Firstly, I would like to back-track to some of the theories that have appeared in the past. In doing so I do not intend to deride any previous ideas, but to add to them as there may not be one mechanism peculiar to the generation of sporadic-E, but several in combination.
The weather-related theories contain some credible arguments but unfortunately they are not all uniquely seasonally coincidental with sporadic-E. The occurrence of jet streams is not necessarily a mid-year phenomenon; neither are atmospheric gravity waves nor wind-shear. This is not to say that these elements are to be dismissed, as each may generate a potential propagation medium or be a trigger in its own right . There is a possibility that jet-streams could cause some primary ionisation which is later acted upon by some other element that increases its potential for carrying signals at, say, 50 MHz. If this were so, then we might expect some propagation at lower frequencies such as 24 or 28 MHz whenever this occurred. The same argument would apply to atmospheric gravity waves and wind shear. Of course in this case the ionisation produced could support propagation at any time of the year on these lower frequencies. The fact that they preferentially do so in summer requires something else to be added to the equation.
The one factor that is obvious is that the Sun is always in prevalence in the hemisphere where sporadic-E predominates, so I must assume that this is relevant. Peaks of solar activity have no bearing on the added intensity of sporadic-E; in fact the opposite has been observed to be the case. In years of minimum solar activity, occurrence and intensity are usually, but not always greater. However, one factor that follows the sun about the equator is the tropical thunderstorm belt. These storms contain an enormous amount of energy; each when released, is the equivalent of fifty atomic bombs. At any one time there are eighteen hundred storms in progress around the world and one hundred lightning strikes each second. The storms are concentrated mainly over land areas in the equatorial regions which limits their occurrence to three blocks, namely Africa, South America and South-East Asia. Only about ten per cent of the tropical storms occur over the sea. So, as you can see, we have an enormous amount of potential energy to play with and to produce elements which I offer as major contributors in the production of sporadic-E.
I have for some time favoured the lightning theory and more so in the last couple of years due to effects observed during thunderstorms. I have experienced the occurrence of sporadic-E shortly after thunderstorms which were detectable on six metres in a south-easterly direction to be followed after several hours by sporadic-E towards the east. Also some local storms together with a spectacular electrical storm, produced sporadic-E afterwards. This was in Malta, where I was told that they often expect to have E propagation after such thunderstorms. Now before you all get agitated and dash to your radios after a thunderstorm, not every storm will produce sporadic-E. This will be clarified shortly. There may be a combination of events necessary for its production, and the storm may merely act as a trigger mechanism. I cannot state that thunderstorms are the sole generator of sporadic-E, but read on.
An article which appeared in Science in 1994  aroused my curiosity as it set my mind back to the age-old theory that thunderstorms could cause sporadic-E. Some of you may recall the television broadcasts about upward-going lightning strikes, or sprites as they were called, which have appeared since that publication. You may think that these strikes do not go high enough to encounter the E-layer. This is partially correct, as in my view, it is not the lightning itself that causes the ionisation.
The Compton Gamma-ray Observatory, carrying the Burst and Transient Source Experiment (BATSE), was launched in 1991 to observe celestial gamma-ray sources. It has detected numerous cosmic gamma-ray bursts and X-ray sources both persistent and pulsed, together with several thousand solar flares. This is not a surprising revelation as we would expect to gather this information as the norm. The observatory being designed to detect events arriving at the earth from space, the sensors are orientated such that they are able not only to detect the sources of such events, but to pinpoint their origin in space. The observatory was not designed to detect earth-borne events and neither its orbit path nor orbital period are favourable to covering large areas of the Earths surface. Nevertheless, the observers were surprised to detect gamma-rays emanating from the Earths atmosphere. By liaising with meteorological agencies they were able in some cases to relate the emanations to thunderstorms. We are fortunate in this country that being obsessed with the weather, we can obtain detailed forecasts and archive information on our weather. A number of other countries also have fairly comprehensive coverage of their weather but, unfortunately, there exist large areas of the globe that do not enjoy this coverage. This is partly due to the fact that some countries do not need this information in such detail as their weather systems are seasonal and to a large extent predictable. The spacecrafts orbit takes it over such areas so precluding correlation of gamma-ray events and thunderstorms in every case.
The critical part of the detection system mentioned above was that as well as the detectors not being in an optimal positions to see the gamma emissions, their trigger level threshold and gating time were not designed to detect these weaker events. Fortunately another experiment aboard the spacecraft, the Orientated Scintillation Spectrometer Experiment (OSSE) was able to confirm them.
