This dataset contains matched in situ reference and EUSTACE temperature estimates used in the validation of version 1.0 of the EUSTACE Globally gridded clear-sky daily air temperature estimates from satellites product and the version 1.0 EUSTACE Global daily air temperature product. The data are made available to allow traceability of the validation. This dataset has been compiled as part of the European Union Horizon 2020 EUSTACE (EU Surface Temperature for All Corners of Earth) project. Note that due to the licensing conditions of some of the validation datasets this data is made available for non-commercial purposes only. Additionally some of the datasets used in validation of the ice data do not allow onward redistribution, so only the ocean and land data have been included in this dataset.

Data for EUCLEIA consists of the following simulations that have been produced with HadGEM3-A (N216 L85): - two 15 member multi-decadal stochastic physics ensemble simulations (ALL and NAT) spanning the period 1960 – 2013 which form the basis for model evaluation assessments including forecast reliability and modelled statistics of extreme events. - two corresponding large ensemble simulations (ALL and NAT) spanning the period 2014 – 2015 to be used for attribution studies. ALL simulations include historical natural and anthropogenic climate forcings which include GHGs, ozone, aerosols and land use. NAT simulations include only historical natural forcings with GHG, aerosol and land use forcings fixed to 1850 levels. The aerosol prescription is essentially the same as for the HadGEM2-ES historical simulations for CMIP5 with post-2005 values being taken from the RCP4.5 scenario. Sea surface temperature (SST) and sea ice (SIC) lower boundary conditions are derived from the HadISST observational dataset which for the NAT simulations have the multi-model mean of a set of coupled model estimates of the anthropogenic contribution removed. Within each set of simulations ensemble members differ only through the operation of the stochastic physics scheme. The multi-decadal simulations share initialisation from ERA-40 reanalysis at 0000Z Dec 1st 1959. Large ensemble simulations are initialised from the atmospheric state of the corresponding multi-decadal experiment at 0000Z Dec 1st 2013.

The data provided here are the numerical simulation data for the multi-decadal experiment (1960 – 2013 inclusive) for the validation of the upgraded Met Office HadGEM3-A based operational event attribution system for EUCLEIA (European Climate and weather Events: Interpretation and Attribution). Improvements include higher horizontal and vertical resolution (N216 L85) and the latest dynamical core (ENDGame) and land surface model (JULES). External forcings are historical natural variability of solar irradiance and volcanic aerosol optical depth as well as historical anthropogenic prescriptions of GHGs, ozone, aerosols and land use change. SST and SIC lower boundary conditions are provided from the HadISST observational dataset. The experiment comprises a 15 member stochastic physics ensemble using kinetic energy backscatter and randomly perturbed physics schemes. All ensemble members share identical initialisation of the atmospheric state from ERA-40 reanalysis at 0000Z December 1st 1959. Atmospheric data are provided at temporal output resolutions of 3-hourly, 6-hourly, daily and monthly; land data are provided at daily and monthly resolutions.

The data provided here are the numerical simulation data for the natural forcings-only version of the EUCLEIA multi-decadal experiment (1960 – 2013 inclusive) for the validation of the upgraded Met Office HadGEM3-A based operational event attribution system for EUCLEIA (European Climate and weather Events: Interpretation and Attribution). Improvements include higher horizontal and vertical resolution (N216 L85) and the latest dynamical core (ENDGame) and land surface model (JULES). External forcings are restricted to just historical natural variability of solar irradiance and volcanic aerosol optical depth. SST and SIC lower boundary conditions are provided from the HadISST observational dataset minus an estimate of the anthropogenic contribution derived from CMIP5 coupled model simulations. The experiment comprises a 15 member stochastic physics ensemble using kinetic energy backscatter and randomly perturbed physics schemes. All ensemble members share identical initialisation of the atmospheric state from ERA-40 reanalysis at 0000Z December 1st 1959. Atmospheric data are provided at temporal output resolutions of 3-hourly, 6-hourly, daily and monthly; land data are provided at daily and monthly resolutions.