The East Asian summer monsoon (EASM) supplies most of the summer rainfall to eastern China. Using paired climate-model ensembles with and without anthropogenic forcings, Burke and Stott (2017) concluded that human influence has dried the total monsoon (about 45 mm less area-average seasonal rainfall) while making the most extreme rainfall events shorter in duration and more intense per day, but they could not verify these model predictions against observations they judged to be imperfect. Subsequent work has now detected a human fingerprint in heavy and even hourly precipitation extremes over China, so the original verification gap has been partially closed. A precise core remains open: (1) whether seasonal total or mean EASM rainfall over the core monsoon region of eastern China carries a detectable anthropogenic signal separable from natural and internal variability; and (2) observational verification and attribution of the predicted structural change of extreme events toward shorter duration with higher per-day intensity. A robust decomposition of the detected anthropogenic signal into greenhouse-gas and aerosol contributions remains an open question in the later literature, but the source's ALL-versus-NAT ensemble design isolates only combined anthropogenic forcing, so that decomposition is related background rather than part of the source's verification gap.

Background

The East Asian summer monsoon (EASM) is the seasonal reversal of winds and rainfall that dominates the climate of eastern China from late spring through summer, delivering most of the region's annual precipitation via a northward-migrating rain band (the mei-yu front). Its variability governs recurrent floods and droughts for a large fraction of the world's population, so attributing observed EASM rainfall changes to specific climate forcings is a central problem in monsoon meteorology.

Burke and Stott (2017) addressed this attribution question with the HadGEM3-A event-attribution system, which compares two large ensembles of atmosphere-only climate simulations driven by observed sea-surface temperatures: an ALL ensemble including all historical forcings (greenhouse gases plus anthropogenic aerosols, along with natural solar and volcanic forcing) and a NAT ensemble including only natural forcings. The difference between the ensembles isolates the modeled effect of anthropogenic forcing. They found that under anthropogenic forcing the EASM becomes generally drier — total seasonal rainfall decreases by roughly 45 mm in the East China area average and dry days increase — while for the most extreme rainfall events, defined as runs of consecutive days each with more than 1 mm of rain, the rainfall per day increases and the number of days over which the rain falls decreases. The model also shows a shift of daily rainfall totals toward the lighter end of the distribution.

Verification against observations was, however, limited by the available data. APHRODITE, a widely used daily gridded rain-gauge product for Asia, was shown in the same study to underestimate heavy rainfall relative to stations and to understate the observed dipole trend pattern known as southern flood-northern drought (SFND): increased summer rainfall and flooding in southern China accompanied by drying in northern China. The source's own simulated total-rainfall decrease is not this dipole pattern: it is more severe in the south of the examined region than in the north, ranging from tens of mm in northeast China to about 100 mm or more in southern areas. Station coverage is sparse in western and parts of interior China. The authors therefore flagged that more detailed, up-to-date observational studies would be required to verify the modeled changes, that the mechanism producing SFND was outside their scope, and that their daily-resolution event definition (consecutive days above 1 mm) could not resolve the true sub-daily duration of rainstorms.

Since 2017, the standard tools of detection and attribution (D&A) have advanced considerably. Optimal fingerprinting — the regression method that compares observed pattern changes with model-simulated signal patterns to detect a forced response above internal variability — has been applied to newly homogenized station-based datasets for China (records corrected for artificial breaks caused by instrument moves and changes). Single-forcing model experiments, in which ensembles are driven by greenhouse gases only or anthropogenic aerosols only (as in CMIP6, the Coupled Model Intercomparison Project Phase 6, and its DAMIP component, the Detection and Attribution Model Intercomparison Project, which coordinates standardized single-forcing experiment suites), allow the anthropogenic signal to be decomposed by forcing agent. Using such tools, later studies detected an anthropogenic signal in the observed increase of heavy summer precipitation over China (Lu et al. 2020), in the intensification of hourly and daily precipitation extremes over eastern China (Chen et al. 2021, with intensification near twice the Clausius-Clapeyron rate, i.e., roughly twice the ~6.5% per degree warming scaling of atmospheric moisture capacity), and in the daily-scale circulation patterns that favor the SFND dipole (Zhou et al. 2021). Robustness tests found human influence detectable in China's extreme precipitation regardless of period, index, or data treatment (Dong et al. 2022).

