地表淺層構造的場址特性，根據地質構造的特性不同，具有很大的異相性。若僅依照PGA或PGV的放大倍率，或是利用GMPE評估經地表場址放大後的震波特性，往往過於簡化，不足以應付強地動模擬時在高頻訊號的解析度。因此，本研究將利用氣象局新一代的井下地震觀測網資料，分析在地表站至井下站間的淺層場址轉換函數，提供非線性的場址效應，藉以提高強地動模擬對地表波形的解析度。 本研究經過第一年的測試：獲得各測站在地表至井下100至300公尺深不等的放大傅立葉頻譜，分析比較非線性解迴旋方法(Projected Landweber deconvolution)及線性水平法(water level method)之穩定性。今年將會使用線性水平法計算各測站於時間域之轉換函數，並與103年：”氣象局井下地震監測站資料分析與應用Ⅱ” 所評估之速度構造以及衰減特性進行比較以及分析。 ;According to the characteristic of geological structures, it shows large discrepancies of near-surface site effect. However, PGA, PGV amplification factors and Ground Motion Prediction Equations (GMPEs) oversimplified the estimation of ground motions, and they are insufficient for getting a better result on waveform simulation at high frequency. In this study, we estimated the transfer functions of near-surface structure by using CWB next generation seismic network datasets. The non-linear site characteristic will be obtained and applied to increase the resolution of waveform simulation as well. In this study, we will apply water level method in borehole seismograph dataset and compare with our previous project results in 2014 after the analyze and stability test of Projected Landweber deconvolution (PLD) method and water level method last year.
