Propagation Prediction Programs: Their Development and Use

Carl Luetzelschwab K9LA

k9la@arrl.net http://mysite.verizon.net/k9la

PVRC Webinar - 20Oct09 - K9LA

What Were Going to Cover

How the ionosphere forms Measuring the ionosphere Solar-ionosphere correlation Variability of the ionosphere Sample prediction Understanding prediction outputs
This presentation will be on the PVRC website
visit http://www.pvrc.org/index.html click on the PVRC Webinars link at the top
PVRC Webinar - 20Oct09 - K9LA 2

How the Ionosphere Forms

PVRC Webinar - 20Oct09 - K9LA

The Atmosphere
Composition of the atmosphere Species that are important in the ionosphere
O (atomic oxygen) O2 (molecular oxygen) N2 (molecular nitrogen) NO (nitric oxide) ionization potential 13.61 eV 12.08 eV 15.58 eV 9.25 eV

78.1% nitrogen, 20.9% oxygen, 1% other gases

maximum wavelength 91.1 nm 102.7 nm 79.6 nm 134 nm

Maximum wavelength is longest wavelength of radiation that

can cause ionization

10.7 cm = 107,000,000 nm

Related to ionization potential through Plancks Constant 10.7 cm solar flux doesnt ionize anything It is a proxy (substitute) for the true ionizing radiation

The sunspot number is also a proxy Visible light = 400 to 700 nm

PVRC Webinar - 20Oct09 - K9LA

True Ionizing Radiation

As radiation progresses down through the atmosphere,
it is absorbed by the aforementioned species in the process of ionization
Energy reduced as it proceeds lower Need higher energy radiation (shorter wavelengths) to get lower

True ionizing radiation can be summarized as follows

.1 to 1 nm and 121.5 nm for the D region

121.5 nm (Lyman- hydrogen spectral line) is the result of a

minimum in the absorption coefficient of O2 and N2
It goes through the higher altitudes easily, and ionizes NO at lower altitudes to give us daytime absorption Sunspots and 10.7 cm solar flux are proxies for the true ionizing radiation
5

1 to 10 nm for the E region 10 to 100 nm for the F2 region

PVRC Webinar - 20Oct09 - K9LA

Measuring the Ionosphere

PVRC Webinar - 20Oct09 - K9LA

Introduction to Ionosondes
To make predictions, you need a model of the

ionosphere Model developed from ionosonde data Most ionosondes are equivalent to swept-frequency radars that look straight up

There are also oblique ionosondes

Transmitter and receiver separated Evaluate a specific path

Co-located transmitter and receiver Also referred to as vertical ionosondes or vertically-incident ionosondes

PVRC Webinar - 20Oct09 - K9LA

What Does an Ionosonde Measure?

It measures the time for a
wave to go up, to be turned around, and to come back down Thus the true measurement is time, not height This translates to virtual height assuming the speed of light and mirror-like reflection The real wave does not get as high as the virtual height
An ionosonde measures time of flight, not altitude, at each frequency
PVRC Webinar - 20Oct09 - K9LA 8

http://digisonde.haystack.edu

Sample Ionogram
daytime data

Red is ordinary wave, green

foF2 foF1 foE

fxF2

electron density profile

is extraordinary wave Critical frequencies are highest frequencies that are returned to Earth from each region at vertical incidence Electron density profile is derived from the ordinary wave data (along with a couple assumptions about region thickness)
Electron density anywhere in the ionosphere is equivalent to a plasma frequency through the equation fp (Hz) = 9 x N1/2 with N in electrons/m3 Note that we dont see layers with gaps in between
9

E region and F2 region have maximums in electron density F1 region is inflection point in electron density D region not measured Nighttime data only consists of F2 region and sporadic E due to TX ERP and RX sensitivity (limit is ~1.8 MHz) PVRC Webinar - 20Oct09 - K9LA

Characterizing the Ionosphere

Ionosphere is characterized in terms of critical
frequencies (foE, foF1, foF2) and heights of maximum electron densities (hmE, hmF2)
Easier to use than electron densities

Allows us to calculate propagation over oblique

paths
MUF(2000)E = foE x M-Factor for E region MUF(3000)F2 = foF2 x M-Factor for F2 region Rule of thumb: E region M-factor ~ 5, F2 region M-factor ~ 3
for more on the M-Factor, visit http://mysite.verizon.net/k9la/The_M-Factor.pdf
PVRC Webinar - 20Oct09 - K9LA 10

Solar Ionosphere Correlation

PVRC Webinar - 20Oct09 - K9LA

11

Whats the Correlation?

