池上斷層沿線全新世河階的演育：斷層活動與河川沉積之交互作用

Abstract

花東縱谷中南段池上地區的階地位在池上沖積扇與池上斷層交會處。來自中央山脈的新武呂溪攜帶大量沉積物在池上地區形成了廣大的沖積扇，並在扇端處受到池上斷層逆衝截過，在斷層上盤可以觀察到一系列不同高度的階地。這與一般常見地形中，河流被抬升並下切後，將舊河道和氾濫平原廢棄遺留在高處所形成的河階不盡相同，而是有著更複雜的沉積與構造活動的交互作用。本研究透過分析階地的地形、組成和年代，欲了解池上地區階地的形成機制，釐清新武呂溪的河流沉積與池上斷層的構造活動是如何共同作用並影響階地地形的形成，同時探討階地地形的演育、沖積扇與河道的變遷、斷層長期滑移速率及其變化等。 在地形分析方面，本研究利用高解析度的5m-DEM進行室內階地判識與對比，並於野外進行確認；初步階地對比是利用階面與現生沖積扇面的比高進行，可分為T1 (17-20公尺) 到T10 (90公尺) 等十階。在階地組成方面，本研究進行詳細的野外調查，並在T1、T2、T4、T5、T7、T10等階地面進行七個槽溝的開挖與分析，重建階地沉積物組成剖面，包含其沉積環境及沉積物來源。本研究亦針對階地中的細粒沉積物進行沉積物物源判別，包括利用X光繞射法分析黏土礦物半定量組成、伊萊石結晶度，以及Itrax-XRF岩芯掃描儀所得之化學元素半定量組成及典型判別分析法(Canonical Discriminant Analysis)等；根據這些項目的分析結果將階地細粒沉積物樣本與中央山脈和海岸山脈的樣本交叉比對，以判斷沉積物的來源。階地組成綜合分析的結果顯示，階地底岩(利吉層)上覆的沉積物可大致可歸為兩類：主要礫石堆積(Primary gravel deposits)與二次堆積(Secondary deposits)。主要礫石堆積為變質岩質的礫石夾透鏡狀砂層，為源自中央山脈的辮狀河河道沉積物。二次堆積覆蓋在主要礫石堆積之上，組成多為礫石與泥質細粒沉積物。礫石岩性均為源於海岸山脈的砂岩，沉積環境包含河道沉積和土石流堆積。泥細粒沉積物之物源判別的結果顯示主要物源以海岸山脈利吉層之成分為主，但少部分階地(富南T5及大坡T4)有間歇性來自中央山脈的訊號，本研究解釋為這些二次堆積形成時階地仍在新武呂溪的影響範圍內。在年代分析方面，碳十四定年結果顯示富南T7階地之二次堆積的年代為5,750-6,880 Cal BP。以二次堆積年代來估計主要礫石堆積的最小年代，本研究推論研究地區之階地均為全新世以來的階地，最老的階地及其主要礫石堆積可能形成於大約七千多年前。本研究亦嘗試在大坡T2階地使用宇宙源核種定年法(Cosmogenic nuclides dating)，以期補足本研究區域主要礫石堆積難以取得碳十四定年樣本的限制。根據同一剖面不同深度的宇宙源10Be分布，利用卡方逆推模型(χ2 inversion model)擬合衰減曲線可得到理論計算之階地曝曬年代，但由於年輕階地的宇宙源10Be含量極少(< 1600 atoms/g)，測量誤差大，且露頭狀況不佳缺少近地表的含量控制，因此年代解釋上仍需小心。 本研究認為在池上斷層上盤由新武呂溪沖積而成的階地其形成與演育的影響因素包括有：斷層持續滑移造成上盤持續相對抬升，以及新武呂溪間歇性對海岸山脈前緣的側蝕並在斷層上盤區域大規模堆積，此兩因素共同作用形成斷層上盤一系列的階地。這些階地的分布同時也記錄了池上沖積扇的發育及新武呂溪的河道變遷歷史：根據主要礫石堆積的分布，本研究認為全新世以來新武呂溪主要影響範圍為沖積扇的北半部，極有可能河道在全新世早期流向為往北，近數百年內才轉往南流。 根據二次堆積的年代、斷層上盤相對抬升的距離以及下盤的沉積速率，本研究推估斷層約七千年來最大平均垂直滑移速率為1.4-1.7 cm/yr。與現今大地測量的結果(2.5-3.0 cm/yr)相比要來得慢許多，暗示著池上斷層在全新世以來的活動速率可能曾經有所改變。

