The International GPS Service (IGS) Ionosphere Working Group

Available online at www.sciencedirect.com

SCIENCE DIRPCT.

doi: lO.l016/SO273- 1 177(03)00029-2

Pergamon

www.elsevier.com/locaIe/asr

THE INTERNATIONAL GPS SERVICE (IGS) IONOSPHERE WORKING GROUP

J. Feltens

Electronic Data Systems (EDS) at Navigation Support O&e, European Space Agency (ESA), European Space Operations Centre (ESOC), Robert-Bosch-Sk 5,

D-64293 Darmstadt, Germany

ABSTRACT

This article is presented on behalf of the International GPS Service (KS) Ionosphere Working Group (Iono_WG) and describes the Working Group’s activities. Detailed information about the IGS in general can be found on the IGS Central Bureau Web page: http://igscb.jpl.nasa.gov. The Iono_WG commenced working in June 1998. The working group’s main task is the routine provision of ionosphere Total Electron Content (TEC) maps with a 2-hours time resolution and of daily sets of Global Positioning System (GPS) satellite and receiver hardware differential code bias (DCB) values. The computation of these TEC maps and DCB sets is based on the routine evaluation of GPS dual-frequency tracking data recorded with the global IGS tracking network. Currently five so called IGS Ionosphere Associate Analy- sis Centers (IAACs) produce with their models and software routinely TEC maps and DCB sets and provide their ionosphere products to IGS Ionosphere Associate Combination Center (IACC) at ESA/ESOC. Once per week the IACC compares then these ionosphere products with a dedicated comparison algorithm. This algorithm is currently being upgraded to be able to compute a combined IGS ionosphere product from the individual IAACs inputs (at the publication date of this paper the upgrade was completed). This new product shall soon be made available to external users through the Crustal Dynamics Data Infor- mation Center (CDDIS), NASA Goddard Space Flight Center, Greenbelt, MD, U.S.A. Beyond the routine provision of ionosphere products, the Iono_WG intends to support the ionosphere community also with other activities, e.g. by using the IGS global tracking network and capabilities to run high-rate data campaigns during events of special relevance for the ionosphere. In such campaigns, dual frequency GPS re- ceivers record, depending on a receiver’s capability, with 1 second or with 3 second sampling rate. A first such campaign was organized during the total solar eclipse on 11 August 1999, a second campaign was run recently under the name “HIRAC/SolarMax” over 7 days from 23 - 29 April 2001. In the medium- and long-term, the working group intends to develop more sophisticated algorithms for deducing mappings of ionospheric parameters from GPS measurements and to realize near-real-time availability of IGS ionosphere products. It is the intent of this paper to give an overview over the lono_WG activities and to point out the importance of GPS for routine ionosphere monitoring. 0 2003 COSPAR. Published by Elsevier Science

Ltd. All rights reserved.

INTRODUCTION The Working Group started its routine activities in June 1998. Several IAACs provide twelve global

TEC maps per day with a 2-hours time resolution and a daily set of GPS satellite DCBs in IONEX format files. IONEX stands for IONosphere map Exchange format and is a specific format in which the IAACs provide their TEC maps and DCB values (Schaer et al., 1998). The routine provision of daily ground station DCBs is under preparation (at the publication date of this paper the ground station DCBs were

A& Space Rcs. Vol. 31, No. 3. pp. 635-644. 2003 0 2003 COSPAR. Published by Elsevier Science Ltd. All rights reserved Printed in Great Britain 0273- 1177103 $30.00 + 0.00

included). The principle of deducing the TEC from dual-frequency GPS-measurements works, in short, as follows: The ionosphere is a dispersive medium, i.e. the amount of delay an electromagnetic signal suffers when travelling through the ionosphere, is proportional to the signal’s frequency. The GPS satellites broadcast their navigation signals on two frequencies, thus allowing in principle to measure ranges (with code or carrier) on these two frequencies, i.e. giving pairs of measured ranges (one on each frequency). Per pair each of the two observed ranges is delayed by the ionosphere to an amount proportional to the TEC experienced in the ionosphere along the signal path and, to the first order, inverse proportional to the square of the signal frequency. The two ranges of one pair can now be combined so that the ionospheric delay is eliminated to an order of 980/o, on the other hand they can also be combined in such a way that the TEC can be derived from them and then used as “TEC observables” for the deduction of ionospheric TEC maps. This is, in a few words, the principle of TEC deduction from GPS measurements. Corrections have to be applied to these “observed” TEC values, e.g. for the integer ambiguity inherent in the GPS carrier measurements, and, since the electronic pathes of the GPS signals are different for the two frequencies in the GPS satellites as well as in the receiver electronics, the so called GPS satellite and receiver DCBs have to be accounted for in the TEC observables modeling. A nice reference describing in detail the method of deriving TEC values and DCBs from GPS measurements is e.g. Newby (1992).

