Mechanisms of Methane Transport Through Mechanisms of Methane Transport Through Populus trichocarpa Populus trichocarpa

. Although the dynamics of methane (CH 4 ) emission from croplands and wetlands have been fairly well investigated, the contribution of trees to global CH 4 emission and the mechanisms of tree transport are relatively unknown. CH 4 emissions from the common wetland tree species P opulus t richocarpa (black cottonwood) native to the Paciﬁc Northwest were measured under hydroponic conditions in order to separate plant transport mechanisms from the inﬂuence of soil processes. Roots were exposed to CH 4 enriched water and canopy emissions of CH 4 were measured. The average ﬂux for 34 trials (at temper-5 atures ranging from 17 to 25 o C) was 2.8 ± 2.2 µ g CH 4 min − 1 (whole canopy). Flux increased with temperature. Compared to the isotopic composition of root water CH 4 , δ 13 C values were depleted for canopy CH 4 where the warmest temperatures (24.4-28.7 o C) resulted in an epsilon of 2.8 ± 4.7‰; midrange temperatures (20.4-22.1 o C) produced an epsilon of 7.5 ± 3.1 ‰; and the coolest temperatures (16.0-19.1 o C) produced an epsilon of 10.2 ± 3.2 ‰. From these results it is concluded that there are likely multiple transport processes at work in CH 4 transport through trees and the dominance of these processes 10 changes with temperature. The transport mechanisms that dominate at low temperature and low ﬂux result in a larger fractionation, while the transport mechanisms that prevail at high temperature and high ﬂux produce a small fractionation. Further work would investigate what combination of mechanisms are speciﬁcally engaged in transport for a given fractionation of emitted CH 4 .


