Storto et al. (2026) Evaluation of a coupled regional reanalysis for the Mediterranean region covering the period 1993–2024

Identification

• Journal: Ocean science
• Year: 2026
• Date: 2026-09-14
• Authors: Andrea Storto, Vincenzo de Toma, Chunxue Yang
• DOI: 10.5194/os-22-2809-2026

Research Groups

• Institute of Marine Science (ISMAR), National Research Council (CNR), Rome, Italy
• National Research Center for High Performance Computing, Big Data and Quantum Computing (ICSC), Bologna, Italy

Short Summary

This study introduces MESMAR-R, a new coupled atmosphere-ocean-hydrology regional reanalysis for the Mediterranean (1993–2024), providing a dynamically consistent reconstruction of past climate. It reveals the region's warming, drying, and changing patterns over three decades, crucial for understanding climate change impacts and extreme events.

Objective

• To introduce and evaluate MESMAR-R, a new high-resolution coupled regional atmosphere-ocean-hydrology reanalysis for the Mediterranean region covering 1993–2024.
• To assess the performance of MESMAR-R against observations and other state-of-the-art reanalyses.
• To investigate the relative impact of data assimilation and interactive coupling on reanalysis skill.
• To analyze long-term oceanic and atmospheric climate trends in the Mediterranean region using MESMAR-R.

Study Configuration

• Spatial Scale: Mediterranean region. Atmospheric grid: approximately 15 kilometers horizontal resolution, 41 hybrid vertical levels. Ocean grid: approximately 7 kilometers horizontal resolution, 72 vertical levels. Hydrology model: 0.5 degree resolution.
• Temporal Scale: 1993–2024 (32 years) for the main reanalysis; 2000–2020 for sensitivity experiments.

Methodology and Data

• Models used:
  • MESMAR (Mediterranean Earth System Model at ISMAR) coupled system.
  • Atmosphere: Weather Research and Forecasting (WRF) model (v4.3.3).
  • Ocean: Nucleus for European Modelling of the Ocean (NEMO) model (v4.0.7).
  • Hydrology: Hydrological Discharge (HD) model (v5.1).
  • Coupler: OASIS3-MCT (v5.0).
  • Ocean Data Assimilation: Incremental three-dimensional variational (3DVAR) scheme.
  • Atmospheric Data Assimilation: Scale-selective spectral nudging towards ERA5.
• Data sources:
  • Observations assimilated (Ocean): Subsurface profiles (XBT, CTD, moored instruments, Argo/glider floats) from UK Met Office EN4 dataset; along-track altimetry data from CMEMS; weak surface relaxation towards CNR-ISMAR Sea Surface Temperature (SST) and EN4 Sea Surface Salinity (SSS).
  • Forcing/Boundary Conditions: Lateral atmospheric forcing from ERA5 reanalysis; ocean boundaries forced by ECMWF ORAS5 reanalysis.
  • Validation Data: Gridded SST, SSS, and Absolute Dynamic Topography (ADT) from Copernicus Marine Service; in-situ profiles from UK MetOffice EN4 dataset; e-OBS gridded near-surface atmospheric variables; Mixed Layer Depth (MLD) data from SeaDataCloud; Ocean Heat Content (OHC) from Copernicus Marine Service Ocean Monitor Indicator; precipitation over ocean from Global Precipitation Climatology Project (GPCP).
  • Reference Reanalyses for Comparison: MEDREA24 and GLORYS12 for ocean; ERA5 and CERRA for atmosphere.

Main Results

• Oceanic Performance: MESMAR-R demonstrates strong skill at the sea surface, achieving the lowest Root Mean Square Error (RMSE) for sea surface height and sea surface salinity, and comparable SST RMSE to regional competitors, outperforming global reanalyses. It exhibits the smallest fresh bias at the sea surface.
• Atmospheric Performance: Near-surface atmospheric variables in MESMAR-R generally show slightly higher RMSE than ERA5 and CERRA, but remain within the same order of magnitude, with localized improvements in dynamically challenging areas.
• Impact of Coupling and Data Assimilation: Data assimilation in both ocean and atmosphere is crucial for optimal coupled performance. Coupling alone (without data assimilation) leads to rapid degradation of subsurface structure, while coupling combined with data assimilation improves surface metrics, particularly salinity, and enhances the representation of SST and Sea Level Anomaly (SLA). Ocean data assimilation is the primary control on basin-scale OHC evolution.
• Climate Trends (1993–2024):
  • Ocean: A strong basin-wide halosteric sea level decrease of approximately -3.87 millimeters per year, consistent with Mediterranean salinification. Total sea level rise averages 3.06 millimeters per year, predominantly driven by mass changes (2.88 millimeters per year). Ocean heat content in the top 700 meters shows a basin-average warming trend of 1.34 Watts per square meter.
  • Atmosphere: Pronounced near-surface warming is observed in both winter (+0.040 °C per year) and summer (+0.047 °C per year). Winter precipitation shows modest drying in the western/central Mediterranean, while summer exhibits widespread drying despite strong increases in precipitable water (+0.100 % per year).

Contributions

• Introduction of MESMAR-R, the first regional coupled atmosphere-ocean-hydrology reanalysis for the Mediterranean region with data assimilation in both ocean and atmosphere, providing a unique and dynamically consistent reconstruction for 1993–2024.
• Comprehensive evaluation demonstrating MESMAR-R's state-of-the-art performance for the ocean component, particularly at the surface and in the upper layers, and reasonable performance for the atmosphere given its simplified assimilation.
• Quantitative assessment of the essential role of data assimilation in both components for achieving optimal coupled reanalysis skill and maintaining physical consistency.
• Provision of a valuable dataset for investigating complex coupled air-sea processes, compound events, marine extremes, and for training deep-learning emulators in the Mediterranean climate change hotspot.

Funding

• Programme CN00000013
• "National Centre for HPC, Big Data and Quantum Computing" Directorial Decree (grant no. 1031 of 17 June 2022)
• PNRR MUR – M4C2 – Investment 1.4 – "National Centers" Directorial Decree (grant no. 3138 of 16 December 2021)

Citation

@article{Storto2026Evaluation,
  author = {Storto, Andrea and Toma, Vincenzo de and Yang, Chunxue},
  title = {Evaluation of a coupled regional reanalysis for the Mediterranean region covering the period 1993–2024},
  journal = {Ocean science},
  year = {2026},
  doi = {10.5194/os-22-2809-2026},
  url = {https://doi.org/10.5194/os-22-2809-2026}
}