A perspective on underwater photosynthesis in submerged terrestrial wetland plants.

BACKGROUND AND AIMS: Wetland plants inhabit flood-prone areas and therefore can experience episodes of complete submergence. Submergence impedes exchange of O(2) and CO(2) between leaves and the environment, and light availability is also reduced. The present review examines limitations to underwater net photosynthesis (P(N)) by terrestrial (i.e. usually emergent) wetland plants, as compared with submerged aquatic plants, with focus on leaf traits for enhanced CO(2) acquisition. SCOPE: Floodwaters are variable in dissolved O(2), CO(2), light and temperature, and these parameters influence underwater P(N) and the growth and survival of submerged plants. Aquatic species possess morphological and anatomical leaf traits that reduce diffusion limitations to CO(2) uptake and thus aid P(N) under water. Many aquatic plants also have carbon-concentrating mechanisms to increase CO(2) at Rubisco. Terrestrial wetland plants generally lack the numerous beneficial leaf traits possessed by aquatic plants, so submergence markedly reduces P(N). Some terrestrial species, however, produce new leaves with a thinner cuticle and higher specific leaf area, whereas others have leaves with hydrophobic surfaces so that gas films are retained when submerged; both improve CO(2) entry.
CONCLUSIONS: Submergence inhibits P(N) by terrestrial wetland plants, but less so in species that produce new leaves under water or in those with leaf gas films. Leaves with a thinner cuticle, or those with gas films, have improved gas diffusion with floodwaters, so that underwater P(N) is enhanced. Underwater P(N) provides sugars and O(2) to submerged plants. Floodwaters often contain dissolved CO(2) above levels in equilibrium with air, enabling at least some P(N) by terrestrial species when submerged, although rates remain well below those in air.

Introduction

Emergent wetland plants are well adapted to waterlogged soils, but can also experience episodes of complete submergence. Complete submergence has an impact on wild species in coastal marshes and river floodplains (Armstrong ), and many rice crops are grown in regions threatened by floods, causing submergence (Jackson and Ram 2003).

Complete submergence impedes the exchange of O2 and CO2 between leaves and the environment (Mommer and Visser 2005; Voesenek ). Light availability to submerged plants also decreases, and markedly so when floodwaters are turbid (Mommer and Visser 2005; Voesenek ). Restricted photosynthesis, but ongoing substrate consumption in respiration or fermentation, causes sugars to become depleted in submerged plants, which in turn can result in damage or even death from substrate exhaustion (Bailey-Serres and Voesenek 2008; Colmer and Voesenek 2009).

The interface between land and water is not well defined as water tables fluctuate with precipitation and evaporation, so that plants experience variable periods and depths of flooding (Sculthorpe 1967). Plants exploit niches across these dynamic flooding gradients, but functional classification of plant types lacks sharp boundaries owing to the continuum of diversity. Notwithstanding these difficulties, plants from the wettest end of the gradient have been classified into two main groups: (i) aquatic plants that primarily live completely submerged and (ii) amphibious plants that live with emergent shoots or develop water forms when submerged (Iversen 1936; cited by Sculthorpe 1967). Emergent wetland plants typically maintain a large portion of their shoots in air, but occasionally become completely submerged. To clearly distinguish these emergent plants from other wetland species with shoot portions in air (e.g. emergent amphibious plants), we refer to this functional group as ‘terrestrial wetland plants’ (present review; Colmer and Pedersen 2008; Pedersen ). Such distinction is important as terrestrial wetland plants typically grow vigorously in waterlogged soils and/or areas with shallow standing water, with the depth limit being determined by capacity for transport of atmospheric O2 to belowground tissues (Sorrell ).

For submerged terrestrial plants, O2 deficiency and escape responses via shoot elongation have been elucidated, revealing sophisticated signalling, changes in gene expression and altered metabolism during submergence (e.g. reviewed by Bailey-Serres and Voesenek 2008,  2010). The capacity for some net photosynthesis (PN) to continue when under water enhances plant tolerance of submergence, as PN provides O2 for internal aeration and sugars for energy metabolism and growth (Mommer and Visser 2005).

The present review examines limitations to underwater PN by terrestrial wetland plants and compares their functioning with aquatic plants. Our focus here on underwater PN as related to the ecophysiology of submergence tolerance adds to the vast knowledge on root adaptations in wetland species. Roots of wetland plants typically contain large volumes of aerenchyma, often a barrier to radial O2 loss, and the ability to tolerate tissue O2 deficits and reduced phytotoxins in waterlogged soils (Armstrong 1979; Jackson and Armstrong 1999; Bailey-Serres and Voesenek 2008; Colmer and Voesenek 2009). Here, we show that underwater PN by submerged terrestrial wetland plants is limited by CO2 availability even though floodwaters commonly contain dissolved CO2 above air equilibrium, and so leaf traits influencing underwater PN are important for submergence tolerance.

The submergence environment during overland floods

Floods differ in seasonal timing, duration, depth and frequency (e.g. Vervuren ). Floodwater properties (e.g. water turbidity and dissolved CO2) that influence plant functioning can also differ substantially; light and CO2 available to submerged plants determine underwater PN and survival (Mommer and Visser 2005; Pedersen ). Thus, the flooding regime and water properties influence plant species distributions in flood-prone areas (Armstrong ; Voesenek ). In this section, we discuss three types of flooding events that can affect terrestrial wetland plants: flash floods, seasonal floods and tidal flooding.

