博士班 Phd
王敘民 Shu-Min Wang
電子郵件Email: shuminwan004@gmail.com
陳威成 Wei-Cheng Chen
電子郵件Email: jasonwuton@gmail.com
李敏豐 Min-Feng Lee
電子郵件Email: s815050505@gmail.com
碩士班 Master
張芹瑜 Qin-Yu Zhang
洪昱翔 Yu-Xiang Hong
電子郵件Email: julian6983@gmail.com
李鑑洲 Jian-Zhou Li
電子郵件Email: z106456466@gmail.com
劉彧愷 Zi-Jie Jian
電子郵件Email: yukailiu0620@gmail.com
簡子傑 Zi-Jie Jian
電子郵件Email: jaychien131@gmail.com
畢業生 Graduate
2023年
廖征暉 Cheng-Hui Liao
基於自然的解決方案(Nature-based Solutions, NbS)概念自2008年提出，近年來已廣泛被應用在水利、海岸與山林治理方面，NbS的核心精神是儘量恢復地球原有的自然生態系統，並依服務能力來應對各式社會挑戰，包含災害治理問題，NbS概念並非無工程措施，NbS是同時考量「水」、「環境」和「人」，讓利害關係者(Stakeholder)可以獲得多重效益，又稱共同利益(Co-benefit)的一個方法，本論文應用NbS研究的課題是海岸保護。 由於NbS仍屬新創名詞，國際間對實施NbS迄今仍未有明確框架或流程，各式問題的影響層面不同，導致各研究提出的設計理念與評估方式雖然相似但並不相同，使得施作NbS難以直接引用。因此，本研究蒐集近年國際間對NbS的定義、原則和標準，嘗試提出一個可應用於海岸保護的NbS治理框架。 本研究共蒐集了18種可應用於海岸保護的NbS解方，逐項闡述其內容，另外，本研究建立選擇NbS的優選方法，利用相關利益者和專家的知識和經驗，以直覺式模糊群體決策方式建置NbS治理評估模式，並確立評估指標的評估方式、權重分析和措施排序流程。透過此流程，決策者能選擇和實施合適的海岸NbS防護措施。最後，本研究以台南黃金海岸保護為例說明論證成果，分析結果證實了本研究所提框架的可行性和應用潛力，未來可擴展至其他海岸段的防護上。 本研究提供了一個與以往國際間實施NbS不同的共識性評估指標和實施框架，有助於推動NbS的普及和實施，有望促進NbS在全球的普及和應用，以應對氣候變遷的挑戰。 The concept of Nature-based Solutions (NbS) was introduced in 2008 to deal with Social challenges and climate change. It has been widely use in water, coastal, and forest management. Although NbS has been widely implemented, there’s no a globally consensus framework, owing to distinct challenges and impacts encountered in each application. This study bridges this gap, proposing an NbS framework for coastal protection within stakeholder’s opinion. The research identifies 18 NbS measures for coastal protection, providing an understanding of each. It establishes an intuitive fuzzy group decision-making model for selecting the best NbS solution, employing the expertise of relevant stakeholders. This framework includes determining assessment methods, analyzing weight, and ranking protective measures. The resulting process enables decision-makers to choose and implement appropriate coastal NbS measures efficiently. The study is substantiated by a case study from the Gold Coast in Tainan, demonstrating the framework’s feasibility and potential for wider application. This research contributes a unique consensus assessment indicator and implementation framework for NbS, differing from previous international efforts. By doing so, it not only encourages the popularization and application of NbS but also fosters its expansion in Taiwan, crucial in meeting climate change challenges.
許家瑄 Chia-Hsuan Hsu
波候(Wave Climate)是指一段時間內特定海洋區域的波高、週期、方向和頻率等統計特徵，近數十年來氣候變遷加劇，台灣周遭海域波候是否有所變化值得探討。本研究採用 WAVEWATCH III 波浪模式(簡稱 WW3)，並使用深水物理套組 ST6 參數，推算 1975 年至 2022 年共計 48 年間在台灣周遭海域的波浪變化，其中的風場輸入係比較了 WRF、ERA5 與 CCMP 三種風場，再和實測風速比較，並考慮時間和空間解析度後，決定同時使用 WRF (2006 至 2022 年)和 ERA5 (1975 至 2005 年)風場為波浪模式之驅動力。 波浪模擬結果以現場浮標觀測資料進行驗證，結果顯示以 WRF 和 ERA5 風場為輸入之波浪模擬結果相近且與資料浮標實測結果一致，使用前述兩個風場來模擬波高的平均絕對誤差分別為 0.33 和 0.41 公尺、相關係數分別為 0.92 和 0.87，對於長期波浪模擬，此結果令人滿意。 本研究分別探討台灣海峽、台灣東部海域以及北南海海域三個區位的波候變化，結果顯示：台灣海峽及北南海海域之平均波高每年增加 0.11 公分，台灣東部海域每年增加 0.32 公分左右；根據每十年的平均波高變化率顯示，氣候變遷對海況的影響隨時間愈來愈顯著，變化率逐年增加。本文進一步分析各季節的波候變化，結果顯示過去五十年來，冬季的波候變化較大，前述三海域的冬季平均波高每年增加 0.4 至 0.5公分；然而，在台灣海峽的夏季平均波高有下降的趨勢，這是否與颱風生成的變化有關值得後續進一步探討。另外，本文還探討波候變化是否有空間差異性，結果顯示不論在台灣海峽、台灣東部海域或北南海海域，代表惡劣海況的大波高範圍有由南向北且逐漸靠近台灣的趨勢。以上的波浪模擬分析結果對瞭解台灣周遭海域過去五十年來的海況長期變化趨勢提供了科學的佐證。 This study collects the open wind dataset from the public and evaluates its applicability to Taiwan waters with buoy observed data. The WAVEWATCH III wave model and the source term ST6 would be applied to calculate the sea conditions in the Taiwan waters and North of South China Sea in recent 50 years. Analysis through the trend of significant wave height in each sea area over the past 50 years is conducted by using linear regression. Among the above calculations, the annual average wave height has increased 0.11 cm per year in the Taiwan Strait and the N.SCS with 0.32 cm per year in eastern Taiwan water. This paper uses the change rate of average wave height per decade to explore the impact of climate change in various sea areas in the past 50 years. It is found that the impact of climate change in each sea area has become more and more obvious along with time. However, Taiwan has usually been hit by typhoons in summer and affected by northeast monsoon in winter. Therefore, this study further analyzed the wave climate changes in different seasons. The results show that wave climate in winter has been affected the most by climate change. The increase of average wave height in various sea areas can reach 0.4 cm to 0.5 cm during winter time. Due to climate change, typhoons gradually move to higher latitudes, and the average wave height in summer in the Taiwan Strait tends to decrease. In addition to the growth trend of wave heights, the distribution of wave heights may also vary with time because of climate change. In this study, the top 10% of the annual average wave heights are adopted to analyze the changes in the distribution of larger wave heights in various sea areas. The results show that no matter the Taiwan Strait, the eastern waters of Taiwan or the North-South China Sea, the areas with relatively high wave heights all move from south to north and gradually approach Taiwan.
陳泓杰 Hung-Jie Chen
極端波浪(extreme wave)指的是在隨機波浪中，一波列中高度較高且不頻繁出現的波浪。本研究參照前人研究結果，將示性波高大於10公尺的波浪定義為極端波浪。過往許多極端波浪的研究多以遙測資料為主，本研究將以實測資料為主，透過分析極端波浪的特性與成因，建立出極端波浪的經驗公式，並推出極端波浪可能的最大值。 本研究蒐集全球各地的實測波浪資料庫，其中以美國NDBC波浪資料庫、以及台灣的波浪資料為主，建立出極端波浪資料庫，收錄接近400筆示性波高大於10公尺的波浪資料，其成因多颱風、颶風等熱帶氣旋所引起，佔極端波浪資料庫中的92%。 透過分析極端波浪的波齡值，本研究發現多數的極端波浪的主要成分皆為風浪，並透過與JONSWAP譜作套配，發現極端波浪多趨近於完全發展波。根據上述條件，利用Young and Vinoth (2013)與Hwang (2016)的颱風風浪經驗公式，比對公式估算值與實測資料的結果，無因次均方根誤差分別為29%與36%，並發現將實測資料帶入兩個公式時，需在風域的計算上做修正，本研究根據實測資料間風域變化，修正Young and Vinoth (2013)中的風域計算，提出等效風域修正係數用以提升經驗公式的計算準確度；根據本研究提出的修正公式，其估算值與實測資料間的誤差降為15%，顯示本研究的修正法能更加準確的估算極端波高。此外利用此修正公式與本研究中的颱風資料估算颱風期間產生的極端波高，其最大值可達到20.7米。 Extreme waves are ocean waves with exceptionally high amplitudes. In this study, waves with significant wave height exceeding 10 meters are defined as extreme waves. By analyzing the characteristics of extreme waves, an empirical formula is developed to estimate the maximum possible extreme wave height. The study compiles a global database of in-situ wave measurements, with a focus on the US NDBC wave database and wave data from Taiwan. The extreme wave database includes nearly 400 records of waves with significant wave height exceeding 10 meters, with 92% of them being associated with tropical cyclones such as typhoons and hurricanes. Analyzing the wave age values of extreme waves, the study finds that most extreme waves are primarily wind waves and tend to approach fully developed waves, as indicated by fitting the JONSWAP spectrum. When compared with Mitsuyasu (1975) directional distribution function, it was observed that the rate of wave energy decay with respect to angles is more gradual. Using the typhoon wind-wave empirical formulas proposed by Young and Vinoth (2013) and Hwang (2016) and comparing their estimations with the measured data, the dimensionless root mean square errors (NRMSE) are found to be 29% and 36%, respectively. It is noted that adjustments are required in the fetch calculations when applying these formulas to the measured data. Based on the variations observed in the fetch from the measured data, the study proposes an equivalent wind fetch correction coefficient to enhance the accuracy of the empirical formula. By incorporating this correction, the NRMSE between the estimated values and the measured data is reduced to 15%, demonstrating that the proposed correction method provides more accurate estimates of extreme wave heights. Moreover, using this corrected formula and typhoon data from this study, the estimated maximum extreme wave height during typhoon events can reach up to 20.7 meters.
