The oceans are the largest global carbon reservoir, holding a total of about 38,000 gigatons of carbon, which is approximately 60 times greater than the total amount of carbon in the atmosphere and approximately 10 times greater than the total amount of carbon in the terrestrial biosphere and soils combined (Ciais et al. 2013). In addition, the fluxes of the Earth’s carbon reservoirs are associated with ocean processes, primarily in conjunction with their connection to terrestrial ecosystems. Among these oceanographic processes, organic carbon fixation by ocean primary producers and carbon exchange at the atmosphere-ocean interface comprise ca. 50% of the anthropogenic carbon (Sabine et al. 2004). Consequently, primary production in pelagic ecosystems accounts for a large percentage of global carbon sinks, and this topic is being actively studied in conjunction with the terrestrial ecosystems. In contrast, less attention has been given to carbon research in coastal ecosystems, likely because they occupy only a small amount (~7%) of the total ocean area (Borges et al. 2005). Furthermore, the input of allochthonous organic material from land and the high density of autotrophic organisms (e.g., vascular plants, macroalgae, phytoplankton, and microphytobenthos) add additional complications to coastal ecosystems, whereas only phytoplankton assemblages generally are considered in the open ocean. Consequently, it is more difficult to fully describe carbon cycles in coastal ecosystems, and more detailed descriptions of the processes important to these cycles are needed if we are to better understand them (Mcleod et al. 2011).
Carbon fixed by photosynthesis can follow several paths, one of which is burial beneath sediments. Although carbon a burial rates are high in coastal areas due to their high primary production, most coastal ecosystems that are dominated by autotrophic organisms (e.g., estuaries, mangroves, and salt marshes) generally act as net CO2 sources to the atmosphere due to overloading of allochthonous organic carbon (Chen and Borges 2009, Mcleod et al. 2011). Indeed, estuary and near-shore ecosystems show especially wide ranges of CO2 effluxes because of diverse sources of photosynthesis and respiration, anthropogenic inputs of carbon and nutrients, and heterogeneity of vegetative assemblages (Borges et al. 2005, Laruelle et al. 2010). In these systems, the high biomass of aquatic vegetation may play a crucial role in coastal carbon sequestration through the conversion of inorganic carbon to organic carbon during photosynthesis (Mcleod et al. 2011, Fourqurean et al. 2012). Consequently, these habitats are often referred to as “hot spots” because of their high carbon sequestration rates associated with rapid primary production. As a result, the global carbon burial budget in vegetated coastal ecosystems, which is estimated to be ~111 Tg C y-1, comprises about one-half of the total carbon buried in the oceans (Duarte et al. 2005) and represents a important portion of total ocean carbon sequestration. Therefore, understanding the rate of autotrophic carbon fixation is essential for better estimation of net carbon fluxes in coastal waters.
Among the oceanic processes that are significant factors connected with oceanic carbon flux, CO2 gas exchange is a relevant parameter associated with carbon flux in oceanic ecosystems in terms of net ecosystem exchange (NEE) (Lovett et al. 2006). NEE, which defines fixed and respired CO2 by individual organisms and populations, or even entire communities, is immensely important for understanding the coastal carbon cycle. Consequently, studies on coastal carbon cycles based on actual measurements of carbon flux parameters at the atmosphere-seawater interface are necessary to understand the roles of vegetation in the coastal carbon cycle.
Despite its importance, carbon mass balance in vegetated coastal ecosystems is difficult to interpret because constructing an adequate carbon model is a daunting job (Ver et al. 1999). In contrast, many conceptual models of carbon cycling have been developed for the open ocean (Randerson et al. 2002, Lovett et al. 2006). These models estimate coastal carbon cycles based on several biological and chemical parameters that are involved in the processes of converting inorganic and organic carbon. However, the organic carbon synthesized by photosynthesis in coastal areas can sink into the benthos, where it can be buried or transported away from the coastal ecosystem to offshore waters. Consequently, budgets of inorganic and organic carbon can be estimated by measuring the components associated with carbon sequestration, such as the net ecosystem production (NEP), the amount of allochthonous carbon input, the transport of carbon-rich material away from shore, and carbon burial rates at the benthos (Duarte et al. 2005).
Based on these perspectives, we focused on ecosystem production (NEP and gross ecosystem production [GEP]) and NEE in a highly productive coastal ecosystem to estimate carbon fluxes driven by autotrophic organisms. In this study, GEP, NEP, and NEE were quantitatively estimated in the Gwangyang Bay coastal ecosystem on the southern coast of Korea, where three communities of benthic autotrophs (eelgrass, macroalgae, and microphytobenthos) and a community of phytoplankton exist. Specifically, the objectives of this study were to 1) estimate GEP and NEP by photosynthetic carbon uptake and respiration of various autotrophs, 2) estimate NEE along the sub-ecosystem types (eelgrass beds, macroalgal beds, unvegetated shallow sediments, deep sediments, and deep rocky habitats), and 3) construct the conceptual box model of CO2 fluxes along the various vegetation types within the local vegetative coastal ecosystem.
