Ionospheric TEC Weather Map Over South America
H. Takahashi
1
, C. M. Wrasse
1
, C. M. Denardini
1
, M. B. Pádua
1
, E. R. de Paula
1
, S. M. A. Costa
2
, Y. Otsuka
3
,
K. Shiokawa
3
, J. F. Galera Monico
4
, A. Ivo
1
, and N. Sant’Anna
1
1
EMBRACE, Instituto Nacional de Pesquisas Espaciais, São José dos Campos, Brazil,
2
GRRP/CGED, Instituto Brasileiro de
Geografia e Estatística, Rio de Janeiro, Brazil,
3
Institute for Space-Earth Environmental Research, Nagoya University, Nagoya,
Japan,
4
Departamento de Cartografia, Universidade Estadual Paulista, Presidente Prudente, Brazil
Abstract Ionospheric weather maps using the total electron content (TEC) monitored by ground-based
Global Navigation Satellite Systems (GNSS) receivers over South American continent, TECMAP, have been
operationally produced by Instituto Nacional de Pesquisas Espaciais’s Space Weather Study and Monitoring
Program (Estudo e Monitoramento Brasileiro de Clima Especial) since 2013. In order to cover the whole
continent, four GNSS receiver networks, (Rede Brasileiro de Monitoramento Contínuo) RBMC/Brazilian Institute
for Geography and Statistics, Low-latitude Ionospheric Sensor Network, International GNSS Service, and Red
Argentina de Monitoreo Satelital Continuo, in total ~140 sites, have been used. TECMAPs with a time resolution
of 10 min are produced in 12 h time delay. Spatial resolution of the map is rather low, varying between 50 and
500 km depending on the density of the observation points. Large day-to-day variabilities of the equatorial
ionization anomaly have been observed. Spatial gradient of TEC from the anomaly trough (total electron
content unit, 1 TECU = 10
16
el m
2
(TECU) <10) to the crest region (TECU > 80) causes a large ionospheric
range delay in the GNSS positioning system. Ionospheric plasma bubbles, their seeding and development,
could be monitored. This plasma density (spatial and temporal) variability causes not only the GNSS-based
positioning error but also radio wave scintillations. Monitoring of these phenomena by TEC mapping becomes
an important issue for space weather concern for high-technology positioning system and telecommunication.
Introduction
Earth’s equatorial ionosphere presents dynamically temporal and spatial variations. During the day time,
formation of the equatorial ionization anomaly (EIA) occurs due to the ionospheric fountain effect over the
geomagnetic equator. After sunset a rapid uplifting of the F layer forms a postsunset equatorial ionization
anomaly (PS-EIA) [Kelley, 2009]. In a certain condition the uplifting produces ionospheric irregularity
(Rayleigh-Taylor instability) generating plasma bubbles along the geomagnetic field lines [Kelley , 2009].
The PS-EIA produces a large plasma density gradient from the trough region (over geomagnetic equator)
to the crest region (10–15° geomagnetic north and south). Looking the gradient in terms of the total electron
content (TEC unit of 1 × 10
16
/m
2
col), it varies from a few total electron content unit, 1 TECU = 10
16
el m
2
(TECU) along the magnetic equator region to 20–50 TECU at the crest region depending on the day and
season [Bagiya et al., 2009]. In the case of plasma bubbles, the TEC gradient is much steep, a difference of
30–50 TECU from the inside to outside of the bubble with a distance of hundreds of kilometers [Takahashi
et al., 2015]. In addition to the spatial gradient of TEC, radio wave propagation inside of the plasma bubbles
is affected by spatial irregularity of plasma density causing radio wave scintillations [Kintner et al., 2007].
The spatial and temporal variations of plasma density (and TEC) cause errors in Global Navigation Satellite
Systems (GNSS)-based positioning services. A difference of 1 TECU, for example, may cause ~0.16 m of
ionospheric range delay as presented in the next section. The demand of information on GNSS positioning
error and radio wave quality has been increased significantly in the last decade, and today it becomes a
crucial matter for operation of many kinds of space-based systems. Off-shore oil plant platform requires
positioning accuracy of less than a few centimeters [International Association of Oil & Gas Producers, 2011].