Introduction
A significant portion of the global cycle of methane (CH 4 ) takes place in terrestrial ecosystems, although it is its importance as a greenhouse gas that garners attention.CH 4 is second only to CO 2 in terms of the radiative forcing of greenhouse gasses at current concentrations (Myhre et al., 2013).Although the rate of increase of CH 4 in the atmosphere has slowed, atmospheric CH 4 concentrations continue to rise (Ciais et al., 2013).Uncertainties in emissions estimates persist, but the largest single source of CH 4 emissions is natural wetlands (Ciais et al., 2013).In one study of a fen ecosystem, up to 90% of CH 4 emission was observed to be through plants (Whiting and Chanton, 1992).CH 4 emitted through trees in flooded systems may contribute a significant fraction of these wetland emissions (Rice et al., 2010;Pangala et al., 2013).Whether in wetlands or in upland soils, it is not yet known if transport of CH 4 through trees is significant in the overall movement of CH 4 through an ecosystem.
The transport mechanisms of CH 4 through trees is potentially a controlling factor for the overall CH 4 flux and should be understood.placed in each hole and held in place by a foam stopper, then the lid was secured to the bucket.The nutrient solution was changed every ten days.Every other day the pH was checked and sodium hydroxide added to the solution if the pH was found to be lower than 5.8.In case of contamination by methanogens of the growing buckets, root water was tested just before being changed.CH 4 was found at just above the detection limits of the gas chromatographs, five orders of magnitude less than the CH 4 concentrations during the experiments described below.The small amount of CH 4 detected may have been dissolved from ambient air; production of CH 4 in the root zone of the growing buckets was concluded to be unlikely.
Saplings were cultivated for at least four months under greenhouse conditions with a photoperiod of 16 hours before experimentation began.Natural light was used unless the intensity dropped below a pre-set PPFD, in which case greenhouse lighting was set to activate.Light levels measured at leaf surfaces at the time of an experiment averaged 72 µmoles light m −2 s −1 but ranged from 3 to 530 µmoles light m −2 s −1 (LI-250A Light Meter, LI-COR Biosciences).Trees grew during the experimentation period and ranged in height from about 75 cm to 1 meter above the bucket lid level of 26 cm.Leaves on each tree were photographed once and leaf area calculated from pixel counts.A correlation of leaf area versus leaf count was done to estimate the leaf area of each tree based on leaf count at the time of each flux experiment (r 2 = 0.79).However, as this was an estimate it was not used for reporting flux per leaf area.
To perform an experiment, one tree was removed from the growing vessel and the tree's roots placed in an acrylic root chamber based on a design used by Rusch and Rennenberg (1998), as depicted in Figure (1).The chamber was filled with deionized water through which natural gas had been bubbled until a high concentration of CH 4 was reached.The concentration of CH 4 in the root chamber was similar to that found in flooded tubs of soil in greenhouse experiments where no plant was present (unpublished data).These concentrations averaged 0.7 µmoles CH 4 mL −1 water (standard deviation ∼ 0.1 µmoles CH 4 mL −1 water), which is in the range of CH 4 water concentrations in a CH 4 flux study done on hydroponically grown rice (Yao et al., 2000).The root chamber was closed around the tree stem with an o-ring between the chamber and chamber lid, then sealed around the stem with modeling clay to separate the water and roots from the ambient air.Electrical tape was used to cover the seams of the root chamber lid.Aluminum foil covered the transparent root chamber to protect the roots from ambient light.After enclosing the tree in the root chamber, the system was left in the greenhouse for at least three hours before samples were taken in the afternoon or early evening.
The water level in the root chamber was marked once the tree roots were enclosed and the chamber sealed.Just before an experiment, the water level was refilled to the initial marked level.The amount of water used to refill the root chamber was taken as a measurement of water transpired by the tree.Also before the experiment, leaf temperature, ambient temperature and relative humidity were recorded.The average ambient temperature was 18 ± 2 o C for these experiments; a set of trials with temperature variation are described in the next section.For sampling, a 100 liter tedlar bag was used to enclose the entire canopy of the tree.Since the bottom of the bag had been cut off and some of the bag material wrapped around the tree stem to seal the canopy, the net volume of the bag was approximately 90 liters.Any variation in CH 4 concentration due to these minor volume variations was later calculated to be dwarfed by the variability observed in flux data.20 mL of canopy air was sampled with a syringe from a septum on the tedlar bag every five minutes for a half hour, beginning as soon as the tedlar bag was closed around the base of the stem.10 mL water samples were taken every ten minutes using a syringe from a sampling Biogeosciences Discuss., doi:10.5194/bg-2016Discuss., doi:10.5194/bg- -60, 2016 Manuscript under review for journal Biogeosciences Published: 21 March 2016 c Author(s) 2016.CC-BY 3.0 License.port on the root chamber.20 mL ambient greenhouse air samples were taken just after the bag was closed and just before it was removed.All samples were then analyzed for CH 4 concentration on an Agilent model 6890 gas chromatograph (GC) with a flame ionization detector (FID) (Khalil et al., 1998a).The rate of CH 4 accumulation in the chamber was analyzed by linear regression to obtain the net CH 4 flux.Water samples were mixed with 20 mL N 2 and agitated by shaking for five minutes.
The water was then expelled and the CH 4 concentration determined by measurement on the GC-FID.The dissolved CH 4 concentration of the water was calculated by the method described in Lu et al. (2000).
Canopy CH 4 flux from each of eight trees was measured twice during November and December of 2010, and a third and fourth time on six of the original trees in March and April of 2011.Some growth occured between the fall and spring trials.
Leaf area was measured between the 2010 and 2011 set of flux measurements and calculated non-destructively as described above.Stem diameters were also recorded.Leaf areas ranged from 0.09 m 2 to 0.49 m 2 with an average of 0.34 m 2 .Leaf counts ranged from 125 to 424 leaves per tree with an average of 289 per tree.The correlation between leaf count and leaf area yielded a determination coefficient of r 2 = 0.79.
Stem diameters were measured also.The diameter measurement was taken at 22 cm from the bottom of the original cutting, which was approximately the height where root growth from the cutting began.Stem diameters ranged from 0.010 m to 0.016 m with an average of 0.013 m.

Temperature Variation Experiment
For the temperature experiment, black cottonwood was grown hydroponically as previously described.Five trees were used, taken as cuttings from one black cottonwood tree growing in the Sandy River Delta, Troutdale, Oregon, on September 19 th , 2011.Experimentation began on January 30 th , 2012, and ran for just over three weeks.Each week, the trees were moved to a greenhouse room with a different temperature while other environmental factors such as day length were kept constant.Leaf counts ranged from 131 to 500 leaves per tree with an average of 329 per tree; this is comparable to the trees used in the previous set of experiments as their sizes were similar.The average day-time temperature was 22.3 o C the first week, 25.2 o C the second week, and 17.1 o C the third week (compared to an average ambient temperature of 18 ± 2 o C for the experiments described in the previous section that measured canopy flux without temperature variation).Natural gas was bubbled through deionized water in the root chamber for approximately 45 minutes before the start of each experiment.The root chamber was then placed on a scale (Mettler Toledo, New Classic MF, model MS1200 1L) in the greenhouse and weighed.After moving the tree from the growing container to the root chamber, the difference in weight was recorded as tree weight.Ambient temperature and relative humidity were recorded once the tree was secured in the root chamber and during the experiment.Canopy samples and root water samples were taken and processed in the same manner as described in the previous section for measuring canopy flux without temperature variation.

