Status of Ionospheric Radio Occultation CHAMP Data

Status of Ionospheric Radio Occultation CHAMP Data Analysis and Validation of Higher Level Data Products Norbert Jakowski1 , Andreas Wehrenpfennig1 , Stefan Heise1 , Christoph Reigber2 , and Hermann L¨ uhr2 1

2

Deutsches Zentrum f¨ ur Luft- und Raumfahrt (DLR), Institut f¨ ur Kommunikation und Navigation (IKN), Kalkhorstweg 53, 17235 Neustrelitz, Germany GeoForschungsZentrum Potsdam (GFZ), Division 1, Kinematics and Dynamics of the Earth; Division 2, Solid Earth Physics and Disaster Research

Summary. The GPS radio occultation technique is a rather simple and inexpensive tool for profiling the electron density of the entire ionosphere from satellite orbit heights down to the bottomside. No other profiling technique (bottomside/topside vertical sounding, incoherent scatter) unifies profiling through the entire ionosphere with global coverage. First results of ionospheric radio occultation (IRO) measurements carried out on board the German CHAMP satellite are reported. To inform potential users we review IRO data products operationally generated in DLR/IKN Neustrelitz and distributed via the Information System and Data Center (ISDC) of GFZ. Methods and algorithms applied for product generation including their limitations in accuracy and spatial resolution are addressed. We present the status and results of product validation with independent data sources. The achieved accuracy of the retrieved electron density profiles is estimated in particular by comparing the IRO results with independent data obtained from vertical sounding stations on Earth and from the Langmuir probe on board CHAMP. It is concluded that CHAMP-IRO measurements have the potential to establish global data sets of electron density profiles for developing and improving global ionospheric models and to provide operational space weather information. Key words: ionosphere, limb sounding, inversion technique, spaceborne GPS, CHAMP

1 Introduction LEO missions such as CHAMP that carry a dual frequency GPS receiver on board, offer a unique chance to improve our knowledge about the ionospheric behavior and to monitor the actual state of the global ionosphere on a continuous basis. The GPS limb sounding technique has been demonstrated to be a powerful tool for remote sensing the Earth’s neutral atmosphere and ionosphere by analyzing GPS radio occultation data obtained from the GPS/MET instrument, flown on the Microlab-1 LEO satellite e.g. [5], [1]. The German CHAMP satellite has been successfully launched by a Russian

CHAMP Ionospheric Radio Occultation

463

COSMOS rocket on July 15, 2000 [7]. The advanced ”Black Jack” GPS receiver developed by the Jet Propulsion Laboratory (JPL) provides on the one hand precise time and orbit information. On the other hand the receiver measures GPS carrier phases in the limb sounding mode starting at CHAMP orbit tangential heights down to the Earth surface. Whereas the sampling rate is 1 Hz for ionospheric radio occultation measurements, the sampling rate turns up to 50 Hz from about 100 km tangential height downward for retrieving neutral gas temperature and water vapor profiles. The first radio occultation measurements performed on board CHAMP in February 2001 were successfully used to retrieve vertical temperature profiles of the global troposphere/stratosphere system [10]. The first ionospheric radio occultation (IRO) measurements carried out on board CHAMP on 11 April 2001 yielded reasonable electron density profiles [4]. In this paper we review the status of IRO data analysis and report validation results achieved so far including all IRO measurements from 11 April 2001 (day 101) until 31 December 2001 (day 365).

