| 摘要: | 傳統地震災害潛勢評估皆應用衰減模式作為評估的依據。而局部場址特性對 PGA 有相當大的影響，因此建立PGA 衰減模式時必須考慮測站的場址特性，以 增進評估的可靠度。由於目前台灣測站的場址特性缺乏可靠的分類參考資料，因 此，本研究之目的將利用實測地震記錄分析測站觀測值對參考衰減模式的系統誤 差，建立台灣地區測站場址特性的PGA 修正函數。研究方法主要分為二個範疇， 首先採用Chien(2001) 使用59 個淺源地震分析所得出的衰減模式做為本研究之 參考之衰減模式。由於其分析時並未對測站場址特性加以分類，因此所得結果為 一般性場址適用。利用參考衰減模式得到各測站預測初步之PGA 值，此為第一 階段的計算結果。第二階段分析時則將其對各測站實際觀測值的系統偏差量作校 正分析。分析同時發現各測站之誤差殘值（NR）與對應之觀測值皆呈半對數之 關係，另外也與地震規模及震源距離之關係做比較，分析結果呈現此誤差殘值與 兩者之關聯性不高。因此，顯示對於系統誤差量的修正以測站場址特性之影響為 主，建立各測站場址特性的修正函數，此為第二階段回歸分析之目的。將各測站 由參考衰減模式預估的PGA 值與實測的PGA 值作分析，其間之關係式為 i att i Obs Y C C Y ) ln( ) ln( 1 0 × + = ，並由回歸分析可得到各測站的修正係數C0、C1 值。 如此即可修正測站場址特性所造成之誤差，得到可靠的地動加速度極值。然每個 地震之發生均有其獨特之震源特性，仍會造成預測之不準度。因此，當地震發生 時，PGA 值藉由二階段計算得到後，若能結合速報系統的實測值，將可更準確地 推估台灣地區地震發生後之地表PGA 分怖，作為早期地震災害潛勢評估。 整體而言，經由上述所採用的二階段計算所得之預測結果與實際觀測的結果 已有較佳之相似性，若加入速報系統參考站資料，進行震源特性修正，則可得更 為接近觀測的PGA 值分佈。由結果顯示，加入速報系統紀錄對各測站於二階段 預測結果作震源特性之修正，可降低預測的標準偏差值，因此可做為早期地震災 害潛勢評估之重要依據；同時在大地震發生時，準確之PGA 分佈圖之預估，將 可協助救災指揮系統作緊急應變之參考，運用得宜應可達到減災之目的。 158 As a result, we can better understand the attenuation characteristics in Taiwan and can reliably predict the peak amplitude for strong ground motion when an earthquake occurs. One of the main factors that affect the attenuation relationship is the existence of site effects. It is believed that the systematic bias mainly comes form the site effects. So, it is necessary to consider the site effects in attenuation relationship of PGA, and the result of predicted in seismic hazard will make best. Generally, site effect is one of the important factors for predicting ground motion. The TSMIP sites still do not be good classified. So, in this paper, we used two step to predicted the best result in seismic hazard. The purpose of this paper is to investigate the classification criterion of local site effect and its influence in the seismic hazard. First step, we got the first PGA of site by attenuation relationship of PGA from Chien(2001), which used more than 3000 seismic records from 59 earthquake events are used to study the attenuation relationship of PGA. The selection criteria are ML > 5.0 and focal depth < 35 km. All events are well recorded by the TREIRS system and TSMIP system. These events occurred in 1995 to 1999, are widely felt in Taiwan. These earthquakes are relocated by this study by using both the TREIRS with TSMIP systems records. The standard deviation of the residuals between the observed and predicted values (i.e. ln(observed)-ln(att. form) ) are 0.7654. Second step, the TSMIP and TREIRS site correction C0、C1, can be simplified by fallowing the law : ln(PGAobs) S=C0+C1*ln(Yatt)S, where (PGAobs) S is the observed PGA value, (Yatt)S is the predicted PGA value obtained by the attenuation relationship laws calacted. The selection criteria are ML > 4.0 and focal depth < 35 km. All data of site are well recorded by the TREIRS system and TSMIP system. The results agree reasonably well with the surface geology from published maps. Generally, the attenuation relationship and site corrections represent a statistically average effect. But, every event possesses its own characteristics, such as the source. Therefore, we used TREIRS system datas from 75 telemetered strong-motion stations to estimated the peak ground motions PDS at TSMIP sites: ) , , ( 1 0 D D R f y PD S s S S × = ,where S S y is the predicted PGA value obtained by the attenuation relationship, R is the ratio by observed value at the TREIRS system station and the predicted value by the attenuation relationship with site correction at the TREIRS system station, D0、D1 are coefficient to correct the amplification by TSMIP system station. With the attenuation relationship and the site corrections data base, we have embarked on a practical but yet important problem of predicting the PGA values as soon as a large earthquake happens in Taiwan. The only input this extreme values prediction calculation system needs is the TREIRS system solution. We expect that the predicted PGA maps will be useful in earthquake emergency response operations. |
