Intra-colonial variability in light acclimation of zooxanthellae in coral tissues of Pocillopora damicornis
- First Online:
- Cite this article as:
- Ulstrup, K.E., Ralph, P.J., Larkum, A.W.D. et al. Mar Biol (2006) 149: 1325. doi:10.1007/s00227-006-0286-4
- 211 Views
We investigated heterogeneity of light acclimation of photosynthesis in sun- and shade-adapted coenosarc and polyp tissues of Pocillopora damicornis. The zooxanthellar community within P. damicornis colonies at Heron Island is genetically uniform, yet they showed a large degree of plasticity in their photo-physiological acclimation linked to light microclimates characterised by fibre-optic microprobes. Microscale scalar irradiance measurements showed higher absorption in polyp than coenosarc tissues and higher absorption in the more densely pigmented shade-adapted polyps than in sun-adapted polyps. The combination of an O2 microelectrode with a fibre-optic microprobe (combined sensor diameter 50–100 μm) enabled parallel measurements of O2 concentration, gross photosynthesis rate and photosystem II (PSII) quantum yield at the coral surface under steady-state conditions as a function of increasing irradiances. Lower O2 levels at the tissue surface and higher compensation irradiance indicated a higher respiration activity in sun-adapted polyp tissue as compared to shade-adapted polyps. Shade-adapted coenosarc and polyp tissues exhibited lower maxima of relative electron transport rates (rETRmax) (84±15 and 41±10, respectively) than sun-adapted coenosarc and polyp tissues (136±14 and 77±13, respectively). Shade-adapted tissues showed stronger decrease of rETR at high scalar irradiances as compared to sun-adapted tissues. The relationship between the relative PSII electron transport and the rate of gross photosynthesis, as well as O2 concentration, was non-linear in sun-adapted tissues over the entire irradiance range, whereas for shade-adapted tissues the relationship became non-linear at medium to high scalar irradiances >200 μmol photons m−2 s−1. This suggests that rETR measurements should be used with caution in corals as a proxy for photosynthesis rates. The apparently high rates of photosynthesis (oxygen evolution rates) suggest that there must be a considerable electron transport rate through the photosystems that is not observed by the rETR measurements. This may be accounted for by vertical heterogeneity of zooxanthellae in the tissue and the operation of an alternative electron pathway such as cyclic electron flow around PSII.
Reef-building corals are sessile marine cnidarians that harbour endosymbiotic microalgae of the dinoflagellate genus Symbiodinium, known as zooxanthellae. The dynamic relationship between production and consumption of O2 in corals is regulated through the photosynthetic activity of the zooxanthellae, the respiration of the zooxanthellae and their hosts and solute exchange across diffusive boundary layers (DBLs) (Shashar et al. 1993; Kühl et al. 1995). Corals are subject to limited gas exchange with the surrounding seawater across a DBL, causing severe hypoxia in the dark due to the absence of light-driven photosynthesis, while O2 supersaturation is common during exposure to high light (Kühl et al. 1995; Ulstrup et al. 2005). The light intensity and spectrum reaching the zooxanthellae, and the solute exchange between the coral and the surrounding seawater are the most important external regulators of photosynthesis (Dubinsky et al. 1984; Patterson et al. 1991; Kühl et al. 1995).
The photosynthetic capacity of zooxanthellae as measured by variable fluorescence is commonly used as a proxy for the health of the coral (Warner et al. 1999; Bhagooli and Hidaka 2003). Decreased zooxanthellar productivity has been correlated to episodes of coral bleaching, a process in which corals lose their symbionts and/or pigments and subsequently may perish (Hoegh-Guldberg 1999; Marshall and Baird 2000). It has also been speculated that genetically distinct zooxanthellae may exhibit differences in their photosynthetic capacity and tolerance towards bleaching conditions (Buddemeier and Fautin 1993; Baker et al. 2004).
Variable chlorophyll a fluorescence analysis has shown differences in photosynthetic performance between upper and lower surfaces of individual colonies (Jones et al. 1998; Ralph et al. 2005), between polyp and coenosarc tissue (Ralph et al. 2002; Hill et al. 2004), and along the length of coral branches (Hill et al. 2004). Kühl et al. (1995) showed significant variation of gross photosynthetic activity between polyp and coenosarc tissue of Favia sp. due to differences in boundary layers and light microclimate within a single polyp. The scalar irradiance, i.e. the spherically integrated quantum flux from all directions about a point, can locally exceed the incident downwelling irradiance (Ed) in corals due to intense scattering and light trapping in the tissue and skeleton (Kühl et al. 1995; Enriquez et al. 2005). The large heterogeneity of the physico-chemical microenvironment within corals can lead to different responses of photo-acclimation and -adaptation in different parts of the coral colony (Falkowski et al. 1990; Jones et al. 1998; Ralph et al. 2002).
