GRACE-FO LRI1B SYSU Dataset
The GRACE (Gravity Recovery and Climate Experiment) and its successor, GRACE-FO (GRACE Follow-On), provide crucial data for global static and time-variable gravity field recovery[1,2]. In particular, the Laser Ranging Interferometer(LRI) instrument on GRACE-FO provide the first laser interferometric range measurements between remote satellites, with the noise level reaching 300pm/√Hz@1Hz, creating new opportunities for high-precision geophysical applications[3].
The Geodesy and Navigation Team at Sun Yat-sen University (SYSU), in collaboration with Innovation Academy for Precision Measurement Science and Technology(AMP), and other partners developed an independently produced GRACE-FO LRI1B data product (SYSU LRI1B S10). We applied meticulous processing in time-tags correction, phase anomaly processing, and scale factor estimation to minimize the impact of outliers such as time-tags anomalies and phase anomalies on the LRI1B data. The precision of LRI1B S10 reaches 1nm/s/√Hz@0.1Hz, which is superior to Jet Propulsion Laboratory's (JPL) LRI1B RL04 product (~5nm/s/√Hz@0.1Hz), and comparable to similar products developed by the German Albert Einstein Institute (AEI). Gravity field recovery results based on LRI1B S10 show that its degree variance beyond degree 30 is better than LRI1B RL04, with better consistency with AEI's results. This data product provides a crucial basis for future GRACE-like gravity satellite missions (e.g., TianQin 2), supporting advancements in data processing, gravity field recovery, and surface mass migration research.
The data processing workflow primarily includes time correction, phase anomaly elimination, phase subtraction, scale factor estimation, light time correction, low-pass filtering, and down-sampling. Appendix A provides a brief introduction to the key steps, and detailed explanations can be found in references [4–6]. The data is stored in YAML format provided by the Science Data System (SDS), with additional details in Appendix B. This dataset covers the entire period of LRI measurements on the GRACE-FO mission from June 2018 to July 2023. It can be directly used for gravity field recovery, mascon solutions, and other scientific applications. Future processed LRI1B data from GRACE-FO will continue to be released.
This dataset was supported by grants from the National Key Research and Development Program of China (2022YFC2204601) and the National Natural Science Foundation of China (12261131504, 42174103). Usage and feedback are welcome.
Authors:
Heng Yin ( yinh39@mail2.sysu.edu.cn )，Yihao Yan，Min Zhong，Wei Feng，Zitong Zhu, Changqing Wang, Jubo Zhu, Defeng Gu
Data download link:
https://zenodo.org/records/13743822
Comparasion of LRI1B biased range rate
Fig. 1 shows the comparison of LRI1B biased range rate spectral density among the SYSU S10, JPL RL04, and AEI V50. For frequencies above 30 mHz, all three products exhibit noise levels below the LRI design level. Both the SYSU S10 and AEI V50 products achieve the precision of 1 nm/s/√Hz @ 0.1 Hz, while the JPL RL04 product, impacted by larger phase anomalies, has a precision of around 5 nm/s/√Hz @ 0.1 Hz. The differences among the three products at 1cpr (cycles per revolution) and 2cpr mainly stem from the variations in the estimation of the scale factors used in each product.
Fig. 1. The comparison of LRI1B biased range rate spectral density among the SYSU S10, JPL RL04, and AEI V50. The red, blue, and green denote the results of JPL RL04, AEI V50, and SYSU S10, respectively. The purple, orange, and brown curves represent the differences between these products, while the black curve indicates the designed noise level for the GRACE-FO LRI.
Appendix A. Overview of LRI1B S10 Data Processing Procedure:
1 Time-tags correction
The GRACE-FO LRI phase measurements use local LRP(Laser Ranging Processor) time frame. To synchronize the LRI phase measurements between the two satellites, the LRP time frames should be convert to the GPS time frames. The time-tags correction method applied in the SYSU LRI1B S10 follows the procedures outlined in [4,5,7].
2 Phase anomaly removal:
The raw LRI phase measurements contain a large number of phase anomalies, such as Phase Jumps(PJs), Cycle Slips (CCs), Single Event Upsets(SEUs) and Momentum Transfer Events(MET) . The SYSU LRI1B S10 product uses the template method to eliminate most PJs. For anomalies like mega PJs, abnormal shape PJs, CCs, SEUs and MET, phase smoothing is applied to ensure that the LRI1B S10 is free from phase anomalies to the greatest extent possible.
3 Light time correction:
To convert the biased range, derived from the LRI phase measurements, into the instantaneous range between the satellites, the light time correction is necessary. The light time correction in LRI1B S10 accounts for both special relativity and shapiro delay from general relativity. Detailed methods can be found in [8].
4 Scale factor estimation:
The laser frequency onboard (vM) experiences a long-term drift relative to the ground-measured frequency(v0, 2.81616393e14 Hz for satellite C and 2.81615684e14 for satellite D). To convert the LRI phase measurements into the range measurements accurately, the exact laser frequency is need to be estimated by multiplying 1/(1+scale factor) to the ground frequency[9]. In LRI1B S10, the daily scale factor(s) is estimated using the least-squares method by minimizing the instantaneous range rate differences between LRI and KBR systems.
