臺灣魩鱙與沿近海經濟魚類之時空相關性分析

Abstract

海洋生態系中，生物間交互關係難以直接觀察，過往研究常透過出現-不出現資料與豐度資料推測物種間關聯。物種共同出現(Co-occurrence)與物種間相似之豐度趨勢可能反映互利、掠食或具相似環境偏好等關係，而物種相互迴避(Co-exclusion)與物種間相反之豐度趨勢則可能代表競爭、躲避掠食者行為或具相異之環境偏好。魩鱙作為鯷鯡科仔稚魚，是臺灣沿近海重要經濟魚類，也是多種海洋魚類的獵食對象，尤見常見經濟性成魚。本研究探討魩鱙與12項經濟魚類的時空關聯性，以判斷潛在的生物交互作用。資料使用1997至2020年漁業年報與2009至2020年魩鱙卸魚聲明，涵蓋主要捕撈魩鱙的9個縣市，將資料轉換為出現-不出現，建立Jaccard相似度(Jaccard similarity)共現性網絡與機率模型(Probabilistic model)共現性網絡評估魚類間共現機率；除魩鱙外之12項魚類皆使用漁業年報進行產量除以漁船數計算單位努力量(Catch per unit effort，簡稱CPUE)，惟漁業年報於2009年後缺少魩鱙漁船數資料，因此將根據資料來源分為第一期(1997至2008年)與第二期(2009至2020年)，單位努力量經標準化後作為魚類相對豐度指標。後續分析進一步建立魚類之間的Pearson相關性(Pearson correlation)網絡探討其空間相關性，並套用時間序列分析檢視魩鱙與其他魚類在不同時間差下之豐度相關性。結果顯示，整體魚類產量逐年下降，桃園市、新竹縣市與台中市以底棲魚類產量最多，新北市、高雄市、屏東市、宜蘭縣與花蓮縣則以大型掠食性魚類產量最高；Jaccard相似度與機率模型共現性網絡未呈現顯著之共現關係，但在Pearson相關性網絡中，第一期之魩鱙與底棲魚類之空間相關性較高，第二期則與大型掠食性魚類有較強之空間相關性；時間序列分析中，第一期之魩鱙與多數底棲魚類(石首魚科、合齒魚科、大棘大眼鯛與刺鯧)、小型大洋性魚類(鯔科)與1項大型掠食性魚類(鯊魚類)在同一年間具有相關性；第二期則與2項底棲魚類(鯛科與石首魚科)、小型大洋性魚類(鯔科)與1項大型掠食性魚類(鯊魚類)有具時間差(一年以上)之相關性，並與2項底棲魚類(大棘大眼鯛及刺鯧)在同一年間具有相關性。綜合上述結果，魩鱙與沿近海經濟魚類之關係隨時空而有所異，可能因為生物交互作用、海洋環境或資料來源不同等因素所影響。本研究對魩鱙與其他經濟魚類之潛在互動關係進行初步探討，有助於為臺灣沿海生態系帶來進一步之了解。

In marine ecosystems, direct observations of biological interactions are challenging. Therefore, presence-absence and abundance data are often used to infer potential interspecific relationships. Co-occurrence and similar abundance trends between species may reflect mutualism, predation, similar environmental preferences, or niche overlap. In contrast, co-exclusion and opposite abundance trends may indicate competition, predator avoidance, or difference environmental preferences. As larval fish of Engraulidae and Clupeidae, engraulid serve not only as prey for marine predators but also as an important fishery resource in Taiwan. This research aimed to investigate the spatiotemporal relationships between engraulid and 12 commercial fishes to see potential ecological interactions. We used production and fish boat number data from fishery report in 1997-2020 and engraulid landing report in 2009-2020, covering 9 districts with consistent engraulid fishing. The data were converted into presence-absence to construct Jaccard similarity co-occurrence networks and probabilistic model co-occurrence networks to assess co-occurrence probabilities. Catch per unit effort (CPUE) was calculated for the 12 fish species (excluding engraulid) through dividing annual catch by the number of boat number from fisheries report. Since boat data for engraulid were not available after 2009, the analyses with engraulid CPUE were divided into two periods: period 1 (1997-2008) and period 2 (2009-2020). Standardized CPUE was used as a proxy for relative abundance. Pearson correlation networks were used to explore spatial correlations, while time series analysis was applied to examine abundance correlations at different time lags. Results showed an overall decline in total fish production over the years. Taoyuan City, Hsinchu County and City, and Taichung City had higher productions of demersal fishes, while New Taipei City, Kaohsiung City, Pingtung County, Ilan County, and Hualien County had greater productions of large predatory fishes. Neither Jaccard nor probabilistic co-occurrence networks showed significant co-occurrence patterns. However, in the Pearson correlation networks, engraulid exhibited stronger spatial correlations with demersal fishes in period 1, and stronger spatial correlations with large predatory fishes in period 2. Time series analysis indicated that in period 1, engraulid had correlations with 4 demersal fishes (Sciaenidae, Synodontidae, Priacanthus macracanthus, and Psenopsis anomala), 1 small pelagic fish (Mugilidae) and 1 large predatory fish (Shark) in the same year. In period 2, lagged correlations (more than 1 year) emerged with 2 demersal fishes (Sparidae and Sciaenidae), 1 small pelagic fish (Mugilidae), and 1 large predatory fish (Shark). Correlation in the same year were observed with 2 demersal fishes (Priacanthus macracanthus and Psenopsis anomala). In summary, the spatiotemporal relationships between engraulid and other commercial fishes varied across time and districts, potentially driven by ecological interactions, environmental conditions, or fishing activities. This study provides insights into potential interactions between engraulid and other offshore commercial fishes, contributing to a better understanding of Taiwan’s coastal ecosystems.