We quantify the effect of radio frequency interference (RFI) on Epoch of Reionization (EoR) power spectra. Specifically, we investigate how the frequency structure of source emission generates contamination in higher-order wave modes that is much more problematic than smooth-spectrum foreground sources. We find that even a single relatively dim RFI source can overwhelm typical estimates of the EoR power spectrum signal on cosmological modes of interest. If total apparent RFI flux density in an integration is kept below 1 mJy, an EoR detection should be feasible given no other systematic. These results emphasize the need for highly effective RFI mitigation strategies for telescopes used to search for the EoR.

We compute the spherically-averaged power spectrum from four seasons of data obtained for the Epoch of Reionisation (EoR) project observed with the Murchison Widefield Array (MWA). We measure the EoR power spectrum over $k= 0.07-3.0~h$Mpc$^{-1}$ at redshifts $z=6.5-8.7$. The largest aggregation of 110 hours on EoR0 high-band (3,340 observations), yields a lowest measurement of (43~mK)$^2$ = 1.8$\times$10$^3$ mK$^2$ at $k$=0.14~$h$Mpc$^{-1}$ and $z=6.5$ (2$\sigma$ thermal noise plus sample variance). Using the Real-Time System to calibrate and the CHIPS pipeline to estimate power spectra, we select the best observations from the central five pointings within the 2013--2016 observing seasons, observing three independent fields and in two frequency bands. This yields 13,591 2-minute snapshots (453 hours), based on a quality assurance metric that measures ionospheric activity. We perform another cut to remove poorly-calibrated data, based on power in the foreground-dominated and EoR-dominated regions of the two-dimensional power spectrum, reducing the set to 12,569 observations (419 hours). These data are processed in groups of 20 observations, to retain the capacity to identify poor data, and used to analyse the evolution and structure of the data over field, frequency, and data quality. We subsequently choose the cleanest 8,935 observations (298 hours of data) to form integrated power spectra over the different fields, pointings and redshift ranges.

The compact configuration of Phase II of the Murchison Widefield Array (MWA) consists of both a redundant subarray and pseudo-random baselines, offering unique opportunities to perform sky-model and redundant interferometric calibration. The highly redundant hexagonal cores give improved power spectrum sensitivity. In this paper, we present the analysis of nearly 40 hours of data targeting one of the MWA's EoR fields observed in 2016. We use both improved analysis techniques presented in Barry et al. (2019) as well as several additional techniques developed for this work, including data quality control methods and interferometric calibration approaches. We show the EoR power spectrum limits at redshift 6.5, 6.8 and 7.1 based on our deep analysis on this 40-hour data set. These limits span a range in $k$ space of $0.18$ $h$ $\mathrm{Mpc^{-1}}$ $<k<1.6$ $h$ $\mathrm{Mpc^{-1}}$, with a lowest measurement of $\Delta^2\leqslant2.39\times 10^3$ $\mathrm{mK}^2$ at $k=0.59$ $h$ $\mathrm{Mpc^{-1}}$ and $z=6.5$.

We introduce a new pipeline for analyzing and mitigating radio frequency interference (RFI), which we call Sky-Subtracted Incoherent Noise Spectra (SSINS). SSINS is designed to identify and remove faint RFI below the single baseline thermal noise by employing a frequency-matched detection algorithm on baseline-averaged amplitudes of time-differenced visibilities. We demonstrate the capabilities of SSINS using the Murchison Widefield Array (MWA) in Western Australia. We successfully image aircraft flying over the array via digital television (DTV) reflection detected using SSINS and summarize an RFI occupancy survey of MWA Epoch of Reionization data. We describe how to use SSINS with new data using a documented, publicly available implementation with comprehensive usage tutorials.

Measurements of 21 cm Epoch of Reionization (EoR) structure are subject to systematics originating from both the analysis and the observation conditions. Using 2013 data from the Murchison Widefield Array (MWA), we show the importance of mitigating both sources of contamination. A direct comparison between results from Beardsley et al. 2016 and our updated analysis demonstrates new precision techniques, lowering analysis systematics by a factor of 2.8 in power. We then further lower systematics by excising observations contaminated by ultra-faint RFI, reducing by an additional factor of 3.8 in power for the zenith pointing. With this enhanced analysis precision and newly developed RFI mitigation, we calculate a noise-dominated upper limit on the EoR structure of $\Delta^2 \leq 3.9 \times 10^3$ mK$^2$ at $k=0.20$ $\textit{h}$ Mpc$^{-1}$ and $z=7$ using 21 hr of data, improving previous MWA limits by almost an order of magnitude.