MATERIALS AND METHODS
Description of the study site
Gwangyang Bay is a shallow, semi-enclosed system, located on the south coast of the Korean Peninsula, whose hydrology is controlled by exchanges with offshore waters through a narrow strait (ca. 4 km wide) (Fig. 1). The maximum tidal range in the bay is approximately 4 m. The shallow coastal area (less than 10 m water depth) occupies more than 50% of the whole Gwangyang Bay area and is dominated by sediments comprised mostly of silt and clay. The primary source of allochthonous carbon and nutrients from land comes from the Sumjin River, although the Hadong coal-fired power plant and the Yulchon Industrial Complex located on the southwest corner of the bay are also potential sources of carbon, nutrients, and pollutants (Fig. 1). The autotrophic communities in Gwangyang Bay are comprised of phytoplankton, microphytobenthos, eelgrass (Zostera marina L.), and several species of macroalgae, all of which are non-uniformly distributed throughout the bay (Kim et al. 1998, Kim and Choi 2004).
Fig. 1.Locations of the study sites in Gwangyang Bay on the southern coast of Korea. Sub-systems are distinguished by the vegetation types and community compositions of autotrophs with vegetation: AG (macroalgae and phytoplankton) and EG (eelgrass, phytoplankton, and microphytobenthos) and without vegetation: AG-out (phytoplankton), EG-out (phytoplankton and microphytobenthos), and UNV (phytoplankton and microphytobenthos).
In this study, the bay ecosystem was grouped into five habitats according to the dominant types of primary producers present and substrate characteristics. Specifically, these were 1) eelgrass beds (EG), 2) macroalgal stands (AG), 3) deep sedimentary habitats (UNV; >10 m), 4) shallow sedimentary habitats (EG-out; ≤10 m), and 5) deep rocky reefs (AG-out) (Fig. 1). The total area of each of these five habitats within a local box scale (30 km2 ) was estimated using a combination of image analysis of aerial photographs (Daum©, Samah©, Goyang, Korea) and shipboard surveys. These areas were 1) eelgrass beds = 1.9 km2 , 2) macroalgal stands = 1.6 km2 , 3) deep sedimentary habitats = 16.8 km2 , 4) shallow sedimentary habitats = 6.3 km2 , and 5) deep rocky reefs = 3.4 km2 . Following this, two eelgrass beds (EG1 and EG2) were selected to estimate primary production by eelgrass, phytoplankton, and microphytobenthos. One macroalgal stand (AG) was selected to determine the spatial distribution of each macroalgal species and phytoplankton and to estimate their individual rates of primary production. Four nonvegetated sites, two shallow sedimentary habitats (EG1-out and EG2-out), one deep sedimentary habitat (UNV), and one deep rocky reef (AG-out), were also selected, and primary production by the resident autotrophs was estimated. Finally, we developed a conceptual box model for carbon flux in the bay based on productivity estimates for the autotrophic organisms made between 20 June and 20 July 2007.
Daily radiation and environment factors
Down-welling irradiance was recorded continuously near the benthos at four of the sampling sites (EG1, EG2, AG, and UNV) using an LI-COR underwater photosynthetically active radiation (PAR) sensor (LI-192) connected to a data logger (LI-1400; LI-COR Inc., Lincoln, NE, USA) that was deployed during in situ incubation experiments on 20-21 June and just above the water surface (air) using a LI-COR PAR sensor (LI-190). The attenuation coefficient of the PAR within the water column was then calculated using Beer’s Law to find the effective irradiances available to the autotrophs in each habitat. The irradiance measured at the eelgrass canopy (approximately 1 m depth) was used to calculate eelgrass production, and the irradiance at the water depth where macroalgal vegetation was most abundant (approximately 4 m depth) was used to calculate macroalgal production. Phytoplankton production at various depths was calculated using the air irradiance data while considering the light attenuation coefficient (Kim et al. 2013a). Additionally, the irradiance profiles at the sediment surface were used to calculate productivity of the microphytobenthos.
Water temperature, salinity, and wind velocity were measured during seawater sampling, and these were used to estimate the pelagic productivity and CO2 gas exchange with the atmosphere. Specifically, temperature and salinity were measured using a HOBO Pendant temperature loggers (Onset Computer Corp., Bourne, MA, USA) and an 8410-A Portasal Portable salinometer (accuracy, ± 0.003; OSIL, Hampshire, UK), respectively. Data for wind velocity were obtained from meteorological observations collected about 14 km from the study sites.
Eelgrass and macroalgae biomass
Eelgrass shoot density was estimated within 13 to 15 quadrats (0.063-m2) that were haphazardly placed within each eelgrass bed by a SCUBA diver. The aboveground biomass of eelgrass for these samples was estimated using the average of the individual shoot weights within each quadrat, and the total area of the eelgrass beds within the bay was determined using a DT-X digital echosounder (Biosonics Inc., Seattle, WA, USA), with data collected using a high frequency transducer (420-kHz single-beam). The echosounder data showed that the eelgrass was distributed homogeneously among the study sites (Kim et al. 2008). Additionally, the biomass of macroalgal species was quantified by collecting all macroalgae within 0.063-m2 quadrats that were placed at 2-m intervals along a three transect line running perpendicular to the coastline at the AG site (n = 18 quadrats). After sample collection, the wet weight was measured to estimate individual species distributions, and the biomass of macroalgae within the samples was determined based on the sum of the different species. The total biomass of macroalgae in the bay was then estimated based on the total area of macroalgal stands and the mass and relative abundance of each species.