Many years of solar data and
worldwide ionosonde data collected The task of the propagation prediction developers was to determine the correlation between solar data and ionosonde data It would have been nice to find a correlation between what the ionosphere was doing on a given day and what the Sun was doing on the same day
PVRC Webinar - 20Oct09 - K9LA ionosonde data solar data

???

12

But That Didnt Happen

http://www.solen.info/solar/
MUF(3000)F2 over Wallops Island (VA) Ionosonde at 1700 UTC
25 20 15

August 2009

MHz
10 5 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31

day of August 2009

No correlation between daily values

August 2009

Zero sunspots Constant 10.7 cm flux

Indicates there are other factors in

Low of 11.6 MHz on August 14 High of 21.5 MHz on August 16

determining the ultimate ionization Comment about CQ WW Phone 2007, 2008

13

PVRC Webinar - 20Oct09 - K9LA

So Now What?
R2 = .0615

Not too good - the developers were forced to come up with a statistical model over a months time frame

R2 = .8637

Good smoothed solar flux (or smoothed sunspot number) and monthly median parameters

PVRC Webinar - 20Oct09 - K9LA

14

day 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

How Do You Determine the raw data Monthly Median?

foF2 5.4 4.3 4.8 4.6 4.7 4.6 4.8 4.4 4.4 0 4.2 4.9 4.2 4.6 4.5 4.9 0 4.4 5.2 4.8 4.9 4.9 4.8 4.7 4.4 4.4 4.3 4.8 4.9 4.9 4.6

put foF2 in ascending order median implies 50%

day 11 13 2 27 8 9 18 25 26 15 4 6 14 31 5 24 3 7 20 23 28 12 16 21 22 29 30 19 1

foF2 4.2 4.2 4.3 4.3 4.4 4.4 4.4 4.4 4.4 4.5 4.6 4.6 4.6 4.6 4.7 4.7 4.8 4.8 4.8 4.8 4.8 4.9 4.9 4.9 4.9 4.9 4.9 5.2 5.4

half of the values below median median

half of the values above median

Variation about the median follows a Chisquared distribution, thus probabilities can be calculated (more on this later)

PVRC Webinar - 20Oct09 - K9LA

15

Correlation Between SF and SSN

Smoothed solar flux 12 = 63.75 + 0.728 R12 + 0.00089 (R12)2 Smoothed sunspot number R12 = (93918.4 + 1117.3 12)1/2 406.37 Using these equations to convert between daily solar flux and daily sunspot number results in a lot of uncertainty PVRC Webinar - 20Oct09 - K9LA 16

Variability of the Ionosphere

PVRC Webinar - 20Oct09 - K9LA

17

What Causes this Variability?

Rishbeth and Mendillo, Journal of Atmospheric and Solar-Terrestrial

Physics, Vol 63, 2001, pp 1661-1680
Looked at 34 years of foF2 data Used data from 13 ionosondes Day-to-day daytime variability (std dev/monthly mean) = 20%

Solar ionizing radiation contributed about 3% Solar wind, geomagnetic field activity, electrodynamics about 13% Neutral atmosphere about 15% [20%]2 = [3%]2 + [13%]2 + [15%]2
PVRC Webinar - 20Oct09 - K9LA 18

Is the Ionosphere In Step?

3000 km MUF
over Millstone Hill and Wallops Island Separated by 653 km = 408 miles Several periods show up when the ionosphere was going opposite ways Worldwide ionosphere not necessarily in step

PVRC Webinar - 20Oct09 - K9LA

19

We Dont Have Daily Predictions

Day-to-day variability just too great We have a good understanding of the solar influence Were beginning to better understand the geomagnetic field
influence

We are lacking a good understanding of how events in the

lower atmosphere couple up to the ionosphere
This is a major reason why prediction programs dont cover 160m (along with the effect of the Earths magnetic field through the electron-gyro frequency)

Its a bit more than just low K = good and high K = bad

160m RF doesnt get as high into the ionosphere as the higher

frequencies

Doesnt help that ionosondes dont measure the lower ionosphere especially at night when we chase DX on the low bands
PVRC Webinar - 20Oct09 - K9LA 20

Summary So Far
Our propagation prediction programs are based on the correlation
between a smoothed solar index and monthly median ionospheric parameters
Monthly median parameters can be represented in different ways

Database of numerical coefficients Equations International Reference Ionosphere Our predictions programs are pretty accurate when the geomagnetic field is quiet Real-time MUF maps seen on the web are kind of a misnomer
If they use a smoothed solar index, then theyre monthly median MUFs If they use todays solar flux or todays sunspot number (maybe even with todays A index), I dont know what they are!

