Ensemble Empirical Mode Decomposition with Adaptive Noise with Convolution Based Gated Recurrent Neural Network: A New Deep Learning Model for South Asian High Intensity Forecasting

Kecheng Peng,Bainian Liu,Yanan Guo,Xiaoqun Cao,Wenlong Tian

doi:10.3390/sym13060931

Abstract

The intensity variation of the South Asian high (SAH) plays an important role in the formation and extinction of many kinds of mesoscale systems, including tropical cyclones, southwest vortices in the Asian summer monsoon (ASM) region, and the precipitation in the whole Asia Europe region, and the SAH has a vortex symmetrical structure; its dynamic field also has the symmetry form. Not enough previous studies focus on the variation of SAH daily intensity. The purpose of this study is to establish a day-to-day prediction model of the SAH intensity, which can accurately predict not only the interannual variation but also the day-to-day variation of the SAH. Focusing on the summer period when the SAH is the strongest, this paper selects the geopotential height data between 1948 and 2020 from NCEP to construct the SAH intensity datasets. Compared with the classical deep learning methods of various kinds of efficient time series prediction model, we ultimately combine the Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) method, which has the ability to deal with the nonlinear and unstable single system, with the Permutation Entropy (PE) method, which can extract the SAH intensity feature of IMF decomposed by CEEMDAN, and the Convolution-based Gated Recurrent Neural Network (ConvGRU) model is used to train, test, and predict the intensity of the SAH. The prediction results show that the combination of CEEMDAN and ConvGRU can have a higher accuracy and more stable prediction ability than the traditional deep learning model. After removing the redundant features in the time series, the prediction accuracy of the SAH intensity is higher than that of the classical model, which proves that the method has good applicability for the prediction of nonlinear systems in the atmosphere.

Highlights

South Asian High (SAH) inThe data period selected in this study is 1948–2020, a total of 73 years of data.ofAccordtensity with the value of hPa maximum potential height and calculated the time series ing to the standard in Section 2.1, we constructed the standard database of SAH intensity data 6 shows the the timetime series datadata of SAH
The results show that the proposed CEEMDAN+ Convolution-based Gated Recurrent Neural Network (ConvGRU) model has higher stability and better performance than the traditional deep learning model, and is more suitable for predicting SAH intensity
Functions (IMFs) and residual components through the CEEMDAN method, the decomposed Intrinsic Mode Function (IMF) was detected by Permutation Entropy (PE) method, and the abnormal IMF

Summary

Introduction

The South Asian High (SAH) is the largest anticyclone in summer and a dominant system in Asia and the dominant circulation feature spanning Southeast Asia to Afghanistan during the summer season (e.g., [1,2,3]), and the intra-seasonal and annual intensity variation of SAH has significant impacts on global heavy precipitation [4,5,6,7] and temperature anomaly [7,8], and it is closely related to extreme weather events, especially the transport of chemical composition in the upper troposphere and lower stratosphere (UTLS) region [9,10,11,12]. Many weather and climate change systems in the field of meteorology and ocean provide a reliable theoretical basis and feasibility for the prediction of the SAH intensity. Studies to explore the day-to-day intensity change of the SAH. We need further studies predicting the intensity variation of the SAH and its impact on extreme weather

Objectives

Methods

Results

Conclusion