147 NetCDF files (~7.2 MB) (1) CESM info: CESM tag: cesm1.3.beta11 CESM compset = FC5 CESM grid = ne30_ne30 ncol = 48602 lev = 30 CESM machine: cheyenne CESM run type = startup (2) 114 Intel runs: new.avgSpatial_annual.000.nc ... new.avgSpatial_annual.115.nc - Note 046 and 099 are not available due to a runtime error. - Run 000 hass no perturbation, and the rest have round-off level perturbations to the initial temperature specified via CAM namelist parameter "pertlim" - intel 17.0.1 (3) 25 GNU runs: new.avgSpatial_gnu.annual.000.nc .... new.avgSpatial_gnu.annual.024.nc - Run 000 hass no perturbation, and the rest have round-off level perturbations to the initial temperature specified via CAM namelist parameter "pertlim" -gnu 6.3.0 (4) 8 Test runs: -NU (new.avgSpatial_single.t.nu.nc0: set dynamics hyperviscosity (horizontal diffusion) to nu = 9X10^14 (original default 1x10^15) -ALBICE00 (new.avgSpatial_single.t.albice00.nc): changed the albedo of bare ice on glaciers (visible and near-infrared albedos for glacier ice) from 0.80, 0.55 to 0.00, 0.00 (in land module) -CPL-BUG (new.avgSpatial_single.t.cpl-bug.nc):set the albedos to zero above 57N latitude (in coupler code) -EXPAND (new.avgSpatial_single.t.expand.nc): a modification to the preq_hydrostatic subroutine where we expanded the calculation of the variable phi (mathematically equivalent modification) -FMA (new.avgSpatial_single.t.fma.nc): fused multiply-add compiler optimization enabled -NO-OPT (new.avgSpatial_single.t.no-opt.nc): changed the Intel compiler flag to remove optimization (i.e., -O0 instead of -O2) -RAND-MT (new.avgSpatial_single.t.rand-mt.nc): substituted the Mersenne Twister pseudorandom number generator (PRNG) for the default PRNG in radiation modules -RH-MIN-LOW (new.avgSpatial_single.t.rh-min-low.nc): set the min. relative humidity for low clouds to cldfrc_rhminl = 0.85 (original default 0.8975) via CAM namelist parameter Note: More info on these perturbations can be found in: Daniel J. Milroy, Allison H. Baker, Dorit M. Hammerling, John M. Dennis, Sheri A. Mickelson, and Elizabeth R. Jessup, “Towards characterizing the variability of statistically consistent Community Earth System Model simulations.” Procedia Computer Science (ICCS 2016), Vol. 80, pp. 1589-1600, 2016 (doi:10.1016/j.procs.2016.05.489). Daniel J. Milroy, Allison H. Baker, Dorit M. Hammerling, John M. Dennis, Sheri A. Mickelson, and Elizabeth R. Jessup, “Towards characterizing the variability of statistically consistent Community Earth System Model simulations.” Procedia Computer Science (ICCS 2016), Vol. 80, pp. 1589-1600, 2016 (doi:10.1016/j.procs.2016.05.489). --------------- Know data issues: --------------- (a) TSMN and TSMX are incorrect in Intel ensemble member 000 (new.avgSpatial_annual.000.nc) - they are both equivalent to TS (b) H2O2_SRF is unnaturally large in years 48 and 49 for Intel ensemble member 43 (new.avgSpatial_annual.000.nc) - the reason is unknown (c) Intel ensemble 115 (new.avgSpatial_annual.000.nc) has a number of corrupt data values (AODVIS, CCN3, ICWMR, SNOWHICE, VD01, VV, ncl_a1_SRF) in year 95 due to a corrupt monthly file) (d) SNOWHICE (snow depth on ice): these spatial averages do not include ocean grid point values - which are fill values) (e) SNOWHLND (snow depth on land): these spatial averages include ocean values which are all 0.0 (would have been better to mask out the ocean values before averaging) (f) Note that 2 monthly output files for Intel had a Nan value at a single grid point. These 2 Nan values were replaced by their surrounding average before the annual and spatial averages were computed: -Intel Ens. 74, year 23, month 6, at grid point 5431 -Intel Ens. 34, year 98, month 9, at grid point 6245 (g) Note: LANDFRAC, PHIS, SOLIN, EMISCLD are constant across the ensemble (not a bug), and EMISCLD is all zeros