propagation prediction programs - development use

7/30/2019 Propagation Prediction Programs - Development Use

1/38

PVRC Webinar - 20Oct09 - K9LA 1

Propagation Prediction Programs:

Their Development and Use

Carl Luetzelschwab K9LA

[email protected]://mysite.verizon.net/k9la


2/38


What Were Going to Cover

How the ionosphere forms Measuring the ionosphere

Solar-ionosphere correlationVariability of the ionosphere Sample prediction

Understanding prediction outputsThis presentation will be on the PVRC website

visit http://www.pvrc.org/index.html

click on the PVRC Webinars link at the top


3/38


How the Ionosphere Forms


4/38


The Atmosphere Composition of the atmosphere

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

Species that are important in the ionosphere

Maximum wavelength is longest wavelength of radiation thatcan cause ionization Related to ionization potential through Plancks Constant

10.7 cm = 107,000,000 nm 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

ionization potential maximum wavelength

O (atomic oxygen) 13.61 eV 91.1 nm

O2 (molecular oxygen) 12.08 eV 102.7 nm

N2 (molecular nitrogen) 15.58 eV 79.6 nm

NO (nitric oxide) 9.25 eV 134 nm


5/38


True Ionizing Radiation

As radiation progresses down through the atmosphere,it is absorbed by the aforementioned species in theprocess of ionization

Energy reduced as it proceeds lower Need higher energy radiation (shorter wavelengths) to getlower

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 aminimum in the absorption coefficient of O2 and N2

It goes through the higher altitudes easily, and ionizes NO at loweraltitudes to give us daytime absorption

1 to 10 nm for the E region 10 to 100 nm for the F2 region Sunspots and 10.7 cm solar flux are

proxies for the true ionizing radiation


6/38


Measuring the Ionosphere


7/38


Introduction to Ionosondes

To make predictions, you need a model of theionosphere

Model developed from ionosonde data

Most ionosondes are equivalent to swept-frequencyradars that look straight up Co-located transmitter and receiver Also referred to as vertical ionosondes or vertically-incident

ionosondes

There are also oblique ionosondes Transmitter and receiver separated Evaluate a specific path


8/38


What Does an Ionosonde Measure?

It measures the time for awave to go up, to be turnedaround, and to come backdown

Thus the true measurement istime, not height

This translates to virtual heightassuming the speed of lightand mirror-like reflection

The real wave does not get ashigh as the virtual height

An ionosonde measures time of flight,not altitude, at each frequency


9/38


Sample Ionogram Red is ordinary wave, green

is extraordinary wave

Critical frequencies arehighest frequencies that arereturned to Earth from eachregion at vertical incidence

Electron density profile isderived from the ordinarywave data (along with acouple assumptions aboutregion thickness) Electron density anywhere

in the ionosphere is

equivalent to a plasmafrequency through theequation fp (Hz) = 9 x N

1/2

with N in electrons/m3 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)

foE

foF1

foF2fxF2

electron density profile

daytime data

http://digisonde.haystack.edu

Note that we dont seelayers with gaps in between


10/38


Characterizing the Ionosphere

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

Allows us to calculate propagation over obliquepaths 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


11/38


SolarIonosphere Correlation


12/38


Whats the Correlation?

Many years of solar data andworldwide ionosonde datacollected

The task of the propagationprediction developers was todetermine the correlation betweensolar data and ionosonde data

It would have been nice to find a

correlation between what theionosphere was doing on a givenday and what the Sun was doingon the same day

solar data

ionosonde data

???


13/38


But That Didnt Happen

August 2009

August 2009 Zero sunspots Constant 10.7 cm flux

No correlation between daily values Low of 11.6 MHz on August 14 High of 21.5 MHz on August 16

Indicates there are other factors indetermining the ultimate ionization

Comment about CQ WW Phone 2007, 2008

MUF(3000)F2 over Wallops Island (VA) Ionosonde at 1700 UTC

0

5

10

15

20

25

1 3 5 7 9 11 1 3 15 1 7 1 9 21 2 3 25 2 7 2 9 31

day of August 2009

M

Hz

http://www.solen.info/solar/


14/38


So Now What?