High-technology agricultures also require a high positioning accuracy [Stafford, 2000]. Ground-based
argumentation system for aircraft landing and takeoff also requires a high positioning accuracy and integrity
of the information [Federal Aviation Administration, 2016]. For such high level of reliability for positioning
systems, irregularity of ionospheric plasma contents is an important issue to monitor and overcome.
Further to these applications, TEC monitoring has become a powerful tool for ionospheric study [Nogueira
et al., 2015] and space weather [Schrijver et al., 2015].
TAKAHASHI ET AL. IONOSPHERIC WEATHER OVER SOUTH AMERICA 937
PUBLICATION
S
Space Weather
FEATURE ARTICLE
10.1002/2016SW001474
Key Points:
• Ionospheric weather monitored by
total electron content (TEC)
• Dynamical development of
ionospheric plasma bubbles (IPB)
• Influence of plasma bubbles on the
GNSS-based positioning system
Correspondence to:
H. Takahashi,
hisao.takahashi@inpe.br
Citation:
Takahashi, H., et al. (2016), Ionospheric
TEC Weather Map Over South America,
Space Weather, 14, 937–949,
doi:10.1002/2016SW001474.
Received 19 JUL 2016
Accepted 7 OCT 2016
Accepted article online 8 OCT 2016
Published online 11 NOV 2016
©2016. American Geophysical Union.
All Rights Reserved.
In order to attend such demands, Space Weather Study and Monitoring Program (Estudo e Monitoramento
Brasileiro de Clima Especial, EMBRACE)/Instituto Nacional de Pesquisas Espaciais has been developing the
ionospheric weather map based on TEC (TECMAP) and GPS signal scintillation map (S4) since 2013.
TECMAPs over the South American continent with a 10 min time resolution are published on the EMBRACE
internet website as a routine base [Estudo e Monitoramento Brasileiro de Clima Especial (EMBRACE) site,
2016]. The purpose of the present work is, first, to present the concept of TEC calculation using GNSS satellite
radio waves. Typical TECMAPs for a normal condition, with plasma bubble activities and during the
geomagnetic storm conditions, are presented. Then the ionospheric error range due to the TEC spatial
gradient will also be discussed.
Observations
How to Calculate TEC
Phase and group delays of radio waves in the ionosphere are dependent on the total electron contents (TECs)
along the ray path and the radio frequency, f (in hertz), with a simple relation [Hoffmann-Wellenhof et al.,
1994]:
d
p
¼ d
g
¼ 40:3 TEC½=f
2
in mðÞ (1)
where d
p
and d
g
represent, respectively, phase delay and group delay and [] indicates a physical content.
One TEC unit measures 1 × 10
16
el/m
2
col, and it corresponds to approximately 0.16 m of the ionospheric
delay.
Using a fact that the delay is a function of the frequency, one can also obtain TEC from a difference of delay
between the two radio waves. GPS satellites, for example, transmit dual frequency radio wave signals from
the orbit at an altitude of ~20,200 km: f
1
: 1575.42 MHz and f
2
:1227.60 MHz.
The signals are composed of binary phase-modulated carrier wave. Observables are the pseudorange (P)
measured by using the code modulation and the accumulated carrier phase (L). Ground-based GPS receivers,
therefore, observe both pseudorange (P
1
and P
2
in meters) and carrier wave phase (L
1
and L
2
in wave number).
