Part II ... by Bob Brown .. NM7M
Chapter 1Friends in Radio Land -
It's no secret that success in DXing means getting signals to and
from a DX station and also having them heard and read at both ends
of the path. But between those two ends, a lot of things happen
in the ionosphere and some of them seem like well-kept secrets.
So the hope is some of that can be dispelled by the discussion
which follows. But we need a beginning and the question is where
to start. Let's take the easy way and cover old ground first, the
matter of ionospheric absorption that was discussed in the second
session of Prop. 101.
So we go back to the idea that RF excites the electrons in going
across the ionosphere, jiggling them at the wave frequency. And
they collide with nearby atoms and molecules, transferring some
energy derived from the waves to the atmosphere. That's how
absorption takes place, mostly down in the D-region. But there's
a frequency dependence we should talk about now, how absorption
varies with the operating QRG and with height, since the collision
frequency of the electrons is not constant; instead, it decreases
with height and that's a help. So it's clear now that ionospheric
absorption is a little more complicated than I first let on back
in Prop. 101.
But one can get a handle on it by looking at the extremes, low
in the D-region, say around 30 km where the collision frequency is
greater than any of the frequencies in our spectrum. In that
circumstance, collisions happen so often the electrons never have
a chance to pick up any energy from the passing RF. On the other
hand, at high altitudes, say around 100 km, collisions are quite
infrequent and the electrons re-radiate most of the energy they
acquire and transfer very little to the atmosphere by collisions.
So it is in between, where wave and collision frequencies are
comparable, that electrons take up RF energy efficiently and then
promptly deliver it over to the atmosphere. So with collision
frequency falling with increasing altitude, 28 MHz RF is absorbed
at lower altitudes than 3.5 MHz RF, as shown below:
Relative Absorption Efficiency per Electron
100 +
+
+ * * * * - 3.5 MHz
+ * o - 28 MHz
10 + * o o *
+ *o o *
+ *o o *
+*o o *
1 *o o *
+ o *
+ o *
.1 + + + + + + + + + + + + + + +
30 40 50 60 70 80 90 100
Ht (km)
That graphic illustrates something that DXers know already, lower
frequency signals are absorbed more than higher ones but it shows
where it all happens. That's news, at least for some.
To go beyond that qualitative result, one must have an analytical
form to represent the curves, call it F(f,h) for frequency f and
height h. Then multiply F(f,h) by the number N of electrons per
cubic meter at height h and include the physical constants to give
the right units, dB/km. When all is said and done, the result is:
Attenuation (dB/km) = 4.6E-2 * N * F(f,h)
But that is only at one place, where the electron density is N.
Our DXer's signal is attenuated by ALL the electrons encountered
along the RF path from point A to point B so that means we need to
know something about the propagation mode, the distribution of
electrons and add up the results, km by km along the path.
That's a tall order but when it's done, it will enable our DXer to
find just how much of the radiated power P survived in going from
A to B. But whether our DXer can be heard still depends on how
well the attenuated signal compares with the noise power getting
to the receiver at B. But I'm getting ahead of myself.
The crude graphic shown above can help in understanding a lot of
simple things. For example, it is possible to identify various
ionospheric disturbances just by the absorption they produce. One
approach is to use an HF receiver to monitor the galactic radio
noise coming in vertically on 30 MHz. Galactic noise gets right
through the F-region as 30 MHz is above its critical frequency,
even at equatorial latitudes where it might reach 20 MHz in a
solar cycle. That instrument is called a riometer, for Relative
Ionospheric Opacity METER, and they are generally deployed at high
latitudes where ionospheric disturbances are most common.
So now, if some disturbance increases the electron density in the
D- or E-region, we see that the galactic noise signal will be
attenuated and indicate the presence of a disturbance. But there
are disturbances and then there are DISTURBANCES. So the graphic
also tells us that anything that disturbs the lower D-region will
produce strong attentuation of the galactic radio noise and,
electron for electron, the attenuation will be much less if the
disturbance produces ionization at much higher altitudes.
The first case would be for polar cap absorption (PCA) events,
like we all experienced in May of '98. In those events, solar
protons produce lots of ionization around 40-50 km altitude and
give rise to tens of dB of additional absorption on 30 MHz and
blackout oblique communication paths going across the polar caps.
Auroral events, say associated with magnetic storms, give rise to
strong ionization above 100 km, where the graphic shows the
absorption efficiency is much lower, and auroral absorption (AA)
events show only a few dB of absorption of galactic noise on 30
MHz. Of course, there are other differences in the two types of
events, how the ionization is distributed in latitude and
longitude and how long they last. More on that later.
One last disturbance, again something that was within our recent
experience with all the flare activity in the summer of '98, is
sudden ionospheric disturbances (SID) from bursts of solar Xrays.
Those Xrays, in the 1-8 Angstrom range discussed earlier, were
incident on the sunlit hemisphere of the earth and literally
swamped the normal distribution of ionization at low altitudes,
giving intense absorption of signals going across the sunlit
region. But experience shows, and the graphic indicates, that the
effects were worst at the lower ends of the spectrum, wiping out
75 meter operations but having little effect on 28 MHz, except
perhaps for some solar noise bursts associated with the flaring.
All this would be quite academic, perhaps, were it not for the
fact that one can use the Internet to see these events in action
or shortly thereafter. Thus, records from the Xray Flux Monitors
on the GOES 8 and 10 satellites are shown at:
http://www.sec.noaa.gov/today.html
giving more meaning to the idea of an SID. And there is data on solar proton and auroral events available too at:http://www.sec.sec.noaa.gov/tiger but those plots involve satellite motions and you have to become something of a "rocket scientist" to get a full understanding of what the records show. We'll get to that later on but the main thing for us in the records is that plots for 0 degrees tell what is going down into the atmosphere, making more ionization and affecting the ionosphere. The 90 degree plots involve particles trapped in radiation belts and are more colorful than informative. While disturbances come and go, affecting our ability to work DX, we really need to know something about the normal situation, say the distribution of ionospheric electrons with height as well as latitude and longitude. That is a big order but, believe it or not, it can be contained in one HD computer disk. I'm talking about the International Reference Ionosphere (IRI), the summary of decades of ionospheric sounding all over the world. So it will provide data on the robust part of the ionosphere at low latitudes where the sun is more overhead and the mid-latitudes where the ionosphere is more seasonal in its properties. But the model is not reliable at high latitudes, say from below the auroral zones and poleward. That region is under the constant influence of the solar wind and electron densities are highly variable, even hour by hour. So that model has its limits. But to bring the model to life, one needs a mapping program to show the vertical and global distribution of ionization. Fortunately, we now have such a program available to amateurs, the PropLab Pro program from Canada. I'll have more to say about that next time. Reference Notes: A better representation of the relative absorption efficiency per electron as a function of height and frequency in the D-region is found in Figure 8.1 in my book, "Little Pistol". And a more detailed discussion of the analytical form, F(f,h), is found in Section 7.4 (Ionospheric Absorption) of Davies' book, "Ionospheric Radio", beginning on p. 214. Also, the variation of collision frequency with height is given in Figure 7.5 on p. 215. |
Chapter 2Friends in Radio Land -
Last time, it was pointed out that further progress on propagation
requires knowledge of how ionospheric electrons are distributed.
Of course, that will be different, day and night, as well as with
seasons and sunspot cycles. Again, it would be easy way to fall
back on something in Prop. 101, say the night-time ionosphere and
continue the discussion from there. But that would involve a
tremendous leap over distance and logic that's not too productive.
