Contacter l'auteur / Contact the author

Recherche dans ce site / Search in this site







HF Propagation tutorial

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

Propagation prediction programs (VII)

Now the past little exercise used old-fashioned tools to do the 5V7A propagation prediction but at a miserably slow pace. Those really drew on three fundamental ideas - the presence of F-region ionization, D-region absorption limiting signal strengths and the geomagnetic field organizing the ionosphere. So using nothing more than the times of sunrise and sunset, those concepts gave a qualitative view of propagation. But without hard numbers, MUFs and signal/noise ratios, that would never meet the needs of the tough decision-making for a DXpedition or a DX contest operation. Hopefuly thanks to propagation prediction programs, we had the opportunity to confirm the opening to Togo late in the eve. 

With computers brought into the matter, the times of sunrise and sunset can be calculated with astronomical precision and DX windows found for working 5V7A on the low bands. The next big problem would be finding the sort of signal strength that could be expected. So a knowledge of the operating modes or hop structures is required, primarily a problem in two dimensions, in the plane of the great-circle path. That sort of thing is done very well by the ray-tracing in the PropLab Pro program or using a VOACAP-based application.

At left the VOACAP interface that uses a complex ionospheric model with tens of functions to predict propagation conditions for a complete communication circuit using not less than 30 parameters in input. This prediction is set for a single path between Brussels (ON) and Brasilia (PY) on September 2002 (SSN = 101, SFI = 146) using a Yagi at the transmitter side with 100 W PEP with a takeoff angle of at least 5, a dipole at the receive site and a QRM level of -150 dBW (or S1). Working in SSB, the S/N required reliability (SNR) is set to 50 dB and the circuit required reliability (SRNxx) to 90%. At right the forecast predict openings between 7 and 14 MHz with signal between S3 and S4, thus weak. Other charts (SNR) confirm the low level of signals with a S/N ratio not higher than 22 dB. Imagine that 2 years later, in 2004 with an SSN close to 25, conditions worstened with a gradual closing of upper bands. Currently only VOACAP-based applications are able to establish forecasts with such an accuracy.

On the higher bands, where MUFs, absorption and E-cutoffs are a concern, computer programs can do a decent job of finding how the ordinary modes would change in the course of a day, say E-hops during the day and F-hops at night as well as mixed modes across sunrise and sunset. But those programs cannot deal with the ionospheric effects from electron density gradients near the terminator or geomagnetic equator so certain modes, like chordal hops and ducting, would not included in their analysis. That's leaves a gap when it comes to having a complete prediction and so computers are fast but will not be as fully quantitative as hoped for in replacing the qualitative efforts used earlier.

As you might expect, the earliest computer program in amateur use, MINIMUF, resembled the scheme with ionospheric maps from the U.S. Dept. of Commerce and just used the control point method for MUFs, via F-region propagation. Neither signal strength nor noise were considered so the method worked best at the top of the amateur spectrum and for very high levels of solar activity. That was unfortunate as amateurs used the same methods at low levels of solar activity, often with misleading or disappointing results.

But MINIMUF fired the imagination of many amateurs and various accessories, including E-layer cutoff calculations, were added to the original code. For example, MINIPROP Version 1 used the F-layer model in MINIMUF and had calculations for E-cutoff and signal strength as well. The early work of Raymond Fricker, MICROMUF 2+ published by Radio Netherlands, was similar but the E-cutoff was regarded as giving values for the LUF, the lowest useable frequency. That's not right as LUF is a D-region matter.

But there was a basic difference between Fricker's MICROMUF 2+ and MINIMUF, how the critical frequency information was obtained. Fricker's F-region algorithm used 13 mathematical functions to simulate the database for critical frequencies from vertical sounding while MINIMUF relied on just one function, adjusted to represent the results of a limited set of oblique soundings.

In another program released at the same time, IONPRED, one of VOACAP precursors, Fricker introduced a novel scheme of hop-testing. Essentially, the program looked at each hop in detail, at the points where the E-layer was crossed and at the highest point where the critical frequency of the F-region was important. So the hop-testing involved determining whether the mode was reliable by seeing if operating frequency was above or below the E-cutoff frequency by 5% and less than the critical frequency for F-region propagation by 5%.

