Methane fluxes measured in a eutrophic peat meadow in The Netherlands dominated by vascular plants showed high spatial and temporal variability. To elucidate this variability as well as the underlying processes, various measurement techniques were used: soil gradients of methane concentrations, the chamber method, and the eddy covariance technique. Additionally, soil temperature at multiple depths, soil water level, root depth, carbon dioxide fluxes, incoming radiation, atmospheric pressure, wind speed, friction velocity latent, heat fluxes, and living biomass were monitored. A comparison of the measurement techniques showed that: (a) the soil gradient method and the chamber method showed comparable methane fluxes only at a site with low water table and shallow roots, while at other sites methane fluxes were underestimated with the soil gradient method by an order of magnitude: (b) a footprint analysis showed that the chamber method and eddy covariance showed similar methane fluxes for the different land elements (dry land, wet land, and ditches plus borders). However, when up-scaling the chamber measurements over time using a regression model based on soil temperature, methane emissions were overestimated by 37% compared to the eddy covariance data. The chamber method was the best technique to assess spatial variability, while eddy covariance was best for assessing temporal variability as well as up-scaling. The soil gradient method for methane fluxes should be used with great care, and probably generates reliable results only in areas with low water table and shallow roots. Both the chamber method and eddy covariance showed significant spatial variability, which was best explained by soil water level in combination with root depth patterns. Together, these variables probably determined the net methane production and the available mechanisms for methane transport to atmosphere. The methane fluxes showed strong temporal variability at different scales: diurnal cycles, significant day-to-day variability, and seasonal variations. Clear diurnal cycles of methane fluxes were observed synchronous to incoming radiation, latent heat and net ecosystem exchange, but not synchronous to temperature. This suggested that stomatal opening and/or pressurized convective throughflow were important mechanisms for gas transport through plants. The variability at the day-today scale was best explained by soil temperature below the water level combined with soil water level. Highest methane fluxes were observed during summer and lowest fluxes during autumn and winter. The vegetation density together with temperature and length of day-light probably determined this seasonality. Also, temporal variability varied spatially, probably due to the water table and root depth. (C) 2009 Elsevier B.V. All rights reserved.