Flash floods occur when heavy rainfall causes water levels to rise rapidly for a variable period of time, especially as run-off moves to low-lying areas (Setter ; Brammer 1990; Ram ). Flash floods in some regions can be more likely to occur during specific seasons, but in other areas flash flooding is not season specific. Seasonal floods are caused by an increase in water flow that surpasses the capacity of rivers in a landscape to discharge the large volumes of water, resulting in overflow of banks and floodplains. The origin of the increased water flow can be seasonal precipitation and/or snow melt (Brammer 1990). Tidal flooding impacts coastal plains and estuarine marshes with depths determined by the moon's cycle (e.g. neap tides and spring tides). Tidal floods involve saline water, whereas overland floods are usually freshwater, with the exception of some inland catchments with salt lakes.

Flooding can occur with various combinations of chemical and physical properties in the water; O2, CO2, temperature, pH and light can all vary (Setter ; Pérez-Lloréns ). Seawater pH is well buffered as it contains HCO3− (2.2 mM; Millero ) and HCO3− also buffers against severe depletion of dissolved CO2. In freshwater floods, HCO3− and CO2 concentrations are highly variable, but dissolved CO2 is commonly above air equilibrium (Table 1). The high CO2 concentrations typically result from respiration by organisms consuming labile carbon compounds (i.e. a net heterotrophic system); in addition, some water bodies receive CO2-enriched groundwater stream flows. By contrast, in net autotrophic systems photosynthesis depletes CO2 and produces O2. So, O2 concentrations in floodwaters can range from severely hypoxic (net heterotrophic) to well above air equilibrium (net autotrophic).

Table 1

Dissolved CO2 and O2 concentrations in various types of floodwaters. Medians with ranges in parentheses.

Environment	CO2 (µM)	O2 (µM)
Terrestrial
Flash flood (1,2) (n = 4)	1040 (3–1953)	150 (‘0’–280)
Seasonal flood (3−6) (n = 6)	365 (47–1600)	79 (‘0’–240)
Tidal flood (7,8) (n = 4)	16 (3–49)	292 (188–522)
Aquatic
Streams and rivers (9,10) (n = 31)	133 (11–836)	n.a.
Ponds (< 1 ha) (11) (n = 7)	59 (<1–374)	n.a.
Lakes (11) (n = 17)	45 (11–210)	n.a.

1,2(Ram ; Setter ); 3−6(Hamilton ; Hamilton ; Richey ; Valett ); 7,8(Pérez-Lloréns ; Winkel ); 9(Sand-Jensen and Frost-Christensen 1998); 10(Jonsson ) 11(Staehr ).

n.a., not available.

O2 was not measured in the water surveys conducted in 9, 10 and 11.

Temperature during flooding events can also vary widely (e.g. ∼6–37 °C; Hamilton ; Valett ; Pedersen ), depending on location and season. Respiration increases at warmer temperatures, which can deplete O2, and O2 concentration is further reduced owing to lower O2 solubility in water as temperature increases. So, the imbalance between O2 demand and supply to submerged terrestrial plants can be further exacerbated as temperature increases.

Flow rates during floods have only been reported, to our knowledge, in three papers: data are available for two flash floods and one seasonal river flood, and flows ranged from 0.002 to 0.3 m s−1. Flow rates affect the thickness of diffusive boundary layers (DBLs) and thereby influence gas and nutrient exchanges with submerged plants (Binzer ; Pedersen ). So, underwater PN can increase with increasing flow velocity since the DBLs become thinner (Jones ), but the response would plateau (cf. O2 supply; Binzer ) or even decline again if flows cause excessive shoot agitation (Madsen ).

Light regimes in floodwaters are dependent on several factors. When floodwaters contain suspended particles or dissolved coloured organic matter (e.g. tannins in Amazonian floodwaters; Parolin 2009), light availability will be reduced. Particle suspension can be highest during early stages of floods and particles often then settle; however, if particles settle on leaves these can still limit light. Waters of high nutrient availability typically support growth of microalgae, with dense populations of both biofilms and phytoplankton leading to lower light penetration to leaf surfaces (Sand-Jensen and Sondergaard 1981; Sand-Jensen and Borum 1991; Lassen ) and consequently also shallower depth limits for plant colonization (Sand-Jensen 1990). Examples of light reductions are available for floodwaters in the rice fields of India and Thailand; the depth at which 50% light remained varied from 0.07 to 0.7 m (Setter ; Ram ).

