Selected daily mean output of decadal prediction system based on MPI-ESM-LR 1.2.01p5, members 17 to 80 (together with standard members 1 to 16 they form the large ensemble set). The daily output of members 1 to 16 (standard ensemble set) has been published in a separate data set 'MPI-ESM-LR 1.2.01 decadal prediction localEnKF daily'. Monthly data can be found in corresponding data sets 'MPI-ESM-LR 1.2.01 decadal prediction localEnKF monthly' and 'MPI-ESM-LR 1.2.01 decadal prediction localEnKF LE monthly'. Data from the 3-hourly output for members r17 to r80 can be found in 'MPI-ESM-LR 1.2.01 decadal prediction localEnKF LE 3hrly atmos surface'.
Hindcasts dkfen4YYYY, members 17 to 80 started November 1st each year 1960-2019, running for 10 years. Initial fields are taken directly from members 1 to 16 of the assimilation, with a perturbation added to the horizontal diffusion at upper stratospheric levels added (ATMO_DISTURBANCE) in the first two months of simulation to members 17 to 80 in chunks of 16 members each.
The output of assimilation asSEIKERAf with 16 members is contained in data sets 'MPI-ESM-LR 1.2.01 decadal prediction localEnKF monthly' and MPI-ESM-LR 1.2.01 decadal prediction localEnKF daily'. The assimilation consists of atmospheric nudging to monthly temperature, vorticity, divergence, sea level pressure from ERA40/ERAInterim/ERA5 and a 16-member oceanic localized ensemble Kalman filter (EnKF) assimilating monthly salinity and temperature profiles from EN4. The EnKF has been implemented offline as a restart wrapper with the parallel data assimilation framework PDAF, developed and maintained by Lars Nerger.
All simulations with CMIP6 external forcing.
Simulations have been started to run during phase II of MiKlip 2018 and have been updated to CMIP6 forcing in 2020/2021 by Sebastian Brune, David Nielsen, Laura Hövel and Vimal Koul.
Output format is close to but not totally matching CMOR standard.