FITS WCS with Terrestrial Satellite Data

[Louis Giglio]

Greetings all,

For the past several months our group has been generating and working
with FITS files containing terrestrial satellite data.  Although FITS
was intended for sharing astronomical data, we've found that it is
quite useful as an archival earth-science data format.  In fact, at
least for our purposes, FITS has turned out to be a very practical
working format as well.  (We've used HDF for both of these purposes in
the past, but have largely dropped it for a host of reasons.  But I
digress...)

Much of our data has been mapped into a variety of cartographic
projections.  Rather than reinventing the wheel, I'd like to use the
proposed WCS conventions of Greisen and Calabretta (1996) to describe
the geographic (as opposed to celestial) coordinates of our mapped
data.

Currently I've been creating experimental terrestrial FITS files that
use the WCS conventions.  The only disadvantage I know of in using WCS
for this purpose is that the left handed spherical coordinate system
is confusing to most folks.  It's easy enough to convert to right
handed longitudes within our display software, however, so I don't
view this as a significant problem.

In the course of working with our experimental FITS files, several
issues have arisen for which I'm hoping to get some advice:


1) CTYPEn keywords

What are reasonable values for the "xLON" and "xLAT" keywords?  The
codes x = G (galactic), E (ecliptic), and H (helioecliptic) have
already been defined by Greisen and Calabrettta.  We've been using use
"XLON" and "XLAT", but we'd prefer to use a convention that will never
conflict with the needs of the astronomical community.  Others have
suggested "PLON" / "PLAT" (for "planetary") and "TLON" / "TLAT" (for
"terrestrial").


2) EQUINOX keyword

For terrestrial data this keyword has no meaning.  Is it OK to omit
it?


3) RADECSYS keyword

Is it correct that this keyword should _not_ be supplied for "xLON"
and "xLAT" type reference frames?  If it should be supplied, what
value should we use?


4) Is anyone else attempting to do this sort of thing?  What
cartographic formats are used for mapped data from unmanned planetary
missions?


5) Am I missing something that makes this a stupid idea?


I'm also interested in hearing general comments and suggestions.


Thanks very much,
Louis Giglio

giglio at hades.gsfc.nasa.gov

NASA/GSFC
Biospheric Sciences Branch
Greenbelt, MD USA