Root xylem CO₂ flux: an important but unaccounted-for component of root respiration

Abstract

Key message

In tree roots, a large fraction of root-respired CO ₂ remains within the root system rather than diffusing into the soil. This CO ₂ is transported in xylem sap into the shoot, and because respiration is almost always measured as the flux of CO ₂ into the atmosphere from plant tissues, it represents an unaccounted-for component of tree root metabolism.

Abstract

Root respiration has been considered a large component of forest soil CO₂ efflux, but recent findings indicate that it may be even more important than previous measurements have shown because a substantial fraction of root-respired CO₂ remains within the tree root system and moves internally with the transpiration stream. The high concentration of CO₂ in roots appears to originate mainly within the root. It has been suggested that plants can take up dissolved inorganic carbon (DIC) from soil, but under most conditions uptake from soil is minimal due to the root-to-soil diffusion gradient, which suggests that most of the CO₂ in root xylem is derived from root respiration. Estimates of the internal flux of CO₂ through root xylem are based on combined measurements of sap flow and internal [CO₂]. Results quantifying root xylem CO₂ flux, obtained for a limited number of species, have raised important concerns regarding our understanding of tree respiration. Taken together, the results of these studies call into question the partitioning of ecosystem respiration into its above- and belowground components, and redefine the energetic costs of tree root metabolism and hence estimates of belowground carbon allocation. Expanding our observations of root xylem CO₂ flux to more species and at longer time scales, as well as improving the techniques used to study this process, could be fruitful avenues for future research, with the potential to substantially revise our understanding of root respiration and forest carbon cycles.

Introduction

A common belief among scientists working on belowground respiration in forests is that root-respired CO₂ diffuses into the soil environment and thereby contributes to CO₂ efflux from soils. Recent results (Aubrey and Teskey 2009; Grossiord et al. 2012; Bloemen et al. 2014a) indicate that a fraction of belowground respired CO₂ may remain within the root system rather than diffusing into the soil. For instance, Aubrey and Teskey (2009) estimated that twice the amount of CO₂ derived from root respiration was transported via the transpiration stream as diffused into the soil environment. As respiring cells in roots lie in close proximity to xylem tissue, respired CO₂ can dissolve in xylem water within the roots and be transported above ground via the transpiration stream rather than diffusing into the soil, implying that we are substantially underestimating the autotrophic component of belowground respiration.

The root xylem CO₂ flux have been estimated only for a few species over short time periods of up to 1 week (Aubrey and Teskey 2009; Grossiord et al. 2012; Bloemen et al. 2014a). However, the results from these studies all indicate that root xylem CO₂ is a very important, but unaccounted-for, component of root and tree respiration. The ramifications of these new results are that the energetic cost of tree root metabolism is much larger than previously assumed and because the largest portion of root-derived CO₂ diffuses into the atmosphere aboveground from tree stems and branches (up to 94 %, Bloemen et al. 2013b), the current understanding of the magnitude of autotrophic respiration in above- and belowground components might be incorrect. Hence, accounting for root xylem CO₂ flux is essential for accurate quantification and partitioning of the autotrophic component of forest ecosystem respiration into its subcomponents. Therefore, there is a need to measure the magnitude of root xylem CO₂ flux in more species and on longer time scales to allow us to accurately account for below- and above-ground tree respiration.

The movement of CO₂ in roots and soil

Respiration occurs in living cells throughout the root, in the epidermis, phloem, and xylem parenchyma in the center of the root (Raven et al. 1999). As the gaseous CO₂ is released by respiration in living cells, a portion dissolves in water and moves into and through the xylem rather than diffusing outwards into the surrounding soil and contributing to soil CO₂ efflux (Fig. 1).