Currently five IAACs contribute with ionosphere products: ??CODE, Center for Orbit Determination in Europe, Astronomical Institute, University of Beme,

Switzerland. ??ESOC, European Space Operations Centre of ESA, Navigation Support Office, Darmstadt,

Germany. ??JPL, Jet Propulsion Laboratory, M/S 138-308, Pasadena, California, U.S.A. ??NRCan, National Resources Canada, Geodetic Survey of Canada Geomatics Canada, Ottawa,

Ontario, Canada. ??UPC, Polytechnical University of Catalonia, Group of Astronomy & GEomatics, gAGE/UPC,

Barcelona, Spain. The mathematical approach used by the different IAACs to establish their TEC maps are quite differ-

ent. Details about the individual IAACs modeling can be found in e.g. Schaer (1999), Feltens (1998), Mannucci et al. (1998), Gao et al. (1994), Hemandez-Pajares M. et al. (1999).

The members of the Iono_WG are (due to a number of 41, the members are listed here without affiliation): Gerhard Beutler, Dieter Bilitza, Ljiljana R. Cander, Mihail Codrescu, Anthea J. Coster, Robert E., Jr. (Rob) Daniell, Patricia H. Doherty, John Dow, Mariangel Fedrizzi, Joachim Feltens, Mariusz Figurski. Werner Gurtner, Manuel Hemandez-Pajares, Pierre Heroux, Ildy Horvath, Norbert Jakowski, Ame Jungstand, Ivan Jelinek Kantor, John A. (Jack) Klobuchar, Attila Komjathy, Jan Kouba, Richard B. Langley, Reinhard Leitinger, Tony Mannucci, Angelyn W. Moore, Steven Musman, Ruth Neilan, Ola 0vsteda1, Kohichiro Oyama, Eurico Rodrigues de Paula, Christoph Reigber, Esther Sardon-Perez, Stefan Schaer, Lawrence Sparks, Peter Stewart, Rene Warnant, Robert Weber, Brian D. Wilson, Xiaoqing Pi, Yunbin Yuan, Rene Zandbergen.

COMPARISONS Once per week the ionosphere products from the different IAACs are compared at the IACC at E&4/

ESOC. A dedicated computer program was established at ESOC from scratch to do that task. In principle this algorithm computes for each of the twelve reference epochs an “unweighted mean” TEC map from the different IAAC TEC maps. Based on the residuals of the different IAAC TEC maps with respect to that “mean” TEC map a weight can be derived for each IAAC. Based on the weights thus obtained, a “weighted mean” TEC map is computed in a second step. By inspecting the residuals of the individual

GPS Ionosphere Working Group 631

IAAC TEC maps with respect to that “weighted mean” TEC map per reference epoch, statistics about the agreement between the different IAACs TEC maps are established.

Fig. 1. The IGS global “weighted mean” TEC maps for 28 March 2000 in units of 1016 m-* [TECU] for UT=Ol:OO (top panel), 0390, and 0590; these are based on the IONEX file for 28 March 2000 containing the COMBINED IGS TEC MAPS and DCBs.

638

TEC hi@ (hoi ht= 450.0 km at 2000 03/X 11sJOzOO w-loctu%.~~~~lcdTEc~md~

Fig. 1 cont. The IGS global “weighted mean” TEC maps for 28 March 2000 in units of 1016 rns2 [TECU] for UT=O7:00 (top panel), 0930, and 11:OO; these are based on the IONEX file for 28 March 2000 containing the COMBINED IGS TEC MAPS and DCBs.


Fig. 1 cont. The IGS global “weighted mean” TEC maps for 28 March 2000 in units of 1016 rns2 [TECU] for UT=13:OO (top panel), 15:00, and 1790; these are based on the IONEX file for 28 March 2000 containing the COMBINED IGS TEC MAPS and DCBs.

