The relationship between El Niño and the Indian monsoon has been a cornerstone of seasonal climate prediction since the 1880s. El Niño suppresses monsoon rainfall; La Niña enhances it. But starting in the 1980s, the correlation weakened. Decades of papers debated whether the physical connection was changing, whether global warming had disrupted the teleconnection, or whether the historical correlation was a statistical artifact.
Sharma et al. (arXiv:2603.16346) argue the connection never changed. The measurement did.
Traditional ENSO indices track Pacific sea surface temperatures — Niño 3.4, SOI, ONI. These capture ENSO's Pacific signature but miss its expression in other ocean basins. When El Niño triggers Indian Ocean warming and Atlantic cooling, those remote responses also affect the monsoon, but they're invisible to Pacific-only indices. As ocean basins warm at different rates under climate change, the Pacific-only proxy increasingly misrepresents ENSO's total influence.
The authors construct a Global-ENSO index using subsurface temperature data across all three tropical oceans. Against this index, monsoon rainfall shows a correlation of 0.8 at an 18-month lead — strong, stable, and unwavering across the entire record. The paradox dissolves: the teleconnection was always there, but the thermometer was reading one ocean when three were speaking.
The resolution is a measurement lesson. A proxy that worked in a stationary climate fails in a non-stationary one because it captures one channel of a multi-channel signal. The physical relationship is preserved in the full data; the apparent instability lives entirely in the projection onto an incomplete basis. What looked like climate change disrupting a fundamental coupling was actually climate change disrupting a measurement convention.