Journal of Global Positioning Systems (2004)
Vol. 3, No. 1-2: 251-258
Performance Evaluation of the Wide Area Augmentation System for
Ionospheric Storm Events
S. Skone, R. Yousuf and A. Coster
Department of Geomatics Engineering, University of Calgary, 2500 University Dr. N.W., T2N 1N4 Calgary, Canada
e-mail: [e-mail] Tel: + 01-403-220-7589; Fax: [phone]
Received: 15 Nov 2004 / Accepted: 3 Feb 2005
Abstract. One of the greatest challenges in developing
accurate and reliable satellite-based augmentation
systems (SBAS) is modeling of ionospheric effects. Wide
area GPS networks are generally sparse (station spacings
of 500-1000 km), and ionosphere models can suffer
degraded performance in regions where large spatial
gradients in total electron content (TEC) exist. Of
particular concern for Wide Area Augmentation System
(WAAS) users is the feature called storm enhanced
density, which is associated with large TEC gradients at
mid-latitudes. This effect is a significant source of error
in the WAAS correction models. The Canadian GPS
Network for Ionosphere Monitoring (CANGIM) consists
of three GPS reference stations in western Canada,
augmented by two additional sites in the northern United
States. In addition to measures of ionospheric activity,
WAAS messages are collected continuously at these sites
and decoded (post-mission) at University of Calgary.
Localization schemes have been developed to compute
WAAS ionosphere corrections for any location in North
America. In this paper, performance of the broadcast
WAAS ionosphere model is quantified through
comparison with truth data from over 400 GPS reference
stations in North America. WAAS ionosphere model
accuracies throughout North America are evaluated for
intense storm events, and compared with WAAS Grid
Ionosphere Vertical Error (GIVE) bounds. Limitations in
the WAAS ionosphere model are identified for enhanced
ionospheric activity and, in particular, the storm enhanced
density phenomenon.
Key words: Ionosphere, WADGPS, WAAS, GPS,
positioning, geomagnetic storm
1 Background
1.1 Storm enhanced density
Storm enhanced density (SED) was originally recognized
in the early 1990’s with the Millstone incoherent scatter
(IS) radar (Foster et al., 2002; Foster and Vo, 2002) and
has been studied in detail with satellite data from the
DMSP and IMAGE satellites, and with TEC data
collected from multiple GPS receivers located across the
US and Canada (Coster et al., 2003a; Coster et al.,
2003b). Analysis of the GPS TEC data shows that during
geomagnetic disturbances, ionospheric plasma is
transported from lower latitudes to higher latitudes,
redistributing plasma across latitude and local time.
Fig. 1 An example of storm enhanced density over North America
during a geomagnetic storm event March 31, 2001

252 Journal of Global Positioning Systems
SED can be described as a narrow plume of greatly
enhanced TEC values (~100 TEC units). A typical SED
plume would extend from the U.S. New England coast
across the Great Lakes region and into central Canada.
An example is shown in Figure 1. Over a period of
several hours, this plume of enhanced TEC moves west
across North America. Since the ionospheric TEC is
generally dependent on local time, it is expected that
similar effects would be observed over Europe. This,
however, has not yet been fully verified. This paper
demonstrates the impact of this feature on WAAS
ionosphere model accuracies in North America.
1.2 Wide area ionosphere model
The presence of large ionospheric gradients associated
with SED is a concern for wide area ionosphere
modeling. In such models, slant line-of-sight TEC
observations are derived from a number of permanent
dual-frequency GPS reference stations. Typically six to
nine TEC measurements are available at each station at a
given epoch. These observations are generally mapped to
an ionospheric shell at a height of 350-400 km (where the
majority of electron density is concentrated) in order to
reduce the model to two dimensions (latitude and local
time). Additionally, the slant TEC observations are
normalized to vertical TEC in order to remove the slant
path dependence and ensure consistency between all
observations. The vertical TEC observations are
combined to define a spatial and temporal model of
ionospheric corrections over the given region (e.g. El-
Arini et al., 1995). Satellite and receiver inter-frequency
biases are also estimated in the ionosphere model
solution.
The WAAS ionosphere corrections are derived from a
permanent network of 25 reference stations in the United
States. A grid of ionosphere corrections is defined at
regular spacings in latitude and longitude (Altshuler et
al., 2002). The grid spacing is 5° for geographic latitudes
of 55° or less, 10° for geographic latitudes in the range
65°-75°, and 10° in latitude and 90° in longitude for the
geographic latitude 85° (DoT, 1999). This grid is
transmitted to single frequency users, where ionospheric
corrections are interpolated from the grid values and
applied to local observations.
