Leaf segments from 3 randomly selected Z. muelleri plants were positioned horizontally in a custom-made flow chamber (Brodersen et al., 2014). Within the chamber, leaf segments were fixed onto a polystyrene plate by needles. The cut ends of the investigated leaf segments were sealed with petroleum jelly prior to experiments to seal the aerenchyma from the surrounding water. A constant flow (~1 cm s⁻¹) of aerated seawater (23°C, salinity = 29) was maintained in the flow chamber via a pump submerged into a seawater reservoir. The applied flow velocity of ~1 cm s⁻¹ is in the lower end of field water velocities (e.g., Gambi et al., 1990; González-Ortiz et al., 2014), but does resemble water movement especially within dense seagrass meadows in closed lagoons such as the conditions in Narrabeen Lagoon. Illumination was provided by a fiber-optic tungsten halogen lamp fitted with a collimating lens (KL-2500LCD, Schott GmbH, Germany). The downwelling photon irradiance (PAR, 400–700 nm) at the leaf surface was measured with a scalar irradiance minisensor (US-SQS/L, Walz GmbH, Germany) connected to a calibrated photon irradiance meter (LI-250A, LI-COR, USA). The leaf segments were illuminated with an incident photon irradiance of 0, 75, 200, and 500 μmol photons m⁻² s⁻¹. Water-column hypoxia was obtained by continuously flushing the seawater in the supporting water reservoir with a mixture of atmospheric air and humidified nitrogen gas. The O₂ concentration of the water reservoir was simultaneously monitored by a submerged Clark-type O₂ microsensor (OX-10, tip diameter of 10 μm, Unisense A/S, Aarhus, Denmark; Revsbech, 1989).

We used Clark-type O₂ microsensors (OX-50, tip diameter of ~50 μm, detection limit ~0.3 μM, Unisense A/S, Aarhus, Denmark; (Revsbech, 1989)) with a fast response time (t₉₀ < 0.5 s) and a low stirring sensitivity (<2–3%) to measure the O₂ concentration at and toward the leaf surface. The O₂ microsensors were mounted on a motorized micromanipulator (Unisense A/S, Aarhus, Denmark) and connected to a microsensor multimeter (Unisense A/S, Aarhus, Denmark) both interfaced with a PC running dedicated data acquisition and positioning software (SensorTrace Pro, Unisense A/S, Aarhus, Denmark). The O₂ microsensors were linearly calibrated from signal readings in 100% air saturated seawater and anoxic seawater (by N₂ flushing and addition of the O₂ scavenger Na₂SO₃) at experimental temperature and salinity. Prior to measurements and calibrations, the microsensors were pre-conditioned with H₂S to prevent drifting calibrations when exposed to H₂S during experiments (Brodersen et al., 2015a). Microsensors were carefully positioned at the leaf tissue surface (defined as 0 μm distance on figures) by manually operating the micromanipulator, while observing the leaf tissue surface and microsensor tip with a boom-stand dissection microscope (AmScope, Irvine, CA, USA). When changing the downwelling photon irradiance, steady state O₂ conditions at the leaf surface re-occurred after ~60 min (data not shown). Microprofiles of O₂ concentration were measured in vertical increments of 100 μm, from the leaf tissue surface to 2 mm distance away (which is in the same order of magnitude as the leaf tissue thickness).

O₂ fluxes across the leaf tissue surfaces were calculated using Fick's first law of diffusion:

where D₀ is the molecular diffusion coefficient of O₂ in seawater at experimental temperature and salinity (2.14 × 10⁻⁵ cm⁻² s⁻¹; cf. tabulated physical parameters for marine systems available at www.unisense.com), and ∂C/∂z is the linear O₂ concentration gradient in the DBL. As we introduced a physical barrier to O₂ diffusion at the abaxial surface by fixing the leaf onto polystyrene with a low O₂ permeability, we take the flux estimated at the adaxial side of the seagrass leaf as representative for the net flux of O₂ across the leaf surface, i.e., J_{O2, tot} = J_{O2, upper-surface} in dark (=respiration) and light (=net photosynthesis; assuming a photosynthetic quotient of 1 mol O₂ produced per mol CO₂ fixed), respectively.