王駿瑋 Chun-Wei Wang
波浪和海流是海洋中兩項重要的物理要素，大部份的研究都分別探討之，然而，波浪和海流同時存在，其間也會產生交互作用相互影響，本研究目的即在於探討在海流出現時對波浪特性之影響。 研究分析主要透過方向波譜(Directional spectrum)進行，然而方向波譜分析方法有很多，本研究就四種常見的方向波譜分析方法進行比較，結果顯示貝斯推估法(BDM)對波浪方向特性整體的解析能力較佳。 本研究分析在高雄興達電廠棧橋海域7204筆陣列式波高儀觀測結果與同步之流速觀測資料。分析結果顯示：當波浪和海流反向(逆流)時，波浪能量產生堆疊，波高變大且與流速成正比，當流速增加0.3 m/s時，波浪能量增加2%；另外，在逆流條件下，逆向流強度提升會導致波浪主頻會從高頻往低頻移動，當流速超過0.6 m/s時，波浪主頻與無海流時主頻相比約會降低31%；至於波浪方向會受到海流的影響而有所偏移，當波浪與海流方向近乎垂直時，會出現最大的波向偏移，約為5.38° (波流向差45°時，波向偏移3.83°)，隨著海流速度的增加，此偏移量愈大。 本研究基於現場實測數據分析獲得上述結果，此結果與過去文獻的定性研究成果類似，但本研究提供了定量數值，對於掌握波流交互作用有了更多的認識。 Waves and ocean currents are two important physical elements in the ocean, waves and ocean currents coexist and interact with each other, thereby influencing another one. This study investigate the impact of the presence of ocean currents on wave characteristics. This study primarily utilizes directional spectrum to examine the wave properties. There are various methods for analyzing directional spectra, we compare four common used methods derived by predecessors. The results show that the Bayesian Directional Method (BDM) exhibits better overall analytical abilities for wave direction characteristics. The study analyzes data from 7204 wave height measurements obtained from an array-type wave gauge located in the vicinity of the Xingda Power Plant Stack Bridge in Kaohsiung. The analysis reveals that waves and ocean currents are in opposite directions (countercurrent), wave energy accumulates, resulting in increased wave height proportional to the flow velocity. A 0.3 m/s increase in flow velocity leads to a 2% increase in wave energy. Moreover, as the flow velocity increases, the waves tend to approach the point of breaking. Under countercurrent conditions, the dominant wave frequency shifts from high frequencies to low frequencies with increasing flow velocity. When the flow velocity exceeds 0.6 m/s, the dominant wave frequency decreases by approximately 31%. Wave direction is also affected by the ocean currents, causing a deviation. The maximum wave direction deviation occurs when the wave direction is nearly perpendicular to the current, at approximately 5.38°(with a 45°difference between wave and current directions, the wave direction deviation is 3.83°). This deviation increases with higher flow velocities. The results are obtained based on the analysis in this study. They are consistent with qualitative findings from previous literature, but this study provides quantitative values, contributing a better understanding of wave-current interactions.
2022年
陳盈智 Ying-Chih Chen
異常波浪會發生在岸邊與海上，因突然出現對岸上的遊客或海上操船或工作平台是一大威脅。為減少人員生命財產損失，發展機率預警系統有其必要。海上與近岸的異常波浪發生機制不同，近年來有許多研究探討海洋異常波浪演化的過程，但仍舊不完全了解；在近岸異常波浪(瘋狗浪)的研究，僅有少數國內學者討論。因發生機制複雜且尚未完全了解，本研究仰賴統計的方法進行預警模式建置，旨在發展海洋異常波浪與海岸瘋狗浪發生機率預測的方法，並進一步建置預警模式。在海洋異常波浪部分，彙整前人研究成果，得到一套異常波浪發生機率估算理論，然該理論採用線性波假設，本研究對估算理論進行修改，並以台灣海域的波浪資料進行率定，得到適合台灣海域的異常波浪發生機率估算理論。估算理論可結合業化波浪模式預報資料，完成一套異常波浪預報模式。預警模式以東吉島連續觀測之波浪資料進行驗證，24小時、36小時和48小時的異常波浪預測機率誤差分別為15.9%、22.9%和20.1%，顯示預警時間越長，預報的誤差越大，但整體來說預警模式具可參考之價值。本研究分析預警模式運作結果，發現當有天氣系統生成時，天氣系統的邊界通常有較高的異常波浪發生機率，且當鋒面通過台灣海峽時，也有異常波浪顯著提升的現象。本研究也分析數個颱風範圍內的異常波浪發生機率，顯示在颱風的第1象限有較高的發生機率，第三象限的發生機率較低，此結果與前人的研究結果相符，顯示異常波浪預警模式預測結果相當合理且可靠。 本研究使用SPH(Smooth Particle Hydrodynamic)粒子法模擬2012年基隆海洋大學外一件瘋狗浪案例，探討波浪越堤後水粒子的運動行為，除了對越堤浪花特性有所掌握，亦發現堤前消波塊會使越堤水量減少，但增加水粒子的水平穿越速度。在瘋狗浪預警模式的部分，本研究蒐集台灣東北角海域的近岸影像，並發展了一套瘋狗浪影像分析程序，從影像中以二值法與均值法萃取浪花邊界，並從浪花的高度與浪花速度定義瘋狗浪，經人工驗證，該方法辨識準確，成功辨識率達86%。本研究在瘋狗浪預警模式建置採用類神經網路方法，以過往瘋狗浪落海事件與浮標資料作為輸入資料與訓練之樣本，輸入資料包含7個與瘋狗浪生成有關之海氣象因子，預警模式可預測未來12、18和24小時之瘋狗浪發生機率。本研究以瘋狗浪落海事件與瘋狗浪影像案例進行驗證，兩者正確率約有8成；同時也參考氣象局長浪即時訊息進行驗證，結果顯示預警模式可以提前發布預警。 本研究建置的兩套異常波浪與瘋狗浪模式，本研究也結合波浪模式進行瘋狗浪與異常波浪機率預測，經驗證兩模式可準確預測發生機率，顯示兩模式已可實際應用。 Freak waves may be generated and occur on the coast and in the ocean. The rapidly appearing freak wave is a huge threat for the people on the coast and on ships. It is necessary to develop a freak wave warning model to reduce the loss of life and property. The mechanics of freak wave generation in the ocean and coast are different. In recent years, many studies have explored the evolution of oceanic freak waves (OFWs), but it is still not fully understood. Only a few researchers have discussed coastal freak waves (CFWs). The mechanics of coastal freak waves are also not fully analytical. Warning models need to rely on stochastic methods for development. The purpose of this study was to propose a method to predict the occurrence probability of oceanic and coastal freak waves. A warning model was constructed by combining the operated wave model and the prediction method. The probability estimation formula for oceanic freak waves was obtained from literature. However, the formula was derived based on the linear wave assumption, in which there is an error under highly nonlinear situations. The new formula to estimate the probability of oceanic freak waves was proposed and calibrated based on the wave data observed in Taiwan waters. The verification of the theory shows that the error has changed from 22.3% to 13.2%. In this study, the operational wave model was applied as the input of the warning model to forecast the occurrence probability of oceanic freak waves. Coupled with the wave model operation, the occurrence probability of OFW could be calculated from the nonlinear parameter (μ4) of the wave, which can be obtained from the wave model output. The warning time of the model was 24, 36, and 48 hours. Verification result of the warning model was obtained by continuous wave observation on Dongji Island. The prediction probability errors of OFW at 24 hours, 36 hours and 48 hours were 15.9%, 22.9% and 20.1%, respectively. The forecast error was greater for longer warning times. The warning model has been tested and operated since 2016. The model output show that when a weather system was generated, the location close to the edge of the weather system usually had a higher probability of OFW. The prediction results of the warning model are quite reasonable and reliable. Near-shore optical images in northeastern Taiwan were collected and analyzed in this study. A coastal freak wave image analysis procedure was also proposed. The coastal freak wave was identified through the splash height and speed. The boundary of the splash was extracted from the image by edge detection technology with a success rate of 86%. The procedure was accurate in identifying CFWs. SPH (Smooth Particle Hydrodynamic) model was applied to simulate a CFW event at National Taiwan Ocean University in Keelung in 2012. The hydrodynamic behavior of the wave overtopping was discussed. The results showed that the armor block in front of the breakwater will reduce the discharge of wave overtopping and increases the horizontal throw speed. In the warning model, a neural network algorithm was used in the construction. The CFW events and buoy data were used as training samples and input data. The model predicts the occurrence probability of a CFW with lead times of 12, 18 and 24 hours. This research was verified by CFW events from media and from optical video. The accuracy rate of the two was approximately 80%. The warning model was also verified by the swell warning announced by CWB. All the verification results showed that the model can provide reliable warning information. CFW and OFW model proposed in this study operated with combination of wave model. The outputs of occurrence probability of CFW and OFW are accurate which mean the model could be used in practical.
