Guidelines for Spitzer observers

The background estimator in Spot can be used to predict the background for Spitzer observations to an accuracy of about 20%. Thus for most purposes, it will be adequate for observation planning. Some exceptions occur as follows.

In the galactic plane, stars will merge together and become confused; at this point one could speak of a background due to starlight, which we have not included. To address this problem, observers should inspect POSS and 2MASS images or consult a model for the density of stars.

For lines of sight passing near young massive stars, the contributions of H II regions and photo-dissociation regions will not be accurately traced by our interstellar cirrus map. To address this problem, observers should inspect the IRAS and MSX images at the wavelength closest to their planned Spitzer observation.

Direct reporting of brightness from other space missions

The archival sources of background measurements that are of greatest interest for Spitzer are IRAS, MSX, and 2MASS. The 2MASS will be useful for identifying bright sources in potential Spitzer fields, and for giving accurate source counts, especially for stellar sources. The MSX galactic plane survey will will be the primary source of background information for galactic plane observers in the near- to mid-infrared.

The IRAS data are served by the IRSKY program in a very user-friendly and accurate manner. Observers are encouraged to look at their fields using IRSKY. They will obtain the sky brightness, after zodiacal light subtraction in 4 wavebands centered on 12, 25, 60, and 100 $\mu$m. They can then add this brightness to the zodiacal light model described above. In this way, the Spot-implemented empirical model for the interstellar medium (which involved a far-infrared spatial template and a single spectral template) can be replaced by actual sky brightness observations with 2-5$^\prime$ resolution. However this method will only work for regions where the sky is bright. This applies to only a relatively small fraction of the sky (< 10%). Also, observers should add the zodiacal light into their background estimate; they should do so using the model in Spot.

The background estimator built into IRSKY/IBIS is already a comprehensive tool that can predict the sky brightness as a function of position and wavelength. It contains a zodiacal light model (Good 1995; fitted to the IRAS data), moderate-resolution maps of the sky, and a means of interpolating from the IRAS wavelengths to any other. Some limitations of the existing tool are

• it does not use the observing date to specify the line of sight for the zodiacal light model, leading to a factor of up to 2 uncertainty in the zodiacal light brightness for a particular observation;
• the IRAS zodiacal light model and data are known to have absolute calibration errors;
• the zodiacal light model was based only on data longward of 12 $\mu$m and cannot be expected to be accurate at shorter wavelengths;
• the interpolation scheme from the IRAS wavelengths to arbitrary other wavelengths is highly uncertain above 20$^\circ$ galactic latitude because the IRAS 12-60 $\mu$m zodiacal-subtracted sky brightnesses are highly uncertain.

For example, near the galactic plane or any bright interstellar cloud, the actual observed brightnesses from the IRAS wavebands will be give a more accurate description of the sky brightness than the model implemented in Spot. This is because our proposed model uses only a single spectral shape for the interstellar medium, whereas in fact the spectrum varies. Therefore, in addition to using model proposed here, we strongly recommend that proposers use IRSKY/IBIS to extract the IRAS observations at 12, 25, 60, and 100 $\mu$m.