氣候變遷透過因環境背景而異的機制，重塑海岸地形及其所形塑的海岸線，然而利用可比較之方法，對北極冰凍圈系統與熱帶三角洲系統進行系統性比較的研究仍相當稀少。本論文建立一套整合遙測與 hydrodynamic modelling 的框架，用以評估兩個對比環境中的海岸線變化（shoreline change）——越南湄公河三角洲（Mekong River Delta， MRD）河口區，以及挪威 Svalbard 的 Kaffiøyra 海岸—冰河區——時間跨度近四十年。本研究以海岸線變化為核心，並將支配其變化的沉積物供應過程（三角洲的河川輸沙，以及 Kaffiøyra 的冰河末端與 outwash plain 動態）作為輔助分析項目。 於 MRD，本研究使用九幅 Landsat 影像（1987–2025）及一套經校正的 Delft3D 模式（NSE 0.66–0.92），量化 223 km 海岸線的變化。Weighted Linear Regression 變化率介於 −32.91 至 +94.22 m/yr 之間，顯示出強烈的幾何控制：凸出的岬角使波浪能量聚焦而侵蝕，凹入的灣澳則使能量發散而堆積。三角洲於約 2005 年後由淨堆積轉為淨侵蝕，最符合人為沉積物匱乏（sediment starvation）之解釋——此歸因係基於多重佐證（converging lines of evidence）——因上游水壩使懸浮沉積物輸送量由 143 Mt/yr（2007）降至 47 Mt/yr（2020）。在 IPCC 海平面上升情境（SSP1-1.9 至 SSP5-8.5）下，上游測站的水位上升呈現非線性放大，於 My Thuan 站在所有情境下均超過海域邊界驅動量的兩倍。 於 Kaffiøyra，本研究使用三十幅 Landsat 影像（1985–2023），同步追蹤八條冰河末端、31 km 海岸線及三處 outwash plain。所有冰河均呈系統性退縮（R² = 0.83–0.96），並與氣溫呈顯著負相關（r = −0.55 至 −0.69，p < 0.01）。Tidewater 冰河平均退縮速率為 41.84 m/yr，超過 land-terminating 冰河（19.52 m/yr）的兩倍，反映了源自 Atlantic Water 入流的額外海洋驅動力。海岸線變化呈現與 wave-refraction 控制一致的空間型態（與 MRD 相仿，惟尚未經模式證實），其中 Aavatsmarkbreen tidewater 冰河末端附近出現極端堆積（+13.02 m/yr），顯示直接的 paraglacial 沉積物輸送。Land-terminating 冰河雖損失 34–69% 的末端面積，卻未偵測到明顯的海岸堆積增強；outwash plain 變化亦甚微（−6.4% 至 +2.7%），顯示其在局部谷地幾何與永凍層（permafrost）控制下，與冰河退縮呈現脫鉤（decoupling）。 比較綜整得出三項主要發現。其一，全球暖化因沉積物路徑的完整性不同，產生相反的沉積物走向：三角洲因人為中斷河川輸沙而流失沉積物，北極海岸則因冰凍圈過程完整而獲得沉積物。其二，wave refraction 在兩個系統中皆依海岸線幾何組織其空間變化型態——於 MRD 已獲因果證實，於 Kaffiøyra 則有觀測支持——而沉積物供應量則決定此幾何脆弱性轉化為可量測海岸線位移的速率。其三，連通性（connectivity，即沉積物路徑將氣候驅動連結至海岸的效率）為海岸脆弱性的關鍵決定因素，且無法僅由氣候驅動推知。因此，海岸脆弱性評估須如同關注氣候驅動一般，同等關注沉積物級聯（sediment cascade).;Climate change reshapes coastal landforms through mechanisms that depend on environmental context， yet systematic comparisons between Arctic cryospheric and tropical deltaic systems using comparable methods remain rare. This dissertation develops an integrated remote sensing and hydrodynamic modelling framework for assessing shoreline change in two contrasting environments — the Mekong River Delta (MRD)， Vietnam， and the Kaffiøyra coastal-glacial region， Svalbard， Norway — over nearly four decades. Shoreline change is the central focus; the sediment-supply processes that govern it — fluvial delivery in the delta， and glacier and outwash-plain dynamics at Kaffiøyra — are analysed as supporting components. For the MRD， nine Landsat scenes (1987–2025) and a calibrated Delft3D model (NSE 0.66–0.92) were used to quantify shoreline change across 223 km of coastline. Weighted Linear Regression rates ranged from −32.91 to +94.22 m/yr， revealing strong geometric control: convex headlands focus wave energy and erode， while concave embayments defocus energy and accrete. The delta′s shift from net accretion to net erosion after approximately 2005 is most consistent with anthropogenic sediment starvation — attributed on converging lines of evidence — as upstream dams cut suspended sediment delivery from 143 Mt/yr (2007) to 47 Mt/yr (2020). Under IPCC sea-level rise scenarios (SSP1-1.9 to SSP5-8.5)， projected water-level rise at upstream stations amplifies non-linearly， exceeding double the ocean boundary forcing at My Thuan under all pathways. For Kaffiøyra， thirty Landsat scenes (1985–2023) tracked eight glacier termini， 31 km of shoreline， and three outwash plains simultaneously. All glaciers retreated systematically (R² = 0.83–0.96) with significant negative temperature correlation (r = −0.55 to −0.69， p < 0.01). Tidewater glaciers retreated at 41.84 m/yr — more than twice the land-terminating rate (19.52 m/yr) — reflecting additional marine forcing from Atlantic Water inflow. Shoreline change showed a spatial pattern consistent with wave-refraction control (comparable to the MRD but not yet model-confirmed)， with extreme accretion near the Aavatsmarkbreen tidewater terminus (+13.02 m/yr) indicating direct paraglacial sediment delivery. Land-terminating glaciers， despite losing 34–69% of terminus area， produced no detectable coastal enhancement， and outwash plains changed minimally (−6.4% to +2.7%)， indicating decoupling from glacier retreat under local valley-geometry and permafrost control. Comparative synthesis yields three findings. First， global warming produces opposite sediment trajectories depending on pathway integrity: the delta loses sediment through anthropogenic interruption of fluvial supply， while the Arctic coast gains it through intact cryospheric processes. Second， wave refraction organises the spatial pattern of change in both systems — causally demonstrated for the MRD， observationally supported for Kaffiøyra — while sediment availability sets the rate at which that geometric vulnerability becomes measurable shoreline movement. Third， connectivity — how efficiently sediment pathways link climate forcing to the coast — is the critical determinant of coastal vulnerability， and cannot be inferred from climate forcing alone. Assessing coastal vulnerability therefore requires reading the sediment cascade as attentively as the climate forcing itself.
