Contacter l'auteur / Contact the author

Recherche dans ce site / Search in this site







HF Propagation tutorial

Composite image of an aurora yielding a power of several GW recorded in the visible and UV part of the spectrum by the Polar satellite.

by Bob Brown, NM7M, Ph.D. from U.C.Berkeley

Effects of the ionization (II)

Right now, there's more than enough ionization up there to support DXing on the low bands, 160 to 40 meters. But the higher bands are still pretty spotty, mainly across low latitudes or in brief bursts of solar activity. But 10 meters will return; trust me.

The discussion so far has dealt with the creation of ionization and how various frequencies in our spectrum make out as far as propagation and absorption are concerned. There's one problem with that discussion, the omission of how, in the course of time, ionization reaches the steady-state electron densities overhead.

So let's turn to that but do it as simply as possible. That means we'll focus on electrons, positive and negative ions. The solar UV and X-rays create those from the oxygen and nitrogen molecules in our atmosphere. I can say it is a big, complicated ion-chemistry lab up there but we'll stay at the generic level, nothing fancy, just electrons and positive ions.

In simple terms, there is a competition between the production and loss of ionization, just like your bank balance where depositing paychecks and paying bills are in competition. So for us, there's a certain number of electrons created per second in a cubic meter of air in the ionosphere by the solar radiation and whatever the number of electrons present, some are being lost by recombining with positive ions to form neutral atoms or molecules again. If the two, gain and loss, are equal, there is a steady-state of ionization; otherwise, there will be a net gain or loss per second from some cause or other.

I haven't said so but the atmosphere is only lightly ionized, say one electron or positive ion per million neutral particles. So electrons have a greater chance to bump into a neutral particle (like in ionospheric absorption) than a positive ion, to recombine to make a neutral atom or molecule. And, of course, there's a vast difference in those rates between the lower parts of the ionosphere, the D-region below 90 km and the F2-region above 300 km. So electrons created by solar UV would be gobbled up rapidly in the D-region but linger on for the better part of a day up in the F2-region.

Solar particles (protons and neutral atoms) hitting head on the upper atmosphere of the earth over the equatorial region. Doc NASA-GSFC.

Good illustrations of the fast processes are found nowadays, solar flares illuminating half the earth with hard X-rays (like those in the 1-8 Angstrom range). They penetrate to the D-region, release electrons which rapidly transfer wave energy to the atmosphere. As soon as a flare ends, the sudden ionospheric disturbance (SID) or radio black-out ends as the electrons in the D-region recombine rapidly and signal strengths return to normal.

The lingering on of electrons in the F2-region is responsible, in part, for the fact that there's still ionization and propagation in hours of darkness. In short, electrons at high altitude recombine slowly after the sun sets. But there's more to the story than that, the role of the earth's magnetic field. Let me explain.

The earth's atmosphere is immersed in the geomagnetic field so any charged particles, say ionization created by solar UV, will then experience a force from their motion in the field. For electrons, that means they will spiral around the field lines when released by UV and not fly off in any direction to another location, higher or lower in the ionosphere. In the propagation business, that is called geomagnetic control, meaning that the earth's field largely determines the distribution of electrons in the ionosphere.

True, the solar UV creates them and they are most numerous where the sun is overhead but they are held on field lines and linger on after dark, to our great advantage.

But the earth's field also creates problems, especially for the low-band operator. It turns out the gyro-frequency of electrons around field lines is about 1 MHz and comparable to frequencies in the 160 meter band. Thus, a more general approach has to be made in the theory of propagation at that frequency, adding the effects of the earth's field on ionospheric electrons. The results are quite complicated, with elliptically-polarized waves on low frequencies where linearly-polarized waves were the story earlier on high frequencies. That is a subject in itself and has to be left for a rainy day. But those are not the only ways that the earth's field enters into the propagation picture. Stay tuned.

Earlier, I said there were other ways that the earth's field enters into the propagation picture. But that's sort of getting ahead of my development so let's backtrack a bit and look at the historical picture.

