 2019), due to its physical properties with a short mixing and water exchange period of around 70 years (Durrieu de Madron et al., 2011). The Mediterranean Sea is therefore defined as a "hotspot for climate change" (Giorgi, 2006), with additional disturbances interacting in this region, both natural and anthropogenic (Lejeusne et al., 2010). Natural disturbances such as more extreme and frequent meteorological events (Parry, 2000;IPCC, 2013) and anthropogenic disturbances due to progressing urbanization, habitat destruction, pollution, overfishing, and species introduction induce additional pressure on the marine ecosystems (Lejeusne et al., 2010). The intrusion of invasive species into the Mediterranean has dramatically increased due to growing aquatic trade, aquaculture, maritime traffic (Raitsos et al., 2010) and most severe, the opening of the Suez Canal in 1869 and its latest widening in 2015 (Galil et al., 2015). More than 600 alien species have already established in the Mediterranean (Zenetos et al., 2017) and this number is expected to increase by 2050 (Sardain et al., 2019).
Although seagrasses occupy only a 0.1% of the ocean surface and their extension is limited to the coastal regions of temperate, sub-and tropical waters (Kennedy et al., 2010), their communities are among the most productive marine ecosystems (Duarte and Chiscano, 1999), modulate biogeochemical processes (Duarte, 1999), and provide valuable ecosystem functions (Costanza et al., 1997). Seagrass meadows form dense meadows providing food, habitat, and nursery grounds for various species (Duarte, 2002) and contribute to the physical protection of the coastal zone by providing wave attenuation and forming dense rhizomes below-ground stabilizing the sediment, thus preventing coastal erosion (Costanza et al., 1997;Fourqurean et al., 2012). Furthermore, they eliminate nutrients from the water column (Barry et al., 2013;Grall and Chauvaud, 2017), P. oceanica meadows are known to be net sinks for Nitrate and Phosphate (Barrón and Duarte, 2009) and contribute greatly to ecosystem resistance to eutrophication (Duarte, 1995;Lloret et al., 2008). Seagrasses have also demonstrated a high capacity of capturing CO 2 , accounting for 20% of the global marine carbon sequestration Duarte and Krause-Jensen, 2017) and are referred to as blue carbon habitats (Macreadie et al., 2019). Seagrass meadows are mainly autotrophic, their productivity exceeds their carbon needs and they can export 24.3% of their production to adjacent ecosystems (Duarte, 2002), or even up to 50-70% (Kennedy et al., 2010). By removing carbon, seagrass canopies also locally mitigate the effects of OA during the daytime, acting as a buffer increasing pH in their canopies (Hendriks et al., 2014).
However, seagrass meadows are susceptible to various threats to their habitat, and face most of the serious threats to marine ecosystems: physical modifications by human activity, decreasing water quality, pressure from invasive species and climate change (Waycott et al., 2009). The Western Mediterranean basin accounts for a total of 510,715 ha of Posidonia oceanica meadows (Telesca et al., 2015), the endemic seagrass species in the Mediterranean Sea, which represents the major coastal ecosystem of the Balearic Islands. Posidonia oceanica is distributed throughout the temperate Mediterranean Sea and is impacted by warming since elevated temperatures cause physiological stress and shoot mortality increases (Diaz-Almela et al., 2007;Marbà and Duarte, 2010). The slow-growing P. oceanica is prone to show a stronger negative response to higher temperatures, due to a low thermal tolerance with optimal temperatures for growth of 15.5-18 • C (Olsen et al., 