How does the submerged environment experienced by terrestrial wetland plants compare with that of water bodies containing permanent aquatic vegetation? In brief, environments supporting healthy stands of submerged aquatic plants, such as the shallow sea, and areas within rivers and lakes, also share many of the above-mentioned constraints to plant growth. Light attenuation in the water column (caused by water itself, dissolved coloured organic matter, phytoplankton and other particles) determines the maximum depth of colonization by aquatic plants. Seagrasses typically grow down to ∼10% of the surface light (Duarte 1991), whereas the depth penetration of plants in freshwater lakes is down to <1% and typically ∼5% of the surface light (Canfield ). The lower light compensation points for the growth of deep–colonizing freshwater plants result from these having higher shoot-to-root ratios than seagrasses. The deepest-growing freshwater plants, such as species of Ceratophyllum and Utricularia, do not produce roots at all (Cook 1990). Similar to terrestrial floodwaters, dissolved inorganic carbon (DIC) in freshwater can also vary widely (e.g. from 0.02 to 5.6 mM in British lakes; Maberly and Spence 1983). Depending on pH, the above DIC concentrations may result in dissolved CO2 levels from near or below air equilibrium (15 µM in freshwater at 20° C) to waters in streams/rivers, ponds and lakes that are typically supersaturated (Table 1); ponds can even contain up to 2000 µM CO2 (133-fold air equilibrium). The temperature in most water bodies fluctuates significantly less than surrounding air due to the much higher specific heat capacity of water compared with air (Hutchinson 1957), but there are exceptions, such as in shallow rock pools with large diel fluctuations (Pedersen ). Finally, the flow velocity in aquatic environments also varies widely, as described earlier for terrestrial floods, from almost stagnant conditions in ponds and deeper areas of lakes to very high velocities in rivers and in surf zones of the sea (2–3 m s−1; Vogel 1994). In fast-flowing water or in wave-zones, the strap-shaped leaves typical of some aquatic plants are highly adaptive as this morphology reduces the pressure drag (Vogel 1994).

In summary, floodwaters faced by terrestrial plants invoke some common constraints of restricted gas exchange and lower light availability, but conditions (O2, CO2, light and temperature) differ between locations and times, posing variable challenges to plant functioning during submergence. Floodwater chemical and physical properties, in addition to the well-recognized importance of seasonal timing, duration, depth and frequency of floods (e.g. Vervuren ), will influence plant growth and survival during submergence.

Net photosynthesis under water

Low CO2 and/or low light can restrict PN by submerged plants (Sand-Jensen 1989). This review focuses on CO2 acquisition. Aquatic species possess leaf traits to enhance DIC supply and thus rates of underwater PN. In Table 2, we compare the leaf traits of terrestrial wetland plants with those of submerged aquatic plants. Below we (i) summarize knowledge of morphological and anatomical leaf traits, and photosynthetic pathways including carbon-concentrating mechanisms (CCMs), and (ii) compare the rates of underwater PN by different types of aquatic and terrestrial wetland plants, as influenced by these leaf traits.

Table 2

Comparison of leaf traits influencing gas exchange and photosynthesis by terrestrial wetland plants when under water and by submerged aquatic plants. Modified from Sculthorpe (1967) with data from additional references as indicated by superscripts: 1(Neinhuis and Barthlott 1997), 2(Colmer and Pedersen 2008), 3(Maberly and Madsen 2002).

Leaf traits	Terrestrial wetland plants	Submerged aquatic plants
Morphology
Leaf size	Medium–large	Small–medium
Dissected/lobed	Rare	Common
Strap-shaped	Rare	Common
Leaf thicknessa	Moderate–thick	Thin
Surface hydrophobicity/leaf gas films1,2	Common	Absent
Hairs/trichomes	Rare	Absent
Anatomy
Stomata	Always present	Absent/non-functional
Cuticle	Always present	Absent/highly reduced
Chloroplasts in epidermal cells	Only in guard cells	Common
Aerenchyma	Variable	Variable
Porosity of lamina	High in thick, low in thin, lamina	High in thick, low in thin, lamina
Supporting fibres	Always present	Rare
Photosynthetic pathway/CCM3
C3	Common	Common
C4	Rare	Rare (but uncertain)
CAM	Absent	Rare
HCO3− use	Absent	Common

aFor data on SLA see Fig. 2. Other leaf features/properties can also differ between terrestrial wetland plants and submerged aquatic plants, such as: venation, lignification, stiffness, surface topography, differences between adaxial and abaxial surfaces, and in the case of some halophytic wetland species, presence of salt bladders and glands.

Fig. 2

Specific leaf area in terrestrial wetland plants, in amphibious homophyllous or heterophyllous wetland plants, and in submerged aquatic plants. The box-whisker plot shows the median, 10 and 90 percentiles, minimum and maximum values, and means are shown as ‘+’. Species and data sources are listed in Appendix 1. Differences amongst means of SLA of the four plant groups were tested by one-way analysis of variance and Tukey's multiple comparison tests. ** P < 0.01. Means with the same letter do not differ significantly at the 95% confidence interval.

Leaf traits of terrestrial wetland plants and submerged aquatic plants

Leaf morphology determines boundary layer resistances to exchange of dissolved gases and ions (Madsen and Sand-Jensen 1991). Boundary layer resistance can limit the rates of CO2 uptake and thus reduce underwater PN in submerged plants as diffusion is 104-fold slower in water than in air (Vogel 1994). Morphological traits (Table 2) that reduce the DBL resistance, by decreasing the distance to the ‘leading edge’ (Vogel 1994), include leaf shapes of small, dissected/lobed and/or strap-like leaves. In addition, aquatic leaves lack trichomes, thus avoiding the development of thicker boundary layers adjacent to their surfaces. Leaves of aquatic species also tend to be thin (Table 2), although there are several exceptions (e.g. isoetids; Sand-Jensen and Prahl 1982). Thin leaves have short internal diffusion path lengths, reducing the overall resistance for CO2 to reach chloroplasts (Madsen and Sand-Jensen 1991; Maberly and Madsen 2002). One example is the lamina of Najas flexilis, which is only two cell layers (Tomlinson 1982). In cases where leaves are relatively thick, CO2 is typically sourced from sediments (e.g. isoetids; Winkel and Borum 2009), and these leaves tend to be of high porosity to facilitate internal gas phase diffusion (Pedersen and Sand-Jensen 1992; Pedersen ; Sand-Jensen ).

