探討魚類群集變化與環境因子間關聯性之統計分析

Statistical Analysis of Temporal Relationships between Fish Community and Environmental Factors

邢惟綱(Wei-Kang Hsing)

指導教授 ： 劉力瑜

共同指導教授 ： 陳虹諺(Hungyen Chen)

國立臺灣大學/生物資源暨農學院/農藝學研究所/碩士(2020年)

https://doi.org/10.6342/NTU202001294

摘要

長時間的魚群取樣資料常被用來從事魚類群集隨時間變化的研究，並且能夠納入環境因子探討魚群變化與外在因素間的關聯性。本研究利用於台灣北部核電場出水口透過三層取樣刺網 (trammel sampling net) 以及水下觀察 (underwater visual census) 所得的魚群取樣資料，並且參考魚類資料庫FishBase計算不同取樣時間多面向的生態指標：物種數、個體數、群集平均最大體長、群集平均偏好水溫、群集平均營養階層、群集中具有漁業利用價值魚類比例、以及群集中珊瑚礁魚類比例。
接著本研究透過主成份分析 (principal component analysis) 配合雙軸圖 (biplot) 探討魚類群集 (fish community) 之不同面向生態指標隨時間變化之趨勢，以及生態指標間、生態指標與環境變數間之線性相關性。納入分析的環境因子包含南方震盪指標 (Southern Oscillation Index, SOI)、海水表層溫度 (sea surface water temperature, SST)、以及亂流混合指標 (turbulent mixing index)。結果發現以個體為平均的生態指標，整體可觀察出早晚年份有差異，代表可能因為人為或是環境因素造成魚類群集的變化，而透過水下觀察觀察到的魚類個體平均最大體長與個體平均養階層卻呈現負相關，這部份有待後續研究。
另外本研究透過小波同調分析 (wavelet coherence analysis) 探討魚類群集 (fish community) 生態指標與環境因子間之短暫連動關係 (transient relationship)。結果發現小波分析的結果並未能反映物種數、個體數、群集平均適應水溫與環境因子具有關聯性，其中一個可能的原因為指標變數雜訊過多，較不適合透過小波分析觀察連動性。

關鍵字

魚類群集 ； 主成份分析 ； 雙軸圖 ； 小波分析

並列摘要

Long-term monitored datasets of fish communities are commonly used to investigate the temporal changes of fish assemblages. With these datasets, environmental factors can also be incorporated to reveal the relationships between changes of fish assemblages and environmental factors. In this study, we used the dataset of fish community captured around the outlet areas of nuclear power plants in northern Taiwan by trammel sampling net and underwater visual census to calculate different aspects of fish community indices: the number of species, the number of individuals, average max length, average preferred temperature, average trophic level, the rate of fish for fishery use, the rate of coral reef fish. Features of different species were referred to FishBase (a fish database).
Next, we used principal component analysis with biplot to investigate the temporal changes of fish community and the linear relationship between indices of fish community and environmental factors, including Southern Oscillation Index, sea surface water temperature and turbulent mixing index. The results show there are differences between early years and late years in terms of the overall individual average indices. This may due to the influence of environmental changes or anthropogenic pressures. But individually average max length and individually average trophic level are negatively correlated recorded by underwater visual census. The reason can be studied in further researches.
We also applied wavelet coherence analysis to investigate the transient relationship between indices of fish community and environmental factors. The results show there is no obvious transient relationship between fish community indices and environmental factors. We speculated too much noise in time series made it difficult to investigate the transient relationship between community indices and environmental variables.

並列關鍵字

fish community ； principal component analysis ； biplot ； wavelet analysis

參考文獻

Allison, E. H., Perry, A. L., Badjeck, M. C., Neil Adger, W., Brown, K., Conway, D., . . . Andrew, N. L. (2009). Vulnerability of national economies to the impacts of climate change on fisheries. Fish and Fisheries, 10(2), 173-196.

Bakun, A., Parrish, R. H. (1982). Turbulence, Transport, and pelagic fish in the California and Peru current systems. California Cooperative Oceanic Fisheries Investigations, 23, 99-112.

Benoit, H. P., Swain, D. P. (2008). Impacts of environmental change and direct and indirect harvesting effects on the dynamics of a marine fish community. Canadian Journal of Fisheries and Aquatic Sciences, 65(10), 2088-2104.

Butchart, S. H., Walpole, M., Collen, B., Van Strien, A., Scharlemann, J. P., Almond, R. E., . . . Bruno, J. (2010). Global biodiversity: indicators of recent declines. Science, 328(5982), 1164-1168.

Casals, J., Jerez, M., Sotoca, S. (2002). An exact multivariate model-based structural decomposition. Journal of the American Statistical Association, 97(458), 553-564.