AbstractDiel variations in oxygen concentration have been extensively used to estimate rates of photosynthesis and respiration in productive freshwater and marine ecosystems. Recent improvements in optical oxygen sensors now enable us to use the same approach to estimate metabolic rates in the oligotrophic waters that cover most of the global ocean and for measurements collected by autonomous underwater vehicles. By building on previous methods, we propose a procedure to estimate photosynthesis and respiration from vertically resolved diel measurements of oxygen concentration. This procedure involves isolating the oxygen variation due to biological processes from the variation due to physical processes, and calculating metabolic rates from biogenic oxygen changes using linear least squares analysis. We tested our method on underwater glider observations from the surface layer of the North Pacific Subtropical Gyre where we estimated rates of gross oxygen production and community respiration both averaging 1.0 mmol O2 m−3 d−1, consistent with previous estimates from the same environment. Method uncertainty was computed as the standard deviation of the fitted parameters and averaged 0.6 and 0.5 mmol O2 m−3 d−1 for oxygen production and respiration, respectively. The variability of metabolic rates was larger than this uncertainty and we were able to discern covariation in the biological production and consumption of oxygen. The proposed method resolved variability on time scales of approximately 1 week. This resolution can be improved in several ways including by measuring turbulent mixing, increasing the number of measurements in the surface ocean, and adopting a Lagrangian approach during data collection.