It is usually assumed that CO2 produced by respiration in woody tissues escapes almost immediately into the atmosphere. Recent measurements indicate that this assumption is often wrong and a large quantity of respired CO2 remains inside the tree. An emerging understanding of the processes involved also indicates that a substantial portion of this CO2 dissolves in xylem sap and is transported away from the site of origin. The amount transported is diurnally and seasonally variable and depends on a complex of factors, including sap velocity, temperature, and pH as well as anatomical barriers to diffusion. We now know that a portion of this CO2 can diffuse to the atmosphere during transport and that another portion can be fixed by photosynthetic cells in woody tissues or leaves. A new approach for estimating woody tissue respiration, involving measurement of both internal and external CO2 fluxes, indicates that the rate of CO2 efflux to the atmosphere is often very poorly correlated to the actual rate of respiration. Quantification of internal CO2 transport, as well as recognition of the factors contributing to its complexity, may help us understand why the apparent rate of stem respiration, as determined by CO2 efflux measurements, has been so variable and difficult to predict for woody tissues within the same tree, among trees of the same species, among different species, and across seasons.
II. Quantity of CO2 in tree stems
A tree can contain a substantial amount of gas. Gas can be found in the intercellular spaces, the pore spaces within cell walls, and the lumens of cells (Gartner et al., 2004). Sachs (1887) calculated that the gas in fir trees (unidentified species) made up 17% of the xylem volume. According to MacDougal et al. (1929), trunks of Quercus sp. and Salix sp. contained approximately 25% gas by volume. Gartner et al. (2004) concluded that the sapwood of temperate softwoods and hardwoods contains 18 and 26% gas by volume, and the heartwood of temperate softwoods and hardwoods contains 50 and 26% gas, respectively.
The composition of gases inside tree stems differs greatly from that in ambient air. The first analysis of gas extracted from a tree stem was made in the early part of the 20th century by Bushong (1907). He withdrew gas from a hole drilled in a Populus deltoides stem and found it consisted of 7% CO2, 1% O2, 61% CH4 and 31% N2. Since then, published reports of stem CO2 concentrations ([CO2]) have been sporadic (Table 1). Measurements of stem [CO2] have ranged from < 1% to over 26%, that is c. 30–750 times higher than ambient atmospheric [CO2], indicating substantial barriers to diffusion of CO2 from the stem into the atmosphere.
Table 1. Carbon dioxide concentrations ([CO2]) measured in situ in stems or branches of trees
Measurements were made using microelectrodes or nondispersive infrared (NDIR) sensors, or in extracted gas samples using chemical absorption (CA), gas chromatography (GC), mass spectrometry (MS), or infrared gas analysis (IRGA). All measurements were made in the gas phase and are expressed as a percentage, except where noted *, in which case they are expressed as mmol l−1 (liquid phase).
In addition to CO2 in the gas phase, a large quantity of CO2 in stems is dissolved in xylem sap. The gaseous phase in the xylem is in contact with the liquid phase (Hari et al., 1991) and an equilibrium exists between the CO2 concentration of the gaseous phase and the total amount of all forms of dissolved CO2 in the liquid phase in the xylem.
According to Henry's law, the partial pressure of a gas over a solution is proportional to the concentration of that gas in the solution (Stumm & Morgan, 1996):
(, total dissolved inorganic carbon in the solution, which is the sum of [CO2]aq, [H2CO3], and ; K1 and K2, the first and second acidity constants, respectively; KH, Henry's constant; pCO2, partial pressure of CO2 over the solution). Using Henry's law, it is possible to convert measured CO2 concentrations in the gaseous phase ([CO2]) to total dissolved CO2 in the liquid phase (, mol l−1). Details of these calculations can be found in McGuire & Teskey (2002). Therefore, it is possible to determine the amount of CO2 dissolved in xylem sap by measuring the [CO2] of gas in holes drilled into the stem and this technique has been commonly used. In this review, CO2 concentrations in the gaseous phase will be indicated as [CO2] (%) and the total amount of all forms of dissolved CO2 in the liquid phase as (mol l−1).
The total amount of dissolved CO2 in xylem sap depends on temperature, pH and the [CO2] of the gaseous phase in contact with the sap (Fig. 1). The constants K1, K2 and KH in Eqn 1 are temperature-dependent. The solubility of CO2 decreases with increasing temperature, so 1%[CO2] is equivalent to 0.73 mmol l−1 at 10°C but only 0.35 mmol l−1 at 40°C (Fig. 1a). pH has a more substantial effect on the quantity of CO2 in solution (Fig. 1b). The primary form of dissolved CO2 is carbonic acid (H2CO3) and the quantity of CO2 in that form is independent of pH. However, at pH values of 5.6 or greater, significant quantities of CO2 can also dissolve as . At yet higher pH, >8.5, formation of a third species allows even more CO2 to dissolve, but tree xylem sap pH has not been reported that high. At 1% CO2, the total quantity of CO2 dissolved in xylem sap at pH 5 and 25°C is 0.35 mmol l−1, but at pH 7 the quantity increases fivefold, to 1.83 mmol l−1. Sap pH in stems has been reported in the range of 4.5–7.4 among different species and can vary seasonally by as much as 2.3 pH units (Table 2).