Isotopic Measurements
In the canopy flux experiments outlined above, both with temperature variation and without, carbon isotopic composition of syringes were taken from the canopy bag, two from the ambient air, and 10 mL water was taken from the root chamber to be mixed with 50 mL N 2 and processed as described above.These samples were injected into stoppered 45 mL glass storage vials.
Syringe samples from these vials were measured by continuous-flow gas chromatography-isotope ratio mass spectrometry on a Thermo Scientific Delta V Advantage IRMS using the method described in Rice et al. (2001).Values for the canopy CH 4 were corrected using the ambient greenhouse samples.
By convention the ratio of heavy to light isotopes of CH 4 is expressed with the delta notation as where R samp = 13 C/ 12 C and R std is the known 13 C/ 12 C ratio of Vienna PeeDee belemnite limestone (VPDB) (Gonfiantini et al., 1995).
Discrimination for or against 13 C in a particular process can be expressed by the ratio of 13 C to 12 C for the process, or C ratio prior to the process and R B is the ratio afterwards.The process may be a chemical reaction, dissolution from water into a gas phase, or diffusion along a gradient.α is known as the fractionation factor.Also convenient is the expression = (α − 1) • 1000 Using equation ( 1) this can be written as It can be shown (Hoefs, 2004) that for small values of , ∼ δ A − δ B .

Stem Experiments
In the interest of separating stem emission from any potential emission occuring from the leaves, two acrylic cylindrical stem cuvettes (20cm in length, 5.8 cm radius) were used to measure flux exclusively emitted from the stem, based on a design used by Rusch and Rennenberg (1998) and depicted in Figure ( 2).These cuvettes were attached around a section of stem 11.7 cm in length, or 23 cm if both stem chambers are used.Experiments with the stem cuvettes were identical to those outlined in the section for canopy flux measurements, except instead of the tedlar bag, the stem cuvettes were attached for a half hour around the tree stem while flux samples were taken every five minutes from each cuvette.The tree was secured in the root chamber for a minimum of 190 minutes and a maximum of 370 minutes before measurements were taken.Cuvettes were placed along the main stem of the tree, one above the other.Water concentration samples were taken as they were for the canopy flux experiments.A total of four trials were conducted, two of which were at different times with the same tree.

Canopy Flux
Concentration of CH 4 in the tedlar bag versus time was graphed for all trials.Of the 28 canopy flux trials without temperature variation performed 2010-2011, 18 had an r-squared value over 0.95 and three had an r-squared value between 0.9 and 0.95.Although the constant-temperature experiments were conducted at a different time than the temperature-varying experiments, since a relationship between temperature and flux was sought the data from these experiments were combined.As described in the Methods section, growth conditions were the same in each case with the exception of temperature variation.
When analyzed, the fluxes were not normally distributed.Therefore, the nonparametric Kruskal-Wallis test was performed in

Isotopic Data
Average root water δ 13 C was -36.7‰ and did not vary significantly (95% C.I. (-36.2‰,-37.2‰)).Compared to the isotopic composition of root water CH 4 , δ 13 C values of canopy CH 4 were depleted on average by 7.8 ± 4.4 ‰.These were binned into the same three temperature groups as the flux data given above; results are given in Table (1

Stem Flux
Results from the stem chamber experiments are summarized in Table (2 Obviously this calculation assumes that flux along the entire tree stem is consistent and represented by that measured in the two chambers; the relevance of these values will be taken up in the discussion.Transpiration, based on water loss from the root chamber, is also reported.

