HR: 1330h
AN: G52B-0046    [PDF]
TI: Assessment of Noise in GPS Time Series for the Stations in mid and Southern Latitudes
AU: * EC, M
EM: ecmalaimani@hotmail.com
AF: National Geophysical Research Institute, UPPAL ROAD, HYDERABAD, AP 500 007 India
AU: N., R
EM: kumar_mala@hotmail.com
AF: National Geophysical Research Institute, UPPAL ROAD, HYDERABAD, AP 500 007 India
AB: The noise characteristics in the time series of daily position estimates for 15 GPS stations from India to Antarctica through Indian Ocean were assessed with 4 years of data using the power spectrum analysis. The analysis reveals that the white noise whose spectral index is 0.4 in addition to flicker noise whose spectral index is between 0 and 1 constitute the total "fractal white" noise in the time series. Combination of white noise (time independent) and coloured or flicker noise (time correlated) enables to characterize the errors in the time series of coordinates. We followed the approach of Ailin Mao et al., and used non-linear least squares to estimate the spectra and detected the signals buried in wide band noise. We present the results of how Fourier spectrum and least squares can be used to estimate the periodogram of time series, depending on whether the data are evenly spaced in time or not. The power spectra analysis of the stations from Hyderabad, India to Maitri, Antarctica shows significant variations (an increasing trend) in the Height component of the time series. The exact geodetic signals thus extracted after improved characterization of errors and noise resulted in improved accuracy of the estimated velocity vectors of the chosen GPS sites by a factor of 10-12, thereby enabling the Indian Plate Kinematics studies to be more reliable and consistent and also to reevaluate the existing Global Plate Motion Models.
DE: 1709 Geodesy
DE: 3220 Nonlinear dynamics
DE: 6944 Nonlinear phenomena
DE: 6974 Signal processing
DE: 8158 Plate motions--present and recent (3040)
SC: Geodesy [G]
MN: 2003 Fall Meeting