LEAP: An Innovative Direction Dependent Ionospheric Calibration Scheme for Low Frequency Arrays

TL;DR

This paper presents LEAP, a novel calibration scheme for low frequency arrays like SKA-Low, which corrects for ionospheric and instrumental phase errors across wide fields without relying on a global sky model, using bandwidth smearing and parallel processing.

Contribution

LEAP introduces a direction-dependent calibration method that leverages bandwidth smearing to measure and correct ionospheric effects independently in multiple directions, suitable for wide-field low-frequency arrays.

Findings

Effective correction of ionospheric phase distortions demonstrated at 88 and 154 MHz.

Method is suitable for wide range of spatial and temporal ionospheric fluctuations.

Parallelizable approach aligns with SKA's compute architecture.

Abstract

The ambitious scientific goals of the SKA require a matching capability for calibration of atmospheric propagation errors, which contaminate the observed signals. We demonstrate a scheme for correcting the direction-dependent ionospheric and instrumental phase effects at the low frequencies and with the wide fields of view planned for SKA-Low. It leverages bandwidth smearing, to filter-out signals from off-axis directions, allowing the measurement of the direction-dependent antenna-based gains in the visibility domain; by doing this towards multiple directions it is possible to calibrate across wide fields of view. This strategy removes the need for a global sky model, therefore all directions are independent. We use MWA results at 88 and 154 MHz under various weather conditions to characterise the performance and applicability of the technique. We conclude that this method is suitable to measure and correct for temporal fluctuations and direction-dependent spatial ionospheric phase distortions on a wide range of scales: both larger and smaller than the array size. The latter are the most intractable and pose a major challenge for future instruments. Moreover this scheme is an embarrassingly parallel process, as multiple directions can be processed independently and simultaneously. This is an important consideration for the SKA, where the current planned architecture is one of compute-islands with limited interconnects. Current implementation of the algorithm and on-going developments are discussed.