Greenhouse gas (GHG) fluxes from tree stems is a relatively new area of research that is gaining recognition as a potentially important source of GHG emissions. Understanding temporal dynamics of tree stem fluxes is particularly important because the magnitude and patterns of these fluxes are poorly understood (especially methane) and the production sources and transport mechanisms of GHG’s in tree stems are still largely unknown. As a result not enough is known about tree stem emissions to determine their contribution to whole ecosystem GHG budgets (Vargas & Barba, 2019). Most studies measuring tree stem GHG fluxes are performed with manual chamber measurements, with a typical frequency of 2-3 weeks which does not provide the resolution required to fully understand the temporal dynamics of these emissions. In this case study, Dr. Barba used the eosAC soil gas flux chambers to investigate the temporal dynamics of fluxes from the stems of oak trees in order to better understand these underlying dynamics.

Study Objectives and Equipment

The goal of this study was twofold. First, Dr. Barba wanted to test if the eosAC automated soil flux chambers were suitable for collecting continuous measurements of tree stem emissions. The eosAC chambers are designed to measure soil respiration, and thus meant to be used in a horizontal position, so vertical positioning of the chambers on tree stems presented a new set of challenges. Second, Dr. Barba wanted to study if pedunculate oak trees could exchange CO2 and CH4 with the atmosphere, and if so, he wanted to use the continuous data he collected to understand the temporal variability of those fluxes.

The Study Site

This study was conducted in Staffordshire, central England, United Kingdom (52.801°N, 2.301°W), at the Birmingham Institute of Forest Research or BIFoR. The experimental site was located within BIFoR’s Free-Air Carbon Dioxide Enrichment (FACE) experimental forest (Hart et al. 2020), but the plot of trees selected was located outside the elevated CO2 plots (see Figure 1 for chamber locations). The forest itself is dominated by pedunculate oak (Quercus robur) in the upper canopy and by common hazel (Corylus avellana) in the understorey. The mean annual air temperature in this region is 10.3 ± 5.4°C.

Figure 1. Aerial view of BIFoR FACE experimental forest. The approximate location of the plot of trees used in this study is shown in the shaded polygon and the locations of the eosAC chambers are represented by black dots (credit: Norbury Park).

Installing the eosAC Chambers

Four eosAC chambers were installed on four different oak trees by Dr. Josep Barba. The trees are mature, with a mean diameter at breast height (DBH) of 60 cm. This particular oak species was selected because not only is it the dominant species in the forest, it is also one of the most widely distributed tree species in Europe.

The chambers were installed on the surface of the tree stems at a height of 70 cm above the ground. One PVC collar was installed on each tree stem 10 days before starting the measurements. Before installation, the collars were trimmed to fit the surface contours of the tree, then silicone was used for sealing the collars with the bark (Figure 2). 

Once the silicone was dry, the chambers were secured to the tree stems with ratchet straps. The final installation is shown in Figure 3.

Figure 2. Cutting PVC collars to fit the contours of the tree stems for installation of the eosAC automates flux chambers.
Figure 3. The eosAC after being installed and secured to the stem of a tree at the field site.

Collecting Measurements

Gas collected in the chambers was measured using an ABB-LGR M-GGA-918 gas analyzer (hereafter referred to as an MGGA), and eosMX multiplexer so all four chambers could be measured using a single gas analyzer. The gas analyzer and the multiplexer were placed inside a trailer in the field location to protect them from the elements (Figure 4). 

The eosLink-MX software was used to schedule chamber measurements and were programmed to measure for 10 minutes per hour. During the study, measurements were collected continuously from October 28 to December 2, 2020 (~ 2400 measurements in total).

Figure 4. The eosMX multiplexer draws gases from the eosAC chambers to be analyzed by the MGGA gas analyzer.

Data Processing and Analysis

The raw data was processed with the eosAnalyze-AC software. The software matches output files from the eosMX multiplexer with output files from the gas analyzer allowing users to view individual chamber measurements and calculate fluxes for individual gas species using either a linear or exponential fit. eosAnalyze-AC also allows users to set custom dead bands and other variables like tubing length and collar depth for more accurate flux calculations. Because CO2 concentration within the chambers during measurements did not always follow a linear pattern, the software’s exponential curve fitting option was selected to calculate fluxes, as it best fit the data (Figure 5).

Figure 5. Calculating fluxes from raw data using the eosAnalyze-AC software using an exponential function. Right and left panels for carbon dioxide and methane, respectively.

Where fluxes were closer to 0, as for CH4 measurements, linear regression fit better than exponential as previously demonstrated for soil and stem measurements (Barba et al. 2019), so a linear fit was used for those measurements (Figure 6). 

Once the fluxes were calculated, the statistical software package R was used to further analyze and plot the data.

