Meteor Scatter Statistics: Theory vs. Practice

Presented here are the results of a three-year effort to collect 6m decodes and compare these experimental results to theoretical predictions. My station, in Claremont, CA is modest: a four-element Yagi at 15 feet, 20 feet of LMR-400 cable to an Icom IC-7100 radio. Audio is fed to a notebook PC running WSJT MSK-144 software. Decodes were received from 99 stations out to 1142 miles. Daily station counts were compiled from PSK Reporter maps providing an interesting byproduct from the main study. Decode SNR’s from three particularly active stations (KE4TH, W0XR and W7OUU were individually averaged. A theoretical model was created in Excel to allow for comparison with the experimental data. For the theoretical model, a sticky problem was that, in general, antenna pointing directions were not known. Thus, averaging was employed. One very cooperative station, KE4TH in Utah, provided extensive test signals with known antenna pointing thus eliminating the uncertainty due to unknown pointing. The theoretical model was based primarily on the sentinel 1997 ITU paper “Communication By Meteor-Burst Propagation.” Another invaluable resource was the QEX paper “The MSK144 Protocol for Meteor-Scatter Communication” by Franke and Taylor. Comparison of the theoretical predictions and the coordinated antenna data average SNR’s for KE4TH yielded agreement to within about 1.5 dB. With the curve of decode probability vs. SNR (Fig. 6) in the Franke and Taylor paper, predicted signal SNR’s can be converted to predicted decode probabilities, thus allowing for ms communication planning a step above just “hoping to get lucky”.