2012), although average optimum temperatures for photosynthesis are typically higher (30-32 • C; Lee et al., 2007). More temperature tolerant species such as Cymodocea nodosa, a seagrass of tropical origin has a higher optimal temperature for growth (24.5 • C) than the temperate P. oceanica (Olsen et al., 2012) and appears to be better adapted to increasing water temperature (Lee et al., 2007;Olsen et al., 2012;Savva et al., 2018). The tropical green alga Halimeda incrassata was reported for the first time in 2011 in Palma Bay (Balearic Islands), introduced most likely by recreational boats (Alós et al., 2016). The alga found favoring conditions to expand rapidly, forming dense meadows over 41% of the monitored area in Palma Bay between 2011 and 2015 (Alós et al., 2016). Still, to the best of our knowledge, a specific growth rate has not been estimated. Benefitting from increasing water temperatures, H. incrassata is more likely physiologically affected by water with a lower pH and less calcium carbonate available for calcification, than increasing temperatures, although through the link between photosynthesis and energy compensation calcification might not be impacted directly (Hofmann and Bischof, 2014). Gradually increasing temperatures with more frequent heatwaves are expected to severely impact the persistence of temperate species with low thermal tolerance (Short and Neckles, 1999;Jordà et al., 2012). Changes in the composition of marine vegetation and structure of seagrass meadows and hence changes in the metabolic activity of these habitats directly affect biogeochemical processes of the coastal zone, impacting the carbon sequestration capacity and the oxygen dynamics, thereby impairing the functioning of these coastal ecosystems (Sand-Jensen and Borum, 1991;Duarte, 1995). These ecosystems are heavily impacted by global warming and increased presence of exotic species (Han and Liu, 2014) and loss of seagrass meadows increases the risk of erosion of historically deposited carbon, releasing it from the sediment and emitting it to the atmosphere (Marbà et al., 2015).
Here we evaluate how three Mediterranean macrophyte species contribute to biogeochemical processes in the coastal zone by evaluating their metabolic activity during the most productive period of the year. We conduct field incubations to estimate oxygen production, directly linked to carbon consumption, and nutrient use comparing seagrass species P. oceanica and C. nodosa and the non-indigenous calcifying alga H. incrassata. We chose to evaluate two distinct time points during growth season, one in spring and one in summer, with expected differences in productivity peaks between the three species and within their preferred temperature optima. The western Mediterranean Sea shows a strong seasonality in temperature and we chose April, when temperatures start to rise and H. incrassata was observed entering an active state while spring is known as the most productive time for P. oceanica, and July, when temperatures start to reach the annual maxima. Additionally, to estimate a growth rate for the tropical H. incrassata, that can serve as an indicator of the mechanism for its rapid expansion in the Mediterranean Sea it is crucial to include the actual growth period. Possible shifts in community composition favoring species more tolerant to warmer temperatures will change the impact of coastal marine vegetation in the Mediterranean Sea on the carbon cycle and nutrient uptake. Our comparison between three macrophyte species with contrasting future trajectories highlights the importance of these vital coastal ecosystems in climate change mitigation and can help to predict changes we might expect in the future.

