Gas samples were collected using a static chamber approach, samples stored in vials (Exetainers), and later analysed using a gas chromatograph with flame ionisation detector for methane detection and an electron capture detector for nitrous oxide quantification. Environmental variables were collected concomitantly, including air temperature, soil temperature, soil pH, soil electrical conductivity, soil dissolved oxygen content, and water table depth. Temperature was determined using a thermocouple; pH, electrical conductivity, and dissolved oxygen using a HACH® rugged outdoor HQ30D multi meter and pH, LDO or EC probe. Water table depth was measured using a depth measure.

Diffusive gas fluxes were determined by using the JMP IN version 11 (SAS Institute, Inc., Cary, North Carolina, USA) statistical package to plot best-fit lines to the data for headspace concentration against time for individual flux chambers, with fluxes calculated from linear or non-linear regressions depending on the individual concentration trend against time. Gas mixing ratios (ppm) were converted to areal fluxes by using the Ideal Gas Law to solve for the quantity of gas in the headspace (on a mole or mass basis) and normalized by the surface area of each static flux chamber. Ebullition-derived methane fluxes were also quantified in our chambers where evidence of ebullition was found. This evidence consisted of either: (i) rapid, non-linear increases in methane concentration over time; (ii) abrupt, stochastic increases in methane concentration over time; or (iii) an abrupt stochastic increase in methane concentration, followed by a linear decline in concentration. For observations following pattern (i), flux was calculated by fitting a quadratic regression equation to the data (P < 0.05), and methane flux determined from the initial steep rise in CH4 concentration. For data following pattern (ii), the ebullition rate was determined by calculating the total methane production over the course of the bubble event, in-line with prior work conducted by the investigators. Last, for data following pattern (iii), a best-fit line was plotted to the methane concentration data after the bubble event, and a net rate of methane uptake calculated from the gradient of the line. Observations following patterns (i) and (ii) were categorized as “ebullition” (i.e. net efflux) whereas observations following pattern (iii) were categorized as “ebullition-driven methane uptake” (i.e. net influx).

Flux data were removed (i.e. filtered) from the dataset if: (i) the change in gas concentration over time was not found to be statistically significant (alpha level of 0.05); (ii) evidence were found that an individual static flux chamber was not gas-tight (evidenced by a negative carbon dioxide flux); (iii) evidence were found that two or more Exetainers had been compromised by incomplete evacuation or leakage; or (iv) other evidence that an individual flux chamber had been compromised (e.g. evidence from field observations that the static flux chamber had been disturbed during the measurement process).