CERES Parameterized Surface Fluxes

CERES computed flux description

  • CRS and SYN products also provide computed untuned fluxes based on the Fu-Liou radiative transfer model using GFDL-GEOS atmosphere, SMOBA Ozone, MODIS-only (CRS), or MODIS/GEO (SYN) clouds and MODIS aerosols using MATCH aerosol properties.
  • The untuned TOA computed fluxes are also “tuned” to the CERES measured fluxes by adjusting cloud, surface and atmospheric profile properties to provide consistent surface, profile and clouds with CERES observed TOA fluxes and are called tuned computed fluxes. For cloudy skies the cloud optical depth, height and amount are minimally adjusted to achieve maximum TOA flux adjustment. For clear-sky surface properties, humidity, lapse rate and aerosols are adjusted.
  • The tuned fluxes can then be used to associate atmospheric profile, cloud and aerosol changes with radiative flux differences, since the surface, profile and clouds are more consistent with the TOA observed flux than the untuned.
  • The computed fluxes can be compared to other climate model (GCM) output. All of the profile inputs have frozen algorithms to avoid having algorithm differences being interpreted as climate trends.
  • The tuned and untuned fluxes are computed for clear-sky (with aerosols), pristine (clear-sky no aerosol), all-sky-no-aerosol, and all-sky (with aerosol conditions at the surface, 500mb, 200mb and 70mb and TOA profile levels hourly and 1° regional level.
  • PAR and UVA and UVB fluxes are also computed as well as broadband SW, LW and WN fluxes
  • Comparison of CERES surface fluxes summarized here
  • More information on CERES Computed flux can be found in the CRS Data Quality Summary (DQS) link to external site

CERES computed untuned and tuned flux concept

Image: CERES computed untuned and tuned flux concept
  • Minimum cloud property adjustment to achieve maximum flux difference
  • Non iterative linear tuning approach to attain consistent clouds and fluxes

CERES SSF (parameterized) surface flux description

Image: CERES SSF (parameterized) surface flux
  • CERES instrument does not measure surface fluxes directly, but uses parameterization (SSF) or computes them using radiative transfer (CRS/SYN)
  • CERES Model A surface fluxes are only parameterized for clear-sky
  • SW surface flux parameterization rely on CERES-TOA flux, MODIS cloud properties, MATCH aerosol climatology, NCEP SMOBA ozone
  • All LW surface parameterization rely on the GEOS-4 profiles and cloud base height
  • Surface fluxes are parameterized hourly using CERES-only temporal interpolation to derive monthly means
  • (Kratz et al. 2009) validates the footprint surface fluxes
  • Model B and computed CERES surface fluxes are validated here link to external site.

Comparison of SSF, CRS and SYN surface fluxes (1)

  • SSF surface fluxes are parameterized using TOA fluxes and MODIS-only cloud properties for the SW. For LW surface fluxes only the cloud base and GEOS-4 profile is used. Only Model B provides consistent algorithm surface fluxes for both clear-sky and all-sky for both SW and LW. (Kratz et al. 2009)
  • CRS/SYN surface fluxes are computed (untuned and tuned) using using GFDL-GEOS atmosphere, SMOBA Ozone, MODIS-only (CRS), or MODIS/GEO (SYN) clouds and MODIS aerosols using MATCH aerosol properties.

23-surface CAVE site flux and SSF (ModelB) and SYN (tuned and untuned) dataset comparison from seasonal monthly means from 2003 to 2005

Surface Mean SW down
(188 W/m2)
SW up
(42 W/m2)
LW down
(334 W/m2)
LW up
(380 W/m2)
  Bias Sigma Bias Sigma Bias Sigma Bias Sigma
ECMWF -4.9 23.8 -9.5 21.8 -0.4 14.3 -0.9 13.9
ISCCP-FD -1.0 20.6 -15.6 20.5 7.1 20.6 0.3 22.5
SRB -2.9 22.4 -18.4 29.9 -0.9 11.2 -2.7 13.9
Model B 0.5 24.0 -15.7 32.4 -0.5 10.3 -7.6 15.4
Terra untuned 4.4 12.3 -13.1 21.8 -5.2 10.4 -5.6 16.4
Aqua untuned 3.3 9.8 -14.7 21.6 -5.6 10.4 -5.3 16.4
Terra tuned 4.6 12.4 -13.1 21.6 -5.2 10.3 -5.0 15.9
Aqua tuned 3.7 9.9 -14.5 21.6 -5.5 10.4 -4.8 15.9
  • The ISCCP-FD, SRB (observed) and ECMWF (modeled) are other surface datasets
  • The SSF surface fluxes are similar than the other surface datasets.
  • The SYN computed fluxes from either Terra or Aqua are similar and have reduced the SW down RMS error by half. Note very little difference after tuning
  • CAVE surface site fluxes are a combination of BSRN, ARM sites, (

Comparison of SSF, CRS and SYN surface fluxes (2)

23-surface CAVE site hourly flux comparison of MODIS and GEO observed and MODIS-only and MODIS/GEO interpolated for July 2004

Terra Surface
SW down
LW down
(334 W/m2)
  Bias Sigma Bias Sigma
MODIS 3.1 22 -0.9 7.0
GEO 2.2 31 -1.6 5.5
MODIS-interp 2.2 37 -2.0 6.3
MODIS/GEO-interp 2.9 32 -1.4 5.7


  • For SW down the SYN product has the best monthly (diurnal complete) fluxes
  • For SW down the CRS product has the best instantaneous fluxes at the CERES Terra or Aqua observations times
  • There seems to be little difference in the SSF/CRS/SYN LW surface fluxes
  • The SSF surface fluxes are parameterized and have the least input, and therefore probably the most stable over time (no GEO and MATCH aerosol input) and are near real-time. SYN/CRS product have a longer processing lag time

CERES Data Product Information

Page Curator: Elizabeth Heckert
NASA Official: Dr. Norman Loeb
Page Last Modified: 03/27/2014 08:20
divider divider
+ Privacy Policy and Important Notices
+ Multimedia Browser Plug-ins
+ Comments or Questions?