Fig. 1

Conceptual model of root-respired CO₂ transport in the transpiration stream. Stars indicate points of respiration in living root tissues, where root-respired CO₂ remains within the root, dissolves in the water taken up by the root, and is transported with the transpiration stream in the xylem vessels (1) or where root-respired CO₂ diffuses from inside the root outward into the soil (2). As a result there are two aboveground fluxes of root-respired CO₂, one that contributes to soil CO₂ efflux and one transported internally via the transpiration stream that contributes to efflux from aboveground tissues (drawing adapted from Cruiziat and Tyree 1990)

Dissolved CO₂ from belowground respiration can be transported upwards with the transpiration stream (Aubrey and Teskey 2009; Grossiord et al. 2012; Bloemen et al. 2013b, 2014a), which contributes to the high internal [CO₂] observed in tree tissues (<1 to over 23 %, Teskey et al. 2008). The internal [CO₂] in the xylem is generally higher than that of forest soil, where [CO₂] is often in the range of 1–2 % (Pumpanen et al. 2003). For example, Teskey and McGuire (2007) simultaneously measured soil [CO₂] at 15 cm depth and internal [CO₂] at the stem base of Platanus occidentalis trees and observed substantially higher [CO₂] in the stem (mean 7.6 %) compared with the soil (mean 1.2 %). Similarly, Ubierna et al. (2009) measured higher [CO₂] in the xylem at the bottom of conifer tree stems compared with [CO₂] in neighboring soil and concluded that the xylem CO₂ flux from belowground had likely resulted from the upward transport of root-respired CO₂, rather than transport of CO₂ derived from the soil. So far, only one study has simultaneously measured [CO₂] directly in roots (as opposed to the stem base) and in soil. In those recently concluded measurements, substantially higher [CO₂] was observed in large roots of Liriodendron tulipifera and Fagus grandifolia compared with the neighboring soil environment (De Bel 2014). From these studies we conclude that under most conditions it should be expected that a decreasing [CO₂] gradient exist from the root to the soil, which explains why almost all of the upward flux of CO₂ in root xylem must be derived from root respiration.

Nevertheless, there is a common misunderstanding regarding the source of CO₂ transported through the xylem from belowground. This probably developed because, until recently, scientists were largely unaware of the high [CO₂] in stems and roots. In the 1960s, crop growers believed that plants could take up and assimilate soil-dissolved inorganic carbon (DIC). In various experiments, plants were irrigated with water enriched with CO₂, thereby increasing the amount of soil DIC available for plant carbon gain (see references in Enoch and Olesen 1993). However, in these experiments the contribution of soil DIC uptake to plant carbon gain was very small, generally less than 1 % (Enoch and Olesen 1993). However, whether soil DIC enters a root depends on the concentration gradient between root and soil. So if soil [CO₂] is higher than that of the root, soil DIC will enter the root. This explains why some studies have shown a substantial contribution of soil DIC to the total [CO2] in the xylem sap, e.g., in willow (Vapaavuori and Pelkonen 1985; Vuorinen et al. 1989; Vuorinen and Kaiser 1997), tobacco (Hibberd and Quick 2002), barley (Stolwijk and Thimann 1957; Vuorinen and Kaiser 1997), peas (Stolwijk and Thimann 1957), summer wheat (Schäfer 1988), tomato (Stemmet et al. 1962), and beans (Amiro and Ewing 1992) (Table 1). Those studies were performed in hydroponic culture and the DIC concentrations were higher than normally found in soil, which reversed the [CO₂] gradient from root to soil, allowing the uptake of a substantial amount of DIC from the hydroponic solution.

Table 1 Studies on the uptake and transport of soil-dissolved inorganic carbon (DIC) in plants classified by type of experiment, method, type of plant/setup, and species

Using data from their field study, Aubrey and Teskey (2009) calculated the potential contribution of soil CO₂ to the quantity of CO₂ in the xylem at the base of the stem. They made the assumption that all the CO₂ in soil solution could enter the root and contribute to the CO₂ present in the xylem. Under that unlikely scenario, they estimated that for Populus deltoides trees growing in the field only 8 % of the total xylem CO₂ flux could have been derived from the uptake of soil DIC. However, since the CO₂ gradient would almost always be in the reverse direction, because xylem [CO₂] was generally in the range of 10–20 % and soil [CO₂] was <2 %, it is highly likely that 8 % is a large overestimation of the actual contribution of soil DIC to the [CO₂] measured in the xylem. Supporting this idea, ¹³C label dissolved in water applied to soil around either Pinus seedlings (Ford et al. 2007) or large conifer trees (Ubierna et al. 2009) showed that soil DIC uptake had only a small effect on aboveground plant carbon gain and CO₂ efflux, respectively. Further, uptake of CO₂ from soil by roots should result in an apparent respiratory quotient (ARQ, ratio of CO₂ emission/O₂ uptake ratio) of <1. Angert et al. (2012) observed an average ARQ of approximately 1 in air sampled from the soil near the roots of different tropical trees, suggesting a lack of uptake of soil DIC by the root system.

