The shift toward large gains over losses was uncertain for offshore Atlantic hurricane starts in 1982-2025

We compared large 24-hour wind gains and losses for offshore Atlantic hurricane starts in 1982-2003 and 2004-2025. We used a fixed NOAA IBTrACS v04r01 file. Each window had five tropical fixes at least 150 km from mapped land and at 0-40 N. Initial wind was at least 65 kt. Gains and losses of at least 30 kt occurred in 66 and 42 of 1,239 early windows and in 137 and 92 of 1,403 late windows. After adjustment for initial wind, latitude and season, the gain rate change was 4.71 (1.51 to 8.26) percentage points. The loss rate change was 2.56 (-0.58 to 5.32) points. The change in gain rate minus loss rate was 2.15 (-2.23 to 6.99) points. Parentheses give conditional 95% season-bootstrap intervals. Thus, the primary test did not resolve a shift toward gains over losses. Both unadjusted tails rose; the adjusted loss change was uncertain. The result does not identify an ocean or climate cause. Sample cuts, coarse adjustment, changing wind estimates and only 44 seasons limit inference. The package contains fixed source rows, window evidence, code, draws, an independent audit and a figure.

North Atlantic · tropical cyclone · rapid intensification · rapid weakening · wind change · paired tails · season bootstrap · IBTrACS