Regional scenarios for climate change impacts assessment require finer spatial scales than those provided by global climate models (coarse resolution, about 300x300km). The ENSEMBLES European project (especially the RT3) provides improved regional models, at spatial scales of 50 km and 25 km square, for both recent past (1961-2000) and future climate scenarios (1950-2050).

Here we retained models in which the spatial resolution is 25km square and covers the European domain. The grid spans 29.125°N to 71.375°N in latitude and 23.875°W to 45.375°E in longitude, with a regular 0.25° step (170x278 points in latitude per longitude). It’s the same resolution than the observed grid, but it spans a smaller area.

Fig 3D relief representation of the topography (RCM common 25x25km grid)

There are two types of simulations:

1)      The Control experiments (ERA40 hereafter): The limited area model (or RCM) is forced at its boundaries by the ERA40 reanalysis (best guess of the observations, which is a mix of simulation and ground measurements, satellite data, boat measurements through complex assimilation mathematical process). Consistent (realistic) external forcing is also applied to the models (greenhouses gases, solar, volcanic, aerosols). All the included models have been run to with the most realistic boundary conditions and external forcing. The experiments span a common period from 1961 to 2000.

2)      The scenario experiments (SRESA1B hereafter). The RCM is forced at its boundaries by a general circulation model (or GCM with a coarser resolution, about 300x300km) forced by the SRESA1B emission scenario

The GCM that has been used to drive the limited area model can vary from an operational centre to another. For instance, the CNRM i.e. the Météo France research centre used its own GCM (ARPEGE-OPA) and so did the Met office for the UK (hadCM3). That can partially explain the spread in the different scenarios with respect to the control run (in which all the RCM are driven by the reanalysis). and provides a better estimate of the uncertainties. The selected common time period spans 1951-2050.

3)      More details about the experiment and the different models set-up are available on the RT3 website

4)      Here is a list of the “actual”  subset of retained models (considering a common grid, and runs available for both the recent climate and the future scenarios...):

C4IRCA3 (MettEire, Ireland), CNRM-RM4.5 (CNRM, Meteo-France), DMI-HIRAM5 (DMI, Denmark),  ETHZ-CLM (ETHZ, Switzerland), ICTP-RegCM3 (ICTP, Italy), KNMI-RACMO2 (KNMI, Netherlands), METNOHIRAM (MET.NO, POLAND) METO-HC(Met Office, UK), MPI-M-REMO (MPI, Germany), OURANOSMRCC (OURANOS, Canada), SMHIRCA (SMHI, Sweden), UCLM-PROMES (UCLM, Spain).