Abstract

Interferometric observations from the radio to infrared and optical are faced with a common challenge: recovering a source’s on-sky brightness from sparsely sampled Fourier data. This is often approached as an image reconstruction problem. In sub-mm observations of protoplanetary discs for example, the most common modeling technique, ‘CLEAN,’ operates in the image plane, precluding a full exploitation of the observations’ frequency information. This in turn places a fundamental limit on the model’s accuracy. In this talk I will reframe the problem as a spatial frequency domain analysis, presenting our open source code that fits the interferometric data directly in the Fourier domain with a high resolution, nonparametric and fast Gaussian process. The technique is currently tailored to (sub-)mm observations of protoplanetary discs, using 1D fits to assess these typically symmetric objects. With synthetic and real observations I will motivate how these fits are recovering disc structures beyond the CLEAN imaging resolution ( ‘sub-beam’ ) in both low and high resolution datasets. Noting scientific applications of our model in the protoplanetary disc context, I will then conclude by outlining the code’s planned extensions and potential utility in other physical and frequency regimes.