Oxygen microelectrodes enable fine-scale studies of gross photosynthetic activity using the light–dark shift technique (Revsbech and Jørgensen 1983) at the surface of coral tissue, where the photosynthesis of the zooxanthellae can be assessed with this technique independent of host or symbiont respiration (Kühl et al. 1995; De Beer et al. 2000). Variable chlorophyll a fluorescence analysis with fibre-optic microprobes (Schreiber et al. 1996; Kühl et al. 2005) can also be used to map photosynthesis and light acclimation at high spatial resolution in corals (Ralph et al. 2002).
Oxygen microsensor techniques enable measurements of gross photosynthesis rate vs. irradiance curves over different time scales. The measurement of gross photosynthesis rate with oxygen microsensors has a spatial resolution of ∼0.1 mm and assumes steady-state O2 distribution before the light–dark shift (Revsbech and Jørgensen 1983; Kühl et al. 1996). Steady-state conditions at the point of measurement thus reflect a balance between O2 transport, consumption and production, which is obtained when acclimation to a particular light intensity has occurred and no further changes to the O2 level are detected. Similarly, the effective quantum yield of photosystem II (PSII) and the relative photosynthetic electron transport rate at steady-state, under a range of irradiances, can be measured with pulse-amplitude-modulated (PAM) fluorometers using the saturation-pulse technique (Schreiber 2004). Measurements of steady-state photosynthesis vs. irradiance curves typically involve incubation for 10–60 min at each irradiance level. This time interval is long enough to allow for photo-acclimation during incubation (Kühl et al. 2001) and may also lead to greater photo-inactivation above saturating irradiance due to the prolonged light exposure.
Another way of investigating the photosynthetic capacity of corals is to measure so-called rapid light curves (RLC), where quantum yields and relative photosynthetic electron transport are measured rapidly over a range of irradiances using only very brief (10–20 s) incubation periods for each irradiance (Schreiber 2004). RLC, obtained over a few minutes, provide insight into the momentary acclimation capacity for light utilisation of PSII (Ralph and Gademann 2005) and should not be interpreted as steady-state photosynthesis vs. irradiance curves.
In this study, we investigated the branching coral Pocillopora damicornis. The population of P. damicornis in the Heron Island lagoon has been found to harbour only one strain of zooxanthellae belonging to clade C (LaJeunesse et al. 2003; Ulstrup et al. 2006), as confirmed by denaturing gradient gel electrophoresis and single-stranded conformation polymorphism, respectively, followed by DNA sequencing. This coral population is therefore suitable to assess intra-colonial variation of zooxanthellar physiology without complications of underlying genetic variation. The branching morphology induces a heterogenous light climate across the colony where basal parts of the colony are shaded by exposed branches at a scale of millimetres to centimetres (Ralph et al. 2005). However, the ambient light climate is also variable at a much finer scale (μm–mm) as a result of structural heterogeneities of the skeleton and tissue, such as those observed between polyp and coenosarc tissue (Kühl et al. 1995; Enriquez et al. 2005). Physiological responses may thus correlate to tissue type as well as to light microclimate (Ralph et al. 2002).
Here we present the first direct microscale comparison of oxygen- and variable fluorescence-based measures of coral photosynthesis, as determined with a novel combined microsensor consisting of an O2 microelectrode and a fibre-optic microprobe (Kühl 2005). We examined the gross photosynthesis and steady-state PSII quantum yields of zooxanthellae harboured in sun- and shade-adapted regions of coenosarc and polyp tissue in individual colonies of the scleractinian coral P. damicornis at Heron Island on the southern Great Barrier Reef. Our results show that variable fluorescence data obtained in corals cannot be directly correlated to productivity, especially at higher irradiances.