So let's talk/walk our way up to higher altitudes, starting from
where we are now, the D-region.
For one thing, the D-region involves a lot of familiar ideas and
we can work from there. For example, below the 90 km level, our
atmosphere is pretty well mixed, about 78% nitgrogen molecules and
21% oxygen molecules, by volume. The remaining 1% is made up of
permanent constituents, like the noble gases as well as hydrogen,
methane and oxides of nitrogen. Of course, every schoolboy knows
about the variable constituents, like water, carbon dioxide, ozone
and various bits of industrial debris, smog, that are found in
around heavily populated regions.
Global weather systems keep the lower atmosphere all stirred up,
in a mechanical sense, but that is not to say that convection from
solar heating is the only influence of the sun. Indeed, as was
discussed earlier, there are electrons and positive ions in the
lower D-region, released by solar EUV and Xrays. When the sun
sets, one might think that all the ionization disappears by
recombination and the region becomes de-ionized and neutral.
Of course, the ionosphere is always electrically neutral, with the
equal numbers of positive and negative charges, but recombination
lowers their numbers. Still some ionization does remain, produced
by other sources; those include UV and Xray photons in starlight,
sunlight scattered by the gas envelope (geocorona) surrounding the
earth and even charged particles, the energetic protons in the
galactic cosmic ray beam.
So it follows that ionospheric absorption would be greatly reduced
after dark but does not go to zero. There is good news in this
discussion, however, as some electrons are taken out of the
absorption loop at night by becoming attached to oxygen molecules.
Those negative ions are so massive that they can't be budged by RF
going by and just do not participate in the absorption process.
And at night, the number of negative ions of molecular oxygen in
the lower D-region grows to large numbers in going downward from
the 85 km level. That is the very reason that those solar proton
or PCA events mentioned in the first session of Prop. 201 show
much less absorption when the sun sets. But when the sun comes
up, solar photons detach electrons from the negative ions and
absorption goes back to the daytime level again. That does not
happen for auroral events and that is another story, about another
region higher in the ionosphere. More on that later.
In any event, the frequency dependence is still in effect for
whatever absorption occurs, taking a heavy toll on low frequency
signals. But that is still not fatal to propagation, even on the
low bands. Thus, everyone knows about broadcast stations coming
in better after dark and those signals can be heard across very
great distances, as many SWLs will testify. And even with more
limited power, 160 meter operators can still work great DX. But
in the last analysis, both SWL and low-band DXers run up against
the same problem, noise. That also has its origins down at low
altitudes so we can deal with that right now, while in the region.
Noise is described as broad-band radiation from electrical
discharges, either man-made or natural in origin. Whatever the
case, being a radio signal, noise will be propagated like any
other signal on the same frequency. That means, for one thing,
that noise signals that are below the critical frequency of the
F-region overhead will be confined to the lower ionosphere,
dissipate down there and not escape to Infinity. By the same
token, noise signals above the critical frequency are lost and
won't bother us very much on the higher HF bands. But the lower
bands do have a problem; so let's talk about it.
Noise of atmospheric origin comes from lightning strikes and will
be seasonal and originate in fairly well-defined areas. Among the
powerful sources of noise are low-latitude regions of South
America, South Africa and Indonesia. But we have our own noise
source, the southeastern states during the summer months. So
broad-band noise originates from those regions and is propagated
far and wide through regions in darkness. But once the sun comes
up, ionospheric absorption takes over and the only noise heard is
of local origin, static crashes from nearby lightning strikes.
The above points are not news to domestic DXers; they are quite
familiar with their own situation and can work within its limits.
But those going on DXpeditions often go into unfamiliar territory
and don't always think about the atmospheric noise problem. So
160 meter operators on DXpeditions have been known to be greeted
by S-9 noise the first time the receiver was turned on. That
evokes instant panic and sets in motion efforts to ameliorate the
problem, say trying different antennas and such. Those don't work
every time and hindsight often proves the problem could have been
avoided, in large measure, by planning the DXpedition for a time
on the winter side of an equinox, not the summer side.
Of course, the other source of noise is quite local, man-made in
origin and coming from various electrical devices. While the
global dimensions of atmospheric noise have been investigated
extensively over the last 50 years or so, the same is true of man-
made noise and it can be categorized as to origin and even given a
frequency dependence.
As for origins, the worst situation is an industrial setting and
then lesser problems are found with residential, rural and remote
sites, in that order. In that regard, the IONCAP propagation
program allows one to select the receiver siting and then takes
that, as well as the bandwidth (in Hz) of the operating mode, into
consideration in calculating the signal/noise ratio that would be
expected for a path. Of course, an operating frequency is put in
for each calculation, giving results for noise power similar to
the rough sort of frequency variation shown below:
Noise Power (dBW/Hz) * - Industrial
-140 - o - Residential
* x - Rural
o * # - Remote
x o *
-160 - x o *
# x o *
. # x o *
. # x o *
-180 - # x o *
. # x o
. # x
. #
-200 - #
+ - - - + - - - + - - - + - - - + - - - + - - - + - - - +
3.5 7.0 10.5 14.0 17.5 21.0 24.5 28.0
Frequency(MHz)
It should be realized that those values for the noise power are
averages throughout a day and subject to considerable variation,
with changes in human activity. So low-band DXers sitting there
in the wee hours of the morning will not hear the buzz of chain
saws or weed-eaters but they might have to put up with other
noise, say sparking heaters in fish tanks or hash from computers,
TVs or various forms of consumer electronics in nearby homes.
Last of all, there are extraterrestrial sources of noise too, from
the galaxy, as noted in regard to riometers, and solar noise
outbursts. Galactic radio noise is quite weak and reception
requires very sensitive receivers at sites well-removed from
sources of man-made noise. But solar noise is another thing and
it can be quite strong at times when solar flares are in progress.
As you'd expect, solar noise can pass through the F-region if its
downward path has an effective vertical frequency that is greater
than the critical frequency of the F-region. Thus, solar noise
would be heard more often at the top of the amateur spectrum,
especially when the sun is at a high angle in the sky. And it can
be quite strong at times, whooshing sounds that rise and fall in
intensity, even capable of overpowering CW and SSB signals on the
higher bands. By way of illustration, solar noise was discovered
by British scientists during WW-II and was first thought to be a
new form of German radar jamming. OK?
Extraterrestrial noise sources are getting a bit far afield so
we'd better get back down in the D-region and move on from there,
going above 90 km and seeing how matters start to change.
Reference Note:
A detailed discussion of radio noise, both atmospheric and man-
made, is found in Section 12.2.4 of Davies book, Ionospheric
Radio. In addition, McNamara shows how to calculate noise power
for the various categories of sites on p.143 of his book, Radio
Amateurs Guide to the Ionosphere; in addition his Appendix A goes
on to show how to find field strengths and S/N values on any path.
|
Chapter 3Friends in Radio Land -
Now we have to move up from the D-region, going above 90 km into
greater heights. In doing that, it is necessary to not only talk
about the ionosphere but also the underlying neutral atmosphere.
A few words about the ionosphere will do for starters since that
is something we've already covered. For example, the collision
frequency of electrons with their neutral surroundings is quite
important in discussing ionospheric absorption. And I mentioned
that falls off with increasing altitude. The same is true of the
collisions between the neutral constituents. So neutral-neutral
collision frequency goes from about 6.9E+10/sec at sea level to
1.2E+4/sec at 90 km, dropping about six orders of magnitude. The
same is true of the number density, going from 2.5E+25 particles
per cubic meter at sea level to 5.9E+19 particles per cubic meter
at 90 km.