640

t= 450.0 km at 2000 g/alzOg d_ lh. LNErl Id

Fig. 1 cont. The IGS global “weighted mean” TEC maps for 28 March 2000 in units of lOI me2 [TECU] for UT=19:00 (top panel), 21:00, and 23:OO; these are based on the IONEX file for 28 March 2000 containing the COMBINED IGS TEC MAPS and DCBs.


The “weighted mean” TEC map for each reference epoch, obtained quasi as by-product when doing the comparison in this way, could be considered as something like a “combination” of the input IAAC EC maps. The same holds for the comparison of DCBs, which is done basically in the same way.

However, the IAACs use very different approaches to establish their TEC maps, resulting in very different temporal and spatial resolutions. These circumstances reflect also in the comparison results; the weighting scheme in the comparison algorithm must be improved. Software upgrades for an improved weighting scheme are currently under work (at the publication date of this paper these upgrades were completed). The next chapter will tell more about the new weighting scheme. As an example of results obtained with the current (old) weighting scheme Figure 1 shows the sequence of IGS “weighted mean” TEC maps of 28 March 2000, a day during a period in the current solar maximum, when the TEC level was very high.

The other important subject of comparisons are the DCBs. When directly comparing the DCB-series of the different IAACs, one can see an overall agreement in the order of about 0.3 ns (0. I ns correspond to an ionospheric range delay in the order of 3 cm). According to S. Schaer, private communication (2000), mean IAAC satellite DCB series show an agreement of about 0.1 ns, while the day-by-day variations are significantly higher. For his analysis Schaer took three months (GPS weeks 1065 - 1077) of daily satellite DCB sets from the IAACs IONEX files and computed from these daily values IAAC-specific mean DCB sets. The five mean DCB sets thus obtained were then compared with respect to each other and also with respect to an overall mean set (AU). Table 1 is an extract of Schaer’s presentation and shows the obtained root-mean-square (rms) errors (in nanoseconds ns). When interpreting these numbers one has to keep in mind that some IAACs estimate their DCB sets together with their TEE maps, while others make separate program runs for this. Some IAACs introduce constraints in their DCBs estimation, while others do not.

Table 1. Agreement of the distinct IAACs satellite DCB sets in [ns] (courtesy S. Schaer)

NRcan Esoc J?% All

CODE 0.122 0.106 0.110 0.370 0.094

NRcan 0.109 0.144 0.371 0.104

0.118 0.373 0.095

0.393 0.117

0.296

VALIDATIONS NRCan and UPC have proposed two self-consistency methods. Both are in principle based on the

analysis of residuals resulting from the comparison of directly from GPS-observables derived TFC values with corresponding TEC values interpolated from the IAACs TEC maps, in order to assess the quality of the distinct IAACs TJZC maps. This is done with GPS data collected at ground stations equally distributed in a global geographic grid. Table 2 lists these globally distributed ground stations.

Table 2. Globally distributed ground stations used for validations with GPS-observables.

Thule, cp = 76.5

I

Ny-Alesund, cp = 78.9

I

TM, cp = 71.6 OteenlatuWenmark i = -68.8 Svalbard, Norway ii = 11.9 Russian Federation ii = 128.9 I

Algonquin Park, cp = 46.0 I

VillqFanca, cp = 40.4 I

Usuda, cp =36.1 Canada E. = -78.1 Spain E, = -4.0 Japan h = 138.4 I

Kourou, cp = 5.3 hZalindi, cp = -3.0 @ezonCity, cp = 14.6 i%x?nch Guyana h = -52.8 Kenya h = 40.2 Phillipines h = 121.1

Santiago, cp = -33.2 Sutherland, cp = -32.4 Perth, cp = -31.8 Chile ?L = -70.7 South @ica h = 20.8 Australia ii = 115.9

o*Iiiggins, Antarctica

cp = -63.3 East Ongle Island. cp = -69.0 CaseY* cp = -66.3 ?b = -57.9 Antarctica h = 39.6 Antarctica )c= 110.5

cp is the approximate latitude, h is the approximate longitude of the station.

The new geographic-dependent weighting scheme for the comparison/combination algorithm is based on the output of these two methods. A detailed description on how these two methods work can be found in Feltens (2000). The “weighted mean” TEC maps coming out with this new weighting scheme will then be the official combined IGS ionosphere product.