Table 2. pH values of xylem sap of trees
In the first seven studies, pH was measured on sap expressed from excised twigs using a pressure chamber. In the last six studies, stems were cored with an increment borer and sap was expressed with a vise. All species have diffuse porous anatomy except P. taeda (tracheid anatomy).
Early methods, which required extraction of gas samples, limited the frequency of sampling. However, when seasonal variation was reported, [CO2] increased during the growing season and decreased during the dormant season (MacDougal, 1927; MacDougal & Working, 1933; Chase, 1934; Eklund, 1990, 1993), giving support to the idea that the source of the CO2 inside the stems was respiration of living cells. For example, MacDougal (1927), who was the first to study seasonal variations, reported that [CO2] in Quercus agrifolia stems increased from 1.4% in January to 15.1% in August. MacDougal & Working (1933) observed high [CO2] in the stems of several tree species during periods of rapid growth. Chase (1934) found similar seasonal variation in [CO2], which followed the annual pattern of mean air temperature. More recently, Eklund (1993) reported that [CO2] in stems of Quercus robur increased in early June to almost 6%, decreased in early July to 3%, and rose to a second peak of almost 9% in late August, while in Acer platanoides, [CO2] consistently remained c. 3–4% throughout the growing season, except for a drop to below 2% in late June.
Most measurements of xylem [CO2] have been confined to the lower part of the stem, so little is known about within-tree variation. Chase (1934) made measurements at six heights up to 10 m in a Populus deltoides stem and found the highest [CO2] in the central part of the trunk, the lowest [CO2] in the lower trunk, and intermediate [CO2] in the upper trunk during the summer months. Overall, stem [CO2] declined substantially through the autumn and winter months, but the central part of the trunk continued to have higher [CO2] than the upper and lower trunk. When [CO2] began to increase in March and April, this pattern changed; the highest [CO2] was found in the lower trunk and the lowest [CO2] was found in the upper trunk. Chase (1934) also measured [CO2] at different depths in the stem of a Populus deltoides tree. He found that [CO2] in the sapwood was always lower than in the heartwood. Teskey & McGuire (2007) found that mean of Platanus occidentalis stems was greater at 3 m height than at 0.1 m during the summer. Over small stem height intervals (< 1 m), stem [CO2] was often observed to increase in Fagus grandifolia, Liquidambar styraciflua (McGuire & Teskey, 2004), Platanus occidentalis (McGuire & Teskey, 2004; Teskey & McGuire, 2007) and Populus deltoides (Saveyn et al., 2007c).
III. Sources of CO2 found in stems
1. Respiration of stem and roots
Most CO2 within the stem originated from respiring cells in the stems and roots. In the stem, live cells are found in the inner bark, vascular cambium, and xylem (Fig. 2). The outer bark, or rhytidome, does not contain living cells, but the inner bark, consisting of an outer layer of peridermal cells and phloem, has a substantial number of living cells, as does the thin layer of cells in the vascular cambium between the phloem and xylem. For simplicity, in this review, ‘inner bark’ refers to cells from the periderm through the vascular cambium. In the xylem, only ray cells in the sapwood are living, and these are the most likely local source of the CO2 found within the stem. The volume of ray cells in the xylem of conifers ranges from 5 to 9%. Hardwoods generally have more ray volume, ranging from 5 to 34%, and in many of these species rays constitute 15–25% of the xylem volume (Panshin & de Zeeuw, 1980). The proportion of live cells in the xylem may change with tree size. In Fagus sylvatica trees, xylem ray volume decreased from approx. 25% to 20% as stem diameter increased from 6 to 16 cm, and from 48% in 1 mm branches to 21% in 8 mm diameter branches (Ceschia et al., 2002). The proportion of live cells decreases rapidly with depth from the cambium in some species (Stockfors & Linder, 1998), but remains constant with depth in other species (Ryan, 1990; Ceschia et al., 2002).
It is not known if the inner bark in woody tissues contributes to the high [CO2] within the stem, but CO2 may move inward from this region under some circumstances. Live cell volume in the inner bark has been reported to range from 56% in Fagus sylvatica (Ceschia, 2001) to 7% in Pinus contorta (Ryan, 1990). Growth rate may also affect live cell volume. In Picea abies, live cell volume increased from 21 to 30% when trees were fertilized with nitrogen (Stockfors & Linder, 1998).
The relative distribution of live cells between the inner bark and xylem in stems depends on species and tree size. In Pinus contorta and Picea engelmanii trees with diameters ranging from 4 to 40 cm, the xylem contained > 80% of the total live cell volume in stems (Ryan, 1990). Similarly, in Fagus sylvatica stems with diameters up to 16 cm, the live cell volume of the xylem and the live cell volume of the entire stem were very similar, indicating that the quantity of living cells in the inner bark was negligible compared with the xylem (Ceschia et al., 2002). By contrast, in 7–10 cm Picea abies trees, only 20–25% of the live cell volume of the stems was found in the xylem (Stockfors & Linder, 1998).