2 IRO measurements and data processing The CHAMP satellite data are received at the DLR Remote Sensing Data Center Neustrelitz and then passed over to the GFZ. Immediately after data dump IRO GPS data are automatically processed in DLR/IKN. The computed data products are made available to the international science community via the Information and Science Data Center (ISDC) of GFZ. To fulfill operational requirements of potential users in science and space weather applications, there has been developed a dynamically configurable processing system in the preparatory phase of the CHAMP mission [9]. The measured GPS data are automatically checked and preprocessed by this system. Due to the modular structure high flexibility is achieved so that retrieval modules can be modified or replaced in the course of the CHAMP mission and supplementary data can be included. Before discussing the data products itself, we briefly review all IRO measurements that have been carried out on board CHAMP in 2001. A comprehensive view on all IRO measurements on board CHAMP and corresponding successfully retrieved electron density profiles is given in Fig. 1. Although the number of IRO measurements amounts to about 200-300 per day, the number of retrieved profiles may be rather small in particular during the first few weeks because the GPS scheduling was not optimized. So, the measurements very often started too late (at tangential heights < 300 km) or ended too early (at tangential heights > 200 km). Since such measurements were automatically rejected by the retrieval algorithm, the number of retrieved profiles may be considerably smaller. Due to several software upgrades the number of successfully retrieved profiles increased to about 50 % of the measurements. Unfortunately, due to the remaining scheduling problems

Norbert Jakowski et al. 4500 4000

number

464

measurements profiles

CHAMP RO Ionosphere 2001

3500 3000 2500 2000 1500 1000 500

month

0 Apr 2001

May 2001

Jun 2001

Jul 2001

Aug 2001

Sep 2001

Oct 2001

Nov 2001

Dec 2001

Fig. 1. Review of monthly IRO measurements on board CHAMP in comparison with retrieved electron density profiles during 2001.

of the GPS receiver software, ionospheric and tropospheric occultations could only be measured in an alternative mode that caused considerable gaps in IRO measurements, as is visible in Fig. 1. From about 21000 measurements obtained in 2001 some 7000 vertical electron density profiles were derived (33%).

3 Data products available via the ISDC of GFZ The radio occultation measurements of the GPS satellites carried out with a sampling rate of 1 Hz include dual frequency L1/L2 carrier phase as well as code phase measurements. The original occultation data are collected over a full day and subsequently stored in a single file defining a level-1 IRO data product in RINEX format. Since the basic information for ionospheric radio occultation studies or retrievals is provided by the total electron content along the radio occultation paths, differential carrier phases are used to generate hourly files of relative TEC data as a value added data product. These data are made available to the user community via the Information System and Data Center (ISDC) of the GFZ. The 1s sampled TEC data are also internally used in the data processing system as input data for the subsequent inversion procedure to retrieve vertical electron density profiles of the ionosphere. To guarantee an operational retrieval of the received data within 3 hours after data dump, additional ionospheric information about the horizontal structure of the ionosphere is not taken into account. Consequently, spherical symmetry of the electron density distribution is assumed for retrieving vertical electron density profiles. Furthermore, to produce stable solutions in the operational mode under quite different geophysical conditions that may seriously violate the assumed spherical symmetry, we took care to develop a


465

robust IRO retrieval technique (see subsequent section). The available products are summarized as follows: 1. IRO GPS data in RINEX format (daily) 2. Geo-located relative TEC along occultation ray paths (hourly) 3. Geo-located electron density profiles (individual IRO events) The operational data are roughly controlled to filter out obviously erroneous data sets. Furthermore, the goodness of the IRO derived electron density profiles is estimated by a quality parameter that is provided in the header of the data products. The data products are made available to the international science community via the ISDC of GFZ.