Productivity and respiration of eelgrass and macroalgae
To estimate rates of photosynthesis and respiration by individual eelgrass and macroalgae, in situ incubation experiments were done in the northeastern part of the bay (beach near EG1) on a clear sunny day (June 20, 2007). Three replicate samples (n = 3) of eelgrass (Zostera marina) and the six most common species of macroalgae were collected from each habitat by a SCUBA diver and moved to the incubation site in an opaque bucket to minimize exposure to light. Leaf #3 (counted from the youngest leaf) from the eelgrass shoot and intact thalli of the macroalgae were selected for in situ incubation under various light intensities after removal of epiphytic organisms by hand because photosynthesis of third leaf is closest approach to mean individual productivity (unpublished data). Specifically, incubation bottles (ca. 280-mL cell culture flasks; Corning Inc., Corning, NY, USA) were individually prepared by filling each bottle with 20-µm-filtered seawater from either the eelgrass (EG-2) or macroalgae (AG) collection sites and inoculating these with 10 µCi of NaH14CO3. The bottles were then wrapped in layers of neutral density screen to create seven different light-transmission intensities (i.e., 0, 13, 17, 30, 43, 67, and 100% of ambient) (Kim et al. 2013a) in order to construct P-I curves. This system was naturally buoyant when placed in the water and was maintained at approximately 10 cm depth, thereby ensuring stable ambient temperatures and allowing waves to naturally agitate the bottle to reduce boundary layer formation. The light intensity at this depth was measured using an LI-1400 data logger with an LI-192 underwater quantum sensor. Each sample was incubated for 3 h, which was sufficient to measure oxygen production and consumption (unpublished data) while minimizing bottle effects such as nutrient exhaustion and / or deterioration of sample vitality (Mateo et al. 2001). After the in situ incubation, photosynthetic activity within each bottle was stopped by the addition of 4% neutral formalin. Small pieces of 14C-labeled tissue (~20 mg)were removed from each bottle and added to individual scintillation vials (20 mL) for analysis of radioactivity by the newly synthesized organic carbon. Here, the plant tissues were first solubilized in darkness with 0.5 M NaOH and 5% Triton X-100 and then oxidized with H2O2 with the addition of 10 mL of Hionic-Flour scintillation cocktail (PerkinElmer, Norwalk, CT, USA). The radioactivity of each sample was then determined using a liquid scintillation counter (Tri-Carb 2100 TR; Packard Instruments Co., Meriden, CT, USA) (Gómez et al. 2007). Following this, carbon uptake by each sample was calculated using an equation derived by Penhale (1977). To calculate carbon uptake rate, total dissolved inorganic carbon (CT) in the seawater was determined by standard method of seawater carbon chemistry analysis (Dickson et al. 2007). The results obtained from these short-term incubation experiments measure gross photosynthesis (Mateo et al. 2001). Dark respiration (Rd) was measured by O2 consumption rates using an optical oxygen sensor spot with a fiber optic connection to Fibox 3 (PreSens, GmbH, Regensburg, Germany) under field conditions (Kim et al. 2013b). Specifically, respiration was calculated by oxygen consumption after 2 h of incubation under dark conditions, and a respiratory quotient of 1.0 was used when converting oxygen to carbon units.
Productivity and respiration of pelagic phytoplankton
The pelagic primary production of phytoplankton at the eelgrass beds (EG1 and EG2) and at the deep rocky reef (AG-out) was determined by the 14C technique in order to estimate the phytoplankton contribution to ecosystem production. Seawater samples were filtered through a 200-µm mesh to remove large zooplankton and suspended particles. The resulting water and phytoplankton were then added to incubation bottles (ca. 80-mL cell culture flask; Corning), which were wrapped in layers of screen as was done with the macroalgae and eelgrass. Each bottle was inoculated with 10 µCi of NaH14CO3 and then incubated at in situ surface water for 3 h. After incubation, samples were filtered using very low vacuum pressure (<50 mm Hg) through glass membrane filters (Whatman GF/F; GE Healthcare Ltd., Buckinghamshire, UK). UltimaGold (PerkinElmer) scintillation cocktail (10 mL) was added to each filter before analysis of the radioactivity. The radioactivity of the synthesized particulate carbon was measured using a liquid scintillation counter. The carbon uptake was then calculated using an equation derived by Strickland and Parsons (1972). To calculate carbon uptake rate, CT in the seawater was determined by standard method of seawater carbon chemistry analysis (Dickson et al. 2007). Areal community production of phytoplankton was calculated with integration of whole depth production with consideration of attenuation coefficient of the PAR within euphotic zone (Taylor 1999). We considered the daily community respiration (CR) in the water column from a literature survey because oxygen consumption rates were not detected using oxygen meters, which were connected with electrochemical and optical oxygen probes, and it accounted for approximately 73% of daily gross production (Williams 1998).
