The interaction of CO₂ concentration and spatial location on O₂ flux and mass transport in the freshwater macrophytes Vallisneria spiralis and V. americana

The ecophysiological processes of aquatic organisms are diverse, and the form and rate of these processes vary within and among species(Phillips and Hurd, 2004; Nishihara et al., 2005; Badgley et al., 2006). In the case of aquatic macrophytes, variations in processes such as nutrient uptake and photosynthesis are influenced by abiotic and biotic conditions (e.g. nutrient history, nutrient concentration, water flow, temperature and light)(Hurd et al., 1996; Nishihara et al., 2005; Badgley et al., 2006). The kinetics of these mass transfer and uptake processes can be linear or nonlinear and can differ among nutrients and species. For example, in the red alga Laurencia brongniartii, the uptake kinetics were linear for nitrate and nonlinear for ammonium(Nishihara et al., 2005),whereas nonlinear kinetics were observed for both nutrients in the brown alga Scytothamnus australis (Phillips and Hurd, 2004). Indeed, physiology ultimately determines the rate of such processes; however, the transport of nutrients to the surface of aquatic macrophytes governs whether the process is kinetically or mass transfer limited (Sanford and Crawford,2000; Nishihara and Ackerman,2006).

Dissolved inorganic carbon (DIC) is available to aquatic macrophytes in the form of CO and HCO₃^–(Madsen and Sand-Jensen,1991). Whereas HCO₃^– is the primary form of DIC available in marine environments due to relatively high pH, the dominant form of DIC in freshwater varies with ambient pH. However, in cases where alkalinities are low, freshwater macrophytes have greater access to dissolved CO₂. Increases in CO₂ concentrations are,therefore, thought to benefit freshwater and marine macrophytes capable of HCO₃^– uptake that have a higher affinity (i.e. sensitivity) for CO₂ (Madsen and Sand-Jensen, 1991; Invers et al., 2001). In general, DIC uptake (i.e. photosynthesis) can be described by a rectangular hyperbola, where for low DIC, photosynthesis rates are directly proportional to the DIC concentration, and for large DIC,photosynthesis rates are saturated(Maberly and Madsen, 1998; Invers et al., 2001; Nishihara and Ackerman, 2006). The delivery of DIC to supply photosynthesis occurs through diffusive and advective transport, and differences in rates occur as a result of the different concentrations and diffusivities of CO₂ and HCO₃^–. Indeed, in calm water, diffusive transport can serve as the primary mechanism by which DIC is transported to a macrophyte, whereas in flowing water advective transport becomes the dominant mechanism of mass transport. Increased DIC mass transfer rates resulting from advection can also supply proportionately more DIC in the form of CO₂ (Nishihara and Ackerman,2006). Therefore, the ecophysiology of aquatic macrophytes is influenced by the mass transport of nutrients, and in conditions where water velocities and ambient concentrations are low, productivity can be limited by mass transport (Stevens and Hurd,1997; Cornelisen and Thomas,2006; Nishihara and Ackerman,2006).

Mass transfer rates are a function of the freestream velocity (U),the molecular diffusivities (D) of DIC, the concentration gradient(ΔC=C_S–C_B) between the bulk water (C_B) and the surface concentrations(C_S), and the geometry of the macrophyte. An increase in U or ΔC will increase mass transfer rates and thus provide more DIC, leading to higher rates of photosynthesis(Nishihara and Ackerman,2006). Mass transfer rates may vary with location, as in the case of a flat plate oriented parallel to the flow, where mass transfer rates decrease monotonically with increasing downstream direction(Schlichting and Gersten,2000). Whether the mass transfer rates of nutrients vary spatially over the surface of a macrophyte, and thus influence physiological processes,remains to be determined.

