GRACE-FO processing at IfE/LUHIgor Koch, Mathias Duwe, Jakob Flury and Akbar Shabanloui Insitut für Erdmessung, Leibniz Universität Hannover/Germany

COST-G Team Meeting 2021 January 11-15, 2021

Online

Gravity field recovery at IfE

▪ GRACE-SIGMA: software package for gravity field recovery from GRACE and GRACE-FO sensor data

▪ Generalized dynamic orbit determination (variational equations approach)

▪ Initially all-MATLAB software [Naeimi et al., 2018]

▪ Several parts of the software converted to C/C++

▪ Mostly vectorized and parallelized

▪ Computation time one solution: app. 3.5 h (18 cores, 128 GB)

2

Overview

3

Orbit pre- adjustment

Gravity recovery- importing data - rotations - forces

Alt. acc. data

0. 1. 2.

iterative No iterations

Kin. orbits

Level 1B

Force modeling

4

Effect Model ReferenceGravity field GOCO06s (d/o: 300) Kvas et al., 2020

Third bodies Moon, Sun, Planets, Ephemerides: DE431 Folkner et al., 2014

Solid Earth tides IERS Conventions 2010 Petit a. Luzum, 2010

Ocean tides FES2014b Carrere et al., 2015

Solid Earth pole tides IERS Conventions 2010 Petit a. Luzum, 2010

Ocean pole tides IERS Conventions 2010 Petit a. Luzum, 2010

Relativistic IERS Conventions 2010 Petit a. Luzum, 2010

Non-tidal AOD1B RL06 Dobslaw et al., 2017

Atmospheric tides AOD1B RL06 Dobslaw et al., 2017

Non-gravitational ACT1B (GRACE C) TU Graz accelerometer data (GRACE D)

JPL, 2018 Behzadpour, 2020

Observations

• K-band range rates (5 s)

• LRI range rates in following releases

• Kinematic AIUB orbits as pseudo-observations (30 s)

5

Orbit screening

• Kinematic positions are compared to GNV1B orbit

• Epochs with 3D-difference >8 cm not considered in estimation

6

0 100 200 300 400 500 600 700 800 900 1000

epoch #

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

0.11

3D

dif.

[m

]

Weights

• KBRR:

2E-07 m/s (fixed)

• Kinematic positions:

diagonal elements of the orbit

covariance matrix (down-weighted by a factor 5)

𝜎KBRR =

𝜎Orbit =

7

GPS down-weighting

8

0 10 20 30 40 50 60 70 80 90 100

degree [-]

10-11

10-10

10-9

10-8

err

or

degre

e s

tandard

devi

atio

n [-]

202007

LUH down-weight

JPL

LUH no down-weight

Parameters

• Per 3h-arc:

• Initial state

• Accelerometer biases

• Empirical parameters [Kim, 2000]

• Per monthly solution:

• Full scale matrix [Klinger a. Mayer-Gürr, 2016]

• Gravity potential (d/o: 96)

9

Empiricals

• Bias A: 90 min

• Bias rate B: 90 min

• Periodic bias E,F: 3 h

• Periodic bias rate G,H: 3 h

10

⇢̇o � ⇢̇c = A+Bt+ (E + Ft) cosu+ (G+Ht) sinu<latexit sha1_base64="8736TZEVsxMVUt0wjXnIdkXCErE=">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</latexit>

0 10 20 30 40 50 60 70 80 90 100

degree [-]

10-11

10-10

10-9

10-8

err

or

de

gre

e s

tan

da

rd d

evi

atio

n [

-]

no emp.

with emp.

Screening / Outlier detection

11

1 2Main adjustment

Pre- adjustment

O-C range rates Post fit range rate residuals

Screening

12

1

Arcs will not be considered in main adjustment

0 50 100 150 200 250

arc number [-]

0

1

2

3

4

5

6

o-c

range r

ate

rm

s [m

/s]

10-6

Screening / Outlier detection

13

1 2Main adjustment

Pre- adjustment

O-C range rates Post fit range rate residuals

Screening

14

2

- Remove suspicious epochs - Recompute solution

Screening

15

2

- Remove suspicious epochs - Recompute solution

0 10 20 30 40 50 60 70 80 90 100

degree [-]

10-11

10-10

10-9

10-8

err

or

de

gre

e s

tan

da

rd d

evi

atio

n [

-]

before screening

after screening

Evaluation

• Spectral noise

• Spatial noise

• Signal content

• Coefficients C20 C30

16

Spectral noise

17

0 10 20 30 40 50 60 70 80 90 100

degree [-]