The calculated net photosynthesis rates (nmol O₂ m⁻² s⁻¹) as a function of the incident photon irradiance (E; μmol photons m⁻² s⁻¹) were fitted with an exponential saturation model (Webb et al., 1974) with an added term, R, to account for respiration (Spilling et al., 2010):

This equation enables estimation of the irradiance at the onset of photosynthesis saturation as E_k = P_max/α, where P_max is the maximal net photosynthesis rate and α is the initial slope of the P_n vs. E curve. The compensation photon irradiance, E_C, was determined as the incident photon irradiance at which the leaf tissue shifted from a net O₂ consumption to a net O₂ production, i.e., the photon irradiance where P_n(E) = 0.

Depth profiles of O₂ concentration in the bulk sediment were obtained as follows. The sediment core was submerged into a ~2 L aquarium, wherein stirring and aeration of the water column was achieved via a Pasteur pipette connected to an air-pump. The surface of the sediment was determined with a boom-stand dissection microscope (AmScope, Irvine, CA, USA) and the O₂ microsensors were carefully positioned at the sediment surface as described above. Microprofiles were performed in vertical increments of 200 μm down to 2 cm depth, i.e., below the O₂ penetration depth. The volume specific O₂ consumption rate of the bulk sediment, R_sed (μmol O₂ m⁻³ s⁻¹), was calculated as:

where J_O2 is the O₂ flux at the seawater/sediment interface (μmol O₂ m⁻² s⁻¹), i.e., the diffusive oxygen uptake (DOU) of the sediment as calculated from Equation (1), and d_O2 is the O₂ penetration depth in the sediment (cm) as shown in Figure S1 (Supplementary Materials).

The O₂ consumption of the fine sediment particles used in the laboratory as well as in situ was determined using a slightly modified approach of Pedersen et al. (2011). The O₂ consumption was separated into total (OX_tot) or biological (OX_bio) O₂ demand in order to determine the chemical O₂ demand as OX_chem = OX_tot – OX_bio.

The total O₂ consumption of the sediment fraction was determined by mixing 50 mL suspended sediment (<63 μm) with 950 mL seawater with a salinity of 28. The solution was immediately transferred into 25 mL glass vials fitted with 2 glass beads to provide mixing and mounted on a rotating wheel (8 rpm) in a constant temperature bath (20.0 ± 0.5°C) (Pedersen et al., 2013). The sediment suspension was incubated for about 1 h (exact times recorded) before the O₂ concentration was measured in each vial using a calibrated sturdy O₂ microsensor (OX500; Unisense A/S, Denmark). Vials with seawater but without suspended sediment served as blanks enabling calculation of the O₂ consumption as μmol O₂ m⁻³ sediment s⁻¹.

The biological O₂ consumption was measured on a sediment suspension, which was initially purged with atmospheric air for 15 min to oxidize reduced metals and sulfide (Raun et al., 2010). After oxidation, the sediment suspension was transferred into 25 mL glass vials and treated as described above.

Two patches (~1 m in diameter) of Z. muelleri were enclosed by custom-made transparent, floating curtains with mixing provided by submerged pumps to simulate water motion outside the enclosures (Narrabeen Lagoon, Australia). One enclosure functioned as a control treatment and the other enclosure as a silt/clay treatment. In the silt/clay treatment, 3 pulses of 375 mL silt/clay particles (see above) were added to the water column per day to mimic a dredging operation. Sediment resuspension was initiated at the beginning of the experiments (afternoon) (pulse 1), just before sunrise (pulse 2) and at midday (pulse 3). Measurements were performed on April 17, 2015 (Series 1) and repeated on April 19, 2015 (Series 2), i.e., there were 27 h difference between Series 1 and Series 2 measurements. Within the enclosures, we measured salinity, light, temperature and O₂ in the water column during measurements of meristematic tissue O₂ and H₂S concentrations. A detailed description of the in situ measurements is given below.