Epoch of Reionization data analysis requires unprecedented levels of accuracy in radio interferometer pipelines. We have developed an imaging power spectrum analysis to meet these requirements and generate robust 21 cm EoR measurements. In this work, we build a signal path framework to mathematically describe each step in the analysis, from data reduction in the FHD package to power spectrum generation in the $\varepsilon$ppsilon package. In particular, we focus on the distinguishing characteristics of FHD/$\varepsilon$ppsilon: highly accurate spectral calibration, extensive data verification products, and end-to-end error propagation. We present our key data analysis products in detail to facilitate understanding of the prominent systematics in image-based power spectrum analyses. As a verification to our analysis, we also highlight a full-pipeline analysis simulation to demonstrate signal preservation and lack of signal loss. This careful treatment ensures that the FHD/$\varepsilon$ppsilon power spectrum pipeline can reduce radio interferometric data to produce credible 21 cm EoR measurements.

We apply two methods to estimate the 21~cm bispectrum from data taken within the Epoch of Reionisation (EoR) project of the Murchison Widefield Array (MWA). Using data acquired with the Phase II compact array allows a direct bispectrum estimate to be undertaken on the multiple redundantly-spaced triangles of antenna tiles, as well as an estimate based on data gridded to the $uv$-plane. The direct and gridded bispectrum estimators are applied to 21 hours of high-band (167--197~MHz; $z$=6.2--7.5) data from the 2016 and 2017 observing seasons. Analytic predictions for the bispectrum bias and variance for point source foregrounds are derived. We compare the output of these approaches, the foreground contribution to the signal, and future prospects for measuring the bispectra with redundant and non-redundant arrays. We find that some triangle configurations yield bispectrum estimates that are consistent with the expected noise level after 10 hours, while equilateral configurations are strongly foreground-dominated. Careful choice of triangle configurations may be made to reduce foreground bias that hinders power spectrum estimators, and the 21~cm bispectrum may be accessible in less time than the 21~cm power spectrum for some wave modes, with detections in hundreds of hours.

We explore methods for robust estimation of the 21 cm signal from the Epoch of Reionisation (EoR). A Kernel Density Estimator (KDE) is introduced for measuring the spatial temperature fluctuation power spectrum from the EoR. The KDE estimates the underlying probability distribution function of fluctuations as a function of spatial scale, and contains different systematic biases and errors to the typical approach to estimating the fluctuation power spectrum. Extraction of histograms of visibilities allows moments analysis to be used to discriminate foregrounds from 21 cm signal and thermal noise. We use the information available in the histograms, along with the statistical dis-similarity of foregrounds from two independent observing fields, to robustly separate foregrounds from cosmological signal, while making no assumptions about the Gaussianity of the signal. Using two independent observing fields to robustly discriminate signal from foregrounds is crucial for the analysis presented in this paper. We apply the techniques to 13 hours of Murchison Widefield Array (MWA) EoR data over two observing fields. We compare the output to that obtained with a comparative power spectrum estimation method, and demonstrate the reduced foreground contamination using this approach. Using the second moment obtained directly from the KDE distribution functions yields a factor of 2-3 improvement in power for k < 0.3hMpc^{-1} compared with a matched delay space power estimator, while weighting data by additional statistics does not offer significant improvement beyond that available for thermal noise-only weights.

Precise instrument calibration is critical to the success of 21 cm Cosmology experiments. Unmitigated errors in calibration contaminate the Epoch of Reionization (EoR) signal, precluding a detection. Barry et al. 2016 characterizes one class of inherent errors that emerge from calibrating to an incomplete sky model, however it has been unclear if errors in the sky model affect the calibration of redundant arrays. In this paper, we show that redundant calibration is vulnerable to errors from sky model incompleteness even in the limit of perfect antenna positioning and identical beams. These errors are at a level that can overwhelm the EoR signal and prevent a detection. Finally, we suggest error mitigation strategies with implications for the Hydrogen Epoch of Reionization Array (HERA) and the Square Kilometre Array (SKA).