Clearly, things thin out as we go up and collisions become much
more infrequent. Of course, you suspected all that but now you
know some of the numbers. But you may have not suspected how
those changes would affect DXing on HF, even VHF. So stay tuned
as I go a bit further; then I will get to the "nuts and bolts".
To go on, I mentioned the atmosphere is lightly ionized and I also
pointed out that recombination was the fate of electrons and
positive ions, especially after dark. But it does go on even in
the sunlight and one process involves recombination of positive
molecular ions of oxygen (O2+) with electrons. When that happens,
the neutral molecule (O2) is re-formed but with excess energy; so
it flies apart, into two oxygen atoms (O). But considering how
lightly ionized things are in the ionosphere, that can hardly be
considered as a strong source of oxygen atoms. OK?
But during the day, the atmosphere is bathed by energetic solar
photons; some, as we know, ionize oxygen molecules and thus can
contribute to the ionosphere. Others dissociate oxygen molecules
into two atoms. But with such a low collision frequency at 90 km,
an oxygen atom can linger around for about a week before finding
another oxygen atom and recombine to form molecular oxygen again.
So the long and short of it is that by the steady illumination
of the atmosphere by the sun, atomic oxygen can build up to become
an important constituent of the atmosphere above 90 km. One step
further tells us the atomic oxygen ions, O+, will be created too
by all those solar photons going by. So how long will those ions
last? Good question; it depends on which process is considered,
perhaps recombination with an electron to form a neutral atom. It
turns out that if recombination were the only possible fate for O+
ions, they'd linger around a long time too. Something else seems
to happen but before getting to that, let's look a bit deeper into
the O+ situation up above 90 km. OK?
The recombination of O+ with an electron is a radiative process,
the excess energy being given off as a photon while the atom
recoils to conserve momentum. But it is slow , I mean VERY SLOW
in the scheme of things. And that seems to be the case for other
similar radiative processes, like with metallic ions. It just
seems to take forever for an electron and metallic ion to get it
together and recombine. But now comes the PUNCH LINE; there are
metallic ions in the upper atmosphere, meteoric debris that has
drifted down and been ionized by solar photons.
And recombination being a slow process, they linger around a long
time. In fact, they can linger around and be caught up in the
occasional weather activity up around 100 km, wind shears. And
being tied, as it were, to field lines, wind shear can compress
them into a thin layer. But their electrons are not far away so
that makes for a thin layer of electrons too. So now you guessed
it; I'm talking about sporadic E layers up around 100 km or so.
The electron population, being squeezed into a thin layer, looks
sort of metallic too when it comes to wave propagation so RF is
really reflected by those layers, the sort of thing we talked
about back in Prop. 101, tilted reflecting layers. In the present
case, the tilt would be that of the magnetic field lines that hold
the charges. But the tilt is not so important to DXers; it's the
presence of a strong, reflecting layer around 100 km altitude.
Sporadic E is known to be a nuisance for HF propagation. By its
presence, it can RF cut off from long paths via the F-region up
around 300 km and thus disrupt long-haul communications. And the
reflecting properties can be so great as to not only reflect RF
from the top of the HF spectrum, to the annoyance of 28 MHz DXers,
but also reflects RF in the VHF portion of the amateur spectrum,
to the joy of the 50 MHz and 144 MHz DXers. I should add that
some contestors love sporadic E as they can go to higher bands and
make many short-haul contacts on bands that would be quite dead
otherwise. All that from the fact that recombination is so slow
for atomic oxygen and metallic ions.
Still speaking about the importance of atomic oxygen in the
atmosphere above the D-region, its build-up by photo-dissociation
of oxygen molecules serves to add it to the "targets" for the
various forms of incoming radiation, photons or charged particles,
that pass through the upper atmosphere. And just to make my
remarks rather "timely", if you saw any bright aurora a couple
weeks ago, at the end of September, the green color you saw was
the 5577 Angstrom spectral line from atomic oxygen. How about
that? I should add that the green aurora "washes out" to become
gray aurora at great viewing distances. That's a property of the
eye, they tell me.
And speaking of great viewing distances, the best atomic oxygen
story I know of has to do with the early days of Rome. It seems a
red glow was seen in the northern sky and the Romans figured it
was the Huns, pillaging villages up north. So they saddled up,
got in their chariots and roared off in the night. No Huns were
found but the sky glowed again the next night. More riding, still
no Huns. Nowadays, we know they were fooled by the red line of
atomic oxygen, 6300 Angstroms found up around 1,000 km. You can
do a simple graphical calculation to find the distance of the
aurora from the Romans. (Using 6,371 for the radius of the earth
and my plastic ruler/compass, I get about 3,300 km; that works out
to about 30 degrees of latitude, putting the aurora up over the
northern coast of Norway. Sounds right to me!)
But back to the ionosphere and the O+ ion. As I indicated, its
recombination with electrons goes very slowly, meaning that it
could undergo other, more likely processes. To make a long story
quite short, an ion-atom interchange can take place in nitrogen
molecules with O+ displacing a N atom and forming a positive
nitric oxide ion, NO+.
So now we have all the principal players in the ionospheric drama,
electrons and negative ions of molecular oxygen as well as all the
molecular ions, oxygen, nitrogen and, now we add, nitric oxide.
It is the physics and chemistry of those ions, in the presence of
the neutral atmosphere, that we have to look to to understand all
the mysteries of HF propagation.
But now with the full cast of characters, we have to work our way
up above 90 km. So the next stop will be the E-region, up around
105 km. During the day, it is one of the levels of the full
electron distribution shown below:
Ht(km) *
*
+ *
| *
| *
300 + *
| F2-Region *
| *
+ *
| *
| *
200 + F1-Region *
| *
| *
+ *
| *
| *
100 + E-Region *
| * * * * * *
| * * * * *
+ * D-Region
|
|
+ - + - - - - + - - - - + - - - - + - - - - + - - - - + - -
1E+1 1E+2 1E+3 1E+4 1E+5 1E+6
electrons/cc
Reference Notes:
A brief discussion of the occurrence of sporadic E layers is given
in Section 3.5 of McNamara's book and a detailed discussion of the
mechanisms related to sporadic E, complete with references, can
be found in the October/November '97 issues of QST.
The Roman aurora story as well as other interesting tales about
the geomagnetic field may be found at the end of the second volume
of "Geomagnetism" by Chapman and Bartels, Oxford University Press,
1940. Great reading!
|
Chapter 4Friends in Radio Land -
We pick up where we left off, going up to the E-region. You will
recall it is the first "step" in the ionosphere that lies above
the D-region, essentially an inflection point in the curve that
outlines the vertical distribution of electrons:
Ht(km)
+ *
| *
| *
100 + E-Region *
| * * * * * *
| * * * * *
+ * D-Region
|
|
0 + - + - - - - + - - - - + - - - - + - - - - + - - - - + - -
1E+1 1E+2 1E+3 1E+4 1E+5 1E+6
electrons/cc
In the early days of ionospheric sounding, that inflection was
enough to give an echo, making it stand out in the records like
the peak of the F-region. And it is there all the time, the most
well-known and studied part of ionosonde records. But there were
also surprises in the same range of the records, sporadic E
layers. But those are known for their irregular and unpredictable
behaviour and make a separate study that will not concern us here.
