Analysis of Small Scale MBR Anisotropy in the Presence of Foreground Contamination

Abstract

Many of the current round of experiments searching for anisotropies in the MBR are confronting the problem of how to disentangle the cosmic signal from contamination due to galactic and intergalactic foreground sources. Here we show how commonly used likelihood function techniques can be generalized to account for foreground. Specifically we set some restrictions on the spectrum of foreground contamination but allow the amplitude to vary arbitrarily. The likelihood function thus generalized gives reasonable limits on the MBR anisotropy which, in some cases, are not much less restrictive than what one would get from more detailed modeling of the foreground. Furthermore, the likelihood function is exactly the same as one would obtain by simply projecting out foreground contamination and just looking at the reduced data set. We apply this generalized analysis to the recent medium angle data sets of ACME-HEMT and MAX. The resulting analysis constrains the one free parameter in the standard cold dark matter theory to be Q=18_-5^+8 microK. This best fit value, although in striking agreement with the normalization from COBE, is not a very good fit, with an overall chi^2/ degree of freedom =208/168. We also argue against three commonly used methods of dealing with foreground: (i) ignoring it completely; (ii) subtracting off a best fit foreground and treating the residuals as if uncontaminated; and (iii) culling data which appears to be contaminated by foreground.