High root [CO₂] indicates the presence of barriers to diffusion

The cause of the high [CO₂] in tree roots, relative to soil, may be analogous to a previously observed oxygen concentrating mechanism in roots of wetland species (Colmer 2003b; De Simone et al. 2003; Aubrey and Teskey 2009). In these species, roots form substantial barriers to gas exchange in the outer cell layers, leading to both increased [CO₂] and low radial O₂ loss (Colmer 2003a; Colmer et al. 2006; Soukup et al. 2007; Abiko et al. 2012). Highly porous aerenchyma tissue in root and shoots of wetland species provides a low resistance pathway for gas movement, enabling internal aeration (Colmer 2003b). Moreover, aquatic plants have mechanisms by which CO₂ derived from sediments and/or from respiration in submerged rhizomes and root tissues moves internally from belowground to the shoot (Brix 1990; Colmer 2003b). Upland trees lack such aerenchyma tissue, making the transpiration stream the principal internal pathway of upward movement of carbon from belowground.

As root systems age, an increasing proportion of the tissue becomes suberized (Kramer and Kozlowski 1979), which, along with inner and outer bark development in secondary roots, may form a substantial barrier for radial diffusion of root-respired CO₂ into the soil. Most research on the effect of suberization on gas diffusion at the root soil interface has been performed on upland crops (Abiko et al. 2012) and grasses (Soukup et al. 2007). Root suberization has been studied in this context on only two tree species, and it was associated with decreased radial oxygen loss and hence increased root [CO₂] (De Simone et al. 2003).

An illustration of the effect of the barriers to diffusion of CO₂ from roots can be seen in studies of root respiration that have measured excised roots (see review by Kuzyakov 2006). In these studies, after excision, root CO₂ efflux rate first increase and subsequently decrease over time (Rakonczay et al. 1997). By excising the roots, barriers to radial diffusion are removed, and a large amount of root-respired CO₂ that was concentrated in the roots can diffuse into the incubation chamber, causing a high apparent respiration rate at the start of the measurement, followed by a rapid decrease (Chapin and Tryon 1982; Rakonczay et al. 1997; Kuzyakov 2006; Marsden et al. 2008b). In particular, in coarse roots there can be a strong flush of CO₂ efflux directly after root excision, as observed in Eucalyptus (Marsden et al. 2008a), because these roots possess more substantial barriers to radial diffusion. Diffusion of root xylem CO₂ can contribute to significantly higher apparent respiration rates after excision than rates observed in intact living roots (Makita et al. 2012).

Current methods to measure root xylem CO₂flux

As first described by Aubrey and Teskey (2009), root xylem CO₂ flux (F _t, mg C h⁻¹) can be estimated from measurements of sap flow (F _s, g h⁻¹) and gaseous [CO₂] (vol %), along with tissue temperature (°C), and xylem pH which allow the calculation of the total quantity of DIC in xylem sap ([CO ^*₂ ], mM)(Fig. 2):

Fig. 2

Schematic of the experimental setup used by Bloemen et al. (2014a) to estimate root xylem CO₂ flux and soil CO₂ efflux, indicating the distances and positions of the stem girdle and equipment installed on the stem and within the soil area occupied by the roots (A _root) to measure the flux of root-respired CO₂ transported in the transpiration stream and soil CO₂ efflux. The black and grey chamber(s) represent the measurement of E _soil with automated and manual chambers, respectively. Chambers were located at 40 cm from the stem. NDIR non-dispersive infrared CO₂ sensor, TDP-10 thermal dissipation probe. The stem thermocouple installed 3 cm above the NDIR sensor and the manual chambers installed 70 cm from the stem for E _soil measurements are not shown (from Bloemen et al. 2014a)

$$F_{\text{t}} = F_{\text{s}} \times [\text{CO}_{2}^{*} ] \times a$$

(1)

With a the atomic weight of carbon. [CO ^*₂ ] is the sum of [CO₂]_aq, [H₂CO₃], [HCO₃ ⁻], and [CO₃ ²⁻], and cannot be measured in situ (Aubrey et al. 2011). Instead, it is calculated based on Henry’s law, which states that the partial pressure of a gas over a solution is proportional to the concentration of that gas in the solution (Stumm and Morgan 1996):

$$[{\text{CO}}_{2}^{*} ] = \left( {1 + \frac{{K_{1} }}{{10^{{ - p{\text{H}}}} }} + \frac{{K_{1} K_{2} }}{{\left( {10^{{ - p{\text{H}}}} } \right)^{2} }}} \right)K_{\text{H}} {\text{pCO}}_{2} ,$$

(2)

where K ₁ and K ₂ are the first and second acidity constants, respectively, K _H is the Henry constant (all of the constants are temperature dependent) and pCO₂ is the partial pressure of CO₂ over the solution, which is equal to measured xylem [CO₂].

