PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
March 21, 2026Environmental Earth Sciences3 citationsOpen Access

Performance assessment of 22 CMIP6 climate models for projecting wind climate and wind power over the Mediterranean Sea

EÇEmre ÇalışırKAKhalid AmaroucheRÇRecep Emre Çakmak

Key Points

  • The objective is to evaluate the performance of 22 CMIP6 climate models in simulating wind climate and estimating wind power in the Mediterranean Sea.
  • Evaluated spatial accuracy using historical wind data from 1985–2014 and near-term projections for 2015–2024.
  • Assessed model performance with metrics like bias and normalized bias compared to ERA5 reanalysis data.
  • Conducted percentile-based analysis to evaluate model accuracy across varying wind conditions.
  • Analyzed near-term wind climate changes under four Shared Socioeconomic Pathways (SSPs).
  • EC-Earth3 identified as the most accurate model with low normalized differences across regions.
  • CNRM-CM6-1-HR overestimates wind power density in higher percentiles in specific basins.
  • Near-term projections show partial agreement between EC-Earth3 and ERA5 under SSP5-8.5, with significant regional discrepancies.

Abstract

This study aims to (1) evaluate the spatial accuracy of 22 CMIP6 climate models in simulating wind climate over the Mediterranean Sea, (2) assess their performance in estimating wind power density across different wind conditions using percentile-based analysis, and (3) examine near-term projected wind climate changes under different SSP scenarios. The models were selected based on the availability of 3-hourly near-surface wind data (10 m height) for historical (1985–2014) and near-term (2015–2024) periods under four SSP scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5) with consistent spatial coverage. The comparison metrics used to compare CMIP6 individual model output against ERA5 reanalysis data are the bias and the normalized bias supported by statistical test to evaluate the related confidence intervals. The best-performing models were identified based on their ability to reproduce both the long-term average and variability of historical wind speeds with low bias and narrow confidence intervals. Furthermore, a percentile-based evaluation (5th to 95th) was conducted to assess model accuracy across the full range of wind conditions, from calm to high-energy periods, ensuring reliable projections of future wind power density distribution changes. The results show statistically significant differences in model performance across sub-basins, with EC-Earth3 emerging as the most accurate and consistent model for most regions, maintaining normalized differences within ± 0.2 for most percentiles (e.g., Aegean Basin St.11, Alboran Basin St.1). EC-Earth3 demonstrates balanced performance across all wind conditions, while CNRM-CM6-1-HR tends to overestimate wind power density at higher percentiles, particularly in the Balearic Basin (St.2, ~ 0.2–0.6) and Tyrrhenian Basin (St.6, ~ 0.6–0.8). Near-term projections (2015–2024) across four SSP scenarios evaluated through linear trend analysis show partial agreement between ERA5 and EC-Earth3, particularly under SSP5-8.5, but significant regional discrepancies elsewhere. This study provides sub-basin-specific guidance for climate model selection in Mediterranean wind resource assessments, supporting renewable energy planning and climate adaptation strategies.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Çalışır et al. (2026) studied this question.

synapsesocial.com/papers/69be35ba6e48c4981c67426fhttps://doi.org/10.1007/s12665-026-12885-6
Ask AI
Helpful
Bookmark
Share
View Full Paper