TUD-L3-5dayEWH-GRACEv2: Novel and Global 0.5°x0.5° 5-day GRACE equivalent water height gridded product

This GRACE Level-3 data product provides gridded Equivalent Water Height (EWH), which are equivalent to (uncorrected) Terrestrial Water Storage Anomalies (TWSA) generated at TU Delft (“TUD”) and available at 5-day to sub-weekly time resolution. The L3Uncorr datasets are derived exclusively from GRACE data:Residual Level-1B (L1B) data (post-fits and geo-fits), provided by Centre for Space Research (UTCSR, TX, USA)*. DOI: 10.18738/T8/VY3U8LCSR RL06 Level-2 data (DDK3-filtered, d.o. 2-60).GGM05C static (long-term mean) field removed.L3 products required additional datasets: (i) NASA GSFC SLR 5x5+C61/S61 coefficients, (ii) geocenter motion (TN-13 CSR RL06) and (iii) Glacial Isostatic Adjustment (GIA). Three different formulations are provided depending on the residual L1B quantity:Based on Line-of-sight Gravity Differences (LGDs): TUD-L3Uncorr-5dayEWH-GRACEv2-LGD-2002_2017-0.5x0.5.nc,Based on residual ranges: TUD-L3Uncorr-5dayEWH-GRACEv2-Range-2002_2017-0.5x0.5.nc,Based on the hybridised variable (LGD+resRange): TUD-L3Uncorr-5dayEWH-GRACEv2-Hybrid-2002_2017-0.5x0.5.nc. These are also provided at the monthly level (TUD-L3[Uncorr]-EWH).The uncertainty grids are also available with the naming convention: TUD-L3[Uncorr]-5dayEWH_UNC-GRACEv2. Important notes for users:"L3Uncorr" solutions refer to 5-day EWH solutions which have NOT been corrected for: GIA, low-degree coefficients (geocenter, C_20, and C_30), and persistent aliasing tides.The datasets span the period April 2002 to June 2017.These gridded EWHs are anomalies with respect to the reference epoch: January 2008. We recommend using the uncertainties TUD-L3-EWH_UNC. This dataset includes C20, C30, geocenter uncertainties (formal errors), and the GIA uncertainty estimate. This dataset is intended for users seeking higher-frequency gravity field variations based solely on GRACE residuals, and is valuable for hydrological and geophysical analyses. The processing steps from residual L1B range rates to LGDs can be found in https://doi.org/10.1029/2025JB031615. Details on the data product structure and use can be found at the TUDelft GRACEcube portal:https://grace-cube.lr.tudelft.nl/L2B_5DAY. A manuscript describing this dataset version and methodology has been submitted to Journal of Geodesy in August 2026 (under review). * We thank CSR for making the dataset publicly available. Without it, this research would not have been possible. Versioning:Updated and final v2, before official v3 (does not match 4TU versioning which is "v4").:Uncertainties of 5-day TUD-L3 solutions are now available. The uncertainties in the Uncorr version suffer from an abrupt jump around 2012. This is not the case for the corrected version. Furthermore, the GIA trend uncertainty is added (units: cm/yr). v2-new (does not match 4TU versioning which is "v3").:Corrected L3Uncorr solutions resulting in resulting in L3 products for:(1) GIA,(2) re-inclusion of low degrees,(3) geocenter motion,(4) and persistent tidal aliasing signals (S2, P1, S1, and R3).Uncertainties of 5-day TUD-L3 solutions are mainly described by TUD-L3Uncorr-5dayEWH_UNC; however, v3 will include an improved L3 uncertainty estimation. v2-old: (1) Significant methodological improvements, including enhanced outlier handling and refined altitude/range correction for improved 5-day signal recovery. (2) Inclusion of full uncertainty grids in TUD-L3Uncorr-5dayEWH_UNC. v1.0(exp): first release (experimental, released for retraceability and transparency).