Measurement of stem respiration of sycamore (Platanus occidentalis L.) trees involves internal and external fluxes of CO2 and possible transport of CO2 from roots


R. O. Teskey, School of Forestry and Natural Resources, University of Georgia, Athens, GA 30602. Fax: 706 542 8356; e-mail:


CO2 released by respiring cells in tree stems can either diffuse to the atmosphere or dissolve in xylem sap. In this study, the internal and external fluxes of CO2 released from respiring stems of five sycamore (Platanus occidentalis L.) trees were calculated. Mean rates of stem respiration were highest in mid-afternoon and lowest at night, and were positively correlated with air temperature. Over a 24 h period, on average 34% of the CO2 released by respiring cells in the measured stem segment remained within the tree. CO2 efflux to the atmosphere consisted of similar proportions of CO2 derived from local respiring cells (55%) and CO2 that had been transported in the xylem (45%), indicating that CO2 efflux does not accurately estimate respiration. A portion of the efflux of transported CO2 appeared to have originated in the root system. A modification of the method for calculating stem respiration based on internal and external fluxes of CO2 was developed to separate efflux due to local respiration from efflux of transported CO2.


An understanding of the respiration rate of tree stems is important for many reasons, including to quantify the carbon cycle of forests (Bolstad et al. 2004), estimate net ecosystem productivity (Maier et al. 2004), evaluate patterns of carbon allocation (Giardina et al. 2003) and examine causes for variation in growth rates or growth efficiencies among trees and stands (Will et al. 2001). However, measurements of stem and branch respiration are often quite variable both among and within trees (Cernusak & Marshall 2000; Ceschia et al. 2002; Bosc, De Grandcourt & Loustau 2003; Wieser & Bahn 2004; Kim & Nakane 2005). The causes of the variability have been difficult to isolate. One hypothesis put forward to explain at least part of this variation is that a portion of respired CO2 can dissolve in sap and move upward in xylem rather than escape to the atmosphere. Therefore, factors affecting rates of transpiration can also affect the amount of CO2 diffusing into the atmosphere. This concept has been proposed and discussed a number of times (Boysen-Jensen 1933; Negisi 1979; Martin, Teskey & Dougherty 1994; Bowman et al. 2005; Gansert & Burgdorf 2005), but direct evidence supporting the hypothesis has been lacking until recently (Levy et al. 1999; Teskey & McGuire 2002). More importantly, the corollary to this concept, that is, that xylem-transported CO2 can diffuse into the atmosphere remote from the site of respiration, has rarely been considered. Recently, direct manipulations of the CO2 concentration ([CO2]) of xylem sap have shown that a portion of transported CO2 escapes from the stem to the atmosphere, and that the rate of efflux is linearly related to xylem sap [CO2] (Teskey & McGuire 2002, 2005).

In this study, we measured the internal and external fluxes of CO2 from respiring cells in tree stems. The objective was to quantify the amount of CO2 that moves in these pathways over a 24 h period. In addition, we measured the changes in [CO2] and CO2 efflux with height in the same stems. We expected that there would be an increase in [CO2] with height that would affect the apparent rate of respiration measured at different heights along the stem. We proposed that both internal and external fluxes of respired CO2 are substantial in magnitude and that they interact with each other and are affected by the environment in complex ways. We also presented methods for measuring internal fluxes of CO2, and a modified procedure for calculating respiration based on both internal and external fluxes.


We calculated stem respiration (RS) of mature sycamore trees (Platanus occidentalis L.) by summing outward flux of respired CO2 (efflux) from the stem to the atmosphere (JOR), upward flux of CO2 transported in flowing sap (JU) and the change in the amount of CO2 stored within the xylem (JX). These fluxes were formerly termed EA, FT and ΔS, respectively (McGuire & Teskey 2004). The three component fluxes were calculated from measurements made on and within the stems of five trees at 1 m above the soil surface during the growing season in 2004. We were limitedby available equipment to measuring one tree at a time. After an initial period of equilibration of at least 8 h, we measured each tree for two consecutive 24 h periods beginning at 0000 h. Data from the two 24 h periods were averaged for each tree, with the exception of Tree 3, which was measured for one 24 h period because of equipment malfunction. Measurements were conducted on 31 July, 13 and 24 August, and 1 and 11 September on Trees 1–5, respectively. Weather conditions were similar during all measurement periods; days were warm and sunny with no precipitation. The trees were growing in Whitehall Forest, a research facility of the University of Georgia located near Athens, GA, USA. Tree diameters ranged from 19.5 to 24.8 cm (Table 1).