Yet these advances leave the source paper's verification gap only partially closed. The most recent attempt to detect anthropogenic influence in China's mean (total) precipitation succeeded in northwest China, northeast China, and the Tibetan Plateau but not in southern China, the EASM core, where models disagree strongly about trends (Wang et al. 2025). Decomposing the anthropogenic signal into greenhouse-gas and aerosol contributions remains formally difficult in the fingerprinting framework (Dong et al. 2022), and single-model process studies give region-dependent answers: greenhouse gases dominate wetting of the southeast while aerosols dominate drying of the southwest (Guo et al. 2023; Xu et al. 2024). And no study has tested the specific structural prediction that anthropogenic forcing shortens extreme-event duration while raising per-day intensity, even though observational work now documents a shift toward short-duration, high-intensity storms in parts of China (Jiang and Tan 2025). Future-projection uncertainty — how aerosol reductions combined with continued greenhouse-gas increases will reshape the monsoon — has meanwhile been narrowed by robust projected increases in EASM rainfall across large ensembles and CMIP6 (Xue et al. 2023), and the field remains under active review (Zhang et al. 2026).

Problem Statement

Complete the detection and attribution of anthropogenic influence on East Asian summer monsoon rainfall over eastern China, as posed by Burke and Stott (2017) but against the improved observations and model ensembles now available. The overarching objective is the observational verification of the source's headline predictions for the anthropogenic response, posed as two detection questions. (1) Determine whether observed multi-decadal changes in seasonal total (mean) monsoon rainfall over the core EASM region of eastern China contain a detectable anthropogenic signal separable from natural external forcing and internal multidecadal variability. (2) Verify observationally whether the structure of extreme rainfall events has changed as the source predicts — events shorter in duration with greater rainfall per day — at daily and sub-daily (hourly) timescales, attributing any change to anthropogenic forcing rather than internal variability, urbanization, or land-use change. A decomposition of the detected anthropogenic signal into greenhouse-gas and aerosol contributions is a related open question in the later literature (see Background) but is not required for completion here, since the source's ALL-versus-NAT ensemble design isolates only combined anthropogenic forcing.

The source fixes the conventions of the required metrics, and they are part of the problem definition: the monsoon season runs from the beginning of April to the end of August; the region is the East China monsoon region with western China excluded, whose seven climatologically similar sub-regions (NE, NCC, NEC, CW, SCC, SEC, SE, as labeled in the source's Figure 1) are available for sub-regional analysis; a rainfall event is a number of consecutive days where each day has total rainfall greater than 1 mm, characterized by its duration (n_days), its total rainfall (n_day_tot), and its mean rainfall per day (intens); and the "most extreme" events are those in the upper tail of the n_day_tot distribution, the source using the 95th percentile defined from the NAT ensemble for events between 1960-1979 and computing likelihoods for 1996-2015 against the 1960-1979 baseline. Because detection verdicts depend on these choices, any analysis must state its region, season, comparison window, baseline, and event definition explicitly; updated comparison windows are admissible, but the source's parameters define the reference metrics against which its predictions were posed. At hourly resolution no source definition exists — the source states that its daily data cannot resolve the "real" duration of individual rainstorms — so any sub-daily event-segmentation criteria (minimum rain intensity and maximum tolerated dry gap within an event) must likewise be stated explicitly.

An answer must establish, with quantified uncertainty, whether observed changes in EASM seasonal-total rainfall and in the duration and intensity structure of extreme events are attributable to anthropogenic forcing, and the magnitude of any attributable change. Because both components are posed as detection questions, a rigorously supported non-detection — or a demonstration that the signal is not separable from natural external forcing and internal variability with the available observations — equally resolves the corresponding question.