The difference of the pseudorange of P
1
and P
2
can provide TEC as [Kantor et al., 2000]:
TEC½
p;i
¼ 9:52 P
2;i
P
1;i
B
i
(2)
where i denotes the satellite PRN (Pseudo Ramdom Noise) number and B
i
is an instrumental bias of
satellite i and a ground receiver. It is a delay originated from the satellite transmitter and ground-based
receiver hardware (response) systems. TEC is an integrated total electron content along the line of sight
between the receiver and a GPS satellite (denoted as Slant TEC). It provides an absolute TEC value, but it
is a coarse data (having a relatively large random error). On the other hand, the difference of the carrier
Table 1. Procedure to Obtain VTEC From RINEX GPS Data
Start
1 GPS data in RINEX format
2 Get: Time series of CODE: P
1
and P
2
(pseudorange in meters)
3 Get: Time series of phase values: L
1
and L
2
(wave numbers)
4 Cycle slip verification and correction
5 Calcul. Difference of pseudorange: D
p
= P
2
P
1
in meters
6 Calcul. Difference of L
1
and L
2
: D
L
=(L
1
/f
1
L
2
/f
2
)*c, where f
1
and f
2
are radio wave frequency
(/s) and c: light velocity in m/s
7 Calcul. Ambiguity: A(t)=D
L
D
p
for each time series
8 Calcul. Averaged A value (A*): using > 1 h of time series of A(t)
9 Calcul. Slant TEC: STEC = 9.52(D
L
A*) B
i
10 Calcul. Instrumental bias (B
i
): for each receiver-satellite couple
11 Calcul. Slant factor: S
i
= τ
1
/τ
0
, from the satellite zenith angle, θ, at the ionospheric height (navigation data)
12 Calcul. Vertical TEC: VTEC = STEC/S
i
End
Space Weather
10.1002/2016SW001474
TAKAHASHI ET AL. IONOSPHERIC WEATHER OVER SOUTH AMERICA 938
phase between L
1
and L
2
is precise and less noisy, but it is a relative value having an ambiguity, not
providing the absolute TEC value.
The difference of the carrier phase
between L
1
and L
2
is also related to TEC as
TEC½
L;i
¼ 9:52 Φ
1; i
Φ
2;i
þ A
i
* (3)
where Φ =(L/f)*c in meters, L is a wave
number in the wave path, f is a frequency,
and c is the velocity of light. A*isaterm
of ambiguity originated from unknown
absolute value of Φ
1
and Φ
2
. It can be esti-
mated by comparing (P
2
-P
1
) and (Φ
1
-Φ
2
)
for each pair of satellite and receiver:
A
i
* ¼< Φ
1;i
Φ
2;i
– P
2;i
P
1;i
>
where <>indicates time average during
the satellite passage (normally 1–4 h). For
the present work, we adopted 1 h aver-
aging to fix the parameter A*. In calcu-
lating A* another ambiguity caused by
carrier wave cycle splitting should be
taken in account. The cycle slip can be
found from the discontinuity of the car-
rier waves. The correction was carried
out by using a same methodology
developed by Belwitt [1990].
The instrumental bias (B
i
), for each recei-
ver and satellite, was obtained by
comparing the uncalibrated TEC values
from all of the satellite and receiver
combinations using the least mean
Figure 1. (a) Ground-based GNSS receiver sites by RBMC, IGS, LISN, and RAMSAC networks over South America. (b) An
example of ionospheric pierce point distribution of the TEC measurement on the night of 19 February 2014.
Figure 2. Spatial resolution map of TEC calculation on the night of 27
September 2014 at 23:00 UT; Blue: 3 × 3 element, light blue: 5 × 5
element, light green: 7 × 7, and brown: 9 × 9 elements. Each element has a
0.5° (~50 km) spatial extension.
Space Weather
10.1002/2016SW001474
TAKAHASHI ET AL. IONOSPHERIC WEATHER OVER SOUTH AMERICA 939
square fitting method. We assumed that
B
i
would be constant during 24 h. The
procedure of calculus was similar to
those presented by Otsuka et al. [2002]
and Camargo [2009].
Once getting the slant TEC (STEC) in
absolute value, we need to obtain a ver-
tical TEC (VTEC) considering geometric
relation between the satellite position,
ionosphere, and ground receiver. The
VTEC is obtained with STEC multiplied
by a slant factor ( S), which is defined as
τ
0
/τ
1
, where τ
1
is the length of the ray
path between altitudes of 250 and
450 km (assumed ionospheric slab) and
τ
0
is the thickness of the ionosphere
(assumed to be 200 km). The VTEC values
are mapped on the ionospheric shell at
a pierce point of the line of sight. The
pierce points are obtained by assuming
the ionospheric peak height to be
300 km altitude [Otsuka et al., 2013].