Table 1.  Stem diameter at 1.5 m, xylem water and gas content, sap pH, soil [CO2] and stem [CO2] at the base of the stem (0.1 m) and at 3 m height of five sycamore trees
TreeStem diameter (cm)Xylem water content (%)Xylem gas content (%)Ratio gas/water contentSap pHSoil [CO2] (%)Stem [CO2] 0.1 m (%)Stem [CO2] 3 m (%)
  1. Different letters next to mean [CO2] indicate statistically significant differences at α = 0.05. Soil [CO2] was measured within 1 m of the base of each tree.

Mean22.5240.724.90.635.81.2 a7.6 b8.9 c

Following McGuire & Teskey (2004), stem respiration (RS) was calculated using a mass balance approach that measured both internal and external fluxes of CO2:


JOR is calculated as


where JO is CO2 efflux to the atmosphere and JO[CO2] is efflux to the atmosphere of CO2 transported in the xylem from lower parts of the tree. JO[CO2] is calculated from stem [CO2] ([CO2]s) measured just below the cuvette, and a tissue-specific linear regression of efflux and stem [CO2],


developed from a separate set of measurements of [CO2]s and CO2 efflux on the five trees used in the study. In Eqn 1, JOR is used as the efflux term instead of JO (McGuire & Teskey 2004) to account for the effect of [CO2]S on the rate of CO2 diffusion from stem to atmosphere.

Following McGuire & Teskey (2004) the three fluxes of CO2 are determined as follows:

Outward flux of CO2 from the stem to the atmosphere (JO) is calculated as


where fA is the rate of air flow (mol s−1) through a cuvette surrounding the stem segment; v is sapwood volume of the segment (m3), and Δ[CO2] is the difference in [CO2] of air flowing into and out of the cuvette (µmol mol−1). The general form of Eqn 4 has been commonly used for plant gas exchange measurements (Long & Hallgren 1985; Maier, Zarnoch & Dougherty 1998).

Upward flux of CO2 in xylem sap (JU) is calculated as


where fS is the rate of sap flow through the segment (L s−1), and Δ[CO2*] is the difference in sap [CO2] measured above and below the cuvette (µmol CO2 L−1 sap), accounting for the concentration of all dissolved products of CO2 in the sap, including bicarbonate and carbonate ions (Stumm & Morgan 1996). JU estimates the amount of CO2 dissolved in sap that was transported away via the xylem from a specified volume of the stem enclosed in the cuvette (v). The calculation compares the [CO2*] of the sap entering the specified volume of stem, that is, CO2 that was already dissolved in the sap, with the [CO2*] of the sap leaving the specified volume, accounting for the rate of sap movement (fS). The theoretical basis for calculating JU is similar to that of the calculation of JO (cf. Eqns 4 and 5).

The change in CO2 stored in the xylem (JX) is calculated as


where [CO2]St0 and [CO2]St1 are means of upper and lower stem [CO2]S (µmol m−3 sapwood) in the stem segment at t0 and t1, respectively, and Δt is the time interval (t1 − t0) (s). JX represents the increase or decrease in mean stem CO2 concentration ([CO2]S) within the specified stem volume during the specified time interval. (In this experiment, the time interval is 10 min). If mean [CO2]S increases over the 10 min interval, JX is positive, indicating that CO2 has been stored in the xylem. If mean [CO2]S decreases, JX is negative, indicating that CO2 has been released from the xylem. Released CO2 leaves the stem volume as either JO or JU. All fluxes are in units of µmol CO2 m−3 sapwood s−1 (see McGuire & Teskey 2004 for additional explanation of the theoretical basis of the calculations).

To measure JO, we used an infrared gas analyzer (IRGA) (model 7000; Li-Cor, Inc., Lincoln, NE, USA) connected to a cuvette that completely surrounded the stem segment at ∼ 1 m above ground level. The cuvette was constructed of clear Mylar film (Ridout Plastics, San Diego, CA, USA) and was sealed to the stem with adhesive closed-cell foam as described in Lai et al. (2002). To improve gas tightness of the cuvette, glazing putty was used to further seal the closed-cell foam to the bark. Compressed air of known near-ambient [CO2] was supplied to the cuvette at 1.0 L min−1 with a mass flow controller (model FMA5514; Omega Engineering, Inc., Stamford, CT, USA). The IRGA was operated in open configuration, and efflux rate was calculated using standard procedures (Long & Hallgren 1985). Cuvette length and stem diameter (minus bark thickness) were measured to determine sapwood volume. Sap flux density (Q) was measured at three xylem depths (15, 50 and 90 mm) with two sets of thermal dissipation sensors (model TDP-100; Dynamax Corp., Houston, TX, USA) (Granier 1987) placed ∼ 20 cm above the cuvette on opposite sides of the stem.