Beyond that, the Iono_WG intends to perform other kinds of validation, which are in short: 1) JPL proposed to make validations by comparing IAACs model vertical TEC values with TEC values derived from the Ocean Topography Experiment (TOPEX) Satellite altimeter data (validations of this type with Jason satellite and Envisat satellite altimeter data are under preparation). This type of validation shall soon be attached to the weekly comparison program runs (at the publication date of this paper the TOPEX validations were attached to the weekly comparison runs). Figure 2 shows in the form of histograms the results of an analysis made at UPC with TOPEX data for each of the five IAACs for the year 2000; the Number of TOPEX observations are shown on the ordinate and the differences {TEC(TOPEX) minus TEC(GPS)} in TEC-units [TECU] (1 [TECU] = 1*1016 electrons/m*) are shown on the abscissa. When interpreting these histograms it should be taken into account that, in these plots, a negative bias means that the estimated TEC with GPS (h c 20200 km) is greater than the TOPEX TEC (h c 1300 km). Then a positive bias means directly a mean underestimation of the TEC with GPS at least equal to the bias (the TOPEX accuracy seems to be 2 TECU).

2) The Deutsche Forschungsanstalt fur Luft.- und Raumfahrt e.V. (DLR) Femerkundungsstation Neustrelitz has proposed to make a validation with ionosondes data: Ionosondes data provide information about the critical frequencies of the different ionosphere layers. For the proposed type of validation the critical frequencyfoF, of the Frlayer shall only be used. The electron density N,F2 is proportional to the square of the critical frequency NmF2 - crd;i)*. By using TEC values derived from the different IAAC TEC-models, equivalent slab thickness values 7EC/NrnF2 could be computed. The equivalent slab thickness is a quite sensible indicator for the GPS-derived TEC-values, especially at low TEC-levels, having typically values of 200 - 400 km over daytime. Systematic studies could find out typical daily and seasonal variations of that parameter, e.g. higher daytime values in summer than in winter.

3) D. Bilitza has proposed to use the International Reference Ionosphere (IRI) as a background model for testing the different algorithms: First the IRI would be used to compute measurements the GPS satellites would make in an IRI ionosphere, then these measurements are fed into the algorithm to be tested, and the outcoming results are compared with the “true” IRI ionosphere. UPC was the first IAAC that did this type of validation with its algorithms.


4) ESA intends to use the different IAACs-models in ESA’s Remote Sensing Satellite (ERS) and Envisat orbit determinations and to analyze the order of orbit determination residuals resulting from the usage of the different IAACs models.

CODE NRCan WA

~32 IMC:UXIE Yl3FMw(15.745.163ob~wP) IGSIAAC EYA.YEA.R2m0(15 745.163obMhabm;

40 .%I -a 10 0 10 20 30

4 30 -20 -10 0 10 20 30 10 TEC(TOP~lEC(GPS)[TEC~

Bias (TPX-GPS) rrEcu1; rfd line

JPL UPC

Fig. 2: TOPEX- vs GPS-TEC for the year 2000 (courtesy Hernandez-Pajares).

SPECIAL ACTIVTIES On the occasion of the solar eclipse on 11 August 1999 the Iono_WG organized and coordinated a

special observation campaign. This event was a unique opportunity to demonstrate the power of the GPS technique in monitoring the ionospheric ionization. As the zone of totality crossed Europe, the rather dense portion of the IGS network provided excellent conditions for monitoring the eclipse: 59 IGS ground stations along the eclipse path from North America over Europe to the Middle East recorded on that day high-rate dual-frequency GPS-data, either with a 1 second or with a 3 seconds sampling, depending on a GPS receiver’s capability. These solar eclipse data were archived at the CDDIS at NASA’s Goddard Space Flight Center and can be accessed atfrp://cddisa.gsfc.nusa.gov/gps/99eclipse (Feltens and Nell, 1999).