5 Low-pass filtering and down-sampling
Down-sampling are applied to reduce the sampling rate of LRI biased range from 9.664 Hz to 0.5 Hz. Similarly, the light time correction is down-sampled from its original 1 Hz to 0.5 Hz. To prevent data aliasing, CRN low-pass filtering is applied before down-sampling[10]. The filter parameters are listed in table 1, where: fs is the LRI1A phase measurements sampling rate. f0 refers to the GRACE-FO orbit frequency. Nc is the convolution number. fc is the cutoff frequency. Nf is the length of the filter window.
Table 1 CRN filtering coefficients
|
fs (C) |
fs (D) |
f0 |
Nc |
fc |
Nf |
|
9.664Hz |
9.664198Hz |
0.176mHz |
9 |
0.25Hz |
747 |
Appendix B: LRI1B File and Data Format Description：
The LRI1B S10 is formatted similarly to the GRACE-FO LRI1B text files provided by the SDS. The files consist of two parts: a header with YAML format and a body containing data content[6].
Header: The header begins with the term “header” and ends with “# End of YAML header”. It contains data description information such as data length, start and end times, file version, and descriptions of the variables included in the file.
Body: Following the header, the data is arranged in columns separated by spaces. The order of the parameters in these columns is specified in the header.
This data format refers to the JPL LRI1B RL04 official product and AEI V50[6,11]. The variable contents are specified as follows.
|
variables name |
column |
unit |
definition |
|
gps_time |
1 |
s |
Seconds past 12:00:00 noon of January 1, 2000 in GPS Time |
|
biased_range |
2 |
m |
CRN-filtered biased inter-satellite range |
|
range_rate |
3 |
m/s |
First time derivative of biased range |
|
range_accl |
4 |
m/s2 |
Second time derivative of biased range |
|
iono_corr |
5 |
|
Estimated scale correction epsilon for biased range, range rate, and range acceleration due to unknown onboard LRI frequency |
|
lighttime_corr |
6 |
m |
Light time correction for biased range |
|
lighttime_rate |
7 |
m/s |
Light time correction for range rate |
|
lighttime_accl |
8 |
m/s2 |
Light time correction for range acceleration |
|
ant_centr_corr |
9 |
m |
Not defined yet, set to 0 |
|
ant_centr_rate |
10 |
m/s |
Not defined yet, set to 0 |
|
ant_centr_accl |
11 |
m/s2 |
Not defined yet, set to 0 |
|
K_A_SNR |
12 |
db-Hz |
CNR of laser ranging for GRACE-FO C satellite |
|
Ka_A_SNR |
13 |
|
Not defined, set to 0 |
|
K_B_SNR |
14 |
db-Hz |
CNR of laser ranging for GRACE-FO D satellite |
|
Ka_B_SNR |
15 |
ns |
Estimated time offset between KBR and LRI |
|
qualflg |
16 |
|
- bit 0 = phase break - bit 1 = phase data was smoothed - bit 2 = phase jump was removed - bit 3 = mega phase jump was removed - bit 4 = sun-blinding period - bit 5 = phase disturbance on one S/C - bit 6 = momentum transfer event - bit 7 = not defined |
*Since LRI1B adopts a format compatible with KBR1B, the variable names follow the conventions used in KBR1B. For example, K_A_SNR and K_B_SNR represent the Carrier to Noise Ratio (CNR) for satellite C and D in LRI1B, respectively, while Ka_B_SNR refers to the time offset between the LRI and KBR systems.
Reference
[1] TAPLEY B D, BETTADPUR S, RIES J C, et al. GRACE Measurements of Mass Variability in the Earth System[J]. Science, 2004, 305(5683): 503-505.
[2] KORNFELD R P, ARNOLD B W, GROSS M A, et al. GRACE-FO: The Gravity Recovery and Climate Experiment Follow-On Mission[J]. Journal of Spacecraft and Rockets, 2019, 56(3): 931-951.
[3] ABICH K, ABRAMOVICI A, AMPARAN B, et al. In-Orbit Performance of the GRACE Follow-on Laser Ranging Interferometer[J]. Phys Rev Lett, 2019, 123(3): 031101.
[4] YIN H, et al. Raw data processing of laser ranging interferometer based on GRACE-FO gravity satellite[J]. Chinese Journal of Geophysics-Chinese Edition, 2024.
[5] MÜLLER L. Generation of Level 1 Data Products and Validating the Correctness of Currently Available Release 04 Data for the GRACE Follow-On Laser Ranging Interferometer[D]. Leibniz Universität Hannover, Hannover, Germany, 2021.
[6] WEN H Y, KRUIZINGA G, PAIK M, et al. Gravity recovery and climate experiment (GRACE) Follow-On (GRACE-FO) level-1 data product user handbook[Z]. 2019.
[7] MISFELDT M. Data Processing and Investigation for the GRACE-FO Laser Ranging Interferometer[D]. Leibniz Universität Hannover, Hannover, Germany, 2019.
[8] YAN Y, MUELLER V, HEINZEL G, et al. Revisiting the light time correction in gravimetric missions like GRACE and GRACE follow-on[J]. Journal of Geodesy, 2021, 95(5): 48.
[9] MISFELDT M, MÜLLER V, MÜLLER L, et al. Scale Factor Determination for the GRACE Follow-On Laser Ranging Interferometer Including Thermal Coupling[J]. Remote Sensing, 2023, 15(3): 570.
[10] THOMAS J B. An Analysis of Gravity-Field Estimation Based on Intersatellite Dual-1-Way Biased Ranging[C]. Pasadena, California, 1999: Technical Report 98-15.
[11] MÜLLER L, MÜLLER V, MISFELDT M. AEI LRI1B and RTC1B Release Notes[Z]. 2022.
