To improve understanding of CH4 fluxes across wetlands globally, we are conducting a synthesis of a growing eddy covariance CH4 flux database, FLUXNET-CH4. FLUXNET-CH4 is a partnership between the Global Carbon Project, Ameriflux, and EuroFlux to compile a global database of existing eddy covariance CH4 flux measurements. Currently, over 60 sites are included in the database, which represents well over 100 site-years of wetland CH4 flux measurements. Methane exchange in wetlands is complex, involving nonlinear processes across multiple spatial and temporal scales. These processes are poorly characterized at ecosystem scales, but are important for accurate representation of CH4 fluxes in process-based models. The analysis of continuous, ecosystem-scale flux measurements across wetlands offers significant promise for furthering understanding of wetland CH4 flux dynamics.

To separate the complex nature of wetland CH4 emissions, we combined multiple statistical approaches to analyze the FLUXNET-CH4 database, including generalized linear models, generalized additive models, wavelet analysis and information theory. Wavelet analysis highlighted that CH4 fluxes exhibited strong variability over a range of time scales, with the variation across time scales differing by wetland type; across wetland types, seasonal variation was greatest in rice paddies, diel variation was greatest in freshwater marshes, and hourly variation was greatest in bogs, fens and wet tundra. Multiple approaches found that soil temperature was the dominant driver of CH4 flux seasonally, whereas wavelet analysis coupled with information theory identified air pressure as an important near-synchronous driver of CH4 emissions at the multi-day scale. Water table fluctuations were also identified as an important asynchronous control on CH4 exchange at the multiday scale. By drawing on the strengths of multiple methods, we aim to continue to provide new insights on the scale-emergent, nonlinear, and lagged processes of CH4 exchange. We acknowledge the FLUXNET-CH₄ contributors for the data provided in these analyses.