The freshwater angiosperms, Vallisneria spiralis L. and Vallisneria americana Michx., are found throughout Europe and North America, respectively, and occupy similar niches in their respective environments (Sculthorpe,1967). Both species can assimilate HCO₃^– (Prins et al., 1980; Madsen and Sand-Jensen, 1991) and can be found in waters high in alkalinity(i.e. 100–1400 mmol m^–3HCO₃^–) (Pip,1984). More importantly, a key morphological trait that differentiates the two is the spiral twist found in V. spiralis(Fig. 1A), unlike the relatively flat leaves of V. americana(Fig. 1B). Little is known about the function of the twist; however, morphological features, which present a more complex geometry than that of a flat surface, have been found to enhance mass transport and physiological processes in other aquatic organisms (Hurd et al., 1996; Falter et al., 2005). V. spiralis and V. americana provide an excellent opportunity to determine whether and how physiological processes are influenced by physical and environmental factors in two closely related species. The objective of this study is, therefore, to determine: (1) how the O₂ flux of V. spiralis and V. americana is influenced by CO₂concentrations and velocity; (2) how O₂ flux and mass transfer rates vary with respect to spatial location; and (3) how a flat and twisted leaf morphology affects O₂ flux and mass transport.

Vallisneria spiralis L. and V. americana Michx. (obtained from Boreal Laboratories) were grown in 20 l aquaria, using nutrient enriched wellwater at 25°C and a soil substrate(Nishihara and Ackerman,2006). The photosynthetically active radiation (PAR) level was maintained at 7.3 μmol photon m^–2 s^–1 (16 h:8 h L:D), as measured in the center of the aquaria with a 4π sensor(QSL2101, Biospherical Instruments, San Diego, CA, USA), to discourage the growth of microalgae and other potential fouling organisms.

The experimental setup is described in detail elsewhere(Nishihara and Ackerman, 2006; Nishihara and Ackerman, 2007). Briefly, a 10 cm×10 cm×100 cm long (water depth: 5–8 cm)flow chamber, with flow straighteners in the first 12 cm, was operated at freestream velocities of 0.5, 0.8, 1.1, 1.8, 2.1, 3.3, 4.1 and 6.6 cm s^–1. Velocity profiles in the empty flume were determined using digital particle image velocimetry (PIV) at the location of the leading edge of the leaf (56 cm downstream of the flow straighteners), and were uniform in shape, especially at the height where the O₂ profiles were obtained above leaf surfaces (3 cm above the flume bottom)(Nishihara and Ackerman,2006). The flow chamber water was maintained at 24°C, and aerated to maintain O₂ and CO₂ saturation, using a mixture of tapwater [pH=9.2±0.2 (mean ± s.e.m.)] and deionized water having a final HCO₃^– concentration of 460 mmol m^–3. PAR (153 μmol photon m^–2s^–1) was provided by a slide projector with a quartz lamp(General Electric, Fairfield, CT, USA) adjusted with neutral density filters. The bulk O₂ concentration and O₂ profiles in the CBL were determined using OXN and OX25 oxygen microsensors (Unisense, Aarhus,Denmark), respectively. The O₂, pH and water temperature were recorded continuously by a computer.

V. spiralis and V. americana leaves were selected from the culture that were: (1) free of obvious epiphytes; (2) at least 8 cm long;(3) twisted only once in a 8 cm span in the case of V. spiralis; and(4) without undulations along the length and width of the leaf in the case of V. americana. The leaves were trimmed to size (8 cm in length), the cut ends were sealed with wax, and allowed to acclimate overnight in the flow chamber water. Leaf sections were glued (cyanoacrylate-based adhesive) to a wire stand, and placed in the working section of the flow chamber,perpendicular to the light source, and parallel to the flow. After each experiment, images of sections of the leaf approximately 1 cm² from the upstream and downstream locations (i.e. 1 cm and 7 cm from the leading edge of the leaf, respectively) were taken and the samples were frozen(–20°C) for later chlorophyll a+b analysis. Chlorophyll_a+b content was determined as described(Nishihara and Ackerman,2006). These two closely related species were chosen in part because their physiologies were assumed to be similar.

