Pakistan and northwestern India experienced devastating floods during June-August 2022 due to unprecedented rainfall, exceeding four standard deviations of the past 40 years. The rainfall anomaly was approximately 2.76 mm/day, resulting in significant loss of life, infrastructure damage, and crop reduction. Successful prediction of such events is critical for mitigation efforts, especially considering the projected increase in extreme flood frequency due to climate change and natural variability. Seasonal climate predictability is generally linked to tropical SST anomalies and associated teleconnection patterns. Previous research explored the predictability of Pakistan's summer rainfall using ENSO and the Indian Ocean Dipole (IOD), with varying results and limited success in explaining the 2022 event. Neither the SINTEX-F2 climate model nor the NMME successfully predicted the 2022 extreme rainfall event, despite correctly capturing the negative IOD and La Niña conditions. This study aims to understand this prediction failure and identify potential improvements in seasonal forecasting.

Existing literature identifies ENSO and IOD as potential sources of predictability for Pakistan's summer rainfall. Ashok et al. (2004) showed positive IODs correlate with high rainfall, suggesting a possible link to the 2022 event given the La Niña conditions. Jeong et al. (2023) attributed the 2022 event to a triple-dip La Niña. However, Hussain et al. (2017) found limited correlation between IOD and precipitation except in coastal areas. Iqbal and Hassan (2018) suggested a shift in the influence of ENSO and IOD on Pakistani floods over time. Studies on the 2010 Pakistan floods highlighted the roles of European blocking, tropical-extratropical interactions, and combined La Niña and negative IOD influences (Hong et al., 2011; Priya et al., 2015). The failure of prediction systems to capture the 2022 event, despite their success in predicting the negative IOD and La Niña, motivates a deeper investigation into other contributing factors.

The study uses the SINTEX-F2 climate model's 108-member ensemble reforecast from May 1991-2022 and NMME data to analyze the 2022 Pakistan rainfall event. The GPCP rainfall data, NOAA OISSTv2 SST data, and NCEP/NCAR reanalysis data are used for evaluation. The study compares the model's ensemble mean precipitation predictions with observed data, calculating the correlation skill. Inter-ensemble correlation analysis is employed to assess the co-variability of precipitation anomalies with SST anomalies and moisture convergence/divergence at 850 hPa. Two sensitivity experiments using SINTEX-F2 were conducted: NAS_OBS (nudging model SST in the NAS to observed values) and NAS_ANOM (nudging to observed SST anomalies). These experiments aimed to test the hypothesis that positive SST anomalies in the NAS contribute to the extreme rainfall. Statistical tests, including paired t-tests, were used to evaluate the significance of differences and correlations.

The SINTEX-F2 ensemble mean prediction for the 2022 Pakistan rainfall captured less than 10% of the observed anomaly, while the NMME captured less than 20%. Both models correctly predicted the negative IOD and La Niña, indicating their skill in representing large-scale patterns. However, inter-ensemble correlation analysis revealed a significant positive correlation between Pakistan rainfall anomalies and NAS SST anomalies within the SINTEX-F2 predictions, suggesting a local influence. The positive SST anomaly in the NAS was underestimated by the SINTEX-F2 but reasonably predicted by the NMME. The sensitivity experiments (NAS_OBS and NAS_ANOM) showed a substantial improvement in capturing the observed anomaly. NAS_OBS captured around 25% and NAS_ANOM approximately 15% of the observed rainfall anomaly. The enhancement of moisture convergence over the NAS due to warmer SSTs was implicated in these improvements. The study also suggests a potential teleconnection pattern between the NAS and East Asia, with increased convection potentially influencing the subtropical Asian jet and leading to hotter-than-normal summers in East Asia. Analysis of previous extreme rainfall events (1994, 2006, 2007, 2010, 2020) suggests a consistent contribution of positive NAS SSTAs to above-normal rainfall, and a scatter plot analysis confirms a significant positive correlation (0.60) between PR anomalies and NAS SST anomalies. Further, NAS SST anomalies appear independent of ENSO and IOD signals, although a positive correlation was observed between NAS SST and IOD within the SINTEX-F2.

The study's findings suggest that the positive SST anomaly in the NAS played a significant but not sole role in the 2022 extreme Pakistan rainfall. While large-scale climate modes (ENSO, IOD) were not the primary drivers of the extreme event, the NAS SST seems to have acted as a regional amplifier. The positive inter-ensemble correlation between NAS SST and the Pakistan rainfall anomalies highlights the importance of accurately resolving regional-scale processes in climate models. The improved prediction skill in the sensitivity experiments demonstrates the potential for better forecasting by improving the model's representation of NAS SST. The observed teleconnection pattern to East Asia is interesting, and warrants further study to quantify its importance and predictability. The study emphasizes the importance of studying the Western Boundary Upwelling System (WBUS) in the western Indian Ocean.

The study successfully demonstrates the importance of the positive SST anomaly in the northern coastal Arabian Sea for predicting extreme rainfall events in Pakistan. While the sensitivity experiments improved prediction skill significantly, they still couldn't fully capture the extreme nature of the 2022 event, indicating a need for further research. Future studies should focus on improving model representation of NAS SST, exploring potential interactions between NAS SST and synoptic-scale processes, and conducting more coordinated multi-model sensitivity experiments. The observed teleconnection pattern between the NAS and East Asia also requires further investigation.

The study primarily focuses on the SINTEX-F2 model, and the findings need to be validated with other climate models. The sensitivity experiments, while improving prediction skill, didn't fully reproduce the extreme rainfall, suggesting additional unmodeled processes may be at play. The analysis doesn't fully explore the role of synoptic-scale variability and high-frequency cyclogenesis, which might have contributed to the extremity of the event. The teleconnection to East Asia is a preliminary finding requiring further investigation to establish its reliability and strength.