The summer monsoon rainfall (July-September) is Pakistan's primary precipitation source. Anomalous monsoon circulation and moisture transport often lead to floods and droughts. Previous research has focused on seasonal and synoptic anomalies, but the role of intraseasonal oscillations (ISOs; 10–90-day periods) in the record-breaking August 2022 Pakistan rainfall event, which affected 33 million people, remains relatively unexplored. Sea surface temperature (SST) anomalies over the tropical Pacific and Indian Ocean (La Niña and negative Indian Ocean Dipole states), easterly wind anomalies over northern India, and southerly wind anomalies from the Arabian Sea have been linked to heavy rainfall in Pakistan. Eurasian heatwaves and blocking events have also been implicated. While studies have examined the influence of synoptic low-pressure systems and western disturbances, a comprehensive understanding of the relative effects of various ISO modes (both tropical and extratropical) and their key processes in the 2022 event remains lacking. Accurate prediction of such extreme events is crucial for disaster mitigation, and the predictability at subseasonal-to-seasonal (S2S) timescales remains unclear. This study examines the impact of tropical-extratropical interactions among various intraseasonal modes and aims to identify which modes are crucial for accurately predicting such events at subseasonal timescales.

Existing literature highlights the influence of various factors on Pakistan's summer monsoon rainfall and extreme events. Studies have explored the roles of seasonal anomalies linked to SSTs in the Pacific and Indian Oceans, particularly La Niña events and negative Indian Ocean Dipole states. These anomalies influence wind patterns, leading to intensified moisture convergence and heavy rainfall. The impact of Eurasian heatwaves and blocking events, through teleconnections, has also been documented. Furthermore, research emphasizes the role of synoptic low-pressure systems, western disturbances, and their interactions in South Asia, suggesting that their increased intensity and duration could amplify flood risks. While the involvement of intraseasonal oscillations (ISOs) in extreme rainfall events is acknowledged, the specific contribution of different ISO modes (both tropical and extratropical) to the 2022 Pakistan event requires further investigation. Previous studies have focused more on the 30–60-day ISO, overlooking the potential contribution of the 10–30-day mode. Furthermore, the predictability of this extreme event at subseasonal-to-seasonal (S2S) timescales, bridging climate prediction and weather forecasting, needs comprehensive assessment.

The study used multiple gridded precipitation datasets: CPC (0.5° × 0.5°), ERA5 (0.25° × 0.25°), and JRA-55 (1.25° × 1.25°). Daily outgoing longwave radiation (OLR) data (2.5° × 2.5°) was used to analyze deep convection, and ENSO status was obtained from the Niño3.4 SST index. ERA5 and JRA-55 data provided wind, vertical p-velocity, specific humidity, and geopotential height. A temporal decomposition approach was employed to separate rainfall variability into synoptic (<10 days), 10–30-day ISO, 30–90-day ISO, and low-frequency background state (LFBS; >90 days) components. The moisture budget equation was analyzed using a scale-decomposed method to examine the contributions of different scale interactions. Wave activity flux (WAF) was calculated to study the propagation of intraseasonal Rossby wave trains. State-of-the-art operational S2S prediction models (CMA, ECCC, ECMWF, HMCR, IAP-CAS, and NCEP) were evaluated to assess their skill in predicting the extreme rainfall event. The analysis focused on identifying the sources of biases in the predictions and determining which ISO modes were most important for skillful subseasonal forecasting.

Analysis of rainfall data revealed that the 2022 Pakistan flood event was characterized by exceptionally high rainfall exceeding climatological means by several standard deviations. The moisture budget analysis showed that enhanced vertical moisture transport was the primary driver of the extreme rainfall. Approximately 56–60% of the rainfall during the flood period was attributed to the 10–30-day and 30–90-day ISO modes. The low-frequency background state (LFBS) contributed nearly 40%, while synoptic variability contributed only about 4%. Both the 10–30-day and 30–90-day ISOs exhibited positive phases during the heavy rainfall period. The northward propagation of the 30–90-day convection from the equatorial Indian Ocean and the southward movement of the mid-latitude 30–90-day variability toward Pakistan played significant roles. The 10–30-day convective anomalies primarily originated from the tropics and moved northward. The spatial and temporal evolution of both 30–90-day and 10–30-day ISOs indicated a coupling of tropical and extratropical modes, particularly the 30–90-day mode. Evaluation of S2S models revealed that models accurately predicting northward-propagating intraseasonal convection (particularly the tropical 30–90-day mode) demonstrated better skill in predicting the extreme event, even if mid-latitude ISO variability was not captured accurately. The biases in predicted rainfall were most strongly correlated with the fidelity of the predicted 30–90-day tropical ISO evolution. Higher-resolution models demonstrated better skill in capturing the enhanced rainfall anomalies.

The findings highlight the critical role of northward-propagating tropical intraseasonal oscillations (ISOs), particularly the 30–90-day mode, in driving the prolonged heavy rainfall in Pakistan during August 2022. The interaction between tropical and extratropical ISOs, especially the merging of northward-propagating tropical convection with southward-moving mid-latitude systems, contributed significantly to this extreme event. The model prediction assessment confirms that the accuracy of predicting the tropical 30–90-day ISO is crucial for skillful subseasonal forecasting of this event. The slight underestimation of rainfall amplitude in the best-performing model members might be related to limitations in accurately capturing synoptic-scale systems. The study suggests that the physical frameworks and settings of the models, including horizontal resolution and convective parameterizations, may significantly affect the accuracy of convective perturbation predictions, warranting further investigation.

The 2022 Pakistan flooding event was primarily driven by the northward propagation of tropical 30–90-day and 10–30-day intraseasonal oscillations (ISOs) from the tropics towards Pakistan. Accurate prediction of these tropical ISOs, particularly the 30–90-day mode, is a major source of subseasonal predictability for extreme rainfall events in this region. Future research should focus on the factors affecting the accuracy of convective perturbation predictions in models, examining the roles of model parameters and investigating the dynamics of tropical ISO diversity in the South Asian summer monsoon region to better understand the frequency of extreme rainfall events in the northern Indian subcontinent.

The study's limitations include potential uncertainties associated with the use of reanalysis datasets and the relatively small number of S2S models considered in the prediction assessment. Although a range of models were used, a more comprehensive evaluation across many more models is warranted for improved generalizability. The precise mechanisms responsible for the simultaneous northward movement of the enhanced 30–90-day and 10–30-day convections require further investigation. Future studies could investigate a wider range of extreme events to better understand their dynamics and predictability.