Skip to main content

General Product Info

CERES Input Data Sources

2000-2005 | | | |

Total Solar Irradiance (TSI) Information

The time series of daily total solar irradiance (TSI) presented here for March 1, 2000 onwards is based primarily on Version-15 data from the Solar Radiation and Climate Experiment (SORCE). The time series is provided  here.

Steps outlined below are used to put together the above time series, extended beyond the period covered by SORCE data. The time series is updated on a monthly basis.

  1. SORCE data are used for the period 25Feb2003 to 30Jun2013. The SORCE data stream was interrupted in mid-July 2013 because of a battery problem on the instrument. 
  2. TSI data for the period prior to 25Feb2003 came from Dr. Greg Kopp of LASP, University of Colorado, Boulder, who had extracted it from a composite dataset from the World Radiation Center (WRC), Davos. The file used, “composite_d41_62_0906.dat” was downloaded from the ftp site: Dr. Kopp had offset WRC-Composite time series to match with the SORCE version available at the time (V09). We offset the WRC-Composite time series further to match with SORCE V15 data following the procedure suggested by Dr. Kopp. According to this procedure, the offset between the SORCE time series and another one is determined by comparing the two time series for the period 25Feb2003 to 31Dec2003. 
  3. The period between 30Jun2013 and 31Oct2014 is covered using a composite dataset available from the Royal Meteorological Institute of Belgium (RMIB). This portion of the RMIB time series (01Jul2013 onwards) is also offset to match SORCE V15 data. In this case, the offset is determined over a 5-year period (01Mar2003 – 29Feb2008) as: Mean of SORCE V15 – Mean of RMIB. 
  4. The period between 01Nov2014 and 01April2019 is covered using SORCE V17, which has been operating in hybrid mode beginning in March 5, 2014. This portion of the SORCE time series (01Nov2014 onwards) is also offset to match SORCE V15 data. Similarly, the offset is determined over a 5-year period (01Mar2003 – 29Feb2008) as: Mean of SORCE V15 – Mean of SORCE V17. 
  5. In the period between 01April2019 and 25February2020, SORCE V17 was superseded by V18. A portion of the time series from 01April2019 through 25February2020 of V-18 is also matched to V-15 using an offset determined as: Mean of SORCE V15 – Mean of SORCE V18. An alternative data gap strategy is also employed for this period as described below. 
  6. Beginning on 11January2018, measurements of TSI are being collected from the TIM instrument (Total Irradiance Monitoring) as part of the TSIS-1 instrument that is flying on the International Space Station (Coddington et al., 2017). Anticipating the transition from SORCE to TSIS, an overlap time series analysis was conducted between Jan 11, 2018 and February 25, 2020, the last day SORCE measurements were collected. The overlap analysis used TIM version 1 as a reference and found that an offset provided the best match to SORCE V15 data as defined as: Mean SORCE V15 – Mean TIM V1. Additionally, during the last period of SORCE measurements there were more frequent gaps in the SORCE TSI. Thus, for the period between 01April2019 and 25February2020, the scaled TIM V1 is used to fill gaps first, then linear interpolation. 
  7. On May, 28, 2020, the TSIS team released Version 3 of TIM instrument (Total Irradiance Monitoring) total solar irradiance times series. TIM V3 began on Jan 11, 2018 and continues collecting measurements through present. This version has a large offset from TIM V2. The TIM V2 and V3 release were not discovered until discontinuities were discovered in the data processed according to #6 above. Simultaneously, it was discovered that SORCE released Version 19 in November 2020 which superseded SORCE V18. The overlapping time period from 01April2018 through 25February2020 is used to access differences and results in a updated offset determined between SORCE V19, V18 and V15 and TIM V1 and V3. Thus, between 01April2018 through 25February2020:
    1. The adjustment of SORCE V19 is computed as: Mean SORCE V15 – Mean SORCE V18 (step #5) + (Mean SORCE V18 – Mean of SORCE V19) which yields the resultant: Mean SORCE V15 – Mean SORCE V19.
    2. Additionally, an analysis between TIM V3 and SORCE V19 showed that an offset correction would suffice. Thus, the TIM V3 scaled to SORCE V15 is used to fill missing SORCE values with the offset: Mean SORCE V15 – Mean SORCE V19 (from Step 7a) + (Mean SORCE V19 + Mean TIM V3) which yields the resultant offset: Mean SORCE V15 – Mean TIM V3.
  8. Only TSIS-1 data is available from TIM V3 starting on 26February2020. Thus, the offset as described in #7b is used from this date forward to scaled TIM V3 to SORCE V15.
  9. Finally, all portions described above are appended end-to-end to produce the entire time series shown in Figure 1.

Figure 1: The CERES TSI time series at 1 AU formed by compositing all the data products identified as described in the text.

CERES Footprint

Figure 1: Area coverage by scan modes showing cross-track scan.
Figure 1: Area coverage by scan modes
showing cross-track scan.

The CERES footprints are 25-km in diameter near nadir, so that there are more footprints on the boundary of a region than inside the region. Moreover, as CERES scans away from nadir, the footprints grow such that they are not small compared to the size of the region, and the distance between footprints in the scan direction increases. If the footprints are large compared to the region, as illustrated in Figure 1, overlap of the footprints with each other and with the boundaries of the region complicates the problem of computing regional averages. The selection of particular footprints to use at the boundaries of the region and the correlation of values of overlapping footprints needs to be considered. Because of these problems, improved techniques for computing regional averages have been developed (Hazra et al. 1993). At present, error studies are underway to define the degree of improvement which these methods provide.