10-11

10-10

10-9

10-8

err

or

de

gre

e s

tan

da

rd d

evi

atio

n [

-]

CSR

GFZ

JPL

ITSG

AIUB

LUH

- w.r.t. mean solution of all centers - C20 zero tide - 2018/06 - 2020/08

Noise over the oceans

18

2018-07 2019-01 2019-07 2020-01 2020-07

date

0.5

1

1.5

2

2.5

3

EW

H R

MS

[cm

]

CSRGFZJPLITSGAIUBLUH

- 2018/06 - 2020/08 - EWH w.r.t. mean solution of all centers - C20 and C30 replaced with SLR - Gaussian filter (400 km) - Climatology model subtracted - (bias, trend, annual, semi-annual signal)

River basin amplitudes

19

20 40 60 80 100 120 140 160 180

basin number

-0.05

0

0.05

0.1

0.15

0.2

0.25

am

plit

ude [m

]

- 2018/06 - 2020/08 - EWH w.r.t. mean solution of all centers - C20 and C30 replaced with SLR - Gaussian filter (400 km) - Fit of climatology model - (bias, trend, annual, semi-annual signal)

0 10 20 30 40 50 60 70 80 90 100

degree [-]

10-11

10-10

10-9

10-8

err

or

de

gre

e s

tan

da

rd d

evi

atio

n [

-]

CSR

GFZ

JPL

ITSG

AIUB

LUH

Greenland mass trends

20

NW CW SW SE NE NO

drainage basin

-45

-40

-35

-30

-25

-20

-15

tre

nd

[G

t/yr

]

NW

SWSE

CW

NE

NO

- 2018/06 - 2020/08 - EWH w.r.t. mean solution of all centers - C20 and C30 replaced with SLR - Gaussian filter (400 km) - Fit of climatology model - (bias, trend, annual, semi-annual signal) - GIA, degree 1, .. not considered

0 10 20 30 40 50 60 70 80 90 100

degree [-]

10-11

10-10

10-9

10-8

err

or

de

gre

e s

tan

da

rd d

evi

atio

n [

-]

CSR

GFZ

JPL

ITSG

AIUB

LUH

Coefficients C20, C30

21

LUH-GRACE-FO-2020 solutions

▪ Solutions are computed operationally

▪ Available a few days after all needed GRACE-FO products are online (L1B, kinematic orbits, alt. acc)

▪ Solutions can be found on ICGEM and LUH data repo: ▪ http://icgem.gfz-potsdam.de/series/03_GRACE_other/LUH/LUH-GRACE-FO-2020

▪ https://data.uni-hannover.de/dataset/luh-grace-fo-2020

22

References

▪ Behzadpour et al. (2020): GRACE-FO accelerometer data recovery within ITSG-Grace2018 data processing, EGU General Assembly 2020.

▪ Carrere et al. (2015): FES 2014, a new tidal model on the global ocean with enhanced accuracy in shallow seas and in the Arctic region, EGU General Assembly 2015.

▪ Dobslaw et al. (2017): A new high-resolution model of non-tidal atmosphere and ocean mass variability for de-aliasing of satellite gravity observations: AOD1B RL06, Geophysical Journal International, Volume 211, Issue 1, Pages 263—269.

▪ Folkner et al. (2014): The Planetary and Lunar Ephemerides DE430 and DE431, IPN Progress Report 42-196.

▪ Kim (2000): Simulation Study of A Low-Low Satellite-to-Satellite Tracking Mission, Dissertation, University of Texas at Austin, 2000.

▪ Klinger and Mayer-Gürr (2016): The role of accelerometer data calibration within GRACE gravity field recovery: Results from ITSG-Grace2016, Advances in Space Research, 58, 1597.

▪ Kvas et al. (2020): GOCO06s – A satellite-only global gravity field model, Earth System Science Data Discussions.

▪ Naeimi et al. (2018): IfE monthly gravity field solutions using the variational equations, EGU General Assembly 2018, Vienna.

▪ Naeimi (2018): A modified Gauss-Jackson method for the numerical integration of the variational equations, EGU General Assembly 2018, Vienna.

▪ Petit and Luzum (2010): IERS Conventions (2010), IERS technical note 36, Verlag des Bundesamts für Kartographie und Geodäsie, Frankfurt am Main.

23