何雅惠 Ya-Hui Ho
台灣海域遊憩活動發展日益蓬勃，以往多以禁止或限制的政策管理水域遊憩活動，近年改施以開放、鼓勵的政策，因此，風險的掌握有其必要性，事先知悉海域遊憩活動環境的風險有助於降低災害之發生，本研究針對海域遊憩活動進行風險評估研究。 本研究藉由模糊德爾菲法建立與海域遊憩活動關聯之風險評估指標與因子，再藉由層級分析法分析各因子權重，計算各因子的相對重要性，並同時考慮危險度與脆弱度進行風險分析，分析因子中，除了納入海氣象水文特性與地文環境特性外，亦考慮了海岸地區的社會經濟、醫療與救災量能以及遊客行為等進行風險評估。 本研究透過海氣象水文與地文觀測資料與數值模擬，以及專家問卷蒐集分析資料，研究結果顯示，對於海灘型態的海岸，共有10項危險度因子和12項脆弱度因子納入分析，危險度因子中權重最高者為「近岸海底坡度」，脆弱度因子中權重最高者為「安全救生設施」。 本研究選擇台灣七處熱門遊憩海岸包含不同海岸型態進行風險評估分析，並比較有無考慮暴露度的風險等級，在遊憩活動上，暴露度可以遊客人數代表之，結果顯示，暴露度的納入對風險等級帶來顯著的影響。風險評估建議應定期更新，資料的齊全性高度影響了分析結果，是未來要有更正確評估應留意之事項。 In the past, comprehensive policies had been strictly implemented on water recreation activities. However, as ocean recreation becomes more and more popular, the policies have become loosen in recent years; Therefore, it is necessary to gain a sufficient understanding of the potential risks. By acknowledging the risks of ocean recreational activities in advance, the occurrence of disaster could also be reduced. This study is concerned with the risk assessment on ocean recreational activities. Risk assessment indicators related to ocean recreational activities are established by the Fuzzy Delphi method. The weight of the indicators is analyzed by the AHP method, which is calculated for the relative importance of each factor. In addition to ocean meteorology and hydrological characteristics and ocean environment, the indicators have also taken social economy, medical response capacity for disaster, and tourist behavior of the coastal area into account for risk assessment. The results show that the beach type of risk assessment includes 10 hazard indicators and 12 vulnerability indicators. In this study, we select seven popular recreational coasts in Taiwan for risk assessment and compare their risks with and without the consideration of exposure index, as the exposure index can be represented by the number of tourists. In addition, it is necessary to be aware of the completeness of data to conduct a more accurate assessment, as the completeness of the dataset can cause a significant impact on the assessment result.
林楨琇 Chen-Hsiu Lin
離岸風電工程在規劃設計時，常採用以歐洲海域資料分析所得之參數為輸入，然而台灣周遭海域之海氣象特性有別於歐洲海域，因此，本研究分析適用於台灣海峽海域之設計參數，主要針對德國聯邦海事和水文局(BSH)之設計標準: 關於專屬經濟海域內離岸結構物之結構設計的最低要求、IEC 61400-3: 風力發電機-第3部分:離岸風電之設計要求以及ISO 19901-1：石油及天然氣工業-離岸結構物的具體要求-第1部分：海氣象設計和運作注意事項，以上規範中所提及之風暴(storm)影響時間、JONSWAP波譜參數及波浪統計比值進行探究。 本文分析位於台灣海峽北段之新竹資料浮標長期實測資料，先提出影響該海域之颱風評選條件，過去24年共有39個颱風對新竹海域海況造成影響，分析結果顯示，颱風期間海面浪高超過最大波高一半之惡劣海況影響時間平均為19.6小時，僅為BSH規範建議值之56%。本研究提出，台灣海峽海域惡劣海象除颱風影響外，還應考慮東北季風影響，東北季風期間之最大示性波高不若颱風期間為大，但該惡劣海況影響時間平均為40.6小時，為颱風期間之兩倍。本文另外針對台灣海峽北部海域之JONSWAP波譜參數進行分析，結果顯示平時海況之譜峰增強因子γ值約為1.68，惡劣海況期間約為1.39，該數值遠小於規範建議值3.3。針對波浪統計參數的比值分析發現，最大波高和示性波高比值為1.46，最大波高對應之週期和尖峰週期的比值為0.91，符合ISO19901-1規範建議範圍，而尖峰週期與示性波高之相關係數為5.8，且指數為0.42，平均週期與示性波高之相關係數為4.4且指數為0.29，兩者皆不符合Goda (2000)建議值。本文最後使用了位於台灣海峽中段彰化外海的浮標資料進行比較分析，結果顯示該海域之惡劣海象影響時間與新竹海域有所差異，顯示台灣海峽之海象變異性頗大，對於離岸風電工程之設計參數，宜採用當地或鄰近海域資料進行分析，過遠測站資料分析結果恐有高度不確定性。 In the design of offshore wind turbines, European metocean parameters are often used to input the parameter setting. However, the metocean characteristics of Taiwanese waters are different from those of European seas. Therefore, this study discusses the storm impact time, JONSWAP spectral parameters and wave statistical ratios from the design standards, which are Bundesamt für Seeschifffahrt und Hydrographie (BSH): Standard Design, Minimum requirements concerning the constructive design of offshore structures within the Exclusive Economic Zone, IEC 61400-3: Wind turbines – Part 3 Design requirements for offshore wind turbines and ISO 19901-1: Petroleum and natural gas industries – Specific requirements for offshore structures – Part 1: Metocean design and operating considerations. This paper analyzes the long-term data of the Hsinchu buoy, and proposes the selection criteria for typhoons. In the past 24 years, a total of 39 typhoons have affected the sea conditions in the waters of Hsinchu. The result shows that the average impact time is 19.6 hours, which is 56% of the recommended value of BSH standard. The author considers that the influence of the northeast monsoon on severe sea state should be taken into account. The average impact time of northeast monsoon is 40.6 hours. In addition, this paper analyzes the JONSWAP spectrum in the northern seas of the Taiwan Strait. The result shows that the peak enhancement factor is about 1.68 in normal sea state and 1.39 in severe sea state. According to the ratio analysis of wave statistical parameters, it is found the ratio of the maximum wave height to the significant wave height is 1.46, and the ratio of maximum period to the peak period is 0.91. The correlation coefficient between the peak period and the significant wave height is 5.8, the power is 0.42, and the correlation coefficient between the mean period and the significant wave height is 4.4, the power is 0.29. Finally, we use the data of Changhua buoy for comparative analysis. The results show the impact time of the severe sea state in this sea area is different from that in the Hsinchu seas.
卓慧丞 Hui-Cheng Cho
碎波角(Peel angle)(亦稱浪崩角)是兩不同時間的碎波邊界連線與碎波前進方向之夾角，是評估衝浪活動適宜性的一個參數，本研究旨在建立一個自動分析碎波角的模式。 本文採用深度學習方法裡的卷積神經網路(Convolutional neural network, CNN)來分析碎波角，本研究透過模式超參數的率定來提高辨識準確率，結果顯示CNN模式中之激活函數選擇Sigmoid函數，並使用2000次迭代以及64的批次大小可獲得最佳模式。本研究分析三個現場觀測海域共30部空拍影片(影像時序列)，對於碎波帶的判識準確性達80%以上，而對於碎波角的分析，獲得其平均誤差為5.9o，在可以接受之範圍。除此之外，考量便利性，本研究亦提出以單張影像取代影像時序列來計算碎波角，結果發現，單張影像與前述影像時序列分析結果相近，兩者誤差約3.9o，此結果確認未來可以單張影像來進行分析。 最後，本文分析台灣10處衝浪熱點海域的碎波角，結果發現各地的碎波角約介於30o~70o之間，其季節變化並不顯著，分析也發現部分海域受海岸地形影響，碎波角有較大空間變異性，民眾從事相關活動時宜留意。 Peel angle is the angle between the boundary line of the breaking wave and the advanced direction of the breaking wave at two separate times. It is a parameter used to assess the suitability of surfing activities. This research intends to build a model for the automatic analysis of peel angles. In this study, the Convolutional Neural Network (CNN) is used to calculate the peel angel. The calibration of the model hyperparameters has been implemented to increase the identification precision. The results indicate that the optimal CNN mode can be obtained by using the Sigmoid activation function,2000 iterations, and a batch size of 64. In three on-site observation sea areas, a total of 30 aerial videos (time-series images) were analyzed, and the identification accuracy of the surf zone was greater than 80%. The average error for the analysis of the peel angle was 5.9 degrees. In addition, replacing time-series images with a single image is proposed in this study in order to determine the angle of breaking wave. The results indicate that the picture and video analysis results are comparable, with a difference of approximately 3.9 degrees. The results of the study confirm that the analysis method of a single image is feasible. Eventually, the peel angle of ten surfing spots in Taiwan is analyzed in this study. The results indicate that peel angles range between 30 and 70 degrees, and seasonal variations are negligible. The investigation also revealed that certain coastal regions are affected by topography and the spatial distribution of the peel angle is different. Therefore, the public should take precautions when engaging in associated activities.