All WAAS ionospheric corrections are associated with a
grid ionosphere vertical error (GIVE) which is also
transmitted to users. The discrete GIVE values (after
translation to meters) range from 0.30 m to 45.0 m. Any
larger values imply “not monitored”. It is crucial that
GIVE values accurately bound the errors in ionosphere
corrections for safety-critical aviation applications.
Investigations have been conducted to reliably detect and
bound WAAS ionosphere errors during periods of
increased ionospheric activity (Walter et al., 2000).
Limitations have been identified for satellite-based
augmentation systems at low latitudes (Doherty et al.,
2002) and high latitudes (Skone and Cannon, 1999).
1.3 CANGIM
The Canadian GPS Network for Ionosphere Monitoring
(CANGIM) currently consists of three primary stations in
Western Canada: Calgary (51.08°N, 114.13°W),
Athabasca (54.72°N, 113.31°W) and Yellowknife
(62.48°N, 114.48°W). These stations allow latitude
profiling of the mid- to high-latitude ionosphere.
Additional sites have recently been installed at MIT
Haystack Observatory, near Boston Massachusetts, with a
second site in northern Minnesota. Future plans include
the addition of three more reference sites to improve
coverage in the western Canadian sector, with stations in
Whitehorse, Churchill and Inuvik (Figure 2). Further sites
will be installed in eastern Canada for extended coverage
of the North American region.
The CANGIM sites are equipped with NovAtel
Modulated Precision Clock (MPC) receivers (NovAtel,
2002) and NovAtel 600 antennas. These units contain
dual-frequency Euro4 cards with an internal integrated
PC and precise oscillator. A comprehensive user-interface
command structure allows direct access via modem or
internet connection. The CANGIM receivers have
specialized firmware which provides scintillation
parameters extracted from 50 Hz L1 phase observations,
in addition to raw GPS code and phase observations, rates
of change of TEC, absolute TEC values, and WAAS
messages. The firmware version (scintw) was developed
by A.J. Systems (Van Dierendonck et al., 1996).
The CANGIM network does not currently operate in real-
time. Communication infrastructure is in place, however,
to allow real-time streaming of data to a central
processing facility at University of Calgary for future
research and development purposes. Algorithms have
been developed (El-Gizawy et al., 2002) to monitor the
level of ionospheric activity, and associated impact on
GPS users, for near real-time applications. After several
hardware upgrades, the three primary stations have been
operating reliably since May 2003. WAAS messages
from the CANGIM are currently archived at University
of Calgary. In this paper, analyses focus on post-
processing of the WAAS ionosphere data from
CANGIM.

Skone et al.: Performance Evaluation of Wide Area Augmentation System 253
Fig. 2 Site locations of the installed MPC units (triangles) and additional
proposed site locations (red diamonds)
2 Geomagnetic storm events
Extreme geomagnetic storms were observed during a
two-month period in 2003. Activity commenced with one
of the most severe storms of the past 15 years, in late
October 2003. A major solar flare developed at
approximately 11:00 UT on October 28. A severe
geomagnetic storm commenced in the Earth’s
environment at 06:00 UT on October 29. Activity
continued for several days, with further coronal mass
ejections at approximately 21:00 UT October 29 and
16:00 UT October 30. Enhanced solar emissions were
again observed in late November 2003, with an extended
geomagnetic storm event on November 20, 2003.
The level of global ionospheric activity during these
events can be quantified using the conventional space
weather index Kp. This index is based on observations of
magnetic field fluctuations at ground-based
magnetometer stations (periods of enhanced ionospheric
activity being characterized by strong electric currents
which are observed as magnetic field perturbations at the
Earth’s surface). The planetary Kp index is derived from
measurements of magnetic field variations at thirteen
global stations at (approximately) equally spaced
longitudes. This index is derived at three-hourly intervals
and values range from 0 (quiet) to 9 (extreme). Such
indices provide an approximate measure of global
ionospheric activity at higher latitudes.
Figure 3 shows the Kp index for the storm period October
29-31, 2003. Kp values of 9 were observed on October 29
and 30. This indicates severe storm events for extended
periods on both days. Communications were disrupted for
commercial aircraft operating in polar regions, and
satellite instruments were shut down to mitigate the
impact of enhanced radiation in the space environment.
Aurora were observed at mid-latitudes in the United
States. Development of strong SED was also observed in
North America.
