Data from: Can a persistence-based prediction of GRACE gravity field coefficients reduce uncertainties at finer temporal scales?
Abstract
This dataset contains the raw and processed data used for the study 'Can a persistence-based prediction of GRACE gravity field coefficients reduce uncertainties at finer temporal scales?'. Raw data includes sets of sample data from the processing levels 1B and 2 of the Gravity Recovery and Climate Experiment (GRACE), where level 1B consists of the samples of satellite measurement products and level 2 consists of the geopotential coefficients. These data were used to investigate the impact of a persistence-based prediction approach in reducing the uncertainties of monthly and finer temporal scale gravity coefficients.
Description of data and files
Folder: Data
This folder contains
- processed data >> gravity field coefficients from energy balance approach (EBA)
- processed data >> gravity field coefficient estimates from the predictive approach
Below is the summary of the folder/file structure and a detailed description of their contents.
Folder summary
- EBA > gravity coefficients at 30-day, 15-day, 7-day intervals; processed from EBA
- GC estimates > estimated gravity coefficients at 30-day, 15-day, 7-day intervals; processed from the predictive approach
Contents in these folders are described below.
Descriptions
1. EBA \ Data
Source: processed gravity coefficients processed from the energy balance approach (EBA)
Tool: The open-source gravity field recovery tool GRACETOOLS is from Darbeheshti et al. (2020). [https://doi.org/10.1007/978-3-030-21718-1_10]
Temporal resolution: 30-day, 15-day, 7-day
Spherical harmonic expansion: 96 x 96 (degree & order 0 to 96, total no. of gravity coefficients = 4753)
Spatial extent: global
Temporal extent: August 2002-July 2013
File count: 3 data files, each composed of C_lm, S_lm, Std_C_lm, and Std_S_lm
File format: .mat
File naming convention:
- 'eba7data.mat' - processed from EBA, 7-day gravity coefficients stored in .mat file
- 'eba15data.mat' - processed from EBA, 15-day gravity coefficients stored in .mat file
- 'eba30data.mat' - processed from EBA, 30-day gravity coefficients stored in .mat file
Variable names:
- 'data_C' > double > num_timesteps x 4753 > 4753 C_lm coefficient time series for num_timesteps = 132 (30-day), 264 (15-day), 528 (7-day)
- 'data_S' > double > num_timesteps x 4753 > 4753 S_lm coefficient time series for num_timesteps = 132 (30-day), 264 (15-day), 528 (7-day)
- 'std_C' > double > num_timesteps x 4753 > 4753 C_lm's standard deviation time series for num_timesteps = 132 (30-day), 264 (15-day), 528 (7-day)
- 'std_S' > double > num_timesteps x 4753 > 4753 S_lm's standard deviation time series for num_timesteps = 132 (30-day), 264 (15-day), 528 (7-day)
Units: none
2. GC estimates \ Data
Source: gravity coefficients estimated from the predictive approach described in the study
Temporal resolution: 30-day, 15-day, 7-day
Spherical harmonic expansion: 96 x 96 (degree & order 0 to 96, total no. of gravity coefficients = 4753)
Spatial extent: global
Temporal extent: August 2002-July 2013
File count: 4 data files, each composed of kf_C, kf_S, kfvar_C, kfvar_S
File format: .mat
File naming convention:
- 'eba7kf.mat' - EBA-based 7-day coefficients, estimated from Kalman filter-based predictive approach, stored in .mat file
- 'eba15kf.mat' - EBA-based 7-day coefficients, estimated from Kalman filter-based predictive approach, stored in .mat file
- 'eba30kf.mat' - EBA-based 7-day coefficients, estimated from Kalman filter-based predictive approach, stored in .mat file
- 'grace30kf.mat' - GRACE-based monthly coefficients, estimated from Kalman filter-based predictive approach, stored in .mat file
Variable names:
- 'kf_C' > double > num_timesteps x 4753 > 4753 C_lm coefficient time series for num_timesteps = 132 (30-day), 264 (15-day), 528 (7-day)
- 'kf_S' > double > num_timesteps x 4753 > 4753 S_lm coefficient time series for num_timesteps = 132 (30-day), 264 (15-day), 528 (7-day)
- 'kfvar_C' > double > num_timesteps x 4753 > 4753 C_lm's variance time series for num_timesteps = 132 (30-day), 264 (15-day), 528 (7-day)
- 'kfvar_S' > double > num_timesteps x 4753 > 4753 S_lm's variance time series for num_timesteps = 132 (30-day), 264 (15-day), 528 (7-day)
Units: none
Tools \ MATLAB packages
- Source: Darbeheshti et al. (2020); Sneeuw et al. (2021)
- GRACETOOLS:
open-source gravity field recovery tool, a MATLAB package to process GRACE level 1B measurements to level 2
developed by Darbeheshti et al. (2020)
accessible at https://www.geoq.uni-hannover.de/en/research/gracetools - SHBUNDLE: SHBUNDLE toolbox (version 4.2021), a MATLAB package to post-process level 2 gravity coefficients developed by Sneeuw et al. (2021); the latest version is provided via download from the Institute of Geodesy (GIS), University of Stuttgart
http://www.gis.uni-stuttgart.de/research/projects/Bundles/
Sharing/access information
This is a summary of links to publicly available data & tools.
- Level 1B and level 2 data from the Gravity Recovery and Climate Experiment (GRACE) mission were used in this work. GRACE Level 1B measurements of JPL release 2.0 (GRACE, 2001) and JPL release 3.0 (GRACE, 2018) were used. GRACE Level 2 geopotential coefficients of maximum spherical harmonic degree and order 96 and release 6.0, produced by three data centres – Center for Space Research-University of Texas (CSR) (University of Texas Center for Space Research, 2018), German Research Center for Geosciences (GFZ) (Dahle, et al., 2018), and Jet Propulsion Laboratory (JPL) (NASA Jet Propulsion Laboratory, 2018) were used.
- NASA PODAAC for GRACE level 2 geopotential coefficients are publicly accessible at
[https://podaac.jpl.nasa.gov/dataset/GRACE_GSM_L2_GRAV_CSR_RL06]
[https://podaac.jpl.nasa.gov/dataset/GRACE_GSM_L2_GRAV_GFZ_RL06]
[https://podaac.jpl.nasa.gov/dataset/GRACE_GSM_L2_GRAV_JPL_RL06] - NASA PODAAC for GRACE level 1B satellite measurements are publicly accessible at
[https://podaac.jpl.nasa.gov/dataset/GRACE_L1B_GRAV_JPL_RL02]
[https://podaac.jpl.nasa.gov/dataset/GRACE_L1B_GRAV_JPL_RL03]
- NASA PODAAC for GRACE level 2 geopotential coefficients are publicly accessible at
- The open-source gravity field recovery tool GRACETOOLS is adopted from Darbeheshti et al. (2020).
[https://www.geoq.uni-hannover.de/en/research/gracetools] - SHC synthesis and post-processing were carried out using the SHBUNDLE toolbox (version 4.2021), a MATLAB package developed by Sneeuw et al. (2021) (a latest version is provided via download from the Institute of Geodesy (GIS), University of Stuttgart.
[http://www.gis.uni-stuttgart.de/research/projects/Bundles/]
Free/open software to read the data
This study used MATLAB [version 9.14.0.2489007 (R2023a) Update 6] for reading the data, preprocessing, and processing.
Alternatively, the given data files can be opened, read, and processed using Python (with NumPy / pandas), R, and GNU Octave
References
Dahle, C., Flechtner, F., Murböck, M., Michalak, G., Neumayer, H., Abrykosov, O., Reinhold, A., & König, R. (2018). GRACE Geopotential GSM Coefficients GFZ RL06 (Version 6.0) [Dataset]. GFZ Data Services. https://doi.org/10.5880/GFZ.GRACE_06_GSM Date Accessed: 2026-03-26
2. Darbeheshti, N., Wöske, F., Weigelt, M., Wu, H., & McCullough, C. (2020). Comparison of Spacewise and Timewise Methods for GRACE Gravity Field Recovery. In J.-P. Montillet & M. S. Bos (Eds.), Geodetic Time Series Analysis in Earth Sciences (pp. 279-315). Springer International Publishing. https://doi.org/10.1007/978-3-030-21718-1_10
3. GRACE. (2001). GRACE_L1B_GRAV_JPL_RL02. Ver. 2. PO.DAAC, CA, USA. [Dataset] accessed [2024-02-16] at https://doi.org/10.5067/GRJPL-L1B02
4. GRACE. (2018). GRACE_L1B_GRAV_JPL_RL03. Ver. 3. PO.DAAC, CA, USA. [Dataset] accessed [2024-02-16] at https://doi.org/10.5067/GRJPL-L1B03
5. Sneeuw, N., Weigelt, M., Devaraju, B., Roth, M., & Antoni, M. (2021). SHBUNDLE Toolbox (4.2021). Institute of Geodesy (GIS), University of Stuttgart: http://www.gis.uni-stuttgart.de/research/projects/Bundles/
University Of Texas Center For Space Research (UTCSR). (2018). GRACE STATIC FIELD GEOPOTENTIAL COEFFICIENTS CSR RELEASE 6.0 [Dataset]. NASA Physical Oceanography Distributed Active Archive Center. https://doi.org/10.5067/GRGSM-20C06 Date Accessed: 2026-03-26
Contact
For questions or regarding further details, please contact: Dinuka Kankanige (kdmalshani@gmail.com)
All the raw data used in the study were publicly available on the NASA PODAAC drive. All the links to access this data, as well as sample data files, are provided in the data repository. Data pre-processing processing were conducted using MATLAB [version 9.14.0.2489007 (R2023a) Update 6]. Additionally, the study used an open-source MATLAB package adopted from Darbeheshti et al. (2020) as the gravity field recovery tool, and a MATLAB package developed by Sneeuw et al. (2021) as the gravity coefficient synthesis and post-processing tool. These tools are publicly accessible, and the links are provided in the data repository.