Increasing global precipitation whiplash due to anthropogenic greenhouse gas emissions

The study addresses how and why rapid transitions between dry and wet precipitation extremes—termed precipitation whiplash—are changing globally under anthropogenic climate forcing. Whiplash events exacerbate risks to ecosystems and society by creating short preparedness windows and cascading impacts (e.g., wildfire risk following wet periods or flooding and erosion following droughts). While prior work has examined wet–dry transitions mostly at seasonal to annual scales or regionally, the global climatology, timing, and changes in sub-seasonal whiplash have remained unclear. The authors propose and apply a metric to detect sub-seasonal wet-to-dry and dry-to-wet transitions globally, quantify historical and projected changes in their frequency, transition duration, and intensity, and disentangle the roles of specific anthropogenic forcings (greenhouse gases, industrial aerosols, and biomass-burning aerosols).

Previous regional studies analyzed transitions and variability of wet and dry extremes largely at seasonal or annual scales using standardized precipitation indices, anomalous seasonal precipitation, or drought–pluvial dipole years. Sub-seasonal circulation drivers of extremes have been examined in specific regions, but a global assessment of sub-seasonal precipitation whiplash has been lacking. Broader literature shows precipitation variability increases in a warmer climate across timescales, driven by thermodynamic moisture increases and dynamic circulation changes, with strong regional heterogeneity. Monsoon variability and seasonal contrasts are expected to intensify with warming, whereas in some regions (e.g., subtropical North America) increased moisture may be offset by weakened circulation variability. Anthropogenic greenhouse gases generally intensify hydrological cycle variability and extremes, while aerosols can counteract some warming effects regionally and historically.

• Datasets: Used daily precipitation from CESM Large Ensemble (CESM-LENS; 40 members, 1920–2100, historical to 2005 and RCP8.5 thereafter), CESM single-forcing ensembles (CESM-XLENS: XAER, XGHG, XBMB; 20/20/15 members with respective forcings fixed at 1920 while others follow historical+RCP8.5), and CMIP6 multi-model ensemble (55 realizations from 22 models; historical to 2014 and SSP5-8.5 for 2015–2100). Observational/reanalysis datasets include ERA5, MERRA-2, JRA-55 (reanalyses), CHIRPS (satellite-based), GPCC and REGEN (gauge-based). All data regridded to 2°.
• Preprocessing: Linear detrending per grid cell to remove long-term forced trends. Compute 30-day rolling sums (sub-seasonal cumulative precipitation). Remove annual cycle via standardization by day-of-year climatology and standard deviation over 1920–2100.
• Event identification: Define dry (wet) extremes as standardized anomalies below 10th (above 90th) percentile thresholds computed from the current period (1979–2019). A whiplash event is a transition from dry-to-wet or wet-to-dry occurring within 30 days. Merge briefly interrupted extremes using a run-theory-like rule; discard extreme spells lasting ≤3 days.
• Metrics: Occurrence frequency (events per year), transition duration (days from last day of first extreme to first day of opposite extreme), and transition intensity (absolute difference between the driest and wettest standardized anomalies within a whiplash event). Average timing within the year computed using circular statistics.
• Change calculations: Relative annual and period-to-period changes computed relative to the current period. Future focus on late-21st-century (2060–2099) under RCP8.5/SSP5-8.5.
• Emergence and uncertainty: Signal-to-noise ratio (S/N) defined as ensemble-mean forced response divided by inter-member standard deviation; emergence when |S/N| ≥ 1. Random resampling used to estimate minimum ensemble size needed for robust detection.
• Anthropogenic attribution: Risk ratios (RR) computed by comparing CESM-LENS with single-forcing runs (e.g., XAER, XGHG, XBMB), and fractional contributions IF_X to future changes (2040–2079 vs 1979–2019).
• Circulation analysis: Composite 500 hPa geopotential height anomalies and vertically integrated vapor transport anomalies around whiplash events, exemplified for Northeast China (NEC); compare current (1979–2019) and future (2060–2099) patterns.
• Regional focus: Six IPCC AR6 monsoon regions (NAmerM, SAmerM, WAfriM, SAsiaM, EAsiaM, AusMCM) analyzed via area-weighted means.

