地球物理探勘多使用反射震測法瞭解地下構造形貌，然而震測剖面展示的是震波走時，需用速度加以轉換成地下深度。本研究使用井下與震測資料以及V0-k方法，建立台灣海峽及西部平原的沈積層速度構造。首先，假設岩層之波傳速度隨著深度加深呈線性遞增的趨勢（即V0-k方法，V0表岩層在海床或地表之震波速度，k表速度隨深度遞增的幅度），接著對每筆資料作速度與深度的迴歸分析，而得到V0（截距）、k（斜率）值，網格化這些值可得到V0與k的空間分佈。V0值的側向變化可反映表層沈積層特性：盆地內的V0值較低；而在西部麓山帶的V0值較高。k值變化則反映側向岩性差異與構造分佈，在難以壓縮之基盤高區具有異常高的k值。將V0、k和深度經過運算得到三維速度分佈，發現垂向速度變化主要受控於沈積物的深埋壓密作用；側向速度變化主要受控於斷陷作用與抬升作用；在垂直剖面的等速度曲線可反映地下構造幾何形貌。最後，利用井下鑽遇的漸新世分離不整合面深度，及其於震測剖面上顯示的時間轉換成深度，測試此三維速度構造之準確度。轉換深度的平均誤差百分比為4.78 %，顯示三維速度構造合理可信。 The method of reflection seismic imaging displays subsurface structures in time rather than in depth. One therefore needs to know the 3-D distribution of subsurface velocity in order to convert seismic time into depth and obtain 3-D subsurface depth structures. I use borehole and seismic data to compute a 3-D sediment velocity structure in the Taiwan Strait and the coastal plain in west Taiwan by using V0-k method. The V0-k method assumes that velocity increases with increasing depth in a linear form, in which V0 is the initial velocity at the seabed or on the ground surface, and k represents the rate of increase of the velocity with increasing depth. Fitting velocity with depth to a linear form for each data set obtains V0 (intercept) and k (slope) pairs, and gridding these values obtains the spatial distribution of V0 and k. The lateral variation of V0 reflects sediment characteristics with lower values in the basin and higher values in the western foothills. The variation of k correlates to the lithology difference and lateral structural variation. The value of k is exceptionally high in the areas of hardly compactible basement highs. Parameters of V0 and k were gridded in a 3-D depth volume. Comparing the velocity depth structures to seismic images, one finds that (1) the vertical velocity variation is mainly controlled by burial compaction of sediments; (2) the lateral velocity variation is caused by stratal offset resulted from major normal or reverse faulting; and (3) the isovelocity curves shown on cross-sections may reflect the geometry of the basement structure. Finally, I use the drilled depths of the Oligocene breakup unconformity (generally < 4 km in depth) and its corresponding depths as converted from seismic data and using the proposed velocity model to test the validity of the 3-D velocity model. The comparison yields a value of 4.78% of mean percentage error between drilled and predicted depths, indicating that the proposed velocity model predicts subsurface velocity reasonably well down to, at least, 4 km in depth.
