Prediction of La Niña Events Based on BP Neural Network Model

Abstract

As an extreme weather phenomenon, La Niña event will cause low water temperature in the equatorial eastern Pacific Ocean and abnormal longitudinal circulation in East Asia, resulting in widespread and continuous abnormal cooling of sea surface temperature in the eastern and central equatorial Pacific Ocean, which will cause frequent precipitation and flooding in the north and high temperature and low rainfall and frequent droughts in the south^[1]. Therefore, it is particularly important to assess the damage and propose response strategies^[2]. For the main countries involved in the global triple La Niña event, this paper selects four representative countries located in different regions to analyze, through descriptive analysis and graphical visualization, and selects the relevant index data of the last three years: monthly average maximum and minimum temperatures, monthly average precipitation rate, monthly average precipitation days, monthly average wind speed and monthly average water temperature, and finds that New York City and Brazil were affected by the triple La Niña event at the same time and were affected to a higher degree, and the temperature changes in Malasila and Jakara were stable, and the temperature effects brought by the triple La Niña event were smaller^[3]. Then, a BP neural network model was established to predict the probability of future La Niña events in each region by modeling pressure, temperature, and salinity data, and it was found that a total of 110 regions would experience La Niña events in the future, with a general probability of 55%-60%..