Gaseous xylem [CO₂] can be measured continuously in situ. Non-dispersive infrared (NDIR) sensors have become small enough that they can be inserted directly into a hole drilled into the xylem of medium- and large-diameter trees (an example is the GMM 221 sensor manufactured by Vaisala Inc., Helsinki, Finland). The NDIR sensors have recently been favored over the micro-electrodes used in previous studies of internal [CO₂] in tree stems (McGuire and Teskey 2002, 2004) because they have shown greater stability, less temperature sensitivity, and better reliability under field conditions (Teskey and McGuire 2007). However, the NDIR sensors that are currently available are larger than the CO₂ micro-electrodes, so their application is limited to larger diameter roots or stems. Further, insertion of an NDIR sensor requires creation of a relatively large hole in the tissue, which may have unknown effects on [CO₂] and sap flow in the xylem.

To calculate [CO ^*₂ ], xylem sap pH must also be measured, because pH has a substantial effect on the quantity of inorganic carbon that can dissolve in xylem sap (see Fig. 1 in Teskey et al. 2008; Erda et al. 2014). Below pH 5.6, most inorganic carbon in xylem sap is in the form of CO₂, but at higher pH significant quantities of CO₂ can dissolve as HCO₃ ⁻ and CO₃ ²⁻. Therefore, higher root xylem CO₂ fluxes are expected for species with higher xylem sap pH, which have been reported in a range of 4.5–7.4 for several species (see Table 2 in Teskey et al. 2008). In contrast to temperature and xylem [CO₂], pH is generally measured non-continuously and destructively (Aubrey et al. 2011) on sap expressed under pressure from detached twigs or stem cores. Due to the low temporal resolution of these measurements, temporal dynamics of pH in plants remain poorly understood (Aubrey et al. 2011; Erda et al. 2014). There appear to be seasonal changes in xylem sap pH, which are important to measure for accurate estimates of [CO ^*₂ ] over time, especially in species with high xylem sap pH (i.e., pH > 6.5) (Erda et al. 2014). The accuracy of currently used destructive techniques has not been directly verified, and could represent an important source of error. A new approach for measuring xylem sap pH, particularly if it was continuous, could be a useful improvement in the accuracy of [CO ^*₂ ] estimates.

Table 2 Fraction of root-respired CO₂ transported via the transpiration stream estimated for different species and for different time periods

Although the operational principles of sap flow measurement techniques are well-established (Smith and Allen 1996), the accuracy of these measurements is also of concern for estimating root xylem CO₂ flux. Sensors measuring sap flow rate and sap flux density for woody species are based on heat transport within the stem (Smith and Allen 1996; Vandegehuchte and Steppe 2013) and can continuously measure tree F _s. Species- and site-specific correction factors might be necessary to obtain accurate estimates of sap flow (Steppe et al. 2010; Sun et al. 2011), which is crucial to ensure that accurate estimates of root xylem CO₂ flux are obtained.

To compare soil CO₂ efflux and root xylem CO₂ flux at the same spatial scale (F _t,scaled, mg C m⁻² h⁻¹), the latter needs to be scaled to the soil area occupied by the root system (A _root, m²; Fig. 2):

$$F_{\text{t,scaled}} = {\raise0.7ex\hbox{${F_{\text{t}} }$} \!\mathord{\left/ {\vphantom {{F_{\text{t}} } {A_{\text{root}} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${A_{\text{root}} }$}}$$

(3)

Aubrey and Teskey (2009) assumed an A _root from the consistent stocking of trees in the plantation used in their study. Bloemen et al. (2014a) used data from root excavation measurements on additional trees to estimate the root-occupied soil surface area in their study on young 9-year-old Quercus robur trees. While root excavation potentially provides the most accurate determination of A _root of younger trees, excavation of the root system of older trees is a difficult task. Therefore, this estimate represents a large source of uncertainty for the comparison of root xylem CO₂ flux relative to soil CO₂ efflux.