With an initial choice of radiation angle, the path structure could be sorted according to E- and F-hops, depending on the outcome of the tests along the way. Fricker also adjusted the height of the F-region according to local time so hop lengths were not constant along a path. As a result, the path could over- or under-shoot the target QTH. If the error was more than 25 km, another radiation angle was chosen and the process started again.

All output parameters that can be displayed in a model like VOACAP for a specified circuit (using Method 20).

In IONPRED, Fricker also calculated the ionospheric absorption, in dB, and added that to the signal loss due to spatial spreading or attenuation and ground reflections.

Another innovative feature of IONPRED was the use of availability of the path, the number of days of the month it would be open for reliable communication. That was something like the FOT-MUF-HPF idea discussed earlier but in the case of IONPRED, the number of days was treated as a continuous variable in contrast to the upper or lower decile approach with the FOT-MUF-HPF method.

As a result, the path could over- or under-shoot the target QTH. If the error was more than 25 km, another radiation angle was chosen and the process started again. 

Nowadays, the method used by Fricker in IONPRED has been improved upon by the use of mode-searching in the MINIPROP PLUS program and in all subsequent applications. Here, the idea is to work up a number of successful modes and then find the one with the greatest signal strength. With computer speeds in the '80s, Fricker's method was extremely time-consuming, to say the least, but nowadays computer speeds are such that the whole process of mode-searching takes a second or two! Hopefully IONPRED was soon corrected, and as you know mutated in IONCAP then VOACAP.

In a sense, the ray-tracing in PropLab Pro is like hop-testing as it just goes forward for a given choice of radiation angle and the calculation stops if the trace is lost to Infinity or stops in the vicinity of the target QTH. As you might expect, the main problem with that approach is that the hops may either fall short or go beyond the target, making it a slow, iterative process to get the path for RF from point A with point B. Beside that, the user would have to evaluate the suitability of the path, whether the number of E-hops would make it too lossy or otherwise. For that reason, I admire how PropLab Pro goes about a problem but it's too slow for an impatient person like me. 

But we can use the ray-tracing in the PropLab Pro program to see paths in both two or three dimensions. It should be said the 2-D case comes fairly close to dealing with the problem in a proper sense by putting in the appropriate ionosphere for each hop on the path, considering date, time and SSN. But it does not take into account terrain, such as the slope of the ground nor the nature of the reflecting surface. Taking one hop at a time, the calculation does takes into account the change in height of the ionosphere but not any tilts or gradients. That is left for the 3-D case.

The three-dimensional ray-tracing is based on solving equations of motion for the ray path, just like Newtonian Mechanics finds the paths of satellites and spacecraft. There are equations for the path advance along and upward in the great-circle as well as the motion perpendicular to that plane. 

The skewing of paths is small in the HF range and thus, it is usually neglected in ray-tracing. That is because refraction goes inversely as the square of the frequency and electron density gradients across paths that occur in the quiet ionosphere are relatively small. The exception to that statement is the auroral zones where large gradients occur.

But at lower frequencies, like 1.8 MHz in the 160 meter band, the refraction or bending of paths becomes larger because of the lower frequency and other effects become important. In particular, the gyration of ionospheric electrons around the geomagnetic field occurs at a rate which is comparable to the signal frequency. So the entire approach to the ionosphere has to be redone, put in more general terms without any approximations. That complete theory was due to Appleton, is called magneto-ionic theory and has been around for about 60 years.

Polarization and RF coupling into the ionosphere

Among the results of the more general theory are that propagation now depends on the angle between a ray path and the local magnetic field; further, the waves which are propagated in the medium are elliptically polarized, another way of saying they consist of two components at right angles to each other and which have a phase difference between them. Beyond that, there are two modes, with opposite senses of rotation of the electric field vector, the ordinary and extra-ordinary waves.

The simple, linearly polarized waves that are so familiar in the discussion of HF signals are just a limiting case of elliptical polarization, when one of the two components at right angles has a very small amplitude compared to the other one. In magneto-ionic theory, that limiting type of polarization results when signals are sent perpendicular to the magnetic field. The other case is circular polarization, when signals are sent along the magnetic field direction. Then, the two components at right angles are equal in amplitude and out of phase by 90.