• Global increase in occurrence: By 2060–2099 relative to 1979–2019, global mean frequency of precipitation whiplash increases by 156 ± 16% (2.56 ± 0.16 times). Over land, increases are 243 ± 22% (3.43 ± 0.22 times).
• Sharper, stronger transitions: Transition durations shorten by about −10 ± 1%, and intensities increase by 13 ± 3% globally, implying faster and more severe swings.
• Spatial hotspots: Largest increases occur in polar and monsoon regions. Timing patterns largely persist, with notable shifts at high latitudes.
• Monsoon regions: EAsiaM +196% total whiplash (≈+98% dry-to-wet; +98% wet-to-dry); SAsiaM +214% (≈+106% each type); AusMCM +206% (≈+97% dry-to-wet; +109% wet-to-dry); WAfriM >+50%; SAmerM +25%; NAmerM ~no significant change. Decreasing transition durations and increasing intensities across monsoon regions except NAmerM.
• Emergence of forced signal: For global/land means, anthropogenic signals in whiplash frequency emerge around 2028 and 2017 (dry-to-wet) and 2033 and 2017 (wet-to-dry) respectively; most land areas show emergence in the 21st century.
• Relation to precipitation totals: Changes in whiplash frequency positively correlate with changes in precipitation totals in ~91% (dry-to-wet) and ~90% (wet-to-dry) of regions. In ~76–75% of the globe, both increase together; in two-thirds of these, whiplash frequency increases exceed precipitation total increases by more than sixfold. Some regions exhibit increased whiplash despite decreasing totals (e.g., margins of wetting/drying zones; parts of southeastern Europe, western NAmerM, Amazon).
• Anthropogenic forcings: • Greenhouse gases (GHG): Increase global whiplash risk by 13 ± 2% by 2028 and 55 ± 4% (59 ± 4% over land) by 2079; account for 87 ± 4% of future (2040–2079 vs 1979–2019) occurrence increases. GHGs contribute to −4 ± 1% shortening of transition duration and +7 ± 1% intensity increase globally by 2079; they explain 40 ± 5% of duration shortening and 80 ± 12% of intensity enhancement. Strong amplification in polar and Pacific–Asian monsoon regions (>50% in AusMCM, EAsiaM, SAsiaM; >120% in polar areas). • Industrial aerosols (AER): Historically offset GHG effects regionally (e.g., 3–10% decreases in EAsiaM, SAsiaM); globally reduce whiplash risk by ~3% by 2028 and even less by 2079 as emissions decline; future residual reductions over parts of SE Asia, E/S Africa, NE South America. • Biomass-burning aerosols (BMB): Small effects overall; ~3–5% reductions in SAsiaM, SAmerM (wet-to-dry only), and AusMCM.
• Circulation mechanisms (NEC example): Whiplash transitions are associated with shifts between anticyclonic and cyclonic anomalies, movement of the western Pacific subtropical high, and modulation of East Asian summer monsoon moisture transport; future patterns are similar but with stronger anomalies, consistent with enhanced hydrological cycle intensity.

The study demonstrates that anthropogenic warming substantially increases the frequency, sharpness, and intensity of sub-seasonal precipitation regime shifts worldwide. By defining and detecting whiplash events at sub-seasonal scales, the authors show that forced changes have emerged or will imminently emerge above internal variability over most land areas. The findings link thermodynamic increases in atmospheric moisture with dynamic circulation shifts to explain more rapid transitions between extremes, particularly in monsoon and high-latitude regions. The strong positive covariation with precipitation totals indicates that as regions get wetter, volatility grows even faster. Aerosol cooling historically muted whiplash increases in some regions, but declining aerosol emissions mean GHG forcing will dominate future volatility. Regionally, mechanisms differ: for example, in the North American monsoon region, increased background moisture is offset by weakened circulation variability, yielding little net change. These results underscore significant implications for water resource management, infrastructure resilience, and ecosystem impacts, as shorter lead times and stronger swings complicate adaptation and risk mitigation.

This work introduces a global, sub-seasonal metric of precipitation whiplash and applies it across observations, a large single-model ensemble, and a multi-model ensemble to quantify historical and future changes. The authors find large, spatially widespread increases in frequency, faster transitions, and stronger intensities by late century, with the most pronounced changes in polar and monsoon regions. Attribution with single-forcing experiments shows greenhouse gases are the dominant driver of increased whiplash risk, while aerosol effects diminish as emissions decline. The study advances understanding of how thermodynamic and dynamic processes combine to intensify precipitation variability at critical sub-seasonal scales. Future research should further untangle regional mechanisms (e.g., monsoon dynamics, WPSH variability, teleconnections), assess the role of short-lived, high-intensity storms at finer spatial-temporal scales, and evaluate compound impacts to better inform adaptation strategies.

The sliding-window, sub-seasonal accumulation method can smooth short-duration, intense rainstorms that also pose major risks, potentially underrepresenting some flood-driving events following droughts. Mechanistic drivers of whiplash vary regionally, and while one region (NEC) is analyzed in detail, comprehensive process attribution across all regions remains for future work. Uncertainties from internal variability and model structural differences persist (especially in CMIP6), and aerosol emission trajectories influence regional projections, though their effects are projected to wane relative to GHG forcing.