The rate of cellular respiration may differ substantially with radial depth in the stem. Measurements of respiratory capacity have not yet been made in situ, and knowledge at this time is limited to the rate at which various excised tissue samples can evolve CO2 or consume O2 under controlled environmental conditions. In some instances, the rate of CO2 efflux of inner bark tissues has been shown to be substantially higher than that of sapwood (Pruyn et al., 2002a,b, 2003). In other studies, the rates of CO2 efflux of inner bark and outer sapwood were similar (Zabuga & Zabuga, 1990; Bowman et al., 2005). In the few studies that examined within-sapwood variation, the rate of CO2 efflux of the outer sapwood was almost always higher than that of inner sapwood (Pruyn et al., 2002a,b, 2003; Bowman et al., 2005), but Zabuga & Zabuga (1990) found no change in the rate of CO2 efflux with sapwood depth in Pinus sylvestris stems. Higher respiratory activity of inner bark and outer sapwood may be related to availability of carbohydrates or nitrogen concentrations (Pruyn et al., 2002b, 2005). The nitrogen concentration of Dacrydium cupressinum stems was highest in the inner bark and declined with increasing depth into the stem (Bowman et al., 2005). Total nitrogen content and total nonstructural carbohydrate content serve as a proxy for enzyme quantity and metabolic substrate availability, respectively. Variation in oxygen and CO2 concentrations in the stem may be an additional source of spatial and temporal variation in respiration. Averaged across five tree species, the mean rate of oxygen consumption of excised sapwood tissue was reduced by as much as 27% when [O2] was reduced from 10 to 5%, and by 14% when [CO2] was increased from 0 to 10%, but the response to various combinations of gas concentrations was complex, and the effect of [O2] on tissue oxygen consumption depended on the [CO2] and vice versa (Spicer & Holbrook, 2007).
Pruyn et al. (2003) measured CO2 efflux from stem cores of 10 tree species in the central Cascade Mountain range in Oregon and found that efflux at 25°C varied three-fold among species. CO2 efflux was inversely correlated to the live bole volumetric fraction (LBVF) (inner bark plus sapwood divided by whole bole volume). Tissues from tree species with < 20% LBVF had 1.3 to three times greater CO2 efflux than did species with > 40% LBVF, suggesting that species that maintain large sapwood volumes also have less metabolically active stem tissues.
The contribution of root respiration to the [CO2] in the stem is uncertain. A recent observation on Plantanus occidentalis trees showed that the [CO2] at the base of the stem was almost as high as the [CO2] at 3 m height, leading to speculation that much of the CO2 in the stem may have originated below ground (Teskey & McGuire, 2007). However, the [CO2] at the base of the stem was too high to be attributed solely, or even mostly, to absorption of CO2 from the rhizosphere. It is known that roots, especially of species resistant to waterlogged or flooded conditions, can tolerate high [CO2] in the root zone, but there is no information on root [CO2] (Greenway et al., 2006). Effective anatomical and morphological features in roots create barriers for radial O2 loss to the soil (Colmer, 2003). In tree roots, suberin formation on the root exodermis has been shown to reduce the efflux of O2 (De Simone et al., 2003) and may serve a similar function for efflux of CO2. Diffusion of O2 from the base of the stem into the root system through aerenchyma is common in many flood-tolerant plants (Colmer, 2003; Armstrong & Armstrong, 2005), and could also be a pathway for CO2 movement upward from the roots to the stem, but aerenchyma tissue is poorly developed or nonexistent in upland tree species.
Salix aquatica was capable of absorbing 14C-labeled dissolved inorganic carbon (DIC) from a nutrient solution and transporting it throughout the plant (Vapaavuori & Pelkonen, 1985; Vuorinen et al., 1989). Ford et al. (2007) demonstrated that 13CO2 dissolved in soil water can be absorbed and fixed in Pinus taeda seedlings. More than 50% of the 13CO2 was recovered above ground, with about two-thirds fixed in stem tissues and one-third fixed in leaf tissues. However, the DIC supplied from the soil contributed < 1% to total seedling C gain. Similarly, Hibberd & Quick (2002) supplied inorganic carbon to roots and stems of Nicotiana tabacum plants as [14C]NaHCO3 and found that 14C accumulated in organic compounds in cells of stems and petioles.
IV. Fate of CO2 released by respiring cells in stems: loss, transport and fixation
1. Loss to the atmosphere
CO2 produced by the living cells of xylem, cambium and inner bark (Fig. 2, arrows 1c, 1b and 1a, respectively) or CO2 imported in the transpiration stream (arrow 1d) can diffuse radially from the stem to the atmosphere. The diffusion coefficient of CO2 in air is 1.6 × 10−5 m2 s−1, whereas it is only 1.6 × 10−9 m2 s−1 in water at 20°C and 101.3 kPa (Nobel, 1999), so radial diffusion of CO2 in the gas phase will predominate. The flux of CO2 diffusing out of the stem into the atmosphere can be characterized by a modified version of Fick's law of diffusion (Pfanz & Aschan, 2001):
(EA, flux of CO2 out of the stem (µmol m−2 s−1); Δ[CO2], difference in CO2 concentration between the gas spaces of the stem and the atmosphere; R, sum of a series of diffusion resistances). The sum of resistances depends on whether the CO2 is located in the inner bark (Fig. 2 (arrow 1a)) or in the xylem (Fig. 2 (arrows 1c and 1d)).
