Please use this identifier to cite or link to this item:
Real-time forecast analysis of swordfish fishing condition for the Taiwanese distant-water longline fishery in the Pacific Ocean
|Authors: ||Chang Chia-Hao|
|Contributors: ||NTOU:Department of Environmental Biology and Fisheries Science|
swordfish;environmental variability;generalized additive model;habitat suitable index
|Issue Date: ||2018-08-08T03:25:22Z
|Abstract: ||摘要 本研究蒐集彙整2009年至2015年太平洋臺灣遠洋鮪延繩釣劍旗魚漁獲資料及衛星遙測海洋環境參數，透過泛加成模式(GAM)及棲地適合度指標(HSI)建構太平洋劍旗魚棲地模式並比較漁況變動，另針對資料解析度進行分析，比較月別五度與週別一度的差異，探討不同資料解析度是否對分析結果產生影響。結果顯示除月相週期外，模式考慮的環境因子皆為顯著(p<0.05)，海表層水溫對劍旗魚釣獲率(CPUE)的影響最為明顯，解析度差異僅對月相週期有很大的影響，顯示該因子僅適用於解析度較高的分析。比較兩種模式的預測分布發現，劍旗魚主要漁場位於太平洋中部熱帶海域，HSI預測的主要漁場較為寬廣，棲地分布在20°N與20°S附近海域，環境特徵符合劍旗魚喜好範圍，GAM預測分布顯示紐西蘭附近海域及美國加利福尼亞州沿岸亦是劍旗魚喜好棲地。整體而言，GAM所預測的劍旗魚棲地分布較吻合名目CPUE的分布，預測的準確度也比HSI來的高，結果支持GAM比HSI更適合用於建構太平洋劍旗魚棲地模式。然而GAM預測分布受到經緯度因子預測能力的限制，未來可考慮海洋分層環境資料於模式分析中，以更瞭解太平洋劍旗魚棲地與時空分布型式。|
Abstract Fishery data of the Taiwanese distant-water tuna longline fleet for swordfish and satellite based remote sensed oceanographic variables for 2009-2015 were collected in this study to investigate the impact of environmental variability on the fishing condition of swordfish in the Pacific Ocean. Empirical models were developed based on generalized additive models (GAMs) and habitat suitable index (HSI), and were used to analyze the feasibility for forecast of fishing condition of swordfish. Impacts of different resolution at monthly 5° and weekly 1° spatial grids were examined in this study to reveal potential effects on the swordfish CPUE analyses. All the marine environmental factors considered in the habitat models were statistically significant (p<0.05), except for the lunar phase. Sea surface temperature can explain a large proportion of deviance on catch rates of swordfish, while the impacts of lunar phase became unclear in the analyses at a coarse resolution. The main fishing ground of swordfish was predicted to locate in tropical waters of the central Pacific Ocean, but the habitat predicted by HSI was limited in specific areas. Results from GAM showed that swordfish occurred in waters near New Zealand and California. However, the areas between 20°N and 20°S were inferred by HSI models to be a preferred habitat for swordfish in the Pacific Ocean. Overall, the fishing ground predicted by GAM was more consistent with the distribution of nominal CPUE and more accuracy in predictions when compared with that from the HSI. GAMs were thus considered as an appropriate approach to develop habitat models for swordfish in the Pacific Ocean. However, the GAMs were limited by spatial and temporal factors in predictions. Marine environmental factors by different layers in the water could be incorporated in the analyses to better understand the swordfish habitats in the Pacific Ocean.
|Appears in Collections:||[環境生物與漁業科學學系] 博碩士論文|
Files in This Item:
All items in NTOUR are protected by copyright, with all rights reserved.