Figure 6. Calculating fluxes from raw data using the eosAnalyze-AC software using a linear regression function. Right and left panels for carbon dioxide and methane, respectively.

Findings Thus Far

The results of this study show that oak tree stems were able to exchange CO2 and CH4 with the atmosphere (Figure 7). Tree stems did emit CO2 during the study period, with emissions showing a downward trend that followed a decreasing trend in air temperature. Despite all trees presented similar seasonal patterns, magnitudes of CO2 emissions were tree specific. For example, tree number one had mean emissions of 0.33 ± 0.10 μmols CO2 m2 s-1 [mean ± sd] and tree number three had mean emissions of 0.96 ± 0.26 μmols CO2 m-2 s-1. We did not expect such differences between trees because all trees were mature, from the same species and at the same location.    

All trees presented both stem CH4 emissions and uptake over the study period, with mean fluxes being close to neutral (-0.018 ± 1.22 nmols CH4 m-2 s-1). These fluxes were lower than anticipated and did not show seasonal trends or dependency on environmental conditions, but that could be normal given that the measurements were performed at the end of Fall, with no tree activity (no leaves on the canopies) and with low temperatures. However, CH4 magnitudes seem to differ between trees, with tree number two mostly showing emissions and tree number 3 showing uptake (Figure 7).

Figure 7. Daily mean fluxes of 24 hourly measurements for CO2 (middle) and CH4 (bottom) compared to recorded atmospheric temperature and soil water content (top). Fluxes of CO2 follow trends in atmospheric temperature much more closely than fluxes of CH4. The shaded areas represent standard deviations of CO2 and CH4 fluxes.
Figure 8. A few days worth of flux measurements collected from tree stems using the eosAC. The plots emphasize the amount of temporal variation observed on the fluxes, even at short (e.g. diurnal and day-to-day) time scales.

Automated hourly measurements revealed similar fluxes within trees for CO2 at short time scales, with no evident diurnal patterns, likely due to the small diel variability of air temperature over the study period, and the time of the year (non-growing season) (Figure 8). CH4 fluxes presented a great variability within days, with no evidences of diurnal cycles. Fluxes changed from emissions to uptake at short periods of time. These results suggest that the commonly used manual measurements (with a typical measurements frequency of 2-4 weeks) might not be suitable for studying temporal trends of stem GHG fluxes (particularly for CH4 fluxes). Additionally, automated measurements could better capture changes in stem CH4 fluxes at short temporal scales, resulting in more accurate mean fluxes when integrated.  

Longer periods of continuous measurements of stem GHG fluxes spanning multiple environmental conditions (also covering the growing season), will be essential for exploring potential seasonal trends of CO2 and CH4 and identifying the drivers of those fluxes. 


Continuous measurements are essential in order to understand temporal variability of CO2 and CH4 tree stem fluxes. Adequately capturing this temporal variability is also essential for reliable mean fluxes. Continuous measurements allow us to both understand temporal dynamics and reliably calculate mean fluxes. For this study, Dr. Barba measured four trees every hour over a month, and he saw that all trees emitted CO2 (0.67 ± 0.30 μmols CO2 m-2 s-1, mean ± sd) with strong differences between trees. All trees were able to either uptake and emit CH4 depending on the moment, but net fluxes equated to a slight uptake (-0.018 ± 1.22 nmols CH4 m-2 s-1).  Planned future work includes: 1) measuring tree stem fluxes continuously for a year to determine if there are CH4 annual cycles; 2) explore spatial variability within a given tree to determine if tree stems present a vertical pattern with stem height; 3) explore if elevated atmospheric CO2 affects stem CH4 fluxes.

Eosense chambers were suitable for measuring tree stem fluxes. The small size of the chambers provides flexibility to adjust the chamber to the roughness and contours of the tree stems, and the design of the chambers enables them to function properly in a vertical orientation. Overall, the eosAC flux chambers, coupled with an M-GGA-918, and controlled by eosMX multiplexer performed well for measuring continuous tree stem GHG fluxes.


Thanks to Josep Barba and Vincent Gauci for sharing their photos, performing the measurements and analyzing the data associated with this study.


Barba, J., Poyatos, R., Vargas, R., 2019. Automated measurements of greenhouse gases fluxes from tree stems and soils: magnitudes, patterns and drivers. Sci. Rep. 9, 1–13.

Hart, K.M.; Curioni, G.; Blaen, P.; Harper, N.J.; Miles, P.; Lewin, K.F.; Nagy, J.; Bannister, E.J.; Cai, X.M.; Thomas, R.M.; et al. Characteristics of free air carbon dioxide enrichment of a northern temperate forest. Glob. Chang. Biol. 2019, 26, 1023–1037.

R Core Team (2019). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL

Vargas, R., Barba, J., 2019. Greenhouse gas fluxes from tree stems. Trends Plant Sci. Vol. 24, no. 4, pp. 296-299.