Total electron content monitoring using triple frequency GNSS data: A three-step approach

Total electron content monitoring using triple frequency GNSS data: A three-step approach J.Spits, R.Warnant Royal Meteorological Institute of Belgium Fifth European Space Weather Week @ Brussels November 17 21 2008 1 / 31

Overview Fifth European Space Weather Week @ Brussels November 17 21 2008 2 / 31

Context Objective Civil frequencies Fifth European Space Weather Week @ Brussels November 17 21 2008 3 / 31

Context Context Objective Civil frequencies Global Navigation Satellite Systems, i.e. GPS and Galileo, emit radio frequencies which can be combined to compute the variability of the ionospheric plasma, and in particular the Total Electron Content (TEC) of the ionosphere Fifth European Space Weather Week @ Brussels November 17 21 2008 4 / 31

Objective Context Objective Civil frequencies Development of a new TEC monitoring technique: using triple frequency measurements developed on simulated data and tested on real data improved in regards with the usual dual frequency technique Fifth European Space Weather Week @ Brussels November 17 21 2008 5 / 31

Civil frequencies Context Objective Civil frequencies GNSS system Carrier signal Frequency (MHz) GPS L1 1575.42 L2 1227.60 L5 1176.45 Galileo L1 1575.42 E5b 1207.14 E5a 1176.45 Fifth European Space Weather Week @ Brussels November 17 21 2008 6 / 31

Simulated data Real data Fifth European Space Weather Week @ Brussels November 17 21 2008 7 / 31

Simulated data Simulated data Real data realistic GNSS measurements GPS and Galileo triple frequency code P i p,k and phase Φi p,k (with k = L1, L2 or L5) including all error sources (ionosphere,troposphere...) validated with real double frequency GPS data used to develop the monitoring technique (before the availability of real triple frequency data) Fifth European Space Weather Week @ Brussels November 17 21 2008 8 / 31

Real data Simulated data Real data triple frequency (L1-E5a-E5b) code and phase measurements from Giove-A satellite for 3 stations (GNOR,GIEN,GKOU) and 7 days (DOY 013-020/2008) used to test and validate the monitoring technique Fifth European Space Weather Week @ Brussels November 17 21 2008 9 / 31

Basic ideas Step 1 Step 2 Step 3 Fifth European Space Weather Week @ Brussels November 17 21 2008 10 / 31

Basic ideas Basic ideas Step 1 Step 2 Step 3 triple frequency technique divided in three steps: 1. extra-widelane (EWL) ambiguities 2. widelane (WL) ambiguities 3. geometric-free system (GF) Fifth European Space Weather Week @ Brussels November 17 21 2008 11 / 31

Basic ideas Basic ideas Step 1 Step 2 Step 3 advantage: the availability of triple frequency measurements allow to form two dual frequency GF phase combinations code measurements are only used in step 1 precision of TEC is not affected by code (hardware and multipath) delays and is therefore improved N.B.: in the dual frequency technique, the use of the GF code combination is required for the ambiguity resolution process precision = 2-3 TECu Fifth European Space Weather Week @ Brussels November 17 21 2008 12 / 31

Step 1 Basic ideas Step 1 Step 2 Step 3 Objective: resolve the EWL ambiguities N 25 which are integer numbers : compute the extra widelane narrowlane (EWLNL) combination C 25 C 25 =Φ i p,l2 Φi p,l5 f L2 f L5 f L2 +f L5 ( f L2 c P i p,l2 + f L5 c P i p,l5)=n 25 + R 25 uses L2 and L5 code and phase measurements Issue: residual terms R 25 including code and phase hardware delays, multipath and noise R 25 < 0.5 cycle? Fifth European Space Weather Week @ Brussels November 17 21 2008 13 / 31

Step 2 Basic ideas Step 1 Step 2 Step 3 Objective: resolve the WL ambiguities N 12 which are integer numbers : compute the differenced widelane (DWL) combination C 125 C 125 =(Φ i p,l1 Φi p,l2) (Φ i p,l2 Φi p,l5 +N 25) λ 25 λ 12 =N 12 + R 125 uses L1, L2 and L5 phase measurements + EWL ambiguities N.B.: not possible with the WLNL combination (see results) Issue: residual terms R 125 causing an error of several cycles on N 12 N 12 = approximated integer values Fifth European Space Weather Week @ Brussels November 17 21 2008 14 / 31

Step 3 Basic ideas Step 1 Step 2 Step 3 Objective: use the results of the first two steps in order to achieve the monitoring of the TEC : resolve the system of two dual frequency Geometric Free phase combinations Φ i p,gf and Φ i (GF system): p,gf { Φ i p,gf = a 12 TEC N i p,l1 +c 12 N i p,l2 Φ i p,gf = a 25 TEC N i p,l2 +c 25 N i p,l5 Fifth European Space Weather Week @ Brussels November 17 21 2008 15 / 31

Step 3 Basic ideas Step 1 Step 2 Step 3 GF system 4 unknowns : TEC and ambiguities N 1,N 2,N 5 we introduce N 25 and N 12 so that there remain 2 unknowns: TEC and N 2 Issue: approximated N 12 so approximated TEC how is it possible to resolve N 12? Fifth European Space Weather Week @ Brussels November 17 21 2008 16 / 31

