Closed- and open-path methane gas analyzers are used in eddy covariance systems to compare three potential methane emitting ecosystems in the Sacramento-San Joaquin Delta (CA, USA): a rice field, a peatland pasture and a restored wetland. The study points out similarities and differences of the systems in field experiments and data processing. The closed-path system, despite a less intrusive placement with the sonic anemometer, required more care and power. In contrast, the open-path system appears more versatile for a remote and unattended experimental site. Overall, the two systems have comparable minimum detectable limits, but synchronization between wind speed and methane data, air density corrections and spectral losses have different impacts on the computed flux covariances. For the closed-path analyzer, air density effects are less important, but the synchronization and spectral losses may represent a problem when fluxes are small or when an undersized pump is used. For the open-path analyzer air density corrections are greater, due to spectroscopy effects and the classic Webb–Pearman–Leuning correction. Comparison between the 30-min fluxes reveals good agreement in terms of magnitudes between open-path and closed-path flux systems. However, the scatter is large, as consequence of the intensive data processing which both systems require.