傅仲偉 Chung-Wei Fu
裂流(rip current)常將泳客捲往外海，造成意外，裂流發生機制複雜，迄今仍難以預測其可能發生的時間與範圍，掌握裂流特性，更加瞭解裂流，有助於未來進行裂流預警。因此，本研究目的為從觀測影像中，辨識出裂流，作為裂流研究資料來源。近年利用人工智慧(AI)研究興盛，本研究係利用AI中的深度學習(Deep Learning, DL)技術辨識裂流，採用的方法以卷積神經網路(CNN)為模式基礎改良的更快速區域卷積神經網路(Faster Region-base Convolutional Neural Network, Faster R-CNN)。 本研究分析國內外裂流衛星影像共220張，另外也蒐集100張無裂流影像，共320張進行訓練與驗證。在Faster R-CNN模式建置的訓練過程中，引用了影像擴增(data augmentation, DA)方法增加訓練資料量，並完成超參數(hyperparameter)之率定，以建置裂流辨識模式。本研究探討模式建置過程中信心門檻值(confidence value)之決定，當設定為0.5時，準確性可達86%；另外，模式訓練過程中使用了影像擴增方法，可使辨識準確率提高14%。透過本研究Faster R-CNN模式的辨識結果與人工診斷結果比較驗證，顯示Faster R-CNN模式辨識準確率可達86%。再與前人研究結果比較，發現本研究研發模式的訓練影像較少即可達到與前人相近的準確性，證實了深度學習演算法具有正確辨識出影像中出現裂流的優異能力。 Rip currents are dangerous currents that result in many accidents by sweeping swimmers and beach goers. The mechanism of rip currents is too complicated to predict when and where rip currents might occur till the present moment. Knowing more characteristics and features about rip currents will be beneficial to their early warning in the future. Therefore, the aim of this study is to recognize rip currents on images and also be researching data. Research with artificial intelligence (AI) is a big hit around the world, thus, this paper uses deep learning (DL), a branch of AI, to identify rip currents. The DL method used is Faster Region-base Convolutional Neural Network (Faster R-CNN) which is an improved version based on another famous model, Convolutional Neural Network (CNN). In this study, the quantity of used images is 220 rip currents images and 100 of non-rip currents images from satellite. In the process of setting up Faster R-CNN, data augmentation (DA) was employed in order to increase the training data amount and finished the calibration of hyperparameters later. This study discuss how large confidence threshold is, as it is 0.5, accuracy would be 86%. Furthermore, using DA on images before training, the accuracy of detecting rip current could improve 14%, better than not using DA. With both the result of Faster R-CNN and identified them by eyes, the accuracy of recognition rip currents would achieve 86%. In addition, comparing with previous study that used the same model, although the research data amount in this paper is less than the previous one, both accuracy are about the same. It shows that this kind of DL algorithm has the excellent abilities to detect rip currents on images correctly.
2021年
梁立翰 Li-Han Liang
臺灣四周擁有豐富的海洋資源，早期漁業捕撈業的興盛使得漁港數量多，然而港口結構物可能破壞海岸沉積物運移的平衡，造成海岸線侵蝕或淤積現象。本研究以安平港的興建為例，探討其對臨近之黃金海岸的影響，並分析假設拆港後對水動力及海岸地形變遷之影響。本研究利用地表水模擬系統(Surface-Water Modeling System, SMS)中的水動力模式(CMS-Wave和CMS-Flow)，以夏、冬兩季代表事件模擬臺南安平港口拆除前後，當地沿岸的水動力變化及地形變遷機制，同時利用漂砂模式(PTM)分析二仁溪河口漂砂的運移軌跡。研究結果顯示，拆港後，在距離原港口1 km範圍內的波高有顯著的變化；平均流速在拆港後於原港口南側2.5 km範圍內亦有明顯差異，尤其在原港口防波堤附近影響最大，隨距離港口越遠影響程度越小，安平港拆除後，由於距離關係，至黃金海岸段的水動力變化已不明顯。地形變遷方面，港口拆除前後主要影響地形侵淤趨勢的位置在距離港口南邊1 km範圍內，在港口拆除前，夏季有淤積趨勢，冬季則無明顯侵淤變化；港口拆除後，夏季的堆積趨勢減小，甚至出現侵蝕現象，而冬季則因少了防波堤的屏蔽效應，開始有明顯的侵淤變化，造成這些現象係由於兩季的代表波浪與沿岸結構物的交互作用，使得帶動沿岸漂砂的水動力不同而導致。本研究亦模擬了二仁溪河口輸砂之漂砂軌跡，本研究區域無論是夏冬兩季漂沙趨勢是緩慢向北移動，本研究認為波浪方向是造成兩季節漂砂沉積結果不同的主要原因，夏季波向多來自西南方，波浪與潮流作用後使漂沙多堆積於河口北側；冬季波向多來自西北方，使得漂砂自河口出流後多堆積於河口南側。
The coastal structures protect the wave attack, but it may break the sediment transport cycle and retreat of the coastline. This study takes the removal of Anping Harbor as an example to explore its impact on the Golden Coast, and analyzes the impact of hydrodynamic and morphodynamic changes due to the sea harbor removal. In this article, the CMS model in the Surface-Water Modeling System (SMS) was used to simulate the hydrodynamic and morphodynamic changes on the representative events of summer and winter. And the Particle Tracking Model (PTM) was also applied to simulate the movement trajectory of the drift sand in the Errenxi River estuary.
The simulation results show that after removed the harbor, the wave height has a significant change within 1 km from the coast, and the average current velocity is also significantly different within 2.5 km south of the harbor, especially near the harbor’s breakwater. The hydrodynamic changes to the Golden Coast section are no longer significant. In terms of morphodynamic changes, the location that mainly affects erosion or deposition before and after harbor removal is within 1 km from the south of the harbor. Before the harbor removed, there is a deposit in summer. After the harbor is removed, the location experiences erosion in summer. Due to the lack of the shielding effect of the breakwater, there is obvious changes in invasion and deposit in winter. These phenomena are caused by the interaction between the waves and coastal structures, which makes the hydrodynamic forces that drive the coastal drift sand different.
This article also simulated the trajectory of sand transport in the Errenxi River estuary. In the study area, the trend of drifting sand is slowly moving northward. The direction of the waves is the mainly factor caused the drift sand deposition in summer and winter.
黃孟威 Meng-Wei Huang
論文題目：台灣鄰近海域極端波浪研究
Thesis Topic: Study on the Extreme Waves in Taiwan Waters
極端波浪(extreme waves)指的是在海洋中發生波高極高的波浪，波浪的成長與風的條件(風速、延時、風域)有關，而風的條件中又以風域(fetch)的資訊與相關研究與最少，因此本研究發展一套風域判釋法，可以根據大氣模式產出的風場資料，自動判釋出風域大小，驗證結果顯示所提出之風域判釋法可以合理地判釋出各種天氣系統下的風域大小。本研究使用所提出之風域判釋法分析台灣鄰近海域在東北季風和西南季風期間的風場資料，分析結果顯示西南季風期間的平均風域大小為860 km，是東北季風期間(410 km)的兩倍以上，但是西南季風期間的平均風速為7.7 m/s，比東北季風期間的平均風速(9.7 m/s)低了20%左右。本文以JONSWAP的波浪成長經驗公式估算季風期間可能發生之極端波浪，考慮各式極端條件進行分析，研究結果顯示，台灣海峽南側在西南季風期間最大波高可能達到2.28 m，是平時的2.17倍；台灣海峽北部海域在東北季風期間最大波高可能達到6.53 m，是平時的3.24倍，以上是單純由季風吹拂所估算出來的極端浪高，若鄰近有其他天氣系統或湧浪，加成的波高可能更大。另外，本文以Young(1988)提出之經驗式估算颱風期間的極端波浪，以過去紀錄最強颱風中心風速(95.8 m/s，2015年派翠沙颱風)估算所產生之極端波高可達23.41 m。
Extreme waves usually occur in the severe sea state. The growth of waves is related to wind conditions including wind speed, duration, and fetch, among them the information and related research of fetch are the least. This study had developed a fetch interpretation method, which automatically determine the length of fetch based on the wind field data from atmospheric models. The verification result shows that the proposed fetch interpretation method is reasonably under various weather systems. This study applied fetch interpretation method to calculate the length of fetch around Taiwan waters during the northeast monsoon and southwest monsoon period. The analysis result shows that the average length of fetch during the southwest monsoon period is 860 km, which is more than twice as length as the period during northeast monsoon (410 km). However, the average wind speed during the southwest monsoon is 7.7 m/s, which is about 20% less than the average wind speed during the northeast monsoon (9.7 m/s). Furthermore, JONSWAP’s wave growth empirical formula was used to estimate the height of extreme waves that may occur during monsoon period. When considering various extreme conditions, the maximum wave height in the south of the Taiwan Strait during the southwest monsoon may reach 2.28 m, which is 2.17 times than usual; during northeast monsoon, the maximum wave height may reach 6.53 m, which is 3.24 times than usual. The above are extreme wave heights estimated only by the blowing of the monsoon. If there are other weather systems nearby, the wave height may be even greater. In addition, this paper uses the empirical formula proposed by Young (1988) to estimate the extreme waves during the typhoon, that the extreme wave height can reach 23.41 m based on the central wind speed of the strongest typhoon ever recorded in the past (95.8 m/s, Typhoon Patricia in 2015).