But those sounders were calibrated in frequency, not electron
density, and thus they provided data on critical frequencies. If
one does a bit of ionospheric theory, the electron density and
critical or plasma frequency are found to be related as follows:
fc = (9*E-6)*SQRT(N)
where fc is in MHz and N in electrons/cubic-meter. Going to the
curve above, the electron density at 100 km is roughly 8E+4
electrons/cc or 8E+10 electrons/cubic-meter, yielding a critical
frequency of 2.6 MHz.
The electron density profile given above is for daytime conditions
so signals incident on the bottom of the ionosphere would pass on
to the F-region overhead if their effective vertical frequency
were above 2.6 MHz. As an illustration, 7 MHz RF launched at
30 degs would have an effective vertical frequency of 3.5 MHz and
make it through to the F-region easily while at 15 degs, the
effective vertical frequency would only be 1.8 MHz and RF would be
blocked or "cut-off" from the F-region. I'm sure you've heard
that term before in connection with propagation programs.
Now I made a couple of points about the positive ion of atomic
oxygen (O+): that its recombination rate is quite low and that it
can undergo ion-atom interchange with molecular nitrogen to yield
a positive ion of nitric oxide (NO+). Just to come up with some
numbers, I checked on the situation here at my QTH, using the
International Reference Ionosphere (IRI) program at local noon for
the recent equinox. The atomic oxygen ion proved to be less than
1% of the positive ions at the 100 km level; also, using some rate
coefficients from ion-chemistry, it turned out that the molecular
ions recombine with electrons at a rate which is 150 time faster
than that for the atomic oxygen ion. OK? See what I mean?
The relative rates will remain the same with solar zenith angle so
that means that at low altitudes in the D-and E-region, the slow
loss rate of O+ by recombination is not important and ionization
largely disappears as molecular ions recombine with electrons when
the sun sets. Put another way, the level of ionization in the
E-region is really controlled by the zenith angle of the sun,
being the greatest when the sun is highest angle in the sky and
quickly disappears by electron recombination when the sun sets.
Of course, the phase of the solar cycle plays a role too so the
experimental studies show that the critical frequency foE of the
E-region during daytime hours is given by the following
expression:
foE = 0.9*[(180+1.44*SSN)*cos(Z)]^(0.25) MHz
where Z is the solar zenith angle, SSN is the sunspot number and
the expression between square brackets it taken to the 1/4 power.
It should be noted that this expression does not apply at high
latitudes where auroral ionization in the same altitude range is
common and would be added to that of solar origin. And it does
not apply at night where there are special conditions just above
the E-region. More on that later.
But beyond those caveats, it should be borne in mind that the data
on which that algorithm is based had some experimental uncertainty
associated with it, say 5%-10% for individual foE entries from the
raw ionosonde records. So it would be a mistake to give any
reliance on the predictions that are inconsistent with the data
input. This holds true throughout all of ionospheric work; the
ionosphere is not a High-Q device and though results derived from
the databases can be given to a large number of figures, not all
of them are really significant. OK?
Now, in your mind's eye, think of a spherical earth and the sun
situated over some point between the Tropic of Cancer and the
Tropic of Capricorn. Circles on the earth's surface centered
on the sub-solar point would be locations having equal solar
zenith angles and thus would have the same value for foE. Of
course, the highest foE value would be at the sub-solar point.
At the time of the recent equinox, when the effective SSN was
about 75, that would give foE as 4.1 MHz for local noon at the
equator. And foE would have the same value at local noon for
times of the summer and winter solstices at the Tropics of Cancer
and Capricorn, respectively, if the SSN remained the same.
If your QTH were on the sunlit hemisphere, you would be able to
find foE for the ionosphere overhead by finding which circle your
QTH was located on. Better yet, if you know about great-circle
navigation, like some boating enthusiasts, you could calculate foE
yourself. All you need to know is the date, time and your own
coordinates to find the solar zenith angle with the aid of the
your hand-held calculator or, better yet, the U.S. Navy Nautical
Almanac computer program; the equation above tells the rest.
This last point brings to the fore that discussions making use
of "Flat Earth Physics" must come to an end. To do things right,
we really need to put in the curvature of the earth and the
ionosphere. So from here on, we'll be treating the ionosphere as
spherical and concentric with the earth. And while we're at it,
we'd better put a bottom on the ionosphere, up there around 60-70
km where the D-region ionization rapidly heads toward zero. If
nothing else, that is needed to find the correct angle for the
effective vertical frequency calculation or the fraction of a path
that goes through ionization in the D-region.
Those who know great-circle navigation can pretty well see how it
would go but other geometers, skilled with a graduated compass and
straight edge, can still see some important facts. For example,
it is fairly easy to show that the angle of approach for RF
incident on a curved ionospheric layer is smaller than for a plane
layer, thus raising the effective vertical frequency and making it
more likely that RF can punch through the region. It's also easy
to show that the slant path through a curved ionosphere is longer
than for a plane layer, thus having RF pass through more electrons
along a path and increasing the amount of ionospheric absorption.
Whether the E-region is a problem or not depends on the operating
frequency. Thus, at the high end of the amateur spectrum where
MUFs of the F-region are important, the operating frequency is
greater than foE and it is possible for RF to go right through the
layer, on to the F-region at greater heights. But that is not to
say that some bending/refraction does not occur in the passage
through the E-region. It is just small compared to the refraction
that brings oblique signals back down to ground level.
At the low end of the amateur spectrum, the E-region is the enemy,
keeping signals on paths with short hops and high absorption. It
is to be avoided at all costs by DXers so their operating times
are all in hours when there is full darkness along the paths of
interest. So come sunset, operations begin and come sunrise, they
come to an end. It's as simple as that but a lot of sleep is lost
in the process.
It is the transition bands, 10-18 MHz, where both the E- and F-
regions are important. Thus, operations are often arranged to
coincide with dawn or dusk on the E-region but while critical
frequencies of the F-region are still high. This is termed "gray
line" operation and is particularly helpful to DXers interested in
long-path propagation. More on that later.
Reference Notes:
Numerical algorithms for critical frequencies are found in most
ionospheric references that have any quantitative aspect to them.
It should be recognized that while the various algorithms may
appear different, they all give good representations of the
experimental data.
An excellent discussion of ionospheric sounding and ionograms is
given in Chapter 5 of McNamara's book, Radio Amateurs Guide to the
Ionosphere. Davies' book, Ionospheric Radio, also has a good
discussion of ionogram scaling and interpretation in Section 4.9.
