DETAILS ON THE IRSKY ESTIMATORS				November 11, 1994
There are three background and confusion estimators in IRSKY, each 
individually selectable. 

The BACKGROUND ESTIMATOR will help determine the source,
background, and zodiacal intensity in your target region.

The CIRRUS CONFUSION ESTIMATOR will help determine the "confusion
noise" due to cirrus emission in a given aperture.

The GALAXY CONFUSION ESTIMATOR will help determine the "confusion
noise" due to extragalactic sources in a given aperture.



*************************************
Estimating the background intensity:
*************************************
The Background Estimator outputs a table with 8 columns of data and
possible warning messages. 

The background region around your target position may be thought of as
including two components: The zodiacal background intensity
(eg.,MJy/sr), and the combined intensity of all other contributions
(astronomical sources, cirrus, etc.). 

The Zodiacal Background column lists a model value computed from the
smooth model prepared by J. Good.  Details of the model are found in
the ISSA Explanatory Supplement, Wheelock et al.  1994, Appendix G.
The value was computed at a fixed solar elongation angle of 90 degrees
and at the point in the Earth's orbit where the orbit crosses the
ecliptic plane.  When the actual solar elongation angle is different
from 90 degrees and/or the Earth is slightly out of the ecliptic plane
(inclination is 1.5 degrees), the true zodiacal contribution may be
different from the values listed by a factor of as much as two (See
also ISOCAM Observer's Manual, Appendix A).

The zodiacal model was used with its full set of parameters, taking
into account the earth's position and the observational solar
elongation angle, to subtract the zodiacal contribution from the IRAS
survey data to produce the Low Resolution Product (LRP), and the
Infrared Sky Survey Atlas (ISSA).  Data values from these products
appear in their respective columns in the table.

The intensity left after the subtraction of the zodiacal background is
represented in each of the columns: 30' LRP, 4' ISSA, and 30' x 30'
mean.  The 4' ISSA column represents the intensity of the central pixel
in the image, at the target coordinates, with the pixel being 1.5' on a
side, but representing about 4' in actual resolution.  At some
positions the subtraction of the zodiacal model has resulted in regions
of negative fluxes in the ISSA and LRP data.

The 30' x 30' mean is taken from the ISSA data, and all pixels greater
than 5-sigma from the mean are removed before averaging. Thus the
central source and bright point sources in the field do not contribute
to the mean.  Similarly the LRP has been processed to exclude the
contribution of bright discreet sources. Thus the 30' x 30' mean ISSA
values and the 30' LRP values are likely to be similar.  If there are
significant differences between these values, it may indicate that one or the
other of the products had trouble removing the contribution of bright
and/or extended sources.


COBE
Where available (currently from -10 to +10 degrees galactic latitude),
an attempt is made to provide a background estimate using the COBE
data. This estimate may be useful if the observing wavelength falls
outside the IRAS range (7-120 microns).  The algorithm for producing a
background estimate from the COBE data:

  1.  Finds the pixel containing the target position (nearest pixel)
      in the COBE DIRBE database, which provides broadband fluxes over
      the entire ISO range.  The COBE DIRBE beam is 0.7 degrees, so be
      aware that the sky is smoothed on this rather large scale.  This
      will lead to large uncertainties in areas with strong structure
      on smaller scales.

  2.  Logarithmically interpolates between the observed COBE DIRBE fluxes
      to obtain an estimate at the wavelength of interest.  No modelling
      is done (unlike the IRAS-based estimate) because the large beam
      means that disparate source regions are likely to be contributing
      to the flux in a spatially varying manner.

The intensity values provided are merely estimates, or educated
guesses.  They are not predictions of what you will see at the position
and wavelength chosen.  They are meant to guide you in planning
observations, but it is up to you to decide how to use them.


INTERPOLATION TO THE SPECIFIED WAVELENGTH:
The background flux at the wavelength of interest is listed in the 5th
row of the table, and is estimated by decomposing the ISSA 4 point
spectrum at the selected location into the best-fitting combination of
two far-infrared emission models taken from the literature.  The models
were drawn from the emission models representing different physical
regions (cirrus, dark clouds, reflection nebula, stars and HII
regions), given by Desert, 1990
The flux density shown is the sum of the two contributing models'
values at the selected wavelength. With this method, the continuum
background within the range 12 to 100 microns can generally be
estimated to within 0.5 MJy/sr at a given wavelength. Outside this
range, background estimates may not be as robust.


INTERPOLATION WARNING MESSAGES:
(These messages may apply to a product including Zodiacal, LRP, ISSA, COBE)

Warning: "Insufficient [product] flux for accurate interpolation"
Meaning: one or more of the IRAS Band (12, 25, 60, 100 um) fluxes 
are below the three sigma noise level, making interpolation difficult.

