| 講演要旨(和文) | | 本研究では、MT法の時間領域でのデータ処理の開発を行った。従来のFFTを用いた周波数領域の解析ではなく、Z変換を用いたディジタルフィルタを適用することで、時間領域での解析が可能になった。また、今まで目視で行なっていた信号の存在区間の判別も、異なるサイトの磁場データを用いて周波数毎に相関を計算することで、自動で判別することが出来た。時間領域・周波数領域で解析した結果を比較すると、時間領域での解析の方が誤差は小さく、より高精度な地下の応答関数を求めることが出来た。 |
|
|
| | 講演要旨(英文) | | Fast Fourier Transform(FFT) has been widely used in data processing in magnetotelluric survey(MT). Although FFT assumes time series to be stationary, MT data is in general non-stationary since the source of MT is the abrupt electromagnetic fluctuation in the ionosphere. We think it may not be optimum to apply FFT method to obtain earth response for processing. We focus on a digital filter called pole on pedestal that extracts and amplifies the signal component of a specific frequency. We developed the data processing of MT data in time domain using this filter. In addition, it is important to select the segment included in the signal. We calculate cross correlations between filtered magnetic data at different sites, and select some segment in which the coherence among the sites has high values. In this way, we choose automatically the segment in the acquired time series. We apply the above-mentioned data processing in time domain to real MT survey data acquired at the Nankai trough. The range of the error-bars in our processing found much smaller than in the other conventional FFT method. We obtain more accurate response functions using novel data processing successfully. |
|
|
|
| |
|