Spatio-temporal Characteristics of the Ionospheric TEC Variation for GPSnet-based Real-time Positioning in Victoria

Paper Menu >>

Journal Menu >>

Journal of Global Positioning Systems (2006)

Vol. 5, No. 1-2:52-57

Spatio-temporal Characteristics of the Ionospheric TEC Variation for

GPSnet-based Real-time Positioning in Victoria

Suqin Wu[1], Kefei Zhang[1], Yunbin Yuan[1,2] and Falin Wu[1]

[1] School of Mathematical and Geospatial Sciences, RMIT University, GPO Box 2476V, Victoria 3001, Australia

[2] Institute of Geodesy and Geophysics, Chinese Academy of Sciences, 54 XuDong Road, Wuhan 43007, China

Abstract. The atmospheric effects, especially the

ionosphere, are the key limiting factors for real-time

high accuracy positioning using the network RTK

technique with a medium-to-long-range baseline

separation. To investigate suitable approaches to

improve ionospheric modeling towards a real-time CM-

level positioning using the Victorian continuously

operating reference stations network (i.e. GPSnet)

system under various ionospheric conditions, this paper

investigates both temporal and spatial variations of the

ionospheric total electrons content (TEC) over Victoria

through analysing GPS dual frequency data from the

GPSnet over a period of two years. Diurnal and seasonal

ionospheric variations, and winter anomaly of the

ionosphere in Victoria are investigated based on GPS-

derived TEC values. Results suggest that the temporal

and spatial TEC variations over Victoria are

complicated. This complex nature of the ionosphere

suggests that it is a challenging task to precisely

represent the behaviours of the ionosphere if only a

single and simple ionospheric model is used for all the

time for RTK uses. It is therefore, necessary to develop

new mathematical models or new procedures for precise

representation of the ionospheric TEC variations in

Victoria using a long period of GPS dual frequency

observations, particularly the predictability of the

ionosphere changes. It is expected that the new approach

will provide a better guidance for the state-wide

network-RTK solutions.

Keywords: Atmospheric modeling, Ionospheric delay,

Ionospheric modeling, TEC, Network RTK

1 Introduction

It is well known that the rapid development of GPS-

based location based services (LBS) has opened new

avenues for ubiquitous positioning and can potentially

revolutionize almost all geospatial practice, such as

surveying, mapping, geographical information systems,

photogrammetry, risk management and emergency

response, vehicle and asset tracking, machine control,

workforce management and logistics. The key

background facility of the LBS is a backbone ground-

based GPS infrastructure - Continuously Operating

Reference Stations (CORS).

Located on the southern seaboard, Victoria is the second

most populous State in Australia. Currently, a Victorian

network RTK system is under intensive development

and it will provide an integration of the post-processing

service currently available and the new real-time

network RTK service across the entire state (Zhang et

al., 2006; Roberts et al., 2004). The main potential

barriers to this development are long base station

separations and the atmospheric effects, especially the

ionospheric effects (Fotopoulos and Cannon, 2001; Gao

and Liu, 2002; Wyllie and Zhang, 2003). The key issue

to achieve real-time CM-level positioning accuracy for

Victorian network RTK is, therefore, the precise

modeling of the atmospheric effects, especially the

ionospheric delay and variability under various

atmospheric conditions (Zhang and Robert, 2003; Han,

1997). It is necessary to carry out systematic research of

atmospheric effects in Victoria and characterise the

nature and fine structure of the local/regional ionosphere

(Yuan, 2002; Huo et al., 2005). This paper investigates

approaches to improve the precision of ionospheric

modeling for real-time positioning.

A detailed analysis of temporal and spatial variations of

the ionospheric TEC over Victoria is conducted using

Wu et al.: Spatio-temporal characteristics of the ionospheric TEC variation

for GPSnet-based real-time positioning in Victoria 53

dual frequency GPSnet data over a period of two years.