Not too good - the developerswere forced to come up witha statistical model over amonths time frame

Good smoothed solar flux(or smoothed sunspotnumber) and monthlymedian parameters

R2 = .0615

R2 = .8637


15/38


How Do You Determine theMonthly Median?

day foF2

1 5.4

2 4.3

3 4.8

4 4.6

5 4.7

6 4.6

7 4.8

8 4.4

9 4.4

10 0

11 4.212 4.9

13 4.2

14 4.6

15 4.5

16 4.9

17 0

18 4.4

19 5.2

20 4.8

21 4.922 4.9

23 4.8

24 4.7

25 4.4

26 4.4

27 4.3

28 4.8

29 4.9

30 4.9

31 4.6

day foF2

11 4.2

13 4.2

2 4.3

27 4.3

8 4.4

9 4.4

18 4.4

25 4.4

26 4.4

15 4.5

4 4.6

6 4.6

14 4.6

31 4.6

5 4.7

24 4.7

3 4.8

7 4.8

20 4.8

23 4.8

28 4.8

12 4.9

16 4.9

21 4.9

22 4.9

29 4.9

30 4.9

19 5.2

1 5.4

raw data

put foF2 inascending order

median implies50%

half of thevaluesbelowmedian

half of thevaluesabovemedian

median

Variation aboutthe median

follows a Chi-squareddistribution, thusprobabilities canbe calculated(more on this

later)


16/38


Correlation Between SF and SSN

Smoothed solar flux12 = 63.75 + 0.728 R12 + 0.00089 (R12)

2

Smoothed sunspot numberR12= (93918.4 + 1117.3 12)

1/2 406.37

Using these equations to convert betweendaily solar flux and daily sunspot numberresults in a lot of uncertainty


17/38


Variability of the Ionosphere


18/38


What Causes this Variability?

Rishbeth and Mendillo, Journal of Atmospheric and Solar-TerrestrialPhysics, 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


19/38


Is the Ionosphere In Step?

3000 km MUFover Millstone Hilland WallopsIsland

Separated by 653km = 408 miles

Several periodsshow up whenthe ionospherewas goingopposite ways

Worldwideionosphere notnecessarily instep


20/38


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


21/38


Summary So Far

Our propagation prediction programs are based on the correlationbetween a smoothed solar index and monthly median ionosphericparameters 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 geomagneticfield 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


22/38


Sample Prediction


23/38


K9LA to ZF

Latitudes / longitudes K9LA = 41.0N / 85.0W ZF = 19.5N / 80.5W

October 2004 Smoothed sunspot number ~ 35 (smoothed solar flux ~ 91)

Are we ever going to see that again? Antennas

Small Yagis on both ends = 12 dBi gain

Power 1000 Watts on both ends

Bands and Path 20m, 17m, 15m on the Short Path Well use VOACAP

When you download VOACAP (comes with ICEPAC and REC533),read the Technical Manual and Users Manual lots of good info


24/38


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) Shortcomings over oceans and in southern hemisphere Most validated

URSI (International Union of Radio Scientists) 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


25/38


VOACAP Input Parameters

System Noise default is residential

Min Angle 1 degree (emulate obstructions to radiation)

Req Rel default is 90% Req SNR 48 dB in 1 Hz (13 dB in 3 KHz: 90% intelligibility)

Multi Tol default is 3 dB

Multi Del 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 visithttp://lipas.uwasa.fi/~jpe/voacap/ by Jari OH6BG (lots of good info) orhttp://mysite.verizon.net/k9la/Downloading_and_Using_VOACAP.PDF


26/38


Prediction Printout13.0 20.9 14.1 18.1 21.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 FREQ

1F2 1F2 1F2 1F2 - - - - - - - - MODE

10.0 4.7 5.5 10.0 - - - - - - - - TANGLE

8.6 8.4 8.4 8.6 - - - - - - - - DELAY

347 222 240 347 - - - - - - - - V HITE

0.50 0.99 0.83 0.46 - - - - - - - - MUFday

123 112 113 124 - - - - - - - - LOSS

28 36 37 27 - - - - - - - - DBU-93 -82 -83 -94 - - - - - - - - S DBW

-168 -163 -166 -168 - - - - - - - - N DBW

75 80 83 74 - - - - - - - - SNR

-27 -32 -35 -26 - - - - - - - - RPWRG

0.90 1.00 1.00 0.89 - - - - - - - - REL

0.00 0.00 0.00 0.00 - - - - - - - - MPROB

1.00 1.00 1.00 1.00 - - - - - - - - S PRB

25.0 8.4 12.5 25.0 - - - - - - - - SIG LW

13.1 4.9 5.3 14.0 - - - - - - - - SIG UP

26.8 12.6 15.7 26.8 - - - - - - - - SNR LW

14.3 7.2 7.8 15.2 - - - - - - - - SNR UP

12.0 12.0 12.0 12.0 - - - - - - - - TGAIN

12.0 12.0 12.0 12.0 - - - - - - - - RGAIN

75 80 83 74 - - - - - - - - SNRxx


27/38


Understanding Prediction Outputs


28/38


Focus on 15m at 1300 UTC13.0 20.9 14.1 18.1 21.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 FREQ

