CMEP logo

Methodology

Back to Diagnostics Plots

Data Processing/Plotting Methods

  • Albedo is computed from the ratio of the average fluxes averaged across the region in question rather than from the average albedo at each point. For example, the global mean albedo is computed from the ratio of the global mean reflected SW flux to the global mean solar insolation rather than from the global mean albedo computed at each gridpoint. {The ratio is always computed in the last step of the derivation.}
  • The PD-PI increment is applied to all comparisons of preindustrial control runs with observations and is defined as the difference in the monthly mean climatology as a f(lon,lat) during the time span of observations in the CESM-LE ensemble mean minus the mean of years 1000-1499 in the CESM-LE 1850 control run. All observations shown (generally in the middle left panels) have been adjusted using the PD-PI increment.
  • Stippling is used to indicate regions where the model departure from either observations or the reference model exceeds 2-standard deviations of internal variability based on the CESM-LE.
  • Hatching indications regions where it exceeds that amount plus observational uncertainty.
  • Observational uncertainty is estimated based on expert guidance from community experts and data producers (e.g. the CERES science team). For some fields, uncertainty is best characterized in percentage terms while for other fields, particularly those where both positive and negative values are possible, absolute uncertainty estimates are preferable.
  • Observational uncertainty is summarized as in: SOLIN (0.1%), ALBEDO (1%), ALBDS (3%), FSNTOA (5%), FLNT (5%), FLUT (5%), RT (15 W/m2), FS (26 W/m2), FSDS (15 W/m2), FSNS (15 W/m2), FLDS (5%), FLNS (15 W/m2), PR (5%), LHFLX (15 W/m2), SHFLX (5 W/m2), TMQ (5%), PSL (200 hPa), Z500 (100m).
  • Regional acronyms includes global mean (GM), northern (NHM) and southern hemisphere means (SHM), ocean (OM) and land means (LM). Area-averages are always based on cosine of latitude weighting.
  • Climatological zonal means are averaged over the globe and correspond to ocean (blue), land (red), and all surfaces (black).
  • For visual clarity, the Y-axis may differ across individual panels on a particular plot.

On Observations


On Plot Layouts

  • Timeseries Plots: Volcanic and other forcing does not include cloud interactions and is estimated from anomalies in clear-sky net top of atmosphere SW radiation (FSNTC) from 30N-30S over ocean with a 12-month running mean applied, baselined by removing the maximum value of the timeseries. This approach eliminates the influence of sea ice on albedo but remains slightly sensitive to changes in water vapor absorption. All smoothing is 'centered' (hence the flat lines evident near endpoints). Shading for the LENS and LME correspond to 1-sigma ranges.
  • Two-D Map Plots: Stippling indicates significant differences at the 95% level from internal variability while hatching indicates exceedance of this and observational uncertainty. Zonal means are for land+ocean (black), land (red), and ocean (blue).

On Model Variables, Abbreviation, Formulation, and Units

  • TS is surface ("skin") temperature (K) and is used in lieu of sea surface temperatures (SSTs), TREFHT is 2m air temperature (K). FSNTOA is net TOA shortwave flux (positive) downward (W/m2) while FLNT is net TOM longwave flux (positive upward). RT is net TOM flux (positive downward; TOM is used as a TOA LW field does not exist) while FS is the net surface flux (positive upward, FS = FLNS-FSNS+LHFLX+SHFLX+LH(snowfall over ocean). RTFS is the sum of RT (positive down) and FS (postive up) and is equal to the sum of atmospheric divergence and tendency.
  • SWCF is shortwave cloud forcing (W/m2, +=warming), LWCF (W/m2, +=warming) and NETCF is net cloud forcing (W/m2, +=warming).
  • PR is total precipitation (CESM equivalent = PRECC+PRECL, mm/day), TMQ is the total precipitable water (mm), LHFLX is the latent heat flux (W/m2), and PSL is sea level pressure (Pa). EP (mm/day) is the difference between evaporation and precipitation (mm/day) and RELHUM992 is the lowest atmospheric model layer relative humidity (%).
  • U10 is the 10-m wind speed (scalar, m/s) and Z500 is the eddy geopotential height at 500 mb (zonal mean removed, units m). OMEGA500 is the 500 mb vertical velocity (hpa/s).
  • Other variables include solar incoming radiation (SOLIN, W/m2), top of atmosphere albedo (ALBEDOO, fraction), surface albedo (ALBDS, fraction), net TOA flux (RT, W/m2), surface downwelling shortwave (FSDS, W/m2) and longwave (FLDS, W/m2) fluxes, and the net downward shortwave (FSNS, W/m2) and net upward longwave (FLNS, W/m2)

FLAGS are used to highlight key discrepancies from observations.

Flags address violations of basic expectations including values that are out of range for global means, NH-SH and OM-LM contrasts in:
  1. FSNT
  2. FLNT
  3. Precipitation
  4. Latent Heat Flux
  5. SW and LW Cloud Forcing
  6. Precipitable Water
and Ocean Heat Content Drift / Trends.

Improvements and Degradations

Improvements and degradations are used to assess notable changes in spatial patterns relative to the reference model (in pattern correlations with observations) exceeding |0.05| for annual means and seasonal differences, and |0.1| for ENSO teleconnections. Based on the LENS, these choices were found to be conservative, exceeding a 2-sigma threshold for all variables. Improvements correspond to increases in the pattern correlation while degradations correspond to reductions.

CMAT Version 1.0