HERMANUS MAGNETIC OBSERVATORY Information and analysis … · times during the year. .....52 Figure 34. Cumulative probability of occurrence of 30 sec rate of TEC index (ROTI) over

Hermanus Magnetic

Observatory

Page 1

Classification: CONFIDENTIAL

Classification: CONFIDENTIAL DOC: 6021-0001-709-A1

HERMANUS MAGNETIC OBSERVATORY

Information and analysis of ionospheric conditions at proposed SKA sites

Doc No: 6021-0001-709-A1

Prepared by: BDL Opperman

PJ Cilliers

L-A McKinnell

Prepared for: F. Adam

Date: 19/07/2005

Hermanus Magnetic Observatory P O Box 32 HERMANUS 7200

Hermanus Magnetic

Observatory

Page 2



APPROVAL PAGE

Name Signature Date

BDL Opperman

Author, HMO

PJ Cilliers

Author, HMO

L-A McKinnell

Author, HMO

DISTRIBUTION LIST

Master Copy: HMO Configuration Control

Copy 1: F. Adam Fanaroff and Associates

Hermanus Magnetic

Observatory

Page 3



Table of Contents

Table of Contents ............................................................................................................. 3 List of Figures................................................................................................................... 4 List of Tables .................................................................................................................... 7 1 Scope ........................................................................................................................ 8 2 Aims and Objectives................................................................................................. 8 3 SKA antennas and GPS receiver locations............................................................... 8 4 Results .................................................................................................................... 10

4.1 Total Electron Content Variation ................................................................... 11 4.1.1 Scope ...................................................................................................... 11 4.1.2 Introduction ............................................................................................ 11 4.1.3 Representative data sets.......................................................................... 11 4.1.4 TEC Model ............................................................................................. 15 4.1.5 TEC variability over SKA sites .............................................................. 20 4.1.6 Variations of foF2................................................................................... 36 4.1.7 Discussion............................................................................................... 41

4.2 Scintillation..................................................................................................... 42 4.2.1 Scope ...................................................................................................... 42 4.2.2 Introduction ............................................................................................ 42 4.2.3 Data......................................................................................................... 44 4.2.4 Approach ................................................................................................ 45 4.2.5 Results .................................................................................................... 45 4.2.6 Discussion............................................................................................... 54

4.3 Large scale phenomena .................................................................................. 54 4.3.1 Introduction ............................................................................................ 54 4.3.2 Results .................................................................................................... 54 4.3.3 Conclusion .............................................................................................. 59

4.4 Energetic particle precipitation rate and proximity to the South Atlantic Anomaly ..................................................................................................................... 59

4.4.1 Introduction ............................................................................................ 59 4.4.2 Results .................................................................................................... 59

4.5 Equatorial Electrojet ....................................................................................... 61 4.5.1 Introduction ............................................................................................ 61 4.5.2 The Equatorial F-Region ........................................................................ 61 4.5.3 Electron Density Variations ................................................................... 63 4.5.4 TEC and IEC .......................................................................................... 64 4.5.5 Conclusion .............................................................................................. 65

5 References .............................................................................................................. 65 5.1.1 Bibliographic References ....................................................................... 65 5.1.2 Internet References ................................................................................. 66

6 Key Project Participants ......................................................................................... 67 6.1.1 Mr BDL Opperman. Co-ordinator......................................................... 67 6.1.2 Dr PJ Cilliers. ......................................................................................... 67 6.1.3 Dr LA McKinnell. .................................................................................. 67

Hermanus Magnetic

Observatory

Page 4



List of Figures

Figure 1. Location of proposed remote SKA antenna locations and nearest

available GPS receivers. ............................................................................. 9 Figure 2. Spatial distribution of dual frequency GPS receivers near the hub of the

proposed Karoo III site (red triangles). These GPS receivers were used in the study of TEC variability. ................................................................ 10

Figure 3. Year maps showing all the GPS receivers of which data was used for each day of the TEC variability analysis by means of SHM over the SKA hub for the period 2000 to 2004. Each element represents one day of data from one receiver.. The colour bar at the bottom of the graph shows the station codes of the GPS receivers used. ....................... 13

Figure 4. TEC values at comparable latitudes from three different GPS receivers. Similar, but time-shifted TEC values are obtained over similar latitudes. 1 TEC Unit equals 1016 electrons.m-2. This illustrates the justification of using data from these three GPS receivers to represent long-term variation in TEC over Madagascar and Mauritius where there are GPS receivers from which data was readily available............... 14

Figure 5. IPP ground trace for slant TEC values obtained for a single GPS receiver (Sutherland, South Africa) for day 170 of 2000. Note relative dense distribution of data points. .............................................................. 16

Figure 6. Diurnal TEC calculated using SHM coefficients estimated from a GPS receiver co-located with an ionosonde at Grahamstown, South Africa. The difference between Ionosonde-TEC and GPS-TEC is attributed to the contribution of the plasmasphere and top-side model uncertainties within the ionosonde data. ........................................................................ 17

Figure 7. Example of GPS signal loss of lock. A significant number of lost observations within the blue ellipses, attributed to a hardware error, were lost during 10-14 UT (330-390° (30°) Sun-fixed longitude ) during summer months of 2000-2001 for the Sutherland GPS receiver. . 18

Figure 8. Example of unreliable TEC value obtained with insufficient data; Data loss typically resulted from signal loss when a single GPS receiver is used. Such large TEC values were omitted in the final data analysis. Normal maximum daytime TEC at SKA hub during the maximum of the solar cycle peaked at about 60 TECU................................................. 19

Figure 9. Dense distribution of TEC IPP ground trace points using data from nine GPS receivers. Signal loss of lock was no longer a problem from 2002 onwards............................................................................................ 19

Figure 10 (a) to (e). Diurnal and seasonal TEC variation at proposed SKA Hub. Peak TEC values decreased as the solar activity level reduced from that near the solar cycle peak in 2000 to that near the solar cycle minimum in 2004...................................................................................... 23

Figure 11. Mean monthly TEC variation with error bars showing 1 standard deviation for 2000-2004 . Note the decrease in the mean TEC values with the decline in the solar activity from 2000 to 2004. ......................... 24

Hermanus Magnetic

Observatory

Page 5



Figure 12. A representation of the diurnal and seasonal TEC variation over Namibia, Botswana and Mozambique for (a) 2003 and (b) 2004. ........... 26

Figure 13. Mean monthly TEC variation with 1 standard deviation for Thohoyandou during 2003-2004. ............................................................. 26

Figure 14. The diurnal and seasonal TEC variation over Lusaka, Zambia for (a) 2002, (b) 2003 and (c) 2004. .................................................................... 28

Figure 15. Mean monthly TEC variation with 1 standard deviation error bars for Zambia for 2001-2003. ............................................................................. 29

Figure 16. A representation of the diurnal and seasonal TEC variation over Madagascar and Mauritius using data from the Reunion GPS receiver for (a) 2003 and (b) 2004.......................................................................... 30

Figure 17. Mean monthly TEC variation with 1 standard deviation error bars for Reunion (representing Mauritius and Madagascar) for 2003-2004.......... 31

Figure 18. The diurnal and seasonal TEC variation over Malindi, Kenya for (a) 2001, (b) 2002 and (c) 2003. .................................................................... 32

Figure 19. Mean monthly TEC variation with 1 standard deviation for Kenya for 2001-2003. ................................................................................................ 33

Figure 20. Diurnal and seasonal TEC variation over Gabon (representing Ghana) for (a) 2001, (b) 2002, (c) 2003 and (d) 2004........................................... 35

Figure 21. Mean monthly TEC variation with 1 standard deviation for Gabon (representing Ghana) for 2001-2004. Note decrease of TEC with decreasing solar activity. .......................................................................... 36

Figure 22. An illustration of the correlation between foF2 and TEC for the years 2001 (top panel) and 2004 (bottom panel), and for Louisvale (left panel) and Grahamstown (right panel). .................................................... 38

Figure 23. The Grahamstown measured foF2 and LAM model predicted foF2 values for the period 1973 to 2004, illustrating the extent of the database as well as the ionospheric variation. .......................................... 39

Figure 24. Contour plots showing the predicted noon local time variation of foF2 over Southern Africa for a solar maximum year 1980 at 4 times during the year. The foF2 contour values in MHz. ............................................. 40

Figure 25. Contour plots showing the predicted noon local time variation of foF2 over Southern Africa for a solar minimum year 1976 at 4 times during the year. The foF2 contour values in MHz. ............................................. 41

Figure 26. Latitudinal variation of diurnal TEC variation for Spring Equinox, September 2002. Highest TEC values are observed at latitudes nearest to the equatorial anomaly (Kenya), while the peak TEC value at Zambia is almost the same as at the SKA hub...................................... 42

Figure 27. Scintillation index S4 as predicted by WBMOD ionospheric scintillation model along the 20ºE meridian over the SKA hub for SSN=0....................................................................................................... 46

Figure 28. Scintillation index S4 as predicted by WBMOD ionospheric scintillation model along the 20ºE meridian over the SKA hub for SSN=75..................................................................................................... 47

Figure 29. Scintillation index S4 as predicted by WBMOD ionospheric scintillation model along the 20ºE meridian over the SKA hub for SSN=150................................................................................................... 47

Hermanus Magnetic

Observatory

Page 6



Figure 30. Rate of TEC scintillation index and Cumulative Frequency of Occurrence for all GPS satellites observed from the equatorial GPS station at Malindi, Kenya (40.19ºE, 3.00ºS) on 22 September 2004........ 49

Figure 31. Elevation and ROTI plots for all satellites observed during a day for a solar maximum year (2001) at 4 times during the year. Each one-day plot shows the occurrence of scintillation over Southern Africa as measured from the Sutherland GPS site. .................................................. 50

Figure 32. Elevation and ROTI plots for all satellites observed during a day for a solar minimum year (2004) at 4 times during the year. Each one-day plot shows the occurrence of scintillation over Southern Africa as measured from the Sutherland GPS site. Note that the scale on the ROTI axis of the 2004 plots has a maximum at 0.4 TECU/min while the 2001 plots have their maximum at 2 TECU/min. The cause of the gap in the autumn equinox data is not known. ......................................... 51

