Document Type


Date of Original Version





Preliminary studies with an ampule analyzing unit and infrared (IR) detector showed that procedures for standardization and determination of total carbon dioxide (ΣCO2), while often precise, lacked the accuracy required to estimate the net productivity and respiration of aquatic ecosystems during studies in which sampling over diel cycles was used. Scaling down sample and standard volumes to the ~l range and the use of a commercial sodium carbonate standard without dilution before and after replicate sample injections gave accurate results as shown by comparison with indirect (pH-alkalinity) ΣCO2 determinations with a standard error of ±3 μmoles in the laboratory and ±6 μmoles at sea for 8 to 10 replicates. This was sufficient to detect a diurnal consumption and nocturnal production of CO2 which were inversely correlated with O2 variation in a salt marsh, an estuarine mesocosm, and the Caribbean Sea.