Study Site
The study site at Punta Negra (39.5266 • N, 2.5515 • E), a small bay on the south coast of Mallorca (NW Mediterranean, see Supplementary Figure 1) with water depths ranging from 2 to 6 m has no direct inflow of sewage waters. The site represents the predominant coastal setting in the Mediterranean Sea and was chosen because of its accessibility and presence of all three study species. The site is dominated by P. oceanica meadows, with meadows of C. nodosa and H. incrassata present on sand and death matte of P. oceanica. For on-site monitoring, a temperature and light logger (HOBO Pendant UA-002-64; Onset, United States) was installed inside a H. incrassata patch on February 12th, 2019 at 2.8 m depth, recording data hourly. Light intensity (PAR,  Table 1). This sensor was left on-site from February to the end of the experiment. Additional HOBO sensors were installed attached to multiparametric sensors measuring watercolumn values in each habitat at max. 1 m distance from the benthic incubations (see "Field Incubations") for the duration of the incubations. The distance between different macrophyte habitats was between 5 and 10 m.

Incremental Growth Rate
The growth rate of H. incrassata was estimated with the Alizarin Red-S dyeing technique (Dustan, 1975;Wefer, 1980;Multer, 1988), adjusted for H. incrassata according to Payri (1988). Two aquaria (92 L Volume) were deployed upside down on two patches of H. incrassata (March 25th to 27th, 2019) and a concentrated solution of Alizarin Red-S was injected at sunrise during two consecutive days, for maximum incorporation during daylight. The aquaria were removed at sunset to ensure recirculating of the water column and prevent hypoxic conditions. This led to an effective staining period of 24 h.

Field Incubations
We targeted the most productive season for seagrass productivity (spring) and the subsequent summer for field incubations to facilitate a comparison between species as Halimeda incrassata is not active during colder periods. During April (11th-12th) and July (2nd-3rd) 2019, we conducted closed in-situ incubations in patches of P. oceanica, C. nodosa, and H. incrassata, each for 24 h. Simultaneously, Hydrolabs (HL4; OTT HydroMet) were installed in each habitat at max. 1 m distance from the incubations, to record temperature, pH, conductivity, depth, and oxygen (O 2 ) to get a record of oxygen dynamics in the water column (Table 1).
For additional calculations of metabolism based on open water oxygen profiles measured with the multiparametric sensors, we considered mixing depth and atmosphere-sea interactions according to Cole et al. (2000), using a modification of a program originally written by Coloso et al. (2008), implemented in MATLAB (v.R2010b, the Mathworks Inc.). For each habitat type, three replicate incubations were set up by SCUBA-divers. PVC-rings (Height: 0.2 m; Radius: 0.09 m) were hammered into the seafloor, allowing for intact patches of macrophytes in the center. Translucent incubation bags (coextruded high barrier multilayer film PA/EVOH/PA/PE), which did not permit gas exchange between the chamber and the water column, with a three-way valve (luer-lock) for sampling were fitted over the rings and fixed with elastic bands. Additionally, planktonic incubations to measure water column metabolic activity in dark and light periods were set up in opaque and translucent glass bottles. Initial samples were collected at sunset from the water column at the start of the incubations for total alkalinity (TA), O 2 , and nutrients. During subsequent dives at sunrise and again at sunset, samples for TA, O 2 and nutrients were taken from all individual incubation bags using 50-ml (polyethylene) acid-washed syringes. Samples taken at sunrise were compared with the first sunset samples to account for the night (period of 11 h in April, 9 h in July), samples at sunset account for the day (period of 13 and 15 h, respectively). O 2 samples were fixed immediately on site according to the protocols by Carpenter (1965) and Labasque (2004) and samples for nutrients were filtered over GF/F filters and stored frozen at −20 • C until further analysis. TA samples were routinely poisoned with mercury (II) chloride (HgCl 2 ).

Biomass Analysis
After each incubation period, macrophyte material from within the incubation rings was collected. The vegetation was rinsed, dried for 24 h at 60 • C to obtain the dry weight and then divided into above and below ground biomass.

Incremental Growth Rate
The site of H. incrassata dyed with Alizarin Red-S was not selected for the incubations but collected separately, 99 days after in-situ staining. Algae were rinsed and placed for 30 min into a 5% Sodium Hypochlorite (NaClO) solution for bleaching. The growth rate was calculated as the ratio of new segments after staining and the sum of all segments (Wefer, 1980). With the days between dyeing H. incrassata until harvesting the biomass, the number of segments produced and the percental increase in segment number per day was calculated.

Analyses of Water Chemistry
Oxygen samples were analyzed by the spectrophotometric Winkler method following recommendations by Labasque (2004) in a spectrometer (Spectronic Helios alpha; ThermoFisher Scientific, United States) at 466 nm and calibrated using a 5-point calibration with Potassium iodide (KIO 3 ) as coloring reagent.
Total Alkalinity was analyzed by open cell potentiometric titration with a Titrando 808 (Metrohm) following the Standard Operation Procedure (SOP) 3b (Dickson et al., 2007). We used certified CO 2 seawater reference material (CRM Batch #136) from Prof. A. Dickson at Scripps Institution of Oceanography (United States) to warrant the quality of the analysis, which had a precision of 4.4 µeq/Kg and an accuracy of <1 µeq/Kg.
With TA and in-situ measured values of salinity, temperature and pH, the CO 2 -concentration was calculated with the CO2SYS software (Pierrot et al., 2006) using pH values in NBS scale (mol/kg H 2 O), K1 and K2 constants from Mehrbach et al. (1973) refit by Dickson and Millero (1987) and the KSO 4 constant after Dickson (1990).