The study of geomagnetism goes back more than 100 years, well before the advent of radio. It was known that the occurrence of magnetic storms was related to the solar cycle and, by the same token, it wasn't long before it was realized that HF propagation was related to it too. The two really came together about 70 years ago when commercial radiotelephone service was established across the Atlantic Ocean. Then it soon became apparent that there were disruptions in service during magnetic storms. You can find all that discussed in the I.R.E. journals in the early '30s.

In that period it was thought that the ionosphere was the result of solar UV, the photons reaching the earth 500 seconds (~8 minutes) after leaving the sun. And while magnetic storms were known to disrupt radio propagation, there was no obvious connection as experience showed magnetic storms occurred a couple days after the flash phase of a large flare on the sun. True, there was the idea of solar material, electrons and protons called "plasma", approaching the earth after a solar outburst and engulfing the geomagnetic field, even compressing it. But the two effects from plasma and UV seemed separable just because of differences in time-of-flight across "empty space" that were associated with the two effects.

But all that changed with the Space Age when it was found that solar plasma was out there all the time, the solar wind, and that it blew past us with differents speeds, 200-1,200 km/sec, as well as different particle densities and even carried magnetic fields along. But for us earth-bound souls, the big surprise was that the solar plasma distorted the earth's magnetic field, essentially taking some field lines on the sunward side and pulling them back behind the earth to form a magnetotail. Moreover, with the solar plasma coming at us, it became clear that a ordered, dipole field did not go on forever, only out to 8-12 earth-radii in the sunward direction and even that depended on solar activity.

So what does this have to do with propagation, you ask. Well remember I said geomagnetic control of the ionosphere means that electrons are held on magnetic field lines, making the earth's field something of a reservoir for ionospheric electrons. But if field lines can be distorted, that would surely affect the density of ionospheric electrons gyrating around them and propagation.

The worst-case scenario is when field lines are dragged way back into the magneto-tail by an increase in solar wind pressure, taking ionospheric electrons with them. That field configuration is sketched crudely below where two compressed field lines are shown in front of the earth, in the solar direction, and two magnetotail field lines in the anti-solar direction as displayed below.

Structure of the geomagnetosphere

0 - Earth

1 - Ionosphère encircling the Earth (not shown)

2 - Plasma sphere

3 - Van Allen belts

4 - Plasma sheet (internal magnetosphere)

5 - Magnetopause

6 - Geomagnetic tail

7 - Polar cone

8 - Geomagnetopause

9 - Solar wind

Clic on image to enlarge. Document BAS.

That would mean a depletion of electrons at F2-region heights and drastic reductions in MUFs, affecting propagation. Fortunately, that fate is reserved primarily for sites at high latitudes, around the auroral zones and poleward.

What I described was what takes place during a major geomagnetic storm. The recovery is a slow process as ionospheric electrons have to be replaced in the usual way, by solar UV and day by day while the sun is up. So it can take days for the bands to recover when a strong magnetic storm reduces MUFs by a large fraction.

Now to be practical again, magnetic activity on earth is caused by interactions of the solar wind out there at the front of the geomagnetic field. The field region around the earth is called the magnetosphere so we're talking about effects on high latitude field lines that go out to the magnetopause, the dividing surface between terrestrial and interplanetary regions. But it must be recognized that this sort of thing is not toggled on and off; it is going on all the time as the solar wind sweeps by. It is just a matter of degree. But how to deal with it in DXing?

The clue comes from an interaction within the magnetosphere, local electrons being accelerated to high energies and then spiralling down field lines to make visible aurora and ionization at E-region heights. Those events are triggered by solar wind interactions at the magnetopause and accompanied by horizontal currents in the E-region that show up in magnetic observations on the ground. It then becomes a matter of using the strength of the local magnetic effects at auroral latitudes, with K- and A-indices like those you hear about on WWV or can check on DX Summit Cluster (OH8X) or DXHeat, to judge the energy input from the solar wind as displayed below.

To read : The Shape of the Sunspot Cycle (PDF), D.Hathaway and al., NASA

WWV data provided by NOAA.

To bring this to a conclusion, good propagation conditions are found when there is a strong UV input to the ionosphere and low magnetic indices, the 3-hour K-index less than 4 and the daily A-index less than 25. Dreadful propagation conditions were found recently in the magnetic storm of August 27 when K reached its limit, 9, and the planetary average of the A-index was 112. But it could have been worse! However, let's look at the brighter side next time, how signals get from A to B.