Xylem [CO2] was measured with solid state non-dispersive infrared (NDIR) CO2 sensors (model GMM222; Vaisala, Inc., Woburn, MA, USA) placed into the xylem 20 mm above and below the cuvette. These sensors replaced the CO2 microelectrodes that we previously used for this measurement (McGuire & Teskey 2004) because they had greater stability, less temperature sensitivity and better reliability under field conditions. To install the NDIR sensors in the stems, 19 mm diameter holes were drilled ∼ 50 mm into the tree. The sensors were inserted into the holes and sealed to the tree surface with flexible rubber sealant (Qubitac; Qubit Systems, Inc., Kingston, Ontario). At each measurement location (above and below the cuvette), two sensors were placed on opposite sides of the tree and data were averaged. Xylem temperature was measured with thermocouples inserted into the stem near the sensors. All sensors and the cuvette were covered with a sheet of reflective bubble insulation (Reflectix; Reflectix, Inc., Markleville, IN, USA) to exclude solar radiation. Measurements were made every 10 s and were averaged and recorded every 10 min with a datalogger (model 23X; Campbell Scientific, Inc., Logan, UT, USA). Data were processed and hourly averages calculated using Excel 2002 (Microsoft Corp., Redmond, WA, USA). Air temperature was measured at a weather station located 300 m from the sample trees.

The NDIR sensors measured the [CO2] in air in a small headspace in equilibrium with xylem sap. Temperature-dependent Henry's Law coefficients were used to convert [CO2] (µmol CO2 mol−1 air) to the concentration of all dissolved products of CO2 in the sap ([CO2*], µmol CO2 L−1 sap) (Butler 1991; Stumm & Morgan 1996; Levy et al. 1999; McGuire & Teskey 2002). The dissolution of CO2 into bicarbonate and carbonate ions increases with increasing pH; therefore, the pH of sap must be considered when calculating [CO2*]. We measured the pH of xylem sap by extracting 5 mm diameter radial cores of sapwood from each tree using an increment corer. The cores were compressed in a vise to express a small quantity of sap, which was transferred immediately with a pipette to a solid state pH microsensor connected to a pH meter (Red-Line Standard sensor, Argus meter; Sentron Europe BV, Roden, the Netherlands). In a preliminary experiment to determine diurnal variation, we measured xylem sap pH every 4 h from 0800 to 2000 h using the same procedures on similar sized sycamore trees, and found < 0.15 pH unit variation within trees over that time period. Based on those results, measurements were made on each tree once in mid-afternoon. Xylem sap pH varied among trees from 5.6 to 6.0 (Table 1). The total concentration of CO2 (gaseous and dissolved) in the stem ([CO2]S) was calculated as follows: water and gas content of the xylem were calculated from the same cores used for pH determination. Core samples were weighed prior to pH measurement and were subsequently dried and reweighed to determine volume proportion of water (W). Volume proportion of gas space (G) in the cores was calculated according to the methods of Gartner, Moore & Gardiner (2004) as


where C is the volume proportion of the core that is cell wall material. C is calculated as


where B is basic density of the core (dry mass/green volume, g cm−3), and P is wood density of pure cell wall material (1.53 g cm−3).

[CO2]S (µmol CO2 m−3 wood) was determined as


where 22.4 L gas = 1 mol gas and 1000 L = lm3 wood. Xylem water content varied among trees from 35.3 to 44.7%, and xylem gas content varied from 17.2 to 33.7% (Table 1).

When diurnal measurements were complete, we made additional measurements on the same trees at three heights from the soil surface: 0.1, 1.5 and 3.0 m. As in the first experiment, we measured each tree on a separate day (13, 15, 21, 23 and 29 September). We placed two CO2 sensors and two thermocouples on opposite sides of a tree at each height. Sap flux density (Q) was measured on opposite sides of the tree at ∼ 1 m height at 25 mm xylem depth with thermal dissipation probes (TDP-30, Dynamax). Sap flux at 25 mm depth was scaled to the entire sapwood radius using radial profiles calculated from measurements made at three depths in the first experiment (Hatton, Catchpole & Vertessy 1990). Soil [CO2] was measured within a 1 m radius of each tree with an NDIR sensor suspended in a polyvinyl chloride (PVC) tube. The tube was 20 cm long and was inserted into the soil to a depth of 15 cm. The sensor was placed near the bottom of the tube, and the top of the tube was sealed with Qubitac putty (Qubit Systems, Inc.). Monitoring indicated that sensor output stabilized within a few hours. All sensors were logged every 10 s and averaged and recorded every 10 min with a datalogger. After a 24 h equilibration period, we used an IRGA (model 7000; Li-Cor, Inc.) in open configuration to make spot measurements of CO2 efflux at mid-day. We attached an 82 cm2 PVC cuvette to the tree surface. A closed-cell foam gasket, petroleum jelly and ratchet cords were used to seal the cuvette to the stem. The cuvette and all sensors were covered with a sheet of reflective bubble insulation that was wrapped around the stem (Reflectix; Reflectix, Inc.) to exclude solar radiation. Air at known near-ambient [CO2] was supplied to the cuvette from a compressed gas tank at a rate of 0.5 L min−1, using a mass flow controller (model FMA5514; Omega Engineering, Inc.). After a 10 m equilibration period in which [CO2] stabilized, IRGA data were manually recorded. Efflux measurements were made sequentially at 1.5 and 3.0 m height and were repeated on opposite sides of the tree. The entire cuvette measurement sequence was repeated a total of two times, and the values for each stem position were averaged. CO2 efflux was calculated on a volumetric basis following procedures in Maier et al.(1998).

