Short-term variation of triple oxygen isotopes and gross oxygen production in the Sagami Bay, central Japan

We studied diurnal and daily variations in triple oxygen isotopes in the euphotic zone of the Sagami Bay during summer 2003 and 2004. There were sharp changes in physical processes, driven by changes in wind speeds, during the study period. Mixed-layer 17D anomalies were negatively correlated with wind speeds, whereas dissolved oxygen (DO) and chlorophyll a (Chl a) were positively correlated, suggesting that both physical and biological processes influence the 17D anomaly. Vertical mixing affected estimation of gross oxygen production (GOP) in the mixed layer, whereas horizontal advection was negligible in the Sagami Bay during the study period. Hourly GOP estimates are consistent with fast repetition rate fluorometer (FRRF) measurements; relatively higher estimates from the former method were due to the storing of past productivity signal. Daily integrated production estimated from 17D anomaly and FRRF were consistent during the entire study period, whereas the estimation from a 18O spike incubation technique showed good agreement with the other two techniques only when the water mass structure was unchanged in the mixed layer. Understanding the distribution of bioactive tracers in the sea is a fundamental challenge in oceanography. Gross oxygen production (GOP) is a fundamental but poorly known characteristic of the planktonic marine ecosystem. Most of our knowledge of primary production is based on data collected by the 14C method (Steeman Nielsen 1952) and other techniques involving incubations in vitro. The 14C method has significant problems associated with it, such as bottle effect (i.e., influence of lack of mixing on growth of bacteria and phytoplankton); loss of labeled 14C as CO2 and DOC; and assimilation of unlabeled CO2 (e.g., Fitzwater et al. 1982; Bender et al. 1987; Howarth and Michaels 2000). Data derived using the 14C method may seriously underestimate oceanic production. The light-dark (LD) bottle incubation method is free from tracer loss, as it measures changes in dissolved oxygen (DO) concentrations, but it is not free of bottle effect (Howarth and Michaels 2000). However, this method assumes equal respiration rates in dark and light. Several studies showed that respiration at light is higher than dark by a factor of 2 : 10 (Bender et al. 1987; Grande et al. 1989, 1991). The 18O incubation method, which is also susceptible to the bottle effect, measures O2 released at the photosystem II (PSII) and therefore directly reflects the GOP. Several investigators (Bender et al. 1987; Kiddon et al. 1995; Bender et al. 1999, 2000) found higher production by 18O incubation than 14C by a factor of two. On the contrary, recently Ostrom et al. (2005) found that primary production by 18O incubation is significantly lower than 14C and LD bottle incubations in the eutrophic and mesotrophic environment and attributed due to consumption of labeled O2 within cells or evolution of extraneous O2. In order to alleviate these problems, Luz and Barkan (2000) applied variations in ratios of all three stable O2 isotopes to estimate changes in global GOP. This method requires no incubation, and average mixed-layer GOP over residence time of oxygen in the mixed layer can be estimated. Luz and Barkan (2000) estimated GOP at the Hawaii Ocean Time-series Station (HOTS) and the Bermuda Atlantic Time-series Station (BATS) using triple oxygen isotopes, and it was comparable with the same measured using the 1 Corresponding author ([email protected]). Acknowledgments We thank the officers and crews of RV Tanse Maru for their help during sampling. We would like to thank Dr. Boaz Luz and Dr. Nathaniel Ostrom for their constructive comments to improve the previous version of this manuscript. This work was funded by the Japan Science and Technology Agency. Limnol. Oceanogr., 51(3), 2006, 1432–1442 E 2006, by the American Society of Limnology and Oceanography, Inc.