Pixelization and Dynamic Range in Radio Interferometry

TL;DR

This paper examines how pixelization, sky curvature, and calibration uncertainties affect the dynamic range of radio interferometric images, proposing a technique to mitigate these effects for improved imaging quality.

Contribution

It introduces a novel method to reduce pixelization and curvature effects in radio interferometry imaging, demonstrated with an implementation in the Obit package.

Findings

The technique significantly improves image dynamic range.

Implementation in Obit shows better results without extra computational cost.

Pixelization and curvature effects are major limits to image quality.

Abstract

This study investigates some of the consequences of representing the sky by a rectangular grid of pixels on the dynamic range of images derived from radio interferometric measurements. In particular, the effects of image pixelization coupled to the CLEAN deconvolution representation of the sky as a set of discrete delta functions can limit the dynamic range obtained when representing bright emission not confined to pixels on the grid. Sky curvature effects on non-coplanar arrays will limit the dynamic range even if strong sources are centered on a pixel in a "fly's eye" representation when such pixel is not located at the corresponding facet's tangent point. Uncertainties in the response function of the individual antennas as well as in the calibration of actual data due to ionospheric, atmospheric or other effects will limit the dynamic range even when using grid-less subtraction (i.e. in the visibility domain) of strong sources located within the field of view of the observation. A technique to reduce these effects is described and examples from an implementation in the Obit package are given. Application of this technique leads to significantly superior results without a significant increase in the computing time.