Help for MASTIR

PURPOSE:

   MASTIR is a program that accepts as input MAS TIR (Channels 42-50)
radiance at sensor image data and outputs upwelling radiance at surface for
those channels.  Also required as input are a valid MODTRAN input file,
to define the atmospheric model, and a MAS TIR spectral response file.


OPERATION:

   MASTIR takes as input the 9 Channels of MAS TIR radiance at sensor data
in units of Watts per square meter per steradian per micrometer, and a file
containing the the MAS TIR spectral response functions. This second
file must be a VICAR labelled file consisting of nine lines (one for each TIR
channel) and 900 samples (one for the spectral response at each wavenumber
from 601 to 1500). The spectral response functions must be normalized, i.e.,
the sum of all values on each line must be 1.0.
   The atmospheric model used is defined by a user supplied MODTRAN input
file.  MASTIR uses these parameters to run MODTRAN twice; once with a nadir
looking geometry, and once for the case of viewing at maximum deflection
from nadir.  From these two runs, the atmospheric transmittance, path
radiance, and downwelling sky radiance at surface are calculated for the nine
MAS TIR bandpasses.  The downwelling sky radiance values are placed in the
VICAR label of the output dataset.
   Values of transmittance and path radiance are calculated for each pixel
location in the scene by interpolation, where it is assumed that each
quantity varies linearly with path length.
   The upwelling radiance at surface is then computed by the following 
formula:

        Rad(at sensor) = RAD(upwelling)*transmittance + Rad(path)

It should be noted that the output quantity (upwelling radiance) of MASTIR
differs from the output quantity of TIMSCAL2 (surface emitted radiance). It
it includes both the surface emitted component plus the surface reflected
component.
   The output dataset consists of 9 channels of halfword (16 bit signed
integer) pixels, in units of milliWatts per square meter per steradian per
micrometer.  Note that these units are 1000 times smaller than the input
units.


WRITTEN BY:  Ron Alley         August 27, 1998

COGNIZANT PROGRAMMER:  Ron Alley

REVISION: Revision 1.2 May 21, 1999