2020年
洪繶雙 I-Shuang Hung
論文題目：紅樹林對波浪衰減之數值模擬研究
Thesis Topic: Numerical Study on Wave Attenuation by Mangroves
近年來利用海岸邊的植被如紅樹林作為海岸防護材料十分常見，利用植被來禦潮除了可能具有防災效果外，尚可帶來生態服務，是一種典型的自然解決方案(NBS)。對植被消波作用造成影響的有很多因子，可分為波浪條件(入射波高、週期)、地文環境(水深、海灘坡度)以及植被特徵(植被種類、密度、寬度、高度)等。本研究使用XBeach非靜水壓模式對波浪傳遞經過紅樹林植被後的減衰情形進行模擬研究，探討幾個重要因子對紅樹林消波的影響。為貼近現況，本研究在模式中建立仿紅樹林特徵之植被模型，分別對根部、莖幹和枝葉的特徵進行參數化設定。本研究探討的影響因子包含有波浪條件(給定規則波和不規則波的波高和週期以及設計波浪條件)、植被特徵(包含紅樹林植被區寬度和密度)，以及地文環境和植被的無因次化參數，即植被和水深之相對高度(稱為相對高度)，各條件的模擬結果以無植被情況為對照組進行比較，並採用波高衰減率(R)和單位植被之波高衰減(r)以量化波浪通過植被之衰減程度。由本研究模擬結果發現，紅樹林對波浪的衰減作用十分顯著，波高衰減率(R)在無紅樹林時皆小於3 %，有紅樹林時皆在40 %以上且最高可達92 %。在本研究探討的影響因子中，推斷紅樹林植被寬度對消波作用的影響程度最大。而在對波浪條件的分析中發現，入射波高越大和波浪週期越長則波浪衰減越多，並且入射波高對紅樹林消波的影響較大。波高衰減至一半時的位置皆在植被區域約30 %內，表示波浪在紅樹林前段的衰減速度較快。
Coastal vegetation such as mangrove forests can be a prominent costal engineering application of attenuating wave energy. They can also bring lots of advantages including ecosystem services, economic values and environmental protection.
As wave propagate through coastal vegetation, the height and energy of the waves will lose. Wave reduction in vegetations depends on many factors: hydrodynamic conditions, vegetation characteristics and environmental factor. The influences of each factor are important for the management of the mangrove area.
This thesis aims to identify the effect on wave reduction by each factor includes incident wave height, wave period, vegetation width, vegetation density and the relative height of vegetation height and water depth. The investigations are carried out using 1-D XBeach numerical model non-hydrostatic mode by setting mangrove into 3 layers vertically from bottom to top. The layers are divided as the roots, stem and the canopy where there is vertically significant for mangroves. The study focuses on the wave height variation in mangrove area with use of wave reduction ratio (R) and average wave reduction per m (r).
Results show that the attenuation of wave height by mangroves is significant. R is above 40% and up to 92% in all cases with mangroves, and less than 3 % in cases without mangroves. 50% of the wave height is attenuated in the first 30% of the mangrove area indicating that the wave attenuates faster in the forepart of mangroves. From the analysis of wave factors, the efficient attenuation of waves is related to incident wave height and wave period. From the analysis of vegetation characteristics, when the mangrove width, density and relative height increases, the wave attenuates more.
葉浩君 Hao-Chun Yeh
論文題目：台灣鄰近海域湧浪研
Thesis Topic: A Study on the Swell at Taiwan Waters
湧浪(Swell)又稱為長浪，其能量消散慢因此可以在海上傳遞相當長的距離，而其週期長、動量大的特性則會令湧浪傳遞至海岸邊時有顯著的破壞力，容易造成結構物受損或激起巨大水花(瘋狗浪)，危害岸邊活動民眾。本研究針對台灣周遭海域觀測到的湧浪進行分析，發現其在地域與季節上有差異，資料分析顯示，颱風前後均會傳來湧浪，颱風離開後傳來的湧浪，其波高較颱風前傳來者大了約40%，影響時長也為颱風前傳來者的3倍，顯示颱風離開後的湧浪對海岸衝擊更大。湧浪不僅出現在颱風前後，季風期間亦常出現湧浪，東北季風引起的湧浪影響時長甚至會長達100小時以上，影響程度不容忽視。近年來，中央氣象局有針對湧(長)浪提出警戒，本研究發現，在長浪警戒發布期間，瘋狗浪實際發生的比率有29%；而非長浪警戒期間，瘋狗浪發生比率為11%；此外也發現湧浪能量越大其BFI值越高，顯示湧浪傳入到近岸區域會增加波浪非線性程度。在瘋狗浪發生時湧浪波高分析發現，瘋狗浪發生時平均湧浪波高為平常時的1.62倍，顯示湧(長)浪是造成瘋狗浪的重要因素之一，但並非為唯一影響因子。本研究還發現，在非長浪警戒期間，若發生瘋狗浪，其通常為單一出現，而在長浪警戒期間，瘋狗浪在一小時內出現超過一次以上機率占80%以上。最後本研究透過風湧浪分離方法，獲得湧浪頻譜，以JONSWAP波譜模型套配有良好的結果，並分別得到颱風期間與季風期間平均湧浪譜模型參數，可進一步作為工程評估之用。
Swell with the characteristic of low energy dissipation can propagate for a long distance. The characteristics of long period and large momentum will make the swell have significant destructive power when it is transmitted to the coast. It is easy to cause structural damage or provoke huge splashes (coastal freak wave), harming the people on the shore. This study analyzes the swells around Taiwan waters. Data analysis shows that swells are transmitted before and after the typhoon. After the typhoon leaves, the swell height is about 40% greater before the typhoon. The impacted period of the swell is also three times as much as that before the typhoon. The typhoon swell impact on coast during typhoon left is more significant than during typhoon coming. Moreover, swell induced by monsoon was found in this study. The impacted period of monsoon swell is up to 100 hours in specific case. It shows that the influence of swells during the monsoon cannot be ignored, and it had the different characteristics from typhoon swells. For example, the swell wave height and swell peak period during the typhoon are larger than those during the monsoon, and monsoon had the longer impacted period. In recent years, the Central Meteorological Administration has warned against long waves. This study found that during the warning time, the occurred rate of the coastal freak waves(CFW) is 29%. And the rate during the non-warning time is 11%. In addition, it is also found that the Benjamin-Feir Index (BFI) is directly proportional to the swell energy. It shows that the contraction of swells to the nearshore area will increase the wave nonlinearity. The average swell height during CFW occurred is 1.62 times than usual, shows that swells are one of the important factors causing CFWs. Furthermore, CFWs occurred more than once per hour during the warning time. At last, in this study obtain the parameters of the average swell spectrum model during typhoon and monsoon respectively. The parameters can be used for engineering assessment in the future.
