Fig. 1: Schematic diagrams of predictive model phases for underwater sound speed estimation.
From: Leveraging satellite observations and machine learning for underwater sound speed estimation
A Training Phase: The model learns from input parameters including month, depth, latitude, longitude, surface temperature, and surface salinity, along with sound speed data derived from Conductivity, Temperature, and Depth (CTD) measurements. These sound speed values are used to establish relationships between the inputs and the target output. B Prediction Phase: The model applies the learned relationships from the training phase to predict underwater sound speed using the same input parameters but without the need for CTD data.
