The oxygen-18 isotope approach for measuring aquatic metabolism in high-productivity waters

Abstract

We examined the utility of δ 18 O 2 measurements in estimating gross primary production (P), community respiration (R), and net metabolism (P : R) through diel cycles in a productive agricultural stream located in the midwestern U.S.A. Large diel swings in O 2 (±200 µmol L −1 ) were accompanied by large diel variation in δ 18 O 2 (±10‰). Simultaneous gas transfer measurements and laboratory‐derived isotopic fractionation factors for O 2 during respiration (α r ) were used in conjunction with the diel monitoring of O 2 and δ 18 O 2 to calculate P, R, and P :R using three independent isotope‐based methods. These estimates were compared to each other and against the traditional “open‐channel diel O 2 ‐change” technique that lacked δ 18 O 2 . A principal advantage of the δ 18 O 2 measurements was quantification of diel variation in R, which increased by up to 30% during the day, and the diel pattern in R was variable and not necessarily predictable from assumed temperature effects on R. The P, R, and P :R estimates calculated using the isotope‐based approaches showed high sensitivity to the assumed system fractionation factor (α r ). The optimum modeled ar values (0.986‐0.989) were roughly consistent with the laboratory‐derived values, but larger (i.e., less fractionation) than α r values typically reported for enzyme‐limited respiration in open water environments. Because of large diel variation in O 2 , P :R could not be estimated by directly applying the typical steady‐state solution to the O 2 and 18 O‐O 2 mass balance equations in the absence of gas transfer data. Instead, our results indicate that a modified steady‐state solution (the daily mean value approach) could be used with time‐averaged O 2 and δ 18 O 2 measurements to calculate P :R independent of gas transfer. This approach was applicable under specifically defined, net heterotrophic conditions. The diel cycle of increasing daytime R and decreasing nighttime R was only partially explained by temperature variation, but could be consistent with the diel production/consumption of labile dissolved organic carbon from photosynthesis.