Eco-hydrological processes in controlling vertical and lateral carbon fluxes in a subtropical salt marsh: Insights from modeling and scale-resolved analysis

Authors:

Abstract
Carbon exchange in salt marshes is shaped by complex interactions among vegetation, climate, and tidal regimes. Here, we explored the spatiotemporal heterogeneity and environmental driving of the autochthons carbon cycle in a subtropical salt marsh using a process-based model. Integrating species-specific phenology, hydrodynamics, and carbon biogeochemistry, the model captures well the seasonal and tidal patterns of gross primary productivity (GPP), ecosystem respiration (ER), methane flux (FCH4), dissolved organic (FDOC), inorganic carbon fluxes (FDIC), and biomass carbon, as well as net ecosystem carbon balance (NECB) in a salt marsh of Yangtze Estuary. Results show the salt marsh functions as net carbon sinks during spring and summer but shifted to sources in autumn and winter due to intensified lateral carbon losses; high marsh zones exhibited greater gaseous carbon import, while low-elevation zones were characterized by elevated dissolved carbon export. SHapley Additive exPlanations analysis revealed the dominance of temperature and tides in regulating NEE and FDOC, respectively, and FCH4 and FDIC exhibited dual dependence on both thermodynamic and hydrodynamic drivers. Wavelet power spectra displayed a divergence in spectral characteristics between vertical and lateral carbon fluxes: NEE and FCH4 displayed smooth energy distributions that aligned closely with temperature, whereas FDOC and FDIC exhibited more variable energy profiles synchronized with tide. Wavelet coherence analyses further illustrated scale-dependent controls, with vertical and lateral carbon fluxes synchronized with diurnal and tidal cycles, respectively. These findings highlight the importance of high-resolution mechanistic studies to enhance carbon accounting and predict long-term sequestration potential in coastal salt marshes.

This publication has 34 references indexed in Scilit: