<p>In this study, we investigate systematic errors in our temporal gravity solutions computed using the improved energy balance approach (EBA) (Shang et al. 2015) by reprocessing the GRACE JPL RL03 L1B data product. Our processing consists of two steps: the first part is the estimation of in-situ geopotential differences (GPD) at the satellite altitude using the energy balance formalism, the second part is the estimation of spherical harmonic coefficients (SHCs) of the global temporal gravity field model using the estimated GPDs. The first step includes daily dynamic orbit reconstruction by readjusting the reduced-dynamic (GNV1B) orbit considering the reference model, and estimating the accelerometer calibration parameters. This is coupled with the alignment of the intersatellite velocity pitch from KBR range rate observations. Due to the strategy of using KBR range-rate in our processing algorithm, the estimation of in-situ geopotential differences (GPD) includes both the systematic errors and the high-frequency noise that result from the range-rate observations. Since estimated GPDs are linearly connected with the spherical harmonic coefficients (SHCs) of the global gravity field model, our temporal models are affected by these errors, especially in high-degree coefficients of the temporal gravity field solutions (from n=25 to n=60).</p><p>In order to increase our solution accuracy, we fit additional empirical parameters for different arc lengths to mitigate the systematic errors in our GPD estimates, thus improving our temporal gravity field solutions. Our EBA approach GRACE monthly gravity field models are validated by comparisons to the official L2 data products, including the official solutions from CSR (Bettadpur et al., 2018), JPL (Yuan et al., 2018) and GFZ (Dahle et al., 2018).</p>
Impact of Different Kinematic Empirical Parameters Processing Strategies on Temporal Gravity Field Model Determination