Those features of propagation were evident in the early days of ionospheric sounding as two echoes were returned for each signal sent upward, the ordinary and extra-ordinary waves, and you will see them on any ionograms that you may inspect. So magneto-ionic theory is a part of the reality of radio propagation. But, for DXers, there is something of a happy simplification as over long distances, the extra-ordinary wave is heavily absorbed and only the ordinary wave needs to be considered.

There is another interesting aspect to propagation down on the 160 meter band, the coupling of RF into the ionosphere. As you know, there is a polarization to the waves emitted by an antenna and on 160 meters, vertical antennas are used most often. That is due to the wavelength being so long that most horizontal dipoles cannot be placed very high, in terms of wavelengths, and thus suffer from high radiation angles, being the so-called "cloud warmers".

Intensity of the horizontal component of the geomagnetic field.

Now in magneto-ionic theory, the polarization of a wave changes continously in the ionosphere as it is propagated through the geomagnetic field. But there are two limiting polarizations, typically at altitudes around 60 km, where the wave enters the ionosphere near point A and where it leaves the ionosphere near point B. When worked out in detail, the theory says that there will be a signal loss, in dB, at entry because of any mismatch between the wave polarization from the antenna and the limiting (elliptical) polarization at entry point A.

For example, signals going in the E-W direction from a vertical antenna at the equator are poorly coupled into the ionosphere because of the polarization mismatch, with vertically polarized waves going against the horizontal field lines. Similarly, there may be signal loss at the exit point B due to any mismatch between the limiting polarization on exit from the ionosphere and the polarization of the antenna at point B.

As indicated, magneto-ionic theory is quite complicated, with elliptically polarized waves and all that, but for signals going from point A to point B, we need not concern ourselves about what goes on high up in the ionosphere between those two points, only the antenna types and the limiting polarizations at the endpoints of the path. That makes life a lot simpler.

Another point about this frequency range; signals can become trapped in the electron density valley above the E-region at night. Thus, if they enter the region, they may be reflected back and forth between the bottom of the F-region and the lower limit at the top of the E-region. That means they'll rattle back and forth between those altitude limits like a ball sliding down a smooth trough. Only if the walls of the trough change in height can the ball get out or, equivalently, can signals get out of the duct if the lower ionosphere changes. In that regard, ducting is undoubtedly responsible for the long-haul DXing done on 160 meters as it avoids repeated ground reflections and traversals of the lower ionosphere which absorb signals at a very high rate.

Reference Notes

A review of various propagation programs can be found in the QST issues for September and October 1996.

Note by LX4SKY. An updated list of programs is available on this site, in the next article : Review of HF propagation analysis and prediction programs, that list not less than 50 applications.

The above discussion gives a very brief summary of the principal aspects of magneto-ionic theory, as it applies to propagation. An analytical summary of the theory is given in Davies' recent book, Ionospheric Radio; however, it really requires a strong background in electromagnetic theory at the level found in university courses in physics and engineering. It should be noted that the method of the theory has a broader application as it represents the first steps toward the study of plasmas in the solar system and in out space.

A discussion and some quantitative aspects of polarization loss on 160 meters are given in my article in the March/April '98 issue of The DX Magazine. In addition, a fuller discussion of magneto-ionic theory and 160 meter DXing is given in Top Band Anthology, published recently by the Western Washington DX Club.

Radio propagation fundamentals

We turn now to other aspects of propagation, from predictions to those circumstances which may disrupt propagation and make predictions go awry. But in doing that, a bit of history would help chart the course.

Dr Hidetsugu Yagi presents his ultimate DXer's gun, the best antenna design ever made for DXing. Still another japanese product of quality, Hi !

First, radio is more than 100 years old now (in 1901 Marconi sent successfully the first wireless message from England to the U.S.A) and the course of events has been onward and upward, in frequency and into the ionosphere. Thus, the earliest signals were down in the kHz region and now technology has advanced to the point where amateurs are operating in the GHz part of the spectrum. But it has been a steady advance in frequency and as we know now, that means signals going higher and higher into the ionosphere as their effective vertical frequency increased.

Amateur operations start in the medium frequency (MF) range with the 160 meter band, around 1.8-2.0 MHz. If one looks into the ray-traces for that band, it is clear that signals in normal communications circumstances stay below the 200 km level most of the time. Of course, ionospheric absorption on that band is so great that DX operations are attempted only on paths in full darkness.

