Matched CalCOFI samples did not resolve a change in the Q3 minus Q1 oxygen contrast at 200 m in 1984-2020

We tested whether oxygen loss changed the Q3 minus Q1 contrast at 200 m off southern California. We reused observed CalCOFI bottle profiles and paired Q1 and Q3 within each station-year. We retained 56 stations and 1,832 paired station-years in 1984-2020. The contrast changed from -0.132 mL/L in 1984-2002 to -0.142 mL/L in 2003-2020. The change was -0.010 mL/L, with a conditional 95% three-year block interval of -0.073 to +0.055. The test did not resolve a seasonal contrast change. All eight sensitivity intervals also included zero. Solubility and oxygen-deficit contrasts did not resolve a change. This does not prove constant seasonality or rule out local changes. Dates within quarters shifted, and missing years, fixed-depth sampling, measurement changes and short temporal blocks limit inference. The package includes reused observed bottles and profiles, source hashes, new paired evidence, code, bootstrap draws, a figure and numerical checks.

CalCOFI · oxygen · seasonal contrast · paired observations · null result · ocean deoxygenation