Step 3 Basic ideas Step 1 Step 2 Step 3 Resolution N 12 = integer values 1 cycle N 12 12 TECu (in GF system) 1. TEC e = rough estimation by dual frequency method 2. TEC e N 12 3. N 12 is precise enough to correct N 12 4. N 12 again in GF system TEC and ambiguities N 1,N 2,N 5 Fifth European Space Weather Week @ Brussels November 17 21 2008 17 / 31

Step 3 Basic ideas Step 1 Step 2 Step 3 Precision of TEC use of GF phase combinations not affected by code (harwdare and multipath) delays but by the corresponding phase delays Fifth European Space Weather Week @ Brussels November 17 21 2008 18 / 31

Simulated data Real data Fifth European Space Weather Week @ Brussels November 17 21 2008 19 / 31

Simulated data STEP 1 - Influence of code multipath and noise on N 25 for GPS and Galileo 0.5 0.5 Simulated data Real data R 25 (cycles) 0.25 0-0.25 R 25 (cycles) 0.25 0-0.25-0.5 286000 293200 300400 307600 Time (s) -0.5 271000 278200 285400 292600 Time (s) 0.5 0.5 0.25 0.25 R 25 (cycles) 0 R 25 (cycles) 0-0.25-0.25-0.5 25200 32400 39600 46800 Time (s) -0.5 7200 14400 21600 28800 Time (s) Fifth European Space Weather Week @ Brussels November 17 21 2008 20 / 31

Simulated data Simulated data Real data Influence 0.25/0.1 cycle resp. for GPS and Galileo Why such a difference? amplitude of code multipath and noise lower for Galileo wavelength of EWLNL combination : 5.861/9.768 m resp. for GPS and Galileo Resolution possible? estimation of the influence of code and phase hardware delays (constant): should not exceed the remaining margin (0.25/0.4 cycle) yes, and easier for Galileo results : we obtain the inital simulated N 25 values Fifth European Space Weather Week @ Brussels November 17 21 2008 21 / 31

Simulated data STEP 2 - Influence of code multipath and noise on N 12 for GPS and Galileo 4 4 Simulated data Real data R 12 (cycles) 3 2 1 0-1 -2-3 R 12 (cycles) 3 2 1 0-1 -2-3 -4 286000 293200 300400 307600 Time (s) -4 271000 278200 285400 292600 Time (s) 4 4 3 3 2 2 R 12 (cycles) 1 0-1 R 12 (cycles) 1 0-1 -2-2 -3-3 -4 25200 32400 39600 46800 Time (s) -4 7200 14400 21600 28800 Time (s) Fifth European Space Weather Week @ Brussels November 17 21 2008 22 / 31

Simulated data Simulated data Real data Influence > 0.5 cycle for GPS and Galileo Why? wavelength of WLNL combination : 0.862/0.814 m resp. for GPS and Galileo Solution? use of a different combination (DWL) results : we obtain the correct N 12 integer values + 2-3 cycles Fifth European Space Weather Week @ Brussels November 17 21 2008 23 / 31

Simulated data Simulated data Real data STEP 3 N 25 and N 12 values in GF system dual frequency estimation of TEC results : we obtain the initial simulated TEC values whatever the ionospheric model used (constant or Klobuchar) whatever the ionization level precision : tests with different mm-level amplitude and different sign values of phase delays influence = several tenth of TECu improvement Fifth European Space Weather Week @ Brussels November 17 21 2008 24 / 31

Real data -18387965.500 STEP 1 - EWLNL combination -18568682.500 Simulated data Real data EWLNL (cycles) -18387965.750-18387966.000-18387966.250 EWLNL (cycles) -18568682.750-18568683.000-18568683.250-18387966.500 0 3600 7200 Time (s) -18568683.500 360000 363600 367200 Time (s) -18359713.500-18606013.500-18359713.750-18606013.750 EWLNL (cycles) -18359714.000 EWLNL (cycles) -18606014.000-18359714.250-18606014.250-18359714.500 7200 10800 14400 Time (s) -18606014.500 0 3600 7200 Time (s) Fifth European Space Weather Week @ Brussels November 17 21 2008 25 / 31

Real data Simulated data Real data Variability? low : < 0.1 cycle in total agreement with simulated results Resolution possible? despite the bias caused by hardware delays? we need to process the further steps and validate the results... Fifth European Space Weather Week @ Brussels November 17 21 2008 26 / 31

Fifth European Space Weather Week @ Brussels November 17 21 2008 27 / 31

Development of a new TEC monitoring technique: using triple frequency code and phase GNSS measurements based on the resolution of integer ambiguities (EWL-WL) using two GF phase combinations precision of TEC improved allowing new opportunities for ionospheric studies Fifth European Space Weather Week @ Brussels November 17 21 2008 28 / 31

Fifth European Space Weather Week @ Brussels November 17 21 2008 29 / 31

Further validation of the TEC monitoring technique with real data: compute a rough estimation of TEC by using the usual dual frequency technique (in progress) use it in the GF system validate the results with external data (ex. IONEX) Fifth European Space Weather Week @ Brussels November 17 21 2008 30 / 31