IMPROVING THE ACCURACY OF IN-WATER TRAVEL-TIME PREDICTIONS FOR SEISMIC EVENT LOCATION USING TWO- DIMENSIONAL UNDERWATER ACOUSTIC PROPAGATION

Ban Treaty Organization includes seismic, hydroacoustic and infrasound sensors that record signals travelling through the solid earth, ocean and atmosphere. Signals may be generated by events including earthquakes, volcanic eruptions and explosions. The times at which signals from the same event arrive at geographically separated sensors can be used to locate the event, provided that travel time from the sensor to all possible event positions is known. The time taken for sound to travel through the ocean from its generating event to the receiver can be approximated as being simply proportional to distance, using an average value for the sound-speed of seawater. However, this approach ignores variations of sound speed with position and time and this can result in location errors for events whose position is determined by the arrival times of waterborne signals. These errors can be reduced if databases containing seawater sound-speed as a function of time, latitude, longitude and depth are used along with a numerical propagation model. This database/model combination allows travel time to be predicted from all candidate locations to receivers of interest. This paper investigates the benefits that can be achieved by the use of a database/model approach, relative to the approach of a single, average sound speed. Improvements are quantified vi