Figure 33. Cumulative probability of occurrence of 30 sec rate of TEC index (ROTI) the level of scintillation over Southern Africa as measured from the Sutherland GPS site for a solar minimum year (2004) at 4 times during the year. ............................................................................... 52

Figure 34. Cumulative probability of occurrence of 30 sec rate of TEC index (ROTI) over Southern Africa as measured from the Sutherland GPS site for a solar minimum year (2004) at 4 times during the year. Note that the scale on the ROTI axis of the 2004 plots has a maximum at 0.25 TECU/min while the 2001 plots have their maximum at 1 TECU/min. ............................................................................................... 53

Figure 35. The percentage occurrence of Spread F at the Grahamstown (33.3°S, 26.5°E) station for the years 2001 (top panel) and 2004 (bottom panel). ....................................................................................................... 55

Figure 36. The percentage occurrence of Spread F at the Louisvale (28.5°S, 21.2°E) station for the years 2001 (top panel) and 2004 (bottom panel). ....................................................................................................... 56

Figure 37. The percentage occurrence of TIDs at the Grahamstown (33.3°S, 26.5°E) station for the years 2001 (top panel) and 2004 (bottom panel). ....................................................................................................... 57

Figure 38. The percentage occurrence of TIDs at the Louisvale (28.5°S, 21.2°E) station for the years 2001 (top panel) and 2004 (bottom panel)............... 58

Figure 39. Plots of energetic particle precipitation over the world, showing (a) proton and (b) electron flux predictions [SPENVIS] ................................ 60

Figure 40. Sun-synchronous view of the H and Z components of the EEJ magnetic effects at a fixed local time (12h00 LT) at an altitude of 450 km. ∆H contour interval is 10 nT for a minimum which depends on the longitude, it is about -50 nT in Africa, -70 nT in South America and -30 nT in India. ∆Z contour interval is 10 nT for a minimum of about -30 nT and a maximum of 30 nT in Africa (Doumouya [2003])............... 62

Hermanus Magnetic

Observatory

Page 7



List of Tables

Table 1. GPS receivers used in TEC variation calculation......................................... 12 Table 2. Periods of identified spread-F in Grahamstown ionosonde data

nearsolstices and equinoxes of the years 2001 and 2004. ........................ 48

Hermanus Magnetic

Observatory

Page 8



1 Scope

This document presents the results of a study on characterising the ionosphere above the proposed Square Kilometre Array (SKA) sites in support of the South African bid to host the SKA. The ionosphere limits the low end of the frequency band for SKA observations due to it being the source of phase and polarization rotation fluctuations at operational ferquencies below 1.5 GHz.

The study described in this document addresses the core requirements as set out in the document “Requirements for the Ionosphere Condition at the Square Kilometre Array (SKA) Radio Telescope Sites, Request for Information and Scope of Work (RfP [2005])”

The study was conducted by the Hermanus Magnetic Observatory (HMO) in collaboration with Rhodes University's Department of Physics and Electronics.

2 Aims and Objectives

The overall aim of this report is to provide detailed results and information arising from the study undertaken to analyse the ionospheric conditions at the SKA antenna sites shown in Figure 1and Figure 2.

3 SKA antennas and GPS receiver locations

Figure 1 shows the location of proposed central and remote SKA antenna locations and the nearest available geodetic grade dual frequency GPS receivers used for the analysis of the total electron content and its variability over the proposed sites. The 9 GPS receivers nearest to the SKA hub are shown as red triangles in Figure 1 and the 6 GPS receivers representative of the remote SKA antennas as green triangles. Table 1 gives the locations of each of the GPS receiver locations.

The heavy red dotted line south of the equator in Figure 1 indicates the southernmost extent of equatorial scintillation, and the heavy green dotted line to the south is the 95th percentile location of the most equator-ward extent of the auroral scintillation region.

The inner green circle centered on the SKA hub shows the intersection of the cone spanned by an elevation angle of 5 degrees above the horizon with the ionospheric F-layer at an altitude of 400 km. The outer green circle shows the intersection of the same cone with the plasmasphere at an altitude of 1000 km. Thus the range of interaction of all ray paths from the SKA-hub through the ionosphere falls well within the mid-latitudes, which is the most benign region for ionospheric scintillations on a global basis.

Hermanus Magnetic

Observatory

Page 9



-20 -10 0 10 20 30 40 50 60-60

-50

-40

-30

-20

-10

0

10

20

Longitude (Deg)

Latitude (Deg)

Magnetic Equator

Crest of equatorial anomaly

90th-percentile auroral oval

Malindi,Kenya Franceville,Gabon

Libreville,Gabon

Zambia

Reunion

Thohoyandou

Figure 2 shows the spatial distribution of dual frequency GPS receivers near the hub of the proposed Karoo III site (red triangles) used in the study of TEC variability together with the 5 spirals connecting the locations of the antennas of the Karoo III SKA site.

Figure 1. Location of proposed remote SKA antenna locations and nearest available GPS receivers.

Hermanus Magnetic

Observatory

Page 10



10 15 20 25 30 35-36

-34

-32

-30

-28

-26

-24

-22

-20

-18

-16

Longitude (Deg)

Latit

ude

(Deg

)

Kuruman

Upington Kimberley

Springbok

De Aar Calvinia

Graaff Reinett Sutherland

Beaufort West

Prieska

Graaff Reinett

Thohoyandou

Figure 2. Spatial distribution of dual frequency GPS receivers near the hub of the proposed Karoo III site (red triangles). These GPS receivers were used in the study of TEC variability.

4 Results

This section contains the results arising from the study as proposed in the Request for information and Scope of Work. Each subsection provides some background on the results obtained.

Hermanus Magnetic

Observatory

Page 11



4.1 Total Electron Content Variation

4.1.1 Scope

This section reports the analysis of the total electron content (TEC) variation over the core and remote sites of the SKA facility over several time-scales including diurnal, annual and variation with solar activity.

4.1.2 Introduction

The Total Electron Content (TEC) along a signal ray path through the ionosphere-plasmasphere region (80-35000 km) is an important indicator of the state and dynamics of the plasma medium. Therefore the variability of the TEC is proportional to the expected fluctuations in the phase delay and polarization rotation measures (RM) of radio astronomy signals

4.1.3 Representative data sets

4.1.3.1 Short term variation

Regular, short term variations such as diurnal and seasonal variations in the Total Electron Content (TEC) parameter of the ionosphere and plasmasphere was addressed by TEC measurements from dual frequency GPS receivers. The TEC represents the combination of the total electron content of the ionosphere over the range 80-1200 km and largest part of the total electron content of the plasmasphere over the range 1200-20200 km. The TEC is measured in TEC Units with 1 TECU equal to 1016 electrons per m2. The TEC was calculated from historic GPS data from the receiver networks in South Africa operated by the Chief Directorate Surveys and Mapping [CDSM) and from receivers in the International GPS Service [IGS] network near the proposed remote SKA antenna locations.

Historic GPS data for South Africa is readily available for some stations since 2000, but relatively few GPS receivers, with accessible historic data, are known to exist at proposed remote locations. A description of available historic GPS data for South Africa and proposed remote antenna locations is presented in Table 1 (RfP [2005]).

Insufficient data at remote antenna locations was addressed by using representative data from locations close to and of similar latitude as the desired location because of ionospheric latitudinal coupling. Similar, but time-shifted TEC variation values (due to longitude difference) is expected from this approach. To illustrate this approach, diurnal TEC variations for September 2003 Equinox obtained from three stations (Reunion (55.57°E, 21.12°S), Lusaka Zambia (28.31°E, 15.32°S) and Thohoyandou South Africa (30.38°E, 23.08°S)) were compared (see Figure 4).

No GPS data was available for Madagascar, Mauritius, Namibia, Botswana, Mozambique or Ghana. To address this data gap, historic data of GPS receivers in

Hermanus Magnetic

Observatory

Page 12



Reunion, Lusaka Zambia and Thohoyandou South Africa were used to represent mean latitudes of Namibia, Botswana and Mozambique (See Table 1, Figure 1 and Figure 4) The closest available data to Ghana was Libreville Gabon (9.67°E, 0.35°N).

Table 1. GPS receivers used in TEC variation calculation

GPS receivers Location

(Lon,Lat)

Representative of SKA antenna locations at

GPS data available since

South Africa SKA Hub centered on (21.39°E, 30.71°S)

Kuruman (kman) (23.43°E, 27.46°S) 412 km from SKA Hub 2002-04-01

Upington (uptn) (21.25°E, 28.41°S) 256 km from SKA Hub 2002-04-01

Kimberley (kley) (24.81°E, 28.74°S) 396 km from SKA Hub 2000-07-01

Springbok (sbok) (17.88°E, 29.67°S) 356 km from SKA Hub 2000-08-01

De Aar (dear) (23.99°E, 30.66°S) 248 km from SKA Hub 2000-09-01

Calvinia (calv) (19.76°E, 31.48°S) 177 km from SKA Hub 2000-09-01

Graaff Reinett(grnt) (24.53°E, 32.24°S) 348 km from SKA Hub 2002-11-01

Sutherland(suth) (20.81°E, 32.38°S) 193 km from SKA Hub 1998-04-15

Beaufort West(bwes) (22.57°E, 32.85°S) 263 km from SKA Hub 2003-10-01

Prieska (pska) (22.75°E, 29.67°S) 174 km from SKA Hub 2004-08-02

Zambia

Lusaka

(28.31°E, 15.32°S)

Zambia, Madagascar

2002-06-02

Thohoyandou (30.38°E, 23.08°S) Namibia, Botswana, Mozambique

2003-06-01

Reunion (55.57°E, 21.12°S) Mauritius, Madagascar, Mozambique

2003-06-19

Gabon

Franceville

Libreville

(13.55°E, 1.63°S)

(9.67°E, 0.35°N)

Ghana

2001-05-24

2000-03-31

Kenya

Malindi

(40.20°E, 3.00°S)

Kenya

1998-01-05

Hermanus Magnetic

Observatory

Page 13



Figure 3. Year maps showing all the GPS receivers of which data was used for each day of the TEC variability analysis by means of SHM over the SKA hub for the period 2000 to 2004. Each element represents one day of data from one receiver.. The colour bar at the bottom of the graph shows the station codes of the GPS receivers used.