Metabolism
To evaluate the metabolism of the enclosed community, Net community production (NCP), community respiration (CR), and gross primary production (GPP) were calculated from changes in oxygen concentrations between dark and light periods. The O 2 − measurements of the planktonic incubations are used to correct the total measured metabolic activity in the benthic incubations for the metabolic activity in the water column. Also, metabolism in the vegetated incubations was corrected for average values from incubations containing only bare sediment, to account for respiration or production by the sediment itself. CR was calculated as the change in O 2 overnight and NCP was calculated as the overall change in O 2 over the incubation period of 24 h. Under the assumption that CR is equal during night and day, GPP was estimated as the difference between NCP and CR. The rates were then corrected for the hours of day light in situ, to obtain metabolism per time unit (h). Then, NCP, CR, and GPP were corrected for the area and volume of each incubation setup, as estimated by injections of 5-ml of a 0.25 mol/L phosphate solution in each bag after the incubations and calculating its dilution after spectrophotometric determination (Hansen and Koroleff, 1999). Finally, we normalized the results by dry weight of the macrophyte material to obtain a final concentration of O 2 in mmol/d/gr. DW.

Calcification Rate
We approached the net calcification rate (g) by applying the alkalinity anomaly technique to derive the net calcification of the community under the assumption that only CaCO 3 incorporation or dissolution influences TA (Smith and Key, 1975). The difference between pre-and post-incubation TA equals TA in µmol/kg, which then was converted into mmol/kg. The net calcification rate g (mmol CaCO 3 /unit time) was then calculated by using following formula: With ρSW (density of Seawater) as 1.03 kg/m 3 and T the time (h) between time periods. We calculated TA for two time periods: TA between initial water samples (sunset) and samples collected at sunrise and between sunrise and samples collected at sunset. After correction for the hours between sampling this permits to investigate whether calcification takes place mostly at night or during the daytime.

Nutrients
Nutrient utilization was evaluated by calculating the changes in nutrient concentration between time periods. Nutrient usage over night was calculated by subtracting the results from samples taken at sunrise from the initial water column samples. Nutrient usage during daylight was calculated by subtracting the nutrient concentration from samples taken at sunset from samples taken at sunrise. Nutrient concentrations in the vegetated incubations were corrected for average values from incubations containing only bare sediment, to account any retention or production by the sediment. A reduction of nutrients in the water column indicates a depletion by the macrophyte, whereas an increase in nutrient concentration within the benthic chamber indicates that the habitat released nutrients.

Statistical Analysis
Two-way ANOVA and Tukey HSD tests were used to distinguish significant differences between macrophyte species and respective incubation setting in spring or summer on the response variables (NCP, CR, GPP, and g). The statistical analyses were conducted in Rstudio (version 1.1.456; RS Team, 2015) using the R platform (R Core Team, 2021).

Field Incubations
There were no strong environmental differences between habitats ( The growth rate of H. incrassata was of 0.93 newly produced segments per day during 99 days after staining, which accounts for a percental increase in segment number per day of 0.9207. This growth rate does not account for an annual growth rate, but as a mechanism for the establishment and expansion of H. incrassata. The time period investigated was from early spring, when temperatures rose and H. incrassata was observed to enter its active state, to summer, when water temperatures approached annual temperature maxima in the Mediterranean Sea, therefore spanning across the crucial period for the success of colonization of H. incrassata.

