I made a spectrogram of the Morse transmission using a
transform width equal to the 180 second dot length, advancing
the transform by 60 seconds each time

 http://abelian.org/vlf/tmp/29499_140303_180_60_rect.png

As expected, the signal is now 'in the yellow'.

Markus advised me that using a shorter transform aids readability
even though it discards some S/N.   Here it is with 90 second
transform advancing by 45 seconds,

 http://abelian.org/vlf/tmp/29499_140303_90_45_rect.png

and sure enough it is now quite easy to read the Morse letters.
There is one 'B' where the dash is broken into two dots by
a noise spike.

I tried Hann and some other windows but rectangular seemed to do
best.

The 10mHz shift was very effective and ought to ID the signal
even if the trace was very faint on a spectrogram.   Fingers
crossed for some reports from further into Europe.

The signal was showing no sign of fading out by 08:00 UT so I
guess it would still come through for another hour or maybe
two.

--
Paul Nicholson
--