Located in the middle-south Longitudinal Valley in eastern Taiwan, the toe of the Chihshang alluvial fan (mainly formed by Xinwulyu River) is cut and uplifted by the thrusting of the Chihshang Fault, which allowed several levels of terraces to form on the hanging wall of the fault. These sedimentation/tectonics context and geomorphic architecture are different from those for river terraces developed on the sides of river channels, where the terraces formation is mainly due to regional uplift (and/or base level drop) and river incision. In this study, through analyzing the morphology, composition and ages of different levels of terraces, we aim at better understanding the mechanism of terraces formation in the Chihshang area. According to our results, we discuss the interactions between Xinwulyu River sedimentation and tectonics of the Chihshang Fault, as well as the implications on the geomorphic evolution, the channel migration of the Xinwulyu River in the valley, the long-term slip rate of the Chihshang Fault and possible change during the past few thousand years. Based on geomorphic analysis from the high resolution 5m-DEM, terraces are distinguished into ten levels according to height difference relative to the present alluvial fan surface, from level T1 (17-19 m) to T10 (90 m). We also carried out field investigation and seven trenches analyses on six terraces surface (i.e., T1, T2, T4, T5, T7 and T10) for the purpose of understanding the sedimentation environments and source of sediments. Especially for the fine grain sediments we conducted source identification by several analyses, including XRD semi-quantitative analysis of clay minerals and measurement of Illite crystallinity, XRF semi-quantitative analysis of chemical elements by using Itrax core scanner and Canonical discriminant analysis of chemical element composition. Through comparing with sample from Central Range and Coastal Range we can identify the source of those fine grain materials. From the results of composition analyses above we found that the terrace sediments overlie the bedrock (Lichi Formation) can be divided into two parts: primary gravel deposits and uppermost secondary deposits. Primary gravel deposits are composed of metamorphic rock gravels and lenticular coarse sands layer derived from the rocks in the Central Range, confirming that the terraces are in principal formed by the Xinwulyu River. Secondary deposits cover above the primary gravel deposits for a few meters. They are composed of either sandstone gravels derived from the Coastal Range or fine-grain silt and clay. Sandstone gravel is interpreted to be fluvial deposits of the Coastal Range tributary rivers or landslide deposits. The results of source identification of fine-grain material show that most of them are composed of materials from Coastal Range (Lichi Formation). However, some terraces (Funan T5 and Tapo T4) show significant amount of the Central Range materials, implying that when secondary deposits were formed, the terraces were still under the influence of Xinwulyu River at a rather low level close to the valley altitude.The 14C dating results indicate the ages of the secondary deposits on Funan T7 terrace are around 5,750-6,880 cal BP. The ages of the secondary deposits provide the minimum ages control of primary gravel deposits. We conclude that all these terraces are Holocene terrace, and the terrace gravel of the oldest terraces might deposit around seven thousand years ago. We also try using cosmogenic nuclides dating method to estimate the age of primary gravel deposits, in which samples for 14C dating usually are difficult to fine in the study area. Exposure age of Tapo T2 profile can be conducted by the χ2 inversion modeling results, but since the 10Be concentrations of young terrace such as T2 are very small (lower than 1600 atoms/g) and the error of measurement is large, also lacking of surface control samples, this age and associated interpretation need to be cautious. We conclude that the terraces in the Chihshang area were formed under influences of continuous Chihshang Fault thrusting movement, episodic lateral erosion and sedimentation of Xinwulyu River, and local deposits from Coastal Range’s tributaries on the hanging wall of fault. The episodic lateral erosion and sedimentation are seemingly accompanied by the river channel migration on alluvial fan. From the uneven distribution of primary gravel deposits we argue that Xinwulyu River might flow to the north in the valley during most of the Holocene, and turn to flow southward in recent hundreds of years. To estimate the long-term slip rate of Chihshang Fault, we use the ages of secondary deposits of the terraces, and the available data of the sedimentation rate on footwall to calculate the maximum vertical slip rate. The result shows that the long-term vertical slip rate during the past seven thousand years is around 1.4-1.7 cm/yr, much lower than the present vertical slip rate from leveling measurement (2.5-3.0 cm/yr). These results imply that the slip rate of the Chihshang Fault might have change significantly during Holocene.