Low-water tails persisted more than high-water tails at three US gauges in retained 1960-2025 days with fixed thresholds

We compared next-day persistence of high daily maxima and low daily minima at six US tide gauges in 1960-2025. We removed observed monthly mean sea level and used fixed site and calendar-month 95th and 5th percentile thresholds. Quality rules retained 137,884 station-days. High-minus-low persistence was -9.08 percentage points at Boston, -8.09 at San Francisco, and -14.67 at Seattle. Conditional intervals adjusted for six gauges were -13.55 to -4.52, -11.54 to -4.47, and -19.61 to -8.78 points. The other three intervals included zero. A paired midpoint comparison changed the contrast in different directions across sites. One- and five-year blocks retained the three negative extrema intervals. These intervals omit threshold estimation, source changes, and missing-data uncertainty. The result depends on selected gauges, daily summary, snapshot, and fixed thresholds. Daily extrema retain tides and cannot identify storm surge, flood duration, physical recovery, or a causal process. Code, input evidence, source hashes, count audits, draws, and repeat instructions are supplied.

water level · daily extrema · tail persistence · NOAA CO-OPS · fixed thresholds