9:30 AM - 9:45 AM
[ACG47-03] Continuous measurements of methane exchange at a temperate secondary forest by the modified gradient method
Keywords:Methane flux, Modified gradient method, Upland forest
Methane (CH4) budget in upland forests is highly uncertain. In this study, we measured CH4 exchange at an upland forest using the modified gradient (GR) method. We present the applicability of the GR method, comparing to CO2 fluxes measured by the eddy covariance (EC) method and CH4 fluxes by the hyperbolic relayed eddy accumulation (HREA) method. Our aim is to show a consistency in CH4 fluxes by different methods for reinforcing knowledge of CH4 budget in forest ecosystems and reducing uncertainties.
We measured fluxes at the Yamashiro forest hydrology research site in Kyoto, Japan during 2016. Turbulent fluxes of momentum, sensible heat, latent heat, and CO2 were measured by the eddy covariance method. We developed a measurement system that the HREA and GR methods could be applied simultaneously. CO2 and CH4 concentration at two height (35, and 25 m) were measured above the canopy. Horizontal wind speed measured at 35, 25, and 22 m for estimating the displacement height. We also examined two diffusion coefficients based on site-specific universal functions for temperature (Φh) and CO2 (Φc). Based on the specification of the available gas analyzer (FGGR-24r-EP, Los Gatos Research, USA), CH4 concentration gradient was expected to be near a detectable limit of the analyzer. Thus, we examined the applicability of nighttime data for the GR method under stable conditions when vertical CH4 concentration was expected to be large.
CO2 fluxes by the EC and GR methods were reasonably agreed for stable and unstable conditions (R2 = 0.66 - 0.68, RMSE = 5.66 - 6.66 gCO2 m-2 d-1); but, no consistency was found in CH4 fluxes from the HREA and GR methods. Using data under stable conditions, CO2 fluxes by the GR method using Φh was 50% overestimated in comparison with those by the EC method. On the other hand, overestimation was not found in CO2 fluxes using Φc. For stable conditions, the value of R2 among CO2 fluxes by two methods increased with integration times; by averaging over 30 days or more, CO2 fluxes by two methods showed a consistency (R2 = 0.86 - 0.91, RMSE = 5.73 - 6.42 gCO2 m-2 d-1). This results suggests that the random errors associated with eddy diffusivity were reduced at the monthly time scale.
Monthly CH4 fluxes by the GR method (0.63 - 1.79 mgCH4 m-2 d-1) and the HREA method (0.58 - 1.96 mgCH4 m-2 d-1) showed similar seasonal variations during the period from June to October and December. A disagreement during the period from January to March was caused by long-term missing data of the EC or the HREA method. The disagreement in April and May was caused by short integration time for determining vertical concentration differences. The consistent seasonal variations among two methods indicates that the GR method under stable conditions could be applicable for measuring CH4 fluxes at this forest.
Based on the GR and HREA measurements, the forest acted as a net annual CH4 source (GR; 172 mgCH4 m-2 yr-1, HREA; 237 mg CH4 m-2 yr-1). Monthly CH4 fluxes by the GR and HREA methods delayed one month to the monthly precipitation during the period from June to October (R2 = 0.97, p < 0.01). This was probably because rainfall turned soils anaerobic conditions gradually, and activations of methanogenic bacteria took time.
We measured fluxes at the Yamashiro forest hydrology research site in Kyoto, Japan during 2016. Turbulent fluxes of momentum, sensible heat, latent heat, and CO2 were measured by the eddy covariance method. We developed a measurement system that the HREA and GR methods could be applied simultaneously. CO2 and CH4 concentration at two height (35, and 25 m) were measured above the canopy. Horizontal wind speed measured at 35, 25, and 22 m for estimating the displacement height. We also examined two diffusion coefficients based on site-specific universal functions for temperature (Φh) and CO2 (Φc). Based on the specification of the available gas analyzer (FGGR-24r-EP, Los Gatos Research, USA), CH4 concentration gradient was expected to be near a detectable limit of the analyzer. Thus, we examined the applicability of nighttime data for the GR method under stable conditions when vertical CH4 concentration was expected to be large.
CO2 fluxes by the EC and GR methods were reasonably agreed for stable and unstable conditions (R2 = 0.66 - 0.68, RMSE = 5.66 - 6.66 gCO2 m-2 d-1); but, no consistency was found in CH4 fluxes from the HREA and GR methods. Using data under stable conditions, CO2 fluxes by the GR method using Φh was 50% overestimated in comparison with those by the EC method. On the other hand, overestimation was not found in CO2 fluxes using Φc. For stable conditions, the value of R2 among CO2 fluxes by two methods increased with integration times; by averaging over 30 days or more, CO2 fluxes by two methods showed a consistency (R2 = 0.86 - 0.91, RMSE = 5.73 - 6.42 gCO2 m-2 d-1). This results suggests that the random errors associated with eddy diffusivity were reduced at the monthly time scale.
Monthly CH4 fluxes by the GR method (0.63 - 1.79 mgCH4 m-2 d-1) and the HREA method (0.58 - 1.96 mgCH4 m-2 d-1) showed similar seasonal variations during the period from June to October and December. A disagreement during the period from January to March was caused by long-term missing data of the EC or the HREA method. The disagreement in April and May was caused by short integration time for determining vertical concentration differences. The consistent seasonal variations among two methods indicates that the GR method under stable conditions could be applicable for measuring CH4 fluxes at this forest.
Based on the GR and HREA measurements, the forest acted as a net annual CH4 source (GR; 172 mgCH4 m-2 yr-1, HREA; 237 mg CH4 m-2 yr-1). Monthly CH4 fluxes by the GR and HREA methods delayed one month to the monthly precipitation during the period from June to October (R2 = 0.97, p < 0.01). This was probably because rainfall turned soils anaerobic conditions gradually, and activations of methanogenic bacteria took time.