2022年1月15日，東加的Hunga Tonga-Hunga Ha’apai火山（HTHH volcano）發生大規模噴發並引發跨洋型態之氣象海嘯（meteotsunami）。臺灣潮位站記錄中觀測到大氣波掃過臺灣後，有著持續數小時振幅逐漸放大之特殊現象，因此本研究為詳細分析臺灣地區記錄到火山引發之大氣波，並解釋造成氣象海嘯增幅現象之原因。 臺灣的中央氣象署（CWA）地面觀測站觀測到以聲重力波（Acoustic-gravity waves， AGWs）形式且包含多種型態之大氣波動，包括了最先抵達且振幅最顯著之大氣蘭姆波（atmospheric Lamb wave），以及陸續抵達多種不同相位波速之AGWs。根據觀測資料對氣壓波的不同種類與特性做詳細分析，依照壓力移動速度、氣壓差、週期將氣壓波分類為：速度約為308.25 m/s且氣壓差約為2 hPa之大氣蘭姆波、以及速度約為287 m/s ~ 227 m/s且氣壓差約為0.5 hPa之4個AGWs。根據臺灣地區之潮位站資料，大氣蘭姆波引發之海嘯初期振幅大約為2 ~ 5公分，但其後4小時之內振幅不減反增，最終增長至10至20公分，放大程度約5倍。但依照模擬結果表明，單純大氣蘭姆波並無法造成這種振幅持續放大之現象。 本研究根據實際大氣觀測資料，分別製作了擬合觀測資料之大氣蘭姆波以及AGWs之線性N形氣壓波模型，並假設氣壓波之速度以等速方式、氣壓波之振幅以臺灣觀測之氣壓值，從火山源傳播出去，傳播過程中振幅保持不變，僅考慮由於波速差異所導致之頻散效應。使用COMCOT海嘯模式來模擬火山爆發後之氣壓傳播與海洋波動。針對事件，模擬了六種不同壓力模型之情境，分別為大氣蘭姆波結合AGWs之整體氣壓波、僅大氣蘭姆波、以及4個各別AGWs。海嘯模擬結果顯示，整體氣壓波之模擬成功再現了潮位站觀測資料上於數小時之內增幅之現象。 進一步探討其中每個氣壓波對於海水變動之貢獻，其顯示大氣蘭姆波引起之海水變動主導潮位站前半段的記錄，而AGWs引起之海水變動主導潮位站後半段之記錄。振幅持續放大之現象與波速較慢之AGWs相關，特別是速度約為227 m/s之AGW，儘管其氣壓差小於大氣蘭姆波。根據海嘯傳遞模擬顯示，由於227 m/s之AGW波速與海洋重力波波速相當，於西北太平洋深海區傳遞時引發普曼共振，使海嘯振幅增大。因此，臺灣潮位站記錄之氣象海嘯增幅現象，為波速較慢之AGWs引起，大氣蘭姆波對於增幅之貢獻有限。 本研究亦針對HTHH引發之典型火山型海嘯進行模擬，使用點源型海嘯源作為初始條件，定義典型火山型海嘯抵達臺灣之時間，並搭配大氣成因之氣象海嘯模擬，為2022年東加HTHH火山海嘯事件做完整解釋。;On January 15， 2022， the Hunga Tonga-Hunga Ha’apai (HTHH) volcano in Tonga erupted violently， generating a transoceanic meteotsunami. Atmospheric waves were observed to pass through Taiwan’s tide gauges， exhibiting a unique phenomenon where the amplitude gradually increased over several hours. This study aims to analyze in detail the atmospheric waves recorded in Taiwan caused by the volcanic eruption and to explain the reasons for the amplification of the meteotsunami. Surface observation stations operated by the Central Weather Administration (CWA) in Taiwan recorded atmospheric waves in the form of acoustic-gravity waves (AGWs)， which exhibited multiple types of waveforms. Among these， the atmospheric Lamb wave was the earliest to arrive and showed the most prominent amplitude. Various AGWs with different phase velocities followed it. Based on observational data， the pressure waves were classified according to their propagation speed， pressure difference， and period. The atmospheric Lamb wave propagated at approximately 308 m/s and was associated with a pressure difference of about 2 hPa. In contrast， four AGWs were identified with propagation speeds ranging from approximately 287 m/s to 227 m/s and pressure differences of around 0.5 hPa. Based on tide gauge data from Taiwan， the initial tsunami wave height triggered by the atmospheric Lamb wave was approximately 2 to 5 centimeters. However， during the following four hours， the amplitude increased rather than decreased and eventually reached 10 to 20 centimeters. It represents an amplification by a factor of about five. According to simulation results， such sustained amplification cannot be explained solely by the atmospheric Lamb wave. In this study， linear N-shaped pressure wave models representing atmospheric Lamb waves and acoustic-gravity waves (AGWs) were constructed based on actual atmospheric observations. It is assumed that the pressure wave propagates from the volcanic source with a constant speed， and its amplitude is based on the atmospheric pressure values observed in Taiwan. During the propagation process， the amplitude remains unchanged， with only the dispersion effect caused by the differences in wave speeds being considered. These models were used as input for the COMCOT tsunami model to simulate the propagation of atmospheric pressure disturbances and corresponding oceanic responses following the volcanic eruption. Six different pressure scenarios were examined， including a combined Lamb wave and AGWs model， a Lamb wave-only model， and four individual AGWs cases. The tsunami simulation results show that the combined Lamb wave and AGWs scenario successfully reproduced the observed sea level rise within a few hours at tide gauge stations. Further analysis of the contribution of each pressure wave to sea level variations reveals that sea level disturbances caused by the atmospheric Lamb wave dominate the early portion of the tide gauge records. In contrast， those induced by AGWs dominate the latter portion. The observed sustained amplification of tsunami amplitude is associated with the slower-propagating AGWs， particularly the one traveling at approximately 227 m/s， despite its pressure anomaly being smaller than that of the Lamb wave. According to tsunami propagation simulations， the AGW traveling at 227 m/s induces Proudman resonance in the deep waters of the northwestern Pacific Ocean， as its phase speed is comparable to that of oceanic gravity waves， leading to an amplification of the tsunami amplitude. The phenomenon of the meteotsunami amplification observed at Taiwan′s tide gauge stations is primarily caused by the slower-moving AGWs， with limited contribution from atmospheric Lamb waves to the amplification. This study also simulates a typical volcanic tsunami generated by the HTHH eruption， using a unit-source tsunami model as the initial condition to define the arrival time of the volcanic tsunami in Taiwan. Combined with the meteotsunami simulation induced by atmospheric pressure waves， this provides a comprehensive explanation of the 2022 HTHH volcanic tsunami event.