The nature of a downward lightning strike should be familiar to most people. As it strikes the ground it tends to disperse outwards like the roots of a tree just below the surface, preferring the paths of high conductivity. This also produces an ionised layer a metre or so above the surface where the dispersal is taking place. Equally, we can imagine an upward-going lightning strike fanning out as it disperses into the first conducting medium that it encounters. As there is no ground up there to provide a solid conducting medium, it will tend to splay out over a large area and extend upwards to a great altitude. As indicated above, the intensity of a lightning discharge is immense. The enormous energy released by a single discharge will, like any other discharge, produce radiation components. These will include the gamma-rays mentioned already, as well as X-rays which are of the hard variety. By hard we mean they will be in the short-wave part of the spectrum and hence very energetic. Also generated will be Extreme Ultra-Violet radiation (EUV).
To avoid atmospheric absorption, the flashes of gamma radiation have to occur at altitudes above 30km. Discharges observed from aircraft and space shuttles have been seen to emanate from altitudes of 40 to 80km, each discharge having a vertical component extending 10km upwards and 10 to 50 km horizontally. The intense field could not only cause runaway electrons, but bremsstrahlung X-rays  to be generated. To accelerate the electrons sufficiently the field would have to exceed 500V per metre over a path of several kilometres to produce the megaelectron-volt electrons, and hence, the gamma-ray bursts. The glow discharges observed from planes and space-shuttles could be sufficient to generate those field intensities. This, as some of you will by now have surmised, brings us dangerously close to the E-layer!
The ionisation which occurs as we approach sunspot maximum has several causes:
Solar radiation in the Extreme Ultra-violet (EUV) and X-ray wavelengths
Galactic Cosmic rays (GCR)
energetic particles from the Sun and the Earths radiation belts.
in the lower atmosphere, radioactive radiation from rocks.
The EUV and X-rays provide the most important source of photo-ionisation above 60km. Radiation of all wavelengths reaches the F2-layer above 200kms, then the F1 layer at 140-200kms. The deep penetrating part of the EUV and X-rays then encounter the E-layer at 90-140kms and finally, the D-layer at 70-90kms. During the hours of darkness this still continues because of resonant scattering from the illuminated side of the Earth, stellar continuum radiation (affecting the E-region) and galactic cosmic rays (affecting the D-layer).
The chemical processes involved in ionisation are quite complex and I will endeavour to give a fairly simple explanation of some of them. Some of the constituents in the ionosphere are molecular ions of nitric and nitrous oxides and some are atomic ions of oxygen, nitrogen, hydrogen and helium together with water. Now these are the main components, but others less important to the present discussion are present also. Because of gravitational separation the lighter components predominate with increasing altitude. This means that as we rise up through the ionosphere the heavier molecular ions give way to atomic ions with finally, in the F-layer, ions of oxygen, helium then hydrogen and free electrons at the uppermost reaches. Now, those molecules that are present in the upper reaches are subjected to the more intense radiation, and such is the photon energy that dissociation from the molecular state to ionisation of the atoms can occur all in one step. Keep this product in mind as you read later about the gamma-ray and X-ray radiation.
In the F-layer the ion chemistry is primarily the conversion of O+ as a primary ion into secondary molecular ions that recombine with electrons. In the E-layer the primary products are molecular ions which are rapidly consumed in secondary reactions. In this region we also find metallic ions which are reckoned to be the remnants of meteor showers and seem to arrange themselves in narrow layers of one to three kilometres thickness. These are what are assumed to be the carrier for sporadic-E. Metal atoms have low ionisation potentials which means that they can be ionised by exothermic charge transfer reactions alone and also being monatomic, are not easily neutralised. This also implies that subjecting them to only moderate amounts of radiation will cause ionisation, hence the possibility of ionisation by ionospheric wind-shear or jet streams.
Because of the greater density of the atmosphere, at the lower levels the molecules are physically closer and recombination after ionisation occurs quickly. This is the reason why I stated earlier that the gamma radiation has to start above 30km. At higher altitudes the density is much less so that the mean free path lengths between molecules and free electrons is much greater so that recombination takes much longer. The ionised state therefore lasts much longer and replacement ions appear as fast as, if not faster than they recombine, so giving us hours of HF propagation.