Metabolism
Metabolic rates, CR, GPP, and NCP confirmed P. oceanica to be a highly productive habitat (Figure 1), with the highest O 2 turnover rates of all habitats. Both, CR [F (2,2) = 28.24] and GPP [F (2,2) = 35.45] are significantly higher than for C. nodosa and H. incrassata (p < 0.05), although NCP showed no significant difference between species (p = 0.445). No significant difference was observed between seasons (p = 0.210 for CR, p = 0.099 for GPP and p = 0.266 for NCP). However, P. oceanica showed increased respiration activity ( 5.322 mmol O 2 /d/m 2 ) and increased GPP ( 3.042 mmol O 2 /d/m 2 ) in summer and decreased NCP of the habitat ( 1.657 mmol O 2 /d/m 2 ). In contrast, C. nodosa and H. incrassata showed increased NCP in summer of 10.266 and 8.284 mmol O 2 /d/m 2 , respectively, due to an increase in GPP ( 9.311 and 12.815 mmol O 2 /d/m 2 ). Calcification rate for C. nodosa decreased by 0.332 mmol O 2 /d/m 2 , whereas for H. incrassata CR increased by 5.155 mmol O 2 /d/m 2 .
A normalization of the results per gram dry weight suggests that in fact C. nodosa has the highest O 2 turnover rates per biomass unit (Figure 2). Although no statistical significance between species nor season (p > 0.05) was found, P. oceanica showed less productivity in summer with a decrease in NCP of 0.012 mmol O 2 /d/gr. DW, whereas C. nodosa and H. incrassata

Multiparametric Sensors
The whole ecosystem metabolic rates calculated with oxygen data from three multiparametric dataloggers measuring in the water column (Supplementary Figure 2) did not allow for distinction of species-specific productivity even if they were situated in distinct habitats (p > 0.05). The net ecosystem metabolic rates were considerably higher than those obtained from enclosed habitats in the benthic chambers, compared to the adjacent sensor with CR [F (1,2) = 1117 in spring, and F (1,2) = 893.8 in summer, p < 0.001 in both seasons] and GPP [F (1,2) = 266.7, p < 0.01

Nutrients
Water column nutrient concentrations, as a measure for all three macrophyte habitat as the site is shallow and well-mixed, of nitrogen oxides [NOX; Nitrate (NO 3 − ) and Nitrite (NO 2 − )], ammonium (NH 4 + ) and phosphate (PO 4 3− ) were significantly higher (p < 0.05) in spring than in summer ( Table 1). The utilization of NOX, phosphate and ammonium within the three different habitats also differed between seasons (Figure 4). In spring (Figure 4A), P. oceanica depleted inorganic nutrients throughout the incubation period, but at higher rates during night (4.055 ± 0.13 µM/m 2 for NOX, 0.379 ± 0.41 µM/m 2 of phosphate and 2.127 ± 1.01 µM/m 2 for ammonium). Cymodocea nodosa showed a depletion of NOX (0.562 ± 0.26 µM/m 2 ) and ammonium (0.372 ± 0.27 µM/m 2 ) during night, but a release at higher rates during day (0.577 ± 0.28, 0.991 ± 0.64 µM/m 2 , respectively). Halimeda incrassata indicated the highest rates of releasing phosphate during the night (0.421 ± 0.28 µM/m 2 ) and NOX during the day (1.061 ± 0.37 µM/m 2 ), yet depleting ammonium at a considerable rate of 1.095 ± 0.95 µM/m 2 during the day. In summer ( Figure 4B), P. oceanica showed primarily a depletion of inorganic nutrients throughout the incubation period, yet at higher rates during night (1.593 µM/m 2 for NOX and 1.411 µM/m 2 for ammonium). Generally, rates for nutrient usage in P. oceanica were lower than in spring, whereas rates for C. nodosa were higher in summer. Cymodocea nodosa, in contrast to spring, showed primarily a depletion of nutrients (0.535 ± 0.48 µM/m 2 for NOX and 2.849 ± 1.92 µM/m 2 for ammonium) during day and a release during night (0.389 ± 0.46 µM/m 2 , 2.119 ± 1.81 µM/m 2 , respectively). Rates of nutrient release in H. incrassata were also lower in summer than in spring.
Dissolved organic nitrogen (DON) in the surrounding water was significantly lower (p < 0.05) in spring than in summer, whereas dissolved organic phosphorus (DOP) was lower (p < 0.05) in summer ( Table 1). Highest rates of dissolved organic nutrient usage were observed for P. oceanica (Figure 5), yet differences between the different macrophytes are apparent. In spring (Figure 5A), both seagrass species showed a depletion of DON during day (11.645 ± 17.14 µM/m 2 for P. oceanica and 10.413 ± 8.56 µM/m 2 for C. nodosa). At night, P. oceanica showed a release of 13.827 ± 9.67 µM/m 2 , whereas C. nodosa still indicated a depletion of DON (13.369 ± 7.59 µM/m 2 ). The green alga H. incrassata also showed a depletion of DON during night (2.699 ± 3.85 µM/m 2 ), but a release (13.522 ± 3.44 µM/m 2 ) during the day. DOP usage was similar for all three macrophyte, showing primarily a depletion during night but a release during day, with rates higher for both seagrasses (1.339 ± 0.25 µM/m 2 for P. oceanica and 1.428 ± 0.24 µM/m 2 for C. nodosa) than for H. incrassata (0.872 ± 0.89 µM/m 2 ). In summer (Figure 5B), all three habitats released DON during night (at rates of 18.512 µM/m 2 for P. oceanica, 6.469 ± 3.85 µM/m 2 for C. nodosa and 8.211 ± 8.59 µM/m 2 for H. incrassata) and showed a tendency toward depleting DON during the day (P. oceanica with the highest rate of 3.937 µM/m 2 ). The release of DOP during day was not observed in summer, except for C. nodosa (1.609 ± 2.69 µM/m 2 ).