Productivity and respiration of benthic microbial mats
To estimate primary production and respiration by microphytobenthos in the soft sediments, sediment samples were collected from vegetated shallow sedimentary (EG2-in), unvegetated shallow sedimentary (EG2-out), and unvegetated deep sedimentary (UNV) regions. For this work, transparent acrylic tubes (ID 5 cm) were inserted ca. 12 cm deep into the sediment by a SCUBA diver without disturbing the surface sediments. Five intact sediment cores at each sampling sites were then moved quickly to the laboratory (within 5 h), where they were held under simulated field conditions of temperature, salinity, and light-dark cycle (23°C, S = 32, and 12 : 12 LD cycle, respectively). Additionally, the samples were acclimated to low light levels provided by daylight fluorescent lights (20 µmol photons m-2 s-1). Oxygen production by photosynthesis was measured with vertical oxygen microprofiles that were obtained using a Clark-type oxygen microelectrode (tip diameter 25 µm; OX-25; Unisense, Aarhus, Denmark), and net areal production and respiration of the microbial mat were calculated based on each oxygen microprofile after 20 min of exposure to seven different of light intensities, which were provided by a halogen illuminator (KL 2500 LCD; Schott, Mainz, Germany) (Kühl et al. 1996). The net community production (NCP) and respiration (CR) were calculated as the sum of the upward and downward O2 fluxes across the diffusive boundary layer based on Fick’s first law of one-dimensional diffusion equation with the consideration of diffusion coefficient along the salinity and temperature (Jørgensen and Revsbech 1985), and oxygen-based NCP and CR were converted to carbon-based units using a photosynthetic (1.2) and respiratory quotient (1.0).
Net / gross community and ecosystem production
Photosynthetic parameters of eelgrass, macroalgal species, and microalgal communities (maximum photosynthetic rate, PB max; photosynthetic efficiency, α; and saturating irradiance, Ek) were estimated by fitting of light response curves using the inhibition model of Platt et al. (1980). The mean values of these parameters and the hourly irradiance profiles were substituted again into the mathematical formulation of Platt et al. (1980) to calculate community productivity (Appendix 1). Areal biomass of vegetation was also considered in order to calculate daily areal community production. Irradiance data were obtained at the center of the eelgrass canopy (1 m depth) and at the water depth where macroalgal species was distributed dominantly (3 m depths) to estimate the production by each species. Estimates were also obtained at <3 m depths for unvegetated and shallow sedimentary habitats and at >3 m depth for deep sedimentary habitats in order to estimate the community production of the benthic microbial mats. These were summed over a 24-h period using irradiance data from different depths and at each site to derive the daily community production (Kim et al. 2013a). In addition, the whole integration of NCP / gross community production (GCP) driven by the autotrophs represents the NEP / GEP.
Appendix 1.Examples of daily variations in atmosphere and underwater irradiance along the studied sites (gray line, yellow filled) and net carbon uptake rate (colored line) of eelgrass (A), macroalgae (B), phytoplankton community (C), and microphytobenthos (D) based on the diurnal irradiance changes and photosynthetic parameters on 20 June 2007.
Determination of seawater pCO2 and NEE (CO2 exchange)
To determine the carbon chemistry in the seawater, surface seawater samples were collected from each site around midday. The samples were transferred quickly to 500-mL Pyrex bottles, without introducing air bubbles, and then immediately poisoned with 100 µL of saturated HgCl2 (Dickson et al. 2007). Concentrations of total dissolved inorganic carbon (CT) and total alkalinity (AT) were determined in the laboratory using coulometric and potentiometric acid titrations in a VINDTA system (Marianda, Kiel, Germany). The accuracy and precision of the CT and AT measurements were checked daily against seawater reference materials with known CT and AT values (certified by A. Dickson, Scripps Institution of Oceanography, San Diego, CA, USA). The precisions of the measurements were approximately ±2 µmol kg-1 for CT and ±1.5 µmol kg-1 for AT. The seawater partial pressure (pCO2) values were calculated from the measured CT and AT values using the carbonic acid dissociation constants of Mehrbach et al. (1973). This set of thermodynamic constants has proved to be the most consistent with laboratory (Lee et al. 1996, Millero et al. 2006) and field (Lee et al. 2000, Millero et al. 2002) measurements of carbon parameters over the oceanic ranges of temperature and salinity. The CO2 exchange at the air-sea interface was calculated using the equation of CO2 exchange = k × s × (pCO2-air - pCO2-sea), where k is the gas transfer velocity (cm h-1), s is the CO2 solubility in seawater, and pCO2-air and pCO2-sea is the difference in partial pressure of CO2 between air and seawater (Weiss 1974). Partial pressure of CO2 in the air was determined by mean of annual values at the Kosan, Jeju Island (386 ppmv). The gas transfer velocity was computed as k = 0.31 × U102 × (Sc/660)-0.5, where U10 is the wind velocity (m s-1) at 10 m in height and Sc is the Schmidt number evaluated for a given temperature and water density [Sc = 2,073.1 - 125.62(temp.) + 3.6276 × (temp.)2 - 0.043219 × (temp.)3 ] (Jähne et al. 1987, Wanninkhof 1992).