From the inner bark to the atmosphere, diffusion resistances are mainly determined by the composition of the cortex, phloem, peridermal layers (phelloderm, phellem and phellogen) and rhytidome (i.e. the outer bark) (Ziegler, 1957). These tissues protect the tree against water loss, insects, and pathogens, but also limit the exchange of CO2 and O2 with the atmosphere. Resistance depends mainly on the amount of lignins, suberins, lipids, and waxes and on thickness (Schonherr, 1982; Lendzian, 2006) and appears to differ substantially with species and tissue age. Mean CO2 conductance of the bark of small branches of Pinus monticola was 0.68 mmol m−2 s−1 (Cernusak & Marshall, 2000) and of the bark of 6-yr-old Betula pendula trees was 0.50 mmol m−2 s−1 (Wittmann et al., 2006). However, CO2 conductance of bark of Syringa vulgaris stems was substantially higher and proportional to age and bark thickness, ranging from 33, 18 and 10 mmol m−2 s−1 for current-year (thinnest bark), 1-yr-old and 3-yr-old (thickest bark) stems, respectively (Pilarski, 1994).
In most tree species, the cambium located between the xylem and inner bark significantly inhibits lateral gas movement (Kramer & Kozlowski, 1979). This may explain the large difference in [CO2] between xylem and bark tissues: xylem [CO2] has been reported to be as high as 26% (Table 1), whereas published values of [CO2] in bark tissues range between 0.06 and 0.17% (Cernusak & Marshall, 2000; Wittmann et al., 2006). In some tree species adapted to waterlogged soil, the cambium may contain small intercellular spaces permitting radial CO2 diffusion (Hook & Brown, 1972; Buchel & Grosse, 1990). The permeability of the cambial zone may change seasonally. In Pseudotsuga menziesii, actively dividing cambium cells and newly generated xylem and phloem cells, which have thin cell walls, appear to be able to exchange CO2 and O2 more efficiently in spring than in December, when cells have thicker cell walls and increased protoplasmic viscosity (Joseph & Kelsey, 2004).
Presumably, CO2 diffusion through inner and outer bark is also facilitated by lenticels, cracks and wounds (Grosse, 1997; Langenfeld-Heyser, 1997). Groh et al. (2002) found that lenticels from Betula potaninii and Sambucus nigra were significantly more permeable to O2 than phellem without lenticels. Groh et al. (2002) also demonstrated that the O2 permeability of lenticels changed seasonally, and was higher during the growing season and lower in autumn and winter. However, data on CO2 permeability are lacking, but we do know that removing some of the diffusional barriers in the bark by wounding dramatically increases the rate of CO2 efflux (Teskey & McGuire, 2005).
The xylem may be an even greater barrier to free gas exchange than the cambium. Sorz & Hietz (2006) found that minimum diffusion coefficients for O2 in water-saturated xylem were always lower than the diffusion coefficients in water alone, indicating that the cell walls presented a major barrier to gas diffusion.
Since diffusion occurs much faster in air than in water, diurnal and seasonal variation in the ratio of gas to water content in the xylem, cambium, and bark can also have an important influence on diffusion rates. During periods of diurnal stem shrinkage resulting from changes in water content, liquid water films are lost from gas spaces and the stem tissue may become more permeable to gas exchange (Hook et al., 1972). The volumetric water content of the sapwood also changes seasonally. For example, it ranged from nearly 50% during winter and early spring, to 20% during the growing season in Acer saccharum trees (Pausch et al., 2000). Sorz & Hietz (2006) measured radial diffusion of O2 in wood at different water and gas contents and found that the diffusion coefficient increased five- to 13-fold in Picea abies, Taxus baccata and Quercus robur, 36-fold in Fraxinus excelsior and c. 1000-fold in Carpinus betulus and Fagus sylvatica at 40% gas volume compared to the diffusion coefficient at 15% gas volume.
