Homogeneous tropospheric path delays from GNSS re-processing by GOP

Transcription

1 Homogeneous tropospheric path delays from GNSS re-processing by GOP Jan Douša, Pavel Václavovic, Michal Eliaš, RIGTC, The Czech Republic June 6-8, 2016, Budapest, Hungary

2 Outline 2 Introduction GNSS meteorology GNSS re-analysis GNSS climatology Conclusions

3 GNSS model for troposphere 3 GNSS model: (carrier-phase observations expressed in range units) L c c N sat sat sat sat rec rec rec rec ion tro Tropo model: symmetric part asymmetric part sin A tro STD mf H ZHD mfw ZWD mfg G N cos A GE

4 Troposphere path delays in GNSS analysis 4 Tropospheric delay in GNSS is often modelled using zenith total delay (ZTD) parameters mapped into an actual slant delay with mapping functions D ( z) mf z ZHD mf z ZWD... T H W ZTD ZHD ZWD separation of zenith total delay was introduced by Davis (1985) hydrostatic component non-hydrostatic ( wet) component Z k H 6 1 d D 10 P 0 gr m k ZWD 10 k dh 6 ' 3 e 2 T m T h0 0 by using T m - mean temperature Mean ZHD value : 2.3 m in zenith ~ 90% of the total path delay Mean ZWD value : 0-50 cm in zenith ~ 10% of the total path delay T m h h0 e dh T e 2 dh T

5 GNSS meteorology 5 Developing since 1997: MAGIC, COST 716, TOUGH, E-GVAP I-III The EIG EUMETNET GNSS Water Vapour Programme, E-GVAP Phases I-III ( ) Project coordinating the near real-time delivery of data from ~2400 GPS sites delivering > 14M ZTDs pcm Focus is on GPS-only hourly processing, delivering only ZTD in 90mins. Operational assimilation at a few European National Met Services, many others under testing. Use of E-GVAP ZTDs has proven positive impact on NWP forecast skill. Surface T and P used for conversion to Integrated precipitable Water Vapour (IWV). Active Quality Control (AQC) in place. MoUs in place with EUREF and EUPOS.

6 COST ES GNSS4SWEC 6 COST Action ES1206: GNSS for Severe Weather and Climate (GNSS4SWEC) Chair: Dr Jonathan Jones, MetOffice Co-chair: Dr Guergana Guerova, Uni Sofia WG1 Chair: Dr Jan Dousa, GOP Co-chair: Dr Galina Dick, GFZ WG2 Chair: Dr Siebren de Haan, KNMI Co-chair: Dr Eric Pottiaux, ROB WG3 Chair: Dr Olivier Bock, IGN Co-chair: Dr Rosa Pacione, ASI

7 WG1 - Benchmark 7 Preparation phase: design & data collection May-June floods of Danube/Moldau/Elbe rivers GNSS: ~500 stations (AT, CZ, DE, PL) SYNOP: ~200 stations (AT, CZ, DE, PL) NWM: regional (Aladin-CZ), global (ERA-Interim, NCEP GFS) RAOBS: E-GVAP + two high-resolution (CZ) WVR: Potsdam, Lindenberg (DE) RADAR images: Brdy, Skalka (CZ) Reference products GNSS: Bernese (GOP), EPOS (GFZ) NWM: G-Nut/Shu (GOP), DNS (GFZ) User phase contributions, evaluations feedbacks, interpretations Douša J, Dick G, Kačmařík M, Brožková R, Zus F, Brenot H, Stoycheva A, Möller G, Kaplon: Benchmark campaign and case study episode in Central Europe for development and assessment of advanced GNSS tropospheric models and products, Atmosph Meas Tech, Online Discussion, 2016

8 NWM assessment within Benchmark 8 GNSS & NWM inter-comparisons (ZTD - zenith total delay) May 1 June 30, GNSS4SWEC Benchmark Aladin-CZ - GOP ERA-Interim - GOP GFS - GOP

9 Tropospheric gradients in dense network May 31, 2013[18UTC] of GNSS4SWEC Benchmark 9 GNSS GOP GNSS GFZ ERA Interim NCEP GFS