Warning: [product] flux too spikey for interpolation 
Meaning: One of the four IRAS bands has over 5 times the flux of any other band,
or the spectrum is double-peaked, making modeling unreliable.

Warning: [product] fluxes not well modeled.
Meaning: The interpolation routine was not able to consistently model 
the IRAS bands themselves with its two component model, so extrapolating
to other bands is unreliable. 

Warning: too few pixels ....
Meaning: There are too few ISSA pixels for an accurate flux estimate.

Warning: "Position outside COBE data set " 
Meaning: The only COBE data currently available between +10 and -10 
degrees from the galactic plane.



REFERENCES:
Desert, F-X., Boulanger, F., and Puget, J., 1990, Interstellar
Dust Models for Extinction and Emission, A & A 237, 215.

Van Buren, D., 1992, Spectral Decomposition of ISSA Fluxes, IPAC Internal
Memorandum.
[contains information on interpolation method used]
This paper is online in PostScript form on IPAC's publicftp machine:
	 ftp.ipac.caltech.edu
in
	/pub/irsky/doc
Look for the file:
	interpolate.txt


Wheelock, S. L. et al., 1994 IRAS Sky Survey Atlas Explanatory 
Supplement, JPL Publication 94-11 (Pasadena:JPL).
[contains information on the attributes of the ISSA images and the 
zodiacal model]

Information on the COBE data is available via anonymous ftp from
	nssdca.gsfc.nasa.gov 
in the COBE directory tree.

****************************
Cirrus confusion computation:
****************************
The ISSA images are of insufficient resolution to directly compute the
confusion due to cirrus structure for typical-sized beams.  Instead, we
use an empirical model of the confusion noise based on the near
infrared brightness of the target area.

An estimate of the brightness of the cirrus emission is made at the
wavelength of interest. To do this, the mean values in each of the four
bands within a 30' square aperture are computed and an interpolation to
the wavelength of interest is performed.  If the background cannot be
determined due to insufficient flux, or a pathological 4-band spectrum,
an attempt is made to extrapolate the background flux from 100 um to
the wavelength lambda by scaling the 100 um flux by F(lambda)/F(100),
where F(.) is the model dust spectrum taken from Desert, 1990.  If the
flux found in the 100 um ISSA image is insignificant (below the noise
level), the program will not calculate the confusion.

Next the parameters necessary for estimating the factor E0 required by
the Gautier et al. formulation are prepared.  To bracket the confusion,
we chose two specific cases from the Gautier work. 
Those are alpha = -2.6 and alpha= -3.2, which bracket the exponent of the 
spatial power spectrum falloff of cirrus structure. The reference
subtraction method assumed is the double aperture chopping method (also
called double position-switching). The beam separations (between source
and reference beams) are preselected to be two beam diameters, but the
user may reset this.  Likewise the selection of the aperture width is
predetermined by the selection of an ISO instrument, or may be set by
the user.

Computation of the confusion noise is based on Gautier, T. N.  et al.
1992, AJ, and Helou and Beichman 1990. 


REFERENCES:
Desert, F-X., Boulanger, F., and Puget, J., 1990, Interstellar
Dust Models for Extinction and Emission, A & A 237, 215.

Gautier, T. N.  et al.  1992, A Calculation of Confusion Noise Due
to Infrared Cirrus, AJ, 103 1313.
[provides details of the cirrus confusion calculation]

Helou and Beichman 1990, Proc. 29, Liege Intn. Astr. Coll., ESA SP 314.
[provides an overview of the cirrus confusion calculation]


*************************************
Extragalactic confusion computation:
*************************************
The ISSA plates are of insufficient resolution to directly compute the
confusion due to extragalactic sources in typical size beams.  To
estimate the confusion noise contributed by the background of faint
galaxies we use a set of modeled galaxy source counts to estimate the
statistical distribution of sources within an aperture. Then we use the
variance of the total flux expected from such a sample as an estimate
of the confusion noise. The method is detailed in Helou and Beichman,
1990.

In this method, source counts are taken from the analytic fits to
infrared galaxy counts (Soifer et al., 1987) and are extrapolated to two
sets of functions of source counts vs. limiting flux. Each set contains
functions computed for wavelengths which cover the ISO wavelengths
range). The two sets of functions differ in that each set is computed
under a different cosmological model so as  to bracket the results of
the calculation. The program finds the smallest flux at which the flux
variance of the faint source background will contribute an uncertainty
of less than 20%, and equates this variance to the  confusion noise.
The two bracketing cosmological models are:  Model 1: q0=0.5, H0=100, and
source density~(1+z)^5; and Model 2: Euclidean.


REFERENCES:

Helou and Beichman 1990, Proc. 29, Liege Intn. Astr. Coll., ESA SP 314.
[provides details of the extragalactic confusion calculation]

Soifer et al.  1987, The IRAS Bright Galaxy Sample, Ap J 320, 238.
[provides the luminosity functions for extragalactic confusion calculations]