Figure 4 shows Kp indices for the period November 19-
21, 2003. Values greater than 7 were observed for an 18-
hour period commencing 09:00 UT November 20, with a
Kp index of 9 during the interval 18:00-21:00 UT. This
indicates high ionospheric activity for an extended period
on November 20, with severe storm effects in the evening
hours (UT). Aurora were again observed at mid-latitudes,
in the United States and Europe, with SED developing in
North America – starting at 18:00 UT and continuing
through 22:00 UT. The impact of these storm effects
(October and November 2003) is evaluated in terms of
WAAS ionosphere model accuracies in Section 4.
Fig. 3 Kp values for October 29-31, 2003 (NOAA SEC)
Fig. 4 Kp values for November 19-21, 2003 (NOAA SEC)
3 GPS TEC maps
SED was observed in maps of the vertical TEC derived
for North America for both the October and November
2003 events. These maps were produced by combining

254 Journal of Global Positioning Systems
dual-frequency data from more than 400 ground-based
GPS receivers around the United States and Canada. The
data used to produce these maps included observations
from both the International GPS Service (IGS)
(http://igscb.jpl.nasa.gov/) and the Continuously
Operating Reference Stations (CORS). The CORS
network is coordinated by the U.S. National Geodetic
Survey (http://www.ngs.noaa.gov/CORS/). The data
were accessed via publicly available data archives on the
World Wide Web ftp://cddisa.gsfc.nasa.go v /gps/ and
http://sopac.ucsd.edu.
The line-of-sight TEC values were converted to vertical
TEC values using a simple mapping function, and
associated to an ionospheric pierce point latitude and
longitude, assuming a peak ionospheric height of 400 km.
GPS satellite and receiver inter-frequency biases were
estimated and removed from the data. To minimize
mapping function errors in modeling the vertical TEC,
the elevation angle was restricted to be greater than 30
degrees. The satellite or SV biases used were determined
by the JPL modeling efforts (Wilson et al., 1999). The
receiver biases were determined by an in-house process.
Maps of TEC were prepared at 5-minute intervals to be
consistent with WAAS ionosphere grid update intervals.
The vertical TEC has been binned in 2° x 2°
latitude/longitude bins. No smoothing is used; the high
level of detail is due primarily to the persistence of a
well-organized connected structure and to the large
quantity of data processed.
Large gradients in the TEC distribution are observed over
the continental United States into Canada during the
October 2003 event. A plume of TEC associated with
SED is apparent at 22:04 UT on October 29 (Figure 5).
The plume of SED has moved north and westward over a
period of several hours, leading to the increased TEC
values and gradients pictured in Figure 5. While
observations are sparse at the high latitudes, it appears
that the SED feature may extend over the polar region
into Europe.
Similar to the October event, the presence of SED is
clearly observed over North America on November 20,
2003. Figure 6 shows the distribution of TEC over North
America at 19:00 UT and 21:00 UT. The plume of SED
has moved westward during the two-hour period, with a
narrow region of enhanced electron content through the
eastern United States into western Canada. The evolution
of this feature in global maps of TEC demonstrates that
the electrons observed at 19:00 UT flow over the pole,
from North America into Europe.
Fig. 5 GPS TEC map derived from 400+ reference stations, 22:04 UT
October 29, 2003
Fig. 6 GPS TEC maps derived from 400+ reference stations, for 19:00 and 21:00 UT, November 20, 2003

Skone et al.: Performance Evaluation of Wide Area Augmentation System 255
4 Results and analysis
Accuracies of the WAAS ionospheric corrections are
analysed in the region of SED during the October and
November 2003 storm events. To evaluate the
performance of WAAS ionosphere models during these
events, maps of the vertical TEC predicted by WAAS are
compared with high-resolution TEC maps produced from
the more than 400 GPS receivers in North America (e.g.
Figures 5 and 6). As stated previously, the WAAS map is
produced using dual-frequency observations from 25
reference stations. It is expected that this sparse network,
as compared with the 400+ stations used to produce
Figures 5 and 6, may not reflect some of the localized
features associated with the SED. The intention is to
evaluate 1) the extent to which the WAAS ionosphere
model captures the SED feature, and 2) the extent to
which GIVE values bound the observed WAAS model
errors.
4.1 October 29, 2003
The WAAS map for 22:00 UT on October 29 is shown in
Figure 7. This figure can be compared directly with
Figure 5. The SED feature in the western United States is
captured, to a large extent, in the WAAS ionosphere map.
The magnitude of TEC values is decreased at latitudes
above 45°, however, as compared with Figure 5. This is
likely due to the spatial distribution of WAAS reference
stations, in which there is no direct coverage of Canadian
latitudes (although there is coverage at higher latitudes on
the west coast of North America – in Alaska). With the
lack of observations in the latitude range 45°-55°, the
SED feature is not fully resolved in the WAAS
ionosphere model. Discrepancies in relative spatial
variations within a given map are attributed to the
different spatial resolution achieved for the two networks
used in the processing.