The analysis includes diurnal, seasonal and annual

variations together with winter anomalies of the

ionosphere based on a series of GPS-based TEC

estimates. A complex ionospheric TEC pattern over

Victoria is revealed. This demonstrates the difficulty to

precisely represent the local/regional ionospheric

behaviours in Victoria if only a simple, single

ionospheric model is used. As a result, new development

of adaptable mathematical models and algorithms is

necessary for a precise representation and ultimate

removal of the ionospheric effects.

2 Methodologies

The ionosphere is part of the atmosphere and located

approximately between 60 km to 1000 km above the

Earth’s surface. The ionosphere is usually “condensed”

onto a very thin spherical shell at a certain altitude (H)

between 300 km and 400 km above the Earth’s surface

in GPS research and applications communities. This is

because the majority of the ionosphere free electrons are

distributed at these altitudes.

In this research, a thin ionospheric shell at a fixed

altitude of 350 km is adopted (i.e. H=350 km). The

vertical TEC is parameterized by using a spherical

harmonics (SH) formula which is referred to a solar-

geographical framework, and a trigonometric single-

layer model (SLM) mapping function (i.e. the cosine

function of the variable zenith angle) is used to project

the vertical TEC onto slant directions along the line-of-

sight between a receiver and a satellite (Odijk, 2002;

Komjathy, 1997; Hofmann-Wellenhof et al., 1997). A 3

degrees and orders spherical harmonics (with 16

unknown parameters/coefficients to be estimated) is

adopted in the calculation. GPS dual frequency data is

used for the analyses. The unknown ionospheric

parameters, along with the satellite and station’s

instrumental biases (IBs), are estimated using the least

squares technique. The IBs for the same base station and

satellite pair are considered as a constant value for a day

as both the receiver and the satellite hardware is

relatively stable in the period of one day.

As for the strategy of data selection, calculation, and

analysis, certain periods of the GPSnet data are selected

and processed to calculate daily, hourly, and/or two-

hourly vertical TEC values for every base station to

analyse temporal ionosphere variations. The temporal

ionosphere variations include diurnal variations, seasonal

variations and winter anomalies. To analyse the spatial

variations of the ionosphere over Victoria, the GPS data

from 13 GPSnet stations are used for the calculation and

analyses. Figure 1 shows the current constellation and

status of rural, regional and metropolitan GPSnet stations

(Asmussen, 2005).

In this paper, a preliminary investigation using the IGS

and GPSnet co-located station called MOBS in

Melbourne is conducted for the temporal ionosphere

variations. The computation procedures and results using

the MOBS data will be presented. This forms the first

stage of the research. Detailed analyses over other

GPSnet stations will be conducted at the second stage of

the research, and real-time predictability and modelling

and correction generation and distribution will form the

third stage of the research.

Fig. 1 GPSnet base station network at rural (top) and metropolitan

areas (bottom) (Asmussen, 2005)

3 Data processing, results and analyses

3.1 The GPSnet -Victorian CORS system

GPSnet is a state-wide CORS geospatial infrastructure

that provides high accuracy and homogenous location

information in Victoria primarily for post-processing

service (Millner et al., 2004). It also provides the first

state-wide DGPS correction service capable of achieving

sub-meter accuracy and single base station RTK service

within a nominal 20 km of selected GPSnet stations in

Australia (see Fig. 1). GPSnet correction information is

available from wireless Internet enabled devices and

forms part of the integrated Vicmap products and

services (Millner et al., 2004). The high accuracy

network RTK correction is obtained with a suitable dual

frequency GPS receiver via an internet connection. The

simplest method of receiving the correction signals

54 Journal of Global Positioning Systems

broadcasted over the Internet is through a GPRS enabled

mobile phone. Any connection to the Internet that can be

made from a GPS service, such as through a wireless

laptop computer, or by use of a compact flash card with

a SIM insert, will allow receipt of the GPS correction

signal. However the current main constraint identified

for high accuracy network RTK solutions is the

atmospheric effects, especially the ionospheric delay

under unfavourable conditions.

