Global Navigation Satellite System data

processing at near real time

Martin Imrisek

Slovak University of Technology in BratislavaDepartment of Theoretical Geodesy

&

Slovak Hydrometeorological Institute

martin.imrisek@stuba.sk

September 19, 2017

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 1 / 20

Introduction

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 2 / 20

Data processing

GNSS data processing may be divided from latency point of view on:

Final processing,

Near real time processing,

Real time processing.

GNSS data processing may be divided also from processing strategypoint of view on:

Precise network positioning,

Precise point positioning.

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 3 / 20

Routine processing

GPS and GLONASS,

network processing at near real time,

MAX–OBS baseline strategy,

32 min latency,

59 stations are being processed every hour,

40 stations are from Euref network,

6 stations are from IGS network,

13 stations are from Austrian, Czech, Hungarian and Slovaknational networks of permanent GNSS stations.

http://space.vm.stuba.sk/pwvgraph/

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 4 / 20

Routine processing

10.0 ° E 12.5 ° E 15.0 ° E 17.5 ° E 20.0 ° E 22.5 ° E 25.0 ° E 27.5 ° E 30.0 ° E

42.5 ° N

45.0 ° N

47.5 ° N

50.0 ° N

52.5 ° N

BAIA

BASV

BBYS

BISK

BUCU

BUTE

CFRM

CAKO

CLIB

CPARCRAK

DEVA

DOPL

GANP

GOPE

GRAZ

GSR1

HUVO

KAME

KATO

KOLSKOSE

KRA1

KUZALIE1

LINZ MOP2

OROS

PADO

PENC

PEMB

PORE

POUS

POZE

PTBB

RISO

SKLM

SKNR

SKSE

SKSKSKSL

SKTN

SPRN

SULP

SUT1

TELG

TUBO

TUWI

USDL

UZHL

VACO

VELS

VEN1

WROC

WTZR

ZOUF

ZYWI

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 5 / 20

Result comparison

IGS stations

Final Precise point positioning

Bucuresti, Romania – BUCU

Ganovce, Slovakia – GANP

Graz, Austria – GRAZ

Uzhgorod, Ukraine – UZHL

Wettzell, Germany – WTZR

Zimmerwald, Switzerland – ZIMM

Differences are shifted about 2150 mm.

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 6 / 20

Result comparison

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 7 / 20

Result comparison

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 8 / 20

Result comparison

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 9 / 20

Result comparison

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 10 / 20

Result comparison

Table: Statistic of differences

BUCU GANP GRAZCorrelation +0.988 +0.988 +0.989Minimum [mm] -38.0 -33.8 -30.1Maximum [mm] 18.6 18.5 23.1Average [mm] -2.13 -1.29 -1.26Standard deviation [mm] 6.08 6.22 6.25

UZHL WTZR ZIMMCorrelation +0.990 +0.989 +0.985Minimum [mm] -18.7 -25.0 -23.0Maximum [mm] 19.7 16.6 22.5Average [mm] -0.36 -0.59 0.14Standard deviation [mm] 5.82 5.92 5.78

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 11 / 20

ZTD data assimilation

AROME – Hungarian domain,

grid point resolution 2.5 km and 60 pressure levels,

GPS data assimilation is respecting ASSIMILATION OF GPSMOISTURE INFORMATION – Mohamed Zied SASSI,

32 white listed stations (more incoming),

static bias correction,

only one GNSS data assimilation cycle at 12UTC 18th August2017 was performed,

NOGNSS and +GNSS 36 hours forecast.

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 12 / 20

ZTD data assimilation – technical check

Figure: Differences in specific humidity at 50th pressure level betweenguess and analysis at 12UTC 18th August 2017.

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 13 / 20

ZTD data assimilation – technical check

Figure: Differences in relative humidity between guess and analysis atSKLM and GANP permanent stations at 12UTC 18th August 2017.

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 14 / 20

ZTD data assimilation – technical check

Figure: Differences in temperature between guess and analysis at SKLMand GANP permanent stations at 12UTC 18th August 2017.

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 15 / 20

ZTD data assimilation – technical check

Figure: Differences in specific humidity between guess and analysis atSKLM and GANP permanent stations at 12UTC 18th August 2017.

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 16 / 20

ZTD data assimilation – technical check

Figure: 24h accumulated rainfall from 00UTC 19th August 2017.NOGNSS (top left), +GNSS (top right), differences (bottom left) andINCA (bottom right).

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 17 / 20

Conclusions

Local GNSS processing designed for SHMI purposes,

obtained ZTD are comparable with IGS final PPP product,

first 3DVAR data assimilation studies on SHMI,

result is not convincing from one assimilation cycle.

Future perspectives :

multiple cycle data assimilation case study,

add more data types to assimilation.

Big Thanks to Alena Trojakova and Mate Mile for support.I would be grateful for any feedback or advice.

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 18 / 20

Thank you for your attention.

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 19 / 20

New generation of permanent GNSS stations

Martin Imrisek (SUT & SHMI) GNSS data processing September 19, 2017 20 / 20