5.4.3 Oxygen

In the thermocline (~100 to 1,000 m), a decrease in the O₂ concentration has been observed between about the early 1970s and the late 1990s or later in several repeated hydrographic sections in the North and South Pacific, North Atlantic, and Southern Indian Oceans (Figure 5.12; see summary table in Emerson et al., 2004, and Section 5.3). Section 5.3 reports on a number of O₂ decreases that fit the overall message of Section 5.4. The reported O₂ decreases range from 0.1 to 6 μmol kg^–1 yr^–1, superposed on decadal variations of ±2 μmol kg^–1 yr^–1 (Ono et al., 2001; Andreev and Watanabe, 2002). In all published studies, the observed O₂ decrease appeared to be driven primarily by changes in ocean circulation, and less by changes in the rate of O₂ demand from downward settling of organic matter. A few studies have quantified the contribution of the change in ocean circulation using estimates of changes in apparent CFC ages (Doney et al., 1998; Watanabe et al., 2001; Mecking et al., 2006). In nearly all cases, the decrease in O₂ could entirely be accounted for by the increased apparent CFC age that resulted from reduced rate of renewal of intermediate waters. Changes in biological processes were only significant at the coast of California and may result from assumptions in the method (Mecking et al., 2006).

Figure 5.12. Changes in oxygen concentration (μmol kg^–1) along two sections in the North Pacific (see map, bottom panel). Top left panel: Difference (1999 minus 1985) along 47°N. Top right panel: Difference (1997 minus 1984) at 152°W. Blue colours indicate a decrease and yellow colours indicate an increase in oxygen over time. The differences were calculated using density as the vertical coordinate. After Deutsch et al. (2005).

It is unclear whether the recent changes in O₂ are indicative of trends or of variability. Recent data in the Indian Ocean have shown a reversal of the O₂ decrease between 1987 and 2002 in the South Indian Ocean of similar amplitude to the decrease observed during the previous decades (McDonagh et al., 2005). Variability has been observed on decadal time scales in the North Atlantic large enough to mask any potential trends (Johnson and Gruber, 2007).

In the upper 100 m of the global ocean surface, decadal variations of ±0.5 μmol kg^–1 in O₂ concentration were observed for the period 1956 to 1998 based on a global analysis of 530,000 oxygen profiles, with no clear trends (Garcia et al., 2005). However, the near-surface changes in O₂ concentration are difficult to interpret. They can be caused by changes in biological activity, by changes in the physical transport of O₂ from intermediate waters or by changes in temperature and salinity. Because there is less confidence in the early measurements and the reported changes cannot be explained by known processes, it cannot be said whether the absence of a long-term trend in surface O₂ is realistic or not.