Climate warming contributes to the record-shattering 2022 Pakistan rainfall

Pakistan lies at the western edge of the South Asian summer monsoon’s pluvial region, with most of the country—especially southern coasts and Indus lowlands—receiving little rain; only the northern mountains receive substantial orographic precipitation. From mid-June to late August 2022, multiple deluges combined with unprecedented glacial melt produced catastrophic flooding that exceeded the infamous 2010 event and all instrumental records over the last 50 years, causing close to 2000 fatalities, displacing over 2.1 million people, inundating more than 75,000 km², and inflicting at least $30 billion (USD) in damages. The purpose of the study is to identify the multiscale physical processes and forcing factors that initiated the 2022 flood and to assess whether such a rare event could occur from natural variability alone or required contributions from anthropogenic climate change.

Prior studies attribute Pakistan’s heavy rainfall episodes primarily to South Asian monsoon low-pressure systems (LPSs), which often originate near the Bay of Bengal and travel inland, triggering deep convection and heavy precipitation ahead and south of their centers. Upper-tropospheric dynamics, including divergence linked to extratropical troughs over northern Pakistan, have contributed to past extremes (e.g., 1988, 2010, 2013). At planetary scales, Pakistan rainfall and its synoptic drivers are modulated by large-scale monsoon circulation influenced by tropical SST anomalies, particularly ENSO and the Indian Ocean Dipole (IOD). While tropical SST patterns (La Niña, negative IOD) and upstream blocking have been proposed for 2022, these anomalies were not unprecedented. Other studies suggest climate warming has already enhanced Pakistan rainfall intensity by more than 50%, but mechanisms remain uncertain and likely act via gradual background changes. This study builds on that literature by linking synoptic LPS behavior to large-scale moisture transport trends and anthropogenic forcing.

Observations: Gridded daily precipitation from CPC Unified Gauge-Based Analysis (CPC-UNI), CMAP, GPCC, and CHIRPS. Reanalysis fields (six-hourly and monthly) from MERRA-2 at 0.5° × 0.625° for 1980–present; chosen for better consistency with gauge-based datasets relative to JRA-55, NCEP/NCAR, and ERA5. Monthly SST from NOAA ERSST v5. Model data: Monthly CMIP6 simulations including historical all-forcing, single-forcing (greenhouse gases only), and future projections under SSP2-4.5 and SSP5-8.5. Multi-model ensemble means used to estimate externally forced signals; ensembles with at least five members used for extreme occurrence analysis. LPS objective tracking: Built on TempestExtremes using 6-hourly MERRA-2 reanalysis. Synoptic sea level pressure (SLP) extracted via 21-day high-pass filter; spectral filtering applied (vorticity retain total wavenumber 5–42; SLP 5–63). LPS candidates identified by local 850 hPa relative vorticity maxima >4 × 10^-5 s^-1 plus a closed-contour SLP criterion (≥0.2 hPa increase within 5°). Candidates linked across time with a maximum translation of 5°; non-transiting features removed. Retained cyclones last ≥2 days with ≥1 track point in 10°N–35°N, 60°E–95°E. Trajectories validated against manual identification from weather charts and Pakistan Meteorological Department. LPS metrics computed: genesis density, track density (converted to LPS-days by dividing 6-hourly counts by 4), translation velocity. Daily rainfall flagged as LPS-induced when an LPS was within 500 km during ±1 day. Moisture budget and transport: Column-integrated moisture budget evaluated for Pakistan box (20°N–32.5°N, 60°E–72.5°E). Net moisture convergence decomposed into zonal and meridional boundary fluxes of integrated vapor transport (IVT); time series of net, zonal (uq), and meridional (vq) components computed. Cross-equatorial moisture transport over the Arabian Sea defined as area-averaged IVT in a specified boxed region and related to LPS presence over Pakistan. Diagnostics: Composites constructed for days with LPS within Pakistan, seasonal anomalies for 2022, and ENSO/IOD phase composites (negative minus positive phases). Trends estimated via linear regression; statistical significance assessed by two-sided Student’s t-test. Correlations computed between normalized cross-equatorial transport and LPS-days over Pakistan; additional correlations for Bay of Bengal moisture import. Extreme occurrence probabilities estimated as the percentage of summers exceeding the 98th percentile of 1951–1980 Pakistan summer mean rainfall for periods 1985–2014 and 2020–2049 in each ensemble member.