Hermanus Magnetic

Observatory

Page 14



To further compensate for data gaps, foF2 values (plasma frequency of the peak electron density value of the F2 ionospheric layer) were modelled over the whole region of interest using a proven neural network-based global foF2 model [Oyeyemi 2004]. TEC is well correlated with foF2, as will be illustrated in paragraph 4.1.6 and hence modelled foF2 can be used to represent TEC where no GPS TEC data are available.

Compressed GPS data were obtained from the archives of the Chief Directorate Surveys and Mapping [CDSM] and from the International GPS Service [IGS]. CDSM data files are significantly larger than IGS due to the higher sampling rate (1s intervals compared to 30 s for IGS data). The IGS data sets were downloaded via ftp. CDSM data were transfered to large capacity hard drives (> 250GB) at CDSM and transported to the HMO.

0 5 10 15 20 25-5

0

5

10

15

20

25

30

35

40

Time of day [UT]

TEC

[TE

CU

]

Comparitive TEC variation values at Equinox (September 2003)

ThohoyandouZambiaReunion

Figure 4. TEC values at comparable latitudes from three different GPS receivers. Similar, but time-shifted TEC values are obtained over similar latitudes. 1 TEC Unit equals 1016 electrons.m-2. This illustrates the justification of using data from these three GPS receivers to represent long-term variation in TEC over Madagascar and Mauritius where there are GPS receivers from which data was readily available.

Regular, long-term variations in TEC (solar cycle dependence) and occurrence of short-temr ionospheric regularities such as spread F and travelling ionospheric disturbances (TIDs) were also investigated using archived ionosonde data from the Grahamstown (33.33°S, 26.50°E) and Louisvale (28.50°S, 21.20°E) ionosonde stations.

Hermanus Magnetic

Observatory

Page 15



4.1.4 TEC Model

TEC was quantified from the linear combination of code and phase pseudorange observations registered at each GPS receiver according to proven and published methods (Cilliers [2004], Hoffmann-Wellenhoff [2001]). A satellite elevation mask of 15º above the horizon was employed to reduce the noise due to multipath scattering. The raw GPS data was decimated to one minute intervals for the calculation of TEC values. TEC measurements were corrected for satellite and receiver differential clock biases (DCB’s) using daily satellite clock biases supplied by [CODE] and receiver biases estimated by a least squares method. A representative TEC value was interpolated at the proposed centre of the SKA hub using a Modified Single Layer Mapping (MSLM) function and Spherical Harmonic Analysis (SHM) techniques. These techniques are similar to that employed by [CODE] to construct Global Ionospheric Maps [GIM] (Schaer [1999]): The SHM expansion of TEC is given by

{ }0 0

( , ) [sin( )] sin( ) cos( )

Sun-fixed longitude= geographic latitude

Normalized associated Legendre functions, desired SHM coefficients

N n

nm nm nmn m

nm

nm nm

TEC P a m b m

Pa b

λ φ φ λ λ

λφ

= =

= +

=

==

∑∑

Slant TEC values calculated at the geographic locations of sub Ionospheric Pierce Points (IPP) (Figure 5) were used with the elevation-dependent cosec mapping function (from slant to vertical TEC) as inputs to derive the SHM coefficients anm, bnm and the receiver/satellite biases. The sun-fixed longitude coordinate system compensated for time-dependence of data. Figure 5 illustrate the IPP ground trace over twenty-four hours for a single receiver. Figure 6 shows the diurnal TEC estimated from this data. The estimated GPS-TEC values correspond well with ionosonde measurements (Figure 6).

Hermanus Magnetic

Observatory

Page 16



0 50 100 150 200 250 300 350 400-40

-38

-36

-34

-32

-30

-28

-26

-24

-22TEC distribution for Day 170 of 2000

Sun-fixed longitude [deg]

Latit

ude

[deg

]

Figure 5. IPP ground trace for slant TEC values obtained for a single GPS receiver (Sutherland, South Africa) for day 170 of 2000. Note relative dense distribution of data points.

Hermanus Magnetic

Observatory

Page 17



Figure 6. Diurnal TEC calculated using SHM coefficients estimated from a GPS receiver co-located with an ionosonde at Grahamstown, South Africa. The difference between Ionosonde-TEC and GPS-TEC is attributed to the contribution of the plasmasphere and top-side model uncertainties within the ionosonde data.

4.1.4.1 Model shortcomings and result validation.

GPS receivers sporadically experience loss of satellite signal lock, resulting in data loss as illustrated in Figure 7. This signal lock may be attributed to scintillation or hardware-related problems. For prolonged periods of data loss (several hours), the SHM-TEC solution becomes unreliable as illustrated in Figure 7 and Figure 8. Normal summer day-time TEC values typically peak around 60 TECU. During the analysis, suspect values were checked against geomagnetic activity (Kp index) of the day and signal loss occurrences. Obvious outliers were omitted in the final analysis. This problem was typically encountered during the summer months of 2000 and 2001 for South Africa when the only available GPS data in close proximity to the SKA hub was from the Sutherland (20.81°E, 32.38°S) receiver, which proved to have a hardware-related problem resulting in loss of data during the midday-hours as illustrated in Figure 7. The Sutherland receiver was replaced in February 2002. More receivers came online from

Hermanus Magnetic

Observatory

Page 18



2002, resulting in a denser data distribution (Figure 9) and consistently more reliable TEC was obtained.

0 50 100 150 200 250 300 350 400-40

-38

-36

-34

-32

-30

-28

-26

-24

-22TEC distribution for Day 358 of 2001

Sun-fixed longitude [deg]

Latit

ude

[deg

]

Figure 7. Example of GPS signal loss of lock. A significant number of lost observations within the blue ellipses, attributed to a hardware error, were lost during 10-14 UT (330-390° (30°) Sun-fixed longitude ) during summer months of 2000-2001 for the Sutherland GPS receiver.

Hermanus Magnetic

Observatory

Page 19



0 5 10 15 20 250

20

40

60

80

100

120

140TEC variability at GPS Receiver(s) in : SKA Hub Day 358 of 2001

Time [Decimal hours UT]

TEC

[TE

CU

]

Figure 8. Example of unreliable TEC value obtained with insufficient data; Data loss typically resulted from signal loss when a single GPS receiver is used. Such large TEC values were omitted in the final data analysis. Normal maximum daytime TEC at SKA hub during the maximum of the solar cycle peaked at about 60 TECU.

Figure 9. Dense distribution of TEC IPP ground trace points using data from nine GPS receivers. Signal loss of lock was no longer a problem from 2002 onwards.

Hermanus Magnetic

Observatory

Page 20



4.1.5 TEC variability over SKA sites

4.1.5.1 Seasonal and diurnal variation over South Africa

The location of the centre of the SKA hub in South Africa (21.39°E, 30.71°S) is used as a reference point to describe TEC variability. TEC was calculated at this location using historic GPS measurements registered at the nearest nine dual frequency GPS receivers in the Chief Directorate Surveys and Mapping (CDSM’s) network and at one receiver in the Hartebeesthoek Radio Astronomy Observatory (HartRAO) IGS network. Table 1 gives the locations of the GPS receivers and the distances between the centre of the hub and the nine receivers nearest to it.

Figure 10 (a) to (e) illustrate the diurnal (time of day) and seasonal (time of year) variation of TEC at the SKA hub in TECU for each of the five years 2000 to 2004.

Annual TEC variation at SKA Hub: South Africa 2000

Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

(a)

Hermanus Magnetic

Observatory

Page 21




Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

(b)

Hermanus Magnetic

Observatory

Page 22




Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

(c)


Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

(d)

Hermanus Magnetic

Observatory

Page 23




Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

(e)

Figure 10 (a) to (e). Diurnal and seasonal TEC variation at proposed SKA Hub. Peak TEC values decreased as the solar activity level reduced from that near the solar cycle peak in 2000 to that near the solar cycle minimum in 2004.

Hermanus Magnetic

Observatory

Page 24



Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec0

10

20

30

40

50

60

70

TEC

[TE

C u

nits

]

Mean monthly TEC variation at SKA Hub (21.4E, 30.7S) for 2001.


10

20

30

40

50

60

70

TEC

[TE

C u

nits

]



10

20

30

40

50

60

70

TEC

[TE

C u

nits

]



10

20

30

40

50

60

70

TEC

[TE

C u

nits

]



10

20

30

40

50

60

70

TEC

[TE

C u

nits

]


Figure 11. Mean monthly TEC variation with error bars showing 1 standard deviation for 2000-2004 . Note the decrease in the mean TEC values with the decline in the solar activity from 2000 to 2004.

Hermanus Magnetic

Observatory

Page 25



4.1.5.2 Namibia, Botswana and Mozambique (using Thohoyandou data)

No GPS data was available for Namibia, Botswana or Mozambique. Data from the CDSM GPS receiver at Thohoyandou, Limpopo Province, South Africa (30.38°E, 23.08°S) was used to represent TEC variation at the latitudes of the SKA antennas in Namibia (Nam-0 at 25.78°S, Nam-1 at 22.18°S and Nam-2 at 19.6°S) Botswana (Bot-0 at 25.72°S, Bot-1 at 24°S, Bot-2 at 21.35°S and Bot-4 at 19°S) and Mozambique (20°S). Data was available from Thohoyandou for the years 2003 and 2004. Contour plots illustrating the TEC variations are shown in Figure 12.

Annual TEC variation over Thohoyando 2003

Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

5

10

15

20

25

30

35

40

45

50

(a)

Hermanus Magnetic

Observatory

Page 26



Annual TEC variation over Thohoyando 2004

Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

5

10

15

20

25

30

35

40

45

50

(b)

Figure 12. A representation of the diurnal and seasonal TEC variation over Namibia, Botswana and Mozambique for (a) 2003 and (b) 2004.

Figure 13. Mean monthly TEC variation with 1 standard deviation for Thohoyandou during 2003-2004.


5

10

15

20

25

30

35

40

TEC

[TE

C u

nits

]

Mean monthly TEC variation for Thohoyandou (30.4E, 23.1S) for 2003.


5

10

15

20

25

30

35

40

TEC

[TE

C u

nits

]

Mean monthly TEC variation for Thohoyandou (30.4E, 23.1S) for 2004.