吳柏緯 Bo-Wei Wu
論文題目：船難事件發生時之海象分析
Thesis Topic: Analysis on the Sea States of Shipwrecks
全世界有近九成的貨物是依靠船運送往世界各角落，隨著船隻活動日漸頻繁，船難的發生頻率也逐年提升。海氣象狀況是發生船難的原因之一。台灣周遭海域時常發生船難，本研究蒐集2004年至2019年間一共120起因海氣象因素造成之船難事件作為本文研究對象。本研究從波譜的角度探討船難成因，根據波譜型態，本文獲得船難發生時，波浪之頻譜主要出現單雙峰交互演化以及雙成分單峰譜兩種型態。本研究參考前人研究，歸納因海象造成之船難可能與波浪能量集中有關，以兩個波譜參數來描述波能集中情形，包含譜尖參數(Qp)與方向分散參數(σθ)，它們分別代表波譜在頻率域以及方向域上能量的集中程度，Qp愈大或σθ愈小都代表波浪能量愈集中。本文將船難事件的波譜參數繪製在Qp-σθ二維關係圖上，分析結果顯示，出現第一種單雙峰譜交互演化的船難海象時，在事件發生前，Qp、σθ兩參數關係位於第二象限(Qp小σθ大)，在船難事件發生時，其關係轉移至第四象限(Qp大σθ小)，此證實了船難發生時波能有顯著集中的趨勢，隨後再回復船難前的海象特徵。而另外一種船難則出現在雙成分單峰譜海況下，船難發生時的波浪能譜為單頻，但能量卻來自兩個方向，因兩成分波頻率接近使波能在波浪系統間傳遞效應更為顯著，有些案例在船難發生前後兩成分波波能有一增一減的趨勢發生。上述兩種船難海象都屬於波浪間的交互作用，只是交互用用的機制不同，但都是造成船難的海象特徵。本文也分析了Qp、σθ在船難發生前後變化率，發現兩參數的變化率平均分別為34%與36%，期望後續研究可以運用此變化率來建立海況危險程度之指標。
With the number of maritime transport increased, navigation safety has become a major concern. Ship accidents usually cause huge economic losses, human casualties and environmental pollution. Meanwhile, news often broadcasted that marine accidents occurred frequently around Taiwan waters. The marine weather and meteorological conditions play a significant role in maritime safety. However, the conditions of typical sea states that trigger shipping accidents are not well understood. This study collected 120 marine accidents in a 16-year (2004~2019) ship accidents dataset. First, we classified the types of wave spectra into one-peaked spectra and double-peaked spectra when the accidents happened. In this thesis, using the trajectory in the Qp σθ diagram to graphically represent the temporal change of the directional spectrum during ship accidents. The higher Qp and lower σθ represent the more concentrated wave energy in frequency fetch and direction fetch, respectively. Under the unimodal sea conditions, the diagram scatter plot between Qp and σθ move from upper left to lower right. This results showed the positive correlation between the shipwreck and seas with more concentrated wave energy. Under the bimodal sea conditions, the shipwreck occurred at the diagram of upper right. The frequency of two components is nearly the same, but comes from different direction. The wave-wave interaction between two components with same frequency lead to one of the reasons for shipwrecks. To conclude, no matter the number of wave components, shipwrecks occurred with the spectrum more concentrated at frequency during ship accidents. The study also discusses the Qp and σθ rate of change, and found that the rate of change is 0.3 approximately. Future study could use the change rate between Qp and σθ to build the early warning of the freakish sea state.
2019年
張紘聞 Hung-Wen Chang
論文題目：應用WWIII 波浪模式於極端波高模擬之研究
Thesis Topic: A Study on the Extreme Wave Height Simulations by WAVEWATCH III
颱風是種變化劇烈的天氣系統，颱風圈內的風力極強，風向風速隨地而異，波浪分布因此到處都不同，由於最大波高在海事工程設計上有著重大的意義，因此模擬颱風期間最大波高成了很重要的課題。本研究應用WAVEWATCH III波浪模式(簡稱WW3)，針對過去9件實際量測到颱風期間最大波高超過10公尺的案例進行波浪推算。透過不同物理套組與各國(台灣、法國、美國)單位所提供之套組參數設定對所選定之案例進行模擬，藉以了解WW3在模擬颱風極端波高之能力。根據模擬與浮標實測資料比較結果顯示，WW3內定的ST2、ST4、ST6套組在整體颱風波浪模擬上都有不錯的表現，但在極端波高模擬時有低估現象，其中以ST6套組模擬結果相對誤差14%為最小。而使用各國作業化單位所建議的套組參數進行模擬時，颱風整體波高結果顯示，ST2套組中風浪成長項的風速修正參數(c0)與ST6套組中風浪成長項的阻力係數修正參數(FAC)的調整在波高量值上有顯著的影響，而對於極端波高模擬以美國海軍研究所使用的ST6套組參數(本文簡稱ST6_NRL3)有最佳的計算結果，相對誤差為10%。此外，本文也探討了WRF風場的準確性以及其對於颱風極端波高模擬的敏感度分析，結果顯示，WRF風場與浮標測得最大波高時之風速平均相對誤差為26%，且有高度敏感性，可知風場準確性對於極端波高的模擬結果較模式本身參數設定之影響為大。
During typhoon events, accurate prediction of waves is essential to the ocean engineering design. It can reduce the damage to coastal infrastructure and protect the safety of people’s lives. However, due to the complex sea state and air-sea interaction, it is still challenging to improve the prediction accuracy in the simulation of extreme typhoons. In this study, we will find out the appropriate solutions of the errors to improve the accuracy of extreme wave height calculation. The WAVEWATCH III wave model (WW3) was used to calculate the wave in the past 9 typhoon cases which the maximum wave height is larger than 10 meters. The source term packages and parameter settings from various countries (Taiwan, France, and USA) had been discussed to understand the ability of WW3 in these cases. According to the comparison between WW3 results and buoy measured data, it is shown that the three packages (ST2/4/6) of WW3 have good performance in the overall typhoon wave simulation. However, there is an underestimation in extreme wave height simulation, in which the relative error 14% of ST6 simulation results is the smallest. When using the settings suggested by the national operation unit, the results of the overall wave height trend of typhoons show that the adjustment of the wind speed correction parameter (c0) in ST2 and the drag coefficient correction parameter (FAC) in ST6 has the most significant influence on the simulation of the extreme wave height during the typhoon. For extreme wave high simulation, the parameters adjust by United States Naval Research Laboratory in ST6 (this article is referred to as ST6_NRL3) have the best calculation results, the relative error is 10%. Furthermore, the study also discusses the accuracy of wind field provided by WRF (The Weather Research and Forecasting). The average relative error of wind speed when the maximum wave height measured by WRF wind field and buoy is 26%. It shows that the wind field accuracy has a greater impact on the extreme wave height simulation than adjusting WW3 parameters.
林沛臻 Pei-Chen Lin
論文題目：利用衛星高度計資料研究颱風波高分布
Thesis Topic: A Study on the Wave Height Distribution during Typhoon Period by Altimeter Data
本文利用衛星高度計資料分析颱風最大風速半徑八倍內之波高空間分布，分析澳大利亞海洋資料網(AODN)所提供之178663筆衛星高度計資料，並與位於外洋海域的台東外洋浮標和東沙浮標資料進行驗證。研究結果顯示：衛星高度計反算之波高和實測值之相關性達0.9以上，週期相關性達0.8以上，顯示使用衛星高度計在波高分布的研究上具有可靠性。本文以颱風期間最大波高以及最大風速半徑為無因次參數從事颱風波高分布研究，首先驗證了Willoughby and Rahn(2004)所提出的颱風最大風速半徑估算經驗式最合理，本文再分析依颱風強度與前進速度分類探討颱風最大風速半徑八倍內之波高分布，結果顯示，颱風波高分布型態受颱風移動速度影響大，當移動速度較小時，最大波高出現在颱風第一象限；當移動速度較快時，最大波高則出現在第四象限，此結果與前人在較強颱風且移動速度慢條件下所提出之波高分布不同，可能原因是過去之颱風波高分布係使用模型風場模擬所得，本文則是使用衛星高度計觀測之實際風場，分析所得之波高分布應具有較高可信度。
The spatial distribution of wave height within eight times the maximum wind speed radius of the typhoon had been discuss in this study by using altimeter data. The satellite altimeter data provide by Australian Ocean Data Network portal (AODN). The AODN data base collects 13 altimeter data. The total length of data is 33 years. The buoy data (Taitung Open sea and Dongsha) had been used to verified wave height, wave period and wind speed obtain from AODN. The maximum wind speed empirical formulas were discussed in this study. Comparison with JTWC best track typhoon data, the best empirical formula was proposed by Willoughby and Rahn (2004). According to the classification of typhoon intensity and forward speed, the wave height distribution show that the pattern is greatly affected by the typhoon forward speed. When the moving speed is slow, the maximum wave height appears in the forward right quadrant of the typhoon; when the forward movement increase, the maximum wave height appears in the rear right quadrant. Compared with Young (1988), the wave height distribution obtained by the satellite altimeter is different under the condition of strong typhoon and slow moving speed. The possible reason is that Young (1988) obtain the typhoon wave height distribution by using wave model with analytic model of typhoon wind field. The wave data collected from altimeter leads to a high reliability of the wave height distribution.
