A Ground Plane Artifact that Induces an Absorption Profile in Averaged Spectra from Global 21-cm Measurements - with Possible Application to EDGES

TL;DR

This paper identifies a systematic artifact caused by ground plane resonances in global 21-cm measurements, which can mimic broad absorption features like the one reported by EDGES, challenging interpretations of the data.

Contribution

It introduces a physical model of ground plane resonances affecting 21-cm spectra and demonstrates their potential to explain observed absorption features.

Findings

Ground plane resonances can produce broad absorption features in spectra.

Resonance models fit EDGES data nearly as well as empirical models.

Ground plane artifacts may account for the EDGES 78 MHz absorption feature.

Abstract

Most of the current Global 21-cm experiments include ground screens that help moderate effects from the Earth. In this paper, we report on a possible systematic artifact within the ground plane that may produce broad absorption features in the spectra observed by these experiments. Using analytical approximations and numerical modeling, the origin of the artifact and its impact on the sky-averaged spectrum are described. The publicly released EDGES dataset, from which a 78 MHz absorption feature was recently suggested, is used to probe for the potential presence of ground plane resonances. While the lack of a noise level for the EDGES spectrum makes traditional goodness-of-fit statistics unattainable, the rms residual can be used to assess the relative goodness of fits performed under similar circumstances. The fit to the EDGES spectrum using a model with a simple 2-term foreground and three cavity-mode resonances is compared to a fit to the same spectrum with a model used by the EDGES team consisting of a 5-term foreground and a flattened Gaussian signal. The fits with the physically motivated resonance and empirical flattened Gaussian models have rms residuals of 20.8 mK (11 parameters) and 24.5 mK (9 parameters), respectively, allowing us to conclude that ground plane resonances constitute another plausible explanation for the EDGES data.