The Little Pistol's Guide to HF Propagation - K9LA

129The Little Pistol's Guide to HF Propagation

The Little Pistol's Guideto HF Propagation

by Robert R. Brown, NM7M

$10.00


Published byWORLDRADIO BOOKS

P.O. Box 189490Sacramento, CA 95818

The Little Pistol's Guideto HF Propagation

by Robert R. Brown, NM7M

2 The Little Pistol's Guide to HF Propagation

The Little Pistol’s Guideto HF Propagation

Robert R. Brown, NM7M

Copyright March, 1996

WORLDRADIO BOOKSP.O. Box 189490

Sacramento, CA 95818


Publisher’s noteRadio signals — in the air everywhere......sometimes.As we drive home from work, the thought that comes to mind is, “I hope the

band is open tonight.”It may turn out to be a night of great signal reports, or you might wonder if

the receiver is broken.When I was a small child, I came across some pre-WWII copies of QST. In the

DX column were pictures of amateurs in exotic locales of the world. It seemed wecould talk to people who lived in those lands that Lowell Thomas presented intheaters, and that the National Geographic wrote about.

I was hooked! I still am. The adrenaline level jumps a little when a “rare”country is on the air. With a child’s imagination back then I could picture thoseradio waves striking the ocean on the way to Sumatra. Others hit and were re-flected from a moonlit Iowa cornfield on the way to Europe.

I’m older now, and I still think of radio waves that way.Worldradio is very pleased to present the writings of a true expert (who writes

so we can understand it) regarding the when, why and how radio waves really getfrom one spot to another. Bob Brown, NM7M, brings to this task the knowledge ofa true scholar and the ability to communicate.

— Armond Noble, N6WR, Publisher


CONTENTSPREFACE

PART ONE

Chapter 1 Introduction .................................................................................... 9 Chapter 2 Elementary Considerations ........................................................ 11 Chapter 3 The Sun and its Radiation ........................................................... 13 Chapter 4 Elements of Propagation ............................................................. 15 Chapter 5 Solar Data ...................................................................................... 20 Chapter 6 Geomagnetic Data ........................................................................ 25 Chapter 7 More on Propagation ................................................................... 28 Chapter 8 Now to Signal Strength ............................................................... 34 Chapter 9 Noise............................................................................................... 38 Chapter 10 Now the Details ............................................................................ 41 Chapter 11 Moving On..................................................................................... 46 Chapter 12 Now to Three or More Hops ...................................................... 52 Chapter 13 Going Further and in More Detail ............................................. 58 Chapter 14 Paths Beyond 10,000 km.............................................................. 65 Chapter 15 Long-Path Propagation ............................................................... 73

PART TWO

Chapter 16 Toward a New Era........................................................................ 83 Chapter 17 Solar Wind and Flares .................................................................. 89 Chapter 18 Magnetic Storms and Aurorae ................................................... 93 Chapter 19 Propagation Conditions, Bad and Good................................. 100 Chapter 20 Historical Features ..................................................................... 105 Chapter 21 Statistical Considerations .......................................................... 107 Chapter 22 Solar/Terrestrial Environment ................................................. 112 Chapter 23 HF Propagation in a Nutshell .................................................. 116

References ............................................................................................................ 120Subject Index........................................................................................................ 122


When it comes to DXing, it is my firmopinion that a person can contact any place onthe face of the earth with just a 100-Watt rigand a tri-bander. Having said that, let me ad-mit I haven’t worked all the countries in theworld myself, but I do sport a WAZ certificateon the wall of my shack and think it supportsmy statement. But if that’s not enough to con-vince you, I have a friend who made the DXCCHonor Roll with just 5 Watts QRP. That wasDan Walker, WG5G; his accomplishment is amatter of public record. So, being a Little Pis-tol can have its rewards, but it takes operatingskills and an ability to take best advantage ofwhat the bands have to offer in the way of HFpropagation.

I can’t help you with the former; thatcomes with experience. But I think I can shedsome light on the latter. In that regard, I havemy own way of introducing and developingthe subject. And since DXing is an intellectualpursuit, I like to think my approach has thekind of logical framework needed to guide aDXer from simple beginnings up to the presentunderstanding of the ionosphere and HFpropagation. So if you want to know and un-derstand what’s going on “upstairs” whenyou’re in the pursuit of DX, read on.

Please don’t think I’m putting you “onhold” in the early sections dealing with solarand geomagnetic data. I do that just to makesure you’re aware of where to get the informa-tion you”ll need to take a measure of currentconditions, anywhere from quiet to stormy.That done, you can judge when it’s best to re-treat from the bands and read on in what I’vewritten or, with promising indicators and well-founded hope, push ahead and go for another“New One,” leaving your reading to anotherday.

Now since DX is synonymous with dis-tance, the logic of my approach in the discus-sion takes you along, one ionospheric hop at a

time, out to about 10,000 km distance, a quar-ter of the way around the globe.

But those efforts, often aimed at reachinginto inaccessible zones, are affected by the stateof the high-latitude ionosphere, and that bringsthe discussion into the Space Age, with the newmodel of the earth’s magnetic field. From there,it goes on to the solar wind, solar flares andcoronal streams, holes and mass ejections, theemphasis being their role in disturbing HFpropagation rather than the theory and intri-cacies of solar physics.

When you’re finally done with what I’vewritten, I think you’ll see better how HF propa-gation really works and it won’t seem quite sopuzzling. And you’ll quickly recognize the el-ements of order (or chaos) in what’s going onat any given time. Once that’s true, you’re onyour way to being able to deal appropriatelywith the circumstances, knowing better whento “go for it” or when to retreat from the bandsand regroup for another day.

But more to the point, you’ll be in a posi-tion to do some real sharpshooting — DXingwith a specific target in mind. That’s the ulti-mate intellectual side of DXing and when suc-cessful, you can not only point with pride atthe QSL on your wall, just like a Big Gunwould, but also offer a critique of the propaga-tion conditions that were involved. With that,you can show how reason triumphs, even inDXing, and never be mocked for mindless pur-suit of DX, blasting away with disregard ofconditions and others on the band.

At this point, having said all the above,I’d suggest you read up through the sectionson solar and geomagnetic data. When that’sdone, you can figure out what to do next, readon or get on the bands and chase some DX. Butwhatever your choice, see it through; it pays.

Guemes Island, WAJanuary 1995

PREFACE


PART ONE


In writing about propagation for the LittlePistol, I’m really writing about two differentsubjects, the first being the ionosphere and theother DXing. So I’ll have to deal with those top-ics, perhaps just one at a time, but in the lastanalysis, the emphasis will be on the iono-sphere and how it relates to the Little Pistol’sDXing.

Now I started in ham radio before WWIIand I don’t recall that term, “Little Pistol,” be-ing used back then. True, there was great in-terest in DXing in the late ’30s, DXCC gettingits start in ’37. But I don’t remember peoplestrutting around, being pointed at as being a“Big Gun,” or anyone feeling sheepish for be-ing just a Little Pistol. Perhaps my own per-sonal sphere of interaction was too small or toolocal. On the other hand, I did know of someoperators who’d qualify now as being a BigGun, say Reg Tibbetts, W6ITH, in the San Fran-cisco Bay Area and the late Don Wallace,W6AM, in Southern California.

When you come right down to it, the dif-ference between a Little Pistol and a Big Gunis one of degree, not interests. So I would haveto think that the operative term in describing aLittle Pistol is “modest,” say one having a mod-est station or modest accomplishments whenit comes to DXing. But modest is not the termto use in talking of the Little Pistol’s ambitions;they’re larger and grander than that, even tothe point of wanting to become a Big Gun inthe future.

Before getting to what the Little Pistolneeds to know about the ionosphere to makethat transition, let’s look at his modest stationand see what’s typically involved. First, theLittle Pistol usually operates on the upperbands at the “barefoot” level, running about100 Watts RF output from a transceiver withsome “bells and whistles.” And the antennasystem would be something like a triband Yagiwith three elements, up there between 30 and

50 feet above ground. A 40-Meter antennawould be included in the setup, say a trappedvertical or a single inverted-V with its apex uparound 35 feet.

By contrast, a Big Gun would have a stoutlinear amplifier after a flashy transceiver, withbells and whistles galore. As for power levels,they could be from 100 Watts output right upto the full legal limit, at the flip of a switch.And antennas would be monoband Yagis,stacked or otherwise, on several towers at 50ft. or more above ground. There might be a 40-Meter Yagi included as well but certainly 80-Meter and 160-Meter dipoles, well up off theground, for low-band DXing. So I think youcan see there’s nothing modest about a BigGun’s setup.

As for accomplishments, the Little Pistolhas the “DX Bug” and may already have gonebeyond the first step, getting the DXCC Award,and is looking to eventually have enough DXin the log to qualify for the DXCC Honor Roll.That takes “a heap of doing,” as the sayinggoes, but every Little Pistol knows of a Big Gunup there in “Honor Roll Country” and usuallyhas a role model close at hand. So it becomes anumbers game for the Little Pistol, the differ-ence between the number of DXCC countriesactually confirmed and the number neededcurrently to qualify for the Honor Roll. Howthat game is played out is another matter, inpart related to our main topic, HF propagation.But there’s another important side related topaperwork.

In that regard, no matter what job I’ve had,it seemed like I always had to push a broom atone time or another and I also had to keeprecords. So it is for the Little Pistol, keeping theshack in good order and maintaining an accu-rate log of stations worked, QSLs received oroutstanding, and following the DX scene, per-haps from reading a DX bulletin or monitor-ing the DX packet cluster. To those chores, I’d

1: INTRODUCTION


have to add another — keeping tabs on thecurrent state of solar/geomagnetic conditions.When taken with the knowledge of the iono-sphere that I’ll be laying out next, those will bevital to the Little Pistol’s advancement towardbecoming a Big Gun.

As for what information about the iono-sphere will help the Little Pistol in reaching hisgoals, I plan to offer a full and serious discus-sion of the subject for the Little Pistol to think

about and study. However, I will I assume theLittle Pistol has gone through the basic ideas,as presented in the ARRL Handbook, to get tohis present position in the DX hierarchy. Thatspares me going through what I’d term “intro-ductory material.” Since DXing is truly an in-tellectual pursuit, my task is to build on thatkind of fundamental discussion and bring outadditional, finer points that will help in DXing.


2: ELEMENTARY CONSIDERATIONS

The Little Pistol knows that RF is reallyelectromagnetic waves, time-varying electricand magnetic fields which can propagatethrough a vacuum with the speed of light. Butin the LP’s efforts at DXing, those waves areactually propagated through a material me-dium, the weakly ionized portion of the upperatmosphere, termed the ionosphere. Indeed,with good fortune, the LP’s signals go up intothe ionosphere and then returned to earth at agreat distance from where they started.

In some circles that’s called ionosphericreflection but a better term is refraction, orbending of the ray path followed by the waves.All that results from the waves going througha medium which consists of free electrons andwhose number density, so many electrons percubic meter, increases with increasing altitude.And the Little Pistol knows there are severalregions that make up the ionosphere, the F-re-gion being the highest and peaking in numberdensity around 300-400 km, then the E-regionaround 110 km and finally the D-region below90 km.

The present knowledge of a complex iono-spheric structure is in contrast to the idea ofone region of ionization developed by Chap-man in the early ’30s. Then, as shown in Fig-ure 2.1, it was assumed that ultraviolet (UV)quanta from the solar spectrum would encoun-ter an increasing number of ionizable atomsand molecules on penetrating the atmosphereand would produce ionization at an increas-ing rate.

But solar quanta would suffer absorptionin the process and the radiation intensity woulddecrease closer to the earth’s surface. As a re-sult, a peak in the production rate of ioniza-tion would be reached and then decline atlower altitudes.

Actually, UV photons in the solar spec-trum have enough energy to not only ionizemolecules of nitrogen and oxygen, as men-tioned above, but also to dissociate them into

Figure 2.1 A schematic representation ofthe production of ionization in the at-mosphere.

their constituent atoms. That last statementwouldn’t involve too much of a stretch of theLittle Pistol’s imagination but to get into thevarious chemical reactions which might takeplace is another matter. Needless to say, “What-ever can happen will happen” in the part ofthe upper atmosphere reached by solar UV andthe LP shouldn’t be surprised at the result, anatmosphere whose chemical composition var-ies with altitude.

More specifically, from ground to the 100-km level, the atmosphere is well mixed by con-vection and turbulence and has a rather homo-geneous composition of the major constituents,nitrogen (78%) and oxygen (21%), but withsome trace constituents (such as water vapor,carbon dioxide and ozone) which are not ho-mogeneously distributed. But above the 100-km level, atomic oxygen, resulting from disso-ciation of molecular oxygen by the solar UV,assumes greater relative importance with in-creasing altitude.

All this discussion of the neutral atmo-sphere may seem strange to the Little Pistol.“After all, radio propagation involves charged


particles — electrons — right?” I won’t say“Wrong!” But how about “Not exactly?” Itturns out that ionospheric electrons findthemselves in the middle of a great big chem-istry laboratory up there with all the electrons,positive ions, atoms and molecules mixing itup, as it were.

Indeed, under the right (or wrong) cir-cumstances, some of the atoms, molecules orions can affect the electron density adversely,to the detriment of propagation so importantto the Little Pistol. So it behooves the LP to payattention to those ideas and come to grips withthe current view of the ionosphere; to wit, thehigher reaches of the ionosphere contain freeelectrons and positive ions from nitrogen mol-ecules, oxygen molecules and atomic oxygenand as well as minor, but important, positiveions formed by ion-molecule interactions.

I think it’s fair to say the Little Pistol un-derstands that the density number of electronsin the ionosphere varies with altitude but theLP probably doesn’t know that the heightvariation of the electrons depends on a com-petition between rates of their production andloss. But the LP could understand that the ratesfor those processes vary with altitude, withsolar illumination (for production) and changesin chemical composition (for loss).

Now the Little Pistol understands thematter in our atmosphere and ionosphere areparts of our environment, held to the earth byits gravitational field. And the LP’s DXingprobes the ionosphere, showing changes in thedegree of ionization from time to time. But thechanges in the neutral atmosphere escape theLP’s attention as it’s not directly responsive toHF radio signals. Again, imagination takes overand the LP knows that both parts of the envi-

ronment vary as a result of changes in solar il-lumination.

But what the LP probably doesn’t under-stand is that the ionosphere is also controlled,even influenced, by the earth’s magnetic fieldand its variations. In the static circumstance,geomagnetic control of the ionosphere may bea new one for the LP, at least when discussedin any detail. The LP is far more likely to haveheard of, even experienced, the effects of iono-spheric storms which go along simultaneouslywith geomagnetic storms and affect HF propa-gation. Those circumstances are not beneficialto LP’s DXing and their explanation will turnout to be rather complicated. So the LP had bestpay attention to those discussions as becom-ing a Big Gun may depend heavily on dealingwith those occasions.

I want to conclude this section on elemen-tary considerations by having the reader notethat so far, everything has all been qualitativein nature. Of course, some quantitative aspectshave been implied — the LP’s operating fre-quencies, changing seasons and times of day,transmitter powers and antenna gains. Thoseare all characteristic of operations at the LittlePistols’s QTH.

Now what needs to be added to the dis-cussion are the quantitative aspects, particu-larly those which go with DX paths across theearth’s surface, their location relative to thegeomagnetic field and the effects of solar illu-mination, rapid on a daily basis and slow overa solar cycle. That additional discussion willturn knowledge of the ionosphere into a quan-titative decision-making tool and, I believe,serve the Little Pistol well so let’s get to it,starting with the sun.


The solar radiation that we’re all familiarwith is in the visible portion of the spectrum,with wavelengths from 400-700 nanometers (1nm = 1E-9 Meter), in going from violet to red.Using more classical units, the visible spectrumlies between 4,000 and 7,000 Angstroms (1 A =1E-8 cm). The fact that the frequency of radia-tion is the speed of light (300,000,000 meters/sec) divided by the wavelength allows one tofind the frequency for visible radiation, around1E+14 Hz or almost a billion times greater thantypical HF Amateur Radio frequencies (3-30MHz).

While the visible spectrum goes throughthe “atmospheric window,” the radiation re-sponsible for creating the ionosphere does not,being absorbed at high altitudes as it ionizesand dissociates atoms and molecules in theupper atmosphere. For our purposes, dealingwith HF radio propagation, the amazing thingis that only a tiny fraction, about .001%, of thesolar radiation incident on the earth’s atmo-sphere is the source of energy for ionosphericprocesses. Ponder that for a moment or two!One-thousandth of one percent — simplyamazing!

That energetic part of the spectrum liesbelow the visible spectrum, in the extreme ul-traviolet (EUV) and X-ray range. And for theradiation to ionize or dissociate constituents inthe atmosphere, its energy must be equal to orgreater than the ionization potential or bind-ing energy of the atoms or molecules. But thoseare atomic and molecular processes and so wehave to digress a moment to give the energyunit that’s appropriate to use, the electron-Voltor eV, instead of the Joule, the unit for mechani-cal energy in the M.K.S. system.

Now an electron volt is the energy that anelectron gains in going through a potential dif-ference of 1 Volt, that is to say it’s equal to theelectron’s charge Q (1.6E-19 Coulomb) multi-plied by the potential difference V, 1 Volt or 1Joule/Coulomb, so 1 eV equals 1.6E-19 Joule.

Going to the energy associated with photonsin the solar spectrum, we use the Planck’s Law,the energy of a photon being given by Planck’sConstant h (6.6E-34 Joule-sec) multiplied by itsfrequency f in cycles per second (or Hz). Andrewriting that expression to use eV for units ofenergy, Planck’s Law would give the energy ofa photon in eV as the product of 4.1E-15 andthe frequency in Hz or 1240 divided by thewavelength in nanometers.

Now a quick look at one’s high schoolchemistry text shows that the ionization poten-tial for hydrogen, the simplest atom of all, is13.6 eV. As you might expect, the ionizationpotential of atomic oxygen, found in the up-per atmosphere, is about the same and that foratomic nitrogen is about 1 eV higher. For theimportant diatomic molecules in the upper at-mosphere, those of oxygen, nitrogen and ni-tric oxide (NO), their ionization potentials are12.5 eV, 15.5 eV and 9.5 eV, respectively. On thatbasis, the part of the solar spectrum that’s ef-fective in creating the ionosphere is wave-lengths of about 100 nm (1000 A) or shorter.And the same energy range would apply to thephotodissociation of oxygen and nitrogen mol-ecules into their constituent atoms.

At this point, the discussion has becomequantitative, dealing with the parts of the so-lar spectrum which ionize and dissociate at-oms and molecules in the upper atmosphere.But it should be noted that the “quiet sun” notonly emits EUV photons in the 100 nm range,as was discussed above, but also X-rays atshorter wavelengths, say 10 nm or even 1 nm.So, they, too, contribute to ionizing the upperatmosphere but to degrees in accordance withtheir relative strength in the solar spectrum andthe abundance of targets in the upper atmo-sphere.

All those remarks apply to what is termedthe “quiet sun” but as the Little Pistol wellknows, there are times when the sun is ratherdisturbed, sending out bursts of radiation

3. THE SUN AND ITS RADIATION


across the spectrum, say radio noise, light, andX-rays. Obviously, the energetic photons insuch outbursts may change the amount of ion-ization in the ionosphere, affecting the LP’sDXing. So the question comes up, “Does theLittle Pistol himself have any way of knowingabout variations of the flux of photons, ener-getic or otherwise, from the sun?”

The answer to that question is a qualified“Maybe!” and depends on the part of the spec-trum under discussion. So if the Little Pistol isoperating at the top of the HF spectrum, sayon the 10 Meter band, solar noise bursts mightbe heard, sort of a “whooshing” sound, and ifstrong enough, “solar QRN” could even inter-fere with the LP’s pursuit of DX. Another pos-sibility is when the LP is in a contact where thepath goes across the sunlit part of the earth;then, increases in the EUV and X-ray flux thatgo with solar flares could make signals fade,maybe even to the extent of being the victimsof a blackout. As for increases in visible light,as an indication of a disturbed state for the sun,that is extremely unlikely. The light output ofthe sun is remarkably constant and only a fewsolar flares have been seen in white light; muchmore sophisticated optical devices are needed

to observe changes in the visible portion of thesolar spectrum.

The above response to the question ofvariations in the flux of solar photons assumesthat the Little Pistol is the only observer. Ofcourse, that’s not the case as there’s a wholeindustry, as it were, just watching and listen-ing to the sun. And the information that’s col-lected is disseminated promptly and widely soif the LP misses some activity on the sun, it’llbe available as part of the solar data for the day.So if the LP wants to know about any changesin the solar radiation which might affect DXing,it’s just a matter of getting hold of the solarrecord and looking at it. And that’s the recordkeeping that I alluded to earlier; in the LP’scase, it’ll be part of the discipline that goes withDXing, keeping tab of solar/terrestrial condi-tions.

Before getting to those matters, we haveto start the discussion of propagation and thenbring forth some its finer points for the LittlePistol. When that is done, we can get to how itis affected by changes in the various physicalparameters that go to make up an appraisal ofsolar/terrestrial conditions.


There’s no need to tell the Little Pistolwhat goes to make up HF propagation as LPhas been practicing that already. Thus, LPknows signals are brought back to earth, hopafter hop, by ionospheric refraction. More tothe point, as far as the LP’s DXing goes, is toshow more about how ionospheric hops takeplace and the factors which influence them.Some of the factors are right here at groundlevel — signal loss and polarization changeswith ground reflections — while others are athigher altitudes, up in the regions that go tomake up the ionosphere.

Now the Little Pistol’s RF leaves groundlevel as expanding waves, the vertical distri-bution of signal intensity depending on thetype of antenna the LP is using and the groundsurface nearby. The usual representation ofantenna patterns, both vertically and horizon-tally, gives the signal strength in various direc-tions, along ray paths which are perpendicu-lar to the advancing wave fronts. At this pointantennas and their patterns are not our con-cern; rather, it’s the ray paths of RF throughthe ionosphere and how they’re affected in go-ing through or approaching regions of the iono-sphere.

To begin that discussion, the Little Pistolneeds to understand that ionospheric refrac-tion depends on the reradiation of RF by theelectrons that go to make up the ionosphere.In the simplest terms, radiation by the LP’santenna results from oscillatory or acceleratedmotions of electrons in his antenna. As the RFwaves expand outward and encounter freeelectrons in the ionosphere, they’re set in mo-tion at the same frequency by the electric fieldE associated with the wave. Like the electronsin the LP’s antenna, the free electrons radiateRF because of their oscillatory motions. Inshort, ionospheric electrons reradiate RF pass-ing by, resulting in an advancing wave frontwhose direction depends on the spatial distri-bution of the electrons.

Now if the density of free electrons in theionosphere did not vary with height, there’dbe no refraction, i.e., no bending or change ofthe ray directions that describe the advance ofthe RF wave front. There’s an optical counter-part to that; light advancing in a straight linein going through a transparent substance witha constant index of refraction. In the iono-spheric case, the electron density increases withaltitude and unless the frequency of signals istoo high, the RF is continuously (but not con-stantly) refracted back toward earth.

It should be noted that statement applieswhether the ray path is ascending or descend-ing through the ionosphere and results fromthe fact that the electron density increases withaltitude above the earth. A crude way of say-ing the same thing is that RF is always beingbent away from regions of higher electron den-sity. The optical counterpart, involving a me-dium of atoms and molecules with bound elec-trons, is just the opposite, light always beingbent toward regions of higher index of refrac-tion.

Returning to the ionosphere, Little Pistolhas enough experience to know that hops arenot always the same, say day and night or atdifferent times of the year, or even in the courseof a solar cycle. But that’s all anecdotal infor-mation and what the LP needs is somethingmore quantitative to go on, what factors affectthe numbers that are associated with the iono-sphere.

The place to start, of course, is with thevertical profile of the electron density, so manyelectrons per cubic meter, as a function ofheight above ground. Two such profiles areshown in Figure 4.1, one for daytime conditionsat the maximum of a solar cycle and the otherat the minimum. Of importance to the LittlePistol are the heights of the various regions —from below 100 km to above 300 km for theD-, E-, F1- and F2-regions, respectively — andthe magnitudes of various electron densities,

4: ELEMENTS OF PROPAGATION


and their critical frequencies, say foE, foF1 andfoF2.

In that regard, theoretical considerationsshow that if RF is sent vertically upward to-ward a region with a number density of N elec-trons per cubic meter, the highest frequency fc(in MHz) that is returned from that level as anecho is given by the following equation

fc = (9•E-6)• SQRT(N)With that as a guide, ionospheric sound-

ers probed the electron density overhead bydirecting RF pulses vertically upward and si-multaneously sweeping the sounder frequencyfrom about 0.5-20 MHz. The time of flight ofechoes was displayed on the Y-axis of an oscil-loscope and the instantaneous frequency on theX-axis. When a critical frequency is exceeded,say for the E-region, the RF pulses pass throughit toward the next region, and is shown by asudden increase in the time for an echo to re-turn as the RF pulses go on to higher altitudes.

As the technique was developed, the vari-ous regions of the ionosphere were established,along with representative values of critical fre-quencies and altitudes. For example, the E-re-gion is present during daylight hours and thecritical frequency foFE ranges from 0.5 MHzto 4.5 MHz, depending on local time and lati-tude. And the F1-region is also present duringdaylight and its critical frequency foF1 rangessimilarly from about 0.5 MHz to 6 MHz. TheF2- region is more important for LP’s DXingbut its critical frequencies are quite variable anddo not show any simple relation to the degreeof solar illumination.

To go on, as the sounding technique wasperfected, geographic mapping of the criticalfrequencies was begun and expanded afterWWII to show on a global scale how illumina-tion affects the ionosphere, with the time of dayand the seasons. In addition, vertical sound-ings were carried out during different phasesof solar cycles and that dimension was addedto the database of ionospheric characteristics.

To illustrate those points, consider Figures4.2 and 4.3 which show global maps of the F2-region, with contours of the critical frequencyfoF2 in MHz, at two levels of solar activity. Fig-ure 4.2 shows the foF2 map for 0600 UTC at

Figure 4.1 Sample electron density profiles:(a) daytime and (b) nighttime. From Davies[1990]

from a thousand electrons per cubic centime-ter down in the D-region to a few million elec-trons per cubic centimeter at the F-region peak.

Those profiles apply during daytime con-ditions at mid-latitudes. At night, the F1- andF2-regions coalesce, the E-region density fallsand the D-region disappears. Different profilesare found in the polar and equatorial regionsbut they need not concern the Little Pistol atthis point. And looking at those profiles, itmight be added that LP’s experience is prima-rily with the hops below the F-region peak.That’s because the LP has been prudent, usingoperating frequencies which did not penetratethe F-peak and, as a result, the hop lengths ofthe LP’s RF were determined by the radiationangle of the RF and the height of the highestionospheric region called into play.

But one cannot control all the RF radiatedby antennas so some high-angle rays from LP’santenna might penetrate the F-region peak andventure into the topside of the ionosphere. Thatbrings up the idea of critical frequencies, thehighest frequencies which are returned by theionosphere when probed by RF at vertical in-cidence. That technique, ionospheric sounding,was used in the late 1920s to explore the mainfeatures of the ionosphere, the various regions


Figure 4.2 Global foF2 map for 0600 UTC, March 1976. (SSN=12) From Daviesand Rush [1985]

the spring equinox of 1976 when the sunspotnumber was 12 and Figure 4.3 gives a similarmap for 1979 when the sunspot number was137. For that time of day, the subsolar point ison the geographic equator at 90° East longitudeand the illuminated portion of the earth is be-tween 0 and 180° E while that in darkness isbetween 180° E and 360° E.

It is of importance to note the foF2 maplacks symmetry about the geographic equator

Figure 4.3 Global foF2 map for 0600 UTC, March 1979. (SSN=137) From Daviesand Rush [1985]

even though solar illumination is symmetricalat the equinox. In addition, it is of interest tosee that F2-region critical frequencies do notdrop to vanishingly small values in the darkregions beyond the sunset terminator, lying at180° East longitude. Both the lack of symme-try of the foF2 map across the equator and thepresence of ionization in the dark ionosphereresult from geomagnetic control of the iono-sphere and will be discussed in a later section.


Figure 4.4 Map of foFE, March and June1958. From Davies [1965]

As noted above, the E-region is presentduring daylight hours and examples of the lati-tude and local time variations of its critical fre-quency foFE are shown in Figure 4.4, the up-per portion of the figure for the equinox andthe lower portion for the summer solstice whenthe subsolar point is at 23.5° N latitude. It isseen that foFE goes to small values beyond theterminator which separates regions in sunlightand those in darkness.

Information on the critical frequencies ofthe ionospheric regions is obtained from thestudy of ionograms. Height information, how-ever, turns out to be more complicated as find-ing the true height above ground at which anRF pulse is returned depends on having aknowledge of the electron density profile.However, a virtual height of a region may beobtained by calculating the distance a pulsewould travel at the speed of light in a vacuumin half the time for a pulse to return to groundlevel. These distinctions will become clearerwhen actual ray paths are worked out fromionospheric profiles.

And that brings us to a more detailed rep-resentation of the main structure of the iono-sphere, from the ground to just below the peak

Figure 4.5 Ionospheric structure for daytimeconditions at mid-latitude. From Davies[1990]

of the F2-region, in Figure 4.5. Ionospheric pro-files like that involve a combination of electrondensities derived from critical frequency dataand heights of the peaks and ledges from stud-ies of iono-grams. Of particular importance isthat the data are brought together in a smoothfashion using mathematical models. The lackof discontinuities or sudden changes in elec-tron density makes it possible to follow raypaths through the ionosphere without having

Figure 4.6 Variations of foFE, foF1 and foF2with sunspot number for midday conditionsat middle latitudes. From Davies [1965]


any abnormal kinks or sudden deviations inray directions.

Earlier it was mentioned that ionospherichops are not always the same, changing for anumber of reasons. Now, having seen close upwhat an ionospheric profile looks like, it’s clearthat hops may differ because of changes in criti-cal frequencies and even heights of the regions,say the F-region peak.

Changes with solar activity, in the courseof a solar cycle, are one cause and Figure 4.6shows how foFE, foF1 and foF2 at midday varywith sunspot number. Taking 11 MHz as thevalue for foF2 when the sunspot number is 180

and 6 MHz for foF2 for a sunspot number of20, the expression given earlier relating criticalfrequency and electron density shows that theelectron densities between the extremes differsby a factor of about 3.4.

When we get to tracing ray paths throughthe ionosphere, it will become apparent howhop lengths differ for those two extremes. Butthat is in the abstract and the Little Pistol needsto know more practical things, say what is thesunspot number and just where that value fitsin the solar cycle in progress. So at this pointwe need to look into the means that LP canemploy to obtain solar data.


Here in the USA the principal source ofsolar data, as it relates to HF propagation, isthrough NOAA in Boulder, CO. It comes inthree different forms: (1) WWV broadcasts at18 minutes after each hour, (2) mailed weeklyas “The Preliminary Report and Forecast ofSolar Geophysical Data” or (3) by telephone atany time, just by calling the NOAA/SESC PBBSat (303) 497-5000 and downloading the SolarReport file. But before we look at those sources,let’s look at the information itself.

The solar data that will interest the LittlePistol in the first instance is that which is givendaily through one of those means. Confiningourselves to the electromagnetic spectrum,NOAA provides daily information on the so-lar radio noise flux at 10.7 cm, the count of sun-spots visible on the solar disk facing the earthand the average background X-ray flux in the1-8 Angstrom range, as recorded by the GOES-7 satellite at synchronous altitude.

The 10.7-cm solar noise flux is broadcasthourly by WWV on 2.5, 5, 10, 15 and 20 MHzand is given as a number which expresses in-tensity in solar radio flux units (1 s.f.u. = 1E-22 Watts per square meter per Hz bandwidth).Typical values of the 10.7-cm flux are anywherefrom 65- 350, depending on the level of solaractivity.

At this point a word of warning is calledfor. Solar photons with a wavelength of 10.7cm have an energy of only 1.1E-5 eV, fallingbelow the energy required to ionize any atmo-spheric atom or molecule by a factor of onemillion (!). Thus, 10.7-cm radiation contributesnothing to ionizing the upper atmosphere.Knowledge of its current value comes from thefact the atmosphere is transparent to 10.7-cmradiation. With the radiation passing freelythrough the atmospheric window, any changesin its value in s.f.u. shows the growth or decayof active regions on the sun or their passageacross the solar disk facing the earth, no mat-ter what the terrestrial weather conditions.

5: SOLAR DATA

That word of warning was given, as un-informed operators often think inappropriatelyabout increases in the daily value of the 10.7-cm solar flux, taking them to represent signifi-cant changes in the solar ionizing flux and thusimprovements in propagation conditions. Thatis just not the case as the relationship betweensolar EUV radiation and the nonionizing 10.7-cm flux is a statistical one, and a very loose con-nection at that.

In addition to the 10.7-cm noise flux, thecount of sunspots, seen as dark regions on thesolar disk, also serve as an indicator of solaractivity. In that regard, the occurrence of a num-ber of terrestrial phenomena have been foundto be correlated with sunspot numbers: mag-netic storms, visual aurora, and long-distanceradio communication, to mention those of di-rect interest to the Little Pistol. All of those arewhat might be termed “gross phenomena” andwhen more detailed observations are made,other correlations are found. We’ll get to themin due time. Since both the 10.7-cm solar noiseflux and sunspot number provide measures ofsolar activity, they may be plotted as daily val-ues, giving a graphical representation ofchanges in solar activity and a means of antici-pating future activity. By way of illustration,Figure 5.1 shows daily values of the 10.7-cmsolar flux and the sunspot count during a pe-riod of recurrent activity in mid-1990. Exami-nation of that figure shows the sunspot num-ber has more irregular variability than does the10.7-cm solar flux.

Another way of representing long-termchanges in solar activity is to plot monthly av-erages of the two quantities. The monthly av-eraging process serves to smooth the data, atleast compared to a plot of daily values; evensmoother representation involves a 13-monthaveraging process. By way of illustration, the13-month smoothed average value of the sun-spot number (140) for July 1990 is obtained byadding the monthly average sunspot numbers


Figure 5.1 Variations of the 10.7-cm solar flux and daily sunspot count in1990, showing 27-day recurrences. Data from NOAA/SESC.

Figure 5.2 A comparison of monthlyaverages and smoothed sunspot numbersstarting in 1900. From NOAA/SESC Report25 October 1994.

from February 1990 to December 1990 and one-half the average sunspot numbers for January1990 and January 1991. That sum is for the 11months centered on July 1990 and one-half ofthe first and last months of a 13-month period.Finally, the smoothed sunspot number for July1990 is obtained by dividing the full sum by12, for the effective number of months used inthe smoothing process.

