A Study of the Algorithms for the Detection of Tsunami Using an Ocean Bottom Pressure Recorder

Abstract

AbstractThe National Institute of Ocean Technology (NIOT), an autonomous organization under the Ministry of Earth Sciences, government of India, is engaged in developing and installing systems for tsunami detection and reporting. This involves high-precision bottom pressure recorders (BPRs) installed on the ocean floor, which can detect water level changes in the order of a few centimeters. Data are logged and recorded subsea by instruments located close to the BPRs. The detection of abnormal changes in the water level is required for detecting a tsunami event. This paper describes algorithms incorporated in most BPRs for detecting a tsunami by predictive methods such as Newton’s Extrapolation and Kalman predictor techniques. The most widely used tsunami detection algorithm is based on Newton’s extrapolation. The tsunami detection technique based on the Kalman prediction algorithm developed by NIOT can be an alternative for the existing technique. This paper describes both the algorithms and analyzes their effectiveness during tsunami event detection using MATLAB software. It is found that the Kalman algorithm has a better detection performance over the Newton extrapolation technique for tsunami wave amplitudes up to 300 mm. The Newton extrapolation technique has a better detection performance for tsunami wave duration of less than 10 min. For tsunami wave durations greater than 10 min, the Kalman algorithm has a better detection performance. As the wave durations of most of the recorded tsunamis are greater than 10 min, the Kalman algorithm could be a viable substitute for tsunami detection.