台灣東北部地區高山湖泊的演化過程與沉積作用之探討，以三星妹池為例

Abstract

東亞季風帶是全球季風系統中，一個相當重要的子系統。它影響了將近全球四分之一的人口。台灣地處東亞季風系統中，其特殊的地理位置提供了絕佳的條件，使其在時間尺度上研究西太平洋的季風以及颱風等天氣變化之紀錄。然而，從台灣各個湖泊中檢索到之全新世古氣候記錄中存在許多差異。這可能是因為過往的研究中缺乏了湖泊沉積學的觀察，限制了我們對控制湖泊沉積的機制的理解，因此在解釋湖泊記錄時引入了不確定性。 為了解存在於湖泊沉積中的古氣候記錄，必須建立一個可靠的年代模型去解釋隱藏在湖泊沉積物中所顯現出的意義。本研究旨在了解位於台灣宜蘭縣太平山區（24°30'N, 121°35'E, 海拔2095公尺）的小型濕地湖泊，三星妹池的湖泊演化過程，以及隱藏在湖泊沉積物中之意義。本研究透過鑽取湖泊岩心作為主要的研究材料。使用來自溼地中不同位置的多個沉積岩心，包括已經被草木覆蓋的樹林沼澤、緊鄰水邊之沼澤區和池子中，等各種不同的沉積環境，目的在於建立一個合理的湖泊形成與演化的過程。 本研究於湖泊沉積岩心中應用了多種不同的分析指標，包含:碳14、鉛210、銫137等定年分析，總有機碳和總氮量分析，生物指標分析，與X射線螢光光譜分析(XRF)。然而，來自濕地不同部分的岩心在經過多種的實驗方法中，發現了年代與沉積速率存在了令人困惑之差異。由多個放射性碳14定年測定的結果證實了這種倒轉並不是由於採樣之人為因素或是實驗分析偏差，而是真實反映了沉積序列的逆轉。可能解釋這種年代反轉的假設包括人為採樣疏失，浮島的翻轉，山崩等。臺灣其他山區湖泊的案例顯示，由水生植物、苔蘚植物和沉積物混合而成的浮島在颱風期間可以移動數百米。如果浮島翻轉，可能會導致反轉之年代序列。建立的穩健年代模型和湖泊演替模型將對後續的氣候研究有所幫助。

Monsoon rainfall in East Asia is a subsystem of the global monsoon system and affects nearly one-quarter of the world's population. Taiwan, an island in the subtropical East Asian monsoon system, provides the unique opportunity to study monsoon and typhoon variability over the western subtropical Pacific on both historical and geological time-scales. However, the Holocene paleoclimate records retrieved from various lakes in Taiwan have shown discrepancies in their recorded climate variability. This could partly be caused by the lack of modern sedimentological observations, limiting our understanding of the mechanisms controlling lake sedimentation and thus introducing uncertainty in the interpretation of lake records. The aim of this study is to understand the complex processes that led to the formation of a small wetland, Sanxingmei Pond, which is located in the Taiping Mountains, Yilan County, NE Taiwan. (24°30' N, 121°35' E, 2095 meters above sea level). This study shows how lake succession can explain differences in lithology between nearby coring sites. Moreover, the study emphasizes the need for robust age models to understand paleoclimate records archived in lacustrine sediments. Convoluted sedimentary layers caused by floating islands are hypothesized as the preferred explanation for significant age reversals observed in some cores. This study uses several sediment cores from different parts of the wetland, including wooded swamps, fens, and open ponds, to construct a robust model for lake formation and succession. A multi-proxy approach was applied in these cores, including 14C, 210Pb, 137Cs, bulk organic analysis (TOC, TN), and X-ray fluorescence (XRF) data. Cores from different parts of the wetland display puzzling differences in maximum ages and sedimentation rates. Multiple radiocarbon dates demonstrate that observed age reversals are not a sampling or analytical artifact, but actually reflect an inversion of the sediment sequence. Possible hypotheses explaining age reversals include typhoons, landslides, and overturned floating fens. Other mountain lakes in Taiwan provide examples of these floating islands where a mixture of aquatic plants, bryophytes, helophytes, and sediment can move hundreds of meters during typhoons. If a floating island is blown by typhoon winds it may become overturned, resulting in an inverse age sequence. An improved understanding of age model and lake succession will be helpful for future paleoclimatic research utilizing lacustrine records.