Daily radiation and environment factors
The daily atmospheric solar irradiance in the bay ranged from 3 to 33 mol photons m-2 d-1 during the study period (20 June to 20 July 2007) (Appendix 2A). Based on the estimated light attenuation coefficient, the surface irradiance was reduced to about 8-12% at the eelgrass canopy depth in EG1 and EG2 and to about 15-32% at the biomass-rich macroalgal site (AG). The irradiance at the shallow sediment (bottom of the eelgrass beds) was 5-8% of surface irradiance (Appendix 2B), and at the deep unvegetated sedimentary site (UNV), only 1% of surface irradiance reached the sediment surface. In contrast, water temperature and salinity were relatively consistent, ranging from 20 to 25°C and from S = 32.2 to 32.5, respectively. The mean wind velocity during the study period was 1.27 m s-1.
Appendix 2.(A) Monthly variation in daily irradiance based on total daily light intensity. (B) Example of diurnal irradiance changes along the vegetation types obtained at the bottom of sampling sites on 20 June 2007 (EG1 and EG2, ca. 3 m; AG, ca. 4 m in depth; and UNV, ca. 9 m).
Eelgrass and macroalgal biomass
At eelgrass bed EG1, the aboveground eelgrass biomass was estimated to be 805.3 ± 138.6 g FW m-2. The area covered by the eelgrass and its total biomass was estimated to be ca. 15,960 m2 and 13 wet tons in June 2007, respectively (Table 1). At eelgrass bed EG2, the mean aboveground eelgrass biomass was estimated to be 975.0 ± 158.8 g FW m-2 and the area covered by the eelgrass and its total biomass was estimated to be ca. 52,800 m2 and 51 wet tons, respectively.
Table 1.All values are presented as the mean ± standard error; number of quadrats = 13-15.
At the macroalgal vegetation site (AG), eleven macroalgal species, of which two were green algae, three were brown algae, and six were red algae, were distributed along the subtidal rocky shore (Table 2). Among these, six species (Ulva australis, Gelidium amansii, Gracilaria textorii, Schizymenia dubyi, Chondria crassicaulis, and Chondrus ocellatus) comprised 91% of the total macroalgal biomass and made up ca. 2 wet tons in the macroalgal vegetation area of 4,300 m2 . More than half of the macroalgal biomass (64%) was distributed in shallow subtidal regions (ca. 2 m depth), and 81% of the biomass was at a water depth of <4 m. Among the green alga, U. australis had the highest individual biomass (mean of 83.12 g FW m-2, range from 0 to 627.47 g FW m-2) and comprised 18% of the total biomass of macroalgae, mostly in the upper littoral zone. Two of the red algae, G. textorii and C. ocellatus, had considerably higher biomass than other species, ranging from zero to 1,320.69 g FW m-2 (mean of 190.59 g FW m-2) for G. textorii and from zero to 1,012.66 g FW m-2 (mean of 116.19 g FW m-2) for C. ocellatus, which represented about 41 and 25% of the total macroalgal biomass, respectively (Table 2).
Table 2.The total biomass of macroalgae was calculated by the species biomass and area of macroalgal bed (4,300 m2 ). All values are presented as means; number of quadrats = 18. a Species was used for a photosynthesis experiment.
Productivity and respiration of vegetation (eelgrass and macroalgae) and microbial communities (phytoplankton and microbial mats)
Photosynthesis varied among the eelgrass and macroalgal species, and its characteristics were clearly identified by photosynthetic parameters of the P-I curves (Appendix 3A-E). Specifically, the maximum photosynthetic rate (PB max), photosynthetic efficiency (α), and saturating irradiance (Ek) of eelgrass Z. marina were 3.28 ± 0.21 mg C g-1 FW h-1, 0.008 ± 0.000 mg C g-1 FW h-1 / µmol photons m-2 s-1, and 241 ± 46 µmol photons m-2 s-1, respectively (Table 3). For the macroalgae, U. australis had the highest PB max (8.31 ± 0.87 mg C g-1 FW h-1) as well as the highest respiration rate (Rd = 0.46 ± 0.04 mg C g-1 FW h-1). Five species of red algae, which accounted for 78% of the total macroalgal biomass, all had much lower PB max values, which ranged from 0.54 to 2.45 mg C g-1 FW h-1 and were 70-93% lower than those of U. australis. In addition, the highest α was observed for U. australis (0.032 ± 0.008 mg C g-1 FW h-1 / µmol photons m-1 s-1), and it was 2-10 times higher than those of other macroalgal species. Other red algae had low α and Rd comparable to the green alga and eelgrass, which varied from 0.003 to 0.017 for α and from 0.08 to 0.15 mg C g-1 FW h-1 for Rd. Z. marina, C. crassicaulis, and G. amansii had high Ek of more than 200 µmol photons m-1 s-1 . Eelgrass and most macroalgae showed the photosynthetic characteristic of photoinhibition, with irradiance stronger than saturating irradiance, except for C. crassicaulis. The degree of photoinhibition varied among the species, with G. textorii showing the highest photoinhibition under strong light intensity (gross photosynthesis was almost compensated).