Materials and methods
Sampling and experimental setup
Samples were collected in the Heron Island lagoon (<2 m depth) adjacent to Heron Island Research Station (152°06′E, 20°29′S) during January 2005. Sun- and shade-adapted fragments of P. damicornis colonies were collected and maintained in flow-through aquaria under shaded conditions <50 μmol photons m−2 s−1 for 1 day. Samples were held securely in plasticine within a custom-built flow chamber 25 cm × 10 cm × 10 cm at a flow velocity of ∼1 cm s−1. A combined O2 microelectrode and tapered fibre-optic microsensor with a combined sensor diameter of 50–100 μm (Kühl 2005) was used to conduct simultaneous measurements of O2 concentration, gross photosynthesis rate and PSII fluorescence yield (maximum quantum yield in the dark, effective quantum yield in the light) at the surface of the coral.
Scalar irradiance spectra (450–750 nm) were measured with a fibre-optic microprobe ∼0.1 mm in diameter (Lassen et al. 1992) connected to a high spectral resolution diode-array detector (PMA-11, Hamamatsu Photonics, Hamamatsu, Japan). The tip of the microprobe was placed on the surface of the coenosarc tissue or inserted into a polyp using a micromanipulator and a dissecting microscope. The probe was advanced to the tissue at a zenith angle of ∼135–145° relative to the vertically incident light. A fibre-optic light source (Schott KL-2500, Mainz, Germany) with a 250-W halogen lamp (Xenophot, Osram, Germany) and fitted with a collimating lens was used for homogeneous, vertical illumination of the coral samples. Once the polyp tentacles engulfed the probe (usually after a few minutes), the measurement was performed. Scalar irradiance measurements were related to the incident downwelling irradiance measured over a black non-reflective surface to reduce stray light and reflection of light. Three haphazardly chosen areas of coenosarc and polyp tissues in sun- and shade-adapted regions, respectively, were measured. All measurements were performed in a darkened room at <1 μmol photons m−2 s−1. The scalar irradiance measurements were used to correct the known levels of actinic incident irradiance used in the microsensor measurements of gross photosynthesis rate and variable fluorescence to tissue scalar irradiance.
Fluorescence yields were measured on both dark-acclimated (maximum Fm, minimum Fo) and light-acclimated (effective maximum Fm′, minimum Ft) samples (see Schreiber 2004 for review). The steady-state measurements were performed after a 10-min initial dark period to obtain the maximum quantum yield of PSII, Fv/Fm = (Fm − Fo)/Fm, and after each of the following seven actinic irradiance steps which provided an estimate of the effective quantum yield of PSII, ΦPSII = (Fm′ − Ft)/Fm′. Non-photochemical quenching, NPQ = (Fm − Fm′)/Fm′, was determined at an intermediate scalar irradiance. NPQ for sun- and shade-adapted coenosarc tissues was determined at 358 and 328 μmol photons m−2 s−1, respectively, whereas NPQ for sun- and shade-adapted polyp tissues was determined at 406 and 290 μmol photons m−2 s−1, respectively. The rETR was calculated as rETR = E0 × ΦPSII (Ralph et al. 2002), where E0 is the scalar irradiance. Information about the partitioning of light energy between PSI and PSII, as well as the absorption cross-section of PSII would be necessary in order to determine absolute ETR values (Schreiber 2004; Ralph and Gademann 2005). The settings of the Microfibre-PAM were adjusted so that a fluorescence signal above 100 units on dark-acclimated samples was obtained (photomultiplier gain 30, output gain 8 and measuring light 8).
The functions of Platt et al. (1980) were fitted to rETR vs. scalar irradiance and gross photosynthesis vs. scalar irradiance data derived from steady-state measurements in order to quantitatively compare their descriptive parameters [fluorescence—maximum rETR (rETRmax), initial slope (αf) and minimum saturating scalar irradiance (Ekf); gross photosynthesis rate—maximum gross photosynthesis rate (Pgmax), initial slope (αPg) and minimum saturating scalar irradiance (EkPg)]. Fitted curves were derived from the empirical data (Sigmaplot v6.1, Systat, Richmond, CA, USA). For further details see Kühl et al. (2001) and Ralph et al. (2002).
Sun- and shade-adapted surfaces of coenosarc and polyp tissues were tested for statistical independency using Pearson’s correlation. Assumptions of normality and equal variances were satisfied, thus validating the use of a one-way analysis of variance for all comparisons of rETRmax, αf, Ekf, Pgmax, αPg and EkPg as well as for Fv/Fm, ΦPSII and NPQ. The non-parametric Kruskal–Wallis test was performed where normality and equal variance were not achieved. Statistical analyses were performed using SPSS software v11.0.0 (USA).