The current solar maximum represents another unique chance to establish such a high-rate tracking database for ionosphere analyses. The two regions of major interest are in this case: (1) the polar regions and (2) low latitudes including the crest regions at both sides of the geomagnetic equator. For this purpose the “HIRAC/SolarMax (High-Rate Campaign under Solar Maximum Conditions)” campaign was organ-

ized by the Iono_WG and lasted from 23 - 29 April 2001. On 26 April a large solar flare was observed which impacted the ionosphere on 28 April. A list of the stations and a map can be accessed as pdf files via ~p://cddisa.g~fc.nasa.gov/pub/gps/Olsolamzax/solarmcrx_table.pdf and jlp://cddisa.gsfc.nasa.gov/ pub/gps/Olsolarmar/solatma__map.pdj: This IGS activity were coordinated with other ionospheric observation programs or measurement campaigns using ionosondes, the European Incoherent Scatter radar facility (EISCAT), high resolution magnetometers, etc. to obtain a comprehensive view of the geomagnetic and ionospheric state. The high-rate GPS and GLONASS (Global Navigation Satellite System) data are also archived at the CDDIS and are available atfrp://cddisa.g.~fc.nu.su.gov/gps/OZsolannau (Feltens et al., 2001). Similar campaigns are also planned for the future.

CONCLUSIONS AND LOOK INTO THE FUTURE The IGS Ionosphere Working (Iono_WG) commenced its activities in June 1998. Its main task is the

routine establishment and provision of ionosphere products using IGS global GPS tracking data. The worldwide IGS GPS and GLONASS ground stations network (about 180 sites) provides the unique opportunity for a global ionosphere monitoring on a routine basis. Currently five so called Ionosphere Associate Analysis Centers (IAACs) contribute to this routine processing. The individual IAACs ionosphere products are compared once per week at the Ionosphere Associate Comparison Center (IACC) at ESOC. The comparison algorithm currently used is, however, not optimal with regard to the kind of weighting scheme used. A new comparison program version with an improved weighting scheme, based on self-consistency test validations, is currently under testing and will be available in August 2001 (at the publication date of this paper the implementation of the improved weighting scheme was completed). Once these software tests are completed, it is intended to start with the routine delivery of an official IGS ionosphere product.

The next important task will be a significant reduction of the time interval between the recording of the GPS observables and the delivery of ionosphere products derived from these GPS data; currently these are 11 days.

In the medium and long-term the lono_WG models shall be extended and improved, e. g. with regard to special models for certain regional and local areas, or at the IAACs the development of more complex mathematical ionosphere models. Final goal is an independent IGS ionosphere model.

REFERENCES Feltens, J., Chapman profile approach for 3-d global TEC representation, in Proceedings of the 1998 IGS Analysis Centers

Workshop, pp. 285-297, ESOC, Darmstadt, Germany, 1998. Feltens, J., Die Ionospharen-Arbeitsgruppe des Internationalen GPS Service (IGS), in Proceedings Nutionuler Workshop z,um

Weltruumwetter, pp. 161-169, Neustrelitz, Germany, 2000. Feltens, J., and C. Nell, GPS data collected during August 1999 solar eclipse, Crust& Dynamics Data Information System

(CDDIS) bulletin, 15(l), 1999. Feltens, J., N. Jakowski, and C. Nell, High-Rate SolarMax IGSlGPS campaign “HIRAC/SolarMax”, Crustal Dynamics Data

Information System (CDDIS) bulletin, 16(3), 2001. Gao, Y., P. Heroux, and J. Kouba, Estimation of GPS receiver and satellite Ll/L2 signal delay biases using data from CACS, in

Proceedings of the International Symposium on Kinematic Systems in Geodesy, Geomatics and Navigation, pp. 109-l 17, Banff, Canada, 1994.

Hemandez-Pajares M., J.M. Juan, and J. Sanz, New approaches in global ionospheric determination using ground GPS data, Journal of Atmospheric and Solar Terrestrial Physics, 61, 1237-1247, 1999.

Mannucci, A. J., B. D. Wilson, D. N. Yuan, C. H. Ho, U. J. Lindqwister, and T. F. Runge, A global mapping technique for GPS-derived ionospheric total electron content measurements, Radio Science, 33,565582, 1998.

Newby, S.P.. An assessment of empirical models for the prediction of the transionospheric propagation delay of radio signals, University of New Brunswick, Department of Surveying and Engineering, New Brunswick, Canada, Technical Report No. 160, 1992.

Schaer, S., W. Gurtner, and J. Feltens, IONEX: The IONosphere Map Exchange format version 1, in Proceedings of the 1998 IGS Analysis Centers Workshop, pp. 233-247, ESOC, Darmstadt, Germany, 1998.

Schaer, S., Mapping and predicting the Earth’s ionosphere using the Global Positioning System, Dissertation, Astronomical Institute, University of Beme, Beme. Switzerland, 25 March 1999.

Documents

The International GPS Service (IGS) Ionosphere Working Group