Hermanus Magnetic

Observatory

Page 27



Annual TEC variation over Zambia 2003

Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

4.1.5.3 Namibia, Botswana and Mozambique (using Zambia data)

Data from the IGS GPS receiver located at Lusaka, Zambia (28.31°E, 15.32°S) was used for the years 2002 to 2004. Results are shown in Figure 14.


Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

(a)

Hermanus Magnetic

Observatory

Page 28




Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

(c)

Figure 14. The diurnal and seasonal TEC variation over Lusaka, Zambia for (a) 2002, (b) 2003 and (c) 2004.

Hermanus Magnetic

Observatory

Page 29




10

20

30

40

50

60

70TE

C [T

EC

uni

ts]

Mean monthly TEC variation for Zambia (28.3E, 15.3S) for 2002.


10

20

30

40

50

60

70

TEC

[TE

C u

nits

]


Figure 15. Mean monthly TEC variation with 1 standard deviation error bars for Zambia for 2001-2003.

4.1.5.4 Madagascar and Mauritius (using Reunion data)

No GPS data was available for Madagascar or Mauritius, but an IGS GPS receiver was located at Reunion (55.57°E, 21.12°S) and logged data in 2003 and 2004. Results are shown in Figure 16. Results from the GPS receivers in Zambia and Thohoyandou representing Namibia, Botswana and Mozambique, are also representative of Reunion as illustrated in Figure 4.


10

20

30

40

50

60

70

TEC

[TE

C u

nits

]


Hermanus Magnetic

Observatory

Page 30



Annual TEC variation over Reunion 2003

Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

(a)

Annual TEC variation over Reunion 2004

Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

(b)

Figure 16. A representation of the diurnal and seasonal TEC variation over Madagascar and Mauritius using data from the Reunion GPS receiver for (a) 2003 and (b) 2004.

Hermanus Magnetic

Observatory

Page 31




5

10

15

20

25

30

35

40

TEC

[TE

C u

nits

]

Mean monthly TEC variation for Reunion (55.6E, 21.2S) for 2004.

Figure 17. Mean monthly TEC variation with 1 standard deviation error bars for Reunion (representing Mauritius and Madagascar) for 2003-2004.

4.1.5.5 Kenya

Data from the IGS GPS receiver located at Malindi, Kenya (40.20°E, 3.00°S) was used to illustrate the diurnal and seasonal TEC variations for the years 2001 to 2003. Results are shown in Figure 18.

Annual TEC variation over Kenya 2001

Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

70

80

90

100

(a)


5

10

15

20

25

30

35

40

TEC

[TE

C u

nits

]

Mean monthly TEC variation for Reunion (55.6E, 21.2S) for 2003.

Hermanus Magnetic

Observatory

Page 32




Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

70

80

90

100

(b)


Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

70

80

90

100

(c)

Figure 18. The diurnal and seasonal TEC variation over Malindi, Kenya for (a) 2001, (b) 2002 and (c) 2003.

Hermanus Magnetic

Observatory

Page 33




10

20

30

40

50

60

70

80

90

TEC

[TE

C u

nits

]

Mean monthly TEC variation for Kenya (40.2E, 3S) for 2001.


10

20

30

40

50

60

70

80

90

TEC

[TE

C u

nits

]


Figure 19. Mean monthly TEC variation with 1 standard deviation for Kenya for 2001-2003.

4.1.5.6 Ghana (using GPS data from Gabon)

Data from the IGS GPS receiver located at Franceville, Gabon (1.63°S, 13.55°E) was used to illustrate the diurnal and seasonal TEC variations for the proposed antenna location in Ghana for the years 2001 to 2004. Results are shown in Figure 20 (a)-(d).


10

20

30

40

50

60

70

80

90

TEC

[TE

C u

nits

]


Hermanus Magnetic

Observatory

Page 34



Annual TEC variation over Gabon 2001

Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

70

80

90

100

(a)


Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

70

80

90

100

(b)

Hermanus Magnetic

Observatory

Page 35




Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

70

80

90

100

(c)


Hour of day [UT]

Mon

th o

f yea

r

0 2 4 6 8 10 12 14 16 18 20 22 24

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0

10

20

30

40

50

60

70

80

90

100

(d)

Figure 20. Diurnal and seasonal TEC variation over Gabon (representing Ghana) for (a) 2001, (b) 2002, (c) 2003 and (d) 2004.

Hermanus Magnetic

Observatory

Page 36




10

20

30

40

50

60

70

80

90

TEC

[TE

C u

nits

]

Mean monthly TEC variation for Gabon (13.6E, 1.6S) for 2001.

Figure 21. Mean monthly TEC variation with 1 standard deviation for Gabon (representing Ghana) for 2001-2004. Note decrease of TEC with decreasing solar activity.

4.1.6 Variations of foF2

One of the most important parameters to quantify the electron density in the ionosphere is the value of the maximum electron density in the ionosphere (NmF2), a quantity which is directly related to the critical frequency of the F2 layer (foF2) through

The foF2 is accurately determined by means of ionosondes. There exists a large database of foF2 data for the Grahamstown, South Africa (33.33°S, 26.50°E) ionosonde


10

20

30

40

50

60

70

80

90

TEC

[TE

C u

nits

]



10

20

30

40

50

60

70

80

90

TEC

[TE

C u

nits

]



10

20

30

40

50

60

70

80

90

TEC

[TE

C u

nits

]


32(Hz) 9 2(electrons/m )foF NmF=

Hermanus Magnetic

Observatory

Page 37



station. This foF2 database is significantly larger on a time scale, spanning some 30 years, than the database of Southern African GPS TEC data which spans only 5 years. The units of foF2 are MHz. The TEC in turn is directly related to the peak electron density NmF2 which makes the largest contribution to TEC. Figure 22 shows the correlation between foF2 and TEC for the Grahamstown (33.33°S, 26.50°E) and Louisvale (28.50°S, 21.20°E) ionosonde stations as recorded at local noon for 2001 and 2004. The correlation between foF2 and TEC allows one to exploit the larger database to illustrate the variation in ionospheric behaviour over the course of a solar cycle. In addition, the ionospheric group of the HMO have been working on developing regional and global models for the prediction of ionospheric behaviour [Oyeyemi 2004]. In particular, the LAM model is a regional model for the prediction of the bottom-side electron density profile over South Africa. Figure 23 shows the variation of the measured foF2 data from 1973 to 2004 inclusive, with the predictions made by our in-house regional model overlaid. In addition to the LAM model, a neural network based global model for foF2 is being developed for the purpose of predicting foF2 at any point on the globe for given input parameters. This model will provide us with, in particular, the power to predict foF2 across the African continent to include areas that do not have an ionosonde or GPS dual frequency receiver. As an illustration of our capabilities in this area, Figure 24 and Figure 25 show contour plots of predicted foF2 over Africa for years of solar maximum and solar minimum respectively, at 4 times during the year (summer and winter solstices, autumn and spring equinoxes).

Hermanus Magnetic

Observatory

Page 38



y = 0.0848x + 7.487

4

6

8

10

12

14

16

0 10 20 30 40 50 60 70 80

TEC, [TECU]

foF2

, [M

Hz]

Louisvale -- Year = 2001, Hr = 10h00

y = 0.1024x + 6.4905

4

6

8

10

12

14

16

0 10 20 30 40 50 60 70 80

TEC, [TECU]

foF2

, [M

Hz]

Grahamstown -- Year = 2001, Hr =

y = 0.1427x + 5.2645

4

5

6

7

8

9

10

11

12

0 5 10 15 20 25 30 35 40

TEC, [TECU]

foF2

, [M

Hz]

Louisvale -- Year = 2004, Hr = 10h00

y = 0.1622x + 4.7286

4

5

6

7

8

9

10

11

12

0 5 10 15 20 25 30 35 40 45

TEC, [TECU]

foF2

, [M

Hz]

Grahamstown -- Year = 2004, Hr =

Figure 22. An illustration of the correlation between foF2 and TEC for the years 2001 (top panel) and 2004 (bottom panel), and for Louisvale (left panel) and Grahamstown (right panel).

Hermanus Magnetic

Observatory

Page 39



3

5

7

9

11

13

15

1973 1978 1983 1988 1993 1998 2003

Year

foF2

, [M

Hz]

measured

predicted

Figure 23. The Grahamstown measured foF2 and LAM model predicted foF2 values for the period 1973 to 2004, illustrating the extent of the database as well as the ionospheric variation.

Hermanus Magnetic

Observatory

Page 40



Figure 24. Contour plots showing the predicted noon local time variation of foF2 over Southern Africa for a solar maximum year 1980 at 4 times during the year. The foF2 contour values in MHz.

10.511

11

11.5

11.5

11.5

12

12

12

12

12

12

12.5

12.5

12.5

12.5

12.5

12.512.5 12.5

12.513

13 13 13

1313

13

13

13 13

13.5 13.5 13.5

13.513.5 13.5

13.5

13.5

13.5

13.5

14

14

14

1414

14

1414

14.5

14.5

foF2 YYYY=1980 DDD=267 HH:MM=12:00

15° W 0° 15° E 30° E 45° E 60° E

60° S

45° S

30° S

15° S

0°

89

9

10

10

10

11

11

11

12

12

1212

1212

12 12

13

1313

13 13


15° W 0° 15° E 30° E 45° E 60° E

60° S

45° S

30° S

15° S

0°

11

11

11.5 11.511.5

11.5

11.5

1212

12 12

12

12

12

12

12.5 12.5 12.5

12.5 12.5

12.5

12.5

12.5

13

13 13 13

13

13 13

13

1313

13.513.5

13.5

13.5 13.5

13.513.5

1414

14

14 14

14.514.5

14.5 14.5


15° W 0° 15° E 30° E 45° E 60° E

60° S

45° S

30° S

15° S

0°

Autumn Equinox

88.5

9

9

9

9.5

9.5

9.5

1010

10

1010

10.5 10.5

10.5

10.5

10.5

11

11

11

11

1111

11.511.5

11.5

11.511.5

11.5

11.5

11.5

1212 12

12

12

12 12

12.5 12.512.5

12.512.5 12.5

13 1313

1313 13

13.5 13.513.5

13.513.5 13.5


15° W 0° 15° E 30° E 45° E 60° E

60° S

45° S

30° S

15° S

0°

Winter Solstice

Spring Equinox Summer Solstice

Hermanus Magnetic

Observatory

Page 41



Figure 25. Contour plots showing the predicted noon local time variation of foF2 over Southern Africa for a solar minimum year 1976 at 4 times during the year. The foF2 contour values in MHz.