Differences in Fv/Fm, and ΦPSII and non-photochemical quenching (NPQ) of sun- and shade-adapted coenosarc and polyp tissues of Pocillopora damicornis determined at 358 and 328 μmol photons m−2 s−1 for sun- and shade-adapted coenosarc tissues, respectively, and at 406 and 290 μmol photons m−2 s−1 for sun- and shade-adapted polyp tissues, respectively. Averages (n = 5 ± SE) are shown
Quantitative parameters derived from fitted relative electron transport rates (rETR) and gross photosynthesis rate curves as a function of PAR
While microsensors for oxygen and variable chlorophyll a fluorescence have previously been employed separately to investigate the heterogeneity of zooxanthellae photosynthesis in hospite (Kühl et al. 1995; Ralph et al. 2002), this study provides the first combined microscale measurements of scalar irradiance, steady-state O2 concentration, gross photosynthesis rates and PSII quantum yields and rETR in coral tissues. This is a major technical advance in coral photobiology. Variable fluorescence-based studies of corals are often used to infer impacts on the regulation of photosynthesis, yet hitherto no studies have actually compared steady-state photosynthesis rates with fluorescence-derived measures of coral photosynthesis obtained at similar spatial scales and within the same tissue type.
Shallow-water corals such as those in the Heron Island lagoon are adapted to high irradiance, which may reach >2,000 μmol photons m−2 s−1 under light fleck conditions. However, zooxanthellae live in a diffuse light field where they receive a major part of the light for photosynthesis as scattered light through their calcium carbonate skeleton and within the tissue (Kühl et al. 1995; Enriquez et al. 2005). Multiple scattering in corals induces changes in the light field due to photon trapping (Kühl and Jørgensen 1994), which can increase the absorption efficiency of pigments significantly (Enriquez et al. 2005). The pigmentation of the coral tissue is critical to light-associated measurements such as variable chlorophyll a fluorescence measurements. Due to a strong contribution of multiple scattering by the skeleton (Enriquez et al. 2005), a highly pigmented area may induce less scattering of light than a more translucent tissue region. Similarly, fluorescence in lightly pigmented tissues may also be affected by multiple scattering. How these phenomena affect the measured fluorescence yields is yet to be determined.
The absorption maximum of chlorophyll a at 675 nm in the sun- and shade-adapted polyps was similar, but at other wavelengths the absorption varied significantly: overall absorption was higher in shade-adapted than in sun-adapted polyps and coenosarc tissue (Fig. 3). The percentage of incident downwelling light absorbed would be expected to correlate with the product of the zooxanthellae areal density and the cellular pigment content (Falkowski et al. 1984; Stambler and Dubinsky 2004) although multiple scattering by the skeleton enhancing the light field should also be taken into account (Enriquez et al. 2005).
Surface scalar irradiance measurements of coenosarc tissues ranged from 120 to 180% of the incident irradiance between 450 and 750 nm, with little spectral difference observed between sun- and shade-adapted coenosarc tissues or sun-adapted polyp tissues. However, the differences in scalar irradiance measured between coenosarc and polyp tissues as well as between sun- and shade-adapted polyp tissues (Fig. 3), highlight the importance and need for light measurements to accurately determine the light microclimate, which is likely to be significantly different from the light climate measured on the tissue surface.
The O2 concentration at the coenosarc tissue surfaces (Fig. 4a) saturated at high irradiances, whereas polyp tissues did not fully saturate within the irradiance range of the experiment (Fig. 4b). This indicates that polyp tissues harbour zooxanthellae, which can retain photosynthetic activity at higher irradiance, e.g. in deeper parts of the polyp, where light is attenuated. In addition, oxygen produced in the thicker polyp tissues has a longer diffusion path and must cross a thicker DBL than oxygen produced in the coenosarc tissue and this may prolong the onset of O2 saturation measured at the tissue surface (Kühl et al. 1996; Stambler and Dubinsky 2004). The retraction of polyp tissues at high irradiances (Brown et al. 2002) and the immobility of coenosarc tissues suggest that zooxanthellae harboured in coenosarc tissues are less shaded and must have greater capacity to cope with high irradiances than polyp tissues (Ralph et al. 2002). The maximum O2 concentration obtained in sun-adapted polyp tissues was lower than in shade-adapted polyps (Fig. 4a) possibly due to a lower density of zooxanthellae in sun-adapted tissues (Falkowski et al. 1984) resulting in lower absorption of incident light (Fig. 3; Stambler and Dubinsky 2004).
