衛星影像應用於臺灣秋刀魚漁業之漁海況變動研究

Studies on the Fishing and Oceanographic Conditions of Taiwanese Pacific Saury (Cololabis saira) Fishery Using Multi-Sensor Satellite Images

曾振德(Chen-Te Tseng)

指導教授 ： 孫志陸

國立臺灣大學/理學院/海洋研究所/博士(2012年)

https://doi.org/10.6342/NTU.2012.01345

摘要

本論文處理分析2006–2010年多重衛星遙測影像（海面水溫、海洋水色及基礎生產力等），利用區域性及全域性空間分析模式，建立西北太平洋臺灣遠洋秋刀魚（Pacific saury, Cololabis Saira）漁獲分布與海洋環境因子之相關性。同時，也利用不同的影像處理及邊緣偵測技術，以及地理資訊系統之空間整合套疊與地理統計分析等方法，探討重要水文特徵（水溫鋒面及特定等值線等）對秋刀魚魚群分布及移動洄遊特性之影響，並推測秋刀魚之漁場分布。結果顯示，漁季初期（6至8月）平均單位努力漁獲量（catch per unit effort, CPUE）較低，其CPUE為10.8公噸/艘/日，盛漁期為9至11月，最高CPUE為10月份之23.1公噸/艘/日，總平均CPUE為15.3公噸/艘/日。臺灣遠洋秋刀魚作業漁場，主要分布於北緯37–48度、東經145–165度，其月別漁獲重心呈現明顯的緯度向變化，受亞熱帶環流鋒面及亞極區環流鋒面相互作用，主要漁場分布於黑潮（Kuroshio）及親潮（Oyashio）的交匯海域。另外，本論文亦利用Cayula–Cornillon直方圖邊緣偵測技術，處理分析衛星海面水溫影像，萃取出水溫鋒面（潮境）分布位置，藉以探討秋刀魚漁獲海域與水溫鋒面之相關性。結果顯示，6至8月份，作業海域內水溫鋒面較少，秋刀魚CPUE也偏低。相反地，9至11月份，水溫鋒面明顯增加，秋刀魚CPUE也隨著提高。同時也發現，秋刀魚的漁獲位置愈靠近水溫鋒面分布海域，其CPUE也比較高，CPUE與最近的水溫鋒面距離，呈現顯著負相關。

此外，本論文亦利用經驗累積分布函數(empirical cumulative distribution function)，處理分析秋刀魚可能棲息海域（potential saury habitat）之海洋環境因子主要分布範圍。結果顯示，臺灣遠洋秋刀魚作業漁場內，魚群可能棲息海域之海面水溫值為14–16 oC，海洋水色值為0.4–0.6 mg m-3，基礎生產力為600–800 mg C m-2 day-1，同時符合前述三項海洋環境因子之條件值，經圖層套疊分析，可萃取出秋刀魚可能棲息海域（熱點）之時空分布。另外，利用地理權重迴歸（geographically weighted regression）模式，處理分析秋刀魚漁獲分布與海洋環境因子的空間非平穩型相關性。結果顯示，利用地理權重迴歸模式，可提高海洋環境因子對秋刀魚漁獲空間分布的解釋變異百分比，略高於目前常用的泛加成模式（generalized additive model）分析方法，故建議進行漁獲分布與海洋環境因子相關性分析時，應同時考量變數間之空間異質性（spatial heterogeneity）及相依性（spatial dependence）。再者，本論文亦探討不同海面水溫上升情境下（包含正常年及1、2、4°C水溫模式），秋刀魚可能棲息海域受氣候變遷影響之時空變動特性。結果顯示，秋刀魚的可能棲息海域，會隨著海面水溫上升，有逐漸往北推移的現象。其中，於水溫上升4°C模式下，秋刀魚可能棲息海域，往北推移的幅度最大，其南邊界線會往北推移至北緯46.15度。

