Prediction of fish catch in the Danube River based on long-term variability in environmental parameters and catch statistics
Authors:Smederevac-Lalić, Marija M.
Kalauzi, Aleksandar J.
Regner, Slobodan B.
Naunović, Zorana Z.
Hegediš, Aleksandar E.
Article (Published version)
© 2017 Elsevier B.V.
MetadataShow full item record
The effects of physical factors on fish catch in the Serbian part of the Danube River were studied for period of six decades. The data on total catch for the Danube River from river kilometre 845 to river kilometre 1433 were collected from Statistical Office of the Republic of Serbia, while water level and water temperature data were collected from 16 water gauge stations along the investigated part of the Danube River for the period 1948–2009. Cross-correlation functions have been used to analyse the functional connection between Danube water level, water temperature and fish catch while ARMA model which combines cyclic (deterministic) and random (stochastic) components of the analysed sequences was used for the forecasts. The cross-correlation function showed negative correlation between water level and temperature as well as between water temperature and catch and positive correlation between water level and catch. The Danube water level and catch were coherent at the periods of 2.06, 4.13, 6.2, 10.33, 20.66 years, while the cross correlation function between these time series did not show phase lag. The results of reconstruction and forecast of water level, temperature, and catch of fish in the Danube River, obtained by summing the cyclic and stochastic components, was used for the forecast till 2029. In 2016, seven years after, the initial forecasts were made, validity of the model was checked by obtaining data for water temperature, water level and fish catch in the Danube River for the period 2010–2015. Model gave the best prediction for water temperature; average standard error was 1.6 times higher for predicted value than for model value while for fish catch and water level they were 1.96 and 4.97, respectively. Methods used in this work could be powerful tool for prediction of fish catch and serve as the basis for better fisheries management.
Keywords:Large river; Fishery; Water temperature; Water level; Solar cycle
In: Science of The Total Environment (2017), 609: 664-671