4.1.7 Discussion

The variation of calculated TEC values for each of the proposed antenna locations follow expected trends: Decreasing with solar cycle from solar maximum (2000-2001) towards solar minimum (2004-2005); diurnal variation peaking at approximately 14:00 Local Time and seasonal variation with maximum values around equinoxes and minimum values during winter months. In addition, TEC increases from mid-latitudes towards the equator with maximum values observed by near equatorial receivers like Kenya and Ghana (Figure 26). From the TEC-foF2 correlation equations (Figure 22) it is expected that foF2 values for maximum equatorial TEC values of 100 TECU should not exceed 20 MHz, well below the minimum operational SKA frequency of 70MHz.

Autumn Equinox Winter Solstice


66.56.5

6.57

7

7

7.5

7.5

7.5

8 8

88

8

8.58.58.5

8.5

8.5

8.5

99 9

9

9

9

9.5

9.59.5

9.59.5

9.510

10

10

1010

10.5

10.5

10.5


15° W 0° 15° E 30° E 45° E 60° E

60° S

45° S

30° S

15° S

0°

3.544.5 4.5

4.55 5

5

5

5.5 5.5

5.5

5.5

6

6

6

6.5

6.56.5

7

7

77

7.57.5

7.5

7.5 7.5

8 88

8

8 88.5


15° W 0° 15° E 30° E 45° E 60° E

60° S

45° S

30° S

15° S

0°

5.56

66.5 6.5

6.57 7

7

7

7.5

7.5

7.5

8

8

8

8.5 8.5

8.5

8.5

8.58.5

99

9

9

99

9.59.5 9.5

9.5

9.5

9.510

10

10


15° W 0° 15° E 30° E 45° E 60° E

60° S

45° S

30° S

15° S

0°

5.5

66

6

6.5

6.5

6.5

7

7

7

7.5

7.5

7.5

7.57.5

7.58

8

8 8 8

8

8

8

8.5

8.58.5

8.5

8.58.5

99

9

99

9.59.5

9.59.5

10 10

10


15° W 0° 15° E 30° E 45° E 60° E

60° S

45° S

30° S

15° S

0°

Hermanus Magnetic

Observatory

Page 42



0 5 10 15 20 250

10

20

30

40

50

60

70

80

90

100

Time [UT]

TEC

[TE

CU

]

TEC variation at Equinox (September 2002)

SKA Hub (21.4E, 30.7S)Zambia (28.3E, 15.3S)Kenya (40.2, 3.0S)

Figure 26. Latitudinal variation of diurnal TEC variation for Spring Equinox, September 2002. Highest TEC values are observed at latitudes nearest to the equatorial anomaly (Kenya), while the peak TEC value at Zambia is almost the same as at the SKA hub.

4.2 Scintillation

4.2.1 Scope

This section presents an analysis of ionospheric scintillation over the proposed SKA sites by means of plots and cumulative distributions of parameters related to ionospheric scintillation. Scintillation data is presented for various times of the year, namely at the equinoxes and solstices, at various frequencies (150 MHz and 1.5 GHz) and for various times during the solar cycle (2001 and 2004).

4.2.2 Introduction

4.2.2.1 Origins of Ionospheric scintillation

Ionospheric scintillation is a rapid random change in the amplitude and phase of radio signals as they pass through the ionosphere. Ionospheric scintillation is caused by random small-scale plasma density irregularities in the ionosphere which result in diffraction and scattering of trans-ionospheric radio signals. Ionospheric scintillation is the major restriction for low frequency (< 100 MHz) imaging of astronomical radio sources since the decorrelation it introduces limits the

Hermanus Magnetic

Observatory

Page 43



improvement in angular resolution that can be obtained by ionospheric phase correction using self-calibration and long integration times in large aperture i.e. long-baseline (>5 km) interferometers. At such low frequencies ionospheric phase and amplitude fluctuations become so large that even the strongest calibration sources become decorrelated and self-calibration techniques fail (Kassim et al, [1993]). Scintillation is most common in the E and F regions of the ionosphere at equatorial latitudes of about 10º north and south of the magnetic equator, but occurs to a limited extent in mid-latitude (20º-50º latitude) regions. Ionospheric scintillation primarily occurs during night time between dusk and midnight and typically lasts for fractions of a second to several seconds, depending on the size and velocity of the ionospheric irregularities. The size of the irregularities can range from hundreds of metres to hundreds of kilometres. In the low latitudes the ionospheric irregularities are associated with decreased ionospheric conductivities associated with plasma bubbles and plumes, plasma drifts and the pre-reversal enhancement of the equatorial eastward electric field. In the high latitudes the localized irregularities are associated with auroral particle precipitation. Mid latitude scintillations do occur but with much lower incidence that at low and high latitudes. Ionospheric scintillation in the mid latitudes is enhanced during adverse geomagnetic storm conditions. There is often a correlation between intensity and phase fluctuations associated with scintillations (Fremouw et al [1980]).

4.2.2.2 Detection and quantification of Ionospheric scintillation

Ionospheric scintillation can be detected through the occurrence of spread-F in ionograms, decreases in the signal to noise ratio in GPS receivers as well as by means of dedicated Ionospheric Scintillation Monitors (ISM) which continuously sample the phase and power of trans-ionospheric radio signal from satellites at rates of up to 50 samples per second. The incidence of spread-F can thus be used as evidence for the likely occurrence of ionospheric scintillations. In our search for likely times to find ionospheric scintillations in GPS data, we were guided by incidents of reported spread F in the ionograms from the Grahamstown ionosonde. Ionospheric scintillation is quantified by means of various parameters including the amplitude scintillation index S4, the standard deviation of the phase variation σΦ., and the standard deviation of the rate of change of TEC index (ROTI).

The S4 index is the ratio of the RMS intensity in a data sample normalized by the average intensity in the same sample. In weak-scatter theory, this number ranges from 0.0 (no scintillation) to 1.0 (full scintillation; all power is scattered). In practice, S4 ranges from 0.0 (assuming no system noise) to 1.7 for extreme equatorial scintillation.

The S4 index, corrected for the effects due to ambient noise, is defined in terms of the signal power P in a sample and the mean signal to noise ratio S/No:

Hermanus Magnetic

Observatory

Page 44



21

2

22

4 /195001

/100

+−

−=

oo NSNSP

PPS

where and oNS / represent 60 second mean values. The direct measurement of the S4 index requires the use of dedicated Ionospheric Scintillation Monitors which observe the amplitude and phase of the L1 frequency (1.57542 GHz) GPS signals at a rate of 50 Hz under computer control and records several scintillation related parameters. The L1 frequency is close to the hydrogen line frequency of 1.420 MHz (21 cm band) commonly used in radio astronomy. The rate of TEC index (ROTI) is based on the standard deviation of the rate of change of TEC (ROT) and is defined by

1222ROTI ROT ROT

= −

The ROTI has been demonstrated to be roughly proportional to S4 (Pi et al, [1997]). The relationship between ROTI and S4 varies between 2 and 5 depending on the ionospheric projection of the satellite velocity and the drift velocity of the ionospheric irregularity. The ROTI is measured in TECU/min (1 TECU = 1016 electrons/m2) over a time interval of typically 1 to 5 minutes to give a time resolution comparable to radio astronomy integration times. The TEC and ROTI can be calculated from the difference of the GPS observables L1 and L2. Statistics about measured L2 loss-of-lock is available as part of the RINEX data recorded by some GPS receivers and is often used as an estimate of phase scintillation. For an undisturbed ionosphere, dual frequency GPS receiver observations can show a flat total electron content (TEC) distribution over several minutes. When spread F occurs on an ionogram, an irregularity appears in the ionosphere, and the calculated TEC will fluctuate in addition to the quick variations of the received carrier-to-noise ratio (Du et al [2000]).

4.2.3 Data

For this study GPS-derived TEC derived from L1-L2 was employed for investigating rapid, non-regular variations of the ionosphere. To characterise the scintillations over the hub of the proposed SKA, GPS data from Sutherland was used. Sutherland has the nearest IGS GPS station to the proposed SKA hub. The GPS station at Sutherland (SUTH) records GPS observables L1, L2, P1 and C1 continuously over 24 hours at 30 s intervals. Data from the GPS receiver at Malindi in Kenya, near the magnetic equator, was used to characterise the scintillations near the equator. These GPS data was downloaded from the IGS ftp mirror site at HartRAO.

Hermanus Magnetic

Observatory

Page 45



4.2.4 Approach

No ISM scintillation receivers have been deployed in the region of interest. We had been advised that the SKA budget will not provide for a scintillation receiver to be purchased and deployed together with the RFI measurement equipment at selected sites. To characterise the ionospheric scintillations we made use of scintillation models for radio transmission at 150 MHz and GPS data at the L-band frequency of 1.5 GHz. The GPS-based scintillation index (ROTI) is derived from short term TEC (L1-L2) variation in GPS observables. The GPS parameter we will use is the standard deviation of the rate of change of TEC (ROTI) expressed in TECU/min averaged over 5 minutes. The ROTI is derived from Sutherland GPS TEC sampled at intervals of 30 seconds. No raw data of ISM recordings of index S4 were available by means of which we could determine the scale factor between S4 and the values of ROTI that we calculated, hence TEC variations are represented in TECU/min for ease of comparison with similar data. The ROTI is determined and plotted for typical quiet (low solar activity, 2004) and active periods (high solar activity, 2001).