Recent studies have established a relationship between the internal stem [CO2] concentration and stem CO2 efflux. By directly manipulating the [CO2] in the xylem, it was possible to demonstrate that CO2 efflux was linearly related to the xylem [CO2] in large intact stems of Quercus alba and Liriodendron tuliperfera (Teskey & McGuire, 2002) and in excised shoots of Liquidambar styraciflua and Platanus occidentalis (Fig. 3) (Teskey & McGuire, 2005). These experiments indicated that although there are substantial barriers to diffusion in the stem, some dissolved CO2 diffuses outward to the atmosphere as it is transported upward in the xylem. However, there is some evidence that this relationship may not exist under all conditions or in all species. In young Pinus taeda tree stems, when transpiration was reduced by removing foliage, xylem [CO2] increased, but there was no apparent change in CO2 efflux (Maier & Clinton, 2006). The authors suggested that during spring, the meristematic region might respire at a much higher rate than the xylem parenchyma and thus would be the major source of CO2 efflux during periods of active growth (Fig. 2, arrows 1a,b, 3a,b). Saveyn et al. (2007b) measured xylem sap [CO2] and stem CO2 efflux in a young Quercus robur tree subjected to soil water depletion and found that [CO2] was positively correlated with CO2 efflux in the predrought period, but this correlation became less clear during early drought stress. Wittmann et al. (2006) concluded that the contribution of xylem CO2 to efflux was negligible in small branches (< 2 cm diameter) of Betula pendula because small branches have a high surface area to volume ratio, so the proportion of live cells in the inner bark would be much larger than the proportion in the xylem. In stems of three individuals of a single Populus deltoides clone, Steppe et al. (2007) found a strong linear relationship between xylem [CO2] and efflux, but the slope varied among the stems, indicating substantial differences in the relationship between xylem [CO2] and efflux, even within individual ramets of the same clone. This result suggests that barriers to diffusion of CO2 may vary among stems and perhaps in different parts of the same tree.
The rate of diffusion of CO2 from the stem to the atmosphere will depend on the concentration gradient, barriers to diffusion, and temperature, since the diffusion coefficient of a gas depends on temperature raised to the power 1.8 (Nobel, 1999). Stockfors & Linder (1998) calculated time lags between changes in stem temperature and stem CO2 efflux in Picea abies. The shortest lag was observed in June and the lag increased in duration through the summer to a maximum in September. Perhaps these results were influenced by seasonal changes in tissue permeability or the effect of temperature on the rate of diffusion, or both.
2. Transport of dissolved CO2 in xylem sap
Observations of midday depression in CO2 efflux from stems has led to speculation that reduced carbohydrate availability (Azcon-Bieto et al., 1983) or stress caused by water deficits (Lavigne, 1987; Kakubari, 1988; Wang et al., 2003; Daudet et al., 2005; Saveyn et al., 2007a,b) lowered the rate of cellular respiration within stems and reduced the rate of CO2 efflux. In some instances, corticular photosynthesis (Sprugel & Benecke, 1991) could also contribute to reduced efflux during daylight hours. An alternative explanation was that a portion of the CO2 released by respiration of local cells dissolved in xylem sap and was transported away in the transpiration stream (Geurten, 1950; Negisi, 1975). Negisi (1979) provided evidence that CO2 efflux from the stem was inversely correlated with sap velocity. In a manipulative experiment on excised stems of Pinus densiflora trees, he found that CO2 efflux was inversely proportional to sap velocity, decreasing by 50% when sap velocity increased from 0 to 0.4 cm min−1. Under field conditions, CO2 efflux of mature stems of Betula pendula can be reduced by 40% at high rates of transpiration (Gansert & Burgdorf, 2005). A strong negative correlation between stem CO2 efflux and sap flux density was found in Dacrydium cupressinum trees, but, interestingly, the slope of the relationship differed substantially from tree to tree, suggesting that diffusional gradients or stem [CO2] varied among these individuals (Bowman et al., 2005). A smaller reduction in CO2 efflux (7%) was observed in Pinus taeda seedlings during periods of high transpiration compared with periods of low transpiration (Martin et al., 1994). In these studies, as well as others in which midday depression in CO2 efflux was observed (Edwards & McLaughlin, 1978; Kaipiainen et al., 1998), it was not possible to separate the relative importance of the effect of CO2 transport on CO2 efflux vs factors that reduced the actual rate of respiration.
Recently, McGuire et al. (2007) demonstrated that under constant temperature conditions, and thus presumably a constant rate of respiration, changes in sap velocity caused large differences in CO2 efflux in detached Platanus occidentalis branch segments (Fig. 3). At extremely high sap velocities (> 10 cm min−1), CO2 efflux accounted for < 10% of the total CO2 released by respiring cells in the branch. At a more realistic maximum sap velocity (1 cm min−1) (Wullschleger & King, 2000), efflux accounted for < 50% of the total CO2. However, correlations between sap flow and CO2 efflux are not always negative. In a field study, Levy et al. (1999) found a positive relationship between the rate of sap flow and CO2 efflux, which they speculated was caused by water at high [CO2] rising from the root system. In some instances, CO2 efflux from stems was not affected by sap flow rates (Edwards & Wullschleger, 2000; Maier & Clinton, 2006). Under uncontrolled field conditions, diurnal changes in tissue temperature, water stress, carbohydrate availability and sap velocity vary in complex ways and can affect rates of both woody tissue respiration and CO2 efflux, making it difficult to separate the importance of individual causal factors. Additionally, tissue size and time of year may affect the relative contributions of CO2 diffusing from the xylem and CO2 released by local respiring cells to CO2 efflux from woody tissues.