3.2 IGS data and GPSnet data pre-processing

GPS dual frequency data from both the MOBS IGS

station and GPSnet stations is used for this research. The

sampling interval is 30s. An elevation cut-off angle of 20

degree is adopted to reduce potential multipath effects

for this experiment. Tests for data quality are performed

before calculating the TEC values, especially for

cleaning possible outliers in the measurements. The

geometry-free (L4) linear combination is formed to

eliminate the effects of the geometry, clock errors, and

the troposphere delay. Again, both the satellite and

receiver’s instrumental biases are considered as a

constant parameter for a day and the unknown

ionospheric TEC model parameters and the unknown

IBs are determined using the least squares technique.

3.3 Results and analyses

3.3.1 Temporal TEC variations

Diurnal variations

GPS dual frequency measurements from MOBS in 2004

are first processed with the method described in Section

2. A time series of the vertical TEC (VTEC) at MOBS

for the year is obtained. The solar-geographical reference

frame is adopted to estimate the TEC model parameters.

The calculation results on the diurnal ionospheric TEC

variations in 2004 are shown in Fig. 2. There are 12 sub-

frames in Fig. 2, and the X-axis and Y-axis represent the

universal time for the day and the VTEC value

respectively in each sub-frame. Each sub-frame

represents VTEC variations in the middle day in the

corresponding month (15th of each month). The figure

shows a time series of 12 one-day-a-month diurnal

VTEC variations in 2004. By comparing all the sub-

frames, one can see that there are significant VTEC

differences in both magnitudes and patterns over the 12

month period.

In Melbourne, local time 5:00~9:00am is the morning,

9:00~17:00 is daytime, 17:00~ 21:00 is dusk, and

21:00~5:00am is nighttime, respectively. From Fig. 2,

one can see that the maximum TEC occurs at

approximately 04:00 universal time (or 14:00 local

time), and the minimum TEC at 17:00 universal time (or

03:00 local time).

Fig. 2 Twelve months daily diurnal VTEC values over the MOBS IGS station, 2004 (15th of each month)

(24 hours snapshots of daily VTEC changes)

0510 15 20 25

200

TECU

0510 15 20 25

200

TECU

0510 15 20 25

200

TECU

0510 15 20 25

200

TECU

0510 15 20 25

200

TECU

0510 15 20 25

200

TECU

0510 15 20 25

200

TECU

0510 15 20 25

200

TECU

0510 15 20 25

200

TECU

0510 15 20 25

200

TECU

0510 15 20 25

200

TECU

0510 15 20 25

200

TECU

(Oct)

(Sept) (Nov) (Dec)

(May) (Jun) (Jul) (Aug)

(Jan) (Feb) (Mar) (Apr)

Wu et al.: Spatio-temporal characteristics of the ionospheric TEC variation

for GPSnet-based real-time positioning in Victoria 55

The daytime TEC values are obviously larger than the

nighttime values. However, the occurrence time of the

maximum/minimum ionospheric TEC values varies with

seasons.

Also, the diurnal ionospheric TEC variations associated

with different seasons and months, and both daytime and

nighttime ionospheric TEC variations are different. This

roughly represents the characteristics of the variations of

the ionospheric TEC (and hence the ionospheric

variations) over Victoria. A more detailed result can be

found in Fig. 3 for diurnal variations of the ionospheric

VTEC over MOBS in 2004.

Seasonal variations and winter anomalies

Fig. 3 shows the TEC values obtained from the MOBS

station in 2004. In this figure, X-axis and Y-axis denote

the universal time and the months respectively. This

result can be considered as an approximate representation

of a time series of the corresponding TEC snapshots for

metropolitan Melbourne and regional Victoria. It shows

both diurnal and seasonal ionospheric TEC variations in

2004. From Fig. 3, one can see that in the spring

(September, October and November) and autumn (March,

April and May) the daytime ionospheric TEC values are,

generally speaking, greater than that of the other two

seasons: summer (December, January and February) and

winter (June, July and August). However in the summer

season, the TEC values in December are as great as the

Spring and Autumn and much greater than that of the

other two months of the same season (Jan., Feb.), One

can also see, from each day of November and December,

high TEC values in a day last longer than that of all the

rest months. The results from the winter season shows

that the majority days of the TEC values are very small

and much smaller than that of the other three seasons.

