颱風通過時，海洋響應並不侷限於海表冷卻，而可能透過風應力、垂向混合、近慣 性運動與地形調制，進一步影響次表層溫鹽結構與密度界面變化。本研究以 2025 年 9 月 樺加沙颱風為個案，針對南海北部、呂宋海峽與巴士海峽周邊海域，探討強颱風對上 層至次表層海洋結構及內波型擾動之影響。研究中以 HYCOM 資料作為背景場與外部參 考，並使用 sbPOM 建立區域三維斜壓海洋模式，結合 ERA5 10 m 風場換算之海表風應 力作為大氣動量強迫，分析颱風通過前、中、後不同深度之溫度場、密度場與海表面高 度變化。 結果顯示，樺加沙颱風通過期間，0 m 溫度場主要呈現海表冷卻與冷尾跡，反映風 生混合、冷水上抬與海表動量輸入對表層水團的直接影響。相較之下，50 m 溫度場開 始出現較明顯的冷暖斑塊與帶狀結構，表示颱風擾動已向混合層下方與上部溫躍層傳 遞。100 m 與 150 m 溫度場則呈現較清楚的東西向溫度差異、冷暖交替及沿呂宋海峽至 南海北部方向延伸的條帶狀訊號，顯示次表層結構可能受到內潮、近慣性響應與地形導 引共同調制。HYCOM 與 sbPOM 比對結果指出，兩者在大尺度冷暖分布與颱風後次表 層異常上具有一致性；其中，HYCOM 適合作為事件尺度背景參考，sbPOM 則可補充 較細緻的局部波列、地形控制與斜壓結構變化。 0–50 m 密度場進一步顯示，颱風主要影響期之密度變化並非僅侷限於 0–10 m 表 層，而是在 20–50 m 亦出現等密度線彎曲、密度梯度壓縮、低密度水舌與高低密度水團 交錯等現象。特別是在 40–50 m，密度界面於颱風通過期間變形最為明顯，且在颱風離 開初期仍保留殘留梯度，顯示內部海洋響應具有滯後性。由於固定深度平面圖仍難以完 全區分內波、水平平流、近慣性響應與地形效應，本文採保守判讀，將其稱為內波型溫 度與密度擾動。 綜合 0–150 m 溫度場與 0–50 m 密度場結果，樺加沙颱風不僅造成海表冷卻，也使 上層至次表層熱鹽結構發生重組。100–150 m 溫度場中的冷暖交替，結合 20–50 m 密度 場中的等密度線彎曲、梯度壓縮及殘留異常，顯示颱風能量可能透過風生混合、近慣性 運動、背景內潮與地形調制向內部海洋傳遞。本文整合水平溫度與密度場、Hovmöller 圖、頻譜分析及剖面等值線變形結果，認為 樺加沙 期間研究區出現具有內波特徵的溫 度與密度擾動；然而，仍不宜將所有條帶狀結構直接視為確定內波，其確切機制仍需結 合流速場、相位關係與現場觀測資料進一步驗證。;Typhoon-induced ocean responses are not limited to sea surface cooling， but may further affect subsurface thermohaline structures and density interfaces through wind stress， vertical mixing， near-inertial motions， and topographic modulation. This study investigates the impacts of Typhoon RAGASA on the upper-ocean to subsurface thermal and density structures， with particular attention to internal-wave-like disturbances over the northern South China Sea， the Luzon Strait， and the Bashi Channel. HYCOM data were used as the background field and external reference， while a regional three-dimensional baroclinic ocean model was established using sbPOM. Surface wind stress was derived from ERA5 10 m wind fields and applied as atmospheric momentum forcing. Temperature fields， density fields， and sea surface height were analyzed at different depths before， during， and after the passage of the typhoon. The results show that during the passage of Typhoon RAGASA in September 2025， the 0 m temperature field was primarily characterized by sea surface cooling and a cold wake， reflecting the direct influence of wind-induced mixing， upward displacement of colder subsurface water， and momentum input on surface water masses. In contrast， the 50 m temperature field began to exhibit more pronounced cold and warm patches as well as banded structures， indicating that the typhoon-induced disturbance had extended downward below the surface layer and into the upper thermocline. At 100 m and 150 m， clearer east–west thermal contrasts， alternating cold and warm anomalies， and banded signals extending from the Luzon Strait toward the northern South China Sea were observed. These features suggest that the subsurface thermal structure may have been jointly modulated by internal tides， near-inertial responses， and topographic effects. The comparison between HYCOM and sbPOM indicates that the two model products show general consistency in the large-scale thermal distribution and post-typhoon subsurface anomalies. However， HYCOM is not a direct observation， but an observation-constrained， data-assimilative ocean analysis product. Therefore， HYCOM is used in this study as an event scale background reference rather than as an absolute truth. Under this framework， sbPOM is evaluated based on whether it can reasonably reproduce the major thermal structures represented by HYCOM， while further resolving regional details related to localized wave-like structures， topographic modulation， and baroclinic adjustment. The 0–50 m density field further reveals that density variations during the main typhoon impact period were not confined to the upper 0–10 m surface layer. Instead， isopycnal bending， density-gradient compression， low-density tongues， and the interleaving of high- and low 第 3 頁 density water masses were also observed at 20–50 m. The deformation of the density interface was particularly evident at 40–50 m during the typhoon passage， and residual density gradients remained during the early post-typhoon stage， indicating a delayed internal ocean response. Since fixed-depth horizontal maps alone cannot fully distinguish internal waves from horizontal advection， near-inertial responses， and topographic effects， this study adopts a conservative interpretation and refers to these features as internal-wave-like temperature and density disturbances. Overall， the combined results from the 0–150 m temperature field and the 0–50 m density field indicate that Typhoon RAGASA not only produced sea surface cooling， but also reorganized the thermohaline structure from the upper ocean to the subsurface layer. The alternating cold and warm anomalies at 100–150 m， together with isopycnal bending， density gradient compression， and residual anomalies at 20–50 m， suggest that typhoon energy may have been transferred into the ocean interior through wind-induced mixing， near-inertial motions， background internal tides， and topographic modulation. Based on the integrated interpretation of horizontal temperature and density fields， Hovmöller diagrams， spectral analyses， and cross-sectional isotherm and isopycnal deformation， this study suggests that internal-wave-like temperature and density disturbances occurred in the study region during Typhoon RAGASA. However， these results should not be interpreted as direct confirmation that all banded structures were generated by internal waves， but rather as evidence of typhoon induced subsurface disturbances with internal-wave-like characteristics.