The effect of the experimental factors (i.e. species-leaf configuration,upstream vs downstream measurement location on the leaf surface,CO₂ concentration and velocity) on O₂ flux, mass transport and the thickness of the concentration boundary layer(δ_CBL) were investigated by profiling the O₂concentration above the leaves (z=0–0.5 cm) of each leaf at x=1 and 7 cm downstream from the leading edge of the leaf (i.e. the upstream and downstream locations). The CO₂ concentration was adjusted to 1.71 mmol m^–3 and 17.1 mmol m^–3by adding 50 mmol m^–3 Tris buffer and appropriate amounts of HCl (i.e. to obtain a pH of 7.5 and 8.5, respectively) to the water in the flow chamber (Stumm and Morgan,1996). During the course of the experiments pH and HCO₃^– concentrations remained stable, regardless of aeration, and no negative effects of the buffer on photosynthesis were observed. At least 5 min elapsed prior to determining the O₂profiles for each species-leaf configuration, measurement location, and CO₂ concentration at all U and a randomized block design was used to minimize the effects of acclimation time. A total of nine V. spiralis and six V. americana leaves were analyzed at 1.71 mmol m^–3 CO₂ and six V. spiralis and six V. americana leaves were analyzed at 17.1 mmol m^–3CO₂. Eight V. americana leaves were also reconfigured from a flat to twisted configuration by gently twisting the leaf to mimic V. spiralis, and maintaining the abaxial surface to the O₂microsensor. In this case, experiments were conducted at 17.1 mmol m^–3 CO₂ at the downstream location in the flat and twisted configurations.

Statistica 6.1 (Statsoft, Inc., Tulsa, OK, USA) and R (R Development Core Team, 2006) were used to analyze the data. Linear and nonlinear regressions were used as appropriate, and an F-test was used to determine whether these were significant, ANCOVA was used to determine whether differences in slopes of the linear regressions were significant, and Tukey's test was used to examine multiple comparisons. For regression analyses of Sh_x and δ_CBL, determined from O₂ profiles, log(Re_x) was the covariate and species-leaf configuration (i.e. V. spiralis, V. americana and V. americana in the twisted configuration), position and CO₂concentration were the factors. In the case of PIV determined Sh_x and δ_DSL,log(Re_x) was the covariate and species-leaf configuration and measurement location were the factors. Significance was defined at P=0.05 and values are reported as mean ± s.e.m., except for the nonlinear regressions where s.e.m. is asymptotic.

The O₂ flux at the upstream and downstream locations of V. spiralis and both leaf configurations of V. americana increased monotonically with increasing velocity at 17.1 and 1.71 mmol m^–3 CO₂ (Fig. 2). The O₂ flux was also similar to those previously determined for V. americana (dotted line in Fig. 2)(Nishihara and Ackerman, 2006)and is comparable to other studies involving aquatic macrophytes(Madsen and Sand-Jensen,1991).

The relationship between the O₂ flux and velocity, determined at 17.1 mmol m^–3 CO₂, can be described by a rectangular hyperbola for both species and measurement locations, where J_max is the maximum O₂ flux and V is the half-saturation velocity (Fig. 2A; Table 1). An F-test revealed that there was no significant difference(F_(30,34)=0.43, P=0.79) between a common V model (unique J_max) and the full model (unique V and J_max for each species and position),therefore the common V model (r²=0.32, F_(6,34)=3.94, P=0.0064) was applied to describe O₂ flux at 17.1 mmol m^–3 CO₂(Fig. 2A and Table 1). Overall,O₂ flux was higher for V. spiralis compared to V. americana. J_max determined from the model was greatest for V. spiralis (P<0.05); however, there was no measurement location effect (P>0.05) within species. V_satat 17.1 mmol CO₂ was 4.5±1.2 cm s^–1, which was similar to an earlier investigation of V. americana(V_sat=4±1 cm s^–1 at 460 mmol m^–3 DIC) (Nishihara and Ackerman, 2006).

