Coastal Fish Production in the Northern Baltic Sea
Environmental conditions play a crucial role in the distribution and abundance of fish species in any area, but the relative importance of various environmental factors is very difficult to determine. The coastline of Finland in the northern Baltic Sea offers a unique natural experimental setting that can be used to assess the relative importance of various environmental factors for the species occupying it. The area includes major variations in several crucial environmental factors: salinity, temperature regime, represented by winter ice duration, coastline characteristics, and eutrophic status. Furthermore, Finland has collected extensive and spatially representative data of water quality and environmental factors, as well as a long and extraordinarily spatially detailed data set of commercial catches of several fish species. This model attempts to correlate the environmental data to the commercial catches of fish species, assuming that the commercial catches reflect, to some reasonable degree, the productivity of that species in that area (compared to other areas and combinations of environmental factors, not to other species).We use a Bayesian network approach to examine the sensitivity of the species to the environmental factors. All of the probability distributions in the model are derived from data, with no priors or expert judgment. The data includes 36 statistical rectangles across the coast of Finland, and 21 years. The variables were discretized to approximate equal frequency of the marginal distribution, but striving also for ecological significance of the breakpoints where possible.
|Total P (mg / m^3)|
|Total N (mg / m^3)|
|Chlorophyll (mg / litre)|
|Salinity: per mile|
|Ice Cover: average duration of ice cover per year (days)|
|Shore Density: length of coast line (mk) / water surface (ha)|
Model Output (average catch (kg) / water surface (ha))
|Herring||Probability||Probability (bar view)|
|0 - 3.5||3.5 - 13||13 - 70|
|Vendace||Probability||Probability (bar view)|
|0 - 0.000015||0.000015 - 0.007||0.007 - 2.7|
|Flounder||Probability||Probability (bar view)|
|0 - 0.00003||0.00003 - 0.01||0.01 - 0.17|
|Pike||Probability||Probability (bar view)|
|0 - 0.015||0.015 - 0.06||0.06 - 0.5|
|Roach||Probability||Probability (bar view)|
|0 - 0.0035||0.0035 - 0.025||0.025 - 1.1|
Observe how herring catches increase and vendace catches decrease as salinity goes up. Also observe how vendace is especially sensitive to ice cover duration. Vendace is a small salmonid freshwater species that in the Baltic Sea occusts mostly in the most northern part, ehere salinity is low and yearly ice coverage is long.
Also pike and roach catches drop to the lowest class when the ice cover duration is the shortest. This is because they are coastal species, and the lowest ice cover durations occur in the open areas. However, pike and roach catches are not so sensitive to the two higher classes of ice cover duration. The same can be seen by changing the shore density.
[Uusitalo et al, 2011] Uusitalo, L., Kuikka, S., Kauppila, P., Soderkultalahti, P. and Back, S. (2011), Assessing the roles of environmental factors in coastal fish production in the northern Baltic Sea: A Bayesian network application. Integr Environ Assess Manag. doi: 10.1002/ieam.180