Prediction of fish migration based on LSTM model

Abstract

Aiming at the problem of fish migration with ocean temperature changes, this paper uses the LSTM model to predict the migration trajectory of the fish. Firstly, using the global ocean temperature data set to predict the most suitable ocean surface temperature for fish to survive, and setting a sampling point in the longitude direction of 54.05°N$\sim$60.05°N to obtain the ocean surface temperature of the area in the past 50 years, preparing for further forecast; Secondly, using the LSTM model to model the ocean surface temperature data and predict the ocean surface temperature in the next 50 years, thus deriving the suitable living area of fish for survival and regarding the area closest to this temperature as the current survival address of the fish. Finally, mapping out the migration route of the fish in the next 50 years. It is verified that this method has small errors, high reliability and accuracy, and can fit the migration route of fish schools well.