台灣花蓮北部地區的新構造與近期地殼變形運動

Abstract

台灣位於歐亞大陸板塊與菲律賓海板塊的交界帶，一個產生弧陸碰撞的地理位置；台灣因此造山隆起，造就了遍佈的活動斷層且頻繁的地震活動。原本逆衝到歐亞大陸南海板塊上的菲律賓海板塊，在花蓮北部地區轉為向北隱沒到歐亞大陸板塊之下，使得這個區域正好介於北邊向東南運動（宜蘭蘇澳至新城），及南邊向西北運動（海岸山脈）兩不同運動方向塊體中間之構造轉換帶上，因此較之其他地區有更複雜的地表活動構造。本區域主要之活動構造為米崙斷層及近期有明顯地表變動的北埔線型。1951 年花蓮發生規模7.3之大地震，地表沿著米崙斷層產生破裂，上盤相對於下盤有1.2公尺的垂直抬升及2公尺的左移動量；而北埔線型在近幾十年來，一直因地形的特徵被描述可能為一活動構造。 本研究藉由台灣四十公尺數值高程模型及GPS-RTK地形剖面的測量，獲得花蓮北部地區的地形資料，並利用構造地形學之概念進行地形之分析，判定構造可能分佈之位置，並推測其可能之地下形貌。為了更瞭解及探討花蓮地區近期的地殼變形運動，本研究佈置了密集的GPS觀測網，包括31個定期觀測點，並於2007年4月至2009年2月間完成5次測量，得到近2年來花蓮地區水平方向之地殼變形速度場，並利用速度場向量進而進行此區域的水平面應變分析。另外並利用D-InSAR技術來觀測此地區垂直之地表變形量，補足GPS短期觀測在高程上精度較差之缺憾。 由地形崖分佈位置的地形分析，可知道米崙斷層南延後方向可能由東北轉向東南而在花蓮南濱附近出海，但地形崖可能受到河流侵蝕而有向東偏移的現象，所以斷層位置也可能在地形崖之西邊。分析北埔地區北寬南窄，崖北低南高的地表約3公里長的隆起線型，本研究另稱其為北埔構造脊。綜合構造脊走向及當地應力方向，推測其為擠壓及逃脫構造架構下，壓縮平移（Transpression）斷層系統活動所造成的正型開花構造。 GPS及D-InSAR地殼變形觀測結果顯示，相對於澎湖白砂S01R，花蓮地區地殼移動以13-41 mm/yr向西北方移動，方位角介於295-360度之間，明顯是受到菲律賓海板塊西北聚合作用的影響。而從全球參考框架（ITRF2000）下來看GPS的速度向量，則可以觀察到花蓮地區地殼變形以0.5-16 mm/yr的速率向方位角89-271度方向移動。本研究推測此現象是由於海岸山脈在海底的地下向西北方向推擠而發展出之擠壓-脫曳之地體構造，花蓮北部地區塊體因擠壓而逃逸至東北方向相對較自由之空間。此外，由GPS及應變分析資料可明顯觀察到花蓮北部地區塊體除了有向東活動外，也呈明顯順時針旋轉的傾向。米崙斷層可能為因此機制而發展出之斷層，使塊體能更容易向上抬升及向東北方向被擠出，在此狀況下，米崙斷層為一後期形成之高角度以左移為主之活動斷層，與海岸山脈斷層平行而不相連。 在米崙斷層的活動上，GPS顯示有較明顯的左移運動，速率約為3-8 mm/yr，垂直抬升之變動量則較小，約為1 mm/yr。間震時期的慢活動速率可能代表米崙斷層為一鎖住之斷層，能量以地震事件釋放為主。此外，測量資料也顯示美崙台地內部也有明顯變形之現象，說明內部可能有其它次要活動構造的存在。

Eastern Taiwan has long been well-known as sitting at the collision suture between the Eurasian and Philippine Sea plates which is characterized by numerous earthquakes and several active faults. The Hualien City is located at the transition point at the northern end of the Longitudinal Valley, whereas the Philippine Sea plate westward thrusts upon the Eurasian plate in the south, while dives under the Eurasian plate from the latitude of Hualien northwards. At least two major active faults have been documented in this relatively complex area of transition zone, in addition to the Coastal Range fault on the northwestern side of the Coastal Range. The Milun fault, on which it ruptured during the 1951 M=7.3 Hualien earthquake, has been mapped running through the city of Hualien, showing coseismic slips of 1.2 and 2 m for its vertical and sinistral slip respectively during the 1951 earthquake. The Beipu fault or lineament, by contrast, was only revealed by its geomorphic feature. In this study, we used 40m-DTM and several high-resolution topographic profiles obtained from GPS-RTK measurement to characterize the geomorphic features related to the aforementioned two structures. Besides, we deployed repeated campaign measurements of a dense network of GPS stations. We carried out measurements for five times during April 2007 to February 2009, in order to obtain the horizontal GPS velocity and subsequent strain analysis. Furthermore, we carried out D-InSAR analysis to observe and monitor the vertical crustal deformation. Several pieces of the geomorphic evidence show that the Milun fault extends southward and probably turns southeastward into the sea. Nevertheless, it remains problematic because GPS velocities show that the surface trace of the Milun fault might lie farther west to the geomorphic scarp, which seems to be modified by river erosion that made the scarp retrieve to the east. As for the Beipu structure, base on its geomorphic features, we tend to interpret it as a positive flower structure, presumably resulted from left-lateral strike slip on the Beipu fault, although the GPS horizontal velocities and InSAR analysis indicated little tectonic rate occurred along the Beipu structure. The GPS results show that Hualien area moves northwestward at a rate of 13-41 mm/yr relative to the stable reference of the station S01R at Penghu, and the orientation is 295-360 degree. On the other hand, the same results under ITRF2000 reference framework illustrate a much intriguing deformation pattern, with a likely eastward extrusion of the Meilun Tableland and the Beipu area, north of the Coastal Range. The Milun fault shows a small sinistral slip at the rate about 3-8 mm/yr; however, our D-InSAR results show there is no significant vertical offset across the Milun fault in the studied period, with a rate of about 1 mm/yr. We thus interpret that the Milun fault probably is now locked, so that the stress were accumulated during the interseismic period, and it would release to cause earthquake and surface movement in the future. The strain analysis from the GPS measurements demonstrate that the northern half of the Hualien area shows a strong clockwise rotation, whereas the southeast part shows a moderate counter-clockwise rotation. We suspect this phenomena is a result of collision-extrusion tectonics that is initiated by the northwest compression of the Coastal Range under the sea east off the Haulien area to push the Meilun Tableland and the Beipu area not only upward but also eastward. The Milun fault may probably developed from this mechanism to let the block more easily to move to free space in the up and east directions. In this case, the Milun fault is more likely a high-angle sinistral fault that is subparallel to the Coastal Range fault, and these two faults might not necessarily connect to each other.