Importance of root xylem CO₂ flux in trees

The transport of CO₂ from roots into stems and branches has the potential to affect many processes involved in carbon gain and loss. First, the transport of CO₂ out of roots through the xylem directly affects soil-efflux-based estimates of root respiration and therefore total belowground respiration. Three studies have addressed this issue (Aubrey and Teskey 2009; Grossiord et al. 2012; Bloemen et al. 2014a). Each study used different methodology, but all three reported that the autotrophic component of root respiration (R _a) or soil respiration (R _s) was greater when root xylem CO₂ flux was included compared with what would have been estimated by the efflux of CO₂ from the soil surface alone (Table 2).

The first report compared soil CO₂ efflux with xylem CO₂ flux in Populus deltoides trees growing in a plantation (Aubrey and Teskey 2009). During most of the daylight period more CO₂ moved internally through the xylem than fluxed from the soil surface (Fig. 3). Over a 7-day period in mid-summer the flux of CO₂ in xylem sap averaged 0.26 mol CO₂ m⁻² day⁻¹, which was statistically equivalent to the flux of CO₂ from the soil surface, 0.27 mol CO₂ m⁻² day⁻¹. Heterotrophic (R _h) and autotrophic (R _a) contributions to CO₂ efflux from the soil surface were not separated. Assuming that R _h and R _a contributed equally to the total CO₂ efflux from the soil, R _a greatly exceeded R _h (72 versus 28 %) when xylem CO₂ flux was included in the calculation of total belowground respiration.

Fig. 3

a Pattern of mean tree sap flow (blue line) and dissolved CO₂ ([CO₂*]) in xylem sap (red line) measured over a 7-day period in four 9-year-old Populus deltoides trees growing in a plantation. b The mean flux of CO₂ from the root system into the base of the stem of the same trees (green line) in comparison to the mean flux of CO₂ from the soil surface (brown line) over the same time period (from Aubrey and Teskey 2009)

The second report investigated the carbon isotope ratio of soil CO₂ efflux in a Eucalyptus stand planted on a site previously dominated by a C₄ grass, Loudetia simplex (Grossiord et al. 2012). The authors assumed that the heterotrophic ¹³C signal of soil CO₂ efflux reflected the contribution of the C₄ grass roots to older soil carbon. Since the ¹³C signal of C₄ plants is enriched in comparison to C₃ plants, the observed increase in ¹³C enrichment of the soil CO₂ efflux signal during the day, compared with the signal at night, indicated that some of the ¹²CO₂ released from respiration of the Eucalyptus roots was diverted upward in the xylem during periods of active transpiration. Thus, on average over a 4-day period, 17 % of the total daily CO₂ released by root respiration was not accounted for in the measurements of soil CO₂ efflux during the daytime, with a maximum of 24 % at high rates of transpiration.

The third report examined soil CO₂ efflux and F _T in Quercus robur trees subjected to stem girdling compared with non-girdled control trees (Bloemen et al. 2014a). The difference in soil CO₂ efflux between the treatments was used as an estimate of the contribution of R _a. The internal flux of CO₂ in xylem sap was also calculated based on measurements of sap flow and xylem [CO₂]. R _a was substantially higher during daylight hours when CO₂ transported through the xylem was also included (Fig. 4). Averaged over a 5-day period in August, R _a accounted for 27 % of total soil CO₂ efflux. When xylem-transported CO₂ was added, R _a increased to 32 % of total belowground respiration. During periods of peak sap flow, 25 % of total soil CO₂ efflux was attributed to R _a, but when root xylem CO₂ flux was included, the R _a estimate increased to 45 % of total belowground respiration.

