Radio observations of the 2015 solar eclipse

Pieter-Tjerk de Boer, PA3FWM web@pa3fwm.nl

On March 20, 2015, there was a solar eclipse in Europe. Unfortunately, the Netherlands wasn't on the path of totality, and unfortunately it was totally clouded in my town of Enschede. However, the effects of this eclipse on radio propagation could clearly be observed. This page documents my observations, using my WebSDR receive setup at the University of Twente and adapted software.

The following is a wide waterfall picture showing all radio signals between 0 and 1700 kHz (from left to right), and between 04:00 and 11:00 UTC (from top to bottom). The maximum eclipse was at about 09:37 UTC. The brightness of each pixel represents the average signal strength during 1 minute in a 2 kHz bandwidth.
(A very large version of this diagram, showing the full 24 hours, and the full spectrum from 0 to 29 MHz, can be viewed here.)

[waterfall diagram showing 0 - 1.7 MHz, 04:00 - 11:00 UTC]

One clearly sees the temporary enhancement of many mediumwave broadcast signals (between 550 and 1600 kHz) around the time of the eclipse. This can be compared to the top part of the picture, which shows typical night-time propagation.

Also some longwave broadcast signals, between 150 and 290 kHz, were enhanced. This was particularly notable on two signals on 189 and 207 kHz from Iceland, which is right on the other side of the totality region as seen from this receiver's location. I made short audio recordings from these Iceland radio stations on both 189 kHz and 207 kHz; especially the latter is very strong and clear.

By way of exception, the signal on 216 kHz, from Roumoules in southern France, had a dip at the moment of the eclipse.

Here's a plot of the signal strength of these longwave signals (horizontal UTC time, vertical relative strength in dB):
[signal strength of longwave signals]
And these are some of the mediumwave signals:
[signal strength of mediumwave signals]
Note that although all four have a peak at the eclipse time, the peaks are not precisely simultaneous, which must be related to the geographical location of the path to these transmitters w.r.t. the path of the moon's shadow.

Here's a map showing the transmitter locations, their great-circle paths to the receiver, and lines of maximum eclipse times (the tickest lines are at each half hour, the thinner ones every 10 minutes, and thinnest ones at 1 minute intervals; for reference, the thick line crossing Iceland is at 09:40).
[map]
There's a clear correlation between the peak times in the plots, and the timing of maximum eclipse as shown on the map.
[longwave signals plot] [mediumwave signals plot]

The peaks of the Iceland longwave signals are at 09:41 (189 kHz) and 09:44 (207 kHz), differing by 3 minutes, which matches nicely with the difference between their paths' maximum eclipse times, which are 09:38 and 09:42. Apparently, the ionosphere lagged the eclipse here by about 2 minutes.

As for the mediumwave signals, the peaks are less pronounced, even after quite a bit of smoothing in the above plots.
The Scottish signal (810 kHz) has a quite narrower peak, at 09:40, lagging mid-path maximum eclipse by about 3 minutes.
The Czech signal on 954 kHz is the last one to peak, between about 09:49 and 10:00, lagging maximum eclipse on this path by about 8 minutes.

More analysis will be posted here later.