However, the peak values occur in the winter season in

late July and last for a few days. This phenomenon is

called winter anomaly. It is also apparent that the winter

anomaly is not the nighttime phenomenon but the

daytime phenomenon. The corresponding detailed results

for July are plotted in Fig. 4, which illustrates obvious

different diurnal variations among all days in this month.

Fig. 5 shows two types of sunspot number, daily and

monthly sunspot respectively. By comparing the results

above with Fig. 5, it can be found that there exist

relatively large sunspot numbers in July 2004 (see the

arrow in Fig. 5) in comparison with other months of this

year. What is interesting is that, from Fig. 3 it can also be

seen that winter anomaly happened to be in the period of

July as well. However, this may be just coincident as it is

not conclusive. According to Pulido and Garat (1997)

winter anomaly is not necessarily correlated with solar

activity (or sunspot number), especially in southern

hemispheres.

Fig. 3 Diurnal and seasonal ionospheric TEC variations at

MOBS station in 2004

07/01/04 07/06/0407/11/04 07/16/04 07/21/04 07/26/04 07/31/04

MM/DD/YEAR

TECU

Fig. 4 Detailed diurnal variations of the ionospheric VTEC at

MOBS IGS station during July, 2004

2003 2004 2005 2006

120

160

200

2003 2004 2005 2006

Year

120

160

200

Sunspot Number

Daily Sunspot Number

Monthly Sunspot Number

Fig. 5 Sunspot number over the last three years from May 2003

to September 2005

56 Journal of Global Positioning Systems

3.3.2 Spatial variations of the ionospheric TEC

A number of 2-hour snapshots of the ionospheric VTEC

are estimated over the last several years using GPS data

from all GPSnet stations (see Fig. 1). Only the time series

of 12 two-hour snapshots of the VTEC over Victoria

from 14 to 15 April 2003 are shown in Fig. 6 as a typical

example. The 12 two-hour snapshots are taken at 12:00,

14:00, …, 08:00, and 10:00 universal time and are shown

in the 12 sub-figures respectively (Fig. 6). In each sub-

frame, X-axis and Y-axis denote longitudinal and

latitudinal directions, respectively.

From Fig.6, one can see that, apart from the fact that the

VTEC values over the region vary with the time of the

day, the VTEC values over the region corresponding to

each epoch are spatially correlated as well (see the

neighbourhood colours of any specific point in any sub-

frame, no colour “jump” e.g., from blue to yellow or red,

or, from green to red). The characteristics of the spatial

correlation make it feasible to model the ionosphere for

the whole Victoria region. However, from Fig.6, one can

also see more detailed information, for example, some of

the sub-frames show an obvious gradient from south-west

to north-east, whereas others do not show the gradient in

the same direction, or, some of them even do not show a

perfect homogenous gradient. These may bring

difficulties in establishing proper mathematical

expressions to model the ionospheric VTEC over a wide

area like Victoria for high precision, real-time

positioning.

4 Improvement on ionospheric modeling for real-time

applications

There are a number of factors which need to be

considered for high accuracy, real-time positioning

applications. One of the most important factors is the

improvement of computation/distribution efficiency of

the ionospheric corrections and reduction of the time

interval of the ionospheric correction. Another important

issue needing to be addressed is the selection or

development of proper ionospheric models. The

ionospheric modeling over a wide area for real-time is a

very challenging work due to the fact that the variation

characteristics of the ionosphere associated with time and

geographic location may be significantly different, and

attention should be especially paid in some cases, such as

when winter anomaly occurs, or when the gradient of the

ionospheric over the area that is modelled for is not

always homogenous, as so on. one may think that it is

better to select different ionospheric models accordingly

rather than using a single, simple model for all

ionospheric conditions. In fact, as depicted in Figures 2-6,

the ionosphere activities are not always relatively calm,

and again a simple mathematical ionospheric model may

hardly represent the real situation of the ionosphere all

the time even for the mid-latitude areas like Victoria.