The rate of sap flow can also have a substantial influence on xylem sap [CO2]. A tight inverse coupling was observed between sap velocity and sap [CO2] in the stem of a Liriodendron tulipifera tree (Fig. 4) (Teskey & McGuire, 2002). Xylem [CO2] was high during the night, when sap velocity was zero, and decreased during the morning as sap velocity increased. During brief periods of cloud cover during the day, sap velocity decreased and xylem [CO2] increased concurrently. As sap velocity decreased in late afternoon, xylem [CO2] increased in proportion. Similarly, under controlled temperature conditions, xylem [CO2] of Platanus occidentalis branches decreased rapidly as sap velocity increased from 0 to 1 cm min−1 and remained low at higher sap velocities (Fig. 5) (McGuire et al., 2007). Xylem [CO2] and sap flux in young Pinus taeda tree stems were also well coupled (Maier & Clinton, 2006). They found that removing most of the foliage caused a decrease of sap flux to near zero. Simultaneously, [CO2] increased substantially. Similarly, during periods of rain, when sap flux is greatly reduced, xylem [CO2] has been reported to double in stems of Populus deltoides, compared with xylem [CO2] on sunny days (Saveyn et al., 2007c).
The existence of bark and wood photosynthesis and its significance for overall carbon gain has recently been reviewed (Pfanz & Aschan, 2001; Aschan & Pfanz, 2003). Corticular photosynthesis has been reported in a wide variety of tree species (Table 3). These studies indicate that a substantial amount of internal CO2 can be fixed by chloroplast-containing tissues.
Table 3. Reported values of maximum stem CO2 refixation rates
Maximum re-fixation rate (%)
na, not available.
Refixation was estimated from CO2 efflux in the dark (EA(d)) and in the light (EA(l)): % refixation = (EA(d) –EA(l))/EA(d) × 100. Values higher than 100% indicate net CO2 uptake.
Corticular and wood photosynthesis have two advantages over leaf photosynthesis: since no stomata are involved, corticular and wood photosynthesis have little associated water loss; and since [CO2] in woody tissues is high, photorespiration is low (Cernusak & Marshall, 2000). In contrast with leaf photosynthesis, CO2 is not a limiting factor for corticular photosynthesis. However, high [CO2] can reduce the photochemical efficiency of photosystem II and may increase photoinhibition (Manetas, 2004). Corticular and wood photosynthesis may be critical for maintaining sufficient O2 concentrations in stems to avoid anoxia (Pfanz et al., 2002). Corticular photosynthesis can fix up to 100% of the respiratory CO2 emissions (Cernusak & Marshall, 2000) and may occasionally exceed CO2 release; that is, stems may also take up CO2 from the atmosphere (Damesin, 2003; Berveiller et al., 2007) (Table 3). Refixation of internal CO2, especially in young trees, compensates for a portion of respiratory carbon loss.
An important limiting factor for corticular and wood photosynthesis is irradiance. For photosynthesis to occur, adequate photosynthetically active radiation must pass through the epidermal, peridermal, and/or rhytidomal layers to reach the light-harvesting complexes of the chloroplasts. A study of woody tissues of 21 species found that light was conducted axially as well as radially in vessels, fibers, tracheids and parenchyma cells, but light quality was modified and its intensity was attenuated proportional to the distance conducted (Sun et al., 2003). The amount of light transmitted in stems of young trees varies between < 10 and 50% of incident light, depending on species and age of the stem segment (Aschan & Pfanz, 2003). Light transmission and the capacity for fixation usually decrease as stems age and develop a thick layer of outer bark.
V. What do we know about woody tissue respiration?
Beginning in the early part of the 20th century, CO2 efflux from tree stems was suggested as an estimate of stem respiration (Johansson, 1933). In a critique of Johansson's (1933) technique, Boysen-Jensen (1933) raised the concern that CO2 efflux from the stem may not be equivalent to the CO2 released by respiration of woody tissues. Boysen-Jensen (1933) made the point that it was entirely uncertain that the CO2 produced by respiration of a segment of the stem would flux outward through the overlaying bark of that segment. He gave two reasons for this uncertainty: (i) the bark of trees was very often impermeable to air, and (ii) transpiration could easily carry the CO2 upward, where it could flux outward from parts of the tree other than where it was produced. Despite these criticisms, the approach of Johansson (1933) was adopted as the principal field method for estimating stem respiration, presumably because it was nondestructive, allowing measurements on intact trees. With the advent of the use of infrared gas analyzers, the Johansson (1933) technique was modified by providing a stream of air through the cuvette (Pisek & Winkler, 1958), allowing real-time continuous measurements. Since then, CO2 efflux measurements have been reported as stem respiration in hundreds of papers. It is notable that although there is a long record of these measurements, finding consistent relationships between rates of woody tissue CO2 efflux and tissue sizes and types, tree ages, species, and environmental conditions has been difficult (Lavigne et al., 1996). Remarkable variation exists even among trees of the same age growing on the same site. For example, within trees, Sprugel (1990) found 10- to 40-fold differences between CO2 efflux rates in 30-yr-old Abies amabilis trees in the same stand. Similarly, Cernusak & Marshall (2000) found a 20-fold difference in the CO2 efflux among branches of Pinus monticola trees growing on the same site.
