Estimating Ionospheric Faraday Rotation at MeerKAT Using Single-Station GNSS Processing

Time-variable ionospheric Faraday rotation limits the absolute electric vector polarization angle (EVPA) calibration at MeerKAT’s UHF and L-band frequencies. Global Ionospheric Maps (GIMs) provide broad total electron-content (TEC) coverage; however, a local Global Navigation Satellite System (GNSS) receiver offers measurements closer to the telescope at higher temporal resolutions. Local signal processing is complicated by signal-dependent differential code biases (DCBs). This project developed and evaluated a single-station GNSS workflow for estimating receiver DCB and predicting ionospheric Faraday rotation measure (IFRM) for MeerKAT.

Dual-frequency GPS observations from the MK01 GNSS receiver at MeerKAT and 12 reference stations were processed from 29 June 2025 to 12 February 2026. Daily receiver DCBs were estimated from carrier-phase-leveled geometry-free observations using IGS, CODE, CAS, and JPL final ionospheric products. Five GIMs were assessed using differential slant total electron content (dSTEC), which removes constant hardware biases within continuous carrier-phase arcs. The estimated DCBs were then incorporated into the Advanced Long Baseline User Software (ALBUS) single-station G01 model to generate IFRM predictions for three selected MeerKAT observation dates.

The level of DCB agreement varied depending on the product. CODE produced the smallest standard deviation of the DCB residuals (0.112 ns) and the smallest root-mean-square error (0.204 ns) relative to its published values. CODE also gave the smallest dSTEC residual RMS across the selected ten-station South African network at 2.02 TECU and for the MK01 receiver at 1.96 TECU. The product ranking was identical at every station, and the daily MK01 residuals correlated strongly with the regional median, with a Pearson correlation coefficient of r = 0.976–0.991. Storm-day RMS was 15%–22% higher than quiet-day RMS, and low-elevation observations produced the largest residuals. ALBUS solutions from MK01 and the Sutherland receivers followed the same sunrise-driven ionospheric evolution on 18 and 19 August 2025, with median inter-station spreads of 0.12 and 0.13 rad m^-2. The spread increased to 0.44 rad m^-2 on 30 October, when the timing and depth of the predicted IFRM minima differed between stations.

The results identify CODE as the best-supported GIM for this region and demonstrate a path from an MK01 receiver without a published DCB to a local, bias-corrected ALBUS IFRM prediction. They also show that Sutherland can reproduce broad regional behavior without always representing the instantaneous ionosphere above MeerKAT. The absolute accuracy and EVPA-calibration suitability of these predictions remain to be studied through further comparison with rotation measures derived from the radial lunar polarization pattern.