The smoothing obtained by using the 13-month running averages is seen in Figure 5.2,comparing the smoothed sunspot numbers(SSN) with monthly averages for the periodstarting in 1900. Needless to say, that type ofsmoothing is not limited to just sunspot data,also being used now in connection with solarflux values. However, in contrast to sunspotdata which goes back to the middle of the 18thCentury, 10.7-cm radio observations are lim-ited to the period following WWII.

In that connection, the Little Pistol knowsabout solar cycles, the sunspot number chang-ing in a cyclical manner with about 11 yearsperiodicity. Back in February 1990, the weeklyBoulder Report summarized the features ofSolar Cycles 1-22, dating from March 1755 tothe present. Of the 21 cycles already completed,

the average value of the smoothed sunspotnumber was 111.7, the average length of thecycles was 11.0 years (132.3 months), the aver-age time to the maximum was 4.29 years (51.5months) and from the maximum to the end ofthe cycles 6.73 years (80.8 months), respectively.


At the moment, we’re nearing the end ofCycle 22 which started in September 1986 andpeaked in July 1989 with a smoothed sunspotnumber of 158.5. Every month or so the weeklyBoulder Report updates the current values ofthe smoothed data. A summary for Cycle 22and the features of recent solar cycles is givenin Figure 5.3, taken from the Boulder Report of11 October 1994.

In addition to data from the current cycle,that figure shows the time histories of thesmoothed sunspot numbers (top) in severalpast cycles and smoothed 10.7-cm solar radioflux (bottom) for the few cycles since WWII.Beyond data from the past cycles, those figuresalso show predictions for the coming solarcycle.

The last type of solar data that’s availablefrom NOAA is the background X-ray flux in

Figure 5.3 A comparison of smoothed sunspot numbers and 10.7-cm solarflux for Cycle 22 with previous cycles.

the 1-8 Angstrom band. It’s given on a dailybasis in the solar data file on the NOAA PBBSand expressed on a logarithmic scale in Wattsper square meter, a number being used for themultiplier and a letter (A, B, C, M and X) forthe power of 10, starting at 1E-8 for the letterA. Daily values for the background X-ray fluxare also included in the Solar Data Summaryin the weekly Boulder Report, along with dailyvalues for the 10.7-cm flux, sunspot number,and sunspot area. Also included in the weeklyBoulder Report is a figure which shows five-minute averages of the X-ray flux, the examplein Figure 5.4 showing how the X-ray flux var-ies on that time scale and as well as an X-rayburst.

X-ray data, whether from background ora flare outburst, gives a better indication of thegeophysical importance of solar conditions


Figure 5.4 Five-minute averages of the X-ray flux from the GOES satellite.Units are Watts per square meter. Data from NOAA/SESC Report 11 October1994.

than observations taken in the visible range ofthe solar spectrum. And it gives a measure ofionizing radiation coming from the solar disk.Unfortunately, the X-ray data in the BoulderReport is not accompanied by measures of so-lar radiation in the EUV range. However, theenergy of X-ray photons in the 1-8 Angstromband exceeds the energy needed to ionize at-mospheric constituents by about a factor of 100instead of falling short by a factor of 1,000,000as is the case in the 10.7-cm range.

As a result, one can take the X-ray flux asa better measure of the rate of ionization in theionosphere from solar activity. And it is of in-terest to compare the range of the X-ray flux ina solar cycle with the ranges for the other indi-cators of solar activity, the 10.7-cm flux andsmoothed sunspot numbers. In that regard, the1-8 Angstrom X-ray flux varied by about a fac-tor of 100 in going from solar minimum to themaximum of Cycle 22, the background flux ris-ing from the A category to the C category. Overthat span of time, the smoothed 10.7- cm fluxrose by a factor of three, from about 70 to about200 s.f.u., and sunspot numbers rose by a fac-tor of 15, going from about 10 to 150.

The variations of the X-ray flux during theperiod used in Figure 5.1 are now shown inFigure 5.5. As expected, they go along with thechanges in 10.7-cm flux and sunspot count al-though they show a larger dynamic range.

Figure 5.5 Variations in the 1-8 AngstromX-ray flux from the GOES satellite. Units arein Watts per square meter. Data from NOAA/SESC Reports.

All three variables characterize solar ac-tivity on the side of the sun facing the earth.However, it should be noted that sunspots areat the solar surface (photosphere) while X-raysand the 10.7-cm photons are created in the so-lar atmosphere. Indeed, X-ray bursts can origi-nate, even be detected, from high above activeregions which are behind the solar limb andthe same is true of radio emissions associatedwith bursts of solar activity.

The solar spectrum includes not only the


portions mentioned in the reports from NOAAbut others as well. Unfortunately, data on theEUV range is not available from NOAA so so-lar indicators given above are used in taking ameasure of solar/terrestrial conditions. Butgiven the chemical composition at ionosphericheights, it should be remembered the D-region

is produced by what is termed “hard” X-rayswith wavelengths less than 10 Angstroms, theE-region by “soft” X-rays with wavelengths inthe range between 10 and 100 Angstroms whilethe F1- and F2-regions can be ionized by ra-diation in the EUV range, with wavelengthsfrom 100 to about 1000 Angstroms.


In turning from data for the 10.7-cm vis-ible or X-ray parts of the radiation coming fromthe sun to data for magnetic conditions hereon earth, the shift in the discussion of HFpropagation is from slowly varying changes ofthe ionosphere, over days, weeks or months,to more rapid ones which can disrupt HFpropagation and result from major changes insolar activity. Of course, that’s not to say therecan’t be any rapid changes in solar radiationwhich result in disturbances of the ionospherefor an hour or so, say from outbursts of solarX-rays. Dellinger discovered events of that na-ture in 1937, shortwave fade-outs (SWF) or sud-den ionospheric disturbances (SID) associatedwith flare outbursts on the sun.

But that was back in what might betermed the “photochemical era” when both thesteady and disturbed states of the ionospherewere thought to result largely from “action ata distance,” solar photons going across the vastempty space between the sun and earth untilencountering the earth’s atmosphere. With thecoming of the Space Age, spacecraft observa-tions showed that empty space or void actu-ally contains a weak and disordered magneticfield from the sun, the interplanetary field, aswell as streams of solar particles, electrons andprotons, spewed out by the sun. This is calledthe “solar wind” — sometimes in a fairly steadyflow, other times in gusts and even as shockwaves.

Ionospheric disturbances associated withgeomagnetic activity result from changes,namely increases in speed, in the solar wind(or solar plasma) which propagate through theinterplanetary medium until encountering theouter reaches of the geomagnetic field. The factthat geomagnetic variations induced by thatsort of encounter will disturb or distort theionosphere results from ionospheric electronsbeing under geomagnetic control. That waspointed out earlier from the fact that foF2 mapsfor equinoctal periods are not symmetrical

about the geographic equator. As a result, geo-magnetic disturbances from such interactionsaffect the spatial distribution of ionizationthrough the magnetic control.

While that may have come as a surpriseto the Little Pistol, it merely reflects the fact thationospheric electrons, on being released byphotoionization, are not free to follow ballistictrajectories. Instead, they move in their localenvironment and experience a magnetic forcewhich influences their motions. As for details,the magnetic force depends directly on theelectron’s velocity and the geomagnetic fieldstrength it finds itself in.

If the electron’s velocity is perpendicularto the geomagnetic field, the magnitude of theforce is given by the product of the electroncharge (e), its speed (v) and the magnetic fieldstrength (B), when those quantities are ex-pressed in M.K.S. units. The direction of themagnetic force is perpendicular to the plane ofthe velocity and field vectors. The electron goesin a circular path around the field line with anangular velocity equal to the product of theelectron’s charge (e) and the field strength (B),and then divided by the electron’s mass (m),e•B/m.

The magnitude of that angular velocity isabout 8.8•E+6 radians per second for a typicalvalue of the geomagnetic field strength. Whenchanged to the number of rotations or cyclesof motion per second around the field line, it’scalled the electron gyrofrequency and is about1.4 MHz. That number should be kept in mindfor later reference as the effect of the geomag-netic field on propagation becomes importantwhen the radio frequency (f) of a signal ap-proaches or is comparable to an ionosphericelectron’s gyrofrequency.

Of course, the velocity of an ionosphericelectron at its time of release by a photoioniza-tion process need not be perpendicular to thelocal geomagnetic field. In that case, the elec-tron spirals around the field line at a rate which

6: GEOMAGNETIC DATA


depends on the component of its velocity per-pendicular to the field. In addition, the elec-tron advances along the field line, spiralling upor down it, with a speed equal to the compo-nent of the electron’s velocity along the fielddirection. Indeed, if its velocity is parallel tothe field direction, it will not spiral nor experi-ence any magnetic force, simply sliding alongthe field line.

To go any further in this discussion, moreinformation is needed about the geomagneticfield, its shape and how it can be disturbed. Asfor shape, the Little Pistol surely remembershigh school physics and the fact that the earth’sfield resembles that of a bar magnet. That sortof description is termed a “dipole field” andhas magnetic field lines as shown in Figure 6.1.

All that is well and good but in reality, itonly gives a rough approximation to the actualfield of the earth. So, for the Little Pistol’s sake,let me say what’s wrong with the dipole modelthat’s been around so long. First, it was basedon an analysis of surface data, largely frompopulated areas, and had gaps over the vastoceans and the polar regions. And the old datawas collected slowly, taking years to assemblebefore performing the mathematical analysis.As a result, it was incomplete and missed thevariability that is so characteristic of the polarregions.

Having said the dipole model is only anapproximation, the next question is where ithas validity and to what degree. There, thegood news is that it’s a fairly good approxima-tion at low and middle latitudes. For the mo-ment, let’s put off the rest of the discussion butkeep in mind that there’s much more to thestory than such a simple model, especially athigh latitudes and the effects of solar plasmastreaming by the earth.

With that discussion put on hold, let’s getinto the geomagnetic data, principally withregard to disturbances of solar origin, that areavailable here in the USA. As with the solarradiation data, the principal source is NOAAin Boulder, CO. And again, the informationcomes in the same ways: the WWV radiobroadcasts 18 minutes after each hour, by mailin the weekly Boulder Report and any day bytelephone from the NOAA PBBS.

All three of those sources of data give thesame sort of information, values of the daily24-hour A-index and three-hour K-indices forgeomagnetic disturbances. Those indices arebased on data from magnetic sensors, termedmagnetometers, which continuously follow thestrength and direction of the geomagnetic fieldat a number of locations on the earth. Thus,they develop information on the quiet level forthe geomagnetic field, say in its componentsto the north, toward the east and vertical di-rection, as well as observe departures fromthose levels due to disturbances.

Each observatory develops its own scalefor variations above or below the quiet-day lev-els but they’ve been normalized so that all sta-

Figure 6.1 A sketch of dipole magnetic fieldlines.

In the earth’s case, the magnetic axis ofsymmetry of the dipole field and that of theearth’s rotation do not coincide. As a matter offact, both the dipole and rotation axes areknown to wander over the earth’s surface. Atthe moment, the dipole axis is tilted about 11.5degrees with respect to the axis of the earth,passing through the earth’s surface near 78.5°N, 69.0° W and 78.5° S, 110.0° E. That meansthe geomagnetic equator is tilted 11.5 degreeswith respect to the geographic equator andcrosses the latter at 21° E and 159° W longi-tude.


tions report on the same basis. So, in terms ofaverages, the three-hour K-indices range gofrom 0 for very quiet to 9 for extremely dis-turbed and the 24-hour A-indices range from 0(very quiet) to 400 (extremely disturbed).

On the WWV broadcasts at 18 minutesafter each hour, the most recent value of thethree-hour K-index is reported from the mag-netometer at Boulder, CO. In addition, either a21-hour estimate of the Boulder A-index isgiven from 2100 UTC until 2400 UTC for thatday or the past 24-hour A-index is given from0000 UTC until 2100 UTC when the next 21-hour estimate is reported. In addition, a verbaldescription is given for the level of current geo-magnetic activity, using terms like quiet, un-settled, active, minor storm, major storm andsevere storm, as well as a forecast in those sameterms.

Turning to the weekly Boulder Report, itgives a summary of the Daily GeomagneticData from two observatories, Fredericksburg,VA and College, AK, as well as estimates forthe planetary 24-hour A-indices and eight 3-hourly K-indices for days in the past week. Itshould be noted that Fredericksburg data isfrom a mid-latitude station while that fromCollege is from a high-latitude site. A close lookat both sets of data will show that the K-indi-ces from the high-latitude site tend to be greaterthan those from mid-latitudes, an importantfact which will be discussed later.

The estimated planetary A- and K-indi-ces are obtained from a group of five magneticobservatories in the Western Hemisphere, ex-tending from Alaska to England. Those valuesare used as an approximation to the planetaryindices, Ap and Kp, which are derived eachmonth from an analysis of data from 13 obser-

vatories, 11 in the Northern Hemisphere andtwo in the Southern Hemisphere. The esti-mated planetary A- and K-indices are derivedin real time whereas the actual planetary indi-ces, Ap and Kp, are derived at the end of amonth, when the data from the widespreadnetwork of observatories has been collected.

Geomagnetic data on the NOAA PBBS isfound in two files, the Solar Report and thePropagation Report. The Solar Report gives theA-indices for the previous day, AFR fromFredericksburg and AP, as obtained from ob-servations at the five observatories mentionedearlier. In addition, the Solar Report gives esti-mates of AFR and AP for the day in questionas well as predicted values for the next threedays. The Propagation Report, on the otherhand, only gives the values of the KP- and AP-indices for the day in question as well as theforecast of their values for the next three days.

The Solar Report and Propagation Reportfrom the PBBS differ from the weekly BoulderReport in that they offer predictions for the10.7-cm solar flux as well as probabilities ofgeomagnetic activity or storminess at bothmiddle and high latitudes for the next threedays. Those two reports do include discussionsof solar and geophysical activity as well as fore-casts of activity.

Further explanation of the details in thosereports will have to wait until propagationhas been discussed more fully. But in theshort run, following solar and geomagneticdata would be helpful to the Little Pistol inunderstanding solar/terrestrial conditionsand in anticipating changes which mighthave an adverse effect of propagation, par-ticularly increases in magnetic activity tominor and major storm levels.


Now that the origin of the ionosphere andthe possibility of its disturbance by solar activ-ity have been discussed, the ground has beenlaid for a fuller treatment of HF propagation.The discussion will be broken down into threeparts: (1) critical frequencies, (2) signal strengthand (3) noise. Such an elaboration is necessaryas any communication on the HF bands de-pends on a path being open as well as adequatesignal strength and low noise at the receiver.

The role of critical frequencies will be dealtwith first in a plane or flat-earth geometry andthen extended to include earth’s curvature.That discussion will begin with idea of thehighest possible frequency on a path in asimple, plane ionosphere, then make use of ray-tracing to show results for a curved ionosphereand finally determine the geometry for morerealistic propagation paths.

So let’s begin with a one-layer, plane iono-sphere with a maximum electron density Nmaxat its peak, a height Hmax above the earth’ssurface, as shown in Figure 7.1. That sort ofmodel would serve to illustrate conditions atnight when the E-region ionization has fallento low levels and only the F2-region remains.From the discussion in the first section onpropagation, the highest frequency that can besent vertically upward from a transmitter andstill returned to the ground as an echo is termed

the critical frequency (fc) of the layer and givenby

fc = (9•E-6)•SQRT(Nmax)where Nmax is the electron density at heightHmax.

Now consider a transmitter sending RFupward at an angle above the horizon or, moreappropriately, an angle Z with the vertical orzenith direction. The RF enters the bottom ofthe ionospheric layer at the angle Z but followsa curved path because of refraction. In that case,theory shows for oblique incidence the high-est frequency fmax returned to the ground fromthe peak of the layer is greater than fc by a fac-tor 1/cosZ. As an example, take fc as 4 MHzand angle Z as 60 degrees. In that case, the elec-tron density Nmax is 2E+11 per cubic meterand the frequency fmax returned at that angleof incidence is 8.0 MHz, twice the critical fre-quency for vertical incidence.

That simple example shows how HFpropagation works, the return to ground levelof an 8.0-MHz signal at oblique incidence andinvolving a smaller deviation of its path, 60degrees instead of 180 degrees for the 4-MHzsignal at vertical incidence. And that resultsfrom the higher frequency signal encounteringa greater number of electrons along the obliquepath than for RF at vertical incidence at the criti-cal frequency (fc).

7: MORE ON PROPAGATION

Figure 7.1 Ray refraction in a single-layer plane ionosphere.


Thus, HF communication is possible atfrequencies well beyond the critical frequen-cies found by sounding the ionosphere and thelimits are set by the electron density overheadand the radiation patterns of antennas. Ofcourse, lower frequencies (f) will also propa-gate at oblique incidence, the main differencebeing that they won’t penetrate as far into theionospheric layer. Since radiation at a fre-quency fmax given by

fmax= fc/cosZpropagates at a zenith angle Z and reaches thetop of the layer at height Hmax where the elec-tron density is Nmax, it’s clear that its effec-tive vertical frequency is fmax•cosZ. By thattoken, radiation of a lower frequency (f) willhave a lower effective vertical frequency,f•cosZ, and be returned from a lower altitude,where the electron density N may be foundfrom the fact that

f•cosZ = (9•E-6)•SQRT(N)and the physical height found from the elec-tron density profile.

Beyond that important result, Figure 7.1shows the ground range or hop distance forthat angle of incidence on the ionosphere. Anda diagram like that points out the differencebetween the actual height from which RF isrefracted downward and the virtual height forreturn of RF by reflection from an ionospheric“mirror.”

While the idea of mirror reflection findsfurther application in the discussion of propa-gation, it should be remembered that it is more

symbolic than physical. The reason, of course,is that a mirror has no refractive properties andthus fails to represent the principal feature ofthe ionosphere, that propagation paths vary ac-cording to frequency. A better way to deal withionospheric problems, whether by refraction orreflection, is to include the earth’s curvatureand calculate the details of paths in sphericalgeometry, using 6,378 km for the earth’s radius.

Under those circumstances, the one-layerionosphere problem works out much like theplane case done before except that the angle Zfor incidence at the height Ho, where thecurved ionospheric layer begins, has to takeinto account the spherical geometry of theproblem. From the geometry shown in Figure7.2, the zenith angle Z at height Ho will be dif-ferent (smaller) than the zenith angle for theRF at launch from ground level.

In addition to that adjustment, there is acorrection (k) that has to be applied to the fac-tor 1/cosZ because of curvature effects in theionosphere itself. The correction factor thenbecomes k/cosZ, where k varies linearly withdistance, from 1.05 for a 1,000- km path to about1.20 for a 3,000-km path. With those changes,one can make a good estimate of the maximumfrequency fmax returned to ground level by acurved ionosphere with a critical frequency (fc)and electron density profile peaking at heightHmax.

As indicated earlier, frequencies lowerthan fmax will not penetrate as far into theionospheric layer shown in Figure 7.1 nor have

Figure 7.2 Ray refraction in a single-layer, curved ionosphere.


as great hop lengths. The next figure will showthose results but before getting to it, a fewwords of explanation are needed. In particu-lar, ray-path calculations can be carried outeasily using a curved ionosphere but it is notconvenient to display the results in sphericalcoordinates. Instead, a rectangular format isused, the horizontal axis showing the distancealong the curved earth and the vertical axis theheight above ground level. And quite fre-quently the distance scales along the two axeswill differ. Those features should be borne inmind in reviewing the following figures.

Figure 7.3 Ray traces for launch at 5°elevation, 2-20 MHz in 1-MHz steps.

With those remarks, look at Figure 7.3which shows the result when rays launched at5 degrees above the ground horizon go into acurved ionosphere with properties like thatshown in Figure 7.1 (fc=4 MHz, Hmax=300 km,Ho=125 km). The paths were traced out insmall steps by means of a computer programand to obtain that display, the lowest frequencywas taken as 2 MHz and the frequency thenstepped upward 1 MHz at a time to 20 MHz.The ray traces first show the bending or refrac-tion at the bottom of the layer for the lowestfrequency and show deeper penetration of thelayer as the frequency in increased. The 13-MHz ray comes close to the electron densitypeak at 300 km altitude and then rays from 14MHz to 20 MHz pass up through the F-layerpeak and go into the topside ionosphere.

The ray traces in Figure 7.3 also show how

hops at a given launch angle increase in lengthdue to greater penetration with increasing fre-quency. Next look at Figure 7.4 where thelaunch angle has been increased to 15 degreesabove the ground horizon. It is seen that thehop lengths are shortened by almost a factorof 2 at the steeper launch angle and the high-est frequency for which the F-region supportsforward propagation dropped from 13 MHz to10 MHz.

Figure 7.4 Ray traces for launch at 15°elevation, 2-20 MHz in 1-MHz steps.

As indicated above, that simple iono-sphere was used to represent a nighttime F-re-gion. If one turns to the foF2 maps in Figures4.2 and 4.3, it is seen that a 4-MHz value forfoF2 would be more appropriate for nighttimeduring a period of low sunspot number. Onecould use higher values for foF2 to explore thematter but at this point, it’s more appropriateto make the discussion more realistic by includ-ing the E-region and showing the role it playsin daylight.

To that end, turn to Figure 7.5 whichshows ray traces for a mid-latitude ionospherearound daybreak and now includes ionizationin the E-region around 115 km altitude. For thatcircumstance, the profile used for electron den-sity in the E-region rose quickly from 90 km toa plateau around 115 km, with a foFE value of2.5 MHz; from there, the electron density rosesmoothly to the F-layer peak at 285 km, withan foF2 value of 8.0 MHz. In this case, thelaunch angle was set at 5 degrees above the


ground horizon while the lowest frequencywas taken as 3 MHz and the frequency thenstepped upward 3 MHz at a time to 48 MHz.

Inspection of that figure shows low fre-quency rays (3-9 MHz) which are returned bythe E-region, then 12-33 MHz rays which arereturned to earth before the F2-peak is pen-etrated and finally the 36-MHz and higher fre-quency rays go through to the topside of the F-region. A careful examination of that figureshows some deviation of the 12-MHz ray as itgoes through the E-region on ascent and de-scent. That has to do with changes in the rateof refraction or bending on going through aregion where the electron density begins tolevel off, as below the E-region ledge, and thenbegins to increase with height again.

Using the same model ionosphere again,now consider the frequency being held con-stant and the radiation angle above the hori-zon varying. First, take the case of a frequencyof 21 MHz and the radiation angle going from5 degrees to 90 degrees in 5-degree steps. Raytraces for that case are shown in Figure 7.6. Thefirst ray trace at 5 degrees elevation goes outto about 3,000 km distance but the next twofall around 2,250 km distance. Then the ray at20 degrees elevation and all others at higherangles penetrate the F-layer peak and pass intothe topside of the F-region.

Now consider the same ionosphere again

Figure 7.5 Ray traces in a mid-latitudeionosphere, showing effects of the E-region,for 3-48 MHz in 3-MHz steps.

Figure 7.6 Ray traces for 21 MHz. The launchangle starts at 5° and is increased in 5-degree steps to 90°.

Figure 7.7 Ray traces for 14 MHz. The launchangle starts at 5° and is increased in 5-degree steps to 90°.

but ray tracing for 14 MHz instead. The resultsare shown in Figure 7.7 and we should notethe similarities and differences between the twocases. First, the horizontal range at 5 degreeselevation is about the same. Next, it is seen thatrays converge again, now around 1,125-kmrange, before then starting to penetrate the F-layer peak. That convergence of rays is a formof focussing and results in stronger signals. Thedistance between the transmitter and thoseconverging rays is called the “skip distance”and the ray convergence illustrates “skip fo-cussing,” a form of signal gain of ionosphericorigin.


By comparing the skip distances for thosetwo cases, 14- and 21-MHz signals, one seesthat skip distance increases with frequency. Ofcourse, that means a receiver located within theskip region will not hear signals from the trans-mitter. That is not an uncommon situation butas the Little Pistol will testify, it proves quiteannoying when trying to work DX, especiallyon 10 and on 15 Meters, as one often fails tohear nearby stations in a pileup that are in con-tact with the DX station.

Another aspect of propagation that’s evi-dent from those ray traces deals with antennapatterns. In particular, high-angle rays fromantennas may not return to earth, dependingon the frequency and the state of the iono-sphere. For the ionosphere under discussion,the limiting radiation angle for RF penetratingthe ionosphere is around 20 degrees for 21 MHzsignals and 30 degrees for 14 MHz signals. Bythat token, a lower limiting angle would beexpected on 28 MHz and antennas for thatband should be as high as possible to keep theradiation pattern low, thus minimizing powerwasted by high-angle radiation going off to in-finity. But I’m sure the Little Pistol alreadyknew that!

On that same subject, RF going throughthe ionosphere, it should be noted that ray di-rection or signal travel may be reversed andthe results obtained above may be used to con-sider reception of signals coming in from space,from satellites or as cosmic radio noise. In ei-ther case, it is seen that signals coming fromthe topside of the ionosphere pass downthrough what is called an “iris,” after the stopin a camera lens aperture or the part of thehuman eye.

For the cases discussed above, the iris hasa radius of about 750 km for 21-MHz signalsand 500 km for 14-MHz signals. Of course, thesize of the iris will depend on the level of solaractivity and the local critical frequency of theF-layer, being larger at around solar minimumand smaller at solar maximum. In any event,the size of the iris shrinks to zero at the localcritical frequency (fc) of the F-layer and signalsat lower frequencies from outer space are notheard coming through the peak of the F-layer.

The Little Pistol knows about the Russiansatellites, RS-10 and RS-12, which use HF sig-nals and how their use is affected by solar ac-tivity. Thus, RS-12, having a 21-MHz uplinkand 29-MHz downlink, is the one most vulner-able to variations in the size of the iris withchanges in solar activity. While the RS-12downlink on 29 MHz may be heard over con-siderable distance or time by a ground station,the access time to the uplink transponder onRS-12 may be much less because of the smalleriris on 21 MHz as compared to that on 29 MHz.

In that regard, Figure 7.8 shows the resultsof some calculations made for a particular RS-12 pass back in 1992. Since the satellite coversabout 440 km distance per minute in its orbitup at 1,000 km altitude, one cannot use just onecritical frequency to characterize the F-regionbelow it. So instead of thinking of the satelliteas overflying an iris of fixed size or openingcentered over the receiving station, the calcu-

lations for access to the satellite’s uplink aredone in terms of the F-layer penetration fre-quencies along the line of sight to the satellite’sflight path.

While that sounds more complicated, it’sreally nothing but an iris effect and the resultsshow what one would expect from the elemen-tary considerations developed above: the fre-quency for F-region penetration falling below

Figure 7.8 F-region penetration frequencyfor the RS-12 satellite at different levels ofsolar activity.


29-MHz downlink frequency of RS-12’s Robot.That comes from Novices accidentally operat-ing right on the 21-MHz frequency of theRobot’s uplink receiver.

Finally, solar and galactic radio noise sig-nals can reach the earth at the upper end of theHF range and over VHF/UHF frequencies.Those signals are monitored by receivers withbeam antennas pointed upward. For propaga-tion purposes, galactic noise signals around 30MHz may be used to indicate times when theionosphere is disturbed by auroral electron andsolar proton bombardment at high latitudes,as well as the occurrence of geomagneticstorms. In those circumstances, cosmic noisesignals penetrating the F-region are attenuatedbecause of greater absorption in the lower iono-sphere, the D-region. That brings us to the nextsubjects in our discussion, signal strength andabsorption.

the satellite’s uplink frequency for longer pe-riods of time, i.e. greater access to the uplinktransponder, when solar activity is low thanwhen high. If one wants to retain the idea ofan iris over the receiving antenna, that meansthe iris would take on a more complicatedshape because of the N-S orientation and ex-tent of the satellite’s path, not just the simpleconical one that goes with a constant F-layerover the antenna.

Another aspect of the same idea, signalsgoing through the iris, is found when trying tomake contacts via the RS-12 satellite, one hears“strange voices” on SSB calling “CQ Contest”on the downlink. That can happen when 21-MHz signals inadvertently reach the satellitethrough the iris of a transmitting antenna whilenormal 15-Meter operation is in progress. Evenmore strange and curious is hearing Novice sig-nals from the 15-Meter CW band right on the


dB = 10•log(P1/P2)and one obtains –113 dB for the power ratio,meaning that LP’s RF is 113 dB below 1 Wattper square meter at that distance.

As a step in improving the calculation,let’s say now that LP’s isotropic radiator is notin free space but a half-wavelength above aperfectly conducting ground. That means thatall the RF that went down below LP’s originalhorizon is now reflected upward by the groundplane and adds to the RF going out above thehorizon.

But that addition is done using the am-plitudes and phases of the electric field strengthE of the rays, both direct and reflected. In thedirections where the field amplitudes add con-structively, signal amplitude would be doubledor the intensity quadrupled. Thus, LP’s signalswould be 6 dB greater in intensity in those di-rections where the direct and reflected rays arein phase, bringing the LP’s signal strength upto –107 dB Watts per square meter in the direc-tion of strongest signals.

If the Little Pistol’s antenna were a dipole,it would have a gain of 2.1 dB over the isotro-pic radiator and if placed a half-wavelengthabove a perfect ground plane, its strongest sig-nals would be at 30 degrees elevation. But wesaid LP’s antenna is a 3-element Yagi so, in-stead, we’ll take its gain over an isotropic ra-diator to be 6 dB, bringing LP’s signals up to –101 dB Watts per square meter. But since therays going off to St. Louis are somewhat belowthat elevation angle, the vertical radiation pat-tern for the Yagi indicates that signal strengthshould be corrected downward by about 1 dB,bringing it to –102 dB Watts per square meter.

At the outset, it was indicated that ideal-ized methods would be used in embarking ona reality Check, so at this point we have topause and look where approximations weremade, as well as whether they could be im-proved on and then taken into account in ar-riving at Little Pistol’s signal strength at St.

Oddly enough, we begin our reality checkon the ability of the Little Pistols of the Worldto make their DX contacts by using simple, ide-alized methods. But that’s the way one startsto make sense out of the complicated radioworld they operate in. So we begin our task bythinking of one Little Pistol who lives in Boul-der, CO, the mecca for those interested in ra-dio. Then we’ll use one of the hops just dis-cussed earlier, off toward St. Louis, MO arounddaybreak, say the 14-MHz ray trace in Figure7.7 that goes out to 1,300 km distance. On the20-Meter band, that hardly qualifies as DX butwe have to start simply now; we’ll get to realDX in due course.

The problem we have to work out now isthe strength of Little Pistol’s signals at that dis-tance after going through the ionosphere thatwas specified earlier, with an foF2 value of 8.0MHz and foFE of 2.5 MHz. I didn’t mention itat the time but that was for a day in mid-Janu-ary and with a smoothed sunspot number(SSN) of 100.

The first step in the process is to considerthe loss of Little Pistol’s signal strength due tosignal spreading over the distance betweenBoulder and St. Louis. We begin with a grosslyunrealistic representation of the situation bythinking of LP using one of those mythical iso-tropic radiators in free space. So if the 100 Wattsfrom LP’s rig were radiated equally in all di-rections, we can find the RF power per squaremeter at 1,300 km distance by simply dividingthe 100 Watts radiated by the area of a sphereof 1,300 km radius. That result is 4.7E-12 Wattsper square meter, certainly a small number.

In the radio world, both large and smallnumbers are dealt with using logarithms (tobase 10). If we use 1 Watt per square meter as areference value, then the power ratio of LP’sRF power (P1) at 1,300 km distance to the ref-erence power level (P2) is 4.7E-12. That powerratio can be expressed in decibels (dB) by us-ing the definition

8: NOW TO SIGNAL STRENGTH


Louis. The first place to start is with the signalspreading calculation. There the ground rangecovered by LP’s signals was used instead of theactual distance along the ray path in the iono-sphere. By going back to the computer pro-gram, one finds that a more realistic distancewould be closer to 1,400 km, making loss dueto signal spreading about 0.6 dB greater.

That correction is one where a geometri-cal calculation is improved on by consideringthe ionospheric path in more detail. There areother such corrections which can be enumer-ated, in two and three dimensions, but whichare difficult to evaluate in magnitude. For ex-ample, there is the question of where the re-ceiving station is located relative to the skipdistance. Thus, any increase in the height of theF-layer peak could shift the skip distance fur-ther toward 1,300 km ground range and, asnoted earlier, contribute gain because of skipfocusing. If the height decreased, the effectwould be just the opposite.

In three dimensions, there’s the questionof how the ionosphere’s spherical shape or cur-vature affects signal intensity. After all, there’ssome focusing or ray convergence obtainedfrom a concave spherical mirror and one canexpect that the same would be true for the iono-sphere. The trouble is that it’s difficult to cal-culate without resorting to huge efforts withray tracing programs.

That’s in contrast to the case in geometri-cal optics for light. There, the geometry is fairlysimple as all rays travel in directions which areclose to the axis of symmetry of the mirrors andadmit the use of straight lines in ray diagramsdealing with the problems. For the ionosphericcase, however, RF rays are off axis in the ex-treme, and elementary methods are just notavailable. With those remarks, we can suggestthere’s some positive gain from ionosphericfocusing but we’ll have to pass when it comesto estimating it, at least for the Little Pistol’ssituation.

Now it wasn’t said in so many words atthe time but the values of foF2 and foFe usedto characterize the path between Boulder, COand St. Louis, MO were really for the midpointof the path, close to Belleville, KS. Thus, in re-

ality, around daybreak the sun is higher in thesky over St. Louis than over Boulder. That sug-gests that there is more in the way of ioniza-tion, and ionospheric refraction, on the east-ern half of the path than the western half. Thatpoint needs no testing but the effect could beexamined by making new ray-tracing calcula-tions which take into account the variation ofthe solar elevation angle for an hour differencein local time, from 90° E to 105° E longitude.But it seems fair to say that when it comes tothe dB of signal loss from path length or spread-ing, it would be a very small effect. So we’lljust leave it there, being aware of the fact theray path is an approximation, for those reasons.

But those very reasons do affect the LittlePistol’s signal strength at the St. Louis end ofthe path. Here, the matter is that of signals’absorption in the D-region of the ionosphere.That region was mentioned early in the discus-sion and the fact is its electron density was thelowest of the ionospheric regions. In addition,it was pointed out the D-region is present onlywhen the atmosphere is illuminated around the60-90-km level and then with at most 1,000 elec-trons per cubic centimeter.

That electron density is too low to give riseto any refraction of HF signals passing throughthe D-region. However, in the daytime, it canweaken signals, low frequencies more thanhigher ones. Essentially, the D-region electronsare excited into oscillatory motions by passingRF but at those depths, the D-region electronscollide with atmospheric constituents andtransfer energy to them. That means the atmo-sphere is heated by a beam of RF passingthrough and, given the conservation energy,that results in the loss of signal strength.

