Upper-range flood outlook misses had shared regional timing at 69 US gauges in 2019-2025 under two randomization tests

We tested whether observed flood-day counts exceeded NOAA annual outlook upper ranges in the same years across gauges. We reused a published stable-threshold sample and kept 69 gauges with complete 2019-2025 records in seven regions. Upper-range misses occurred in 23.4% of 483 gauge-years. The mean regional variance of yearly miss fractions was 0.04629. This was 3.40 times the mean from independent circular shifts of each gauge series. No draw in 19,999 reached the observed value. A second test preserved all gauge series and network year totals but shuffled regional membership. Its mean variance was 0.02796, with a conditional Monte Carlo tail fraction of 0.00045. Thus, the alignment included regional structure beyond the network-wide year pattern under these references. Widening each range half-width by factors of 1.5 and 2 left upper-miss variance ratios of 3.04 and 2.88 against circular shifts. This adds a binary upper-miss diagnostic to prior evidence on signed regional errors; it is not independent confirmation. Seven years, selected gauges, reused observation checks and randomization assumptions limit inference. The tests do not identify a physical cause or estimate future joint flood risk. Code, aggregate counts, random draws, source hashes, a figure and independent arithmetic checks are included.

high tide flooding · forecast verification · regional dependence · NOAA · randomization test · upper-range misses