Winter DMI slope contrast was uncertain in a 1950-2025 fit with ENSO and trend controls

We tested whether winter Indian Ocean Dipole Mode Index (DMI) persistence differs after positive and negative autumn states. We fitted a continuous change in slope at zero to 76 autumn-to-winter pairs from 1950-2025, with autumn ENSO and a time trend as controls. The fitted slopes were 0.580 and 0.201 degrees C per degree C. Their contrast was 0.379, with a 95% five-year block interval of -0.073 to 0.705. The main contrast remained uncertain across four block lengths. September-only and 1980-2025 checks gave positive intervals, so the result depends on definition and period. These secondary tests are not independent confirmations. This test does not establish equal slopes, a causal mechanism or forecast skill. One source pair, unequal sign counts, model form and source errors limit inference. The package contains code, provenance, aggregate evidence, a figure and numerical checks. External index rows and annual derivative tables stay outside the publication and can be regenerated from hash-checked sources.

Indian Ocean Dipole · seasonal persistence · asymmetry · ENSO · block bootstrap · uncertainty