Extraction of sea temperature in the Barents Sea by a scale space multiresolution method - prospects for Atlantic salmon

• Published in: Journal of Applied Statistics
• Main Authors: Pasanen, Leena, Laukkanen-Nevala, Päivi, Launonen, Ilkka, Prusov, Sergey, Holmström, Lasse, Niemelä, Eero, Erkinaro, Jaakko
• Other Authors: University of Oulu. Department of Mathematical Sciences, Knipovich polar research institute of marine fisheries and oceanography, Luke / Luonnonvarat ja biotuotanto / Ekosysteemit ja ekologia / Vaelluskalojen ekologia (4100100314), 4100100314
• Format: Article in Journal/Newspaper
• Language: English
• Published: Taylor & Francis
• Subjects: ilmastonmuutokset; merivesi; lämpö; lohi; aikasarja-analyysi; climate change; sea water; temperature; Atlantic salmon; time series analysis; Barents Sea
• Online Access: http://jukuri.luke.fi/handle/10024/540870

Variation of marine temperature at different time scales is a central environmental factor in the life cycle of marine organisms, and may have particular importance for various life stages of anadromous species, for example, Atlantic salmon. To understand the salient features of temperature variation we employ scale space multiresolution analysis, that uses differences of smooths of a time series to decompose it as a sum of scale-dependent components. The number of resolved components can be determined either automatically or by exploring a map that visualizes the structure of the time series. The statistical credibility of the features of the components is established with Bayesian inference. The method was applied to analyze a marine temperature time series measured from the Barents Sea and its correlation with the abundance of Atlantic salmon in three Barents Sea rivers. Besides the annual seasonal variation and a linear trend, the method revealed mid time-scale (approximate to 10 years) and long time-scale (approximate to 30 years) variation. The 10-year quasi-cyclical component of the temperature time series appears to be connected with a similar feature in Atlantic salmon abundance. These findings can provide information about the environmental factors affecting seasonal and periodic variation in survival and migrations of Atlantic salmon and other migratory fish. 2017