Using Echo Sounder Technology for Detecting and Predicting Local Sea State

Cabled ocean observatories, which support continuously deployed underwater sensors, enable an increasing amount of near real-time, high resolution data to be collected from the marine environment. These data can be used in the design of intelligent systems to predict and measure properties of the marine environment. One environmental factor that is very important to maritime transportation and safety is sea state, that is, the "roughness" of the sea surface. Sea state is dependent on winds, but can vary significantly with location and over time, based on a wide range of factors. We show how publicly available sensor data can be used to detect and predict local sea state. To do this, we will discuss various techniques that can help us deal with a large set of echo sounder data, and aggregate it in a way that allows basic machine learning techniques to categorize the local sea state at a given time according to the Douglas Sea Scale.