Discussion
As a common species in wetland environments that also grows well under non-inundated conditions, it is assumed that transport mechanisms in poplar will be representative of other tree species in similar environments.The versatility of poplar makes it ideal for transport studies in this sense.Although experiments were conducted on trees grown in a hydroponic environment, the physiological adaptations of poplar expressed under inundation in the field should be same; therefore the transport mechanisms themselves should not differ in essence.Hydroponically grown rice has been used for CH 4 transport studies where greater control was necessary than what was possible in the field, yet the influence of physical characteristics of the plant or temperature on CH 4 transport were studied (Hosono and Nouchi, 1997;Yao et al., 2000).
Regardless of the mechanisms, trees having the capability of moving CH 4 from the soil to the atmosphere will do so when CH 4 is present in the root zone, as demonstrated by the fact that in all experiments where CH 4 was present in the root zone, flux was measured from the canopy.Although it has been suggested that the developmental stage of a tree may play a role in the ability of a tree to transport CH 4 (Pangala et al., 2014), the specific CH 4 transport mechanisms themselves are not expected to be altered by more mature tree growth.Therefore, the mechanisms described here should be applicable qualitatively to more mature trees, even though there may be quantitative shifts in CH 4 flux.
For the more thoroughly studied rice plant, correlations between morphophysiological and anatomical characteristics of rice plants and CH 4 emission exist (Nouchi et al., 1990;Das and Baruah, 2008).Although a large amount of CH 4 has been shown to come through emergent wetland plants (Shannon et al., 1996), physiological correlations for tree emission of CH 4 have been studied very little and the role that woody vegetation has in wetland emission of CH 4 is still poorly understood (Vann and Megonigal, 2003;Terazawa et al., 2007) 2013).In the case of Rusch and Rennenberg (1998), this was taken as evidence for the major transport of CH 4 to be by diffusion.While the stem pathway does appear to dominate, the exact mechanisms of transport may or may not include diffusion as discussed below.
An analysis of the possible physical pathways of CH 4 through the tree is helpful in interpreting the isotopic fractionation of CH 4 observed from canopy measurements.CH 4 entering the root system with water can move apoplastically, or without traversing cellular membranes.A membrane barrier at the root endodermis must be crossed before the xylem tissue is entered, which is the main bulk-flow water transport system through roots and stems of trees (Tyree and Ewers, 1991).This xylem pathway is driven by a hydrostatic pressure gradient, as opposed to the water potential that moves water from cell to cell across membrane barriers (Tyree and Ewers, 1991).The bulk flow of water causes no isotopic fractionation.Therefore, CH 4 transported by transpiration would show an isotopic fractionation due to dissolution from water to gas and from the crossing of two cell membrane barriers at the root endodermis.Experimentation is needed to know the isotopic fractionation due to membrane transport.
Alternatively, CH 4 may enter the roots without ever being drawn into the xylem tissue, instead entering air-filled aerenchyma tissue connected to lenticels along the tree's stems.This pathway is believed to be the dominant pathway for CH 4 flux from rice and some wetland plants (Denier Van Der Gon and Van Breemen, 1993;Chanton, 2005).Isotopic fractionation would occur due to dissolution from water to gas and from any concentration-gradient driven diffusion between the root zone and the atmosphere.Transfer of CH 4 from water to air leads to an isotopic fractionation on the order of -0.8‰ (Knox et al., 1992), an order of magnitude smaller than the fractionation measured here.
Molecular diffusion is described by Fick's first law, where molecules move along a concentration gradient by random collisions: J is the flux in mass per area per time, D is the diffusion coefficient in length 2 per time, C is the concentration and z is distance, and dC/dz is the concentration gradient.For a gas moving along a concentration gradient through air, the diffusion coefficient is related to the molecular weights of the constituents: where M 1 is taken to be the molecular weight of air and M 2 for the molecular weight of CH 4 (Mason and Marrero, 1970).The expected fractionation for CH 4 diffusing in air is found from the ratio of the diffusion constants for 12 CH 4 and 13 CH 4 .Using equation ( 4) and taking the molecular weight of air to be 28.9 g mol −1 , the α for diffusion is 1.019.The resulting isotopic fractionation due to molecular diffusion, therefore, is expected to be -19‰ (Chanton, 2005).barriers, bulk flow, and diffusion.However, although gas transport through aerenchyma is usually assumed to be diffusive (Chanton, 2005;Garnet et al., 2005;Machacova et al., 2013;Pangala et al., 2014), this work strongly suggests that molecular diffusion is not the primary transport mechansim.It should be noted that these fractionation values reflect the total combination of transport mechanisms through the tree: if any CH 4 does exit from other parts of the tree while most exits through the stem, the isotopic fractionation measured is from all CH 4 emitted.In other words, the values give indications about the combination of mechanisms along the pathways but does not discriminate between the pathways themselves.
The significant relationship between epsilon and canopy CH  5).This explains the decrease in epsilon with both flux and temperature; however, given the current data it is not possible to separate the influence of these two factors as drivers of epsilon.