Observations of the 21cm Epoch of Reionization (EoR) signal are dominated by Galactic and extragalactic foregrounds. The need for foreground removal has led to the development of two main techniques, often referred to as "foreground avoidance" and "foreground subtraction." Avoidance is associated with filtering foregrounds in Fourier space, while subtraction uses an explicit foreground model that is removed. Using 1088 hours of data from the 64-element PAPER array, we demonstrate that subtraction of a foreground model prior to delay-space foreground filtering results in a modest but measurable improvement of the performance of the filter. This proof-of-concept result shows that improvement stems from the reduced dynamic range requirements needed for the foreground filter: subtraction of a foreground model reduces the total foreground power, so for a fixed dynamic range, the filter can push towards fainter limits. We also find that the choice of window function used in the foreground filter can have an appreciable affect on the performance near the edges of the observing band. We demonstrate these effects using a smaller 3 hour sampling of data from the MWA, and find that the hybrid filtering and subtraction removal approach provides similar improvements across the band as seen in the case with PAPER-64.

21 cm power spectrum observations have the potential to revolutionize our understanding of the Epoch of Reionization and Dark Energy, but require extraordinarily precise data analysis methods to separate the cosmological signal from the astrophysical and instrumental contaminants. This analysis challenge has led to a diversity of proposed analyses, including delay spectra, imaging power spectra, m-mode analysis, and numerous others. This diversity of approach is a strength, but has also led to confusion within the community about whether insights gleaned by one group are applicable to teams working in different analysis frameworks. In this paper we show that all existing analysis proposals can be classified into two distinct families based on whether they estimate the power spectrum of the measured or reconstructed sky. This subtle difference in the statistical question posed largely determines the susceptibility of the analyses to foreground emission and calibration errors, and ultimately the science different analyses can pursue. In this paper we detail the origin of the two analysis families, categorize the analyses being actively developed, and explore their relative sensitivities to foreground contamination and calibration errors.

Interferometric arrays seeking to measure the 21 cm signal from the Epoch of Reionization must contend with overwhelmingly bright emission from foreground sources. Accurate recovery of the 21 cm signal will require precise calibration of the array, and several new avenues for calibration have been pursued in recent years, including methods using redundancy in the antenna configuration. The newly upgraded Phase II of Murchison Widefield Array (MWA) is the first interferometer that has large numbers of redundant baselines while retaining good instantaneous UV-coverage. This array therefore provides a unique opportunity to compare redundant calibration with sky-model based algorithms. In this paper, we present the first results from comparing both calibration approaches with MWA Phase II observations. For redundant calibration, we use the package OMNICAL, and produce sky-based calibration solutions with the analysis package Fast Holographic Deconvolution (FHD). There are three principal results. (1) We report the success of OMNICAL on observations of ORBComm satellites, showing substantial agreement between redundant visibility measurements after calibration. (2) We directly compare OMNICAL calibration solutions with those from FHD, and demonstrate these two different calibration schemes give extremely similar results. (3) We explore improved calibration by combining OMNICAL and FHD. We evaluate these combined methods using power spectrum techniques developed for EoR analysis and find evidence for marginal improvements mitigating artifacts in the power spectrum. These results are likely limited by signal-to-noise in the six hours of data used, but suggest future directions for combining these two calibration schemes.

We present techniques developed to calibrate and correct Murchison Widefield Array (MWA) low frequency (72-300 MHz) radio observations for polarimetry. The extremely wide field-of-view, excellent instantaneous (u, v)-coverage and sensitivity to degree-scale structure that the MWA provides enable instrumental calibration, removal of instrumental artefacts, and correction for ionospheric Faraday rotation through imaging techniques. With the demonstrated polarimetric capabilities of the MWA, we discuss future directions for polarimetric science at low frequencies to answer outstanding questions relating to polarised source counts, source depolarisation, pulsar science, low-mass stars, exoplanets, the nature of the interstellar and intergalactic media, and the solar environment.

