Forecasting the SAM Index Value Using Fourier Series and Neural Networks

Modern science refers to krill as a strategic resource of the twenty-first century, because it is valuable for its biological composition (lysine, valine, leucine and isoleucine, threonine, phenylalanine, methionine). The lipids of Antarctic krill contain many unsaturated fatty acids, phospholipids, and sterols. Krill can be successfully used for organic products intended for human consumption. To forecast the presence of krill in fishing zones, the Southern Annular Mode (SAM) index will be used. The article presents the results of forecasting the SAM index using tools such as a recurrent neural network and Fourier series-based model. The experiments revealed that these forecasting methods unfortunately had a high percentage of error, suggesting the impracticality of this approach. However, the task of forecasting the SAM index remains highly relevant and requires the exploration of new solutions. Russia plans to start krill fishing at the end of 2023, with a specially designed vessel for operations in the Antarctic region. The annual krill catch is projected to reach 450 thousand tons. To achieve these goals, scientifically grounded forecasting mechanisms for indicators directly or indirectly indicating the presence of krill biomass in planned fishing areas need to be developed.