The Indian Summer Monsoon Rainfall (ISMR) is crucial for the economy and food security of South Asia, supporting a significant portion of the global population. Despite advancements in observation and modeling, accurate ISMR prediction remains a challenge. While ENSO has been a primary predictor, its correlation with ISMR has weakened in recent decades, indicating a need to explore other influencing factors. Existing research suggests various potential predictors such as Eurasian snow cover, Pacific Decadal Oscillation (PDO), Indian Ocean Dipole Mode (IOD), Atlantic Niño, and Atlantic tripole, but their mechanisms of influence and causal relationships with ISMR remain debated. Evidence suggests an association between cold NA SSTs and ISMR droughts, and NA SST's link with ISMR on multi-decadal timescales. The study hypothesizes a teleconnection between NA-SST and ISMR and aims to establish the causality between them using advanced nonlinear causal inference techniques, accounting for other potential drivers and their interdependencies.

The literature review extensively covers the history of ISMR prediction, highlighting the long-standing reliance on ENSO despite its weakening relationship with ISMR in recent decades. It summarizes existing research on alternative predictors such as Eurasian snow cover, PDO, IOD, Atlantic Niño, and Atlantic tripole, emphasizing the ongoing debate regarding their physical mechanisms and the fraction of ISMR variability they explain. The review also details the existing evidence for a connection between NA SST and ISMR, including associations with past mega-droughts and multi-decadal timescale relationships. The lack of quantitative nonlinear causal inference in establishing the link between NA-SST and ISMR is highlighted as a key gap in the existing literature, leading to the study's focus on advanced nonlinear causal inference techniques.

The study uses multiple datasets, including the Parthasarathy rainfall data for ISMR, COBE SST2 data for SST fields, NCEP 20CR V3 for atmospheric fields, and ERA-5 and NOAA OI SST weekly data for robustness checks. Multi-decadal modes (MDMs) were extracted using an improved variant of Complete Ensemble Empirical Mode Decomposition (ICEEMDAN). The core methodology involves two nonlinear causal inference techniques: PCMCI+ and Granger causality. PCMCI+ was used to quantify causal relationships between ISMR and potential drivers (ENSO, AMO, PDO, NAO, IOD, Atlantic Niño) considering conditional independence and nonlinearity, examining both contemporaneous and lagged relationships (up to 5 months) within the same summer season. Granger causality, with both linear and nonlinear versions, was employed to test the robustness of the PCMCI+ findings. For intraseasonal analysis, daily data filtered with a 7-day moving average were used with PCMCI+ to investigate the hypothesized teleconnection pathway between NA-SST, NAO, barotropic vorticity (BV), upper-level vorticity (IUV), lower-level vorticity (ILV), and ISMR. Various indices, such as AMO, NAO, Nino3.4, PDO, At-Nino, and IOD, were defined based on box-averaged SST and SLP anomalies. The authors also analyzed CMIP6 model simulations to evaluate model fidelity in simulating the teleconnection between ISMR and NA-SST.

The key findings demonstrate that NA-SST and ENSO are independent drivers of ISMR with comparable influence. PCMCI+ analysis revealed that ENSO negatively influences ISMR, while AMO (represented by NA-SST) positively influences ISMR. The ENSO-ISMR relationship was found to be bidirectional, while the AMO-ISMR relationship was unidirectional, with AMO driving ISMR. The PDO showed no direct link with ISMR, and the NAO's influence on ISMR was mediated through AMO. The At-Nino did not directly impact ISMR, and its relationship was linked to the AMO. The IOD was found to have a non-directional association with ISMR. Intraseasonal analysis using PCMCI+ supported the hypothesized teleconnection, with NA-SST driving BV, which influences IUV, ultimately affecting CI-ISMR. Granger causality analysis corroborated these findings, although with slightly different lag structures, confirming the independent roles of AMO and ENSO. Analysis of various SST datasets and multiple causal inference methods confirmed the robustness of the key results. Examining the linear correlation between JJAS NA SST and ISMR for years when ENSO was near neutral revealed a significant correlation, indicating NA-SST can explain up to 20% of ISMR variability. CMIP6 model analysis found that some models simulated the AMO as an independent driver of ISMR, but there were model-dependent biases in simulating the strength of the ENSO-ISMR relationship. Composites of SST anomalies during non-ENSO drought and flood years showed consistent patterns, highlighting the complementary roles of NA SST and ENSO in influencing ISMR.

The study's findings challenge the long-held assumption of ENSO as the primary driver of ISMR variability, revealing the significant and independent contribution of NA-SST. The identification of a robust, unidirectional causal link from NA-SST to ISMR, operating through an atmospheric bridge, expands the understanding of ISMR predictability. The comparable strength of influence of NA-SST and ENSO on ISMR highlights the need to incorporate extra-tropical SSTs into ISMR prediction frameworks. The intraseasonal analysis reveals a complex teleconnection pathway involving the NAO and vorticity fields, providing mechanistic insight into the NA-SST's influence on ISMR. The results suggest that the potential skill of ISMR prediction could be significantly enhanced by considering NA-SST, especially during ENSO transition periods. The study's implications extend beyond ISMR, suggesting a broader influence of extra-tropical SST on tropical climate.

This research conclusively demonstrates the significant and independent role of North Atlantic SST in influencing Indian Summer Monsoon Rainfall, alongside ENSO. The findings advocate for a reassessment of the prevailing causal framework, highlighting the importance of extra-tropical factors in ISMR predictability. Future research should focus on improving climate model simulations of extra-tropical-tropical teleconnections to enhance the accuracy of ISMR seasonal forecasts. The study strongly suggests moving beyond the legacy of TOGA to incorporate the influence of extra-tropical SST in the conceptual understanding of tropical climate predictability.

The study acknowledges potential limitations associated with the use of historical datasets, including data sparsity and uncertainties in the early observational records. While multiple SST datasets were used to improve robustness, some sensitivity to dataset choice was noted for less prominent causal links. The reliance on specific causal inference methods introduces assumptions (such as causal sufficiency and Markov condition for PCMCI+) that could influence results. Finally, limitations inherent in climate model simulations (such as biases in representing ENSO and NA SST variability and teleconnections) may impact the generalization of some findings.