We present new radio continuum observations of NGC253 from the Murchison Widefield Array at frequencies between 76 and 227 MHz. We model the broadband radio spectral energy distribution for the total flux density of NGC253 between 76 MHz and 11 GHz. The spectrum is best described as a sum of central starburst and extended emission. The central component, corresponding to the inner 500pc of the starburst region of the galaxy, is best modelled as an internally free-free absorbed synchrotron plasma, with a turnover frequency around 230 MHz. The extended emission component of the NGC253 spectrum is best described as a synchrotron emission flattening at low radio frequencies. We find that 34% of the extended emission (outside the central starburst region) at 1 GHz becomes partially absorbed at low radio frequencies. Most of this flattening occurs in the western region of the SE halo, and may be indicative of synchrotron self-absorption of shock re-accelerated electrons or an intrinsic low-energy cut off of the electron distribution. Furthermore, we detect the large-scale synchrotron radio halo of NGC253 in our radio images. At 154 - 231 MHz the halo displays the well known X-shaped/horn-like structure, and extends out to ~8kpc in z-direction (from major axis).

The Detection of redshifted 21 cm emission from the epoch of reionization (EoR) is a challenging task owing to strong foregrounds that dominate the signal. In this paper, we propose a general method, based on the delay spectrum approach, to extract HI power spectra that is applicable to tracking observations using an imaging radio interferometer (Delay Spectrum with Imaging Arrays (DSIA)). Our method is based on modelling the HI signal taking into account the impact of wide field effects such as the $w$-term which are then used as appropriate weights in cross-correlating the measured visibilities. Our method is applicable to any radio interferometer that tracks a phase center and could be utilized for arrays such as MWA, LOFAR, GMRT, PAPER and HERA. In the literature the delay spectrum approach has been implemented for near-redundant baselines using drift scan observations. In this paper we explore the scheme for non-redundant tracking arrays, and this is the first application of delay spectrum methodology to such data to extract the HI signal. We analyze 3 hours of MWA tracking data on the EoR1 field. We present both 2-dimensional ($k_\parallel,k_\perp$) and 1-dimensional (k) power spectra from the analysis. Our results are in agreement with the findings of other pipelines developed to analyse the MWA EoR data.

The Murchison Widefield Array (MWA) has collected hundreds of hours of Epoch of Reionization (EoR) data and now faces the challenge of overcoming foreground and systematic contamination to reduce the data to a cosmological measurement. We introduce several novel analysis techniques such as cable reflection calibration, hyper-resolution gridding kernels, diffuse foreground model subtraction, and quality control methods. Each change to the analysis pipeline is tested against a two dimensional power spectrum figure of merit to demonstrate improvement. We incorporate the new techniques into a deep integration of 32 hours of MWA data. This data set is used to place a systematic-limited upper limit on the cosmological power spectrum of $\Delta^2 \leq 2.7 \times 10^4$ mK$^2$ at $k=0.27$ h~Mpc$^{-1}$ and $z=7.1$, consistent with other published limits, and a modest improvement (factor of 1.4) over previous MWA results. From this deep analysis we have identified a list of improvements to be made to our EoR data analysis strategies. These improvements will be implemented in the future and detailed in upcoming publications.

We present deep polarimetric observations at 154 MHz with the Murchison Widefield Array (MWA), covering 625 deg^2 centered on RA=0 h, Dec=-27 deg. The sensitivity available in our deep observations allows an in-band, frequency-dependent analysis of polarized structure for the first time at long wavelengths. Our analysis suggests that the polarized structures are dominated by intrinsic emission but may also have a foreground Faraday screen component. At these wavelengths, the compactness of the MWA baseline distribution provides excellent snapshot sensitivity to large-scale structure. The observations are sensitive to diffuse polarized emission at ~54' resolution with a sensitivity of 5.9 mJy beam^-1 and compact polarized sources at ~2.4' resolution with a sensitivity of 2.3 mJy beam^-1 for a subset (400 deg^2) of this field. The sensitivity allows the effect of ionospheric Faraday rotation to be spatially and temporally measured directly from the diffuse polarized background. Our observations reveal large-scale structures (~1 deg - 8 deg in extent) in linear polarization clearly detectable in ~2 minute snapshots, which would remain undetectable by interferometers with minimum baseline lengths >110 m at 154 MHz. The brightness temperature of these structures is on average 4 K in polarized intensity, peaking at 11 K. Rotation measure synthesis reveals that the structures have Faraday depths ranging from -2 rad m^-2 to 10 rad m^-2 with a large fraction peaking at ~+1 rad m^-2. We estimate a distance of 51+/-20 pc to the polarized emission based on measurements of the in-field pulsar J2330-2005. We detect four extragalactic linearly polarized point sources within the field in our compact source survey. Based on the known polarized source population at 1.4 GHz and non-detections at 154 MHz, we estimate an upper limit on the depolarization ratio of 0.08 from 1.4 GHz to 154 MHz.