Let's leave a curved ionosphere to later and do some "Flat-Earth Physics" to see how signals get from point A to point B. For that we start with a simple model of the ionosphere in which the electron density increases upward and peaks at about 300 km altitude. That's something like a night-time ionosphere.

Now it may seem strange but one can draw an analogy between the flight of a baseball and RF going up through that ionosphere. For the baseball, high school physics teaches you how to calculate how high a baseball would go if thrown vertically upward. In college, the ball is thrown or hit upward at an angle. The method is the same in both cases: the ball rises until the increase in its potential energy in the earth's gravitational field is equal to the kinetic energy it had from its initial vertical motion.

Neglecting friction, the baseball's path is a parabola that is symmetrical about its highest point and the ball returns to the ground at the same angle to the vertical as it was launched. While not really parabolic in shape, the flight of RF through that simple ionosphere is similar, reaching a peak altitude that is determined by the frequency and launch angle, symmetrical about the peak and returning to ground at the same angle. How does that happen? Let me explain.

The flight of a baseball and the path of RF in a simple ionosphere are determined by gradients, of the gravitational energy of the ball in the first case and the electron density distribution in the second one. There is a gradient of either of those quantities if there's a change in value with altitude, say gravitational energy or electron density greater at higher altitudes than lower at altitudes. The gradients are responsible for the bending or curvature of the paths in the both cases and, numerically, they are given by the change in value per km change in altitude. OK?

In spite of all the "Home Run Fury" these days, let's leave the baseball part of the analogy and focus on what happens to RF. So we see that hops, with RF rising and then returning to ground, are the result of the vertical gradient of the electron density in the ionosphere. On reflection at ground level, angles of incidence and reflection are equal and the path continues upward again.

But there can be horizontal gradients as well, say across the terminator where there is more ionization on the sunlit side than the side in darkness. So if RF signals were sent initially parallel to the terminator, one would expect the RF to be bent away from the sunlit side, with its higher level of ionization, and toward the darkness. Right? That's skewing, pure and simple, with the RF refracted away from the region of greater ionization.

The height a baseball reaches depends on its speed and direction; for RF, that translates into frequency and launch angle. But one sees that from different arguments. Let me add a few words there. At any height in the ionosphere, there are electrons and positive ions. If, by mystical powers, you could grab a handful of each and then pull them apart, they would be attracted to each other by the electrical forces between unlike charges and on release, they'd swish back and forth, carrying out an oscillatory motion. The frequency of that motion is called the plasma frequency and it depends on the density or number of particles per unit volume, N.

For the ionosphere, where ionization increases with height, the plasma frequency increases too. For our night-time case, the peak electron density in the F-region might correspond to a plasma or critical frequency of 7 MHz for the F-region. Now vertical ionospheric sounding shows that pulses of RF below 7 MHz would be returned to ground while any above 7 MHz would penetrate the peak of the ionosphere and go on to Infinity.

At left, the refraction of a radio wave in a plane stratified medium. Since the plasma frequency (critical frequency) increases with height, n becomes smaller and the wave gradually bends toward the horizontal. At right a side-view of propagation paths of SuperDARN rays in the ionosphere at a frequency of 12.45 MHz for elevation angles from 5 to 50°. The dashed line indicates the electron density profile. Documents realized by Andreas Schiffler on a home computer using a ray tracing program solving a set of differential equations.

For oblique propagation, we have to find the effective vertical frequency of the RF, just like the vertical component of the baseball's velocity. For RF, it's found the same way, multiplying the frequency by the cosine of the zenith angle at launch. So, in the "Flat Earth" approximation, 7 MHz RF launched from ground at 30° above the horizon (or 60° from the vertical) would have an effective vertical frequency of 3.5 MHz. OK, the "baseball analogy" would say that the RF going off obliquely would rise until it reached a height where the local plasma frequency is 3.5 MHz and then return to ground. Of course, it would be on a curved path, the RF would be moving parallel to the earth's surface at the top of the path and returning to ground at the same angle as when launched, just like the baseball problem.

