The timing and magnitude of methane emissions from landscapes to the atmosphere are critical inputs to terrestrial carbon cycle models. Methane production, oxidation, and transport can be affected by non-cyclical, episodic water level changes (1-10 times per year; 1-5 days duration) driven by rain, management, storms or tides. These changes may provoke sudden burst-like emissions of up to 11% of annual emissions as floodwaters are mixed, drained or dried out. They may also invoke longer periods of dampened fluxes as microbial communities and emission pathways recover from oxygenation. Episodic events can generate “hot moments” that accumulate to higher annual emissions, and require higher-frequency monitoring data to predict.

We use a new, synthesized database of CH4 fluxes measured by eddy covariance, FLUXNET-CH4, to test the predictability of methane-emissions dynamics in flooded landscapes. The research has several objectives: (1) describe and generalize the timing and magnitude of CH4 response to when water level drops below the soil surface and (2) describe the time for methane emission rates to recover following reflooding after dry periods of different durations. Next we explore whether predictive algorithms developed for water table depth can be generalized to terms such as soil moisture, oxygen level, reduction potential or pressure (which may impact ebullition). There are challenges to universalizing a response to water table using this synthesis database:

1. Lack of consistent measurements of water level.
2. Lack of significant response to water level dynamics.
3. Differentiating the effect of different sources of water level change

In the latter, lateral flooding, tides, precipitation, and irrigation may each drive a different magnitude of response, due to temperature, salinity, or other characteristics.

The work should help improve predictions in methane-producing landscapes with variable water levels, such as rice paddies, floodplain wetlands, coastal regions, and peatland complexes. The research can support decisions regarding water table management in controlled wet landscapes. It may also guide measurement strategies given the simultaneous importance and expense of monitoring for “hot moments”.

We acknowledge the FLUXNET-CH4 contributors for the data used for these analyses.