In addition to these morphological traits, leaves of aquatic species also have anatomical traits that further reduce diffusive resistances for CO2 to reach chloroplasts (Table 2). Aquatic leaves lack, or have very reduced, cuticles. Diffusion across the cuticle is the main pathway of dissolved gas exchange as the leaves lack stomata, or if present, the stomata are non-functional (Pedersen and Sand-Jensen 1992). The diffusion path length to chloroplasts is also minimized by having these organelles in all epidermal cells, and in sub-epidermal cells the chloroplasts are positioned towards the exterior (Table 2).

Submerged aquatic plants also display physiological adaptations to increase the CO2 concentration at Rubisco, the site of carboxylation (Table 2)—these are referred to as CCMs (Maberly and Madsen 2002; Raven ). In submerged aquatic plants, CCMs include HCO3− use (Prins and Elzenga 1989), C4 (Magnin ), C3–C4 intermediates (Keeley 1999) and Crassulacean Acid Metabolism (CAM) photosynthesis (Keeley 1998). Carbon-concentrating mechanisms increase PN in CO2-limited submerged environments, and have also been suggested to diminish photorespiration (Maberly and Madsen 2002). Photorespiration results from the oxygenase activity of Rubisco and is promoted by a low CO2:O2 ratio (Ogren 1984), a condition common in leaves when under water (Bowes 1987). The low CO2 availability in aquatic environments would in itself lower the CO2:O2 ratio. Moreover, O2 in submerged leaves can be high as escape is slower than production in PN; O2 escape is not only hampered by DBLs but also by the relatively low O2 solublity in water; CO2 is 28.5-fold more soluble than O2 at 20° C (Baranenko ). Reduced photorespiration in a submerged aquatic CAM plant has been recently demonstrated (Pedersen ), supporting the view that CCMs do reduce photorespiration in aquatic species.

By contrast with aquatic species, leaves of terrestrial wetland plants lack most of the features described above (Table 2) and so suffer from large diffusion limitations to CO2 supply for PN when under water, unless they possess leaf gas films (Raskin and Kende 1983; Colmer and Pedersen 2008; Pedersen ) or produce submergence-acclimated leaves (Mommer and Visser 2005). Below, we evaluate underwater PN by leaves of terrestrial wetland plants and then consider the occurrence and functioning of leaf gas films.

Net photosynthesis of aquatic and submerged terrestrial wetland plants; leaf traits enhance CO2 supply

The most comprehensive comparison of underwater PN by aquatic and terrestrial wetland plants is the study by Sand-Jensen . Thirty-five species of four life forms (terrestrial, amphibious homophyllous, amphibious heterophyllous and aquatic species) were compared (listed in Appendix 1). Inclusion of data from other studies in the present analysis was constrained by differences in techniques and conditions used for underwater PN measurements, e.g. CO2, temperature and light (Appendix 2).

Classifications of wetland plants into functional groups are convenient, but are also imperfect as the boundaries are not sharp (see Introduction). As examples, some terrestrial wetland species produce new leaves when submerged and these can display some acclimation to the underwater environment (Mommer ). Similarly, homophyllous amphibious plants can also display some acclimation, e.g. thinner cuticles and modestly thinner leaves when formed under water (Nielsen 1993), but these changes are far more subtle than those displayed by heterophyllous amphibious plants. Not surprisingly, different authors have classified some species into different life forms. Here, our focus is on the comparison of underwater PN of leaves formed (i) in air by terrestrial wetland species, (ii) under water by amphibious homophyllous species, (iii) under water by amphibious heterophyllous species and (iv) under water by aquatic species.

An additional noteworthy feature of the study by Sand-Jensen was documentation of dissolved CO2 levels in lowland stream habitats. Underwater PN was measured at ambient and at elevated CO2 concentrations, to provide rates of relevance to the field situation as well as CO2-saturated PN for aquatic leaf types. The level of elevated CO2 used (∼800 μM, being ∼50-fold air equilibrium) would have saturated PN by the aquatic leaf types. It is uncertain whether rates were CO2 saturated for some of the terrestrial leaf types, which can require as much as 75-fold of air equilibrium CO2 when submerged (Colmer and Pedersen 2008).

We compare the rates on the dry mass basis (Fig. 1A and B) used by Sand-Jensen and also on a projected leaf surface area basis (Fig. 1C and D); conversions used specific leaf area (SLA) data in the literature (Fig. 2; Appendix 1). Data for SLA were not available for three of the aquatic and three of the terrestrial wetland species in Sand-Jensen , so these six were omitted from the present analysis (Appendix 1).

