Attention-Based CNN-BiLSTM Model for La Niña Events

The La Niña events, as a global climate pattern, exerts a significant impact on the environment, agriculture, water resources, and social infrastructure. Given its complexity, existing prediction methods often have limitations in terms of accuracy. This study introduced an advanced deep learning model that incorporates Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory networks (BiLSTM), and Attention Mechanisms (AM) with the aim of improving the prediction accuracy of La Niña events. By analyzing sea surface temperature data in the equatorial region from 1950 to 2021, this model successfully forecasted the Oceanic Niño Index (ONI), accurately identified La Niña events, and revealed the specific impacts of La Niña on China's climate and agriculture through factor analysis. The results demonstrated that this model excels not only in terms of prediction accuracy but also in its ability to deeply analyze the influence of La Niña on regional climate, particularly during significant events in China in 2020 and 2011. This study validated the effectiveness of deep learning in predicting complex climate phenomena and provided new perspectives and tools for future research in related fields.