4 Data retrieval technique Since a number of ionospheric phenomena are accompanied by strong spatial plasma density gradients, and the path through the ionosphere is in the order of 1000 - 2000 km, the spherical assumption of the Abel inversion technique does not hold in general. To overcome this methodological restriction, tomographic solutions are required. Although, as mentioned above, the operational retrieval assumes spherical symmetry, a tomographic approach is established [3]. The applied tomographic approach has the advantage that additional information from ground based GPS measurements, models and/or other sources such as peak electron densities obtained at vertical sounding stations can easily be included in the reconstruction of the electron density profile, at least in post-processing. The measured differential GPS phases provide the total electron content along the ray path through the spherically layered voxel structure as the sum of the product ne x dsi , where ne is the mean electron density of voxel i and dsi corresponds with the ray path element in this voxel. Since the ray path elements are known from the ray path geometry, the electron density of different shells can successively be derived from the 1s sampled measurements when the tangential point of occultation rays comes closer and closer to the Earth down to the bottom of the ionosphere. Practically, the solution starts with the first measurement at the greatest tangential height by using an adaptive model for the topside ionosphere and plasmasphere above the CHAMP orbit height. The adaptive model consists of a Chapman layer whose topside is extended by a slowly decaying exponential term with a scale height of 10000 km. Crucial model parameters such as the plasma scale height at the upper boundary are determined by 6 iterations in order to ensure a smooth transition between model and measurements. The model assisted technique is required due to the rather low orbit height of CHAMP (430 km) which comes close to the peak density height in particular at low latitudes.

466

Norbert Jakowski et al.

5 Validation results Before any retrieved electron density profiles can be used in ionospheric physics, the accuracy and reliability of the retrievals have to be estimated. This can be done by comparison with measurements obtained by quite different techniques such as vertical sounding, incoherent scatter radar or in situ measurements. Principally, it has to be stated that the IRO derived electron density profiles provide a unique measure. The limb sounding technique provides profiles that describe the mean vertical electron density distribution over a fairly large area with a diameter of about 2000 km. Thus, a comparison with precisely located vertical sounding data is expected to yield differences because they are are strongly influenced by the local ionospheric structure. In Fig. 2 IRO retrieved electron density profiles are compared with corresponding vertical sounding (digisonde) data obtained in Juliusruh (54.6◦ N; 13.4◦ E) and Athens (38.0◦ N; 23.5◦ E) around noon times on 27 and 29 April 2001, respectively. The plots indicate the capabilities for an estimation of the systematic differences between both techniques. Related studies are on-going. In this paper we confine our attention to the profile key parameters such as the peak density NmF2 or peak plasma frequency f0F2 and the peak height hmF2. The deviations from vertical sounding data (coincidence radius: 8◦ , time window: 20 minutes) are estimated by using globally distributed stations that were available in the SPIDR data basis [6]. The computed absolute and percentage deviations are shown in Fig. 3. The distribution of the deviations is based on 267 and 119 values for f0F2 and hmF2, respectively. Both the shift of the absolute distribution function (0.07 MHz for f0F2 and -1.1 km for hmF2) as well as that of the percentage distribution (2.17 % for f0F2 0.6 % for hmF2 ) are rather small indicating a nearly symmetrical scatter. The width of the distribution corresponds to a RMS deviation of 16.6% and 12.7% for f0F2 and hmF2, respectively, agreeing with former estimates [8], [4]. It is

Fig. 2. Comparison of IRO derived plasma frequency profiles with corresponding vertical sounding profiles obtained at the stations Julisruh (left) and Athens (right).


467

Fig. 3. Distribution of absolute (left) and percentage (right) differences between IRO derived f0F2 (N=267) and hmF2 (N=119) values and corresponding vertical sounding observations taken as reference. The center of the difference distribution functions are shifted by +0.07 MHz/2.17 % for f0F2 and by -1.1 km/0.6 % for hmF2.

assumed that a better agreement between IRO and VS derived key parameters can only be achieved in case of a rather ’smooth’ ionosphere. Further data will be collected and analyzed to improve the statistics. A very effective way of checking the topside electron density profiles is a comparison with the Planar Langmuir Probe (PLP) data obtained along the satellite track. The PLP is part of the Digital Ion Drift Meter and was provided by the US Air Force Research Laboratory, Hanscom, MA. Fig. 4 shows

Fig. 4. Scatter plot of in situ measurements of the electron density obtained from the Langmuir probe versus IRO derived electron densities in CHAMP orbit height. The occultation measurements are separated from the CHAMP orbit plane by less than 7◦ ( < 800km).