1F2 1F2 1F2 1F2 - - - - - - - - MODE

10.0 4.7 5.5 10.0 - - - - - - - - TANGLE

8.6 8.4 8.4 8.6 - - - - - - - - DELAY

347 222 240 347 - - - - - - - - V HITE

0.50 0.99 0.83 0.46 - - - - - - - - MUFday

123 112 113 124 - - - - - - - - LOSS

28 36 37 27 - - - - - - - - DBU-93 -82 -83 -94 - - - - - - - - S DBW

-168 -163 -166 -168 - - - - - - - - N DBW

75 80 83 74 - - - - - - - - SNR

-27 -32 -35 -26 - - - - - - - - RPWRG

0.90 1.00 1.00 0.89 - - - - - - - - REL

0.00 0.00 0.00 0.00 - - - - - - - - MPROB

1.00 1.00 1.00 1.00 - - - - - - - - S PRB

25.0 8.4 12.5 25.0 - - - - - - - - SIG LW

13.1 4.9 5.3 14.0 - - - - - - - - SIG UP

26.8 12.6 15.7 26.8 - - - - - - - - SNR LW

14.3 7.2 7.8 15.2 - - - - - - - - SNR UP

12.0 12.0 12.0 12.0 - - - - - - - - TGAIN

12.0 12.0 12.0 12.0 - - - - - - - - RGAIN

75 80 83 74 - - - - - - - - SNRxx

monthlymedianMUF

time

MUFdayfor 15m

signalpower


29/38


15m Openings at 1300 UTC

20.9 MHz (monthlymedian) Enough ionization on

half the days of themonth

21.2 MHz

Enough ionization on.46 x 31 = 14 days ofthe month

14 MHz and below Enough ionization every

day of the month

24.9 MHz

Enough ionization on 1day of the month

28.3 MHz Not enough ionization

on any dayWe cant predict which days are the good days

MUF Plot

5

10

15

20

25

30

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

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

MH

z


30/38


15m Signal Power

-94 dBW (monthly median) = -64 dBm Assume

S9 = -73 dBm (50 microvolts into 50) one S-unit = 5 dB

typical of receivers Ive measured

except below S3 or so its only a couple dB per S-unit -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

S9 to 15 over 9 for this path Rule of thumb actual signal power for any path could be from a

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

Dont make assumptions about your S-meter measure it

S9+10 -63 dBmS9 -73 dBmS8 -78 dBmS7 -83 dBmS6 -88 dBmS5 -93 dBmS4 -98 dBmS3 -103 dBm

S2 -108 dBmS1 -113 dBm


31/38


VOACAP vs W6ELProp


32/38


Whats Different with W6ELProp?

Underlying concept is still the correlation between asmoothed solar parameter and monthly medianionospheric parameters

For foF2, W6ELProp uses equations developed byRaymond Fricker of the BBC VOACAP uses database of numerical coefficients to describe

worldwide ionosphere Another option is IRI (PropLab Pro)

W6ELProp rigorously calculates signal strength using

CCIR methods VOACAP calibrated against actual measurements

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


33/38


Comparison - MUF

K9LA to ZF, October 2003

0

5

10

15

20

25

30

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

time, UTC

MUF,MH

z

W6ELProp

VOACAP

Close, but there aredifferences especiallyaround sunrise andsunset

The difference is how theF2 region is representedin the model VOACAP is database of

numerical coefficients

Frickers equations inW6ELProp simplified

this to 23 equations (1main function + 22modifying functions)


34/38


Comparison Signal Strength

In generalW6ELProppredicts highersignal strengths

VOACAP ismore realisticwith respect tosignal strength

K9LA to ZF, October 2003

-90

-85

-80

-75

-70

-65

-60-55

-50

-45

-40

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

time, UTC

signalpower,d

Bm

W6ELProp

VOACAP


35/38


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


36/38


SM to KH6 in Dec at SM Noon

Path on SM end isperpendicular to theterminator RF from SM

encounters the Dregion right aroundthe terminator

But the solar zenithangle is high

Rest of path is indarkness

A index and K indexare important for thisover-the-pole path Were at zero for a

couple days


37/38


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 Dont forget the predictions in the 21st Edition of the ARRL Antenna

Book CD by Dean N6BV 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

Interested in validating a prediction? Visit mysite.verizon.net/k9la/Validating_Propagation_Predictions.pdf


38/38

PVRC W bi 20O t09 K9LA 38

Q & A

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

And stay tuned for another PVRCpropagation webinar the topic will be

Disturbances to Propagation

propagation prediction programs - development use

Documents