Fig. 1

Underwater net photosynthesis ( Net photosynthesis was measured at 15° C and is expressed per leaf dry mass (A and B) or per projected leaf area (C and D) at ambient CO2 levels (90–400 µM in the natural habitats; A and C) or at elevated CO2 levels (800 µM; B and D). Species and SLA data sources are listed in Appendix 1. Our analysis focused on the study by Sand-Jensen as it is the most comprehensive available; addition of other data was constrained by differences in techniques and conditions used (e.g. CO2 and temperature; Appendix 2). Terrestrial, leaves formed in air by emergent wetland plants; homophyllous, leaves formed under water by amphibious wetland plants; heterophyllous, leaves formed under water by amphibious wetland plants; aquatic, leaves formed under water by submerged aquatic plants (cf. Sculthorpe 1967). Rates on a mass basis (A and B) were converted to an area basis (C and D) using the published SLA data (Fig. 2, Appendix 1). The box–whisker plot shows the median, 10 and 90 percentiles, minimum and maximum values, and means are shown as ‘+’; the dot in the terrestrial column indicates an outlier. Differences amongst means of the four plant groups within each panel were tested by one-way analysis of variance and Tukey's multiple comparison tests. ** P < 0.01 and *** P < 0.001. Means with the same letter within each panel do not differ significantly at the 95% confidence interval.

The overall beneficial effects of aquatic leaf traits (Table 2) for underwater PN, as well as the generally poor performance of leaves of terrestrial plants, were clearly demonstrated in Sand-Jensen . These authors highlighted that (i) underwater PN on a mass basis increased from terrestrial, then amphibious, to truly aquatic leaf types and (ii) Danish lowland stream waters are commonly supersaturated with CO2, allowing even some terrestrial species to have adequate PN for growth when submerged in these habitats.

The higher PN by aquatic leaf types per unit mass with near-ambient CO2 concentrations (∼90–400 μM) demonstrates the higher C-return per unit of dry mass investment by these leaf types in an underwater environment as compared with terrestrial types (Fig. 1A). When external CO2 was supplied at an elevated level of ∼800 μM (Fig. 1B), underwater PN values by the aquatic and heterophyllous amphibious leaves still exceeded those of the terrestrial and homophyllous leaf types. The low rates by terrestrial leaves even with elevated CO2 further demonstrate the large diffusion limitations for CO2 entry that restrict underwater PN.

Expression of underwater PN rates on a surface area basis, the units typically used in terrestrial plant physiology (whereas in aquatic sciences, rates are typically expressed per unit dry mass), interestingly, removes differences between the terrestrial and aquatic leaf types, at both ambient and elevated CO2 (Fig. 1C and D). The order of magnitude of higher SLA (Fig. 2) of aquatic and many amphibious leaf types clearly sets an upper limit for PN on an area basis. Maximum PN, however, would rarely be achieved in most aquatic environments owing to light and CO2 limitations (Sand-Jensen 1989; Kirk 1994) so that the lower CO2-saturated rates of PN on an area basis for aquatic leaves would be of little consequence for their life under water.

Comparisons of the rates of underwater PN by terrestrial wetland plant leaf types with those achieved by aquatic leaf types are informative with respect to performance when submerged (Fig. 1), but here we also consider how these rates under water compare against those in air. For the terrestrial wetland species in Fig. 1, we could only find data on PN in air for three (Carex elata, Ranunculus repens and Phragmites australis; Appendix 1); PN in air at ambient CO2 was 12.5–17 µmol m−2 s−1. When submerged with CO2 at levels near ambient (but well above air equilibrium in these habitats), the mean PN under water was only 9% of that in air (cf. Fig. 1C). Thus, underwater PN is greatly reduced when terrestrial wetland species become submerged.

The analyses presented above for underwater PN by leaves of terrestrial wetland plants involved experiments in which leaves growing in air were tested under water. Several terrestrial wetland species produce new leaves when submerged, and these can display some acclimation to the underwater environment (e.g. thinner cuticles and thinner leaves; Mommer ). Acclimated leaves have decreased resistances against CO2 and O2 movement across the cuticle and epidermis (Mommer and Visser 2005; Mommer ). The best example is the several-fold reduction in cuticle resistance and thus the 69-fold higher underwater PN at an external CO2 concentration of 250 μM by Rumex palustris (Mommer ). Although a study of seven terrestrial wetland species established the formation of a thinner cuticle as a common response when submerged, and demonstrated enhanced underwater gas exchange, the degree of this response was not correlated with submergence tolerance among these species (Mommer ). These anatomical, and in some cases morphological (e.g. R. palustris leaves are also more elongated), changes in submerged leaves of terrestrial species are much more subtle than the altered leaf development displayed by amphibious heterophyllous species which produce true aquatic leaf types when under water (Nielsen 1993).

In summary, PN by terrestrial wetland plants is reduced markedly when they are submerged. Leaves of terrestrial wetland plants generally lack the numerous beneficial leaf traits for underwater PN possessed by aquatic plants, although new leaves can display some acclimation (e.g. thinner cuticles and higher SLA). In addition, as discussed in the next section, some leaves of terrestrial wetland species retain a gas film when submerged, a trait that also enhances underwater PN.

Leaf gas films enhance the net photosynthesis of submerged terrestrial wetland plants

Many terrestrial wetland plants have water-repellent (i.e. hydrophobic) leaf surfaces, resulting in self-cleaning by water droplets as these run off leaves (Neinhuis and Barthlott 1997). Leaf water repellence has been assessed by measurement of water droplet contact angles with the surface (Adam 1963; Brewer and Smith 1997; Neinhuis and Barthlott 1997)—angles of 140° or more indicate a hydrophobic surface whereas angles of 110° or less indicate a wettable surface. Water repellence (i.e. surface hydrophobicity) is determined by the micro- and nano-structures of the surface, as well as wax crystals (Wagner ; Bhushan and Jung 2006).