It is assumed here that the high-resolution map (Figure 5)
represents “truth”. Figure 8 shows the difference between
the high-resolution TEC map and the WAAS TEC map in
Figure 7, converted to L1 range delay. Any bias between
the two TEC maps has been removed in generating
Figure 8 - such that the mean spatial value in Figure 8 is
0 m. Larger differences are observed in the northwest, in
the region of SED. These larger positive errors indicate
the extent to which the WAAS model underestimates the
northward extent of the enhanced plume of TEC. Errors
in ionospheric corrections are as large as 10 m at latitudes
of 45°-48°.
Fig. 7 TEC map produced from WAAS ionosphere grid at 22:00 UT on
October 29, 2003
Fig. 8 Difference between GPS high-resolution TEC and WAAS TEC
for 22:04 UT on October 29, 2003 (converted to L1 range delay)
In order to assess whether the WAAS ionosphere errors
are bounded appropriately, the WAAS GIVE values are
compared with the WAAS model errors plotted in Figure
8. The GIVE values are plotted in Figure 9. These values
are valid for approximately the same time at which
Figures 5, 7 and 8 are valid. GIVE values greater than 45
m are observed throughout most of North America,
indicating that the ionosphere corrections are “not
monitored” and users are denied ionospheric corrections
over a large region. In the northeastern United States,
GIVE values of 10-20 m are observed. These error
bounds exceed the WAAS model errors computed in
Figure 8 for this region – indicating that the GIVE values
appropriately bound the observed ionospheric model
errors during this event.

256 Journal of Global Positioning Systems
Fig. 9 GIVE values for the WAAS ionosphere grid at 22:00 UT on
October 29, 2003
4.2 November 20, 2003
The WAAS maps for 19:00 and 21:00 UT on November
20 are shown in Figure 10. This figure can be compared
directly with Figure 6. The SED feature developing in the
eastern United States at 19:00 UT is captured at lower
latitudes in the WAAS ionosphere map. Similar to the
October storm event, TEC values decrease significantly
at latitudes above 45°, as compared with the high-
resolution “truth” maps (Figure 6). This is again due to
the lack of WAAS reference stations in Canada, where
WAAS observability is degraded at Canadian latitudes.
The SED feature is not fully resolved in the WAAS
ionosphere model, as compared with the fine-scale truth
maps. The larger WAAS grid spacing (5° × 5°) does not
allow for full resolution of the SED gradients.
Figure 11 shows the difference between the high-
resolution truth TEC maps and the WAAS TEC maps in
Figure 10, with the TEC values converted to L1 range
delay. Again, any bias between the two TEC maps has
been removed in generating Figure 11 - such that the
mean spatial value in Figure 11 is 0 m. Larger differences
are observed at the higher latitudes at 19:00 UT, where
the region of SED has extended through Canada and over
the pole. WAAS model errors as large as 10 m are also
observed in the northern United States at 19:00 UT,
where the WAAS model underestimates the SED
enhancement in this region. At 21:00 UT, WAAS model
errors of 5-10 m are observed in a narrow region
extending northwest through the central United States,
consistent with the plume of SED in Figure 6. This again
reflects deficiencies in the WAAS model in resolving
such localized features.
Fig. 10 TEC maps produced from WAAS ionosphere grid at 19:00 and 21:00 UT on November 20, 2003

Skone et al.: Performance Evaluation of Wide Area Augmentation System 257
Fig. 11 Differences between GPS high-resolution TEC and WAAS TEC for 19:00 and 21:00 UT on November 20, 2003 (converted to L1 range
delay)
Fig. 12 GIVE values for the WAAS ionosphere grid at 19:00 and 21:00 UT on November 20, 2003
In order to assess whether the WAAS ionosphere errors
are bounded appropriately, the WAAS GIVE values are
plotted in Figure 12. These values are valid for
approximately the same time at which Figures 6, 10 and
11 are valid. No GIVE values greater than 45 m are
observed within North America during this event, such
that the availability of WAAS ionospheric corrections is
maintained throughout the storm period. The GIVE
values are in the range 10-16 m throughout North
America, and these bounds exceed the errors plotted in
Figure 11 – indicating that the WAAS model errors are
bounded in all regions during this event.
5 Conclusions
Severe geomagnetic storm events took place October 29-
31, 2003 and November 20, 2003. The impact of these
events was observed globally, with aurora in Europe and
North America and the development of SED at mid- to
high-latitudes. In this paper, investigations were
conducted to evaluate performance of the WAAS
ionosphere model during these storm events. High-
resolution “truth” TEC maps were derived using dual-
frequency observations from more than 400 GPS
reference stations in North America.