While I bought my copy of the International Reference Ionosphere,
as I recall, you can download IRI from a website on the Internet;
so try:
ftp://nssdc.gsfc.nasa.gov/pub/models/ionospheric/iri
This will give you a numerical view of the ionosphere, in tabular form, and provides the basis of many calculations. |
Chapter 5Friends in Radio Land - So far, we've been down in the D- and E-regions, talking about how electron collisions are responsible for absorption or attenuation of signals. Also, we got into comparing the effective vertical frequency of a signal with the critical frequency of the E-region to determine whether the signal would be blocked or go up into the F-region. We even have an algorithm for the critical frequency for the E-region, at least when the sun is up. Now, at this point, any progress up into higher regions of the ionosphere has to wait until we settle some pressing questions: about paths from point A to B and how, when the sun is up, they are affected by ionization in the E-region. Put another way, we have to do some mapping - showing details of the path from point A to B and where it lies relative to the regions which are sunlit. Of course, mapping brings up the question of coordinates and how RF is propagated. Coordinates are easy; you just need a good atlas. But those are not always easy to find. For example, I spent a small fortune on a new atlas from the National Geographic Society only to learn that it did not have any information on coordinates. I mean "NONE!" I did get a Rand McNally atlas, "Today's World", as a birthday present and found that it had coordinate grids in it, 1 degree latitude by 1 degree longitude. I suppose that can be considered "Good enough for Government Work" or ionospheric propagation but I rely on Goode's World's Atlas that high schools used years ago. As for paths, they are taken, to a first approximation in radio work, as being along great-circles on the globe. That would be good except for the fact that I pointed out earlier that RF can suffer lateral deviations, skewing one way or the other, due to gradients of the electron density across the path. But in the HF range, that skewing is relatively minor so we can, at least for a start, go with the idea of great-circles being appropriate to show where RF goes. In simplest terms, a great-circle is the trace on a sphere that results when it is sliced by a plane that also goes through the center of the sphere. Perhaps the best known great-circle is the terminator which divides the earth into regions which are sunlit and those which are not. So the sun illuminates half the earth and if you take the trace of that boundary, it also happens to be the intersection of a plane and the spherical earth. OK? Now radio paths are different in that they are only parts of the great-circle on the earth, that from A to B. That is called the short-path from A to B and the spherical arc can be up to about 20,000 km in length. But how does that path appear on maps is an interesting question; it depends on the type of projection. Now I should say at the outset that if you look in the early part of any atlas, you will be treated to a discussion of the various types of map projections. The one we see often is the Mercator or rectangular projection. There, distortions increase with latitude and what are in reality two points, the North and South Poles, are ultimately distorted into lines at the top and bottom of the map. The division of sunlit and dark regions, given by the terminator, shows up as something resembling a sine curve, at least for times of the year away from the equinoxes. And, depending on length, a radio path will have that curved character too. What is needed for our purposes is both a path and the terminator, for the date and time of interest. The part of the path in darkness will not suffer absorption to any extent while the part in the sunlit region is at risk, ionospherically speaking. Those who operate on the low bands, 40 meters down to 160 meters, are interested only in times when the entire path is in darkness. While sunrise/sunset tables are of some help, this is really where mapping becomes important. But, first, pause and look at sunrise/sunset tables, like the ones in the ARRL Operating Manuals. Assuming that a path falls fully within the dark hemisphere, operating times without the peril of severe absorption depend on whether the path is to the west or east of primary QTH. For a path toward DX to the west, there will be total darkness on the path after DX sunset and until the sun rises at your QTH. For DX to the east, it is just the opposite, from your sunset until the sun rises in the east. I have to say the use of tables is tedious and give not much resolution in time and locations, really a poor substitute for a mapping program. But some people still use them. The mapping program I like best is one included in the MINIPROP PLUS propagation program. The entries are simple, date and time, and coordinates of the terminii. Usually one's coordinates are default to the calculation and the far terminus is either given by the call prefix, districts, if the country happens to cover a large area, or actual coordinates. The program then gives a Mercator map, with the terminator and sun clearly shown, and both short-and long paths. It also gives the times of sunrise and sunset at each end and it is a simple matter to find when the path would open and close as well as the number of hours of darkness. In that projection, paths and the terminator are sine-like curves and the terminator moves east to west with time. There are other programs, like Geochron, Geoclock and DXAID, in which the position of the terminator actually advances as you watch it in real-time. Some people swear by that option but I'm not very excited by it, being more interested in what I'm hearing on the air. There is another type of map which I find most helpful in my propagation work, the azimuthal equidistant projection. You see that type of map in the back of the ARRL Operating Manual, with the first one centered on W1AW. In contrast to the Mercator projection, where distortions increase in going toward the poles, the azimuthal equidistant map is centered on one point and the distortions increase with distance toward the antipodal point on the opposite side of the earth. In fact, the antipodal point is distorted into a circle, in contrast to the straight lines for the geographic poles in the Mercator projection. The advantage of the azimuthal equidistant map is that all great- circle paths going out from a QTH in the center are given by straight lines. In addition, the distance along the path is linear, out to the antipodal distance of 20,000 km. But the disadvantage of the azimuthal equidistant map is that it has to be created for each QTH. There is another projection in which ALL great circles are straight lines, no matter where on the map. That is the gnomonic projection, used occasionally in propagation work. The gnomonic projection is centered on one geographic pole or the other and its disadvantage is non-linearity, with distortions which increase in going to lower latitudes and the maps usually only cover 30-45 degrees of latitude going equatorward from the poles. Myself, I prefer the azimuthal equidistant projection in the DXAID program as it includes auroral zones based on the model used to display the NOAA auroral maps on the Internet. The NOAA auroral maps on the Internet are given in terms of auroral activity while the maps in DXAID use K-indices for the corresponding levels of magnetic activity. So in using it, one can tell whether a path is more tangential to the auroral zone, for a given level of magnetic activity, or actually passes across the polar cap. With that kind of knowledge, one understands conditions far better just on hearing a signal. In spite of that preference for propagation purposes, I have to admit that I find the shape and motions of the terminator a bit odd in the azimuthal equidistant map projection, something that I have a hard time getting used to. In contrast to that, I have no problem with the terminator in the Mercator projection, its changes with time seem quite natural. So I have to say that each projection has its function as well as virtues and that one really needs a familiarity with both to deal with propagation problems. Having said all of that, we have to move on, above the E-region and into ionization that's largely responsible for propagation, toward the F-region peak. That will take us right into the matter of propagation predictions by bands, from fundamentals as well as computer programs. Of course, I've already made the point that a full-service propagation program would include noise, say as signal/noise ratios. Now, I think you can understand it when I say a person interested in propagation cannot get along without a good mapping program. In the ideal case, both the forecasting and mapping programs would be on the same computer disk. Failing that, at least both ought to be readily available to a DXer. Reference Notes: The MINIPROP PLUS program by W6EL has been available for some years but now it is undergoing a change, from being DOS-based to a Windows program. When the transformation has been completed, I imagine it will be advertised again in QST. In any event, the Mercator projection maps in this program are extremely agile and fast, making it easy to make rapid comparisons of paths in time. Like MINIPROP PLUS, DXAID has excellent graphics, particularly the azimuthal equidistant mapping version with auroral zones included. It also has a propagation module that is based on the F-layer algorithm due to Raymond Fricker of the BBC. At the present time, the algorithm is being improved, making it more comparable to predictions that would be obtained from the International Reference Ionosphere. Earlier tests show Fricker's work, in MINIPROP and other programs, comes closer to mimicing propagation predictions by IONCAP than other programs available at the time. Now, it should be even better. The ultimate test of paths is found in ray-tracing and the PropLab Pro program from Solar Terrestrial Dispatch is the only one that is presently available. The program not only traces propagation paths but also provides details on the distribution of electrons, globally or vertically, and gives a foundation for all ionospheric work. Myself, I would be absolutely LOST without PropLab Pro. 73, Bob, NM7M |