The transfer of energy from the wave tothe atmosphere takes place in two steps, thefirst part when the D-region electron gains en-ergy from the passing wave and the second partwhen it undergoes a collision and transfersenergy to a molecule. Deep in the atmosphere,where ionospheric electrons have a high colli-sion frequency (fcoll) with atoms and mol-ecules (about two billion or 2E+9 collisions persecond), RF waves have little chance to impartany energy to the electrons before a collision


occurs. So the absorption of RF energy is verylow for all HF frequencies down around 30 km.

But higher in the D-region, say around 90km, the collision rate is much lower (about300,000 or 3E+5 collisions per second), evenbelow typical amateur operating frequencies.As a result, an ionospheric electron makesmany oscillations, reradiating RF energy fromthe incident wave, before it finally makes a col-lision and transfers some energy to the atmo-sphere.

In between those two extremes, the ab-sorption of energy from the passing waves goesthrough a maximum, essentially when the col-lision and operating frequencies are about thesame. For that circumstance, both the first andsecond steps of the absorption process areworking at comparable rates. The numericaldetails of D-region absorption can be workedwith the aid of laboratory experiments whichsimulate the electron collisions and a model forthe atmospheric region. From that, one can findthe absolute efficiency for energy absorption perelectron and how it varies with height in theD-region and wave frequency.

In that regard, the relative absorption effi-ciency per electron at D-region heights and forthe five harmonically-related amateur bands inthe HF spectrum is shown in Figure 8.1. In a

qualitative sense, the Little Pistol is aware ofthose results, knowing it’s tough to work realDX on the 7-MHz band in broad daylight butmuch easier at night. Under nighttime condi-tions, the D-region electron population falls tolow levels as the main source of ionization, theflux of “hard” solar X-rays, is no longer inci-dent and the only sources of ionization are scat-tered sunlight and the low flux of galactic cos-mic rays passing through the region.

As for working out the absorption loss forthe LP’s path in Figure 7.7, it’s just a matter offollowing the ray path through the D-region,on ascent from Boulder and descent to St.Louis, and then weighting the number of elec-trons encountered in a square-meter columnalong the way according to their absolute ab-sorption efficiency at each altitude and addingup the losses. When that’s done, the result isanother 1.7 dB loss, bringing LP’s signalstrength to –103.7 dB Watts per square meter.If we add in the 0.5 dB correction for signalspreading, LP’s signal strength would be –104.3dB Watts per square meter.

It should be noted, however, that the ab-sorption result noted above was not obtainedwith the ray tracing program but another, sepa-rate calculation which took into account the factthat the actual D-region illumination differedin the vicinity of those two sites separated by15 degrees of longitude or one hour of localtime. Thus, no further adjustment is needed,say along the lines suggested earlier for the F-region, and we can go on to the next item re-lated to the Little Pistol’s signals, the antennaat the other end of the path which determinesthe signal strength reaching the input of thedistant receiver.

Let’s assume the receiving antenna is ahalf-wave dipole at a half-wavelength over aperfect ground and oriented so that its axis (orwire direction) is perpendicular to the direc-tion of LP’s incoming signals. There are hardways and easy ways to find how much of LP’sRF power arrives at the input of the receiver.The hard way is to find the electric fieldstrength of LP’s signals at the dipole and thenfind the voltage induced. The easy way, andthe one we’ll use, is to think of the receiving

Figure 8.1 The relative absorption efficiencyper electron as a function of height in theD-region and at different frequencies.


antenna as a collector of RF that’s passing by;in that approach, one merely has to know thecollecting area or aperture for the dipole.

Again, we start simply and ideally andwork our way up toward reality. So if the re-ceiving antenna were an isotropic radiator, elec-tromagnetic theory shows its collecting areawould be 0.079 times the square of the wave-length. That idea is not new, dating back towork by Friis in 1946 and you’ll even find it inthe ARRL Antenna Book. So for the LP’s 14-MHz signals, the effective collecting areawould be 36.3 square meters if the isotropicreceiving antenna were in free space. If we ex-press the collecting area in logarithms, then theRF power collected by that isotropic receivingantenna would be the sum of –104.3 dB for theradiation flux (Watts per square meter) and theantenna area (15.6 dB square meters) or an RFpower of –88.7 dB below 1 Watt at the antennafeedpoint.

Now we make some additional correc-tions, +6 dB for the receiving antenna beingover a ground plane and +2.1 dB for it being adipole instead of an isotropic radiator. Thatbrings the signal power at the feedpoint up to–80.6 dB Watt or –80.6 dBW. Finally, the lastcorrection, for the fact that the radiation is com-ing in below the 30-degree angle, the most fa-vorable direction of a dipole; for the ray fromBoulder to St. Louis, that amounts to a correc-tion of –0.5 dB, bringing the RF power to –81.1dBW.

All of the discussion above has been on alofty plane, using the results for idealized situ-ations and assuming the ionosphere can beconsidered as a stable, smooth distribution ofionization. But, as the saying goes, we have to“get down to earth” in the discussion, puttingin a correction for the fact that both the trans-

mitting and receiving antennas are not overperfectly conducting ground planes. In orderto do that, we have to resort to the effects ofreal ground on vertical antenna patterns, boththe LP’s 3-element Yagi and the receiving an-tenna.

Each type of real ground has its own elec-trical conductivity and a dielectric constant andin terms of quality, they range from excellent(sea water) to poor (sand or ice) when it comesto simulating the ideal ground that was usedin the discussion above. There’ll be more onthe subject of ground when the discussion turnsto the surface reflections that contribute tomulti-hop paths. At this juncture, the impor-tant point is that the finite conductivity of aground surface leads to losses on wave reflec-tion in the vicinity of an antenna, whether usedfor transmitting or receiving purposes.

So, for the moment, a correction of–2 dB, taken from the ARRL Antenna Book, willbe used for the effect of an average ground onradiation coming from a dipole and also on theLP’s 3-element Yagi. That brings the RF poweravailable at the feed point of the receiving an-tenna to -85 dBW, on rounding to the nearestinteger value. Assuming no other losses, thatpower at the receiver must be compared withnoise power from atmospheric and man-madesources to see if LP’s signals are readable.

So now we should turn to a discussion ofnoise. Before doing that, however, it should benoted that the ideas expressed above are theoutline of how longer, more complicated pathswould be analyzed for signal strength. It’s justa matter of taking the hops one at a time andadding up the results. Of course, the degree ofsolar illumination will vary along a path andthat will have to be taken into account.


We’ve already considered two of the as-pects needed for successful communication —signals getting through the ionosphere andtheir strength. Now the question is whether thesignal strength is great enough to compete withthe noise at the receiving end. That is an im-portant question but before turning to it, oneshould bear in mind that the analysis to thispoint has been for an ionosphere whose criti-cal frequencies, foFE and foF2, were specified,not derived or predicted quantities.

In reality, those characteristics of the iono-sphere will change in the course of a day as thesun goes across the sky. That being the case,one would think that the specified quantitieswould be of value for some limited period oftime or, if used for an hour, be considered tohave some variability. But if the critical frequen-cies used in calculations were obtained fromsome sort of prediction scheme, it should beunderstood that the values have some measureof uncertainty which should be considered ininterpreting the results. Put another way, it isthe height of innocence to think that any mea-sured physical quantity, from the ionosphereor whatever, is without uncertainty and to treatit as though it’s without any probable error.

In the present discussion that idea, of un-certainty, really comes to the fore for the firsttime with noise, particularly that of man-madeorigin. Since noise power is a quantity thatwe’d have difficulty measuring ourselves, weuse the results from various scientific studies.And in doing so, the first question deals withthe noise power in a receiver, its average value,and then its variability.

To begin the discussion, we note noise isalways present in the output of a receiver, thenoise reaching an antenna being amplifiedalong with the incoming signal. Of the threetypes of radio noise — galactic, atmosphericand man-made — consideration will be givenonly to the last two as galactic noise is negli-gible in all but the most remote sites. That re-

sults in the HF range as the ionospheric iris overa receiving antenna closes as the frequencydrops below 30 MHz. But that is not to say thatsolar noise outbursts cannot be heard near thetop of the HF spectrum during periods of so-lar activity but being sporadic, they don’t fit inthe discussion of HF prediction methods.

Of the other two types, atmospheric andman-made noise, the second one is probablymore familiar, as most DX operators have beentroubled at one time or another by impulsivebursts of noise from power tools, spiky igni-tion noise or the steady buzz of neon signswreaking havoc on the bands. And noise, be-ing broadband by nature, varies at one’s re-ceiver output according to the receiver band-width in use, say 500 Hz for CW to 2,400 Hzfor SSB in amateur situations and 6,000 Hz forAM.

While man-made noise may be quite site-or time-specific, studies show that averagenoise powers may be categorized by popula-tion levels or locations — industrial, residen-tial, rural and remote unpopulous. In that re-gard, the IONCAP program uses the follow-ing for the noise power relative to one Watt at3 MHz and with 1 Hz bandwidth: –140 dBWfor industrial, –145 dBW for residential, –150dBW for rural and finally –164 dBW for ruralunpopulous sites. Those values are from a 1974report by the U.S. Department of Commerce.

In order to find the noise power to com-pare with the signal strength calculated earlier,the type of site must be identified and the band-width (b) specified so it may be used to multi-ply by (or added logarithmically to) the noisepower at 1 Hz bandwidth. In addition, a cor-rection must be applied to find the noise powerfor the 14-MHz operating frequency since thestudies showed noise power falls with increas-ing frequency, as shown in Figure 9.1.

If we assume the receiving site in St. Louisis in a residential setting and SSB signals arebeing received, the –145 dBW value at 3 MHz

9: NOISE


Figure 9.1 Average man-made noise power as a function of frequency fordifferent environments.

must be reduced by about 18 dBW because ofthe lower noise power at 14 MHz. However, abandwidth of 2,400 Hz means another adjust-ment upward by 34 dB to cover the greateramount of noise received, giving a final valuefor the average noise power of –129 dB above1 Watt or –129 dBW. When compared to the –85 dBW signal from the LP’s antenna, it meansthe LP’s signal-to-noise (S/N) ratio could beas high as 44 dB in a residential area of St. Louis.And using the above method, one can see thatthe S/N ratio would be lower by 5 dB in anindustrial setting and higher by 5 dB in a ruralarea.

While surveys provide data on the aver-age level of noise power across the spectrumin the various settings, they also give informa-tion on noise variability. As one might expect,the man-made noise level in populated regionsalso shows daily variation, being greatest dur-ing waking hours of local time. That variationruns around +/–10 dBW when averaged overpopulated regions and the hours in the day.

Since all the noise calculations so far haveused quantities which came from given or av-erage values, results should be made more re-alistic by including their statistical fluctuations.In that regard, the +/–10 dBW variation in av-

erage noise power would cause S/N ratios tovary or fluctuate, even fall 10 dB below thevalues given, making reception difficult since6 dB of signal or noise amounts to 1 S-unit.

Beyond variations in noise power, thereare other statistical aspects of the S/N ratio tobe considered, the various forms of fading af-fecting signal strength; every DXer knowsthose problems well. How does that happenand how much can it affect the S/N ratio dur-ing a contact? Those are good questions andwill be treated later as they’d take the discus-sion too far afield at this point. Right now, wehave to conclude the discussion of noise byturning to atmospheric noise.

As the term implies, atmospheric noisehas its origin in the meteorological processes,from lightning strokes near and afar. Those dis-charges give rise to a broad emission of radionoise which is propagated just like any othersignal, the higher frequencies largely confinedbelow the F-region peak and the lower frequen-cies absorbed by the D-region during the day.

The geographic distribution of lightningstrikes varies with the season, particularlystrong sources being located in South America,South Africa and Indonesia during December,January and February and more “equator-


ward” during the months of June, July andAugust. In the USA, thunderstorm activity isconcentrated in the southeastern and mid-western states during the summer months.Those features are reflected in the distributionof noise with seasons, atmospheric noise be-ing most intense in those regions during themonths indicated.

As for magnitudes, global noise mapshave been produced by the CCIR for eachthree-month period of the year and in four-hour blocks of local time (LT). In that repre-sentation, map contours show atmosphericnoise power available (in dB) from a losslessantenna relative to a thermal noise source at288 degrees Kelvin. Those values differ fromthe ones for man-made noise in that a 1 Hzbandwidth is used at a frequency of 1 MHzinstead of 3 MHz. The information on atmo-spheric noise from CCIR also includes how thenoise varies with frequency.

Going to Little Pistol’s contact with St.Louis, around sunrise in January, that’s in theseason when atmospheric noise is the lowestand the same is essentially true for the time ofday. When one works out the details, the at-mospheric noise power in St. Louis would be

about –125 dBW, something like 1 S-Unit lowerthan the level for man-made noise in a residen-tial area.

While it should be borne in mind that at-mospheric noise, like man-made noise, has astatistical fluctuation to it, in this particular casethe man-made noise would seem to be the mostimportant factor, at least at the St. Louis end ofthe path. Of course, the Little Pistol will havenoise of some sort at his QTH too so that wouldhave to be taken into account in consideringthe other side of any QSO.

To summarize the discussion in the lastthree chapters, for any successful communica-tion to take place, three conditions must be met:signals must get through on the operating fre-quency, signal strength must be adequate andthe noise level must be well below the signalstrength. Noise power, especially of man-madeorigin, can be important at any time and on anyfrequency while signal absorption dominatesat the low end of the HF spectrum. Of course,solar activity is important for the choice of op-erating frequency, especially for the high endof the HF spectrum. But all three considerationsare important at any time: MUF, signal strengthand noise power.


In discussing the propagation of LittlePistol’s signals from Boulder to St. Louis, themethods used were straightforward and to thepoint, dealing with the questions of critical fre-quency, signal strength and the presence ofnoise. Now, in preparation for going on to morecomplicated paths and with more than one hopto them, it’s necessary to consider some detailswhich were either omitted or will come upshortly. The first matter deals with wave po-larization.

As everyone knows from the experimentsof Heinrich Hertz more than a century ago,electromagnetic waves are characterized ashaving some form of polarization. The usualapproach is to consider the case of propaga-tion where the electric and magnetic fields ofthe wave are in the plane of the wavefront andperpendicular to the direction of advance of thewave. For wave propagation through vacuumor an isotropic, homogeneous dielectric, thetwo fields, E and H, take their simplest form,perpendicular to each other and linearly po-larized.

That last term, linear polarization, meansthe electric field E, for example, is parallel to aline or axis, changing its magnitude and direc-tion at the operating frequency and the sameis true for the magnetic field H. Graphically, inwave propagation, the fields for an advancingwave may be represented by arrows whoselength and direction vary with time and spaceas shown in Figure 10.1. In that figure, an in-stantaneous view is given of how the electric

field, Ey, varies in the y-direction and the mag-netic field, Hz, in the z-direction for a waveadvancing in the x-direction, to the right in thefigure.

In the case of radio waves, their polariza-tion depends on the transmitting antenna, atleast at the outset. For a dipole, the electric fieldis parallel to the direction of the wire for wavesradiated perpendicular to the wire’s orienta-tion. If the wire is parallel to the ground, thewaves are termed horizontally polarized. Bythe same token, vertically polarized waves areradiated if the axis of the antenna, say a quar-ter-wave vertical, is perpendicular to the earth’ssurface. That’s as simple as it gets but it doesn’tlast long.

The reason for that statement goes backto ionospheric sounding, the technique usedin the late ’20s to explore the structure of theionosphere and the critical frequencies of thevarious regions. While not mentioned earlier,those studies showed there is not one but twowave echoes coming back from each RF pulsesent upward. One, called the ordinary wave O,is just what would be expected for an iono-sphere that contains only free electrons and theother, the extraordinary wave X, results fromthe fact the ionosphere is immersed in theearth’s magnetic field.

Those observations were obtained fromdisplays which plotted the instantaneous fre-quency of RF pulses, as the ionosonde sweptfrom 1-20 MHz, and the times of arrival of theechoes from overhead, using the X- and Y-axes

10: NOW THE DETAILS

Figure 10.1 Electric and magnetic field vectors for a plane-polarized,sinusoidal electromagnetic wave.


signals will be propagated as both O- and X-waves, the next step is to represent the wave’soriginal linearly polarized E-field as the sumof two elliptical waves.

But let’s make life easy for ourselves, us-ing the special case of circular polarization in-stead of the more general elliptical polariza-tion that theory would suggest. We can do thatif the signals are propagated along or close tothe direction of the geomagnetic field. That’llbe good enough for our present purposes andleads to less confusion. So, going to Figure 10.3,the electric field E of the wave which leavesthe LP’s antenna oscillating in a direction par-allel to the earth’s surface is now representedas the sum of two electric fields which rotatein opposite directions, i.e., circularly polarizedwaves.

of an oscilloscope tube. The two echo traces, Oand X, from a sweep in frequency are illustratedin Figure 10.2, an idealized representation ofan ionogram. Of interest is the fact that the twotraces are separated in frequency by about one-half the local electron gyro-frequency. In termsof the propagation of the Little Pistol’s signals,that means the X-wave is returned from a re-gion of the F-layer where the critical frequencyfxF2 is about 0.5 MHz higher than foF2 for theO-wave.

Those results indicate the ionosphere isdoubly refracting, just like a piece of mica or acalcite crystal, and with two different wavepolarizations propagating through it. Whenelectromagnetic theory was applied to theproblem, it was found that the O- and X-wavesare elliptically polarized and with oppositesenses of rotation. If that’s not complicatedenough, features of their propagation dependon the direction of wave travel relative to thegeomagnetic field!

A full discussion of those matters is verycomplicated and has to be left to the experts.For the present discussion, we need the funda-mental points and how the Little Pistol’s DXingwould be affected. To that end, we should thinkof the signals leaving the LP’s antenna as be-ing horizontally polarized since that’s how theLP’s Yagi is set up. But knowing that the LP’s

Figure 10.2 An idealized ionogram fordaytime conditions showing echo heightsas a function of frequency for O- and X-waves.

Figure 10.3 The representation of a linearlypolarized wave as the sum of two circularlypolarized waves, rotating in oppositedirections at the same frequency.

If you stop and think about it, those tworotating field vectors add up to the originalhorizontal E-field from the LP’s antenna, theperpendicular components always being inopposition and cancel while the componentsparallel to the original direction always add.Thus, the oscillating horizontal E-field is re-placed by two circularly polarized waves, ro-tating in opposite directions around their di-rection of travel.

As with any doubly refracting material,we can expect the two circularly polarizedwaves to travel at speeds which are slightlydifferent. The simplest illustration of the point


would be for S-to-N propagation at the mag-netic equator. In that case, theory indicates thecircularly polarized X-wave rotates in the coun-terclockwise (c.c.w.) direction, when lookingnorthward along the magnetic field direction,while the O-wave has a clockwise (c.w.) rota-tion.

Beyond the two different senses of rota-tion for the O- and X-waves, theory indicatesthat the O-wave travels slower than the X-waveso in the time required for waves to reach agiven location, the c.w. O-wave will have ro-tated more about the geomagnetic field direc-tion than the c.c.w. X-wave. If the two instan-taneous rotating fields are combined to obtainthe equivalent linear polarization at that loca-tion, as in Figure 10.4, it’s seen the plane of po-larization has rotated a finite amount about thedirection of propagation in the course of wavetravel.

Figure 10.4 Rotation of the plane ofpolarization from two circularly polarizedcomponents traveling at slightly differentspeeds through the ionosphere.

Obviously, there are other cases of iono-spheric propagation which could be examined,say propagation perpendicular to the geomag-netic field. But in Amateur Radio there are nofixed paths, and directions of propagation rela-tive to the geomagnetic field vary constantly,even on ascent and descent along a path. Sothe idea of rotation of the plane of polarizationof RF is appropriate for discussions of HFpropagation; but the case cited above, termedlongitudinal propagation, is more for generalillustration than any specific application.

With all the different directions of propa-gation as well as path lengths in use duringamateur operations, it’s quite unlikely that theoriginal polarization of a wave would remainby the time propagation carries it to the receiv-ing end of a path. And with large-scale changesin the height of the upper regions of the iono-sphere taking place in the course of a day,reaching minimum heights at local noon in thewinter hemisphere and maximum heights inthe summer hemisphere, it would seem thatjust about any kind of wave polarization mightbe incident on a receiving antenna.

That being the case, for calculation pur-poses, it’s assumed that incoming signals at areceiver are essentially unpolarized at any time,power divided equally between horizontal andvertical polarizations, and a receiving antennadoes not respond to components of the incom-ing radiation which are perpendicular to itsown polarization. Thus, a vertical antennawould not respond to the horizontal compo-nents of incoming unpolarized radiation anda horizontal antenna would not respond tovertical components. All that amounts to a po-larization coupling loss of 3 dB, to be added tothe other losses along a path and bring the LP’ssignal to –88 dBW in residential St. Louis anda S/N ratio that is corresponding lower at 32dB.

At this point, we have to consider extend-ing these methods to longer paths, now includ-ing surface reflections between hops. However,one thing is lacking, the effects of ground re-flections. For that, it’s assumed that reflectingsurfaces are smooth and characterized by elec-trical conductivities and dielectric constants.

When the problem is worked out in de-tail, it turns out that the plane of polarizationrotates in proportion to the average magneticfield and columnar electron content along thewave path and inversely with the square of theoperating frequency. That effect is called “Fara-day rotation” and results in the Faraday fad-ing of satellite signals as their plane of polar-ization rotates relative to simple linearly po-larized antennas.


Figure 10.5 Reflection coefficients on 14MHz for typical ground as a function of anglewith the horizon for both horizontal andvertical polarization.

Next, Fresnel’s equations from electromagnetictheory are used to find the reflection coefficient,the ratio of the reflected E-field to that incidenton the surface, for different operating frequen-cies, polarizations and angles of incidence. Inthat regard, Figure 10.5 shows how the reflec-tion coefficient on 14 MHz varies with radia-tion angle for typical ground material and thetwo polarizations.

The striking feature of that figure is thebehavior for vertical polarization where thereflection coefficient goes through a minimumaround 26 degrees. That angle is called thepseudo Brewster angle after Brewster’s anglefor reflections from a pure dielectric in the op-tical case where the reflection coefficient forvertical polarization actually goes to zero. Un-der those circumstances, unpolarized light be-comes horizontally polarized on reflection atthe Brewster’s angle for the material.

If radio waves are considered as unpolar-ized, as suggested above, one can calculate theintensity loss on ground reflection by averag-ing the reflection losses in intensity for hori-zontally and vertically polarized waves. Sincethe earth has vast oceans and ice caps as wellas continents, one needs to consider where re-flections might take place along a path and in-

clude the ground losses which result. For thethree types of surface — ice cap, ground, seawater — the intensity losses on 14 MHz areshown in Figure 10.6. There, it is seen how lowlosses result from ocean reflections and highlosses from reflections on polar ice caps. Be-yond those results, calculation shows a weakvariation of ground losses with frequency, be-ing slightly lower on 7 MHz and higher on 21MHz for the various materials.

While on the subject of reflections, itshould be noted that reflection surfaces are notalways smooth or parallel to the local horizon.While the angles of reflection from a surfacewill always equal the angles of incidence ofwaves on it, the radiation angle of the outgo-ing waves will depend on the inclination of thesurface. That will be important if the size orextent of a reflecting surface is many wave-lengths across. If the scale of variations of sur-face regularity is smaller, then the reflectionprocess becomes more diffuse in nature, not themirror reflection usually considered. That re-sults in more of a scattering of incoming radia-tion, an additional loss to be considered everytime a ray path reenters the ionosphere after areflection.

Earlier, it was pointed out that some fo-cusing gain might be expected from iono-

Figure 10.6 Reflection loss on 14 MHz as afunction of angle with the horizon for threedifferent surfaces.


spheric refraction because of ray convergenceresulting from the concave nature of the F-re-gion. By the same token, some de-focusing losswould be expected because of ray divergencefrom the convex nature of the earth’s surface.Those effects would produce a slight gain, saya dB or so, over a path with several hops as thenumber of ionospheric refractions along a pathof N hops exceeds the number of surface re-flections, N-1, by one.

The last point to consider, no matter howmany hops on a path, is fading and its sources.That question comes up when one wonders ifa contact can be maintained once it’s made, inshort a question of reliability. In that regard,there are fades and FADES, the latter taking asignal down to the noise level. One can copewith the former but a path starts to break downwhen a signal disappears, dropping by morethan 3 S-units. So whether a path is reliable ornot depends on the most optimistic estimateof the S/N ratio and whether it could suffer adrop of 18 dB or so and still be in the rangewhere a signal would be readable.

Fading really has its origins within thehigher portions of the ionosphere, movementsof the refraction region affecting propagationalong a path. Thus, a rise of the F-region mightshift ray paths so that the skip region around atransmitter has suddenly increased, placing thereceiver in the skip zone. That is more of amatter of propagation failing, the path no

longer being open for that operating frequency,and amounts to what is termed “MUF fading,”the maximum useable frequency for the pathfalling below the transmitter frequency in useand then perhaps rising again.

Of course, there is the other side to thatcoin, the shift in ray paths placing the receivercloser to or farther away from where it wouldbenefit from skip focusing. That would still befading in the sense that the signal strengthwould rise and fall but is far less catastrophicthan skip fading.

Another form of fading results fromchanges in wave polarization. Thus, if the elec-tron content along a path changes for any rea-son, there will be a rotation in the plane of po-larization at the receiving antenna with a cor-responding change in signal strength. Such fad-ing may be observed even on single-hop paths.Of course, having antennas with both horizon-tal and vertical polarizations and the ability toswitch the receiver rapidly between themwould minimize fading of that kind. But itshould be noted that technique is generally lim-ited to simple antennas, say dipoles and quar-ter-wave or half-wave verticals.

At this point, the essentials have been cov-ered and it’s time to stop and take the first ofseveral overviews of HF propagation. The nextone will illustrate points that’ll become impor-tant as the hop structure and propagation cir-cumstances become more complicated.


Now that a one-hop path has been exam-ined, from its main features to some of the finerdetails, the next step is to move on to longerhops and the more general aspects of propa-gation. While one-hop paths could be part ofthe DX scene for the Little Pistol, they’re usu-ally not much longer than 3,500 km because ofthe range of F-layer heights. That being thecase, any longer path will involve two or morehops, span more than one time zone and go offin all sorts of interesting directions. No won-der Little Pistol finds them exciting.

Just to show what possibilities are outthere, look at Figure 11.1, which shows an azi-

On the subject of map projections, distor-tions are important to note, the Mercator pro-jection representing the two geographic polesas straight lines instead of points; in addition,the distortions in that projection become pro-gressively greater for regions farther away fromthe equator. In contrast to the Mercator projec-tion, distortions in the equidistant azimuthalprojection become greater in going away fromthe center of the map and, as noted above, theextreme is when the antipodal point, 20,000 kmdistant, is represented by a circle. In any event,with the Mercator projection, great circle pathsat low latitudes look like segments of sinecurves but paths which go through high lati-tudes may look like segments of the termina-tor at the equinoxes.

While those maps present a wealth of in-formation, they convey nothing about the na-ture or topography of the terrain along anygreat circle path. That being the case, thosemaps must be supplemented by other informa-tion about reflection surfaces as signal losseswill vary along a path, differing by as much as3-6 dB between sea water and ground reflec-tions. While the earlier discussion involved achosen path, Boulder to St. Louis, other possi-bilities present themselves and it is importantto know where the great circle paths fall on amap, and for more than just concerns aboutpoor reflection surfaces. But those questions areeasily settled by a distance-heading calculationwhich starts with the coordinates of the trans-mitter and receiver and brings forth the greatcircle beam heading and distance. Knowingthat F-region hops are in the 3,500 km range,it’s a simple task to decide how many hops willbe involved, the heading for the path and thesignificant features which would be encoun-tered by RF going along the route in question.

The simple one-hop path treated earlierhad critical frequencies, foFE and foF2, givenat the outset, for daybreak on the path betweenBoulder, CO and St. Louis, MO in mid-Janu-

11: MOVING ON

Figure 11.1 Azimuthal equidistant projectionmap centered on Boulder, CO.

muthal equidistant projection map centered onBoulder, CO. In that map projection, great circlepaths are straight lines and the beam headingsfor DX are evident, as the map is centered on aline passing through the north and south poles.In addition, that map projection gives distancesfrom its center on a linear scale, grid lines cor-responding to locations which are multiples of5,000 km from Boulder and all the way out toBoulder’s antipodal point at 40° S, 75° E, 20,000km distant and not far from Amsterdam Island(FT8Z).


ary. The next question has to do with how onefinds similar ionospheric features for anotherpath, say one requiring two hops in going frompoint A to point B, and another time. Thatbrings up the $64 question — how to predictcritical frequencies of the ionosphere.

Of course, if one had maps like those inFigures 4.2 and 4.3, it would be a simple task— just work out the geometry of the path us-ing an azimuthal equidistant projection map,locate the midpoints of its hops and then usethe contours on an ionospheric map to find thecritical frequencies involved. Sound simple? Itis, but not quite as indicated.

For one thing, the height of the F-layerwould not be the same at the local times of thetwo midpoints of the path. That means that ifRF were launched at some angle, the hoplengths of the two sections of the path woulddiffer in length, the result being that the RFmight undershoot or overshoot the receiver atthe end of the path. So how is the launch anglearrived at? There are two choices, the hard wayand the elegant way.

The hard way involves using the idea ofmirror reflections at the virtual heights alongthe path. By working out the geometry alongthe curved earth, mirrors and all, it is possibleto find the launch angle which connects thetransmitter and the receiver. But how much dovirtual heights differ from true heights of re-fraction on a path and how close do pathlengths agree for the two cases?

In regard to the first question, Figure 11.2shows virtual heights for mirror reflection andtheir variations with local time using a typicalionosphere, in this case during the month ofDecember. For a curved ionosphere, it turns outhop lengths at radiation angles between 3 and15 degrees differ by about 1% or so when mir-ror reflection and ray tracing are compared.That should be quite satisfactory for the prob-lem at hand so given the local times at the mid-points, a launch angle can be found by usingvirtual heights at midpoints of the hops.

The next step in finding whether the pathfrom point A to point B is open or not for thatlaunch angle is to find the values for foF2 atthe midpoints of the hops and the time in ques-

Figure 11.2 F2-layer heights for December.

tion. That means using an ionospheric map.When those are in hand, the problem then be-comes one of finding the effective vertical fre-quencies, mentioned earlier, for the launchangle that takes the path from point A to pointB.

While the radiation angles will be thesame along the path, even after ground reflec-tion, the differences in critical frequency andvirtual height between hops will give differ-ent frequencies, f1 and f2, which may be re-turned to earth at oblique incidence. Clearly,the lower of the two is the one that ensures thatRF will proceed from the transmitter to the re-ceiver and is the maximum useable frequencyor MUF for the path. Any higher choice of fre-quency will result in RF being lost to infinityon the hop with the lower value of foF2.

If that procedure were carried out for ev-ery hour of the day, the daily variation of theMUF would be obtained and indicate timeswhen one amateur band or another might betried in making contacts from point A to pointB. All that’s needed is ionospheric maps for theF-region critical frequency foF2- and F2-layerheights for the month in question; the rest isjust arithmetic. And that’s the tedious methodthat DXers used before PC programs came onthe market, the National Bureau of Standardsof the U.S. Department of Commerce provid-ing all the ionospheric maps, nomographs andgraphical aids needed.

Nowadays, HF propagation programs


contain information from which ionosphericdata may be derived and once the user indi-cates a path, the date and sunspot number, itonly takes the computer a few seconds to dothe type of calculation outlined above and putthe results on the screen or the printer. So thehard way now becomes easy. The elegant way,mentioned above, is to carry out ray traces atthe operating frequency, using model iono-spheres appropriate for each hop or local time,and step through radiation angles in an itera-tive process until the resulting ray trace fallswithin a target size, say no more than 25 km,at the end of the path. That’s not an MUF cal-culation but does indicate whether a band isopen or not at the time in question. Higherbands may be tried and limits set on the fre-quencies which may be used in the course of aday.

That sort of ray tracing can be carried outin the new Canadian computer program,SKYCOM, and used in many different ways toillustrate how HF propagation varies on a path,say with time, launch angle or frequency. Andthe program is quite flexible in the sense thatany number of hops can be dealt with. Whilecomputing time is a factor, the limiting one isusually the curiosity and patience of the user.

But ray tracing brings up an importantpoint about using computers in ionosphericcalculations. In particular, ray tracing and MUFcalculations require some sort of experimentaldatabase for critical frequencies, how they varyover the globe with seasons and sunspot num-bers. In that regard, a user working withSKYCOM may employ either the URSI modelionosphere in calculations or the model iono-sphere proposed by CCIR but most other MUFprograms use just the CCIR database. In addi-tion, SKYCOM gives critical frequency datafrom either model for any geographical loca-tion, date and sunspot number.

An important question is how the CCIRdata is actually used in making MUF calcula-tions. For example, the German programMINIFTZ4 carries out what seems to be acoarse interpolation of CCIR data in findingcritical frequencies for its calculations. On theother hand, the methods of Raymond Fricker,

formerly of BBC External Services, are basedon the CCIR data but give the spatial and tem-poral variations of foF2 data by fitting themwith a number of mathematical functions.Two of Fricker’s better known F-layer algo-rithms, MICROMUF 2+ and MAXIMUF, giverather comparable results but they differ incomplexity as the former algorithm uses 13functions and the latter 26 functions. The lastalgorithm, MAXIMUF, with some modifica-tions, has been used in HF prediction programssuch as MINIPROP and IONSOUND.

While MINIFTZ4 and MAXIMUF arebased on CCIR data, the amount actually usedwas limited to that from times around the equi-noxes and solstices. MINIFTZ4 carries out in-terpolations in latitude, longitude and timeusing the CCIR data itself while MAXIMUFuses the data from the equinoxes and solsticesbut MAXIMUF’s functions were optimized togive the best fit of the variations of foF2 withlatitude, time of day and year.

The American HF prediction program,IONCAP, is well known and, after its develop-ment by the Central Radio Propagation Labo-ratory of the U.S. Department of Commerce, itwas first distributed in 1978. The method bywhich IONCAP represents the spatial and tem-poral properties of the ionosphere is termed“numerical mapping” and uses a series ofmathematical functions with coefficients ad-justed for each month and time of day. In thatregard, the IONCAP program includes datafiles to generate coefficients for each month ofthe year. While IONCAP may be used to findMUF information, it’s much more flexible, hav-ing a total of 29 different methods to predictthe performance of HF radio systems.