DISCUSSION
Marine macrophytes play a foundational role in coastal ecosystem productivity, though not all seagrass and macroalgae taxa have the same potential for eliminating CO 2 from the water column. The habitat of P. oceanica is a highly productive coastal ecosystem with a high potential for eliminating CO 2 from the water column. Posidonia oceanica's large biomass and the formation of vast and dense meadows make it an invaluable carbon sink with a high capacity of carbon storage beneath its  meadows. Although C. nodosa appears to have a more efficient metabolism when normalized by biomass, the superior biomass and physiology of P. oceanica provides a larger photosynthetic active surface than canopies of C. nodosa. However, the insitu biomass of the macrophytes was rather low and values for metabolic response rates [>160 mmol O 2 /d/m 2 for GPP and >125 mmol O 2 /d/m 2 CR in P. oceanica meadows Champenois and Borges, 2012;Eggert, 2012); >160 mmol C/d/m 2 for GPP and >80 mmol C/d/m 2 for CR in C. nodosa after conversion from O 2 metabolic rates (Egea et al., 2020)] in previous studies exceeded the values measured in this study. The Mediterranean Sea has experienced a decline in seagrass coverage (Delgado et al., 1999), estimated to be between 13 and 38% of initial coverage of P. oceanica since 1960 (Marbà et al., 2014). Physiological stress due to progressing coastal eutrophication, enhanced organic loading of the water column and respective increased light attenuation are contributing to this decline (Brocke et al., 2015), and in fact, the available underwater light at our study site in Punta Negra was generally lower than in other locations in the Western Mediterranean (González-Correa et al., 2008). Additionally, the intrusion of non-indigenous species like H. incrassata have resulted in higher competition for nutrients and light (Davis and Fourqurean, 2001). Although the temperature measured in summer (25.68-27.12 • C) is favorable to the settlement and colonization of H. incrassata, the growth rate of 0.93 new segments per day for the time period investigated is lower than for the congenerous H. tuna in a tropical reef system in Florida, probably related to environmental differences between these studies (Vroom et al., 2003). However, environmental differences to these studies on sites where Halimeda spp. are endemic are apparent and the growth rate for this study might be an underestimation due to its exponential growth pattern.
H. incrassata maintains its basic functioning in an inactive state in low temperatures (15.09-17.38 • C measured in spring) (Thorhaug, 1976) and benefits from a higher temperature in summer, increasing photosynthesis, and promoting calcification (Campbell et al., 2016). Evidently, we see the expected better metabolic performance of H. incrassata in summer, whereas the water temperature exceeded the preferred growing temperature range for P. oceanica (Olsen et al., 2012), which could reduce its growth to cope with thermal stress (Marín-Guirao et al., 2018). The observed decline in autotrophy in summer is in line with the observations of thermal stress in summer (García et al., 2013), specifically when seawater exceeds 28 • C associated with the occurrence of heat waves , whereas the more tolerant C. nodosa seems to be less affected.