In baseball, there's friction and that changes the flight of a baseball. We don't put "friction" in the RF problem. Instead, the electron density at a given height may vary along the path direction, say become smaller. That would serve to "tilt" levels of the ionosphere upward and weaken the density gradient. As a result, there would be less refraction or bending after the peak altitude than before, and that tilt serves to increase the length of a hop and change the RF angle on return to a lower value.

In reality one would expect some change in electron density along any path, increasing as a path goes into sunlit regions or decreasing when going into the dark. So even if nothing else changed, one would not expect hop lengths nor radiation angles to always remain exactly the same all along a path.

The above approach, equivalent to mirror reflections of RF, is Newtonian in the sense that the analogy treats a RF path like that of a particle (baseball) and not a wave. When the Maxwellian or wave approach is carried out, one finds that refraction is the same except that the effects vary inversely with the square of the wave frequency. So in a given part of the ionosphere, 80 meter RF paths are refracted or bent much more than 10 meter RF paths, either vertically or horizontally. OK?

MUF and RF attenuation

MUF/LUF and signal strength estimation for July 25, 2004 calculated with DX ToolBox for a power of 100W PEP from ON to TA.

OK, now we have the idea of critical frequencies and hops so it is no big deal to work out how propagation on a path may be open or closed for DXing on a given frequency. But to do that, we need at least map of where the RF is headed and an idea of how many hops would be involved. Beyond that, some ionospheric details are required, the critical frequencies along the path at the date and time in question.

If one gets into the mathematics of all this, it turns out that hops via the F-region may reach about 3,500 km and half that via the lower E-region. So using those ideas, one can estimate the hop situation, at least as long as there is not a mixture of E- and F-hops. So consider a path from my QTH in the Northwest to London, some 7,500 km in length. That would work out, to a first approximation, to 3 F-hops of 2,500 km each.

Now what about the critical frequencies at the peaks of the hops; how high are they and what bands might be open to me, say at 1200 UTC?

To answer that question, one would need some sort of database, an array of observations from which an estimate could be obtained by interpolation, or a mathematic simulation of the database that could be used to calculate the critical frequencies. Actually both methods are used in modern propagation prediction programs but either way, appropriate numerical values could be obtained for the peaks of the hops. But what to do with that data?

For a one-hop path, the matter is simple; the effective vertical frequency of the RF that is launched must be less than the critical frequency for the path to be completed. No problem. For two hops, the effective vertical frequency of the RF must be less than the SMALLEST of the critical frequencies of the two hops to have a complete path. And the operating frequency that gives the highest effective vertical frequency that can complete the path is called the Maximum Useable Frequency (MUF) for the path at that time and for the corresponding solar conditions.

But the path from my QTH to London involves 3 hops; what's the story there? Historically, the idea was handled like the 2-hop path, using the critical frequencies at the first and last hop to determine the MUF. The idea was that if propagation failed, it usually would be due to conditions at one end of the path or the other. Anyway, this is called the "control point" method and is used in most simple propagation programs. More sophisticated approaches would use critical frequencies at each and every hop and the lowest would be the important one that limits propagation.

Propagation charts for the 20-m band on August 2, 2004 22:00 UTC for paths from respectively New York to London and vice versa. In both case of course the number of hops is 3 as listed in the blue upper bar. The estimation map are different because at that time the position of the Sun is simply not the same over both QTH. Estimation calculated with DX ToolBox by LX4SKY.

It should be noted that the control point method would be quite satisfactory for MUF calculations so long as the critical frequency of the middle hop is not less than those at either end of the path. That would be the case for paths going across the more robust ionosphere at low latitudes where the sun is more overhead during a day. But MUF calculations using two control points for high latitude paths, like from the Northwest to London, can be misleading as the critical frequency for the middle hop (over Northern Canada and Greenland for the path to G-land) could be lower than at the end points and thus propagation not supported across the entire path using the MUF from control points.

The MUF calculations play an important part in propagation predictions but it must be remembered that signal strength, in comparison with noise, is an important consideration. As noted earlier, ionization and MUFS are more important for the higher ends of the amateur spectrum and signal/noise considerations for the lower end. In any event, for communication a path must be open or available and signals must be readable and reliable.

