Warm and cold monthly SST persistence differences were uncertain in three ocean regions in 1982-2025

We tested whether warm sea surface temperature anomalies have more next-month persistence than cold anomalies in three fixed ocean regions in 1982-2025. We used NOAA ERSSTv5 and OISST v2.1 monthly means, removed a seasonal cycle and linear trend, and fitted a common warm-minus-cold slope contrast with season-specific baseline slopes. The contrasts were -0.024 and -0.031 in the Northeast Pacific, -0.100 and -0.087 in the subpolar North Atlantic, and 0.104 and 0.026 in the Tasman Sea, for ERSST and OISST respectively. All six 95% intervals from 24-month block resampling included zero. Six-test adjusted intervals and 12- and 36-month block intervals also included zero. Excluding possible sea-ice cells had little effect in the Atlantic. Trend and extreme-value checks changed some estimates. The test does not establish a persistence difference or equal persistence. It measures conditional monthly dependence, not daily heatwave duration or a physical cause. Short records, shared source observations, native-grid masks and uncertainty conditional on preprocessing limit inference. Code, spatial means, source hashes, bootstrap evidence, audit checks and a figure are included.

sea surface temperature · persistence · warm cold asymmetry · block bootstrap · null result · ERSSTv5 · OISST