We did not quantify respiration in our study. However, respiration in sun-adapted polyps is typically greater than in shade-adapted polyps (Coles and Jokiel 1977; Gattuso and Jaubert 1990) resulting in lower net oxygen production. While polyp tissues were severely O2 depleted after 10 min of darkness, the coenosarc tissues retained a minimum O2 level of 36±26% saturation (Fig. 4a, b). Polyp tissues are highly mobile and may, therefore, exhibit higher respiration than coenosarc tissue. The differences in oxygen dynamics between coenosarc and polyp tissue are, however, also due to boundary layer effects. The interaction of flow and coral topography causes heterogeneous and complex DBLs over coral tissues (Kühl et al. 1995). Thicker boundary layers occurring over polyp tissues as compared to coenosarc tissues (Shashar et al. 1993) cause a more severe mass transfer limitation resulting in faster oxygen depletion and lower oxygen levels in the polyp tissue during darkness (Fig. 4). Corals can alleviate such mass transfer limitation by expanding their polyp tissues during night-time (Patterson 1992).
The higher compensation irradiance of sun-adapted than of shade-adapted tissues also indicates that more O2 is consumed through respiration in high-light adapted tissues (Fig. 4a, b) (Coles and Jokiel 1977; Gattuso and Jaubert 1990). Both coenosarc and polyp tissues, irrespective of orientation, were supersaturated with O2 in moderate light (<190 μmol photons m−2 s−1) and the steady-state O2 curves (Fig. 4a, b) remained stable showing no indication of increasing oxygen consumption and/or decreasing photosynthesis (Fig. 5) at high scalar irradiance. Kühl et al. (1995) found no sign of photoinhibition in an Acropora species even at irradiances >2,000 μmol photons m−2 s−1. This indicates that zooxanthellae can remain actively photosynthesising over a wide irradiance range, but see our caution about ETR measurements below. Whether this capacity is mainly physiological or is at least in part due to host factors such as protective pigments (Salih et al. 2000), alleviation of inhibitory irradiances due to tissue contraction and other issues related to the optical properties of the tissue, remains to be studied in further detail. The population of P. damicornis colonies at Heron Island harbours a genetically uniform but physiologically flexible zooxanthellar community with regard to photo-acclimation. Such flexibility could be of vital significance following bleaching events where corals have lost symbionts due to combined changes to light and temperature conditions (Baker et al. 2004).
Steady-state relationships of gross photosynthesis rate, O2 concentration and rETR
The linear correlation between gross photosynthesis rates and O2 concentration (Fig. 7) coupled with the difference observed in O2 concentration with increasing scalar irradiance in sun-adapted polyp tissues (Fig. 4b), confirms that oxygen consumption is relatively higher in sun-adapted compared to shade-adapted polyp tissues. This is also reflected in the slope of the correlation line of gross photosynthesis rate in sun-adapted polyps as a function of O2 concentration, which was more than twice that of shade-adapted polyps (Fig. 7b).
In contrast to the high gross photosynthesis rates and O2 concentration observed at high scalar irradiances, shade-adapted coenosarc and polyp tissues showed a decline in rETR over the course of the steady-state light curve, which would indicate that electron transport through the photosystems became severely restricted (Fig. 5b, d). The lack of correlation observed (Figs. 5, 6) suggests that variable chlorophyll a fluorescence measurements underestimated photosynthesis productivity in corals. Lack of correlation has previously been shown for free-living algae (Geel et al. 1997), a macroalga (Longstaff et al. 2003) as well as for higher plants (Harbinson et al. 1990) but in these cases rETR was higher than oxygen rates and several other studies have shown a good correlation between rETR and either 14C- or O2 production-based measurements of photosynthesis. Thus, the mismatch between rETR and gross photosynthetic O2 production in corals is an important observation. Variable chlorophyll a fluorescence-derived measures of photosynthesis are often used as a proxy for photosynthetic competence and primary production, e.g. in connection with coral bleaching studies. In the following we discuss some possible mechanisms for the observed lack of correlation.