Table 3.Data units for eelgrass and macroalgae: PB max, mg C g-1 FW h-1; α, mg C g-1 FW h-1 / μmol photons m-2 s-1; Ek, μmol photons m-2 s-1;Rd,mg C g-1 FW h-1. Data units for phytoplankton and microphytobenthos: PB max, mg C m-2 h-1; α, mg C m-2 h-1 / μmol photons m-2 s-1; Ek, μmol photons m-2 s-1; CR,mg C m-2 h-1. All values are presented as the mean ± standard error. ND, not determination.
Appendix 3.Photosynthesis vs. irradiance curves (P-I curves) of eelgrass (Zostera marina) (A), six species of macroalgae Ulva australis (B), Chondria crassicaulis (C), Chondrus ocellatus (D), Gelidium amansii (E), Gracilaria textorii (F), and Schizymenia dubyi (G), phytoplankton communities obtained from in situ experiments using the classical 14C method (H), and microphytobenthos communities obtained from laboratory experiments using the oxygen profiling method (I). Vertical bars represent ± standard error of the means (n = 3-5).
The photosynthetic characteristics of phytoplankton obviously differed by vegetation types, but those of the microbial community were not significantly identified with habitat types (Appendix 3H & I). Higher areal PB max max and α of phytoplankton were observed in the eelgrass beds, with an average of 93% higher for PB max max and 117% higher for α in the eelgrass beds (EG1 and EG2) than in AG. The Ek values of phytoplankton assemblages from the three habitats ranged from 111 to 141 µmol photons m-2 s-1, and these values were not very different among the vegetation types. The photosynthetic parameters of areal PB max max for the microphytobenthos were highest in the eelgrass bed (EG2, 19.99 ± 8.27 mg C m-2 h-1) and lowest on the outside of the eelgrass bed (EG2-out, 13.40 ± 1.70 mg C m-2 h-1). The α of microphytobenthos at the eelgrass bed was similar to that of the deep sedimentary habitat (mean value, 0.110 mg C m-2 h-1 / µmol photons m-2 h-1), and this value was 44% higher than the value outside of the eelgrass bed (EG2-out). The range of Ek of microphytobenthos was 141 to 197 µmol photons m-2 s-1 irrespective of the habitat type. Additionally, sediment respiration, which is included in microphytobenthos and microbial respiration, did not show a large difference between habitat types, ranging from 7.07-7.54 mg C m-2 h-1. Together, the photosynthesis and respiration of dominant macrophytes and microbial communities are important factors for estimating community and ecosystem production.
Community / ecosystem production and NEE
GCP was high in the vegetation communities, but a high rate of respiration is a potential countervailing factor for high GCP of the vegetation system. Specifically, the eelgrass and macroalgal communities exhibited extremely high NCP and GCP compared with those of the microalgal communities (phytoplankton and microphytobenthos) (Fig. 2). Additionally, the GCPs at the eelgrass beds were 6.51 ± 0.67 and 5.70 ± 0.62 g C m-2 d-1 in EG1 and EG2, and the NCPs of the eelgrass accounted for 35 and 10% of GCP at EG1 and EG2, respectively. The NCP at the macroalgal vegetation (AG) site accounted for 63% of GCP, and the GCP and NCP values were 4.83 and 3.05 g C m-2 d-1, respectively. The pelagic phytoplankton community was a net autotrophic system (mean value, 0.288 g C m-2 d-1), whereas the benthic microbial mat showed net heterotrophic communities (mean value, -0.156 g C m-2 d-1) irrespective of community type.
Fig. 2.The average daily net community production (NCP) (A) and daily gross community production (GCP) (B) of autotrophic organisms at the seven sites during 20 June to 20 July 2007 (31 days). Daily net ecosystem production (NEP) and gross ecosystem production (GEP) are calculated by integration of NCP and GCP along the autotroph compositions. All values are presented as mean ± standard error.
The high rates of GCP and NCP in the vegetation systems also led to higher values of NEP and GEP (Fig. 2). According to the NEP and P : R ratio (GEP to ecosystem respiration, Re), all study sites showed net autotrophic ecosystems, ranging from 0.09-3.31 g C m-2 d-1 for NEP and from 1.10-2.33 for P : R ratio (Table 4, Fig. 2). GEP was highest at the eelgrass habitat (EG1, 7.06 ± 0.68 g C m-2 d-1) due to the high productivity and biomass of eelgrass. Otherwise, the lowest GEP was observed at the unvegetated rocky shore (AG-out, 0.97 ± 0.04 g C m-2 d-1), where only phytoplankton production was considered. The highest NEP was found in AG (3.31 ± 0.45 g C m-2 d-1), along with the highest P : R ratio (2.33). Also, the eelgrass beds showed ca. 78% lower NEP (2.26 ± 0.67 and 0.73 ± 0.63 g C m-2 d-1 in EG1 and EG2, respectively) compared to GEP, even though these sites showed the highest GCP due to high ecosystem respiration.