In contrast, the relationship between the O₂ flux at 1.71 mmol m^–3 CO₂ and U varied with location along the leaf, but not with species (Fig. 2B). O₂ profiles were difficult to determine at the higher velocities for V. spiralis and V. americana at this CO₂ concentration because the concentration gradient was too thin to profile. Specifically, eight downstream O₂ profiles were obtained at 4.1 and 6.6 cm s^–1 for V. spiralis,compared to V. americana where five profiles were obtained at 3.3 and 4.1 cm s^–1 and none at 6.6 cm s^–1 at the upstream location; only four profiles at 4.1 cm s^–1 and two profiles at 6.6 cm s^–1 were obtained at the downstream location. Nevertheless, the O₂ flux at the upstream location was greater than the downstream location when U<V_sat=1.8±0.9 cm s^–1(Fig. 2B). The O₂flux was nonlinear at the upstream location, and Eqn 3 was fit to the data. An F-test revealed that the most appropriate model for this location was a reduced form (common V and J_max in Eqn 3; r²=0.26, F_(2,13)=2.94, P=0.026), since neither a full model nor a common V model was significant (F_(11,13)=0.26, P=0.79 and F_(11,12)=0.48, P=0.50, respectively)(Table 1). In contrast, at the downstream location, O₂ flux appeared to increase linearly with U, and a linear model was applied to the data. In this case, an ANCOVA revealed that the slopes for the two species were not significantly different (F_(1,11)=0.23, P=0.64) and species had no effect on O₂ flux (F_(1,12)=0.023, P=0.88), therefore a common linear model was applied to these data(r²=0.67, F_(1,13)=29.61, P<0.0001) (Table 1).

The Sh_x for the O₂ profiles increased with Re_x for all species-leaf configurations, CO₂concentrations, and measurement locations(Fig. 3). Measurement location had a significant effect on the slope of the regressions(F_(2,58)=9.94, P=0.0002) and the individual regressions for each species-leaf configuration and measurement location were significant (r²>0.68, P<0.05; see Table 2A). There were no significant differences among slopes (i.e. logB) at the upstream location(P>0.05), which was similar to previous results for V. americana (0.45±0.04)(Nishihara and Ackerman, 2006)and for theoretical laminar mass transport over a flat plate (0.5)(Schuepp, 1993). Whereas the intercepts were not significant, they were much greater than the theoretical value for parameter A for a constant surface flux boundary condition in a laminar boundary layer [i.e. log(0.464Sc^–0.33)=–1.18(Schlichting and Gersten,2000)]. In contrast, the slopes at the downstream location were heterogeneous (P<0.05; Table 2A), although a multiple comparisons test, indicated that the slopes for V. spiralis (0.87±0.09) and V. americana(0.82±0.07) in the twisted configuration at 17.1 mmol m^–3 CO₂ were similar (P>0.05), as were V. americana at 17.1 mmol m^–3 CO₂(0.70±0.10) and V. spiralis at 1.71 mmol m^–3CO₂ (0.69±0.05). The slopes determined at the downstream location were of similar order to the theoretical value for mass transport in a turbulent concentration boundary layer over a flat plate (0.8)(Kays et al., 2005). The log-transformed intercepts (P>0.05) were not significantly different from zero; however, they were greater than predicted by a theoretical turbulent concentration boundary layer [i.e. log(0.030Sc^–0.33)=–2.37(Schlichting and Gersten,2000)].

The shear velocities determined from the PIV ranged from 0.040±0.002 to 0.7±0.1 cm s^–1, and the u^*values at the upstream locations were always greater than those measured at the downstream location at each velocity(Table 3). Moreover, u^* increased with freestream velocity for all species-leaf configurations and measurement locations. The Sh_xdetermined from hydrodynamic measurements (i.e. PIV results) also increased with increasing Re_x(Fig. 4). The slope of the Sh_x–Re_x regression did not vary with measurement location (F_(1,39)=0.21, P=0.65),which is in contrast to the Sh_x determined from the O₂ concentration boundary layer. However, there was a species-leaf configuration effect (F_(2,40)=4.53, P=0.017),where the slope for the regressions at the downstream location of V. americana in the twisted configuration was significantly different from all the other slopes (P<0.05), except for the slope determined at the downstream location of V. spiralis(Table 2B). The intercepts determined at the upstream locations were significantly different from zero(P<0.05), in contrast to the downstream locations(Table 2). Regardless, these values were also larger than that predicted by the turbulent concentration boundary layer.

