StatusReport ofthe IGS ClockProducts - BIPM · Existing IGS Core Products SERIES Product Interval...

Status Report of the

IGS Clock Products

2015 September 17 – 18

Michael J. ColemanClock Products CoordinatorInternational GNSS ServiceU.S. Naval Research Laboratory

Washington, DC, USA

Consultative Committeeon Time and Frequency

Bureau International

des Poids et Mesures

Sevres, Cedex, France

Existing IGS Core Products

SERIES Product Interval AccuracyIssue

Latency

Time

Ultra Rapidgps orbits 15 min ∼ 3.0 cm Daily atgps clocks ∗ 15 min ∼ 150 ps 03:00 09:00 3 – 9

(igu)polar motion 6 hrs ∼ 40 µas 15:00 21:00 Hourslength of day 6 hrs ∼ 10 µs utc

Rapidgps orbits 15 min ∼ 2.5 cmgps clocks † 5 min ∼ 75 ps (rms) Daily at 17 – 41

(igr)polar motion 1 day ∼ 40 µas 17:00 utc Hourslength of day 1 day ∼ 10 µs

Finalgps orbits 15 min ∼ 2.5 cmglonass orbits 15 min ∼ 5.0 cm Weekly ongps clocks † 5 min ∼ 75 ps (rms) Wednesday 11 – 17

(igs)polar motion 1 day ∼ 30 µas or Thursday Dayslength of day 1 day ∼ 10 µs

* Only satellite clocks are reduced in the Ultra Rapid products.

† Both satellite and station clocks are reduced but not all IGS network station clocks.

M. Coleman (IGS // US NRL) IGS Clock Products 2015 September 17 – 18 2 / 12

Production Line

Timescale

Generation

Analysis Center

Coordinator

Kevin Choi, NGS

Analysis Centers

Data

Centers

IGNInstitut

GeographiqueNational

KAISKorean Astronomyand Space Science

Institute

CDDISCrustal DynamicsData Information

Systems

SIO

USNO WHU (Wuhan University)

GOP (Geodetic Observatory Pecny)

CODE Center for Orbit Determination in Europe

SIO (Scripps Institute of Oceanography)

JPL (Jet Propulsion Lab) ESOC

NRCan (Natural Resources Canada)

GFZ (GeoForschungsZentrum)MIT

NGS (National Geodetic Survey)

GRG (Groupe de Recherches en Geodesie Spatiale)


IGS Time

The current IGS timescale is formed for both the IGS rapid and final products and isderived using a standard Kalman Filter approach.

Kalman Filter was introduced in IGS2.0 version. This second version wasimplemented in 2011 with repreocessing computed from 2010.

Specifications for IGST

Consists of an average of 50 – 60 positively weighted member clocks.

Timescale is generated and exchanged with the IGS data clock reference for theRapid and Final products.

Stability performance as low as E-16 for longer averaging intervals.

Steers to UTC via AMC2 or USNO. Goal is towards UTC(USNO).

Timescale Features

Institutes a harmonic component into model for all satellite clocks.

Automated break detection responds to phase and frequency breaks in clockdata as well as day–by–day boundaries.

Steering achieved using Linear Quadratic Gaussian Steering Algorithm.


General IGS Clock Model

Filter uses a four state clock model for all station clocks. The total phase stateincorporates additional white noise process.

For satellite clocks, two fixed period harmonics are estimated and added to thetotal phase state. Hence, some clocks have 8 states.

∫

dt

White Noise

⊕

∫

dt

White Noise

⊕

∫

dt

White Noise

⊕

uf

r

up

f

uθ

pur

NoiseWhite

Drift Frequency Phase

θ

⊕

∼

∼

(a1 + ua1 ) cos(2πω1t) + (b1 + ub1) sin(2πω1t)

(a2 + ua2 ) cos(2πω2t) + (b2 + ub2) sin(2πω2t)

Periodic #1

Periodic #2


IGS(R)T Offset from UTC

52000 53000 54000 55000 56000 57000−50

0

50

Nan

osec

onds

UTC−GPST (Circular T) IGST−GPST IGRT−GPST

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

52000 53000 54000 55000 56000 57000−50

0

50

Nan

osec

onds

UTC−IGS(R)T*

UTC−IGST UTC−IGRT

52000 53000 54000 55000 56000 57000−50

0

50

MJD

Nan

osec

onds

IGRT − IGST

* Assumes that GPS Time via Circular T is equivalent to GPST

U.S

. Nav

al R

esea

rch

Labo

rato

ry 0

4−S

ep−

2015

, 17:

51


GPS Clock Frequency Weight Contribution

2010 2011 2012 2013 2014 2015 20160

20

40

60GPS Clock Contribution to IGS(R)T

Clock

Weight%

IGRIGS

2010 2011 2012 2013 2014 2015 20160

20

40

60

80Size of IGS(R)T Clock Membership

Number

ofClocks

GPS

IIF

Broadcast

Start


Ground Stations Reduced in Combinations

These tables show some common stations that are weighted in the timescale over thefirst 257 days of year 2015.

