Empirical Dynamic Modeling Reveals Complexity of Methane Fluxes in a Temperate Salt Marsh.

Methane dynamics within salt marshes are complex because vegetation types, temperature, oscillating water levels, and changes in salinity and redox conditions influence CH4 production, consumption, oxidation, and emissions. These non‐linear and complex interactions among variables affect the traditionally expected functional relationships and present challenges for interpreting and developing process‐based models. We employed empirical dynamic modeling (EDM) and convergent cross mapping (CCM) as a novel approach for characterizing seasonal/multiday and diurnal CH4 dynamics by inferring causal variables, lags, and interconnections among multiple biophysical variables within a temperate salt marsh using 5 years of eddy covariance data. EDM/CCM is a nonparametric approach capable of quantifying the coupling between variables while determining time scales where variable interactions are the most relevant. We found that gross primary productivity, tidal creek dissolved oxygen, and temperature were important for seasonal/multiday dynamics (rho = 0.73–0.80), while water level was most important for diurnal dynamics during both the growing and dormancy phenoperiods (rho = 0.72 and 0.56, respectively). Lags for the top‐ranked variables (i.e., gross primary productivity, dissolved oxygen, temperature, water level) occurred between 1 and 5 weeks at the seasonal scale and 1–24 hr at the diurnal scale. The EDM had high prediction capabilities for intra‐/inter‐seasonal patterns and annual CH4 sums but had limitations in representing large, infrequent fluxes. Results highlight the importance of non‐linearity, drivers, lag times, and interconnections among multiple biophysical variables that regulate CH4 fluxes in tidal wetlands. This research introduces a novel approach to examining CH4 fluxes, which will aid in evaluating current paradigms in wetlands and other ecosystems. Plain Language Summary: The movement of methane gas in salt marshes is complex and influenced by various factors such as plant types, temperature, water level, and changes in water salinity and oxygen levels. These factors interact in complex ways, making predicting the outcome of methane production and movement difficult. We applied a new method of studying methane dynamics using Empirical Dynamic Modeling and Convergent Cross Mapping. After analyzing 5 years of ecosystem‐scale measurements of methane fluxes, we show that the amount of oxygen in the water, the temperature, and the amount of light received by plants are crucial for understanding regulating methane fluxes within days and across seasons. The interconnections among these variables are complex, and methane fluxes may have delayed responses (i.e., lag times), highlighting the importance of these interactions. This research improves our comprehension of how environmental factors interact to affect methane fluxes in wetlands. Key Points: Methane dynamics in salt marshes are complex and influenced by multiple variablesEmpirical dynamic modeling and convergent cross mapping are novel approaches to characterizing CH4 dynamicsResults highlight drivers, non‐linearities, lag times, and interconnections among multiple biophysical variables for CH4 dynamics [ABSTRACT FROM AUTHOR]