Layered ocean sound speed model sharpens seafloor positioning

Researchers in China have developed a layered sound speed gradient model for GNSS-A seafloor positioning that accounts for depth-dependent ocean variation. The method cut simulated errors from decimeters to millimeters and improved field results, with potential uses in seafloor deformation monitoring and underwater acoustics.

Why it matters: - GNSS-A positioning is used to track seafloor tectonic movement and support global reference frames. - Existing methods can miss depth-dependent sound speed changes in the ocean, which can push seafloor coordinate errors into the centimeter range. - The new layered model improves positioning precision and could also help extract ocean structure information from routine surveys.

What happened: - Researchers from Shandong University in Weihai and the Chinese Academy of Surveying and Mapping in Beijing developed a layered horizontal sound speed gradient model for GNSS-A positioning. - The study was published online May 29, 2026 in Satellite Navigation. - The paper is identified by DOI: 10.1186/s43020-026-00197-w. - The original source URL is the published paper.

The details: - The model divides the water column into depth intervals, with each layer assigned its own time-varying eastward and northward sound speed gradient. - Adjacent layers are connected through depth-weighted continuity constraints to stabilize the joint inversion of seafloor transponder coordinates and sound speed parameters. - Example layer divisions in the study included 0–300 m, 300–800 m, and 800–3000 m. - Each layer’s horizontal gradient was represented with B-spline basis functions. - The constraints were weaker in the upper ocean, where variability is higher, and stronger at depth, where conditions are smoother. - The layered design was intended to avoid unphysical oscillations between layers while preserving real vertical structure. - In simulations, the conventional single-layer model produced 3D positioning errors of 66–231 mm. - The layered model reduced those simulation errors to 2.6–5.6 mm. - In a representative MYGI field campaign, the acoustic travel-time residual standard deviation fell from 0.1243 ms to 0.0787 ms. - The estimated transponder coordinates closely matched independent GARPOS solutions, with most component-wise differences reduced to the millimeter level. - Long-term analysis across four seafloor geodetic sites — MYGI, KAMS, CHOS, and FUKU — showed stable multi-epoch coordinate time series over nearly a decade. - Those long-term results were broadly consistent with GARPOS and showed less scatter than conventional single-layer models in several components.

Between the lines: - The conventional GNSS-A approach effectively averages horizontal sound speed gradients across the whole water column, which works best only when gradients are nearly depth-invariant. - The new method treats the ocean as layered, reflecting the fact that temperature and salinity often change sharply near the surface and more smoothly at depth. - That makes the inversion more physically realistic and reduces systematic bias in acoustic travel-time residuals. - The approach also turns positioning data into a source of oceanographic information, because layer-wise gradients reveal how sound speed variability changes with depth.

What's next: - The researchers plan to test adaptive layer selection and add ocean model data to strengthen constraints. - The layered framework could be extended beyond GNSS-A to other underwater acoustic positioning systems, including Long Baseline and Ultra-Short Baseline systems. - Further development could support longer-term seafloor monitoring and more accurate marine geodetic reference frames.

The bottom line: - A layered sound speed model gives GNSS-A a more realistic view of the ocean and can push seafloor positioning from decimeter-scale errors toward millimeter-level precision.