Since it is now well-established that a large amount of respired CO2 remains within tree stems and that CO2 is transported upward in the transpiration stream, we ask the question: how accurately do efflux measurements represent the rate of actual respiration of woody tissues?
To address this question, McGuire & Teskey (2004) developed a mass balance approach that accounts for all fluxes of CO2 from and within a stem segment: efflux to the atmosphere through bark (EA), dissolved CO2 entering (IT) and exiting (ET) the stem segment in flowing sap, and the increase or decrease in xylem CO2 concentration over time (ΔS). Therefore,
RS = EA + FT + ΔS(Eqn 3)
where FT (transport flux) =ET–IT; RS is µmol m−3 sapwood s−1) and
and fA is the rate of air flow through a cuvette surrounding the stem segment, v is the volume of the segment, Δ[CO2] is the difference in [CO2] of air entering and exiting the cuvette, fS is the flow rate of sap through the segment, Δ is the difference in of the xylem at the top and bottom of the segment, T1 and T0 are mean at T1 and T0, respectively, L is amount of water in the segment, and T is time interval. To calculate respiration from these equations, diurnal measurements of CO2 efflux, sap flux, xylem and xylem water content were made. Teskey & McGuire (2007) modified the EA calculation to account for efflux of CO2 transported into the stem segment from lower parts of the tree, as
where is calculated from a tissue-specific linear regression of efflux and measured at the bottom of the stem segment.
Using this approach, stem respiration was measured on five Platanus occidentalis trees during summer (Teskey & McGuire, 2007). Transport flux (FT) was the most important internal CO2 flux, and it was a major component of stem respiration during the daytime hours when sap was flowing, accounting for as much as 70% of the CO2 released from respiring cells at midday. At night, FT was negligible because sap flow ceased. However, at night, CO2 efflux to the atmosphere greatly overestimated stem respiration because it consisted of both CO2 produced by local cells and CO2 that had been transported upward in the transpiration stream from the roots and lower part of the stem.
Over a 24 h period, 55% of the CO2 efflux to the atmosphere consisted of CO2 that had been transported from lower in the tree. On average, over the 24 h period, when efflux was corrected for the flux of imported CO2, it represented 66% of the total respiration of the living cells in the measured segment. Perhaps more importantly, mean hourly uncorrected efflux, which included transported and locally derived CO2, was not related to stem respiration calculated by the mass balance approach (Fig. 7). Using a modified version of this technique, Bowman et al. (2005) also concluded that there was no significant relationship between CO2 efflux and the respiratory activity of stems of Dacrydium cupressinum trees.
Undoubtedly, estimates of CO2 efflux from stems to the atmosphere have contributed to our understanding of tree and ecosystem carbon budgets. However, we conclude that we have gained little specific knowledge about woody tissue respiration from these measurements. Some of the CO2 released by respiring cells does diffuse into the atmosphere at the point of origin, but the amount is often obscured by the simultaneous diffusion of CO2 that has been transported in the xylem from woody tissues lower in the tree or from the roots and rhizosphere. These processes cannot be separated easily because the concentration of CO2 in the xylem varies substantially both diurnally and seasonally, and is influenced by many factors simultaneously, including temperature, sap pH, barriers to diffusion in the xylem and bark, rates of sap flow, and the number and rate of activity of live cells in various tissues along the pathway. For woody tissues, CO2 efflux to the atmosphere is not equivalent to respiration.
VI. Why do trees have the capacity for internal capture of CO2?
The entrapment of [CO2] within tree stems may not have evolved as a carbon-conserving strategy, but rather may have its origins in the development of barriers designed to minimize water loss to the atmosphere. Regardless, a consequence of these barriers is that the CO2 released from respiration of living cells is prevented from rapidly diffusing out of the stem. Trees take advantage of this source of CO2 by fixing a portion of it in green woody tissues or leaves before it can diffuse to the atmosphere. Aschan & Pfanz (2003) concluded that internal recycling of CO2 is an important function of stems.
Although the [CO2] within tree stems is very high relative to that of the atmosphere, the overall quantity of CO2 available to be fixed by stem photosynthesis is often small compared with that fixed from the atmosphere via leaf photosynthesis. For example, carbon gain from corticular photosynthesis in midsummer has been reported to be 10–15% of total carbon assimilation (see reviews by Schaedle & Foote, 1971; Pfanz & Aschan, 2001). In some instances, a more substantial fraction of total plant carbon gain has been contributed by woody tissue photosynthesis. In a group of woody desert species, twig photosynthesis contributed over 80% of the annual plant carbon gain due in part to the relatively small investment in leaves in these species (Comstock et al., 1988). Similarly, Foote & Schaedle (1978) concluded that when leaves were absent on Populus tremuloides, corticular photosynthesis contributed substantially to the carbohydrate supply of the tree.