Fig. 4

Estimates of the autotrophic component of belowground respiration (R _a) calculated by the conventional approach based solely on measurements of soil CO₂ efflux (black dashed line), or calculated by the new approach, based on combined measurements of E _soil and the scaled flux of root-respired CO₂ transported in the transpiration stream (black line). The shaded areas represent standard deviation (from Bloemen et al. 2014a)

Although there are currently only three studies available to discuss, it is important to note that all of the studies found that a significant quantity of root-respired CO₂ entered the xylem and moved upward in the transpiration stream. The results of these studies indicate that this process is an important component of our mechanistic understanding of carbon flows in trees and forests. Likewise, the variation observed among these studies suggests that there is much more to be learned about the process. There are many possible explanations for the variation in the quantity of CO₂ transported in xylem sap reported in these three studies, regardless of potential variation due to weather conditions (Saveyn et al. 2008). Root systems with large barriers to diffusion will build up high internal [CO₂] and more CO₂ will dissolve in xylem sap. Species with inherently high sap pH, such as Populus deltoides, will have higher sap [CO ^*₂ ] than species with lower sap pH. The movement of CO₂ in the xylem was greatest at times when transpiration was highest, suggesting that variation in rates of transpiration among species may also be an important source of variation in the amount of CO₂ transported in sap. The different methods used in the three studies likely played a role in the variable results. The studies were also short in duration (4–7 days) which prevented the identification of seasonal differences in root xylem CO₂ flux.

Xylem-transported CO₂ also has an effect on estimates of R _a and the belowground R _a/R _h ratio. This ratio is important for understanding forest net ecosystem productivity and the contribution of autotrophs and heterotrophs to the forest carbon cycle. However, it is widely recognized that the estimate of R _h is hampered by methodological limitations (Subke et al. 2006). While R _h estimates are uncertain, the R _a estimate is also incorrect unless CO₂ transport in the xylem is included. In addition to the direct measurements described above, indirect evidence of the effects of xylem-transported CO₂ on R _s is mounting. Soil CO₂ efflux was strongly negatively correlated with evapotranspiration in a grassland ecosystem (Balogh et al. 2014) and a daytime depression in R _s was correlated with transpiration in a Pinus sylvestris stand, implying that xylem-transported CO₂ contributed to the reduction in R _s (Subke et al. 2009). Yang et al. (2012) found that in six temperate forest types in China, R _s was well correlated with soil temperature in the dormant season but not in the growing season, and suggested that transpiration and xylem transport of CO₂ during the growing season caused the difference.

Another effect of root-derived CO₂ in xylem sap occurs when it is transported into the shoot and increases estimates of stem and branch respiration. The upward movement of dissolved CO₂ in the xylem was first demonstrated in a Quercus alba tree in which a solution containing a high concentration of dissolved CO₂ was infused into the stem. The xylem sap [CO₂] above the injection point doubled compared to the initial value and there was a corresponding increase in CO₂ efflux from the stem (Teskey and McGuire 2002). Since then, there have been several studies documenting the contribution of internally transported CO₂ to the efflux of CO₂ from stems (e.g., McGuire and Teskey 2004; Gansert and Burgdorf 2005; Teskey and McGuire 2005; Steppe et al. 2007; Yang et al. 2012). One of those studies suggested that a large portion of the CO₂ that diffused from the stem over a 24-h period had been transported in sap from roots to the point of efflux (Teskey and McGuire 2007), which was first substantiated in a study on Populus deltoides trees in a plantation (Aubrey and Teskey 2009). The potential for long distance transport of CO₂ in the xylem was made much clearer when a large quantity of ¹³CO₂ was infused into the base of the stem immediately above the root collar of Populus deltiodes trees (Bloemen et al. 2013b). After 2 days the trees were harvested and the tissues were examined for the presence of the ¹³C tracer. The abundance of ¹³C had significantly increased in all tissues up to the top of the crown, including stem, branches, twigs, and leaves. However, when calculating the amount of ¹³C recovered compared with the amount introduced into the trees, it was found that 83–94 % of the ¹³CO₂ had been lost, i.e., it had diffused into the atmosphere from the stem, branches, and twigs as it was transported upward. This result is a clear indication that xylem-transported CO₂ from roots and locally produced CO₂ both contribute to CO₂ efflux from the stem. Overall, xylem transport of root-respired CO₂ causes an underestimation of R_a, and it also causes an overestimation of stem, branch, and twig respiration determined from measurements of CO₂ efflux. There is some indication that xylem-transported CO₂ may also diffuse out of leaves, and therefore affect estimates of leaf respiration (Bloemen et al. 2013a, 2014b), but more evidence is needed before a conclusion about that can be reached.