The elevated NCP and decline in CR in summer suggests a resource reallocation toward growth rather than energy storage. Whole ecosystem metabolic rates calculated with oxygen data from the multiparametric sensors showed trends of increased productivity in summer in all habitats. Since the hydrolabs were installed 0.2 m above the seafloor, and the measurement of the oxygen profile in the water column is the net result of mixing of water bodies passing through various benthonic as well as pelagic compartments with lateral water movement the multiparametric sensors represent ecosystem values. The distinction between the three macrophyte species and each habitat can only be made with a system preventing this lateral movement like the benthic chambers. However, P. oceanica is known to export a substantial amount of its productivity toward adjacent ecosystems (Duarte, 2002;Kennedy et al., 2010) and the increase in NCP measured by the multiparametric sensors is possibly due to the export of oxygenated water out of P. oceanica meadows. Apart from warming, the Mediterranean Sea shows a trend of decreasing pH of −0.0044 ± 0.00006 pH units per year (Flecha et al., 2015). Seagrass species might have the ability to increase productivity with decreasing pH, using CO 2 as an alternative carbon source; the photosynthetic activity and function of seagrass meadows causes fluctuations of 0.2-0.7 pH units (0.05-0.19 pH units in this study) on a daily scale, making them a valuable buffer system against OA (Hendriks et al., 2010(Hendriks et al., , 2014. The major source of carbon for H. incrassata is the precipitation of bicarbonate to build up structural components, hence a higher calcification rate that can be observed in summer (Invers et al., 1997(Invers et al., , 2001. However, for the vegetated sites with lower shoot density of C. nodosa and H. incrassata, we cannot rule out a substantial contribution of bacterial activity in the sediment to calcification rates, a correction for bare sediment did not lead to reliable results due to the large deviation within incubation settings. Although H. incrassata finds favorable conditions in the study site, its physiology is most likely to be negatively affected by progressing OA due to changes in the water chemistry and further limitation of carbonate (Nelson, 2009;Price et al., 2011;Hofmann and Bischof, 2014). Low water temperatures and elevated levels of CO 2 in spring present the additional risk of dissolution, although non-living segments of Halimeda spp. might be resistant to dissolution under elevated CO 2 (Peach et al., 2017). However, epiphytic coverage, although not investigated here, and potential nutrient limitations on plant growth cannot be excluded.
Nutrient availability is setting an overall limit for the development of seagrasses (Lucea et al., 2003) and the oligotrophic waters around the Balearic Islands are characterized by enriched nutrient concentrations during spring and low concentrations in summer (Alcoverro et al., 1997(Alcoverro et al., , 2001, corresponding with our findings. C:N:P ratios in the water column in our study exceeded the Seagrass Redfield Ratio of 500-550:25-30:1 (Duarte, 1990;Fourqurean et al., 2007). Low ratios in spring are likely to result from increased demand for nutrients due to favorable conditions for growth (Fourqurean et al., 2007), whereas increasing demand for N in summer is likely to lead to P limitation. In addition, our results suggest that the methodology used in this study might have led to P limitation within the incubation bags, most likely due to the limited volume of the incubation chamber (2.27 L for P. oceanica, 1.27 L for C. nodosa, and 1.11 L for H. incrassata). N:P ratios for P. oceanica were considerably higher (5.178 in spring and 3.741 in summer) than reported in previous work (Barrón and Duarte, 2009). Posidonia oceanica represents an important nutrient filter and acts as a net sink of inorganic nutrients, although associated microand macroalgae within the meadows contribute to nutrient depletion (Touchette and Burkholder, 2000). Posidonia oceanica depends on bioavailable NOX and ammonium as nitrogen sources and