All of the discussion up to this point has dealt with propagation from a conventional viewpoint - determined by the ionosphere that is overhead and, in turn, one controlled by the level of solar activity. Obviously, propagation is a complicated process and it may seem a bit naive but we try to make all our predictions on a given date using using databases which rest on only a few numbers - sunspot number and magnetic indices. It is not surprising that predictions are not 100% reliable. Such high expectations would deny the variability of the original data input from ionospheric sounding and not reflect the roles of dynamic solar variables.

So far, this brief summary of the principal points that are involved in HF propagation has been largely centered on words and concepts. More advanced topics require a good deal of graphics so I will make appeal from time to time to a figure or two in one or more of the reference books given earlier. While figures are the best way to convey some of the material, I will also try to put the ideas in simple words that will carry most of the meaning.

To me, the study of ionosphere and propagation changed markedly with the advent of the Space Age. Thus, with the International Geophysical Year (IGY) in '57, high-altitude balloons, rockets and satellites began to probe the regions where only radio waves had been before. So the "Photochemical Era", where solar photons and atmospheric processes were thought to control the dynamics of the ionosphere, gave way to the "Plasma and Fields Era" we're in now, where the interaction of the solar wind with the earth's field and the atmosphere are the controlling factors for propagation.

In simple terms, hams no longer look out the window for their local weather, determined by the day, time and season, but now turn to the Internet to get a daily report on the Space Weather. In a sense, propagation and DXing just became less mysterious and even more interesting. That's what we'll be pointing toward in next pages, preparing for all the details in last section.

It's no secret that success in DXing means getting signals to and from a DX station and also having them heard and read at both ends of the path. But between those two ends, a lot of things happen in the ionosphere and some of them seem like well-kept secrets. So the hope is some of that can be dispelled by the discussion which follows. But we need a beginning and the question is where to start. Let's take the easy way and cover old ground first, the matter of ionospheric absorption that was discussed in the second session.

So we go back to the idea that RF excites the electrons in going across the ionosphere, jiggling them at the wave frequency. And they collide with nearby atoms and molecules, transferring some energy derived from the waves to the atmosphere. That's how absorption takes place, mostly down in the D-region. But there's a frequency dependence we should talk about now, how absorption varies with the operating QRG and with height, since the collision frequency of the electrons is not constant; instead, it decreases with height and that's a help. So it's clear now that ionospheric absorption is a little more complicated than I first let on back in the introduction.

But one can get a handle on it by looking at the extremes, low in the D-region, say around 30 km where the collision frequency is greater than any of the frequencies in our spectrum. In that circumstance, collisions happen so often the electrons never have a chance to pick up any energy from the passing RF.

On the other hand, at high altitudes, say around 100 km, collisions are quite infrequent and the electrons re-radiate most of the energy they acquire and transfer very little to the atmosphere by collisions.

So it is in between, where wave and collision frequencies are comparable, that electrons take up RF energy efficiently and then promptly deliver it over to the atmosphere. So with collision frequency falling with increasing altitude, 28 MHz RF is absorbed at lower altitudes than 3.5 MHz RF, as shown at right. That graphic illustrates something that DXers know already, lower frequency signals are absorbed more than higher ones but it shows where it all happens. That's news, at least for some.

To go beyond that qualitative result, one must have an analytical form to represent the curves, call it F(f,h) for frequency f and height h. Then multiply F(f,h) by the number N of electrons per cubic meter at height h and include the physical constants to give the right units, dB/km. When all is said and done, the result is:

Attenuation (dB/km) = 0.046  x  N  x  F(f,h)

But that is only at one place, where the electron density is N. Our DXer's signal is attenuated by ALL the electrons encountered along the RF path from point A to point B so that means we need to know something about the propagation mode, the distribution of electrons and add up the results, km by km along the path.

That's a tall order but when it's done, it will enable our DXer to find just how much of the radiated power P survived in going from A to B. But whether our DXer can be heard still depends on how well the attenuated signal compares with the noise power getting to the receiver at B. But I'm getting ahead of myself.