Going to the high frequency (HF) range, 3 - 30 MHz, signals go higher toward the F-region peak around 300-400 km and darkness becomes less of a necessity near the top part of the spectrum. In fact, solar radiation is needed to bring the level of ionization up to the level required for propagation.

Historically, in the time that operating frequencies rose, the range of DX contacts increased and it became apparent that the solar cycle played a role in propagation. Moreover, various disturbances became apparent. So the early '20s had amateurs opening up trans-Atlantic operations and that was commercialized in the late '20s with the advent of radiotelephone circuits to Europe. In that time, it was found that the communication links failed during geomagnetic storms. Those could last for days but there were also strange blackouts that lasted anywhere from just a few minutes up to an hour. In 1937, those short wave fadeouts (SWF) were found to be associated with solar flares. Moreover, it was becoming apparent that the disruptions to magnetic storming came a day or so AFTER solar flares.

From all that, it became clear that the sun was a major player in the field of radio propagation and scientists began looking into the details. The SWF problem was fairly simple, just being the release of electrons in the ionosphere from the photoelectric effect of solar X-rays. The magnetic storm effect was a more subtle problem as it implied some slower process, not X-rays moving across the solar system at the speed of light. In that regard, those geophysicists who studied the earth's magnetic field proposed that there was a stream of matter sent out from the sun and then its encounter with the geomagnetic field was the triggering mechanism. From the time delays between flares and storms, first estimates were made of the speed of the solar matter. More than that, they could not say at the time.  

Today thanks to satellites we are monitoring the sun activity 24 hours a day as well as the status of the geomagnetosphere. At left, variation of the geomagnetic field components on April 6, 2000 by 1600 UTC when the solar wind shock wave impacted the earth geomagnetosphere. This spectacular event produced bright aurora in the forecoming hours. A right two cross-sections of the geomagnetosphere observed in the ecliptic plan on April 6, 2000 during an after that event. Click on images to run the animations. Documents SPIDR-NGDC-NOAA and PIXIE/S.M.Petrinec.

Now that brings up the question of just how far out geomagnetic field lines extend from the earth. Of course, that goes to the model of the geomagnetic field in use at the time. That was, in simple terms, the sort of thing you get if you stuff a bar magnet into the earth and look at how the field lines extend past the surface of the earth. In short, the model back in the '40s and '50s was that for a centered dipole field that was tipped with respect to geographic coordinates, the dipole axis piercing the earth's surface at 79.3 N, 71.8 W at the north pole and the south pole through the corresponding antipodal point.

That was the field used when the first Pioneer space shots took place after the IGY, an experiment looking at the strength and orientation of the earth's field as the spacecraft moved out, away from the earth. That flight produced a REAL surprise, with data showing the earth's field varying slowly and in an orderly fashion as the spacecraft moved outward but then suddenly, when it reached something like 8 earth radii, the field became weaker and less organized, almost random in its orientation. Clearly, the orderly dipole field no longer described the situation at those distances, giving way to the presence of an interplanetary magnetic field. And what was previously considered as empty space, except for meteoritic dust and debris, was also found to contain of plasma (protons and electrons) that was streaming away from the sun.

Now, before exploring that extreme, we should look at the dipole field and see what could be expected from it. As you know, say from your high school physics course, the field lines pass out of the southern hemisphere and then after going out some distance, they return and enter the northern hemisphere of the earth. That was the classical picture; so let's see what it says, at least until we get into trouble with the Pioneer data.

Now the magnetic dipole has a system of coordinates of its own, related to the direction of its axis relative to the geographic axis and equatorial plane. With the dipole orientation given above, one can work out the magnetic coordinates of any point on the earth. For example, my location at 48.5 N and 122.6 W is one that corresponds to 54.4 N, 62.1 W in the dipole coordinates. OK?

But let's look at the dipole and its field lines. They go out from the southern hemisphere and come back down into the northern hemisphere. But how far do they go out? That would be important when it comes to thinking about the collision of solar plasma and the dipole field, suggested by the geomagneticians. It's not hard to work out where the magnetic field lines cross the plane of the geomagnetic equator and there is a simple relation between that distance and the magnetic latitude where the field lines start:

L = 1 / cos j

with j as the magnetic latitude and L is the distance, measured in earth radii (Re). 