Now its time to run a sample prediction

We dont have daily predictions

PVRC Webinar - 20Oct09 - K9LA 21

Sample Prediction

PVRC Webinar - 20Oct09 - K9LA

22

K9LA to ZF
Latitudes / longitudes October 2004 Antennas Power
K9LA = 41.0N / 85.0W ZF = 19.5N / 80.5W Smoothed sunspot number ~ 35 (smoothed solar flux ~ 91) Are we ever going to see that again? Small Yagis on both ends = 12 dBi gain 1000 Watts on both ends 20m, 17m, 15m on the Short Path When you download VOACAP (comes with ICEPAC and REC533), read the Technical Manual and Users Manual lots of good info
PVRC Webinar - 20Oct09 - K9LA 23

Bands and Path

Well use VOACAP

VOACAP Input Parameters

Method
Controls the type of program analysis and the predictions performed Recommend using Method 30 (Short\Long Smoothing) most of the time Methods 1 and 25 helpful for analysis of the ionosphere

Coefficients

CCIR (International Radio Consultative Committee)

URSI (International Union of Radio Scientists)

Shortcomings over oceans and in southern hemisphere Most validated

Rush, et al, used aeronomic theory to fill in the gaps Groups

Month.Day

10.00 means centered on the middle of October 10.05 means centered on the 5th of October
Defaults to URSI coefficients PVRC Webinar - 20Oct09 - K9LA 24

VOACAP Input Parameters

System
Noise Min Angle Req Rel Req SNR Multi Tol Multi Del default is residential 1 degree (emulate obstructions to radiation) default is 90% 48 dB in 1 Hz (13 dB in 3 KHz: 90% intelligibility) default is 3 dB default is .1 milliseconds

Fprob

Multipliers to increase or reduce MUF

Default is 1.00 for foE, foF1, foF2 and 0.00 for foEs

For more details on setting up and running VOACAP, either visit http://lipas.uwasa.fi/~jpe/voacap/ by Jari OH6BG (lots of good info) or http://mysite.verizon.net/k9la/Downloading_and_Using_VOACAP.PDF
PVRC Webinar - 20Oct09 - K9LA 25

Prediction Printout
13.0 20.9 1F2 10.0 8.6 347 0.50 123 28 -93 -168 75 -27 0.90 0.00 1.00 25.0 13.1 26.8 14.3 12.0 12.0 75 14.1 1F2 4.7 8.4 222 0.99 112 36 -82 -163 80 -32 1.00 0.00 1.00 8.4 4.9 12.6 7.2 12.0 12.0 80 18.1 1F2 5.5 8.4 240 0.83 113 37 -83 -166 83 -35 1.00 0.00 1.00 12.5 5.3 15.7 7.8 12.0 12.0 83 21.2 1F2 10.0 8.6 347 0.46 124 27 -94 -168 74 -26 0.89 0.00 1.00 25.0 14.0 26.8 15.2 12.0 12.0 74 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 FREQ MODE TANGLE DELAY V HITE MUFday LOSS DBU S DBW N DBW SNR RPWRG REL MPROB S PRB SIG LW SIG UP SNR LW SNR UP TGAIN RGAIN SNRxx 26

PVRC Webinar - 20Oct09 - K9LA

Understanding Prediction Outputs

PVRC Webinar - 20Oct09 - K9LA

27

time

Focus on 15m at 1300 UTC

14.1 1F2 4.7 8.4 222 0.99 112 36 -82 -163 80 -32 1.00 0.00 1.00 8.4 4.9 12.6 7.2 12.0 12.0 80 18.1 1F2 5.5 8.4 240 0.83 113 37 -83 -166 83 -35 1.00 0.00 1.00 12.5 5.3 15.7 7.8 12.0 12.0 83 21.2 1F2 10.0 8.6 347 0.46 124 27 -94 -168 74 -26 0.89 0.00 1.00 25.0 14.0 26.8 15.2 12.0 12.0 74 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 FREQ MODE TANGLE DELAY V HITE MUFday LOSS DBU S DBW N DBW SNR RPWRG REL MPROB S PRB SIG LW SIG UP SNR LW SNR UP TGAIN RGAIN SNRxx 28

13.0 20.9 1F2 10.0 monthly 8.6 median 347 MUF 0.50 123 28 MUFday -93 for 15m -168 75 signal -27 power 0.90 0.00 1.00 25.0 13.1 26.8 14.3 12.0 12.0 75

