An Adaptive Nonlinear Iterative Method for Predicting Seafloor Topography From Altimetry‐Derived Gravity Data

Abstract

The ocean covers 71% of the Earth's surface. At present, only about 20% of the seafloor topography (ST) has been directly measured by ships, and most areas are predicted from satellite altimetry‐derived gravity products. In this study, an adaptive nonlinear iterative (ANI) method is proposed to address two major problems in gravity ST inversion: linear approximation and empirical seafloor density contrast (SDC). In ANI, the SDC is adaptively estimated as an output, while higher‐order Parker expansion and modified Bott's iteration are combined to recover nonlinear topography. We apply our new method using the DTU21GRA altimetric gravity model and single‐beam bathymetry to predict the ST in a part of the South China Sea. Results reveal that the average SDC in the study area is 1.24 g/cm3, which compares well to CRUST1.0. The root‐mean‐square (RMS) error between the nonlinear model and single‐beam checkpoints is 102.1 m, which is improved by 34.5%, 29.2%, and 18.3% compared with the non‐gravity model, topo_24.1, and linear model, respectively. The RMS error between the nonlinear model and multibeam bathymetry is 91.0 m, which is better than the linear model. Analysis of two‐dimensional profiles shows that the nonlinear model reveals more terrain details than the linear model.

A regional digital bathymetric model fusion method based on topographic slope: A case study of the south China sea and surrounding waters

“…Smith and Sandwell (1994) calculated a seafloor topography model of the Southern Ocean using the remove-restore technique based on the correlation seen between marine gravity anomalies and seafloor topography at heights of 15–160 km [ 8 ]. With the support of theory and practice, many schoars have used ocean gravity anomalies to invert seafloor topography [ [9] , [10] , [11] , [12] ], and a large number of regional and global seafloor topography models have emerged. Currently, most models have achieved global coverage, with the main difference being vertical accuracy.…”