The forecasting chain is based on the 00 UTC, GFS forecasts at 0.5 degree resolution. The chain comprises the hydrostatic model BOLAM, which is driven directly by the global model, and the non-hydrostatic model MOLOCH (horizontal resolution 0.02 degrees), which is nested in cascade using a 1-way nesting procedure. BOLAM run starts at 00 UTC, MOLOCH is nested at 09 UTC. MOLOCH domain is smaller than official DPHASE domain. A 39-h MOLOCH forecast is provided daily. Only a sub-set of TIGGE list is provided (see DS). More information available here: http://www.isac.cnr.it/~dinamica/ Grid description: lat-lon Arakawa C grid. Rotated equidistant grid. DDOM: xfirst: -2.69 yfirst: -1.84 xsize: 340.0 ysize: 290.0 xinc: 0.02 yinc: 0.02 xnpole: -171.0 ynpole: 44.7

The forecasting chain is based on the 18 UTC, ECMWF forecasts at 0.25 degree resolution. The chain comprises the hydrostatic model BOLAM, which is driven directly by the global model, and the non-hydrostatic model MOLOCH (horizontal resolution 0.02 degrees), which is nested in cascade using a 1-way nesting procedure. BOLAM run starts at 18 UTC, MOLOCH is nested at 00 UTC. MOLOCH domain is smaller than official DPHASE domain. A 48-h MOLOCH forecast is provided daily. Only a sub-set of TIGGE list is provided (see DS). More information available here: http://www.isac.cnr.it/~dinamica/ Grid description: lat-lon Arakawa C grid. Rotated equidistant grid. DDOM: xfirst: -3.5 yfirst: -14.0 xsize: 330.0 ysize: 290.0 xinc: 0.021 yinc: 0.02 xnpole: -170.0 ynpole: 32.5

The forecasting chain is based on the 00 UTC, GFS forecasts at 0.5 degree resolution. The chain comprises the hydrostatic model BOLAM, which is driven directly by the global model, and the non-hydrostatic model MOLOCH (horizontal resolution 0.02 degrees), which is nested in cascade using a 1-way nesting procedure. BOLAM run starts at 00 UTC, MOLOCH is nested at 09 UTC. MOLOCH domain is smaller than official DPHASE domain. A 39-h MOLOCH forecast is provided daily. Only a sub-set of TIGGE list is provided (see DS). More information available here: http://www.isac.cnr.it/~dinamica/ Grid description: lat-lon Arakawa C grid. Rotated equidistant grid. DDOM: xfirst: -2.69 yfirst: -1.84 xsize: 340.0 ysize: 290.0 xinc: 0.02 yinc: 0.02 xnpole: -171.0 ynpole: 44.7

The forecasting chain is based on the 18 UTC, ECMWF forecasts at 0.25 degree resolution. The chain comprises the hydrostatic model BOLAM, which is driven directly by the global model, and the non-hydrostatic model MOLOCH (horizontal resolution 0.02 degrees), which is nested in cascade using a 1-way nesting procedure. BOLAM run starts at 18 UTC, MOLOCH is nested at 00 UTC. MOLOCH domain is smaller than official DPHASE domain. A 48-h MOLOCH forecast is provided daily. Only a sub-set of TIGGE list is provided (see DS). More information available here: http://www.isac.cnr.it/~dinamica/ Grid description: lat-lon Arakawa C grid. Rotated equidistant grid. DDOM: xfirst: -3.5 yfirst: -14.0 xsize: 330.0 ysize: 290.0 xinc: 0.021 yinc: 0.02 xnpole: -170.0 ynpole: 32.5

This is the 4000-yr control run with MPI-ESM-CR v. 1.2, carried out with fixed preindustrial CO2 concentrations. It provides the initial conditions for the abrupt 2% CO2 increase forcing and 1% ramp-up ensembles in the same project. Initial conditions are distance by 200 years in order to ensure reasonable decorrelation. The experiment is performed with MPI-ESM model, coarse resolution (CR: T31). The project is aimed as a testbed for the Green’s functions computed via the 2xCO2abrupt experiment. This is a model application of the linear response theory, as described in Lembo et al. 2020 (see references).

MPI-ESM model (T31 resolution) v. 1.2: step forcing experiment with abrupt doubling of CO2 concentrations A step forcing experiment with instantaneous doubling of CO2 concentrations at t=0. The experiment is performed with MPI-ESM model, coarse resolution (CR: T31). It consists of an ensemble with 20 runs over 2000 years, starting from a se top perturbed initial conditions representative of pre-industrial CO2 levels. The experiment is aimed at computing the Green’s function of observable, whose convolution with the time evolution of a forcing allows for prediction of the long-term evolution of the forcing in response to a given forcing. This in the context of the linear response theory (see Lembo et al. 2020).

MPI-ESM model (T31 resolution) v. 1.2: ramp forcing experiment with 1% CO2 concentration increase every year until doubling with respect to pre-industrial values. A ramp forcing experiment with linear increase in CO2 concentrations by 1% every year, starting from t=0: when doubled concentration is achieved, the concentration is kept steady until the end of the run (1000 years). The experiment is performed with MPI-ESM model, coarse resolution (CR: T31). It consists of an ensemble with 20 runs, starting from a set of perturbed initial conditions representative of pre-industrial CO2 levels (same as for the 2xCO2abrupt experiment, see doi:10.26050/WDCC/2xCO2abrupt). The experiment is aimed as a testbed for the Green’s functions computed via the 2xCO2abrupt experiment. This is a model application of the linear response theory, as described in Lembo et al. 2020.