Global Navigation Satellite System–Acoustic (GNSS-A) positioning technique is an important tool for monitoring the submarine tectonic movement and seismic. The submarine positioning accuracy however is seriously affected by the sound speed variation. This contribution investigates the influence of sound speed variation on the seafloor geodetic positioning and proposes diversified acoustic delay models, including two five-parameter models and two three parameter models. It shows that the five-parameter zenith delay model can be degenerated into a three-parameter zenith delay model in the single-point positioning case or under the single-layer sound speed field (SSF) assumption. The proposed zenith delay models are verified by the Japanese opened seafloor geodesy observation-array (SGO-A) data and the sound speed gradient relative to the reference sound speed profile (SSP) is obtained. Experimental tests show that, both the proposed five-parameter zenith delay and three-parameter zenith delay models can achieve a three-dimensional positioning precision at centimeter-level and they can be used to produce a more stable long-term horizontal coordinate time series relative to the GNSS-A ranging combined positioning solver (GARPOS V1.0.0).