Fig. 6 VTEC snapshots over Victoria from April 14 to April 15, 2003. The 12 sub-frames from left to right represent 12 2-hour

VTEC snapshots from 12:00, 14:00, …, 8:00, and 10:00UTC, respectively (unit: TECU)

142.5 14

143.5 144 144.

145 145.5 14

-38

-37.5

-37

-36.5

-36

-35.5

-35

-34.5

142.

143 143.

144 144.

145145.

-38

-37.5

-37

-36.5

-36

-35.5

-35

-34.5

142.

143143.

144 144.

145 .

-38

-37.5

-37

-36.5

-36

-35.5

-35

-34.5

142.

143143.

144 144.

145 .

-38

-37.5

-37

-36.5

-36

-35.5

-35

-34.5

142.5 14

143.5 144 144.

145 145.5 14

-38

-37.5

-37

-36.5

-36

-35.5

-35

-34.5

142.

143 143.

144 144.

145145.

-38

-37.5

-37

-36.5

-36

-35.5

-35

-34.5

142.

143143.

144 144.

145 .

-38

-37.5

-37

-36.5

-36

-35.5

-35

-34.5

142.5 14

143.5 144 144.

145 145.5 14

-38

-37.5

-37

-36.5

-36

-35.5

-35

-34.5

142.5 14

143.5 144 144.

145 145.5 14

-38

-37.5

-37

-36.5

-36

-35.5

-35

-34.5

142.

143 143.

144 144.

145145.

-38

-37.5

-37

-36.5

-36

-35.5

-35

-34.5

142.5 14

143.5 144 144.

145 145.5 14

-38

-37.5

-37

-36.5

-36

-35.5

-35

-34.5

142.5 14

143.5 144 144.

145 145.5 14

-38

-37.5

-37

-36.5

-36

-35.5

-35

-34.5

12:00UTC 14:00UTC 16:00UTC 18:00UTC

20:00UTC 22:00UTC 00:00UTC 02:00UTC

04:00UTC 06:00UTC 08:00UTC 10:00UTC

Wu et al.: Spatio-temporal characteristics of the ionospheric TEC variation

for GPSnet-based real-time positioning in Victoria 57

5 Conclusions

This preliminary research investigates both temporal and

spatial characteristics of the ionospheric TEC variations

in the State of Victoria based on GPS measurements.

The purpose of this research is to investigate an

appropriate model / algorithm / approach for improving

local ionospheric modelling to support network RTK

service for wide areas.

Diurnal, seasonal variations and winter anomalies of the

regional ionosphere in Victoria are investigated and

preliminary results are given. It is concluded that proper

ionospheric modeling are needed to support regional

network RTK positioning over a wide area. The new

approaches may involve a selection of different

ionospheric mathematical models for different

ionosphere variations of different time of day, time of

month, time of year. The availability of the long

observation data from GPSnet is a rich resource and an

ideal testbed for further development of network RTK

algorithms. It is essential to study the regular and

irregular variations, medium and long term trends, and

statistical properties of irregular variations which will

provide a detailed picture of super-short (mins) to short

(1 hour) scale ionospheric changes. We are aiming to

develop new network RTK algorithms that are capable

of providing reliable, seamless, high-accuracy

positioning services across the state of Victoria.

Acknowledgements

This research is supported by the Victoria Partnership for

Advanced Computing through its e-Research Grant

Scheme (Round 9) and The Australian Research Council

grant through its Linkage Scheme (LP0669259 and

LP0455170). The research consortium (LP0455170)

consists of RMIT University (leading institution),

University of NSW, University of Melbourne,

Department of Sustainability and Environment (Victoria)

and Department of Lands (NSW). The research

consortium (LP04669259) consists of RMIT University

and EOS Space Systems Pty Limited. Dr Ron Grenfell

from the School of Mathematical and Geospatial

Sciences, RMIT University is thanked for his

contribution for an early version of the paper which was

presented at GPS/GNSS2005 HK conference.