Detection of the Epoch of Reionization HI signal requires a precise understanding of the intervening galaxies and AGN, both for instrumental calibration and foreground removal. We present a catalogue of 7394 extragalactic sources at 182 MHz detected in the RA=0 field of the Murchison Widefield Array Epoch of Reionization observation programme. Motivated by unprecedented requirements for precision and reliability we develop new methods for source finding and selection. We apply machine learning methods to self-consistently classify the relative reliability of 9490 source candidates. A subset of 7466 are selected based on reliability class and signal-to-noise ratio criteria. These are statistically cross-matched to four other radio surveys using both position and flux density information. We find 7369 sources to have confident matches, including 90 partially resolved sources that split into a total of 192 sub-components. An additional 25 unmatched sources are included as new radio detections. The catalogue sources have a median spectral index of -0.85. Spectral flattening is seen toward lower frequencies with a median of -0.71 predicted at 182 MHz. The astrometric error is 7 arcsec. compared to a 2.3 arcmin. beam FWHM. The resulting catalogue covers approximately 1400 sq. deg. and is complete to approximately 80 mJy within half beam power. This provides the most reliable discrete source sky model available to date in the MWA EoR0 field for precision foreground subtraction.

21 cm Epoch of Reionization observations promise to transform our understanding of galaxy formation, but these observations are impossible without unprecedented levels of instrument calibration. We present end-to-end simulations of a full EoR power spectrum analysis including all of the major components of a real data processing pipeline: models of astrophysical foregrounds and EoR signal, frequency-dependent instrument effects, sky-based antenna calibration, and the full PS analysis. This study reveals that traditional sky-based per-frequency antenna calibration can only be implemented in EoR measurement analyses if the calibration model is unrealistically accurate. For reasonable levels of catalog completeness, the calibration introduces contamination in otherwise foreground-free power spectrum modes, precluding a PS measurement. We explore the origin of this contamination and potential mitigation techniques. We show that there is a strong joint constraint on the precision of the calibration catalog and the inherent spectral smoothness of antennae, and that this has significant implications for the instrumental design of the SKA and other future EoR observatories.

We present the 21 cm power spectrum analysis approach of the Murchison Widefield Array Epoch of Reionization project. In this paper, we compare the outputs of multiple pipelines for the purpose of validating statistical limits cosmological hydrogen at redshifts between 6 and 12. Multiple, independent, data calibration and reduction pipelines are used to make power spectrum limits on a fiducial night of data. Comparing the outputs of imaging and power spectrum stages highlights differences in calibration, foreground subtraction and power spectrum calculation. The power spectra found using these different methods span a space defined by the various tradeoffs between speed, accuracy, and systematic control. Lessons learned from comparing the pipelines range from the algorithmic to the prosaically mundane; all demonstrate the many pitfalls of neglecting reproducibility. We briefly discuss the way these different methods attempt to handle the question of evaluating a significant detection in the presence of foregrounds.