SigmaPlot 2004 v. 9.01 (Systat Software Inc., San Jose, CA, USA) and SigmaStat 3.11 (Systat Software) were used to compare data with regression procedures and means comparison tests.


Vertical comparisons of stem CO2 concentration and CO2 efflux

Midday measurements of soil gas phase [CO2] adjacent to each of the five trees ranged from 0.8 to 1.5%, with an overall mean of 1.2% (Table 1). In comparison, the [CO2] in the gas phase in the xylem was substantially higher, ranging from 5.6 to 10.1% (mean 7.6%) at the base of the stem and 6.9–12.0% (mean 8.9%) at a height of 3 m. Although there was substantial tree-to-tree variation in [CO2]S at the base of the stem, within every tree there was a close correspondence between the [CO2]S at the base of the stem and the [CO2]S higher in the stem (Fig. 1). The tree-to-tree variation in xylem [CO2]S was correlated with stem sapwood area [[CO2]S = 1.8 + 28.0 (sapwood area, m2), R2 = 0.68, n = 5, P = 0.09].

Figure 1.

The relationship between [CO2]S at the base of the stem and [CO2]S measured at 1.5 (filled symbols) and 3 m (open symbols) heights in the stem xylem. Values for each tree are shown as separate symbols (filled ▴ hidden below open ▵). [CO2]s, stem CO2 concentration.

Mean stem CO2 efflux from each tree at 1.5 and 3.0 m height was linearly related to mean [CO2]S measured at the same height (Fig. 2). A similar relationship was observed in a controlled experiment in which xylem sap [CO2] was manipulated experimentally while stem respiration remained constant (Teskey & McGuire 2005). Direct manipulation of xylem [CO2] provided a means of separating the effect of xylem-transported CO2 on CO2 efflux from that caused by changes in rates of local tissue respiration. The slopes of the regression lines fit through the two data sets were nearly equal (Fig. 2), suggesting in both instances that a portion of the xylem-transported CO2 diffused from the stem as it moved upward and contributed, along with CO2 from local respiring cells, to the total quantity of CO2 that fluxed from the stem to the atmosphere.

Figure 2.

Mean CO2 efflux was linearly related to stem [CO2]S. Filled circles represent spot measurements of CO2 efflux and [CO2]S at 1.5 and 3.0 m heights on the stem. Open circles were recalculated data from Teskey & McGuire (2005, Fig. 2a). They represent spot measurements made on sycamore sapling stems after direct manipulation of the [CO2] in the stem. Stem [CO2] was manipulated without changing the rate of stem respiration by severing the stems from the roots and placing them in water with different [CO2] that was absorbed by the stems. P < 0.001 for both regression equations. [CO2]s, stem CO2 concentration.

Diurnal patterns of stem respiration (RS) and its component fluxes

In this experiment, both internal and external CO2 fluxes were calculated over 24 h periods on summer days. Mean hourly RS, calculated as the sum of mean hourly JOR, JU and JX for the five trees, showed a distinct diurnal pattern, with the lowest rates between 0000 and 0800 h (Fig. 3a). Mean hourly RS was correlated with mean hourly air temperatures (TA) [RS = 13.93 + 3.39 (TA), R2 = 0.81, n = 24, P < 0.001]. Maximum RS occurred in late afternoon, when none of the contributing fluxes were at their maxima, but were either increasing (JOR, JX) or near their peak rate (JU). Peak RS occurred when stem temperature was very close to the diurnal maximum, [CO2]S was at the diurnal minimum, and air temperature was decreasing (Fig. 3).

Figure 3.

Mean hourly data from five sycamore trees measured continuously over 24 h periods in the summer. (a) stem respiration (RS), CO2 efflux to the atmosphere (JOR), internal CO2 transport flux (JU), and CO2 storage flux (JX); (b) xylem sap flux and the total concentration of CO2 in the stem xylem ([CO2]S); and (c) stem and air temperatures.