References

Asmussen H. (2005) An Introduction of GPSnet, VICpos &

MELBpos, invited presentation at School of Mathematical

and Geospatial Sciences, RMIT University, 18 Oct.

Fotopoulos G., Cannon M. (2001) An Overview of Multi-

reference Station Methods for Cm-level Positioning,

GPS Solutions, 4(3):1-10.

Gao Y., Liu Z. (2002) Precise Ionosphere Modeling Using

Regional GPS Network Data, Journal of Global

Positioning Systems, 1(1): 18-24.

Han S. (1997) Carrier Phase-Based Long-range GPS

Kinematic Positioning, PhD thesis, School of Geomatics

Engineering, The University of New South Wales,

Sydney, Australia.

Hofmann-Wellenhof B., Lichtenegger H. and Collins J. (1997)

GPS Theory and Practice, Springer Wien New York,

Fourth Edition.

Huo X., Yuan Y., Ou J., Wen D. and Luo X. (2005) The

Diurnal Variations, Seasonal Variations and Winter

Anomalies of the Ionospheric TEC Based on GPS Data

in China. Progress in Natural Science, 15(1): 56-60.

Komjathy A. (1997) Global Ionospheric Total Electron

Content Mapping Using the Global Positioning System,

PhD thesis, Technical Report No.188, Department of

Geodesy and Geomatics Engineering, the University of

New Brunswick, New Brunswick.

Millner J., Hale M., Standen P. and Talbot N. (2004) The

Development and Enhancement of GPS/GNSS

Infrastructure to Support Location Based Service

Positioning Systems in Victoria, Proceedings of

GPS/GNSS conference, Sydney, 16 pages.

Odijk D. (2002) Fast Precise GPS Positioning in the Presence

of Ionospheric Delays, PhD thesis, Dept of Mathematical

Geodesy and Positioning, Delft University of Technology,

Delft, The Netherlands.

Pulido A.M. and Garat E.F. (1997) The Ionospheric F2

Region Winter Anomaly and Its Dependence on Solar

Activity in the Northern and Southern Hemispheres,

GEOFISICA INTERNACIONAL, 36 (2), also available at

http://serpiente.dgsca.unam.mx/

serv_hem/revistas/fisica/1997/02/martnez.html.

Roberts, C., Zhang, K., Rizos, C., Kealy, A. and Ge, L. (2004)

Real-time Atmospheric Modelling for Centimetre-level

Positioning Based on Global Navigation Satellite System

(GNSS) Continuously Operating Reference Station

Networks, Journal of GPS, Vol.3, No.1-2, pp.218-225.

Yuan Y. (2002) Study on Theories and Methods of Correcting

Ionospheric Delay and Monitoring, PhD thesis, Institute

of Geodes and Geophysics, Chinese Academy of Science.

Wyllie S. and Zhang K. (2003) A Comparison of Ionospheric

Models for Precise Positioning in Victoria, Proceedings

of the 6th International Symposium on Satellite

Navigation Technology, Melbourne, Australia, pp1-18,

Zhang K. and Roberts C. (2003) Network-based Real-time

Kinematic Positioning System: Current Research in

Australia, Proceedings of Geoinformatics and Surveying

Conference, pp 1~12.

Zhang K., Wu F., Wu S., Rizos C., Roberts C., Ge L., Yan T.,

Gordini C., Kealy A., Hale M., Ramm P., Asmussen H.,

Kinlyside D. and Harcombe P. (2006) Sparse or Dense:

Challenges of Australian Network RTK. Proceedings of

IGNSS Conference 2006, July 18-21, 2006, Queensland,

Australia.