4.2.5 Results

4.2.5.1 Scintillation model results.

The estimates of ionospheric scintillation shown in the figures below were calculated by means of the WBMOD v15.01 ionospheric scintillation model (Secan [2005]). The WBMOD model is designed to show only variations in scintillation with sunspot number (SSN) and local time in the mid-latitude region. The scintillation climatology model, WBMOD, estimates the worst-case intensity-scintillation levels at a frequency of 150 MHz. The plots present estimates of ionospheric scintillation for three levels of the solar cycle (based on the sunspot number, SSN) at six local times for a location centred on the proposed SKA hub in South Africa. The SSN values used in the model are 0, 75, and 150 (the model caps SSN at 150). The local times used are 00h00, 04h00, 08h00, 12h00, 16h00, and 20h00 (mid-latitude scintillation maximizes at local midnight in the mid-latitudes). Each calculation is for a far-distant "transmitter" at a frequency of 150 MHz and provides the 95th percentile S4 value (the value of S4 that is expected to be exceeded only 5% of the time). Each plot shows the variation of the 95th percentile of the intensity scintillation index, S4, vs. elevation angle from horizon to horizon along the 20º longitude meridian passing through the centre of the proposed SKA hub. The parameter “Extended elevation angle” used for the x-axis in the plots, is an extension of elevation angles from 0º to 90º which refer to ray paths to the north along the 20º meridian line, to elevation angles beyond 90º which refer to elevations to the south, such that 180º corresponds to the southern horizon.

Hermanus Magnetic

Observatory

Page 46



S4 model SSN=000

0.00

0.10

0.20

0.30

0.40

0.50

0 30 60 90 120 150 180

Extended elevation angle [deg]

S4 (9

5th

perc

entil

e)

00h00

04h00

08h00

12h00

16h00

20h00

The WBMOD model is rather rudimentary at mid-latitudes, primarily because scintillation is much less severe at mid-latitudes than elsewhere and less data has been collected there. For example, while there is a very complex seasonal variation in the equatorial region there is none in mid-latitudes. The main driver in the level of mid-latitude scintillation is the amount of solar-produced plasma there is available in the ionosphere. This is strongly controlled by the solar flux (modelled here by the SSN) and the local time. Thus, the plots below cover the two primary variations which determine mid-latitude scintillation (Secan [2005]).

Figure 27. Scintillation index S4 as predicted by WBMOD ionospheric scintillation model along the 20ºE meridian over the SKA hub for SSN=0.

Hermanus Magnetic

Observatory

Page 47



S4 model SSN=075

0.00

0.10

0.20

0.30

0.40

0.50

0 30 60 90 120 150 180


S4 (9

5th

perc

entil

e)00h00

04h00

08h00

12h00

16h00

20h00


S4 model SSN=150

0.00

0.10

0.20

0.30

0.40

0.50

0 20 40 60 80 100 120 140 160 180


S4 (9

5th

perc

entil

e)

00h00

04h00

08h00

12h00

16h00

20h00


Hermanus Magnetic

Observatory

Page 48



4.2.5.2 GPS derived scintillation indexes during periods of spread F occurrence.

The occurrence of spread-F in the Grahamstown ionograms near significant dates during the year, both near solar maximum (2001) and near solar minimum (2004), were taken as indicative of a possibility of detecting ionospheric scintillations in GPS data. Table 2 shows the dates and times considered for GPS scintillation analysis, together with the geomagnetic storm index Kp for the particular periods.

Table 2. Periods of identified spread-F in Grahamstown ionosonde data nearsolstices and equinoxes of the years 2001 and 2004.

Year Solar

cycle DoY Date Significance GPS week Time of Spread F

Kp-index

083 Mar 24 Autumn equinox 1106 03:30-04:30 3.5 180 Jun 29 Winter solstice 1120 04:30-05:00 0 275 Oct 3 Spring equinox 1134 00:00-04:00 5

2001 Max

357 Dec 23 Summer solstice 1146 02:00-03:00 1 100 Apr 9 Autumn equinox 1265 01:00-01:30 3 176 Jun 24 Winter solstice 1276 00:30-03:00 1 266 Sep 22 Spring equinox 1289 19:30-04:00 5

2004 Min

356 Dec 21 Summer solstice 1302 03:00-05:00 2 Kp-indexes were derived from the Space Physics Interactive Data Resource (SPIDR) site http://spidr.ru.ac.za/spidr/ The figures below present the ROTI for GPS signals for the dates indicated in Table 2. The particular IGS station used for these figures is Sutherland (20.81ºE, 32.38ºS). The coverage of Sutherland GPS observations used for these scintillation measurements spans a circular area with a diameter of about 1300 km corresponding to ray paths above an elevation angle of 15º. The time scale is universal time (UT). The spread-F occurrence typically occurs near midnight and during the early morning hours (UT). The standard deviation used in ROTI is calculated over a period of 5 minutes for the data sampled at 30 s intervals with the condition that no cycle slip happens during the interval. During the solar maximum scintillations with ROTI exceeding 2 TECU/min over a 5 minute period occur less than 0.1% of the time. On 22 September 2004 an extended duration of spread F occurred in the Grahamstown ionosonde data for a period of 8.5 hours from 19:30 to 04:00. A mild geomagnetic storm (Kp-index =5) was in effect. The scintillation index ROTI at Sutherland as shown in Figure 32 never exceeded ROTI>0.1 TECU/min for the duration of the spread F occurrence. During the same time the scintillation index ROTI at the equatorial station

Hermanus Magnetic

Observatory

Page 49



Malindi, Kenya, as shown in Figure 30, exceeded ROTI >0.25 TECU/min, considered weak scintillation, only once during the period of spread F. The cumulative frequency of occurrence of ROTI >0.25 TECU/min was less than 0.02% over the 24-hour period.

Figure 30. Rate of TEC scintillation index and Cumulative Frequency of Occurrence for all GPS satellites observed from the equatorial GPS station at Malindi, Kenya (40.19ºE, 3.00ºS) on 22 September 2004.

0 0.25 0.5 0.75 110-3

10-2

10-1

100ROTI for Mali on 2004-09-23

ROTI [TECU/min]

Cum

ulat

ive

Freq

. of O

ccur

ence

0 3 6 9 12 15 18 21 240

50

100Elevations of satellites seen from Mali on 2004-09-23

Ele

vatio

n [d

eg]

0 2 4 6 8 10 12 14 16 18 20 22 240

0.5

1ROTI for Mali on 2004-09-23

UT [hours]

RO

TI [T

EC

U/m

in]

Hermanus Magnetic

Observatory

Page 50



4.2.5.2.1 Rate of TEC index (ROTI)

4.2.5.2.1.1 Year 2001 (Near solar maximum)

Figure 31. Elevation and ROTI plots for all satellites observed during a day for a solar maximum year (2001) at 4 times during the year. Each one-day plot shows the occurrence of scintillation over Southern Africa as measured from the Sutherland GPS site.



0 3 6 9 12 15 18 21 240

50

100Elevations of satellites seen from suth on 2001-12-24

Ele

vatio

n [d

eg]

0 2 4 6 8 10 12 14 16 18 20 22 240

0.5

1

1.5

2ROTI for suth on 2001-12-24

UT [hours]

RO

TI [T

EC

U/m

in]

0 3 6 9 12 15 18 21 240

50


Ele

vatio

n [d

eg]

0 2 4 6 8 10 12 14 16 18 20 22 240

0.5

1

1.5


UT [hours]

RO

TI [T

EC

U/m

in]

0 3 6 9 12 15 18 21 240

50


Ele

vatio

n [d

eg]

0 2 4 6 8 10 12 14 16 18 20 22 240

0.5

1

1.5


UT [hours]

RO

TI [T

EC

U/m

in]

0 3 6 9 12 15 18 21 240

50


Ele

vatio

n [d

eg]

0 2 4 6 8 10 12 14 16 18 20 22 240

0.5

1

1.5


UT [hours]

RO

TI [T

EC

U/m

in]

Hermanus Magnetic

Observatory

Page 51



4.2.5.2.1.2 Year 2004 (Near solar minimum)

Figure 32. Elevation and ROTI plots for all satellites observed during a day for a solar minimum year (2004) at 4 times during the year. Each one-day plot shows the occurrence of scintillation over Southern Africa as measured from the Sutherland GPS site. Note that the scale on the ROTI axis of the 2004 plots has a maximum at 0.4 TECU/min while the 2001 plots have their maximum at 2 TECU/min. The cause of the gap in the autumn equinox data is not known.



0 3 6 9 12 15 18 21 240

50


Ele

vatio

n [d

eg]

0 2 4 6 8 10 12 14 16 18 20 22 240

0.1

0.2

0.3

0.4ROTI for suth on 2004-06-25

UT [hours]

RO

TI [T

EC

U/m

in]

0 3 6 9 12 15 18 21 240

50


Ele

vatio

n [d

eg]

0 2 4 6 8 10 12 14 16 18 20 22 240

0.1

0.2

0.3


UT [hours]

RO

TI [T

EC

U/m

in]

0 3 6 9 12 15 18 21 240

50


Ele

vatio

n [d

eg]

0 2 4 6 8 10 12 14 16 18 20 22 240

0.1

0.2

0.3


UT [hours]

RO

TI [T

EC

U/m

in]

0 3 6 9 12 15 18 21 240

50


Ele

vatio

n [d

eg]

0 2 4 6 8 10 12 14 16 18 20 22 240

0.1

0.2

0.3


UT [hours]

RO

TI [T

EC

U/m

in]

Hermanus Magnetic

Observatory

Page 52



4.2.5.2.2 Cumulative probability of occurrence of scintillation.

4.2.5.2.2.1 Year 2001 (Near solar maximum)

Figure 33. Cumulative probability of occurrence of 30 sec rate of TEC index (ROTI) the level of scintillation over Southern Africa as measured from the Sutherland GPS site for a solar minimum year (2004) at 4 times during the year.



0 0.25 0.5 0.75 110-3

10-2

10-1


ROTI [TECU/min]

Cum

ulat

ive

Freq

. of O

ccur

ence

0 0.25 0.5 0.75 110-3

10-2

10-1


ROTI [TECU/min]

Cum

ulat

ive

Freq

. of O

ccur

ence

0 0.25 0.5 0.75 110-3

10-2

10-1


ROTI [TECU/min]

Cum

ulat

ive

Freq

. of O

ccur

ence

0 0.25 0.5 0.75 110-3

10-2

10-1


ROTI [TECU/min]

Cum

ulat

ive

Freq

. of O

ccur

ence

Hermanus Magnetic

Observatory

Page 53



4.2.5.2.2.2 Year 2004 (Near solar minimum)

Figure 34. Cumulative probability of occurrence of 30 sec rate of TEC index (ROTI) over Southern Africa as measured from the Sutherland GPS site for a solar minimum year (2004) at 4 times during the year. Note that the scale on the ROTI axis of the 2004 plots has a maximum at 0.25 TECU/min while the 2001 plots have their maximum at 1 TECU/min.