Hourly mean transport flux (JU) was dependent on sap flux rather than changes in [CO2]S, and it exhibited the greatest diurnal variation of the CO2 fluxes, ranging from near zero at night to a maximum of 51.2 µmol m−3 s−1 at 1300 h (Fig. 3a,b). JU was positively correlated with air temperature [JU = −105.28 + 5.41(TA), R2 = 0.95, n = 24, P < 0.001], likely because sap flux and air temperature were highly correlated.

The change in the amount of CO2 stored in the stem (JX) also had a diurnal pattern. CO2 was lost from the storage pool (negative JX) during the day (∼ 1000–1600 h) and added to the storage pool (positive JX) in the evening, night and early morning (∼ 1700–0900 h). This pattern was inversely related to the diurnal pattern of sap flux, that is, when sap flux was at its maximum, JX was at its minimum, suggesting that flowing sap diluted the CO2 concentration in the xylem. There was also a slight increase in JX in the first 2 h after sap movement began in the morning, likely due to a buildup of CO2 in the roots and lower stem at night. JX was negatively correlated with sap flux [JX = 5.21 − 3.06 (Q), R2 = 0.58, n = 24, P < 0.001] and air temperature [JX = 32.61 − 1.42 (TA), R2 = 0.48, n = 24, P < 0.001].

The third component of RS, CO2 efflux to the atmosphere from local cell respiration (JOR), was separated from total CO2 efflux (JO) using Eqn 2 and the relationship between CO2 efflux and [CO2]S measured on the same trees (Eqn 3 & Fig. 2). JOR exhibited less diurnal fluctuation than JU or JX (Fig. 3a). However, JOR exhibited a consistent midday decline from ∼ 1100–1800 h, which was associated with high values of JU. Mean hourly JOR also followed a diurnal pattern similar to that of mean stem temperature (TS) (Fig. 3a,c) and was well correlated with it [JOR = −73.06 + 4.60 (TS); R2 = 0.75, n = 24, P < 0.001]. At night, JOR was substantially greater than JU or JX. However, during daylight hours from 1100 to 1600 h, JU exceeded JOR.

Rates of RS, JOR, JU and JX were averaged for 4 h periods to assess differences in the contributions of JOR, JU and JX to RS (Table 2). Mean RS ranged from 78.1 µmol m−3 s−1 in late afternoon (1600–2000 h) to 43.5 µmol m−3 s−1 in the early morning hours (0400–0800 h). Throughout the 24 h period, JOR was a consistently large flux, but from 0800–2000 h, the JOR and JU fluxes were not statistically different. Mean JOR remained relatively constant throughout the 24 h period, ranging from 36.9 µmol m−3 s−1 (1200–1600 h) to 44.6 µmol m−3 s−1 (2000–0000 h). However, JU did not remain constant, as it is dependent on sap flux. JU ranged from 0.40 µmol m−3 s−1 at night (0400–0800 h) to 46.5 µmol m−3 s−1 in the afternoon (1200–1600 h). The latter value represented 63% of total RS during that period. JX was smaller than the other two fluxes most of the time, except at night when JU was near 0. During most of the daylight hours (0800–1600 h), JX was negative, that is, CO2 was escaping from the storage pool. The relative contributions of JU, JX and JOR to RS varied considerably over the diurnal period (Fig. 4). The effect of the release of stored CO2 on RS between 0800 and 1600 h, and the dependence of JU on sap flux are particularly apparent.

Table 2.  Means and SEs of stem respiration (RS), CO2 efflux to the atmosphere (JO), efflux to the atmosphere of locally respired CO2 (JOR), efflux to the atmosphere of xylem-transported CO2 (JO[CO2]), CO2 transport flux (JU), CO2 storage flux (JX), stem temperature (TS), air temperature (TA), the total concentration of CO2 in the stem, including carbonate forms ([CO2]S) and sap flux density (Q)
Time of day
(µmol m−3 s−1)
(µmol m−3 s−1)
(µmol m−3 s−1)
(µmol m−3 s−1)
(µmol m−3 s−1)
(µmol m−3 s−1)
(mol m−3)
(cm h−1)
  1. Four hour means calculated from five sycamore trees measured continuously for 24 h. Significant differences (α = 0.05) in 4 and 24 h means among component fluxes JOR, JO[CO2], JU and JX are indicated by different letters.