李丞哲 Cheng-Che Lee
論文題目：利用空拍機影像反算表面流場
Thesis Topic: Video-based Measurement of Sea Surface Current using UAV
海面連續影像序列提供了時間域以及空間域的海表面特徵資訊，是極具發展潛力的海洋環境資料，本研究討論利用空拍機影像所拍攝得到的海面影像時序列作為解析海表面特徵之可行性。本文透過影像分析技術對空拍機拍攝之海面影像時序列做三維傅立葉轉換，再利用頻散關係式(dispersion relation)對影像譜進行濾波，疊代計算得表面海流。本研究在探討觀測準確性之前對空拍機觀測海流遭遇之參數設定與環境因子問題進行探討，包含空間對位校正控制點所需之數量與擺放位置、控制點座標精準度、飛行高度以及風速和光線等因子，結果顯示控制點的佈設位置比數量影響更大，本文建議將控制點均勻佈設在陸域範圍，控制點數量至少四處以上，兩控制點間距建議大於五十公尺，另外本研究証實手機GPS定位座標即可滿足空拍機從事海流觀測之需要。本研究另發現當風速大於8m/s時，太陽仰角過高時，觀測結果不佳。本研究規劃空拍機和微波雷達及ADCP同步觀測來做定性和定量驗證，結果顯示空拍機觀測結果和雷達觀測結果一致，和ADCP量測結果之比對顯示流速均方根誤差為0.0985m/s，流向誤差為12.9º。由定量和定性上的結果顯示，以空拍機影像反算表面流場技術可實際應用於現場觀測，需要更多的實驗來獲得定量驗證結果以提高使用信心。
The images sequence provides information on the sea surface characteristics in the time and spatial domain. In this study, UAV had been used to capture the images sequence. An image analysis method based on the three-dimensional Fourier transform was also developed for obtain the sea surface current information which could be calculated iteratively by dropper shift from wavenumber-frequency spectrum. The influence factors and observation setting of the UAV, including the quantity and position of the ground control points, the coordinate accuracy, the flight altitude, the wind speed and light, were also tested. The results show that the position of the ground control points is more affected than the quantity. We recommended that the ground control points be evenly distributed around the land area. In addition, GPS position from mobile phone (iPhone XR) can meet the needs of the UAV for ocean current observation. This study also found that the observation results are not good under the conditions of wind speed greater than 8m/s or sun elevation angle too high. For verification and validation of sea surface current calculated from UAVs images, X-band radar and ADCP had been deploy and compare with. The results show that the UAV observation results are consistent with the radar, and the comparison with ADCP shows that the RMSE of the current speed is 0.0985 m/s and the direction error is 12.9º.
林家亘 Chia-Hsuan Lin
論文題目：蒲福風浪關係修正之研究
Thesis Topic: A Study on the Modified Beaufort Wind-Wave Correlation
海氣象預報的方式至今有很多，包含諸多常用的經驗方法，或是數值模式預報等。然而漁民、海上作業員、海上遊憩業者及漁類養殖業者，較常根據蒲福風浪關係表由風級來推算浪級，屬於一種快速的經驗法則應用方式。因此本文旨在根據現場的實測海象資料及中央氣象局漁業氣象預報資料，來探討蒲福風浪關係在台灣海域環境下的適用性。本文分析對象包含台灣海峽內之新竹海域、北部龍洞海域以及台東外洋海域。本研究取得中央氣象局漁業氣象預報資料與新竹、龍洞及台東外洋浮標的實測資料，首先針對漁業氣象預報中的平均風級與陣風進行比對，發現漁業氣象預報在三個海域會因不同季節而有不同程度的誤差，而風級的預測偏差可能導致浪級的預測不準確。如夏季期間預測風級較易有高估現象，以陣風為例，其最大有高估達88.4%(龍洞海域)；冬季期間預測風級則較容易低於實測風速量測值，以陣風為例，最大有低估達54.8%(台東外洋海域)。因此本文以新竹、龍洞及台東外洋三測站實測資料與傳統蒲福浪級比較。分析結果顯示，當風級大於(等於)3時，實測浪高均小於傳統蒲福風級所對應的浪高，且發現三個海域風浪成長規模皆不同，主要是各海域水深與地理位置的差異，影響該海域的風浪成長。本研究也就各海域分別提出修正之蒲福風浪關係，當使用經驗推算可快速得到更貼近的實際海況，因此本文認為所提出之修正蒲福風浪關係較適用於台灣鄰近海域。
Taiwan’s marine meteorology is rather complicated due to 1) complicated landform around Taiwan sea area, 2) typhoons in summer and autumn seasons, and 3) Kuroshio current that flows from Philippine to Japan via the east of Taiwan sea area. To study the complexity of Taiwan’s marine meteorology, this research analyzed and compared fishery forecasting data released by Central Weather Bureau (CWB), Taiwan and in-situ measurement data retrieved from data buoys at Hsinchu, Longdong, and Taitung open ocean, respectively. The comparison discovered that the forecasting data had a significant difference with the measured data. Furthermore, the analyses of this research disclosed that the measured wave-height was significantly lower than that of Beaufort wind-wave correlation, when the wind scale was higher than (or equal to) Beaufort number 3. Accordingly, in this study, a new modified Beaufort wind-wave correlation based on the regression of in-situ measurement data is proposed as a more applicable reference for ocean workers in Taiwan sea area.
2018年
楊博堯 Po-Yao Yang
論文題目：台灣環島實測暴潮分析
Thesis Topic: Analysis on Storm Surge in Taiwan Coastal Ocean
颱風暴潮是台灣常見的自然災害之一，前人關於暴潮的研究多半針對特定海岸或是特定颱風進行分析，對於台灣海岸全面性的暴潮特性分析較為缺乏，為了能更了解台灣環島暴潮偏差之特性，本研究系統性的分析環島潮位資料，獲得合計3334個站-颱風之歷年暴潮偏差。分析結果顯示，發生超過1公尺的暴潮偏差大都在北部海岸，主要受西北颱引起。在台灣南部地區觀測到的暴潮偏差量值雖相對較小，但它們約為各區平均潮差的八成，相較於北部及中部海岸相同分析僅約四到五成，暴潮實質的影響對南部海岸較為顯著。另外，本文透過頻率分析計算各地暴潮偏差重現期，結果顯示南部海岸歷年最大暴潮偏差對應的重現期普遍大於當地海堤設計標準，這再度顯示了南部海岸對於暴潮災害之防範須特別謹慎。本文另外探討風場對於暴潮模擬之影響，使用WRF風場及Holland模型風場對於過去兩場曾對台灣西南部造成較大暴潮的颱風事件進行模擬。首先比較兩種風場，結果顯示觀測點距離颱風越遠，兩種風場風速峰值出現時間差可能越大，低估風速峰值現象也越明顯，比較之下WRF有更好的模擬結果，但WRF風場與實際峰值仍有誤差約18.3%，峰值出現時間差約4小時。進而使用WRF風場模擬暴潮，結果顯示西南區潮位站模擬之暴潮偏差峰值誤差40%，峰值出現時間差平均為7.8小時，造成暴潮模擬誤差的可能原因之一為風場本身存在著誤差，比對風速資料發現風速普遍低估，峰值也普遍提早出現，顯示風場對於暴潮模擬具有一定的影響力。
The storm surge is one of the common natural disaster in Taiwan. In the past, researchers focused on storm surge studies at specific regions or typhoons, a comprehensive analysis around Taiwan coastal ocean is still missing. In order to better understand the characteristic of storm surge around Taiwan, we through the data analysis and numerical simulation. Statistical results show that the average storm surge height in the south is about 80% of the average local tidal range, and the percentage is higher than those in the north and central. Furthermore, we used frequency analysis to calculate the return period on storm surge in each region. The results show that the return periods of the biggest storm surges in the southern regions are generally greater than local design criteria. This paper also discussed the relationship between wind field and storm surge simulation. We used WRF wind field and Holland wind field to simulate two cases which had significant influence in the southwestern part of Taiwan in the past. The results show that the WRF wind field has better performance and the error at storm surge peak value is about 40% and the error of the time of peak appear is about 7.8 hours in southwest of Taiwan. The simulation results were checked with wind data. It was found that there exists a certain error in the wind field. Therefore, we can conclude that the wind field has a significant influence on storm surge simulation.
吳祥煜 Hsiang-Yu Wu
論文題目：台灣海域實測雙峰波譜模型之研究
Thesis Topic: A Study on Bimodal Spectrum Fitting for Taiwanese Waters
海洋中常同時存在風浪和湧浪，在頻率域上，這種交錯海象常以雙峰譜形式表現。本研究分析台灣周圍海域的雙峰波譜發現，雙峰譜出現的頻率約為8.8%，當雙峰譜出現時，平均波高為1.66m，較常年平均約低15%，顯示交錯環境下波高發展受限。同時，本研究也發現雙峰譜並非只出現在颱風期間，非颱風期間的雙峰譜出現頻率更高。為了模式化雙峰波譜，本研究利用四個前人提出的雙峰波譜模型：六參數譜(SP model)、Guedes Soares譜(GS model)、Torsethaugen譜(TT model)、Mackay譜(MK model)，套配到實測雙峰譜，結果顯示六參數譜有最佳表現。然而，分析發現模型套配過程中，當海況中湧浪能量超過風浪能量的50%以上，以及兩波峰頻率差小於0.09 Hz等條件下，前人的雙峰譜模型容易變形簡化為單峰譜，此結果並不合理，因此本研究嘗試提出新的譜模型。本文先證實了高斯分布適合來表示湧浪成分波譜，而對於風浪成分則使用JONSWAP-Glenn形式的譜模型，結合兩者成為一個新型雙峰波譜模型，本文稱之為GJ譜。透過實測數據的比對分析，根據六個評鑑參數，證實本文所提的GJ譜模型表現優於前人所提的幾個模式，與SP譜模型比較，雙峰譜變形簡化為單峰譜的比率由6%降至3%，本研究提出了一個表現更好的雙峰譜模型。
Co-exist of wind wave and swell is common in the real ocean. In frequency domain, this crossing sea is expressed by a bimodal spectrum. This study works on the bimodal analysis for Taiwanese Waters. It is found the occurrence probability of bimodal is only 8.8%. The significant wave height in bimodal sea state 1.66m is 15% lower than annual average wave height. In addition, bimodal has higher occurrence probability in non-typhoon condition. To formulate the bimodal spectra, four models presented by previous researchers are evaluated. They are Six-Parameters Spectrum Model, Guedes Soares Spectrum Model, Torsethaugen Spectrum Model and Mackay Spectrum Model. The candidate models are fitting to real bimodal spectra. The best model is the Six-Parameters Spectrum Model according to several evaluation parameters. However, it is found the bimodal spectrum deformed to uni-modal when swell energy above 50% of wind wave energy and the frequency difference less than 0.09 Hz. This is unreasonable. This study proposed a new model that is composed of a Gaussian function and a JONSWAP-Glenn model. This new model has been verified by field data and proofed to have better performance. The deformation rate reduces from 6% to 3% by using the proposed model. This study proposed a useful bimodal spectrum model for future application.
