CMIP6-MedPlus: downscaled climate projections for the Mediterranean region
Domaine:
climategeospatialType de record:
The CMIP6-MedPlus dataset provides CMIP6 downscaled daily precipitation and mean, minimum, and maximum temperature at 0.25° resolution for an extended Mediterranean domain over the period 1986–2100. The area ranges from 18.5°W to 45.25°E in longitude and 18.5°N to 53.25°N in latitude, covering southern and central Europe, North Africa, and portions of the Middle East. The dataset derives from statistical downscaling of an ensemble of CMIP6 climate models (Table 1) from the ScenarioMIP experiment. For each model, two Shared Socioeconomic Pathways (SSPs) have been considered: SSP1-2.6 and SSP3-7.0.
The climate data archive has been developed through a dedicated statistical downscaling framework, which comprises a two-stage procedure. First, a Convolutional Neural Network (CNN) was trained to spatially downscale the coarse-resolution climate fields, using ERA5 data as the high-resolution reference to guide the learning process. Subsequently, the CNN outputs were bias-adjusted using the Quantile Delta Mapping (QDM) method, which preserves changes in the distribution while correcting systematic model biases, with ERA5 serving as the reference dataset.
This work was supported by OurMED PRIMA Program project funded by the European Union’s Horizon 2020 research and innovation under grant agreement No. 2222 (https://www.ourmed.eu/).
Table 1. List of the selected GCMs.
| Model | GCM institute |
| EC-Earth3-Veg | EC-Earth-Consortium |
| GFDL-ESM4 | NOAA-GFDL |
| MPI-ESM1-2-HR | DKRZ |
| MRI-ESM2-0 | MRI |
| NorESM2-MM | NCC |
Dataset structure
The dataset contains four climate variables:
pr – daily precipitation
tas – daily mean temperature
tasmax – daily maximum temperature
tasmin – daily minimum temperature
For each variable, there are five NetCDF (.nc) files, each containing downscaled projections from one of the climate models included in the multi-model ensemble. The files follow the naming convention {Variable}_{ModelName}. For example, the precipitation data produced using the EC-Earth3-Veg model are stored in pr_EC-Earth3-Veg.nc.
Each NetCDF file provides daily climate projection data over the extended Mediterranean region at a spatial resolution of 0.25°. The file contains gridded values for the selected variable covering both the historical period (1985–2014) and the future period (2015–2100) under the two SSP scenarios. The main components of the file are organized as follows:
Dimensions
· lat (140): indicates 140 latitude points ranging from 53.25°N to 18.5°N, representing the north–south extent of the domain.
· lon (256): indicates 256 longitude points ranging from −18.5°E to 45.25°E, representing the west–east extent of the domain.
· lat_bnds (141) and lon_bnds (257): define the boundaries of the latitude and longitude grid cells.
· time_hist (10957): daily timestamps for the historical period (1985-01-01 to 2014-12-31), stored as days since 1850-01-01.
· time_ssp (31411): daily timestamps for the future period (2015-01-01 to 2100-12-31), also stored as days since 1850-01-01.
Coordinates
· lat and lon: latitude and longitude values representing the centre of each grid cell.
· lat_bnds and lon_bnds: latitudinal and longitudinal bounds defining the edges of each grid cell.
· time_hist and time_ssp: time coordinates for historical and future periods, expressed as days since 1850-01-01.
Data variables
· {variable}_hist (time_hist, lat, lon): daily values of the selected variable in the historical period.
· {variable}_ssp126 (time_ssp, lat, lon): daily projected data under the SSP1-2.6 scenario.
· {variable}_ssp370 (time_ssp, lat, lon): daily projected data under the SSP3-7.0 scenario.