Enigmatic marine ecosystem metabolism measured by direct diel ΣCO2 and O2 flux in conjunction with DOC release and uptake

In an attempt to evaluate CO₂ changes as an index of net ecosystem metabolism, CO₂ and dissolved organic carbon (DOC) by infrared (IR) analysis and O₂ by the Winkler method were followed over 12 diel cycles in a salt marsh, a simulated estuarine ecosystem, and the mixed layer of the northwestern Caribbean Sea. Each ecosystem exhibited replicable diel cycles, net production during the photoperiod, and a significant inverse correlation between CO₂ and O₂ changes. Daily rates of system production and respiration calculated from CO₂, however, exceeded those from O₂ by 1.5 to 3.5 fold in nearshore waters and by 2 to 6 fold in the open ocean. Net total system apparent production based on O₂ and CO₂, respectively, were 2 345 and 3 604 mg C m^-3d^-1 for the salt marsh, 348 and 625 mg C m^-3d^-1 for the estuarine ecosystem, and 53 and 306 mg C m^-3d^-1 for the oceanic ecosystem, both parameters exceeding ¹⁴C data in the literature for similar environments by one to two orders of magnitude. The IR CO₂ productivity estimates are compatible with the diel cycles in DOC. In the marsh and Caribbean Sea, maxinium DOC concentrations were usually observed in the evening following a gradual accumulation during the photoperiod, while minimal values occurred in the early morning. In all ecosystems the average net release of DOC lagged CO₂ uptake and O₂ production and represented 22.7 and 43.3% of the carbon fixed as estimated from CO₂ uptake and O₂ production, respectively. If the consistently higher CO₂ measurements are not a systematic error nor due to atmospheric diffusion, then diel variation in O₂ and CO₂ may not always be quantitatively coupled due in part to habitat-dependent factors such as the nonphotosynthetic incorporation of CO₂ and chemosynthetic removal of O₂.