Zhang et al. (2026) Deceleration of water cycle for global semi-arid regions driven by climate warming

Identification

• Journal: Climate Dynamics
• Year: 2026
• Date: 2026-03-27
• Authors: Qiang Zhang, Zesu Yang, Ping Yue, Liang Zhang, Jian Zeng
• DOI: 10.1007/s00382-026-08136-w

Research Groups

• Key Open Laboratory of Arid Climatic Change and Disaster Reduction of CMA, Institute of Arid Meteorology, CMA; Key Laboratory of Arid Climatic Change and Reducing Disaster of Gansu Province, Lanzhou, China.
• Climate Change and Resource Utilization in Complex Terrain Regions Key Laboratory of Sichuan Province; Chengdu Plain Urban Meteorology and Environment Observation and Research Station of Sichuan Province, Sichuan Provincial Engineering Research Center for Meteorological Disaster Prediction and Early Warning, School of Atmospheric Sciences, Chengdu University of Information Technology, Chengdu, Sichuan, China.
• Center for Earth System Modeling Research, Lanzhou University, Lanzhou, China.

Short Summary

This study reveals a systematic deceleration of the water cycle in global semi-arid regions from 1982 to 2022, driven by climate warming, which contradicts the expected global intensification and is attributed to water-limited evapotranspiration responses and a positive feedback loop among water cycle components.

Objective

• To detect trends in the components of the water cycle in global semi-arid areas under global warming.
• To explore the coupling processes among local water cycle variables and their responses to climate warming, contrasting with previous studies attributing changes primarily to atmospheric circulation.

Study Configuration

• Spatial Scale: Global semi-arid regions (defined by an aridity index between 0.2 and 0.5), analyzed on a 0.5° × 0.5° grid.
• Temporal Scale: 1982–2022 (41 years).

Methodology and Data

• Models used:
  • Priestley-Taylor-JPL (PT-JPL) model (for evapotranspiration estimation and factorial analysis).
  • UTrack-atmospheric-moisture model (for moisture tracking).
  • Global Land Data Assimilation System (GLDAS) land surface models (for ensemble evapotranspiration and soil water content).
  • Convergent Cross Mapping (CCM) (for causality analysis).
• Data sources:
  • Precipitation: PRECipitation REConstruction over Land (PREC/L), GLDAS, NCEP, ERA5, MERRA2, CRU.
  • Temperature: CPC GHCN_CAMS Gridded 2 m temperature.
  • Relative Humidity: Calculated from CPC 2 m temperature and CRU vapor pressure; also CRU, GLDAS, NCEP, ERA5, MERRA2, HadISDH.
  • Soil Water Content: Ensemble average of four GLDAS land surface models; also NCEP, ERA5, MERRA2, CCI, GLEAM.
  • Evapotranspiration (ET): Ensemble mean of seven datasets (MERRA-v2, ERA5, four GLDAS offline products, PT-JPL); also PT-JPL, ERA5, GLDAS, MERRA2, GLEAM.
  • Runoff: GRUN (observation-based global gridded runoff dataset); also GLDAS, NCEP, ERA5, MERRA2, GloFAS.
  • Moisture Divergence: ERA5 monthly averaged data.
  • Moisture Trajectories: UTrack moisture trajectory dataset (derived from ERA5).
  • Model Forcing (PT-JPL): Temperature (GLDAS), relative humidity (GLDAS), wind speed (GLDAS), Normalized Difference Vegetation Index (NDVI) from GIMMS and MODIS.
  • Validation Data: Eddy covariance observations from FLUXNET and HiWATER (for ET validation).

Main Results

• Global semi-arid regions experienced significant warming from 1982 to 2022, with an average rate of 0.28 °C per decade.
• The water cycle in the vast majority of global semi-arid areas is systematically slowing down, with all key components exhibiting declining trends.
• Water fluxes decreased: evapotranspiration (ET) at -7.4 ± 2.6 mm per decade, precipitation at -8.2 ± 9.2 mm per decade, and runoff at -3.4 ± 3.6 mm per decade.
• Water storage decreased: soil water content at -0.37 ± 0.35 % per decade and air relative humidity at -0.69 ± 0.48 % per decade.
• Intensified water constraints are the main cause of the slowdown. Rapid temperature rise causes a reduction in relative humidity and ET, triggering the deceleration.
• Decreased relative humidity (increased vapor pressure deficit, VPD) directly suppresses ET by limiting stomatal opening (negative response of stomatal conductance to VPD).
• A positive feedback loop exacerbates the impacts: rising temperature decreases relative humidity, which suppresses ET. Decreased ET and relative humidity diminish cloud cover, reducing precipitation, which in turn lowers soil water content and runoff. Reduced soil water content further limits ET and enhances air temperature.
• The decline in local moisture recycling dominates the overall weakening of the water cycle in most semi-arid regions, despite an increase in external moisture transport.
• The aridity index (AI) in global semi-arid regions decreased significantly (p < 0.001) at a rate of -0.0011 per year, with 76% of this decrease attributed to enhanced potential ET and 24% to reduced precipitation, indicating exacerbated aridification.

Contributions

• Provides robust evidence for a systematic deceleration of the water cycle in global semi-arid regions under climate warming, contradicting the widely accepted global acceleration trend.
• Identifies the primary mechanism for this deceleration: water-limited ET response to warming, driven by decreased relative humidity and a positive feedback loop among water cycle components (temperature, relative humidity, ET, cloud cover, precipitation, soil water content, runoff).
• Highlights the critical role of internal moisture recycling degradation in the weakening of the semi-arid water cycle, complementing the "wet-gets-wetter, dry-gets-drier" paradigm.
• Offers scientific support for understanding the expansion of semi-arid areas and informs water resource management and drought risk mitigation strategies in these vulnerable regions.

Funding

• National Natural Science Foundation of China (Grant No. 42230611 and 42005071).

Citation

@article{Zhang2026Deceleration,
  author = {Zhang, Qiang and Yang, Zesu and Yue, Ping and Zhang, Liang and Zeng, Jian},
  title = {Deceleration of water cycle for global semi-arid regions driven by climate warming},
  journal = {Climate Dynamics},
  year = {2026},
  doi = {10.1007/s00382-026-08136-w},
  url = {https://doi.org/10.1007/s00382-026-08136-w}
}