Global Bias-Corrected CORDEX Dataset at Quarter Degree Resolution

Area of Science:

• Climate Science
• Environmental Data Science

Background:

• Existing Coordinated Regional Downscaling Experiment (CORDEX) products suffer from inconsistent methods, varied references, and limited coverage, hindering regional climate analysis.
• Comparability across different CORDEX domains is challenging due to methodological heterogeneity and diverse observational datasets.

Purpose of the Study:

• Introduce GloBCORD-QD, a novel quasi-global, high-resolution climate dataset.
• Provide a consistent framework for historical and future climate projections to enhance climate impact assessments.

Main Methods:

• Generated a quasi-global dataset at 0.25° resolution using the ISIMIP3BASD quantile mapping approach.
• Applied trend-preserving bias correction to CORDEX-CORE simulations from four Earth System Models.
• Utilized CHELSA-W5E5 as a unified observational reference for five key climate variables.

Main Results:

• GloBCORD-QD significantly reduces systematic biases and improves distributional performance (Perkins Skill Score from 0.66 to 0.97).
• The dataset successfully preserves climate change signals across diverse variables, regions, and emission scenarios (RCP2.6, RCP8.5).
• Integration of multiple CORDEX domains into a unified framework enables robust spatial comparisons while retaining regional detail.

Conclusions:

• GloBCORD-QD provides a methodologically consistent, high-resolution dataset crucial for robust and comparable climate impact assessments.
• The dataset supports improved uncertainty quantification in climate projections, aiding evidence-based adaptation and policy development.
• Facilitates cross-regional and cross-sectoral climate impact studies by offering a unified quasi-global perspective.