The error in VTEC increases with increas-
ing the zenith angle of the satellite-
looking direction. This is mainly due to
multipath error in the low elevation
angle and the error in the instrument
bias estimation. In the present method,
satellites with the zenith angle less than
60°, i. e., the elevation angle more than
30° are used in calculation. The process
of data handling and processing algo-
rithm was based on that has been devel-
oped by Otsuka et al. [2002]. Table 1 lists
the sequence of data processing from
the RINEX original data to TEC data.
GNSS Ground-Based Receivers Over
South America
There are four main GNSS receiver net-
works on the South American continent.
Dual frequency GNSS receivers have
been operated by IBGE (Brazilian
Institute for Geography and Statistics).
IBGE releases a package of RINEX format
data (RBMC) from approximately 120
sites over Brazil, with 15 s time intervals
[Brazilian Institute for Geography and
Statistics Website, 2016]. RBMC real-time
data are available through Networked
Transport of RTCM via Internet Protocol.
In addition to the RBMC data, the data
from IGS (International GNSS Service)
[International GNSS Service Website,
Figure 3. Color shaded TEC maps over South America on the evening of
31 August 2014, from (a) 21:00, (b) 22:00, (c) 23:00, and (d) 00:00 UT,
showing development of the postsunset equatorial ionization anomaly.
The black line indicates the geomagnetic equator, and the dashed line
indicates solar terminator at around 300 km altitude.
Space Weather
10.1002/2016SW001474
TAKAHASHI ET AL. IONOSPHERIC WEATHER OVER SOUTH AMERICA 940
2016], LISN (Low-latitude Ionospheric
Sensor Network) [Low-latitude Ionospheric
Sensor Network website, 2016], and
RAMSAC (Red Argentina de Monitoreo
Satelital Continuo) [Red Argentina de
Monitoreo Satelital Continuo Website, 2016]
were also collected under collaboration
scheme and processed at EMBRACE. Some
additional information of these GNSS recei-
ver networks has been reported elsewhere
[Takahashi et al., 2015].
Figure 1a demonstrates geographic loca-
tions of the GNSS receiver sites used in the
present work. All of the area of Brazil, Peru,
Chile, and Argentina are covered in this
way. Only a few observation sites are avail-
able in the central part of South America
(Bolivia and Paraguay). The total number
of GNSS receivers over the entire continent
is around 140. Normally, one receiver looks
up simultaneously four to six GNSS satellites
within the elevation angle larger than 30°.
Therefore, a high density of TEC observation
can be achieved. Figure 1b presents, as an
example, ionospheric pierce points over
South America at 01:20 UT, 19 February
2014. The total number of observation
points is 646 in the present case. It can be
seen that there is a relatively high-density
observation area extending between
20°–30°S and 40°–60°W. Spatial resolution of
50 to 150 km can be achieved in this area.
However, observation points were less in the
low-latitude regions (0 to 15°S). In some areas,
thedistancelargerthan500kmfromone
pointtotheotheroccurred,whichmakesit
difficult to obtain the TEC spatial variations.
Spatial Resolution of TECMAP
TEC data were mapped on the ionospheric
shell at 300 km altitude with a horizontal
cell of 0.5° × 0.5° in latitude and longitude.
In order to optimize the spatial resolution
of TEC and to cover the entire area with
TEC values, we first calculated a running
average of the three cells; this corresponds
to 3 × 3 elements covering an area of
~160 × 160 km
2
. If no data were found in
the area, the running average area expands
to 5 × 5 elements, which corresponds to
approximately 260 × 260 km
2
. In this way,
the running average element expands up
to 21 × 21 elements, which corresponds
to approximately 1000 × 1000 km
2
as an
Figure 4. Same as Figure 3 but for 25–26 September 2014, showing
development of the plasma depletions.
Space Weather
10.1002/2016SW001474
TAKAHASHI ET AL. IONOSPHERIC WEATHER OVER SOUTH AMERICA 941