CERES Nested 1.0° Processing Grid

Latitude Segment # of zones
in segment
extent (°)
# of
# of regions
in segment
Equator to 45° 90 360 32400
45° to 70° 50 180 9000
70° to 80° 20 90 1800
80° to 89° 18 45 810
Total 180 44012

After processing the CERES nested equal area grid output is transformed into an 360 longitude by 180 latitude equal angle grid. For nested regions greater than 1°, the equal angle regional values are replicated.

CERES Surface Type IDs

The map below shows the Earth subdivided into 18 different scene types. Seventeen were defined by the International Geosphere Biosphere Programme (IGBP) and an 18th (Tundra) was added to differentiate barren regions in high latitudes from barren deserts in the tropics and sub-tropics. The original data set was provided at 1km resolution by the USGS and subsequently degraded into 1/6 degree and 1-degree equal angle maps for use in CERES processing. These data are not intended for the study of land use changes over time but offer a snapshot of the Earth’s surface to provide initial estimates of surface emissivity.

Eighteen surface/biome type categories used by CERES are:

Surf Index Surface Type Surf Index Surface Type
1 Evergreen-Needleleaf-Forest 10 Grassland
2 Evergreen-Broadleaf-Forest 11 Wetland
3 Deciduous-Needleleaf-Forest 12 Cropland
4 Deciduous-Broadleaf-Forest 13 Urban
5 Mixed-Forest 14 Crop-Mosaic
6 Closed-Shrublands 15 Permanent-Snow
7 Open-Shrubland(Desert) 16 Barren/Desert
8 Woody-Savanna 17 Water
9 Savanna 18 Tundra
A complete description of the land surface types as defined by IGBP.

Surface/biome types 1 ‐ 17 correspond to those defined by the International Geosphere-Biosphere Programme (IGBP).

Click on a specific region on the map to display the region information.

Or, enter the specific lat/lon and select "Submit".

Geodetic Zone Weights Information

CERES Ed2.6 and higher products use geodetically weighting to compute global means. This spherical Earth assumption gives the well-known So/4 expression for mean solar irradiance, where So is the instantaneous solar irradiance at the TOA. When a more careful calculation is made by assuming the Earth is an oblate spheroid instead of a sphere, and the annual cycle in the Earth’s declination angle and the Earth-sun distance are taken into account, the division factor becomes 4.0034 instead of 4. Consequently, the mean solar irradiance for geodetic weighting is ~1361/4.0034 = 340.0 W/m2, compared to 1361/4 = 340.3 W/m2 for spherical weighting.

Spherical earth zonal weighting uses the sin(lat1 lat2), where lat1 and lat2 are the boundaries of the zone geodetic weighting assumes an oblate spheroid, with the equatorial radius = 6378.137 km, and the polar radius = 6356.752 km.

A FORTRAN program used to calculate a global mean from zonal means given that the Earth is not a true sphere is provided here.

The CERES geodetic 1.0-deg zonal weights are provided here.

From the National Geospatial-Intelligence Agency,

δA = Cpδy

δy = δφ360 CM

CM = 2πRM

RM = α(1 − ε 2)ω3

ε = [ƒ(2 − ƒ)] 1/2

δy = 2π360 α(1 − ε 2)ω3 δφ

CP = 2πRP

Rp = acosφω
Cp = 2π acosφω
δA = 2π acosφω 2π360 α(1 − ε 2)ω3 δφ
δA = 4π2 α2(1 − ε2)360 cosφω4 δφ

ω = (1 − ε2 sin2 φ1/2

ω4 = (1 − ε2 sin2 φ2

δA = 4π2 α2(1 − ε2)360 cosφ(1 − ε2 sin2 φ2 δφ
ωt = cosφ(1 − ε2 sin2 φ2

Σ cosφ(1 − ε2 sin2 φ2

CERES Instruments in Cross-Track Scan Mode

There are two CERES instrument onboard both the Terra and Aqua satellites. One is typically in cross-track mode and the other in either the RAPS (Rotating Azimuth Plane Scan) or FAPS (Fixed Azimuth Plane Scan) mode. The cross-track instrument is recommended by the CERES Science Team since the spatial distribution of footprints is uniform.

Also over time, the instrument in RAPS mode has increased spectral darkening of the transmissive optics. Click here for examples of instrument scanning spatial sampling. Refer to the Operations Tables for further details on instrument scan modes.

CERES Instrument Temporal Coverage

Spacecraft Instrument(s) Launch Date Start Date End Date
TRMM PFM 11/27/1997 12/27/1997 05/29/2001
Terra FM1 & FM2 12/18/1999 02/25/2000 Still in operation
Aqua FM3 & FM4 05/04/2002 06/19/2002 Still in operation
S-NPP FM5 10/28/2011 01/27/2012 Still in operation
J01 FM6 11/18/2017 01/6/2018 Still in operation

Product Processing Level Definitions

Level 3B:
Level 3 data products that are adjusted within their range of uncertainty so as to satisfy known constraints on the climate system (e.g., consistency between average global net TOA flux imbalance and ocean heat storage).

Level 3:
Data products are the radiative fluxes and cloud properties that are spatially averaged into uniform regional and zonal grids and globally and also temporally averaged into daily, monthly hourly, or monthly means.

Level 2:
Data products are derived geophysical variables at the CERES footprint resolution as the Level 1B source data. They include the Level 1B parameters, along with the retrieved or computed geophysical variables such as radiative broadband fluxes and their associated MODIS cloud properties.

Level 1B:
Data products are processed to sensor units. The BDS product contains CERES footprint filtered broadband radiances, geolocation and viewing geometry, Sun geometry, satellite position and velocity, and all raw engineering and status data from the instrument.