A Transformer-Based Deep Learning Model for Successful Predictions of the 2021 Second-Year La Nina Condition

doi:10.1029/2023GL104034

IOCAS-IR > 海洋环流与波动重点实验室

	A Transformer-Based Deep Learning Model for Successful Predictions of the 2021 Second-Year La Nina Condition
	Gao, Chuan1,2 ; Zhou, Lu 1,3; Zhang, Rong-Hua2,3,4
	2023-06-28
发表期刊	GEOPHYSICAL RESEARCH LETTERS
ISSN	0094-8276
卷号	50 期号:12 页码:10
通讯作者	Zhang, Rong-Hua([email protected])
摘要	A purely data-driven and transformer-based model with a novel self-attention mechanism (3D-Geoformer) is used to make predictions by adopting a rolling predictive manner similar to that in dynamical coupled models. The 3D-Geoformer yields a successful prediction of the 2021 second-year cooling conditions that followed the 2020 La Nina event, including covarying anomalies of surface wind stress and three-dimensional (3D) upper-ocean temperature, the reoccurrence of negative subsurface temperature anomalies in the eastern equatorial Pacific and a corresponding turning point of sea surface temperature (SST) evolution in mid-2021. The reasons for the successful prediction with interpretability are explored comprehensively by performing sensitivity experiments with modulating effects on SST due to wind and subsurface thermal forcings being separately considered in the input predictors for prediction. A comparison is also conducted with physics-based modeling, illustrating the suitability and effectiveness of 3D-Geoformer as a new platform for El Nino and Southern Oscillation studies. Plain Language Summary The tropical Pacific experienced the prolonged cooling conditions during 2020-2022 (often called a triple La Nina), which exerted great impacts on the weather and climate globally. However, physics-derived coupled models still have difficulty in accurately making long-lead real-time predictions for sea surface temperature (SST) evolution in the tropical Pacific. With the rapid development of deep learning-based modeling, purely data-driven models provide an innovative way for SST predictions. Here, a transformer-based deep learning model is used to evaluate its performance in predicting the evolution of SST in the tropical Pacific during 2020-2022 and explore process representations that are important for SST evolution during 2021, including subsurface thermal effect and surface wind forcing on SST, the crucial factors determining the second-year prolonged La Nina conditions and turning point of SST evolution. A comparison is made between the completely differently constructed physics-derived dynamical coupled model and the pure-data driven deep learning model, showing they both can be used for predictions of SST evolution in the 2021 second-year cooling conditions. This indicates that it is necessary to adequately represent the thermocline feedback in predictive models, either in dynamical coupled models or purely data-driven models, so that El Nino and Southern Oscillation predictions can be improved.
关键词	the 2021 second-year cooling condition a transformer-based deep learning model 3D multivariate prediction subsurface thermal effect comparison with dynamical models
DOI	10.1029/2023GL104034
收录类别	SCI
语种	英语
资助项目	National Natural Science Foundation of China[42176032]; Strategic Priority Research Program of the Chinese Academy of Sciences (CAS)[XDB42000000]; NSFC[42030410]; Strategic Priority Research Program of CAS[XDB40000000]; Startup Foundation for Introducing Talent of NUIST; Laoshan Laboratory[LSKJ202202402]
WOS研究方向	Geology
WOS类目	Geosciences, Multidisciplinary
WOS记录号	WOS:001058648500004
出版者	AMER GEOPHYSICAL UNION
WOS关键词	EL-NINO ; OSCILLATOR MODEL ; ENSO ; PACIFIC ; DISPLACEMENTS ; FORECASTS ; PROGRESS
引用统计	被引频次：21[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://ir.qdio.ac.cn/handle/337002/181698
专题	海洋环流与波动重点实验室
通讯作者	Zhang, Rong-Hua
作者单位	1.Chinese Acad Sci, Inst Oceanol, Ctr Ocean Mega Sci, Key Lab Ocean Circulat & Waves, Qingdao, Peoples R China 2.Laoshan Lab, Qingdao, Peoples R China 3.Univ Chinese Acad Sci, Beijing, Peoples R China 4.Nanjing Univ Informat Sci & Technol, Sch Marine Sci, Nanjing, Peoples R China
第一作者单位	中国科学院海洋研究所
推荐引用方式 GB/T 7714	Gao, Chuan,Zhou, Lu,Zhang, Rong-Hua. A Transformer-Based Deep Learning Model for Successful Predictions of the 2021 Second-Year La Nina Condition[J]. GEOPHYSICAL RESEARCH LETTERS,2023,50(12):10.
APA	Gao, Chuan,Zhou, Lu,&Zhang, Rong-Hua.(2023).A Transformer-Based Deep Learning Model for Successful Predictions of the 2021 Second-Year La Nina Condition.GEOPHYSICAL RESEARCH LETTERS,50(12),10.
MLA	Gao, Chuan,et al."A Transformer-Based Deep Learning Model for Successful Predictions of the 2021 Second-Year La Nina Condition".GEOPHYSICAL RESEARCH LETTERS 50.12(2023):10.