Conclusions
Based on this and previous work, multiple mechanisms appear to be likely for CH 4 transport through trees.Most, if not all, CH 4 is transported through tree stems and not with transpiration.At least two mechanisms may be at work in this process: one dominant at low flux (and low temperature) leading to a substantial fractionation between root water and emitted CH 4 , and another mechanism dominant at higher flux (and higher temperature) with little or no fractionation.Molecular diffusion cannnot be treated as the sole transport mechanism responsible for plant transport of CH 4 ; isotopic data suggests it is not even a dominating mode of transport.As for CH 4 emissions from upland soils, because there is evidence against significant CH 4 transport through leaves with transpiration and stem emission is not supported in the literature for those conditions, it appears unlikely that CH 4 produced in the soil would be emitted in large quantities before it was oxidized.
Accurately predicting emissions feedbacks in the soil-tree-atmosphere system will depend on knowing the rate at which a particular transport mechanism changes for a given environmental parameter, e.g., temperature.An increase in CH 4 flux corresponding to temperature may be linked to a temperature-dependent mechanism within the tree, as suggested here by the fractionation data.While one part of the system may experience a strong positive feedback, such as increased CH 4 production in the soil with temperature, the feedback on transport mechanisms of CH 4 through trees could mitigate the overall effect.
Biogeosciences Discuss., doi:10.5194/bg-2016-60,2016 Manuscript under review for journal Biogeosciences Published: 21 March 2016 c Author(s) 2016.CC-BY 3.0 License.Statistical analysis includes only the 18 trials with r-squared values over 0.95, leaving out trials where inconsistent factors may have altered fluxes.Of the 17 trials conducted in the 2012 temperature variation experiment, the 16 trials with an r-squared above 0.95 are included in the analysis below.These results indicate that for the canopy, emissions are normally linear over the half-hour sampling period.Flux is reported for the entire canopy, that is, the total change in CH 4 concentration in the sampling chamber.The average temperature for the 18 trials in 2010-2011 was 18.3 o C; the average flux was 3.0 ± 2.6 µg CH 4 min −1 .In the 2012 temperature variation experiment, average daytime temperature the first week was 22.3 o C, 25.2 o C the second week, and 17.1 o C the third week.The average flux during the 2012 temperature experiment was 2.5 ± 1.7 µg CH 4 min −1 .The average flux for the warmest temperature, 25.2 o C, was 3.2 ± 1.4 µg CH 4 min −1 ; the average flux at the next highest temperature, 22.3 o C, was 3.2 ± 2.2 µg CH 4 min −1 ; at the coolest temperature of 17.1 o C, the average flux was 1.6 ± 0.8 µg CH 4 min −1 .
place of ANOVA.Fluxes were binned into three groups, the first ranging from 16 o C to 19.1 o C, the second from 20.4 o C to 22.1 o C, and the third from 24.4 o C to 28.7 o C. Fluxes were found to vary significantly with temperature (p = 0.0012).A linear (robust) regression of CH 4 flux versus temperature is shown in Figure (3) (R 2 = 0.61; the p-values of the coefficients f(x) = 2.53x10 −7 x -2.67x10 −6 are 7.0x10 −4 and 0.60x10 −1 ).
Figure(5).The value of decreases with increasing flux and increasing temperature: the root water values are stable from saturating the root chamber with CH 4 , and canopy δ 13 C values become more enriched with increasing temperature and increasing flux.Both flux and temperature have a statistically significant correlation with epsilon (R 2 for epsilon versus flux is 0.37; the p-values of the coefficients f(x) = -1.64x106 x + 11.8 are 7.12x10 −6 and 5.33x10 −12 ; R 2 for epsilon versus temperature is also 0.37; the p-values of the coefficients f(x) = -0.88x + 25.1 are 4.51x10 −5 and 2.57x10 −7 ).
).The lower stem chamber is referred to as S1 and the upper chamber as S2.Placement of the stem chambers (lower side) ranged from 0.03 m to 0.46 m along the main tree stem.6 Biogeosciences Discuss., doi:10.5194/bg-2016-60,2016 Manuscript under review for journal Biogeosciences Published: 21 March 2016 c Author(s) 2016.CC-BY 3.0 License.The raw flux is given from each stem chamber: fluxes ranged from 0.015 to 1.7 µg CH 4 min −1 with an average of 0.65 µg CH 4 min −1 .Fluxes per enclosed stem area are simply the raw flux divided by the average stem area enclosed by the chamber (0.0058 m 2 for the first trial, 0.0048 m 2 for the following three).The estimated total stem flux is calculated by dividing the total main tree stem length by the length enclosed by the two stem chambers, then multiplying by the sum of the two stem chamber fluxes.
. It should be mentioned that more recently, Pangala et al. (2014) did find a strong correlation between lenticel density and CH 4 flux from stems in the tree species Alnus glutinosa.It is clear from the data here that temperature and flux, δ 13 C and flux and δ 13 C and temperature are significantly related.Further study is needed to establish rigorous links between CH 4 flux from trees with physiological parameters.Little evidence was found to indicate leaf flux in general.Calculations dividing CH 4 flux by estimated leaf area indicated that any leaf emission would be below detection limits; regardless, twice leaf emissions were tested using a leaf cuvette instead of a canopy bag under the same experimental conditions as described above.No emission was measured.CH 4 emission measured in the stem experiment and scaled to the stem area for the tree yields a total CH 4 flux value in the range of measured fluxes for Biogeosciences Discuss., doi:10.5194/bg-2016-60,2016 Manuscript under review for journal Biogeosciences Published: 21 March 2016 c Author(s) 2016.CC-BY 3.0 License. the entire canopy.The average flux for the 18 canopy emission trials in 2010-2011 was 3.0 ± 2.6 µg CH 4 min −1 .From Table(2), the estimated total stem fluxes from the main stem of the tree based on stem chamber measurements were 3.1, 3.8, and 12 µg CH 4 min −1 .This suggests that a considerable amount of the total CH 4 emitted exits from stems after traveling through the tree, possibly via air-filled aerenchyma tissue.A decrease in flux was observed with height of the stem chamber along the stem, consistent with findings byRusch and Rennenberg (1998);Terazawa et al. (2007);Pangala et al. ( Biogeosciences Discuss., doi:10.5194/bg-2016-60,2016 Manuscript under review for journal Biogeosciences Published: 21 March 2016 c Author(s) 2016.CC-BY 3.0 License.The isotopic fractionations of ∼3 -10‰ measured here indicate that CH 4 transported through the tree follows neither a purely bulk flow pathway nor a purely diffusional one, as has been suggested for rice and other wetland plants.The value can be interpreted to indicate two or more mechanisms of transport, possibly some combination of movement through membrane 4 flux (Figure (4)) can be explained by different transport mechanisms dominating depending on the amount of CH 4 emitted.Given at least a small amount of CH 4 in the root zone, a low flux occurs that is dominated by a particular transport mechanism or mechanisms with a larger isotopic fractionation.When various ecophysiological parameters, such as temperature, combine to result in a high flux, more CH 4 is emitted by a different transport mechanism or mechanisms with a correspondingly smaller fractionation as shown in Figure ( Biogeosciences Discuss., doi:10.5194/bg-2016-60,2016 Manuscript under review for journal Biogeosciences Published: 21 March 2016 c Author(s) 2016.CC-BY 3.0 License.