SED was observed to develop during the late hours UT
on October 29, 2003. Increased TEC values were
observed at approximately 22:00 UT over the west coast
of North America. A high-resolution TEC map was
compared with the WAAS ionosphere values, and the
WAAS model was observed to underestimate TEC at the
higher latitudes extending into Canada. This is attributed
primarily to the lack of WAAS reference stations (and,

258 Journal of Global Positioning Systems
therefore, ionosphere observations) at Canadian latitudes.
WAAS model errors as large as 10 m were observed at
geographic latitudes of 45°-48°. The GIVE values were
flagged as “not monitored” throughout most of North
America during this event, indicating that the ionospheric
storm threat had been detected. While it is important to
bound the ionosphere model errors reliably, the very large
GIVE values significantly over-bound the errors and deny
WAAS service throughout an extended region.
An extreme geomagnetic storm was also observed to
develop during the late hours UT on November 20, 2003.
A plume of enhanced electron content (SED) developed
in the eastern United States at approximately 18:00 UT,
extending northward into Canada, and moving westward
over the central United States by 21:00 UT. High-
resolution truth TEC maps were again compared with the
WAAS ionosphere predictions, and the largest WAAS
model errors were observed near the region of SED. The
WAAS model was deficient in resolving fine-scale
structure and gradients associated with the SED, and
WAAS model errors as large as 10 m were observed in
the United States. All WAAS ionosphere model errors
were, however, bounded by the GIVE values for this
event.
Acknowledgements
The authors wish to acknowledge the International GPS
Service (IGS) and the Continuously Operating Reference
Station (CORS) network, for raw GPS data used in the
processing. Introductory material has been published
previously in Proceedings of the European Navigation
Conference GNSS 2004.
References
Altshuler E, Cormier D, and Go H (2002) Improvements to the
WAAS ionospheric algorithms, Proceedings of the ION
GPS 2002, Portland, OR, 24-27 September
Coster A J, Foster J C, Erickson P, Sandel B, and Rich F
(2003a) Monitoring space weather with GPS mapping
techniques, Proceedings of the ION National Technical
Meeting 2003, Anaheim, CA, 22-24 January 22-24
Coster A J, Foster J C, and Erickson P J (2003b) Monitoring
the ionosphere with GPS, GPS World, 40-45
Doherty P H, Dehel T, Klobuchar J A, Delay S H, Datta-Barua
S, de Paula E R, and Rodrigues F S (2002) Ionospheric
effects on low-latitude space-based augmentation
systems, Proceedings of the ION GPS 2002, Portland, OR,
September 24-29
DoT (1999) Federal Aviation Administration Specification
Wide Area Augmentation System (WAAS), FAA-E-
2892B, Change 1
El-Arini M B, Conker R, Albertson T, Reegan J K, Klobuchar J
A, and Doherty P (1995) Comparison of real-time
ionospheric algorithms for a GPS wide-area
augmentation system (WAAS), NAVIGATION: Journal of
the Institute of Navigation, 41(4):393-413
El-Gizawy M, and Skone S (2002) A Canadian ionospheric
warning and alert system, Proceedings of the ION GPS
2002, Portland, OR, 24-27 September
Foster J C, Erickson P J, Coster A J, Goldstein J, and Rich F J
(2002) Ionospheric signatures of plasmaspheric tails,
Geophysical Research Letters, 10.1029/2002GL015067
Foster J C, and Vo H B (2002) Average characteristics and
activity dependence of the subauroral polarization
stream, Journal of Geophysical Research,
0.1029/2002JA009409
NovAtel (2002) Modulated Precision Clock User Manual
Skone S and Cannon M E (1999) Adapting the wide area
ionospheric grid model for the auroral region, Canadian
Aeronautics and Space Journal, 45(3):236-244
Van Dierendonck A J, Hua Q, Fenton P, and Klobuchar J
(1996) Commercial ionospheric scintillation monitoring
receiver development and test results, Proceedings of the
ION 52nd Annual Meeting, Cambridge, MA, 19-21 June
Walter T, Hansen A, Blanch J, Enge P, Mannucci T, Pi X,
Sparks L, Iijima B, El-Arini B, Lejeune R, Hagen M,
Altshuler E, Fries R, and Chu A (2000) Robust detections
of ionospheric irregularities, Proceedings of the ION GPS
2000, Salt Lake City, UT, 19-22 September
Wilson B D, Yinger C H, Feess W A, and Shank C (1999) New
and improved - the broadcast inter-frequency biases, GPS
World, 10(9):56-66