Date: Fri, 30 Aug 2002 14:14:19 -0700
From: John Carpenter <jmc@astro.caltech.edu>


Statement of problem:
    A key goal of our project is to quantify the observational characteristics
    of disks (as traced by the IRAC, MIPS, and low-res spectra) and determine
    how these characteristics evolve in time. The observed characteristics will 
    map, in a manner to be determined, the physical properties of disks
    (composition, spatial structure, frequency) and ultimately the diversity
    and evolution of planetary systems.  Rather than focus on explaining the 
    observed properties of every individual star in our sample, it will be 
    informative to quantify the mean and dispersion in the observed disk 
    characteristics as a function of stellar properties (primarily age),
    and to then interpret the broad trends within the data. The analysis can 
    be broken down into 3 steps: 
    (1) quantify the disk characteristics based on the SEDS and low-res 
        spectra;
    (2) determine how these characteristics correlate (if at all) with stellar
        properties;
    (3) interpret any broad trends with the physical models developed by
        our group.
    This abstract deals mainly with quantifying the observed disk 
    characteristics and looking for trends with the data itself (steps 1 and 2).

Method of Attack
    It is not clear to me yet what is the appropriate metric to quantify the
    the disk properties, but I have some ideas. The simplest, but inadequate, 
    measure is the fraction of stars that have an 24um excess. This will be 
    looked at certainly, but that alone does not fully exploit the data. 
    Instead one should be able to group the data by the shape of the SEDS and 
    the spectral features. This is usually done by visual inspection of the 
    data, but I believe it is worth exploring more quantitative measures so
    that the observations can be readily compared with our models. 
    Specifically this analysis seems well suited for principal component 
    techniques which will identify any groups inherent in the data.

Required data
    This project will require the IRAC, MIPS, and low-res spectra, and also
    Martin's models in order to remove the underlying stellar photosphere.
    I do not see a need at first to make a distinction between 
    "post-accretion" and "debris" disks, and it would be vastly preferable 
    to analyze all feps targets together. In that manner, any differences 
    between young post-accretion disks and old debris disks will fall out of 
    the analysis, and in an ideal world, we would be able to make a 
    quantitative distinction between the two.

Duration
    This is a long-term project. In the short term, one can explore and 
    fine-tune the analysis techniques on the theoretical models before
    SIRTF launch, and then optimize the analysis once real data starts
    coming in. In the long term I foresee a paper that will run a global
    analysis on all stars in our sample and analyze the data on a purely
    empirical basis. A separate paper, probably led by one of the modelers 
    on our team, would interpret any broad trends that are identified.

Level of commitment
    High

List of co-investigators
    I have yet to discuss this abstract with anyone, but I would suspect 
    Meyer (?), Hillenbrand (?), Backman (?) and others would be
    interested.