Figure 1 .
Figure 1.Depiction of experimental setup with dimensions.Four springs (not shown) are evenly spaced aroung the chamber lid, attaching to the chamber body and holding the lid down by tension.Modeling clay is used to seal between the chamber and the tree stem; electrical tape is run along the seams of the lid to prevent any leaking.

Figure 2 .Figure 3 .Figure 4 .Figure 5 .
Figure 2. Depiction of stem chamber attached to tree.Stem chamber design is based on(Rusch and Rennenberg, 1998).Actual lid pieces are square bases into which the cylindrical halves of the chamber fit, sealed with an o-ring.Not depicted are the four springs, one at each corner of the square lid base, that attach to the opposite lid piece and hold the chamber together by tension.Modeling clay is used to seal between the chamber and the tree stem; electrical tape is run along the middle seam of the chamber to prevent any leaking.Two wires running from the chamber fan and sealed through the lid with glue connect to a 12-volt battery external to the chamber.
Stem fluxes.Stem chamber height refers to the height of the lower side of the stem chamber along the main tree stem, above the top of the root chamber.Main stem heights were 0.90 m, 0.49 m for the second and third trials, and 1.62 m for the final trial.The second and third trials were performed on the same tree.*In the second stem experiment, the clay seal over the root chamber came in contact with the clay seal over the lower part of S1, creating a leak into the stem chamber and an artificially high flux.min −1 ) (µg CH4 min −1 m −2 ) (µg CH4 min −1 ) (mmol H2O s −1

Table 1 .
δ 13 C Values binned by temperature for the 2010-11 canopy flux experiments and 2012 canopy flux experiments with temperature