Kim et al. (2025) An integrated approach for characterizing and selecting climate change scenarios based on variability and extremeness

Identification

• Journal: Scientific Reports
• Year: 2025
• Date: 2025-11-20
• Authors: Jaeyoung Kim, Moon‐Hwan Lee, Joong‐Bae Ahn, Seung Beom Seo
• DOI: 10.1038/s41598-025-24707-z

Research Groups

• Water Management Division, Land and Environment Research Group, Korea Environment Institute, Sejong, Republic of Korea
• International School of Urban Sciences, University of Seoul, Seoul, Republic of Korea

Short Summary

This study proposes a novel integrated approach for selecting optimal Global Climate Model (GCM) and Shared Socioeconomic Pathway (SSP) combinations for aquatic environment impact assessments by quantifying and integrating climate change variability and extremeness into a single metric. The method effectively captures the full range of climate scenarios with a minimal number of combinations, demonstrating that expected trends in variability and extremeness under severe warming are not always consistent across all GCMs.

Objective

• To develop a novel integrated approach for selecting optimal GCM-SSP combinations to assess the impact of climate change on the aquatic environment, specifically focusing on the comprehensive spatial and temporal ranges of climate projections, variability, and extremeness.

Study Configuration

• Spatial Scale: Republic of Korea (approximately 100,000 km²), with projected data extracted at 601 sites (Thiessen polygons for rain gauges), downscaled to a 1 km × 1 km resolution.
• Temporal Scale: Historical simulations for 1981–2010 and future projections for 2011–2100. Extreme Climate Indices (ECIs) calculated annually from 2021 to 2100 (80 years). Past daily data for entropy calculation used from 2000 to 2019.

Methodology and Data

• Models used:
  • 7 Global Climate Models (GCMs) from CMIP6: CanESM5, EC-Earth3, GFDL-ESM4, IPSL-CM6A-LR, MIROC6, MPI-ESM1-2-LR, MRI-ESM2-0.
  • 5 Shared Socioeconomic Pathways (SSPs): SSP119, SSP126, SSP245, SSP370, SSP585, resulting in 35 GCM-SSP combinations.
  • Downscaling method: Improved GIS-based regression technique by Eum et al. (2018) to 1 km × 1 km resolution.
  • Scenario selection algorithm: Katsavounidis-Kuo-Zhang (KKZ) algorithm.
  • Quantification methods: Entropy analysis (Shannon entropy, Entropy weight method, Multivariate entropy method) for variability, and frequency analysis for extremeness.
  • Statistical methods: Pearson correlation and Variance Inflation Factor (VIF) for ECI optimization.
• Data sources:
  • Downscaled daily projected data from GCM-SSP combinations for precipitation and air temperature (maximum, minimum).
  • 26 Extreme Climate Indices (ECIs) from ETCCDI, reduced to 8 representative ECIs (Rx5day, R10mm, CDD, CWD, TXx, TNn, WSDI, DTR) after correlation and VIF analysis.
  • Historical daily climate data for representative site selection and ECI optimization.

Main Results

• The integrated approach successfully transformed high-dimensional climate data (ECI type, time, location) into a single metric representing unique variability and extremeness for each GCM-SSP scenario.
• Entropy analysis for variability showed that MPI had the highest mean entropy (2.16), with IPSL-SSP370 exhibiting the maximum entropy and MRI-SSP119 the minimum.
• Extremeness quantification revealed Can5-SSP585 had the highest extremeness score, while MIR6-SSP585 had the lowest.
• The integrated values (variability × extremeness) ranged from 1.35 to 3.54, with Can5-SSP585 (3.54) and GFDL-SSP585 (3.39) showing the highest values, and MIR6-SSP370 (1.35) the lowest.
• While variability and extremeness generally increased under severe global warming scenarios (e.g., SSP585), this trend was not consistently observed across all GCMs and SSPs, highlighting the need to consider all GCM-SSP combinations.
• The KKZ algorithm effectively selected a minimal set of scenarios (e.g., MPI-SSP126 as centroid, Can5-SSP585 as maximum extreme, MIR6-SSP370 as minimum extreme) that captured the full range of integrated variability and extremeness.
• Comparative methods (envelope-based ΔT–ΔP, k-means clustering, PCA-based selection) were less effective in simultaneously capturing both the maximum and minimum extreme conditions of variability and extremeness.

Contributions

• Introduces a novel integrated approach that quantifies and combines climate change variability and extremeness into a single, interpretable metric for GCM-SSP scenario selection.
• Offers a robust methodology for dimensionality reduction of complex, high-dimensional climate projection data across spatial, temporal, and index domains.
• Provides a systematic framework for selecting a minimal yet representative set of GCM-SSP scenarios that span the full range of projected climate impacts, including average and extreme conditions.
• Emphasizes treating each GCM-SSP combination as a distinct scenario, addressing inherent uncertainties and limitations of relying on single GCMs or SSPs.
• The framework is expandable to include other weather variables (e.g., wind speed, solar radiation) and transferable to other environmental domains.

Funding

• Korea Environment Industry & Technology Institute (KEITI) through the Climate Change R&D Project for New Climate Regime funded by the Korea Ministry of Environment (MOE) (RS-2022-KE002152).

Citation

@article{Kim2025integrated,
  author = {Kim, Jaeyoung and Lee, Moon‐Hwan and Ahn, Joong‐Bae and Seo, Seung Beom},
  title = {An integrated approach for characterizing and selecting climate change scenarios based on variability and extremeness},
  journal = {Scientific Reports},
  year = {2025},
  doi = {10.1038/s41598-025-24707-z},
  url = {https://doi.org/10.1038/s41598-025-24707-z}
}