Matched CalCOFI oxygen anomalies at 200 m showed no clear synchrony change between 1984-2002 and 2003-2020

We tested whether interannual oxygen anomalies became more synchronized across southern CalCOFI stations at 200 m. We rebuilt a fixed bottle archive and used 5,822 annual station-quarter values from 6,420 casts at 56 stations. Each station pair used common years, with separate linear trends removed in each period. We averaged 4,234 pair correlations within quarters, then gave the four quarters equal weight. Mean correlation was 0.115 in 1984-2002 and 0.120 in 2003-2020. The change was +0.0056, with a conditional 95% three-year block interval of -0.0863 to +0.0881. Whole-year draws kept all stations and quarters together. All feasible planned check intervals included zero. The 300 m check lacked eligible April-June pairs. Thus, the test did not resolve a change in shared oxygen variation. It does not show constant synchrony, stable hypoxia risk or a physical cause. Sparse seasonal coverage, quarterly sampling, short periods, method changes and resampling assumptions limit the result. The package gives an attributed annual panel, source hashes, new code, all accepted weights and estimates, a figure and independent numerical audits. Raw data remain external.

CalCOFI · dissolved oxygen · spatial synchrony · partial correlation · matched sampling · block bootstrap · null result · ocean conditions