Upper midpoint tails had a larger two-week recurrence contrast at Seattle under two baselines in 1960-2025

We compared upper and lower daily water-level midpoint tails at six US gauges in 1960-2025. The recurrence contrast is the repeat probability at 13-16 days minus that at 6-9 days. After year-month median removal, the upper-minus-lower contrast was 23.44 percentage points at Seattle and 10.26 at San Francisco. Conditional intervals adjusted for six sites were 20.03 to 26.39 and 6.11 to 14.77 points. An annual and seasonal baseline gave 18.74 points at Seattle, with an interval of 14.88 to 22.36. It left the San Francisco difference uncertain. The other four primary site intervals included zero. The Seattle difference stayed positive at 90th and 99th percentile thresholds and with one- and five-year blocks. This extends campaign work on upper tails and next-day persistence. The daily midpoint retains tide variation and is not measured surge. Baseline choice, selected gauges, fixed thresholds, overlapping prior data, and dependence assumptions limit inference.

water level · tail asymmetry · recurrence · daily midpoint · NOAA CO-OPS · tide gauge