Trees are among the most productive plants on earth, yet they have the most massive support structures for their foliage of any plant form. Recycling of carbon released by the respiration of woody tissues reduces the carbon cost of these support structures. Stem photosynthesis can reduce the loss of CO2 from woody tissues to the atmosphere by 50 to > 100% (Table 3), suggesting that on a 24 h basis, 20–40% recovery of respired carbon may be feasible, at least in canopy branches. Also, the overall carbon gained from refixation has likely been underestimated if CO2 transported in xylem sap is also being refixed, since this carbon would not be accounted for by the current methods of calculating refixation; that is, by the difference in the rate of CO2 diffusion to the atmosphere in light and dark. Evidence for fixation of transported CO2 in the xylem was observed in an experiment in Alnus glutinosa in which supplementing the [CO2] in the xylem produced an increase in O2 output from the stem (Armstrong & Armstrong, 2005). Indirect evidence has also been seen in stem [CO2] measurements made in Fagus sylvatica and Populus deltoides in the dormant season (Saveyn et al., 2006; M. A. McGuire et al., unpublished). In both instances stem [CO2] decreased when the stem was illuminated, and increased when the stem was in the dark.
Recycling of respired CO2 may be particularly important during periods of environmental stress. Similar to desert species with low leaf areas (Comstock et al., 1988), corticular photosynthesis may become a substantial component of a tree's carbon gain during periods of severe water stress when daily leaf carbon gain may be negative, or when there has been substantial leaf abscission. The segmentation hypothesis of Zimmermann (1983) suggests that under water stress, the hydraulic architecture of trees allows them to excise extremities; first leaves, then twigs, then branches, etc., so that the main stem survives the longest. In a stress situation when a tree has few or no leaves remaining, corticular photosynthesis may be a substantial aid to survival.
Internal carbon recycling may have been a more important contributor to overall carbon gain at times when atmospheric [CO2] was lower. The [CO2] in the atmosphere has been estimated as low as 185 µmol mol−1 during the last glacial maximum 21 000 yr ago (Otto-Bliesner et al., 2006). Under those conditions, maximum rates of leaf net photosynthesis may have been only one-half of present values. Leaf transpiration rates would also have been higher as a result of either higher stomatal density or conductance (Ainsworth & Rogers, 2007), which would cause greater sap velocity and a higher proportion of CO2 transport in xylem sap (McGuire et al., 2007). Photosynthetic cells in the stem, surrounded by water at high [CO2], are presumably CO2-saturated, and may have contributed more to overall woody plant carbon gain under those circumstances. However, this hypothesis is speculative, as we are not aware of any studies that have examined it.
Corticular and wood photosynthesis may be more critical for maintaining sufficient O2 concentrations in stems to avoid anoxia than for carbon gain per se (Pfanz et al., 2002). The [O2] in xylem is often half that of ambient air, or c. 10% (Gansert et al., 2001; Spicer & Holbrook, 2005) and there are reports of it reaching concentrations of 5% or less (Jensen, 1969; Eklund, 2000). However, we are not aware of any comparisons between the quantity of oxygen released by stem photosynthesis and that supplied by either sap transport or diffusion through bark. Sap flux is considered the major source of O2 in the xylem (Gansert, 2003; Spicer & Holbrook, 2005) but the importance of O2 generated by stem photosynthesis may increase when sap flux is diminished. Gansert (2003) reported that when sap flow was reduced, the O2 deficit in Betula pubescens stems increased dramatically; under these conditions any O2 provided by stem photosynthesis would help in maintaining adequate rates of respiration.
There is a high concentration of CO2 in the xylem of tree stems that may be a consequence of anatomical and morphological adaptations to limit the loss of water from the xylem. Stem photosynthesis utilizes a portion of this CO2, providing aeration to living cells in the xylem, and offsetting the carbohydrate cost of building and maintaining the massive woody structural system of trees. A consequence of the retention and transport of CO2 in stems is that the respiration of woody tissues has not been accurately estimated by measurements of CO2 efflux to the atmosphere. To determine the actual rate of stem respiration, internal CO2 fluxes as well as CO2 efflux to the atmosphere must be measured or accounted for in some other way. Although the picture is not yet complete, the mass balance approach may provide a more accurate estimate of stem respiration than other methods. However, whatever methods are used, it is clear that we must move beyond our current reliance on CO2 efflux as a measure of woody tissue respiration.
The authors wish to thank the Research Foundation – Flanders (FWO) for the postdoctoral fellow funding granted to KS and for the travel funding granted to AS and KS. The authors also wish to thank the Special Research Fund (BOF) of Ghent University for the PhD funding granted to AS. Support for this project is gratefully acknowledged for ROT and MAM from the National Research Initiative of the USDA Cooperative State Research, Education & Extension Service (2003-35100-13783), the National Science Foundation (0445495) to ROT; and the Global Forest Foundation. We also thank Dr Barbara Lachenbruch for guidance on stem live cell volume data.
Quantity of CO2 in tree stems
Sources of CO2 found in stems
Fate of the CO2 released by respiring cells in stems: loss, transport and fixation
What do we know about woody tissue respiration?
Why do trees have the capacity for internal capture of CO2?