Spitzer Documentation & Tools
MIPS Instrument Handbook



4.1.3        Uncertainties and Repeatability

We derived the flux conversion using both primary and secondary stars by fitting the relationship between predicted flux and DN.  We have to rely on models to predict the fluxes at MIPS bands. Kurucz models are used for A stars, templates from Cohen et al. (2003 and references therein) are used for K stars, and direct transfer of solar spectrum  (Tobiska et al. 2000) for G stars. We extrapolated the Cohen templates and Kurucz models to far infrared using the Engelke function (Engelke 1992) for A and K stars.  The derived conversion factors are shown in Table 4.10.


The stellar photometry using the primary calibrator of HD159330 shows that the flux is repeatable at the 0.4% level.  However, additional errors in the absolute calibration come from uncertainty in model prediction and aperture and color corrections. 


Estimated uncertainty on point sources extracted from pipeline products is 4% for 24 microns, 7% for 70 microns, and 12% for 160.  Additional analysis can achieve 2% at 24 microns, 5% at 70 microns, and 9% at 160 microns.  Extended source measurements have been consistent with these measurements.

Accuracy, Stability, and Repeatability

As with all other aspects of performance, the stability and repeatability of the 24 micron array is excellent.  Photometry is repeatable to 1% or better and the absolute calibration is better than 5% determined (see section 4.1.1) using observations of solar analog, A, and K giant stars. 


The 70 micron calibration is based (see section 4.1.2) on solar analog, A, and K giant stars.  Using all the calibration star measurements, the accuracy of the 70 micron measurements is around 10%.  These calibration measurements include stars with predicted fluxes ranging from 30 mJy to 5 Jy on backgrounds between 4 and 25 MJy/sr.  One of the calibration stars (HD 163588) was measured every MIPS campaign and the repeatability of this star is better than 7%.  The main difference between these two accuracies is likely the result of the nonlinearities still to be corrected (see the Gordon et al. papers on data reduction or Rieke et al. 2004, ApJS, 154, 25, for more details). 


Unlike the other two arrays, the calibration of the 160 micron data is based on a variety of objects (see section 4.1.2) including asteroids.  The resulting multiple calibrations agree to within 12%.  Because these were not the planned calibration sources (see sections 4.1.2 and 7.2.6), the 160 micron flux calibration is more uncertain than for the other two MIPS arrays. Repeatability of measurements of the same object is ~10%.