We present first results from radio observations with the Murchison Widefield Array seeking to constrain the power spectrum of 21 cm brightness temperature fluctuations between the redshifts of 11.6 and 17.9 (113 and 75 MHz). Three hours of observations were conducted over two nights with significantly different levels of ionospheric activity. We use these data to assess the impact of systematic errors at low frequency, including the ionosphere and radio-frequency interference, on a power spectrum measurement. We find that after the 1-3 hours of integration presented here, our measurements at the Murchison Radio Observatory are not limited by RFI, even within the FM band, and that the ionosphere does not appear to affect the level of power in the modes that we expect to be sensitive to cosmology. Power spectrum detections, inconsistent with noise, due to fine spectral structure imprinted on the foregrounds by reflections in the signal-chain, occupy the spatial Fourier modes where we would otherwise be most sensitive to the cosmological signal. We are able to reduce this contamination using calibration solutions derived from autocorrelations so that we achieve an sensitivity of $10^4$ mK on comoving scales $k\lesssim 0.5 h$Mpc$^{-1}$. This represents the first upper limits on the $21$ cm power spectrum fluctuations at redshifts $12\lesssim z \lesssim 18$ but is still limited by calibration systematics. While calibration improvements may allow us to further remove this contamination, our results emphasize that future experiments should consider carefully the existence of and their ability to calibrate out any spectral structure within the EoR window.

Experiments that pursue detection of signals from the Epoch of Reionization (EoR) are relying on spectral smoothness of source spectra at low frequencies. This article empirically explores the effect of foreground spectra on EoR experiments by measuring high-resolution full-polarization spectra for the 586 brightest unresolved sources in one of the MWA EoR fields using 45 h of observation. A novel peeling scheme is used to subtract 2500 sources from the visibilities with ionospheric and beam corrections, resulting in the deepest, confusion-limited MWA image so far. The resulting spectra are found to be affected by instrumental effects, which limit the constraints that can be set on source-intrinsic spectral structure. The sensitivity and power-spectrum of the spectra are analysed, and it is found that the spectra of residuals are dominated by PSF sidelobes from nearby undeconvolved sources. We release a catalogue describing the spectral parameters for each measured source.

In this paper we present observations, simulations, and analysis demonstrating the direct connection between the location of foreground emission on the sky and its location in cosmological power spectra from interferometric redshifted 21 cm experiments. We begin with a heuristic formalism for understanding the mapping of sky coordinates into the cylindrically averaged power spectra measurements used by 21 cm experiments, with a focus on the effects of the instrument beam response and the associated sidelobes. We then demonstrate this mapping by analyzing power spectra with both simulated and observed data from the Murchison Widefield Array. We find that removing a foreground model which includes sources in both the main field-of-view and the first sidelobes reduces the contamination in high k_parallel modes by several percent relative to a model which only includes sources in the main field-of-view, with the completeness of the foreground model setting the principal limitation on the amount of power removed. While small, a percent-level amount of foreground power is in itself more than enough to prevent recovery of any EoR signal from these modes. This result demonstrates that foreground subtraction for redshifted 21 cm experiments is truly a wide-field problem, and algorithms and simulations must extend beyond the main instrument field-of-view to potentially recover the full 21 cm power spectrum.

Detection of the cosmological neutral hydrogen signal from the Epoch of Reionization, and estimation of its basic physical parameters, is the principal scientific aim of many current low-frequency radio telescopes. Here we describe the Cosmological HI Power Spectrum Estimator (CHIPS), an algorithm developed and implemented with data from the Murchison Widefield Array (MWA), to compute the two-dimensional and spherically-averaged power spectrum of brightness temperature fluctuations. The principal motivations for CHIPS are the application of realistic instrumental and foreground models to form the optimal estimator, thereby maximising the likelihood of unbiased signal estimation, and allowing a full covariant understanding of the outputs. CHIPS employs an inverse-covariance weighting of the data through the maximum likelihood estimator, thereby allowing use of the full parameter space for signal estimation ("foreground suppression"). We describe the motivation for the algorithm, implementation, application to real and simulated data, and early outputs. Upon application to a set of 3 hours of data, we set a 2$\sigma$ upper limit on the EoR dimensionless power at $k=0.05$~h.Mpc$^{-1}$ of $\Delta_k^2<7.6\times{10^4}$~mK$^2$ in the redshift range $z=[6.2-6.6]$, consistent with previous estimates.

