Refining gravity anomaly data of coastal areas by combining XGM2019e-2159 and SRTM/GEBCO_2024 residual terrain model with forward modeling method

Abstract

As one of the Earth's fundamental physical fields, the gravity field model's accuracy is considerably constrained in areas with sparse coverage or data gaps. In coastal areas, satellite altimetry data are affected by land contamination and errors from tidal models, while shipborne gravity measurements fail to obtain valid gravity data in nearshore regions. Therefore, gravity field models' accuracy in coastal areas is relatively lower. Additionally, due to the truncation of global gravity field models at specific degrees, truncation errors prevent the acquisition of high-precision gravity anomaly (GA) information. To address this issue, a gravity forward modeling-based refinement method using the residual terrain model (RTM) is proposed for coastal regions. In this study, a mass center offset correction (MCOC) is introduced as the main innovation to mitigate prism position shift errors induced by the rock-equivalent topography (RET) method. Specifically, a high-resolution detailed topography model of the study area, referred to as the SRTM–GEBCO model, was first constructed by merging the land digital elevation model SRTM V4.1 with the ocean bathymetric model GEBCO_2024. Based on this model, the reference topography was then derived using a spatial filtering approach based on the moving average method. The RTM is then discretized into regular grid prisms, and the GA generated by the RTM at target points is computed in the spatial domain using the prism integration method to refine the XGM2019e-2159 gravity anomaly (XGM-GA) model. For computation points located in coastal areas, the RET method is adopted to avoid distinguishing between the densities of land and ocean prisms during forward modeling, and the MCOC is further applied to correct the resulting positional bias. To validate the feasibility of this method, this study focuses on a selected region along the U.S. West Coast (125–122° W, 39–42° N) and refines the XGM-GA model. The experimental results are validated using NGS99 measured GA data. The results indicate that, after RTM correction, the root mean square difference between the modeled and measured GA decreases from 14.58 to 8.25 mGal overall and from 14.98 to 7.97 mGal in coastal areas. Furthermore, power spectral density analysis of the XGM-GA model before and after RTM correction reveals a significant enhancement in short-wavelength energy in the corrected model. The results demonstrate that gravity forward modeling based on the RTM is effective in compensating for the truncation errors of the XGM-GA model, thereby improving its accuracy.