Mesoscale eddy trajectory prediction in the Northeastern South China Sea based on an deep learning framework

Abstract

This study presents an Eddy LSIN (LSTM-Informer) framework for the accurate prediction of mesoscale eddy trajectories in the northeastern South China Sea. This comprehensive framework is designed to improve mesoscale eddy trajectory prediction, and it employs a multimodal spatiotemporal eddy representation (STER) dataset, a temporal Informer (T-Informer) network for long-sequence modeling, and a geographically aware combined mean squared error (CMSE) loss function. Mesoscale eddies are critical dynamic processes in oceans, and accurately predicting their trajectories is of great importance in oceanographic research. Leveraging the physical characteristics of mesoscale eddies, the framework is based on the construction of a high-resolution dataset that integrates sea level anomalies, relative vorticity, and kinetic energy features. An autoencoder is employed for feature extraction, resulting in the STER dataset, which reduces prediction errors by approximately 70% (mean squared error) and 82% (mean absolute error)compared to the unprocessed sla-t-u-v dataset. The framework incorporates a long short-term memory model for short-term prediction and a temporally encoded T-Informer model for long-term prediction. To balance numerical accuracy and spatial precision, CMSE is introduced by combining the mean squared error with the Haversine distance-based Geographic error. Experimental results show that the framework achieves average trajectory center errors of approximately 8 km for 1-day predictions, 18 km for 7-day predictions, and 39 km for 20-day predictions. Compared with conventional models, the T-Informer improves the long-sequence prediction accuracy by around 40% and effectively mitigates cumulative errors. Overall, the Eddy LSIN framework demonstrates high accuracy and robustness in short- and long-term mesoscale eddy trajectory prediction and offers a promising approach for advancing predictive oceanography.

Keywords

Mesoscale eddy / Deep learning / Trajectory prediction / Northeastern South China Sea

References

[1] AdamsDK, McGillicuddyDJJr, ZamudioL, ThurnherrAM, LiangX, RouxelO, et al.. Surface-generated mesoscale eddies transport deep-sea products from hydrothermal vents. Science, 2011, 332(6029): 580-583.

[2] AlfordMH, PeacockT, MacKinnonJA, NashJD, BuijsmanMC, CenturioniLR, et al.. The formation and fate of internal waves in the South China Sea. Nature, 2015, 521(7550): 65-69.

[3] ChanPW, LamCC, HuiTW, GaoZ, FuH, SunC, et al.. The effects of upper-ocean sea temperatures and salinity on the intensity change of tropical cyclones over the western North Pacific and the South China Sea: an observational study. Atmosphere, 2024, 156674.

[4] CheltonDB, SchlaxMG, SamelsonRM. Global observations of nonlinear mesoscale eddies. Prog Oceanogr, 2011, 91(2): 167-216.

[5] CheltonDB, SchlaxMG, SamelsonRM, de SzoekeRA. Global observations of large oceanic eddies. Geophys Res Lett, 2007, 3415L15606.

[6] ChenG, HouY, ChuX. Mesoscale eddies in the South China Sea: mean properties, spatiotemporal variability, and impact on thermohaline structure. J Geophys Res-Oceans, 2011, 116C6. C06018

[7] GanJ, LiuZ, HuiCR. A three-layer alternating spinning circulation in the South China Sea. J Phys Oceanogr, 2016, 46(8): 2309-2315.

[8] GeLY, HuangBX, ChenXY, ChenG. Medium-range trajectory prediction network compliant to physical constraint for oceanic eddy. IEEE Trans Geosci Remote Sens, 2023, 61: 1-14.

[9] Graves A, Mohamed AR, Hinton G (2013) Speech recognition with deep recurrent neural networks. In: 2013 IEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, pp 6645–6649. https://doi.org/10.1109/ICASSP.2013.6638947

[10] HochreiterS, SchmidhuberJ. Long short-term memory. Neural Comput, 1997, 9(8): 1735-1780.