Similar data acquisition equipment and microsensor as described above were used for the field measurements of internal O₂ partial pressure (pO₂) and H₂S concentrations ([H₂S]) in the meristematic tissue of Z. muelleri over diel cycles. Internal H₂S concentrations were measured with Clark-type H₂S microsensors (H2S-25, tip diameter of ~25 μm, 90% response time <10 s, detection limit ~0.3 μm, Unisense A/S, Aarhus, Denmark; Jeroschewski et al., 1996; Kühl et al., 1998) that were linearly calibrated in anoxic, acidic (pH 4) Na₂S solutions of known H₂S concentrations (0, 50, and 100 μM). Within the enclosures, the microsensors were mounted on micromanipulators that were supported by stabilized aluminum spears at a water depth of ~1 m. The O₂ and H₂S microsensors were simultaneously inserted into the briefly-exposed shoot base of the target plants close to the basal leaf meristem, which was then re-buried ~2 cm into the sediment to re-establish the biogeochemical gradients (Pedersen et al., 2004). Positioning of the O₂ microsensors was done by observing the sensor signals during insertion until a constant signal was recorded (Borum et al., 2005). The H₂S microsensors were inserted via a similar approach, using a combination of sensor signal responses to light exposure and positioning the electrodes at approximately the same depth into the leaf meristem tissue as the O₂ microsensors. The intra-plant O₂ and H₂S concentrations were measured simultaneously inside one plant in the control treatment and one plant in the silt/clay treatment, and then replicated.

Diel changes in ambient incident photon irradiance (continuously measured via Odyssey light loggers; Dataflow Systems, Christchurch, NZ), water-column pO₂ (via O₂ micro-optodes; OXF500PT, Pyroscience, Aachen, Germany; connected to a 4-channel Firesting meter, PyroScience, Germany), and water-column temperature (via HOBO temperature data loggers; UA-002-08, Onset Computer Corporation, Bourne, MA, USA) were recorded over ~24 h within the enclosures. All sensors were calibrated according to the manufactures instructions, mounted on a metal spear and positioned at leaf canopy height. Logging (1 Hz) by all data loggers was synchronized with the logging of microsensors used for the intra-tissue measurements.

All microsensors are temperature sensitive (e.g., Kühl and Revsbech, 2001) and thus the measurements of internal pO₂ and [H₂S] obtained by the calibrated O₂ and H₂S microsensors were temperature corrected using the following equations (available at www.unisense.com):

where S_amb is the sensor signal measured in situ (mV), S_air is the calibration signal of the sensor determined at known partial pressure and temperature (e.g., 100% air saturation; in mV), Z is the zero current of the sensor measured at known partial pressure and temperature (i.e., 0% air saturation; in mV), P₀ is the known partial pressure used to define S_air (kPa), k is the temperature coefficient of the respective sensor (~0.02°C⁻¹; exact values for individual sensors can be provided by the manufacturer, www.unisense.com), T_cal is the known calibration temperature (°C), and T_amb is the ambient temperature (°C) continuously measured in situ.

where G is the slope of the calibration curve that represents the sensitivity of the sensor (μmol L⁻¹ mV⁻¹), S is the signal of the sensor (mV), S₀ is a constant that describes the zero current (μmol L⁻¹), k is the temperature coefficient of the respective sensor (~0.02°C⁻¹), T_cal is the known calibration temperature (°C), and T_amb is the ambient temperature (°C) continuously determined in situ.

These final sensor calibrations were done after the in situ experiments using the temperature data obtained in the respective enclosures by the submerged HOBO temperature data loggers (HOBO, Onset Computer Corporation, Bourne, MA, USA).

To enable comparison of sediment activity, we determined the O₂ demand and characteristics of the added silt/clay particles (<63 μm) and the bulk sediment without seagrass biomass. The O₂ was depleted within the upper 1.2 mm of the bulk sediment and the sediment remained anoxic with depth (Figure S1). The volume-specific O₂ consumption rate of the bulk sediment was estimated to 374 ± 33 μmol O₂ m⁻³ s⁻¹ (Table 1). In contrast, the fine sediment particles consumed 1319 ± 6 μmol O₂ m⁻³ s⁻¹ when taking both the biological and chemical O₂ demand into account. The biological O₂ demand of the silt/clay particles was 1254 ± 29 μmol O₂ m⁻³ s⁻¹ resulting in a chemical O₂ demand of 65 μmol O₂ m⁻³ s⁻¹ (Table 1). Hence, the chemical O₂ demand of the fine sediment particles can thus most likely be neglected.