In its original form, IONCAP operatedwith computing systems which used IBM-punched cards for its data input. That formatwas carried over to the version of IONCAPdeveloped for personal computers and is quiteslow, tedious, even exasperating to use. Nowa-days, one can enjoy all the benefits of IONCAPwithout those troubles as the CAPMAN pro-gram, based on IONCAP, employs a driver toallow data input without the problems that gowith the punched-card format.


be open from around 0700 UTC to 1600 UTCwhile the 28-MHz band would be open fromabout 0800 UTC to 1500 UTC. As a pattern,that’s the way the ionosphere works, the lowerfrequency band opening first and closing last.It’s just a matter of solar illumination, at leastwhen it comes to opening.

And if one took those MUF curves at facevalue, it would seem that the 7-MHz bandwould be open 24 hours a day! And the sameis true for the 3.5-MHz band. In that regard, Ican only say it’s just amazing what MUF datawill suggest by itself, especially when used in-correctly. The problem, as you well know fromthe earlier discussion of the Little Pistol’s path,is MUF curves do not contain any sort of infor-mation about the factors that go into signalstrength, such as absorption and reflection lossas well as transmitter power and antenna gains,and noise, mostly from man-made sources. Inshort, MUF curves involve wave refraction athigh altitudes and are only one-third of thestory, not covering any of the ionospheric ef-fects at low altitude or the noise factors that gointo HF propagation.

One can use the present case, the pathfrom Frankfurt to Riyadh, to show how propa-gation works, going beyond the MUF curvesin some detail. Thus, there are actually twothings which defeat full 24-hour usage of the7-MHz and 3.5-MHz bands, ionospheric ab-sorption and the E-region cutoff frequency.Absorption was discussed earlier, the situationwhere signal strength is reduced because ofwave energy transfer to the atmosphere by theelectron-molecule collisions in the D-regionwhen the RF goes by.

And since ionization in the D-region re-sults from illumination by the sun, absorptionprocesses begin with sunrise and end with sun-set. While not mentioned earlier, during day-light on a path, absorption varies inversely withthe square of the frequency for ordinary or O-waves at the upper end of the HF range. At thelower end of the HF spectrum, extraordinaryor X-waves suffer even greater absorption withfrequency because magneto-ionic effects be-come more important when the radio frequen-cies are comparable to electron gyro-frequen-

Because of the large number of coefficientsused in its numerical mapping, IONCAP is feltto give the best representation of ionosphericsituations and has become the standard againstwhich the results of other programs are com-pared. As an example, Figure 11.3 shows MUFcurves for IONCAP and two of Fricker’s algo-rithms, MICROMUF 2+ and MAXIMUF, for a4,330-km path in the month of January, head-ing 117 degrees east of north from Frankfurt,Germany to Riyadh, Saudi Arabia. While thereis general agreement in shape, one can see thatsignificant departures, as much as +/-5 MHz,may exist between the results for MUF valuesfrom those two algorithms and those fromIONCAP.

There are other prediction programs lesssophisticated than those noted above, sayFricker’s MINI-F2 and MINIMUF. Their pre-dictions are far less satisfactory as they’re bothbased on single-function algorithms, unable todeal effectively with the wide range of paths,seasons and solar conditions and still give re-sults which are comparable to IONCAP andthe other programs.

Returning to the MUF curves in Figure11.3, one might conclude from their shape thatthe 14-MHz band would be open from about0600 UTC to 1800 UTC. Then, depending onwhich curve is chosen, the 21-MHz band would

Figure 11.3 MUF curves from threepropagation programs for the path fromFrankfurt, Germany to Riyadh, Saudi Arabia.


cies. For frequencies like 3.5 MHz and 7-MHz,paths are more vulnerable to absorption thanhigher frequencies such as 21 and 28 MHz.

Returning to the present case, D-regionabsorption is so great from 0400 UTC to 1500UTC as to make the path unworkable on 3.5MHz. In the same fashion, the path drops outon 7 MHz from around 0500 UTC until 1400UTC. But only part of that is due to signal ab-sorption; the remainder is due to signals beingcut off from the F-layer by the rise in criticalfrequency foFE of the E-region with solar illu-mination. In that regard, the screening of lowfrequency signals from the F-region by E-re-gion ionization was illustrated earlier in Fig-ure 7.5.

So the simple propagation mode that in-volves two F-hops during darkness changes toone with mixed modes, an F-hop on the darkend of the path and an E-hop in sunlight orjust E-hops when the path is fully illuminated.The latter suffer from heavy D-region absorp-tion and bring signal levels down to the pointwhere the path is closed for all intents and pur-poses. Those changes in propagation mode areshown symbolically in the four parts of Figure11.4: going from a 2F mode in (a) to a mixedmode, 1F1E, in (b), then the 2E mode suffers ablackout because of great absorption, so thento a mixed mode, 1E1F, in (c) and finally the 2Fmode again in (d).

In analytical terms, the changes in the E-cutoff frequency and ionospheric absorptionalong the path are shown in Figures 11.5 and11.6, respectively. The first of those two figuresshows the E-cutoff frequency at each end of thepath. Since the E-cutoff frequency does not ex-ceed 10 MHz, signals on the 10-MHz band andall higher bands are not affected by the E-layer,at least as far as being cut off.

However, the 7-MHz and 3.5-MHz bandsare seriously affected, as noted above. FromFigure 11.5, one sees that the sunrise on the lowlatitude part of the path is earliest and thus theE-cutoff condition sets in there first. And sincethe higher latitude end of the path is about 40degrees farther west in longitude, sunset on thepath is later there and the E-cutoff conditionends there last. Those results are the basis for

the mode changes in Figure 11.4.The other effect of solar illumination on

the path is signal absorption. That takes placewhen signals traverse the D-region, on ascentfrom the transmitter and descent to the receiveras well as just before and after the ground re-flection. In that regard, Figure 11.6 gives thetotal absorption along the path for the five har-monically-related amateur bands. It should benoted that the absorption values were calcu-lated without regard to whether the band wasopen or not. The results show that total absorp-tion along the path rose with solar illumina-tion and reached peaks of 12 dB (2 S-Units) orless on 14 MHz and the higher bands but itreached 50 dB (8 S-Units) or more on the lowerbands, essentially shutting them down for com-munication purposes.

To those losses the signal loss on groundreflection must be added. The 2F mode is themost durable one, either at night on the lowerfrequencies or daytime on 14 MHz and above.Ground loss is independent of solar illumina-tion and for the radiation angle (13°) whichconnects the end points of the path, reflectiontakes place on the ground in western Turkeywhere the loss would be about 4 dB of 14 MHz.The ground reflection points for the mixedmodes would be different than the one givenabove for the 2F mode but those modes are onlyof brief duration and will not be considered

Figure 11.4 Symbolic representation of thepropagation modes for the Frankfurt/Riyadhpath.


Figure 11.6 Ionospheric absorption for theFrankfurt/Riyad path as a function offrequency.

Figure 11.5 E-region cutoff frequencies atthe two ends of the Frankfurt/Riyadh path.

further.The last aspect to consider for the present

path is noise, either atmospheric or man-madein origin. With regard to atmospheric noise, thelevel at the northern end of the Frankfurt/Riyadh path in January would be comparableto that found at the same time in the easternstates in the USA while at the southern end ofthe path, the noise would be comparable to thatin the Caribbean, but certainly nothing intenselike the levels of noise in South America orSouth Africa. Man-made noise levels at the twoends of the path would probably be higher thanatmospheric values, at a residential level as aminimum and more likely industrial since bothlocations are major cities.

In any event, propagation for other twohop paths could be worked out as done earlierbut now using the ideas and details outlinedabove: MUF values, E-cutoff frequencies, D-region absorption, ground loss and man-madenoise. The only details still to be settled areoperating frequency, transmitter power levelsand antenna gains at both ends of the path. Butlest anyone has missed the point, let me stressagain that whether propagation on a path isworkable or not depends on whether three con-ditions are met, all at the same time: an openband (MUF), good signal strength and lownoise levels. As we say in the trade, “MUF isnot ENUF!”


The discussion up to this point has dealtwith nothing more than two hops, paths whichgo out about 7,000 km from the Little Pistol’sQTH. Thus, looking at the map centered onBoulder in Figure 11.1, one sees with two hopsthe LP’s signals could reach out to WesternEurope and North Africa, Asiatic Russia, someislands in the Pacific and the northern part ofSouth America. But to reach out any farther,say beyond the Balkans, into Central Africa, tothe Orient and the rest of South America, an-other hop would be needed. But what wouldthat mean in terms of the ideas developed sofar?

Let’s take a round number, 10,500 km, fora path length; for that distance, 50% greaterthan the length of the two-hop path, the LittlePistol’s signal strength would fall by about 3.5dB due to additional signal spreading. That’snot even a full S-Unit; not to worry. But anothersurface reflection would be involved, maybeoff of sea water for a change. In any event, thatwould be another 1-4 dB loss, whether day ornight. And then there’s D-region absorption,maybe something like another 6 dB loss on 14MHz if the path were sunlit or essentially noth-ing with darkness of the path.

Putting the losses all together, the sumwould range from about 5 dB to 14 dB or about1 to more than 2 S-Units lower signal at the DXend of the longer path. The good news is thatthe noise would probably be about the sameunless the Little Pistol was trying to make aDX contact with Indonesia, South America orSouth Africa in their summer thunderstormseason. Those directions would cover about 270degrees of the field of view from Boulder, as-suming the Little Pistol’s QTH is not too closeto the Flatiron Range on the outskirts of Boul-der. The rest of the compass directions wouldextend from Siberia to Central Europe, goodDX paths all.

With DXing out to those distances, therange of longitude or number of time zones

along a path increases. And in trying to worksome DX, it’s not out of the question to thinkthat one end of a path could be sunlit and theother in darkness. Of course, that raises ques-tions about critical frequencies after sunset. Butanother question would be how critical fre-quencies vary over great distances, say acrossthe northern polar cap or to the south, acrossthe equator. For answers to those questions, weneed to turn to ionospheric maps for foF2, likethose in Figures 4.2 and 4.3.

Take Figure 4.3 to start with; that’s from atime of high solar activity and makes for a moreoptimistic and cheerful discussion. The left halfof that figure is the part of the earth in sunlightand the right half is in darkness. Of course, fullillumination of the part of the earth facing thesun continues as it rotates on its axis; the onlything that changes is the part in daylight, thesun moving across the sky, east to west.

Now the point has been made that theionosphere is created by solar UV so one hasto think that the creation of an ionosphere is acontinuing process, centered at the subsolarpoint and thus extending over half the earth’satmosphere. Unlike the oceans which have tre-mendous inertia, the part of the atmosphereclose to the earth co-rotates with it. Thus, whilesolar illumination of the earth is steady in time,fresh atmosphere rotates into solar viewaround dawn, then it’s carried past the noonmeridian and finally out of solar view at thedusk. Relative to us fixed on the ground, theionosphere, represented by the foF2 contoursin Figure 4.4, advances as a whole across thesky, just like the sun. But there are interestingfeatures of the contours in Figure 4.3, firstcrowded around the Greenwich Meridianwhere the sun is rising at 0600 UTC and stillgoing off into darkness, beyond the sunset lineat 180° E longitude.

The rapid rise in foF2 values at sunrise isdue to the buildup of electron density at F-re-gion heights with the onset of photoionization.

12: NOW TO THREE OR MORE HOPS


But the slow decay of ionization in foF2 valuesafter sunset is quite different from the rapiddisappearance of ionization in the E-region,shown by foFE contours in Figure 4.4. Thosechanges, taking place at two different altitudes,should be discussed at this point so we will di-gress to indicate their differences and theirsimilarities.

In its ionized state, a small volume of theionosphere contains a number of negative elec-trons, important to propagation, and an equalnumber of positive ions, keeping the volumeelectrically neutral. Normally, the ionization isproduced by the effects of solar photons in theUV and X-ray range. At low altitudes, the posi-tive ions will come from nitrogen and oxygenmolecules and at high altitudes there will alsobe positive ions from oxygen atoms.

In that volume, the electrons, ions andneutral constituents will all be colliding andinteracting. Electrons can recombine with thepositive ions and produce neutral structuresagain. With molecular ions, the recombinationmay result in the molecule being dissociatedinto atoms; for example, an oxygen molecularion can recombine with an electron and resultin the formation of two oxygen atoms. Thatprocess is called dissociative recombination.Another possibility is an atomic oxygen ion canrecombine with an electron and result in a neu-tral oxygen atom and the emission of a lightphoton. That process is called radiative recom-bination.

If that wasn’t complicated enough, a posi-tive oxygen atomic ion (O+) can exchangeplaces with a nitrogen atomic ion (N+) in a ni-trogen molecular ion (N2+) and create a nitricoxide ion (NO+). That, in turn, can recombinewith an electron and produce a nitrogen atomand an oxygen atom. In short, it’s a big chem-istry lab up there and anything that can hap-pen between ions and molecules will happen.

Now it turns out that recombination goesslowly when an electron encounters an oxygenatomic ion (O+) and much faster when it en-counters any of the molecular ions. But theoxygen atomic ion is found mainly up in the F-region and the molecular ions are found downin the D- and E-regions. From the latter, one

can see why the D- and E-regions disappear sorapidly with sunset and from the former, onecan understand why the electron density in theF-region builds up rapidly at sunrise to highconcentrations and, in part, why it lingers longafter sunset.

In those few paragraphs, some of the mainideas behind ionospheric chemistry have beenpointed out. Obviously, the reactions betweenelectrons and positive ions can affect propaga-tion by shifting the electron density one wayor the other. As a matter of fact, the “winteranomaly” for the F-region, higher foF2 valuesin the winter when the sun is lower in the sky,find its explanation when those simple reac-tions are combined with atmospheric chemis-try. Actually that is great for DXing as with thesun lower in the sky, the MUFs are higher yetthere is less D-region absorption. But we’ll getto that shortly.

Other interesting features of the foF2 mapin Figure 4.3 are the low values of foF2 in thepolar regions and high values found at equa-torial latitudes, particularly in the afternoonand early evening hours of local time. The lowvalues of foF2 in the polar regions will be seento have a solar origin, involving the interac-tion of the geomagnetic field and solar plasmastreaming by, while the high values at low lati-tudes result from the redistribution of ioniza-tion by tidal effects in the atmosphere.

At this point in the discussion, features ofthe ionospheric maps across greater distancesare of interest in that the Little Pistol can ex-plore more of them with longer paths whichreach beyond the middle range of latitudes.Thus, there is propagation on transpolar andtransequatorial paths to consider. And there arethe effects of solar activity to add to the dis-cussion, how changes in sunspot numbers mayaffect the reliability and reach of propagationpaths. As a matter of fact, once we get into pathsof three or more hops, the discussion soon be-comes wide open, almost to the limit of round-the-world (RTW) signals. So let’s turn to it,starting with the discussion of MUF predic-tions.

On two-hop paths, the maximum useablefrequency (MUF) at a given time is obtained


by finding the limiting frequency for each hopand then taking the lower of the two values.The same idea is used for longer paths but con-sideration is limited to hops at the two ends ofthe path, not in the middle. That approach goesback to WWII when Newbern Smith in the USAand K. W. Tremellen in the U.K. developed theidea of “control points” for a path.

Their argument was that propagation ona path failed most often at one end or the other.Thus, the approach was to find the MUF val-ues at two control points on the great circlepath, at 2,000 km from each end of the path,and use the lower of the two to represent theMUF for the path as a whole. An additionalpair of points at 1,000 km from the ends of thepath may be used to examine whether the E-region plays any role in controlling propaga-tion, as discussed earlier.

That technique is used extensively in MUFprograms for computers and gives reliable re-sults for two hops when the role of the E-re-gion is included. Going to three or more hops,results vary in reliability, depending on thepath under consideration. For example, forthree or more hops that stay within the mid-latitude range or go south toward or across theequator, the method proves to be quite reliable.That is the case as critical frequencies foF2 atthe control points near each end are lower thanthat of the middle hop at a lower latitude, asmay be seen from Figure 4.3.

But looking at that figure, one sees thatfoF2 values at high latitudes are lower than atmid-latitudes. As a result, the MUF on a middlehop of a longer, poleward path could be lowerthan the operating frequency and the full pathfails to support propagation even though datafrom the control points at the ends would sug-gest otherwise. An illustration of those featuresmay be seen in MUF calculations for a pathfrom San Francisco to London for the samemonth, March. That path involves a great circledistance of 8,732 km, up across Hudson Bay inCanada and the southern tip of Greenland. Thepath can be broken down into three equal partsand when the IONCAP prediction program isused to find MUF values for each, the resultsshown in Figure 12.1 are obtained. There, the

Figure 12.2 MUF variations as a functionof time for the entire path from SanFrancisco to London.

Figure 12.1 MUF variations as a functionof time for three hops on a path from SanFrancisco to London, England.

first hop in Western Canada and the third hopin the North Atlantic show strong daily varia-tions in their MUF values but the second hopis somewhat more poleward than the hop overthe North Atlantic. As a result, its MUF varia-tion in not as great and controls propagationfor part of the day, as seen in Figure 12.2 whichgives the MUF for the entire path but calcu-lated using three control points as well as theusual pair of points.

In the case of three hops on a path which


Figure 12.3 MUF variations as a functionof time for the path from Boulder toAsunçion, Paraguay.

crosses the geographic equator, the critical fre-quencies at the ends of the path are the con-trolling factors. But for that type of path, thevalue of the MUF may be significantly largerthan for another path of comparable lengthwhich stays within one hemisphere.

That is the case because of the higher criti-cal frequencies at low latitudes and is illus-trated by MUF curves derived using IONCAP,shown in Figures 12.3 and 12.4. There, twopaths are used, both from Boulder in the monthof January but one path (8,798 km) going acrossthe equator to Asunçion, Paraguay (Figure12.3) and the other path (8,061 km) to Madrid,Spain at almost the same latitude as Boulder(Figure 12.4).

The striking feature of Figure 12.3 is themagnitude of the peak MUF value at 1600UTC, almost twice that shown for the north-erly path in Figure 12.2. However, since thecalculations were for the month of January, itis not clear from the MUF curve for the pathas a whole if the difference is due to winterconditions north of the equator or summerconditions to the south. The foF2 map in Fig-ure 4.3 is of no help either, being appropriatefor an equinox instead of close to a solstice, asin Figure 12.3.

But that’s where the results for the sec-ond path in Figure 12.4 come in. The great

circle path goes from Boulder to Madrid, Spain,starting and ending at 40 degrees north lati-tude and never going above 47 degrees northlatitude. The two MUF curves are for July andJanuary; the first shows only a broad, feature-less variation during summer while the secondshows a large, striking peak in MUF duringwinter.

The winter increase in the foF2 frequency,as shown by MUF values in Figure 12.4, istermed the “winter anomaly” as it is just op-posite to the behavior of other critical frequen-cies like foFE which decrease with lower solarelevation in winter. By way of explanation, thewinter peak in MUF means more ionizationexists in the F-region or less ionization is lostby recombination in the winter. Thus, the win-ter anomaly has to do with the ratio of the pro-duction and loss of electrons. Production is di-rectly related to the presence of atomic oxygenin the region and losses are due to recombina-tion of molecular ions, or equivalently, thenumber of molecules.

Now the study of atmospheric chemistryhas shown that the ratio of the number of oxy-gen atoms per unit volume to that for nitrogenand oxygen molecules varies during the yearat F-region altitudes. In fact, that ratio is greaterin winter than in summer, thus favoring theproduction of electrons from O-atoms over

Figure 12.4 MUF variations as a function oftime for the path from Boulder, CO to Madrid,Spain in January and July.


Figure 12.5 Ionospheric variability shown byfoF2 soundings from Slough, England. FromPiggott and Rawer [1972]

losses of electrons by recombination in molecu-lar interactions. With the balance tipped in thatmanner, the electron density at F-region heightsis greater in the winter months and, by the sametoken, so are MUFs for propagation.

Those variations in critical frequenciestake place on a long time scale — months. Thequestion then is whether there are other largevariations of critical frequencies on a shortertime scale, say from one day to the next at agiven hour of the day. The answer there is inthe affirmative and the effect is greatest in thewinter months. That brings up the larger ques-tion of statistical fluctuations of critical frequen-cies, a very important subject for Amateur Ra-dio.

But first, to anchor the discussion in thereality of experimental observations, look atFigure 12.5 which shows foF2 values recordedfrom ionospheric sounding at Slough, Englandin January 1969. For the month of January, thereare 31 measured values of foF2 for each hourof the day and the data points in the figure tellthe story, the ionosphere showing considerablevariation in its characteristic features. As is thecustom in dealing with such situations, the firstquantity that’s derived for an hour of data isthe monthly median value. That value dividesthe data set, half of the data points for that hourlie at higher values of foF2 and half at lowervalues and the line in the figure traces the me-dian values through the hours of a day. But thatonly settles the middle of the range of foF2 val-ues for a particular hour of the day in themonth; the spread of values above and belowthat median value remain to be determined.Again, use is made of statistical definitions andterminology, the upper and lower decile val-ues; they are the dividing lines for the upper10% of the 31 data points for the hour in ques-tion, say the three highest values of foF2, andthe lower 10%, the three lowest values of foF2.

A close examination of the data in Figure12.5 shows the monthly median value at 1600UTC is 7.6 MHz, the upper decile value is 9.2MHz and the lower decile value is 6.8 MHz.Another way of expressing the meaning of theupper and lower decile points is to say that 90%of the observations at that hour were above the

lower decile value, 10% above the upper decilevalue and 80% of the observations between thetwo decile values. If the ionosphere overSlough, England were the control point for aone-hop path, at 1600 UTC, the observed maxi-mum useable frequencies on the path wouldhave exceeded those derived from the lowerdecile, median and upper decile values of foF2on 90%, 50% and 10% of the days, respectively.

Clearly, the daily variation of foF2 valuesshown in Figure 12.5 means that predictionsbased on them would show the same variabil-ity. So, noting the large change in foF2 betweenday and night as well as the near constantspread in daily foF2 values, it follows that thepercentage variation in predictions would beparticularly large during a winter night. Butexperience shows that the percentage variationat night decreases as the seasons change, be-ing smallest during summer. In regard to varia-tions in predictions, the IONCAP programgives three different values for the maximumuseable frequency at a given hour on a path:the highest possible frequency (HPF) from theupper decile value mentioned above, the MUFfor the median (50%) value of foF2 and the FOT


(an abbreviation of optimum working fre-quency from French) which is taken as 0.85 ofthe median MUF.

In practical terms, that means if an oper-ating frequency falls on the HPF of a path, thechances are one-in-ten that the band will beopen at the time in question. The chances are50-50 if the band frequency falls on the me-dian MUF and close to nine-in-ten if it fallsaround the FOT. But it should be borne in mindthat “open” has nothing to do with signalstrength or noise, the other two quantitieswhich determine the reliability of propagationat a given hour.

In that regard, an interesting exercise inpropagation is to compare how long a path isavailable just in terms of operating frequencyand how long signals are strong enough for realcommunication. The latter is usually a smallfraction of the former, meaning that a lot of timecould be wasted pursuing a DX contact just onthe basis of MUF data alone when signalstrength information would be more to thepoint, achieving success on a path.

Of course, the remarks in that last para-graph imply that the propagation program inuse is one that includes signal strength, per-haps even noise, in its predictions. In the earlydays of computer predictions, the amateurcommunity relied heavily on programs whichonly had a MUF capability. That was adequateon the higher bands, say 10 and 15 Meters,during times of high solar activity but fails mis-erably at low levels of activity, toward solarminimum. Now, with computers havinggreater memory and MUF programs comingcloser to IONCAP’s predictions, full-servicepropagation programs are the norm, not theexception.

But progress in that direction has come ata price in another, the loss of awareness of theionosphere as a whole. Earlier, when more te-dious, manual methods were used in makingHF predictions, the user had access to iono-spheric maps and could see how the iono-sphere changed with time, say shifting north-ward, largely within the bounds of the termi-

nator, with the coming of summer and thensouthward with the onset of winter. And thenthere are the changes in solar activity, the iono-sphere puffing up with the rise in activity to-ward solar maximum and then the deflationon going toward solar minimum.

The idea of puffing up may be seen bycomparing Figures 4.2 and 4.3, in that order,and that of deflation may be sensed by think-ing that an ionosphere at solar maximum, asin Figure 4.3, is then followed 6-8 years laterby one at solar minimum, as in Figure 4.2, whenthe cycle ends. But the exciting thing for Ama-teur Radio is that the whole process will startall over again, a renewal with the next cycle,with everything that goes with solar activityto be repeated again, DX in all its glory!

Beyond those ideas, the decline in the useof ionospheric maps resulted in the ionospherebeing taken for granted, thought of more as alarge, distended mass of ionization, not some-thing with detail or structure to it like the val-ley of foF2 contours around sunrise or themountain of ionization around the equator (ac-tually the geomagnetic equator). That meant ashift away from understanding the details ofionospheric refraction, as shown by ray trac-ing, to a reliance more on symbolic methods,using mirror reflections.

By not using the variation of refractionwith frequency, the symbolic methods onlyshow how gross deviations in paths result fromtilts or changes in electron density of the iono-sphere, either from variations in layer heightor differences in critical frequency along a path.But the role that frequency plays is importantto the magnitude of simple effects like hoplength, and it separates other refractive effects,chordal hops and ducting in propagation overgreat distances. So without a good map andthe methods to make use of it, it’s impossibleto develop a good feeling for the lay of the landin ionospheric terms and the quantitative sidesof important features of propagation for greatdistances. That being the case, we turn to iono-spheric maps to see what they can tell us.


13: GOING FURTHER AND IN MORE DETAILWhen one travels, whether on foot, by

automobile or boat, even by radio, it pays tohave a map or chart to go by. As discussed ear-lier, radio propagation involves two types ofmaps, one for the E-region and the other forthe F-region, and they may be thought of asstacked, with one above the other, at about 115km and around 300 km altitude. But just whichregion controls propagation at a given timedepends on the frequency involved and wherethe region in question is encountered.

In some circumstances, say frequenciesabove 14 MHz, the E-region is of minor impor-tance, only deviating RF rays slightly on pen-etration and not screening them off from theF-region. Thus, one could consider only theeffects from ionization in the F-region and usea foF2 map to set the level of ionization at theF-layer peak; how the ionization is distributedwith altitude below the F-peak is another mat-ter and depends on the time of day.

That brings up another aspect of iono-spheric maps: half of the E-region map is es-sentially blank or empty, solar UV not reach-ing the dark side of the earth and E-region ion-ization on the day side being lost rapidly atsunset. That being the case, if one thinks of themaps as being stacked, HF propagation by thepart of the F-region above an empty or blankpart of the E-region map is pretty much thesame at low frequencies in the HF spectrum aswhen the higher part of the spectrum freelypenetrates the illuminated portion of the E-re-gion. Thus, while operations of the 3.5-4.0-MHz band typically involve short hops dur-ing the day, longer paths are possible acrossthe ionosphere at night. Beyond that, any fur-ther details would depend on how foF2 con-tours are organized and the distribution of ion-ization with altitude.

The ionospheric maps considered so farinvolved symmetrical illumination of the earthat the spring equinox. But the latitude of thesubsolar point shifts during a year, reaching

23.5° N and 23.5° S at the summer and wintersolstices, respectively. Since differences in so-lar illumination at the two solstices go accord-ing to which hemisphere contains the subsolarpoint, consider the summer solstice aroundJune 21; any results for that time would essen-tially apply also around December 21 in theSouthern Hemisphere.

Let’s start with the ionospheric map forthe E-region and take the time as 0000 UTC,when the sun is on the International Date Lineout in the Pacific Ocean. Now, instead of deal-ing with the ionosphere in abstract terms, theland masses below it will be included. So con-sider Figure 13.1 which shows the Mercatormap of foFE contours, centered over the WestCoast and with contours in 0.5-MHz steps,from 1.0 MHz to 3.5 MHz. Solar illuminationis centered on 23.5° north latitude and 180° lon-gitude and the 1.0-MHz contour falls close tothe Mercator projection of the terminator, theline dividing sunlit regions from those in dark-ness.

Now consider the foF2 contours for thesame circumstances, shown in Figure 13.2. Oncomparing the two maps, one sees the influ-ence of sunlight on the F-region, sharp andstrong around the sunrise portion of the ter-minator but declining beyond the line of dark-ness. By way of interpretation of the experi-mental data, it’s clear that all F-region ioniza-tion is not lost with sunset, as in the E-region.Instead, it lingers for quite a while, as shownby the lower critical frequencies foF2 in the re-gion of darkness.

Beyond the shape of the foF2 map and itsrelationship to the terminator, there are inter-esting features within the contours, the highvalues of foF2 on contours at low latitudes and,curiously, in the late afternoon region of localtime. Those prove to be geomagnetic in originbut for the moment, the higher foF2 values atlow latitudes go to explain the MUF curves forthe Boulder-Asunçion path in Figure 12.3. That


Figure 13.1 Global foFE map for 21 June (SSN=100).

figure was for the month of January but if onethinks of swapping hemispheres in Figure 13.2to get a foF2 map for December, the results arefairly obvious.

Next, looking at mid-latitude contours inthe summer and winter regions of the iono-sphere in Figure 13.2, one can understand theresults obtained earlier on the path from Boul-der to Madrid, shown in Figure 12.4. In par-ticular, one notes that foF2 values vary onlyslowly with longitude in the Northern Hemi-

sphere, explaining the weak variation in MUFon the path around at 40° N latitude and withtermini separated by about 100 degrees of lon-gitude. During winter months in the NorthernHemisphere, the situation would be similar tothat in the lower portion of Figure 13.2. Thus,the strong peak in the MUF on the Boulder/Madrid path finds its explanation the closelyspaced foF2 contours around the dusk merid-ian.

Now this section was touted as “Going

Figure 13.2 Global foF2 map for 21 June (SSN=100).


Further and in More Detail” but we’ve cometo the point where the going gets tough. Thatis the case as the paths discussed so far essen-tially all fell within the main parts of an foF2map, contained within the bounds of the ter-minator and something which is fairly easy tokeep clear in one’s mind. But going to pathswhich reach beyond 10,000 km distance, one-quarter the way around the world, and keep-ing appropriate ionospheric features straightin one’s memory becomes quite a challenge.Beyond that, operating frequencies and theirrelation to critical frequencies are not the onlyparts of the propagation equation; signalstrength and noise are other important quanti-ties too. And the statistical variations in criti-cal frequencies and in the propagation of noiseare important as well.

While knowledge of MUF values requirescalculations with a propagation program, ei-ther on a computer in the shack or from thecurves published monthly in QST, qualitativeaspects of signal strength can be obtained bylooking at the amount of illumination on a pathand how it varies in the course of a day. That isdone by looking at how the path lies relativeto the terminator. A quick and easy way of do-ing that is with the DX EDGE, a series of plas-tic slides with the shape of the average termi-nator in every month, which can be slid alonga Mercator map of the world to simulate theearth’s rotation in a day. More elegant, but in-volved ways of doing the same thing are withcomputer programs like GEOCLOCK,EARTHWATCH or DXAID. EARTH-WATCHuses a Mercator projection while GEOCLOCKand DXAID show the terminator position inthe azimuthal equidistant projection. The lat-ter fit quite well with path information in thesame format, as would be obtained by the LittlePistol using the map in Figure 11.1.

In simple cases, say the path is shortenough to fall within the sunlit part of the earthsometime in the day, signal strengths will bethe lowest when the path is fully illuminated,and will improve at times when one end or theother goes into darkness or the whole path isin the dark. But that approach only gives quali-tative information on signal strength and the

operating frequency must be taken into con-sideration, as suggested in the earlier discus-sion that went with Figure 8.1. And those timeswould have to be compared with MUF datafor the times when the path would be avail-able for operation. Finally, the matter of reli-ability would be settled by considering thepresence of noise, greatest during wakinghours from man-made sources or seasonal fromatmospheric origin.

A good example of how those matters canbe handled from both the qualitative and quan-titative standpoints is found by considering the3Y0PI DXpedition in February ’94. At the time,the smoothed sunspot number was 35 and thepath from the Little Pistol’s QTH to Peter I Is-land (68.8° S, 90.5° W) relative to the termina-tor is shown in Figures 13.3 and 13.4. In thoseprojections the sun’s position is shown about12.6 degrees south of the equator and the ter-mini are at the ends of the path. Local noon onthe path is about 1845 UTC, as indicated in Fig-ure 13.3.

From the earlier discussion, some tenta-tive conclusions may be drawn just from theterminator in Figure 13.3. For example, sunriseon the path would be around 1300 UTC andwith that, critical frequencies foF2 would show arapid increase along the sunrise portion of theterminator. The ionospheric absorption on thepath would peak around 1845 UTC and withsunset on the path around 0100 UTC, absorptionwould fall to low values while critical frequen-cies in the summer hemisphere would slowlydecay toward minimum values around dawn.

Qualitatively, propagation on the lowerbands, from 3.5 MHz to 10 MHz, would be con-trolled by ionospheric absorption while MUFvalues would be the determining factor from14 MHz to 28 MHz. Thus, low-band openingswould be limited to during darkness, after 0100UTC to before 1300 UTC on 3.5 MHz and some-what longer openings on 7 MHz and 10 MHz,because of somewhat lower absorption at thehigher frequencies. The higher bands wouldopen after sunrise on the path, about 1300 UTC,and continue past sunset, the duration depend-ing on the statistics of ionospheric critical fre-quencies for a sunspot number of 35.


Figure 13.4 Azimuthal equidistant map showing the path from Boulder, CO toPeter I Island.

Figure 13.3 Mercator map showing the path from Boulder, CO to Peter IIsland.

Of those qualitative considerations, thelow-band features would be fairly firm andcould serve as a guide in trying to make a con-tact with Peter I Island . But the high-band fea-tures suffer from quantitative uncertainty andcould only be firmed up with data on thespread in frequencies between the FOT andHPF for the path. Clearly, though, with a lowlevel of solar activity, the highest bands, 24

MHz and 28 MHz, would be more uncertainand spotty. So it would seem that operationswould be most likely limited to transitionbands, between 10 MHz where absorption is alarge factor and 21 MHz where high values offoF2 are necessary.

From the quantitative standpoint, the pre-dictions of Methods 9 and 22 from IONCAPcan be used to give the necessary details to fill


out what was outlined above. Method 9 calcu-lates the MUF and two other critical frequen-cies for the path and are shown in Figure 13.5.Method 22 takes those features over to providea measure of the mode availability and which

is expressed as a probability, in decimal form(0-1.00), that the operating frequency is belowthe maximum useable frequency at the hourin question. Similarly, Method 22 gives a mea-sure of the path reliability and it is expressed

Figure 13.5 Critical frequencies for the path from Boulder, CO to Peter I Island.