[11] HughesCW, AshER. Eddy forcing of the mean flow in the Southern Ocean. J Geophys Res-Oceans, 2001, 106(C2): 2713-2722.

[12] Hurlburt HE, Chassignet EP, Cummings JA, Kara AB, Metzger EJ, Shriver JF et al (2008) Eddy-resolving global ocean prediction. In: Ocean Modeling in an eddying regime. Geophysical Monograph Series, pp 353–381. https://doi.org/10.1029/177GM21

[13] LeCunY, BengioY, HintonG. Deep learning. Nature, 2015, 521(7553): 436-444.

[14] LiJ, WangG, XueH, WangH. A simple predictive model for the eddy propagation trajectory in the northern South China Sea. Ocean Sci, 2019, 15(2): 401-412.

[15] LinX, DongC, ChenD, LiuY, YangJ, ZouB, et al.. Three-dimensional properties of mesoscale eddies in the South China Sea based on eddy-resolving model output. Deep Sea Res Part I Oceanogr Res Pap, 2015, 99: 46-64.

[16] LiuS, YangY, TangD, YanH, NingG. Association between the biophysical environment in coastal south China sea and large-scale synoptic circulation patterns: the role of the Northwest Pacific subtropical high and typhoons. Remote Sens, 2021, 1316. 3250

[17] LodiseJ, ÖzgökmenT, GonçalvesRC, IskandaraniM, LundB, HorstmannJ, et al.. Investigating the formation of submesoscale structures along mesoscale fronts and estimating kinematic quantities using Lagrangian drifters. Fluids, 2020, 53159.

[18] MaC, LiS, WangA, YangJ, ChenG. Altimeter observation-based eddy nowcasting using an improved conv-LSTM network. Remote Sens, 2019, 117. 783

[19] MasonE, PascualA, McWilliamsJC. A new sea surface height-based code for oceanic mesoscale eddy tracking. J Atmos Ocean Technol, 2014, 31(5): 1181-1188.

[20] McWilliamsJC. Submesoscale, coherent vortices in the ocean. Rev Geophys, 1985, 23(2): 165-182.

[21] OeyLY, EzerT, ForristallG, CooperC, DiMarcoS, FanS. An exercise in forecasting loop current and eddy frontal positions in the Gulf of Mexico. Geophys Res Lett, 2005, 3212L12611.

[22] QiuB, ChenS. Eddy-induced heat transport in the subtropical North Pacific from Argo, TMI, and altimetry measurements. J Phys Oceanogr, 2005, 35(4): 458-473.

[23] RaspS, ThuereyN. Data-driven medium-range weather prediction with a Resnet pretrained on climate simulations: a new model for WeatherBench. J Adv Model Earth Syst, 2021, 132. e2020MS002405

[24] Salehinejad H, Sankar S, Barfett J, Colak E, Valaee S (2017) Recent advances in recurrent neural networks. Preprint at arXiv:1801.01078

[25] SahooBB, JhaR, SinghA, KumarD. Long short-term memory (LSTM) recurrent neural network for low-flow hydrological time series forecasting. Acta Geophys, 2019, 67(5): 1471-1481.

[26] Shi XJ, Chen ZR, Wang H, Yeung DY, Wong WK, Woo WC (2015) Convolutional LSTM network: a machine learning approach for precipitation nowcasting. In: 29th Annual Conference on Neural Information Processing Systems. Montreal, pp 1–12. https://doi.org/10.48550/arXiv.1506.04214

[27] SiegelDA, GranataTC, MichaelsAF, DickeyTD. Mesoscale eddy diffusion, particle sinking, and the interpretation of sediment trap data. J Geophys Res-Oceans, 1990, 95(C4): 5305-5311.

[28] Staudemeyer RC, Morris ER (2019) Understanding LSTM–a tutorial into long short-term memory recurrent neural networks. Preprint at arXiv:1909.09586

[29] SunWJ, DongC, TanW, LiuY, HeY, WangJ. Vertical structure anomalies of oceanic eddies and eddy-induced transports in the South China Sea. Remote Sens, 2018, 105. 795

[30] SunZY, HuJY, ZhengQA, GanJP. Comparison of typhoon-induced near-inertial oscillations in shear flow in the northern South China Sea. Acta Oceanol Sin, 2015, 34: 38-45.