Chapter 6Friends in Radio Land - Now we have to get down to cases, dealing with the ionosphere above the D- and E-regions. But the transition is a smooth one, going from a well-mixed region largely made up of molecules and molecular ions to a region where collisions are less frequent, atoms become more abundant and constituents start to be sorted out by their chemical weight. We'll never really get up to the case where hydrogen is the dominant constituent but that is the idea, gravitational separation, in the upper reaches above us. The ionization in the E-region is under solar control and was shown by the critical frequency depending on solar zenith angle. Now, in going higher, toward the F-region peak, solar control does continue, up to the F1-region at about 200 km altitude. So the critical frequency foF1 during daytime is expressed similarly: foF1 = [4.3 + 0.01*SSN]*[cos(Z)]^(0.2) MHz As shown earlier, the electron density in the F1 region is greater than the E-region and the same is true of the critical frequency. And constant frequency contours will be centered about the sub- solar point. But at large zenith angles, the algorithm is less reliable and at night, the ionization in the F1-region decreases to low values. It does not go to down to a vanishing level but, instead, there is a "valley" in the electron density above the night-time E-region, as shown below: Ht(km) + * | * | * 200 + * | * | * + * | * Valley | * 100 + * E-peak | * * * * * * * | + D-Region | | 0 + - + - - - - + - - - - + - - - - + - - - - + - - - - + - - 1E+1 1E+2 1E+3 1E+4 1E+5 1E+6 electrons/cc The origin of the valley is complex, related to the change from molecular ions of oxygen and nitrogen down low to the appearance of atomic oxygen and the ion-atom interchange above 90 km that produces the molecular ion of nitric oxice (NO). Again, the ionization in darkness has the same origin as the E-region. Whether day or night, the ionization in the D-region is just not great enough to significantly bend or refract HF signals. On the other hand, during the day, ionization in the E-region can cut off signals from reaching the F-region. In short, signals like that go off on low-angle, shorter E-hops during the day. At night, HF signals will just pass through the weak ionization that remains in the E-region, shown above, just as if it were not there. That's another way of saying that the night-time value for foE is very low, even less than 0.5 MHz, and the region is no impediment to the advance of HF signals. On the other hand, that's NOT the case for signals in the 160 meter band. That will be VERY interesting but let's do some other things first. For example, let's look at how critical frequencies vary with sunspot number so we can put effects of the various ionospheric regions in perspective. For one thing, with the different heights for the regions, E-region around 100 km while the F1-region is around 200 km and the F2-peak up around 300 km, the frequency data will show how signals penetrate into the ionization overhead. That has a bearing on the lengths of the hops that result or, in more meaningful terms, on our ability to work DX on the various bands. So let's look at a crude representation of some mid- latitude critical frequency data for daytime conditions: 11 * MHz | 10 * F2 | 9 * F2 | 8 * F2 | 7 * | F2 6 * F1 | 5 * F2 F1 | 4 * F1 E | E 3 * E | 2 * | 1 * - + - + - + - + - + - + - + - + - + - + - + - + - + - + - + 20 40 60 80 100 120 140 Sunspot Number This crude graphic requires that you use your mind's eye to make connections between data points but the results is pretty clear: the lower E- and F1-regions which are under solar control show only modest changes in critical frequency or electron density as the sunspot number increases with solar activity. The F-region, on the other hand, shows large changes in critical frequency and is not under solar control, without any simple algorithm involving the solar zenith angle like the E- and F1-regions. The best way to illustrate the difference between solar control of the E-region and the situation with the F-region is through the use of maps showing the iso-frequency contours for the two regions. So the map below illustrates the situation for 0600 UTC on the spring or fall equinoxes. Of course, the sun is on the equator and at 0600 UTC, it is located at 90E longitude. The iso- frequency contours are illustrated below, circles centered on the sub-solar point (but distorted by the Mercator projection). Accordingly, the left side of the figure is the sunlit portion of the earth, the right side is in darkness and terminator consists of two straight lines at 0E and 180 E longitude. 90N o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o . * * * * * | . . * * | 0600 UTC . 60N . * o o o * | Fall or Spring . . * o o * | Equinox . . * o o * | . 30N . * o x x x o * | Solar Control . . * o x x o * | E-region . . * o x S x o * | . Eqtr 0++++o++++o++( U )++o++++o++++o++++o++++o++++o++++o++++o++++0 . * o x N x o * | . . * o x x o * | . 30S . * o x x x o * | * - 2 MHz . . * o o * | o - 3 mHz . . * o o * | x - 4 MHz . 60S . * o o o * | . . * * | . . * * * * * | . 90S o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o 0 60E 120E 180E 240E 300E 360E As noted above, the situation is similar for the F1-region except that the critical frequencies are somewhat higher. But the idea of solar control is clear from this type of figure; the ionization is where the sun shines and essentially nothing in darkness! Now as far as the F-region is concerned, its peak is up around the 300 km level and depends on the season, time of day and sunspot number. But at those heights, the electron collision frequency is low and the recombination rate of electrons with the positive ions (O2+ and NO+) is quite low. So ionization continues to exist after sunset; also, the geomagnetic control of the ionosphere is shown by the fact that the F-region map for critical frequency foF2 is organized better by geomagnetic coordinates rather than the usual geographical coordinates. The maps shown below are admittedly crude, of necessity, but they convey how the shape of geomagnetic dip equator compares with the iso-frequency contour of the F-region at low latitudes: 90N o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o . | . . | . 60N . | . . Geomagnetic Dip Equator | . . (exaggerated in scale) | . 30N . | . ._ _ _ _ _ _ _ _ _ | _ _. . S - - - - - - | - - . Eqtr 0++++o++++o++( U )++o++++o++++o- - o++++o++++o++++o++--o++++0 . N | - - - - - - . . | - - - - . 30S . | . . | . . | . 60S . | . . | . . | . 90S o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o 0 60E 120E 180E 240E 300E 360E 90N o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o . | . . x x x x x x x | Sample Contour . 60N . x x | 10 MHz for SSN = 137 . . x x | 0600 UTC . . x x | Fall or Spring . 30N . x F-region x| Equinox . . x |x x x . . x S | x x x . Eqtr 0+x++o++++o++( U )++o++++o++++o++++o++++o++x+o++++o++++o++++0 . x N | x x . . x | x . 30S . x x x|x x . . x x | x x x x x . . x x | x x x x . 60S . x x x x x x x x | . . | . . | . 90S o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o 0 60E 120E 180E 240E 300E 360E The sunlit and dark hemispheres are the same as before but it is seen that F-region continues after sunset, particularly at low latitudes and along the direction of the geomagnetic dip equator. Such critical frequency maps demonstrate that the ionosphere is controlled by the geomagnetic field at great heights but down lower, the distribution of ionization is under solar control. The transition occurs in going up through the F1-region. As for DX propagation, it is controlled in quiet times by the geomagnetic field but it doesn't take much imagination to think that any sort of disturbance of the field would upset DXing. More later! Reference Notes: Critical frequency maps of the E- and F-regions can be seen in my Little Pistol book. In addition, they will be found in books by McNamara and Davies. Excellent critical frequency maps are obtained from the PropLab Pro program. In fact, that program gives a full complement of ionospheric maps and in several projections. |