Superhydrophobic leaves retain a microlayer of gas when submerged, referred to as ‘gas envelopes’ (Setter ) and/or ‘leaf gas films’ (Colmer and Pedersen 2008). We prefer the term ‘gas film’ because although leaves of some species retain a gas layer on both sides (i.e. enveloped in gas), others retain a gas layer on only one side due to differences in hydrophobicity between adaxial and abaxial surfaces (Colmer and Pedersen 2008; Winkel ). Gas films on leaves have been observed in field situations for several terrestrial wetland species when submerged in lakes, ponds, river edges and rice fields on floodplains: rice (Setter ); P. australis, cover of New Phytologist, Volume 177(4); Spartina anglica (Winkel ); and own observations (A. Winkel, T. D. Colmer and O. Pedersen). Information on the persistence of gas films on leaves with time following submergence is scant; gas films remained for at least 2 weeks (i.e. evaluation was terminated at 2 weeks) on leaves of Phalaris arundinacea, P. australis and Typha latifolia (all with gas films on both sides) and Glyceria maxima (gas film on only the adaxial side) in a controlled environment (Colmer and Pedersen 2008), but for some other species gas films diminish within a few days (own unpublished data; A. Winkel, T. D. Colmer and O. Pedersen).

Gas films on submerged leaves enhance CO2 fixation, as first demonstrated for rice (9- to 10-fold increase; Raskin and Kende 1983). The beneficial effect of leaf gas films to underwater PN has also been shown for other terrestrial wetland species; at 50 μM dissolved CO2, gas films increased underwater PN by 1.5- to 6-fold in leaves of four wetland species (Colmer and Pedersen 2008). Data demonstrating the beneficial effect of leaf gas films on underwater PN are shown for several species in Fig. 3. Apparent resistance to CO2 entry, at environmentally relevant CO2 concentrations in the submergence water, was ∼5-fold less in leaves with gas films compared with those with gas films removed (rice and P. australis; Pedersen ).

Fig. 3

Underwater net photosynthesis in terrestrial wetland plants with or without leaf gas films and when gas films were removed. Measurements for six species were conducted with 50 µM CO2 at 20 °C and photosynthetically active radiation (PAR) of 600 µmol m−2 s−1; the exceptions were Oryza sativa (30 °C; PAR 350 µmol m−2 s−1) and S. anglica (15 µM CO2; PAR 550 µmol m−2 s−1). These reflect the higher temperatures in tropical rice fields (O. sativa) and the lower CO2 in seawater that submerges Spartina marshes. Gas films were removed from leaf surfaces by brushing with 0.05% Triton X-100. Species lacking leaf gas films were also brushed with Triton X-100 and showed no, or only a slight, reduction in PN. Data from Colmer and Pedersen (2008), Pedersen and Winkel .

Leaf gas films provide an enlarged gas–water interface to promote gas exchange with the surrounding floodwater (CO2 uptake during light periods; O2 uptake during dark periods) (Colmer and Pedersen 2008; Pedersen ). In addition to the enlarged gas–water interface, leaf gas films might also enable stomata to remain open when leaves are submerged. By contrast, for leaves without gas films, stomata are hypothesized to close upon submergence (Mommer and Visser 2005), so that CO2 and O2 must then transverse the cuticle (Mommer ). The beneficial effect of leaf gas films on underwater PN was not only demonstrated by the marked decreases when these were removed (Fig. 3), but also leaves with this feature had higher rates of underwater PN than leaves from species without leaf gas films (Fig. 3). Thus, leaf gas films appear to enable rates of underwater PN by terrestrial leaves similar to those achieved by submergence-acclimated leaves of terrestrial wetland plants (data and discussion in Colmer and Pedersen 2008). Terrestrial species possessing leaf gas films would benefit from enhanced underwater PN during short to medium periods of submergence, depending on persistence of the films. By contrast, for species lacking leaf gas films but that produce new acclimated leaves under water, these new leaves take several days to produce so that PN would likely be less during the initial submergence period, but continued production of acclimated leaves would benefit these species during medium to prolonged submergence.

Detailed knowledge on leaf gas films is available only for rice (one cultivar only; Pedersen ). Measurements using O2 microelectrode profiling determined that gas film thickness varied from <10 to 140 μm; positional differences mainly resulted from ridges on leaves (i.e. gas films thinner at the tops of ridges, thicker between adjacent ridges). Using a ‘buoyancy method’ to measure gas volumes on the surfaces, and within, submerged leaves, showed that tissue porosity was 19% (v/v) and the gas volume of the films was 3.8 times more than the gas within the rice leaf. Diffusive boundary layer widths adjacent to submerged leaves with gas films were surprisingly larger than those adjacent to submerged leaves without gas films, so the enlarged water–gas interface provided by the gas films would have been the major mechanism that reduced resistance to gas exchange of the leaves when under water. At dissolved CO2 concentrations of relevance to field conditions (15–180 μM; e.g. in Thailand, Setter ; India, Ram ), underwater PN was enhanced 4- to 4.9-fold by gas films on leaves of rice (Pedersen ). Underwater PN by leaves with gas films and CO2 at near-ambient concentrations was 22% of PN in air. When gas films were removed artificially from leaves of completely submerged rice, tissue sugar levels and growth were both reduced. Thus, leaf gas films contribute to submergence tolerance of rice by enhancing CO2 entry for underwater PN.