PVRC Webinar - 20Oct09 - K9LA

15m Openings at 1300 UTC

MUF Plot
30 25

20.9 MHz (monthly

median)
Enough ionization on half the days of the month Enough ionization on .46 x 31 = 14 days of the month Enough ionization every day of the month

MHz

20 15 10 5 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

21.2 MHz

14 MHz and below 24.9 MHz 28.3 MHz

MUFday
(multiply by 31 to get the number of days in the month)

Enough ionization on 1 day of the month

Not enough ionization on any day

We cant predict which days are the good days

PVRC Webinar - 20Oct09 - K9LA

29

15m Signal Power

-94 dBW (monthly median) = -64 dBm Assume
typical of receivers Ive measured
S9 = -73 dBm (50 microvolts into 50) one S-unit = 5 dB
except below S3 or so its only a couple dB per S-unit
S9+10 -63 dBm S9 -73 dBm S8 -78 dBm S7 -83 dBm S6 -88 dBm S5 -93 dBm S4 -98 dBm S3 -103 dBm S2 -108 dBm S1 -113 dBm

-64 dBm = 10 dB over S9 Variability about the monthly median from ionospheric texts (for

example, Supplement to Report 252-2, CCIR, 1978) Signal power could be from one S-unit higher to two S-units lower on any given day on this path

Rule of thumb actual signal power for any path could be from a
Dont make assumptions about your S-meter measure it
PVRC Webinar - 20Oct09 - K9LA

S9 to 15 over 9 for this path

couple S-units higher to several S-units lower than median on any given day

30

VOACAP vs W6ELProp

PVRC Webinar - 20Oct09 - K9LA

31

Whats Different with W6ELProp?

Underlying concept is still the correlation between a

smoothed solar parameter and monthly median ionospheric parameters For foF2, W6ELProp uses equations developed by Raymond Fricker of the BBC

W6ELProp rigorously calculates signal strength using

CCIR methods
VOACAP calibrated against actual measurements

VOACAP uses database of numerical coefficients to describe worldwide ionosphere Another option is IRI (PropLab Pro)

For more details on setting up and running W6ELProp, visit http://mysite.verizon.net/k9la/Downloading_and_Using_W6ELProp.PDF

PVRC Webinar - 20Oct09 - K9LA 32

Comparison - MUF
K9LA to ZF, October 2003
30 25

Close, but there are

MUF, MHz

20 15 10 5 0 1 3 5 7 9 11 13 15 17 19 21 23

W6ELProp VOACAP

differences especially around sunrise and sunset The difference is how the F2 region is represented in the model
VOACAP is database of numerical coefficients Frickers equations in W6ELProp simplified this to 23 equations (1 main function + 22 modifying functions)

time, UTC

PVRC Webinar - 20Oct09 - K9LA

33

Comparison Signal Strength

K9LA to ZF, October 2003
-40 -45 -50 -55 -60 -65 -70 -75 -80 -85 -90 1 3 5 7 9 11 13 15 17 19 21 23

In general
W6ELProp VOACAP

W6ELProp predicts higher signal strengths VOACAP is more realistic with respect to signal strength

signal power, dBm

time, UTC

PVRC Webinar - 20Oct09 - K9LA

34

The Mapping Feature in W6ELProp

This is a great tool for low band operating Recently on the topband reflector SM2EKM told

of a 160m QSO with KH6AT in late December at local noon Without digging any farther, this sounds like a very unusual QSO

PVRC Webinar - 20Oct09 - K9LA

35

SM to KH6 in Dec at SM Noon

Path on SM end is
perpendicular to the terminator
RF from SM encounters the D region right around the terminator But the solar zenith angle is high

Rest of path is in

darkness A index and K index are important for this over-the-pole path
Were at zero for a couple days

PVRC Webinar - 20Oct09 - K9LA

36

Summary
We dont have daily predictions Predictions are statistical over a months time frame All prediction software is based on the correlation between a
smoothed solar index and monthly median ionospheric parameters Many good programs out there with different presentation formats and different bells and whistles
VOACAP predictions to/from more than 170 locations Only give signal strength and dont include the WARC bands Choose the one you like the best
VOACAP considered the standard Several use the VOACAP engine Dont forget the predictions in the 21st Edition of the ARRL Antenna Book CD by Dean N6BV

Interested in validating a prediction?

Visit mysite.verizon.net/k9la/Validating_Propagation_Predictions.pdf

PVRC Webinar - 20Oct09 - K9LA

37

Q&A
This PowerPoint presentation is at http://mysite.verizon.net/k9la

And stay tuned for another PVRC propagation webinar the topic will be Disturbances to Propagation

PVRC Webinar - 20Oct09 - K9LA

38