- Magnitude and rarity: In summer (June–September) 2022, Pakistan’s country-wide average precipitation reached 3.95 mm/day, exceeding the 1980–2021 mean (1.03 mm/day) by 283% and by approximately 7 times its interannual standard deviation. The largest anomalies were over the climatologically dry southern plains (<0.6 mm/day climatology), with a corridor of heavy rainfall extending from central India to southern Pakistan. - Event structure: Six intense rain pulses occurred from mid-June to end-August 2022. Except for the first mid-June episode, each pulse was triggered by LPSs migrating northwestward from the Bay of Bengal across central India toward Pakistan. LPS genesis in 2022 was below average, but LPS intensity and longevity were stronger-than-average, with many storms traveling farther west and spending record numbers of days over Pakistan, yielding record-high LPS activity there. - Cross-equatorial moisture transport: A conveyor belt of moisture originated in the South Indian Ocean and crossed the equator off Somalia toward NW India/Pakistan, preconditioning LPS survival and heavy rainfall. The 2022 seasonal mean cross-equatorial transport remained anomalously strong even without transient LPS signatures. The normalized cross-equatorial transport correlates with the normalized number of LPS-days in Pakistan (r = 0.51 including 2022; r = 0.47 excluding 2022). Moisture import over the Bay of Bengal showed little relation to Pakistan LPS activity (r ≈ -0.26 without 2022). - Moisture budget trends (1980–2022): The meridional moisture inflow into Pakistan increased by 17.46% per decade (+1.29 × 10^8 kg s^-1 per decade), reaching a historical high in 2022; zonal moisture outflow trended upward by 3.56% per decade (-0.44 × 10^8 kg s^-1 per decade), but was weaker than the trend projection in 2022. The combination produced a large anomalous convergence in 2022. The long-term upward trend in cross-equatorial transport is robust across reanalyses and consistent with climate model projections. - Role of internal variability (ENSO/IOD): Summer 2022 featured a prolonged La Niña and a strong negative IOD. Pakistan rainfall is negatively correlated with ENSO (r = -0.37) and shows no significant correlation with IOD (r ≈ 0.05 without 2022). ENSO–IOD composites indicate anomalous easterly moisture flux over central India that tends to reduce Pakistan’s mean eastward outflow, consistent with some contribution from natural variability. However, these composites lack the strong cross-equatorial transport over the Arabian Sea seen in 2022. - Model-based attribution and projections: CMIP6 historical all-forcing simulations show a modest but spatially coherent forced increase in Pakistan summer rainfall (about 4–5% per decade) anchored by enhanced cross-equatorial moisture transport. Single-forcing runs indicate GHGs alone produce wetting and increased cross-equatorial transport. The occurrence of Pakistan seasonal extremes (>98th percentile of 1951–1980) shows little change in 1985–2014 (0–9.3%) but increases substantially in 2020–2049 under SSP2-4.5, with some models projecting further increases under SSP5-8.5 (differences not statistically distinct). Models agree on a wetting trend due to dynamical enhancement of monsoonal cross-equatorial transport and more frequent LPS activity, though the magnitude is uncertain (e.g., 1.7–19% under SSP2-4.5). While the frequency/intensity of La Niña and negative IOD events remains roughly unchanged, their rainfall impacts over Pakistan increase in a warming climate. - Overall: The 2022 extreme was facilitated by both internal variability (La Niña and negative IOD reducing zonal outflow) and anthropogenic warming (enhanced cross-equatorial moisture transport providing fuel for LPS convection and enabling westward propagation).

The study addresses whether the 2022 Pakistan flood arose purely from natural variability or required anthropogenic influence. Analyses show that while internal variability (La Niña with negative IOD) favored conditions reducing Pakistan’s mean moisture export, the event’s severity hinged on an exceptionally strong cross-equatorial moisture transport over the Arabian Sea, which is part of a long-term upward trend consistent with anthropogenic warming. This background moistening and dynamical enhancement sustained LPS intensity and longevity, enabling storms to travel farther west into Pakistan and persist longer over dry regions, thereby producing unprecedented seasonal accumulations. Observational trends, reanalyses, and CMIP6 models converge on a mechanism in which greenhouse gas–forced increases in cross-equatorial moisture transport and LPS activity substantially raise the likelihood of rare, season-scale extremes over Pakistan. Thus, anthropogenic warming set the stage upon which internal variability produced the record-shattering outcome. The implications for the field are that future flood risk assessments for Pakistan and the broader South Asian monsoon region must account for dynamical changes in moisture transport and LPS statistics, not just thermodynamic moisture increases.

This work provides a multiscale explanation for the record-shattering Pakistan rainfall of summer 2022. It demonstrates that six LPS-driven rain pulses, sustained and intensified by anomalously strong cross-equatorial moisture transport over the Arabian Sea, produced the extreme seasonal totals. The background state conducive to such transport has been trending upward for decades, consistent with anthropogenic forcing, while internal variability (La Niña and negative IOD) modulated moisture fluxes to further favor the event. Model analyses indicate greenhouse gases alone can drive increased cross-equatorial transport and a wetting trend, and that the probability of seasonal extreme rainfall over Pakistan will rise substantially in coming decades even under moderate emissions. Future research directions include: resolving whether the observed cross-equatorial transport trend exceeds natural variability with longer records and detection–attribution methods; improving model representation of LPS genesis, intensity, and propagation via higher resolution and better physics; quantifying the sources of inter-model uncertainty in projected wetting magnitudes; and developing event-based risk frameworks that combine changing background moisture transport with modes of variability such as ENSO and IOD.

- Attribution nuance: While evidence indicates a causal role for enhanced cross-equatorial transport, disentangling cause–effect between LPS strength and moisture transport remains challenging. - Natural variability bound: It is not yet established whether the observed upward trend in cross-equatorial moisture transport exceeds natural variability. - Data and dataset differences: Rainfall magnitude in 2022 varies among observational datasets, though all show an exceptional event. - Model limitations: CMIP6 models do not reproduce events as extreme as 2022 (4–7σ) and have coarse resolution that limits realistic simulation of LPS characteristics, contributing to uncertainty in projected magnitudes. - Uncertainty sources: The magnitude of future wetting and extreme occurrence increases varies substantially across models; a comprehensive analysis of uncertainty sources is beyond the scope of this study.