Carbon dioxide transported in xylem from roots also contributes to total tree carbon gain. As the CO₂ moves though the xylem and diffuses through the bark, it encounters chlorophyll-containing cells that can use it for photosynthesis. In the Bloemen et al. (2013b) study, up to 17 % of the ¹³CO₂ that moved through the shoot was refixed. Of that 17 %, 3 % was fixed in stems, 11 % in branches and 3 % in leaves. The amount of refixation of root-respired CO₂ is likely to be species-dependent. We speculate that species with thick bark that allows little light to reach photosynthetic cells in the stem likely recycle little transported CO₂, while species with thin bark and green stems and branches likely recycle a much larger amount.

In addition to photosynthetic refixation, there is some evidence that the phosphenolpyruvate carboxylase (PEP_C) pathway may be used to recycle xylem-transported CO₂. Recent investigations have shown the importance of PEP_C for non-photosynthetic carbon fixation in stems (e.g., Berveiller et al. 2007; Berveiller and Damesin 2008), but there are few recent studies in roots. Experiments using radioactive carbon isotopes first showed fixation of CO₂ in roots in the mid-twentieth century. Detached barley (Hordeum vulgare) roots were submerged in a solution containing H¹¹CO₃ ⁻, took up the solution for a 92-min period, and retained almost 5 % of the ¹¹C label as reduced carbon (Overstreet et al. 1940). The exudate from the unsubmerged cut tops of the roots also contained reduced ¹¹C, suggesting that fixed carbon was transported in the xylem stream. Poel (1953) supplied excised barley roots with a solution enriched with NaH¹⁴CO₃ and found that the bicarbonate was taken up and reduced. Products were organic acids and amines, consistent with fixation in the tricarboxylic acid cycle, which is now known to be supplied with carbon compounds catalyzed by PEP_C. Jackson and Coleman (1959a, b) provided evidence that PEP_c catalyzed the majority of CO₂ fixed in bean (Phaseolus vulgaris) roots. In an experiment using cell-free enzyme preparations from stem and root wood of Robinia pseudoacacia, Höll (1974) found that ¹⁴CO₂ was transformed into organic compounds in the presence of PEP, but not when PEP was absent, indicating that PEP_c was the carboxylating enzyme. The author concluded that in living trees, respiration was the likely source of CO₂ for the reaction, but that cells surrounding the xylem could also use CO₂ from the transpiration stream. More recently, Hibberd and Quick (2002) reported that C₃ tobacco showed characteristics of C₄ photosynthesis in cells that surround the xylem. They suggested that malate derived from PEP_c activity in the roots was transported through the xylem and decarboxylated at sites of photosynthesis in the stem and petiole, representing a more spatially separated version of the C₄ photosynthetic pathway. All of these studies have indicated the importance of PEP_c-catalyzed carbon fixation in plant tissues, but the magnitude of this fixation pathway in the carbon economy of trees has never been quantified.

Conclusion and future prospects

In trees, a large portion of respired CO₂ remains within the root system and is transported away from the site of respiration via the transpiration stream rather than diffusing into the soil. The decreasing CO₂ gradient from inner root tissues to the soil indicates that under most conditions the internal CO₂ transported through root xylem is mainly derived from root respiration and not from soil DIC uptake. Measurements of [CO₂] in both tree roots and neighboring soil for different tree species will be essential to increase recognition of the importance of root-respired CO₂ to the internal CO₂ flux in trees. Although current studies are limited to a small number of species and short measurement periods, they all indicate that root xylem CO₂ flux represents an important unaccounted-for component of root respiration and call into question our contemporary understanding of tree carbon cycling. These studies highlight the need to measure root xylem CO₂ flux, as well as the amount of root-respired CO₂ diffusing through the soil, to accurately determine R _a and R _s. Since a substantial fraction of the root-derived CO₂ may diffuse from stem and branch tissues into the atmosphere, current methods for estimating autotrophic respiration in above-and below-ground tissues based on CO₂ efflux into the atmosphere are likely to be inaccurate. Laser-based ¹³CO₂ efflux measurements in combination with stem isotope labeling as performed by Bloemen et al. (2013b) could be used to accurately quantify aboveground efflux of xylem-transported root-respired CO₂. Moreover, studies of root xylem CO₂ flux on additional species, in varied ecosystems, and over longer time periods have the potential to redefine our understanding of the energetic costs of tree root metabolism and our assessment of the components of tree and forest carbon cycles.