The δ_CBL determined from the O₂ profiles and the δ_DSL determined from the PIV measurements decreased with increasing velocity and Re_x for both species-leaf configuration and measurement location(Fig. 5). Whereasδ _CBL determined from the O₂ profiles appeared similar in shape and thickness at both locations(Fig. 5A,C), theδ _CBL was thinner at the upstream location when normalized to distance from the leading edge of the leaf (i.e. plotted vs Re_x) (Fig. 5B,D). ANCOVA revealed that measurement location affected the rate of change of the thickness of the δ_CBL with respect to Re_x [i.e. log(δ_CBLx^–1Sc^–0.33)varied with Re_x] and significant differences were detected between measurement locations (F_(2,64)=8.53, P<0.0001). The slopes of the regressions at the upstream location were homogeneous (P>0.05), whereas they were heterogeneous at the downstream location (P<0.05; Table 4). The slopes of V. spiralis at both CO₂ concentrations were similar(P>0.05), as were both leaf configurations of V. americana (P>0.05) at both CO₂ concentrations(Table 4). There was some overlap between the species-leaf configurations, where the slopes V. spiralis at 1.71 mmol m^–3 CO₂ were similar to those of V. americana and the slopes of V. americana at 1.71 mmol m^–3 CO₂ were similar to those of V. spiralis. The δ_CBL among species-leaf configuration,measurement location and CO₂ concentration were similar(P>0.05); however, the average δ_CBL values at both measurement locations were thinner at the upstream (48±2%) and downstream locations (21±1%) compared to the theoreticalδ _CBL (based on Eqn 7).

The diffusive sublayer thickness (δ_DSL) determined from Eqn 8, using the measured shear velocities, was similar in thickness to theoretical estimates of theδ _DSL determined from the 1/7 power law(White, 1999) at the leading edge (106±6% of the theoretical δ_DSL; Fig. 5E) whereas at the trailing edge the δ_DSL was 51±2% of the theoreticalδ _DSL (Fig. 5E,F).

The rates of physiological processes were found to vary among closely related aquatic angiosperm species and the rates were influenced by mass transport. The complexity of these interactions is evident in this study,where the effects of CO₂ concentration, freestream velocity and measurement location along a leaf surface affected the photosynthesis rates of V. spiralis and V. americana(Fig. 2). The effects of increasing CO₂ concentrations on photosynthesis are well documented, but the spatial variation is not(Invers et al., 2001; Lobban and Harrison, 1994; Schippers et al., 2004; Nielsen et al., 2006). In this study, O₂ flux was similar in both species at low CO₂concentrations. However, increasing the CO₂ concentration only enhanced the photosynthesis rates of V. spiralis, with no effect on V. americana. It is evident that V. spiralis is better able to respond to the higher CO₂ concentrations and increase its O₂ flux.

The difference in O₂ flux between V. spiralis and V. americana is likely due to physiological(Fig. 2A) rather than morphological differences, given that the twisted configuration of V. americana did not enhance O₂ flux through increases in the mass transfer coefficient (see below). Moreover, given that DIC was held constant and that the HCO₃^– concentrations at both CO₂ concentrations were similar (i.e. 460 mmol m^–3HCO₃^–), the increase in O₂ flux with increases in CO₂ concentration implies that the flux of CO₂ is greater for V. spiralis than V. americana. Regardless of the ability to use both CO₂ and HCO₃^– (Prins et al., 1980; Madsen and Sand-Jensen, 1991), the enhancement of O₂ flux in V. spiralis suggests differences in CO₂ affinity between the two species, namely that V. spiralis has a higher affinity for CO₂. Such variations in CO₂ affinity are not uncommon,and in general freshwater macrophytes that depend on CO₂ as the sole source of carbon have the highest affinity for CO₂(Madsen and Sand-Jensen,1991). Freshwater macrophytes such as V. spiralis and V. americana, which are able to use HCO₃^–, are intermediate in their affinity for CO₂, whereas marine macrophytes that exist in relatively high pH waters have the lowest CO₂ affinity(Madsen and Sand-Jensen,1991).