0 0.2510-3

10-2

10-1


ROTI [TECU/min]

Cum

ulat

ive

Freq

. of O

ccur

ence

0 0.2510-3

10-2

10-1


ROTI [TECU/min]

Cum

ulat

ive

Freq

. of O

ccur

ence

0 0.2510-3

10-2

10-1


ROTI [TECU/min]

Cum

ulat

ive

Freq

. of O

ccur

ence

0 0.2510-3

10-2

10-1


ROTI [TECU/min]

Cum

ulat

ive

Freq

. of O

ccur

ence

Hermanus Magnetic

Observatory

Page 54



4.2.6 Discussion

None of the modelled scintillation data show any significant scintillation (scintillation index above S4=0.25) even during high SSN conditions. None of the GPS data show significant scintillations even during spread-F occurrences during either 2001 or 2004. This confirms that scintillations are rare in the mid-latitudes. Scintillation is expected to be significant only near the equatorial anomaly and the auroral oval. Variability associated with the scintillation occurring at the equatorial anomaly is not relevant for the proposed RSA SKA site because of large distance from equatorial anomaly and from the auroral oval.

4.3 Large scale phenomena

4.3.1 Introduction

Large-scale static and dynamic structures in the ionosphere give rise to horizontal gradients in the TEC which cause irregularities in transionospheric waves. These irregularities appear as spread F (SF) or Travelling Ionospheric Disturbances (TID) on ionograms recorded with bottomside ionosondes.

For the purpose of this study 2 years worth of half hourly ionograms from the ionospheric field stations, Louisvale (28.5°S, 21,2°E) and Grahamstown (33.3°S, 26.5°E), were manually scrutinised for occurances of both the SF and TID phenomena. The years 2001 and 2004 were chosen for this study since the former year was close to a period of maximum solar activity and the data availability for both years was higher. Statistics of the probability of occurrence of SF and TID were generated.

4.3.2 Results

The figures that appear in this section are contour plots illustrating the percentage occurrence of these irregularities at the Grahamstown and Louisvale stations in 2001 and 2004. The percentage measure was determined as a relationship between the number of ionograms in a particular month exhibiting the chosen irregularity against the total number of ionograms recorded in that month at each hour. The time resolution for the ionograms was half hourly and, although there were some data gaps present, the data availability for the chosen period of time was good. In this study a total of 28811 and 33646 ionograms for Louisvale and Grahamstown respectively were examined. For the Louisvale station it was found that only 5.54% and 2.12% of all examined ionograms exhibited spread F and TID irregularities respectively. At Grahamstown the percentages were 9.26% and 2.43% for spread F and TID respectively.

Hermanus Magnetic

Observatory

Page 55



From Figure 35 and Figure 36 it can be seen that when spread F occurs it is more likely to occur in the early morning hours, with the highest probability occurring at sunrise. Incidences of TID are more likely to occur around midday in the winter months.

Figure 35. The percentage occurrence of Spread F at the Grahamstown (33.3°S, 26.5°E) station for the years 2001 (top panel) and 2004 (bottom panel).

Hermanus Magnetic

Observatory

Page 56



Figure 36. The percentage occurrence of Spread F at the Louisvale (28.5°S, 21.2°E) station for the years 2001 (top panel) and 2004 (bottom panel).

Hermanus Magnetic

Observatory

Page 57



Figure 37. The percentage occurrence of TIDs at the Grahamstown (33.3°S, 26.5°E) station for the years 2001 (top panel) and 2004 (bottom panel).

Hermanus Magnetic

Observatory

Page 58



Figure 38. The percentage occurrence of TIDs at the Louisvale (28.5°S, 21.2°E) station for the years 2001 (top panel) and 2004 (bottom panel).

Hermanus Magnetic

Observatory

Page 59



4.3.3 Conclusion

The above study has shown that the probability of occurrence of the ionospheric irregularities giving rise to SF and TID is relatively small and confined to certain times of the day and months. We therefore conclude that the ionosphere is stable over the proposed South African SKA sites. In the event that South Africa hosts the SKA, a more extensive study could be conducted to determine the predictability of these occurrences.

4.4 Energetic particle precipitation rate and proximity to the South Atlantic Anomaly

4.4.1 Introduction

The South Atlantic Anomaly (SAA) is the region where the Earth’s inner van Allen radiation belt makes its closest approach to the planet's surface. For a given altitude, the radiation intensity is higher over this region than elsewhere. It is produced by a "dip" in the Earth's magnetic field. The South Atlantic Anomaly gives rise to a higher precipitation rate of energetic particles in this region than at equal latitudes elsewhere. This in turn affects the likelihood and frequency of scintillation.

4.4.2 Results

Figure 39 illustrates the proton and electron flux globally and illustrates the higher concentration of particle precipitation over the South Atlantic Anomaly region. From this it is believed that the potential effect of the South Atlantic Anomaly (SAA) at the proposed latitudes of the SKA will be small and in most cases negligable. However, in anticipation of the SKA being hosted by South Africa, the HMO has a masters student who has just started a project investigating the particle precipitation rate over South Africa.

Hermanus Magnetic

Observatory

Page 60



(a)

(b)

Figure 39. Plots of energetic particle precipitation over the world, showing (a) proton and (b) electron flux predictions [SPENVIS]

Hermanus Magnetic

Observatory

Page 61



4.5 Equatorial Electrojet

4.5.1 Introduction

The Equatorial Electrojet (EEJ) was first discovered in 1922 when measurements of the horizontal north-south component of the geomagnetic field, H, showed a diurnal range more than double the value expected. It was found that this enhancement was caused by a varying electric current of about 300 km in width flowing in a very narrow zone near the magnetic dip equator. Within a narrow belt of about 3° geomagnetic latitude (Ø) the Earth’s magnetic field lines are almost perfectly horizontal. Any accumulated charge is unable to leak away to other latitudes and strong polarization electric fields build up, notably between altitudes of 70 km and 140 km, where the electrons are magnetized, but the ions are still collisional. This serves to enhance the Cowling conductivity (Cowling [1932]) in the east-west direction and the result is the intense band of electric current flowing at dynamo altitudes. This electrojet current reveals itself as a sharp enhancement in H in a narrow latitude range of ±5°. This electric field is transferred almost undiminished to higher altitudes, to affect profoundly the entire region in the latitude range ±25° (Campbell and Matsushita [1967], Rajaram [1977]). Figure 40, taken from Doumouya et. al., [2003], illustrates the EEJ effects in terms of the H and Z components at a fixed local time of 12h00 and an altitude of 450km.

The aim of this section of the document is to describe the EEJ in terms of its potential influence on the ionosphere over the proposed Square Kilometer Array (SKA) sites.

4.5.2 The Equatorial F-Region

The electric field E, which is transferred to the higher altitudes, interacts with the horizontal geomagnetic field to set up an upward directed E x B drift plasma, notably in the F-region. The consequences are the following:

• Enhanced F-region altitudes over the Equator compared to other latitudes.

• Decreased F-region ionization between 09h00 and 15h00 at the Equator and the low latitudes, known as the ‘noon bite-out’.

• Two peaks in F-region ionization at subtropical latitudes around ±20°, which appear to have their source in ionization that is uplifted at the equator. This is known as the F-region equatorial anomaly. (Rajaram [1977]):

The strength of the EEJ can be described by the parameter ∆H, which can be obtained from the magnetic data recorded during the International Equatorial Electrojet Year (IEEY) in West Africa. The electric field (E), which drives the EEJ, plays a major role in the variation of the thickness and the height of the F2 layer. However, the variation in the shape of the bottomside F2 layer is not sensitive to the electric field.

Hermanus Magnetic

Observatory

Page 62



Figure 40. Sun-synchronous view of the H and Z components of the EEJ magnetic effects at a fixed local time (12h00 LT) at an altitude of 450 km. ∆H contour interval is 10 nT for a minimum which depends on the longitude, it is about -50 nT in Africa, -70 nT in South America and -30 nT in India. ∆Z contour interval is 10 nT for a minimum of about -30 nT and a maximum of 30 nT in Africa (Doumouya [2003]).

Hermanus Magnetic

Observatory

Page 63



The equatorial anomaly is formed by the action of the force produced by the electric field E and the magnetic field B. This E x B force acts in an upward direction during the daytime under magnetically quiet conditions. Electrons move upwards along the magnetic field lines away from the equator, causing a depletion of electrons in the equatorial region, and the F2 peak to move to a higher altitude. This movement also affects the thickness of the F2 layer. It is also known that the strength of the EEJ, ∆H, is proportional to E, which implies that the peak density height, hmF2, and the thickness of the F2 layer are greatly influenced by the electric field whose strength is indicated by ∆H. In addition, it has been shown that when the upward ionization drift velocity increases the thickness of the F2 layer should also increase (Obrou et. al. [2003]).

4.5.3 Electron Density Variations

At equatorial latitudes the F region vertical drifts (i.e. the east west electric fields) control the F region ionization over the equator. These drifts are driven by the F region vertical electric field, which is coupled along the magnetic field lines to the E region at slightly higher latitudes. During the daytime hours the eastward electric field in the F region, induced by the E region dynamo, results in an upward movement of ionization due to the E x B drift, and then produces the well-known F region ionization anomaly. The eastward electric field reverses during the nighttime causing the reversal of the ionization drifts to a downward direction. Before the reversal, the vertical drifts at times show a post sunset enhancement believed to be caused by the F region dynamo. Post sunset enhancements are routinely observed in F2 peak density and total ionospheric electron content at low latitudes. These enhancements also exhibit seasonal and solar cycle variations (Bhuyan et. al. [2003]).

Electron density measured at an altitude of 500 km during a low solar activity period showed an asymmetric equatorial ionization anomaly (EIA). At the equinoxes, the electron density maximizes at about 10°N and 5°S, and is higher in the Northern Hemisphere compared to that in the Southern Hemisphere. The crest to trough ratio in the Northern and Southern Hemispheres are 1.5 and 1.25, respectively. During the December solstice, the ionization peak in the Southern Hemisphere is higher than that in the Northern Hemisphere. At the time of the June solstice, the EIA is not well developed. The northern anomaly crest is at 5°N while the location of the Southern crest is beyond 10°S. The position of the northern anomaly crest shifts towards the equator from about 10°N during the equinoxes to 5°N at the December solstice (Bhuyan et. al. [2003]).