0000–040044.1 (7.8)72.1 (10.7)38.8 a (8.7)33.3 a (2.6)0.4 b (0.2)4.9 c (1.3)24.7 (0.8)19.6 (0.9)3.2 (0.3)43.3 (13.2)
0400–080043.5 (7.2)72.0 (10.6)38.2 a (8.5)33.7 a (2.8)0.4 b (0.2)4.8 c (1.4)24.3 (0.9)19.2 (0.9)3.3 (0.3)68.9 (25.9)
0800–120063.0 (15.0)70.7 (9.9)36.9 a (7.7)33.8 a (3.0)28.6 a (12.0)−2.5 b (2.6)23.6 (0.9)24.7 (1.1)3.3 (0.3)2342.3 (403.5)
1200–160073.7 (17.9)68.8 (9.1)36.8 a (7.2)32.0 a (2.7)46.5 a (13.2)−9.6 b (2.9)24.1 (1.0)28.2 (1.3)3.2 (0.2)4344.5 (391.5)
1600–200078.1 (17.5)73.1 (9.9)41.6 a (8.2)31.6 a (2.2)32.6 a (12.4)3.9 b (2.7)24.9 (1.1)25.3 (1.1)3.2 (0.2)2926.1 (564.2)
2000–000057.4 (10.4)77.0 (10.8)44.6 a (9.2)32.4 a (2.1)5.0 b (3.3)7.5 b (1.4)25.1 (1.0)20.9 (1.0)3.3 (0.2)336.9 (154.5)
24 h mean60.0 (5.9)72.3 (1.1)39.5 a (1.3)32.8 a (0.4)18.9 b (8.0)1.6 c (2.6)24.5 (0.2)23.0 (1.5)3.3 (0.0)1677.0 (734.1)
Figure 4.

Relative contribution of storage flux (JX, black filled bars), CO2 transport flux (JU, unfilled bars) and CO2 efflux from local tissue respiration (JOR, hatched bars) to stem respiration over 4 h period during the day.

Average RS over the 24 h period was 60.0 µmol m−3 s−1 (Table 2). Mean total CO2 efflux to the atmosphere (JO) was 72.3 µmol m−3 s−1, or 21% greater than RS. Mean calculated JOR was 55% of mean CO2 efflux (JO) (Table 2), which meant that about one-half of the CO2 diffusing from the stem to the atmosphere at 1 m stem height was not due to the respiration of local cells. A comparison of mean hourly measurements of JO and mean hourly calculations of RS showed no correspondence between CO2 efflux from the stem (JO) and the calculated rate of stem respiration (RS) (Fig. 5). The difference was especially large at night, when JO overestimated RS by as much as 63% in the 4 h after midnight (Table 2).

Figure 5.

Relationship between hourly mean values of total stem CO2 efflux to the atmosphere (JO) and stem respiration (RS = JOR + JU + JX).


Early studies of the structure and function of trees revealed the existence of high concentrations of CO2 within stems. Bushong (1907) first reported [CO2] of 7.2% in Populus deltoides stems. MacDougal & Working (1933) measured [CO2] in stems of an additional eight tree species and found all the trees had high [CO2], ranging from 1.4 to 26.3%. Chase (1934) provided further confirmation of high internal CO2 concentrations in a study of five other tree species. Sporadic studies since then in other species, and using various measurement methods, have consistently confirmed those early observations of high stem [CO2] (Jensen 1967; Eklund 1990; Hari, Nygren & Korpilahti 1991, Eklund 1993; Levy et al. 1999; McGuire & Teskey 2002).

Data from the current study support the conclusion that a substantial portion of the CO2 found within tree stems originated in the root system. High [CO2] at the base of the stem, averaging 7.6%, could have resulted from the accumulation of respired CO2 in xylem sap, which was being transported from the root system, or from a high rate of respiration at that point in the stem. CO2 efflux measurements at the base of the stem and at other locations did not provide evidence of exceptionally high rates of respiration at the stem base, supporting the proposal that the origin of the CO2 was the root system. High [CO2] has been reported in tree roots (Rakonczay, Seiler & Kelting 1997). Soil [CO2] (mean 1.2%) was much lower than [CO2] measured at the base of the stem. Uptake of small amounts of dissolved inorganic carbon by roots has been observed in other species (Cramer & Lips 1995), but the low [CO2] of the soil in this study indicate that only a small portion of the CO2 in the xylem at the base of the stem could have originated in the soil solution.

We did not find the expected increase in [CO2]S with stem height because of the buildup of CO2 transported from lower in the stem. The small change in [CO2]S with stem height suggests the possibility that cellular respiration in the stem was contributing relatively less than root respiration to the total CO2 in the stem. However, other factors may have influenced [CO2]S. At night, when no CO2 is transported upward in flowing sap, local stem respiration alone causes an increase in xylem [CO2]S and efflux (Teskey & McGuire 2002). During the day, CO2 is transported from the roots, but increasing rates of sap flux may cause a decrease in [CO2]S (and efflux) as a result of the dilution effect of the large quantity of water coming from the soil at relatively low [CO2*]. Additionally, the balance between the amount of CO2 absorbed in xylem sap versus that which diffuses to the atmosphere likely depends on the location of the respiring cells within the stem, bark permeability, sap flux rate, sap pH and temperature. Temperature may be particularly important because of its direct effect on the rate of respiration and its effect on solubility of gases.

