Sun et al. (2026) Pronounced trends in vegetation phenology responses to flash droughts across China (2003−2023)

Identification

• Journal: Ecological Indicators
• Year: 2026
• Date: 2026-01-01
• Authors: Tao Sun, Xinhua Zhang, Yujie Cai, Chun Yang, Zhurui Gao, He Meng, Jiannan Zhang, Yijun Guo, Tsetan Sam
• DOI: 10.1016/j.ecolind.2025.114578

Research Groups

• State Key Laboratory of Hydraulics and Mountain River Engineering, Sichuan University, Chengdu, China
• Key Laboratory of Water Safety for Beijing-Tianjin-Hebei Region of Ministry of Water Resources, China Institute of Water Resources and Hydropower Research, Beijing, China
• Division of International Cooperation, China Institute of Water Resources and Hydropower Research, Beijing, China
• Water Conservancy Team of Sangzhuzi District, Shigatse City, China

Short Summary

This study quantified the spatiotemporal patterns and interannual trends of vegetation phenological responses (Start of Season - SOS, End of Season - EOS) to pre-season flash droughts across China from 2003 to 2023, revealing divergent regional sensitivities and dominant climatic drivers. It found that while SOS advancement due to flash droughts is decreasing, both advancement and delay of EOS are increasing, with these trends varying across hydroclimatic regions.

Objective

• To identify and characterize flash drought events based on soil moisture dynamics in China.
• To examine the spatial and temporal patterns of vegetation phenology (SOS and EOS) and their responses to pre-seasonal flash droughts (pre-SOS and pre-EOS).
• To assess the interannual trends and driving mechanisms of vegetation phenological responses to pre-seasonal flash droughts by analyzing the influences of flash drought characteristics, climatic factors, and hydrological conditions.

Study Configuration

• Spatial Scale: Entire territory of China.
• Temporal Scale: 2003 to 2023 (21 years).

Methodology and Data

• Models used: No explicit process-based models were used for simulation. Data products from reanalysis systems (ERA5, GLDAS v2.1, GLEAM v4.2a) were utilized.
• Data sources:
  • Soil Moisture: Multi-source average of ERA5 (volumetric soil water layer 1, 2, 3), Global Land Data Assimilation System Version 2.1 (GLDAS v2.1) (SoilMoi010cminst, SoilMoi1040cminst, SoilMoi40100cminst), and Global Land Evaporation Amsterdam Model Version 4.2a (GLEAM v4.2a) (root-zone soil moisture). All datasets cover 2003-2023.
  • NDVI: MODIS MOD13Q1 (16-day temporal resolution, 250 m spatial resolution) for 2003-2023.
  • Temperature: ERA5 (daily temporal resolution, 0.25° × 0.25° spatial resolution).
  • Phenological Observations: 11 Chinese Phenological Observation Network (CPON) sites (2014) for validation.
  • Ecogeographic Zoning: Resource and Environmental Science Data Center of the Chinese Academy of Sciences for regional classification (humid, sub-humid, semi-arid, arid).
  • Methods for Flash Drought Identification: Pentad-mean (5-day) soil moisture percentile, based on Yuan et al. (2023) definition (soil moisture drops from >40% to <20% within four pentads, decline rate >5 percentile points per pentad, event persists at least four pentads).
  • Methods for Phenology Extraction: Savitzky-Golay (SG) filtering, sixth-order polynomial fitting, time-series harmonic analysis, dynamic threshold method (normalized NDVI thresholds of 0.2 and 0.5), rate of change method (first-order derivative of NDVI). Multi-method composite estimate used.
  • Methods for Phenological Response Quantification: SOS anomaly (ASOS) and EOS anomaly (AEOS) comparing drought-affected pixels to mean of non-drought pixels in a 5 × 5 pixel window.
  • Statistical Analysis: Theil-Sen slope method for trends, Mann-Kendall (MK) test for significance (p < 0.05, p < 0.01), Pettitt change point test for abrupt shifts, Partial correlation analysis for drivers (temperature, soil moisture, drought speed, duration, timing).

Main Results

• From 2003 to 2023, the national mean SOS advanced at a rate of 0.323 days per year, with a significant shift detected in 2013.
• The national mean EOS postponed at a rate of 0.146 days per year, with a significant shift detected in 2010.
• 49.1% of pre-SOS flash droughts caused SOS advancement, but the magnitude of this advancement decreased by 0.075 days per year (p < 0.01). This decrease was most pronounced in semi-arid and sub-humid regions.
• SOS delays induced by pre-SOS flash droughts have shortened over time, most pronounced in arid regions.
• 61.5% of pre-EOS flash droughts caused EOS advancement, and 38.5% caused EOS delay.
• The magnitude of EOS advancement in response to pre-EOS flash droughts increased by 0.063 days per year (p < 0.01), particularly in arid and humid zones.
• The magnitude of EOS delay in response to pre-EOS flash droughts also increased by 0.062 days per year (p < 0.01), significantly in the humid zone.
• The declining SOS response is likely driven by slower flash drought onset speeds (though not statistically significant).
• Enhanced EOS responses (both advancement and delay) are mainly attributed to higher drought-period temperatures.
• Earlier pre-SOS flash droughts were more likely to induce significant delays in SOS.
• Pre-EOS flash droughts occurring closer to the EOS were more likely to cause EOS advancement.

Contributions

• Quantified the spatiotemporal patterns and interannual trends of vegetation phenological responses to pre-season flash droughts across China, addressing a critical gap in understanding these rapid-onset events.
• Introduced and validated two metrics (ASOS and AEOS) to quantify phenological anomalies, effectively isolating direct drought impacts from interannual phenological trends.
• Elucidated the dominant drivers of these phenological responses, highlighting the roles of flash drought characteristics (onset speed, timing), climatic factors (temperature), and hydrological conditions (soil moisture).
• Provided new scientific insights into the mechanisms underlying vegetation phenology responses to flash droughts, emphasizing the synergistic effects of multiple factors and their regional variability across hydroclimatic gradients.
• Offered crucial scientific insights for understanding ecosystem adaptation mechanisms under future climate change and for guiding ecological management strategies.

Funding

• National Natural Science Foundation of China (Grant No.: 52370206)
• Science and Technology Project of Shigatse, Tibet (Grant No.: RKZ2025ZD-006)
• Open Funds of the State Key Laboratory of Hydraulics and Mountain River Engineering, Sichuan University (SKHL2303)

Citation

@article{Sun2026Pronounced,
  author = {Sun, Tao and Zhang, Xinhua and Cai, Yujie and Yang, Chun and Gao, Zhurui and Meng, He and Zhang, Jiannan and Guo, Yijun and Sam, Tsetan},
  title = {Pronounced trends in vegetation phenology responses to flash droughts across China (2003−2023)},
  journal = {Ecological Indicators},
  year = {2026},
  doi = {10.1016/j.ecolind.2025.114578},
  url = {https://doi.org/10.1016/j.ecolind.2025.114578}
}