BEYONDPLANCK

Abstract

We present a Bayesian method for estimating instrumental noise parameters and propagating noise uncertainties within the global BEYONDPLANCKGibbs sampling framework, which we applied toPlanckLow Frequency Instrument (LFI) time-ordered data. Following previous works in the literature, we initially adopted a 1/fmodel for the noise power spectral density (PSD), but we found the need for an additional lognormal component in the noise model in the 30 and 44 GHz bands. We implemented an optimal Wiener-filter (or constrained realization) gap-filling procedure to account for masked data. We then used this procedure to both estimate the gapless correlated noise in the time-domain,ncorr, and to sample the noise PSD parameters,ξn = {σ0, fknee, α, Ap}. In contrast to previousPlanckanalyses, we assumed piecewise stationary noise only within each pointing period (PID), and not throughout the full mission, but we adopted the LFI Data Processing Center results as priors onαandfknee. We generally found best-fit correlated noise parameters that are mostly consistent with previous results, with a few notable exceptions. However, a detailed inspection of the time-dependent results has revealed many important findings. First and foremost, we find strong evidence for statistically significant temporal variations in all noise PSD parameters, many of which are directly correlated with satellite housekeeping data. Second, while the simple 1/fmodel appears to be an excellent fit for the LFI 70 GHz channel, there is evidence for additional correlated noise that is not described by a 1/fmodel in the 30 and 44 GHz channels, including within the primary science frequency range of 0.1–1 Hz. In general, most 30 and 44 GHz channels exhibit deviations from 1/fat the 2–3σlevel in each one-hour pointing period, motivating the addition of the lognormal noise component for these bands. For certain periods of time, we also find evidence of strong common mode noise fluctuations across the entire focal plane. Overall, we conclude that a simple 1/fprofile is not adequate for obtaining a full characterization of thePlanckLFI noise, even when fitted hour-by-hour, and a more general model is required. These findings have important implications for large-scale CMB polarization reconstruction with thePlanckLFI data and the current work is a first attempt at understanding and mitigating these issues.