綜合上述結果，本論文已完成探討西北太平洋臺灣遠洋秋刀魚漁獲之時空分布特性，並掌握海洋環境因子對秋刀魚魚群移動及洄遊分布之影響。這些分析結果可提供秋刀魚漁業資源管理策略參考，並作為未來發展秋刀魚漁海況速預報服務的重要基礎資訊。

關鍵字

秋刀魚 ； 衛星遙測影像 ； 漁海況分析 ； 經驗累積分布函數 ； 地理權重迴歸 ； 可能棲息海域

並列摘要

Five years (2006–2010) fishery data coupled with multi-sensor satellite images was examined to determine the habitat characterization for Pacific saury (Cololabis saira) in the northwestern Pacific Ocean (NWP). Results showed that monthly average CPUEs (metric tons/boat/day) ranged from 10.8 in early fishing season (June to August) to 23.1 in October. The overall average CPUE was 15.3. The major fishing grounds located within 37–48oN latitude and 145–165oE longitude with a remarkable latitudinal movement of the monthly mean centers of gravities.

Pacific saury’s habitat preferences in the Taiwanese fishing grounds were determined using the empirical cumulative distribution function. The high CPUEs corresponded to areas where sea surface temperature (SST) ranged from 14–16 oC, Chlorophyll-a (Chl-a) concentrations ranged from 0.4–0.6 mg m-3 and net primary productions (NPP) ranged from 600–800 mg C m-2 day-1. Local areas within the NWP with these similar satellite-derived oceanographic parameters were assumed to be the potential habitat zones of Pacific saury.

Satellite SST data also were used to extract fronts in Pacific saury fishing grounds. The fronts were identified by the Cayula-Cornillon edge detection algorithm. The results show that low frequency of SST fronts is associated with lower saury’s CPUEs during the early fishing season (June to August). Conversely, high frequency of SST fronts is associated with higher saury’s CPUEs during the peak fishing season. Additionally, if the fishing locations of Pacific saury are close to the SST fronts, higher saury’s CPUEs are observed.

The spatial non-stationary geographically weighted regression (GWR), a local modeling technique, was applied to examine the influence of oceanographic variability on the distribution of Pacific saury. The results of the GWR were compared with those of a generalized additive model (GAM). Results indicated that the distribution of Pacific saury is positively related to SST and Chl-a, and the GWR models explained more variability than the GAMs.

Based on the SST preferences in concert with the corresponding fish distributions, monthly potential saury habitats were predicted. Possible changes in potential saury habitats were estimated under 4 scenarios: the present years and with 1, 2, and 4 °C rises in SST due to climate change. Results revealed an obvious poleward shift of potential saury habitats during the influence of increases in SSTs. The southernmost boundary of potential saury habitat located at 40.24°N latitude at the present time shifted to 46.15°N latitude under the scenario of a 4 °C rise in SSTs.

In summary, the results obtained from this study improve our understanding of the variability in the spatial distribution of saury habitats, and could form the basis for fishery management and fishing forecasts.

並列關鍵字

Pacific saury ； satellite remote sensing image ； fishing and oceanographic investigation ； empirical cumulative distribution function ； geographically weighted regression ； potential habitat zone

參考文獻

Huang, W.B. (2007) Body Length, Weight, and Condition Factor of Pacific Saury (Cololabis saira) from the Landed Size-classes of Taiwanese Catch in Comparison with Japanese Statistics. Journal of the Fisheries Society of Taiwan, 34(4): 361–368.

Andrade, H.A., Garcia, C.A.E. (1999) Skipjack tuna fishery in relation to sea surface temperature off the southern Brazilian coast. Fisheries Oceanography, 8(4): 245–254.

Attrill, M.J., Power, M. (2002) Climatic influence on a marine fish assemblage. Nature, 417: 275–278.

Bakun, A., (2006) Fronts and eddies as key structures in the habitat of marine fish larvae: opportunity, adaptive response and competitive advantage. Scientia Marina, 70: 105–122.

Belkin, I.M., O’Reilly, J.E. (2009) An algorithm for oceanic front detection in chlorophyll and SST satellite imagery. Journal of Marine Systems, 78(3): 319–326.