Table 4.Monthly gross ecosystem production: ecosystem respiration (Re) ratio at each study site (20 June-20 July 2007)
Based on the results of the seawater CO2 exchange measurements (NEE), the eelgrass beds (EG1 and EG2), which had high GEP with dense eelgrass, act as CO2 sinks (-0.81 ± 0.17 and -0.45 ± 0.09 mg C m-2 d-1 for EG1 and EG2, respectively). Additionally, EG1-out acts as a CO2 sink despite an NEP that was lower than those in the vegetation systems (-0.34 ± 0.07 mg C m-2 d-1). The other unvegetated sites (except for EG1-out) act as sources of CO2, for which the results ranged from 0.20 to 2.32 mg C m-2 d-1 (Fig. 3). Otherwise, AG acts as a CO2 source to the atmosphere even if this system is characterized by highly productive areas (1.66 ± 0.34 mg C m-2 d-1).
Fig. 3.Comparison of net ecosystem production (NEP) (A) and gross ecosystem production (GEP) (B) vs. CO2 exchange at the airseawater interface (net ecosystem exchange, NEE) for the two eelgrass beds (◆, EG1; ●, EG2), the macroalgal bed (▲, AG), shallow sedimentary region (◇, EG1-out; ○, EG2-out), deep sedimentary region (□), and deep rock region (△, AG-out) during 20 June to 20 July 2007 (31 days). All values are presented as the mean ± standard error.
Conceptual model of the coastal carbon cycle
A conceptual diagram of the coastal carbon cycle, which puts ecosystem production and NEE along the various types of sub-ecosystems, is represented in Fig. 4 within a local box scale (ca. 30 km2 ). The entire Gwangyang Bay system uptakes a total 41.4 tons of inorganic carbon via GEP and releases 35.1 tons of CO2 through Re to coastal waters (Fig. 4A). Specifically, the deep sedimentary region showed the highest GEP of 18.0 ton C d-1. Although the eelgrass bed occupied only a small area, this system represents 73% of the bay’s GEP in the local box scale compared to the deep sedimentary habitats (13.3 ton C d-1). The macroalgal vegetation had the lowest GEP, NEP, and Re (0.6, 0.3, and 0.2 ton C d-1, respectively), presumably because the macroalgal habitat occupied the smallest proportion of the bay’s area. The eelgrass bed has the highest NEP among the sub-ecosystems (2.8 ton C d-1). Consequently, this embayment system has approximately 6 tons of carbon is potentially fixed as organic matter based on NEP, and some portion of photosynthesized organic carbon possibly sequestered to the outer sea. Also, approximately 47 kg of CO2 is emitted into the atmosphere from the seawater, as estimated by NEE.
Fig. 4.Conceptual diagram of the carbon flux driven by coastal primary producers in the northeastern part of Gwangyang Bay (30 km2 ). Mean value of daily gross ecosystem production (GEP) and Re (A) and net ecosystem production (NEP) and net ecosystem exchange (NEE) (B) at different vegetated / unvegetated systems during 20 June to 20 July 2007 (31 days).
Although many studies have attempted to describe coastal carbon cycles, there have been limits to quantify carbon budgets within vegetative coastal ecosystems, and additional efforts are still required (Tokoro et al. 2014). For example, some studies have focused on biological metabolism and carbon mass balance approaches within ecosystem production based on gas exchange and NCP (Bates et al. 1998, Sarma 2003). Additionally, carbon flux has been estimated by comparing net ecosystem production to gas exchange in an estuary system (Pradeep Ram et al. 2003). However, these studies have been largely restricted to unvegetated areas where only phytoplankton and / or microphytobenthos were considered. Otherwise, some studies have focused on biological metabolism to estimate the carbon balance in vegetative coastal habitats. The carbon budgets in vegetation-dominated habitats have thus been estimated by measuring primary production. However, these studies have not considered many of the complicated carbonic parameters, such as gas exchanges with the atmosphere (Frankignoulle and Bouquegneau 1987, Kaldy et al. 2002). To address these deficiencies, our investigation was conducted to establish a conceptual box model for coastal carbon fluxes, with attention to carbon mass balance based on ecosystem production and CO2 gas exchange in various types of vegetated and unvegetated systems.
Coastal vegetated ecosystems are generally characterized as highly productive areas, and our production / respiration data support this. All of the sub-ecosystems in our study were autotrophic in nature (i.e., P : R ratio >1) (Table 4), with production : respiration ratios comparable to those found in other studies (e.g., seagrass bed, 1.19- 1.50; unvegetated habitats, 0.03-34.3) (Duarte and Agusti 1998, Ziegler and Benner 1998). Although Gwangyang Bay was a sufficiently productive environment during the experimental period, as shown by the P : R ratio, not every sub-ecosystem acted as a CO2 sink from the atmosphere. This is not surprising given that many coastal ecosystems with vigorous biogeochemical activities can act as CO2 sources to the atmosphere (Borges et al. 2005). Our NEE results showed that a heavily vegetated bay could act as both a sink and source of CO2 depending on the type of vegetation present (Fig. 3). Specifically, our two eelgrass beds acted as CO2 sinks, presumably because of the high photosynthetic activity of the eelgrass and the near-stagnant seawater. These environmental traits may lead to a reduction of pCO2 in the surface seawater with disruption of seawater mixing between the inside and outside of the eelgrass bed (Madsen et al. 2001). These characteristics also occurred at our unvegetated areas. Specifically, the shallow sedimentary site (EG-out) had moderate rates of CO2 exchanges; however, the high-energy rocky shore (site AG) and deep coastal sedimentary sites (AG-out and UNV) acted as CO2 sources to atmosphere. Contrary to the eelgrass beds, these habitats may lead to the migration of inorganic carbon out of the habitat due to horizontal and vertical turbulent eddies and mixing (Madsen et al. 2001). Thus, pCO2 may not be reduced in the seawater in these habitats even though macroalgae take up large amounts of inorganic carbon from the surrounding seawater. Consequently, NEE can be characterized by complex traits of coastal environments such as seawater movement and organic inflows from land mitigated by the productivity of autotrophs.
