A tri‑basin matching index improves Indian Summer Monsoon Rainfall predictability

Interannual variability of Indian Summer Monsoon Rainfall (ISMR) is vital for South Asian water resources and agriculture, but its prediction is challenged by non-stationary ocean–monsoon interactions. This study proposes a Matching Index (MI), defined as the pointwise product of normalized ISMR and sea surface temperature anomalies, combined with Empirical Orthogonal Function analysis (MI–EOF) to identify state-dependent ocean–monsoon coupling regimes. MI–EOF reveals three dominant regimes: ENSO intensity, distinguishing strong and weak ENSO states modulated by the Indian Ocean Dipole and tropical South Atlantic; ENSO lifecycle, highlighting phase-dependent effects of Indian Ocean Basin warming on ISMR; and ENSO spatial patterns, separating central- and eastern-Pacific ENSO influences on monsoon variability. Based on these physically interpretable regimes, a scenario-dependent reconstruction framework is developed to adaptively select predictors. The framework improves ISMR reconstruction correlation from 0.73 using conventional multivariate linear regression to 0.83, with stable performance throughout the record. Results demonstrate that ISMR variability is fundamentally state dependent and cannot be described by stationary linear relationships. The MI framework provides a physically based approach for diagnosing evolving air–sea coupling and improving monsoon prediction under climate change.