468


Fig. 5. Latitudinal distribution of the IRO derived peak electron density at daytimes (10-15 LT) during the days 100-244 (left) and 285-365 (right) at all longitude zones.

a scatter plot of IRO data against corresponding Langmuir probe data. To obtain a reasonable number of coincidences, the angular distance between the location of the IRO event and the satellite trace where PLP measurements were carried out, amounts up to 7◦ in this figure. Obviously the correlation between IRO and PLP data is quite consistent over a large dynamic range of more than two decades indicating a reasonable estimation of the electron density in the transition region between the adaptive model and IRO measurements.

6 Observations and discussion The first ionospheric radio occultation measurements were carried out on board CHAMP on 11 April 2001. For the entire year more than 7000 vertical electron density profiles distributed on global scale have been derived. This tremendous data set enables already a first physical view on the electron density profiles to get an idea whether well-known ionospheric features become visible. One of the ionospheric key parameters is the peak electron density NmF2 that is correlated with the ionospheric critical frequency f0F2 by the relation N mF 2 = 0.0124 x (f 0F 2)2 in SI units. Fig. 5 (left panel) clearly indicates the strong latitudinal dependence of the ionospheric ionization at day-time hours at all longitude sectors for northern summer conditions. Although there is an enhanced scattering of the data at the southern hemisphere and the summer hemisphere ionization is somewhat higher at mid- up to high-latitudes, the distribution is nearly symmetrically around the equator. When going to northern winter conditions in the right panel, the peak density distribution changes significantly. Again, in conjunction with the enhanced illumination at the summer hemisphere, the mid- to high-latitude range, now


469

Fig. 6. Latitudinal distribution of the IRO derived F2 layer peak heights at daytimes (10-15 LT) during the days 100-244 (left) and 285-365 (right) at all longitude zones.

at the southern hemisphere, indicates an enhanced ionization level compared with winter conditions (left panel). But the mid- to low-latitude range of the southern summer hemisphere is characterized by a significantly smaller peak density than those observed at the winter hemisphere. As Fig. 6 indicates, the corresponding peak density height is generally greater at the summer hemisphere. This is due to the enhanced plasma temperature at the summer hemisphere that is characterized by a stronger solar energy input. This fact correlates also very well with the variation of plasma scale heights Hs(425) derived at the upper boundary during the model assisted retrieval procedure as described in section 4 (Fig. 7). By the way, the rather homogeneous be-

Fig. 7. Latitudinal distribution of the IRO derived plasma scale heights Hs at 425 km height during day-times (10-15 LT) on the days 100-244 (left) and 285-365 (right) at all longitude zones.

470


havior of this plasma scale height in the transition region between model and measurements indicates a stable retrieval procedure. In both figures the summer hemisphere indicates higher plasma scale heights that can be explained by enhanced plasma temperatures. In fact, this effect is emphasized in the computed equivalent slab thickness τb = T EC(hmF 2)/N mF 2 of the bottom side electron density profile (Fig. 8). This slab thickness parameter seems to

Fig. 8. Latitudinal distribution of the IRO derived bottomside slab thickness during day-times (10-15 LT) on the days 100-244 (left) and 285-365 (right) at all longitude zones.

be a very sensitive parameter related to the plasma temperature. In both figures strong gradients between the summer and winter hemisphere exist, indicating that in particular the bottomside ionosphere is very sensitive to solar radiation variations.

