1867 年基隆海嘯為臺灣少數具有多源歷史文獻佐證之儀器觀測前海嘯事件。史料記載金山與基隆一帶曾出現強烈震動、海水退去、海床裸露及後續回湧災害，頭城相關族譜與寺廟遷建紀錄亦顯示其影響可能延伸至蘭陽平原北側。然而，由於缺乏地震儀與潮位資料，其海嘯源位置、生成機制及沿岸水位分布仍具高度不確定性。 本研究整合歷史文獻、海嘯沉積物研究與頭城相關事證，選定金山、基隆及頭城為主要歷史約束點，並以金山 6–7 m、基隆 4–6 m 及頭城 1–2 m 作為工作性高水位相容性區間。研究採用 COMCOT 建立標準化單元海嘯源資料庫，結合影響強度分析法（Impact Intensity Analysis， IIA）與反餘弦相似比對指標（Inverse-Cosine Similarity Index， ICS），分別評估候選源之多測站水位影響能力及歷史水位比例相容性，並進一步檢驗第一類地震情境、高滑移地震、局部等效水體擾動及複合式海嘯情境。 IIA×ICS 綜合分析顯示，臺灣東北近海、基隆外海至棉花峽谷西南側一帶為相對較具歷史與水動力相容性之候選源區。第一類地震情境約 Mw 7 地震情境於三站所產生之水位明顯低於歷史相容性區間；高滑移 C1–C3 情境雖可提高金山與基隆水位，並維持頭城相對較低之空間特徵，但仍無法單獨同時符合三站歷史條件。隨地震背景增強，三站剩餘水位比例亦隨之改變，至 C3 情境時已難辨識出明顯之局部候選源集中區。 進一步於高相容性區配置 36 組局部等效水體擾動源後，在第一類地震情境背景下，U45、U46、U51 與 U52 四組案例可使金山、基隆與頭城之最大正水位同時落入工作性歷史區間，且水位時序呈現先退水後增水之特徵，與史料描述具有一定相容性。完整波形疊加結果則顯示，不同來源之最大正水位不能直接相加，其複合反應受到到達時間與相位關係控制；在 C1 與 C2 較強地震背景下，相位干涉的重要性增加，反而未找到可同時穩定符合三站歷史水位之單一局部來源。 綜合而言，本研究結果顯示，單一第一類地震情境源不足以解釋 1867 年基隆海嘯，而臺灣東北近海之局部水體擾動在三站水位量級、相對分布及先退後湧特徵上具有較高之水動力相容性。較強的地震背景亦不必然提高複合情境之歷史相容性。惟本研究主要評估候選源區與生成機制之相容程度，尚不足以確認實際海嘯源或地震觸發海底山崩、火山活動等特定生成機制。;The 1867 Keelung tsunami is one of Taiwan’s few pre-instrumental tsunami events supported by multiple historical records. Accounts from Jinshan and Keelung describe strong shaking， sea retreat， exposed seabed， and subsequent return flow， while records from Toucheng suggest that the affected area may have extended to the northern Lanyang Plain. However， the tsunami source and generation mechanism remain uncertain due to the lack of instrumental observations. This study integrates historical records， tsunami-deposit studies， and evidence from Toucheng， using Jinshan， Keelung， and Toucheng as historical constraint sites with working water-level ranges of 6–7 m， 4–6 m， and 1–2 m， respectively. A standardized COMCOT unitsource database was combined with Impact Intensity Analysis (IIA) and the Inverse-Cosine Similarity Index (ICS) to evaluate both the multi-station impact and historical water-level compatibility of candidate sources. Earthquake， local equivalent water-body disturbance， and composite tsunami scenarios were further examined. The combined IIA×ICS results identify the offshore region of northeastern Taiwan， from offshore Keelung to the southwestern Mien-Hua Canyon， as the main high-compatibility candidate area. Conventional Mw ~7 earthquake scenarios produce water levels far below the historical ranges， while high-slip C1–C3 scenarios increase the response at Jinshan and Keelung but still fail to satisfy all three sites simultaneously. Among 36 local disturbance sources， U45， U46， U51， and U52 reproduce the working water-level ranges at all three sites under the conventional earthquake background and also show an initial sea retreat followed by a positive surge， broadly consistent with historical descriptions. Waveform superposition further shows that source contributions cannot be evaluated by simply adding their individual maximum water levels， because arrival time and phase strongly influence the composite response. Under the stronger C1 and C2 earthquake backgrounds， phase interference becomes more important， and no single local source consistently satisfies all three historical constraints. Overall， a localized water-body disturbance offshore northeastern Taiwan shows greater hydrodynamic compatibility with the observed water-level magnitudes， relative spatial distribution， and retreat–return sequence than a single conventional earthquake source. Nevertheless， the results indicate only source-region and mechanism compatibility and do not identify the actual tsunami source or confirm a specific triggering mechanism.
