Cross-Comparisons of Sediment Incubation Methods to Bound Stochastic Influences on Denitrification of Natural Waters from Mississippi River Floodplain Wetlands

Flood regimes in large river systems such, as the Mississippi River, are inherently stochastic meaning that floodplain wetlands experience varying hydrostatic pressures of groundwater upwelling (exfiltration) and infiltration from overland flooding. Distinctions in water delivery can drastically alter the oxygen levels and groundwater delivery into wetland sediments where anaerobic microbes remove nitrates through denitrification. Our findings bound conditions for modeling denitrification rates across oxic flood waters versus exfiltration by anoxic groundwaters. Four-by-four factorial laboratory incubation treatments included oxic versus anoxic waters (degassed with helium) introduced by exfiltration or infiltration (Figs. 1A, 1B).

Sediments collected in triplicate from four floodplain wetlands located along Dogtooth Bend segment of the Mississippi River near the Ohio River confluence were incubated with river water. Sediments were incubated at 29 ^oC for 96 h. Nitrogen gas production was measured by membrane inlet mass spectrometry (MIMS). Inflow and outflow waters were analyzed for nitrate, ammonia, phosphate, and dissolved organic carbon while sediments were characterized for their physical traits. In contrast to most studies, that estimate denitrification relative to surface area in incubations only (i.e. address conditions of surface flooding), we also present our findings relative to sediment volumes (i.e. evaluate denitrification rates from exfiltration of groundwater). Regression analyses compared denitrification from surface area versus volume calculations (R² values > 0.89); consequently providing an excellent tool for converting estimates from surface area alone to varying sediment saturation for more rigorous assessments of subsurface interactions.

Average denitrification rates relative to sediment volume were significantly higher in anoxic-deep-injection cores (“AD cores”; 23.83 + 1.94 N mg/m³/d) compared to anoxic-surface-delivery cores (“AS cores”; 19.98 + 1 N mg/m³/d), that also exceeded oxic-deep-injection cores (“OD cores"; 14.96 + 1.78 N mg/m³/d) and oxic-surface-delivery cores (“OS cores”; 10.23 + 1.04 N mg/m³/d). Thus, average denitrification followed an anoxic-injection hierarchy of AD > AS > OD > OS (p-values < 0.003), which was maintained for denitrification relative to surface area. Regarding site-specific distinctions, for sandy sites this hierarchy persisted, and each treatment differed significantly. Sites with clayey sediments had very-low permeability regardless of injection type; thus only oxic versus anoxic treatments differed significantly irrespective of water delivery. By contrast sites with loamy sediments, injection type significantly influenced denitrification rates while neither oxic nor anoxic water treatments differed. An AICc model showed that phosphate, ammonia, temperature variation, dissolved oxygen, and sand content explained 33% (p-value < 0.05) of the variation in denitrification rates across all cores and treatments.

Our findings highlight the greater insights provided from cross-comparison incubation designs to better inform landscape-scale models of denitrification rates across floodplain wetlands depending on the magnitude and duration of flooding.