10 Benchmark GNSS and NWM assessment GNSS reference products Bernese + DD (GOP) ZTDs (1h), GRD(6h) EPOS-8 + PPP (GFZ) ZTDs (15min), GRD(1h) Zenith total delays (ZTD) NWM source (software) Grid resolution Analysis [hour] Forecast [hour] GNSS source (software) Pairs # Excl # Bias [mm] Sdev [mm] 10 NWM-derived parameters G-Nut/Shu (GOP): ZWD + ZHD + T/Tm + vert. model DNS (GFZ): ZWD + ZHD + GRAD + MF RMS [mm] ERA (Shu) 1 deg 6 0 GOP (Bernese) ERA (Shu) 1 deg 6 0 GFZ (EPOS-8) ERA (DNS) 1 deg 6 0 GOP (Bernese) ERA (DNS) 1 deg 6 0 GFZ (EPOS-8) GFS (DNS) 1 deg 6 3 GOP (Bernese) GFS (DNS) 1 deg 6 3 GFZ (EPOS-8) ALADIN (Shu) 4.7 km 6 0,1,2,3,4,5 GOP (Bernese) ALADIN (Shu) 4.7 km 6 0,1,2,3,4,5 GFZ (EPOS-8) Tropospheric horizontal gradients North & East NWM GNSS NS gradients EW gradients source source Pairs (excl) Bias [mm] Sdev [mm] RMS [mm] Pair (excl) Bias [mm] Sdev [mm] RMS [mm] ERA GOP 224 (4) (3) ERA GFZ 224 (3) (3) GFS GOP 224 (5) (4) GFS GFZ 224 (3) (4)

11 Tropospheric gradients in dense network 11 Prepare state-of-the-art homogeneous tropospheric product(s) for climate EUREF Repro2: (2014) 3 ACs process full EUREF Permanent Network (EPN) ASI /E-GEOS (Gipsy/Oasis) GOP Geodetic Observatory Pecný (Bernese V52) MUT Military University of Technology (GAMIT) 2 ACs contributes with EPN subnetworks LPT Swisstopo (Bernese V52) IGN Instituto Geografico National (Bernese V52) EUREF combined solution ASI/E-GEOS EUREF Tropospheric Product Coordinator Others IGS Repro2: waiting for IGS final PPP, available troposphere from CODE, GFZ (TIGA) GRUAN: focus on few stations instrument collocation activities of GFZ & ASI

12 GOP Repro2 solutions ( ) 12 GOP0 GMF, 3 deg cut-off, ZTD(1h) + GRD(6h) GOP1 VMF1, 3 deg cut-off, ZTD(1h) + GRD(6h) GOP2 VMF1, 7 deg cut-off, ZTD(1h) + GRD(6h) GOP3 VMF1, 10 deg cut-off, ZTD(1h) + GRD(6h) GOP4 VMF1, 3 deg cut-off, ZTD(1h) + GRD(6h) + ATL GOP_Repro1, NMF, 3 deg cut-off, ZTD(1h), no GRD! (I05 and I08 ATX/Reference Frame) Coordinate statistics: Solution North [mm] East [mm] Up [mm] GOP-Repro1/I GOP-Repro1/I GOP GOP GOP GOP GOP

13 GOP reprocessing - raw coordinate time-series 13 plots of daily-based independent solutions expressed in a single reference frame! antenna changes ice/snow problem? poor data qual. non-european litosphere desk post-seismic movement post-glacial uplift 13

14 Multi-year combination results 14 estimation of stations velocity and coordinate discontinuities Final horizontal repeatability 1-2 mm Final vertical repeatability 4-6 mm 14

15 GOP Repro2 solutions ( ) 15 GOP0 GMF, 3 deg cut-off, ZTD(1h) + GRD(6h) legacy solution GOP1 VMF1, 3 deg cut-off, ZTD(1h) + GRD(6h) new solution GOP2 VMF1, 7 deg cut-off, ZTD(1h) + GRD(6h) special variant GOP3 VMF1, 10 deg cut-off, ZTD(1h) + GRD(6h) special variant GOP4 VMF1, 3 deg cut-off, ZTD(1h) + GRD(6h) + ATL new solution New: daily/weekly midnight homogenization: ZTD CRD days

