Simon et al. (2025) Evaluation of different bias-corrected EURO-CORDEX databases and the expected future changes in precipitation over Hungary

Identification

• Journal: The Science of The Total Environment
• Year: 2025
• Date: 2025-09-16
• Authors: Csilla Simon, Anna Kis, Csaba Zsolt Torma
• DOI: 10.1016/j.scitotenv.2025.180451

Research Groups

• ELTE Eötvös Loránd University, Institute of Geography and Earth Sciences, Department of Meteorology, Budapest, Hungary

Short Summary

This study investigates projected changes in mean precipitation characteristics and extremes over Hungary using raw and bias-corrected EURO-CORDEX simulations, finding that a newly created HuClim-based bias correction performs best and projects increased annual precipitation (up to 30% in highlands) but fewer wet days by the end of the century.

Objective

• To investigate the projected changes of mean precipitation characteristics and precipitation-related climate indices over Hungary for two future time slices (2021–2050, 2070–2099) with respect to the 1976–2005 reference period.
• To analyze how the choice of the reference dataset and different calibration periods affects the projected changes, specifically comparing raw, MESAN-based, FORESEE-HUN, and a newly created HuClim-based bias-corrected EURO-CORDEX datasets.

Study Configuration

• Spatial Scale: Hungary (approximately 93,000 km²), located in Eastern-Central Europe (45.7°–48.7°N, 15.9°–22.9°E), with a horizontal resolution of 0.11° (approximately 12 km).
• Temporal Scale:
  • Reference period: 1976–2005 (30 years)
  • Validation period: 2006–2022 (17 years)
  • Future time slices: 2021–2050 (near future), 2070–2099 (far future)
  • Daily precipitation data.

Methodology and Data

• Models used:
  • Five EURO-CORDEX Regional Climate Models (RCMs): CCLM4-8-17, HIRHAM5, RACMO22E, RCA4, REMO2009.
  • Driven by four Global Climate Models (GCMs): MPI-ESM-LR, EC-EARTH, HadGEM2-ES, CNRM-CM5.
  • Representative Concentration Pathway (RCP) scenarios: RCP4.5 (medium stabilisation) and RCP8.5 (high-end greenhouse gas concentration).
  • Bias-correction methods: Distribution scaling (for BC-MESAN), Percentile-based quantile mapping (for BC-HUCLIM and FORESEE-HUN).
• Data sources:
  • Raw EURO-CORDEX RCM simulations.
  • Three bias-corrected EURO-CORDEX databases:
    • BC-MESAN: Bias-corrected using MESAN reanalysis data (1989–2010 calibration period).
    • FORESEE-HUN: Bias-corrected using HuClim data (1971–2020 baseline period).
    • BC-HUCLIM: Newly created bias-corrected dataset using HuClim data (1976–2005 calibration period).
  • Reference datasets for evaluation and bias-correction:
    • HuClim: Quality-controlled, gridded (0.1° × 0.1°) daily precipitation data for Hungary from 500 stations (1971–2022).
    • MESAN: Operational mesoscale analysis system (approximately 11 km resolution) integrating weather radar, satellite, and ground-based measurements.
  • Precipitation-related climate indices analyzed: RR1 (wet days), R20 (very heavy precipitation days), RX5day (5-day heavy precipitation periods), RX1day (highest daily precipitation sum), R99p (extremely wet days).

Main Results

• Validation (2006–2022): The HuClim-based bias-corrected simulations (BC-HUCLIM) provided the best performance, showing the greatest improvement in correlation coefficients (0.8–0.95) and normalized standard deviation (0.5–1), especially for light rain (RR1). Raw simulations showed significant overestimation (40–55%) in mountainous areas.
• Mean Annual Precipitation:
  • Under RCP4.5, changes remain below 20% in both future time slices, with minimal differences between databases.
  • Under RCP8.5, changes are more robust, with mean annual precipitation projected to increase by 10% in lowlands and up to 30% in mountainous areas (BC-HUCLIM projects 26–30% increase in Northern Mountains).
• Seasonal Precipitation:
  • Winter precipitation is likely to increase by up to 52% (far future, RCP8.5) across all databases.
  • Summer precipitation shows the greatest uncertainty, with most simulations projecting a decrease (2–20%), while some BC-HUCLIM simulations suggest an increase (3–18%) for 2021–2050.
  • Spring and autumn precipitation are expected to increase, with autumn reaching 25–30% increase by the end of the century.
• Climate Indices (Extremes):
  • The annual number of wet days (RR1) is projected to decrease by 1–7 days per year on average by 2070–2099 under RCP8.5.
  • The frequency of very heavy precipitation days (R20) is expected to increase by 3–7 days per year on average by 2070–2099 (BC-HUCLIM).
  • The number of 5-day heavy precipitation periods (RX5day) may increase by 2–6 cases per year.
  • The highest daily precipitation sum (RX1day) is likely to increase by 20–180% according to BC-HUCLIM and FORESEE-HUN, while raw projections show negligible changes.
  • Extremely wet days (R99p) are expected to increase by 20–35% across the country.
• Topographical Effects: The effect of topography on precipitation changes is more pronounced in HuClim-based bias-corrected simulations, particularly for increases in mountainous regions under RCP8.5.

Contributions

• This study is the first for Hungary to systematically compare the impact of different reference datasets and calibration periods on bias-corrected climate change projections for precipitation.
• It introduces and validates a new, high-performing bias-corrected dataset (BC-HUCLIM) for Hungary, demonstrating the superior accuracy achieved by using the local, high-quality HuClim observational dataset as a reference.
• Provides comprehensive, spatially and temporally resolved projections of mean and extreme precipitation changes over Hungary under two key RCP scenarios, highlighting regional vulnerabilities.
• Emphasizes the critical importance of selecting reliable reference data and evaluating multiple RCMs for robust climate change impact assessments.

Funding

• National Research, Development and Innovation Office – NKFIH, project code OTKA FK-142349.
• Hungarian Scientific Research Fund (for Csaba Zsolt Torma).

Citation

@article{Simon2025Evaluation,
  author = {Simon, Csilla and Kis, Anna and Torma, Csaba Zsolt},
  title = {Evaluation of different bias-corrected EURO-CORDEX databases and the expected future changes in precipitation over Hungary},
  journal = {The Science of The Total Environment},
  year = {2025},
  doi = {10.1016/j.scitotenv.2025.180451},
  url = {https://doi.org/10.1016/j.scitotenv.2025.180451}
}