There was substantial among-tree variation in [CO2]S at the base of the bole. The cause of this variation is not known; however, it generally corresponded with tree size, such that trees with larger diameters had higher [CO2]S. Although the range of tree diameters was small in this study, this relationship suggests that larger trees with larger root systems may accumulate more CO2 in the xylem sap. We speculate that it may be possible to use the measurement of [CO2]S at the base of the tree to indirectly estimate root respiration, or respiratory capacity, of individual trees growing in the field. In addition, high [CO2]S at the base of the stem indicates that root respiration may have been underestimated by previous soil CO2 efflux measurements. It appears that a portion of CO2 released from root respiration dissolves in xylem sap and is transported within the tree, and so is unaccounted for by soil CO2 efflux measurements. To illustrate this effect, we estimated that mean [CO2*] at the base of the stem was 0.01 mol CO2 L−1, and the mean sap flux was 40.3 L d−1 tree−1, so that on average, 0.4 mol CO2 tree−1 d−1 moved from the root system into the stem. If a reasonable estimate of the root respiration is 1–2 µmol m−2 soil surface s−1 (Yuste et al. 2005), then the amount of root-respired CO2 transported in the xylem stream is equivalent to the amount of CO2 that fluxes from 2.3 to 4.6 m−2 of soil surface. The significance of this number depends on the amount of area occupied by the root system of the individual trees, which is unknown. Of course, this calculation is only a rough approximation, and the fate of root-respired CO2 warrants further investigation.

The other part of this study provided an estimate of the mean magnitude of CO2 fluxes in sycamore tree stems diurnally during part of the growing season. The calculation of stem respiration using CO2 fluxes showed complex inter-relationships between internal and external fluxes of CO2 that in turn were affected by other factors including sap flux, [CO2]S and temperature.

Fluxes of the CO2 that remained within the stem (JU + JX) were substantial during daylight hours, accounting for 47% of total respired CO2 in the afternoon. At night, with minimal sap movement, JU + JX accounted for less than 12% of evolved CO2. Therefore, the relative importance of internal and external fluxes reversed from day to night. This cycling is consistent with other calculations of internal and external CO2 fluxes in tree stems (McGuire & Teskey 2004). The relative magnitude of internal and external fluxes differed among these studies, suggesting that the internal and external fluxes may vary with species, tree age and environmental conditions. Positive calculations of JU were made throughout the 24 h period in all trees. Positive JU confirmed that when xylem sap was moving, a portion of respired CO2 was transported in xylem sap. This result is consistent with previous manipulative experiments that demonstrated upward transport of CO2 in xylem sap within stems (Teskey & McGuire 2002, 2005).

The change in storage of CO2 in the stems (JX) was a relatively small flux. Negative JX observed during the day indicate that previously dissolved CO2 was being released from the xylem sap. At night, JX was positive, that is, there was a net increase in CO2 stored in the xylem accounting for 6–8% of total RS. Diurnal changes in stem temperature were correlated with the cycling of JX between positive and negative, and may be the cause of the diurnal pattern of JX. As with other gases, the solubility of CO2 decreases with increasing temperature.

CO2 efflux to the atmosphere (JO) has commonly been used as the measure of respiration of stems and other tissues. This study provide evidence that JO comprises two components: CO2 that is released from the respiration of local tissues (JOR), and the diffusion of CO2 transported from the root system or lower in the stem (JO[CO2]). Transport of CO2 in xylem sap has been previously observed (Stringer & Kimmerer 1990; Teskey & McGuire 2002, 2005), as has the linear relationship between the concentration of transported CO2 in xylem sap and CO2 efflux from the stem (Teskey & McGuire 2002, 2005). In this study, we calculated that transported CO2 (JO[CO2]) made up 55% of the CO2 efflux measured over a 24 h period. This calculation is based on the fact that sap entering the measured volume of stem enclosed in the cuvette contains CO2 that has been transported upward from lower parts of the tree. Additional CO2 respired by local tissue and remaining within the measured volume is accounted for by the transport and storage flux terms (JU + JX).

