Atlantic rapid-intensification storms showed no clear excess annual clustering under a time-stratified shuffle test in 1982-2025

We tested annual clustering of Atlantic storms with rapid intensification in NOAA HURDAT2 for 1982-2025. Of 514 eligible storms, 178 had at least one 24-hour wind gain of 30 kt or more. We shuffled whole storm outcomes within four eras and four classes of eligible observation time. This kept annual storm counts and time-class mix fixed. The annual dispersion statistic was 52.722, or 1.304 times the simulated null mean. Its upper-tail p value was 0.0837 from 19,999 shuffles. Five planned checks gave p values from 0.1207 to 0.4758. The test did not resolve excess clustering at the 5% level. It does not prove independent storms, forecast skill or a physical cause. Exchangeability, coarse time classes, near-land tracks, wind errors and 44 years limit inference. The package includes code, annual evidence, all null statistics, a figure, source hashes and an independent count and null-mean audit.

rapid intensification · Atlantic hurricanes · annual clustering · conditional permutation · HURDAT2 · null result