Net photosynthesis rates increased with increasing incident photon irradiance for both plants with and without leaf silt/clay-cover (Figure 1; showing O₂ fluxes from/into leaves). Moreover, net photosynthesis rates were higher in control leaf segments (no silt/clay added) exposed to hypoxic water conditions, resembling water-column O₂ levels at sunrise, as compared to leaf segments kept in water at 100% air equilibrium (Table 2). Plants with leaf silt/clay-cover exhibited net O₂ consumption already at an incident photon irradiance of ~75 μmol photons m⁻² s⁻¹ owing to reduced light availability for leaf photosynthesis (Figure 1; Table 2). Net photosynthesis rates of the control plants were 3 to 5-fold higher under moderate photon irradiance (200 μmol photons m⁻² s⁻¹) as compared to plants with leaf silt/clay-cover (Table 2). During darkness, a constant diffusive O₂ influx across the leaf surfaces of both plants with and without leaf silt/clay-cover was observed (Figure 1). However, we found a reduction in the O₂ flux into the silt/clay-covered leaves of 28–35% as compared to leaves without silt/clay-cover (Table 2; measured at 100% and 40% air equilibrium, respectively).

During water-column hypoxia, the leaf silt/clay-layer impeded the diffusive O₂ supply resulting in almost anoxic conditions at the leaf tissue surface (~16 μmol O₂ L⁻¹) of plants with leaf silt/clay-cover. This substantially increased the risk of H₂S intrusion into the below-ground tissues during night-time as a result of inadequate internal aeration (Figure 1). The thickness of the DBL surrounding the leaves increased from ~200 μm to ~500 μm in the presence of the leaf silt/clay layer (Figure 2). This resulted in a reduction in the O₂ influx to the leaves from 484 ± 133 nmol O₂ m⁻² s⁻¹ in plants without leaf silt/clay-cover to 419 ± 145 nmol O₂ m⁻² s⁻¹ in plants with an inactivated leaf silt/clay-layer. When coated with a biologically active silt/clay layer, leaves exhibited a further reduction of the O₂ influx to 395 ± 102 nmol O₂ m⁻² s⁻¹ (Figure 2).

The silt/clay-cover on seagrass leaves resulted in a pronounced increase of the plants' compensation irradiance from 53 ± 7 μmol photons m⁻² s⁻¹ for control leaf segments to 145 ± 46 μmol photons m⁻² s⁻¹ for leaf segments with silt/clay cover, both kept in a water column at 100% air equilibrium (Figure 3; Table 3). In a water column with O₂ kept at 40% atmospheric equilibrium, the compensation irradiance increased from 20 ± 8 μmol photons m⁻² s⁻¹ for control leaf segments to 109 ± 47 μmol photons m⁻² s⁻¹ for leaf segments with silt/clay cover (Figure 3; Table 3). The leaf silt/clay-layer effects on plant photosynthesis and respiration lead to a ~2.4-fold increase in the irradiance causing onset of net photosynthesis saturation for plants with leaf silt/clay-cover as compared to plants without leaf silt/clay-cover (Table 3), and to a 49–72% reduction of the leaf surface O₂ concentration in darkness for plants with a leaf silt/clay-cover as compared to plants without a leaf silt/clay-cover (Table 3).

The pO₂ dynamics in the water-column of the control and silt/clay treatment showed similar patterns on a diel basis, with steadily declining pO₂ during night-time reaching minimal water-column O₂ conditions around sunrise, followed by a rapid increase in the water-column pO₂ shortly after sunrise approaching atmospheric saturation (20.6 kPa) or even leading to water-column supersaturation relative to atmospheric pO₂ around midday (Figures 4A,B). Water-column O₂ levels within the enclosures fluctuated substantially during night-time owing to water bodies with varying O₂ content being introduced to the seagrass meadow from non-vegetated areas within the lagoon and/or from the ocean due to tidal water movement. In contrast, water-column temperature remained relatively constant on a diel basis but generally decreased from ~22°C on the first measuring day (Series 1) to ~20°C at the end of the second measuring day (Series 2). Minor fluctuations in the water-column temperature during night-time correlated with the passing of aerated water bodies as observed in the water-column pO₂ measurements (Figures 4A,B). The incident photon irradiance measured at leaf canopy height followed a typical bell-shaped diel curve, with minor fluctuations in the control treatment due to passing cloud cover. This was in strong contrast to the silt/clay treatment, where we measured substantially reduced light conditions as compared to the control treatment, especially in the hours following experimentally manipulated silt/clay re-suspension (Figures 4A,B). Moreover, a pronounced difference in the light availability was observed between measuring days Series 1 and Series 2, where Series 1 represented sunny conditions and Series 2 represented a cloudy late autumn day at Narrabeen Lagoon (Figures 4A,B).

