Citation

Bhardwaj R, Kumar A, Sharma ML. Disaster Adv. 2013; 6(3): 24-31.

Copyright

(Copyright © 2013, Shankar Gargh)

DOI

unavailable

PMID

unavailable

Abstract

A practical approach for mitigating earthquake risk may be the Earthquake Early Warning (EEW). An EEW system is capable of issuing a warning few seconds prior to arrival of approaching ground motion to the populated and other vulnerable area. In this study a simple and robust algorithm based on the exceedance of specified threshold time domain amplitude levels has been discussed for automatic primary wave onset (P-onset) detection and EEW analysis. The algorithm has been applied on combined strong motion database (K-NET + PEER-NGA) and tested on Indian strong motion database which demonstrates satisfactory results. The databases are processed to compute two parameter, root sum of squares cumulative velocity (RSSCV) and cumulative absolute velocity (CAV) respectively and compared with specified threshold levels. In case of automatic P-onset detection, an accuracy of around 71% and 76% has been achieved for combined dataset and Indian dataset respectively, for a time difference of 0.1 sec. The estimated threshold levels for RSSCV and CAV attribute are 10 cm/sec and 41 cm/sec respectively. Percentage of correct P-onset pick as well as correct alarms estimation using RSSCV and CAV attributes has been discussed in details. Individual station vote for RSSCV and CAV values have been used to detect the occurrence of a major earthquake (M >= 6) and provide onsite warning within seconds after the arrival of the primary waves in the area around the station. When the array of stations is available, the approach can be applied to multistation data, a three stations vote method and average of event vote method have been used in this study.