The experiments by Pedersen also elucidated that when rice leaves are in flowing water (15 mm s−1; simulating low flows such as might occur across rice fields), the gas film oscillates and the transition zone between mass flow in the bulk medium and diffusion in the boundary layer was wider, and more variable, than for leaves without a gas film. Oscillations of leaf gas films in flowing water were also noted by Barthlott , and they reported that specialized surface hairs on the leaves of Salvinia molesta can stabilize the gas film, even in fast-flowing water (such as in streams). The leaf surface of S. molesta possesses ‘eggbeater-shaped hairs’ that are hydrophobic except for the tips, a feature that enables gas film formation and retention by ‘pinning’ the water–air interface (Barthlott ). The presence of this feature was suggested to prevent the formation and detachment of bubbles that otherwise could occur when in fast-flowing waters (Barthlott ). This is a very interesting leaf surface feature, although the ecophysiological significance could be debated as S. molesta is a floating plant not typically found in fast-flowing waters; the large gas volume trapped by these specialized structures on the surface of the leaves would contribute significantly to the buoyancy of this floating plant.

In addition to enhanced CO2 uptake for photosynthesis, leaf gas films also improve O2 uptake during darkness from floodwaters into leaves (Colmer and Pedersen 2008; Pedersen ). Thus, leaf gas films enhance leaf O2 status both during the daytime and during nights, with benefits also of improved internal aeration of the entire body of submerged plants. Oxygen derived from PN during light periods, as well as O2 entry from the floodwater into leaves when in darkness, moves internally via aerenchyma to roots of rice (Pedersen ) and rhizomes and roots of S. anglica (Winkel ).

In conclusion, our recent studies of leaf gas films (Colmer and Pedersen 2008; Pedersen ; Winkel ) have supported the hypothesis by Setter , who observed this feature on submerged rice in field situations in Thailand, that gas films provide ‘an interface between the gas and water phases for collection of CO2 and dispersal of O2 during the day or collection of O2 during the night’. This mechanism is analogous to the gas layer (plastron) on some aquatic insects that provides an enlarged gas–water interface between the tracheary system and surrounding water (Thorpe and Crisp 1949; Raven 2008; Pedersen and Colmer 2012). For terrestrial wetland species, the few data available indicate that leaf gas films enable rates of underwater PN similar to those achieved by submergence-acclimated leaves, in both cases being higher than in terrestrial air-formed leaves without these features (data and discussion in Colmer and Pedersen 2008).

Conclusions and future perspectives

Submergence can have adverse effects on terrestrial wetland plants because of restricted gas exchange and low light. Floodwaters are variable in dissolved O2, CO2, light and temperature. Few data are available on key environmental parameters in various submergence environments—yet these factors influence underwater PN, plant growth and survival. Knowledge of floodwater conditions will enhance one's understanding of plant performance during submergence and enable the design of controlled experiments that better simulate particular submergence environments.

Submergence tolerance of terrestrial wetland plants is influenced by leaf traits. Although terrestrial wetland plants generally lack the numerous beneficial leaf traits possessed by aquatic plants, the few studies available demonstrate that some terrestrial species produce new leaves with a thinner cuticle under water and others possess leaf gas films. The improved gas diffusion between leaves and floodwaters enhances underwater PN and so contributes significantly to sugar and O2 supply of submerged plants. However, studies of leaf gas film functioning are in their infancy. Our priorities are (i) quantification of the occurrence and persistence of leaf gas films amongst a wide number of wetland species, and determination of whether this trait is related to species distributions in various flood-prone wetlands (cf. analysis of shoot elongation trait; Voesenek ) and (ii) evaluation of whether rice, or its relatives, possesses variation in leaf gas film formation and persistence, and elucidation of the underlying genetic control of this trait using the array of resources available in rice.

More broadly, there are surprisingly few studies on PN by terrestrial wetland plants when emergent and when submerged. Also lacking are measurements of PN with time after submergence and de-submergence. Future studies should compare the performances of species from various habitats, using a range of appropriate bases of expression of PN rates (area, mass, chlorophyll and leaf N) to facilitate interdisciplinary comparisons by aquatic and terrestrial plant biologists.

Sources of funding

This work was funded by The Danish Council for Independent Research Natural Science grant No. 645-06-0287; The Centre for Lake Restoration, a Villum Kann Rasmussen Centre of Excellence; and Australian Research Council Project LP0882350. AoB PLANTS supported O.P.'s participation at the International Botanical Congress 2011 in Melbourne.

Contributions by the authors

A.W. compiled the literature and drafted the section ‘The submergence environment during overland floods’. T.D.C. and O.P. contributed equally to the remaining sections.

Conflicts of interest statement

None declared.

