head	1.1;
access;
symbols;
locks
	http:1.1; strict;
comment	@# @;
expand	@b@;


1.1
date	2002.12.04.11.12.10;	author MarkAllen;	state Exp;
branches;
next	;


desc
@none
@


1.1
log
@Notes on coordinating Aladin, SED and SExtractor
@
text
@Dec 04, 2002

I've made some notes to help coordinate the way Aladin, 
SED and SExtractor Web Service talk to each other,
with an emphasis on getting the flux conversions 
right. Also included here is a full description of
how we currently understand the EIS photometry.

This is relevant for those preparing data for the
demo, those wanting to plot data as an SED, and
coordination of re-extracted catalogs from the
SExtractor Web service.

SEDs :

A core element of the demo is to be able to plot
SEDs from the cross-matched EIS UBVRIJK and UBVRI 
catalogs, and from the published single band UBVRI
catalogs.

Highly desirable is the ability to also plot
SEDs of catalogs from the SExtractor Web service 
re-extracted EIS catalogs. Also desirable is the ability
to plot other SED data loaded into Aladin (or VizieR).

The current status is that Aladin can provide 
a VOTable of the fluxes, errors, central frequencies
and bandwidths for the EIS UBVRIJK catalog, as
input to the SED tool. We can (almost) also provide the
published EIS single pass band catalogs in the same way.
This all works with the current SED v0.3 from Markus.

(Next parts are particularly relevant to SExtractor WS)

At this point in time, the list:
fluxes, errors, central frequencies and bandwidths
represents the minimum info required in order to plot 
an SED. So, in general the minimum info we require from 
the SExtractor WS is:

Position (RA,DEC)
Flux(Jy) 
Error in Flux(Jy)
Effective or central frequency (Hz)
Bandwidth (Hz)

This will allow us to plot the catalog positions
on an image, and to send the SED info to the SED tool.

In the VOTable that comes back to Aladin from SExtractor, 
we would like these columns to have the following UCDs:

Position RA  (J2000) : POS_EQ_RA_MAIN
Position DEC (J2000) : POS_EQ_DEC_MAIN

Flux (Jy)             : PHOT_FLUX_DENSITY
error in flux         : ERROR
central frequency (Hz) : OBS_FREQUENCY
bandwidth  (Hz)        : INST_FILTER_FWHM

Also, can you please put the header of the image into the 
info field of the VOTable? (Please contact us if this is
not clear)

Note, this should also make local tables loaded into Aladin
(versions > 0.904 avail by Dec 06) as VOTables with these UCDs, 
plottable via the SED tool. (I also note that the SED v0.3 has a 
VOTable load button but don't know yet how that works)


Re-extracted EIS catalogs:

For catalogs re-extracted from the EIS data there is probably going to
be a problem (as I described to Martin by phone on Friday)
that there isn't sufficient information in the image headers alone
to extract fluxes in Jy. 

(Can anyone prove me wrong on this? - If you can verfiy that 
SExtrator does give correct flux in Jy for EIS data, then 
send the catalog with UCDs as described above.)

This is the same problem we've had to deal with here (in Aladin)
for the already available EIS catalogs, where we have
to convert the magnitudes into fluxes. To do this we have
adopted a set of Vega fluxes for the UBVRIJK filters.
(See "Mag-Flux conversions and caveats" included below)

So, how do we get fluxes from the SExtractor WS re-extracted
catalogs ? The SExtractor WS could either do the same conversion,
(mag-flux) using the same Vega fluxes, or somehow get the information required 
to convert the image counts/s to flux. Alternatively SExtractor WS 
could send re-extracted EIS catalogs back to Aladin in 
magnitudes units (utilising the Zero points which are in the 
EIS WFI/ISAAC/SOFI image headers). These magnitudes could then
be converted to fluxes in Aladin using  the same conversion program
as we use for the existing catalogs. Since the conversion parameters 
are likely to be updated, our preference (for the demo at least) is 
to get the re-extracted EIS catalogs in magnitudes. 

SExtractor can provide (along with other copius output) a number
of magnitude values. Looking at what has been done for the published
EIS catalogs (Arnouts 2001) the magnitude listed as "mag_auto"
appears to be the most relevant one for our purposes. 

In the case of a re-extracted catalog in magnitude units, please
provide in the VOTable, the following columns and UCDs

U       (mag) : PHOT_MAG_U
U error (mag) : ERROR
B       (mag) : PHOT_MAG_B
B error (mag) : ERROR
V       (mag) : PHOT_MAG_V
V error (mag) : ERROR
R       (mag) : PHOT_MAG_R
R error (mag) : ERROR
I       (mag) : PHOT_MAG_I
I error (mag) : ERROR
J       (mag) : PHOT_MAG_J
J error (mag) : ERROR
K       (mag) : PHOT_MAG_K
K error (mag) : ERROR


These UCDS will be recognized as data requiring conversion to flux 
using the recipe installed for EIS.
Note that there isn't an effective frequency or bandwidth column
here. This is because the unit converter knows these values from
the UCD.  

Also note that the ERROR UCDs are all the same. If there was a 
VOTable with more than one type of error we would be in trouble.
In Vizier this problem is handled by a naming convention for the
column names, where e_ColName indicates the error for column with name 
ColName. I'd suggest naming columns with this convention for the
the SExtractor WS VOTable output. (future versions of VOTable 
may be able to handle this column association within the table 
structure, but the current version doesn't)

Other units (surface brightness etc.) :

As I undestand it, there is currently no provision for handling
surface brightness info, i.e. fluxes in Jy/arcsec^2, or fluxes in
any other format than F_nu(Jy). So, any such info would need to
be converted to total Jy ("per source").



@