shows net uptake rates in both seasons, whereas it is assumed to be a net source of ammonium and DON on an annual scale (Barrón and Duarte, 2009). Yet, DON and its bioconversion into bioavailable forms by seagrass associated microbiota (Tarquinio et al., 2018) can function as an important alternative nitrogen source for seagrasses (Vonk et al., 2008;Van Engeland et al., 2011). In fact, P. oceanica seems to depend on DON as co-substrate during its optimal growth period in spring, the substantial uptake of DON during the day to meet the demand for biomass production decreases the overall net release of the habitat (Hemminga et al., 1991;Day et al., 2012). A net release of DON of 2.182 µM/m 2 per day in spring is far lower than in summer (14.575 µM/m 2 per day) and average daily rates based on annual rates (12.055 µM/m 2 per day; Barrón and Duarte, 2009). The utilization of DON as co-substrate is hindered in summer due to progressing P-limitation, overall decreased nutrient availability and intensified the competitive pressure by beneficial growth conditions for C. nodosa and H. incrassata. Whereas both tropical species are acting as net nutrient sources in spring, C. nodosa undertakes a shift to a net sink of NOX and ammonium, suggesting the utilization as a source of nitrogen under enhanced growth. H. incrassata presents a net source of inorganic nutrients even under favorable growth conditions in summer and together with a further decline of P. oceanica increases the risk of further eutrophication.
Concluding, temperature and its implications on metabolism is the key driver for the development of biomass of all three macrophytes and therefore for the vegetative structure and carbon sink activity in coastal settings. Yet, the data presented here also confirms the influence of changing environmental parameters between season on the metabolic activity of the macrophytes. A shift toward higher average temperatures as predicted in climate change scenarios (van Vuuren et al., 2011;IPCC, 2013) will lead to unfavorable conditions for P. oceanica, exceeding its temperature optimum and therefore contribute to further regression of P. oceanica (Telesca et al., 2015). Moreover, a rise in SST might facilitate the colonization by sub-and tropical species and ongoing warming facilitates the settlement and expansion of non-indigenous species such as H. incrassata. Furthermore, progressing nutrient limitation in coastal settings and advantages in competition for nutrients, light and space favoring macroalgae contributes to the decline of seagrasses (Van Tussenbroek and Van Dijk, 2007;Thomsen et al., 2009;Han and Liu, 2014). The loss of vital P. oceanica meadows threatens the autochthonous biodiversity, potentially causing a shift in coastal marine vegetation (Cheung et al., 2009;Lejeusne et al., 2010;Burrows et al., 2011). Further regression of P. oceanica and the inherent loss of ecosystem services will severely impact the productivity and biogeochemical cycling of the coastal ecosystem in the Mediterranean Sea.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
IH: conception. IH and SF: design of the study. LM: draft the first version of the manuscript. All authors conducted the relevant field work jointly, contributed to data acquisition and sample analysis, revised the manuscript critically, contributed, and had substantial impact on the final draft of the manuscript.

FUNDING
Funding was provided by projects PRD2018/18, from the Conselleria d'Innovació, Recerca i Turisme of Mallorca (Spain) and RTI2018-095441-B-C21 (SUMAECO) from the Spanish Ministry of Science, Innovation and Universities. SF was supported by a "Margalida Comas" postdoctoral scholarship, funded by the Balearic Islands Government.