If we now look at this mechanism from an earthbound point of view with the ionising mechanisms starting in the atmosphere and the gamma-rays moving upwards through the D-layer and into the E-layer we see a possible generator for sporadic-E. If we have intense, relatively localised ionisation fields being cumulatively produced by successive strikes is it not possible that eventually E-clouds could be formed? There may be a case for assuming that some of these clouds may form in the tropical and sub-tropical areas and drift towards the poles explaining the more pronounced occurrence of E in southerly latitudes. Then again, subjecting a partially ionised layer to a further thunderstorm at a more northerly latitude could be sufficient to trigger propagation.
Now I come to the point where I tentatively offer an explanation of trans-Atlantic propagation on six metres. What if the trans-Atlantic QSOs made on the supposedly multi-hop E mode should turn out to be partially F-layer after all? This may not seem such a wild statement if you will bear with me a while longer.
As I have explained before, normal F-layer propagation is the result of EUV and X-rays. These are hard, which I defined as the short intense penetrative rays that reach down into the D-layer, and the gamma-rays produced by lightning fall into this same category. If the more intense of these rays travel up through the E-layer they will encounter the F-layer with the distinct possibility of localised ionisation. Whether or not this would be sufficient to cause the almost loss-less propagation we often experience is conjecture. Just think of the tiny pin-prick of mans interference with our upper atmosphere; rocket launches are known to have burnt holes in the ionosphere due to the gases of hydrogen, carbon dioxide and water vapour ejected, causing a halving of the electron concentrations through oxygen ions recombining in great quantities (ATLAS-F 1982) . So what price a massive electrical storm now?
There is also a case for a combined effect as the E-layer would also be subject to ionisation and produce a method of propagation akin to E-assisted TEP or ducting across the Atlantic, and incidentally to the east as well if we are to consider all eventualities. The signal, not crossing the equator or region of minimum dip, would not be subjected to the flutter-fading that can happen on that path. The TEP flutter-fading is said to be caused by the rise of a plasma cloud in the equatorial region around sunset thus rapidly altering path-lengths. It is this plasma cloud that supports the TEP propagation.
I have personally not noted any such effect on the trans-Atlantic QSOs, even those which were made late at night and the early hours of the morning, when one would expect this effect as the Sun would be setting somewhere along the path. The propagation which I experienced in October 1987 involving E-assisted TEP to Botswana had some flutter which the southerly stations did not note to any extent and may have been due to the fact that I was located at the northerly extremity of the path.
If we had available ionosonde data during the openings we might see some F-layer reflections but the blanketing effect of the E-layer could preclude detection . We can only measure the conditions at each end of the path as there is naturally a dearth of land between here and North America.
The seasonal aspect of sporadic-E is undoubtedly solar related and the fact that the tropical storm area cycles about the equator adds fuel to the argument. The occasional burst of sporadic-E in December still has to be explained, although a particularly intense event in the southern tropics could result in a northerly drift of a section. It may be coincidental that this occurs about the time in their season as we would expect a maximum in our northern summer season.
Studies, as previously mentioned, have been mainly concentrated in the areas where there is a substantial amateur population and almost twenty-four hour monitoring takes place. For us ever to understand fully the causes of sporadic-E would entail a more detailed coverage of the areas involved. When you take into consideration the area south of the Mediterranean reaching into the north and centre of Africa the number of observers per unit area must be practically nil. The southern African states are sparsely populated by amateurs and, even then, the proportion interested or active on VHF who are able to monitor for more than a few hours a day must be miniscule compared with the large area. The same conditions pertain with the large Japanese concentration to the sparse Australian continent and the North America to South America circuit.
It is questionable whether with the present systems we will ever be able to answer all the questions as to the cause of sporadic-E. One thing that is certain is that you can never be sure that you have found the answer to a problem - you will invariably find an answer, that is, until someone else turns up another answer.
Sporadic E Studies, J Bacon,G3YLA, Rad-Com, May to August 1989
Discovery of Intense Gamma-Ray Flashes of Atmospheric Origin, G J Fishman et al, Science Vol 264 May 1994.
I would like to express my thanks to my colleagues, Dr Isobel C Walker and Dr Roderick Ferguson, at Heriot-Watt University for their help in producing this article.
After submission of this article another source of research in this area has come to my attention in Scientific American, August 1997. In this article by Messrs Mende, Sentman and Wescott they confirm that the discharge events are taking place an altitude of about 90 km on a regular basis. The large potential differences that exist between storm clouds and ground also exist between those same clouds and the ionosphere which infers that the lightning can travel equally well in either direction.
There is much in the way of research to be done in this area and the technical challenges of investigating such short-lived events at such altitudes presents the investigators with enormous problems which will, undoubtedly, take many years to overcome.
To return to the archive page click here