Spatial and temporal heterogeneity of geochemical controls on carbon cycling in a tidal salt marsh

Abstract

Tidal salt marsh ecosystems store copious amounts of carbon (C) within sediments. In order to predict how these C stores may be affected by environmental change, it is critical to assess current CO2 and CH4 production and efflux from these ecosystems. Production and efflux of these greenhouse gases (GHGs) are governed by coupled geochemical, hydrological, physical and biological processes in sediments that are sensitive to local conditions, which can result in large spatial and temporal heterogeneity of GHGs dynamics within the ecosystem. To understand how the drivers of GHGs dynamics vary across salt marsh ecosystems, we coupled solid-phase geochemistry to measurements of porewater chemistry (to ∼1 m), CO2 and CH4 production in sediments and efflux to the atmosphere in a temperate tidal salt marsh for over one year to capture seasonal patterns within two vegetation zones of the marsh landscape that have distinct biogeochemical and hydrologic conditions: Tall Spartina (TS) and Short Spartina (SS). The SS vegetation zone experienced nearly constant inundation, low redox values (−200 to 200 mV), porewater pH 6–7 that did not vary with depth or time, an enrichment of pyrite and goethite with depth and up to 3 mM porewater sulfide. In contrast, the TS vegetation zone on the natural levee proximal to a tidal channel experienced large water level oscillations due to spring-neap tides that resulted in variable but higher redox values (0–700 mV), porewater pH 6–7 at depth but surface (0–3 cm) as low as 4 in the spring, an enrichment of ferrihydrite and a depletion of pyrite at ∼30 cm, and up to 0.8 mM ferrous Fe in porewater. At 50–56 cm, solid phase analyses (STXM-NEXAFS) revealed differential C speciation between the two vegetation zones, with stronger C-Fe spatial association at TS and stronger C-Ca co-association at SS despite both having similar soil pH of 3–4. These results suggest that soil pH may not be strongly predictive of C-mineral control in flooded marsh sediments. Both vegetation zones showed consistent CO2 and CH4 emissions from sediments to the atmosphere throughout the study period with TS having ∼60% higher median CO2 and SS having ∼55% higher median CH4 efflux. Using depth profiling, unexpectedly high concentrations of CO2 (>200 μM) and CH4 (>200 μM) were observed at depths 50–75 cm at both zones that were higher for SS in these sulfate-rich (up to 17 mM) sediments, which suggests methylotrophic methanogenesis occurs deep within the profile of salt marsh sediments away from the tidal channel. Moreover, if we extrapolate our median depth values of CH4 and CO2 to the 5.3 Mha of global salt marshes, this could account for a conservative estimate of ∼70 Gg of unaccounted C stored in gaseous form (i.e., CH4 and CO2) in marsh sediments, which should be considered when attempting to understand the current patterns and future responses of carbon dynamics from these ecosystems.