README file for CAVE Clouds & Radiation Swath (CRS) data files.

Summary:
--------

The Surface and Atmospheric Radiation Budget (SARB) working group of 
the CERES science team runs a radiation transfer model beneath CERES 
observations. The CERES data product, which supplies the footprint 
resolution, instantaneous results, including radiation transfer results 
is called the Clouds & Radiative Swath (CRS) data product. For the CAVE 
data set we have extracted the data records (one record per footprint) 
for all CERES footprints within 25km of each CAVE surface site.

File Format:
-----------

Data are real, direct access, 8-bit binary, of record length 2060 bytes.
They are written using fortran 90 code on an Unix-SGI (i.e. Big-Endian)
Each record contains 515 variables for each CERES observation. Each 
record is made up of 3 parts and represents a large amount of information 
for the CERES footprints. The number of footprints in any given file is 
not constant but a function of the site latitude and the orbital 
characteristics of each satellite.

  1 - "SSF" portion: This first set of 290 variables contains satellite
            ephemeris data, CERES Top of Atmosphere radiance and flux 
            observations and statistics dealing with the properties of 
            clouds observed within the footprint. CERES collects MODIS 
            imager pixel radiances collocated inside the field of view 
            of the CERES instrument. It then weights the imager pixels 
            by the CERES energy distribution function (Point spread function) 
            and derives the properties of the clouds viewed within the field 
            of view of the CERES instrument. This cloud information comes 
            from the CERES team using weighted MODIS pixels, it is not a 
            MODIS data product.

  2 - "CRS" portion: Variables 292 through 454 in the record document the
            SARB radiation transfer results for a footprint. The  radiation 
            transfer results are output at 5 atmospheric pressure levels, 
            though the model is run at much higher vertical resolution for 
            4 scenarios list here:
 
               'Pristine': No aerosol or clouds
               'Clear'   : Include aerosol, no cloud
               'Cloud no Aerosol': include clouds, exclude aerosols
               'All Sky' : complete run including clouds and aerosols

            After these first four scenarios are run the model TOA results 
            are compared to the CERES observed TOA fluxes and radiances. 
            Depending on scene conditions, various inputs are carefully 
            adjusted and the model is run again. This provides constrained 
            (or "tuned") model calculations and it is these tuned values, 
            along with the amount by which they were adjusted, which are 
            reported in the CRS files. The majority of the 163 variables in 
            the CRS portion of the CAVE files is made up of these tuned 
            fluxes, the adjusment made to the original (untuned) fluxes, 
            and the initial and adjustments to the variables potentially
            changed by the tuning process. (The algorithm allows for only
            ONE itteration in the tuning process.)

  3 - "Derived Variables ("DER"): Variables 455 through 515 are derived for 
            these CAVE files and NOT on the official CRS data product which 
            one might download from a data center. They consist mostly of 
            the "Untuned" or original flux calculations from the radiation 
            transfer model along with TOA and surface albedos. All of these 
            variables were derived fromthe preceeding data in the first two
            sections of the footprint record.

  For a more complete description of these files see:

     http://www-cave.larc.nasa.gov/cave/pages/crs.html
 

Basic Data Set:
---------------

The list of 515 variables given for each CERES footprint (1 record per 
footprint), in these CAVE CRS files is given in the readcrs.f90 or
readcrs.pro programs, no need to repeat it here.