Zhou et al. (2026) Reconstruction of temperature, precipitation, and identification of extreme climate events in high mountain Asia over 500 years using multi-method EnKF

Identification

• Journal: Scientific Reports
• Year: 2026
• Date: 2026-01-17
• Authors: Jun Zhou, F. Chen, Yi Zhu, Fuming Xie, Caixia Qin, Hua Tian, Yiyuan Shen, Xia Yang, Yunpeng Duan, Muhammad Mannan Afzal, Shiyin LIU
• DOI: 10.1038/s41598-026-36469-3

Research Groups

• Yunnan Key Laboratory of International Rivers and Transboundary Eco-Security, Yunnan University
• Institute of International Rivers and Eco-security, Yunnan University
• Southwest United Graduate School, Kunming

Short Summary

This study reconstructed gridded temperature and precipitation fields for High Mountain Asia (HMA) over 500 years (1501-2000 CE) using multi-method Ensemble Kalman Filter data assimilation and two proxy datasets, providing a quantitative baseline for contextualizing recent climate changes and identifying extreme events.

Objective

• To reconstruct temperature and precipitation variability across High Mountain Asia (HMA) over the past five centuries (1501-2000 CE) using three distinct paleoclimate data assimilation methods (OEnKF, AOEnKF, HGAOEnKF) and two proxy datasets (Pages2k and an expanded Combined dataset).
• To systematically compare how reconstruction performance is influenced by different assimilation methodologies and proxy datasets at an HMA-wide scale.
• To conduct a comprehensive identification and characterization of extreme climate events over the past 500 years, analyzing their temporal evolution, spatial patterns, and compound nature.

Study Configuration

• Spatial Scale: High Mountain Asia (HMA), defined as 21°N–51°N, 62°E–110°E.
• Temporal Scale: 500 years (1501-2000 CE) for reconstructions; 1951-2000 CE for instrumental validation; 1941-2000 CE for proxy system model calibration.

Methodology and Data

• Models used:
  • Offline Ensemble Kalman Filter (OEnKF)
  • Analog Offline EnKF (AOEnKF)
  • Hybrid Gain Analog Offline EnKF (HGAOEnKF)
  • Prior ensemble derived from the Community Earth System Model-Last Millennium Ensemble (CESM-LME) project.
  • Statistical Proxy System Models (PSMs) based on linear regression.
• Data sources:
  • Proxy datasets: Pages2k dataset (201 records, primarily tree-rings, 1 ice core); Expanded Combined dataset (297 records, 290 tree-rings, 5 ice cores, 2 speleothems).
  • Instrumental reanalyses (for validation): Berkeley Earth, Hadley Centre Climatic Research Unit Temperature version 5 (HadCRUT5), Climatic Research Unit Time Series (CRU TS), Goddard Institute for Space Studies Surface Temperature Analysis (GISTEMP), Twentieth Century Reanalysis version 3 dataset (20CRv3).
  • Calibration data: Berkeley Earth gridded dataset (temperature, 1°×1° resolution); Global Precipitation Climatology Centre (GPCC) data (precipitation, 0.5°×0.5° resolution).

Main Results

• Temperature reconstructions showed skillful performance (correlation coefficient, r = 0.63-0.70) and precipitation reconstructions showed modest skill (r = 0.21-0.26) in proxy-based cross-validation.
• The six reconstructions exhibited high internal consistency (mean inter-reconstruction correlations: 0.93 for temperature, 0.96 for precipitation) and minimal inter-method differences (average 0.04 K for temperature, 0.01 mm/day for precipitation).
• Expanding the proxy network from Pages2k to the Combined dataset enhanced extreme event detection, identifying 39% more extreme cold years and 21% more extreme warm years.
• The reconstructions captured the transition from Little Ice Age (LIA) cooling to modern warming, with the LIA exhibiting multi-phase cooling (prolonged cold period 1620-1680 CE, most extreme cooling 1810-1840 CE at approximately -0.6 K).
• Precipitation showed multi-decadal oscillations rather than a persistent trend, with a notable sustained moistening in the late 20th century (1980-2000 CE).
• Systematic extreme event identification revealed that cold and warm years accounted for 11.0% and 8.6% of the 500-year record, respectively.
• Three major extreme events were identified:
  • The 1641-1644 CE cold event: A temperature-dominated extreme with 80-100% spatial coverage and mean cooling of -0.68 K, linked to volcanic forcing.
  • The 1817-1820 CE compound cold-dry extreme: The most severe multi-year event, characterized by -0.64 K cooling and a -0.13 mm/day precipitation deficit, also linked to volcanic forcing (Tambora eruption).
  • The 1994-2000 CE warm-wet episode: The longest sustained warming in the 500-year record, with +0.83 K warming and a +0.14 mm/day precipitation surplus.

Contributions

• Provides the first systematic comparison of how different paleoclimate data assimilation methodologies and expanded proxy datasets influence reconstruction performance across High Mountain Asia.
• Presents the first comprehensive identification and characterization of extreme climate events (timing, duration, spatial patterns, and compound nature) for HMA over the past 500 years.
• Generates 500-year gridded temperature and precipitation fields for HMA, offering a crucial quantitative baseline to contextualize recent climate changes within pre-industrial natural variability.
• Addresses previous limitations in HMA paleoclimate research, such as spatially fragmented reconstructions, focus on single variables, and reliance on single reconstruction methods or limited proxy datasets.

Funding

• National Science Foundation of China (grant number 42171129 and 42361144874)
• National Key Research and Development Program of China (grant number 2024YFC3013400)
• Scientific Research Fund Project of Yunnan Education Department (grant number 2025Y0071)
• Yunnan University Postgraduate Scientific Research Innovation Fund Project (grant number KC-24248782)

Citation

@article{Zhou2026Reconstruction,
  author = {Zhou, Jun and Chen, F. and Zhu, Yi and Xie, Fuming and Qin, Caixia and Tian, Hua and Shen, Yiyuan and Yang, Xia and Duan, Yunpeng and Afzal, Muhammad Mannan and LIU, Shiyin},
  title = {Reconstruction of temperature, precipitation, and identification of extreme climate events in high mountain Asia over 500 years using multi-method EnKF},
  journal = {Scientific Reports},
  year = {2026},
  doi = {10.1038/s41598-026-36469-3},
  url = {https://doi.org/10.1038/s41598-026-36469-3}
}