It has been reported by Su et. al. [1995] that the electron density at 600 km altitude has its highest value within a broad range of latitudes around the geomagnetic equator. At higher latitudes, the electron density was found to decrease with increasing latitude in both hemispheres and the density was reported to be higher in the summer hemisphere. It was observed that the highest daytime values of electron density occur earlier at high latitudes (Bhuyan et. al. [2003]). Measurements show that at the altitude of 500 km, the asymmetric ionisation anomaly is prevalent during the equinox and December Solstice while the EIA is not so well developed at the June Solstice. From models of the electron density and temperature at 600 km, it has been found that the asymmetry in electron density is higher in the summer hemisphere, and lower, in the winter hemisphere, and is

Hermanus Magnetic

Observatory

Page 64



caused principally by a summer to winter difference in the meridional wind. The meridional wind carries the F region in the summer hemisphere to altitudes where the chemical loss rate is low, and lowers the F region to altitudes of higher chemical loss rate in the winter hemisphere (Su et. al. [1995]). It was further seen that the time of diurnal maximum of electron density varies with season. The diurnal peak in density occurs at about 12h00 LT in summer and at about 14h00 LT in winter (Bhuyan et. al. [2003]).

The electron densities at all locations exhibit large day-to-day variability at any given local time. Since ionization in the F region is directly proportional to solar ionizing radiation, changes in the incoming solar flux at extreme ultraviolet wavelengths might be one of the factors responsible for the variability. Another factor that directly affects the F region ionization at low latitudes is the strength of the EEJ. Electron density in the equatorial F region is affected by the variations in electrojet strength through the ‘fountain effect’. Changes in the electric field result in changes in the E ×B drift and subsequent changes in the amount of plasma lifted up that diffuses downward to low latitudes along the magnetic field lines. It has been seen that the electron density is well correlated with the EEJ, except for winter nighttime, since the EIA is not well developed during winter (Bhuyan et. al. [2003]).

4.5.4 TEC and IEC

It has been noted that while a noon-time bite-out in ionization exists at equatorial stations from altitudes of 180 km to well above hmF2, no such ‘bite-out’ is seen in the equatorial total electron content (TEC). This is surprising in view of the fact that the major contribution to TEC comes from altitudes in the vicinity of hmF2, with less contribution expected from plasmaspheric altitudes upwards. The explanation for this lies in the plasma flow patterns due to the EEJ and these effects can be summarized as follows (Rajaram [1977]):

Equatorial regions loose ionization during the daytime from the region below 1200 km but gain ionization at higher altitudes. Therefore the noontime bite-out, while being seen at levels around NmF2, does not manifest itself in TEC.

Sub-tropical regions in the long run gain ionization at all altitudes from (a) the equatorial flow (b) ionosphere-protonosphere flow. Consequently a strong latitudinal geomagnetic anomaly is seen in TEC during the day time.

The variability in ionospheric electron content (IEC) is best described by the ratio of the standard deviation of daily IEC values from the monthly mean values. For middle and high latitude stations, the day-to-day variability of IEC during the day is < 25% irrespective of location, season and solar activity while at night it is significantly higher, especially in the equinoctial season. For equatorial and low latitudes it was found that the EEJ has a pronounced influence on IEC at stations within and near the crest of the equatorial anomaly. The larger the E field, the larger the EEJ current and the higher the vertical plasma drift within the equatorial F-region. This will enhance the fountain effect mechanism of lifting and diffusing ionization to low latitudes. Therefore, an observed high variability of IEC at anomaly crest latitude and its better correlation with electrojet strength variability is a manifestation of the electrodynamic coupling between equatorial and low latitude ionosphere. In fact, a positive correlation between electrojet

Hermanus Magnetic

Observatory

Page 65



strength and IEC at anomaly crest latitude, negative correlation between the former and IEC at equator and no correlation at intermediate locations were found (Aravindan et. al. [1990]).

4.5.5 Conclusion

From Figure 40 it can be seen that the EEJ effects will be limited to stations within the range 10°S to 30°N in Africa. The proposed SKA sites that fall into this range are Ghana (6.69°E, 0.6°S) and Kenya (36.82°E, 1.29°S). From the literature it is possible to assess the influence that these effects would have on radio waves passing through the ionosphere at these latitudes, and apply corrections accordingly. As explained above the TEC within the ionosphere above the 2 proposed equatorial SKA locations would not be affected since by definition the TEC is the electron content along a ray path, and the areas that are enhanced will be offset against those that are depleted. The major effects would be seen in the bottomside F region where the electron density reaches a maximum. Should South Africa win the bid to host the SKA considerable effort will be put into the understanding of the EEJ and the development of corrective procedures to allow for the influence of the EEJ.

5 References

5.1.1 Bibliographic References

Aquino M.H.O., Rodrigues F.S., Dodson A., Moore T., and Waugh S. Results of Statistical Analysis of GPS Ionospheric Scintillation Data in Northern Europe. Proceedings: Atmospheric Remote Sensing using Satellite Navigation Systems, URSI Matera, 13-15 October 2003, Matera, Italy.

Aravindan, P. and Iyer, K. N., Day-to-day variability in ionospheric electron content at low latitutes, Planet. Space Sci., 38, No. 6, pp. 743-750, 1990.

Bhuyan, P.K., Chamua, M., Bhuyan, K., Subrahmanyam, P., Garg, S. C., Diurnal, seasonal and latitudinal variation of electron density in the topside F-region of the Indian zone ionosphere at solar minimum and comparison with the IRI, Journal of Atmospheric and Solar-Terrestrial Physics, 65, pp 359–368, 2003.

Campbell, W and Matsushita, S., Physics of geomagnetic phenomena, Academic Press, London, 1967.

Cilliers P.J., Opperman B.D.L., and Mitchell C.N. Electron density profiles determined from tomographic reconstruction of total electron content obtained from GPS dual frequency data: First results from the South African network of dual frequency GPS receiver stations, Advances in Space Research, Vol 34/9, pp. 2049-2055, November 2004.

Cowling, T. G., Monthly Notices, Roy. Astron. Sot., 93, 90, 1932.

Hermanus Magnetic

Observatory

Page 66



Doumouya, V., Cohen, Y., Arora, B. R., Yumoto, K., Local time and longitude dependence of the equatorial electrojet magnetic effects, Journal of Atmospheric and Solar-Terrestrial Physics, 65, pp 1265–1282, 2003.

Du J., Caruana J., Wilkinson P, Thomas R, and Cervera M., Determination of Equatorial Ionospheric Scintillation S4 by dual frequency GPS. Proceedings from Workshop on the Application of Radio Science. La Trobe University, Australia, 27−29 April 2000.

Hofmann-Wellenhof B, Lichtenegger H and Collins J. GPS Theory and Practice, Fifth Edition, Springer, ISBN 3-211-83534-2, 2001

Obrou, O.K, Radicella, S.M, and Adeniyi, J.O., The equatorial electrojet and the profile parameters B0 and B1 around midday, Journal of Atmospheric and Solar-Terrestrial Physics, 65, pp 299–304, 2003. Oyeyemi E.O., and Poole A.W.V., Towards the development of a new global foF2 empirical model using neural networks, Advances in Space Research, 34(9), pp. 1966 – 1972, 2004.

Rajaram, G., Structure of the equatorial P-region, topside and bottomside-a review, Journal of Atmospheric and Terrestrial Physics, 39. pp. 1125 – 1144, Pergamon Press, 1977.

RfP, Requirements for the Ionospheric Condition at the SKA Radio Telescope sites: request for Information and scope of Work., April 2005.

Secan, J. WBMOD v15.01 ionospheric scintillation model, NorthWest Research Associates, Inc., Tucson Arizona, 9 February 2005.

Shaer S., Mapping and Predicting the Earth's Ionosphere Using the Global Positioning System, Geodatisch-geophysikalishe Arbeiten in der Schweiz, Volume 59, 1999.

Su, Y.Z., Oyama, K.I., Bailey, G.J., Takahashi, T., Hirao, K., Comparison of the satellite electron density and temperature measurements with plasmasphere ionosphere model., Journal of Geophysical Research, 100, 14591–14603, 1995.

5.1.2 Internet References

[CDSM]. Chief Directorate Surveys and Mapping, Cape Town, South Africa. www.trignet.co.za [CODE]. Centre for Orbit Determination. http://www.aiub.unibe.ch/download/CODE/ [GIM]. Global Ionospheric Maps. http://www.cx.unibe.ch/aiub/ionosphere.html [IGS]. International GPS Service. http://igscb.jpl.nasa.gov/ [SAIDC]. South African Ionospheric Data Centre. Grahamstown, South Africa. http://ionosond.ru.ac.za/ and http://spidr.ru.ac.za/ [SPIDR]. Space Physics Interactive Data Resource. Boulder Colorado, United States. http://spidr.ngdc.noaa.gov/spidr/ and http://spidr.ru.ac.za/

Hermanus Magnetic

Observatory

Page 67



[SPENVIS] http://www.spenvis.oma.be/spenvis/help/background/traprad/traprad.html

6 Key Project Participants

The following three key participants were responsible for executing the project. They were assisted by one technician and two students.

6.1.1 Mr BDL Opperman. Co-ordinator

Applied Physicist. Hermanus Magnetic Observatory.

PhD candidate in Ionospheric Physics (Rhodes University). Competencies. Computational Physics, Ionospheric Physics, GPS, Orbital mechanics, Digital signal processing.

6.1.2 Dr PJ Cilliers.

Ionospheric Research Physicist. Hermanus Magnetic Observatory : Competencies. Computational Electromagnetics, Ionospheric Physics, GPS, Digital signal processing.

6.1.3 Dr LA McKinnell.

Ionospheric Research Physicist. Hermanus Magnetic Observatory.

Research Associate. Department of Physics and Electronics, Rhodes University. Key competencies. Ionospheric physics, neural networks, Ionosondes, computational physics.

HERMANUS MAGNETIC OBSERVATORY Information and analysis … · times during the year. .....52 Figure 34. Cumulative probability of occurrence of 30 sec rate of TEC index (ROTI) over

Documents