UC Santa Barbara

Destriping is a well-established technique for removing low-frequency correlated noise from Cosmic Microwave Background (CMB) survey data. In this paper we present a destriping algorithm tailored to data from a polarimeter, i.e.an instrument where each channel independently measures the polarization of the input signal. We also describe a fully parallel implementation in Python released as Open Source software and analyze its results and performance on simulated datasets, both the design case of signal and correlated noise, and also with the addition of other systematic effects.Finally we apply the algorithm to 30 days of 37.5 GHz polarized microwave data gathered from the B-Machine experiment, developed at UCSB. The B-Machine data and destriped maps are publicly available. The purpose is the development of a scalable software tool to be applied to the upcoming 12 months of temperature and polarization data from LATTE (Low frequency All sky TemperaTure Experiment) at 8GHz and to even larger datasets. © 2013 Elsevier B.V.