References

  1. Claire Burke, Peter Stott (2017), 'Impact of Anthropogenic Climate Change on the East Asian Summer Monsoon', Journal of Climate 30(14), 5205-5220. DOI: 10.1175/JCLI-D-16-0892.1 (author copy: arXiv:1704.00563). https://doi.org/10.1175/JCLI-D-16-0892.1
  2. Chunhui Lu, Fraser C. Lott, Ying Sun, Peter A. Stott, Nikolaos Christidis (2020), 'Detectable Anthropogenic Influence on Changes in Summer Precipitation in China', Journal of Climate 33(13), 5357-5369. DOI: 10.1175/JCLI-D-19-0285.1. https://doi.org/10.1175/JCLI-D-19-0285.1
  3. Yang Chen, Wei Li, Xiaoling Jiang, Panmao Zhai, Yali Luo (2021), 'Detectable Intensification of Hourly and Daily Scale Precipitation Extremes across Eastern China', Journal of Climate 34(3), 1185-1201. DOI: 10.1175/JCLI-D-20-0462.1. https://doi.org/10.1175/JCLI-D-20-0462.1
  4. Baiquan Zhou, Panmao Zhai, Simon F. B. Tett, Fraser C. Lott (2021), 'Detectable anthropogenic changes in daily-scale circulations driving summer rainfall shifts over eastern China', Environmental Research Letters 16(7), 074044. DOI: 10.1088/1748-9326/ac0f28. https://doi.org/10.1088/1748-9326/ac0f28
  5. Siyan Dong, Ying Sun, Xuebin Zhang (2022), 'Attributing observed increase in extreme precipitation in China to human influence', Environmental Research Letters 17(9), 095005. DOI: 10.1088/1748-9326/ac888e. https://doi.org/10.1088/1748-9326/ac888e
  6. Yan Guo, Buwen Dong, Jiangshan Zhu (2023), 'Anthropogenic impacts on changes in summer extreme precipitation over China during 1961–2014: roles of greenhouse gases and anthropogenic aerosols', Climate Dynamics 60, 2633-2643. DOI: 10.1007/s00382-022-06453-4. https://doi.org/10.1007/s00382-022-06453-4
  7. Suyeon Moon, Nobuyuki Utsumi, Jee-Hoon Jeong, Jin-Ho Yoon, S.-Y. Simon Wang, Hideo Shiogama, Hyungjun Kim (2023), 'Anthropogenic warming induced intensification of summer monsoon frontal precipitation over East Asia', Science Advances 9(47), eadh4195. DOI: 10.1126/sciadv.adh4195. https://doi.org/10.1126/sciadv.adh4195
  8. Huiwen Xu, Huopo Chen, Wenyue He, Lefeng Chen (2024), 'Contributions of Anthropogenic Greenhouse Gases and Aerosol Emissions to Changes in Summer Precipitation Over Southern China', Journal of Geophysical Research: Atmospheres 129, e2023JD040331. DOI: 10.1029/2023JD040331. https://doi.org/10.1029/2023JD040331
  9. Tao Wang, Ying Sun, Xuebin Zhang, Xiu-Qun Yang, Heyang Song (2025), 'Detectable Anthropogenic Influence in Mean Precipitation of China', Geophysical Research Letters 52, e2025GL114870. DOI: 10.1029/2025GL114870. https://doi.org/10.1029/2025GL114870
  10. Weizhen Jiang, Yong Tan (2025), 'Characterization of intra-event patterns of heavy rainfall events in China during 2008–2019 by machine learning', Theoretical and Applied Climatology 156, 372. DOI: 10.1007/s00704-025-05607-8. https://doi.org/10.1007/s00704-025-05607-8
  11. Daokai Xue, Jian Lu, L. Ruby Leung, Haiyan Teng, Fengfei Song, Tianjun Zhou, Yaocun Zhang (2023), 'Robust projection of East Asian summer monsoon rainfall based on dynamical modes of variability', Nature Communications 14, 3856. DOI: 10.1038/s41467-023-39460-y. https://doi.org/10.1038/s41467-023-39460-y
  12. Wenxia Zhang, Tianjun Zhou, Xiaolong Chen, Jie Jiang (2026), 'Anthropogenic influences on precipitation changes: research progress and prospects', Chinese Science Bulletin. DOI: 10.1360/csb-2025-5845. https://doi.org/10.1360/csb-2025-5845