The non-linearity observed between combined measurements of O2 and rETR are reflected as a dramatic decrease in rETR of the coral tissue at [O2] ∼250–325 μmol l−1 (Fig. 8). This observation suggests that an alternate pathway for electron transport is entrained rather than being a product of greater respiration, photorespiration (Schnitzler-Parker et al. 2004) or photo-inactivation. It is unlikely that the Mehler ascorbate-peroxidase (MAP) reaction, which has oxygen as a terminal electron acceptor (Geel et al. 1997; Asada 1999) can account for this effect since the net oxygen production from water is offset completely by the uptake of oxygen in the MAP reaction and therefore oxygen exchange is zero. Here, there must be a flow of electrons to a sink that is maintaining a high rate of oxygen evolution which is apparently not measured by the PAM system at high irradiances. Down regulation of absorbed light energy through the xanthophyll cycle would be a sink for light energy but not for electrons. Thus the only other electron pathway that could possibly cause such an alteration to the relationship between O2 synthesis and rETR appears to be cyclic PSII electron flow (Lavaud et al. 2002; Franklin and Badger 2001; Longstaff et al. 2002). Variable chlorophyll a fluorescence monitors only the rate of charge separation of PSII reaction centres leading to electron transport, which may not necessarily correlate directly to O2 evolution. Also, there is no correlation between decreasing rETR at high irradiance (Fig. 5a–d) and increased non-photochemical quenching in coenosarc tissue (Table 1). This also suggests that there is non-cyclic ETR in PSII. Nevertheless, it is curious that oxygen evolution appears to be proceeding at near maximum rates while the electron transport rate approaches zero.
Our combined measurements were done with the tip of the oxygen microelectrode and the microfibre positioned at the tissue surface. Discrepancies between surface measurements of variable fluorescence and “true” photosynthesis rate have been investigated in some detail by researchers using such techniques to assess microphytobenthos productivity (Perkins et al. 2002; Forster and Kromkamp 2004) and our observations could in part be due to similar effects. The steady-state oxygen level measured at the tissue surface is an integral measure of net balance of photosynthesis and respiration in the tissue below. In contrast, the gross photosynthesis measurement has a spatial resolution of about 100 μm (Revsbech and Jørgensen 1983) and thus quantifies zooxanthellar photosynthesis close to the tissue surface. The microfibre probe measures chlorophyll a fluorescence in front of the fibre tip. The volume from which fluorescence is sampled depends on the optical properties of the tissue. In less pigmented more translucent tissues variable fluorescence from a larger tissue volume is sampled as compared with measurements on the surface of tissues with a higher pigment density and/or with the zooxanthellae positioned further from the tissue surface. A decrease in relative ETR at higher irradiance could thus be due to contraction of tissue with zooxanthellae or full inhibition of zooxanthellae close to the tissue surface, while fluorescence from zooxanthellae in deeper layers is not detected by the microfibre. Such mechanisms would, however, be most pronounced in polyp tissue and cannot explain the decrease in rETR observed in the less contractile and more optically transparent coenosarc tissue. Furthermore, our O2 microelectrode measurements showed no inhibition of gross photosynthesis close to the tissue surface. Clearly, there is a need for a closer investigation of this interesting phenomenon in corals.
The ability of coral polyps to contract and expand with environmental and physical stimuli has been shown to affect the local microenvironment (Kühl et al. 1995). The present study showed that, combined microsensor measurements were generally more variable in polyp tissues. The relationship between the estimated rate of relative PSII electron transport and the gross photosynthesis rate as well as O2 concentration was found to be non-linear at high irradiances. Data obtained using variable fluorescence, especially in light stress conditions should thus be interpreted with caution and not be assumed to represent productivity. Further studies are needed to resolve the apparent discrepancy between photosynthetic oxygen production and fluorescence-derived measures of the rate of charge separation at PSII. Meanwhile, we caution against direct extrapolation of variable fluorescence data to actual rates of photosynthesis in corals exposed to high irradiance.
We thank A. Glud for microsensor construction, N. Ralph for fabricating the flow-through chamber, and Dr B. Kelaher for advice on statistical data processing. We thank S.H. Magnusson for assistance with collecting the scalar irradiance data. We acknowledge R. Hill for editorial comments on drafts of the manuscript and the Department of Environmental Science, UTS and all staff at Heron Island Research Station for support. The research was funded jointly by a GBRMPA Science for Management award and a PADI grant to KEU, a UTS institutional grant to PJR, grants from the Australian Research Council to AWDL and PJR, and a grant from the Danish Natural Science Research Council to MK. The research was conducted under GBRMPA permit number G04/12776.1. This is contribution number 207 from the Institute of Water and Resource Management.