On the noise estimation in super dual auroral radar network data

The Super Dual Auroral Radar Network (SuperDARN) currently consists of more than thirty high-frequency (HF, 3–30 MHz) radars covering mid-latitude to polar regions in both hemispheres. Their major task is to map ionospheric plasma circulation which provides information about the interactions between the solar wind and the near-Earth's space plasma environment. One of the major factors defining radar data quality is the signal-to-noise ratio (SNR), which requires an accurate characterization of the HF noise. The standard SuperDARN data analysis software uses the SNR as part of a set of empirical procedures designed to remove low-quality data from further analysis. In this study we found that the currently used empirical algorithm systematically underestimates the noise level by up to 40%. Based on comparison of theoretical and observational noise statistics, we resolve this issue by designing and validating a procedure for accurate background noise level estimation. We then propose a simple SNR threshold to replace the existing criteria for excluding low-quality data. In addition, we show that several aspects of the radar operational regime design, as well as short-lived anthropogenic radio interference, can adversely affect the quality of the noise estimates at selected radar sites, and we propose ways to mitigate these problems.