Daily means missed 46–48% of summer oxygen flags in 108 USGS records at a 5 mg/L limit, 2000–2009 and 2016–2025

We tested whether daily mean oxygen detects days with low daily minimum oxygen. We used approved USGS records for June–August. The fixed panel has 108 sites and 161,165 matched days in 2000–2009 and 2016–2025. All three daily statistics came from the same underlying sensor series on each day. At a 5 mg/L limit, the mean missed 48.4% and 45.6% of minimum flags in the two periods. These shares use equal weights for months, years and sites. The missed-day fractions were 10.95% and 8.43%. Their change was −2.52 percentage points. A nominal 95% interval from spatial and shared-year blocks was −5.58 to +0.88 points. Thus, the change remains uncertain. Checks with stricter coverage and lower limits gave the same uncertainty. An audit found that an earlier SCI field named DOMIN used USGS code 00001, which is maximum. Code 00002 is minimum. The audit removes that field as evidence about minima. It does not retest the earlier mean-oxygen results. The panel is selected and uneven. A minimum flag does not measure time below the limit or harm to species. The package has a corrected source extract, code, provenance, all results, figures and independent checks.

dissolved oxygen · daily minimum · river monitoring · USGS · data audit · low oxygen · matched records · reproducible dataset