We confirm our recent prediction of the "pitchfork" foreground signature in power spectra of high-redshift 21 cm measurements where the interferometer is sensitive to large-scale structure on all baselines. This is due to the inherent response of a wide-field instrument and is characterized by enhanced power from foreground emission in Fourier modes adjacent to those considered to be the most sensitive to the cosmological H I signal. In our recent paper, many signatures from the simulation that predicted this feature were validated against Murchison Widefield Array (MWA) data, but this key pitchfork signature was close to the noise level. In this paper, we improve the data sensitivity through the coherent averaging of 12 independent snapshots with identical instrument settings and provide the first confirmation of the prediction with a signal-to-noise ratio > 10. This wide-field effect can be mitigated by careful antenna designs that suppress sensitivity near the horizon. Simple models for antenna apertures that have been proposed for future instruments such as the Hydrogen Epoch of Reionization Array and the Square Kilometre Array indicate they should suppress foreground leakage from the pitchfork by ~40 dB relative to the MWA and significantly increase the likelihood of cosmological signal detection in these critical Fourier modes in the three-dimensional power spectrum.

The separation of the faint cosmological background signal from bright astrophysical foregrounds remains one of the most daunting challenges of mapping the high-redshift intergalactic medium with the redshifted 21 cm line of neutral hydrogen. Advances in mapping and modeling of diffuse and point source foregrounds have improved subtraction accuracy, but no subtraction scheme is perfect. Precisely quantifying the errors and error correlations due to missubtracted foregrounds allows for both the rigorous analysis of the 21 cm power spectrum and for the maximal isolation of the "EoR window" from foreground contamination. We present a method to infer the covariance of foreground residuals from the data itself in contrast to previous attempts at a priori modeling. We demonstrate our method by setting limits on the power spectrum using a 3 h integration from the 128-tile Murchison Widefield Array. Observing between 167 and 198 MHz, we find at 95% confidence a best limit of Delta^2(k) < 3.7 x 10^4 mK^2 at comoving scale k = 0.18 hMpc^-1 and at z = 6.8, consistent with existing limits.

Detection of 21~cm emission of HI from the epoch of reionization, at redshifts z>6, is limited primarily by foreground emission. We investigate the signatures of wide-field measurements and an all-sky foreground model using the delay spectrum technique that maps the measurements to foreground object locations through signal delays between antenna pairs. We demonstrate interferometric measurements are inherently sensitive to all scales, including the largest angular scales, owing to the nature of wide-field measurements. These wide-field effects are generic to all observations but antenna shapes impact their amplitudes substantially. A dish-shaped antenna yields the most desirable features from a foreground contamination viewpoint, relative to a dipole or a phased array. Comparing data from recent Murchison Widefield Array observations, we demonstrate that the foreground signatures that have the largest impact on the HI signal arise from power received far away from the primary field of view. We identify diffuse emission near the horizon as a significant contributing factor, even on wide antenna spacings that usually represent structures on small scales. For signals entering through the primary field of view, compact emission dominates the foreground contamination. These two mechanisms imprint a characteristic "pitchfork" signature on the "foreground wedge" in Fourier delay space. Based on these results, we propose that selective down-weighting of data based on antenna spacing and time can mitigate foreground contamination substantially by a factor ~100 with negligible loss of sensitivity.

The Murchison Widefield Array (MWA) is a new low-frequency interferometric radio telescope built in Western Australia at one of the locations of the future Square Kilometre Array (SKA). We describe the automated radio-frequency interference (RFI) detection strategy implemented for the MWA, which is based on the AOFlagger platform, and present 72-231-MHz RFI statistics from 10 observing nights. RFI detection removes 1.1% of the data. RFI from digital TV (DTV) is observed 3% of the time due to occasional ionospheric or atmospheric propagation. After RFI detection and excision, almost all data can be calibrated and imaged without further RFI mitigation efforts, including observations within the FM and DTV bands. The results are compared to a previously published Low-Frequency Array (LOFAR) RFI survey. The remote location of the MWA results in a substantially cleaner RFI environment compared to LOFAR's radio environment, but adequate detection of RFI is still required before data can be analysed. We include specific recommendations designed to make the SKA more robust to RFI, including: the availability of sufficient computing power for RFI detection; accounting for RFI in the receiver design; a smooth band-pass response; and the capability of RFI detection at high time and frequency resolution (second and kHz-scale respectively).