Salinity gradients across estuaries influence wetland carbon storage, methane (CH4) biogeochemistry, and plant productivity. Estuarine freshwater wetlands may experience increases in salinity during drought; however, the impact of salinization on greenhouse gas (GHG) emissions is uncertain. We measured ecosystem-scale GHG emissions from a wetland experiencing salinization during the 2011–2017 California drought and used information theory analyses to quantify couplings and causal interactions between CH4 fluxes and two dominant environmental drivers; gross ecosystem photosynthesis and temperature. Machine learning models were then used to estimate salinization-induced changes in CH4 fluxes and plant productivity. We observed dynamic CH4 flux-driver relationships across the salinization disturbance event, where temperature connections strengthened, and productivity connections dampened during salinization. Annual gross ecosystem productivity reduced by 64% during peak salinization, whereas annual CH4 emissions only reduced by 10%, suggesting that other CH4 substrate sources compensated for reductions in recent photosynthate. Our results demonstrate the value of applying information theory and machine learning approaches to ecological analyses and suggest that drought-induced salinization may increase GHG emissions from estuarine freshwater wetlands.