[31] TepetidisN, KoutsoyiannisD, IliopoulouT, DimitriadisP. Investigating the performance of the informer model for streamflow forecasting. Water, 2024, 16202882.

[32] ThiriaS, SorrorC, ArchambaultT, CharantonisA, BereziatD, MejiaC, et al.. Downscaling of ocean fields by fusion of heterogeneous observations using deep learning algorithms. Ocean Model, 2023, 182. 102174

[33] ThoppilPG, FrolovS, RowleyCD, ReynoldsCA, JacobsGA, MetzgerEJ, et al.. Ensemble forecasting greatly expands the prediction horizon for ocean mesoscale variability. Commun Earth Environ, 2021, 2189.

[34] TianXQ, ZhangCQ, LiuH, ZhangBF, LuC, JiaoPF, et al.. Research on air quality in response to meteorological factors based on the informer model. Sustainability, 2024, 1616. 6794

[35] Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN et al (2017) Attention is all you need. In: 31st Annual Conference on Neural Information Processing Systems (NIPS). Available at https://proceedings.neurips.cc/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf

[36] VolkovDL, LeeT, FuLL. Eddy-induced meridional heat transport in the ocean. Geophys Res Lett, 2008, 3520L20601.

[37] WangQ, ZengL, ChenJ, HeY, ZhouW, WangD. The linkage of Kuroshio intrusion and mesoscale eddy variability in the northern South China Sea: subsurface speed maximum. Geophys Res Lett, 2020, 4711. e2020GL087034

[38] WangX, WangH, LiuD, WangW. The prediction of oceanic mesoscale eddy properties and propagation trajectories based on machine learning. Water, 2020, 1292521.

[39] WangXG, LiC, WangXN, TanLN, WuJ. Spatio-temporal attention-based deep learning framework for mesoscale eddy trajectory prediction. IEEE J Sel Top Appl Earth Observ Remote Sens, 2022, 15: 3853-3867.

[40] WangXN, WangXG, YuM, LiC, SongDL, RenP, et al.. Mesogru: deep learning framework for mesoscale eddy trajectory prediction. IEEE Geosci Remote Sens Lett, 2021, 19: 1-5.

[41] XiaoCJ, ChenNC, HuCL, WangK, XuZW, CaiYP, et al.. A spatiotemporal deep learning model for sea surface temperature field prediction using time-series satellite data. Environ Model Softw, 2019, 120. 104502

[42] YangQX, NikurashinM, SasakiH, SunH, TianJW. Dissipation of mesoscale eddies and its contribution to mixing in the northern South China Sea. Sci Rep, 2019, 91. 556

[43] YangQX, TianJW, ZhaoW, LiangXF, ZhouL. Observations of turbulence on the shelf and slope of northern South China Sea. Deep-Sea Res Part I-Oceanogr Res Pap, 2014, 87: 43-52.

[44] Zargar S (2021) Introduction to sequence learning models: RNN, LSTM, GRU. Department of Mechanical and Aerospace Engineering, North Carolina State University. Available at https://doi.org/10.13140/RG.2.2.36370.99522

[45] ZhangZ. Submesoscale dynamic processes in the south China Sea. Ocean-Land-Atmos Res, 2024, 3. 0045

[46] ZhaoW, ZhouC, TianJW, YangQX, WangB, XieL, et al.. Deep water circulation in the Luzon Strait. J Geophys Res Oceans, 2014, 119(2): 790-804.

[47] ZhouH, ZhangS, PengJ, ZhangS, LiJ, XiongH, et al.. Informer: beyond efficient transformer for long sequence time-series forecasting. Proceedings of the AAAI Conference on Artificial Intelligence, 2021.

Funding

Hainan Province Science and Technology Special Fund(ZDYF2023GXJS151)

RIGHTS & PERMISSIONS

The Author(s)