30 MHz riometer absorption. Document IPS.

The crude graphic shown at right can help in understanding a lot of simple things. For example, it is possible to identify various ionospheric disturbances just by the absorption they produce. One approach is to use an HF receiver to monitor the galactic radio noise coming in vertically on 30 MHz. Galactic noise gets right through the F-region as 30 MHz is above its critical frequency, even at equatorial latitudes where it might reach 20 MHz in a solar cycle. That instrument is called a riometer, for Relative Ionospheric Opacity Meter, and they are generally deployed at high latitudes where ionospheric disturbances are most common.  The graphs displayed at right for example have been recorded over the Australian Antarctica base in 2004.

So now, if some disturbance increases the electron density in the D- or E-region, we see that the galactic noise signal will be attenuated and indicate the presence of a disturbance.

But there are disturbances and then there are disturbances. So the graphic also tells us that anything that disturbs the lower D-region will produce strong attenuation of the galactic radio noise and, electron for electron, the attenuation will be much less if the disturbance produces ionization at much higher altitudes.

The first case would be for polar cap absorption (PCA) events, like we all experienced in May of '98. In those events, solar protons produce lots of ionization around 40-50 km altitude and give rise to tens of dB of additional absorption on 30 MHz and blackout oblique communication paths going across the polar caps. Auroral events, say associated with magnetic storms, give rise to strong ionization above 100 km, where the graphic shows the absorption efficiency is much lower, and auroral absorption (AA) events show only a few dB of absorption of galactic noise on 30 MHz. Of course, there are other differences in the two types of events, how the ionization is distributed in latitude and longitude and how long they last. More on that later.

Severe solar X-ray flux recorded on November 4, 2003 by GOES satellites. Document SEC.

One last disturbance, again something that was within our recent experience with all the flare activity in the summer of '98, is sudden ionospheric disturbances (SID) from bursts of solar X-rays. Those X-rays, in the 1-8 Angstrom range discussed earlier, were incident on the sunlit hemisphere of the earth and literally swamped the normal distribution of ionization at low altitudes, giving intense absorption of signals going across the sunlit region. But experience shows, and the graphic indicates, that the effects were worst at the lower ends of the spectrum, wiping out 75 meter operations but having little effect on 28 MHz, except perhaps for some solar noise bursts associated with the flaring.

If this would be quite academic, perhaps, were it not for the fact that one can use the Internet to see these events in action or shortly thereafter. Thus, records from the X-ray Flux Monitors on the GOES 10 and 12 satellites are available at SEC/NOAA, giving more meaning to the idea of an SID.

We'll get to that later on but the main thing for us in the records is that plots for 0 degrees tell what is going down into the atmosphere, making more ionization and affecting the ionosphere. The 90-degree plots involve particles trapped in radiation belts and are more colorful than informative.

While disturbances come and go, affecting our ability to work DX, we really need to know something about the normal situation, say the distribution of ionospheric electrons with height as well as latitude and longitude. That is a big order but, believe it or not, it can be contained in one computer hard disk. I'm talking about the International Reference Ionosphere (IRI), the summary of decades of ionospheric sounding all over the world. So it will provide data on the robust part of the ionosphere at low latitudes where the sun is more overhead and the mid-latitudes where the ionosphere is more seasonal in its properties.

But the model is not reliable at high latitudes, say from below the auroral zones and poleward. That region is under the constant influence of the solar wind and electron densities are highly variable, even hour by hour. So that model has its limits. But to bring the model to life, one needs a mapping program to show the vertical and global distribution of ionization. Fortunately, we now have such a program available to amateurs, the PropLab Pro program from Spacew. I'll have more to say about that next time.

Reference Notes

- A better representation of the relative absorption efficiency per electron as a function of height and frequency in the D-region is found in Figure 8.1 in my book, The Little Pistol's Guide to HF Propagation.

- And a more detailed discussion of the analytical form, F(f,h), is found in Section 7.4 (Ionospheric Absorption) of Davies' book, Ionospheric Radio (IEE Electromagnetic Waves Series, Vol. 31), beginning on p. 214. Also, the variation of collision frequency with height is given in Figure 7.5 on p. 215.

Next chapter

Distribution of ionospheric electrons

Page 1 - 2 - 3 - 4 - 5 - 6 - 7 - 8 -

Back to:


Copyright & FAQ