夏兆威 Chao-Wei Hsia
論文題目：湧浪分析與追蹤研究
Thesis Topic: A Study on the Swell Characteristics and Tracking
海面紛紜波浪是由風浪及湧浪混合而成，其中湧浪具低耗散的特性可傳遞很長距離而不消散，當湧浪傳遞至近岸時，對海岸結構物的衝擊程度大，也易激起浪花威脅岸邊人員活動安全，本研究旨在分析台灣週遭海域湧浪特性並嘗試追蹤湧浪源頭。本研究挑選五個浮標資料進行湧浪資訊分析，並進一步統計以了解四周海域湧浪之分布情形。最大的湧浪波高記錄是在2012杰拉華颱風期間，波高達16.97公尺；在台灣北部與東部海域冬季期間湧浪較高，主要來自東北季風，平均達1.6公尺以上；台灣西南海域湧浪夏季期間湧浪較高，平均達1.15公尺以上，主要來自西南季風與颱風。從湧浪特性分析發現，颱風期間引起的湧浪波高大但影響時間短，平均約80小時；而季風系統引起的湧浪則高度小但影響時間長，最長可達一週以上。本研究挑選20個顯著湧浪案例，使用迴歸線法追蹤湧浪源頭，並以地面天氣圖與模式風場資料進行佐證，結果顯示此方法可以合理掌握湧浪來源。針對颱風期間之湧浪案例，本文發現湧浪多生成於颱風第一象限，且生成位置距離颱風中心約0.4-2.4倍最大七級暴風半徑處。研究發現部分湧浪追蹤結果不合理的案例因為天氣系統移動過快所致，颱風移動速度達18 公里/小時以上就無法成功追蹤，也顯示湧浪源頭追蹤方法之使用限制。
The coexistence of wind wave and swell often occurs in the sea. Swell can transmit to the nearshore for long distance with the characteristic of low-dissipated. Swell is likely to damage the coastal structures and endanger the safety of the people on shore. The purpose of this study is to analyze the characteristics of swell in Taiwan’s waters and applied the swell tracking technology to typical swell events. The analysis results show that the swell height in the northern and eastern seas of Taiwan during the winter is higher than other seasons, with an average of more than 1.6 meters; the swell in the southwestern Taiwan during the summer is higher than other seasons, with an average of more than 1.15 meters. The swell induce by typhoon is high, but the impact time is short. The average duration is about 80 hours. In contrast, the swell induced by the monsoon is relative small but has a long duration of impact and can be up to a week or more. This study used the regression line method to find the source of the swell and corroborated the weather map and the wind field data. The results show that this method is suitable to find the sources of swell. From the case of swell during the typhoon, it was found that the swell was mostly generated in the first quadrant of the typhoon. The study also found that some cases of unreasonable tracking results were due to high speed of weather system movement (18 km/hr or more).
蘇瑋琳 Wei-Lin Su
論文題目：應用資料探勘方法於瘋狗浪研究
Thesis Topic: Application of Data Mining Methods on Coastal Freak Wave Study
臺灣周遭海岸每年發生許多瘋狗浪事件，造成嚴重傷亡。海岸瘋狗浪發生原因相當複雜，至目前仍無一套理論能預測瘋狗浪何時何地會發生。本研究分別採用類神經網路及群集分析方法，從過往實際發生之落海事件建置瘋狗浪機率預警系統。採用過去瘋狗浪相關研究中得知會顯著影響瘋狗浪發生之海氣象參數，做為預警系統之輸入因子，兩種方法均使用相同的分析資料，經資料探勘方法充分訓練後，再進行驗證，結果顯示，不論類神經網路方法或群集分析方法，預警所得之正確率皆達七成以上，皆具有良好預警能力。整體而言，類神經網路系統具有較高的正確率，顯示有較好的預警能力，但群集分析系統之回應率較佳，表示預測結果可信度很高，在實際運作中誤報率較低，也具有一定的參考價值。經由敏感度測試證實，示性波高、波浪尖峰週期與班傑明非線性指數(Benjamin Feir Index , BFI)為影響瘋狗浪發生之顯著因子。本文從少數預警失敗案例發現，這些失敗案例主要集中於波高偏小的時刻，顯示預警系統仍有改善的空間；此外也發現本研究建置之群集分析系統與中央氣象局現行運作測試中的群集分析瘋狗浪預警系統比較，驗證正確率皆在同一個級距，此方法失敗的案例主要是發生在波向風向差因子主導之瘋狗浪事件，較無法準確預測。本文考慮未來預警系統作業化之穩定性，評估使用不同輸入資料建置預警系統之差異，結果顯示以實測海氣象觀測資料建置預警系統最佳。
There are many coastal freak wave events every year in Taiwan, which cause serious casualties. The mechanic of coastal freak wave is quite complicated. There is still no theory to predict when and where coastal freak wave will occur. In this study, data-driven warning systems based on an artificial neural network (ANN) and cluster analysis are proposed to predict the occurrence possibility of a CFW. It shows that, regardless of the neural network method or cluster analysis method, the accuracy rate of warnings is more than 70%, and all have good warning capabilities. The artificial neural network system has a higher accuracy rate and shows good early warning capability. However, the cluster analysis system has a better response rate, indicating that the prediction results are highly reliable and the false alarm rate is higher in actual operations, it also has a certain reference value. Through sensitivity analysis, it was confirmed that the significant wave height (Hm0), wave peak period (TP), and Benjamin Feir Index (BFI) were significant factors influencing the occurrence of coastal freak wave. This article further understands a few number of cases of warning failures, and finds that the warning system still has been improved when the two warning systems aim at wavelet-high coastal freak wave incidents. This study considers the stability of the operation of the warning system, evaluates the differences in the use of different input data to build an early warning system, and the results show that it is best to build an early warning system based on buoy observation data.
2017年
李堉辰 Yu-Chen Lee
論文題目：從方向波譜分離風湧浪之研究—有限吹風延時法
Thesis Topic: Wind Wave and Swell Separation from Directional Wave Spectrum – The Duration-Limited Method
在隨機海洋中，海浪經常是由風浪及湧浪混合而組成。從海洋現場觀測之波浪頻譜中分離出風浪和湧浪的成份，對於波浪理論研究、海浪預報及近岸與離岸工程實務上都有重要之意義。過去風湧浪分離之研究主要從一維波譜進行，然而，風浪或湧浪的判別必須考慮風向和波向之關係，因此，從二維方向波譜來分離風浪和湧浪在物理上更顯合理。Hanson and Phillips (2001)提出了一個基於方向波譜的風湧浪分離方法大致上均可獲得正確結果，NWWIII波浪模式亦採用其方法。然而，本研究發現了Hanson and Phillips (2001)的方法在大風速的情形下高估了風浪成分，這是由於該法假設了無限吹風延時與無限風域的緣故。本文提出一個考慮吹風延時的風湧浪分離修正方法，稱為有限吹風延時法，此方法是基於風浪成長理論所提出。本文選擇了12筆各式氣象、海象條件下的方法波譜進行驗證，分析結果顯示本文的分離結果合理，分離所得的風浪成分符合風浪成長關係式，分離的結果還透過波浪尖銳度分布的檢測與能譜時序列演變的判別，都證實了本文分析結果正確。
The coexistence of wind wave and swell often occurs in the sea. It is necessary to separate wind wave and swell components for engineering design and hazard warnings. Swell separation using a frequency spectrum (1D) was originally proposed and useful. However, the separation of wind wave and swell should consider the misalignment between wave and wind directions. A separation approach based on a directional spectrum (2D) is essential. Hanson and Phillips (2001) proposed a 2D separation method, that has been widely applied, such as being adopted in the WaveWatch III model, and most of the separation results are correct. However, this study found that this may overestimate wind sea energy under strong wind conditions because their method essentially assumed unlimited wind fetch and duration. This study proposed a modification to reflect wind duration when separating wind wave and swell. We used 12 typical directional wave spectra in various marine weather conditions for validation. The wind wave and swell separation by our proposed method obtained reasonable results. The wind wave component is verified to satisfy the wind-generated wave growth function. In addition, the results were re-verified when studying their steepness distribution and spectrum time series.