Extreme Values in Short-Term 20 m Sequential Matched WSPR Observations

Automated amateur radio networks, such as WSPRnet, daily compile data on hundreds of millions of radio contacts. This wealth of information is valuable for researchers exploring and forecasting High-Frequency (HF) propagation and its correlation with solar phenomena. A prerequisite for meaningful investigations is a comprehensive understanding and documentation of the inherent variability present in the data. Prior investigations highlighted the extreme short-term variability in SNR reports from 20-meter sequential matched observations, variability in excess of usual distributional assumptions.  Here, we describe and model those extreme observations.  Using descriptive statistics and logistic regressions, we provide evidence of some temporal and spatial patterns associated with the extreme SNR values and develop predictions for their occurrence.