Now if you conjure up the image of a dipole, surrounded by its magnetic lines of force, you can see that low-latitude field lines do not go out very far from the surface of the earth. But it's a different story for high latitude field lines and if worked out, we obtain the following table :

Mag Lat (degs)  Distance (L in Re)

     10                  1.03

     20                  1.13

     30                  1.33

     40                  1.70

     50                  2.42

     60                  4.00

     70                  8.55

     80                  33.2  

So the high latitude field lines are the ones in harm's way when it comes to the collision between the plasma coming from the sun and the earth's field. And, by the same token, the low-latitude field lines that go out only short distances from the center of the earth are pretty well protected from the direct effects of the collision between solar plasma and the geomagnetic field. Of course, that fits with your operating experience, paths going across the polar cap are far more subject to disruption than those going to low latitudes.

Before getting to the nature of the various propagation effects that originate on the sun, we should note briefly that the view of the earth's field that I gave in the introduction is not quite the full story. In particular, it was suggested that the solar wind blowing by the obstacle of the geomagnetic field is like the flow problem of a bullet in air, but now with the bullet (geomagnetic field) fixed and the air (solar wind) in relative motion. So it was suggested (and verified) that a bow shock in the solar wind was out there in front of the magnetosphere as displayed at right.

Now, to carry the aerodynamics a bit further, it was suggested that the position of the bow shock would vary, moving closer to the earth at higher speeds of the solar wind.

And that proved to be the case, obtained by satellite observations after the original work with Pioneer I. But the geomagnetic field is a bit different than a hard obstacle and it was expected that the field could be compressed at times, particularly if the solar wind came at it as a sudden blast. And, as you guessed, that is the case as shown by magnetic sensors on geostationary satellites. During some severe magnetic storms, those satellites report conditions which put them right in the interplanetary magnetic field, showing that the magnetosphere has been compressed by the solar wind and that the magnetopause was temporarily inside 6.6 Re. Absolutely amazing!

Now, having told you about the troubles of geomagnetic field lines, think back a bit to what I said earlier: they are the things which hold your precious ionospheric electrons in place! So maybe all those disruptions in propagation during magnetic storms are not all that surprising, with field lines being pushed around by the solar wind.

There's more to magnetic storm effects than just compressing the field lines in front of the earth. As I suggested way back in the second page, field lines on the front of the magnetosphere can be dragged into the magnetotail. In that process, the ionospheric electrons of the F-region on those field lines are removed from the front of the magnetosphere and, in essence, are distributed on much longer field lines on the rear of the magneto-sphere. On both counts, the high-latitude F-region suffers a loss in ionization and critical frequencies in the affected regions are reduced. Of course, the sun shines, day in and day out, so with some magnetic quiet, solar illumination will restore the regions and communications across those high latitudes returns to normal.

Those words of explanation will have to suffice as the problems of the magnetosphere are quite complicated, with unfamiliar or non-classical ideas, and are best left for the magnetospheric physics-types to wrestle with. We need not get enmeshed in the details, only be able to recognize when there's a problem and consequences that will follow. In that regard, the records of magnetometers at high latitudes are our best bet as they give vivid portrayals of the storms that develop, thanks to simultaneous, yet secondary effects which result. There, I am thinking of the aurora, both optical and radio, as well as the current systems which build up during a disturbance initiated by the solar wind.

Again, the details need not concern us but the main features are what we note: optical emissions coming from above the 100 km layer, VHF reflections off of auroral displays, ionospheric absorption of signals going across an active auroral zone and strong magnetic disturbances observed on the ground from the current systems which develop along the ionized region. More on this in the the last chapter.

Research Notes

A good historical account of the early days of radio can be found in the first chapter of McNamara's book, "Radio Amateurs Guide to the Ionosphere" published by Krieger Publ.Corp. in 1994. And it's a good book too. Get a copy if you are serious about radio propagation.

Add also the link to this web pages, History of amateur radio, appreciated by ARRL's staff and CQ Magazine's editors too.

Last chapter

Geomagnetic disturbances

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

Back to:


Copyright & FAQ