Chapter 7Friends in Radio Land - Last time, I showed one sample contour of a global map of the F-region, for 10 MHz when the SSN was 137. You can go back to the map to see how it spilled over into the hours of darkness. But that was only one contour. So the question comes down to the rest of the map, what other contours were like and their limits in critical frequency. Looking at the sample contour, it is easy to think that parts of the globe closer to the sub-solar point would have higher values of critical frequency, up to 16-17 MHz. After all, the sun was more overhead for there and the solar UV had less atmosphere to penetrate. But at larger zenith angles, particularly toward the polar regions, the critical frequencies would be lower, going down to 6-7 MHz. All that for a SSN of 137. What about lower SSN, say toward solar minimum? Then, for the region where the critical frequency was 10 MHz earlier, you can just put in 5-6 MHz and at higher latitudes, you can put in 3-4 MHz while at low latitudes, the value is 11-12 MHz. But whatever the SSN, the highest critical frequencies are always found at the lower latitudes. As a practical matter, that is an explanation why contest DXpeditions go toward equatorial regions; the bands are always open there and it is just a matter of how far their signals go poleward before running out of sufficient ionization. So I like to say that the low-latitude regions are the most robust of the ionosphere. But there is a difference between "robust" and "ROBUST", say for solar minimum and solar maximum. Before getting to that, I should point out there are "islands of ionization" at low latitudes, as shown by the additional contours given below: 90N o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o . | . . x x x x x x x | Sample Contour . 60N . x x | 10 MHz for SSN = 137 . . x x | 0600 UTC . . x x | Fall or Spring . 30N . x +++++++++++ x| Equinox . . x + 17 MHz + |x x x . . x +++++++++++ | x x x . Eqtr 0+x++o++++o++(SUN)++o++++o++++o++++o++++o++x+o++++o++++o++++0 . x +++++++++++ | x x . . x + 16 MHz + | x . 30S . x +++++++++++ x x|x x . . x x | x x x x x . . x x | x x x x . 60S . x x x x x x x x | . . | . . | . 90S o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o++++o 0 60E 120E 180E 240E 300E 360E What I have shown is somewhat out of scale, too wide in latitude and poorly positioned in longitude, as you would see if you looked at the original global map of the F-region. But it conveys the idea, islands of strong ionization in the afternoon/evening hours. This is called the "equatorial anomaly" and has profound effects for propagation, giving rise to long, chordal hops on HF and DX on VHF. Those regions are a regular part of the ionosphere, day in and day out, and the high level of ionization there adds to the robustness that I spoke of earlier. A few paragraphs earlier, I made mention of the fact that global maps of the F-region change with solar activity. One way of making these ideas more vivid in one's mind is to think of them like relief maps, with a "frequency surface" that rises or falls in height as critical frequencies change with increasing or decreasing SSN. The quantitative side of that approach can be shown by means of a N-S slice through the global maps that one obtains, say from the PropLab Pro program, for two different sunspot numbers: foF2 (MHz) 15 + Equinox at 0600 UTC, 120E longitude | | ooo | ooo o oo | o o o .. ooo 10 + o .. o o . . ooo | o . . ooo . .. o | o . . . .. oo | ooooo . . . .. oo | ooo .. . .. oo 5 + ooooo ..... ... ooo | .. ... | ...... ... | . - SSN = 10 o - SSN = 100 | o++o++o++o++o++o++o++o++o++o++0++o++o++o++o++o++o++o++o++o++o 90S 60S 30S Eqtr 30N 60N 90N Those N-S cuts across the F-region maps show the two "islands" of the equatorial anomaly as well as the deep notch in between them. Also, it shows again the geomagnetic control of the ionosphere by the asymmetry of the ionosphere at 120E, due to the fact that the magnetic dip equator is about 5 degrees north of the geographic equator at that longitude. Admittedly, the above graphics are pretty crude but they cover the main aspects of the ionosphere - E-, F- and F2-region maps - showing how ionization is distributed and how it varies with changes in solar activity. It is within those regions that we are trying to propagate signals. So we should lay down some great-circles to see where the paths are going relative to the ionization. The test, of course, is if the effective vertical frequency along a path is less than the critical frequency encountered. As long as that's true, propagation will continue; otherwise, the RF will penetrate the F-region and be lost. Looking at the last graphic, you can see that "the test" gets tougher at high latitudes where the critical frequency is on the low side, a few MHz. Thus, there will be angles at which the RF penetrates the ionosphere and is not returned to ground level. That is "skip", discovered by John Reinartz back in the mid'20s, and obviously gets worse at higher frequencies. In that regard, there is one "side light" to that on the higher bands. Thus, it is quite easy to "pass the test" and work to the south on 21 MHz, for example, as the ionosphere is quite "robust" in the N-S direction. But looking at the last figure, one can see that the ionosphere is "puny" in the E-W direction, with very low critical frequencies. As a result, when chasing DX on 21 MHz, skip makes it impossible to hear the station east or west of you that got the South American contact that you were trying for. At this point, our discussion comes down to exploring the aspects of the distribution of ionization, vertically and horizontally. The vertical distribution determines how signals are refracted or bent along a path while the horizontal distribution determines whether a hop is completed or how long it might be. There are two approaches we can follow, the rigorous one would be to trace ray paths through a model ionosphere while the practical one would be to use the model in a propagation program, looking at the critical frequencies at the two control points on a path to see what the MUF would be and whether one's RF passes the test. Ray-tracing takes us back to the analogy between the flight of a baseball and RF across the ionosphere. Mathematically, the flight of the ball is worked out using Newton' Laws, with equations of motion in two or three dimensions. You should not be surprised if I tell you that equations of motion for RF can be worked out, with the ionosphere playing the role of gravity. So, like any baseball or even spacecraft, the methods of mechanics work with RF and the equations of motion solved, step by step, to find the path of RF. In that regard, the PropLab Pro program is outstanding; all you have to do is put in the locations of the terminii, the date and time as well as the sunspot number, and it solves those equations of motion and traces out the path of the RF. Just fantastic! But there is one more thing to add; PropLab Pro also includes the role of the geomagnetic field in the equations of motion. At the upper end of the HF spectrum, that is not important as the QRG is large compared to the electron gyro-frequency about the field lines. But down around 160 meters, the 1 MHz gyro-frequency is comparable to 1.8 MHz and the effects of the magnetic field no longer appear to be negligible in the equations of motion. There are some interesting consequences for wave polarization as well as signal absorption. In addition, signals can get trapped in that valley above the night-time E-region and ducted to great distances with low loss. But we'll get to that later; first, MUF programs. Reference Notes: Originals of all the figures mentioned above can be found in my article, "On the Down-Sizing of the Ionosphere", that appeared in the July/August '94 issue of The DX Magazine. Also, the two main F-region maps are on p. 29 of my book on long-path propagation and also found in Davies' book, "Ionospheric Radio". In addition, there are a number of ray traces shown in my Little Pistol book, illustrating skip and showing how RF hops vary with frequency as well as radiation angle. |
Chapter 8Friends in Radio Land - Now we are in a position to talk about propagation predictions. I say that as you understand that predictions require some sort of representation of ionospheric maps, both E- and F-regions, and a method that looks at how effective vertical frequencies compare with critical frequencies along a great-circle path. I must admit that I have injected "effective vertical frequency" (EVF) into the discussion; you normally don't see that term when you read about propagation. In McNamara's book, he uses another form, "equivalent vertical incidence frequency", in his discussion but I find that just too wordy and besides, my choice of EVF fits the bill and tells the story. I hope you agree. Anyway, we know the test which our RF undergoes as it ascends after launch: if its effective vertical frequency is less than the local critical frequency, it will be contained by the ionosphere and if not, it will go past the F-layer peak and be lost. The propagation prediction business has to do with how that test is carried out - to what approximation or detail the test is made and with what sort of model of the ionosphere. I've already mentioned the control point method in which the test is made at the first and last hops on a path. That method was developed back in WW-II, by Smith in the USA and Tremellen in the UK, and was based on the notion that if a path failed, it was usually at one end or the other. I pointed out that works well as long as any hops in the middle of the path do not have LOWER critical frequencies. Beyond that, you should remember that the method represented a great step forward at the time, even though it was when ionospheric mapping was in its infancy. So the control point method was based on an approximation and its use involved a database which was both limited and uncertain, at least at the