利用合成孔徑干涉雷達與數值模擬分析2018-2021台灣西南部震間變形與構造演化機制

Abstract

台灣西南部目前正由隱沒增積岩體轉為碰撞造山，並於上新世於造山前緣沉積厚層泥岩。泥岩的力學性質造成了此處發生了顯著震間潛移的現象，並且造成了地表結構物的損壞。因此本研究希望透過地表變形的監測與數值方法來了解其變形行為與形成機制，提供未來高地表變形區域的防災與減災參考。 本研究第一部分使用小基線演算法分析 Sentinel-1 C-band雷達2018-2021升軌與降軌影像之時間序列，配合GPS連續站觀測資料，對InSAR的平均速度場做擬和校正，最後透過線性逆推將速度場從視衛星方向投影到東西方向與垂直方向。從結果發現在震間相當活躍的構造有：後甲里斷層、中洲斷層、小岡山斷層、右昌斷層與鳳山斷層，而麓山帶內的旗山斷層則受限於雷達反射不佳而無法觀測其震間行為。變形行為上，可以把西南台灣分為三個主要的構造區塊，分別為右昌斷層以北為震間相當活躍的褶皺逆衝帶；右昌斷層以南，則以鳳山斷層為界，以東之屏東平原無明顯構造活動，以西則為活躍的背斜-泥貫入體發育帶。 研究的第二部分使用有限元素法程式DynEarthSol建立二維模型，探討造成泥岩潛變的動力學機制。透過改變泥岩黏滯度、內摩擦角與內聚力等力學性質與地表侵蝕作用觀察其行為，並可以將泥岩變形區分為三種不同的動力學過程：對稱發育、不對稱發育與斷層發育。從結果觀察到，當泥岩之黏滯度足夠低時，則會發育較對稱的的背斜與泥貫入體；當泥岩的尖峰摩擦角足夠低時，則會發育以塑性變形主導較不對稱的背斜褶皺與小規模斷層；而地表作用的影響則反映了高地表侵蝕速率使對稱背斜更容易發育，其波長較長且單一；同時，當地表侵蝕率下降與黏滯度上升，則會在淺部產生覆瓦狀斷層。對應到西南台灣的地表變形行為，從變形波長與速度場資料觀之，高地表侵蝕率的非對稱背斜發育應為最佳解。

In southwestern Taiwan, the orogeny process is changing from trench subduction-related accretionary wedge to plate collision. Owing to the deposition of the thick layer of mudstone in the Pliocene time, the rheology of mudstone results in interseismic creeping. Also, the gradual deformation causes damage to the infrastructure. Thus, we would like to provide the ground deformation observation through the interferometric synthetic aperture radar (InSAR) method and explore the deformation mechanism with the finite element method, which can be applied to disaster mitigation in these fast deforming areas. In the first part of this study, I exploit the small baseline algorithm (SBAS) to analyze the ascending and descending tracks of Sentinel-1 images from 2018 to 2021. Next, I use the cGPS data to constrain the velocity field derived from SBAS-InSAR observation. Finally, the velocity field can be projected back to the east-west and up-dip direction from the line of sight direction. From the result, some structures are still active in the interseismic time, such as Hochiali fault, Chongchou fault, Hsiaokangshan fault, Yochang Fault, and Fengshan Fault, and some of them are located in the densely populated area. However, we cannot access the interseismic behavior of some faults due to the vegetation, such as Chishan fault. The deformation pattern in southwestern Taiwan can be simply divided into three different regimes by Yochang fault and Fengshan Fault. (1) The north block: it is the active fold-and-thrust belt. (2) the southeast block: it is relatively inactive, and is covered by the thick layer of sediment brought by the Kaoping river; (3) the southwest block: it is left-laterally moving with respect to the southeast block, and there are many active anticline-diapirs in this region. In the second part of this study, I design a 2D numerical model with finite element code, Dynearthsol, to explore the dynamic process of mudstone creeping. By exploring the mechanical property of mudstones, like the viscosity, internal friction angle and cohesion, and the surface diffusion coefficient, we can classify mudstone deformation into three different mechanisms: symmetrical growing, asymmetrical growing, and imbricate thrusting. When the viscosity of mudstone is low enough, the model is prone to grow symmetry diapir. Since the viscosity of mudstone increases, the low friction angle layer plays an important role in the formation of diapirs, and this type of diapir is more or less asymmetry. It is believed that the decreasing of friction angle is related to the high-pressure zone which is a common phenomenon in mudstone areas. On the other hand, models with high viscosity and high friction angle grow only one dominant thrust fault. In the surface process model, if the erosion rate is high, the removal of the shallow strata and deposition of sediment at the low-land area will promote the diapir to grow, while the low erosion model will generate the imbrication of the thrust fault. If we make the erosion rate twice as higher as the deposition rate, a single diapir is able to move closer to the ground surface and the propagation rate of the deformation front will be suppressed. Comparing the first and second parts of this study, we can simply conclude that the low friction and high viscosity model is a good solution to the on-land structure nowadays. Relatively, the low viscosity model could explain the off-shore diapir and the structure profile.