Figure 13.6 Mode availability and path reliability on 3.5 MHz, 7 MHz and 10MHz for the path from Boulder, CO to Peter I Island in February 1994 (SSN=35).


as a probability, in decimal form (0-1.00), thatthe signal/noise ratio exceeds an acceptableminimum value at the hour in question.

The predictions for the mode availabilityand path reliability for the low-band frequen-cies, 3.5 MHz to 10 MHz, are shown in Figure13.6. The upper part of that figure shows theavailability of the path on 10 MHz is reducedsomewhat in the hours before dawn when thecritical frequencies reach their lowest values.After sunrise, the availability goes back to 1.00.The path reliability, on the other hand, is highduring hours of darkness and then drops to lowvalues as absorption increases with the ap-proach on sunrise and remains low duringhours of daylight on the path. Since the oper-ating frequencies on the low bands are gener-ally below the critical frequencies throughoutthe entire day, the qualitative suggestions givenearlier were on a sound basis and confirmedby the more detailed calculations.

The same type of mode availability andpath reliability curves for the higher bands, 14MHz to 28 MHz, are shown in Figure 13.7.There it is seen that mode availability falls withsunset on the path around 0100 UTC, especially

for the higher bands, but then rises to highervalues with sunrise. However, with the sun-spot number of 35, the spread in critical fre-quencies, FOT to HPF, is not large enough tomake the higher bands equally available aftersunrise. And the path reliability curves showthe effects of ionospheric absorption being low-est, according to increasing frequency, duringthe hours around noon on the path.

With operations on the low bands ham-pered by absorption and limited on the highbands by critical frequencies of the ionosphere,it is of interest to examine the mid-bands, say14 MHz and 18 MHz, where the two effects areeach in transition, as it were. To that end, a newquantity, “DX feasibility,” is defined here, theproduct of the mode availability and path reli-ability. That is done in the interest of findingtimes of operation which offer the greatest com-bined promise of success. Since both the sepa-rate quantities range from zero to unity, thatwill be the same range for DX feasibility andthe results for the 14-MHz and 18-MHz bandsare shown in Figure 13.8.

While the DX feasibility factor for the 14-MHz band falls to below 0.2 around local noon

Figure 13.7 Mode availability and path reliability on 14 MHz, 21 MHz and 28MHz for the path from Boulder, CO to Peter I Island in February 1994 (SSN=35).


on the path, when the ionospheric absorptionis the greatest, the average feasibility (0.49) forthe day as a whole exceeds the correspondingvalue (0.39) for the 18-MHz band. Looking atFigure 13.8, one sees that the greatest chancesof making a contact on the 14-MHz band arein the morning when the band opens rapidlyand then in the afternoon hours as the bandslowly closes. But the 18-MHz band is largelyopen during daytime hours on the path.

To continue the discussion, other qualita-tive aspects of HF propagation may be inferredfrom the example in hand. Of particular inter-est are the consequences of the near N-S na-ture of the path from Boulder to Peter I Island. One effect is the rapid rise in MUF with sun-rise as the terminator sweeps across the path.The effect would be even more pronouncedwith a northern terminus in the Midwest, sayin Chicago, where the longitude is about 90°W, and at an equinox when the terminator isalso in the N-S direction.

In that regard, it should be noted that sun-

Figure 13.8 DX feasibility for the 14-MHz and 18-MHz bands for the path fromBoulder, CO to Peter I Island in February 1994 (SSN=35).

rise occurs earlier at F-region altitudes than atground level or in the D-region. As a result, theincrease in D-region absorption on a path thatgoes with sunrise would lag behind the rise infoF2 values. That point is of particular interestin connection with HF propagation coveringvery great distances, more than halfwayaround the earth. But that will have to waituntil some aspects of foF2 maps have been ex-amined in more detail.

In closing the discussion of the ideas be-hind mode availability and path reliability, itwould seem fair to say statistical features inamateur operations are encountered when aband seems closed even though the MUF waspredicted above the operating frequency or anunexpected band opening when the MUF waspredicted to be below the frequency in ques-tion. While one would think that statisticswould be evenhanded and balance out in thelong run, the disappointment in cases like thefirst are more than balanced out by the joys fromthe second. That’s what keeps DXers going.


14: PATHS BEYOND 10,000 KM

Looking at Figure 11.1, the azimuthalequidistant map centered on the Little Pistol’sQTH in Boulder, one sees that some of the rarerDX stations are beyond 10,000 km distance,those beyond the north coast of Africa and thecountries of Southeast Asia. That means DXpaths of at least four hops on short path and agood fraction of them across the northern po-lar region. The paths over the pole are uncer-tain because of their susceptibility to magneticdisturbance and paths toward Africa are diffi-cult because of the “aluminum curtain” alongthe East Coast. That brings up the other possi-bility, sneaking signals into Africa and South-east Asia by the back door, via long-path propa-gation. That last idea proves to be quite practi-cal so let’s turn to a discussion, from simple tocomplex, of long-distance propagation.

To begin, it was pointed out earlier thatit’s often convenient to show how HF propa-gation proceeds along a path by using a sym-bolic method, using reflections of rays fromcurved mirrors at ionospheric heights. Mirrors,of course, do not have any refractive proper-ties so the technique is limited to geometry anddoes not cover how propagation along pathsbehaves at different frequencies.

The simple case of a mirror reflection fora single hop can be extended to two or morehops. And as discussed earlier, mirror heightsmay be different from one hop to the next,making hop lengths different along a path. Ofcourse, critical frequencies may also vary alonga path so the MUF for each hop could be dif-ferent. That last idea was used to introduce theidea of a maximum useable frequency (MUF)for a path as a whole.

No mention has been made of mirrorsbeing tilted within an individual hop; thatchange in geometry would make the ascent anddescent portions of a hop unequal in length.Since mirrors are more symbolic than physi-cal, the question then turns to the physicalmeaning of ionospheric tilts, as they are called.

The answer, of course, is quite simple; the re-fraction or bending of a ray is different on oneleg of a hop from the other or, equivalently, thecritical frequency foF2 and electron densityprofile varies, positively or negatively, alongthe hop.

Consider the case of one hop in the flatearth representation, rays going from left toright. If the far right end of the ionosphericmirror is higher than the nearby left end, thedownleg for the one-hop path will be longerthan the upleg, etc., and the electron densityshould be less in the second half of the hop thanin the first half. But electron density in the iono-sphere translates into critical frequencies so thefoF2 value would have to decrease along thathop.

Considering that ionospheric hops cancover up to 3,500 km or so in horizontal range,increases or decreases of foF2 along a path arenot unreasonable. Indeed, all the ionosphericmaps shown so far have variations like that inmany regions. And the maps also have regionswhere changes in foF2 are rather small. Forexample, Figure 4.3 shows large increases anddecreases in foF2 for paths in the N-S direc-tion, especially in the early evening hours oflocal time (around 210° E longitude). And thereare slow variations in the E-W direction at po-lar latitudes and increases or decreases, de-pending on the direction of a ray path, for pathsin N-S and E-W directions. In short, critical fre-quencies and ionization vary widely over theentire ionosphere. Of course, HF propagationprograms carry that sort of information in theirdatabases and make use of it in making MUFpredictions. But the user does not really comeinto close contact with that reality and its con-sequences, especially if thinking only in termsof simple hops from mirrors which are paral-lel to the earth’s surface. That being the case,it’s worthwhile to illustrate what’s buried deepin propagation programs and note some of theinteresting features that come to the fore, first


qualitatively by using mirror reflections andthen quantitatively by going to ray tracing.

For discussion, consider a path from LosAngeles (34.0° N, 118.3° W) to Kiev (50.5° N,30.5° E) at 0600 UTC in June, during a periodof high solar activity (SSN= 150). The pathlength is about 10,080 km and would involvethree hops. One can break the path into anynumber of small segments but for discussionpurposes, consider breaking each hop into 8segments. Using the F-layer algorithm from theMAXIMUF program, one can calculate foF2values every 420 km along the path.

By taking differences in foF2 between ad-jacent segments and dividing the result by thedistance, one obtains a rough measure of howthe ionosphere varies along the path. Thus, iffoF2 decreases between adjacent steps along apath, the ratio is negative; that corresponds toan upward tilt of the ionospheric mirror alongthe path or less downward refraction of the farend of the hop than when it’s parallel to theearth’s surface. In a similar way, positive ra-tios mean the mirror is tilted downward andmore downward refraction would result.

Ultimately in this sort of discussion, thematter will come down to ray tracing, so thediscussion should be put in appropriate termsat the outset. Now differences in critical fre-quency foF2 are really related to differences inelectron density at the F-layer peak and when

that conversion is made, the ratio can be ex-pressed in more physical terms. In order to dothat one needs the fundamental equation whichrelated peak electron density and foF2 values.From that it is seen that differences in electrondensity are proportional to differences in thesquares of foF2 values.

So the ratio which shows whether themirror is tilted one way or another has thephysical dimensions of MHz2 /km. That num-ber gives a measure of what is termed the elec-tron density gradient and looking at the differ-ences in foF2 values across an ionosphere, onefinds electron density gradients from 0.005 or5E-3 (MHz•MHz)/km up to about 0.1 or 100E-3 (MHz•MHz)/km.

With that discussion concluded, let’s goback to the LA/Kiev path and look at how foF2values and the electron density gradient var-ied step by step along the path, as in Figure14.1. There it is seen that foF2 went through aminimum around the midpoint of the path,from about 7.5 MHz at the start to 5 MHz atthe middle and back up again, while the elec-tron density gradient began weakly negative,with a numerical value around 1E-3(MHz•MHz)/km and slowly increased toabout the same positive value at the end of thepath. All in all, there was little variation alongthe path and that is quite reasonable consider-ing the path was mainly mid- and high lati-

Figure 14.1 foF2 values and gradients in 420-km steps along a path from LosAngeles to Kiev, Ukraine.


14.2.Where the gradient is positive, one can

expect greater downward refraction or shorterhops and just the opposite when the gradientis negative. From Figures 14.1 and 14.2, it’s clearthat ionospheric tilts differ greatly along thetwo paths, being much more significant on thepath crossing the geomagnetic equator. In thatconnection, ionospheric tilts in the equatorialregion have been suggested as the origin ofhigh MUFs and strong signals on transequa-torial paths.

To continue with the discussion, Figure14.3 gives symbolic representation of a chordalhop, one which involves two ionospheric re-fractions without an intervening ground reflec-tion. On that path, the first ionospheric tilt (up-ward) gives less refraction on the downleg thanon the upleg, resulting in the ray missing theearth and going on to another region with anionospheric tilt (downward) which refracts theray back to earth. Without a surface reflectionat the midpoint of the path, signal strengthwould be higher by 1-6 dB and if the chordalsegment of the path went above the D-region,an additional gain in signal strength wouldresult.

Going now to ray tracing, the variation offoF2 along a path, as in Figure 14.2, can be con-verted to a variation of peak electron density.

Figure 14.2 foF2 values and gradients in 375-km steps on a path from LosAngeles to Sydney, Australia.

tudes in a summer month.Now consider a path going in the oppo-

site direction at the same date and time, fromLos Angeles to Sydney, Australia (33.9° S, 151.3°E). That path is about 12,000 km in length andwould involve 4 hops. By breaking the hopsinto 8 segments, each 375 km long, and work-ing out the foF2 values and the electron den-sity gradient along the path, the results shownin Figure 14.2 are obtained. In contrast to theLA/Kiev path across higher latitudes, the pathfrom LA to Sydney goes across low latitudes,the equator and into low latitudes in the South-ern Hemisphere.

As noted earlier, much higher critical fre-quencies are encountered in those regions,reaching values almost twice that found alongthe other path. And the electron density gradi-ent shows much greater variations too, bothpositive and negative and reaching values 7-10 times larger. And the high, double peak infoF2 values along the path as it crosses low lati-tudes is termed the “equatorial anomaly.” Itoriginates around the geomagnetic equator andis the result of a redistribution of photoioniza-tion by the action of an electric field of atmo-spheric origin. That electric field moves equa-torial ionization across magnetic field lines anda fountain effect lifts and carries it to the northand south, giving foF2 peaks as seen in Figure


That, in turn, can be put into mathematicalform and used in the ray-tracing program towork out details of a ray path across the equa-torial anomaly. Before giving those results,however, a few words on ray tracing are in or-der.

First, the technique follows the advanceof a ray, the height and orientation of its seg-ments being plotted, step by step, along thepath. The change in ray direction in going from

changing foF2 values and heights of F-peaksin a multihop path to reflect the fact that solarillumination varies with distance along thepath. Those changes can be made in the ray-tracing program at each surface reflection andused within the next hop but the electron den-sity profile does not change shape within agiven hop. The changes in ray paths from vary-ing illumination are generally small unless theoperating frequency is such that the E-region

Figure 14.3 Symbolic representation of a chordal hop between two tiltedregions in the F-layer.

one step to the next depends on the frequencyand the rate at which the electron densitychanges with position. As ray segments are lo-cated, their orientation determines whether theray is ascending or descending.

In its simplest form, ray tracing assumesthat the electron density in a hop varies onlywith height above ground but not along thehop. The density rises from essentially zero atground level to a small plateau or ledge at theE-region, around 110 km altitude. Above thatlevel, the electron density rises and goesthrough another ledge in the F1-region at about175 km during daytime and reaches the F-peakat about 300-400 km altitude. Numerical val-ues for electron density at the critical heightsare obtained from data based on ionosonderecords and the shape of the electron densitycurve obtained from various model profiles.For ray-tracing calculations, the model profilesmust join smoothly, in magnitude and slope,at every transition.

An improvement in ray tracing is whencritical frequency data is changed from one hopto the next along a path. That usually means

is called into play, making the next part of thepath an E-hop instead of another F-hop.

Ray tracing in that approximation usesone-dimensional methods within a given hopand discontinuous changes in critical fre-quency data in going from one hop to another.The calculations are done using Snell’s Law forwave refraction but in a special form for spheri-cal geometry in which the electron density var-ies only with height. The next step in ray trac-ing is to move up to a two-dimensional ap-proach with continuous variations in criticalfrequency data, both with height and along thepath.

Now leaving the mirror approach, asshown in Figure 14.3, consider a ray trace for apath across an equatorial distribution in ion-ization where the peak electron density variesalong the path in a manner like that suggestedby Figure 14.2. Thus, the highest foF2 value wastaken as 15 MHz and a value of 10.6 MHz usedat the minimum. The critical frequency data forthe F1- and E-regions were taken as 2 MHz and1 MHz, respectively, for late-afternoon or early-evening conditions at equatorial latitudes.


Using a rectangular format to display theresults of calculations in spherical geometry,Figure 14.4 shows the ray trace for a 28-MHzsignal going through that ionosphere and re-sults in a chordal hop that covers almost 7,000km distance. At 28 MHz, the signal goesthrough the E- and F1-regions with practicallyno deviation in direction and the major changesin direction in ray direction are seen to takeplace around 250 km altitude. On descent fromthe first peak, the ray is in a region of decreas-ing electron density and suffers less downwardrefraction than on ascent. As a result, the rayleaves the lower part of the E-region, misseshitting the earth, as in Figure 14.3, and thengoes on for another F-region refraction wherethe increasing electron density refracts it backto ground level.

Earlier, ray traces were shown in whichrays with increasing frequency were sent up-ward into the ionosphere. The result that comesfrom such displays is that higher frequencyrays go deeper into the ionosphere, probing asfar as the region just below the F-layer peak,while lower frequency rays get as far as thelower part of the F-region or are stopped by

change of electron density with height. Thus,28-MHz rays are refracted back to earth moreslowly than 7-MHz rays and 7-MHz rays arerefracted back to earth at lower altitudes than28-MHz rays.

And one other simple result is that raysin the ionosphere below the F-peak are alwaysrefracted back toward the ground as long asthe region is one where the electron density isincreasing with altitude. Where the electrondensity might be constant over a short distance,say at the E-region or F1-region ledges, rayssimply continue onward undeflected, whetherascending or descending, until the electrondensity increases again.

Those ideas are part of the basis of thechordal hop in Figure 14.4, less refraction tak-ing place on the downside of the first part ofthe path because of a smaller variation of elec-tron density with height. The other part of thematter has to do with the earth’s curvature,making it fall away from under the ray as itadvances. Actually, the direction of every stepof an advancing ray results from a competitionbetween the two factors, ionospheric refractionback toward earth and apparent refraction in

Figure 14.4 Ray trace for a chordal hop on 28 MHz.the E-region and refracted downward. In termsof ray-tracing calculations, those results aremanifestations of the general features of iono-spheric refraction, the refraction per step alonga path varying as the inverse square of the fre-quency as well as directly with the rate of

the opposite direction because of the earth’scurvature.

The competition between those two pro-cesses in the same ionosphere will be faster atlower frequencies as the rays are refracted morestrongly. As a result, lowering the frequency


from 28 MHz to 14 MHz replaces the slow,large-scale refraction which gave a chordal hopby fast, small-scale chordal ducting, as in Fig-ure 14.5. There, because of the lower frequency,the ray barely penetrates past the F1-ledge andfor the same reason, it is refracted more rap-idly and oscillates up and down in a heightrange of about 30 km while advancing alongalmost parallel to the earth.

The robust ionosphere in Figure 14.2 wasfor late-afternoon and early-evening conditionsacross the geomagnetic equator. At later times,more toward the dawn hours, critical frequen-cies along the same path are lower and alsoadmit long-distance propagation by chordalducting at lower operating frequencies. Thus,for 7 MHz and lower in frequency, ray tracessimilar to Figure 14.5 will result, the difference

being that with stronger refraction at a lowerfrequency, a somewhat greater number of os-cillations take place along the path. Obviously,those circumstances go to explain, at least inpart, long-distance propagation on the lowerbands.

The discussion above deals with the ef-fects of longitudinal gradients in the iono-sphere, the circumstances when a ray pathcrosses foF2 contours perpendicular to theirorientation. It takes no stretch of the imagina-tion to think that ray paths could also go al-most parallel to foF2 contours. The most obvi-ous case would be a path parallel to the termi-nator around sunrise and the electron densitygradient would be transverse to the ray path.

Earlier, the effects of longitudinal gradi-ents were illustrated using ionospheric mirrors,

Figure 14.5 Ray trace of chordal ducting on 14 MHz.

Figure 14.6 Schematic drawing for a non-great circle hop due to a transversefoF2 gradient.


Figure 14.7 Projection of ray traces for non-great circle (NGC) hops on avertical plane.

tilted one way or the other along a ray path.The same can be done for transverse gradients,the ionospheric mirror now being tilted acrossinstead of along the path. As before, the higherside of the ionospheric mirror would be wherethe electron density is lower, toward darknessaround the sunrise terminator.

By tilting the ionospheric mirror in thatfashion, rays reflected off the mirror would bedeviated toward the high side of the mirror, inthe direction of decreasing electron density andfoF2 values. As a result, where ray paths arenormally expected to lie in the plane of a greatcircle path, they now are deviated out of the

plane, resulting in non-great circle (NGC)propagation. That is illustrated in Figure 14.6where a NGC hop and its projections in boththe vertical and horizontal planes are given.

As might be expected, ray tracing can beapplied to this type of situation but the prob-lem is more complicated in that another dimen-sion has to be added to the calculations, allow-ing now both longitudinal and transverse gra-dients along a path. To illustrate the quantita-tive side of the problem, traces are shown inFigures 14.7 and 14.8 for a sunrise situationwhere 7-MHz rays start along the terminator.

As seen in the first figure, under those cir-

Figure 14.8 The projections on the earth’s surface of sideways deviations ofnon-great circle (NGC) paths due to a transverse foF2 gradient.


cumstances the rays just penetrate the E-regionbefore being refracted downward. The effect ofthe transverse gradient, operating above the E-region, is to deviate the ray path away from theregion of higher ionization (in sunlight) by amodest amount. For the ray going out 2,000 km,the side deviation is about 25 km or an angulardeviation in path direction of 1.5 degrees.

In the dawn situation, the sideward devia-tion remains about the same as long as the raypaths stay below the F1 ledge. At a higher fre-quency, say 14 MHz, ray paths may penetratethe F1-ledge and undergo deviations abouttwice those shown in Figure 14.8 because ofgreater exposure to the transverse gradient. Atan even higher frequency, say 21 MHz, therewould be greater penetration and exposure butthe net deviation would remain small becausethe sideward refraction also goes as the inversesquare of the frequency.

In any event, non-great circle (NGC)propagation results from transverse gradientsin the ionosphere, as shown above for a singlehop, and the effects will be cumulative alonga multihop path. With the frequency depen-dence of refraction, the largest effects wouldbe at low frequencies and at times of darknesswhen ray paths can go past E-region heightsand into the F-region. While various types andmagnitudes of gradients can be found by in-specting an ionospheric map, it should beborne in mind that the ionosphere is a dynamicsystem, more than just the average environ-ment displayed in foF2 maps. Thus, whilepropagation effects from average gradients areto be expected, day in and day out, unusualgradients are not out of the question, especiallyin disturbed times, and propagation beyondthe normal range may be encountered at thosetimes.


15: LONG-PATH PROPAGATIONzon, close to the terminator. Under those cir-cumstances, long-path DXing has a possibilityof being successful if the dawn/dusk iono-sphere still supports propagation. The otherlosses and noise may still be factors, say groundinstead of ocean reflections and noise propa-gated from thunderstorm activity.

Now to a first approximation, paths frompoint A to point B are great circles so the thingto do for long-path DXing is find the long-pathdirection to the DX location. That’s simple; justadd 180 degrees to the beam heading for ashort-path connection. If the Little Pistol inBoulder wanted to make a contact on long path,the thing to do would be to look at a map likethat in Figure 13.4. That figure was obtainedby using the DXAID program and not onlyshows the azimuthal equidistant map centeredon Boulder but also an oval to represent theshape of the terminator at the chosen time. Inthe context of the Peter I Island operation, thefigure shows the sun is north of Australia andthe path to 3Y0 is partially in darkness around0100 UTC.

Now consider a different path, to SaudiArabia. For a short-path connection to SaudiArabia, the Little Pistol’s tribander should beheaded in a direction 29 degrees east of northbut for long path, it would be changed to 209degrees. That would still be along a straightline on the azimuthal equidistant map, goingout at 209 degrees bearing to the antipodalpoint (now represented by a circle at 20,000 kmfrom Boulder) and then continuing back in to-ward the center of the map at the 29 degreeheading until it reaches HZ-land.

At that heading, the long-path distance toHZ-land is 27,880 km, a factor of 2.3 greaterthan the short-path distance, 12,150 km. Be-cause of signal spreading, the mere thought ofgoing long-path means the Little Pistol wouldtake a loss of at least 7.2 dB. But it would beworth the try for a contact as it’s a way to getsignals into HZ-land via the back door, as it

The varied features of the ionosphere, asrepresented by foF2 maps, become more im-portant the longer the path. Thus, for the short-est, single-hop paths, all that’s needed is thecritical frequency nearby but with multihoppaths, the spatial gradients of critical frequen-cies become important as well. In a sense, it’slike rolling a ball down a hill. At the start, theball’s path depends on the slope of the hill butits ultimate destination will be determined byall the hills and valleys beyond the startingpoint.

Turning to long-path propagation, thatbegins when RF goes beyond the antipodalpoint of the transmitter, signals reaching morethan halfway around the world. And long-pathDXing has been a popular mode in spite of thedistances involved and the potential losses insignal from adding more distance, surface re-flections and D-region absorption in extend-ing a path. At times of high solar activity, it is aregular mode of propagation, open every dayexcept at times of magnetic disturbance. Thatis an important point and will be discussed atsome length at the end of this chapter and inthe chapters which follow.

But getting down to long-path propaga-tion, the length of the path is the least concern,the inverse square law only lowering signalstrength by 6-10 dB if a path length weredoubled or tripled in going from short to longpath. More ground losses would be significantfor additional reflections off of dirt but thethreat may not be all that great as consideringthe earth as a whole, about 80% of its surface iscovered by oceans, the next best reflecting sur-face after a conducting metal foil.

That leaves D-region losses as a threat andin the last analysis, it will prove to be the con-trolling consideration in long-path propaga-tion. So knowing that ionospheric absorptionwould be a factor, DXers try to limit its effectby sending signals off into an ionosphere indarkness or where the sun is low to the hori-


Figure 15.1 A Mercator map showing the path from Boulder, CO to Riyadh,Saudi Arabia and the terminator at 1400 UTC on 21 December.

were, and not in direct competition with thehordes of operators on the East Coast and inEurope. And looking at the map in Figure 13.4,it’s clear that most of the path would be oversea water. That minimizes the effects of sur-face reflections and leaves D-region absorptionand the effects of darkness on critical frequen-cies to worry about. So back to maps and theterminator.

The azimuthal equidistant projection isfine for showing great circle paths from one’sQTH but all other great circles not centered atthe same location come out as ovals, the termi-nator in Figure 13.4 being a case in point. Andon a given day, the ovals rotate about the QTHat the center of the map. When it comes to judg-ing how a path is illuminated, that map pro-jection tends to be confusing so consider theproblem again, now using the Mercator mapin Figure 15.1.

That figure shows the path from Boulderto Riyadh as a heavy, sinusoidal-like trace.Unlike an RF path which is a great circle andfixed for all time, the terminator is a great circlebut its shape and projection on a Mercator mapchanges with season, as seen earlier in the dis-cussion of the foFE contours. For the presentdiscussion, long path to HZ-land, the date andtime have been chosen to illustrate how D-re-gion absorption may be minimized. Thus, thedate was taken as 21 December, the winter sol-

stice, and the time as 1400 UTC when the sub-solar point is off the east coast of Brazil at 23.5°S, 330° E.

Looking at Figure 15.1, it’s clear the por-tion of the great circle for the path and the ter-minator are almost in spatial coincidence, withhalf of the path on the dawn side of the termi-nator and the other half along the dusk side,making for a situation with low D-region ab-sorption. That situation gives a gray line pathand is an important part of long-path DXing.But obviously, with the shape of the termina-tor changing markedly with seasons and thegreat-circles for paths being fixed, long-pathDXing will have strong seasonal aspects.

The circumstances in Figure 15.1 showother possibilities for long-path DXing fromBoulder in the winter months, into Europe andAfrica. For those directions, Figure 13.4 showsthat the Little Pistol’s beam headings wouldbe similar to that for Saudi Arabia, to the westof south. And Figure 15.1 shows that with theadvance of time past 1400 UTC, the sun willrise on the first part of the long path off to thewest and set on the other part to the east.

The long-path possibilities for other direc-tions during winter, say toward India and SriLanka, would be less propitious, signals goingoff to the east of south and into the sunlit hemi-sphere. Beyond that, paths to India and SriLanka would go to high latitudes, both geo-


Figure 15.2 The DX Edge setting for 1530 UTC during December.

graphically and geomagnetically. The latterproves particularly important when it comesto the reliability of making long-path contactsto that part of the world as the paths from Boul-der which go to high magnetic latitudes aremore susceptible to disturbance, as will be seenshortly.

There is another time, in June at the sum-mer solstice, when long-path openings are pos-sible from Boulder to HZ-land by gray-linepropagation. The path and terminator wouldbe exactly the same as shown in Figure 15.1,the difference now being that the time is 0200UTC and the subsolar point is southeast of Ja-pan at 23.5° N, 150° E. The gray-line situationwill be the same but since the time is changedby 12 hours, the path at summer solstice is duskto dawn from Boulder whereas the path dis-cussed earlier, at winter solstice, was dawn todusk.

Physically, that means the winter anomalyof the F-region has exchanged hemispheres, theearly morning long path from Boulder in De-cember going into summer in the SouthernHemisphere, with lower foF2 values along thepath, while the early evening long path in Junegoes into winter in the Southern Hemisphere,with higher foF2 values. But perhaps the larg-est difference in the two situations is in con-nection with sociological factors, there beingmore QRM from local operators on the bands

in the USA than at the less populous DX end.That being the case, early morning hours provemore productive than evening hours for long-path DXing.

In an effort to show more of the details oflong-path propagation, consider another set ofpaths, from here in the northwest corner of theUSA. First, take the case of the winter solsticeagain and the time as 1430 UTC, now shownby the DX Edge in Figure 15.2. There, paths toEurope go off into the dark hemisphere andare not open while the first hops from theNorthwest are still in darkness. That is the caseas they’re in the winter hemisphere and foF2values for sites along the path fall to low val-ues when in darkness for 14 hours or so. In thatregard, the opening of a long path to HB-landresults from the rise of foF2 values at sunrise,after 1530 UTC, on points along the great circle,as shown in Figure 15.3.

For that figure, the path was divided into36 sections of 450 km length and foF2 valueswere calculated at each step along the path forevery half hour, starting at 1430 UTC. In orderto be successful in making a long-path contactwith HB-land on 14 MHz, foF2 values in ex-cess of 5 MHz are required along the path. FromFigure 15.3, it is seen that foF2 values weremore than enough along the major portion ofthe path, even displaying the high values inthe equatorial anomaly.


Figure 15.3 Sunrise variation of foF2 values along a great circle path fromWashington to Switzerland in December.

But on the first hop from here in theNorthwest, foF2 values are too low and won’tsupport propagation along the entire path un-til the sun starts to rise on the first hop, around1500 UTC and out about 1,400 km west of LosAngeles. After the sun rises at F-region heights,quickly raising the foF2 values and opening thepath, it rises somewhat later on the D-regionand ionospheric absorption sets in and startsclosing the long-path circuit to HB-land.

The long path from the Northwest to In-dia and Sri Lanka is closed in the wintermonths, just as for Boulder, because it goes offinto the sunlit Southern Hemisphere. However,

it is open in the morning hours of the summermonths, as may be seen from Figure 15.4. Thatfigure is for the summer solstice and 1230 UTC.In that case, the terminator which cuts acrossthe Northwest also goes across the tip of SouthAmerica and then the southern waters of theIndian Ocean, finally coming close to India andSri Lanka.

The paths to those locations are close tothe terminator and propagation to India andSri Lanka opens and closes for the same rea-sons as before, sunrise bringing foF2 values upto where propagation is supported along theentire path and then closes somewhat later

Figure 15.4 The DX Edge setting for 1230 UTC in June.


when ionospheric absorption sets in as the sunrises on the D-region. Those processes still takeplace in the winter hemisphere, where foF2values have fallen during the long night. Thecase for the path to India is shown in Figure15.5 where the rise in foF2 values takes placebetween step 9 (13° S, 103° W) and step 19 (76°S, 37° W) after 1200 UTC.

From the above discussion, however, itwould be a mistake to think that long-pathDXing is only possible along the gray linewhere solar illumination is weak. For any typeof successful propagation, three conditionsmust be met: a MUF value sufficient to sup-port the frequency in use, good signal strengthat the receiving station and a low level of noise.During times of high solar activity, say aroundsolar maximum, critical frequencies do not fallto low levels in the long hours of darkness.

In that regard, experience here in theNorthwest during the peak phase of Cycle 22showed that long-path contacts with Africawere frequently possible in the morning hoursduring the northern summer, as in Figure 15.4,by sending signals on long-path directions al-most at right angles to the terminator. Withsolar activity at that level, foF2 values in thedark hemisphere were still sufficient to sup-port the propagation. But during the winter sol-stice when the gray line was more favorablypositioned, as in Figure 15.2, it proved very

difficult to contact South Africa because of allthe atmospheric noise propagated from tropi-cal thunderstorms during the southern sum-mer season.

So far, the discussion has dealt with long-path propagation opening due to the growthin foF2 values with sunrise on the F-region andits closing due to the increase in absorption asthe sun rises later on the D-region at lower al-titudes. Beyond those ideas, it should bepointed out that some of the extreme long-pathcircuits profit from chordal hops across theequatorial anomaly, the resulting signals beinggreater for the lack of losses that would haveoccurred from an intervening surface reflection.That would be true of paths with a significantlongitudinal ionospheric gradient; other longpaths lacking that benefit would be weaker insignal strength, limited by the losses on typi-cal earth-ionosphere hops.

At the higher part of the HF spectrum, sayfrom 14 MHz and above, chordal hops may besimple, as in Figure 14.4, or show some duct-ing, as in Figure 14.5. At lower frequencies, sig-nals do not penetrate as deeply into the F-re-gion and chordal ducting would dominatewhen it comes to long hops on DX paths. Thatwould be particularly true for frequencies of10 MHz or lower as hops would be shorter andmore numerous, quickly dissipating RF energywere it not for long, efficient ducted hops.

Figure 15.5 Sunrise variation of foF2 values along a great circle path fromWashington to India and Sri Lanka in June.


Figure 15.6 Parallels of magnetic latitude for a centered dipole.

Finally, as remarked earlier, great circlepaths are fixed with respect to the earth butthe terminator varies in position and shape,moving along with the hours of the day fromEast to West on a Mercator map or changingshape with the seasons. In connection withlong-path propagation, the most polewardreach of a path is important for reasons otherthan simply their relationship to the termina-tor. In particular, it is important to know whichregion of geomagnetic field was involved in thepoleward excursion of a path as propagationdisturbances of magnetic origin vary in degreeaccording to the highest latitude on a path.

In that regard, regions on the earth areclassified according to a system of geomagneticcoordinates based on the dipole model dis-cussed earlier in connection with geomagneticdata. That model resulted from a large-scaleanalysis of magnetic observations and has anaxis of symmetry which passes through thecenter of the earth, tilted about 11.5 degreeswith the geographic axis. With an equatorialplane perpendicular to the magnetic axis andpassing through the center of the earth, thegeomagnetic coordinate system has parallelsof magnetic latitude as shown in Figure 15.6.

The auroral zones, where luminous dis-plays are seen most often, generally fall be-tween 60 and 70 degrees magnetic latitudewhile the polar plateau lies above 70 degreesand subauroral latitudes are just below 60 de-

grees magnetic latitude. With that system, thepaths from here in the Northwest to southernAfrica would be classified as auroral as theirmost southerly excursions in latitude rangebetween 61 and 69 degrees south geomagneticlatitude for termini from Pt. Elizabeth toSwaziland. Only Cape Town, at the most south-ern tip of South Africa, is in the subauroral cat-egory. After that, paths to Zambia and SaudiArabia fall in the polar category as their mostsoutherly excursions in latitude range between71 and 85 degrees south geomagnetic latitude.

The category of the path to a particularDX site depends on the location of the trans-mitter. Thus, from here in the Northwest, thepath to India goes through the auroral zonewhile from Boulder, it goes across the polarplateau. Similarly, from here in the Northwest,the path to Spain crosses the polar plateau butfrom Boulder, it passes through the auroralzone. In terms of geographic distinctions, thosepath differences would be related to the solarillumination and all that follows from it, reflec-tion surfaces, D-region absorption and foF2values.

