1. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 1 Ionospheric Parameter Estimation Using GPS Attila Komjathy, Lawrence Sparks and Anthony J. Mannucci Jet Propulsion Laboratory California Institute of Technology M/S 238-600 4800 Oak Grove Drive Pasadena CA 91109 Email: Attila.Komjathy@jpl.nasa.gov

2. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 2 IONOSPHERE ESTIMATION USING GPS Using GPS signals to measure the ionosphere Understand purpose and operation of reference stations Understand how ionospheric corrections are formed Forming ionospheric measurements from GPS observables Data quality and editing Calibration of GPS data This lecture covers: Why?Topics

3. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 3 Introduction Currently the largest error sources in GPS positioning is that of ionospheric refraction causing signal propagation delays  What can be done? If we have a dual-frequency GPS receiver, then the ionospheric effect can be almost totally accounted for What if we have a single-frequency receiver? –We can ignore the effect and live with the consequences  –We can minimize it using various processing techniques –We can model it using empirical ionospheric models such as the GPS single-frequency Broadcast model, IRI2000 model, PIM, etc. –We can measure it using nearby dual-frequency receiver observations (pseudorange only, carrier-phase only, pseudorange/carrier-phase combined) and apply it as a correction to the single-frequency observations. What is the error in positioning accuracy caused by the ionosphere and how can we reduce it?

4. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 4 Illustration for GPS and Ionosphere IRI-95 profile

5. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 5 Broadcast Ionospheric Model

6. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 6 Broadcast Model: Seasonal Variation

7. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 7 Broadcast Model: Solar Cycle Dependence

8. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 8 International Reference Ionosphere

9. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 9 IRI Model: Seasonal Dependence

10. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 10 The Accuracy of Broadcast Model

11. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 11 GPS Observation Equations GPS pseudorange observation equation: GPS carrier phase observation equation: Range, clock, ambiguity, ionosphere, troposphere, satellite bias, receiver bias, multipath, noise

12. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 12 Generating GPS Ionospheric Observables precise but ambiguous less precise but unambiguous phase-leveled ionospheric observable

13. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 13 GPS Ionospheric Measurements Code measurement Phase measurement

14. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 14 Leveling the Phase Using Code Measurements The level is computed as: where E is the elevation angle. The uncertainty on the level is computed in a rather rough way using a combination of  th (E) and observed pseudorange scatter : The level is computed by averaging PI-LI using an elevation-dependent weighting. Higher elevation data is weighted more heavily. (The weighting is based on historical Turborogue PI-LI noise/ multipath data giving a historical PI-LI scatter of  th (E) where E is elevation.) The TEC sigma in the JPL Processed Data files are the level uncertainty.

15. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 15 Supertruth Data for Three Threads

16. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 16 The Impact of Arc Lengths

17. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 17 Major Error Source: The Code Multipath

18. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 18 Global Ionospheric Mapping: GIM is the slant TEC; is the thin shell mapping function for shell 1, etc; is the horizontal basis function (C 2, TRIN, etc); are the basis function coefficients solved for in the filter, indexed by horizontal (i) and vertical (1,2,3 for three shells) indices; are the satellite and receiver instrumental biases. For three shells, our model is where For single shell, our model is

19. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 19 WAAS Ionospheric Models WAAS planar fit ionospheric model is Pseudo-IGP approach: IPP treated as if it were collocated with IGP are the planar fit parameters, are the distances from the IGP to the IPP in the eastern and northern directions, respectively. are the additional planar fit parameters describing quadratic and cross terms. WAAS-type quadratic fit ionospheric model is where

20. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 20 All-Site GPS Data Processing Algorithm Bias Fixing Algorithm using all available GPS stations worldwide: is the biased phase-levelled ionospheric observable is the thin shell mapping function for shell 1, etc; is the horizontal basis function (C 2, TRIN, etc); are the satellite and receiver instrumental biases. GIM TEC prediction GIM satellite bias estimate Biased TEC observation are the basis function coefficients solved for in the filter, indexed by horizontal (i) and vertical (1,2,3 for three shells) indices;

21. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 21 Single Vs. Three-Shell Model Limitations The concept of multi-shell GIM:Single-shell 2-D maps Does not capture small-scale variations in the ionosphere Multi-shell is more realistic and accurate than the single-shell approximation

22. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 22 Receiver Bias Estimation Precision

23. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 23 Slant TEC Bias-Fixing Method Estimated bias time series: errors caused by GIM, multipath, noise, sub-daily bias drift Bias-removed slant TEC Location of station

24. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 24 Coverage of Daily IGS Network and Regional Networks (10 degree elevation mask; 450 km shell height)

25. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 25 An Example of the Diurnal Variation of TEC for a Geomagnetically Quiet Day Components in TECU, TECU/hour, TECU/km Example for Single Shell Model Results

26. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 26 JPL’s GIM Multi-Shell Model

27. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 27 TOPEX Validation for 12 Oct 2001, Track 10

28. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 28 Recent GIM Validation Using Jason-1

29. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 29 Global Point Plots

30. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 30 Point Plot Differences

31. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 31 October 30, 2003 DST -390 nT at 2315 UT on October 30 2nd Interplanetary Coronal Mass Ejection

32. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 32 November 20, 2003 Storm Details are difficult to interpret Quiet ionosphere following the storm 5-day average of quiet ionosphere removed: structures are easier to detect

33. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 33 An Example for Repeatibility of Estimated Satellite Biases: Multi-Shell versus Single-Shell Multi-shell significantly improves repeatibility in daily bias estimates –We compare bias averages over 7–10 days –Scatter (std. dev.) over a week improved by factor of 2 to 4 Satellite biases –7-day scatter improved from 2–6 cm to 8–24 mm This may indicate reduction of systematic errors in bias estimation 6 cm 0 cm

34. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 34 An Example for Repeatibility in Estimated Receiver Biases:Multi-Shell versus Single-Shell 0.6 m 0 m Receiver biases 7-day scatter improved from 8–64 cm to 0.5–19 cm Larger scatter due to stations in low latitude sector Systematic error? Examine long time-series of biases Look for shifts in ionospheric delay level for all biases simultaneously

35. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 35 Comparison of Single and Multi-Shell Results for ENG1 Postfit Residuals Prediction Residuals ENG1 = English Turn, LA Improvement at low elevation angles

36. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 36 Comparison of Single and Multi-Shell Results for MBWW Postfit Residuals Prediction Residuals MBWW = Medicine Bow, WY Improvement at low elevation angles

37. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 37

38. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 38 Six-satellite COSMIC constellation Launched April 14, 2006 Low-Earth Orbiter GPS Electron Density Profile COSMIC coverage: 2500 profiles/day COSMIC Ionospheric Weather Constellation

39. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 39

40. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 40 Global Assimilative Ionospheric Model Data Assimilation Process Driving Forces Driving Forces Mapping State To Measurements Mapping State To Measurements Physics Model Physics Model Kalman Filter State and covariance Forecast State and covariance Analysis Adjustment Of Parameters 4DVAR Innovation Vector Kalman Filter –Recursive Filtering –Covariance estimation and state correction Optimal interpolation Band-Limited Kalman filter 4-Dimensional Variational Approach –Minimization of cost function by estimating driving parameters –Non-linear least-square minimization –Adjoint method to efficiently compute the gradient of cost function –Parameterization of model “drivers”

41. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 41 Input Data Types Ground GPS Data (Absolute TEC) >200 5-min. to Hourly Global GPS Ground Stations Assimilate >300,000 TEC points per day (@ 5 min rate) per day Space GPS Data (Absolute or Relative TEC) CHAMP (@ 440 km) SAC-C (@ 700 km) IOX (@ 800 km) GRACE (@ 350 km) Topex/Poseidon (@1330 km) (Upward looking only) Jason 1 (@1330 km) (Upward looking only) C/NOFS & COSMIC constellation UV airglow data (135.6 nm radiance) LORAAS on ARGOS, GUVI on TIMED SSUSI/SSULI on DMSP TIP on COSMIC Other Data Types TEC from TOPEX/JASON Altimeters Ionosonde bottomside profiles DMSP in situ CHAMP in situ GRACE cross-links

42. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 42 Kalman Assimilation Runs for June 26, 2006 Three runs: GAIM Climate (no data) Ground GPS TEC (200 sites) Ground + COSMIC links (upward & occultation) Resolution:2.5 deg. Lat. 10 deg. Lon. 40 km Alt. No. of grid cells: 100,000 Sparse Kalman filter: Update & propagate covariance Truncate off-diagonal covariance that is beyond physical correlation lengths

43. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 43 GAIM Assimilation Using Ground and COSMIC Data

44. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 44 GAIM Validation Using Jason-2 Vertical TEC Ground-data only Ground and space data

45. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 45 GAIM vs. Abel HmF2 Comparison

46. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 46 GAIM Driven By Ground GPS Only versus JASON VTEC June – Nov. 2004: 137 days

47. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 47 COSMIC Demo 2007 COSMIC data: 120+ minutes latency CDAAC: COSMIC Data Analysis and Archiving Center at UCAR Start of orbit End of orbit; data downloaded Data received at CDAAC Limb TEC available Profiles (Abel) available GAIM 3-D global electron density grids 15-minute cadence Global ground network data: 5-minute and 1-hour latency

48. Workshop on the Future of Ionospheric Research for Satellite Navigation, Dec 4-15, 2006, ICTP, Trieste, Italy 48 What You Have Learned 1.Ionosphere is the largest error source in GPS positioning 2.Empirical models can be used to mitigate effects 3.Dual-frequency GPS data can be used to solve for the ionospheric effect 4.Error sources affecting GPS-based ionospheric estimation: arc length,leveling, biases, multipath, noise, etc. 5.Global Ionospheric Mapping techniques: single vs multi- shell approaches: ionospheric delay and biases estimation 6.Validation of maps, point plots, movies, etc.