The NEP (integration of pelagic and benthic NCPs) in the coastal ecosystem is a key factor associated with carbon export to the region outside of the coastal system (Duarte et al. 2005), and it is particularly important in the aspect of carbon sequestration because a significant portion of the exported carbon can be moved to the open ocean or become buried in sediment. Especially, submerged aquatic vegetation showed significantly high ecosystem production (Kaldy et al. 2002). Thus, the production of benthic autotrophs is an important component of the coastal ecosystem with respect to regulating the carbon cycle, and it could enhance the organic carbon sequestering efficiency (Mcleod et al. 2011). The areal NEPs partitioned by the five types of coastal sub-ecosystems (eelgrass bed, macroalgal vegetation, shallow sedimentary, deep sedimentary, and deep rocky) were substituted to construct the conceptual diagram of carbon flux in the local coastal system (see Fig. 4). The estimated NEP value in the local box scale (30 km2 ; northeastern part of Gwangyang Bay) was about 6.3 ton C d-1, which could be expressed as the organic carbon mass balance (Fig. 4A). Especially, the NEP in vegetated habitats (eelgrass and macroalgal beds) exhibited 33% of GEP in a whole are of studied site (30 km2 ), even though these habitats occupied only 12% of the total area. If vegetated coastal ecosystems produce an overload of photosynthesized organic carbon, it may lead to an increase in the rate of carbon sequestration (Barrón et al. 2003). There are two important pathways for understanding the fate of fixed organic carbon in the coastal ocean (Ver et al. 1999). A portion of the exported carbon will be either 1) transported to the open ocean (dissolved and / or particulate organic carbon) or 2) buried in the sediment. The organic detritus on the upper layer of the sediments could af-fect the oxidation and recycling of nutrients (Giusti and Marsili-Libelli 2005) and be exported to the marginal sea through the “continental shelf pump mechanism” (Tsunogai et al. 1999). However, quantification of the organic carbon cycle in the vegetated coastal ecosystem via these two pathways was not addressed by our research, thus these factors remain as topics for future research to clarify the coastal organic carbon cycle. Another minor possible pathway of exported carbon is that it is 3) emitted to the atmosphere. Our NEE results quantified the CO2 emitted to the atmosphere and showed that our local coastal area emits approximately 47.4 kg d-1 CO2 into the atmosphere, which is about 200 times smaller than NEP. This level is small compared to ecosystem production, but CO2 gas exchanges at the atmosphere-seawater interface have yet to be explored in relation to the ecosystem production of various vegetation systems.
This study would be remarkable indeed, since many measurable parameters were considered in relation to coastal carbon flux such as biomass, productivity of various coastal autotrophs, and CO2 exchanges with in situ light profiles. However, spatial-temporal factors are not reflected in this study because of difficulties associated with their investigation. For example, carbon input sources from the inland and heterogeneities of vegetation standing crop and productivity are still needed to be clarified for understanding coastal carbon fluxes in vegetative embayment systems, but these parameters are highly influenced by varied time-space scales (Randerson et al. 2002). Especially, regulation of NEP is highly depending on gross primary production and ecosystem production with monthly time scale and regional space scale research, but river carbon export is relatively low priority. Although, this study conducted with limited time-space scales, river carbon input source should be considered to complete whole conceptual carbon flux model within the vegetative bay system to extent time-space scale. Also, targeted study area showed high magnitude of tidal cycle, thus it crucial affects on productivity of coastal autotrophs because light intensity and microalgal abundance are highly relied on this factor. We could not apply the complicated tidal cycles to design conceptual model of coastal carbon cycling, it should be clarify through the future investigation.
In summary, this study provides an example of how to use a mass balance approach to integrate estimates of carbon flux with measurements of photosynthesis and CO2 gas exchange in a complicated coastal vegetated system. Based on our results, we quantified carbon flux over the short-term in a bay ecosystem using the model proposed. Although, our conceptual box model does not consider all the complicated biological activities and environmental fluctuations, it is sufficient for evaluating short-term and local-scale estimations of carbon flux. Additionally, we presented a conceptual box model that can be applied to other vegetated systems in order to estimate coastal carbon cycling. To achieve this aim, additional parameters about carbon chemistry in seawater need to be considered. These include the fate of dissolved inorganic carbon, particulate / dissolved organic carbon, and hydrological parameters and carbon mineralization for the long-term and large-space scale.