The ratio of JO[CO2]to JOR was relatively constant because mean [CO2]S did not fluctuate greatly over the 24 h period. In other tree species, stem [CO2]S has exhibited much larger diurnal changes (McGuire & Teskey 2004). Those trees would likely also have shown larger diurnal fluctuations in the relative contributions of JO[CO2] and JOR to JO. Several factors are likely to affect the relative proportions of JO contributed by locally respired CO2 versus transported CO2: temperature, soil moisture, sap flux, barriers to gas diffusion through bark, [CO2]S and metabolic rates of live cells in the stem. These factors can make the ratio of JOR to JO[CO2] highly variable within days, across seasons and among species. The slope of the relationship between CO2 efflux and [CO2]S can also be expected to change with species, but it may also change within a species at different times of the year.

Stem respiration can be estimated in situ by measuring the evolution of CO2 from respiring cells, but the internal and external fluxes of CO2 are complex and must be separated and summed together to provide accurate estimates. Efflux to the atmosphere is only one component, and alone it is insufficient for estimating stem respiration because it is a combination of transported and locally derived CO2. Bowman et al. (2005) demonstrated that measured stem CO2 efflux in the daytime is lower than efflux predicted from the relationship between nighttime efflux and temperature, and that the reduction in daytime efflux is negatively correlated with sap flux. They used these relationships, along with the close correspondence between JO and RS at night, when JU is near zero (McGuire & Teskey 2004), to estimate rates of stem respiration throughout the day. Our new evidence indicates that caution may be needed when using this approach because JO appears to consist of a combination of locally respired CO2 and transported CO2. The rate of release of the former depends on metabolic processes, and the rate of release of the latter is determined by factors affecting diffusion and solubility. When [CO2]S is high, there is a potential for a large error when equating JO with RS even when sap flux and JU are 0.

The first estimate of stem respiration of intact trees was based of the rate of CO2 efflux to the atmosphere (Johansson 1933). Early researchers in this subject, including Johansson, were aware of the reports of high [CO2] within tree stems. In a critique of the Johansson (1993) paper, Boysen-Jensen (1933) speculated that Johansson's experimental approach was flawed because respired CO2 could be transported within the stem away from the site of respiration. Boysen-Jensen (1993) argued that this transport meant that not all of the CO2 released from the respiring cells would be accounted for by efflux measurements alone. Our study supports that contention. Our measurements also agree with later observations of reduced rates of apparent stem respiration when transpiration rates increased (Negisi 1979; Hari et al. 1991; Martin et al. 1994; Levy et al. 1999; Bowman et al. 2005; Gansert & Burgdorf 2005). The cause of this reduction can be attributed to an increase in JU with increasing sap flux and the associated decrease in [CO2]S caused by the dilution effect.

In most studies, the respiration rate of woody tissues has been defined as CO2 efflux, ignoring the potential for respiratory CO2 to dissolve in xylem sap and to be transported within the tree. We have presented evidence that CO2 efflux from the stem is not an accurate measure of respiration. This conclusion does not mean that CO2 efflux is not a useful measurement. CO2 leaving a stem or branch represents respired CO2, although not necessarily CO2 released from respiration of the local tissue. Just as the origin of the CO2 that fluxes from the soil cannot be determined exactly, the CO2 diffusing from the stem can be of soil, root or stem origin. Measures of CO2 efflux from woody tissue represent losses of carbon from the plant or ecosystem. It is important to characterize these losses. For example, Damesin et al. (2002) estimated that one-third of the total CO2 efflux (including soil efflux) from a Fagus sylvatica forest was released from tree stems and branches. Although Damesin and co-workers attributed this efflux to branch and stem respiration (as is typical), we contend that the flux includes a measure of CO2 loss from the ecosystem that occurred through the stems and branches.

In conclusion, this study has shown that a portion of respired CO2 remains within the tree and moves in xylem sap, adding to the complexity of the relationship between CO2 efflux and respiration. Over a 24 h period, 34% of the CO2 released by respiring cells in sycamore stems was transported internally away from the site of evolution. Inaddition, transported CO2 affected total CO2 efflux to the atmosphere by contributing about half of the CO2 released from the stems at the point of measurement. The relative amounts of CO2 contributed from local and transported sources to total efflux from the stem can be expected to vary diurnally and seasonally depending on many factors including sap velocity, temperature, [CO2]S, live cell volume, tree diameter and species. We conclude that CO2 efflux to the atmosphere is not equivalent to woody tissue respiration of intact trees. We recommend adopting the terminology ‘CO2 efflux’ rather than ‘respiration’ to depict measurements of CO2 release from woody tissues to the atmosphere.


The project was supported by the National Research Initiative of the US Department of Agriculture (USDA) Cooperative State Research, Education and Extension Service, Grant No. 2003-35100-13783; National Science Foundation Grant IOB 0445495; and the Global Forest Foundation.