As already mentioned, the effect of low CO₂ concentrations was similar between the species; however, the relationship between O₂flux and velocity varied spatially. Specifically, O₂ flux was saturated at V_sat >1.8±0.9 cm s^–1 at the upstream location, but O₂ flux continued to increase with velocity at the downstream location(Fig. 2B). This linear relationship suggests that photosynthesis is mass transfer limited at the downstream location. However, O₂ flux did not vary spatially at the high CO₂ concentration, where the O₂ flux saturated at both measurement locations for both species(Fig. 2A). Evidently, the behavior of nutrient uptake varies spatially on leaf surfaces at low nutrient concentrations. Spatial heterogeneity in mass transport has not been incorporated in current models that couple physiology and mass transport processes (e.g. Sanford and Crawford,2000). Although, interactions of nutrient concentrations and mass transport have been observed in the nutrient uptake rates of seagrasses(Cornelisen and Thomas, 2006),marine algae (Hurd et al.,1996) and corals (Badgley et al., 2006), and in photosynthesis rates of marine macrophytes(Wheeler, 1980), periphyton(Larkum et al., 2003) and V. americana (Nishihara and Ackerman, 2006), little is known about how freestream velocities affect the spatial variability in nutrient uptake.

The saturating behavior of O₂ flux at the upstream location and the linearly increasing behavior of O₂ flux at the downstream location suggest that upstream uptake and associated decrease of nutrients in the concentration boundary layer impact physiological processes occurring downstream (e.g. Chambré and Acrivos, 1956; Ackerman et al.,2001). The mechanism(s) responsible for the downstream decrease in O₂ flux observed in this study has yet to be determined; however,it is likely that the surface concentrations of nutrients decrease with downstream distance from the leading edge as a result of nutrient depletion in the concentration boundary layer.

The consistency of parameter B in the relationship between Sh_x and Re_x at the upstream location(0.47–0.52) demonstrates that the concentration boundary layer was laminar at that location (Fig. 3, Table 2)(Schlichting and Gersten,2000). Conversely, parameter B differed significantly among the downstream locations (0.69 to 0.95; Fig. 3, Table 2), but none of these values were significantly different (P>0.05) from the theoretical value of mass transport in a turbulent concentration boundary layer (B=0.8) (Schlichting and Gersten,2000). Based on these observations and those of an earlier study(Nishihara and Ackerman,2006), it is apparent that the concentration boundary layer of V. spiralis and V. americana shifts from a laminar to turbulent regime between Re_x of 700 to 1500. These values are considerably lower than the transitional velocity for momentum over a smooth flat plate (Re_x=3×10⁵)(Schlichting and Gersten,2000), but are similar to mass transport phenomena in terrestrial plant leaves in a turbulent freestream (e.g. Re_x=1860)(Schuepp, 1993).

Given that surface corrugations on a flat surface were demonstrated to increase mass transfer rates in engineering applications(Tzanetakis et al., 2004), the twisted morphology of V. spiralis was predicted to enhance mass transport. However, the similar shape obtained by twisting V. americana did not enhance mass transfer rates(Fig. 3A,B). This is consistent with the observation that undulations in the blades of the marine macrophyte Macrocystis integrifolia did not to enhance inorganic nitrogen uptake(Hurd et al., 1996). However,the assimilation of DIC has been recently reported to be related to plant architecture in aquatic angiosperms(Nielsen et al., 2006) and similar morphological effects on nutrient uptake have been reported in corals(Lesser et al., 1994; Helmuth et al., 1997). It is possible that the effect of the twist in V. americana was not detected, due to limitations in the microsensor technique (see below). Clearly, the influence of morphological features in mass transport remains an equivocal issue (cf. Thomas and Atkinson,1997).

