These pages explain how SARB surface maps are utilized to provide surface spectral albedo and emissivity for running the Fu & Liou code operationally.

The Surface and Atmospheric Radiation Budget (SARB) working group will make accurate calculations of atmospheric column heating rates on a global scale. To make such profiles the SARB group will use the delta-two stream (for SW, 2/4 for LW) radiation transfer code developed by Fu & Liou. Shortwave spectral albedos are required in fifteen bands from 0.2 to 5.0 microns. Longwave spectral emissivities are in required in twelve bands between 2850 and 0 cm-1. The spectral albedo and emissivities are determined via a table lookup based on the scene type underlying the CERES footprint. These values will be adjusted due to solar zenith angle, cloud condition and surface conditions within the field-of-view of the instrument. To define the surface of the globe the SARB group will use the 17 scene types defined by the International Geosphere Biosphere Programme (IGBP) plus Tundra, fresh snow, and Ice scene types. These types are assumed known on a 10 minute (1/6 of a degree) equal angle map covering the globe.

Except for water surfaces (which are treated separately) surface spectral properties for the Fu & Liou code are selected based on observed scene type. Imager data from the same satellite (TRMM - VIRS, TERRA & AQUA - MODIS) are collocated inside the CERES footprint and on the CERES scene type map. This determines the percentage of each scene type within the CERES footprint. The imager data is convolved with the CERES point spread function providing an energy weighting function for each scene type. A table lookup determines spectral albedo (emissivity) for each scene type which are then weighted by the scene type percentages from the imager and integrated giving a spectral albedo (emissivity) curve for the entire footprint. If the footprint is clear, a TOA to surface parameterization is used to determine broadband albedo and this is used to adjust the spectral curve up or down such that the spectral integral of the albedo is equal to the observed broadband albedo. If snow is found overlaying land adjustments are made according to the results Betts and Ball, 1997.

For partly cloudy and overcast conditions there is no retreieval surface broadband albedo. To compensate, a history map of the clear sky albedo is maintained. This "most recent history" map supplies a data base from a recently observed albedo can be chosen for correcting the scene dependent spectral albedos. For cloudy conditions between 5% and 50% cloud the historical values will be adjusted to the current solar zenith angle. For 50% to 100% cloud condition, a diffuse angle of 53deg will be used as the solar zenith angle with which to adjust the broadband albedo.

Table 1 shows the CERES scene types and the source of the spectral albedo curve for each type. (Text List of Table 1.) The determination of the spectral emissivities is documented in NASA Technical Paper "Surface Emissivity Maps for Use in Satellite Retrievals of Longwave Radiation". A complete description of the land surface types as defined by IGBP are given below in Appendix A.

Details on Spectral Albedos

From various sources each scene type is assigned a spectral albedo and emissivity. As the spectral intervals defined in literature do not match those required a set of weights were developed using MODTRAN surface insolation to interpolate (extrapolate) spectral albedos to the required limits.
For example, the original Briegleb et al. shortwave spectral intervals are:

   0.2-0.5, 0.5-0.7, 0.7-0.85, 0.85-4.00 microns,

with Fu & Liou shortwave spectral bands:

Band  1        2        3          4          5          6
   0.2-0.7, 0.7-1.3, 1.3-1.80, 1.80-2.50, 2.50-3.50, 3.50-4.00 microns.

Laboratory measurements of individual leaves, and the 2.1um channel reflectances from MODIS, verify that the extrapolation of the 0.85-4.0 micron albedo value into the near IR Fu & Liou bands is inappropriate. To correct this in the SARB spectral curves a set of ratios were calculated based on several plots from Bowker et al.(1985). Two ratios were defined:

  R1 = refl(1.6um)/refl(1.0um)
  R2 = refl(2.1um)/refl(1.0um)

  Fu(3)*R1, Fu(4 through 6)*R2.

Leaf types used to calculate the ratios from Bowker et al. and the IGBP type adjusted are given in the following table. The spectral adjustments for Pine Needles, American Elm and Sycamore were based on the Bowker et al. measurements of living vegetation. A dead or decaying canopy would be expected to have smoother spectral features than are shown in the figures. We have not yet included seasonal variations in canopy reflectance.

     IGBP Type (#)         Bowker et al. Spectral Curve
     -------------         ----------------------------
( 1)Evergreen Needleleaf   Ponderosa Pine Needles (#62) 
( 2)Evergreen Broadleaf    Am. Elm(#55) & Sycamore(#66) (averaged)
( 4)Deciduous Broadleaf    Am. Elm(#55) & Sycamore(#66) (averaged)
( 6)Closed Shrubs          Mesquite(#80 & #81)  	(averaged)

Spectral albedo from the CARE (1998) experiment include Grass and Cropland and an example of the original curve and the interpolated values in the Fu & Liou bands are shown in the figure below.