Stations with IGS & UTC (Partial List)

Station IGST IGRT

IGS UTC Location Days wp wf Days wp wf

BJNM NIM Beijing CHINA 59 4.31 4.30 0 -- --

IENG IT Torino ITALY 38 2.31 0.89 239 3.77 3.85

NIST NIST Boulder CO USA 0 -- -- 212 4.80 4.80

OPMT OP Paris FRANCE 223 4.37 4.38 1 4.86 4.96

PTBB PTB Braunschweig GERMANY 233 4.44 4.44 249 4.85 4.84

SFER ROA San Fernando SPAIN 225 4.36 2.21 6 5.16 3.53

SPT0 SP Boras SWEDEN 130 3.69 3.68 194 4.66 4.66

TWTF TL Chung-Li TAIWAN 119 4.32 4.31 113 5.07 5.07

USNO USNO Washington DC USA 73 3.58 1.09 58 4.37 4.36

WAB2 CH Bern SWITZERLAND 231 1.64 1.97 238 2.13 0.11


Case Observation: IENG Rapid Product Estimates

Data is mostly present for the rapid product.

U.S. Naval Research Laboratory, 01−Sep−2015 16:35:17

−1.5

−1

−0.5

0

0.5

1

1.5

2

phas

e (n

s)

0

2

4

6

ph. s

igm

a (n

s)

Sigmas

0 5 10 15 20 25 300

5

10

MJD − 57235

wφ (

%)

−6

−4

−2

0

2

4

6

8

freq

uenc

y (n

s/da

y)

−0.06

−0.04

−0.02

0

0.02

0.04

0.06

drift

(ns

/day

2 )

phase + WHPH + periodic(− 0.0061t2− 0.019t+ 19) phase (− 0.0061t

2− 0.019t+ 19) frequency (− 0.0051t

1+ 0.081) drift ( 0.058) 2 phase reset(s)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 1AUG2015

Filter States & SigmasIENG

01−Aug−2015 through 31−Aug−2015

0

5

10

fr. s

igm

a (n

s/da

y)

0

0.1

0.2

0.3

0.4

dr. s

igm

a (n

s/da

y2 )

0

0.5

1

wf (

%)

0

0.5

1

wd (

%)


Case Observation: IENG Final Product Estimates

Data is largely absent for the final product.

U.S. Naval Research Laboratory, 11−Sep−2015 09:40:44

−0.1

−0.05

0

0.05

0.1

0.15

0.2

0.25

phas

e (n

s)

0

5000

10000

15000

ph. s

igm

a (n

s)

Sigmas

0 5 10 15 20 250

1

2

MJD − 57235

wφ (

%)

−5

0

5

10

freq

uenc

y (n

s/da

y)

−0.04

−0.02

0

0.02

0.04

0.06

0.08

0.1

0.12

drift

(ns

/day

2 )

phase + WHPH + periodic(− 1t2+ 23t− 1.2e+02) phase (− 1t

2+ 23t− 1.2e+02) frequency (− 1.6t

1+ 18) drift ( 0.39)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30AUG2015

Filter States & SigmasIENG

01−Aug−2015 through 29−Aug−2015

0

50

100

150

fr. s

igm

a (n

s/da

y)

0

0.5

1

dr. s

igm

a (n

s/da

y2 )

0

0.5

1

wf (

%)

0

0.5

1

wd (

%)


Proposed Improvements

UTC(k) Laboratory Inclusion

UTC labs occasionally fall from IGS(R)T due to Analysis Center computationchanges and/or drop outs. This is especially more prominent in IGR.

Request that Analysis Centers utilize some collection of UTC(k) labs to allowimproved IGS(R)T stability and UTC steering (as best possible / if possible).Request at next Governing Board Meeting in [ December 2015 @ AGU ]

Goal via steering to these references is to maintain

∣

∣utc – igs(r)t∣

∣ < 10 ns

Satellite Clock Models

Improvements to the initialization of clock states and covariances are beingtimescale filter. This will assist with false starts for GPS clocks and/or resets.

Increase the contribution of GNSS clocks in IGS(R)T; in particular, includefrom Galileo, Glonass and Beidou as data becomes available after MGEXproject complete. [ Longer term timeline ]

GPS (in future, GNSS) clocks are nearly always part of the IGS combinations.


IGS Information / Contact

For more information on the IGS core products, experiments and working groups:

www.igs.org

For suggestions or requests for the Clock Products:

[email protected]


StatusReport ofthe IGS ClockProducts - BIPM · Existing IGS Core Products SERIES Product Interval...

Documents

Transcript of StatusReport ofthe IGS ClockProducts - BIPM · Existing IGS Core Products SERIES Product Interval...