Concentration boundary layers can be estimated through direct measurements of the scalar (e.g. O₂) profile, or indirectly by measuring the hydrodynamic boundary layer and multiplying by a scaling factor (e.g. Sc^–0.33) (Dade,1993). It is common to describe the thickness of the concentration boundary layer that forms around the surface of macrophytes and refer to it as`boundary layer resistance' to mass transport(Wheeler, 1980; Stevens and Hurd, 1997; Larkum et al., 2003). In this case, the δ_CBL is given byδ _CBL=J/(C_S–C_B),by assuming that the surface concentration is zero (i.e. the surface is a perfect sink) (Wheeler, 1980; Stevens and Hurd, 1997; Larkum et al., 2003). However,the δ_CBL determined by this method, may overestimateδ _CBL (Stevens and Hurd,1997; Larkum et al.,2003) as previously demonstrated(Nishihara and Ackerman,2006), where the δ_CBL values determined from O₂ profiles were >63% smaller than the δ_CBLvalues determined by using the boundary layer resistance model. These observations suggest that the two assumptions (i.e. spatially homogeneous flux and a perfect sink condition) invoked to determine the δ_CBLvalues may be inappropriate. For example, saturating nutrient uptake rates at high velocities indicate kinetic limitation (i.e. mass transfer rates >nutrient uptake rates), which would occur if the nutrient is in excess and the surface concentration is >0 (Nishihara and Ackerman, 2006). It is apparent that indirect estimates of the concentration boundary layer from hydrodynamic theory overestimate mass transfer rates in aquatic macrophytes.

The importance of direct measurements of the concentration boundary layer led to the application of O₂ microsensors in boundary layer research, such as those used in this and other studies(Glud et al., 1994; Larkum et al., 2003; Nishihara and Ackerman, 2006; Nishihara and Ackerman, 2007). Evidence suggests that microsensors directly affect the concentration boundary layer and revealed that the concentration boundary layer, determined by theδ _DSL, was reduced to 55–75% of the theoreticalδ _DSL (Glud et al.,1994). It is difficult to assess whether the decrease in the concentration boundary thickness was caused by the microsensor, given that the hydrodynamic boundary layer was not characterized in that study(Glud et al., 1994). However,it has been suggested that the microsensor has little impact on the concentration boundary layer at low Reynolds number (Re_d)of the microsensor [∼6 (Hondzo et al.,2005)], which ranged from 0.14 to 1.8 in this study. Therefore,near the surface of the leaf, where velocities are lower than the freestream velocity (i.e. viscous flow, Re_d <1), the flow remains attached to the microsensor. Nevertheless, δ_CBL measured with the microsensors were ⩽48% of the theoretical δ_CBL at the upstream location, which is less that what is predicted by microsensor-induced compression (Glud et al.,1994). Therefore, the thinner δ_CBL may also result from the influence of biological and chemical processes on O₂ flux and concentration boundary layer thickness(Nishihara and Ackerman,2006). Given the difficulties in measuring the O₂profiles at high velocities (i.e. >3.3 cm s^–1 in this study) and the possible compression effect of the microsensor, the technique used in this study may be too coarse to resolve differences between the flat and twisted leaf morphologies at moderate to high velocities.