But more significant in categorizing pathsare differences in the geomagnetic field linesthey go across. With the ionosphere under geo-magnetic control, low latitude field lines go outa few earth radii but the high latitude iono-sphere has its electrons gyrating around the feetof magnetic field lines which go out to great


distances, many earth radii, above the lowerionosphere itself. Since geomagnetic distur-bances are strongest at high latitudes, they areaccompanied by ionospheric disturbances ofsignals propagating across those regions.

That last statement would seem to sug-gest that a discussion of propagation distur-bances would be in order now, at least forpaths going across high latitudes. That is cer-tainly the case but the scope of the discussion

proves to be much greater than would seemapparent from all that has been developed sofar. The matter goes to a larger subject, the en-vironment surrounding the earth, and it beginswith a major revision in the model for the geo-magnetic field. That result came from explora-tion in the Space Age and we turn now to thatin the next section, continuing the discussionof HF propagation.


PART TWO


begin the discussion where we left off, withlong-path propagation, and look at how a newunderstanding of the geomagnetic field, par-ticularly at high magnetic latitudes, emergedwith space exploration.

The earlier view of the geomagnetism wasthat the earth was surrounded by a field likethat of a dipole, as in Figure 6.1, and orbitedthe sun in a region containing little more thana handful of planets, meteors and interplan-etary dust. That view has changed drasticallyand one of the first new ideas that impactedon propagation was the modern understand-ing of the shape and dynamics of the earth’smagnetic field. It is best explained by examin-ing the current view of the geomagnetic field,shown in Figure 16.1.

That figure shows the magnetic regimesurrounding the earth, the magnetosphere,with the earth’s field compressed by the solarwind on the dayside and drawn out on thenightside. In essence, the earth’s field is con-fined to a cavity carved out of the stream ofsolar plasma flowing by, something like thewake behind an object fixed in a stream. In the

16: TOWARD A NEW ERA

The discussion up to this point has dealtlargely with the effects of solar UV radiationincident on the atmosphere, the chemicalchanges resulting from photodissociation andthe creation of the ionosphere by photoioniza-tion. That was the focus of the early work onpropagation, an era before WWII when photo-chemistry dominated the scene. True, distur-bances of propagation did occur, with short-wave fade-outs and ionospheric storms. Thefade-outs (SWF) were associated with solarflares, and ionospheric storms were coincidentwith magnetic storms but there was no unify-ing understanding of the origin of the effects.

That had to wait until the end of WWIIwhen the Space Age got under way. Then, high-altitude balloons, rockets and satellites beganprobing the upper reaches of the atmosphereand ionosphere, regions where only radiowaves had been before. In less than two de-cades, a picture emerged which brought to-gether the diverse aspects of radio propagation,the formation of the ionosphere and its distur-bance.

The discussion which follows will dealwith the new ideas that appeared on the sceneas the “photochemical era” came to an end andwas replaced by the “plasma and fields era”that we’re in now. In considering the new un-derstanding of HF propagation, it should benoted that we’re moving away from a quasisteady state scene where the basic ionosphericproperties change on a time scale characteris-tic of solar cycles. Now the discussion shifts towhere changes occur on a more rapid scale, thetime it takes UV and bursts of solar X-rays toreach the earth or the time for puffs of ionizedmatter to arrive at the earth’s orbit.

The photochemical era really dealt withthe physics and electromagnetic theory of thebottom side of the ionosphere. The new erastarted at the F-region peak in ionization andwent outward through the topside ionosphere,finally reaching the sun. In that regard, we can

Figure 16.1 The magnetosphere formed bythe interaction of the solar wind and thegeomagnetic field. From Davies [1990]


well as drift slowly in longitude because thegeomagnetic field weakens with increasing lati-tude. And when the source of F-region elec-trons is turned off at sunset on the atmospherebelow, the electron content in the reservoirstarts to decrease as electrons recombine withpositive ions at one end or the other of theirbounce motion or are scattered out of trappedorbits and spiral down to the atmosphere be-low and become lost.

That description shows how the F-regionis maintained at night, unlike the E-regionwhich decays rapidly, and it holds as long asfield lines are closed, going from one hemi-sphere to the other. Looking at Figure 16.1, itwould not apply to field lines in the polar capwhich go back into the tail of the magneto-sphere. F-region electrons may be released onthe feet of those field lines but their popula-tion would not build up nor be maintained bydurable trapping as the polar field lines are notclosed in the usual sense.

The distinctions between closed and openfield lines are not sharp nor steady in time as itis found that the location of the outer bound-ary of the plasmasphere varies with geomag-netic activity. Indeed, it is on that basis that theassociation of ionospheric storms with geomag-netic storms is understood, high-latitude fieldlines which were closed on the front of the mag-netosphere suddenly being swept back into themagnetotail with the onset of magnetic activ-ity, taking their supply of electrons with them.When that happens, the field lines at a high-latitude site are no longer closed in the usualsense and unable to retain electrons spirallingup from ionospheric heights. As a result, F-re-gion critical frequencies drop drastically andpropagation is no longer supported at the samelevel as before. In addition, there is an effecton the electron density profile, not only lower-ing the peak value but also raising the heightof the F2-maximum somewhat, as shown inFigure 16.2.

In order to complete the discussion, dis-tance scales and time scales must be added. Asfor distances, the ionospheric electron popula-tion held on magnetic field lines extends outto about 4-5 earth-radii (Re). Beyond that, out

figure, the magnetic axis is tilted forward andthe geometry of field lines vary, depending ontheir magnetic latitude of origin

In that configuration, there is a low-lati-tude toroidal arrangement of dipole-like fieldlines, then an asymmetrical region of field lineswhich still go from one hemisphere to the otherand finally field lines from polar regionswhich extend away from the direction of thesolar wind, back into the magnetotail. Thosedistinctions emerged as the geomagnetic fieldwas examined by magnetometers on space-craft, the compressed field extending out 10-12 earth-radii on the sunward side and themagnetotail reaching back many tens of earth-radii in the opposite direction.

It’s on those various types of field linesthat ionospheric electrons find themselveswhen released by photoionization. While theirorigin is different than the electrons in theearth’s radiation belt, the physics of their dy-namical motions is the same, resulting in elec-trons gyrating around magnetic field lines. Butif the ionospheric electrons have a componentof motion along a field line, they may spiraleither down or up the field line. With down-ward spiralling motion, electrons go to loweraltitudes and may recombine with positive ionsin the time they spend there or, failing that, theymay be reflected upward by the convergingmagnetic field lines, just like electrons trappedin the radiation belt.

On the other hand, with upward spiral-ling motion, electrons move into higher alti-tudes where the ionosphere thins out evenmore and may go out several earth-radii towhere field lines cross the magnetic equator.After that, the electrons will continue spiral-ling around field lines and go down towardthe ionosphere below, in the opposite hemi-sphere. In short, electrons starting with photo-ionization in one hemisphere may carry outbounce motions between hemispheres and be-come trapped in the earth’s field.

As a result, the geomagnetic field acts asa reservoir, termed the plasmasphere, holdingelectrons released initially in the sunlit atmo-sphere at the feet of field lines. The spirallingelectrons will bounce between hemispheres as


Figure 16.3 Distribution of magnetic storms(Ap>40) by month.

years by NOAA. Thus, in the previous 5 solarcycles, there were 1,033 major magnetic stormswith Ap of 40 or higher. That works out to about206 storms per solar cycle, 19 storms per yearand corresponds to significant ionospheric dis-turbances for 10-15 percent of the time.

There is a seasonal dependence for mag-netic storminess, as shown in Figure 16.3. Thus,it is seen that magnetic storms are most fre-quent at the equinoxes, a time when the geo-magnetic field is facing more directly into thesolar wind direction. There is also a solar cyclevariation for magnetic storminess but that willbe discussed later in connection with solar ac-tivity and its effect of HF propagation.

Returning to time scales, beyond thosewhich would be associated with drasticchanges in field lines and the depletion of iono-spheric electrons, the question of ionosphericreplenishment needs to be considered. In thatregard, the source is solar UV and it createsionospheric electrons at a steady rate, magneticstorm or not. Thus, the greater the magneticdisturbance and depletion of ionospheric elec-tron content, the longer the time required forsolar UV to restore the F-region electron popu-lation and return critical frequencies back tonormal. That goes a long way in explainingwhy HF propagation is poor for so long afterthe onset of a large magnetic storm.

Figure 16.2 Electron density profiles duringquiet and storm conditions. From Norton[1969]

to 8-10 Re, the trapping or containment of elec-trons is not as durable, field lines being subjectto disturbances that are propagated inwardfrom the impact of solar wind on the bound-ary of the magnetosphere.

The trapping of ionospheric electrons ison the dipole-like field lines at lower latitudesand the outer limit of containment translatesto an upper limit of 60-65 degrees magnetic lati-tude at the earth’s surface. Poleward of thatlimit, critical frequencies foF2 show large varia-tions and propagation across the polar caps ismore uncertain than at lower latitudes and failsbadly during magnetic storms.

As for time scales, several are involved inthe discussion of ionospheric disturbances. Forone, there’s the time scale of the magnetic dis-turbance itself which brings on the ionosphericchanges. The study of geomagnetism showsmany types of disturbances, at different lati-tudes and with different time scales. Limitingthe discussion to magnetic storminess, thosetimes are listed in the Boulder Report or WWVbroadcasts and are characterized by high A-andK-indices, major storms having values of Aabove 40 and K above 5.

The time scale for onset of magneticstorminess is measured in hours and stormdurations may be in days. The frequency ofstorms is seen from data collected over the


Of course, there are less drastic iono-spheric disturbances than the large-scale lossesof propagation which occur during major mag-netic storms when solar wind pressures andinteractions are more severe. Those smallerevents are associated with more modest dis-turbances, say Ap-indices below 40, and theyoften show a 27-day recurrence tendency thatgoes with the solar rotation period. Those per-turbations of the F-region are from magneticdisturbances moving inward, changing foF2critical frequencies after the impact on the mag-netosphere of the enhanced streams in the so-lar wind.

The 27-day recurrence tendency points torelatively stable streams of solar wind plasmawith small angular widths, fixed relative to thesun and rotating with it. Thus, the model forthe recurrent disturbances is one with streamsof solar plasma rotating around every 27 days,sweeping past the slower moving earth and thestreams perturbing the geomagnetic field linesas well as the F-region electrons bound to them.The stability of the streams is such that someof them persist for months on end, sometimesfor almost a year.

A contemporary example of data on thistype of phenomena is given in Figure 16.4where the A-index from the Fred-ericksburgmagnetometer is plotted against time for threesolar rotations. Clearly, magnetic activity peaks

about 3-4 days after the arbitrary 27-daymarker. In terms of Amateur Radio operation,those days are the most disturbed magneticallyand would show the most disturbed propaga-tion, but still below storm conditions. More tothe point, however, the days when the A-Frindex was below 10 would offer the best propa-gation conditions as magnetic disturbance ofthe F-region would be a minimum then.

By way of summary, the above remarksshow the differences between features of thecurrent model of the geomagnetic field and themore elementary dipole model in use beforeWWII. But those remarks do not explain howthe solar wind compresses the field nor themechanism that operates during magneticstorms, moving field lines on the front of themagnetosphere back into the tail region, car-rying their supply of ionospheric electrons withthem. That discussion will deal not just withthe flow of ionized material leaving the sun butalso its close relationship to the interplanetaryfield. The matter proves to be quite involvedand only the essentials will be given here, leav-ing the details to those who want to pursue thefull range of solar/terrestrial relationships.

To begin the discussion, ionized materialcoming from the sun, solar plasma, has an ex-tremely high electrical conductivity. In fact, theconductivity is so high that any electric fieldor separation of charge that might be induced

Figure 16.4 Recurrence of magnetic activity, as given by A-Fr. From NOAA/SESC data.


ern Hemisphere now are merged with the in-terplanetary field and being swept down-stream, as indicated by the arrows.

The merging region at the front of themagnetosphere is indicated by an X-like sym-bol. There is another merging region in themagnetotail, marked in a similar fashion. In thismodel new field lines, on being swept back,exert inward pressure on other lines in the tailand oppositely directed field lines, above andbelow the equator in the magnetotail, maymerge and become closed field lines again.

The figure, of course, only gives a two-dimensional picture of three-dimensional pro-cesses. Thus, the transport of field lines is re-ally a convection process, field lines from thefront of the magnetosphere being swept backinto the tail and then, after reconnection, mov-ing forward again. For a steady-state situation,the rate of field line removal on the front of themagnetosphere must be balanced by the trans-port of field lines forward after reconnection.But those processes require time and experi-ence shows that field line erosion can occur tothe extent that the front of the magnetosphereis inside the orbital location (6.6 Re) of the syn-chronous satellites. Those excursions are brief,measured in minutes, and then the magneto-pause moves outward again.

Qualitatively this model can explain thenew shape of the geomagnetic field as well asthe erosion and transfer of field lines from thefront of the magnetosphere to the rear. Quanti-tatively a number of other aspects have to beadded: the flow speed of solar plasma and itsvariation, the strength and general orientationof the interplanetary field and how the loca-tion and rates of magnetospheric processes, forerosion of field lines at the front and recon-nection in the tail, vary with them.

The above discussion, only a few para-graphs in length, deals with a topic which haskept magnetospheric physicists busy for twodecades. Needless to say, there’s much moreto the story and some further mention will bemade of the model, mainly in connection withthe ionospheric aspects of auroral electron andsolar proton bombardment. That will be donebecause it’s important to have an appreciation

Figure 16.5 A schematic view of magneto-spheric structure due to the interactionbetween the solar wind and the geomagneticfield. From Akasofu [1978]

by a change in magnetic field conditions in theplasma is immediately counteracted or neutral-ized by charge flow. So electric fields do notexist in the flowing plasma and the magneticfield is constant, frozen in the plasma.

It was suggested by J.W. Dungey in 1962that the solar plasma, carrying the interplan-etary field, interacts with the terrestrial mag-netic field in an interesting way. Earlier, theclassic geomagnetic model had field lines comeout of the Southern Hemisphere, cross the geo-magnetic equator and then go downward intothe Northern Hemisphere. Dungey’s proposalwas that the interplanetary field could mergewith the terrestrial field when it had a south-ward component relative to the earth. In thattype of circumstance, interplanetary field linesbeing carried past the earth by plasma flowwould join continuously with the terrestrialfield lines. But with the plasma motion, terres-trial field lines would then be carried in thedirection away from the sun by the solar windplasma.

A snapshot of the process is shown in Fig-ure 16.5, simplified to show field lines in thenoon/midnight plane and with the magneticaxis perpendicular to the plasma flow. There itis seen that high-latitude field lines from theSouthern Hemisphere that would have crossedthe equatorial plane and closed with the North-


of the phenomena which have a negative im-pact on propagation, probably 15-20 percent of

the time, and to understand the full meaningof warning messages related to the problems.


17: SOLAR WIND AND FLARESThe solar wind, as the name implies,

comes from the sun and is solar coronal mate-rial which has expanded out into interplanetaryspace along extended solar field lines. Therange of speeds of the solar wind is from 300-1,000 km/sec, with an average around 400 km/sec. On that basis, the average transit time fromthe sun to the earth is about 4.3 days.

The solar wind consists of protons andelectrons in equal numbers, keeping it electri-cally neutral, and their energies are low, about1/4 electron-Volt (eV) for the electrons and theproton energies are higher (about 500 eV) inproportion to the proton/electron mass ratio.Typical number densities in the solar wind are5 protons per cubic centimeter. On that basis,if one thinks of the earth’s magnetosphere as atarget or collector whose aperture is about 10Re in radius, the power input available fromthe solar wind would be about 2E+10 Watts foraverage conditions. That figure is small com-pared to the 1.8E+17 Watts that the earth re-ceives as sunlight.

Solar flares take place in the vicinity ofactive regions on the sun and are described assudden brightenings in radiation, as typicallyobserved by solar astronomers in hydrogenemissions. Increases in radiation may takeplace in other parts of the spectrum, bursts ofenergetic X-rays giving rise to ionospheric ef-fects, such as sudden shortwave fade-out(SWF) or sudden ionospheric disturbance(SID), from ionization released in the D-regionof the sunlit portion of the atmosphere. An ex-ample of an X-ray burst is found in Figure 5.4,showing that flare X-ray fluxes may reach lev-els like 3E-5 Watts per square meter and con-tinue on for hours after a rather sudden onset.

The relation of SWF events to solar flareswas discovered by Dellinger in 1937. A classicexample of the SWF effect on a radio link isshown in Figure 17.1 where 9.570-MHz signalson a 600-km path fade out for the better part ofan hour. In terms of Amateur Radio operations,

Figure 17.1 Shortwave fade-out on 9.57 MHzat 1915 UTC on 24 November 1937. FromSnyder and Bragaw [1986]

effects of that type are well known, especiallyat the lower frequencies where D-region ab-sorption is the greatest. They occur most fre-quently around the maximum of a solar cycleand if the absorption is intense and of long du-ration, they serve as indicators of flare activityand possible magnetic storm effects on HFpropagation, delayed by some 20-40 hours.

Flares are listed in the weekly BoulderReport according to location on the solar diskand the times in UT when the flare began,peaked in brightness and ended. Those timesmay be given in terms of hydrogen emissionsfrom ground-based observations or X-raybursts at satellite altitude. In addition, they areclassified according to brightness in hydrogenemissions (faint, normal or bright) and X-rayflux in Watts per square meter in the 1-8 Ang-strom range. Finally, flares are listed in impor-tance according to the area in square degreesof solar surface at maximum brightness. Allthat information is summarized in the UsersGuide for the Boulder Report.

Both optical and X-ray flares vary in num-ber during a solar cycle, the optical flares be-ing far more numerous and in phase or stepwith the sunspot number. The Space Environ-


ment Laboratory of NOAA has compiled flarestatistics on the last three solar cycles and findsthe number of optical flares varies between15,000 and 19,000 per entire solar cycle andwith flare rates around 1,000 per month at thepeak of a cycle. But X-ray flares, those whichare most likely to affect the ionosphere, are farless numerous, something like 50-80 per solarcycle and maximum flare rates the order of 15per month. Obviously, with so few X-ray flaresper cycle, it is impossible to chart any sort ofsolar cycle dependence.

There is another aspect to the flare pro-cess which affects the D-region of the iono-sphere, the acceleration of solar protons to en-ergies in the MeV (1E+6 eV) range, even to theBeV (1E+9 eV) range. Those flare events are notas numerous as X-ray flares, between 20 and40 per solar cycle and with peak rates the or-der of 2-4 per month. Beside the difference inenergy, keV for X-rays and MeV for protons,proton events are longer lasting in their effects,often measured in days as compared to an houror so for X-ray events.

In reporting flare events, the Space Envi-ronment Laboratory distinguishes between“proton events” and “PCA events.” The formerare small events, where the solar proton flux isweak and only detected at satellite altitudes.In that regard, the proton detectors on theGOES satellites record proton fluxes above four

thresholds: >1, >10, >30 and >100 MeV. An ex-ample of satellite recordings is shown in Fig-ure 17.2, with three proton flare events withina single week. Considering the long durationof the proton fluxes at satellite altitude, thequestion comes up whether that is characteris-tic of the acceleration process at the source orprotons are leaking out of a magnetic contain-ment region. The latter is the favored interpre-tation, with acceleration taking place duringthe optical flare itself.

Solar protons with energies above 10 MeVmay penetrate to D-region altitudes and cre-ate ionization there by collision processes.However, that is limited largely to high-lati-tude regions as the geomagnetic field acts asan energy filter, admitting particles mainlywhere the field lines are close to the verticaldirection and deflecting away incoming pro-tons from other regions where the field is morehorizontal. Hence, the term PCA for polar capabsorption events as D-region absorption takesplace all across the polar cap during solar pro-ton bombardment.

The ionization of the D-region continuesas long as solar protons penetrate to thosedepths but the resulting electron density de-pends on the degree of illumination. In particu-lar, D-region electrons remain free during day-light hours and may give rise to huge absorp-tion effects (measured tens of dB in the HF

Figure 17.2 Examples of solar proton events recorded at satellite altitude.From NOAA/SESC Report 29 May 1990.


range) but at night, electrons attach themselvesto oxygen molecules and the massive negativeions are far less effective than free electronswhen it comes to signal absorption.

In Amateur Radio operations, the exist-ence of a PCA event is revealed by a blackoutof long duration of signals coming across thepolar caps, say from northern Europe to theWest Coast or from Asia to the northern partsof the East Coast. Around the equinoxes, therewill be some nighttime recovery of signalswhen the polar caps go into darkness but dur-ing polar summer, signals recover slowly as theproton flux declines.

Beyond absorption effects in the HF range,radiation effects from energetic solar protonevents pose a threat to high-altitude aircraftflight, disturb VLF navigation and communi-cation systems that depend on the D-region.The additional ionization which results froman influx of solar protons looks like a sunriseto those systems, perhaps out of character fora site in darkness if the protons have been de-flected into that region by the geomagneticfield.

Solar flares do generate noise in the radiospectrum, all the way from the HF range (10MHz) to the microwave range (30 GHz). Whileemissions vary significantly from one flare tothe next, there is a sweep aspect to frequencies

in the solar radio spectrum, illustrated in Fig-ure 17.3. In that diagram, the emissions whichmay be heard at a given time lie along a verti-cal line, say a broadband continuum of noisein the 10-300 MHz range about 40 minutes af-ter the flare began. Similarly, along a horizon-tal line, one would note noise bursts which driftdown in frequency, rapidly past 100 MHz veryearly in the noise storm and then more slowlyabout 10 minutes into the noise emissions. Inpresent practice, discrete frequencies, 245 MHzand 2,695 MHz, are monitored and when thepeak fluxes reach certain threshold levels abovebackground, NOAA will issue an Alert mes-sage to interested parties.

As a result of solar noise emissions, ama-teur operators on the higher HF bands may alsohear the sort of whooshing sound that goeswith flares. Of course, being in the HF portionof the spectrum, solar radio noise bursts do notcontribute to ionization processes like the ac-companying X-rays do. Thus, except for beinga warning signal for flare activity, solar noisebursts are more of a nuisance than anythingelse.

Earlier in this section, it was mentionedthat flares may be followed by magnetic storms,some 20-40 hours later. That was the classicalapproach to magnetic storms, dating back tothe 19th century when it was found that a mag-

Figure 17.3 Dynamic spectrum of solar radio noise emissions. From NOAAPreliminary Report and Forecast User Guide.


netic storm occurred about a day after a large,bright flare. So, up to the 1960s, the idea wasput forward that a blast wave was generatedin the flare process, ionized material movingat 1,000-2,000 km/sec then arrives at earth’sorbit and somehow triggers a magnetic storm.That was in addition to the electromagneticemissions (radio, light and X-rays) which camepromptly. The idea of a blast wave had some

support from the fact that many disturbancesstart with a sudden storm commencement(SSC), suggesting the arrival of a puff of solarmaterial. On the other hand, a good number ofmagnetic disturbances of storm proportionsbegan with gradual commencements (GC) sothere was still some room for new ideas andwe turn to them next.


As noted earlier, the earth’s magnetic fieldis monitored with the aid of magnetometers,devices which follow changes in the magnitudeand direction of the field. Under quiet condi-tions, the components of the field vary by onlya few tenths of a percent in the course of a day.But during magnetic storms, departures fromquiet day levels may reach several percent andlast for times ranging from hours to days. Andstorms can be either sudden (SSC) or gradual(GC) in commencement.

But the association between magneticstorms and solar flares has not always beenone-to-one as some large flares are not followedby magnetic storms. Then, too, magneticstorminess occurs in the absence of significantflare activity. The one idea that now seems be-yond dispute is that the sun evaporates ion-ized matter, solar plasma, part of the solar co-rona becoming so hot that even the sun’s grav-ity is unable to retain it. That steady sort ofemission is the source of the solar wind, stream-ing outward at about 400 km/sec.

The earlier idea that associated flare ac-tivity with magnetic storms suggested thatdenser, faster-moving solar plasma left the vi-cinity of a flare site and if directed toward theearth, it would be responsible for the magneticstorm that ensued by some sort of interactionwith the geomagnetic field. That associationwas made stronger by the fact that many posi-tive associations were made between flaresnear the center of the sun, i.e., the central me-ridian of the solar disk, and magnetic storms.But there were also instances where the asso-ciation or storm prediction was not fulfilled.

And there were similar successes and fail-ures in connection with energetic solar protonscoming from large flares, success often com-ing with the prompt arrival of solar protonsfrom a flare site west of the central meridian.But then there were exceptions, large flareswest of the central meridian and no protons orelse protons but quite delayed in their arrival

at the earth’s orbit. And some proton eventshad no association with flare activity on thedisk at all, perhaps coming from an active re-gion that had rotated past the western limb.

All of the above discussion was in con-nection with ground-based observations. Butit was based on older techniques, before recentadvances in the technology of solar astronomy.There the solar coronagraph was perfected af-ter WWII to the point that the solar coronacould be examined without the aid of eclipses,even with instruments flown on satellites. Andthere were advances in the technology of spec-troheliographs as well, giving solar images inselected wavelengths. Those new techniquesprovided opportunities to look for associationsbetween magnetic storms and other solar phe-nomena, especially what are termed “coronalholes” and “coronal mass ejections” (CME).

Now a few words about coronal holes.Coronal holes are regions where solar field linesgo off into space and from which solar windmaterial may readily escape. The speed ofstreams from those regions is high by solarstandards, around 600 km/sec, and it isthought that the flow of material from the re-gions extracts so much energy that the regionsare cooled down and appear darker, like sun-spots, in comparison with nearby regions. Theregions seem to be long lived, on occasions last-ing many 27-day solar rotations, and give riseto recurrent forms of modest magnetic activ-ity, as in Figure 16.4, or even geomagneticstorms of some significance.

A series of coronal hole maps is shown inFigure 18.1; those maps are coronal hole con-tours from 1,083 nm (10,830 Angstrom) imagesof the sun taken by the spectroheliograph atKitt Peak, NM. The rotation of the coronal hole,day by day, toward a central meridian positionis seen from the figure and the high-speed so-lar wind stream from the region contributed,in part, to the change from quiet conditions for26-28 October to major storm levels on 29-30

18: MAGNETIC STORMS AND AURORAE


with flare activity show that CME’s usually liftoff before the flare activity begins. But the lead-ing edges of CMEs have speeds from aroundthat of the ambient solar wind, 400 km/sec, upto 1,200 km/sec. The higher range of speedssuggests that when a CME moves outward,into normal solar wind from the sun, it maydrive a shock wave ahead of itself in the sur-rounding solar plasma.

In substance those are the circumstancesnow suggested to lead up to sudden com-mencement magnetic storms; all that’s neededis solar wind in the vicinity of the earth withthe interplanetary field it carries pointingsouthward. When a CME-driven shock wavereaches the earth’s orbit, geomagnetic fieldlines merge with the field lines in the shockedplasma and a magnetic storm gets under way,complete with strong ionospheric effects. Thescheme has no need to be associated with flareactivity although it may be present after a CMEhas started. Of course, the time delay and mag-nitude of the effects which result from a CMEdepend on its speed, with delays ranging from1-4 days.

During the recent ’94 CQ WW DX SSBContest all those factors came into play, as mayseen by reading the sections of the Solar Re-port issued at 2200Z on 29 October on theNOAA/SESC BBS. The Report is quite long soonly the Geophysical Activity Summary for theperiod 28/2100Z to 29/2100Z is given and theGeophysical Activity Forecast for the next dayis quoted:

“The geomagnetic field began the periodquietly. However, a sudden impulse of 24gamma (24 NT) was observed at 29/0025Z.This is attributed to a long duration (CME)event that occurred on 25/1000Z for a lengthy86-hour delay. The field became active shortlyafter the impulse and minor to severe stormconditions followed in the 0900-1500Z interval.Some high-latitude sites experienced K-indicesof 9 in the 1200-1500Z interval. A brief magne-topause crossing was observed at GEOS-7 be-tween 29/1430-1433Z. At the end of the period,the field calmed to unsettled levels. Energeticelectron fluxes (capable of disabling controlson some spacecraft) began the period at high

October 1994. In that latter period, a magneto-pause crossing was observed briefly by theGOES spacecraft, at 1430-1433 UTC on 29 Oc-tober.

In contrast with the long-lived solarstreams associated with coronal holes, there areother emissions of coronal material, sporadicor noncurrent in nature and not flare related.From coronagraph observations at satellite al-titudes, it appears that large quantities of solarmaterial are ejected sometimes into interplan-etary space, thus amounting to coronal massejections (CME) when they occur. And cur-rently, studies in relation to other forms of so-lar activity suggest that CMEs have a close as-sociation with eruptive prominences or disap-pearing filaments, having comparable widthsat the sun but generally removed from activeregions.

And observations indicate that CMEsoriginate in regions where solar field lines areclosed, away from regions around coronalholes where solar wind plasma is expandingoutward. What associations in time there are

Figure 18.1 Coronal hole maps for 24-30October 1994. From NOAA/SESC Report


Figure 18.2 Annual mean sunspot number and magnetic storms per year,1932-1987.

levels but dropped to normal background lev-els by 29/0300Z.

"The geomagnetic field should be activeto major storm levels for 30 Oct. As the currentdisturbance ebbs, it will be replaced by a coro-nal hole related one. Unsettled to minor stormconditions would be experienced through 03Nov as a result of that high speed stream. En-ergetic electron fluxes should return to highlevels around 01 Nov.”

Those remarks pretty well summarizewhere matters stand now as far as the initia-tion of magnetic storms is concerned. Ofcourse, storms have been known for more thana century and the same is true of the impor-tance of sunspot counts when it comes to tak-ing a measure of solar activity. As noted ear-lier, solar flares occur in numbers proportionalto the rise and fall of sunspot numbers. That isnot true of magnetic storms, as may be seen inFigure 18.2 where both the distribution of mag-netic storms and sunspot count are shown for5 solar cycles. Except for Cycle 19, an excep-tional cycle in many ways and shown in themiddle of the figure, it’s seen that magneticstorm activity is delayed relative to the peakin sunspot numbers.

A more detailed view of the relationshipof storm activity and the phase of a solar cyclemay be seen in Figure 18.3 where storm data

for the 5 cycles are superimposed and plottedyear by year in terms of the percent of stormsin the entire cycle. While there is no unique re-lationship evident, the delay of storm activityrelative to solar maximum is clear and the sta-tistics show that for those 5 cycles, the mag-netic storms after each solar maximum out-numbered those before by a 3-to-1 ratio. By wayof interpretation, the post-maximum peak inmagnetic storm activity is related, in part, tothe fact that high-speed streams from coronalholes are more frequent after solar maximum.

With the initiation of a magnetic storm,large-scale changes take place within the mag-netosphere — high-latitude field lines aredrawn back into the magnetotail, solar protonsbombard more of the polar cap if a PCA eventis in progress, energetic electrons rain down atauroral latitudes and result in auroral luminos-ity and ionization at altitudes above 100 km.All that signifies the dissipation of energy butthen the question becomes one of the energyinput from the solar wind and its transfer, es-pecially to the far side of the magnetospherewhere the aurora occurs.

Those are some of the questions whichkeep magnetospheric physicists busy and evenafter three decades, there is no universal agree-ment on any detailed theory. But that shouldn’tbe surprising considering the great magnitude


Figure 18.3 Percent of magnetic storms per year during Solar Cycles 17-21.

and scope of the problems and the fact that thevast expanses involved make data samplesseem meager, even with a vigorous space re-search program. But progress has been made,stemming from the revision of the model forthe geomagnetic field and the realization thata transfer of energy takes place from the solarwind to the magnetosphere whenever the in-terplanetary field has a southward component.The rest is in the details which remain to besettled: whether the energy input from the so-lar wind is dissipated continuously or possi-bly stored, to be released at a later time, andhow low-energy electrons in the magneto-sphere are accelerated from about 1 eV to the10 keV or so required to produce auroral dis-plays.

Those are good questions but may be ofless interest to those concerned primarily aboutHF propagation. For them, the ionospheric as-pects of the various phenomena are more rel-evant so let’s turn first to auroral ionization. Inthat, the electron influx during auroral displaysnot only excites atoms and molecules to emittheir characteristic radiations but also gives riseto intense ionization by collisions at the end ofthe electron’s range and X-ray fluxes which canpenetrate deep into the atmosphere.

Just as with the quiet ionosphere undersolar illumination, the electron density thatbuilds up with auroral bombardment depends

on a competition between ionizing and lossprocesses. First, the rate of electron productionin a region depends on the velocity and flux ofthe electrons as well as its atmospheric com-position and density. At auroral heights, atomicoxygen is an important constituent, as shownby the green color in auroral displays. The otherparts of the auroral spectrum depend on thepresence of molecules of nitrogen and oxygen.

Next, loss processes take place because ofelectron recombination with positive ions andare fairly rapid around 100 km altitude, in con-trast to the higher F-region where the rate ofrecombination is much slower. But with highincoming fluxes, significant electron densitiesbuild up and strong ionospheric absorption ofsignals may occur during auroral displays,even showing rapid changes in absorptionmuch like changes in auroral luminosity. Anexample of such variations is shown in Figure18.4 where the intensity of green auroral emis-sions during an auroral event after local mid-night is compared with the ionospheric absorp-tion of 27.6 MHz cosmic radio noise, both alongthe same field of view from the vertical direc-tion.

Even though ionospheric absorptionevents show large, rapid variations during au-roral displays around local midnight, auroralabsorption is evident at other times of daythroughout the year. The former events are as-


Figure 18.4 Variations in ionosphericabsorption and auroral intensity for dir-ections 12 degrees north and south ofCollege, AK. From Ansari [1964]

sociated with the breakup phase of auroral dis-plays while the other, more enduring absorp-tion events are associated with well-defineddecreases in the local magnetic field on mag-netometer records or extend into morninghours and beyond.

In that regard, Figure 18.5 shows the dailyvariation of auroral absorption at College,Alaska during the 5-year period centered onthe peak of activity in Cycle 19. The variationsare divided between the five most magneti-cally disturbed (upper curves) and five mostquiet (lower curves) days for summer, winterand equinoctial periods and show how the ab-sorption of 27.6 MHz varied with local time.Clearly, absorption is not limited to just whenauroral displays might be visible and wouldhave to be reckoned with when it comes to HFpropagations as results indicate that absorptioneffects extend over a considerable range of lo-cal time or longitude.

But the region of ionization giving rise tothe absorption of signals during auroral activ-ity is more limited in latitude, just as is the casefor auroral displays, and may be seen in Fig-ure 18.6 where the auroral absorption zone inAlaska is compared with the visual auroralzones (incidence and occurrence) from all-sky

camera data. From those data, it is seen thatauroral absorption (AA) events are restrictedto a fairly narrow range of geomagnetic lati-tudes, in contrast to PCA events which coverthe entire polar cap. Another difference be-tween the two types of absorption events is thatPCA events are slowly varying and of longduration, measured in days, while AA eventsshow rapid changes, as in Figure 18.4, and lastonly for hours at a time.