7 Conclusions We have reported the status of the generation of ionospheric data products computed from GPS radio occultations on board CHAMP since 11 April 2001. Although the validation work is not yet finished, ionospheric data products are available via the ISDC. Potential users should be aware of the principal averaging character of IRO derived electron density profiles. If compared with localized vertical sounding measurements, the F2 layer peak electron density f0F2 and the corresponding height hmF2 agree within a spread of 17 % and 13 % , respectively. This agreement might be improved, if additional information, e.g. on horizontal gradients or local densities, is included in the retrieval procedure. Horizontal gradients can be deduced from horizontal TEC maps as they are produced from ground based GPS measurements in DLR/IKN since 1995 (e.g. [2], http://www.kn.nz.dlr.de ). Furthermore, the


471

in situ electron density data of the PLP may be used as an additional input to improve the retrieval results in comparison with ionosonde data, at least in post-processing. An essential improvement is expected, if IRO retrievals and tomographic reconstructions of the topside electron density distribution (see Heise et al. in this issue) are combined to get a comprehensive view on the entire vertical electron density structure of the ionosphere from the bottomside up to GPS orbit heights. The discussion of the early results has shown that the IRO retrieved electron density profiles and deduced parameters provide a consistent description of the general ionospheric behavior. This fact indicates that IRO data should have a big potential for studying a number of ionospheric phenomena on global scale and for improving ionospheric models. Since IRO derived ionospheric electron density profiles may be retrieved operationally, satellite missions carrying a GPS receiver on board such as CHAMP are very promising candidates for space weather monitoring tasks.

Acknowledgements The authors are very grateful to all the colleagues from the international CHAMP team. We thank Jens Mielich (IAP, Germany) and Ioanna Tsagouri (NOA, Athens) for providing vertical sounding data for comparison. This study was carried out under grant number 01SF9922/2 of the German Federal Ministry of Education and Research (BMBF).

References 1. Hajj GA and Romans LJ (1998) Ionospheric electron density profiles obtained with the Global Positioning System: Results from the GPS/MET experiment. Radio Sci, 33, 175-190 2. Jakowski N (1996) TEC Monitoring by Using Satellite Positioning Systems. In: Kohl H, R¨ uster R, Schlegel K (Eds) Modern Ionospheric Science. EGS, Katlenburg-Lindau, ProduServ GmbH Verlagsservice, Berlin, pp 371–390 3. Jakowski N (1999) Capabilities of radio occultation measurements onboard LEO satellites for ionospheric monitoring and research. In: Vernon A (Ed) Proc. 4th COST 251 Workshop ’The Impact of the Upper Atmosphere on Terrestrial and Earth-Space Communications’, 22-25 March, Funchal, Madeira, Portugal, pp116-121 4. Jakowski N, Wehrenpfennig A, Heise S, Reigber Ch, L¨ uhr H, Grunwaldt L, Meehan T (2002) GPS Radio Occultation Measurements of the Ionosphere from CHAMP: Early Results. Geophys Res Lett, accepted 5. Kursinski ER, Hajj GA, Schofield JT, Linfield RP, Hardy KR (1997) Observing Earth’s atmosphere with radio occultation measurements using the Global Positioning System. J Geophys Res, 102, D19, 23429-23465 6. O’Loughlin KF (1997) SPIDR on the Web: Space Physics Interactive Data Resource on-line analysis tool. Radio Sci, 32, 2021-2026

472


7. Reigber Ch, L¨ uhr H, and Schwintzer P (2000) CHAMP mission status and perspectives. Suppl to EOS, Transactions, AGU, 81, 48, F307 8. Schreiner WS, Sokolovskiy SV, Rocken C (1999) Analysis and validation of GPS/MET radio occultation data in the ionosphere. Radio Sci, 34, 949-966 9. Wehrenpfennig A, Jakowski N, Wickert J (2001) A Dynamically Configurable System for Operational Processing of Space Weather Data. Phys Chem Earth, 26, 601-604 10. Wickert J, Reigber Ch, Beyerle G, K¨ onig R, Marquardt Ch, Schmidt T, Grunwaldt L, Galas R, Meehan T, Melbourne WG, Hocke K (2001) Atmosphere sounding by GPS radio occultation: First results from CHAMP. Geophys Res Lett, 28, 3263-3266