Hi Rik, and the Group, thanks for your efforts at comparing the signal as
received here and in Belgium. I think although it seems at first sight that
there is a vast difference between the plots, I can see some similarities
and possibly some reasons for the differing shapes of the plots. The story
starts....

The series of X-ray flares on the 24/25 th of November were accompanied by a
large cloud of plasma. Some of the faster particle effects can be seem I
believe by the increase in deep fades on plots starting on the 26 and 25th
Nov evenings (My dates are plot-start dates) I believe the main shock from
of the CME arrived at about 0330z on 28th. This produced a high level of
particle injection into the D-layer, driving the nightime D-layer into
strong absorption at 136kHz. Note the K-index rises to storm levels at about
midnight on 26/27th.

The radio effects at LF are known to lag the index slightly. The plot for
29th November shows severe attenuation all night.  Propagation during this
period would be by 2 or more like even 3 hops  ( a rather lossy process)
with the 'reflection' layer being the bottom of the D-layer at about 50kms.
At this time CFH is often 'visible' (on a waterfall display) for most of the
day.  An interesting effect now appears. If you look at successive nights
the attenuation increases as the night progresses (at least for the first
4-5 hours) I believe this may be that as the photo-dissociated ions decay,
the reflection layer moves steadily upwards into the D-layer but then the
signal must pass through more of the 'absorbing'  part of the layer to reach
the reflection level. The injected ions have a different nature to the ions
producted by solar UV, do not decay as quickly, and seem to mainly act as
absorbers.

This 'indifferent' low signal propagation continues until 2 December. Note
that the average level of the signal seems to increase slowly each night,
but there are no violent deep fades. The 3rd December starts with the signal
decreasing steadily for the first 4 hours from the initial peak, then a very
rapid fading starts to set in, leading to a good peak between 0500 and
0700z. I believe this is the beginging of what I call the "multipath period"
of the event. The absorption has decreased sufficently for the signal to be
able to significantly penetrate the D-layer and be 'reflected' from two
different levels. This leads to destructive interference ( the 'optical'
definition of interference) and will produce a 'fringe' pattern in the
receiving region. The fringe pattern will move as the altitude of the
reflecting layers change.

A crude calculation suggests that the fringe peaks could be as widely
separated as 25kms (about 10 wavelengths) due to the acute angle of arrival.
I believe in some circumstances, late in this part of the event, some of the
signal may be reaching the E-layer. This might mean the interaction of 3
separate paths ( different numbers of hops? ) and a very complex fading
pattern. The result is a totally different temporal pattern of the fading
and also, depending on the phase coincidence of maybe three paths, a
significantly different peak levels at different sites.

I believe that this multipath fading period of the event could be the best
opportunity for 'catching the wave' in the surfers terms, as 3 coincident
wave paths interfering constructively could yield a 9dB improvement over a
single path signal strength.

My 5th Dec plot shows an series of increasing peaks to a mamoth 46dBu
between 0530z and 0700z. I have some slight quarms about the signal level
comparison for plots before and after the 5 Dec due the removal of the
modulation. It is estimated that this should not make a difference of more
than 3dB, which is considerably less that the massive levels we have seen
recently.

Both Rik's and my plot show a slowly declining series of peaks as the night
progresses. What is interesting is to look in detail and see the fast fading
(around about 2db p-p) with a period of about 10mins, visible on both plots.
The other interesting point is that a lot of the peaks on my plot correspond
to dips on Rik's and vice-versa. I do have all the raw data from my plots
available for re-analysis.

I am certainly interested to see the shape of any plots that can be produced
from differnt locations. It may be possible to shed some more light on this
problem of multipath....and the perenial question of "How many hops?"

The above ideas are based on a study (slightly mind numbing!) of several
papers on LF propagation, but mainly on a tutorial by John Belrose in the
1975 AGARD series, and a number of articles in the 1990 CCIR Reports and
Recommendations. Very little modern work  (after 1940) actually covers this
kind of propagation, over such a long path, in detail, because it only has a
nuisance value to the professionals. My ideas may be controversial, possibly
even totally wrong, but at least its a start, and they do seem to line up
with at least three solar event periods logged during my monitoring..

My thanks to Vaino for a lot of encouragement and advice, and Rik for his
work displaying the plots on his web site.

Cheers de Alan G3NYK   JO02PB
[email protected]