It should be noted that the magnitude ofauroral absorption events around 30 MHz mayreach several dB, but not tens of dB. But at au-roral heights just above the 100-km level, therelative absorption efficiency of electrons isquite low, as shown in Figure 8.1, and that im-plies high electron densities, around 1E+6 percubic centimeter, over a small range of altitude.That is the origin of auroral E-layers that theVHF enthusiasts enjoy so much.

For most HF operators, however, auroralE-layers amount to another absorbing regionthat’s below the F-layer, extends E-W alongauroral displays and is narrow in the N-S di-rection. And there’s no day/night effect forabsorption events at auroral heights like thatfound in the D-region during PCA events. That

Figure 18.5 Daily variation of auroralabsorption at College, AK for 5 magneticallydisturbed (upper curve) and 5 quiet (lowercurve) days from September 1957 to August1962. From Basler [1963]


Figure 18.7 Auroral absorption for differentsolar elevation angles as a function ofhorizontal magnetic disturbance. FromBrown and Barcus [1963]

Figure 18.6 A comparison of the auroral absorption zone in Alaska withvisual auroral zones from all-sky camera data. From Basler [1963]

is seen in Figure 18.7 where auroral events ob-served during various levels of local magneticdisturbance showed no day/night variation inabsorption, when plotted according to the so-lar elevation angle at the time of observation.

Another aspect of auroral events thatwould be of interest to the HF operator, par-ticularly one engaged in making long-path con-tacts across high latitudes, is that auroral phe-nomena show similar effects in conjugate re-gions, i.e., regions on the earth connected bymagnetic field lines. Thus, auroral electronsmay spiral down field lines from themagnetotail and produce similar effects at sitesin the two hemispheres which are separatedby as much as 14,000 km. An example of a con-jugate auroral absorption event is shown inFigure 18.8 where the absorption of 27.6 MHzcosmic radio noise during a magnetic storm isseen to be comparable at conjugate auroralsites, Kotzebue, Alaska (KL7) and MacquarieIsland (VK0), Australia.

Finally, it should be pointed out that theabove discussion, while not all that compli-cated, goes beyond what is experienced in typi-cal Amateur Radio situations, dealing with re-sults of observations obtained by looking upat a small region of the sky instead of obliquely,off to the horizon, as in Amateur Radio opera-tions. And, when you get right down to it, thatdifference is what makes propagation a chal-lenge, not only for what is attempted but for

the range of circumstances, even conditionsthat may exist along a path.

To see what I mean, think of a path for adistant contact you might attempt to make: itstarts in sunlight, goes across the magneticequator, through mid-latitudes and the auroralzone in the opposite hemisphere and back to-ward a station in darkness in the original hemi-sphere. That has the all the makings of a dusk-to-dawn long-path contact. But what about allthe ionospheric phenomena that would be in-


of contact is made; it happens all the time. Moreto the point, it’d be satisfying to understandhow it all works and then be able to make it areality, on the right frequency and at the righttime. That would make any Little Pistol happy,no doubt about it.

Figure 18.8 Time variations of auroral absorption during an event recordedat three riometer sites in Alaska and one in the Southern Hemisphere. FromLeinbach and Basler [1963]

volved along the way — first, transequatorialpropagation, then from mid-latitude to auroralzone conditions and finally ionospheric gradi-ents, to name a few.

That sounds like an ionospheric smorgas-bord or something that’s been put together bya committee. But it’s no miracle when that sort


Whether the Little Pistol thinks propaga-tion is good or bad depends on his knowledgeof what propagation should be like under thebest of circumstances and what it happens tobe, given the state of solar/terrestrial condi-tions. As for “should be,” that goes to the prop-erties of an undisturbed ionosphere for thesmoothed sunspot number (SSN) which char-acterizes the phase of the solar cycle. Knowingthe SSN, the LP could check out the rosy sce-nario for various DX possibilities by runninghis HF propagation program to see what pre-dictions it comes up with for openings, at leastMUF values and workable S/N ratios.

Better yet, LP could have done his home-work in advance and knowing the season, al-ready have a prediction in hand for the earli-est DX opening of the day on 14 MHz. Thatwould be off to the East of his QTH (Boulder,CO), say, when the West Europeans might startcoming in. For example, with an SSN of 50,West Europeans start to come through aftersunrise, around 1500 UTC in the winter. Lis-tening then, it would be a question of what isheard and what it means.

Suppose the LP points his beam towardnorthern Scandinavia and listens for signals,say from LAs, SMs or OHs; from Boulder, thosepaths go across the northern auroral zone, closeto the polar cap, and the presence or absenceof those signals means something for the stateof the magnetosphere. If the Scandinavianswere heard, that would suggest a quiet mag-netosphere with nothing in the way of auroralactivity. But if they were not heard within areasonable time, the next thing to do is checkon signals from northern Europe, say Fs andDAs; those paths go across the southern tip ofGreenland and are at the lower edge of theauroral zone. If nothing is heard from them,that raises the spectre of a magnetic storm,minor or major.

To check that out, the LP would have to

look for signals from southern Europe, say EAs;paths like that go below the auroral zone anddepend on F-region ionization that’s held onfield lines deeper in the magnetosphere. Ifnothing’s heard from those paths, it’s either amatter of a major magnetic storm in progressor something’s wrong with LP’s antenna. Tocheck on the latter point, LP would turn thebeam toward South America and see what’scoming through.

For LUs and PYs, the path from Bouldergoes off to the southeast, maybe 1,000 kmlonger than to Europe and in full sunlight at1500 UTC. With sunlight on the path, LP couldexpect signals on the weaker side but thereshould be no problem with the MUF for thepath as only the robust equatorial ionosphereis involved. So the LP would have an answer,one way or the other.

Now I have to confess to dragging out thatdiscussion, perhaps unduly, for pedagogicalpurposes, taking you through the steps in theanalysis of band conditions, something like thequalitative analysis scheme you used in yourhigh school chemistry course to identify anunknown compound. It’s a valid approach butthere are shortcuts in almost anything we at-tempt and so it is with propagation, the LP justpointing the beam toward Europe and listen-ing for calls on the band. On 14 MHz, a tribandYagi has a forward pattern which is broadenough, typically 70 degrees between –3 dBpoints, to include directions from Finland toSpain. So with one beam setting and some se-rious listening, the LP may come up with ananswer as to how propagation shapes up orwhether storm conditions have already takenover.

But there are other shortcuts for the LP touse; the simplest is just to tune in the WWVbroadcast at 18 minutes after every hour andlisten to the statements about solar and geo-magnetic activity. If there was a magnetic storm

19: PROPAGATION CONDITIONS, BAD & GOOD


in progress, the LP would hear about it. Butwhat do the statements about solar activity andmagnetic activity mean?

On WWV broadcasts, solar activity is de-scribed as being between very low and veryhigh, with low, moderate and high boundedby those extremes. Magnetic activity uses anA-index derived from the magnetometer atBoulder and estimates are given for the aver-age of the A-index for the current day and alsothe K-index in the last three-hour interval.

Solar activity is expressed in terms re-lated to the number and type of X-ray flaresobserved at satellite altitude. The type or classof flare is given with alphabetic designations— A, B, C, M and X — the X-ray flux increas-ing in that order. For example, “moderate ac-tivity” would be solar conditions where thesatellite X-ray detector recorded 1-4 M-Classflares while “very high activity” would meanfive or more 5M-Class flares. No mention isgiven for optical flare activity, coronal holes orcoronal mass ejections.

The description of magnetic activity be-gins with conditions which are “quiet” or “un-settled”; that means the estimated A-index forthe day is less than 16 and the most recent three-hour K-index was probably less than three.That’s a very good omen! But when the geo-magnetic field is said to be “active,” the A-in-dex is then between 16 and 30 and the K-indexis up to 4.

It’s at that point when the warning lightshould go on in LP’s head, especially if the A-index is in the high end of the range. That’sclose to some form of serious disturbance, thefirst being “minor storm” when the A-index isbetween 30 and 50 and the K-index is up to 5.At those times, one can clearly notice problemswith HF propagation and they only get worseif followed by “major-storm” conditions whenthe A-index is between 50 and 100 and withthe K-index 6 or above, or “severe storm,” then

the A-index is greater than 100 and K-indicesare 7 or greater, the highest possible being 9.

The information broadcast on WWV at 18minutes past each hour, and on WWVH at 45minutes past each hour, is a 45-second “Geo-physical Alert Message” and can also be ob-tained by telephone, calling (303) 497-3235 atany time. The message on that line is a taperecording giving solar and geophysical activ-ity and indices for the most recent 24 hours andfor the next 24 hours, updated every three hours.

If a computer and modem are available,the price of a phone call to Boulder is betterspent by calling the Public Bulletin Board Sys-tem (PBBS) at (303) 497-5000 and getting morecomplete information and forecasts. In callingthe PBBS, use a protocol with a conventional8-bit word, one stop bit and no parity. The PBBSwill operate at 300, 1200 and 2400 baud. A wideselection of options is offered on the menu thatis presented but for LP’s purposes, the Solarand Propagation Reports would be the ones tocopy and can be downloaded in less than aminute with a 1200-baud modem.

The long-distance charge for a call to theNOAA PBBS will be minimal, especially if thecall is made at night or in the morning, beforeworking hours. For here in the Northwest, thecurrent Solar and Propagation Report is in theshack every morning for a cost less than $5 amonth. Your charges may be different so checkwith your phone company for the best timesto call and toll charges.

In order to see what is included in thoseReports, consider first the Solar Report for 3October 1994. It consists of six (6) sectionswhich will be quoted as received from thePBBS. On reading through it note the contrastbetween geomagnetic and solar conditions —severe storm levels, surely with major disrup-tion of HF propagation, and very low solar ac-tivity, no problem. So read on; the report fol-lows:

IA. Analysis of Solar Active Regions and Activity from 02/2100Z to 03/2100Z: Solar activity was very low.

IB. Solar Activity Forecast: Solar activity is expected to be at very low levels.


IIA. Geomagnetic Activity Summary from 02/2100Z to 03/2100Z:The geomagnetic field has been at quiet to unsettled levels until 02/2230Z when activityincreased to minor to severe storm levels. This activity is believed due to a favorably positionedrecurrent coronal hole feature. The greater than 2 MeV energetic electron flux has been mostlynormal to moderate. At about 03/1400Z the flux exhibited a rapid rise to peak in the high rangeat about 03/1600Z.

IIB. Geophysical Activity Forecast: The geomagnetic field is expected to be at major storm conditionson day one and then becoming minor storm levels the second day of the forecast. Active condi-tions are expected on day three.

III. Event Probabilities 04 Oct 06 OctClass M 01/01/01Class X 01/01/01Proton 01/01/01PCA Green

IV. Penticton 10.7 Cm FluxObserved 03 Oct 074 Predicted 04 Oct-06 Oct 074/074/07490-Day Mean 03 Oct 078

V. Geomagnetic A IndicesObserved AFr/Ap 02 Oct 012/009Estimated AFr/Ap 03 Oct 070/065Predicted AFr/Ap 04 Oct-06 Oct 050/060-035/040-020/030

VI. Geomagnetic Activity Probabilities 04 Oct-06 Oct A. Middle Latitudes

Active 20/30/30Minor Storm 30/30/30Major-Severe Storm 30/20/20

B. High LatitudesActive 20/20/30Minor Storm 30/40/30Major-Severe Storm 40/30/20

So that was the bad news and forecasts ofthings to come, from the solar and geophysi-cal standpoint. Now what about propagation?That’s covered in a separate report. However,

Propagation Reports are given every six (6)hours and include a forecast for the next six (6)hours. That being the case, consider the follow-ing:

Primary HF Radio Propagation Report issued at 03/0520Z Oct 94.

Part I. Summary 03/0000Z to 03/0600Z Oct 94 Forecast 03/0600Z to 03/1200Z Oct 94

Quadrant I II III IV0 to 90W 90W to 180W 180 to 90E 90E to 0

Region Polar U2/-25 U2/-25 U3/-20 U3/-20Auroral U2/-30 U2/-30 N4/-20 N4/-20Middle N4/-20 N4/-20 N5 N5Low N5 N5 N6 N6Equatorial N6 N6 N7 N7

Part II. General Description of Propagation Conditions Observed during the 24-hour period ending 02/ 2400Z, and Forecast of Conditions for the next 24 hours.


Conditions were normal until the last 6 hours of the day when geomagnetic storming in the polar andauroral areas caused degradation, mostly in the night sectors, multipathing and absorption caused verypoor to fair conditions in those areas.

Forecast: Conditions should get worse, especially in the night sectors, and may cause some transauroral/polar paths to be unuseable. Multipathing, absorption, sporadic-E and fading may occur, and MUFsshould be depressed significantly. The storming may persist for several days.

Part III. Summary of solar-flare induced ionospheric disturbances which may have caused shortwavefades in the sunlit hemisphere during the 24-hour period ending 02/2400Z Oct 94. . .Start End Confirmed Freqs Affected . . . NoneProbability for the next 24 hours. . . . Nil

Part IV. Observed/Forecast 10.7-cm Flux and K/ApThe observed 10.7-cm flux for 02 Oct 94 was 074The forecast 10.7-cm flux for 03, 04, 05 Oct 94 are 076, 077, and 077.The observed K/Ap value for 02 Oct 94 was 02/09.The forecast K/Ap values for 03, 04, 05 Oct 94 are 03/15, 04/25, and 04/30.Satellite X-ray Background: A1.0 (1.0 E Minus 05 ergs/cm sq/sec).The effective Sunspot Number for 02 Oct 94 was 20.0.

Secondary HF Radio Propagation Report issued at 03/1200 Z Oct 94.

Part I. Summary 03/0600Z to 03/1200Z Oct 94Forecast 03/1200Z to 03/1800Z Oct 94

Quadrant I II III IV

0 to 90W 90W to 180W 180 to 90E 90E to 0Region Polar U3/-20 U2/-25 U2/-25 U3/-20

Auroral N4/-20 U2/-30 U2/-30 N4/-20Middle N5 N4/-20 N5 N5Low N6 N5 N5 N6Equatorial N7 N6 N7 N7

Part II. Conditions affecting HF Propagation: Geomagnetic storming which began at 03/00Z has degraded propagation, particularly in the higher latitude night sectors. Forecast: Expect little change ingeomagnetic conditions over the next 24 hours. Multipathing, absorption and fading are likely for theentire period.

Before getting to propagation matters, it’sclear, on comparing the Solar Report andPropagation Reports, that propagation fore-casters underestimated the potential for mag-netic activity, forecasting Ap of 15 for 03 Octwhile the estimated value during 03 Oct was65. Their only solace would be that the mag-netic field was quiet, as they had forecasted,until the last 6 hours of 03 Oct 94. But even ifthe storming held off, it would have ruinedtheir forecast for 04 Oct.

The only part of this report that needs ex-planation is where conditions are specified ac-cording to region. In that regard, the Propaga-

tion Report distinguishes between normal (N),fair (U) and poor (W) conditions. Thus, the con-ditions in higher latitudes were fair to normal.But there are degrees of each, 1 through 9:

1. Useless 2. Very Poor 3. Poor 4. Poorto Fair 5. Fair 6. Fair to Good 7. Good8. Very Good 9. Excellent

and also predictions of MUF values, when 20%or more from seasonal means. Those predic-tions, either positive or negative, are for single-hop, 4,000-km paths with midpoints located inthe relevant grid block.

With those distinctions in mind, we cango on to the update:


would be quiet for the lack of storm conditions,and then the number of those which would bestormy with corresponding perturbations ordisruptions of propagation conditions.

While the thrust so far might seem to beslanted or biased toward the bad side of propa-gation conditions, those figures tell anotherstory related to the good side. For example,over the three-year period covered, the data inFigure 19.2 show that, on the average, propa-gation conditions were free of any sort of stormdisruption 84.3% of the time. That’s the goodnews, only tempered by the fact that variationsabout the average reached a promising 95.9%during the first part of ’91 and a grim 69.1% inthe next three-month interval during themiddle of ’91.

The data in Figure 19.1 indicate that therewas an average of 6.7 energetic flares per monthduring that three-year period. Those disrup-tions, SWFs or SIDs, are short in duration, moreof an annoyance than anything else. So thegood news is still that about 84% of the time,right there during Solar Maximum whenpropagation is the greatest, the bands wereopen to the Little Pistol for DXing. That wassomething to celebrate at the time or to hope for inlooking forward to when Cycle 23 comes around.

That was the bad news and one shouldn’tdwell on it too long; after all, most of the time,propagation is not disrupted that much. Butthere are some points worth considering be-fore going on to the good news. For example, agreat deal of information was available to theforecasters, both solar and propagation, rightat their fingertips, far more than the Little Pis-tol had. That would limit any forecasting ef-fort by the LP if based only on the numbersfrom WWV. At best, the LP could plot valuesof the solar flux and A-index but beyond fore-casting 27-day recurrences of those values andwhat they imply, little more could be said ordone without more information. However,even 27-day recurrences are variable and un-certain, being found more frequently on thedescending side of a cycle. And disruptionsfrom energetic X-ray flares vary too, being morefrequent near the peak of a cycle. That is shownby Figure 19.1 which displays variations in thenumber of energetic X-ray flares in three-monthintervals across the peak of activity in Cycle22. But no matter what the source, propaga-tion disturbances are the main concern here.

In that regard, Figure 19.2 shows geomag-netic variations for the same period, the per-cent of days in three-month periods which

Figure 19.2 Levels of magnetic activity inthree-month intervals, 1 April 1989 to 31March 1992.

Figure 19.1 The number of M5 or greater X-flares in three-month periods, 1 April 1989to 31 March 1992. From NOAA/SESC Report.


In talking about the expectation of Cycle23, the next in the series starting back in 1750,it’s worth pausing for a bit and taking a look atwhat the historical record has to say about so-lar cycles. The Little Pistol, of course, is hop-ing for another humdinger, maybe even asgood as Cycle 19, the greatest of all time — re-corded time, that is. But what was the range ofsunspot numbers in those previous cycles?That’s worth looking at, even briefly, so turnto Figure 20.1; that shows the range of the maxi-mum and minimum sunspot counts, up to themaximum of the present one, Cycle 22.

Looking at that figure, the Little Pistol cansee that things have not always been the sameas they are now and, as a corollary, probablywon’t be the same in the future. At the moment,we’re on the downside of Cycle 22 with an SSN

to 1715. Of course, there is some question aboutthe reliability of the actual sunspot counts backin those times. After all, it was just at the verybeginning, telescopes and sunspots being newon the scene. But there are other indications,say almost a total lack of auroral sightings inthe period, to support the idea. So that’d be achilling experience, radio propagation essen-tially carried out at minimum levels of solaractivity.

But think what life would be like on theHF bands without the geomagnetic field. Wecurse it roundly now when magnetic stormsare in progress but when things calm down,we’d be lost without it as it’s the vital agentthat keeps the F-region in place when the sungoes down. Without it, there’d be the D- andE-regions when the sun is up in the sky andsome sort of higher region that would persistsomewhat after sunset, until electron-ion re-combination reduced the electron density to alow level due to just the ionization from scat-tered starlight and cosmic rays. As a guess, it’dprobably be like the F1-region is now, presentin daytime but less of a DXer’s choice than theF2-region. In any event, that’d be DXing, prettymuch a daytime affair.

While the geomagnetic field is not goingto disappear, the Little Pistol should know thatit does wander around the globe. A fascinat-ing paper by Fraser-Smith (1987) deals withpole wandering in recent times, the North Polein the centered-dipole approximation movingfrom 70° N, 263° E in 1831 to 77.4° N, 257° E in1984. Since the earth’s field controls ionosphericelectrons to a large degree, it is not out of thequestion that, given time, the foF2 maps thatwe depend on now will have to be redonesometime in the next century, as the pole wan-dering continues its course.

Of course, there’s the whole science ofpaleomagnetism which gives glimpses of oldgeomagnetic orientations from magnetism fro-zen into various artifacts when they solidified

20: HISTORICAL FEATURES

Figure 20.1 Maximum and minimumsunspot numbers for Cycles 1-22.

around 25. Back in September ’93, the SSN was50, at about the maximum level in Cycles 5 and6, so the LP could review his DX log and seewhat might be expected if Cycle 23 came inwith a maximum SSN of 50 or so. Grimthought.

But even worse would be another maun-der minimum, a sunspot drought with SSN lessthan 20 that lasted about 75 years, from 1640


in earlier geological times. Those are extremesthat we need not worry about in our lifetimesbut are amusing to think about. For example,what would HF propagation be like if thenorth-seeking pole were located on the Inter-national Date Line at the geographic equator.That places the south-seeking pole at the equa-tor and on the Greenwich meridian. So whatwe now call transpolar propagation to Europewould be transequatorial propagation, geo-magnetically speaking, and one can’t help butwonder what the cold polar temperatureswould do to the chemistry which bears on iono-spheric physics. That’ll take a while to figureout, maybe to be worked on during a cold,wintry day or two.

Of course, the ultimate historical momentfor radio was when Marconi carried out hisexperiment across the Atlantic in 1901 whenthe SSN was only 3. That moment and otherslike it are interesting to contemplate but willnot be recreated in real time again. About thebest that can be done is to explore propagationin other times using some standardized eventor touchstone, making use of what can be donewith computers. In that way one wouldn’t haveto go at the slow pace of real time; instead, witha computer, one could have a model ionosphereto play with, almost like a toy, and satisfy one-

self on how it really worked. In that regard, Iuse the CQ WW DX Contest format in a com-puter game I devised, SOLAR MAX, and cansee how game scores are affected by sunspotnumbers, seasons, transmitter powers orQTHs. Some use SOLAR MAX to check outband plans for contests but my interests aremore historical in nature and I can make greatleaps back in time and space with SOLARMAX.

In concluding this short section, it is im-portant that some sort of numerical compari-son be put forward, a number to use in talkingabout historical matters. So turning to Figure4.6, giving critical frequencies at noon and atmid-latitudes as a function of SSN, the LittlePistol could expect about a 30% drop in criticalfrequencies in Cycle 23 if it came in with a maxi-mum SSN of 50, like Cycles 5 and 6. But at themoment, NOAA and others are not of thatmind and are predicting that the next solarmaximum will be in March 2000 with a maxi-mum SSN of 108, with an uncertainty in SSNof +/-20. That’d be a bit below the average(111.7) of all the earlier cycles and critical fre-quencies only about 10% below those in Cycle22. That’s the current prediction so all the LittlePistol has to do is wait and see how accurate itis when the real thing comes around.


In the early days of radio, it was a matterof observing the various phenomena as theybecame evident and trying to find their origin.At the low frequencies first used, day-nighteffects were immediately apparent and the roleof solar illumination, although not fully under-stood, was deemed important. With the startof ionospheric sounding, a major step forwardwas taken and the study of radio propagationbecame more of a science, measures of the iono-sphere overhead taken in the course of a dayrather than being observed off obliquely, acrossvast distances and in different directions.

And then there were disruptions — radiostorms — and they soon became associatedwith magnetic disturbances. Since there was along history connecting terrestrial magneticactivity and solar activity, as given by sunspotcounts or measures of the area of sunspots, itshould come as no surprise that aspects of ra-dio propagation were explored too, to seewhether they were associated with solar activ-ity and if so, how they varied with it.

The critical frequencies — foFe, foF1 andfoF2 — of the ionosphere are a case in point.With the introduction of the ionosonde, thevertical structure of the ionosphere was ex-plored and daily variations of the critical fre-quencies soon found. But those data were nottaken on a continuous basis; that would havepresented a huge problem, just in the analysisof the data alone. So continuous monitoringwas never really attempted and periodic sam-pling used instead, the critical frequencies ob-served every hour or sometimes more fre-quently, depending on the need. But that meantthere were gaps in the records of the criticalfrequencies, nothing really known betweensamples or, put another way, nothing knownon how the critical frequencies varied at timescales shorter than the sampling period.

If critical frequencies of the ionospherevaried little, say from day to day at a given

hour, the gaps in the data record would be oflittle concern and no major causative agentneed be found. But that proved not to be thecase so there was little recourse at the time butcontinue with sampling, compiling a datarecord of critical frequencies and their limits.Causative agents would have to wait untilmore was known about how the ionosphereworked. Nowadays, we have big words andphrases for that sort of undertaking, buildinga database, and the contents of the databaseare described with the language of statistics.

And therein lies a problem as manypeople are not familiar with that language norwhat it means. True, there seems to be a uni-versal feeling for an average, like a batting av-erage in baseball, but outside of that context,there seems to be little appreciation of whatgoes into an averaging process or other mea-sures of performance, especially what they sug-gest about the range of variability for futureevents based on past performance.

There’s a worry here and to see what it isfor the ionospheric situation, consider the geo-magnetic data shown earlier in Figure 19.2.People might take those results for propaga-tion disruptions due to magnetic storminessduring Cycle 22 and expect the same to applyin Cycle 23. That would not be unreasonable ifthey took the larger view and accepted the ideathat variations about the average are all partof the picture. But to expect that HF propaga-tion will be the same from one three-monthperiod (the sampling interval) to the next, onlydisrupted by magnetic storms about 15% of thetime, is unrealistic considering the databasefrom which the result was originally derived.There were variations about the average,storminess for as much as 31% or as little as5% of the time in the three-month periods thatmade up the record. Those have to be reckonedwith.

Critical frequencies foF2 and MUFs are

21: STATISTICAL CONSIDERATIONS


Figure 21.1 Typical foF2 values for a given hour in the 30 days of a month.

other cases in point, discussed earlier in con-nection with Figure 12.5. To illustrate that ideaagain, consider the data in Figure 21.1, foF2values measured at a given hour in the 30 daysof a month. The data make up a fairly simpledistribution of foF2 values and its median(50%) value is 7.85 MHz. That means that halfthe data points lie above that value and half liebelow it. If that data were used in connectionwith propagation along a path and the obliquegeometry of the path called for the verticalvalue foF2 to be increased by a factor of 3.5,the median value of the maximum useable fre-quency (MUF) would be 27.5 MHz. Thus, 50%of the time, the MUF for the path would beabove 27.5 MHz and 50% of the time it wouldbe below 27.5 MHz.

The two decile values, marking the topand bottom 10% of the foF2 distribution, are8.7 MHz, the upper decile value, and 7.0 MHz,the lower decile value. On that basis, 10% ofthe time the maximum useable frequency forthe path would be 30.5 MHz or greater and 10%of the time the maximum useable frequencywould be 24.5 MHz or lower in frequency. An-other way of saying the same thing might bethat for 80% of the time, the maximum useablefrequency for the path would lie above 24.5MHz and below 30.5 MHz.

All that seems fairly straightforward as

long as one has the data in sight or in mind.But when it comes to MUF calculations with apropagation program, the Little Pistol does notsee the data, much less even think that therewas a distribution of foF2 values in the data-base which the computer uses to predict theMUF value. And therein lies a problem, eventwo or three.

Taking first things first, say the LittlePistol’s propagation program came up with anMUF of 15 MHz for a path to Timbuktu and tohis great disappointment, he found the pathwas closed. What’s the problem? If the solar/geophysical scene were quiet, it’d probably justbe a case of statistics, the LP not coming to gripswith the fact that MUF values are not alwaysthe same and have some statistical variation oftheir own. So chances were that his disappoint-ment was due to the fact that he tried to oper-ate at a time when the real time foF2 valuealong the path to Timbuktu was significantlybelow the median value in the propagationprogram’s database.

But there’s a nagging question — just howdid the Little Pistol come up with the MUFvalue that was so disappointing? Was his MUFcalculation on firm ground or maybe, justmaybe, the Little Pistol got a questionable re-sult as he took a 10.7-cm flux value from thedaily WWV broadcast and put that in his


propagation program. That’d be a silly thingto do as propagation programs are supposedto convert smoothed 10.7-cm flux values to cor-responding smoothed sunspot numbers beforemaking any calculations.

If LP thought about it, a smoothed valueinvolves 13 months of data so the use of asingle, daily value of 10.7-cm flux would be avery poor substitute for a smoothed averageof 395 separate values. So maybe the LP wasdisappointed for two reasons, the first for notunderstanding that any prediction is foundedon data with a statistical spread in values, andsecond, and possibly more importantly, thatindividual data entries will never give reliableresults when smoothed data input is called for.

If the Little Pistol made that mistake, onewonders what prompted him to do it in thefirst place, a change in the 10.7-cm flux valuefrom one day to the next? But as discussed backin the section on solar data, the 10.7-cm radia-tion is far from energetic enough itself to affectthe ionosphere; it’s merely an indicator of ac-tive regions on the solar disk so any smallchanges are of no particular importance. Butlarge changes have meaning, say those shownin Figure 5.1, showing the appearance and dis-appearance of active regions or real hot spotson the solar disk.

So propagation conditions become a mat-ter of degree, the general level of the solar ac-tivity and the sort of conditions to expect be-ing obtained from smoothed indicators, butafter that the Little Pistol should think in termsof statistical fluctuations rather than lookingfor changes in propagation due to any short-term variations.

Looking back now to the beginning of thissection and the start of the discussion of statis-tical aspects of propagation, you’ll notice it allstarted in connection with magnetic distur-bances during Cycle 22. That’s not exactly bignews as it’s been known since back in 1937.However, only in the last thirty years or so hasit been understood how magnetic activity caninfluence propagation through changes in theelectron populations held on field lines, par-ticularly at the higher latitudes. That’s one ofthe major contributions to propagation stud-

ies from the plasma-and-fields era, when thenew model of the geomagnetic field came onthe scene and took on an important role in un-derstanding more of the relationship betweenthe dynamics of the field and the ionosphere.

The rest of the statistical discussion givenabove dealt with propagation just as if it werestill a matter from the photochemical era,speaking of variations in the critical frequen-cies foF2 as though changes in solar UV, fromactive regions coming and going across thesolar disk, were all that mattered. As for theother important side of ionospheric matters, thegeomagnetic field, while recognized as a caus-ative agent, it is not treated as one nor giventhe status of a parameter in the database.

That’s another way of saying that the so-lar UV database remains intact and individualcritical frequencies used in MUF predictions arenot modified because of magnetic activity,whether forecast or in progress. Even if an at-tempt were made to include it, one of the prob-lems with measures of magnetic activity is thatthey vary to some extent according to latitudeand longitude. Thus, variations of the magneticelements are much greater in magnitude andmore frequent at auroral zone latitudes thanmiddle latitudes. And their algebraic signs maydiffer, negative excursions in the horizontalcomponent at auroral latitudes at the same timeas positive changes at middle latitudes. Thus,while it would be desirable to add magneticactivity as a variable in MUF calculations, theactual choice to be used is not clear; as a resultthe foF2 databases currently used in MUF workare essentially those for quiet magnetic condi-tions.

So the scheme largely used now in HFpropagation matters is to consider any signifi-cant level of magnetic activity as a threat topropagation, affecting it in minor ways untilthe K-index reaches 4. Any increase in K-indexbeyond that takes on more serious proportions,depending on whether a minor or major stormis in the offing. For minor storms, propagationwould probably be affected adversely for a dayor so on the highest HF bands, 21 and 28 MHz,but a major storm could take out the higherbands completely on all but transequatorial


paths for the better part of a week and evenhave an impact on the 7-MHz band. That be-ing the case, alternatives are sought by trans-ferring activity to the lower bands until themagnetic activity has subsided.

Harking back to the control-point methodof calculating MUF values, the threat or possi-bility of magnetic activity affecting propaga-tion can be added, after the fact, by using ameasure like the K-index from a nearby obser-vatory to adjust an overall MUF value. In es-sence, that approach would add magnetic ac-tivity to propagation considerations, not di-rectly as a parameter like sunspot number butas a correction factor, the magnitude approach-ing a local veto for situations of storm propor-tion.

Presently, that approach is used by themost recent versions of the MINIPROP PLUSand CAPMAN programs to modify the MUFfor a path, stations in the USA using the K-in-dex from the magnetic observatory at Boulder,CO, as broadcast by WWV. Since the K-indexchanges every three hours, the method isstrictly short term when it comes to anychanges in MUF predictions. Unfortunately, thedetails of the method are not well known asthe original publication was in a laboratorytechnical report rather than in the scientific lit-erature.

There is another physical variable whichplays a role in HF propagation, the electric fieldE. In our everyday experience, we are awareof electric fields from meteorological circum-stances, static discharges during dry weatheror thunderstorm activity. In that setting, chargeseparation can give rise to potential differencesgreater than 100,000 Volts. The situation is dif-ferent in the case of the magnetosphere as itresembles a dynamo, where a moving conduc-tor goes across a magnetic field and drives acurrent by virtue of its motion. For the earth’smagnetosphere, the conducting material isplasma coming from the sun and it movesacross the earth’s field, giving rise to a poten-tial difference the order of 300,000 Volts acrossthe magnetosphere. In turn, that drives elec-trical currents which circulate and dissipateenergy within the magnetosphere.

While magnetospheric electric fields aremapped along magnetic field lines holdinghighly conducting material, such fields are notstrong enough to be detected near ground levelin the presence of fields of meteorological ori-gin. But they are not without their influence,the most notable case being electron bombard-ment of the atmosphere at auroral latitudes,producing visual displays showing the particleinflux and significant ionization above the 100-km level. While not fully understood at themoment, auroral electrons apparently are ac-celerated by electric fields parallel to geomag-netic field lines, taking them from low ener-gies the order of 1 eV up to tens of keV.

As far as propagation is concerned, theobvious effect would be signal absorption as aresult of the additional ionization deposited bythe incoming electrons. Since magnetosphericelectric fields are not found at ground level,they can hardly be included as parameters inmaking primary MUF calculations; instead,they play a secondary role in propagation byincreasing the absorption of signals goingacross the auroral zones.

As for a statistical side, there are variationsin the occurrence of aurora, the latitude limitsof the auroral zone with local time, magneticactivity and the phase of a solar cycle. In thatregard, the polar cap is inside the auroral ovaland is in the range of latitudes where protonbombardment is fairly uniform during PCAevents. Across auroral latitudes, there is a falloffof D-region ionization during PCAs and exceptfor severe geomagnetic disturbances, it doesnot reach into the middle latitudes.

Polar cap absorption (PCA) is anotherphenomenon that plays a secondary role in HFpropagation. Before getting to its statisticalside, we should note that the origin of ener-getic solar protons is on the sun but when theyreach the earth’s orbit, their impact on the at-mosphere is determined by the state of the geo-magnetic field at the time. For example, if alarge solar flare accelerates protons to the MeVenergy range and they cross interplanetaryspace to finally arrive at the earth, they areguided, according to energy or momentum, bythe earth’s magnetic field. If the field were quiet


when the first protons arrived, their impactwould be inside a small auroral oval. But later,with a magnetic storm of some severity thenin progress, the incoming protons would finda polar cap that’s open wider and thus affectthe D-region at lower latitudes than before.

