Detecting of Precursor Type Small Amplitude Anomalies In Geomagnetic Time Series
Abstract
We present the method of detecting the precursor type small amplitude anomalies in geomagnetic field differential time series. Analysis of multiannual investigations of geomagnetic field dynamics on the geophysical network in seismoactive Tran- scarpathian trough manifested that developing of seismical activity in the region often is accompanied by anomalous changes in differential magnetic field. Duration of these anomalies 0,5-4 month. Their amplitudes up to 7 nT. Anomalies are of baylike shape. Their peaks are located nearer to the close. The moments of local earthquakes lie on the stage of field normalizing. The anomalous changes between various pairs of geo- magnetic poins depends on the location of earthquakes epicenters. The abovewritten is a base of formulation the problem of regional seismic regime geomagnetic precur- sors model construction. Such a model has to rest on correlations between energy, time, locations of catalogued earthquakes and anomalous differential magnetic field amplitudes and durations for all pair of points in the network. The main difficulty of these correlations determination depends on rather small amplitudes of precursors type anomalies. Often they can not de determined on noise background in visual. The pre- sented method of such anomalies detecting includes next main stages. Elimination of bad data by analysis of intraquartiles distances. Correlation coefficients between dif- ferential time series and half period of running assymetric sine is calculated. The sine assymetry consist of unequal real time duration at 1-st and 2-nd its quarteperiods. The algorithmus central part consist of two inner cycles: cycle of sine halfperiod change and cycle of peak moving. Construction of sine-correlation filters for investigated time series on the base of maximum correlation coefficients (by absolute value). Convolu- tion of constructed filters with investigated time series. The revision of method by help of artifically created time series presents a rather good results. The precursor type dif- ferential geomagnetic anomalies can be assurely detected on the background of noise with dispersion several times larger then amplitude of would be detected anomalies.
- Publication:
-
EGS General Assembly Conference Abstracts
- Pub Date:
- 2002
- Bibcode:
- 2002EGSGA..27.1367G