9.5.1.4 IRAS vs. Spot vs. IRSKY: Which one should you use?

For some overview information on infrared backgrounds, see the Infrared Compendium, a web-based reference at the SSC website:

http://ssc.spitzer.caltech.edu/documents/compendium/

Specifically, there is a page on backgrounds under the overview section, along with pages on resolution, diffraction, and confusion issues. For a detailed description of exactly what Spot is doing, see this document:

http://ssc.spitzer.caltech.edu/documents/background/

In brief, this is what Spot does:

1. Takes a smooth empirical function fit to IRAS and DIRBE data (from Schlegel et al. 1998, ApJ, 500, 525), and derives a brightness distribution.
2. Employs a ``standard'' ISM spectrum (or rather, ratios of fluxes at Spitzer and IRAS wavelengths), scaled to the right intensity as just derived, and infers fluxes at the Spitzer wavelengths.
3. Adds above to a 3-d model of the zodiacal dust based on DIRBE data, also derived for Spitzer-specific wavelengths.

Thus, Spot is:

1. Using a model based on data from more than one mission.
2. Trying to compensate for the fact that Spitzer's view of the zodiacal light is changing in time as it moves away from the Earth (in fact, literally through clumps of dust).
3. Making an educated guess as to the brightness at Spitzer (rather than IRAS) wavelengths.

The first point suggests that one should also consult the IRAS images, as (a) model assumptions may in reality be incorrect, (b) the model will smooth out sharp variations in the real data in confused or complex regions of sky, and (c) Spot has assumed a single temperature for the ISM (the ``standard'' ISM spectrum), which breaks down in HII regions such as the ONC. The second point is particularly important if observations of your target(s) are likely to be dominated by zodiacal dust; Spot knows specifically where Spitzer will be as a function of time, and this can matter a great deal- there is an Earth-trailing blob of zodiacal dust, and Spitzer is going right through it. For the true novice, the third point may be specifically valuable, but it is also important (specifically in this chapter's case of the ONC where the ISM is very bright) to remember that one should always check the IRAS images to see actual data for the region in question. In the example we have here, the ISM clearly dominates, but in other regions this may not be the case. The ISSA plates have had the zodiacal light subtracted out so if you are worried about background in the ecliptic plane but not the Galactic plane, this may affect you. See the chapter on moving targets for more on some of these issues. IRSKY is a big workhorse of a program that has been used for years and does a lot of things. The background estimator that is in IRSKY spits back quite a bit of output. It gives you an average over a region of sky of the IRAS data at each wavelength (and interpolates to arbitrary wavelengths using a somewhat tricky method). It returns estimates of brightness from IRAS and COBE data after having subtracted off a zodiacal light model (based on IRAS data) with known limitations, certainly not taking into account Spitzer's position as a function of time. And, it does not understand point sources; a bright point source can distort the mean brightness in a region. Going to a bigger region might give a better idea of the diffuse emission. A median would perhaps get rid of a giant point source. An average would include the point source and not give a measure of the diffuse emission. It would just go down as the inverse square of the aperture until the point source is no longer important, unless there is lots of structured emission, in which case it is hard to even define what the background means, as is likely the case in the ONC. In the case of the ONC, it's all one big smear! Spot helpfully returns separate contributions for each of the major sources of background so you can pick and choose what you want to take from it. For example, if you are worried about the contribution from the ISM (if the ISM is very bright, like $>100 \; {\rm MJy/sr}$, you should be worried), pull the zodiacal light estimates alone out of Spot and add them to the estimates from your cursor position on ISSA images in Spot (your best option) or estimates from IRSKY (assuming you understand what is resolved and what is not), or experiment with sending slightly different RA and Dec coordinates to Spot. In the end, though, keep the original IRAS images in mind. These are actual data, and nothing beats actual data! Depending on where your target is, you could also consult Midcourse Space Experiment (MSX) data, which were taken in and near the Galactic plane at 8.28, 12.13, 14.65, and 21.3 microns. Spot will allow us to download MSX images of a target (obviously, only if it is in the region mapped by MSX). Alternatively, we can retrieve MSX data (or any other data, including our own) as FITS files from the web (e.g., from http://irsa.ipac.caltech.edu/), and load those into Spot to play with them. For IRAC wavelengths, starlight originating from stars merging together and becoming confused in the Galactic plane may also be an issue. For this, consult (actual data again!) 2MASS or DSS images, or if necessary, a star-count model, to determine if it is a problem for any given region. If our proposed region resembles any other region already observed and released into the archive, we can even load actual Spitzer data into Spot, and then investigate and scale it appropriately for our proposed target. You can read more details about the observations which are currently available ( e.g. First Look Survey, Legacy etc), by downloading the archive tool (Leopard) and accessing the Spitzer Archive. Additional information may be found at:

http://ssc.spitzer.caltech.edu/archanaly/archive.html

In the current example, a contribution from starlight is expected on top of the contribution from the ISM, given the images we retrieved from 2MASS above. The contribution from the ISM still dominates, however.