The interaction between physiology and mass transport among species is complex, and elucidating the mechanisms underlying these processes should provide insight on how environmental factors influence the biology of aquatic organisms. In addition, the diversity in these interactions makes it difficult to predict how environmental alterations including climate change will affect aquatic ecosystems. Presently, global CO₂ levels are increasing,leading to the acidification of freshwater and marine ecosystems(Harley et al., 2006). Macrophytes that are physiologically limited by present CO₂ levels may see a dramatic increase in productivity(Schippers et al., 2004). However, in marine systems, where the majority of macrophytes have low CO₂ affinity (Madsen and Sand-Jensen, 1991), the effects are likely to be small except in the limited number of marine angiosperms (i.e. seagrasses)(Invers et al., 2001). Furthermore, coupled with increased advective transport, the increase in CO₂ mass transfer rates could provide a mechanism for macrophytes with high CO₂ affinity to displace species (e.g. V. americana) that are less sensitive to increases in CO₂. In areas where water velocities promote high mass transfer rates, small increases in CO₂ could have a significant impact on the distribution of aquatic macrophytes. However, in marine systems, where the majority of macrophytes have low CO₂ affinity(Madsen and Sand-Jensen,1991), the effects are likely to be small. Clearly, further research is required to elucidate the effects of mass transport on ecophysiological processes.

Physiological and mass transport mechanisms are coupled and in biological systems such as aquatic macrophytes these processes vary spatially. The O₂ flux was higher in V. spiralis than in V. americana when CO₂ concentrations were high, but were similar at the lower CO₂ concentration. Importantly, the O₂ flux varied spatially on the leaf surface at low CO₂ concentrations,where O₂ flux (i.e. photosynthesis) was kinetically limited at the upstream location and mass transfer limited at the downstream location. Further studies are needed to elucidate the effects of morphology on mass transport and the mechanisms underlying the spatial heterogeneity of O₂ flux and mass transfer rates. Mass transport relationships must be considered to properly evaluate how changes in environmental conditions affect the productivity of aquatic ecosystems.

List of abbreviations and symbols

 
• δCBL

  concentration boundary layer thickness (m)

 
• δDSL

  diffusive sublayer thickness (m)

 
• ΔC

  concentration gradient (mmol m^–3)

 
• θ

  dimensionless concentration

 
• κ

  von Karman constant

 
• ν

  molecular diffusivity of momentum (m²s^–1)

 
• ρ

  density of water (kg m^–3)

 
• τ

  boundary shear stress (Pa)

 
• a,b

  constants to Eqn 1(m^–1)

 
• A,B,C

  constants to Eqn 6

 
• C

  nutrient concentration (mmol m^–3)

 
• C_S

  surface concentration (mmol m^–3)

 
• C_B

  bulk water concentration (mmol m^–3)

 
• CBL

  concentration boundary layer

 
• chl_a+b

  total chlorophyll a + b concentration

 
• DSL

  diffusive sublayer

 
• D

  molecular diffusivity (m² s^–1)

 
• DIC

  dissolved inorganic carbon

 
• J

  O₂ flux (mmol m^–2s^–1)

 
• J_max

  maximum O₂ flux (mmol m^–2s^–1)

 
• J_chla+b

  Chlorophyll a+b normalized O₂ flux [mmol m^–2 s^–1 (g_chla+bm^–2)^–1]

 
• k_c

  mass transfer coefficient (m s^–1)

 
• PAR

  photosynthetically active radiation (μmol photon m^–2 s^–1)

 
• P_max

  maximum oxygen flux [mmol m^–2 s^–1(g_chla+b m^–2)^–1]

 
• PIV

  particle image velocimetry

 
• Re_d

  microsensor Reynolds number

 
• Re_x

  local Reynolds number

 
• Sc

  Schmidt number

 
• Sh

  _x local Sherwood number

 
• u

  ^* shear velocity (m s^–1)

 
• U

  freestream velocity (m s^–1)

 
• V

  half-saturation velocity (m s^–1)

 
• V

  _sat saturation velocity (m s^–1)

 
• x

  distance from the leading edge of the leaf (m)

 
• z

  height above leaf surface (m)

The authors would like to thank Patrick Ragaz for assistance with the PIV. This research was supported in part by funding from the University of Guelph and the Natural Sciences and Engineering Research Council of Canada to J.D.A.