The internal, meristematic pO₂ of both control plants and plants experimentally exposed to suspended silt/clay decreased steadily from early in the afternoon throughout the night. A minimum internal, meristematic pO₂ was reached shortly after sunrise. Thereafter, a rapid increase in meristematic pO₂ occurred as a response to increasing solar irradiance resulting in photosynthetic O₂ production (Figures 4C,D). Control plants as well as silt/clay-treated plants exhibited lower pO₂ relative to the water-column during night-time with tissue pO₂ fluctuations correlating with changes in water-column pO₂ (Figures 4A–D). A clear discrepancy in the meristematic pO₂ between control plants and leaf silt/clay-treated plants was measured during light-limitation in the early morning hours (06:30–09:00) (Figure 4C) with relatively lower pO₂ in silt/clay-treated plants indicating a silt/clay-induced reduction in light availability.

The meristematic below-ground tissues of both control and silt/clay-treated plants turned anoxic, or severely hypoxic, late at night. Meristematic pO₂ of silt/clay-treated plants reached anoxia from around 05:00–06:30 in Series 1 and already from 23:30 in Series 2, while the control plants only were exposed to anoxic conditions in the meristematic tissue for short time periods (<1 h; Figures 4C,D). Simultaneous measurements of internal, meristematic H₂S concentrations revealed phytotoxic H₂S intrusion into silt/clay-treated plants during night-time in Series 2 from around 23:30 correlating with the recorded period of meristematic tissue anoxia (Figures 4C,D). Internal H₂S levels reached a maximum of 8.3 μmol H₂S L⁻¹ around 08:00 in the morning and then started to decrease shortly after sunrise in response to photosynthetic O₂ production leading to disappearance of H₂S in the meristem by 10:30. No H₂S intrusion was detected into the control plants.

During night-time, tissue pO₂ was derived from O₂ in the surrounding water diffusing into the leaves and spreading via aerenchyma to below-ground tissues (Pedersen et al., 1998; Colmer, 2003; Brodersen et al., 2015a). The critical water column O₂ level was defined as the water column pO₂ below which oxic conditions in the meristematic tissue could no longer be sustained, and this critical O₂ level was estimated by plotting the internal pO₂ determined in situ against water-column pO₂ (Figure 5). In Series 1, the meristematic tissue of the silt/clay-treated plant became anoxic at a water-column pO₂ of ~5.5 kPa during night-time as compared to ~8.7 kPa in the control plant (Figures 5A,C); a tendency that dramatically changed during prolonged exposure to suspended silt/clay particles (i.e., in Series 2) where the silt/clay-treated plant became anoxic already at a night-time water-column pO₂ of ~13 kPa as compared to ~6.4 kPa in the control plant (Figures 5B,D). These in situ findings aligned well with the lower O₂ influx into leaves with silt/clay-cover, as compared to control leaves, determined in the controlled laboratory experiments during darkness (Figures 1–3; Tables 2,3).

The silt/clay-induced shading effects on the intra-plant pO₂ during natural light exposure of the seagrass leaf canopy was evaluated by plotting the in situ meristematic pO₂ as a function of incident photon irradiance (Figure 6) revealing an ~45% reduction in meristematic pO₂ in plants exposed to suspended silt/clay as compared to control plants, seen as a decrease in α, i.e., the slope describing the internal O₂ evolution as a function of photon irradiance, from 0.14 to 0.08 (Figure 6).