	Sources of information
	Underwater PN (used in Fig. 1)	SLA (used in Figs 1 and 2)	PN in air (used in text)
Terrestrial n= 10
Equisetum palustre	Sand-Jensen et al. (1992)	Andersson and Lundegård (1999)	Data not available
Phragmites australis	Sand-Jensen et al. (1992)	Colmer and Pedersen (2008)	Hellings and Gallagher (1992)
Epilobium hirsutum	Sand-Jensen et al. (1992)	Nielsen (1993)	Data not available
Carex elata	Sand-Jensen et al. (1992)	Meziane and Shipley (2001)	Busch and Lösch (1998)
Poa pratensis	Sand-Jensen et al. (1992)	Meziane and Shipley (2001)	Data not available
Chrysosplenium alterniflorum	Sand-Jensen et al. (1992)	Wang et al. 2009	Data not available
Ranunculus repens	Sand-Jensen et al. (1992)	Nielsen (1993)	Lynn and Waldren (2002)
Solanum dulcamara	Sand-Jensen et al. (1992)	Flynn et al. (2006)	Data not available
Barbarea stricta	Sand-Jensen et al. (1992)	Nielsen (1993)	Data not available
Cardamine amara	Sand-Jensen et al. (1992)	Nielsen (1993)	Data not available
Amphibious (homophyllous) n = 7
Catabrosa aquatica	Sand-Jensen et al. (1992)	Nielsen (1993)	Not considered
Glyceria maxima	Sand-Jensen et al. (1992)	Colmer and Pedersen (2008)	Not considered
Myosotis laxa	Sand-Jensen et al. (1992)	Lenssen et al. (2003)	Not considered
Veronica anagallis	Sand-Jensen et al. (1992)	Nielsen (1993)	Not considered
Veronica beccabunga	Sand-Jensen et al. (1992)	Nielsen (1993)	Not considered
Berula erecta	Sand-Jensen et al. (1992)	Nielsen (1993)	Not considered
Myosotis palustris	Sand-Jensen et al. (1992)	Nielsen (1993)	Not considered
Amphibious (heterophyllous) n = 5
Callitriche cophocarpa	Sand-Jensen et al. (1992)	Nielsen (1993)	Not considered
Callitriche stagnalis	Sand-Jensen et al. (1992)	Tom Vindbæk Madsen, personal communication	Not considered
Sparganium emersum	Sand-Jensen et al. (1992)	Colmer and Pedersen (2008)	Not considered
Sparganium erectum	Sand-Jensen et al. (1992)	Nielsen and Sand-Jensen (1989)	Not considered
Sagittaria sagittifolia	Sand-Jensen et al. (1992)	Dina Ronzhina, personal communication	Not considered
Aquatic n = 7
Lemna trisulca	Sand-Jensen et al. (1992)	Dina Ronzhina, personal communication	Not considered
Potamogeton perfoliatus	Sand-Jensen et al. (1992)	Spence et al. (1973)	Not considered
Elodea canadensis	Sand-Jensen et al. (1992)	Madsen et al. (1996)	Not considered
Potamogeton crispus	Sand-Jensen et al. (1992)	Nielsen and Sand-Jensen (1989)	Not considered
Potamogeton pectinatus	Sand-Jensen et al. (1992)	Nielsen and Sand-Jensen (1989)	Not considered
Batrachium peltatum	Sand-Jensen et al. (1992)	Nielsen (1993)	Not considered
Batrachium aquatile	Sand-Jensen et al. (1992)	Nielsen and Sand-Jensen (1989)	Not considered

Source	Species tested	CO2 (µM)	Temperature (°C)	PAR (µmol m−2 s−1)	Notes
Nielsen (1993)	Barbarea stricta, Batrachium aquatile, Berula erecta, Callitriche cophocarpa, Cardamine amara, Catabrosa aquatile, Epilobium hirsutum, Glyceria maxima, Hydrocotyle vulgaris, Littorella uniflora, Lobelia dortmanna, Lotus uliginosus, Montia fontana, Myosotis palustris, Polygonum amphibium, Ranunculus repens, Sparganium emersum, Sparganium erectum, Veronica anagallis-aquatica, Veronica beccabunga	100	25	600	Also PN rates in air, although some seem unusually high
Sand-Jensen and Frost-Christensen (1998)	Myosotis palustris, Sparganium emersum	20 and 280	12 and 24	400	Also effects of initial O2 concentration and temperature on underwater PN
Sand-Jensen and Frost-Christensen (1999)	Berula erecta, Menta aquatica, Myosotis palustris, Veronica anagallis-aquatica	100 and 700	15	350	Also initial slope at PN rate-limited CO2 concentrations and Pmax
Vervuren et al. (1999)	Arrhenatherum elatius, Phalaris arundinacea, Rumex crispus	2,200	20	740	Also PN rates under water after 30 days of submergence
Nielsen and Nielsen (2006)	Lobelia cardinalis, Nesaea crassicaulis	40 and 1500	20	1200	Also PN rates in air
Mommer et al. (2007)	Rumex palustris	10–10 000	20	400	Also full CO2 response curve and PN rate in air
Pedersen et al. (2006)	Halosarcia pergranulata (syn. Tecticornia pergranulata)	20–6800	20	1500	Also PN rate in air
Colmer and Pedersen (2008)	Acorus calamus, Glyceria maxima, Phalaris arundinacea, Phragmites australis, Sparganium emersum, Typha latifolia	50 and 500	20	600	Also full CO2 response curve for Phragmites australis
Pedersen et al. (2009)	Oryza sativa	15–2000	30	350	Also full CO2 response curve under water and PN rate in air
Pedersen et al. (2010)	Hordeum marinum	18–2000	20	350	Also full CO2 response curve under water and PN rate in air