|Abstract: ||海表粗糙度是控制海洋邊界層的主要因素，可以由拖曳係數來表示。另外，均方傾度（mean square slope，MSS）也是一種表征海表面粗糙度的參數。因此，本研究目的在於以ADCP觀測資料估計MSS的算法，並討論其有效性和準確性。第二，藉助岸基通量塔以渦動方差法量測海表拖曳係數。第三，探討控制大氣邊界層特性的海表粗糙度，風速和波浪參數，並討論這些參數之間的内在聯係。|
本研究中，我們發現15公尺海表風速小於10 m/s時，拖曳係數與風速呈現正相關，當風速在10m/s到15m/s之間時，拖曳係數與風速成負相關。另外，MSS與風速呈現相關性，相關係數為0.7，p值是2×〖10〗^(-20)。與此同時，根據第二類回歸模型，拖曳係數與MSS呈現正相關，然而相關係數為0.01. 拖曳係數與均方傾度之間的對應關係離散程度較大意味著，除均方傾度之外，仍有其他參數可用來解釋拖曳係數與均方傾度的關係。因此，為探究哪些其他參數對拖曳係數的與均方傾度的關係有影響，我們做了8組實驗，每個實驗被設計爲測試一組參數來討論結果。這些參數是，潮流方向，示性波高，示性波浪周期，主波向，風速，風與波浪之間的夾角，風生波浪海表粗糙度雷諾數（wind sea surface roughness Reynolds number，Rb），以及波齡。最後，在我們探究風與波浪參數的内在關係后，我們發現Rb和波齡最適合被用以分類討論拖曳係數與均方傾度的關係，以及被用以分類討論拖曳係數與風速的關係。
;The sea surface roughness which is a dominating factor controlling marine boundary layer (MBL) can be described by the drag coefficient〖(C〗_d). Additionally, the mean square slope also can be a parameter of the sea surface roughness. Therefore, this study proposes to implement an algorithm that estimates the mean square slope (MSS) from ADCP measurement and to discuss its effectiveness and correctness. Second, to implement the observation of sea surface drag coefficient using the eddy covariance method from coastal based flux tower. Third, to investigate the characteristics of sea surface roughness, wind speed, and wave parameters, which control the MBL and discuss the inter-relationship among the characteristics.
The C_d was estimated from turbulence measurement from a 3-D ultrasonic anemometer using the eddy-covariance method. The instrument installed on a flux tower above 15 m of mean sea level in TaiCOAST station located on the western coast of Taiwan. While around 1.4 km to the northwest from the Yongan coast, in-situ wave data was obtained by ADCP that mounted at the bottom sea from 26 May 2011 to 22 June 2011. Then, the wave directional spectrum calculated from water elevation data was transformed into the wavenumber spectrum to estimate the MSS.
In this study, we found the drag coefficient was positively correlated to wind speed when u ̅_15 was less than 10 m/s, while the trend became negative when u ̅_15 was between 10 m/s and 15 m/s. Next, the MSS shows a strong dependency on u ̅ with a correlation coefficient of 0.7 and p-value of 2×〖10〗^(-20). While the C_d was positively correlated to MSS base on model II regression, however, the dots were scattered so that the correlation coefficient r was 0.01. The scattered dependency of C_dN on MSS indicated there were other parameters except for MSS to explain the pattern of C_dN. Thus, to investigate what are the other parameters that have more influence on C_dN, we made 8 experiments, each experiment including one parameter, to discuss the result. And, these parameters were tidal current direction, H_(1⁄3), T_(1⁄3), D_p, wind speed, the angle difference between wind and wave direction, R_b, and wave age. Finally, after we did inter-relationship between wind-wave parameters, we found that R_b and β were suitable parameters to categorize the pattern of drag coefficient in relationship with wind speed as well with MSS.