SED Data Model: Questions and Comments

[Ed Shaya]

There is yet another argument for purposefully taking spectral data 
without regard to absolute calibration.  Many physical quantities rely 
only on the ratio of equivalent widths in a given spectrum.   Gas 
collision temperature and gas density are examples.  The equivalent 
width is the ratio of the area of a line to the continuum flux at the 
center of  the line so the sensitivity at that wavelength cancels out of 
the EW.   Studies of gas physics, which provide a major contribution to 
the spectral archives, need not and very often do not have calibrations.

However, I would argue that some effort needs to be made to ensure some 
measure of an error estimate is on archived data.  It just is very hard 
for anyone to do valid science without this.  Calling the astronomer is 
not acceptable because most humans die.  To do Poison statistics one 
needs to at least know how many photons are represented by a count 
(DN).  So that is necessary, usually suffiicient, and easy.  In cases in 
which this is not available or appropriat, the noise can be estimated by 
assuming a certain section of the spectrum is actually smooth and so the 
fluctuations are a measure of the noise.  This is tricky however, 
because often at higher resolution or better S/N it turns out that a 
"smooth" section is a forest of lines.  Only someone who really 
understands the physics of the object can properly assess this, ie the 
observer or his/her analysis team.

Igor Chilingarian wrote:

>Hello,
>
>On Thu, 3 Feb 2005, Gilles DUVERT wrote:
>
>  
>
>>Hi,
>>
>>DM is not just the sum of its parts. One has to make choices. The need
>>is to insure data QUALITY and MAINTAINABILITY. Keep things simple. DM
>>should describe the physics of the data, that unite, not the habits of
>>the data producer, that divide.
>>
>>So:
>>
>>A measure is a value and an error. If there is no flux calibration, the
>>measurement does not exist, and a data model shall not be concerned by this.
>>DO NOT PERMIT TO WRONG OR UNRELIABLE DATA TO BE OUT IN THE VO!!!
>>    
>>
>
>Value and error have nothing to do with the flux calibration. And this doesn't
>always affect the physics. Detector of any type (at least those I know to be
>used in optical astronomy) doesn't give an output in a physical flux, but only
>in counts per channel (or pixel, or just total counts), and no error is given:
>usually it is assumed __by astronomer__ to correspond to Poisson photon
>statistics or some other primitive model. And probably everybody agrees on the
>fact, that it is NOT POSSIBLE to have reliable absoulute measurements for flux
>while you're below the atmosphere. All the calibrations are completely
>MODEL-DEPENDENT: you estimate some extinction law, the same atmosphere
>transparency during observing object and calibration frames, and so forth.
>
>Beside this there is a problem of calibration for photometry as well. If you
>take a photometric point, say Johnson J magnitude with some error-bar
>(favourite example by Pedro), it doesn't bear you any useful information,
>because nobody knows what Johnson J is! There is __a kind of__ transparency
>curve for a filter, but it varies quite a lot in different realization of the
>filters used in different observatories. What astronomer will do to find out
>the details (zero-point, transmission curve, and so on) -- he'll give a phone
>call to the person who published the data.
>
>Does it mean that according to your proposition, we must reject all the
>ground-based spectral data, and photometry except SDSS :)? What will remain?
>The processed contain of MAST + SDSS + some other space mission archives? And
>all the astronomers will be happy with this?
>
>  
>
>>USE FREQUENCY, this is the core of reality, insensitive to media. Does a
>>bracket gamma line takes a cold because the weather that day was cold
>>and humid and it did not take its scarf? Leave the gory details to the
>>producer of the data. Leave the end-user convert metadata frequencies to
>>whatever unit it pleases himself.
>>    
>>
>I'll make the comments concerning this point in the next message.
>
>  
>
>>>4. Resolution and line spread functions
>>>
>>>      
>>>
>>Let the DM support standard functions (sigma, gauss, lorentz) that fit
>>99% of needs, and leave the transform of the data to this functions to
>>the  expertize of the data producer. He knows best.
>>    
>>
>It is absolutely impossible, because it requires deconvolution that never
>works in practice. So, Jonathan's idea about representing the functions and
>evaluating them is a good solution. For a moment, the DM allows only FWHM - an
>integral characteristic of the line-spread-function, that is unsufficient in
>case you want to make detailled analysis -> you again have to give a phone
>call (or e-mail) to the data provider to find out the details.
>
>With best regards,
>						Igor
>
>P.S.: Sorry, if you have this message more than once.
>
>  
>