16 GRD: comparison to ERA-Interim (statistics) 16 Parameters: ZTD + N/E-gradients Software: using GFZ s ray-tracing software (Zus et al, 2012) NWM: global 1 1deg ECMWF s ERA-Interim reanalysis (Dee et al. 2011) Period: , initially pre-selected 30 EPN stations Statistics: mean over all selected stations Solution ZTD [mm] Pairs # Excl # GT0 (Rep1) ± GOP ± GOP ± GOP ± GOP ± GOP ±

17 ZTD(GNSS Repro1/Repro2): GOP x EUREF 17

18 18 GRD: comparison to ERA-Interim (time-series)

19 19 GOP-TropDB - background motivation Tropospheric parameters from space geodetic techniques are side products along with estimating geodetic parameters of main interest However, various approximations still needed for tropospheric modeling: Separating hydrostatic/wet parts: ZTD = ZHD + ZWD Mapping to the zenith: STD (ele) = mf h ZHD + mf w ZWD Asymmetry modeling: STD (ele,azi) = STD(ele) + mf g (G N cosa + G E sina) Motivations complementary benefits for: Geodesy: assessments of tropospheric approximations as well as other models Geodesy: exploitation of external tropospheric parameters for positioning Meteorology: assimilation of ZTDs/STDs/gradients into numerical weather models Climatology: monitoring of long-term trends in integrated water vapor Goals: Intra-technique comparisons (GNSS or other space geodetic techniques) Inter-technique comparisons (independent evaluations) Additional functionalities QC, conversions, TS analyses, archive,

20 GOP-TropDB - structure 20 3 rd generation system, driven by PostgreSQL (9.3/9.4) User and public data sets User and public outputs User and public functions User and public operations

21 GOP-TropDB interactive web service 22

22 GNSS climatology 23 From GNSS observations to IWV, trends & variations L c c N sat sat sat sat rec rec rec rec ion Δ tro ZTD Noise for Geodesy Signal for Meteorology ZWD ZTD ZHD 0.16 m ZHD = f(p s ) 2300*P s [mm] IWV T ZWD ZWD T m = f(t s ) or NWM integrated

23 observation processing post-processing 24 WG3: Standards and Methods for Climate rinex data ZTD params sitelog metadata PROCESSING ZTD formal errors auxiliary log info data (orbit, clock, ERP) models (loading, antenna) RMS, AMB Settings (cutoff, tropo model) Standards IGS (observations) IERS (for geodesy) WG3 X (for climate) SCREENING Data (temp model) Methods (tests) X IWV CONVERSION HOMOGENISATION POST-PROCESSING IWV IWV params formal errors log info Data (Ps, Tm, Ki) Methods (interpolation) Data (reference) Methods (statistic) Resampling Filtering, averaging Courtesy of O.Bock X X X (depend on application)

24 IWV trend and variations 25 Require data cleaning, homogenization and temporal modeling

25 Conclusions 26 GOP-Repro2 finished (July 2015) Period: , progressively increase # of stations Cover: EUREF Permanent network (total >250 European stations) Results: outperformed any previous Repro1, e.g. GOP-Repro1/I08 5 variants - several strategies applied, assessed, best approach selected Comparisons with ECMWF s ERA-Interim re-analysis Best agreement GO4-3 deg, VMF1 and atmospheric non-tidal loading Accuracy: 8mm (ZTD) and 0.5mm (ZTD horizontal gradients) GOP-TropDB automated comparisons & interactive web service Special focus to support GNSS4SWEC/WG3 and climate study Combination of parameters (ZTD+GRD, COORD) across midnights/weeks Significant effort to remove problematic stations from the solution Tropospheric parameters at epochs HR:30 and HR:00 (interpolated) Ongoing development towards product cleaning and homogenization

26 GOP three targets 27 GOP Reprocessing EUREF Permanent network (EPN) Software development Solution variants GOP-TropDB Tropospheric product intra-/inter-technique comparison Data conversions (format, ZTD IWV) Data dissemination Software development Product preparation Time-series analysis Data cleaning & flagging Data homogenization Software development