outset. Nowadays, the database has improved quite a bit but still will undergo some revisions in the future as the Internation Reference Ionosphere is updated from time to time. I really don't know the details of the first uses of the control point method but I am familiar with some at the present time. For example, the pioneer program in amateur radio circles was MINIMUF, with source code first published in QST in December '82. That method used M-factors, numbers between 3 and 4, for division of the QRG to obtain EVF for comparison with critical frequencies at about 2,000 km from the ends of the path; for that, MINIMUF used a database founded on oblique ionospheric sounding. One can fault the source code of MINIMUF for not taking into account the earth's field, leaving out the equatorial anomaly and organizing the ionosphere only with geographic coordinates. Beyond that, the database was rather limited in scope. But MINIMUF caught the imagination of the amateur radio community and all sorts of accessories were attached to MINIMUF, ionospheric absorption and man-made noise, to mention just a few. MMINIMUF's shortcomings, the lack of geomagnetic control in the method and no consideration of radiation angle, placed it in a poor position to compete with other programs that came along and corrected those deficiencies. Here, I have in mind the work of Raymond Fricker of the BBC External Services. In the mid-80s, he published programs like MICROMUF and MAXIMUF which included the role of the geomagnetic field and put in radiation angles so one could compare MUF predictions for more than just the lowest mode. Somewhat later, the Germans introduced a program, FTZMUF2, that used a grid point method to obtain critical frequencies from the CCIR database and used interpolation to obtain the spatial and temporal data for making predictions. They went on to show that FTZMUF2 gave a better representation of the CCIR-Atlas data for 3000 km MUFs than did MINIMUF. Beyond that, they incorporated FTZMUF2 in their own MUF prediction program, MINIFTZ4. But Fricker used an entirely different approach when it came to the database for his calculations; he used mathematical functions to simulate the CCIR database, now in the International Reference Ionosphere. Then he used the functions to calculate foF2 at the midpoints of the first and last hops in his programs, MICROMUF 2+ and MAXIMUF, as in the control point method. Those were the propagation prediction programs available until the IONCAP program was brought down to a smaller size where it could be incorporated in a table-top computer. Then the effort was to make comparisons between the various programs and IONCAP, to see which one was in best agreement. Of course, that assumed that IONCAP was the proper standard for comparison, the best model. In any event, the upshot of the comparisons, made by Tad Sargent, W0MPG at NOAA in '87, was that Fricker's programs were closest in agreement with IONCAP, then came MINIFTZ4 and MINIMUF was the poorest. But, mind you, those programs involved some form of approximation and the database that IONCAP relied on was really unknown, except for being based on vertical sounding. All those comparisons were made using various paths, Boulder to London, to Hawaii, to Mexico City, you name it. But that really begged the basic question as to how well the underlying database compared with the International Reference Ionosphere (IRI), the best representation available at the present time. In that connection, I undertook a study of how the mathematical F-layer algorithm in Fricker's MAXIMUF compared with IRI, not just for a path or two but over the entire world. Thus, foF2 values were calculated at intervals of 5 deg in latitude and 5 degs in longitude from Fricker's mathematical functions and compared with corresponding values from IRI. That method showed where Fricker's values were low, where high and an overall measure of his methods. The result was that Fricker's method, when used to make a map of the F-region, gave good agreement over the entire globe with the values from IRI, point by point, but the agreement could even be improved considerably by the simple offset of 1 Mhz added to the foF2 values calculated by his methods. Put another way, Fricker's foF2 map was very much like the map from IRI, with details such as the islands of ionization showing up as well as various aspects of geomagnetic control, but the critical frequencies were a bit low. All in all, I found it amazing! And that approach proves to be just another way of testing F-layer algorithms, seeing if they can make a good ionospheric map or not. MINIFTZ4's algorithm gets good marks in that regard but with problems from its interpolation methods while MINIMUF's F-region map has little resemblance to a real ionosphere on a global scale. That accounts for some of its erratic predictions for DXing. Unfortunately, the F-layer algorithm of IONCAP is not available so comparisons with the IRI remain to be done. Perhaps some of the IONCAP developers will do that in the future. But whatever the outcome, IONCAP is still a valuable program and provides some of the other aspects of propagation prediction that are important. Thus, in addition to having methods for calculating MUFs, it deals with the range of values of critical frequencies resulting from the statistical variations in the sounding data. Here, I refer to statistical terms like the median as well as the upper and lower decile values of critical frequencies from the sounding data. In a propagation setting, the median value of the data at a particular hour during a month would be one such that half the observed values lie above it and half fall below it. If a median value is used in propagation calculations, one obtains what is termed the Maximum Useable Frequency (MUF) for the path. The upper and lower decile values of critical frequency have to do with the 90% and 10% limits. Thus, the upper decile value during a month of observation is a frequency which is exceeded only 10% of the time, 3 days, while the lower decile value during a month is a frequency which is exceeded 90% of the time, 27 days. When those values are used in propagation calculations, one then obtains the Highest Possible Frequency (HPF) and the Frequency of Optimum Transmission (FOT) for the path. A sample of that kind of calculation is given below (in MHz), for a path from Boulder, CO to St. Louis, MO in the month of January and when the SSN is 100: GMT FOT MUF HPF GMT FOT MUF HPF 1 10.7 13.6 17.4 13 6.4 7.5 8.4 3 7.4 9.6 12.0 15 13.0 15.3 17.1 5 5.7 6.9 8.7 17 16.6 19.3 22.0 7 6.1 7.4 9.7 19 18.1 21.1 24.0 9 6.5 8.0 9.4 21 17.7 20.6 23.5 11 5.0 6.1 7.2 23 15.9 18.5 21.1 Looking at those numbers, you can see that the HPF and FOT values lie about 15% above and below the MUF values. That should put you on notice; if the propagation program you use gives only MUF values, the real-time values for the ionosphere could differ by as much as +/-15%. And that is only from the statistical variations in the basic data; there are still the approximations in the method to worry about as well as geophysical disturbances. But those remarks apply mainly to the higher HF bands; down on 80 and 160 meters, ionization is not a concern on oblique paths. Instead, noise and ionospheric absorption limit what can be done. And propagation programs are useless for those bands as the main criterion is darkness along paths, not MUFs. But the role of the geomagnetic field is important and affects the modes that are possible. All that in due time. As for geophysical disturbances, those will be our main effort in Prop. 301 and need not concern us at this point. We are really concentrating on the undisturbed ionosphere and its properties or modes, variable though they may be. And while still talking about the IONCAP program, it is worthwhile to note that its methods deal not only with the statistics of F-layer ionization, through MUFs and the like, but also down lower where absorption and noise become have their origin. So IONCAP has F-region methods which give not only the availability of a path, the fraction of days in a month it is open on a given frequency, but also D-region methods which give the reliability of a mode, the fraction of time the signal/noise ratio exceeds the minimum required for the mode. This was not meant to be something just in praise of IONCAP but for me, it is the best that I have at my disposal. True, there are other programs based on it and you will have to judge for yourself whether those programs meet your requirements or not. You should read the reviews out there, in QST and the DX Magazine, to get a feeling for what they can offer you in your pursuit of DX. If possible, check with a user to see if the program matches your goals or needs for DXing. At this point, we've come to where ionospheric disturbances from the impact of the solar wind on the magnetosphere are of real importance. Needless to say, they add to the uncertainties that have been cited above. But in contrast to the statistical side of propagation, there are clues that help deal with the geophysical side of propagation. That will be our task in future sessions. However, this is a good time to pause and pull things together with questions, discussion and the like. So review what I've put before you, then think about what else you want to know or what I left out. Ask questions, of me or your friends. I will be here, available on e-mail, but will take a "Time Out" until November 5. 73, Bob, NM7M |
| Part III >> |