Statistically PCA events are not numerous,an event or so per month at solar maximum,but AA events are found frequently duringmagnetic storms throughout a solar cycle. Boththose particle influxes reach the atmosphereand give rise to ionospheric absorption forpassing signals but there are differences, pro-tons events going deeper into the atmosphereand producing larger absorption effects, in tensof dB in the HF range, and infrequent in com-parison to AA events which produce only a fewdB of absorption at the same frequency.

And transpolar paths are rarely able to goacross the entire polar cap without a groundreflection and thus, during a PCA event, sig-nal strength may suffer severely because of theabsorption along the path coming down to areflection and rising again through the D-re-gion. That is not the case for AA events as theabsorbing region is more narrow and ring-likeat auroral latitudes so it is possible for a pathto skip over it, the signal being at F-region alti-tudes in crossing the auroral zone, thus escap-ing the absorption effects at E-region altitudes.

All in all, there’s little reason to think PCAevents could be factored into the primarypropagation calculations; as a result, they aremore like threats to propagation, if predicted,and a grim reality if actually present. Auroralevents are less of a problem from the absorp-tion standpoint but may occur during magneticactivity and thus affect critical frequenciesrather than signal strength.

The one bright side of PCA events for HFpropagation is that they show day/night ef-fects in the ionospheric absorption they pro-duce. That has to do with the fact that, in dark-ness, ionospheric electrons form negative ionswith oxygen molecules and because of having

a greater mass, they are far less effective in ab-sorption processes. By way of contrast, auroralabsorption does not show a day/night effectas negative ion formation is much less prob-able above the 100 km level where auroral ion-ization is deposited.

Everything considered, the two forms ofparticle influx into the high-latitude ionosphereresult in absorption effects, PCA events last-ing for days at a time, covering the polar cap,and offer a major obstacle to propagation acrosshigh latitude regions. On the other hand, AAevents are more frequent but rather restrictedin latitude and duration, lasting for hours, andgiving rise to absorption events of about 1/4the magnitude of that for an average PCAevent.

While auroral absorption events, originat-ing within the magnetosphere, are found withcomparable absorption levels at conjugate re-gions in the two hemispheres, solar protonswhich produce PCA events originate from out-side the magnetosphere and do not enjoy equalaccess to the two polar regions. So, in addition,along with their day-night effects, they have adifferent impact on propagation in the twohemispheres.

By way of example, consider the PCAevent of 11 June 1991. It was a major event andproduced some 17 dB of absorption of cosmicradio noise on 30 MHz at Thule, Greenland butonly 1.5-dB absorption at the Amundsen-ScottBase (South Pole). At the time, Thule was infull sunlight while the South Pole was in dark-ness. The difference in D-region absorptionmay be due to differences in access of solar pro-tons to the polar caps via the magnetic field aswell as the day/night effect mentioned above.In any event, the difference in absorption be-tween the two hemispheres is striking and, ifnothing else intervened, the southern polar capmight have been workable for HF propagationbut paths going across the northern polar capwere definitely closed by the PCA event.


The discussion so far has pointed to thefactors which affect propagation, the magni-tude and variability of their effects. Of neces-sity, it has been a piecemeal approach, one itemdiscussed at a time. At this point, let’s standback and look at the big picture, as shown inFigure 22.1. Of course, that amounts to some-thing like a snapshot of the solar/terrestrial en-vironment and for that reason, it gives only atwo-dimensional representation of the situa-tion and as a snapshot, it only represents onemoment in time.

The Little Pistol, on viewing that figure,might be intimidated, thinking that all theforces of Nature are in conspiracy against himin his quest for DX. But those forces are sel-dom out all at one time. In fact, it took years ofwork to develop that picture and DXing wasgoing strong all the time. Moreover, it wasfound that the cast of players was constantly

storminess, shown in Figure 19.2, suggestedthat not only would propagation be good atthe peak of a solar cycle, like it was in Cycle22, but also relatively undisturbed, on average,something like 85% of the time. That might bea bit too optimistic, maybe 75% is a more con-servative figure to work with, but it representsa starting point in singling out the cast of play-ers on scene during reasonable propagationconditions.

So for those times and circumstances, amore appropriate version of Figure 22.1 wouldnot show either energetic particles and burstsof X-rays. As a matter of fact, a version of thatfigure based on the understanding of propa-gation before WWII would only show the sun,emitting UV radiation, and an atmosphere im-mersed in a dipole field out at the earth’s or-bit. The rest of the cast came on stage later.

With time the solar wind and the inter-

22: SOLAR/TERRESTRIAL ENVIRONMENT

Figure 22.1 Solar/terrestrial environment. From Rosenthal and Hirman [1990]

changing, a burst of X-rays this day or a blastin the solar wind that day, and only in direst ofcircumstances would all those forces be arrayedat once. So even though the Little Pistol knowswhat lurks out there, he shouldn’t approachDXing with any fear or trepidation.

For example, the discussion of magnetic

planetary field were found to be constantly “inthe act” but with a more complicated geom-etry than the spirals shown in Figure 22.1. Inthat regard, the interplanetary field has its ori-gin on the sun and is dragged out by the solarwind, as it evaporates from the sun’s corona.Thus, field lines extend out on one side of an


equatorial surface and return to the sun on theother side, like a simple dipole field.

But with the 27-day solar rotation andvariations in the pressure of the solar wind, theequatorial surface is not flat but more like aballerina’s skirt during a pirouette. On that ba-sis, the vector direction of the interplanetaryfield observed at the earth’s orbit depends onwhether the earth is above or below the equa-torial surface as well as the undulations orwaviness of field lines and the surface, asshown in Figure 22.2.

In any event, in that scenario, the pressureof the solar wind compresses the geomagneticfield and the level of ionization in the iono-sphere is determined by the solar UV reachingthe earth’s orbit. The electrons released in theionosphere are held on magnetic field lines andtheir spatial organization depends on thespread of UV across the sunlit hemisphere anddetails of the earth’s magnetic field.

For ionospheric purposes, solar UV is glo-bal in scope and the primary variable to be con-cerned about, with activity or the average mag-nitude of its effects described by the smoothedsunspot number. And for something like 75%

So even though having a global coverage andcapable of giving rise to rather dramatic effects,the fact that they are so transient relegates themto a minor status, not even as a secondary vari-able in radio propagation.

The effects of energetic particles, namelysolar protons, have been known for a long time,even before WWII when their terrestrial effectswere termed “polar radio blackouts.” Whilethe energies of solar X-rays are in the keV rangeand solar protons in the MeV range, they bothpenetrate as far as the D-region where iono-spheric absorption efficiency is the greatest.The X-rays suffer exponential attenuation bythe atmosphere in reaching the D-region whilesolar protons are lost mainly because of ioniz-ing collisions with atoms and molecules alongtheir range in the atmosphere.

The effects of energetic particles, PCAevents, are not global in scope, important onlyfor propagation paths across the polar caps. Butwhile infrequent, the fact their effects may belarge in magnitude, reducing HF signals bymany dB, and of considerable duration, saydays at a time, makes them something to bereckoned with when it comes to propagation,

Figure 22.2 Interplanetary field lines near the earth’s orbit.

of the time, HF communications could be dealtwith effectively just by using the smoothedsunspot number, in conjunction with a goodfoF2 database, in a propagation program.

Other players of solar origin may comeon stage from time to time. Those would in-clude bursts of solar X-rays and energetic pro-tons, both in connection with the flare processwhich takes place above active regions. Theflare X-rays are global in scope, just like solarUV, but since the additional ionization theyproduce then disappears quickly when the flareends, they are only bit players in the drama.

but only at the status of a secondary factor.In Figure 22.1, the spiral trajectories of the

energetic particles about interplanetary fieldlines are shown in projection. The size of thespirals depends on the momentum or energyof the particles, less energetic particles follow-ing field lines closely. Those circumstances rep-resent possible motions when the interplan-etary field is organized or well ordered. Dur-ing some PCA events, energetic particles arrivepromptly at the earth after a flare. That usu-ally takes place when the flare site is locatedwest of the central meridian and the earth is


well connected to the sun by the interplanetaryfield.

On other occasions, when flare sites areoff to the east of central meridian, the flux ofenergetic particles at the earth’s orbit increasesmore slowly and may give rise to terrestrialeffects of longer duration. Those circumstancessuggest interplanetary field lines have a wavi-ness or small-scale structure to them and thatsome energetic particles diffuse or staggeracross the field lines. Either way, energetic par-ticles coming from the sun have to be reckonedwith but at the secondary disturbance level, notas primary factors in HF propagation.

At this point, it should be noted that bothsolar X-ray bursts and proton events involveenergetic radiation, in the keV and MeV energyrange, only minor magnetic effects, if any, re-sult from the ionization created. That is in sharpcontrast with geomagnetic disturbances whichresult from low energy solar plasma, in the eVenergy range, reaching the earth’s orbit fromfast streams out of coronal holes or plasmaemerging from coronal mass ejections. The so-lar plasma from those origins arrives at theearth over a broad region, 10-15 Re across, andmay represent a sudden, large power input tothe magnetosphere. The ensuing interactionswill then start a magnetic storm, modifying theconfiguration of field lines throughout themagnetosphere.

But the main effect as far as propagationis concerned is ionospheric storming, from criti-cal frequencies foF2 falling sharply because ofF-region ionization being carried away into themagnetotail. Another effect would be the fur-ther opening of the polar cap, giving greateraccess to any solar protons coming from thesun and thus expanding the region where aPCA event is in effect. Of course there is alsothe matter of energization of magnetosphericelectrons to auroral energies, up to tens of keV,and their less frequent effects at E-region alti-tudes, ionization which results in additionalrefraction and absorption of signals.

All in all, the effects which accompany amagnetic storm are the most important as faras propagation is concerned. They are majoreffects in that they disrupt the ionosphere at

the F-region altitudes instead of down at D-re-gion altitudes, as with solar X-rays or energeticprotons. From an operator’s standpoint, mag-netic storms are a threat to propagation andrepresent something that must be accepted andworked around when they occur.

Finally, to pull all the discussion together,the Little Pistol would help his cause, DXingto greater fame and glory, by not only beingkeenly aware of the current solar/terrestrialenvironment but also following the predictionsof professional forecasters, say at NOAA, whohave a wide range of current observations todraw on. While the numbers from WWVbroadcasts, or log charts from them, make goodconversation, they really won’t help muchwithout additional input. So the Little Pistolneeds to know where he is in the solar cycle,something of the recent solar and geophysicalactivity, even about disturbances, and what ispredicted for the immediate future.

A first step in that direction would be tofollow the information published by NOAA/SESC in the weekly Boulder Report. That doesinclude weekly updates of the twenty-sevenday outlook for magnetic activity, an impor-tant item, but more often than not, the reportis held up in the mail for several days, reduc-ing the immediate value of that information.Of course the narrative text, giving highlightsand forecasts of solar and geomagnetic activ-ity, is important and valuable, too, especiallyany mention of effects from coronal holes orcoronal mass ejections.

But the very best thing for the Little Pis-tol to do is check in with the NOAA PBBS anddownload both the Solar and Propagation Re-ports. As mentioned earlier, it is relatively in-expensive to obtain all that data on a daily ba-sis and that makes for more effective planningwhen it comes to DXing, either new ones orDXpeditions coming on station.

Of course a full-service propagation pro-gram is a must for primary planning purposes.The information from Boulder will give thesecondary information needed to anticipatemagnetic disturbances. A graphic aid, with amap display and terminator, is of great help,especially in seeing where a path goes and if it


would be in harm’s way during any type ofsolar/terrestrial disturbance.

With the above approach, the Little Pistolwill surely advance on the ladder toward theDXCC Honor Roll, hopefully at a more rapidpace from his better understanding of the so-lar/terrestrial environment and with all theaids at his disposal. But the Little Pistol shouldnot forget that solar activity and all that goeswith it are renewable resources, coming backagain every 11 years. We don’t even have torecycle solar activity; it cycles itself, again andagain. If the Little Pistol wants to play thatgame, it’s in his best interest to understand howit works and then play with all the skill he can

muster. But the good part is that if he doesn’tmake the DXCC Honor Roll during this solarcycle, there’ll surely be another one before toolong and he can pick up where he left off.

So the fact that the sun renews itself and,in the process, it can reinvigorate us is excitingto contemplate. I’d hope that the Little Pistolwould see the marvel and mystery in it all.More than anything else, however, I’d hopethat he understands DXing is an intellectualpursuit, worth thinking hard about and work-ing at just for the pure joy of it. If he does that,all the time and effort I’ve put in this book willbe worth it.


23: HF PROPAGATION IN A NUTSHELL

This chapter is for those readers who wantto hear the “answer” before thinking about the“question.” Either they’re devoted to detec-tive stories or the popular “answer and ques-tion show” on TV, Jeopardy. So in what fol-lows will be a summary, terse statements with-out too much in the way of explanation, justwhat’s important and how propagation reallyworks. It can be read with amusement, puzzle-ment and profit by all.

But first, we have to lay the groundwork.So to begin, it all goes back to the sun emittingultraviolet light and X-rays while spewing outstreams of ionized matter, largely protons andelectrons. The UV and X-rays make it to Earthin 500 seconds flight time and come more orless continuously while the streams of chargedmatter, solar plasma, may vary considerablyin their flow but travel at about 400 km/secinstead of 300,000 km/sec, as is the case for theUV and X-rays. And all that, solar photons andplasma, reach earth at a distance of 150,000,000km from the sun.

The earth is one of the terrestrial planets,with an atmosphere and a magnetic field. Theatmosphere is well mixed up to an altitude ofabout 90 km, consisting mainly of nitrogen andoxygen molecules. While solar X-rays penetratebelow 90 km and ionize some of those mol-ecules, at higher altitudes solar UV modifiesthe chemical nature of the atmosphere by pho-todissociation. Thus, oxygen atoms begin toappear above 90 km from the dissociation ofoxygen molecules and solar UV ionizes themas well as the oxygen and nitrogen molecules.The positive ions and electrons released in thatfashion go to make the ionosphere.

The flux of ionizing radiation from the sunvaries slowly with solar activity and ionizationof the earth’s atmosphere takes place continu-ously, the distribution of ionizing processesover the sunlit portion of the earth essentiallyfixed along the earth/sun line. But atmosphereclose to the earth, being thin and without tre-

mendous inertia like the oceans, co-rotates withthe earth. As a result, fresh atmospheric mate-rial rotates into solar view at dawn, then is car-ried past the noon meridian and finally out ofsolar view at dusk.

The electron density which results at iono-spheric heights is due to a competition betweenproduction, mainly from photoionization bysolar UV, and loss of ionization, by recombina-tion of electrons with positive ions and trans-port processes. The electron density showssome structure from chemical changes withincreasing altitude, distinct regions and finallya peak, before disappearing slowly at great al-titudes.

The D-region is the lowest, having an elec-tron density on the order of a thousand elec-trons per cubic centimeter in the range 60-90km during the day and vanishing at night. TheE-region is something like a ledge, with an elec-tron density on the order of 50,000 electronsper cubic centimeter during daytime and re-maining fairly constant for several kilometersaround 110 km height. This is the region whereatomic oxygen and its positive ion start to be-come important. The F-region, on the otherhand, up around 300 km, contains on the or-der of a million electrons per cubic centimeterand decays slowly at night.

The spatial distribution of ionization iscontrolled by the earth’s magnetic field as elec-trons, on being released by solar UV, are con-strained in their motion, gyrating around thefield lines instead of flying off freely on ballis-tic trajectories. Thus, ionospheric data is betterorganized in terms of geomagnetic coordinatesthan geographic coordinates. In addition, thereare significant features of geomagnetic originin the global distribution of electrons over theionosphere as well as temporary distortions ofthe distribution from the effects of geomagneticstorms.

That’s a fairly concise summary of whatthe quiet ionosphere is like and how it is being


created constantly by solar UV. Just how it re-ally works and what disturbs it are other sub-jects. Unfortunately, not everybody in AmateurRadio is interested in those matters, being will-ing to bumble along without much understand-ing. In some ways they’re like joggers who suitup, go to the front door and dash out into theelements, not even paying attention to the sea-son or the weather. Of course there are betterways of doing things and that’s what this dis-cussion is all about.

First it should be understood that the iono-sphere is a three-dimensional distribution ofionization, electrons and positive ions. At theoutset the details of how the ionization is dis-tributed are really not so important for us,merely that it is possible to map it out acrossthe globe and then use those ionospheric mapsas well as some basic knowledge to under-stand, even predict, propagation.

On that subject, radio propagation is noth-ing but the paths followed by electromagneticwaves that are trapped below the peak of theionosphere. But with high enough frequencyor launch angle above the horizon, radio wavescan penetrate the F-region peak and passthrough the top side of the F-region. That canlead to positive results when a satellite is onthe receiving end, or a negative result, whenthe radio waves are simply lost into space andnothing is accomplished.

During daytime the lowest parts of theionosphere (D- and E-region) limit propaga-tion of waves with frequencies below 10 MHzto local contacts and those bands are usedthen largely for ragchewing and local emer-gency traffic. In short, the lower bands arenot very useful for DXing when paths aresunlit.

In the absence of solar illumination, theD- and E-regions disappear and nighttimepropagation on the lower bands, say 3.5 MHzand 7 MHz is limited primarily by band con-gestion and noise, of atmospheric origin orfrom man-made sources. Put another way,even at solar minimum there will be enoughionization in the F-region to support DXingon those bands as long as no portion of a pathis in sunlight. But the ugly villain, noise,

propagates across dark paths, too, and mustbe reckoned with.

At the top of the HF spectrum, say on the24-MHz or 28-MHz bands, propagation de-pends more on the level of solar activity asionospheric absorption is much less there, vary-ing as the inverse square of the frequency. As aresult, at high levels of solar activity, the pur-suit of DX on those frequencies is quite feasibleon paths which are sunlit and then it becomesa question of finding when the MUF on a pathis above the chosen operating frequency. Andin going from 3.5 MHz to 28 MHz, man-madenoise falls by 25 dB so that is less of a factorthan on the lowest band. But at low levels ofsolar activity, toward solar minimum, DXingon those bands is limited largely to low-lati-tude and transequatorial paths.

The middle of the HF spectrum, say from10 MHz to 21 MHz, is what might be termed a“transition region” as all three factors whichmake for success in DXing —open bands,strong signals, weak noise —are at risk. Inshort, for frequencies in the middle of the HFspectrum, DX paths must be less exposed tosolar illumination and that means in the courseof a day, MUFs on paths of interest may fallbelow the operating frequency more often. Ofcourse, noise is always a concern, at one end ofa path or the other.

The 14-MHz and 21-MHz bands seem tohave the greatest activity when it comes toDXing, especially before and after solar maxi-mum, and support the full range of paths acrossthe globe. Given the degree of congestion onthose bands, success in DXing requires thatoperating times be chosen carefully, especiallywith regard to solar/terrestrial conditions andthe sociological factors which make the bandsmore congested.

All of that discussion was more specificas far as how the ionosphere works but ratherqualitative in other respects, just a few of therelevant quantities given, say ionosphericheights and some typical electron densities.And nothing was mentioned as to how it wouldbe disturbed although it should come as nosurprise that the sun will be tracked down ul-timately as the origin of the disturbances. That


makes for an interesting situation where thesource of the ionosphere and its disturbingagent are one and the same.

In proceeding in the discussion of how theionosphere is formed and disturbed (or de-formed), questions and distinctions involvingenergy come up — first, solar UV photons withabout 10 electron-Volts or 10 eV of energy cre-ating the ionosphere by ionizing the atoms andmolecules in the atmosphere and then distur-bances of the ionosphere resulting from burstsof solar X-rays with energies around 1,000 eV,(1 keV), or solar protons with energies greaterthan 1,000,000 eV, (1 MeV).

But instead of those energetic bursts be-ing the most frequent factors in ionosphericdisturbances, it turns out the puffs and blastsof plasma in the solar wind give rise to thegreatest problems in propagation. On thoseoccasions, the dynamic interaction of solarplasma with the geomagnetic field reshapes thefield, effectively removing ionization from thelower ionosphere and then inhibiting propa-gation. In terms of energy, the power input ofsolar plasma to the earth is huge and leads to amechanism, still not fully understood, bywhich electrons of solar origin, with about 1eV in energy, are raised to energies like 10 keVin auroral displays, all within the confines ofthe geomagnetic field.

While propagation can be characterizedas good or bad, with all shades in between, itis better described as being normal (N), fair (U)or poor (W), as done in Propagation Reportsfrom the NOAA PBBS. At this QTH, the lasttwo abbreviations are taken to mean “Ugly”and “Wretched,” typical of conditions foundduring minor and major magnetic storms, re-spectively. But for practical purposes, the ques-tion is not so much about how bad propaga-tion conditions can become but how often arethey in bad shape or how long the disturbanceswill last. In that regard, the good news is thatnormal propagation conditions can be expected75%-85% of the time. But there are variationswithin events, some major storms disruptingthe upper bands for days at a time while mi-nor storms can be brief, lasting a day or less.

It’s in those normal, calmer periods that

most amateurs do their DXing and the prob-lem then becomes one of finding the most effi-cient or effective way of going about it. But,residing in the USA, even a casual glance at anazimuthal equidistant map should make itclear to any operator that a good part of thetime DXing will involve trying to get signalsacross the polar cap to DX stations on the otherside of the globe.

Given that, whether one can chase DX ona given day or not depends on the state of af-fairs at high magnetic latitudes. So a weathereye should be cocked in the direction of the sun,watching for signs of trouble. Light, travelingfaster than solar plasma, can give optical sig-natures which signal or warn of solar distur-bances to come. Knowing what to expect andwhen to expect it turn out to be factors whichcontribute to greater effectiveness in amateuroperations, in general, and DXing, in particu-lar, minimizing time wasted on dead or dyingbands. That’s where the information in the So-lar and Propagation Reports from NOAA PBBSis extremely helpful.

Now having said that about avoidingtimes of disturbance, the question becomeshow to make the most of normal conditions.For that operations must be supported by sev-eral things: (1) a global database for critical fre-quencies, by smoothed sunspot number, sea-sons and time, (2) a full-service computer pro-gram to use the database, giving informationon MUFs, signal strengths (easily converted toS-Units) and noise, atmospheric or manmade,(3) a source of reliable data and current infor-mation on solar/terrestrial conditions, and (4)an understanding of the various aspects of HFpropagation, say paths across the polar regionsor the geomagnetic equator as well as long-pathpropagation.

The first two supporting items are fairlyeasy to come by and the third one is a matterof choice, either basing one’s operations ona minimal input of information and trustingto luck or using the options available fromNOAA to make a more refined approach topropagation and DXing. In any event, it isimportant to have an awareness of the gen-eral level of solar activity and recent solar/


terrestrial events and conditions.The last item on that list takes time and is

obtained largely by on-the-air experience,supplemented by output from the propagationprogram in use. Thus, everyone has some feel-ing when the JAs are coming through but whatabout the JTs? That’ll take a bit of study, look-ing for times when those signals would be read-able, over the local noise level.

And then there’s the matter of sharp-shooting, getting ready for a rare one like the3Y0PI operation in February ’94. There areseveral ways of preparing for operations likethat, DX well off the beaten track. First there’sthe matter of beam heading. That’s easy. Thenthere’s the matter of the MUF, also known as“mode availability.” That can be obtainedfrom knowledge of the E- and F-regions buthas nothing to do with signal strength or sig-nal/noise ratios. Both those are determined,in part, by absorption processes in the D-re-gion and can be obtained by calculations of“path reliability.” That’s interesting but to

plan times for operation would require go-ing back and forth between them, availabil-ity and reliability, keeping track of which ishigh and which is low, etc. It’s easier to workout a new quantity that’s termed “DX feasi-bility,” which simply combines the two prob-abilities into one; from that probability, a planof action becomes pretty obvious. Of course,a quick review of mob psychology wouldhelp too.

Since the possibilities are infinite when itcomes to DX and propagation, this discussioncould go on forever. But one thing would bemissing, figures or curves which illustrate theideas and carry far greater meaning than just afew words. In that regard, the chapters of thisbook contain more than 80 figures. So maybeit’s time to add music to the words of thistheme, starting at the beginning and workingthrough the fuller discussion that has been puttogether for the benefit of your friend and mine,the Little Pistol. So turn to it!


Akasofu, S.I., Dynamics of the Magneto-sphere, Proceedings of the ChapmanConference, D. Reidel Publishing Co.,Dordrect, Holland, 1978.

Anari, Z.A., The Aurorally AssociatedAbsorption of Cosmic Radio Noise atCollege, Alaska, J. Geophys. Res. Vol. 69,p. 4493, 1964.

ARRL Handbook for Radio Amateurs,American Radio Relay League,Newington, CT, 1994.

ASAPS, Advanced Stand Alone PredictionSystem, IPS Radio and Space Services,Australia, 1992.

Basler, R., Radio Wave Propagation in theAuroral Ionosphere, J. Geophys. Res.Vol. 68, p. 4665, 1963.

Brown, R.R., Long-Path Propagation, AStudy of Long-Path Propagation in SolarCycle 22, Anacortes, WA, 1992.

Brown, R.R. and J.R. Barcus, Day-NightRatio for Auroral Absorption EventsAssociated with Negative MagneticBays, J. Geophys. Res. Vol. 68, p. 4175,1963.

Chapman, S. and J. Bartels, Geomag-netism, Vols. 1 and 2, Oxford UniversityPress, New York, 1940.

Davies, K., Ionospheric Radio Propaga-tion, U.S. Department of Commerce U.S.National Bureau of Standards, Mono-graph 80, Washington, D.C., 1965.

Davies, K., Ionospheric Radio, PeterPerigrinus Ltd., London, 1990.

Davies, K. and C.M. Rush, High-Fre-quency Ray Paths for IonosphericLayers with Horizontal Gradients,Radio Sci., 20, p. 95-110, 1985.

Fraser-Smith, A.C., Centered; and Eccen-tric Geomagnetic Dipoles, 1600-1985, J.Geophys. Res. Vol. 25, No. 1, p. 1-16,1987.

Fricker, R., H.F. Propagation, A Microcom-puter Method for Predicting the FieldStrength of H.F. Broadcast Transmis-sions, BBC External Services, Engineer-ing, London, England, 1988.

Handwerker, J., IONSOUND, An Iono-spheric Propagation Prediction Programfor IBM PC’s and Compatibles, Lexing-ton, MA, 1994.

Leftin, M., Ionospheric Predictions, Tele-communications Research and Engineer-ing Report, Vol. 1-4, Institute of Tele-communications Sciences, Boulder, CO,1971.

Leinbach, H. and R.P. Basler, IonosphericAbsorption of Cosmic Radio Noise atMagnetically Conjugate Auroral ZoneLocations, J. Geophys. Res. Vol. 68, p.3375, 1963.

Lucas, D., CAPMAN, Computer AssistedPrediction Manager, Boulder, CO, 1994.

McNamara, L. F., Radio Amateurs Guideto the Ionosphere, Krieger PublishingCo., Melbourne, FL, 1994.

Norton, R.B., The Middle-Latitude F-region during Some Severe IonosphericStorms, Proc. IEEE, Vol. 57, p. 1147, 1969.

REFERENCES


Oler, C., SKYCOM, High FrequencyIonospheric Signal Analyst, Solar Terres-trial Dispatch, Stirling, Alberta, Canada,1994.

Piggott, W.R. and K. Rawer, URSI Hand-book of Ionogram Interpretation andReduction, Report UAG-50, World DataCenter A for Solar-Terrestrial Physics,Boulder, CO, 1975.

Rosenthal, D.A. and J.W. Hirman, ARadio-Frequency Users Guide to theSpace Environment Services Geophysi-cal Alert Broadcasts, NOAA TechnicalMemorandum, ERL SEL-80, SpaceEnvironment Laboratory, Boulder, CO,June 1990.

Shallon, S., MINIPROP, An HF Propaga-tion Prediction Program, Los Angeles,CA, 1994.

Snyder, W.F. and C.L. Bragaw, Achieve-ment in Radio, Seventy Years of Radio

Science, Technology, Standards andMeasurement at the National Bureau ofStandards, National Bureau of Stan-dards, Boulder, CO, October 1986.

Teters, L.R., J.L. Lloyd, G.W. Haydon andD.L. Lucas, IONCAP, IonosphericCommunications Analysis and Predic-tion Program, U.S. Department ofCommerce, Washington, D.C., 1983.

PROPAGATION AIDS

DXAid by Peter Oldfield, 251 CheminBeaulne, Piedmont, Quebec J0R 1K0

DX Edge from Xantek, Inc., P.O. Box 834,Madison Square Station, New York, NY10159

Geoclock by Joseph Ahlgren, 2218 N.Tuckahoe St., Arlington, VA 22205


NEEDS TO BE RENUMBERED WHEN PAGES ARE FIXED

10.7-cm flux, 2011-year cycle, 21, 9527-day recurrence, 21, 86, 93, 103A-Index, 27absorption, 36, 49, 50absorption coefficient

absolute, 36relative, 36

angleincidence, 28radiation, 30, 31, 32, 44,reflection, 43zenith, 28, 30

Angstrom, 13antenna aperture, 37antipodal point, 46atmospheric composition, 11aurora, 93auroral

absorption, 97, 110, 111displays, 96emissions, 96ionization, 96latitudes, 78zone, 78, 98

Brewster angle, 44chordal hops, 68, 69, 77chordal ducting, 70, 77collision frequency, 35conjugate points, 98, 111control points, 54, 110corona, 93coronal

hole, 93, 94mass ejection (CME), 93-94

cosmic radio noise, 32cosmic rays, 36critical frequency, 16, 28

foFE, 16, 18foF1, 16, 18foF2, 16, 18

D-region, 11, 13, 35day/night ratio, 98defocusing, 45dipole

equator, 26field, 26, 103, 112pole, 26

dissociation, 13disturbance

ionospheric, 16magnetic, 85solar, 16

DX feasibility, 63, 64E-cutoff frequency, 49, 50, 51E-region, 11, 16, 28, 31echo, 16, 28, 41effective vertical frequency, 74electron density profile, 16, 18electron energy, 89electron-Volt, 13electron-ion collisions, 105electron-neutral collisions, 35equatorial anomaly, 75, 77equatorial ionosphere, 67extraordinary wave, 42, 49F1-region, 11, 15F2-region, 11, 15fading, 45Faraday rotation, 43focusing, 35frequency

highest possible (HPF), 56, 61 maximum useable (MUF), 47, 49,

54, 62optimum working (FOT), 56, 61optical, 13

geomagneticactivity, 85, 100, 109control, 12, 25storm, 85, 93, 100, 104, 109

gradient, 66, 70, 71, 77grayline path, 37great circle path, 46, 73ground reflection

loss, 44, 50, 111scatter, 44

gyrofrequency, 25heating, 35height

reflection, 18, 47virtual, 18, 29, 47

index of refraction, 15interplanetary field, 25, 112, 113ion chemistry, 53ion production, 4, 53ion loss, 12, 53ionization potential, 13ionogram, 18, 42ionosonde, 16, 41

SUBJECT INDEX


ionosphere plane, 28 curved, 29

ionospheric profile, 15, 18ionospheric tilts, 57, 65ions

positive, 12, 53negative, 111

irregular ground, 44K-Index, 11, 30long-path propagation, 73, 74lower decile, 56, 108magnetic

conjugate, 98disturbance, 93equator, 57,118field, 94latitude, 78longitude, 78

magnetic stormsgradual commencement (GC), 93sudden commencement (SC), 93

magnetometer, 26magnetopause, 83, 87, 94magnetosphere, 84, 87, 95, 110,magnetotail, 84, 95, 114map

azimuthal equidistant, 46global foFE, 16, 55, 58, 59global foF2, 16, 17, 57, 58, 59Mercator, 46, 74, 75-77

mirror reflections, 29, 65mode availability, 62-63, 119modes

ionospheric, 50mixed, 50

MUFfailure, 54focusing, 45

negative ions, 111NOAA PBBS, 20, 27, 101noise

atmospheric, 38galactic, 38man-made, 38

noise power, 39non-great circle path, 71oblique

incidence 28, 29propagation, 28, 29sounding, 29

path reliability, 62, 63, 119paths

multi-hop, 46, 52, 72

single hop, 28, 29PCA events, 90, 110, 111, 113photodissociation, 13, 116 photoionization, 13, 67, 116plane geometry, 28, 65planetary indices, 26plasma

magnetospheric, 84solar, 84

plasmapause, 84plasmasphere, 84polar cap, 90polar cap absorption, 110, 111polar paths, 78polarization

circular, 42elliptical, 42horizontal, 41vertical, 41

polarization loss, 43propagation programs

IONCAP, 48IONSOUND, 48MAXIMUF, 48MICROMUF 2+, 48MINI-F2, 49MINIFTZ, 48MINIMUF, 49MINIPROP, 48

raybending, 11, 28path, 28

reflection, 11, 29tracing, 19, 27, 29-32, 68, 70

recombinationdissociative, 53, 116radiative, 53

reflectioncoefficient, 44loss, 44

refraction, 11, 15, 28, 35, 45, 47, 57, 65, 114refractive index, 15secant law, 29short-path, 29shortwave fade-out (SWF), 16, 83signal strength, 28, 34signal/noise ratio, 39skip

fading, 31, 45focusing, 31, 45zone, 31, 45

SKYCOM, 48solar

corona, 112


chromosphere, 112flares, 25, 89flux (10.7 cm), 20photosphere, 112plasma, 83, 86, 87, 93, 116proton event, 87, 90radio emissions, 91wind, 25, 86, 87, 89, 112

shortwave fade-out (SWF), 16, 83, 89south-north propagation, 43spherical geometry, 29sudden ionospheric disturbance (SID), 25, 89sunspot

cycle, 21, 22, 95, 105maximum, 95, 105minimum, 95, 105

number, 16, 20, 21, 95, 105thunderstorms, 44trans-equatorial propagation, 53transpolar paths. 53, 111upper decile, 56, 108UV, 11wave

extraordinary, 41, 49front, 41ordinary, 41-43, 49unpolarized, 44

winter anomaly, 55, 75WWV broadcasts, 20, 100X-rays

solar, 13, 20, 22, 89, 101, 103, 113


Worldradio

P.O. Box 189490, Sacramento, CA 95818

BOOKS

The Little Pistol's Guide to HF Propagation - K9LA

Documents

Transcript of The Little Pistol's Guide to HF Propagation - K9LA