TY - JOUR

T1 - GRACE Follow-On accelerometer data recovery by high-precision environment modelling

AU - Huckfeldt, Moritz

AU - Wöske, Florian

AU - Rievers, Benny

AU - List, Meike

N1 - Publisher Copyright: © 2024 COSPAR

PY - 2024/6/15

Y1 - 2024/6/15

N2 - The Gravity Recovery and Climate Experiment Follow-On (GRACE-FO) satellites are equipped with high-precision three-axis accelerometers to measure all non-gravitational accelerations acting on the satellites. Precise knowledge of these perturbations is a major prerequisite for successful Gravity Field Recovery (GFR). Unfortunately, after only one month in orbit the accelerometer on one of the two satellites produced decreasingly accurate measurements. Due to this, the GRACE-D accelerometer data has to be replaced by artificial data for the use in the GFR process. In most existing approaches, which are called transplants, this data is constructed from GRACE-C measurements. We are developing a GRACE-D accelerometer product that relies mainly on modelled acceleration data instead of a full transplant. By using physical models to generate acceleration data, we are able to better understand the accelerometer signal and its characteristics, which is beneficial for subsequent processes involving accelerometer data. We present the evaluation of our modelled data in terms of comparison to the official Science Data System (SDS) accelerometer data. For the first GRACE-FO satellite real accelerometer is used as a reference, but for the second satellite the main comparisons are done with respect to the ACH1B transplant data product. To get information about environmental changes and long-term effects data sets covering one year and nearly the whole mission duration are presented. In terms of gravity field solutions the performance of purely modelled accelerometer data for GRACE-D and a minimalistic transplant approach of estimated density values is compared to transplant products of other processing centres for the year 2019 as well as one month of 2023. Through this, the modelled solution can act as a cross-reference to the transplant, contributing to an improved implementation of the external perturbation characteristics and the influence of higher solar activity on our models and the subsequent gravity field solutions can be determined. Comparison of the artificial acceleration data to the real ACT1B data for GRACE-C showed that models related to the atmospheric drag are the limiting factors in our high-precision environment modelling approach. Nevertheless, it was still possible to generate monthly gravity field solutions with a combination of ACT1B data for GRACE-C and artificial data for GRACE-D that showed hydrological signals. Application of the transplant that substitutes the limiting modelled thermospheric density with estimated values obtained at GRACE-C positions reduces the residuals between artificial and real acceleration data significantly. Estimation of gravity fields showed that the improvements due to the transplant directly transfer into the gravity field solutions. Our minimalistic and physically motivated transplant is shown to be a valuable alternative to the transplant data from TUG and JPL. The GROOPS software is utilized to generate monthly gravity field solution to guarantee comparability of the presented solutions for the geodesy community. The presented results not only show the quality of our transplant product but also enabled us to identify potential areas of improvement in the data generation process and our in–house gravity field recovery tool. The ZARM GRACE-D transplant product, all radiation accelerations and estimated density is available for download at: https://www.zarm.uni-bremen.de/zarm_daten/.

AB - The Gravity Recovery and Climate Experiment Follow-On (GRACE-FO) satellites are equipped with high-precision three-axis accelerometers to measure all non-gravitational accelerations acting on the satellites. Precise knowledge of these perturbations is a major prerequisite for successful Gravity Field Recovery (GFR). Unfortunately, after only one month in orbit the accelerometer on one of the two satellites produced decreasingly accurate measurements. Due to this, the GRACE-D accelerometer data has to be replaced by artificial data for the use in the GFR process. In most existing approaches, which are called transplants, this data is constructed from GRACE-C measurements. We are developing a GRACE-D accelerometer product that relies mainly on modelled acceleration data instead of a full transplant. By using physical models to generate acceleration data, we are able to better understand the accelerometer signal and its characteristics, which is beneficial for subsequent processes involving accelerometer data. We present the evaluation of our modelled data in terms of comparison to the official Science Data System (SDS) accelerometer data. For the first GRACE-FO satellite real accelerometer is used as a reference, but for the second satellite the main comparisons are done with respect to the ACH1B transplant data product. To get information about environmental changes and long-term effects data sets covering one year and nearly the whole mission duration are presented. In terms of gravity field solutions the performance of purely modelled accelerometer data for GRACE-D and a minimalistic transplant approach of estimated density values is compared to transplant products of other processing centres for the year 2019 as well as one month of 2023. Through this, the modelled solution can act as a cross-reference to the transplant, contributing to an improved implementation of the external perturbation characteristics and the influence of higher solar activity on our models and the subsequent gravity field solutions can be determined. Comparison of the artificial acceleration data to the real ACT1B data for GRACE-C showed that models related to the atmospheric drag are the limiting factors in our high-precision environment modelling approach. Nevertheless, it was still possible to generate monthly gravity field solutions with a combination of ACT1B data for GRACE-C and artificial data for GRACE-D that showed hydrological signals. Application of the transplant that substitutes the limiting modelled thermospheric density with estimated values obtained at GRACE-C positions reduces the residuals between artificial and real acceleration data significantly. Estimation of gravity fields showed that the improvements due to the transplant directly transfer into the gravity field solutions. Our minimalistic and physically motivated transplant is shown to be a valuable alternative to the transplant data from TUG and JPL. The GROOPS software is utilized to generate monthly gravity field solution to guarantee comparability of the presented solutions for the geodesy community. The presented results not only show the quality of our transplant product but also enabled us to identify potential areas of improvement in the data generation process and our in–house gravity field recovery tool. The ZARM GRACE-D transplant product, all radiation accelerations and estimated density is available for download at: https://www.zarm.uni-bremen.de/zarm_daten/.

KW - Accelerometer

KW - Environment modelling

KW - Geodesy

KW - Gravity field

KW - Transplant

UR - http://www.scopus.com/inward/record.url?scp=85190091444&partnerID=8YFLogxK

U2 - 10.1016/j.asr.2024.03.068

DO - 10.1016/j.asr.2024.03.068

M3 - Article

VL - 73

SP - 5783

EP - 5805

JO - Advances in space research

JF - Advances in space research

SN - 0273-1177

IS - 12

ER -